Available online at www.sciencedirect.
com
Sedimentary Geology 202 (2007) 1 4
Editorial
Concepts and applications of fluvial sedimentology: Towards
interdisciplinary analysis of process and product
1. Introduction
This special issue contains a series of papers presented
at the Eighth International Conference on Fluvial
Sedimentology (ICFS) held in Delft, The Netherlands,
between 7 and 12 August, 2005. Conferences in the ICFS
series have been held every four years since 1977, and
seek to stimulate the exchange of ideas among the
community of scientists with common interests in rivers
and their deposits, both modern and ancient. The 8th ICFS
continued the tradition of providing a state-of-the-art
overview of the full range of sedimentological research on
fluvial systems. As a result of the excellent attendance
(300 participants from 25 countries), the scientific
sessions comprised a wide range of presentations on the
broad themes of sediment dynamics and deposits. This
special issue contains seventeen papers, all of which were
originally presented at the conference.
The Eighth ICFS focused on process-based modelling
of fluvial systems, on the application of quantitative
sedimentological data in modelling of the subsurface, and
on the interaction between fluvial and marine processes
through studies of fluvio-deltaic system dynamics. The
scientific committee considered the latter topic particu-
larly relevant in view of the conference venue. Holland,
the western part of the Netherlands, is a densely populated
area largely located below sea level, which would not
have existed without extensive manipulation of the nat-
ural environment. Large-scale engineering efforts in the
Dutch coastal lowlands, aimed at reducing the constant
risk of fluvial and marine flooding, started around one
thousand years ago and continue to this day. The societal
relevance of such engineering efforts will continue to
grow in the upcoming decades, as indicated by recent
predictions of climate change, which imply locally sig-
nificant increases of fluvial peak discharges as well as
global sea-level rise. Since an estimated 70% of the
0037-0738/$ - see front matter 2007 Elsevier B.V. All rights reserved.
doi:10.1016/j.sedgeo.2007.07.008
www.elsevier.com/locate/sedgeo
world's population lives at or near the coast, geoscientific
research aimed at increasing our understanding of fluvio-
deltaic system dynamics will become a top priority in the
near future.
2. Ancient systems
Prediction of fluvio-deltaic stratal architecture as a
function of various controlling factors has traditionally
been based on qualitative studies of sedimentary facies
and stratal geometries in outcrops. Subsurface data from
wells are typically correlated by means of seismic. Both
types of geological interpretations draw heavily on an-
alogue studies of modern processes and systems. Detailed
interpretation of a much wider range of fluvial environ-
ments is now possible as further studies of rivers in more
remote areas have become available (e.g. dryland rivers).
The traditional deterministic view, according to which the
rich diversity of fluvial deposits could be attributed to
hydraulic conditions that were directly related to specific
depositional environments, has been gradually abandoned
in favour of a dynamic view on fluvial systems and
deposits, with an increased focus on complex responses
and indirect controls. In his keynote lecture, Janok
Bhattacharya reviewed the controls on Distributary
channels, fluvial channels and incised valleys and illus-
trated that branching patterns in fluvio-deltaic systems
reflect the interaction of multiple marine and fluvial
processes, such as water depth, wave climate, tidal
regime, and discharge (variability). He presented several
examples of revised interpretations of ancient deltaic
deposits, based on insights gained from the study of
modern systems.
Nearly one half of the papers contained in this special
issue deals in one way or another with analysis of
ancient deposits. The contributions of Allen and
Fielding, Fielding et al., Hillier et al., Juhasz et al.,
2 Editorial
and Paredes et al. present classic syntheses of alluvial-
system stratigraphy and sedimentology, based on out-
crop and subsurface data, from which inferences are
drawn about the roles of base-level variations, climatic
regimes, sediment supply, and tectonic activity. The
study of Jones and Hajek focuses on the stratigraphic
expression of avulsion styles, whereas the studies of
Gonzalez-Acebron et al. and Ochoa et al. illustrate the
relations between provenance, dispersal, diagenesis and
hydrothermal metamorphism of fluvial sands in an
intracratonic rift basin.
3. Quaternary systems
Late Quaternary systems especially those still
active today provide sedimentologists with a close
link between process and product. Studies of (sub)recent
(Quaternary) systems are one of the cornerstones of
sedimentology, because they provide high-quality data
and permit detailed interpretations of the responses to
changes in boundary conditions, such as palaeoclimate
(discharge and sediment supply) and neotectonics (local
gradients). The boundary conditions can often be
reconstructed from independent evidence (palaeocli-
mate proxies) and/or Global Circulation Models, which
provide researchers with powerful tools to investigate
fluvio-deltaic system dynamics, often giving access to
depositional records over a much longer timescale than
is possible from records of human observations.
Analysis of (sub)recent systems may provide conclusive
evidence for the spatial variability of responses to e.g.,
peak discharges, sea-level change, vegetation, perma-
frost, and volcanic eruptions, to name just a few
possibilities. Another distinctive quality of (sub)recent
systems is the absence of diagenesis, which is of
particular interest in applications of modern analogues
in groundwater and hydrocarbon reservoir modelling.
A better understanding of the complex relationships
between controls and responses will ultimately lead to
improved interpretations of the stratigraphic record. The
keynote lecture by Torbjrn Trnquist, entitled How
trivial is the identification of fluvial incision and ag-
gradation, and how complex are their controls? illus-
trated the plethora of possible explanations of aggradation
and incision in the deglacial Rhine system, and showed
that comparisons with analogue flume-tank experiments
may go some way towards reducing the range of feasible
scenarios.
Four papers in this special issue deal with (sub)recent
fluvio-deltaic systems. Nador et al. and Gao et al. report
on the evolution of Weichselian fluvial systems in response
to climate and tectonics. Aguzzi et al. present a detailed
palaeo-environmental analysis of a deep borehole through
a Late Quaternary incised-valley fill. Berzesio et al.
illustrate the development of high-resolution hydrogeolo-
gical reservoir models of the shallow subsurface by inte-
gration of geological and geophysical techniques.
4. Numerical models
The main reason for pursuing numerical modelling as
an avenue of research in fluvial sedimentology is that
prediction of fluvio-deltaic stratal architecture as a
function of the boundary and initial conditions in a
particular geological setting has great economic value and
provides an objective method for testing outcrop-based
conjectures on the palaeogeographic evolution of the
Earth. Ongoing research in this field is aimed at handling
the complexity of responses to rising sea (base) level,
climate change, and the internal (autogenic) dynamics of
fluvio-deltaic systems.
Rudy Slingerland critically reflected on the current
state of process-based stratigraphic modelling in his
keynote address Process-based stratigraphical model-
ling of fluvial systems: Why is progress so slow? and
identified three areas in which more research is needed.
Firstly, there is a need for more accurate numerical
models of fluvial morphodynamic processes. For
example, we still do not know the necessary and suffi-
cient conditions that create avulsions and bifurcations,
nor can we predict sedimentary facies on floodplains. It
also is vexing that the fractal behaviour of fluvial sys-
tems clouds the question of what processes can and
should be considered subgrid-scale. Secondly, we need
a better understanding of inherent non-linear and chaotic
effects. There is a strong localization of key quantities
such as material flux and sedimentation rates in fluvio-
deltaic systems, which is proving difficult to model.
Perhaps as a consequence, our models are not yet cap-
able of self-organization and spontaneous formation of
complex patterns. Thirdly, we are not testing our models
well enough. We need community datasets of high-
resolution 3D seismic data combined with extensive
core and well logs and objective, community-run inter-
model comparisons.
Huib de Vriend reviewed modelling of sediment
transport and river morphodynamics under the heading
Present-day river modelling. Developments in this
field are closely coupled with new observation techni-
ques, which are taking us from a data-starved to a data-
rich environment. Techniques such as multibeam sound-
ing and Acoustic Doppler Current Profiling enable a
detailed study of flow fields and bedform patterns in large
rivers. Repeated measurements at the same locality permit
 Editorial
expansion of such data sets to four dimensions.
Knowledge of bedform dynamics is essential for hy-
draulic interpretations of sedimentary structures, which
will increase our understanding of the geological record,
provided that their preservation potential can be assessed.
A remarkable point to emerge from his talk is the fact that
morphodynamic models used for engineering purposes
usually describe the time-evolution of the river bed,
because engineers rarely need to consider the fluvial
stratigraphic record. Geoscientists, however, have to live
with what has been left behind by the river, i.e. the internal
structure of the deposits, which may be difficult to
interpret in the absence of detailed knowledge of the
processes. A logical consequence of the above is that a
different presentation of model output may already
contribute towards bridging the gap between these two
different disciplines, and allow data- and model-driven
approaches to complement each other.
In his keynote address Subsurface fluvial modelling:
fashions, fads and classic black numbers, Colin North
covered the application of quantitative sedimentological
data to modelling of subsurface reservoir architecture.
Since subsurface models used in the oil industry are
typically constructed from sparse data, they must be sup-
plemented by modern or ancient analogues and conceptual
Fig. 1. Approaches and objects of study covered by the International Conference on Fluvial Sedimentology.
3
models derived from various sources. This has spurred a
renewed interest in outcrop studies aimed at collecting
quantitative analogue data. In contrast to the process-based
models discussed above, subsurface models do not include
the time-evolution of a system, but aim at mimicking the
stratigraphy observed at control points (wells) by means of
stochastically optimized interpolation. The evolution of
this stochastic approach to geometric modelling in the
Norwegian oil industry was described by Kevin Keogh in
his keynote address, which is included in this special issue
(Keogh et al.).
The other four modelling papers span the full range
from highly abstracted models of erosion and deposition
based on budget-capacity concepts (Tipper) to process-
based modelling of sediment transport (Reesink and
Bridge) and morphodynamics (Van Maren). The paper
by Vanlierde et al. illustrates the use of a semi-empirical
model to predict the contribution of authigenic pre-
cipitates to the suspended load of a fluvial system.
5. Outlook
Fig. 1 seeks to depict the different approaches, objects
of study, and spatio-temporal dimensions currently falling
under the ICFS umbrella. It is obvious that the term fluvial
4 Editorial
sedimentology must be loosely interpreted to cover the
full range of activities presented at these conferences,
which are continuously expanding due to new attempts to
solve long-standing and fundamental problems. In theory,
the dynamics of fluvial systems can be unraveled if we are
capable of acquiring large amounts of high-quality data
covering the full range of spatial and temporal scales over
which such systems evolve. However, data sets of this
type are simply not available, and we need to resort to
other methods to expand our horizon, such as extraction
of proxy data from the geological record. To properly
interpret the stratigraphic record of ancient systems in the
light of such proxies, one would have to understand the
effect of the geological filter through which the
stratigraphic record passed, which brings us to the prob-
lem of preservation.
A possible answer to this lies in process-based
numerical modelling of alluvial systems and stratigraphy,
which is indeed rapidly developing, but has not yet
reached the point at which these questions can be
adequately addressed. One of the fundamental problems
in this field is the difficulty of conditioning models to data.
We cannot yet answer the basic question Is a stratigraphic
model good enough? because we have no concept of the
natural variability (and predictability) of fluvial systems
over long timescales. This reflects our inability to conduct
realistic experiments, e.g., replicate runs of continental-
margin evolution starting from the same set of initial
conditions. We can hardly be blamed for this, as we do not
have the means to organize such experiments, nor the time
to await their outcome. What we should be doing is find
ways around this, for instance by exploiting the simi-
larities between real-world systems and their down-scaled
representation in flume tanks. Instead of focusing our
attention on the scaling issues involved in the translation
of flume experiments to the real world, we might use the
flume as a tool to investigate the natural variability and
predictability of system evolution. This would require less
restrictive assumptions based on second-order similarity,
i.e., not the systems themselves are to be compared, but
the experimental variability of the small-scale system is
regarded as a proxy of natural variability. We believe that
the rate of progress in fluvial sedimentology depends to a
large extent on our willingness and ability to integrate
different types of data- and model-driven approaches. We
are looking forward to the next ICFS, which will be held
in 2009 in Tucuman, Argentina.
Acknowledgements
We thank all the individuals and organizations who
contributed to the 8th ICFS, which was hosted by Delft
University of Technology, and co-organized by staff
members from Vrije Universiteit Amsterdam, Utrecht
University, Alterra (Wageningen Agricultural University),
TNO (National Geological Survey), RWS-RIZA (Direc-
torate-General for Public Works and Water Management),
NCR (Netherlands Centre for River studies), and Shell E
and P. The Conference would not have been possible
without financial support from Total E and P Nederland,
Shell, RWS, Statoil, IAMG, IAS, SEPM, Wintershall,
Panterra, Elsevier, KNGMG, NWO, KNAW, and TU
Delft. We wish to thank the reviewers for their valuable
time and efforts, and gratefully acknowledge the support
from the SG editorial office (Sheilagh Douma). We thank
SG Editor Chris Fielding for his final reviews of the
contributions to this special issue.
Gert Jan Weltje*
Robert M. Hoogendoorn
Delft University of Technology,
Faculty of Civil Engineering and Geosciences,
Department of Geotechnology, P.O. Box 5048,
NL-2600GA Delft, The Netherlands
Corresponding author.
E-mail address: g.j.weltje@tudelft.nl (G.J. Weltje).
Susan B. Marriott
University of the West of England,
Faculty of Environment and Technology,
Coldharbour Lane, Bristol BS16 1QY,
United Kingdom
Jaco H. Baas
University of Wales Bangor, School of Ocean Sciences,
Centre for Catchment to Coast Research,
Menai Bridge, Anglesey LS59 5AB, United Kingdom
 Available online at www.sciencedirect.com

Sedimentary Geology 202 (2007) 5 34

www.elsevier.com/locate/sedgeo

Sedimentology and stratigraphic architecture of the Late Permian

Betts Creek Beds, Queensland, Australia
Jonathan P. Allen , Christopher R. Fielding
Department of Geosciences, 214 Bessey Hall, University of Nebraska-Lincoln, Lincoln, NE 68588-0340, United States

Abstract

The Late Permian Betts Creek Beds form a succession of coal-bearing alluvial-coastal plain sediments in a basin marginal
setting within the northeastern Galilee Basin, Queensland, Australia. The unit is 5060 m in thickness at Porcupine Creek
National Park where outcrop is laterally continuous for several kilometers. Eight facies have been identified within the
formation and can be grouped into 2 facies associations: (A) channel deposits and (B) floodbasin deposits. The channel
association consists of conglomerate (A1) and trough cross-bedded multistorey sandstone facies (A2), both interpreted as
deposits of low-sinuosity river systems, tidally influenced fluvial channels (A3), interbedded sandstone and siltstone (A4)
interpreted as the abandonment fill of the alluvial systems, and diamictite (A5) interpreted as debris flows. The floodbasin facies
association is composed of sandy siltstone (B1; proximaldistal floodbasin), carbonaceous siltstone (B2; mire), and bioturbated
siltstone (B3; estuarine) facies.
The overall sediment body architecture can be resolved into 6 unconformity-bounded cycles interpreted as sequences in the
genetic sense, which are sheet-like in geometry and, in general, consist of amalgamated multistorey, multilateral braided fluvial
deposits at the base overlain by extensive sheet-like overbank mudstones, carbonaceous shales, and coals. Sea-level change is
interpreted as the primary control on the sequence architecture of the formation, while the internal stratigraphic architecture,
however, varies between sequences and is a function of a combination of sea level, tectonic, and autogenic controls.
2007 Elsevier B.V. All rights reserved.

Keywords: Alluvial architecture; Late Permian; Eastern Australia; Sea-level change; Cyclicity

1. Introduction act as major controls on the architecture of alluvial

While the past ten years have seen remarkable
progress and advances in our understanding of the
major controls on the large-scale sediment body archi-
tecture of alluvial successions, these controls are still
not clearly understood (Blum and Trnqvist, 2000).
Outcrop, laboratory, and computer modeling studies
have suggested that sea level, tectonics, and climate
Corresponding author. Fax: +1 402 472 4917.
E-mail address: jallen19@bigred.unl.edu (J.P. Allen).
0037-0738/$ - see front matter 2007 Elsevier B.V. All rights reserved.
doi:10.1016/j.sedgeo.2006.12.010
successions (e.g. Shanley and McCabe, 1993; Mackey
and Bridge, 1995; Olsen et al., 1995; Heller and Paola,
1996, and see Blum and Trnqvist, 2000; Bridge, 2003
for reviews). However, analyses of the stratigraphic
architecture of coal-bearing alluvial plain sequences
are more limited. While some studies have high-
lighted the effects of sea level (Bohacs and Suter,
1997), tectonics (Fielding, 1987), and climate (Field-
ing and Webb, 1996; papers in Pashin and Gastaldo,
2004) on coal-bearing strata, few have successfully
deconvolved signals of multiple controls within in-
dividual successions.
6 J.P. Allen, C.R. Fielding / Sedimentary Geology 202 (2007) 534

The Betts Creek Beds in north-central Queensland, and autogenic controls on the sediment body architecture
Australia provide an excellent opportunity to examine the of coal-bearing alluvial plain systems. In this study, the
interplay of both allogenic (sea level, tectonics, climate) effects of sea level and tectonics are dominant controls on

Fig. 1. Location maps: (a) location of the Galilee Basin in relation to other major Permian basins in eastern Australia; (b) northern Galilee Basin detailing
major structural features (modified from Hawkins and Green, 1993); (c) geology of the Hughenden area (modified from Vine and Paine, 1974).
 J.P. Allen, C.R. Fielding / Sedimentary Geology 202 (2007) 534 7

Fig. 2. Revised lithostratigraphy for the northeastern and eastern margins of the Galilee Basin with the adjacent Bowen Basin. Tectonic phases are
shown for the Galilee (from Van Heeswijck, 2004) and Bowen (from Fielding et al., 2001) Basins. Bowen Basin stratigraphy compiled from Lang
et al. (2001); Fielding et al. (2001), and Anthony (2004).

the sediment body architecture and the relationship
between these and autogenic controls can be differentiated.
2. Regional geology
2.1. Basin setting
The Galilee Basin is a large intracratonic basin of
Late Carboniferous to Triassic age located in central
Queensland (Fig. 1a). Basement comprises the Late
DevonianEarly Carboniferous Drummond Basin to the
east (Evans, 1980), the early Paleozoic metasedimentary
and intrusive rocks of the Thomson Orogen in the center
(Evans, 1980), and Precambrian to early Paleozoic
metamorphic rocks west of the Cork Fault (Vine, 1976;
Hawkins and Green, 1993). Much of the Galilee Basin is
deeply buried beneath JurassicCretaceous and Cen-
ozoic sediments of the Eromanga Basin, with the only
significant surface exposure along the eastern and
northeastern basin margin (Fig. 1a, b). Therefore,
determining the lithostratigraphy of the basin is heavily
dependent on limited subsurface data. The basin is
divided into southern and northern regions at approxi-
mately 24 S (Hawkins and Green, 1993). This paper
focuses on exposures near the northeastern margin of the
Galilee Basin.
The northern Galilee Basin is bounded to the east by
the Drummond Basin, to the north by the early Paleozoic
LolworthRavenswood Block, and to the west by the
Precambrian Mount Isa Block and Georgina Basin
(Evans, 1980; Scott and Hawkins, 1992; Hawkins and
Green, 1993; Fig. 1b). Two depressions, the Koburra
8 J.P. Allen, C.R. Fielding / Sedimentary Geology 202 (2007) 534
Trough and the Lovelle Depression, are separated by the
early Paleozoic Maneroo Platform (Fig. 1b).
In general, the Paleozoic basin fill can be divided into
two sections (Evans, 1980; Fig. 2). The Late Carboni-
ferous to Early Permian basin fill is characterized by
fluvial and lacustrine sediments of the Joe Joe Group,
some of which are glaciogenic (Jochmus Formation and
Boonderoo Beds). The late Early Permian was a time of
widespread peat development represented by the
Aramac Coal Measures. The basin experienced a period
of non-deposition and/or gentle uplift (Evans, 1980)
before the accumulation of widespread fluvial and coal
deposits of the Betts Creek Beds and correlatives (Vine,
1976; Evans, 1980). During the Late Permian, sedi-
ments from the Bowen Basin spilled across the Spring-
sure Shelf into the Galilee Basin (Fielding et al., 2001),
and may have buried the bounding ridges, thereby
connecting not only the Bowen and Galilee Basins, but
also the Cooper Basin (Hawkins, 1978). Triassic strata
Fig. 3. Location map showing study area within Porcupine Creek Gorge and locations of architectural panels and measured sections.
in the basin are dominantly fluvial (Rewan Group,
Warang and Clematis Sandstones, and Moolayember
Formation; Hawkins and Green, 1993).
The Galilee Basin fill was initially regarded as entirely
non-marine (Vine and Paine, 1974; Vine, 1976), but
Evans (1980) later suggested that marine conditions
entered the southern Koburra Trough during the Late
Permian. Furthermore, Scott and Hawkins (1992) and
Hawkins and Green (1993) recognized bioturbated paralic
facies along the eastern margin which they considered
equivalent to the Peawaddy Formation of the Denison
Trough. The present study recognizes several horizons of
bioturbation and indications that paralic sedimentation
extends farther northward along the basin margin.
2.2. Tectonic and structural setting
There have been several interpretations of the
structural evolution of the Galilee Basin including:
 J.P. Allen, C.R. Fielding / Sedimentary Geology 202 (2007) 534 9

Table 1
Summary of the general characteristics and interpretation of the lithofacies for the Betts Creek Beds
Association/ Lithology Sedimentary structures Biota Geometry Interpretation
Facies
AChannel deposits
A1 Poorly sorted, clast- Trough cross-bedding, None recorded Extensive sheet Low-sinuosity
supported, polymictic, scour surfaces, geometry, paleoflow alluvial channels
cobble conglomerate clast imbrication to 142
A2 Medium- to very Trough cross-bedding, Glossopteris, Extensive sheet Low-sinuosity
coarse-grained scour surfaces, flat lamination, Vertebraria, geometry, incised alluvial channels
sandstones ripple cross-lamination, Calamites lateral margins,
parting lineation, convolute paleoflow to 192
bedding, load casting
A3 Medium- to very Trough cross-bedding, Thalassinoides and Sheet geometry, Tidally
coarse-grained bidirectional/bimodal Planolites burrows paleoflow to 188 influenced
sandstones cross-bedding, flat lamination, alluvial channels
ripple cross-lamination,
wave ripples
A4 Interbedded to Ripple cross-stratification, None recorded Confined to Abandonment
interlaminated sandstone interlamination structures channel forms fill of alluvial
and siltstone channels
A5 Clast-rich sandy Massive unstratified, Rare macerated Confined to Debris flows
diamictite, basement, rare clast imbrication plant material channel forms
volcanic, and and crude bedding
intraformational clasts

B Floodbasin deposits
B1 Interbedded to Small-scale cross-stratification, Planolites burrows, Overall extensive Overbank
interlaminated ripple cross-lamination, Glossopteris, Vertebraria, sheet geometry, small
sandstone and siltstone interlamination structures, Calamites channel forms, some
convolute bedding, load casting depositional dips
B2 Carbonaceous Compositional banding Glossopteris, Araucarioxylon, Overall extensive Mire
shale and coal Vertebraria, Calamites, sheet geometry
unidentified seed-fern,
in situ stumps
B3 Bioturbated (BI = 05) Ripple cross-stratification, Diplocraterion habichi, Extensive sheet Estuary/coastal
interbedded sandstone interlamination structures, Conichnus, Thalassinoides, geometry, small embayment
and siltstone synaeresis cracks, Planolites, Palaeophycus channel forms
rare clastic dikes burrows

(1) a pull-apart basin formed by shearing along the
northeastern and eastern margin of Australia (Evans,
1980), (2) a pericratonic basin (Vine, 1976; Veevers
et al., 1982), (3) a basin formed by convective
downwelling and regional downwarp (Middleton and
Hunt, 1989), and (4) a foreland basin (De Caritat and
Braun, 1992). Subsequently, Van Heeswijck (2004) re-
evaluated the structural development of the Drummond
and Galilee Basins and concluded that the develop-
ment of the Galilee Basin was similar to the
development of the adjacent and coeval Bowen
Basin (Van Heeswijck, 2004). Fielding et al. (2001)
developed a paleogeographic synthesis for the Bowen
GunnedahSydney Basin System (BGSBS) in which
major events and sequences were found to be
synchronous from north Queensland to southernmost
New South Wales. The BGSBS is considered to have
developed sequentially by crustal extension (Late
CarboniferousEarly Permian), passive thermal sub-
sidence (mid-Permian), and then foreland crustal
loading (Late PermianMiddle Triassic) in response
to pulsed thrust sheet propagation during the Hunter
Bowen Event (Fielding et al., 2001). Van Heeswijck
(2004) recognized a similar three-phase history within
the Galilee Basin, concluding that the Galilee and
Bowen basins were closely related in their develop-
ment (Fig. 2).
2.3. Regional geology of the Betts Creek Beds
The Betts Creek Beds (BCB) and their correlatives
represent the final development of coal-forming
10 J.P. Allen, C.R. Fielding / Sedimentary Geology 202 (2007) 534
conditions within the Galilee Basin. Age control for the
Galilee Basin relies on palynological and plant macro-
fossils (White, in press; Evans, 1967). Evans (1967)
proposed palynological stages for eastern Australia,
placing the BCB in his Stage 5 (late Early PermianLate
Permian). Nonetheless, several authors contend that
Stage 5 represents only the Late Permian (Hawkins,
1978; Hawkins and Green, 1993), and this is the view
taken in this paper.
The BCB vary greatly in thickness, reaching a
maximum thickness of 390 m (Evans, 1980).
Hawkins (1976) described the entire succession as
an eastward-thickening lens, thinning rapidly west-
ward and to the north. The BCB crop out in the
Hughenden area and in Porcupine Creek (Galah)
Gorge (Fig. 1c, 3), where the maximum thickness is
60 m. The only published study of the BCB outcrop
was part of a mapping survey in which the BCB were
interpreted as fluvial and paludal sediments, with
alluvial fan, piedmont, and mudflow deposits
(Vine and Paine, 1974). The present study focuses
on the BCB sensu stricto in Porcupine Creek Gorge
(Fig. 3).
3. Study area and methods
The BCB are well exposed for several kilometers
in > 75 m high cliffs along Porcupine Creek Gorge
65 km northeast of Hughenden in northern
Queensland, Australia (Fig. 3). The nature of the
exposures is ideal for a three-dimensional architec-
tural study. The area studied as well as the location of
cliff exposures (referred to as panels) used for
architectural analysis and measured sections are
shown in Fig. 3. Photomosaics were taken for all
exposures in order to map out bounding surfaces and
large-scale architectural elements. Twenty, detailed
measured sections were logged at the centimeter scale
and the resulting data were coupled to the corre-
sponding photomosaics. Major bounding surfaces
identified first in photomosaics and measured sec-
tions were walked out in the field in order to
correlate surfaces between outcrop panels and
accurately describe the geometry of the formation.
A laser rangefinder was used to delineate major
bounding surfaces and to measure thicknesses of
major units that were inaccessible. Regional dip,
however, allowed the entire formation to be investi-
gated and detailed at some point along the outcrop
belt. Paleocurrent measurements and hand specimens
from a lithologic suite representative of the BCB
were also collected.
4. Lithofacies and architectural elements
Eight lithofacies are recognized within the BCB
(Table 1). These facies have been assigned to two
associations that reflect the architecture and major
depositional environments of the BCB and are (A)
channel deposits and (B) floodbasin deposits.
Eight architectural elements and two higher-order
lithosomes have been identified within the succession of
the BCB (Table 2). The two higher-order lithosomes,
channel and overbank, were established in order to
avoid confusion within the architectural classification.
Architectural features are observed at different scales
and orders, and should therefore have a hierarchy for
their classification, just as bounding surfaces have a
hierarchy that establishes the scale of feature that they
enclose (channel-belt, channel, macroform, etc.). There-
fore, lithosomes are defined as architectural features that
are made up of architectural elements, which in turn are
composed of smaller, internal features.
Architectural elements are distinguished by a variety
of characteristics including bounding surfaces, lithofa-
cies composition, geometry, vertical and lateral associa-
tions, and paleocurrent distributions. Bounding surfaces
were assigned a numerical rank after Miall (1996).
Bounding surfaces of first-and second-order have not
been detailed on any of the architectural panels, because
they cannot be distinguished. The criteria for recognition
of architectural elements are summarized in Fig. 4 and
Table 2, and were adapted and modified after Miall
(1985, 1988, 1996). For ease of comparison between this
study and others, elements have been given codes in the
fashion of Miall (1996). Several elements, however,
warranted a different code from those set out by Miall
(1996) and will be discussed below, along with modified
elements. Those elements that have already been defined
(DA, LA, SF, OF, LE, CS) will not be detailed further.
4.1. Channel (CH) lithosome
Channel lithosomes as discussed above contain
within-channel elements. CH lithosomes are not clearly
defined within the study area and only a few channel
margins have been observed. This lack of definition may
be due to the abundance of outcrop exposure roughly
parallel to paleoflow. It is not expected that channel
margins would be observed with such orientations;
instead a sheet-like channel geometry is observed.
Where outcrop orientations are normal to paleoflow,
several channel margins can be delineated, but only two
architectural panels (2, 5) are observed with this
orientation. Where channel margins are exposed, they
 J.P. Allen, C.R. Fielding / Sedimentary Geology 202 (2007) 534 11

Table 2
Classification and description of architectural elements recognized within the Betts Creek Beds
Architectural Element Facies Geometry Bounding surfaces
Composition
Channel lithosome
Channel (CH) A1, A2, B1, B3 Variable scale from 20 to >400 m wide Basal 5th order bounding surfaces.
and 17 m deep. Erosively based with concave-up margins.
Most margins are not clearly defined
within the study area.

Within-channel elements
Nested channel A2, A3 Sheet-like complexes extending >1 km Bound by 4th order surfaces that may
fill element (NC) laterally, and 111 m thick (commonly <6 m). coalesce into 5th order surfaces. Erosively
Individual NC tabular or lenticular elements irregular/planar to slightly concave bases,
range from 10 to >500 m long and 15 m thick. in some places channel-form. Upper surfaces
usually erosively irregular or planar, flat or
slightly convex where not eroded. Internal
3rd order surfaces.
Downstream-accretion A2 Wedge-shaped body 2 m thick and >40 m long. Bound by 4th order surfaces with dipping
macroform (DA) Overlies OF element, overlain by NC elements. internal 3rd order surfaces. Planar to
slightly concave base, erosively irregular
to planar tops.
Lateral-accretion A2 Tapered wedge-shaped bodies up to 80 m wide Bound by 4th order surfaces with dipping
macroform (LA) and 16 m thick. Overlain by NC elements. internal 3rd order surfaces. Planar to slightly
concave base, erosively irregular to planar tops.
Channel fill (CF) A4, A5 Variable scale from 2050 m wide and Bound by 5th or 4th order basal surfaces
27 m deep. Overlie NC and OF elements, and 5th order upper surfaces. Sharp, erosive,
overlain by CH or NC elements. concave (channel-form) bases, erosionally
irregular, concave upper surfaces.
Scour fill (SF) A1, A2 Lenticular bodies extending up to 50 m long Erosively irregular to concave basal surface,
and 2 m deep. Concentrated at the erosively planar to gradational top surface.
base of a unit.

Overbank lithosome
Overbank fines (OF) B1, B2, B3 Extensive sheet-like bodies, extending >2000 m Surfaces usually planar or gently
laterally, 0.12.5 m thick. Overlain by CH and concavoconvex, sharp to transitional bases.
NC elements, associated with LE and Tops sharp to erosionally irregular bounded
CS elements by 5th or 4th order surfaces.

Overbank elements
Levee (LE) B1 Sheet-like to lenticular bodies 40100 m long
with internal inclined accretionary layers extending
<50 m down-dip, 15 m thick, passing laterally
into OF elements. Commonly associated with CS
and small-scale CH elements. Overlain
by NC elements.
Crevasse splay (CS) A2 Lenticular bodies, extending <40 m laterally and
0.12 m thick. Associated with LE and OF
elements. Erosively overlain by NC elements.
Architectural elements adapted and modified from Miall (1985, 1996).
Bound by 4th order surfaces. Sharp,
planar non-erosive bases. Interbedded
sandstone and siltstone pass down
depositional dip into interlaminated
sandstone and siltstone. Top usually
truncated by concave or flat surfaces.
Bound by 4th order surfaces. Erosively sharp,
planar base, with planar to undulatory tops.

probably represent the temporary margins of migrating
channels within a wide channel belt.
4.2. Nested channel fill (NC)
This element broadly equates to the SB element of
Miall (1996), and is composed of sediments of Facies A2.
4.3. Channel fill (CF)
Two variations of this element are recognized within
the BCB. The most common CF elements are infilled
with fine-grained sediments of Facies A4 and are
interpreted as representing channels undergoing aban-
donment. This variant of CF element is similar to the FF
12 J.P. Allen, C.R. Fielding / Sedimentary Geology 202 (2007) 534

Fig. 4. Architectural elements and lithosomes recognized within the Betts Creek Beds. Cross-sections are the same horizontal and vertical scale.
Bounding surfaces are numbered according to rank. Element classification after Miall (1988, 1996).

(C) element of Miall (1996). The other variant of CF
element is composed of volcaniclastic sediments of
Facies A5, interpreted as channels that were occupied by
debris flows during periods of explosive volcanism (see
below for interpretation). This variant of CF element is
also characterized by broader, shallower channels.
5. Facies and architectural analysis
5.1. Facies association A channel deposits
5.1.1. Facies A1 conglomerate
Facies A1 consists of extraformational, clast-sup-
ported, polymictic, cobble conglomerate. Clasts reach
80 cm in a-axis length, but average 8 cm. They are
poorly sorted, subrounded to rounded, platy to spher-
oidal, and composed predominantly of quartz, gneiss,
granite, and tuff. The matrix is coarse, sub-angular,
quartz sand. Facies A1 unconformably overlies the Late
CarboniferousEarly Permian Boonderoo Beds and is
restricted to the basal part of the formation. Upper
contacts are sharp and the facies is overlain by either
Facies A2 or B1. The conglomerate fines upwards from
a cobble conglomerate to interbedded very coarse
sandstone and pebble conglomerate.
Facies A1 is characterized by massive to crude bedding
arranged in sheet-like to lenticular bodies with upper and
lower surfaces defined by SF elements with localized
trough cross-bedding. Several locations where Facies A1
occurs appear to be composed of a mosaic of coalesced SF
elements. Rare CH elements are also observed. Trough
cross-bedding is as much as 0.8 m thick, but commonly
<0.4 m. Paleocurrent directions from trough cross-beds
trend southeastward (mean = 142; Fig. 5a). Imbricated
large clasts in the basal portion of the conglomerate
indicate paleoflow to the south (Fig. 5b), slightly oblique
to the paleoflow indicated by trough cross-bedding.
5.1.1.1. Interpretation. The sheet-like geometry,
abundance of SF elements with subordinate trough
 J.P. Allen, C.R. Fielding / Sedimentary Geology 202 (2007) 534
Fig. 5. Rose diagrams summarizing paleocurrent data for Facies A1
collected from (a) trough cross-beds, (b) clast imbrication.
cross-bedding, and relatively low dispersion of paleo-
current data are highly compatible with deposition
within low-sinuosity alluvial channels. Facies A1 has
many characteristics that have been widely recognized
in modern and ancient gravelly braided streams (Smith,
1974; Boothroyd and Ashley, 1975; Miall, 1977; Rust,
1984; Ramos et al., 1986; Smith, 1990).
The bulk of Facies A1 represents deposits from
longitudinal gravel bars and sheets in a relatively shallow
setting marked by high flow velocities (cf. Williams and
Rust, 1969; Smith, 1974; Rust, 1978; Ramos et al., 1986;
among others). Trough cross-bedding records the
avalanching of clasts on the downward or lateral flanks
of longitudinal bars under diverging flow (Rust, 1984;
Collinson, 1996). Associated coarse- to medium-grained
sandstones were deposited from the lateral margins of
bars during falling-stage flow (cf. Massari et al., 1993).
Where trough cross-bedding is not associated with bar
deposits, it is interpreted as three-dimensional dunes
migrating under low flow regimes.
13
Abundant SF elements are attributed to the complex
pattern of rapidly changing channels and bars, which are
characteristic of gravelly braided streams. Many of the
bar deposits have scoured upper and lower surfaces
suggesting that the deposits originate from several bars
rather than a single depositional event (cf. Ramos et al.,
1986). They are therefore considered compound bars
constructed in channels wherein flows competent
enough to move coarse bedload are relatively frequent
(Smith, 1974; Karpetta, 1993).
5.1.2. Facies A2 coarse- to medium-grained, multi-
storey sandstone
Facies A2 is contained within coarse- to medium-
grained sandstone bodies up to 11 m thick. These bodies
are erosionally based and sheet-like in overall geometry.
Basal contacts of sandstone bodies are sharp and
undulatory 5th order surfaces, with up to 8 m or more
of relief, and can be traced over the entire study area
(>7 km). These contacts are commonly lined with well-
rounded quartz, siltstone, and tuff clasts averaging 6 cm in
length. Pervasive soft-sediment deformation, commonly
in the form of load casts and convolute bedding (Fig. 6a),
is also common along basal contacts. Upper contacts of
sandstone bodies are mostly sharp 5th order surfaces,
although some are gradational. Upper contacts are locally
marked by interbeds of fine-grained, ripple cross-
laminated or flat-laminated sandstone and siltstone.
Where lateral sandstone-body margins are observed,
they are wedge-shaped pinchouts that are incised as
much as 5 m into previously deposited sediments,
commonly of Facies B1 (Fig. 6b). Interbeds of fine-
grained sandstone and siltstone appear at the margins of
sandstone bodies and bedding becomes increasingly
undulatory and deformed. Overall vertical grain-size
variation is typically irregular or minimal, although some
sandstone bodies show a slight (? crude) upward fining.
Most sandstone bodies in the BCB are multistorey,
displaying well-defined storeys, although they become
more amalgamated and composite in the upper two-
thirds of the formation. Storeys are typically < 3 m in
thickness, but locally reach up to 10 m, and can be
traced for hundreds of meters laterally. Storey bound-
aries are typically sharp and undulatory, but some are
erosional channel surfaces with as much as 8 m of relief
(commonly < 3 m) that incise into underlying fine-
grained sediments as well as other storeys (Fig. 6c). In
some instances, overlying storeys may truncate and
completely cut out underlying storeys. Basal storey
boundaries are lined with granule or pebble lags. Some
storeys are separated from each other by <60 cm of fine-
grained sediments of Facies B1.
14 J.P. Allen, C.R. Fielding / Sedimentary Geology 202 (2007) 534

Fig. 6. Facies A2 and A3. (a) Sandstone bed (Section 7) showing strong soft-sediment deformation (convolute bedding and loading). Note sharp,
undulatory lower contact with underlying sediments of Facies B1. Notebook = 20 cm. (b) View of Section 1 showing the amalgamated nature of
composite sandstone bodies within the upper part of the formation. Lateral pinchout (arrow) into Facies B1. Note the complex interbedded nature of
Facies B1 near sandstone body margin. Section oblique to paleoflow. Section approximately 15 m. (c) Multistorey sandstone (Panel 3) separated by
Facies B1, B2. Note the channel form truncating an underlying parting of Facies B1 and cutting out an underlying sandstone storey. Section oblique to
paleoflow. Geologist = 1.9 m. (d) Bedding plane exposure (Section 14) showing symmetrical wave ripples from the top of a Facies A3 storey. Ripple
crests are arrowed. Wavelengths approximately 50 cm. Notebook = 20 cm. (e) View of Section 19. Notebook = 20 cm. (f) Line drawing interpretation
of (e) detailing herringbone cross-stratification.

The internal architecture of the sandstone storeys in thickness, several hundred meters long, and are
consists predominantly of stacked, tabular or lenticular defined by erosional, slightly concave-up 4th order
NC elements arranged in a multistorey and multilateral surfaces that may coalesce into higher 5th order
framework. Individual NC elements are typically 15 m surfaces. Basal surfaces are planar in most cases, but
 J.P. Allen, C.R. Fielding / Sedimentary Geology 202 (2007) 534
Fig. 7. Photomosaic, line interpretation, and measured sections of Panel 5 (see Fig. 3 for location). Measured sections detail vertical and lateral facies relationships (see Fig. 13 for symbol key). Panel
oriented perpendicular to paleoflow.

15
16 J.P. Allen, C.R. Fielding / Sedimentary Geology 202 (2007) 534

sandstone bodies. In several areas, convex-up cross-

bedding (humpback cross-bedding; cf. Saunderson
and Lockett, 1983) is observed in which topset, foreset,
and bottom set beds are all preserved.

Fig. 8. Cross-bed thicknesses of Facies A2 and A3.

in several architectural panels bases exhibit channel-

form geometries. Because only two architectural panels
are oriented normal to paleoflow, lateral dimensions of
NC elements are not easily discerned, but where
observed these elements range from 50150 m wide
to > 200 m (Fig. 7). Thin, discontinuous fine-grained
partings are commonly observed separating NC ele-
ments, especially lower in the formation. SF elements up
to 2 m in thickness are common at the bases of
sandstone storeys, but may be found throughout the
interval.
LA and DA elements are rare in the lower two-thirds
of the formation, but increase in abundance in the upper
one-third (Fig. 7). Where observed, accretionary
elements are typically between 26 m in thickness and
have a distinctive wedge-shaped geometry bounded by
4th order surfaces with internal 3rd order surfaces.
In some instances Facies A2 is associated with the
finer-grained facies of Association B. These occur as CS
elements. CS elements are typically observed as
lenticular bodies, although one CS element is observed
as a clear channel feature with wings extending out on
either side (Fig. 7). This element has a maximum
channel fill of 1.6 m and is 35 m wide not including
the wings. At the margins, this CS element becomes
increasingly interbedded with ripple cross-laminated
sandstone and siltstone.
Sedimentary structures within Facies A2 are dom-
inantly trough cross-bedding in sets up to 2.0 m thick,
commonly < 0.5 m (Fig. 8). Trough cross-beds are
gently curved in both flow-parallel and transverse
sections and at times appear planar. Trough cross-beds
are typically lined with gravel, locally display rhythmic
grain-size alternations and are preserved as a complex Fig. 9. Rose diagrams summarizing paleocurrent data collected from
mosaic of troughs rather than as coherent cosets. Cross- current generated structures within the Betts Creek Beds from (a)
bed set thickness usually decreases upward through Facies A2, (b) Facies A3, (c) Facies B3.
 J.P. Allen, C.R. Fielding / Sedimentary Geology 202 (2007) 534
Table 3
Calculated values for mean dune height (hm) and mean channel
bankfull depth (d) using mean thickness of cross-sets (sm) using the
methodology of Bridge and Tye (2000)
Mean Standard Mean Estimated
Cross-set deviation dune water
Thickness (ssd) height depth (d)
(sm) (hm)
Facies A2 (n = 315) 0.40 m 0.29 m 1.18 m 7.0611.76 m
Facies A3 (n = 153) 0.43 m 0.25 m 1.25 m 7.5112.52 m
In addition to trough cross-bedding, diffuse, flat
lamination is common in the lowermost storeys.
Discontinuous siltstone partings are common through-
out Facies A2. Partings are typically < 2 cm and are
of limited lateral extent. In some instances, partings
preserve plant material referable to Glossopteris sp.,
Vertebraria, and Cordaites sp.
Soft-sediment deformation (loading and convolute
bedding) is commonly associated with the basal contacts
described above. Horizons of soft-sediment deformation
may be traced laterally over < 100 m's, but these hori-
zons normally vary in both vertical extent and degree of
deformation. Cross-beds are also observed to pass later-
ally into convolute bedding in several places.
Paleocurrent directions measured from trough cross-
beds are predominantly southward (mean = 198;
Fig. 9a), with little dispersion about the mean.
5.1.2.1. Interpretation. The sheet-like external and
internal geometry, overall low dispersion of paleocur-
rent directions, dominantly coarse-grained sediments,
rare evidence of systematic lateral accretion (especially
in the lower portion of the formation), and the
erosionally based, channelized nature of Facies A2 all
indicate deposition within low-sinuosity, most likely
braided, alluvial channels (cf. Bridge, 1985; Bristow
and Best, 1993). On the basis of external and internal
geometry of Facies A2, channels were up to 8 m deep,
but typically 35 m, and sediment accumulated by the
lateral switching of channels across a broad channel
belt. The thickness of individual NC elements is not
considered to be representative of channel depth, due to
the nature of erosional surfaces separating NC elements.
It is suggested that the preserved sandstone bodies are
remnants of larger channel sandstones (cf. Martinsen
et al., 1999). An estimation of channel depth, however,
may be ascertained from cross-bed thickness. A number
of experimental studies have related the set thickness of
cross-strata and dune height to the relationship between
dune height and water depth (Leclair et al., 1997;
Bridge, 1997; Leclair and Bridge, 2001) and these have
17
been applied to both modern and ancient sediments
(Bridge and Tye, 2000; Leclair and Bridge, 2001;
Adams and Battacharya, 2005). In order to calculate
mean bedform height (hm), cross-sets are measured and
a mean (sm) and standard deviation (ssd) are calculated.
If the ssd/sm ratio is 0.88 ( 0.3) mean dune height can
be estimated (Leclair and Bridge, 2001) as:
bism =1:8 1
hm 5:3 b 0:001 b2 2
The relationship between dune height and channel
depth (d) shows a wide range of values that range from
3 < d/hm < 20 (Bridge and Tye, 2000), however it appears
that d/hm values average between 6 and 10 (Allen, 1982;
Bridge and Tye, 2000).
Table 3 shows the estimated bankfull flow depth for
the BCB, which ranges from 7.111.8 m. These values
are in general agreement with observations from erosional
relief observed along channel margins and from bar
accretionary units that can also be used to estimate
channel flow depth (Bridge, 2003). Overall, the channel
systems within the BCB can be interpreted through
paleocurrents, architectural features, and quantitative
estimates as being >400 m wide and 712 m deep.
Channels had considerable erosional capability
that is indicated by the excision of underlying strata.
The nested character of the channels (abundance of
NC elements) suggests that they were repeatedly
eroded and filled, and flood-plain sediments were
largely removed (cf. Reinfelds and Nanson, 1993).
The dominance of sandstone-filled channels and lack of
fine-grained channel fills are consistent with a braided
river interpretation (Bridge, 1985). The fine-grained
partings that separate some storeys and occur as dis-
continuous lenses throughout Facies A2 are interpreted
as remnants of the slack-water bar-top component
of intra- or interchannel islands or small floodplain
deposits (cf. Reinfelds and Nanson, 1993) in a multi-
channel system. Small-scale channels incised into the
upper contacts of sandstone bodies are interpreted as
cross-bar channels.
The dominance of trough cross-bedding and the rarity
of evidence for systematic lateral accretion within this
facies, especially within the lower part of the formation,
suggest that channel floors were relatively devoid of
macroforms, and instead were covered by moderately-
sized dunes that migrated in a predominantly downstream
direction. Considering the complexity of trough cross-
bedding preserved as mosaics rather than coherent cosets,
the dunes were most likely low-relief, three-dimensional,
slightly sinuous dunes. Paleocurrent data also suggest that
18 J.P. Allen, C.R. Fielding / Sedimentary Geology 202 (2007) 534
dune migration had a minor across-channel migration
component. Lateral accretion is observed within the upper
part of the formation and records lateral accretion on braid
bars in these areas. The general paucity of lateral accretion
in the lower part of the formation compared to the upper
part suggests that braiding became more pronounced
during the history of the Betts Creek alluvial system.
The convex-up crossbedding (humpback dunes) is
similar to structures produced experimentally (Simons
et al., 1965; Saunderson and Lockett, 1983) and found
within ancient sequences (Allen, 1983; Re, 1987; Re
and Hermansen, 1993; Fielding and Webb, 1996;
Hjellbakk, 1997). In each of these cases, these structures
were representative of the transition from dune to upper-
stage, plane-bed flow regimes. The abundance of chaotic
and convolute bedding is interpreted as autokinetic
(produced purely by sedimentary processes; sensu
Leeder, 1987) in origin and suggests rapid sediment
accumulation rates (Allen, 1982; Owen, 1996; Re and
Hermansen, 2006). Rapid sediment accumulation rates
are further supported by the occurrence of SF elements
found at the bases of sandstone storeys. Where SF
elements are preserved, they are interpreted as channel-
lag deposits that were transported within the thalweg of
major channels. The coarse nature of the lag deposited
within SF elements indicates that flows were at times
powerful.
Where this facies is associated with Association B as
CS elements, it is interpreted as most likely representing
crevasse splays.
5.1.3. Facies A3 bidirectional trough cross-bedded
sandstone
Facies A3 is similar to Facies A2 in overall character;
however, key differences warrant a separate facies desig-
nation. Facies A3 occurs in sandstone bodies 12 m in
thickness with overall sheet-like geometries. Lower and
upper contacts are commonly sharp, but locally lower
contacts may be gradational. Facies A3 everywhere
overlies Faces A2. Locally, upper bedding plane contacts
preserve symmetrical wave ripples (Fig. 6d) trending
016, and Thalassinoides and Planolites burrows. In
some areas, upper contacts are also associated with
small-scale channels of Facies B1 and B3.
Internally, sandstone bodies of Facies A3 are
composed of NC elements with very similar geometries
to the NC elements found in Facies A2. Trough cross-
bedding is the dominant sedimentary structure, although
small-scale trough cross-bedding, flat lamination, ripple
cross-lamination, and local parting lineation and
carbonaceous/coaly laminae are common near the
upper ( 1 m) bed contacts with either overlying
sandstone storeys of this facies or fine-grained sedi-
ments of Facies B1-3.
Paleocurrent data from trough cross-bedding show a
bidirectional distribution with an overall southward
direction (mean = 188; Fig. 9b) and in some instances
bipolar distributions (herringbone cross-stratification;
Fig. 6e, f).
5.1.3.1. Interpretation. Facies A3 is interpreted as the
deposits of tidally influenced alluvial channels. The
overall similarities and close association with Facies A2
suggest that Facies A3 was deposited under similar
conditions. Channel depth estimates within this facies
are 712 m, similar to those from Facies A2 (Table 3).
Paleocurrent distributions, along with the presence of
herringbone cross-stratification, suggest that the upper
storeys of sandstone beds have been strongly influenced
by reversing flows, most likely related to tidal processes.
Tidal processes have been documented to occur 0
300 km inland from a coeval shoreline in modern envi-
ronments (Gelfenbaum, 1983; Allen, 1991). However,
in this case the increasing mud component, ripple cross-
lamination, wave ripples, and marine trace fossils (i.e.
Thalassinoides) suggest deposition in near-coastal envi-
ronments. The association with small channel fills of an
estuarine character (Facies B3, see below) is consistent
with such an interpretation.
5.1.4. Facies A4 interbedded sandstone and siltstone
Facies A4 is composed of interbedded sandstone and
siltstone in varying proportions, in units up to 3 m thick.
This facies is preserved as CF elements and occurs
predominantly in the upper half of the formation. Lower
contacts are sharp and erosional, and always occur
above sediments of Facies A2. Upper contacts are
erosional, sharp, undulatory, and generally concave-
upward boundaries with overlying Facies A2 sediments.
Although sedimentary structures are not observable in
detail due to inaccessibility, structures representing low-
energy conditions from waning flows such as ripple
cross-lamination and interlamination structures (flaser,
wavy, lenticular, pinstripe bedding) are thought to
dominate within this facies.
Sandstone beds are < 1 m in thickness, exhibiting
lateral variability with beds < 10 m in length. Sandstone
to siltstone proportion decreases upwards.
5.1.4.1. Interpretation. The occurrence of Facies A4
above Facies A2 and its limitation to CF elements
suggest that these sediments were formed during
abandonment of river channels. This facies, therefore,
represents the abandonment fill of the low-sinuosity
 J.P. Allen, C.R. Fielding / Sedimentary Geology 202 (2007) 534 19

Fig. 10. Facies A5. (a) Overall characteristic fabric of Facies A5 showing clast-rich nature (Section 12). (b) Large ( 1.7 m) siltstone raft within the
lower portion of this facies. Note the irregular outline of the clast and that the siltstone still preserves internal lamination. Notebook = 20 cm.

alluvial channel deposits of Facies A2. Facies A4 is
everywhere overlain by Facies A2, suggesting that
abandoned channels within the floodplain were reacti-
vated by subsequent channels (cf. Reinfelds and
Nanson, 1993). This further supports the interpretation
that sediments of A2 accumulated by lateral switching
of active channels across a broad channel belt.
5.1.5. Facies A5 diamictite
Facies A5 is a distinctive, light gray, clast-rich (10
25% clasts), sandy diamictite (Fig. 10a; c.f. Moncrieff,
1989). Clasts range from granules to boulders, < 1.7 m in
length. Clasts are basement (granite, gneiss, quartz),
first-cycle volcanic tephra, and intraformational siltstone
and sandstone. Intraformational clasts are irregular
and typically preserve internal bedding (small-scale
trough cross-bedding, ripple cross-lamination, flat
lamination) and soft-sediment deformation (Fig. 10b).
Clasts are poorly sorted and matrix-supported. Facies
A5 is typically unstratified, but crude bedding and
clast imbrication (indicating flow to the southeast)
have been observed at Section 19 (see Fig. 3 for loca-
tion). Wispy silt laminae, coaly traces, and macerated
plant material are common in the upper part of this
facies.
Facies A5 is rare, laterally discontinuous, and occurs
exclusively within CF elements that may extend for
several hundred meters. The maximum thickness of
Facies A5 at outcrop is 3 m. It commonly overlies Facies
B1 and basal contacts are sharp and undulatory. Upper
contacts are sharply defined by the overlying Facies A2
or locally by Facies B1.
20 J.P. Allen, C.R. Fielding / Sedimentary Geology 202 (2007) 534

5.1.5.1. Interpretation. Facies A5 is interpreted as
reworked pyroclastic material deposited by debris flows.
The extremely poor sorting, lack of internal stratifica-
tion, and presence of matrix-supported clasts are all
characteristic of subaerial debris flows documented from
both ancient and modern environments, where deposi-
tion is en masse rather than by grain-by-grain sedimenta-
tion (Nemec and Steel, 1984; Smith, 1986; Smith and
Lowe, 1991; Major, 1997). Intraformational clasts with
irregular shapes and displaying internal stratification
suggest that debris flows were relatively viscous and
lacked turbulence (cf. Jorgensen and Fielding, 1999;
Kataoka and Nakajo, 2002) in order for these features to
be preserved during transport and deposition.
It is believed that debris flows formed in association
with contemporaneous volcanism, with source area(s) to
the north and/or east. During this time interval wide-
spread volcanic activity was occurring on the north-
eastern margin of the Bowen Basin as a result of
foreland tectonics and the establishment of a continental
volcanic arc (Fielding et al., 2001), and may have
provided the volcaniclastic detritus that sourced debris
flows. Paleocurrent data from the Bowen Basin show
that during the Late Permian (263258 Ma) alluvial
conditions were established that led to an overfilling
situation in which first-cycle volcanic detritus spilled
across the Springsure Shelf into the Galilee Basin
(Fielding et al., 2001). The absence of coeval primary
proximal volcaniclastic facies associated with Facies A5
suggests that the location of volcanism was far away and
most likely extrabasinal, which is consistent with an
interpretation that volcanism in the Bowen Basin
sourced debris flows. While debris flows generated on
alluvial fans rarely extend further than 10 km, volcanic
debris flows can travel > 100 km away from the source
(Smith, 1986; Mothes et al., 1998). Facies A5, therefore,
records evidence for distal debris-flow deposition as the
result of large-scale volcanism several hundred kilo-
meters away. The paucity of airfall tuff beds within the
study area is noteworthy. Although no coeval primary

Fig. 11. Facies B1 and B2. (a) View of proximal Facies B1 at Section 10. Note the loading associated with the basal contacts of sandstone beds.
Some sandstone beds pinch out laterally (arrow). Notebook = 20 cm. (b) View of distal Facies B1. Scale = 15 cm. (c) Facies B1 at Section 13
displaying rhythmic alternations of fine-grained sandstone and siltstone. Scale bar = 50 cm. (d) View of Faces B2 at Section 13 showing the abrupt
contact with overlying Facies A2. Note the small channel form incised and filled with Facies B1 (arrowed). Section is oblique to paleoflow. Section
approximately 8 m.
 J.P. Allen, C.R. Fielding / Sedimentary Geology 202 (2007) 534 21

volcaniclastic sediments occur within the northeastern
Galilee Basin, explosive volcanism was ongoing during
the Late Permian in the Bowen Basin. Airfall from such
explosive volcanism could reach the study area, but
recent climate models predict westerly trade winds
(offshore) during the Permian winter and southerly
directed winds during summer (Gibbs et al., 2002).
Under these circumstances, airfall from volcanism
would not be expected in the northeastern portion of
the Galilee, because prevailing surface winds were
blowing in the opposite direction.
All BCB debris-flow deposits are confined to CF
elements that have sharp, concave-up, channelized bases.
These debris flows probably exploited pre-existing
topography by occupying existing channels. Debris
flows within volcanic settings have been shown to follow
drainage paths and be partially or completely confined to
channels (Pierson and Scott, 1985). The intraformational
siltstone and sandstone clasts in the BCB are most likely
the result of floodbasin and channel sediment entrainment
during the occupation of channels by debris flows. Further
support for this interpretation lies in the characteristics of
Facies A5 at Section 19, where clast imbrication and
crude bedding appear. Imbrication is consistent with a
debris-flow interpretation (Smith, 1986), in which
particles at the base of debris flows display a directional
fabric formed by pulsed shearing within the flow body
(Mulder and Alexander, 2001). The direction of imbrica-
tion predominantly to the southeast is comparable to the
overall paleoflow measured from facies within the

Fig. 12. Facies B3. (a) Photograph is oblique to bedding showing the overall bioturbated fabric within Facies B3 (BI = 45). Diplocraterion habichi
(Dh) is preserved on bedding surfaces, while Conichnus (C) traces are preserved on a vertical section. (b) Cross-sectional view of Thalassinoides
(Th) burrow.
 22
J.P. Allen, C.R. Fielding / Sedimentary Geology 202 (2007) 534
Fig. 13. Measured sections oriented in a NS transect (parallel to paleoflow) detaling the overall sheet-like architecture of sequences and the internal variablility between individual sequences (see text
for details). See Fig. 3 for section locations. Note that correlations are traceable across continuous cliff-exposures within the study area.
J.P. Allen, C.R. Fielding / Sedimentary Geology 202 (2007) 534 23

Fig. 13 (continued ).
24 J.P. Allen, C.R. Fielding / Sedimentary Geology 202 (2007) 534
channel association, suggesting flow was confined by pre-
existing topography.
5.2. Facies Association B floodbasin deposits
5.2.1. Facies B1 interbedded to interlaminated
sandstone and siltstone
Facies B1 comprises intervals up to 6 m thick of
interbedded to interlaminated sandstone and siltstone.
The interbedded/interlaminated nature and the proportion
of sandstone to siltstone within this facies vary markedly
depending on where in the formation it is found (laterally
and vertically) and the relationships to other facies. As a
result, this facies is typically a complex mosaic of
sandstone and siltstone that varies laterally and vertically.
Two endmembers (proximal and distal) are
described below in order to give an adequate description
of the range of characteristics found within Facies B1. In
general, sandstone beds are medium- to fine-grained and
have sharp, basal and upper contacts. Siltstones are
medium to dark gray and are characterized by a number
of interlamination structures including pinstripe (linsen),
lenticular, and wavy lamination. Siltstones also contain
significant amounts of tuffaceous material.
The proximal endmember (Fig. 11a) is recognized
where Facies B1 is connected to channel margins of
Facies A2. Here this facies consists of interbedded, fining-
upwards 70:30 to 50:50 sandstone:siltstone,. Sandstones
range from medium to very fine, but are mostly medium-
grained. Individual beds are typically 210 cm thick.
Both upper and lower bed contacts are sharp and
undulatory, and lower contacts show varying amounts
of soft-sediment deformation mainly in the form of
loading and localized convolute bedding (Fig. 11a). Many
individual sandstone beds are laterally persistent, but
some have been observed as lensoid bodies <5 m wide.
Sandstones contain small-scale cross-stratification and
ripple cross-lamination. Siltstone beds are 15 cm thick
and contain mainly pinstripe and lenticular bedding, as
well as abundant ripple cross-lamination (locally occur-
ring as climbing ripples). No trace fossils have been
observed within the proximal endmember.
Where associated with channel sandstones of Facies
A2, the thickly interbedded nature of Facies B1 is
restricted laterally, and with increasing distance from the
channel margin, sand percentage and grain-size decrease
and the unit as a whole passes and interfingers into
thinly interbedded and increasingly more interlaminated
sandstones and siltstones of the more distal end-
member (Fig. 11b). This lateral change occurs over a
few tens of meters. In several locations the proximal
variant of Facies B1 occurs within LE elements with
inclined accretionary layers dipping 510 at the
margins of channel bodies (Fig. 7).
Siltstones of the distal endmember of Facies B1 contain
the same features as mentioned above but additionally
show abundant coaly laminae, rare Planolites, and
starved ripples. These siltstones also preserve fragments
of Glossopteris, Vertebraria, and Cordaites. This end-
member of facies B1 occurs within OF lithosomes and is
laterally persistent over several hundred meters.
Additionally, Facies B1 appears locally as medium to
dark gray massive to fissile siltstone with an increasing-
upward organic content. Plant remains of Glossopteris,
Vertebraria, Cordaites, and other unidentifiable plant
fragments are common on bedding planes. This variant
of Facies B1 is commonly found as partings within or
separating sandstone beds of Facies A2 and A3, and
typically passes vertically into Facies B2.
At one locality, Facies B1 displays rhythmic lamina-
tions of heterolithic fine-grained sand and silt (Fig. 11c).
5.2.1.1. Interpretation. Facies B1 is interpreted as
floodbasin deposits that range from environments
proximal to channel margins to distal locations not
influenced by active channel processes. The proximal
endmember of Facies B1, occurring laterally adjacent to
channel margins and composed of interbedded sand-
stone and siltstone, represents a proximal floodbasin
setting, with sediment deposited by initially high energy
sheet flows from waning flood currents. These condi-
tions produced the sharply bounded nature of sandstone
bodies, and then as flow energy decreased, low-energy
currents and suspension fallout sedimentation occurred.
Where this facies occurs as LE elements, it is interpreted
as levee deposits as the geometry is similar to other
examples (cf. Fielding et al., 1993; Michaelson et al.,
2000; Ray and Chakraborty, 2002). Depositional dips in
Facies B1, representing the alluvial-ridge topography,
the lateral fining into more distal floodbasin deposits,
and close association with channel margins, are all
consistent with a proximal overbank/levee interpretation
(cf. Fielding et al., 1993).
Thicker and more laterally extensive sandstone beds
within Facies B1 are similar to splay deposits (cf.
Farrell, 1987; Tye and Coleman, 1989; Jorgensen and
Fielding, 1996), and most likely represent the proximal
portions of crevasse splays. However, for the most part,
sandstone beds are not laterally persistent, and crevas-
sing is not thought to have been a major process.
The distal endmember of Facies B1, due to its lateral
and vertical relationships, suggests processes similar to
those of the proximal end member. The paucity of
sandstone beds and abundance of interlamination
 J.P. Allen, C.R. Fielding / Sedimentary Geology 202 (2007) 534
structures representative of a lower flow regime
suggest that sedimentation was primarily due to
suspension fallout. In these cases, Facies B1 was not
affected by active channel processes but deposition
occurred mainly during sheet-flood events in shallow
ponds. The rarity of trace fossils suggests that these
ponds were anoxic or dysoxic. Here, and where
associated with Facies B2, Facies B1 represents a
transitional environment from shallow, stagnant
ponds, to peat-forming mires where abundant plant
material and coaly laminae occur.
At one locality, Facies B1 shows evidence of
influence from tidal processes. The deposits of fine-
grained sandstone and siltstone in rhythmic alternations
are interpreted as tidal rhythmites (Fig. 11c).
5.2.2. Facies B2 carbonaceous siltstone/coal
Facies B2 is composed of laterally extensive
carbonaceous siltstone and coal in units up to 2.5 m
thick in the study area, but reaching up to 10 m thick in
cored sections (GSQ Hughenden 3-4R). Basal contacts
are gradational from Facies B1, whereas upper bound-
aries are sharp and erosionally truncated by Facies A2
(Fig. 11d). At the lower contacts, Facies B2 is variably
interbedded with Facies B1. Where well-developed, this
facies becomes increasingly coal-rich upward. This
facies displays lateral continuity over several kilometers
in extensive OF lithosomes. Abundant plant remains are
preserved at several horizons, including Glossopteris,
Vertebraria, Araucarioxylon-like wood, and Cordaites.
In situ tree stumps up to 20 cm wide are preserved
at the top of Facies B2 in one area. Stumps are
rooted into a carbonaceous horizon that also contains
well-preserved Glossopteris foliage, and are buried by
sediments of Facies B1 displaying centroclinal cross-
stratification.
In several places, sharply bounded sandstone bodies of
Facies A2 up to 2.5 m thick occur within this facies.
5.2.2.1. Interpretation. Facies B2 is the result of peat
accumulation in coastal wetland environments. Peat
accumulation was evidently in situ from the abundance
of roots (Vertebraria). The number of channelized sand-
stone bodies (Facies A2) within this facies suggests that
the wetland may have been crossed by a network of small
channels. The extensive sheet-like geometry of Facies
B2 suggests that the mires were predominantly low-
lying, rheotrophic environments, rather than raised and
rain-water fed swamps (McCabe, 1984). The association
of Facies B2 with Facies B3 (see below) suggests that
these were coastal wetland environments situated on a
coastal plain near estuarine embayments.
25
5.2.3. Facies B3 bioturbated interbedded sandstone
and siltstone
Facies B3 comprises intervals of bioturbated inter-
bedded very fine- to fine-grained sandstone and siltstone
no more than 50 cm and in most places < 10 cm thick.
Medium, gray, sandy siltstone is the predominant
lithology within this facies with siltstone:sandstone
proportions varying from 50:50 to 90:10 depending on
location. Lower contacts are typically sharp. Facies B3 is
everywhere observed overlying sediments of Facies A3.
Upper contacts are commonly gradational into Facies
B1, but are locally sharply bounded by Facies B1 and
B2. This facies occurs at several stratigraphic horizons
and can be traced laterally over at least 3 km within OF
lithosomes.
Sandstone beds are sharply bounded and some
are discontinuous, passing laterally into sandy siltstone.
Sandstones are < 3 cm in thickness and commonly ripple
cross-laminated (current-, wave-, and combined flow-
generated). Siltstones are characterized by a variety of
interlamination structures ranging from pinstripe (lin-
sen) to lenticular and wavy lamination, and thicker beds
are internally flat or ripple cross-laminated. Measure-
ments from ripple cross-lamination indicate paleoflow
to the southwest (mean resultant = 190; Fig. 9c),
although azimuths range from northwest to southeast.
Small synaeresis cracks are common throughout this
facies, and other soft-sediment deformation structures
occur locally. At one location (Section 6) clastic dikes
composed of very fine- to fine-grained sandstone were
observed at the lower contact of this facies with
underlying Facies A3. Sandstone dikes are variable in
width ranging from 1020 cm and up to 50 cm in length.
No preferred orientation was observed.
The intensity of bioturbation is variable throughout
Facies B3, but typically decreases upwards. Bioturba-
tion varies from location to location and within indi-
vidual beds. The Bioturbation Index (BI, Bann et al.,
2004: where BI = 0 is no bioturbation and BI = 6 repre-
sents complete bioturbation) ranges from 0 to 5.
The trace fossil suite is low in diversity and domi-
nated by Diplocraterion habichi (Fig. 12a), Conichnus
(Fig. 12a), and Thalassinoides (Fig. 12b), with minor
Planolites and Palaeophycus sp. D. habichi are most
often observed in plan view as small (15 mm) paired
vertical burrows. The other traces were readily identified
in both plan and cross-sectional views. The trace fossils
have reduced sizes, except in a few instances where
especially large individuals (< 15 cm) of Conichnus
were noted. Trace fossils occur throughout Facies B3
and do not show any preferential clustering at any one
stratigraphic horizon.
26 J.P. Allen, C.R. Fielding / Sedimentary Geology 202 (2007) 534

Small-scale channel forms (2 20 m) were observed 5.2.3.1. Interpretation. Facies B3 was deposited
in several areas incised into sediments of Facies A2, A3, within estuaries or coastal embayments. The dominance
B1, B3 and infilled by Facies B3. of ripples (wave-and tide-influenced) and wavy and

Fig. 14. Composite measured section of the Betts Creek Beds detailing vertical facies variations and sequence stratigraphic interpretations. Key as in
Fig. 13.
 J.P. Allen, C.R. Fielding / Sedimentary Geology 202 (2007) 534
lenticular bedding implies deposition within a shallow-
water environment. Bipolar and bimodal paleocurrent
distributions also reflect tidal processes.
Synaeresis cracks suggest deposition in a perma-
nently subaqueous, brackish-water setting. Cracks found
throughout Facies B3 have been interpreted as the
product of synaeresis rather than desiccation due to the
lack of well developed polygonal patterns on bedding
planes, elongate to trilete planform, ptygmatically folded
cross-section, and the absence of other evidence that
would suggest subaerial exposure (soil development,
roots, etc.). Synaeresis cracks form in response to tension
as muddy, water-saturated sediment loses water to an
overlying fluid layer causing an increase in the strength
of interparticle forces within the clays (cf. Plummer and
Gostin, 1981; Allen, 1982). They typically form in
muddy environments where there are large fluctuations
in salinity (Burst, 1965). Salinity variations during or
immediately following the deposition of muddy sedi-
ments induced the formation of synaeresis cracks in
experiments (Plummer and Gostin, 1981; Allen, 1982).
The appearance of synaeresis cracks in association with
sandstone beds in Facies B3 suggests that sand
deposition was accompanied by an influx of saline water.
The trace fossil assemblage is also consistent with a
brackish-water interpretation because it contains a
mixture of size-reduced, simple dwelling and feeding
structures (deposit-feeders dominant, with less abundant
suspension feeders) characteristic of opportunistic
suites. It is a restricted expression of a mixed Sko-
lithos/Cruziana Ichnofacies that is characteristic of
inshore coastal environments (cf. Bann et al., 2004).
Typical estuarine trace fossil assemblages from both
modern and ancient settings are characterized by (1) low
diversity, (2) ichnotaxa typically found in marine
environments, (3) reduced size compared to open
marine counterparts, (4) simple burrows constructed
by trophic generalists, and (5) vertical and horizontal
burrows common to both Skolithos and Cruziana
Ichnofacies (Howard and Frey, 1973; Ekdale et al.,
1984; Beynon and Pemberton, 1992; Pemberton et al.,
1992; Bann and Fielding, 2004; Bann et al., 2004).
These criteria can all be recognized within Facies B3.
The preponderance of diminutive traces found within
Facies B3 is considered to be an adaptive response to
salinity-induced physical and chemical stresses. Size
reduction is most noticeable in organisms that maintain
constant contact with surficial brackish waters, whereas
deeper, deposit-feeding structures display less size
reduction (Bann and Fielding, 2001). Larger forms of
Conichnus are believed to represent sites at a distance
from the brackish water interface.
27
The combined sedimentological and ichnological
data, therefore, point to a highly stressed brackish water
environment (i.e. coastal/estuarine embayment) that was
influenced by both wave and tidal processes.
6. Stratigraphic and sequence architecture
One of the most distinctive features in the vertical
stacking patterns of the Betts Creek Beds is a series of 6
unconformity-bounded cycles each 1014 m thick
(Fig. 13). These cycles could represent autogenic cycles,
high-frequency parasequences (Van Wagoner et al.,
1988; Naish and Kamp, 1997), or true sequences formed
in low-accommodation settings (Kidwell, 1997; Field-
ing et al., 2000; Bann et al., 2004). Given that the BCB
represent the entire Late Permian Galilee Basin fill
( 20 myr), it is likely that these cycles represent true
sequences. The BCB display the typical sequence
architecture characteristic of the examples cited above
that were preserved in low-accommodation settings
where they are thin, condensed, and show considerable
top truncation of the highstand systems tract.
Although autogenic cycles may result in regular
sedimentary cycles, it is unlikely that semirandom
processes such as river avulsion and mire growth and
accumulation could be the controlling force on the BCB
cycles. Given the period of time represented by the
BCB, the facies distribution would be expected to be
truly random if autocyclicity were the controlling
mechanism. Evidence suggesting that such autocyclic
processes operate on timescales of < 103 years (Calder
and Gibling, 1994), supports an allocyclic sequence
development for the cycles within the BCB rather than
autogenic cyclicity.
The recognition of these cycles as sequences and the
ability to separate the sequences into conventional
systems tracts (LST, TST, HST), rely on subtle
indicators, such as tidally influenced fluvial deposits
(Facies A3) and thin, discrete horizons of bioturbated
estuarine deposits (Facies B3) interpreted as flooding
surfaces (Fig. 14). If not examined in detail, the BCB
could easily be mistaken for the archetypal amalga-
mated stacked fluvial deposits characteristic of updip
basin margin settings (i.e. Shanley and McCabe, 1993;
Miall, 1996).
6.1. Sequence boundaries
Major erosion surfaces interpreted as sequence
boundaries are defined by the bases of coarse-grained,
laterally extensive, braided fluvial sandstones that erode
into underlying coastal plain coals. These surfaces can
 28
J.P. Allen, C.R. Fielding / Sedimentary Geology 202 (2007) 534
Fig. 15. Photomosaic and line interpretation of Panel 7 (see Fig. 3 for location), detailing the variability of the internal sequence architecture. Note the occurrence of CF elements within Sequence 4 and
the amalgamated nature of NC elements within Sequences 5 and 6. Panel oriented parallel to paleoflow.
 J.P. Allen, C.R. Fielding / Sedimentary Geology 202 (2007) 534
be traced continuously in the cliff exposures over the
entire study area, and therefore represent regionally
extensive erosional surfaces. In addition, they can be
distinguished from other, more local, erosional surfaces
caused by autogenic processes. The erosional relief on
these surfaces can only be observed locally where
channel margins are observed, and is typically < 5 m.
There is no evidence that they have any significant
regional erosional relief and are more-or-less planar in
nature at the scale of several km's (Fig. 13).
These surfaces differ from scoured erosional surfaces
characteristic of fluvial environments resulting from
autogenic processes, because they are more prominent
and widespread. The sequence boundaries within the
BCB are similar to other interpreted sequence bound-
aries in alluvial strata (Van Wagoner et al., 1990; Shanley
and McCabe, 1991, 1993, 1994; Gibling and Bird, 1994;
Olsen et al., 1995; Yoshida et al., 1996), in that there is a
sharp juxtaposition between braided fluvial sandstones
and carbonaceous coastal-plain deposits marking a
change in the amount of accommodation space. How-
ever, the sequence boundaries within the BCB are planar
and show minimal erosional relief, similar to sequence
boundaries observed in the Ericson Sandstone by
Martinsen et al. (1999). A planar sequence boundary is
to be expected in low-gradient settings such as basin
margins. Where the slope of the fluvial profile near the
coastline is similar to the bathymetric profile of a shallow
shelf, only minor fluvial incision will occur and incised
valleys tend to be relatively wide (Shanley and McCabe,
1994; Posamentier and Allen, 1999). Under such
circumstances the fluvial systems may simply be
extended. This is in contrast to areas where there is a
distinct difference in slope between the alluvial/coastal
plain and shallow marine shelf, which favors valley
incision (Shanley and McCabe, 1994; Posamentier and
Allen, 1999). Flume studies (Wood et al., 1991; Koss
et al., 1994), theoretical models (Schumm, 1993;
Posamentier and Allen, 1999), and outcrop studies
(Shanley and McCabe, 1991, 1994; Martinsen et al.,
1999) suggest that planar sequence boundaries and
fluvial sheet sandstones are common in low-accommo-
dation settings during base-level fall and early rise.
6.2. Flooding surfaces
As mentioned above, flooding surfaces (FS) within
the BCB are represented by the thin deposits of Facies
B3. FS occur as sharp contacts between tidally
influenced, alluvial-channel facies and estuarine/coastal
embayment facies (Facies B3) marking an increase in
water depth. Where FS occur between Facies A3 and B3
29
(TST deposits) and Facies B1, B2 (HST deposits), they
may represent the maximum flooding surface (MFS;
Fig. 14). A MFS interpretation, however, is tentative:
nonetheless, these intervals are interpreted as having
formed under conditions of greatest water depths and
minimal sediment supply.
6.3. Sequence architecture
Sequences show an overall, extensive sheet-like
architecture which varies little across the study area
(Fig. 13). Sequences consist of a basal, sheet-like,
braided fluvial sandstone complex predominantly made
up of stacked tabular or lenticular NC elements arranged
in a multistorey and multilateral framework that ero-
sionally overlies sediments of the underlying sequence
(typically OF elements; Fig. 15).
Overlying the sandstone complexes are extensive,
sheet-like, floodbasin and coastal plain OF lithosomes
locally containing LE and CS elements. OF elements
contain a mosaic of facies that vary both laterally and
vertically, but always begin with Facies B3 at the base,
overlain by Facies B2 and B1. OF lithosomes are ero-
sionally overlain by a sheet-like NC sandstone complex
that marks the base of the overlying sequence (Fig. 15).
6.3.1. Variation in sequence architecture
The sequences of the BCB show variation in internal
architecture vertically through the formation. The se-
quences can be divided on this basis into 4 groups
(Sequence 1; Sequences 2, 3; Sequence 4; Sequences 5,
6) within each of which architecture is similar, but
between which, architecture differs.
6.3.1.1. Sequence 1. Sequence 1 is composed entirely
of Facies A1, and is markedly different from the
overlying sequences. It is typically 11 m in thickness
and thins slightly southward. This sequence consists
mostly of SF elements, although some CH features are
also observed. SF elements are broad and shallow
(2 50 m) and form a complex mosaic that characterizes
this basal sequence of the formation.
6.3.1.2. Sequences 2, 3. Sequences 2 and 3 are
characterized by the overall architecture outlined
above. These sequences contain the best developed
OF lithosomes (Fig. 15), but this may possibly be the
result of bias in data collection due to the fact that they
are easily accessed from creek level, whereas other
sequences are in most places inaccessible vertical cliff
exposures. Only two accretionary elements (DA, LA)
have been observed in these lower sequences, where
30 J.P. Allen, C.R. Fielding / Sedimentary Geology 202 (2007) 534
they overlie an OF lithosome. The two elements are at
the same stratigraphic horizon and they may represent
the same architectural feature.
6.3.1.3. Sequence 4. Sequence 4 is similar to
Sequences 2 and 3; however, this sequence contains
the vast majority and best preservation of CF elements
(Fig. 15). As a result of the abnormally higher
proportion of CF elements within this sequence as
compared to the other sequences, Sequence 4 has a
much higher proportion of mud-to-sand.
6.3.1.4. Sequences 5, 6. These sequences are similar
in overall nature to Sequences 2 and 3; however, there
are some key differences. The basal sandstone com-
plexes within these sequences have a much more
amalgamated character in comparison to those of under-
lying sequences (Fig. 15). NC elements are laterally
restricted and are typically <100 m in length. This is in
contrast to NC elements of the previous sequences that
in some instances measure > 500 m in length. Observa-
tions indicate a general paucity of fine-grained partings
between individual NC elements.
Along with NC elements, basal sandstone complexes
also consist of LA and DA elements. Accretionary
elements typically average 3 m in thickness, although
in Sequence 6, LA elements measuring 67 m in
thickness are observed (Fig. 7).
7. Discussion
The above architectural analysis highlights several
different controlling forces on the sediment body
architecture of the Betts Creek alluvial system. The
overall sediment body architecture of the BCB was
affected by allogenic controls, and we argue that sea level
is the only control whose signal can account for the
genetic sequence architecture of the formation. The
recognition and interpretation of Facies B3 as an estuarine
deposit and the subsequent development of a sequence
stratigraphic framework for the BCB demonstrate that
changes in relative sea level had a profound control on the
overall sediment body architecture of the formation.
Given the basin-marginal context and evidence for a
steady, slow subsidence regime throughout accumulation
of the BCB, it is considered most likely that absolute
(eustatically driven) sea-level fluctuations were largely
responsible for the observed sequence patterns. While
climate has major influences on the stratigraphic
architecture of fluvial systems (Fielding and Webb,
1996; Blum and Trnqvist, 2000; Ray and Chakraborty,
2002; Bridge, 2003), and fluvial systems reflect the
influence of both sea level and climate change (Blum and
Trnqvist, 2000), the effects of sea-level change, in this
study, are believed to have overwhelmed any recogniz-
able climatic signature. Poor age constraints would also
make any estimates of a climatic signal difficult.
The low-accommodation, basin-marginal setting, the
overall lack of evidence to suggest changes in tectonic
subsidence rates (stratigraphic discordance, changes in
paleocurrent direction, increased soft-sediment defor-
mation, etc.), along with evidence for periodic marine-
influenced conditions argue persuasively for a strong
sea-level control on the genetic fluvial architecture. The
relatively homogeneous overall sheet-like, sequence
architecture, therefore, is reflective of sea-level change
and suggests that the magnitude of change responsible
for creating these sequences was relatively consistent
during the Late Permian. The BCB were deposited at the
end of the Late Paleozoic Ice Age, and recent evidence
from eastern Australia suggests that during BCB
deposition, Gondwana experienced at least two discrete
periods of glaciation (Fielding et al., 2006; Birgenheier
et al., 2006). It is believed that the sequences within the
BCB are the result of glacioeustatic sea-level changes
caused by the expansion and contraction of polar ice
volumes on Gondwana.
While sea-level change is invoked to explain the
sequence architecture, other controls must also be called
upon in order to explain the internal variability of the
sequences illustrated in previous sections.
The tectonic setting of the Galilee and adjacent
Bowen Basins provides insights into one of the controls
affecting the internal stratigraphic architecture of the
BCB. During the Late Permian, both basins were ex-
periencing a period of foreland loading in response to
thrust sheet propagation from the Hunter-Bowen Event
(Fielding et al., 2001; Van Heeswijck, 2004). This
period of foreland loading is associated with widespread
and explosive volcanism in the north-northeastern por-
tion of the Bowen Basin, which is believed to have
sourced the volcanic detritus observed in the BCB.
Sudden influxes of volcanic debris have been shown to
drastically alter fluvial sedimentation patterns in several
studies (Besly and Collinson, 1991; Fielding et al.,
1993; Nakayama and Yoshikawa, 1997; Kataoka and
Nakajo, 2002). This influx of fine-grained material is
thought to have increased the preservation of abandoned
channel deposits that has resulted in architecture with a
higher proportion of CF elements and a fluvial system
with an increased mudstone-to-sandstone proportion as
observed in Sequence 4. Braided systems with high
mud-to-sand proportions have been documented from a
variety of tectonic settings including extensional basins
 J.P. Allen, C.R. Fielding / Sedimentary Geology 202 (2007) 534
(Mack and Seager, 1990; Mack and James, 1993),
intermontane basins (Lorenz and Nadon, 2002), and
foreland basins (Bentham et al., 1993; Willis, 1993;
Adams and Battacharya, 2005). While tectonic pro-
cesses in the basin-marginal setting may not have been
an important control on the fluvial style of the BCB,
tectonic processes and explosive volcanism in the
source area (i.e. the Bowen Basin) may have been (cf.
Mack and James, 1993; Lorenz and Nadon, 2002).
While the majority of the BCB has a very low
mudstone-to-sandstone proportion, this section of the
formation adds to a growing number of examples of low
net-to-gross fluvial systems that challenge classic
braided fluvial facies models (Walker and Cant, 1984).
The internal architecture of the BCB was also
influenced by more localized autogenic controls. The
increased amalgamation of channel sandstones observed
in the higher sequences (Sequences 5, 6) is believed to
be a result of increased rates of avulsion. Considering
that avulsion rates are influenced by a number of
variables (sedimentation rate, sinuosity increase, base
level, climate, tectonic uplift, etc.; see Jones and
Schumm, 1999), there are several possible explanations
for an increased rate of avulsion in the upper portion of
the BCB. However, the most likely explanation is the
interplay between changes in base level (sea level in this
scenario) and tectonics, and the subsequent effects on
sedimentation rates these processes would have caused.
Mackey and Bridge (1995) and Heller and Paola (1996)
have shown that increased rates in sediment supply
increase rates of avulsion frequency. Increased rates of
sediment supply caused by periodic explosive volcan-
ism are envisaged to have caused increased topographic
relief of channels above the floodplain, which led to a
greater increase in avulsion frequency. These increased
rates of sediment supply may also explain the increase in
channel depth and prevalence of braiding (as indicated
by the occurrence of accretionary elements in higher
sequences) in this part of the formation.
8. Conclusions
The Betts Creek Beds represent deposition on an
alluvial/coastal-plain setting and consists of two major
facies associations: channel and floodbasin. The channel
facies association includes the deposits of low-sinuosity,
braided rivers, tidally influenced rivers, and fluvially
reworked, pyroclastic deposits. The floodbasin facies
association consists of proximal to distal floodplain,
swamp/mire, and estuarine/coastal embayment deposits.
The architecture of the formation is characterized
by 6 sequences that are each 1014 m in thickness,
31
condensed, and top-truncated. The internal architecture
of the sequences is characterized by amalgamated sheet-
like, multistorey, multilateral braided fluvial sandstones
overlain by estuarine mudstones and coastal-plain
carbonaceous shales and coals. The overall sequence
architecture was controlled by changes (presumed
eustatic) in sea level, while the internal architecture,
which is more variable and complex, was the result of a
combination of sea level, tectonic, and autogenic controls.
The fact that well-developed, albeit thin and
condensed, sequences can be preserved and recognized
in a low-accommodation, basin-marginal succession is
quite striking. Traditionally, such settings have proved
difficult to analyze by sequence stratigraphic methods,
because key surfaces are not easily discerned from more
local erosional surfaces. The nature of the BCB outcrop
and the recognition of discrete marine-influenced units
interpreted as flooding surfaces allow for a high-
resolution sequence framework to be established in a
unit that could easily be mistakenly interpreted as
archetypal amalgamated, stacked fluvial deposits char-
acteristic of updip basin-margin settings.
Acknowledgements
The authors would like to thank the reviewers whose
critical comments considerably improved the quality of
this manuscript. This research was supported by the Mr.
and Mrs. J.B. Coffman endowment in sedimentary
geology at the University of Nebraska-Lincoln. We
would also like to thank the Queensland Parks and
Wildlife Service for allowing research to be undertaken
and samples collected at Porcupine Creek Gorge.
References
Adams, M.M., Battacharya, J.P., 2005. No change in fluvial style
across a sequence boundary, Cretaceous Blackhawk and Castlegate
Formations of Central Utah. U.S.A. J. Sediment. Res. 75,
10381051.
Allen, J.R.L., 1982. Sedimentary Structures: Their Character and
Physical Basis. Elsevier, Amsterdam. 593 pp.
Allen, J.R.L., 1983. Studies in fluviatile sedimentation: bars, bar-
complexes and sandstone sheets (low-sinuosity braided streams) in
the Brownstones (L. Devonian), Welsh Borders. Sediment. Geol.
33, 237293.
Allen, G.P., 1991. Sedimentary processes and facies in the Gironde
estuary: a recent model for macrotidal estuarine systems. In:
Smith, D.G., Reinson, G.E., Zaitlin, B.A., Rahmani, R.A. (Eds.),
Clastic Tidal Sedimentology. Memoir Can. Soc. Petrol. Geol.,
vol. 16, pp. 2940.
Anthony, D.P., 2004. A review of recent conventional petroleum
exploration and in-field gas reserves growth in the Denison
Trough, Queensland. Eastern Australasian Basins Symposium II.
Petrol. Explor. Soc. Austr. Adelaide, pp. 276295.
32 J.P. Allen, C.R. Fielding / Sedimentary Geology 202 (2007) 534
Bann, K.L., Fielding, C.R., 2001. Applications of ichnology to
hydrocarbon exploration-examples from the Permian of eastern
Australia. In: Hill, K.C., Bernecker, T. (Eds.), Eastern Australasian
Basins Symposium 2001. Spec. Publ. Petrol. Explor. Soc. Austr.
AustIMM, Melbourne, pp. 603612.
Bann, K.L., Fielding, C.R., 2004. An integrated ichnological and
sedimentological comparison of non-deltaic shoreface and sub-
aqueous delta deposits in the Permian reservoir units of Australia.
In: McIlroy, D. (Ed.), The Application of Ichnology to Palaeoen-
vironmental and Stratigraphic Analysis. Geol. Soc. Spec. Publ.,
vol. 228, pp. 273307.
Bann, K.L., Fielding, C.R., MacEachern, J.A., Tye, S.C., 2004.
Differentiation of estuarine and offshore marine deposits using
integrated ichnology and sedimentology: Permian Pebbley Beach
Formation, Sydney Basin, Australia. In: McIlroy, D. (Ed.), The
Application of Ichnology to Palaeoenvironmental and Strati-
graphic Analysis. Geol. Soc. Spec. Publ., vol. 228, pp. 179212.
Bentham, P.A., Talling, P.J., Burbank, D.W., 1993. Braided stream
and floodplain deposition in a rapidly aggrading basin: the
Escanilla Formation, Spanish Pyrenees. In: Best, J.L., Bristow,
C.S. (Eds.), Braided Rivers. Geol. Soc. Spec. Publ., vol. 75,
pp. 177194.
Besly, B.M., Collinson, J.D., 1991. Volcanic and tectonic controls of
lacustrine and alluvial sedimentation in the Stephanian coal-
bearing sequence of the MalpasSort Basin, Catalonian Pyrenees.
Sedimentology 38, 326.
Beynon, B.M., Pemberton, S.G., 1992. Ichnological signature of a
brackish water deposit: an example from the Lower Cretaceous
Grand Rapids Formation, Cold Lake Oil Sands area, Alberta. In:
Pemberton, S.G. (Ed.), Applications of Ichnology to Petroleum
Exploration; A Core Workshop. Soc. Econ. Paleont. Miner. Core
Workshop, vol. 17, pp. 199222.
Birgenheier, L.T., Fielding, C.R., Frank, T.D., Rygel, M.C., Roberts,
J., 2006. Lithologic consequences of glacially influenced deposi-
tion: a case study from the Late Paleozoic of New South Wales,
Australia. AAPG Annual Convention Abstracts, p. 10.
Blum, M.D., Trnqvist, T.E., 2000. Fluvial response to climate and
sea-level change: a review and look forward. Sedimentology 47,
248.
Bohacs, K., Suter, J., 1997. Sequence stratigraphic distribution of
coaly rocks: fundamental controls and paralic examples. AAPG
Bull. 81, 16121639.
Boothroyd, J.C., Ashley, G.M., 1975. Process, bar morphology, and
sedimentary structures on braided outwash fans, northeastern Gulf
of Alaska. In: Jopling, A.V., McDonalds, B.C. (Eds.), Glaciofluvial
and Glaciolacustrine Sedimentation. Spec. Publ. Soc. Econ.
Paleont. Miner., vol. 23, pp. 193222.
Bridge, J.S., 1985. Paleochannel patterns inferred from alluvial
deposits: a critical evaluation. J. Sediment. Petrol. 55, 579589.
Bridge, J.S., 1997. Thickness of sets of cross strata and planar strata as
a function of formative bed-wave geometry and migration, and
aggradation rate. Geology 25, 971974.
Bridge, J.S., 2003. Rivers and Floodplains: Forms, Processes, and the
Sedimentary Record. Blackwell, Oxford. 491 pp.
Bridge, J.S., Tye, R.S., 2000. Interpreting the dimensions of ancient
fluvial channel bars, channels, and channel belts from wireline-
logs and cores. AAPG Bull. 84, 12051228.
Bristow, C.S., Best, J.L., 1993. Braided rivers: perspectives and
problems. In: Best, J.L., Bristow, C.S. (Eds.), Braided Rivers.
Geol. Soc. Spec. Publ., vol. 75, pp. 291304.
Burst, J.F., 1965. Subaqueously formed shrinkage cracks in clay.
J. Sediment. Petrol. 35, 348353.
Calder, J.H., Gibling, M.R., 1994. The Euramerican Coal Province:
controls on Late Paleozoic peat accumulation. Palaeogeogr.
Palaeoclim. Palaeoecol. 106, 121.
Collinson, J.D., 1996. Alluvial sediments. In: Reading, H.G. (Ed.),
Sedimentary Environments: Processes, Facies, and Stratigraphy.
Blackwell Science, Oxford, pp. 3783.
De Caritat, P., Braun, J., 1992. Cyclic development of sedimentary
basins at convergent plate margins: structural and tectono-thermal
evolution of some Gondwana basins of eastern Australia.
J. Geodyn. 16, 241282.
Ekdale, A.A., Bromley, R.G., Pemberton, S.G., 1984. Ichnology; the
use of trace fossils and stratigraphy. Soc. Econ. Paleont. Miner.
Short Course, vol. 15. 317 pp.
Evans, P.R., 1967. Upper Carboniferous and Permian palynological
stages and their distribution in eastern Australia. Record 1967/99.
Bur. Min. Res. Geol. Geophys., Aust.
Evans, P.R., 1980. Geology of the Galilee Basin. In: Henerson, R.A.,
Stephenson, P.J. (Eds.), The Geology and Geophysics of North-
eastern Australia. Geolog. Soc. Aust., pp. 299305.
Farrell, K.M., 1987. Sedimentology and facies architecture of
overbank deposits of the Mississippi River, False River region,
Louisiana. In: Ethridge, F.G., Flores, R.M., Harvey, M.D. (Eds.),
Recent Developments in Fluvial Sedimentology. Spec. Publ. Soc.
Econ. Paleont. Miner., vol. 39, pp. 111120.
Fielding, C.R., 1987. Coal depositional models for deltaic and alluvial
plain sequences. Geology 15, 661664.
Fielding, C.R., Webb, J.A., 1996. Facies and cyclicity of the Late Permian
Bainmedart Coal Measures in the Northern Prince Charles Mountains,
MacRobertson Land, Antarctica. Sedimentology 43, 295322.
Fielding, C.R., Falkner, A.J., Scott, S.G., 1993. Fluvial response to foreland
basin overfilling: the Late Permian Rangal Coal Measures in the
Bowen Basin, Queensland, Australia. Sediment. Geol. 85, 475497.
Fielding, C.R., Naish, T.R., Woolfe, K.J., Lavelle, M.A., 2000. Facies
analysis and stratigraphy of CRP-2/2A, Victoria Land Basin,
Antarctica. Terra Antarctica 7, 323338.
Fielding, C.R., Sliwa, R., Holcombe, R.J., Jones, A.T., 2001. A new
paleogeographic synthesis for the Bowen, Gunneday, and Sydney
basins of Eastern Australia. In: Hill, K.C., Bernecker, T. (Eds.),
Eastern Australasian Basins Symposium. Spec. Publ. Petrol.
Explor. Soc. Austr. AustIMM, Melbourne, pp. 269279.
Fielding, C.R., Bann, K.L., MacEachern, J.A., Frank, T., Rygel, M.C.,
2006. Stratigraphic stacking patterns through a protracted Icehouse
interval disclose alternating glacial and non-glacial periods: the
Permian of Eastern Australia. AAPG Annual Convention
Abstracts, p. 32.
Gelfenbaum, G., 1983. Suspended sediment response to semidiurnal
and fortnightly tidal variations in a mesotidal estuary: Columbia
River, U.S.A. Mar. Geol. 52, 3957.
Gibbs, M.T., Rees, P.A., Kutzbach, J.E., Ziegler, A.M., Behling, P.J.,
Rowley, D.B., 2002. Simulations of Permian climate and
comparisons with climate-sensitive sediments. J. Geol. 110, 3355.
Gibling, M.R., Bird, D.J., 1994. Late Carboniferous cyclothems and
alluvial paleovalleys in the Sydney Basin, Nova Scotia. Geol. Soc.
Amer. Bull. 106, 105117.
Hawkins, P.J., 1976. Facies analysis and economic significance of the
Late Permian strata in the northern Galilee Basin: Queensland
Govt. Min. J. 77, 1532.
Hawkins, P.J., 1978. Galilee Basinreview of petroleum prospects.
Queensl. Gov. Min. J. 79, 96112.
Hawkins, P.J., Green, P.M., 1993. Exploration results, hydrocarbon
potential, and future strategies for the northern Galilee Basin.
APEA J. 33, 280296.
 J.P. Allen, C.R. Fielding / Sedimentary Geology 202 (2007) 534
Heller, P.L., Paola, C., 1996. Downstream changes in alluvial
architecture: an exploration of controls on channel stacking
patterns. J. Sediment. Res. 66, 297306.
Howard, J.D., Frey, R.W., 1973. Characteristic physical and biogenic
sedimentary structures in Georgia estuaries. AAPG Bull. 57,
11591184.
Hjellbakk, A., 1997. Facies and fluvial architecture of a high-energy
braided river; the upper Proterozoic Seglodden Member,
Varanger Peninsula, Northern Norway. Sediment. Geol. 144,
131161.
Jones, L.S., Schumm, S.A., 1999. Causes of avulsion: an overview. In:
Smith, N.D., Roger, J. (Eds.), Fluvial sedimentology VI. Spec.
Publ. Intern. Assoc. Sediment., vol. 28, pp. 171178.
Jorgensen, P.J., Fielding, C.R., 1996. Facies architecture of alluvial
floodbasin deposits; three-dimensional data from the Upper
Triassic Callide Coal Measures of east-central Queensland,
Australia. Sedimentology 43, 479495.
Jorgensen, P.J., Fielding, C.R., 1999. Debris-flow deposits in an
alluvial-plain succession: the Upper Triassic Callide Coal Measures
of Queensland, Australia. J. Sediment. Res. 69, 10271040.
Karpetta, W.P., 1993. Sedimentology and gravel bar morphology in an
Archaean braided river sequence: the Witpan Conglomerate
Member (Witwatersrand Supergroup) in the Welkom Goldfield,
South Africa. In: Best, J.L., Bristow, C.S (Eds.), Braided Rivers.
Geol. Soc. Spec. Publ., vol. 75, pp. 369388.
Kataoka, K., Nakajo, T., 2002. Volcaniclastic resedimentation in distal
fluvial basins induced by large-volume explosive volcanism: the
Ebisutoge-Fukuda tephr, Plio-Pleistocene boundary, central Japan.
Sedimentology 49, 319334.
Kidwell, S.M., 1997. Anatomy of extremely thin marine sequences
landward of a passive-margin hinge zone: Neogene Calvert Cliffs
succession, Marlyand, USA. J. Sediment. Res. 67, 322340.
Koss, J.E., Ethridge, F.G., Schumm, S.A., 1994. An experimental
study of the effects of base-level change on fluvial, coastal plain,
and shelf systems. J. Sediment. Res. 66, 354363.
Lang, S.C., Grech, P., Root, R., Hill, A., Harrison, D., 2001. The
application of sequence stratigraphy to exploration and reservoir
development in the CooperEromangaBowenSurat basin
system. APPEA J. 41, 223250.
Leclair, S.F., Bridge, J.S., 2001. Quantitative interpretation of
sedimentary structures formed by river dunes. J. Sediment. Res.
71, 713716.
Leclair, S.F., Bridge, J.S., Wang, F., 1997. Preservation of cross-strata
due to the migration of subaqueous dunes over aggrading and non-
aggrading beds: comparison of experimental data with theory.
Geosci. Can. 24, 5566.
Leeder, M.R., 1987. Sediment deformation structures and the
palaeotectonic analysis of sedimentary basins, with a case-study
from the Carboniferous of northern England. In: Jones, M.E.,
Preston, R.M.F. (Eds.), Deformation of Sediments and Sedimen-
tary Rocks. Geol. Soc. Spec. Publ., vol. 29, pp. 137146.
Lorenz, J.C., Nadon, G.C., 2002. Braided-river deposits in a muddy
depositional setting: the Molina Member of the Wasatch Formation
(Paleogene), west-central Colorado, USA. J. Sediment. Res. 72,
376385.
Mack, G.H., James, W.C., 1993. Control of basin symmetry on fluvial
lithofacies, Camp Rice and Palomas formations (Plio-Pleistocene),
southern Rio Grande Rift, USA. In: Marzo, M., Puigdefabregas, C.
(Eds.), Alluvial Sedimentation. Spec. Publ. Intern. Assoc. Sedi-
ment., vol. 17, pp. 439449.
Mack, G.H., Seager, W.R., 1990. Tectonic control on facies
distribution of the Camp Rice and Palomas formations (Plio-
33
Pleistocene), southern Rio Grande Rift, USA. Geol. Soc. Amer.
Bull. 102, 4553.
Mackey, S.D., Bridge, J.S., 1995. Three dimensional model of alluvial
stratigraphy: theory and application. J. Sediment. Res. 65, 731.
Major, J.J., 1997. Depositional processes in large-scale debris flow
experiments. J. Geol. 105, 345366.
Martinsen, O.J., Ryseth, A., Helland-Hansen, W., Flesche, H.,
Torkildsen, G., Idil, S., 1999. Stratigraphic base level and fluvial
architecture: Ericson Sandstone (Campanian), Rock Springs
Uplift, SW Wyoming, U.S.A. Sedimentology 46, 235259.
Massari, F., Mellere, D., Doglion, C., 1993. Cyclicity in non-marine
foreland-basin sedimentary fill: the Messinian conglomerate-
bearing succession of the Venetian Alps (Italy). In: Marzo, M.,
Puigdefabregas, C. (Eds.), Alluvial Sedimentation. Spec. Publ.
Intern. Assoc. Sediment., vol. 17, pp. 501520.
McCabe, P.J., 1984. Depositional environments of coal and coal-
bearing strata. In: Rahmani, R.A., Flores, R.M. (Eds.), Sedimen-
tology of Coal-bearing Sequences. Spec. Publ. Intern. Assoc.
Sediment., vol. 7, pp. 1342.
Miall, A.D., 1977. A review of the braided river depositional
environment. Earth-Sci. Rev. 13, 162.
Miall, A.D., 1985. Architectural-element analysis: a new method of facies
analysis applied to fluvial deposits. Earth-Sci. Rev. 22, 261308.
Miall, A.D., 1988. Architectural elements and bounding surfaces in
fluvial deposits: anatomy of the Kayenta Formation (Lower
Jurassic), southwest Colorado. Sediment. Geol. 55, 233262.
Miall, A.D., 1996. The Geology of Fluvial Deposits: Sedimentary
Facies, Basin Analysis, and Petroleum Geology. Springer-Verlag,
Berlin. 586 pp.
Michaelson, P., Henderson, R.A., Crosdale, P.J., Mikkelsen, S.O.,
2000. Facies architecture and depositional dynamics of the Upper
Permian Rangal coal measures, Bowen Basin, Australia.
J. Sediment. Res. 70, 879895.
Middleton, M.F., Hunt, J.W., 1989. Influence of tectonics on Permian
coal-rank patterns in Australia. Int. J. Coal Geol. 13, 391411.
Moncrieff, A.C.M., 1989. Classification of poorly sorted sedimentary
rocks. Sediment. Geol. 65, 191194.
Mothes, P.A., Hall, M.L., Janda, R.J., 1998. The enormous Chillos
Valley lahar: an ash-flow-generated debris flow from Cotopaxi
volcano, Equador. Bull. Volcanol. 59, 233244.
Mulder, T., Alexander, J., 2001. The physical character of subaqueous
sedimentary density flows and their deposits. Sedimentology 48,
269299.
Naish, T.R., Kamp, P.J.J., 1997. Sequence stratigraphy of sixth-order
(41 k.y.) PliocenePleistocene cyclothems, Wanganui Basin, New
Zealand: a case for the regressive systems tract. Geol. Soc. Amer.
Bull. 109, 978999.
Nakayama, K., Yoshikawa, S., 1997. Depositional processes of
primary to reworked volcaniclastics on an alluvial plain; an
example from the Lower Pliocene Ohta tephra bed of the Tokai
Group, central Japan. Sediment. Geol. 107, 211229.
Nemec, W., Steel, R.J., 1984. Alluvial and coastal conglomerates; their
significant features and some comments on gravelly mass-flow
deposits. In: Koster, E.H. (Ed.), Sedimentology of Gravels and
Conglomerates. Memoir Can. Soc. Petrol. Geol., vol. 10, pp. 131.
Olsen, T., Steel, R., Hgseth, K., Skar, T., Re, S.L., 1995. Sequential
architecture in a fluvial succession: sequence stratigraphy in the
Upper Cretaceous Mesa Verde Group, Price Canyon, Utah.
J. Sediment. Res., Sect. B Stratigr. Glob. Stud. 65, 265280.
Owen, G., 1996. Experimental soft-sediment deformation; structures
formed by the liquefaction of unconsolidated sands and some
ancient examples. Sedimentology 43, 279293.
34 J.P. Allen, C.R. Fielding / Sedimentary Geology 202 (2007) 534
Pashin, J.C., Gastaldo, R.A. (Eds.), 2004. Sequence Stratigraphy,
Paleoclimate, and Tectonics of Coal-Bearing Strata. AAPG Stud.
Geol., vol. 51. 238 pp.
Pemberton, S.G., MacEachern, J.A., Frey, R.W., 1992. Trace fossil
facies models: environmental and allostratigraphic significance. In:
Walker, R.G., James, N.P. (Eds.), Facies Models: Response to Sea
Level Change. Geol. Assoc. Can., pp. 4772.
Pierson, T.C., Scott, K.M., 1985. Downstream dilution of a lahar;
transition from debris flow to hyperconcentrated streamflow. Water
Resour. Res. 21, 15111524.
Plummer, P.S., Gostin, V.A., 1981. Shrinkage cracks: desiccation or
synaeresis. J. Sediment. Petrol. 51, 11471156.
Posamentier, H.W., Allen, G.P., 1999. Siliciclastic Sequence Strati-
graphyConcepts and Applications. Soc. Econ. Paleont. Miner.
Conc. Sediment. Paleont. 7, 210 pp.
Ramos, A., Sopea, A., Perez-Arlucea, M., 1986. Evolution of
Buntsandstein fluvial sedimentation in the northwest Iberian
Ranges (central Spain). J. Sediment. Petrol. 56, 862875.
Ray, S., Chakraborty, T., 2002. Lower Gondwana fluvial succession of
the PenchKanhan valley, India: stratigraphic architecture and
depositional controls. Sediment. Geol. 151, 243271.
Reinfelds, I., Nanson, G., 1993. Formation of braided river flood-
plains, Waimakariri, New Zealand. Sedimentology 40, 11131127.
Re, S.-L., 1987. Cross-strata and bedforms of probable dune to upper-
stage plane-bed origin from a Late Precambrian fluvial sandstone,
northern Norway. Sedimentology 34, 89101.
Re, S.-L., Hermansen, M., 1993. Processes and products of large,
Late Precambrian sand rivers in northern Norway. In: Marzo, M.,
Puigdefabregas, C. (Eds.), Alluvial Sedimentation. Spec. Publ.
Intern. Assoc. Sediment., vol. 17, pp. 151166.
Re, S.-L., Hermansen, M., 2006. New aspects of deformed cross-
strata in fluvial sandstones: examples from Neoproterozoic
formations in northern Norway. Sediment. Geol. 186, 283293.
Rust, B.R., 1978. Depositional models for braided alluvium. In: Miall,
A.D. (Ed.), Fluvial Sedimentology. Memoir Can. Soc. Petrol.
Geol., vol. 5, pp. 605625.
Rust, B.R., 1984. Proximal braidplain deposits in the Middle Devonian
Malbaie Formation of eastern Gasp, Quebec, Canada. Sedimen-
tology 31, 675695.
Saunderson, H.C., Lockett, F.P.J., 1983. Flume experiments on
bedforms and structures at the dune-plane bed transition. In:
Collinson, J.D., Lewin, J. (Eds.), Modern and Ancient Fluvial
Systems. Spec. Publ. Intern. Assoc. Sediment., vol. 6, pp. 4958.
Schumm, S.A., 1993. River response to baselevel change: implications
for sequence stratigraphy. J. Geol. 101, 279294.
Scott, S.G., Hawkins, P.J., 1992. Coal geology of the northern Galilee
Basin and its implications for coalseam methane investigations. In:
Beamish, B.B., Gamson, P.D. (Eds.), Symposium on Coalbed
Methane Research and Development in Australia, vol. 5, pp. 85102.
Shanley, K.W., McCabe, P.J., 1991. Predicting facies architecture
through sequence stratigraphy an example from the Kaipar-
owits Plateau, Utah. Geology 19, 742745.
Shanley, K.W., McCabe, P.J., 1993. Alluvial architecture in a sequence
stratigraphic framework: a case history from the Upper Cretaceous of
southern Utah, USA. In: Flint, S.S., Bryant, I.D. (Eds.), The
Geological Modeling of Hydrocarbon Reservoirs and Outcrop
Analogues. Spec. Publ. Intern. Assoc. Sediment., vol. 15, pp. 2156.
Shanley, K.W., McCabe, P.J., 1994. Perspectives on the sequence
stratigraphy of continental strata. AAPG Bull. 78, 544568.
Simons, D.B., Richardson, E.V., Nordin Jr., C.L., 1965. Sedimentary
structures generated by flow in alluvial channels. In: Middleton,
G.V. (Ed.), Primary Sedimentary Structures and their Hydro-
dynamic Interpretation. Spec. Publ. Soc. Econ. Paleont. Miner.,
vol. 12, pp. 84115.
Smith, N.D., 1974. Sedimentology and bar formation in the upper
Kicking Horse River, a braided outwash stream. J. Geol. 81,
205223.
Smith, G.A., 1986. Coarse-grained nonmarine volcaniclastic sediment:
terminology and depositional process. Geol. Soc. Amer. Bull. 97,
110.
Smith, S.A., 1990. The sedimentology and accretionary styles of an
ancient gravel-bed stream: the Budleigh Salterton Pebble Beds
(Lower Triassic), southwest England. Sediment. Geol. 67,
199219.
Smith, G.A., Lowe, D.R., 1991. Lahars: volcano-hydrologic events
and deposition in the debris flow-hyperconcentrated flow
continuum. In: Fisher, R.V., Smith, G.A. (Eds.), Sedimentation
in Volcanic Settings. Spec. Publ. Soc. Econ. Paleont. Miner.,
vol. 45, pp. 5970.
Tye, R.S., Coleman, J.M., 1989. Depositional processes and
stratigraphy of fluvially dominated lacustrine deltas; Mississippi
delta plain. J. Sediment. Petrol. 59, 973996.
Van Heeswijck, A., 2004. The structure and hydrocarbon potential of
the northern Drummond Basin and northeastern Galilee Basin,
central Queensland, Australia. Eastern Australasian Basins
Symposium II. Petrol. Explor. Soc. Austr. Adelaide, pp. 319330.
Van Wagoner, J.C., Posamentier, H.W., Mitchum, R.M., Vail, P.R.,
Sarg, J.F., Loutit, T.S., Hardenbol, J., 1988. An overview of the
fundamentals of sequence stratigraphy and key definitions. In:
Wilgus, C.K., Hastings, B.S., Kendall, C.G.St.C., Posamentier,
H.W., Ross, C.A., Van Wagoner, J.C. (Eds.), Sea Level Changes
An Integrated Approach. Spec. Publ. Soc. Econ. Paleont. Miner.,
vol. 42, pp. 3539.
Van Wagoner, J.C., Mitchum, R.M., Campion, K.M., Rahmanian, V.D.,
1990. Siliciclastic Sequence Stratigraphy in Well Logs, Cores, and
Outcrops: Concepts for High-Resolution Correlation of Time and
Facies. AAPG Meth. Explor. Ser. vol. 7, 63 pp.
Veevers, J.J., Jones, J.G., Powell, C.M., 1982. Tectonic framework of
Australia's sedimentary basins. APEA J. 22, 283300.
Vine, R.R., 1976. Galilee Basin. In: Leslie, R.B., Evans, H.J., Knight,
C.L. (Eds.), Economic Geology of Australia and Papua New
Guinea, Australas. Inst. Min. Metal. AIMM Monograph Series,
vol. 7, pp. 316321.
Vine, R.R., Paine, A.G.L., 1974. Hughenden Queensland, 1:250,000
Geological Series Sheet SG/55-1, Explanatory Notes. Bur. Min.
Res. Geol. Geophys. Aust. 39 pp.
Walker, R.G., Cant, D.J., 1984. Sandy fluvial systems, In: Walker, R.G.
(Ed.), Facies Models, 2nd edition. Geosci. Can. Reprint Series,
vol. 1, pp. 7189.
White, M.E., in press. 1963 plant fossil collections from Hughenden
area, Great Artesian Basin. Record 1964/64 Bur. Min. Res. Geol.
Geophys. Aust.
Williams, P.F., Rust, B.R., 1969. The sedimentology of a braided river.
J. Sediment. Petrol. 39, 649679.
Willis, B.J., 1993. Ancient river systems in the Himalayan foredeep,
Chinji Village area, northern Pakistan. Sediment. Geol. 88, 176.
Wood, L.J., Ethridge, F.G., Schumm, S.A., 1991. Influence of
subaqueous shelf angle on coastal plain-shelf slope systems: an
experimental approach. AAPG Bull. 74, 1349.
Yoshida, S., Willis, A., Miall, A.D., 1996. Tectonic control of nested
sequence architecture in the Castlegate Sandstone (Upper Cretac-
eous), Book Cliffs, Utah. J. Sediment. Res. 66, 737748.
 Available online at www.sciencedirect.com

Sedimentary Geology 202 (2007) 35 57

www.elsevier.com/locate/sedgeo

Possible climate variability in the Lower Old Red Sandstone Conigar

Pit Sandstone Member (early Devonian), South Wales, UK
Robert D. Hillier a,, Susan B. Marriott b , Brian P.J. Williams c , V. Paul Wright d
a
Department of Geology, National Museum and Galleries of Wales, Cardiff, CF10 3NP, UK
b
School of Geography and Environmental Management, University of the West of England, Bristol, BS16 1QY, UK
c
Department of Geology & Petroleum Geology, University of Aberdeen, Aberdeen, AB24 3UE, UK
d
School of Earth, Ocean and Planetary Sciences, University of Wales, Cardiff, CF10 3YE, UK

Abstract

The Early Devonian Conigar Pit Sandstone Member (CPSM) of the Freshwater West Formation (Lower Old Red Sandstone,
Lochkovian age) at Freshwater West in south Wales comprises a heterolithic, predominantly alluvial suite (mudstones, fine to coarse-
grained sandstones, conglomerates) with varying depositional architecture. A number of discrete facies associations are observed. The
fine-grained, inclined- and non-inclined heterolithic association is dominant within the CPSM. It represents dryland sinuous channelized
flow (IHS bedsets) and unconfined flow (non-inclined bedsets) at terminal and intermediate floodouts deposited under semi-arid
conditions. Mudstones were deposited in two distinct environments associated with this semi-arid river system. Laminated and
burrowed, reddish brown mudstones were deposited in shallow semi-permanent lakes or pools on the floodplain. Massive mudstones
were deposited as within-channel muddy braid-bars, with sedimentary structures being lost during subsequent compaction. The
mudstones exhibit Vertisol pedogenesis reflecting the seasonal wetting and drying associated with the semi-arid climate. Periods of
intense desiccation are indicated by deep desiccation cracks and associated rubbly surfaces. Possible, though unproven tidal conditions
influenced a small percentage of the heterolithic channels at Freshwater West. The extent of these channels is uncertain.
The coarser-grained multi-storey sandstone association was deposited by low-sinuosity rivers with a fluctuating, but perennial
discharge. Associated with these laterally extensive sandbodies was a high water table with surface ponding (wetlands) that enabled the
preservation of plant detritus. During high discharge events, flow expanded over the contemporaneous floodplain depositing the sheet
sandstone association within splay complexes. The interpretation that perennial discharge may have been triggered by basin-wide
climate change challenges the long-held view that the Lower ORS climate was continuously semi-arid in nature. The spatial/temporal
extent of this climate change is uncertain, as it is possible that the multi-storey sandbodies were deposited by allogenic rivers draining a
distant, possibly uplifted source area. A decrease in the observed frequency and maturity of Vertisol profiles, and a corresponding
increase in multi-storey sandbodies upward through the succession may reflect a long-term, episodic pattern of climate change.
2007 Elsevier B.V. All rights reserved.

Keywords: Devonian; Fluvial; Dryland; Climate change; Anglo-Welsh Basin; Old Red Sandstone

1. Introduction

The Old Red Sandstone (ORS) comprises Upper

Corresponding author.
E-mail address: robhillier@aol.com (R.D. Hillier).
0037-0738/$ - see front matter 2007 Elsevier B.V. All rights reserved.
doi:10.1016/j.sedgeo.2007.05.006
Silurian to Lower Carboniferous, mainly continental
deposits bordering the present North Atlantic (Friend
et al., 2000; Barclay et al., 2005). The ORS represents
 36
R.D. Hillier et al. / Sedimentary Geology 202 (2007) 3557
Fig. 1. A). Outcrop map of Lower ORS Anglo-Welsh Basin in south Wales and the Welsh Borderland. RF Ritec Fault, BF Benton Fault, LlWBF LlandyfaelogWelsh Borderland Fault
System. B). ORS outcrop map of southern Pembrokeshire (based on Allen et al., 1982; Williams et al., 1982; Hillier and Williams, 2007).
 R.D. Hillier et al. / Sedimentary Geology 202 (2007) 3557 37

the syn- to post-orogenic depositional response (molasse)
to the Caledonian Orogeny, being modified by synchro-
nous tectonism and volcanicity. The southernmost expo-
sures of the ORS in the UK are found within the Anglo-
Welsh Basin, cropping out over an area of approximately
20,000 km2 in southern Wales and the Welsh Borderland
(Fig. 1). Studies of the Lower ORS (Ludlow to Emsian,
Fig. 2) have generated significant sedimentological
advances, for example the now classic analysis of high-
sinuosity fluvial channels (Allen, 1964, 1965, 1970,
1974a), and calcic Vertisols within dryland environments
(Allen, 1973a, 1974b, 1986). Early studies of the Lower
ORS Anglo-Welsh Basin-fill concentrated on the deposi-
tional architecture of alluvial sandstone bodies (Allen,
1970; Allen and Williams, 1978; Tunbridge, 1981; Allen
and Williams, 1982; Allen et al., 1982; Williams et al.,
1982; Allen, 1983). Parallel work detailed interfluvial
depositional patterns (Allen and Williams, 1979) and
corresponding geomorphic and pedogenic processes
(Marriott and Wright, 1993, 1996; Wright and Marriott,
1996). Subsequent research has linked both channel and
floodplain processes and their spatial/temporal relation-
ships, characterizing a dynamic mudstone-dominated,
dryland deposystem (Owen, 1995; Owen and Hawley,
2000; Love and Williams, 2000; Marriott and Wright,
2004; Williams and Hillier, 2004; Marriott et al., 2005).
In this paper, we describe and interpret the Conigar
Pit Sandstone Member (CPSM) of the Freshwater West

Fig. 2. Stratigraphic Column of the Lower ORS Anglo-Welsh Basin. Key: TTB Townsend Tuff Bed; CPC/BFL Chapel Point Calcretes and
Bishop Frome Limestone Members. LFC Llyn-y-fan Conglomerate; LCF Llanishen Conglomerate Formation. Modified after Williams et al.
(2004), Hillier and Williams (2007).
38 R.D. Hillier et al. / Sedimentary Geology 202 (2007) 3557

Formation at Freshwater West, southwest Wales
(Fig. 2). The CPSM is a heterolithic suite characterized
by a variety of sand- and mudstone-dominated intervals
with varying architecture. We speculate as to the origin
(s) of the facies associations observed, and comment on
controlling mechanisms.
2. Palaeogeography of the Lower Old Red Sandstone
of the Anglo-Welsh Basin
The Anglo-Welsh Basin developed externally to the
Caledonian mountain front on the southern margin of
the newly amalgamated Lower Palaeozoic Laurussian
continent (Friend et al., 2000). It is dominated by
continental fluvial systems that transported sediment
from the Caledonian mountains to the contemporaneous
coastline of the Rheic Ocean to the south and east during
its evolution (Simon and Bluck, 1982).
The northern limit of the main Lower ORS outcrop is
the BentonLlandyfaelogWelsh Borderland fault sys-
tem (Fig. 1), but it is likely that the basin had a greater
aerial extent connecting the Anglesey, Clun Hills and
Long Mountain ORS outliers (Soper and Woodcock,
2003; Hillier and Williams, 2006). The basin was not
significantly deformed until Emsian/Eifelian times
during the Acadian Orogeny (Friend et al., 2000;
Dewey and Strachan, 2003; Soper and Woodcock,
2003). Palaeomagnetic data from the Lower ORS of
south Wales indicates a palaeogeographic location
within tropical latitudes (approximately 17 5S, Chan-
nel et al., 1992). Palaeoclimate interpretations have
mainly come from studies of palaeosols preserved
within the ORS. Analysis of calcic Vertisols in the
Anglo-Welsh Basin led Allen (1974b, 1985) to suggest
that the Lower ORS accumulated under warm to hot,
semi-arid conditions with seasonal rainfall. Studies of
depositional styles in fine-grained deposits indicate that
the fluvial systems had similar characteristics to modern
ephemeral dryland rivers (e.g. Marriott and Wright, 2004;
Marriott et al., 2005). The Anglo-Welsh Basin has

Fig. 3. Restored fence diagram illustrating isopach variations within the Lower ORS of southern Pembrokeshire. Section 1, north of Milford Haven,
2 Pembroke Peninsula, 3 Freshwater West, 4 Freshwater East to Manorbier. Sections not drawn to scale, thicknesses in brackets are estimated from
incomplete or structurally complex sections. Modified from Williams and Hillier (2004).
 R.D. Hillier et al. / Sedimentary Geology 202 (2007) 3557
provided important information on the evolution of early
land plant assemblages (Edwards and Richardson, 2004).
As the preserved plant material is allochthonous, habitats
remain uncertain, though vascular plants are believed to
have colonized areas close to perennial water stores, with
interfluvial areas possibly providing refugia for cryptos-
pore-producing plants (Edwards and Richardson, 2004).
The absence of a dense vegetation cover would have led to
extremes of sediment yield following precipitation,
causing widespread flash flooding.
3. Lower ORS stratigraphy
The Lower ORS stratigraphy to the south of the Ritec
Fault (Fig. 1) was first described by Dixon (1921), who
assigned most of the succession to the Red Marls.
These were renamed the Milford Haven Group by Allen
and Williams (1978), and divided into three lithostrati-
graphic formations, in ascending order the Freshwater
Fig. 4. Geological map of Freshwater West (modified after Barclay et al., 2005, reproduced with permission from the Joint Nature Conservation
Committee).
39
East, the Moor Cliffs and the Freshwater West
Formations. The formations were subdivided mainly
on sedimentological criteria, rather than on faunal or
floral remains. Isopach variation within the Lower ORS
strongly suggest SW Pembrokeshire was influenced by
extensional tectonics (Fig. 3). It is likely that the Benton
Fault exerted considerable control on basin geomor-
phology and drainage evolution (Williams and Hillier,
2004), with maximum accommodation space being
developed in the hangingwall of the fault. The Lower
ORS thins eastward, indicating a decreasing throw
along the Benton Fault in this direction (Fig. 3).
The top of the Moor Cliffs Formation is defined by the
top of the Chapel Point Calcretes Member, a sequence of
calcic Vertisols of varying maturities reflecting a period of
extended hiatus and aridity across the Anglo-Welsh Basin
(Allen, 1974b, 1985). A change in depositional style,
textural grade and provenance mark the onset of
sedimentation immediately after formation of the Chapel
40 R.D. Hillier et al. / Sedimentary Geology 202 (2007) 3557

Fig. 5. Graphic log measured through the CPSM at Freshwater West. The log starts immediately above the Chapel Point Calcretes Member of the
Moor Cliffs Formation. For key, see Fig. 9.
 R.D. Hillier et al. / Sedimentary Geology 202 (2007) 3557 41

Fig. 5 (continued ).

Point Calcretes Member. A decline in the abundance of zircon, tourmaline and rutile in the CPSM, implies either
garnet, mica and metamorphic grains and a change to progressive unroofing of a new provenance, or significant
sedimentary and igneous grains, with an increase in modification of the drainage pattern (Allen and Crowley,
42 R.D. Hillier et al. / Sedimentary Geology 202 (2007) 3557
1983; Allen, 1985). It is possible that parts of the former
Lower Palaeozoic Welsh Basin and Irish Sea areas were
being inverted at this time, with palaeocurrents from
fluvial channels flowing consistently from the north and
northwest (Fig. 5; Allen, 1985).
The Freshwater West Formation is subdivided into
the lower Conigar Pit Sandstone Member (CPSM, the
subject of this paper), and the overlying Rat Island
Mudstone Member (Williams et al., 1982). A number of
alluvial facies associations have been detailed within the
CPSM (Williams et al., 1982; Williams and Hillier,
2004), and these will be discussed in turn below. Thin,
grey mudstones have yielded an assemblage of
moderately abundant sporomorphs (cryptospores and
trilete spores) and phytodebris assigned to the middle
part of the Emphanisporites micrornatusStreelispora
newportensis (MN) Spore Zone of Early Devonian
(mid-Lochkovian) age (Higgs, 2004). The overlying Rat
Island Mudstone Member is dominated by red mud-
stones with thin, sheet-like sandstones.
The Freshwater West Formation contains a rich,
diverse terrestrial ichnofauna dominated by Beaco-
nites barretti (Morrissey and Braddy, 2004). Associ-
ated trackways reflect a gregarious arthropod epifauna
of arachnids (Paleohelcura), and myriapods (Diplich-
nites and Diplopodichnus). Additional ichnotaxa
comprise arthropod foraging and resting traces,
worm burrows and faecal pellets characteristic of
the Scoyenia ichnofacies (Morrissey and Braddy,
2004).
4. The Conigar Pit Sandstone Member at Freshwater
West
Foreshore and cliff sections at Freshwater West on
the Pembroke Peninsula (Figs. 3 and 4) provide the most
complete, continuously exposed ORS section in the
Anglo-Welsh Basin (Williams et al., 1982; Barclay
et al., 2005). The exposures form the southern limb of
the CastlemartinCorse Anticline, where the CPSM
attains a thickness of 308 m (Fig. 5).
In southern Pembrokeshire, the base of the CPSM is
typically defined by a coarse-grained, extraformational
conglomerate, generally worked into cross-bedded or
planar laminated sheets. These form the channel lags to
upward-fining and -thinning inclined heterolithic strat-
ification (IHS). This association is, however, absent at
Freshwater West, possibly due to its elevated deposi-
tional location high on the Benton Fault hangingwall
(Williams and Hillier, 2004). A number of discrete
facies associations are observed within the CPSM at
Freshwater West.
4.1. Fine-grained inclined and non-inclined heterolithic
association
This association was first described in detail at
Freshwater West by Allen (1974a).
It comprises mainly red heterolithic, interbedded
mudstones and very fine- to fine-grained sandstones
(organized into planar or low-angle, upward-fining
bedset cycles up to 3 m thick (Fig. 6A, B). Bedset
bases generally overlie planar to undulatory erosion
surfaces and are defined by cm- to dm-thick lags of
intraformational rip-up clasts and/or reworked calcrete
nodules. Sandstone bedsets are invariably cm to dm
thick, dominated by two- and three-dimensional current
ripple cross-lamination commonly arranged in climbing
laminasets (Fig. 6B, C). Some are thin ( 10 cm) and
consist of pinstripe lamination (fine-sand and mudstone
couplets). Beds generally fine upward into overlying
cm- to dm-thick, red mudstones which display desicca-
tion cracks and/or calcrete nodules (Fig. 6B). The
bedsets are commonly inclined to the depositional
horizontal by 10 to 20 (IHS bedsets, Fig. 6A). Internal
reactivation surfaces are common, as are heterolithic,
non-inclined bedsets with similar internal sedimentary
structures (Fig. 6D). Both the inclined and non-inclined
bedsets either fine upwards into the overlying mud-
stones (Fig. 6E), or terminate at well-defined erosion
surfaces. This association is commonly profusely
bioturbated, in particular by large meniscate, back-filled
B. barretti burrows (Morrissey and Braddy, 2004).
Pale to dark grey, IHS bedsets also occur within this
outcrop (e.g. 14.217.9 m, 259.6262.6 m Fig. 5). They
reveal current ripple cross-lamination and commonly
have mud drapes on foresets, with bidirectional flow
being interpreted from closely-spaced laminasets.
Rhythmic, thickening and thinning of laminae are
observed within some bedsets. Framboidal, nodular
pyrite is common, as is the preservation of plant remains
(tea-leaf).
The IHS bedsets are interpreted as the deposits of
sinuous, shallow ephemeral channels (Fig. 7). As the
cross-laminated sandstones are thin, and commonly
bounded by desiccated mudstones, it is probable that
discharge was short-lived. Similar units are interpreted
elsewhere within the ORS as ancient analogues of
deposits found in the present day semi-arid, Channel
Country of central Australia (Marriott and Wright, 2004;
Marriott et al., 2005). There, a low gradient, multi-
channel drainage network comprises sinuous, anasto-
mosing channels that coexist with braided channels
along their courses (Rust and Nanson, 1986; Nanson
et al., 1986; Rust and Nanson, 1989; Gibling et al.,
 R.D. Hillier et al. / Sedimentary Geology 202 (2007) 3557 43

Fig. 6. The inclined and non-inclined heterolithic association. All measurements referenced to Fig. 5 log. (A) Inclined-heterolithic-stratification
(IHS), 126 m, scale 25 cm. (B). Upward-fining and -thinning bedsets, with large desiccation crack. 265.5 m, scale 2 cm. (C). Bedset comprising
climbing current ripple cross-lamination typical of association, 236 m, scale 2 cm. (D). Non-inclined heterolithic bedsets, 242.3 m, scale 25 cm. (E).
Upward-fining/thinning cycle, base of trough cross-bedded sandstone at 263.75 m, scale 25 cm. (F). Laminated fine-grained sandstone/mudstone
couplets interpreted as suspension fallout within ponded water.

1998; Tooth, 2000a,b). At low discharges, flow is
confined to sinuous, trunk-stream channels with depo-
sition on active accretionary benches. During higher
stage discharge, the channels undergo sudden and
substantial changes in planform (Tooth and Nanson,
2000). The channel zone expands, forming a zone of
braided/anastomosing channels that cannibalize sedi-
ment from the underlying floodplain. Much of the fine-
grained sediment is produced by the weathering of
argillaceous rock in the upper catchment reaches
(Maroulis and Nanson, 1996). A similar situation
could be envisaged for the Lower ORS deposystem as
the rocks of the former deep-marine, mud-prone Lower
Palaeozoic basins may have been partly exposed during
44 R.D. Hillier et al. / Sedimentary Geology 202 (2007) 3557

Fig. 7. Schematic block diagram illustrating dryland deposystem and alluvial architecture.

the Early Devonian, and liberated particles would have
aggregated within soils, and continued to develop within
floodplain Vertisols lower in the drainage basin as a
result of frequent wetting and drying cycles (Rust and
Nanson, 1989). Intraformational mud-chip conglomer-
ates and calcrete nodule lags reflect the reworking of
contemporaneous floodplain deposits. The occurrence
of desiccated mudstone layers that can be traced to the
base of the lateral-accretion sets clearly points to the
complete desiccation of the river.
The non-inclined nature of some heterolithic bedsets
may be explained by deposition at intermediate or
terminal floodout zones (Fig. 7; Tooth, 1999, 2000a,b).
The thin bedsets of pinstripe lamination possibly reflect
deposition in ephemeral floodplain ponds (Fig. 6F).
The presence of the grey IHS bedsets suggests
possible periodic influence of tidal conditions within the
CPSM at Freshwater West. The occurrence of pyrite and
the preservation of plant debris allude to anaerobic
groundwater. Although pyrite can develop in anoxic
terrestrial environments, its development is also
favoured in modern tidal marshes where sulphate-
bearing marine water infiltrates sediments and reacts
with ferrous iron in the presence of sulphur-reducing
bacteria to form pyrite. Such sulphidization processes
generate potentially acid sulphate soils observed in tidal
marshes of the Surinam coastline (Augustus and Slager,
1970), the Jutland coastline (Madsen and Jensen, 1988),
the Mekong Delta (Brinkman et al., 1993), and
Australian estuaries (Lin et al., 1995; Smith et al.,
2003; Wilson, 2005).
Possible tidal influence has been speculated elsewhere
in the Anglo-Welsh Basin (Allen, 1973b; Allen and
Williams, 1982; Almond et al., 1993; Barclay et al., 1994;

Fig. 8. Photographs of sandstone-dominated facies associations. All measurements referenced to Fig. 5 log. (A). Sheet sandstone comprising trough
cross-stratification and planar-laminated bedsets. 263.7 m, scale bar 2 cm. (B). Sheet sandstone-dominated by planar lamination 84 m, scale bar
25 cm. (C). Sheet architecture to sandbody, 156 m, scale bar (circled) 25 cm. (D). The Beaconites Bed of Williams et al. (1982) note profuse
development of meniscate, backfilled Beaconites barretti burrows, 156 m, scale bar 2 cm. (E). Stacked trough cross-stratification within multi-storey
sandstone, 254 m, scale bar 25 cm. (F). Downclimbing toesets to trough cross-stratification (highlighted), 305 m, scale bar 2 cm. (G). Laminated fine-
sandstone and organic-rich layers (tea-leaf) with associated pyrite nodules (arrowed). 258 m, scale 2 cm. (H). Tide-influenced, grey/green IHS
bedset, 133 m, scale bar 25 cm.
 R.D. Hillier et al. / Sedimentary Geology 202 (2007) 3557 45

Hillier and Williams, 2004; Williams and Hillier, 2004). the case for tidal influence unproven, pending further
The absence of definitive marine micro- and macro-fossils research. The occurrence however of tidal influence
or their traces within the CPSM at Freshwater West leaves within the CPSM should come as no surprise as marine
46 R.D. Hillier et al. / Sedimentary Geology 202 (2007) 3557
incursion might be expected within a low-gradient
deposystem marginal to the Rheic Ocean.
4.2. Sheet sandstone association
Sheet sandstones are common throughout the Fresh-
water West Formation (Allen and Williams, 1979;
Williams et al., 1982; Williams and Hillier, 2004),
comprising green to grey, cm- to dm-scale sandstones
that are predominantly fine- to medium-grained and
moderately sorted. They differ from the non-inclined
heterolithic association in their increased lateral extent, bed
thickness, preserved set thickness and overall coarser
grain-size. Bed bases are planar to undulatory, and mildly
erosional, commonly strewn with a thin lag of mudstone
rip-up clasts or reworked calcrete nodules. Dm-thick
laminasets are dominated by plane-bed, low-angle plane-
bed and planar and trough cross-stratification within
coarser-grained beds (Fig. 8A, B), and current ripple-
drift lamination within finer-grained units. Sandstone
sheets may contain single or amalgamated laminasets,
the latter commonly separated by thin (cm-thick),
discontinuous intraformational conglomerate lags. Bedset
tops are abrupt and planar, or fine upwards into overlying
siltstones which are invariably pedogenically modified and
desiccated. Beds may be traced across the length of
exposure (10 s of m) with minor thickness variations
(Fig. 8C). An example is the Beaconites Bed of Allen
et al. (1981) at Freshwater West which attains a maximum
observed thickness of 120 cm, thinning to 80 cm over a
distance of 5.5 m. Width/thickness ratios at Freshwater
West exceed 70. Alternatively, some beds infill broad (10 s
of m), shallow (cm to dm), scoured depressions (Williams
and Hillier, 2004). Bioturbation, dominated by backfilled,
meniscate B. barretti burrows, is common (Fig. 8D).
The sheet sandstone association was presumably
sourced from multi-storey sandstone complexes de-
scribed below, the sheets representing unconfined lobes
and high width-to-depth ratio channels developed as
splays off the main channel belts. Such splays are
described in association with ephemeral sand-bed rivers
of central and eastern Australia (Williams, 1970;
O'Brien and Wells, 1986; Tooth, 2005). It is possible
that the CPSM splays prograded into standing bodies of
water within semi-permanent, or seasonal waterholes
or billabongs marginal to the main channel axes. If so,
the splays may be, in part, analogous to the non-inclined
heterolithic stratification, but have been interpreted as a
distinct facies association due to their perceived genetic
association with the multi-storey sandstone complexes
(see below). Thicker cross-stratified bedsets may
represent deposition in the main splay channels, with
bedforms changing to current ripples during flow
deceleration associated with transmission loss or at
splay margins. The high width-to-depth ratio of the
sandstone sheet splays within the CPSM equates to the
stage I splays of Smith et al. (1989).
Few modern examples have been described of splays
from coarse-grained, low-sinuosity rivers. The most
detailed study comes from the perennial Niobrara River
in Nebraska (Bristow et al., 1999). Here, splays are
channelized in their proximal portions, emerging at a high
angle to the main river channel, before turning sub-
parallel to the main channel in their distal reaches. Dunes
in the channelized areas become smaller as the channels
diverge, becoming dominated by current ripples with
occasional mud drapes in distal areas. The largest splay
described covers an area of 200 m by 1000 m, is up to
2.5 m thick, and developed over a 5-year interval. It
contains two to three upward-fining bedsets that represent
discrete flood events or channel switching across the
splay. Bedsets contain planar or trough cross-stratification
that pass upward into ripple and horizontal lamination
reflecting reduction in flow velocity associated with
abandonment of splay channels (Bristow et al., 1999).
4.3. Multi-storey sandstone association
Within the CPSM, the thickest sandbodies comprise
amalgamations of grey/green, fine- to medium-grained
sandstone. These bodies are up to 10 m thick at
Freshwater West (Figs. 5 and 9), and are sheet-like over
the scale of the exposure (10 s of metres). Here
sandstone body bases are mildly erosional (dm-scale
erosion) into underlying sediments. Lags of reworked
calcrete nodules and mudstone rip-up clasts dm-thick
are commonly worked into cross-sets that overlie the
basal erosion surface. The sandstone bedsets are
dominated by trough cross-stratification, with set
thickness typically of dm-scale (Figs. 8E, 9). Individual
storeys (m thickness) are dominated by trough cross-
stratification commonly with well-defined asymptotic
foresets, implying significant suspended-load fallout
during dune migration. Rarely observed are toesets with
thin mud drapes. Storeys often fine-upward into
laminasets of fine-grained sandstone with climbing
current ripple cross-lamination, again signifying high
rates of sedimentation. The laminasets commonly
contain Beaconites. Many laminasets exhibit red colour
mottling that accentuates the ripple lamination and some
contain cm-scale, horizontally-elongated calcrete
nodules over which laminae are draped. Blocks of red,
mottled fine-grained sandstone are common in lags at
the base of the overlying green, coarse-grained storeys.
 R.D. Hillier et al. / Sedimentary Geology 202 (2007) 3557
Fig. 9. Graphic logs through multi-storey sandbody, scale in metres above base of CPSM.

47
48 R.D. Hillier et al. / Sedimentary Geology 202 (2007) 3557
Amalgamation surfaces between the storeys are gener-
ally planar, or define broad, shallow scours (Fig. 8E).
These commonly have a lag of angular intraformational
mudstone clasts, locally of boulder grade. In the topmost
sandbody within the CPSM, low-angle bounding
surfaces separate toesets of downward-climbing cross-
sets (Fig. 8F). In some finer-grained units, sparse,
apparently in situ, cm-scale framboidal pyrite nodules
are present (Fig. 8G). These also occur as reworked
clasts in some of the basal lags.
The cross-bedded strata are likely to have been
deposited as bars in a large, sand-dominated, multi-
channel, low-sinuosity river (cf. descriptions of Miall,
1977; Bridge et al., 1996; Ashworth et al., 2000; Best
et al., 2003; Sambrook Smith et al., 2005; Fig. 10). The
development of such braid-bars is strongly stage-
dependent, with thickness of cross-sets scaling with
flow depth. Observed amalgamation surfaces within the
CPSM multi-storey sandbodies reflect the migration of
lunate dunes across bar surfaces or at their margins. The
trains of climbing current ripples developed at lower
flow stages. Downclimbing dunesets reflect the down-
stream accretion of dunes across sand bars, and down
their lee slopes or lateral margins. Similar mid-channel,
downstream accreting macroforms are described by
Fig. 10. Schematic block diagram of depositional environments and alluvial architectures of the multi-storey sandbodies and associated sheet
sandstone splays.
Allen (1983) from the Brownstones Formation in the
Welsh Borderland.
Although not observed in planform, the multi-storey
units are similar architecturally to models proposed for
sand-dominated braided systems (e.g. Miall, 1977,
1978; Walker and Cant, 1984; Bridge, 1993; Miall,
1996). The lack of associated overbank fines and the
dominance of trough cross-stratification may simply
reflect deposition from a braided channel system
characterized by lateral and downstream-migrating
bars (Fig. 10). However, single channel systems can
also produce multi-storey sheets if accommodation
space development is low, and sediment supply and
avulsion frequency is high (Holbrook, 1996; Makaske,
2001; Adams and Bhattacharya, 2005). Rapid avulsion
of individual channels would cannibalize pre-existing
sediment, and remove any fine-grained deposits that
accumulated within interchannel areas.
It is likely that during low discharge small, ponds
developed within topographic lows on bar surfaces. This
is evident from the preservation of thin, discontinuous
planar-laminated bedsets of fine sandstones and organic-
rich layers, with locally abundant plant debris (tea-
leaf, Fig. 8G). The close association with framboidal
and concretionary pyrite alludes to anoxic conditions
 R.D. Hillier et al. / Sedimentary Geology 202 (2007) 3557
associated with a high water table. Thin, laterally
extensive calcrete nodule horizons within the multi-
storey sandstone bodies may represent groundwater
calcretes deposited at the capillary fringe.
It is probable that the multi-storey sandbodies were
deposited by discharge that was perennial, although
pulsed, conditions common in modern tropical climate
zones with wet-dry seasonality though they lack the
lithofacies heterogeneity associated with short-term
variation in discharge (Miall, 1996). The general
uniformity in the thickness of cross-stratified bedsets
(dm-scale) suggests the rivers did not experience
extremes of discharge variation or sediment transport
rates, although differing stages of flow are demon-
strated by the presence of climbing current ripples
across bar surfaces. Fluvial systems containing
abundant medium- to large-scale bar forms are
predominantly perennial rivers (e.g. the Platte and
William Rivers, Smith, 1970; Blodgett and Stanley,
1980; the South Saskatchewan River, Cant and
Walker, 1978). Indicators of waterlogging (iron
mottling, organic preservation, pyrite nodules) may
reflect the infiltration capacity of the sandbodies, with
baseflow associated with perennial discharge maintain-
ing the water table height. Nowhere within the CPSM at
Freshwater West is there an interdigitation of multi-
storey sandbodies and deposits of the fine-grained
inclined and non-inclined heterolithic association. This
may reflect the probability that these facies associations
did not coexist spatially or temporally within the
drainage basin. The grain-size differences between the
two associations may support this interpretation. The
moderately to well-sorted nature of the sandstones
within the multi-storey units reflects progressive sort-
ing, arguably over a considerable distance.
The lack of preserved channel margins indicates
channel widths greater than that of the exposure
(N70 m). Although multi-storey sandstone bodies are
sheet-like in geometry over the scale of the exposure, it
is impossible to correlate them between sections 1 km
apart (Williams and Hillier, 2004). As discussed above,
it is possible that a genetic relationship with the sheet
sandstone association exists. The multi-storey sandstone
bodies were deposited within the main channel belt,
with the thinner sheet sandstones representing deposi-
tion on interfluves during avulsion events (Fig. 10).
Similar geometries are documented in ancient semi-arid
channel/interfluve deposits (Stear, 1983, 1985). Al-
though not observed at Freshwater West, we speculate
that the splays might emanate at well-defined channel
breaches, possibly having a wing geometry with
respect to the main channel (Nadon, 1994; Kraus, 1996).
49
4.4. Mudstone association
4.4.1. Descriptions of mudstone facies
Mudstone units comprise about 30% of the CPSM at
Freshwater West, occurring as mm-scale laminae within
the sandstone associations or as discrete units up to 4 m
thick revealing evidence of depositional environment
and post-depositional alteration. The mudstone associ-
ation can be divided into several facies, based on
sedimentary structures and other features present:
Laminated facies these are dark, reddish brown
and range in thickness from 50 to 600 mm. They are
most commonly found overlying laminated fine- or very
fine-grained sandstone as part of fining-upward units or
as thin beds within the heterolithic facies association. In
their lower sections, many contain elongate carbonate
concretions up to 20 mm by 40 mm that extend between
laminae and locally distort them. In some units the
laminae are accentuated by carbonate stringers or crystal-
laria (Fig. 11A). The laminated facies is commonly
extensively bioturbated, with vertical burrows up to
5 mm in diameter and up to 120 mm long. The burrows
are generally filled with sediment that is slightly coarser
and more micaceous than the surrounding matrix. In some
units, carbonate concretions formed around the burrows as
prominent tubules (Fig. 11A), up to 120 mm in diameter
and with a central core of reddish brown sediment.
Massive facies these units lack any primary
sedimentary structures. They are dark reddish brown
and have scattered reduction spots and patches. Some
have thin (1030 mm) basal lags of matrix-supported
intraformational, calcrete and rip-up mud clast con-
glomerates and others have patches of reworked floating
calcrete clasts and sparse sand grains. Some have tiny
(12 mm) dark red clasts and red mottling, similar to
that described for the multi-storey sandbodies.
Pedogenically modified facies these mudstones
are dark, reddish brown or greyish purple and have
no evident lamination or other primary sedimentary
structures. Most have weakly developed, intersecting
pedogenic slickensides (wedge-shaped peds) and sparse
cm-scale calcrete nodules (Fig. 11B, Stage I calcrete per
Machette, 1985) throughout. There are a few units
where larger calcrete nodules (Stage II and III of
Machette, 1985) are present in discrete horizons, but
these tend to be confined to the lowermost 135 m of the
exposed section at Freshwater West.
Other characteristic features in the mudstones
desiccation cracks are common, observed both as
polygonal features on bed surfaces and as infilled cracks
extending in some places down at least 1 m through the
mudstone beds. Crack infills comprise both sandstone
50 R.D. Hillier et al. / Sedimentary Geology 202 (2007) 3557

Fig. 11. Mudstone and pedogenic fabrics, all depths related to Fig. 5 log. (A). Laminated mudstone with vertical burrows, drab haloes and calcrete
tubules. Some laminae picked out by carbonate stringers (arrowed). Note also vertic and horizontal drab haloes. (B). Massive mudstone with
pedogenic slickensides (wedge-shaped peds) 247 m. (C). Mudstone units with brecciated zone showing cracking and reworking (arrowed). 28.5
29.5 m. Scale is 5 cm. (D). Mottling in fine-grained sandstone. Pencil is 150 mm.

and mudstone, the latter being picked out by slightly
different red colouration or reduction staining. Rubbly,
brecciated zones are also observed between mudstone
beds (Fig. 11C). Some of the mudstones are discrete
units of the laminated, massive or pedogenically
modified facies. However some are thicker units, up to
4 m thick, made up from different facies. Several of the
mudstones also show drab haloes. These are elongate or
reticular patterns of bluish grey staining that appear to
pick out burrows (or possible root traces), pedogenic
slickensides and bedding planes or laminae (Fig. 11A).
Representatives of all the mudstones described here can
be found in the heterolithic association, interbedded
with sandstones and intraformational conglomerates.
4.4.2. Interpretation and depositional environments
The mudstones appear to have been deposited in two
distinct environments of a river system influenced by a
semi-arid climate. Similar laminated and burrowed,
reddish brown mudstones have been described from the
Moor Cliffs Formation (Marriott and Wright, 2004),
where it was envisaged that they were deposited in
shallow, semi-permanent lakes or pools on the flood-
plains. Finely laminated muds could result from over-
bank suspension fallout during floodplain inundation
from many minor flood events. However, the size and
number of the burrows present imply that subaqueous
conditions probably persisted for longer periods than
minor flooding would allow. Semi-permanent, flood-
plain pools or pans, such as those on the Pongolo
floodplain, KwaZuluNatal, South Africa, many of
which are several hectares in size (Heeg and Breen,
1982), form wetland areas and suffer large variations in
water level. Following inundation the pans are replen-
ished and vegetation colonizes damp areas surrounding
them, increasing in coverage as pan water level falls
during the dry season. The vegetation enters the detritus
system to be devoured by aquatic decomposers when
water levels rise again (Furness and Breen, 1980, 1982).
In extended periods of drought pans, dry out completely.
Naturally we cannot be certain that the system was as
evolved during the early Devonian. In the CPSM there is
no in situ evidence of colonization by vegetation other
than the spores described by Higgs (2004), although
plant debris (tea-leaf detritus) is present in the green/
grey sandy associations, being preferentially preserved
 R.D. Hillier et al. / Sedimentary Geology 202 (2007) 3557
by the interpreted anoxic, high water table. Burrowing is
however extensive in many of the laminated units,
which implies that productivity was high as the
burrowers must have exploited a food source. The
massive mudstones have few sedimentary structures that
would indicate their origin, though some have thin lags
of intraformational calcrete-clast conglomerate and
pockets of floating sand grains and tiny calcrete clasts.
They resemble both at outcrop and in thin section
massive mudstones described from the Rat Island
Mudstone Member of the Freshwater West Formation
(Marriott et al., 2005) and the Moor Cliffs Formation
(Marriott and Wright, 2004; Wright and Marriott, 2007),
respectively above and below the CPSM. These units
were interpreted as within-channel deposits of bedload-
transported, mud aggregates as muddy braid-bars or
ephemeral channel fill, or as possibly deposited as a
muddy flood sheet. Sedimentary structures that would
identify bar forms are absent, having been lost during
compaction when the mud aggregates lose some of their
coherence (Maroulis and Nanson, 1996), though they
may be preserved if protected by a framework of
quartz grains (see Retallack, 2005, for observations on
the preservation of clay aggregates). Sand-sized, mud
aggregates are commonly found in areas where smectite
is dominant, as they are produced during the shrinking
and swelling cycles that characterise formation of gilgai
in Vertisols (see below). The aggregates are readily
entrained from soil surfaces by wind and water (Rust
and Nanson, 1989). The Channel Country of central
Australia is a modern analogue for the type of semi-arid
environment where low-gradient hinterlands restrict the
input of coarse, exotic sediment, encouraging reworking
and redistribution of sediment within the depositional
basin (Gibling et al., 1998). The massive mudstones are
found mainly in the lowermost 90 m of the outcrop at
Freshwater West, interspersed with thin, intraforma-
tional conglomerates and fine sandstones, and IHS units.
This suggests that this period of deposition at least was
dominated by a muddy river system (see discussion
below).
Both massive mudstones and laminated mudstones
may have been subjected to desiccation and pedogenic
processes when exposed. Pedogenesis would have
overprinted any primary sedimentary structures by the
formation of new textural and structural features. The
curved, slickensided surfaces (wedge-shaped peds) are
typical features of Vertisols and develop in soils rich in
swelling clays, giving the soil surface regularly spaced
mounds known as gilgai (Wilding and Tessier, 1988)
that provide local topography on generally flat flood-
plains. In calcic Vertisols, beneath the structural horizon
51
containing the slickensides, calcrete nodules occur at the
depth of seasonal wetting (wetting front). The size and
rate at which the nodules develop depend on the nature
of the substrate (Gile et al., 1966; Machette, 1985), and
on the residence time that sediment remains within soil-
forming horizons (Leeder, 1975; Marriott and Wright,
1993). This means that calcrete nodules will develop
relatively slowly in mud-rich alluvial sediment and
ultimately their development will be governed by the
rate at which new increments of sediment are added to
the soil surface by overbank flooding (or fallout from
dust storms in arid and semi-arid environments)
(Marriott and Wright, 1993; Wright and Marriott,
1996; Marriott and Wright, 2006).
In the CPSM, pedogenic calcretes are poorly devel-
oped, mostly only to Stage I (Machette, 1985), and consist
of discrete nodules 1020 mm in diameter, most of which
are sparsely distributed within horizons containing
slickensides. This tends to indicate that relatively regular
increments of sediment were added to the surface, since
the soil horizons moved upwards so that the calcic horizon
overprinted the zone with slickensides to form a cumulate
soil profile (Marriott and Wright, 1993, 2006). There are
very few palaeosols showing a greater degree of calcrete
development within the CPSM.
In addition to evidence for desiccation, the CPSM
shows many features associated with water logging
caused by raised water tables or surface inundation. The
bluish-grey staining associated with drab haloes in some
mudstones is probably due to local reduction, but unlike
circular greenish reduction spots and the staining of the
top of a few mm of mudstone units associated with
waterlogging from overlying waterlain deposits, the
drab haloes were likely to have been altered because of
changes to Eh potential around burrows (or root traces)
(Retallack, 2001). This would be due to the incorpora-
tion of organic material into the burrow by the burrow
maker, such as faeces or mucus, or by anaerobic decay
of organic material concentrated along prismatic ped
faces or laminae. The saturation that produced the drab
haloes is not likely to have persisted for long periods as
gleying of the unit would have occurred.
The presence of horizontally-elongate calcrete
nodules in the laminated mudstones and some sandstone
beds is probably an indication of intermittently raised
water tables, which can result from deflation of the
floodplain surface or increased infiltration rates (Huerta
and Armenteros, 2005). Longer periods of saturation,
either due to perennial river channels or in the
subsurface to raised water tables and aquifers, may
give rise to slight acidification of groundwater and
changes in Eh that occur particularly in the presence of
52 R.D. Hillier et al. / Sedimentary Geology 202 (2007) 3557
organic matter. This can result in reduction, mobilization
and reprecipitation of iron and manganese sesquioxides
in well-drained sediment (Leenheer, 2002). These
processes could have given rise to the colour mottling
and segregation seen in some of the sandbodies and
some mudstones in the CPSM (Fig. 11D).
5. Discussion
The CPSM at Freshwater West comprises a hetero-
lithic, broadly upward-coarsening succession with
variable alluvial architecture (Fig. 5). Controlling
influences on the varying facies associations described
above will be discussed in turn.
5.1. Climate: Short-term variations
The climate within the Lower ORS of the Anglo-
Welsh Basin has long been interpreted as warm to hot,
semi-arid with a seasonal rainfall. This interpretation
stems largely from the interpretation of Vertisols in
mudstone-dominated successions (e.g. Allen, 1974b;
Allen and Williams, 1979, 1981, 1982; Allen, 1985;
Marriott and Wright, 1993). Indicators of ephemeral
discharge during periods of semi-arid climate abound
within the fine-grained inclined and non-inclined
heterolithic association. These are typified by the
event nature of fluvial sedimentation, controlled by
seasonal, low annual rainfall, high rates of potential
evaporation, and sparse vegetation cover (Nanson et al.,
2002). This facies association is typical of endogenic
fluvial systems that are sourced within a dryland
environment. Discharge in such drainage networks
comprises discrete events, being commonly punctuated
by long periods of zero flow. Following precipitation
events, discharge diminishes downstream due to high
rates of infiltration and evaporation loss (Tooth, 2000a,b),
with increased flow only being seen in areas immediately
downstream of tributary confluences. Once flows exceed
bankfull stage, transmission losses can exceed 90%
(Nanson et al., 2002), promoting deposition at terminal
splays. Channel size rapidly decreases downstream in
response to transmission losses, ultimately with the
transported load being deposited at terminal floodouts
when competence is lost. Periods of intense desiccation
and subsequent pedogenesis may follow. The interpreta-
tion of pulsed, perennial flow during deposition of the
multi-storey sandstone complexes, and the lack of any
observations at Freshwater West to link these complexes
to the finer-grained inclined and non-inclined heterolithic
association (both temporally and spatially) requires
significant changes in discharge. It is possible that the
multi-storey units represent allogenic rivers that were
sourced almost entirely from outside of the dryland region
(Miall, 1996; Nanson et al., 2002), similar to the present-
day River Nile that travels through the Sahara Desert but
has perennial base flow sourced from relief rainfall and
snowmelt from the East African highlands. The juxtapo-
sition of humid upland source areas with arid or semi-arid
distal drainage networks is common (Miall, 1996), with
upland areas generating relief rainfall and downwind rain
shadow regions. Evidence to support an external source
within the CPSM is primarily that of grain-size: the multi-
storey units are coarser-grained compared to the hetero-
lithic associations. The moderate to well-sorted nature of
the former presumably indicates progressive sorting along
a lengthy drainage network. It is possible that high-relief
areas were periodically generated within the source region
of the drainage network due to activity on the Benton
Fault during the early Devonian, giving rise to localized
variations in climate, especially rainfall, and acting as the
source for coarser-grained detritus.
An alternative hypothesis explaining the changes
observed in the CPSM is that the punctuated episodes
of perennial flow may have been generated by periods of
increased rainfall and discharge sourced within the former
dryland region. Such variability in climate would need to
have been of high frequency to account for the rapid
variation in facies associations observed. Rapid recent
climate fluctuations within tropical regions are well
documented. For example, the Sahara Desert region
experienced a monsoonal climate between 14 ka and
5.5 ka, with increased rainfall maintaining a permanent
vegetation and dense drainage network. Rapid aridity
followed, with desert conditions establishing within a
century (De Menocal et al., 2000).
Shorter, century-long variations in the intensity of
rainfall associated with the Indian Summer Monsoon have
been identified in Holocene alluvial deposits along the
southern Himalayan front (Bookhagen et al., 2006).
Enhanced rainfall intensity led to increased sediment flux
from source areas, resulting in significant deposition in the
low gradient, distal reaches of the Sutlej River drainage
basin. Periods where rainfall intensity decreased led to
reduced sediment flux and transport capacity within the
drainage network promoting localized incision and terrace
formation.
5.2. Climate: long-term variations
Analysis of the vertical section through the CPSM at
Freshwater West reveals an upward increase in the
amount of coarse-grained sheet sandstones and multi-
storey complexes, and a corresponding decrease in the
 R.D. Hillier et al. / Sedimentary Geology 202 (2007) 3557
frequency and development stage of Vertisol profiles
(Fig. 5). This pattern may equate with the progressive
unroofing of a sand-prone provenance area. The trend
however, must reflect an upward increase in periods of
perennial flow, together with an upward decrease in
seasonal (wet/dry) climate variation. Similarly, the
change to predominantly mudstone-dominated inclined
and non-inclined heterolithic strata within the overlying
Rat Island Mudstone Member may also reflect a shift to
semi-arid dominated climate conditions.
It is possible that the low frequency and weak
development of Vertisol profiles in the upper half of the
CPSM section at Freshwater West is due to a decrease in soil
residence time related to varying accommodation controls.
This theme will be addressed below. The possibility of long-
term climate change either within the source region of the
drainage basin, or across the southern margin of Laurussia
requires further substantial investigation.
5.3. Accommodation controls
It is generally accepted that variations in accommoda-
tion (uplift, subsidence and incision) affect the style and
location of fluvial systems at the macroscale (e.g. Sloss,
1991; Leeder, 1993; see discussions in Ethridge et al., 1998;
Hickson et al., 2005). However at the channel belt scale
there is considerable disagreement as to the relative and
absolute importance of climate, tectonics and eustasy on
continental deposits. Indeed, the situation may be con-
founded by the process of geomorphic convergence,
whereby different causative processes may lead to similar
end products (Ethridge et al., 1998).
Changes in fluvial style between the high-sinuosity,
heterolithic associations and the low-sinuosity, multi-
storey channels may be the response to variable
sediment accommodation-space development. Fine-
grained fluvial systems appear to develop within low-
gradient drainage basins (Orton and Reading, 1993)
similar to that observed within the arid to semi-arid
interior of Australia, or that envisaged for the Lower
ORS Anglo-Welsh deposystem (Allen and Williams,
1979, 1981). Owen and Hawley (2000) have advocated
that tectonic controls generated changes in alluvial
grade and architecture within the CPSM age-equivalent
St. Maughans Formation in central South Wales. They
identified sandstone facies with architectural elements
identical to those of the CPSM multi-storey sandbodies
and upward-fining channel fills deposited by migrating
compound bars with well-developed abandonment
phases and coarse-grained, high-energy, channelized
lenses interpreted as sheet flows. This variable nature of
deposition is interpreted as representing distinct phases
53
of fluvial development reflecting tectonic controls
triggered by fault movement along the Carreg Cennen
Disturbance to the north. Owen and Hawley (2000)
discounted a climatic origin for the change in grade and
architecture due to the common occurrence of reworked
calcrete nodules within the sandstone bodies, interpreted
as derived from contemporaneous floodplain Vertisols.
The textural maturity of the multi-storey sandbodies
within the CPSM, however, implies significant progres-
sive sorting and therefore probable lengthy drainage
networks. The Anglo-Welsh Basin is interpreted as
having a much broader aerial extent, with the present
outcrop pattern being an erosional remnant following the
Emsian Acadian Orogeny (Soper and Woodcock, 2003).
Drainage networks were possibly larger than those
proposed by Owen and Hawley (2000), and the multi-
storey sandbodies could have been sourced by allogenic
rivers or a climate change within the basin itself.
6. Conclusions
The CPSM at Freshwater West comprises a multi-
stage, multi-channel, predominantly alluvial suite that
is heterolithic in nature. A number of discrete facies
associations are observed. The fine-grained inclined
and non-inclined heterolithic association is the domi-
nant association within the CPSM. It represents
ephemeral, sinuous channelized flow (IHS bedsets)
and unconfined flow (non-inclined bedsets) at terminal
and intermediate floodouts. The mudstones were depos-
ited in two distinct environments associated with this river
system. Laminated and burrowed, reddish brown mud-
stones were deposited in shallow, semi-permanent lakes
or pools on the floodplain, with laminated mudstones
resulting from suspension fallout during floodplain
inundation. Massive mudstones are interpreted as having
been deposited as within-channel, muddy braid-bars, or
possibly as muddy flood sheets. Sedimentary structures
that would identify specific depositional processes are
absent, having been lost during subsequent compaction.
Vertisols reflect seasonal wetting and drying associated
with the semi-arid climate. Periods of intense desiccation
are indicated by deep desiccation cracks and associated
rubbly surfaces.
Coarser-grained, multi-storey sandstone complexes
contain amalgamated storeys deposited by low-sinuosity
rivers with a fluctuating, but perennial discharge. Associ-
ated with these laterally extensive sandbodies was a high
water table with surface ponding that enabled the
preservation of plant detritus. During high discharge
events, flow expanded over the contemporaneous flood-
plain depositing sheet sandstones within splay complexes.
54 R.D. Hillier et al. / Sedimentary Geology 202 (2007) 3557
We speculate that this sheet sandstone association is
connected to the multi-storey sandbodies in a wing or
steer's head geometry. The interpretation of perennial
discharge possibly invokes a change in climate, challeng-
ing the long-held view that the Lower ORS climate was
entirely semi-arid. However the extent of this climate
change is unclear. Uplift of a distant source area may also
have instigated the rivers in which the multi-storey
sandbodies were deposited.
The upward decrease in the frequency and maturity
of Vertisol profiles, and the corresponding increase in
multi-storey sandbodies may reflect a long-term,
episodic pattern of climate change. Alternatively, the
pattern may reflect the progressive unroofing of a sand-
prone provenance, with the low frequency and weak
development of Vertisol profiles in the upper half of the
CPSM being due to a decrease in soil residence time
related to varying accommodation-space development.
Possible tide-influenced channels comprise a small,
but important volume of the CPSM at Freshwater West
though their extent is uncertain.
Further investigation of CPSM localities is needed to
identify temporal/spatial patterns of architectural change
and tidal influence within south Pembrokeshire, and to
distinguish between the relative controlling processes.
The study then needs to expand, to incorporate other
age-equivalent sections in order to identify larger-scale
climate patterns and depositional processes.
Acknowledgements
The manuscript benefited significantly from the
constructive reviews provided by W.J. Barclay and Dr.
S. Gould. The Joint Nature Conservation Committee is
thanked for permission to publish Fig. 5.
References
Adams, M.M., Bhattacharya, J.P., 2005. No change in fluvial style across
a sequence boundary, Cretaceous Blackhawk and Castlegate
Formations of central Utah, USA. J. Sediment. Res. 75, 10381051.
Allen, J.R.L., 1964. Studies in fluviatile sedimentation: six cyclothems
from the Lower Old Red Sandstone, Anglo-Welsh Basin. Sedimen-
tology 3, 163198.
Allen, J.R.L., 1965. The sedimentation and palaeogeography of the
Old Red Sandstone of Anglesey, North Wales. Proc. Yorks. Geol.
Soc. 35, 139185.
Allen, J.R.L., 1970. Studies in fluviatile sedimentation: a comparison
of fining-upwards cyclothems with reference to coarse member
composition and interpretation. J. Sediment. Petrol. 40, 298323.
Allen, J.R.L., 1973a. Compressional structures (patterned ground) in
Devonian pedogenic limestones. Nature 243, 8486.
Allen, J.R.L., 1973b. A new find of bivalve molluscs in the uppermost
Downtonian (Lower Old Red Sandstone) of Lydney, Gloucester-
shire. Proc. Geol. Assoc. 84, 2729.
Allen, J.R.L., 1974a. Studies in fluviatile sedimentation: lateral variation in
some fining-upwards cyclothems from the Red Marls, Pembrokeshire.
Geol. J. 9, 116.
Allen, J.R.L., 1974b. Studies in fluviatile sedimentation: implications of
pedogenic carbonate units, Lower Old Red Sandstone, Anglo-Welsh
outcrop. Geol. J. 9, 181208.
Allen, J.R.L., 1983. Studies in fluvial sedimentation: bars, bar-complexes
and sandstone sheets (low-sinuosity braided streams) in the Brown-
stones (L. Devonian), Welsh Borders. Sediment. Geol. 33, 237293.
Allen, J.R.L., 1985. Marine to fresh water: the sedimentology of the
interrupted environmental transition (Ludlow-Siegenian) in the
Anglo-Welsh region. Philos. Trans. R. Soc. Lond., B 309, 85104.
Allen, J.R.L., 1986. Pedogenic calcretes in the Old Red Sandstone facies
(late Silurianearly Carboniferous) of the Anglo-Welsh area,
southern Britain. In: Wright, V.P. (Ed.), Palaeosols: their Recognition
and Interpretation. Blackwell Science, Oxford, U.K., pp. 5686.
Allen, J.R.L., Crowley, S.J., 1983. Lower Old Red Sandstone fluvial dispersal
systems in the British Isles. Trans. R. Soc. Edinb. Earth Sci. 74, 6168.
Allen, J.R.L., Williams, B.P.J., 1978. The sequence of the earlier
Lower Old Red Sandstone (Siluro-Devonian), north of Milford
Haven, southwest Dyfed (Wales). Geol. J. 13, 113136.
Allen, J.R.L., Williams, B.P.J., 1979. Interfluvial drainage on Siluro-
Devonian alluvial plains in Wales and the Welsh Borders. J. Geol.
Soc. (Lond.) 136, 361366.
Allen, J.R.L., Williams, B.P.J., 1981. Sedimentology and stratigraphy
of the Townsend Tuff Bed (Lower Old Red Sandstone) in South
Wales and the Welsh Borders. J. Geol. Soc. (Lond.) 138, 1529.
Allen, J.R.L., Williams, B.P.J., 1982. The architecture of an alluvial suite:
rocks between the Townsend Tuff and Pickard Bay Tuff Beds (early
Devonian), southwest Wales. Philos. Trans. R. Soc. Lond., B 287,
5189.
Allen, J.R.L., Elliot, T., Williams, B.P.J., 1981. Old Red Sandstone and
Carboniferous fluvial sediments in South Wales. In: Elliot, T. (Ed.),
Field Guides to Modern and Ancient Fluvial Systems in Britain and
Spain. Proceedings of the 3rd International Symposium on Fluvial
Sedimentology, Keele, pp. 139.
Allen, J.R.L., Thomas, R.G., Williams, B.P.J., 1982. The Old Red
Sandstone north of Milford Haven. In: Bassett, M.G. (Ed.),
Geological Excursions in Dyfed, Southwest Wales. National
Museum of Wales, Cardiff, pp. 123149.
Almond, J., Williams, B.P.J., Woodcock, N.H. 1993. The Old Red
Sandstone of the Brecon Beacons to Black Mountains area. In:
Woodcock, N.H., Bassett, M.G. (Eds.), Geological Excursions in
Powys. University of Wales Press, National Museum of Wales, Cardiff,
pp. 311330.
Ashworth, P.J., Best, J.L., Roden, J.E., Bristow, C.S., Klaassen, G.J.,
2000. Morphological evolution and dynamics of a large, sand braid-
bar, Jamuna River, Bangladesh. Sedimentology 47, 533555.
Augustus, P.G.E.F., Slager, S., 1970. Soil formation in swamp soils of
the coastal fringe of Surinam. Geoderma 6, 203211.
Barclay, W.J., Rathbone, P.A., White, D.E., Richardson, J.B., 1994.
Brackish water faunas from the St. Maughans Formation: the Old
Red Sandstone section at Ammons Hill, Hereford and Worcester,
UK, re-examined. Geol. J. 29, 369379.
Barclay, W.J., Browne, M.A.E., McMillan, A.A., Pickett, E.A.,
Stone, P., Wilby, P.R., 2005. The Old Red Sandstone of Great
Britain. Geological Conservation Review Series Joint Nature
Conservation Committee, vol. 31. 393 pp.
Best, J.L., Ashworth, P.J., Bristow, C.S., Roden, J., 2003. Three-
dimensional sedimentary architecture of a large, mid-channel sand
braid bar, Jamuna River, Bangladesh. J. Sediment. Res. 73,
516530.
 R.D. Hillier et al. / Sedimentary Geology 202 (2007) 3557
Blodgett, R.H., Stanley, K.O., 1980. Stratification, bedforms, and
discharge relations of the Platte braided river system, Nebraska.
J. Sediment. Petrol. 50, 139148.
Bookhagen, B., Fleitmann, D., Nishiizimi, K., Strecker, M.R., Thiede,
R.C., 2006. Holocene monsoonal dynamics and fluvial terrace
formation in the northwest Himalaya, India. Geology 34,
65016604.
Bridge, J.S., 1993. Description and interpretation of fluvial deposits: a
critical perspective. Sedimentology 40, 801810.
Bridge, J.S., Smith, N.D., Trent, F., Gabel, S.L., Bernstein, P., 1996.
Sedimentology and morphology of a low-sinuosity river: Calamus
River, Nebraska Sand Hills. Sedimentology 33, 851870.
Brinkman, R., Bao Ve, N., Kim Tink, T., Phuoc Hau, D., van Mensvoort,
M.E.F., 1993. Sulfidic materials in the Western Mekong Delta,
Vietnam. Catena 20, 317331.
Bristow, C.S., Skelly, R.L., Ethridge, F.G., 1999. Crevasse splays from
the rapidly aggrading, sand-bed, braided Niobrara River, Nebras-
ka: effect of base-level rise. Sedimentology 46, 10291047.
Cant, D.J., Walker, R.G., 1978. Fluvial processes and facies sequences in
the sandy braided South Saskatchewan River, Canada. Sedimentol-
ogy 25, 625648.
Channel, J.E.T., McCabe, C., Woodcock, N.H., 1992. An Early
Devonian (pre-Acadian) magnetization component recorded in
the Lower Old Red Sandstone of South Wales (UK). Geophys.
J. Int. 108, 883894.
De Menocal, P.D., Ortiz, J., Guilderson, T., Sarnthein, M., 2000.
Coherent high and low-latitude climate variability through the
Holocene warm period. Science 292, 667673.
Dewey, J.F., Strachan, R.A., 2003. Changing SilurianDevonian
relative plate motion in the Caledonides: sinistral transpression to
sinistral transtension. J. Geol. Soc. (Lond.) 160, 219229.
Dixon, E.E., 1921. Geology of the South Wales Coalfield: Part XIII:
the country around Pembroke and Tenby. Mem. Geol. Surv.
Britain, sheets 244 and 245, H.M.S.O. 220 pp.
Edwards, D., Richardson, J.B., 2004. Silurian and Lower Devonian
plant assemblages from the Anglo-Welsh Basin: a palaeobotanical
and palynological synthesis. Geol. J. 39, 375402.
Ethridge, F.G., Wood, L.J., Schumm, S.A., 1998. Cyclic variables
controlling sequence development: problems and perspectives. In:
Shanley, K.W., McCabe, P.J. (Eds.), Relative Roles of Eustasy,
Climate, and Tectonism in Continental Rocks Spec. Pub. Soc.
Econom. Pal. and Miner., vol. 59, pp. 1729.
Friend, P.F., Williams, B.P.J., Ford, M., Williams, E.A., 2000.
Kinematics and dynamics of Old Red Sandstone basins. In: Friend,
P.F., Williams, B.P.J. (Eds.), New Perspectives on the Old Red
Sandstone Geol. Soc. London Spec. Pub., vol. 180, pp. 2960.
Furness, H.D., Breen, C.M., 1980. The vegetation of seasonally flooded
areas of the Pongolo River floodplain. Bothalia 13, 217231.
Furness, H.D., Breen, C.M., 1982. Decomposition of Cynodon dactylon
(L.) Pers. in seasonally flooded areas of the Pongolo River floodplain:
pattern and significance of dry matter and nutrient loss. Hydro-
biologia 97, 119126.
Gibling, M.R., Nanson, G.C., Maroulis, J.C., 1998. Anastomosing
river sedimentation in the Channel Country of central Australia.
Sedimentology 45, 595619.
Gile, L.H., Peterson, F.F., Grossman, R.B., 1966. Morphological and
genetic sequences of carbonate accumulation in desert soils. Soil
Sci. 101, 347360.
Heeg, J., Breen, C.M., 1982. Man and the Pongolo floodplain.
South African National Scientific Programmes Report, vol. 56.
Council for Scientific and Industrial Research, Pretoria, South
Africa.
55
Hickson, T.A., Sheets, B.A., Paola, C., Kelberer, M., 2005. Experimental
test of tectonic controls on three-dimensional alluvial facies
architecture. J. Sediment. Res. 75, 710722.
Higgs, K.T., 2004. An early Devonian (Lochkovian) microflora from
the Freshwater West Formation, Lower Old Red Sandstone,
southwest Wales. Geol. J. 39, 359374.
Hillier, R.D., Williams, B.P.J., 2004. Sedimentation and tectonics: the
marine Silurianbasal Lower Old Red Sandstone transition in
southwest Wales. Geol. J. 39, 237256.
Hillier, R.D., Williams, B.P.J., 2006. The alluvial Old Red Sandstone
basins, In: Brenchley, P.J., Rawson, P.F. (Eds.), The Geology of
England and Wales, Third edition. Geol. Soc., London, pp. 155171.
Hillier, R.D., Williams, B.P.J., 2007. The Ridgeway Conglomerate
Formation of SW Wales, and its implications. The end of the
Lower Old Red Sandstone? Geol. J. 42, 5583.
Holbrook, J.M., 1996. Complex fluvial response to low gradients at
maximum regression: a genetic link between smooth sequence-
boundary morphology and architecture of overlying sheet sandstone.
J. Sediment. Res. 66, 713722.
Huerta, P., Armenteros, I., 2005. Calcrete and palustrine assemblages on
a distal alluvial floodplain: a response to local subsidence (Miocene
of the Duero Basin, Spain). Sediment. Geol. 177, 253270.
Kraus, M.J., 1996. Avulsion deposits in Lower Eocene alluvial rocks,
Bighorn Basin, Wyoming. J. Sediment. Petrol. 66B, 354363.
Leeder, M.R., 1975. Pedogenic carbonates and floodplain sediment
accretion rates: a quantitative model for alluvial arid-zone lithofacies.
Geol. Mag. 112, 257270.
Leeder, M.R., 1993. Tectonic controls upon drainage basin develop-
ment, river channel migration and alluvial architecture: implica-
tions for hydrocarbon reservoir development and characterisation.
In: North, C.P., Prosser, D.J. (Eds.), Characterization of Fluvial and
Aeolian Hydrocarbon Reservoirs Spec. Pub. Geol. Soc. London,
vol. 73, pp. 722.
Leenheer, J., 2002. Processes controlling attenuation of dissolved
organic matter in the subsurface. In: Aiken, G.R., Kuniansky, E.L.
(Eds.), U.S. Geological Survey Artificial Recharge Workshop
Proceedings, Sacramento, California, April 24, 2002: USGS
Open-file Report 02-89.
Lin, C., Melville, M.D., Hafer, S., 1995. Acid sulphate soillandscape
relationships in an undrained, tide-dominated estuarine floodplain,
Eastern Australia. Catena 24, 177194.
Love, S.E., Williams, B.P.J., 2000. Sedimentology, cyclicity and
floodplain architecture in the Lower Old Red Sandstone of SW
Wales. In: Friend, P.F., Williams, B.P.J. (Eds.), New Perspectives
on the Old Red Sandstone Geol. Soc. London Spec. Pub., vol. 180,
pp. 371388.
Machette, M.N., 1985. Calcic soils of the southwestern United States.
In: Weide, D.L. (Ed.), Soils and Quaternary Geology of the South
West United States Geol. Soc. of Amer. Spec. Papers, vol. 203,
pp. 121.
Madsen, H.B., Jensen, N.H., 1988. Potentially acid sulphate soils in
relation to landforms and geology. Catena 15, 137145.
Makaske, B., 2001. Anastomosing rivers: a review of their classification,
origin and sedimentary products. Earth-Sci. Rev. 53, 149196.
Maroulis, J.C., Nanson, G.C., 1996. Bedload transport of aggregated
muddy alluvium from Cooper Creek, central Australia. A flume
study. Sedimentology 43, 771790.
Marriott, S.B., Wright, V.P., 1993. Palaeosols as indicators of geomorphic
stability in two Old Red Sandstone alluvial suites, South Wales.
J. Geol. Soc. (Lond.) 150, 11091120.
Marriott, S.B., Wright, V.P., 1996. Sediment recycling on Siluro-
Devonian floodplains. J. Geol. Soc. (Lond.) 153, 661664.
56 R.D. Hillier et al. / Sedimentary Geology 202 (2007) 3557
Marriott, S.B., Wright, V.P., 2004. Mudrock deposition in an ancient
dryland system: Moor Cliffs Formation, Lower Old Red
Sandstone, southwest Wales, UK. Geol. J. 39, 277298.
Marriott, S.B., Wright, V.P., 2006. Investigating paleosol completeness
and preservation in mid-Paleozoic alluvial paleosols: a case study
in paleosol taphonomy from the Lower Old Red Sandstone. In:
Alonso-Zarza, A.M., Tanner, L. (Eds.), Paleoenvironmental
Record and Applications of Calcretes and Palustrine Carbonates
Geol. Soc. of Amer. Spec. Papers, vol. 416, pp. 4352.
Marriott, S.B., Wright, V.P., Williams, B.P.J., 2005. A new evaluation
of fining-upward sequences in a mud-rock dominated succession
of the Lower Old Red Sandstone of South Wales, UK. In: Blum,
M.D., Marriott, S.B., Leclair, S. (Eds.), Fluvial Sedimentology VII
Int. Assoc. Sedimentol. Spec. Publ., vol. 35, pp. 517529.
Miall, A.D., 1977. A review of the braided river depositional
environment. Earth-Sci. Rev. 13, 162.
Miall, A.D., 1978. Lithofacies types and vertical profile models in
braided river deposits: a summary. In: Miall, A.D. (Ed.), Fluvial
Sedimentology Can. Soc. Pet. Geol. Mem., vol. 5, pp. 597604.
Miall, A.D., 1996. The Geology of Fluvial Deposits, Sedimentary
Facies, Basin Analysis, and Petroleum Geology. SpringerVerlag,
Berlin 582 pp.
Morrissey, L.B., Braddy, S.J., 2004. Terrestrial trace fossils from the
Lower Old Red Sandstone of southwest Wales. Geol. J. 39,
315336.
Nadon, G.C., 1994. The genesis and recognition of anastomosed fluvial
deposits: data from the St. Mary River Formation, southwestern
Alberta, Canada. J. Sediment. Res., Sect. B Stratigr. Glob. Stud. 64,
451463.
Nanson, G.C., Rust, B.R., Taylor, G., 1986. Coexistent mud braids and
anastomosing channels in an arid-zone river: Cooper Creek, central
Australia. Geology 14, 175178.
Nanson, G.C., Tooth, S., Knighton, A.D., 2002. A global perspective on
dryland rivers: perceptions, misconceptions and distinctions. In: Bull,
L.J., Kirkby, M.J. (Eds.), Dryland Rivers: Hydrology and Geomor-
phology of Semi-arid Channels. John Wiley & Sons, London,
pp. 1754.
O'Brien, P.E., Wells, A.T., 1986. A small, alluvial crevasse splay.
J. Sediment. Res. 56, 876879.
Orton, G.J., Reading, H.G., 1993. Variability of deltaic processes in
terms of sediment supply, with particular emphasis on grain size.
Sedimentology 40, 475512.
Owen, G., 1995. Senni Beds of the Devonian Old Red Sandstone,
Dyfed, Wales: anatomy of a semi-arid floodplain. Sediment. Geol.
95, 221235.
Owen, G., Hawley, D., 2000. Depositional setting of the Lower Old
Red Sandstone at Pantymaes Quarry, central South Wales: new
perspectives on the significance and occurrence of Senni Beds
facies. In: Friend, P.F., Williams, B.P.J. (Eds.), New Perspectives on
the Old Red Sandstone Geol. Soc. London Spec. Pub., vol. 180,
pp. 389400.
Retallack, G.J., 2001. Soils of the Past: an Introduction to
Paleopedology. Balckwell Science, Oxford 404 pp.
Retallack, G.J., 2005. Earliest Triassic claystone breccias and soil-
erosion crisis. J. Sediment. Res. 75, 663679.
Rust, B.R., Nanson, G.C., 1986. Contemporary and palaeochannel
patterns and the late Quaternary stratigraphy of Cooper Creek,
southwest Queensland, Australia. Earth Surf. Process. Landf. 11,
581590.
Rust, B.R., Nanson, G.C., 1989. Bedload transport of mud as pedogenic
mud aggregates in modern and ancient rivers. Sedimentology 36,
291306.
Sambrook Smith, G.H., Ashworth, P.J., Best, J.L., Woodward, J.,
Simpson, C.J., 2005. The morphology and facies of sandy braided
rivers: some considerations of scale invariance. In: Blum, M.D.,
Marriott, S.B., Leclair, S.F. (Eds.), Fluvial Sedimentology VII Int.
Assoc. Sediment. Spec. Publ., vol. 35, pp. 145158.
Simon, J.B., Bluck, B.J., 1982. Palaeodrainage of the southern margin
of the Caledonide mountain chain in the northern British Isles.
Trans. R. Soc. Edinb. Earth Sci. 73, 1115.
Sloss, L.L., 1991. The tectonic factor in sea-level change: a counter
veiling view. J. Geophys. Res. 96, 66096617.
Smith, N.D., 1970. The braided stream depositional environment:
comparison of the Platte River with some Silurian clastic rocks,
north central Appalachians. Geol. Soc. Amer. Bull. 81, 29933014.
Smith, N.D., Cross, T.A., Dufficy, J.P., Clough, S.R., 1989. Anatomy
of an avulsion. Sedimentology 36, 124.
Smith, J., van Oploo, P., Marston, H., Melville, M.D., Macdonald,
B.C.T., 2003. Spatial distribution and management of total actual
acidity in an acid sulphate soil environment. Catena 51, 6179.
Soper, N.J., Woodcock, N.H., 2003. The lost Lower Old Red
Sandstone of England and Wales: a record of post-Iapetan flexure
or Early Devonian transtension? Geol. Mag. 140, 627647.
Stear, W.M., 1983. Morphological characteristics of ephemeral stream
channel and overbank splay sandstone bodies in the Permian
Lower Beaufort Group, Karoo Basin, South Africa. In: Collinson,
J.D., Lewin, J. (Eds.), Modern and Ancient Fluvial Systems
Special Publication of the International Association of Sedimen-
tologists, vol. 6, pp. 405420.
Stear, W.M., 1985. Comparison of the bedform distribution and
dynamics of modern and ancient sandy ephemeral flood deposits in
the southwestern Karoo region, South Africa. Sediment. Geol. 45,
209230.
Tooth, S., 1999. Floodouts in Central Australia. In: Miller, A.J., Gupta, A.
(Eds.), Varieties of Fluvial Form. John Wiley & Sons, pp. 219247.
Tooth, S., 2000a. Process, form and change in dryland rivers: a review
of recent research. Earth-Sci. Rev. 51, 67107.
Tooth, S., 2000b. Downstream changes in dryland river channels: the
Northern Plains of arid central Australia. Geomorphology 34,
3354.
Tooth, S., 2005. Splay formation along the lower reaches of ephemeral
rivers on the Northern Plains of arid central Australia. J. Sediment.
Res. 75, 636649.
Tooth, S., Nanson, G.C., 2000. Equilibrium and nonequilibrium
conditions in dryland rivers. Phys. Geogr. 21, 183211.
Tunbridge, I.P., 1981. Old Red Sandstone sedimentation an
example from the Brownstones (highest Lower Old Red Sand-
stone) of south central Wales. Geol. J. 16, 111124.
Walker, R.G., Cant, D.J., 1984. Sandy fluvial systems, In: Walker, R.G.
(Ed.), Facies Models, Second edition. Geoscience Canada, Reprint
Series, vol. 1, pp. 7189.
Wilding, L.P., Tessier, D., 1988. Genesis of Vertisols: shrink swell
phenomena. Vertisols: their Distribution, Properties, Classification
and Management, Technical Monograph, 18 SMSS. Texas A & M
University, pp. 5581.
Williams, G.E., 1970. Flood deposits of the sand-bed ephemeral
streams of central Australia. Sedimentology 17, 140.
Williams, B.P.J., Hillier, R.D., 2004. Variable alluvial sandstone
architecture within the Lower Old Red Sandstone, southwest
Wales. Geol. J. 39, 257275.
Williams, B.P.J., Allen, J.R.L., Marshall, J.D., 1982. Old Red Sandstone
facies of the Pembroke Peninsula, south of the Ritec Fault. In:
Bassett, M.G. (Ed.), Geological Excursions in Dyfed, Southwest
Wales. National Museum of Wales, Cardiff, pp. 151174.
 R.D. Hillier et al. / Sedimentary Geology 202 (2007) 3557 57

Williams, B.P.J., Marriott, S.B., Hilllier, R.D., 2004. Preface: the
Lower Old Red Sandstone of the Anglo-Welsh Basin: an
introduction. Geol. J. 39, 233236.
Wilson, B.P., 2005. Elevations of sulfurous layers in acid sulfate soils:
what do they indicate about sea-levels during the Holocene in
eastern Australia? Catena 62, 4556.
Wright, V.P., Marriott, S.B., 1996. A quantitative approach to soil
occurrence in alluvial deposits and its application to the Old Red
Sandstone of Britain. J. Geol. Soc. (Lond.) 153, 907913.
Wright, V.P., Marriott, S.B., 2007. The dangers of taking mud for
granted: lessons from Lower Old Red Sandstone dryland river
systems of South Wales. Sediment. Geol. 195, 91100.
 Available online at www.sciencedirect.com

Sedimentary Geology 202 (2007) 58 71

www.elsevier.com/locate/sedgeo

Sedimentology of the Whiteclay Gravel Beds (Ogallala Group) in

northwestern Nebraska, USA: Structurally controlled drainage
promoted by Early Miocene uplift of the Black Hills Dome
Christopher R. Fielding a,, Hannan E. LaGarry b , Leigh Anne LaGarry b ,
Bruce E. Bailey c , James B. Swinehart b
a
Department of Geosciences, 214 Bessey Hall, University of Nebraska-Lincoln, Lincoln NE 68588-0340, USA
b
Conservation & Survey Division, School of Natural Resources, 113 Nebraska Hall, University of Nebraska-Lincoln,
Lincoln NE 68588-0517, USA
c
University of Nebraska State Museum, 307 Morrill Hall, University of Nebraska-Lincoln, Lincoln NE 68588-0338, USA

Abstract

The newly recognized Whiteclay Gravel Beds (WGB) of the Miocene Ogallala Group crop out as a narrow, discontinuous
ribbon of sands and gravels in Dawes and Sheridan Counties, northwestern Nebraska, USA. The WGB are exposed in a series of
municipal gravel quarries and natural exposures that define a linear trench in underlying strata at least 20 m deep and up to 300 m
wide, with short, southeast-trending reaches separating generally longer east-trending sections. This gravel-filled trench can be
recognized from the NebraskaSouth Dakota border near Whiteclay, Nebraska southeastward to east of Gordon, Nebraska, a
distance of 30 km. The outcrop belt of the WGB is coincident in location and trend with the Whiteclay Fault Zone. Where
exposed in quarries, the walls of the trench are steep-sided, vertical, or locally overhanging. Polished surfaces, slickensides, and
parallel joint sets are common in the walls of the trench near Whiteclay, but uncommon in those to the east. The narrow belt defined
by this trench is filled by stratified gravel (b 2.0 m, typically b 0.3 m) of sedimentary lithologies derived from various Cenozoic
units (but principally Anderson Ranch Formation), and sand. Relatively small amounts of unrounded granitic, volcanic, and quartz
gravel are preserved in places. Cross-bedding and clast imbrication indicate palaeoflow towards the east.
The WGB are interpreted to have formed in response to tectonic upheaval associated with uplift of the Black Hills of South
Dakota in Early Miocene times. Fault rupture topography facilitated formation of a steep-sided canyon, or valley, up to 20 m deep,
being virtually straight with sharp bends at intervals of several km. An alluvial channel belt developed in the floor of the valley,
filling the available accommodation space with coarse sand and gravel via aggradational stacking of the deposits of successive
channels and channel belts. Channel belts were probably braided, with individual channels up to 4 m deep and a few tens of metres
wide. The multi-storey character of the deposit indicates multiple episodes of cutting and filling. The coarse grain-size of the fill
suggests energetic discharge with frequent bankfull flows, even though the system had a relatively low gradient (0.004). An
abundance of reworked fossil debris is derived from several stratigraphic units, clasts of which have been identified in the fill. The
presence of a contemporary merychippine horse and a primitive species of the oreodont Brachycrus constrain formation of the
WGB to a short interval within the Early Miocene (c. 17.5 Ma). The mammal fauna suggests that this stream was a valuable source

Corresponding author.
E-mail address: cfielding2@unl.edu (C.R. Fielding).

0037-0738/$ - see front matter 2007 Elsevier B.V. All rights reserved.
doi:10.1016/j.sedgeo.2006.12.009
 C.R. Fielding et al. / Sedimentary Geology 202 (2007) 5871
of water, while fragments of aquatic organisms such as turtles and fish indicate perennial discharge. The WGB provides a crucial
window into a pluvial period in the Miocene that is largely unpreserved elsewhere in the basin, facilitated in part by fault rupture
topography.
2007 Elsevier B.V. All rights reserved.
Keywords: Nebraska; Cenozoic; Structurally controlled drainage; Fluvial system; Incised valley fill
1. Introduction
Continental stratigraphy is typically fragmentary,
with generally only a small fraction of elapsed time
recorded in preserved deposits, and more time being lost
in omission surfaces of varying magnitude. According-
ly, the reconstruction of palaeoclimate and other
palaeoenvironmental conditions from such archives is
challenging, and often time-space frameworks contain
large gaps in knowledge due to lack of record.
Recognition of discrete stratigraphic windows can
help to fill in these gaps in knowledge in certain
circumstances. This paper describes and interprets one
such serendipitous find in the Miocene continental
record of the High Plains in northwest Nebraska, USA.
It allows new insight into a discrete period of the Early
Miocene that is otherwise unrecorded in the region, and
which is in many ways out of context. This deposit is,
furthermore, a well-exposed and quite unequivocal
example of structurally-controlled drainage, allowing
new insights into the characteristics of such fluvial
milieux.
2. Regional geological setting
Regional uplift of the Rocky Mountains and Great
Plains region and overfilling of the foreland basin
during the Laramide Orogeny forced the retreat of the
Cretaceous Interior Seaway (6065 Ma), resulting in the
subaerial exposure and weathering of rock units of Early
Cretaceous to Eocene age across the northern Great
Plains (e.g., Heller et al., 1988). In northwestern
Nebraska (Fig. 1), Tertiary strata rest unconformably
on pedogenically-modified Cretaceous Pierre Shale
(Retallack, 1983; Terry, 1998). The Hartville, Laramie,
and Black Hills uplifts provided sediment for rivers
which flowed east-southeast across the region (Clark,
1975; Stanley and Benson, 1979; Swinehart et al.,
1985). Uplift of the Black Hills is generally believed to
have ended in the Eocene prior to deposition of the
White River Group (Lisenbee and DeWitt, 1993).
Overlying the Pierre Shale (Late Cretaceous) are, in
ascending order, the volcaniclastic claystones and
59
siltstones of the White River Group (late Eocene
Oligocene), the volcaniclastic sandstones of the Arikaree
Group (late OligoceneEarly Miocene), and the terrige-
nous clastic deposits of the Ogallala Group (Miocene).
These strata are some of the most highly fossiliferous
rocks exposed in North America (Fig. 2). All of these
units contain incised fluvial systems (palaeovalley fills)
that drained broadly away from the orogenic mountain
range eastward across the overfilled foreland (Swinehart
et al., 1985).
The Ogallala Group (Fig. 2) consists principally of
mudrocks and sandstones representing several distinct
episodes of aeolian and fluvial sedimentation generally
resulting from the uplift of the Rocky Mountains to the
west. Carbonate nodules formed in palaeophreatic zone
environments, and palaeosols are common, as are root
traces and vertebrate fossils. A variety of Formation and
Member level units has been defined within the Ogallala
Group (Fig. 2), with the newly-defined Whiteclay Gravel
Beds an as yet unassigned unit near the base. The unit
maps as a discrete, eastsoutheast-trending, linear belt of
coarse-grained clastic sediment that appears to be
discordant to both interpreted Cenozoic drainage axes in
the region (Swinehart et al., 1985) and modern fluvial
drainage patterns (Fig. 1). The trend and location of the
outcrop belt, however, are closely coincident with the
surface trace of the Whiteclay Fault Zone, which is
estimated to displace Cenozoic stratigraphic units by up to
70100 m in the study area (Souders, 1981).
The Whiteclay Gravel Beds have been exposed at the
surface 1) by natural landscape processes, which have
formed inverse topography in many areas in which the
gravel-rich alluvial facies form linear topographic highs
(Fig. 3A), and 2) by commercial and municipal excava-
tions for road base materials (Fig. 3B).
3. Sedimentology of the Whiteclay Gravel Beds
(WGB)
3.1. Methods
The WGB were mapped by a combination of
ground investigation and examination of remotely-
60 C.R. Fielding et al. / Sedimentary Geology 202 (2007) 5871

Fig. 1. Maps showing the location of the study area in central-northern USA. Note the surface traces of faults propagating outward from the Black
Hills Dome. The study area is the northwestern part of Nebraska.

sensed data. Gravel excavations and natural exposures
were visited, vertical sections logged, and facies re-
lationships recorded in two and three dimensions using
maps and photomosaics. Palaeocurrent data were
collected from selected sedimentary structures, notably
cross-bedding in sands and clast imbrication in
gravels. Clast imbrication measurements were restrict-
ed mainly to large clasts, and only in situations where
the original horizontal surface could be established
independently. Maximum clast size was assessed at
several localities by measuring the a-axis of the twenty
largest visible clasts and calculating the mean of each
population.
3.2. External geometry of the WGB lithosome
The WGB are preserved as a linear, broadly
westnorthwesteastsoutheast-trending belt between
100 and 300 m wide, inset predominantly into the
older Miocene Anderson Ranch Formation (Hunt, 2002;
Fig. 2). The belt has been mapped by tracing strings of
natural and artificial exposures across the landscape
physically and by means of remote imagery. In this way,
a single, linear tract can be traced from the Nebraska/
South Dakota State line at Whiteclay, NE, eastsoutheast-
ward to northeast of the town of Gordon, NE, a distance
of c. 30 km (Fig. 4). Our mapping has established the
 C.R. Fielding et al. / Sedimentary Geology 202 (2007) 5871 61

course of the channel in considerable detail. Many propagating away from the southern Black Hills uplift
sections of the channel are virtually straight, with of South Dakota (Figs. 1, 4). The channel occupies a
apparently abrupt corners forming an en echelon pattern position within the surface trace of the Whiteclay Fault

Fig. 2. Cenozoic stratigraphic column for northwestern Nebraska (modified from Swinehart et al., 1985).
62 C.R. Fielding et al. / Sedimentary Geology 202 (2007) 5871
Fig. 3. Photographs illustrating principal styles of exposure of the
Whiteclay Gravel Beds. A) Natural inverted topography exposure on a
ridge top, Collins property. B) Municipal excavation for road base,
northeast of Gordon.
Zone near Whiteclay (including the Pine Springs Ranch
locality: Fig. 4), but some distance to the south of, and
parallel to, the fault trace further east (Fig. 4).
Since the WGB are preserved at or near the present
land surface, it is possible that the original gradient of
the formative stream may be preserved. Comparison of
surface elevations representing the top of the channel fill
at various locations reveals some interesting patterns. At
the Pine Springs Ranch locality, close to the west-
northwestern end of the mapped belt, the WGB occur at
an anomalously low elevation (1060 m), whereas the
remainder of the major exposures show a progressive
decline in elevation from Reeves' Pits (1200 m) east-
southeast to the Gordon Pits (1110 m), giving an
approximate gradient of 0.004 (Fig. 4). The progressive
decline in elevation from Reeves' Pits to the Gordon Pits
is interpreted as a more or less unmodified fluvial
gradient, whereas the anomalously low elevation of the
WGB at Pine Springs is interpreted to be the result of
post-17.5 Ma tectonic activity. Evidence of post-WGB
structural activity is preserved in the Pine Springs Pits,
in the form of fractured, sheared and slickensided WGB
gravels, vertical to locally overturned gravel-basement
contacts, and small-scale fault offsets of strata (Fig. 5),
and the locality lies within the mapped surface trace of
the Whiteclay Fault Zone. Furthermore, a tremor of
Richter Magnitude 4.5 near Whiteclay, NE, in May
2003, and other historical events indicate that the
Whiteclay Fault Zone is still active in modern times.
At most localities, the margins of the WGB tract are
steep-sided, with walls dipping at 6090, being locally
overturned as noted above (Fig. 5). Numerous widespread
exposures of the contact confirm that this is an original,
unmodified relationship, i.e., that the channel that formed
the WGB had precipitous banks, up to 20 m high. Locally,
the channel wall is terraced (Fig. 5A), with narrow
(b 10 m), flat shelves generally on the upper slope. Where
commercial excavations have exposed the floor of the
channel, such as at Gordon, it is an irregular, though
generally flat-lying surface.
3.3. Internal facies characteristics
The WGB comprise a heterogeneous assemblage of
generally poorly-consolidated gravels, sandy gravels and
sands, with local carbonate-cemented conglomerates and
minor, pedogenically-modified silts mainly near the top of
the channel fill (Fig. 6). Other than in the occurrence of silt
at the top of the section, the WGB display no obvious
fining-upward trend. The gravel-sized clasts (b 2 m in
long axis diameter) are composed principally of carbon-
ate-cemented siltstone and sandstone eroded from the
Cenozoic host rocks (principally Anderson Ranch
Formation), younger upper Arikaree Group and basal
Ogallala Runningwater Formation (Fig. 2; Cook, 1965;
Skinner and Johnson, 1984; Bailey, 2001). Locally,
granules and pebbles of more exotic, igneous and
metamorphic lithologies are found, and sand grade facies
comprise a mixture of intrabasinal lithic grains and
monomineralic quartz and feldspar. Locally, clasts of
cemented gravel were found within the gravels them-
selves, suggesting some recycling of intraformational
debris. A variety of fossil fragments were found within the
channel deposits, notably vertebrate teeth, but including
jaw and other bone material of a variety of mammals, fish
debris, turtle carapace fragments, and petrified wood.
In many places, the contact with the host siltstone and
fine grained sandstones is a smooth, polished surface.
Elsewhere, angular to rounded blocks and larger enclaves
(up to 10 10 2 m, at Pine Springs: Fig. 7) of the host
material were found near the foot of the channel wall,
encased or surrounded by channel fill gravels and sands.
3.3.1. Facies 1 crudely stratified and imbricated
pebble to boulder gravel
Facies 1 comprises wedges and sheets of moderately to
well-sorted, well-rounded, clast-supported, pebble to
 C.R. Fielding et al. / Sedimentary Geology 202 (2007) 5871 63

boulder gravels (Fig. 6A). Locally, Facies 1 is the dominant
constituent, whereas where the lithosome is more
heterogeneous, Facies 1 typically occurs at the base of
the section (for example, at the Gordon pits: Fig. 8). Clasts
are mainly oblate to triaxial (bladed) in form with lesser
equiaxial and prolate (rod-like) shapes (Zingg, 1935), and
are composed of intrabasinal, carbonate-cemented silt-
stone and sandstone. Gravels are typically clast-supported
but most contain a subordinate, well-sorted medium-
grained sand matrix. Maximum Particle Size (MPS) was
calculated as the mean of the 20 largest clasts visible at a
given exposure, and varies from 46 cm at Pine Springs to
41 cm at Collins' pits. There appears to be no consistent
alongstream trend evident from MPS data (Fig. 4).
Facies 1 typically displays little or no apparent
stratification, other than crude bedding evident locally
from clast long-axis alignment or particle size differen-
tiation (Fig. 6A). The most common sedimentary
structures are b-axis (and less commonly, a-axis) clast
imbrication, particularly of the larger oblate and bladed
clasts, and cluster bedform remnants (Brayshaw, 1984)
are common around the largest clasts. No systematic or
persistent lateral or vertical variations in grain-size were
noted within Facies 1. Palaeocurrent data derived from
imbrication of large clasts indicate a generally east to
eastsoutheast palaeoflow direction with little dispersion
about the mean direction, being typically parallel or sub-
parallel to the trend of the adjacent channel margin.
3.3.2. Facies 2 flat-stratified gravel, sandy gravel
and interstratified sand and gravel sheets
Facies 2 is locally abundant (for example, in the Pine
Springs Pits: Fig. 7B) but is generally a subordinate
component of the WGB. It comprises sheets of variable
texture and fabric, often interstratified with Facies 1 and 3.
Gravel in Facies 2 is grossly similar to that in Facies 1, but
is typically somewhat finer-grained, better-sorted, and is
stratified (Figs. 6B, 7B). Stratification is typically defined
by clast size variations, clast elongation, and by variations
in sand content. The principal stratification types
preserved are a crude flat or low-angle planar stratifica-
tion. Some cluster bedforms were noted, and clast

Fig. 4. Map showing the mapped course of the Whiteclay Gravel Beds (WGB) fluvial system (grey line) and the surface trace of the Whiteclay Fault
Zone (black line, ticks point to downthrown block). Note the generally eastsoutheast trend of the channel belt, with numerous straight reaches
terminated by abrupt bends. MPS refers to Maximum Particle Size data, calculated for each locality by taking the mean of the 20 largest visible
clasts. Circular histogram shows all palaeocurrent data collected from the WGB in this study.
64 C.R. Fielding et al. / Sedimentary Geology 202 (2007) 5871

Fig. 5. Photos to illustrate channel margin features of the Whiteclay Gravel Beds. A) Exhumed southern wall of the channel belt at the Gordon Pit 1, showing a
steep, possibly terraced bank with pods of gravel in contact with the host strata. Geologist for scale 1.8 m. B) Close-up of small-scale fault offsets in the host
rock lithology at Pine Springs Ranch quarry, suggesting tectonic activity postdating emplacement of the fluvial gravels. C) View of the southeastern wall of the
Pine springs Ranch quarry (Fig. 7) facing southeast, showing the precipitous nature of the channel belt wall. Face is c. 10 m high. D) View of the same corner
as is shown at the left end of C), from the opposite direction, showing how the bank of the channel belt has been overturned locally by neotectonic activity.

imbrication is evident among the more gravel-rich
intervals. Stacks of normally graded beds each c. 0.3 m
thick were noted in this facies at some localities.
Furthermore, some zones preserve an open-framework
gravel fabric. Palaeocurrent data indicate similar patterns
to those described for Facies 1.
A local variant of Facies 2 comprises poorly sorted
fine-to coarse-grained sand with dispersed finer-grained
gravel, locally in sufficient concentration to form a clast-
supported fabric but more commonly matrix-supported
(Fig. 6B). Clasts in this variant are mostly small pebble
grade, and often angular to subangular rather than the
more typical rounded shapes. This sub-facies displays
similar flat stratification and cross-stratification to other
variants of Facies 2, into which it is in many places
laterally and vertically transitional.
3.3.3. Facies 3 cross-stratified, gravel, sandy gravel
and interstratified sand and gravel channel fills
Facies 3 comprises intervals of large-scale trough
cross-stratified gravel, sandy gravel and interstratified
gravel/sand that appear to define simple channel fills
2050 m wide and 24 m deep (e.g., Figs. 6C, D, 8).
Internally, this facies contains trough-form bedding
surfaces concordant with the basal erosion surface and
scaled with the overall channel form, comprising sets of
trough cross-bedding on various scales with some zones
of open-framework gravel. Palaeocurrent relationships
are similar to the other facies.
3.3.4. Facies 4 trough cross-stratified sand
Facies 4 is overall a subordinate component of the
WGB, and is confined to wedges, sheets and channel
fills that are typically interstratified with other facies
(e.g., Figs. 6C, 8). It comprises well-sorted, brown-
coloured, typically medium-grained sand locally with
dispersed small gravel. Trough cross-stratification is
ubiquitous with sets ranging from 0.10.5 m in
thickness, and flat stratification is also common.
Dispersed angular to subrounded, clay and silt clasts
(up to 72 cm), cemented gravel clasts, and vertebrate
bone debris are locally abundant. Palaeocurrent data
 C.R. Fielding et al. / Sedimentary Geology 202 (2007) 5871 65

show similar relationships to those derived from Facies
1 and 2.
3.3.5. Facies 5 pedogenically modified silt, and
interstratified sandsilt
Facies 5 is typically found at the top of the channel
fill (e.g., Figs. 6C, 8), and comprises one or more
interval of silt or interstratified sandsilt with pedogenic
fabrics and features (peds, blocky or prismatic structure,
rhizotubules, colour mottling). Complicating this is the
common overprinting by modern soils, with extant plant
roots penetrating through the interval where present.
3.4. Facies Interpretation
The channel-confined character of the WGB as a
whole, the coarse-grained, unbioturbated facies assem-
blage dominated by unidirectional flow structures, and
the abundance of fossil debris of land mammals and
aquatic organisms (turtles and fish), together strongly
indicate a fluvial environment of deposition. Palaeocur-
rent data indicate an eastward direction of water and
sediment dispersal, as do changes in elevation data from
the eastern exposures (Fig. 4). The strong coincidence in
location and trend between the WGB and the Whiteclay
Fault Zone suggest a causal relationship, and the trace of
the channel course (extensive straight reaches with
abrupt bends) indicates a strong structural control on
channel geometry and orientation. The low dispersion
about the mean for palaeocurrent data sets at individual
localities suggests that channels were more or less
straight, perhaps wandering somewhat between steep-
sided canyon walls. Imbrication data, in particular, are
generally regarded as reliable indicators of palaeoflow

Fig. 6. Photos of the principal Whiteclay Gravel Beds lithofacies. A) Facies 1 at Pine springs Ranch quarry, showing crude stratification and local
imbrication of large clasts. Cliff is c. 4 m high. Palaeoflow was from right to left. B) Facies 2 at Reeves' property, showing well-developed flat
stratification defined by variations in texture and fabric, with imbrication of some larger clasts (palaeoflow was from right to left). Note also the
abundance of more matrix-rich, small, angular pebble gravel. C) Alternations of Facies 3 (channel-form gravels) and Facies 4 (cross-stratified sands)
at Gordon Pit 1. Palaeoflow was from left to right. Geologist for scale 1.8 m. D) Close-up of contact visible in centre-left of C), showing deeply
scoured base of Facies 3 and imbrication of the larger clasts. Notebook for scale 19 cm. Palaeoflow was from left to right.
66 C.R. Fielding et al. / Sedimentary Geology 202 (2007) 5871

direction (Bluck, 1974; Rust, 1975), and in the WGB
show very close alignment with local channel bank
orientations.
The coarse grain-size and rounding of the channel fill
sediments suggest high energy discharge, which would
have been constrained by the high, steep banks. The well-
developed grain size segregation among Facies 13, and
the scarcity of mixed-grain-size lithologies (diamicts,
matrix-supported gravels) suggests that in general, flows
carried relatively low sediment concentrations, and were
able to sort effectively the wide range of grain sizes
delivered into the Whiteclay river. The matrix-supported
variant of Facies 2, on the other hand, seems to indicate
that some flows carried higher sediment concentrations
that were at times deposited en masse, preventing
effective segregation of grain-size populations into well-
developed stratification. The abundance of fossil debris of
aquatic organisms, the lack of any indicators of semi-arid
or arid conditions (pedogenic features such as rhizotu-
bules, calcrete development, strong horizonation, etc.)
seems to suggest that the WGB accumulated under
conditions of relatively constant water supply.
The majority of coarse debris in the WGB, including
all of the coarsest material, is derived from formations

Fig. 7. A) Map showing exhumed course of Whiteclay Gravel Beds channel belt at the Pine Springs Ranch quarry. Note the abrupt change in direction
at the eastern end of the nearly straight excavation, and the location of a raft of the host lithology in the floor of the quarry that is surrounded on all
sides by gravel of the Whiteclay Gravel Beds. This block is interpreted as a raft of the host material that was either derived from bank collapse, or
(more likely) undermined from the channel floor. Note also the low dispersion in palaeoflow direction indicators, and parallelism with the channel belt
walls over the entire extent of the exposure. B) View of the southeastern wall, facing southeast (see A for the location), showing alternation between
Facies 1 (crudely stratified, coarse gravels) and Facies 2 (flat-stratified gravels and sandy gravels) in the face, and the precipitous channel belt wall
exposed by the excavation. Face is c. 10 m high. C) Interpreted raft of siltstone believed to have been undermined from the channel floor and entrained
within the flow of the formative channel. Geologists for scale 1.8 m.
 C.R. Fielding et al. / Sedimentary Geology 202 (2007) 5871 67

exposed in the channel walls, suggesting that much of the
sediment load carried by formative channels was derived
from channel incision, collapse and undermining of
channel banks, and from reworking of channel sedi-
ments. However, the subordinate component of small
gravel derived from crystalline rocks indicates that the
stream system was somehow connected to a hinterland of
exposed crystalline basement rocks.
The complex internal architecture of the WGB (e.g.,
Fig. 8) indicates that the formative channel was not cut
and filled in a single event, but rather was active over a
protracted period and filled in a composite fashion. This
impression is reinforced by the fact that the WGB
lithosome is not a single architectural form (i.e., one CH
element using the terminology of Miall, 1985), but
rather is a mosaic of numerous elements, including
channels (CH), gravel sheets (GB), sandy bedforms
(SB) and minor matrix-supported gravels (SG: all codes
from Miall, 1985).
Facies 1 is interpreted as arising principally from the
movement of diffuse gravel sheets and low amplitude
macroforms along the floor of a low-sinuosity, perhaps
braided channel (Smith, 1974; Hein and Walker, 1977;
and see review by Bridge, 2003). Facies 2 is interpreted
in a similar way, but in cases where the sediment
population at a given time and point on the channel
contained significant proportions of both gravel and
sand. The less well-sorted, angular gravel-rich variant of
Facies 2 is interpreted to record periods when the
sediment concentration in flows was unusually high,
perhaps following a nearby bank collapse or some other
abrupt introduction of large volumes of new material.
Clast imbrication and cluster bedforms are common
features of modern gravelly, low-sinuosity rivers (e.g.,
Smith, 1974; Rust, 1975; Church and Gilbert, 1975;
Lunt et al., 2004). Both Facies 1 and 2 may be remnants
of low-amplitude unit bars of various kinds (cf. Smith,
1974; Hein and Walker, 1977; Bluck, 1979), the quality
and extent of exposure not being sufficient to allow
confident diagnosis of channel/barform relationships
(cf. Lunt et al., 2004).
Facies 3 is interpreted as the fills of thalweg channels
that were the principal axes of water and sediment
transport in the Whiteclay river system. From their

Fig. 8. Photomosaic and interpretive line drawing of the northwest face of the Gordon Pit 1, showing facies relationships and palaeocurrent data. Note
that the left end of the photomosaic shows the exposure at a greater distance, owing to difficulties in acquiring orthogonal images. Palaeoflow was
from left to right. Geologist for scale (thrice) 1.8 m.
68 C.R. Fielding et al. / Sedimentary Geology 202 (2007) 5871
dimensions, channels were of the order of 24 m deep
and 2050 m wide, and evidently migrated over the
limited width of the channel floor environment (100
300 m). The internal architecture of the channels, with
concentric bedding surfaces concordant with the basal
erosion surface, is similar to the interpreted braided
channel confluence pool facies documented from the
Rhine Valley Pleistocene gravels by Siegenthaler and
Huggenberger (1993), and may be examples of the
nested channel cuts described from Cretaceous strata
by Holbrook (2001). At a smaller scale, the cross-
bedding is indicative of sediment movement in channel
floor dunes of varying dimensions, and the zones of
open-framework gravel may be interpreted as resulting
from leeside separation of gravel from sand in the front
of some bedforms (cf. Lunt and Bridge, 2005).
The wedges and sheets of well-sorted, medium-
grained sand that comprise Facies 4 are somewhat
anomalous in the context of Facies 13, but nonetheless
are commonly found interfingering with the gravelly
channel fills of Facies 3. Given this, they are interpreted
as channel fill deposits formed at times of minimal
gravel supply, and may reflect times of decreasing water
and sediment dispersal through the system. The
abundance of mud clasts within the sands indicates
that the process of channel widening by bank collapse
was still active during these periods, and that such soft
clasts were perhaps most likely to be preserved when
transported with sand rather than with gravel. Sand
moved along the channel floor mainly as small, three-
dimensional dunes at these times, possibly in a manner
similar to some modern braided streams in the region
(e.g., Smith, 1971; Blodgett and Stanley, 1980; Bridge
et al., 1986).
Facies 5 (pedogenically modified silts and fine-
grained sands), where evident, is always at the top of the
channel fill and forms a ragged blanket over the coarser-
grained facies described above. Accordingly, Facies 5 is
interpreted as deposits of shallow, gentle flows during
the final filling and abandonment of the WGB river
system, and soil formation following abandonment and
stabilisation of the alluvial surface.
3.5. Facies architecture
The available data indicate that the WGB comprise a
series of discrete channel fills and other architectural
elements of the order of metres thick and up to a few
tens of metres wide, contained (or nested) within a
larger-scale erosion surface that is up to 20 m deep and
up to 300 m wide. Although the quality and extent of
exposure do not allow full disclosure of architectural
relationships, the dataset suggests that the WGB
lithosome is a small valley fill that is one channel-belt
in width and composed of stacked, smaller-scale
channel fills. The WGB lithosome satisfies the criteria
set out by Fielding and Gibling (2005) and Fielding (in
press) for recognition as a valley fill, namely 1. the
master erosion surface is at a scale that is out of context
with enclosing fluvial strata, 2. the master erosion
surface can be traced as a discrete entity over a
considerable distance, 3. the depth of incision recorded
by the master erosion surface is several times the depth
of scour (thickness) recorded by component channel
fills, and 4. the fill enclosed by the master erosion
surface comprises numerous smaller-scale channel
bodies. Regrettably, the nature of the exposure does
not allow the delineation of bounding surface relation-
ships at the level of detail necessary to further evaluate
architectural relationships (cf. Holbrook, 2001).
4. Discussion
The single channel thread, straight course with abrupt
bends, confinement within steep banks and coincidence
with the Whiteclay Fault Zone all strongly point towards
a structural control on the WGB fluvial system. The
orientation and location of the channel system southeast
of the Black Hills Dome of South Dakota suggest that it
formed via reactivation of Black Hills uplift stresses in
Early Miocene times. The character of the channel belt
fill indicates that the formative channel was a moderate-
sized stream that wandered to a limited degree across a
narrow belt 100300 m wide, and was competent to
carry sediment up to coarse gravel grade in mainly clear
water flows. All this is suggestive of a robust discharge
regime including substantial peak flows on a regular
basis. This is out of context within the Ogallala Group as
a whole, which preserves extensive low-gradient, large
meandering river systems (Runningwater and Valentine
Formations, Fig. 2). Although it is possible that other, as
yet undiscovered gravelly channel fills are preserved
within the Ogallala Group of western Nebraska, it is
nonetheless clear that the WGB are anomalous in
indicating such an energetic flow regime. The WGB is
interpreted as recording an otherwise unknown pluvial
period within the Lower Miocene that is associated with
a previously unrecognized fossil mammal assemblage,
presumably attracted into the area by the reliable supply
of water and food. The fauna, which includes a large
variety of probable reworked material, is nonetheless
distinctive because of the occurrences of two new taxa, a
primitive species of the oreodont Brachycrus and a
merychippine horse. These, together with merycodonts,
 C.R. Fielding et al. / Sedimentary Geology 202 (2007) 5871
suggest correlation with the Hemmingfordian North
American Land Mammal Age, and allow an age
estimate of 1717.5 Ma for the WGB (Tedford et al.,
2004).
At the Gordon Pit 1, at least two cycles of channel fill
gravel (Facies 3) overlain by cross-bedded sand (Facies
4) are evident (possibly three: Fig. 8), suggesting some
repetitive process of channel establishment and partial to
complete infilling. These crude fining upward cycles
could potentially be climatically-driven oscillations,
with initial channel incision caused by an increase in
annual precipitation, runoff and consequent vertical
erosion, gravel deposition reflecting a period of
abundant coarse sediment supply from channel widen-
ing and bank collapse, and subsequent sand deposition
during the latter stages of channel filling. The final stage
in the filling of the WGB channel belt is the formation
and pedogenic modification of silts and fine-grained
sands, perhaps reflecting a return to the more semi-arid
climate regarded as characteristic of the Lower Miocene
in this region.
There has been considerable interest in the char-
acteristics of alluvial channel bodies in structurally
controlled settings (e.g., Schumm et al., 2002). Model-
ling studies have been carried out to attempt to simulate
alluvial stacking patterns in two and three dimensions
(Allen, 1978, 1979; Leeder, 1978; Bridge and Leeder,
1979; Bridge and Mackey, 1993). Documented exam-
ples of modern structurally-controlled drainage include
rivers confined by grabens and half-grabens, and
narrower, straight, single thread channels along fault
trends (e.g., Alexander and Leeder, 1987, 1990; Peakall,
1998; Peakall et al., 2000; Schumm et al., 2002). Cases
from the rock record that have been interpreted as
structurally controlled drainage rely generally on
indirect evidence, such as the stacking patterns predicted
by modelling studies together with some facies
indicators (e.g., Fielding, 1984; Alexander, 1986;
Fielding and Johnson, 1987; Leeder et al., 1996). The
value of the WGB case is that it is an ancient example of
a clearly structurally controlled fluvial system that is
fully exposed at the surface, allowing it to be mapped in
three dimensions. The WGB system is unlike most
described examples, being evidently facilitated by linear
fault rupture topography, rather than by ground tilting
effects (e.g., Peakall et al., 2000; Schumm et al., 2002).
5. Conclusions
The newly recognized Whiteclay Gravel Beds
(WGB) of the Ogallala Group in northwest Nebraska,
USA, constitute a previously unknown window into
69
climate and tectonic activity in the Lower Miocene of
the High Plains of central-western North America. The
WGB were deposited in a narrow (100300 m wide),
linear tract bounded by steep-sided walls and inset into
older Cenozoic strata, that can be traced over at least
30 km at the present land surface as a single channel
belt. This belt has a distinctive planform, with long
straight reaches ending in abrupt corners, and is
coincident at least in part with the Whiteclay Fault
zone. The channel is therefore interpreted to have
followed fault rupture topography that propagated in an
eastsoutheastward direction away from the Black Hills
Dome of South Dakota at about 17.5 Ma. This is some
considerable time after the generally-regarded date of
uplift of the Black Hills Dome, and indicates a degree of
tectonic rejuvenation in the region.
The WGB comprise a composite gravel-dominated,
incised valley fill up to 20 m thick, confined within steep
walls. Five facies were recognized, and interpreted as a
variety of mainly low-amplitude gravel and sand sheets
in channel floor environments. Component channels
were 2050 m wide and 24 m deep, and probably
braided across the narrow channel belt. Gravel channel
fills and sheets are typically overlain by well-sorted sand
bodies, and this vertical succession is repeated in some
exposures, implying a repetitive pattern of multiple
channel incision and infilling over time, perhaps related
to cyclical variations in runoff or to pulses of tectonic
activity.
The diverse vertebrate fauna found in the Gordon pits
includes local first occurrences of merychippine horses,
merycodonts, and the oreodont Brachycrus, and indi-
cates the WGB were deposited near the boundary
between the early and late Hemingfordian North
American Land Mammal Age, c. 17.5 Ma (Tedford
et al., 2004), suggesting movement along the Whiteclay
Fault Zone at this time. The unit provides a window into
a pluvial period in the Early Miocene that is unrecorded
elsewhere in the region.
Acknowledgements
We thank the staff of the Nebraska National Forest
for support, cooperation, and the use of office facilities
during this research. Field work was funded in part by
the University of Nebraska State Museum (UNSM)
Highway Salvage Paleontology Program, the Nebraska
National Forest from 19911995, and the University of
NebraskaLincoln Conservation and Survey Division
STATEMAP projects for 19962005. This work was
conducted under Special-Use Permits from the
Nebraska National Forest to Bruce E. Bailey of the
70 C.R. Fielding et al. / Sedimentary Geology 202 (2007) 5871
University of Nebraska State Museum (User No. 1044-
1), Hannan E. LaGarry (User No. 2033-4), and Leigh
Anne LaGarry (User No. 2033-6) of the University of
NebraskaLincoln Conservation and Survey Division.
M.R. Voorhies is thanked for confirming the biochro-
nologic age determination. R. Hillier and J. Holbrook
are thanked for constructive reviews of the submitted
manuscript.
References
Alexander, J., 1986. Idealised flow models to predict alluvial
sandstone body distribution in the middle Jurassic Yorkshire
Basin. Mar. Pet. Geol. 3, 298305.
Alexander, J., Leeder, M.R., 1987. Active tectonic control of alluvial
architecture. In: Ethridge, F.G., Flores, R.M., Harvey, M.D. (Eds.),
Recent Developments in Fluvial Sedimentology. Spec. Publ. Soc.
Econ. Paleont. Miner., vol. 39, pp. 243252.
Alexander, J., Leeder, M.R., 1990. Geomorphology and surface tilting
in an active extensional basin, SW Montana, USA. J. Geol. Soc.
(Lond.) 147, 461467.
Allen, J.R.L., 1978. Studies in fluviatile sedimentation: an exploratory
quantitative model for architecture of avulsion-controlled alluvial
suites. Sediment. Geol. 21, 129147.
Allen, J.R.L., 1979. Studies in fluviatile sedimentation: an elementary
geometrical model for the connectedness of avulsion-related
channel sand bodies. Sediment. Geol. 24, 253267.
Bailey, B.E., 2001. Miocene stratigraphy of the Niobrara River valley,
Western Cherry County, Nebraska. Unpublished MS thesis,
University of Nebraska, Lincoln, 92p.
Blodgett, R.H., Stanley, K.O., 1980. Stratification, bedforms and
discharge relations of the Platte River system, Nebraska. J. Sediment.
Petrol. 50, 139148.
Bluck, B.J., 1974. Structure and directional properties of some val-
ley sandur deposits in southern Iceland. Sedimentology 21,
533544.
Bluck, B.J., 1979. Structure of coarse grained braided stream alluvium.
Trans. R. Soc. Edinb. 70, 181221.
Brayshaw, A.C., 1984. Characteristics and origin of cluster bedforms
in coarse-grained alluvial channels. In: Koster, E.H., Steel, R.
(Eds.), Sedimentology of Gravels and Conglomerates. Can. Soc.
Petrol. Geol. Mem., vol. 10, pp. 7785.
Bridge, J.S., 2003. Rivers and Floodplains: Forms, Processes and
Sedimentary Record. Blackwell Science, Oxford. 491pp.
Bridge, J.S., Leeder, M.R., 1979. A simulation model of alluvial
stratigraphy. Sedimentology 26, 617644.
Bridge, J.S., Mackey, S.D., 1993. A revised alluvial stratigraphy model.
In: Marzo, M., Puigdefabregas, C. (Eds.), Alluvial Sedimentation.
Spec. Publ. Int. Ass. Sediment., vol. 17, pp. 319337.
Bridge, J.S., Smith, N.D., Trent, F., Gabel, S.L., Bernstein, P., 1986.
Sedimentology and morphology of a low-sinuosity river: Calamus
River, Nebraska Sand Hills. Sedimentology 33, 851870.
Church, M., Gilbert, R., 1975. Proglacial fluvial and lacustrine
environments. In: Jopling, A.V., MacDonald, B.C. (Eds.), Glacio-
fluvial and Glaciolacustrine Sedimentation. Spec. Publ. Soc. Econ.
Paleont. Miner., vol. 23, pp. 22100.
Cook, H.J., 1965. Runningwater Formation, middle Miocene of
Nebraska. Am. Mus. Novit. 2227, 18.
Clark, J., 1975. Controls of sedimentation and provenance of sediments
in the Oligocene of the central Rocky Mountains. In: Curtis, B.F.
(Ed.), Cenozoic history of the southern Rocky Mountains. Geol. Soc.
Am. Mem., vol. 144, pp. 95117.
Fielding, C.R., 1984. A coal depositional model for the Durham Coal
Measures of NE England. J. Geol. Soc. (Lond.) 141, 919931.
Fielding, C.R., in press. Sedimentology and stratigraphy of large river
deposits: recognition in the ancient record, and distinction from
incised valley fills. In: Gupta, A. (Ed.), Large Rivers:
Geomorphology and Management. Wiley Interscience, Chichester.
Fielding, C.R., Gibling, M.R., 2005. Distinguishing between channel
and valley fills: definitions, diagnostic criteria and dimensional
data. 8th Int. Conf. on Fluvial Sedimentology, Delft, August 712,
2005, p. 101 (Abstract volume).
Fielding, C.R., Johnson, G.A.L., 1987. Sedimentary structures
associated with extensional fault movement from the Westphalian
of NE England. In: Coward, M.P., Dewey, J.F., Hancock, P.L.
(Eds.), Continental Extensional Tectonics. Geol. Soc. Spec. Publ.,
vol. 28, pp. 511516.
Hein, F.J., Walker, R.G., 1977. Bar evolution and development of
stratification in the gravelly, braided, Kicking Horse River, British
Columbia. Can. J. Earth Sci. 14, 562570.
Heller, P.L., Angevine, C.L., Winslow, N.S., Paola, C., 1988. Two-
phase model of foreland-basin sequences. Geology 16, 501504.
Holbrook, J., 2001. Origin, genetic relationships, and stratigraphy over
the continuum of fluvial channel-form bounding surfaces: an
illustration from middle Cretaceous strata, southeastern Colorado.
Sediment. Geol. 144, 179222.
Hunt, R.M., 2002. New amphicyonid carnivorans (Mammalia,
Daphoeninae) from the Early Miocene of southeastern Wyoming.
Am. Mus. Novit. 3385.
Leeder, M.R., 1978. A quantitative stratigraphic model for alluvium
with special reference to channel deposit density and interconnec-
tedness. In: Miall, A.D. (Ed.), Fluvial Sedimentology. Can. Soc.
Petrol. Geol. Mem., vol. 5, pp. 587596.
Leeder, M.R., Mack, G.H., Peakall, J., Salyards, S.L., 1996. First
quantitative test of alluvial stratigraphic models: southern Rio
Grande rift, New Mexico. Geology 24, 8790.
Lisenbee, A.L., DeWitt, E., 1993. Laramide evolution of the Black
Hills Uplift. In: Snoke, A.W., Steidtmann, J.R., Roberts, S.M.
(Eds.), Geology of Wyoming. Geol. Surv. Wyoming Mem., vol. 5,
pp. 374412.
Lunt, I.A., Bridge, J.S., 2005. Formation and preservation of open-
framework gravel strata. 8th International Conference on Fluvial
Sedimentology, Delft, August 712, 2005, Abstracts Volume,
Delft University of Technology, p. 180.
Lunt, I.A., Bridge, J.S., Tye, R.S., 2004. A quantitative, three-
dimensional depositional model of gravelly braided rivers.
Sedimentology 51, 377414.
Miall, A.D., 1985. Architectural-element analysis: a new method of facies
analysis applied to fluvial deposits. Earth-Sci. Rev. 22, 261308.
Peakall, J., 1998. Axial river evolution in response to half-graben
faulting: Carson River, Nevada. J. Sediment. Res. 68, 788799.
Peakall, J., Leeder, M.R., Best, J.L., Ashworth, P., 2000. River
response to lateral ground tilting: a synthesis and some implica-
tions for the modeling of alluvial architecture in extensional basins.
Basin Res. 12, 413424.
Retallack, G.J., 1983. Late Eocene and Oligocene paleosols from
Badlands National Park, South Dakota. Geol. Soc. Am. Spec. Pap.
193 (82 p.).
Rust, B.R., 1975. Fabric and structure in glaciofluvial gravels. In:
Jopling, A.V., MacDonald, B.C. (Eds.), Glaciofluvial and Glacio-
lacustrine Sedimentation. Spec. Publ. Soc. Econ. Paleont. Miner.,
vol. 23, pp. 238248.
 C.R. Fielding et al. / Sedimentary Geology 202 (2007) 5871
Schumm, S.A., Dumont, J.F., Holbrook, J.M., 2002. Active Tectonics and
Alluvial Rivers. Cambridge University Press, Cambridge. 276 pp.
Siegenthaler, C., Huggenberger, P., 1993. Pleistocene Rhine gravel:
deposits of a braided river system with dominant pool preservation.
In: Best, J.L., Bristow, C.L. (Eds.), Braided Rivers. Geol. Soc.
Spec. Publ., vol.75, pp. 147162.
Skinner, M.F., Johnson, F.W., 1984. Tertiary stratigraphy and the Frick
collection of fossil vertebrates from north-central Nebraska. Bull.
Am. Mus. Nat. Hist. 78, 215368.
Smith, N.D., 1971. Transverse bars and braiding in the lower Platte
River, Nebraska. Geol. Soc. Amer. Bull. 82, 34073420.
Smith, N.D., 1974. Sedimentology and bar formation in the upper Kicking
Horse River, a braided outwash stream. J. Geol. 82, 205223.
Souders, V.L., 1981. Geology and groundwater supplies of southern
Dawes and northern Sheridan Counties, Nebraska. Nebraska
Conservation and Survey Division. IANR Open File Report,
vol. 29. University of Nebraska, Lincoln. 125 p.
Stanley, K.O., Benson, L.V., 1979. Early diagenesis of High Plains
Tertiary vitric and arkosic sandstone, Wyoming and Nebraska. In:
Scholle, P.A., Schluger, P.R. (Eds.), Aspects of Diagenesis. Soc.
Econ. Paleont. Miner. Spec. Publ., vol. 26, pp. 401423.
71
Swinehart, J.B., Souders, V.L., DeGraw, H.M., Diffendal Jr., R.F.,
1985. Cenozoic paleogeography of western Nebraska. In:
Flores, R.M., Kaplan, S.S. (Eds.), Rocky Mountain Section,
Soc. Econ. Paleont. Miner. Cenozoic Paleogeography of west-
central United States: Rocky Mountain Paleogeography Sym-
posium, vol. 3, pp. 209229.
Tedford, R.H., Albright III, L.B., Barnosky, A.D., Ferrusquia-Villafranca,
I., Hunt Jr., R.M., Storer, J.E., Swisher III, C.C., Voorhies, M.R.,
Webb, S.D., Whistler, D.P., 2004. Mammalian biochronology of the
Arikareean through Hemphillian interval (late Oligocene through
early Pliocene epochs). In: Woodburne, M.O. (Ed.), Late Cretaceous
and Cenozoic mammals of North America. Columbia University
Press, pp. 169231.
Terry, D.O., 1998. Lithostratigraphic revision and correlation of the lower
part of the White River Group. In: Terry, D.O., LaGarry, H.E., Hunt,
R.M. (Eds.), Depositional environments, lithostratigraphy, and bio-
stratigraphy of the White River and Arikaree Groups (late Eocene to
Early Miocene, North America). Geol. Soc. Am. Spec. Pap., vol. 325,
pp. 1438.
Zingg, T., 1935. Beitrag zr Schotteranalyse. Schweiz. Mineral.
Petrogr. Mitt. Bd. 15, 39140.
 Available online at www.sciencedirect.com

Sedimentary Geology 202 (2007) 72 95

www.elsevier.com/locate/sedgeo

Tectonic versus climatic control on the evolution of fluvio-deltaic

systems in a lake basin, Eastern Pannonian Basin
Gy. Juhsz a,, Gy. Pogcss b , I. Magyar c , G. Vakarcs c
a
Department of Basin Analysis, Geological Institute of Hungary, Budapest, Stefnia st.14, H-1143 Hungary
b
Department of Applied and Environmental Geology, Etvs Lrnd University, Budapest, Pzmny P. stny 5, H-1117 Hungary
c
MOL Hungarian Oil and Gas Co., Budapest, Batthyny t 45, H-1039 Hungary

Abstract

The Pannonian Basin is a back-arc lake basin, situated inside the Central European Carpathian loop, surrounded by the orogenic
belts of the Eastern Alps, Carpathians and Dinarides. Integrated stratigraphical and sedimentological research reported here focuses
on palaeogeographical evolution, and particularly the response of the fluvio-deltaic systems to climatic and tectonic controls in the
Late-Neogene sedimentary succession, in the eastern part of the Pannonian Basin. This work has been carried out using subsurface
data including regional composite seismic profiles and well-logs. Dip sections were studied to define the stratigraphical
architecture, depositional environments and sequence stratigraphy, and to interpret the sedimentary evolution of the area.
The eastern Pannonian Basin is characterized by sediment input from the northeast which was one of the main routes of the fluvial-
deltaic systems entering and progressively filling Lake Pannon from 11.62.6 Ma. The formations studied were deposited in
environments that ranged from the deep-water basin plain, slope, delta front, and coastal plain to alluvial plain, causing the deposition of
characteristic lithofacies associations, and characteristic biofacies. These form laterally continuous lithofacies units that can be correlated
through the basin. All depositional facies units have considerable thicknesses (3001000 m) in the deepest parts of the sub-basins.
Sequence stratigraphical interpretation was carried out correlating 3rd and 4th-order sequences (Ma to ky in range) through a
regional composite seismic profile network in the dip direction. Stratigraphical architecture and sedimentary facies indicate that the
3rd-order cycles were controlled by tectonic driving forces, while 4th-order cycles were probably driven by large-scale climatic
changes within the Milankovitch band. In the 3rd-order time interval Pa-3 between 9.1 and 6.8 Ma, a major tectonostratigraphic
event occurred, highlighting the role of strike-slip faulting-related transtensional subsidence in the Derecske sub-basin,
superimposed on the regional post-rift thermal subsidence of the Eastern Pannonian region. The substantial base level fall at SB Pa-
4 at 6.8 Ma (over 200 m, perhaps amplified by subsequent tectonic movements) can be related to the onset of a new compressional
phase of structural evolution in the basin.
2007 Elsevier B.V. All rights reserved.

Keywords: Late-Neogene; Integrated stratigraphy; Fluvio-deltaic system; Sedimentary evolution; Controlling factors; Pannonian Basin

1. Introduction

Corresponding author.
The Pannonian Basin forms the westernmost Late-
E-mail addresses: juhasz@mafi.hu (Gy. Juhsz), Neogene remnant of the Paratethys. It is a back-arc basin
pogacsasgy@ludens.elte.hu (Gy. Pogcss), immagyar@mol.hu surrounded by the Carpathian arc, the Alps and the
(I. Magyar). Dinarides. After the Pannonian Basin lost its connections
0037-0738/$ - see front matter 2007 Elsevier B.V. All rights reserved.
doi:10.1016/j.sedgeo.2007.05.001
 Gy. Juhsz et al. / Sedimentary Geology 202 (2007) 7295 73

to the sea (ca. 11.5 Ma), the salinity decreased to brackish
and the basin evolved into a large lake. The post-rift Late-
MiocenePliocene sedimentary sequence referred to as
the Pannonian s.l. stage (Jmbor, 1989) forms a 2nd
order sedimentary cycle filling the central part of the
Pannonian Basin. Rivers carried large volumes of
sediments from the emerging, surrounding mountains.
Two main deltaic systems entered and filled the basin,
although there were several other minor sources scattered
along the margins. The Pannonian Basin is characterized
by a complicated history of tectonic activity and dif-
ferential subsidence, resulting in the formation of several
sub-basins.
The study area is focused on the eastern part of the
Pannonian Basin and comprises two sub-basins, the
Derecske Trough and the eastern part of the Bks Basin
including the more elevated basement highs between
and surrounding them, and their hinterlands (Figs. 1, 2).
The sediments in the study area represent a large delta
complex that prograded from the northeast and was one
of the major deltaic systems infilling Lake Pannon. The
best studied sub-basin has been the Bks Basin in

Fig. 1. Location and structural contour map at the base of the post-rift Late-Miocene sediments with the regional profiles (with roman numerals).
74 Gy. Juhsz et al. / Sedimentary Geology 202 (2007) 7295
Fig. 2. Stratigraphical chart of the studied sedimentary succession (modified after Magyar et al., 1999; Vakarcs, 1997).
which a large amount of work was undertaken in the late
1980s. It is well-documented by Mattick et al. (1994),
Molenaar et al. (1994), Phillips et al. (1994), and
Pogcss et al. (1994) in Teleki et al. (1994). Close to
the Derecske Trough a major discordance was soon
recognized, and later the significance of the Pa-4
sequence boundary was emphasised on seismic profiles
(Vakarcs and Vrnai, 1991; Vakarcs et al., 1994;
Vakarcs, 1997; Lemberkovics et al., 2005), while
PlioceneQuaternary strike-slip faults on the Hungarian
Plain were discussed at a regional scale (Pogcss et al.,
1989a,b; Rdly, 1990). The sedimentary architecture of
the Hungarian Plain, and the extreme thickening of the
deltaic depositional unit in the eastern part of it, as well
as the sequence stratigraphical features were assessed on
the basis of borehole data (Juhsz, 1992, 1993). Most of
the earlier working groups have had special areas of
interest and approach, in tectonics, stratigraphy, sedi-
mentology or palaeontology, even in basin-scale
investigations, the results of which are discussed in the
relevant sections below. This paper and its predecessor
studies represents the first time that the authors have
adopted an integrated approach, using and combining
most of the available subsurface geological and
geophysical information and interpretations (Juhsz
et al., 2004b, 2005a,b).
The aim of this paper is to compile a detailed and
refined, integrated stratigraphical framework, to develop
a better understanding of the extreme variability of the
fluvio-deltaic architecture and sedimentary evolution of
the study area, and to define the controlling factors on
the evolution of the fluvio-deltaic system of the study
area.
2. Methods and dataset
The Late-Neogene sedimentary succession in the
eastern part of the Pannonian Basin was studied by
subsurface geological methods. The stratigraphical and
sedimentological interpretations at different scales were
integrated in the basin study. The work was based on
regional composite seismic profiles, analysis of well-
logs (Serra, 1985) and core-sample data. Regional seis-
mic profiles were both oriented in the dip direction,
which highlights sediment supply routes into the basin,
and strike-oriented (Fig. 1). The regional profile of Vakarcs
(Vakarcs et al., 1994; Vakarcs, 1997) which served as the
starting point of our interpretation (profile III), follows
the axis of the sub-basins through the narrow, elongate
Derecske trough and the eastern part of the Bks basin.
However, there are limited well data along this profile,
therefore additional regional composite profiles on the
flanks of the trough and near basement highs, on both
sides, were also included in this study to improve borehole
control on the seismic data. One of these profiles, profile II
(Fig. 1), north of the master profile (profile III), follows the
 Gy. Juhsz et al. / Sedimentary Geology 202 (2007) 7295
trace of a major strike-slip fault from the transtensional
basin (Derecske trough) to the flanks of the transpressional
high. Geological information on the Late-Miocene
Pliocene sedimentary succession was provided by several
hundred exploration boreholes. Well-logs (mostly Gamma
Ray, Spontaneous Potential and Resistivity plots) were
interpreted, and core samples and cuttings were investi-
gated for sedimentological and palaeontological
information.
2.1. Problems of timing
Dating of the key surfaces has been based on
palaeomagnetic measurements from the boreholes Dva-
vnya-1, Vszt-1, Kaskanty-2, and Tiszapalkonya-I
(Elston et al., 1990; Lantos et al., 1992; Elston et al.,
1994). The first two wells were drilled in the southern, deep
basinal areas, where continuous sedimentation took place
through the Late Miocene to present. The polarity zones
were interpreted from the Brunhes Chron back to the
beginning of the Nunivak Subchron of the Gilbert Chron
close to the bottom of the boreholes (C3n2n, 4.63 Ma). The
two other boreholes, in turn, are located in marginal
positions, in the northern and in the western part of the
central Pannonian Basin, where an unconformity between
the Upper Miocene and Pliocene was observed (Jmbor,
1989; Juhsz et al., 1996). The youngest datum possibly
identified in the Tiszapalkonya boring was the beginning of
C3Ar (7.14 Ma), some 150 m below the unconformity
surface. In the Kaskanty boring, the youngest event
identified was the beginning of C3An2n (6.73 Ma), about
50 m below the unconformity. Correlation between the four
magnetostratigraphic boreholes was attempted on seismic
profiles (Pogcss et al., 1994), and the time interval
between the top of the Kaskantyu and bottom of
Dvavnya and Vszt records was distributed through
the sediment thickness (e.g. Vakarcs et al., 1994).
This dating method encompasses a high degree of
uncertainty. The youngest reliably dated volcanics in the
central Pannonian Basin sedimentary succession are as old
as 9.6 Ma (Balogh et al., 1986); interpretation of the
polarity record from younger sediments lacks geochrono-
metrical support. In the absence of more reliable dating,
however, the writers tentatively accept and adopt the
magnetostratigraphical interpretation proposed by Elston
et al. (1994), transferred to the recent geomagnetic polarity
time scale (GPTS) of Lourens et al. (2004).
3. Geological and tectonic setting
The average surface elevation of the eastern part of the
Pannonian Basin is 85120 m above sea level and the
75
ground surface shows only minor topographic variation.
The total sediment thickness of the NeogeneQuaternary
in the central Pannonian Basin (MakHd trough,
Bks, Derecske basins) amounts to 57 km.
The subsidence of the Pannonian Basin was an integral
part of the complex orogenic evolution of the European
Alpine Belt (Royden, 1988; Horvth and Tari, 1999; Tari
et al., 1999; Hmor et al., 2001). The multi-phase
NeogeneQuaternary subsidence of the Pannonian
Basin generated discrete depocentres, which filled with
sediments eroded from the surrounding orogenic moun-
tain belts. Extension within the Carpathian loop and
compression (thrusting and folding) in the surrounding
orogenic belts were related to three factors: collision of the
African/Adriatic and European plates; evolution of
subduction (roll back, hinge retreat, migration of slab
detachment) along the Carpathian arc (Nemcok et al.,
2005), and mantle upwelling (asthenospheric doming).
The pre-Neogene basement of the Pannonian Basin
consists of two different PalaeozoicMesozoic tecto-
nostratigraphic terranes, the ALCAPA (Pelso) terrane
and the Tisza terrane. Both terranes reveal a complex
history as part of the Tethyan realm: Permian continental
to oceanic rifting, Triassic to Jurassic spreading and
drifting, followed by Late CrateceousPaleogene sub-
duction and continental collision (Haas et al., 1999;
Brezsnynszky et al., 2000). Rotational contact and
welding of the ALCAPA and Tisza tectonostratigraphic
terranes took place during the Early Miocene (Balla,
1984; Horvth, 1993; Csontos, 1995; Fodor et al.,
1999). Juxtaposition of these terranes was contempora-
neous with the onset of Pannonian Basin formation.
The early syn-rift phase (2418 Ma) of the basin was
characterized by NWSE and NESW striking rifts
forming a system of long, narrow grabens connected by,
and connected to, the Carpathian thrust belt by transform
faults (Royden, 1988). Syn-rift grabens were character-
ized by fluvial depositional environments (Hmor, 1995;
Hmor et al., 2001). According to Csontos (1995), the
observed fault pattern indicates radial shortening in the
external parts of both the Tisza and ALCAPA terranes.
Fodor et al. (1999) proposed repeated dextral transpres-
sion within the ALCAPA, and associated compression
within the Tisza terrane. Thrusts and imbricated struc-
tures, indicating shortening, were identified in the
Pannonian Basin from seismic data (Pogcss et al.,
1991), under Middle Miocene and younger sediments.
Shortening events (Detzky et al., 2000) inverted the
former basins (e.g. Szolnok Flysch Trough). There was a
causal connection between syn-rift and post-rift palaeo-
geographical evolutionary stages and subduction and/or
mantle upwelling related volcanic activity. Eruption of
76 Gy. Juhsz et al. / Sedimentary Geology 202 (2007) 7295
crust-sourced acidic rhyolites was associated in space and
time with major phases of crustal stretching (Szabo et al.,
1992). Acidic magmatic centres were located along major
tensional ruptures. Major changes in stress field led to the
generation of a new set of ruptures and resulted in further
rhyolite eruptions.
The style of early syn-rift extension changed to a wide
rift style (Tari et al., 1999) between 1814 Ma. Fodor et al.
(1999) proposed two distinct events during this period: the
first characterized by a NESW to ENEWSW tensional
or strike-slip stress field, and a second event dominated by
EW to SENW tensional, strike-slip type field. Pull-
apart along transform faults led to deepening and widening
of the former narrow fluvial grabens and opening of
several new rifts and/or pull-apart basins within the
Pannonian Basin. Half grabens, bounded by listric faults
and crestal collapse grabens related to flat-ramp listric
faults (Kkai and Pogcss, 1991), represented the
depocentres. Overall Middle Miocene palaeogeography
was dominated by widespread marine facies belts
separated by areas of erosion and non-deposition.
Termination of the syn-rift phase, in late Middle Badenian
time (14 Ma), is indicated by a regional unconformity.
Large displacements along listric faults resulted in
tilting of originally nearly horizontal strata and/or
structural inversion, and led to the formation of a
regional unconformity between the syn-rift Middle
Miocene and the post-rift Upper Miocene sediments
(Juhsz et al., 1999; Hmor et al., 2001). Royden (1988)
proposed that sedimentation within each sub-basin was
influenced by the proximity of individual basins to the
thrust front of the Carpathians. These basins, located
close to the thrust front, contain thick, fault-bounded
syn-rift sedimentary section, but are overlain by a thin
post-rift section. Basins located in more internal
positions within the CarpathianPannonian region
were starved during the syn-rift phase. These differences
can be related to the differences in thermal subsidence
rate of the basement after the extension and to the
proximity of each basin to sediment sources in the
rapidly uplifted, surrounding Carpathians.
The post-rift phase includes the Late Badenian,
Sarmatian, and Pannonian s.l. times (ca. 144? Ma).
Palaeogeographical and structural conditions were
controlled by the Leithaian tectonic cycle (Hmor,
1978); from west to east, step-by-step termination of
thrusting in the Carpathians (Jiricek, 1979), and uplift of
the AlpCarpathianDinaridic realm. Fodor et al.
(1999) propose early stage basin inversion, ENE
WSW or EW compression at about 10 Ma. Compres-
sion was later followed by renewed extension, strike-
slip faulting and post-rift thermal subsidence.
In the early post-rift phase, the overall rate of creation
of accommodation exceeded the rate of sediment
supply, thus water depth increased. Maximum water
depth was reached in the Early Pannonian (10 Ma).
Clinoform patterns in seismic reflection data indicate
more than 1000 m water depth in Early Pannonian time.
In a later stage of basin evolution (Late Pannonian), the
overall rate of (post-rift thermal) subsidence was less
than the rate of sediment supply and water depth
decreased. The Late-MiocenePliocene (Pannonian s.l.)
sedimentary succession of the post-rift phase is the main
focus of both hydrocarbon and water exploration and
production.
The basin inversion phase was thought to have begun in
PlioceneQuaternary time (ca. 4 ?0 Ma) but our new
results demonstrate that inversion could have begun much
earlier (see below). The palaeogeographical and structural
evolution was controlled by the Rhodanian tectonic cycle.
Uplift of the AlpineCarpathian arc continued, and
compression also characterized the emerging inner
mountain ranges (inselbergs). Later stage basin inversion
(Horvth and Cloetingh, 1996; Gerner et al., 1999) was
characterized by NWSE to NS maximum stress axes,
strike-slip and compression in the Alpine/Dinaric junction,
and a tensional stress field and subsidence in the central
part of the Pannonian Basin (Fodor et al., 1999). Late
PliocenePleistocene NESW orientated strike-slip faults
have been identified on seismic profiles in several areas
within the Pannonian Basin (Pogcss et al., 1989a,b;
Detzky et al., 2000; Bada et al., 2000).
4. Stratigraphy and depositional facies
The Late-MiocenePliocene (Pannonian s.l.) post-
rift sediments in the study area unconformably overlie
Middle Miocene (marine Badenian or restricted marine
Sarmatian) sediments of the syn-rift phase. The abrupt
environmental changes from Sarmatian to Pannonian,
including a decline in salinity, are reflected by the
disappearance of marine biota and the diversification of
endemic mollusc, ostracod and dinoflagellate species
(Papp et al., 1985). Stratigraphical subdivision of the
Pannonian s.l. deposits is mostly based on dinoflagel-
lates from the deep lacustrine facies, molluscs in the
shallow lacustrine facies, and mammals in the fluvial
and terrestrial facies (Fig. 2). The resolution of Lake
Pannon biostratigraphy is comparable to that of coeval
marine biostratigraphical systems (Magyar et al., 1999).
Resource industries, in particular hydrocarbon and
water exploration and production, have preferred to
apply lithostratigraphy rather than biostratigraphy
because it was easier to adopt. Detailed and thorough
 Gy. Juhsz et al. / Sedimentary Geology 202 (2007) 7295 77

Fig. 3. General sedimentary succession in a well in the study area, representing the average lithology, facies and depositional environments, as well as
the corresponding lithostratigraphic units and their range of thickness. Poor core recovery is indicated by the core samples, while the majority of the
wells do not have any core-log. The facies units are time-transgressive in the study area, sedimentation was more or less continuous, therefore the
PlioceneMiocene boundary is set inside or at the base of the fluvial succession. The extremely thick shoreface to coastal plain facies unit is most
common in the Pa-3 3rd-order depositional sequence which formes the major part of the study area while in the Pa-4 sequence the facies unit thins out
abruptly (see later). Large-scale CU successions (1030 m) on the logs indicate shoreface and delta front progradation. Legend: see Fig. 4.
78 Gy. Juhsz et al. / Sedimentary Geology 202 (2007) 7295
core investigations proved that in the post-rift Panno-
nian s.l. sedimentary succession, the lithostratigraphical
units approximately correspond to the main depositional
facies units or lithofacies associations which are time-
transgressive and were deposited in specific deposition-
al environments (Brczi and Phillips, 1985; Brczi et al.,
1988; Rvsz et al., 1989; Jmbor, 1989; Juhsz, 1991,
1992, 1994; Phillips et al., 1994; Molenaar et al., 1994).
This is a direct result of the lateral continuity and
considerable thickness of the deposits throughout the
basin.
From bottom to top, the shoaling-upward sedimen-
tary succession of a deep basin can be recognized, from
deep lacustrine to extremely thick heterolithic fluvio-
deltaic facies assemblages. In the study area, the fluvio-
deltaic succession represents the major part of the entire
sedimentary infill. Moreover, in a considerable area, the
shoreface, delta front and coastal plain sediments which
form an extremely thick lithostratigraphic and lithofa-
cies unit, can reach 1000 m in thickness. Below we give
a brief description of the main lithofacies assemblages
or units, while their main features and sedimentary
Fig. 4. Characteristic core-logs and sedimentary structures of the main facies units. Legend: 1. Sand, 2. Silt, 3. Argillaceous marl, 4. Bioturbation,
5. Cross-bedding, 6. Gravel beds, 7. Concretes, 8. Deformed beds.
structures are shown in Figs. 3 and 4. As the lithological
characteristics of each unit are very different, they can
be clearly recognized on the logs. Core-logs are scarcely
available, however the analysis of sporadic core samples
of 28 m thickness individually, gave an indication of
the depositional facies and a fairly good correlation with
the well-logs (Figs. 3, 4).
The deep lacustrine basinal succession is represented
by basinal marls, deep-water turbidites and overbank
fines, fine-grained slope apron deposits from the flanks
of the trough, and offshore muds deposited on the basin
slopes (Fig. 5). The initial hemipelagic marls of a
starved basin (Endrd Formation) are overlain by deep-
water sands (D facies unit on Figs. 3, 4, 5, Szolnok
Formation) only in the deepest parts of the two studied
sub-basins. The thick turbidite succession is built up of
fine-grained sandstones and siltstones with interbedded
marls, partly infilling the deep basinal areas. They
include sediments of different types of gravity and mass
flows. The distribution of deep-water sands in the
Derecske trough indicates that channelized turbidites
flowed along the axis of the basin. Thus the sandy part
 Gy. Juhsz et al. / Sedimentary Geology 202 (2007) 7295
Fig. 5. Three-dimensional depositional model of the Late-Miocene sedimentation in the eastern part of the Hungarian Plain (modified after Juhsz, 1992).
of the turbidites can be found only in the axial zone, but
in a great thickness (near 1000 m). Moving towards the
flanks of the narrow trough sandy turbidites are not
seen, only an extremely thick muddy succession. This
suggests that the overbank fines of the channelized
turbidites, as well as fine-grained slope-apron deposits,
derived from the steep flanks, can be found in the
boreholes of the studied profiles. Towards the SW,
entering the Bks Basin, sandy turbidites are wide-
spread, suggesting that their course changed to lobate-
type fans, that interfinger with turbidite systems derived
from the NW (Juhsz, 1992; Molenaar et al., 1994).
The uppermost part of the deep lacustrine (basinal)
succession, comprising argillaceous marls and silt-
stones (C lithofacies unit, Algy Formation), is
widespread throughout the basin. It was deposited on
well-developed clinoforms, i.e. on the basin slope and
the slope of the deltaic system approaching the area
from the NE. This unit was previously interpreted as
sediments of the delta slope. The scale and dimensions
of this slope, however, are much greater than that of a
delta slope, as the height of the clinoforms varies from
100 to 700 m in the study area. The thickness of this
lithofacies unit depends on the palaeo-water depth
which implies that it was between 100 and ca. 6700 m.
According to the seismic record, the slope angle was
variable in space and time. The ongoing subsidence and
differential compaction caused the upper surface of this
lithofacies unit to often follow the basement morphol-
ogy. The silty marl succession contains 2 to 40 m thick
sandy intercalations of different origin. Evidence of
79
density current activity and significant downslope
slumping is seen in the sedimentary structures.
The slope marls are overlain by a thick, well-
developed sandy succession (B lithofacies unit, jfalu
Formation), deposited mostly in shoreface, delta front,
delta-plain and coastal plain environments. The predom-
inant lithology is rather sandy, but sands are mostly
interbedded with siltstones, claystones and marls, as well
as locally with lignite beds (Figs. 3, 4). Coarsening-
upward silty to sandy successions were deposited on the
delta front while delta-plain and coastal plain sediments
are characterized by sandy distributary channel fills and
interbedded claystone and sandstone successions as
overbank flood deposits, as well as bay and marsh
sediments. It is widespread in the study area and its
thickness varies in extremes here, from 20 m to more than
1200 m in the deepest boring, due to palaeogeographical
and tectonic changes during its deposition. The principal
fluvial palaeoflow direction was towards the SW in this
part of the basin. Due to the fact that the preservation
potential of deltaic sediments is much higher than that of
sediments deposited in other parts of the shoreline, the
bulk of this unit comprises deltaic sediments. Delta front
sediments dominate the SW portion of the delta complex,
whereas the proportion of delta-plain sediments gradually
increases towards the NE.
In the distal and central part of the wide morpholog-
ical shelf formed in this part of the basin, the thickness
of delta front sediments can reach 1000 m. They consist
of an alternation of offshore or lagoonal marls and
coarsening-upward delta front or shoreface successions.
 80
Gy. Juhsz et al. / Sedimentary Geology 202 (2007) 7295
Fig. 6. Profile II: 3rd and 4th-order sedimentary cycles in a characteristic seismic profile on the northern flanks of the Derecske trough, following the trace of a major strike-slip fault in the basin. Note
the characteristic progradational-retrogradational pattern on a seismic profile in the 3rd-order cycle between 6.89.1 Ma and the character of SB Pa-4 shown in detail below. The major transtensional
and transpressional zones are visible in the profile. Legend: 1. Third-order sequence boundary, 2. 4th-order sequence boundary, 3. Position of the shelf-edge, 4. Middle-Miocene syn-rift sediments
(suffered inversion), 5. Late-MiocenePliocene post-rift sediments.
 Gy. Juhsz et al. / Sedimentary Geology 202 (2007) 7295 81

These successions or parasequences form depositional more delta-plain and coastal plain sediments intercalated
cycles, comprising a series of thickening and coarsen- with the delta front deposits occur. Gradually up-dip to
ing-upward, as well as thinning and fining-upward sand the NE these are interdigitate with fluvial sediments of
bodies (see below). Moving towards the NE more and great thickness.

Fig. 7. Profile IV: 3rd and 4th-order sedimentary cycles on a characteristic seismic profile in the southern part of the study area, marking sediment
infill from the NE direction. Note the thickening back nature of the 4th-order cycles in the upper part of the section, both individually and as a
package, indicating structural changes. Note also the character of the 3rd-order SB Pa-4. Here the seismically transparent offshore facies is missing
above the wedge-shaped seismic facies unit with high-amplitude reflectors, probably eroded in the course of the strong relative base level fall at SB
Pa-4. The transgressive feature, however, can be detected in the stratigraphical column of the boreholes.
82 Gy. Juhsz et al. / Sedimentary Geology 202 (2007) 7295
Fluvial sediments (A lithofacies unit, Zagyva
Formation) can reach a considerable thickness both in
the NE and SW parts of the study area (a few 100 m to
1000 m), but in some intervening areas, corresponding
with the thickest deltaic deposits, they thin out
completely. The fluvial sediments are much older in
the NE than in the SW of the area. The fluvial succession
incorporates several facies, such as channel fill,
floodplain and in some cases floodbasin (wetland)
sediments. Very few core samples of this unit are
available since these sediments have not yet been the
focus of hydrocarbon exploration. Therefore our
understanding is based principally on well-logs, with
the exception of the 3 fully cored, but shallow boreholes
penetrating max. 1000 m into Quaternary and Panno-
nian sediments at the edges of our study area.
5. Sequence stratigraphical analysis
Sequence stratigraphical analysis was based mainly
on seismic profiles, and the results were cross-checked
with well-log data where available. 3rd- as well as 4th-
order cycles were found and correlated throughout the
study area. The 3rd-order sequence boundaries have
been adopted from Vakarcs et al. (1994) and Vakarcs
(1997), the timing correction follows Berggren et al.
(1995) and their main profile was considered the starting
point for this study (profile III, Fig. 1). The 4th-order
cycles were recognized independently on the studied
regional cross-sections. Analysis of the seismic profiles
and the stacking pattern of the logs made it possible to
identify the transgressive, highstand, lowstand, and in
some cases, the falling-stage systems tracts (TST, HST,
LST, and FSST, according to Posamentier and Vail,
1988; Van Wagoner et al., 1988, 1990; Hunt and Tucker,
1995; Plint and Nummedal, 2000).
In the study area three 3rd-order sequences the Pa-2,
Pa-3, and Pa-4 sequences were identified (Fig. 6). Their
sequence boundaries form major unconformities that
were easily recognized on the basis of the onlapping
features. They are all type I sequences, their common
features are the anomalously thick lowstand prograding
complex and relatively thin transgressive and highstand
deposits, thought to be caused by the extremely wide
(100200 km) shelf zone (Tari et al., 1992; Vakarcs et al.,
1994). The Pa-3 sequence lies in the central part of the
study area, and it was the focus of interest, therefore was
studied in detail. Its LST is well-developed and extremely
thick, and the transgressive surface is easily detected. The
HST, however, is very thin if present, as it was mostly
eroded during the succeeding rapid relative lake-level fall
in the last stage of its development. 3rd-order scale FSST
were identified by recognizing large-scale facies changes
on the basis of well-log stacking patterns, discussed in
detail below. Its recognition on seismic profiles was not
clear because of the high rate of erosion.
4th-order cycles were identified on the dip profiles
where progradation of the shoreface was clear. In some
4th-order cycles the resolution of the seismic image was
not sufficiently high enough to identify all the systems
tracts. Nevertheless, in several cases seismic resolution
was sufficient to recognize 4th-order FSST based on the
offlapping nature of the seismic pattern. The general
sequence stratigraphical picture displays the 3rd- and
4th-order cycles in NESW trending profile in dip
directions (Figs. 6, 7). The detail of profile IV represents
the thickening back nature of the TST and HST of the
3rd-order cycle Pa-3.
6. Integrated stratigraphy and hierarchical stacking
patterns of depositional sequences
6.1. Tectonically driven 3rd-order cycles diagnostic
features, characteristics and evidence
Integrated stratigraphic interpretation was focused on
the Pa-2, Pa-3, and Pa-4 3rd-order sequences to
establish a refined overall picture of the stratigraphic
architecture and the palaeogeographical evolution.
Although Lake Pannon did not have a real continental
shelf, a wide morphological shelf developed during
deposition of the Pa-2 and Pa-3 sequences, and
progradation displays practically a shelf-margin setting.
Instead of the terms outer-, inner shelf, applied to sea-
submerged continental shelves, the authors prefer to use
the terms distal, central and proximal for the lacustrine
settings of the Pannonian Basin.
The main diagnostic features and peculiarities
observed on the parallel dip profiles that need
explanation are as follows:
A wedge-shaped body of high-amplitude seismic
facies, surrounded by a seismically transparent, low-
amplitude facies on the parallel dip-oriented seismic
profiles.
The most spectacular feature in the study area is 3rd-
order SB Pa-4. Therefore one of the goals was to
highlight the character and facies changes through it.
An important diagnostic feature is the onlapping
nature of bottomset clinoforms towards the Szegha-
lomFldes basement high as well as the Middle
Miocene sediments from the east characterized by
folded or curved reflectors (indicated as 4th-order
sequence boundaries) in the lower part of 3rd-order
 Gy. Juhsz et al. / Sedimentary Geology 202 (2007) 7295 83

Fig. 8. Spatial distribution of the same phenomena on quasi-parallel dip profiles changing from N to S, respectively (see on map). The wedge-shaped
seismic facies of high-amplitude reflectors thins northwards, thickest in the deepest part of the Derecske trough. It is overlain by a seismically
transparent facies wedge with low-amplitude reflectors which thins NE on the individual profiles and disappears (i.e. progressively more eroded) in
the southern profile. According to sedimentological interpretation, the seismic facies of high-amplitude reflectors represents delta front/shoreface
sediments while the transparent seismic facies above represents transgressive offshore shaly sediments. Therefore it represents the diagnostic major
progradationretrogradation feature overlain by SB Pa-4. Circles mean position of shelf-edge and its change in time.
84 Gy. Juhsz et al. / Sedimentary Geology 202 (2007) 7295

Fig. 9. The expression of the wedge-shaped seismic facies of strong reflectors in the lithological column of a well, representing a major
progradationalretrogradational pattern. The sediments comprise a series of shoaling-upward shoreface/delta front facies successions that could be
termed parasequences. The individual delta front/mouth bar sand bodies thicken and coarsen upwards representing an overall progradation, then thin
and fine upwards representing an overall retrogradation. The profile is a detail of regional profile II. For location see Fig. 6. Legend: 1. fine-grained
sandstone, 2. siltstone, 3. argillaceous marl.

cycle Pa-3 while the reflectors become smoother
upwards on profile II (Figs. 6, 10).
6.1.1. The wedge-shaped high-amplitude seismic facies
and its surroundings signs of a major transgression
The wedge-shaped high-amplitude, great continuity
seismic facies thins northwards and is thickest in the
deepest part of the Derecske trough (Fig. 8). It is
overlain by a seismically transparent facies wedge with
low-amplitude reflectors (TST of 3rd-order cycle Pa-3)
which thins north-eastwards on the individual profiles
and disappears (i.e. it is progressively more eroded) in
the southern profile. Correlating with the available
boreholes (Fig. 9) it is quite obvious that the high-
amplitude seismic facies represents delta front/shoreface
sediments, while the transparent seismic facies above it
represents transgressive offshore marls. The overall
picture reflects a major progradingretrograding fluvio-
deltaic wedge, thickening back towards the alluvial
plain, and reaching an extreme (2000 m) maximum
thickness. It had been understood that the terminations
of the high-amplitude reflectors practically pinpoint the
shelf-edge of the prograding delta downwards. On the
upward side it was possible to follow the backstepping
of the shelf-edges deep inland, representing the
transgressive surface. This phenomenon is also recorded
in the borings, indicating a major transgression in the
study area far to the NE in the course of deposition of the
Pa-3 sequence. The magnitude of this transgression was
not known before as offshore sediments and coarsening-
upward delta front facies can be found even in the
northeastern part of the basin immediately under the
Quaternary sediments at 200300 m depth, covering
thick fluvial sediments.
The sharp change of accommodation to sediment sup-
ply ratio can be well observed at the 3rd-order sequence
boundaries in the studied profiles. Progradation, as well as
aggradation in the LST of the studied 3rd-order cycle Pa-3,
means that strong sediment supply could keep up with
basin subsidence, whereas during the TST and HST strong
tectonic subsidence (and possibly lower rate of sediment
supply) created space for a transgressive phase.
6.1.2. The most spectacular surface SB Pa-4 indicates
considerable base level fall
The SB Pa-4 (ca. 6.8 Ma), at the end of the 3rd-order
cycle Pa-3, is the strongest erosional and onlapping
feature in the study area. It indicates a very strong
relative base level fall of ca. 200300 m in the eastern
part of the basin, with sudden large-scale facies changes.
The significant relative lake-level fall produced an FSST
consisting of an offlapping succession of shoreface/delta
front sand body wedges. These forced regressive sand
bodies or stranded parasequences can be identified on
 Gy. Juhsz et al. / Sedimentary Geology 202 (2007) 7295 85

Fig. 10. Sedimentological and sequence stratigraphical profile referring to characteristic facies changes after the 3rd-order SB Pa-4 (6.8 Ma). The
major progradationalretrogradational pattern of the 3rd-order cycle Pa-3 can be seen on the NE part of the profile. Note the offlapping forced
regressive stranded parasequence sand bodies just under the SB Pa-4 settled during the large-scale relative lake-level fall (FSST). The major channel
belt of an axial river can be traced at the boundary zone between the emerging (NE) and subsiding (SW) areas in the basin after the SB Pa-4. For
location see Fig. 6. Legend: 13: Upper-MiocenePliocene (Pannonian s.l.) post-rift sediments: 1. Fluvial sediments, 2. Delta front and coastal plain
sediments, 3. Delta-slope and basin-plain sediments, 4. Reworked sand bodies, 5. 3rd-order sequence boundary, 6. 4th-order sequence boundary, 7.
Incised channel fills (channel belt), 8. Middle Miocene syn-rift sediments.

the logs and reconstructed on dip-oriented sedimento- during and after the lake-level fall on the top of the HST
logical cross sections (Fig. 10). Northeast of these and most of it was eroded. At this time rivers began to
forced regressive sand bodies, strong erosion occurred incise, although on the studied profiles extensive major

Fig. 11. Depositional facies versus 3rd- and 4th-order cycles in the SE part of the basin, in more elevated position, according to well and seismic data
(profile V). Note that the major transgression can be recognized here in the thickness of the delta front facies unit as well as in the transition of the
fluvial and the deltaic facies unit due to the strong erosion on the SB Pa-4. Legend: see Fig. 10.
 86
Gy. Juhsz et al. / Sedimentary Geology 202 (2007) 7295
Fig. 12. Characteristic stratigraphical and sedimentological column in the distal (Well A), central (Well B) as well as proximal (Well C) part of the wide morphological shelf of the study area. Well C
represents at the same time the deepest part of the Derecske trough with an extremely (ca. 1700 m) thick fluvio-deltaic succession, here shown only the uppermost part of it. Over a few 100 m thick
fluvial sediments the thick transgressive succession of delta front sediments returns, consisting a series of individual coarsening-upward parasequences, referring to the major transgression at the area
at about 7.27.5 Ma (TST of Pa-3). For location see profile II (Fig. 6).
 Gy. Juhsz et al. / Sedimentary Geology 202 (2007) 7295
large-scale IVF systems were not recognized in this part
of the basin. However, in the deep basinal areas deeply
incised canyons can be identified corresponding with
SB Pa-4.
The rate of erosion varied in different parts of the
sub-basins. Erosion was strongest in the southern part of
the study area (profile V, Fig. 11), where the offshore
sediments of the major transgression above the thick
delta front facies unit are missing or can be traced only
in the vicinity of the wells. Here the seismic profiles also
do not exhibit the transparent seismic facies overlying
the wedge-shaped high-amplitude seismic facies. De-
positional facies versus 3rd- and 4th-order cycles in the
southern part of the study area can be seen on the
sedimentological profile, according to borehole and
seismic data. Due to the strong erosion at SB Pa-4 the
major transgression can be recognized on this profile
only in the transition between the fluvial and the deltaic
facies unit (Fig. 11). Erosion was most negligible in the
NE part (profile I), where even the progradation of the
HST (upper delta) can be seen (Fig. 8).
6.1.3. Sedimentological and sequence stratigraphical
features in different zones of the wide morphological shelf
During the major progradingretrograding phase of
the 3rd-order tectonostratigraphic cycles Pa-2 and Pa-3, a
wide morphological shelf was formed, characterized by
particular fluvio-deltaic architecture. To characterize the
whole shelf zone that can be seen on seismic data, the
distal, central, and proximal parts of the shelf zone are
distinguished, each having different sedimentological and
stratigraphical character. The stratigraphical column of
each zone will be discussed briefly by boreholes indicated
or projected on profile II (Figs. 6, 12).
The distal part of the shelf was characterized by
offshore marls intercalated with coarsening-upward
fringing sandmarl successions. The signs of the
major progradationretrogradation are represented in a
set of shoaling-upward (thickening and coarsening-up)
delta front successions that would be termed para-
sequences and which indicate the overall progradation.
This is overlain by a set of deepening upward (thinning
and fining-upward) delta front successions related to
overall retrogradation, although they comprise several
4th-order cycles. The major transgression is character-
ized here by a thick offshore marl succession. The
offshore marls are overlain by a thin forced regressive
delta front sand body at SB Pa-4. Above the SB Pa-4
there is a thick fluvial succession, displaying fining-
upward fluvial cycles (Fig. 12A).
The central part of the shelf can be demonstrated by
the sedimentological and sequence stratigraphical
87
column of a borehole which proves 1000 m of delta
front/shoreface sediments. This thick succession con-
sists of several progradingretrograding cycles of
thickening and coarsening-upward, as well as thinning
and fining-upward delta front successions that would be
termed parasequences. In the uppermost part the major
transgression (the 3rd-order MFS of Pa-3 sequence) can
be seen at 300 m depth, and above it the Pa-4 sequence
boundary (Fig. 12B).
The proximal part of the shelf is characterized by one
of the deepest wells from the Derecske trough. It has an
extremely thick fluvio-deltaic succession (N 1200 m).
The major progradationretrogradation pattern is well
characterized in the lithological and facies column. The
delta front and coastal plain sedimentary succession
interfingers with alluvial deposits of a considerable
thickness (300 m). The latter is overlain by 350 m thick
delta front sediments. At the top, immediately below the
Quaternary boundary, the offshore marls of the TST of
Pa-3 sequence occur (Fig. 12C).
6.2. Climatically driven 4th-order cycles characteristics
and evidence
6.2.1. Sigmoidal and wedge-shaped 4th-order deltaic
cycles
The well-developed 3rd-order sequence Pa-3 con-
tains an extremely thick succession of deltaic sediments,
representing mainly delta front and delta-plain as well as
coastal plain environments. Due to the enormous yield of
sediment carried by rivers emerging from the hinterland
of the Eastern Carpathians, sedimentation was able to
keep up with the available accommodation space created
by the coeval subsidence of the area. Rather thick sig-
moid and wedge-shaped 4th-order deltaic cycles were
deposited during this period.
The signs of overall progradational to aggradational
movement of the shoreline is seen in the LST of the 3rd-
order Pa-3 cycle, built up by 4th-order cycles which are
characteristically sigmoidal in profile, thinning up-dip
onto the coastal plain and downlapping towards the basin
plain (Figs. 6, 7). An aggradational feature can also be
observed towards the coastal plain and alluvial-plain
areas for more than 100 km, showing that accommoda-
tion was created continuously through subsidence of the
whole basin. Above these cycles, the pattern changes to
weakly aggradational along the profiles and finally
backstepping of the shoreline is marked by wedge-
shaped retrogradational cycles pinpointing the trans-
gressive or ravinement surface in the upper part of the
Pa-3 3rd-order sequence. A definite thickening back of
the geometry of the cycles can be seen towards the
88 Gy. Juhsz et al. / Sedimentary Geology 202 (2007) 7295

Fig. 13. 4th-order fluvial cycles above the 6.85 Ma SB Pa-4 interpreted on the basis of borehole data (detail of profile II). They correspond fairly well to the
cycles identified on the seismic profile. Similarly the channel-belt facies recognized on the logs is well visible on the seismic in the form of lots of smaller
insized channels (see Fig. 10). The profile is a detail of regional profile II. Legend: 1. fluvial sediments, 2. deltaic sands, 3. offshore marls, 4. reworked sand
bodies.

coastal plain areas meaning that a depocentre shifted
backward caused by structural changes (Fig. 7).
At this time, the thickest deltaic succession was
deposited in the basin, forming a wide and shallow-water
morphological shelf. Shifting from retrogradational to
progradational phases the deltas passed through the
shallow shelf zone each time, and even if the relative
lake-level variation was low, they reached the shelf-edge
and became shelf-margin deltas. During transgression
and highstand, a wide delta front zone was formed,
therefore this wide shelf was infilled quickly in each
cycle by delta front deposits. This is the reason why the
greatest proportion of the delta front sediments remains
in the SW and central part of the shelf zone, as shown in
the stratigraphic column of the boreholes (Fig. 12B).
6.2.2. Large-scale fluvial cycles
The most prominent architectural feature of the
fluvial succession within the Pa-4 and Pa-3 3rd-order
cyles is the presence of fining-upward, large-scale
fluvial cycles (sensu Miall, 1996) within the Milanko-
vitch band. The cycles, varying between 40120 m
thick, were interpreted on the basis of well-logs, but the
same scale cycles are also seen in seismic profiles,
showing a good correlation with the logs (Figs. 10, 11).
These fluvial cycles can be correlated over longer
distances, even on the basis of well-logs (Fig. 13). The
large-scale fluvial cycles consist of a set of fining-
upward facies successions, beginning with incised
channel fills of the LST or LAS (low accommodation
space), overlain by sediments of the TST or IAS
(increasing accommodation space) of the proximal
floodplain settled during rising base level (terminology
after Schumm, 1993; Shanley and McCabe, 1994, 1998;
Weissmann et al., 2002; Juhsz et al., 2004a). In the
alluvial plain areas, the uppermost parts of the cycles
consist of aggrading floodbasin sediments of the HST or
HAS (high accommodation space), deposited during
 Gy. Juhsz et al. / Sedimentary Geology 202 (2007) 7295
high base level. In the channelbelt zones amalgamated,
multi-storey channel fill sand bodies begin the cycle,
overlain by thinner or thicker sediments deposited in the
proximal floodplain of the TST. Here the aggrading
HST sediments are missing, and were probably mostly
eroded by the subsequent channel incision during the
next base level lowstand.
The major channel belt of an axial river can be traced
at the boundary zone between the emerging and
subsiding areas in the basin in the Pa-4 3rd-order
sequence, above the SB Pa-4, flowing almost at a right
angle to the profile (Fig. 10). The major sediment
dispersal direction changed somewhat, from NE to SW
earlier, to NNE to SSW in the Pa-4 cycle. The position
of the channel belt was probably determined by
structural movements in this zone.
7. Discussion
7.1. Sedimentary evolution and tectonic imprint
The sediments in the eastern part of the Pannonian
Basin represent a large delta complex that prograded
from the NE towards the SW, and was one of the major
delta systems infilling Lake Pannon. On the regional
profiles three 3rd-order sequences the Pa-2, Pa-3, and
Pa-4 sequences were identified. They are character-
ized by a particular fluvio-deltaic architecture, including
extremities in the thickness of the delta front and fluvial
facies, an anomalously thick lowstand prograding
complex, and relatively thin transgressive and highstand
deposits at a third-order scale. However, the majority of
the delta front and coastal plain sediments of the 3rd-
order LST belongs to the TST and HST at a 4th-order
scale. This was a source of misunderstanding and a
divergence of opinion in early works interpreted from
different databases, i.e. by logs and sedimentology, and
by seismic evidence (Juhsz, 1993; Vakarcs et al.,
1994). The 3rd-order sequence boundaries form major
unconformities and their correlative conformities in the
basin. During the deposition of Pa-2 and Pa-3 3rd-order
sequences, a wide morphological shelf was formed, and
progradation displays a typical shelf-margin setting in
spite of the fact that it was a huge lacustrine basin.
Extensive large-scale Incised Valley Fill (IVF)
systems were not recognized in this part of the basin,
but smaller (2040 m depth) incised channels are
observed on seismic data and up-valley multi-storey
channel fills were identified from well-logs within the
fluvial succession. In the deep basinal areas towards the
Bks Basin, however, a few deeply incised canyons
can be seen connected to SB Pa-4. The associated IVF
89
systems were probably eroded, similarly to most of the
HST deposits, during and after the strong relative lake-
level fall. Collapse of the shelf-edge is a common
feature in both the 4th- and 3rd-order cycles.
By the time delta progradation reached the study area
from the NE during deposition of the Pa-2 3rd-order cycle
at about 1011 Ma, the Derecske trough had already
become a narrow, somewhat shallower trench (Fig. 14).
Channelized turbidites flowed along the axis of the trench,
while the overbank fines of the channelized turbidites as
well as fine-grained slope apron sediments deposited on
the flanks. Most of the turbidites in the deepest part
(N 1000 m in thickness) belong to Pa-1, but a smaller part
belongs to the Pa-2 3rd-order cycle. Overall progradation
to aggradation of the deltaic system continued in the basin
until 7.58 Ma, the huge amount of sediment supply
keeping pace with thermal subsidence, and a wide
morphological shelf was formed. The trough was almost
infilled when progradation changed through aggradation
to a strong retrogradation. During the TST and HST of the
3rd-order cycle Pa-3 rapid tectonic subsidence (and
possibly lower rate of sediment supply) created space
for a transgressive phase.
A major transgression occurred with the MFS at about
7.27.5 Ma far inside the fluvial plain areas when relative
lake-level rose at a considerable rate (200300 m). From
this time the backwards thickening of the cycles can be
observed towards the already filled deep basinal area.
Seemingly increase of the rate of subsidence and/or tilting
of the area took place. At the same time, in the bottomset of
the clinoforms the timeline terminations onlap towards the
Szeghalom basement high from the E by gently folded and
curved reflectors showing that the palaeotopography was
much flatter during the time of progradation (Figs. 6, 10).
In the upper part of the Pa-3 cycle (TST and HST during
the major transgression) this feature is absent, demonstrat-
ing the syn-sedimentary nature of the uplift of the
SzeghalomFldes high. As a consequence of this large-
scale progradationalretrogradational cycle, the Pa-3
third-order cycle deposited the thickest deltaic succession
in the Pannonian Basin, especially in this area (Fig. 11).
A possible explanation for these phenomena is
differential subsidence and uplift and/or tectonic
reactivation of the basement structures. In the vicinity
of the Derecske Trough, several gently dipping thrust
sheets could be detected (Pogcss et al., 1991), most of
them reactivated during the Early to Middle Miocene
extensional rifting phase (Windhoffer et al., 2005).
Windhoffer and Bada (2005) suggested the presence of a
flat-ramp-flat thrust system where the flat segments
were not reactivated. They assumed the flat segments
were cut by normal faults rooting in the reactivated ramp
90 Gy. Juhsz et al. / Sedimentary Geology 202 (2007) 7295

Fig. 14. Sedimentary evolution of the Late-Miocene (Pannonian s.l.) sedimentary succession in the eastern part of the Hungarian Plain. Note that
during the 2nd phase (B: overall progradation) progradation and retrogradation of the 4th-order cycles alternated with each other, each time passing
through the shallow shelf zone and reaching the shelf-edge, becoming a shelf-edge delta from time to time. The major transgression was caused by
a strong structural change and tilting back towards the NE edges of the sub-basin. Finally, at the end of the infilling of the Derecske trough, a major
relative lake-level fall can be identified (ca. 6.8 Ma) probably caused by the onset of a compressional/inversion phase or a tilting back to the west of
the whole eastern area. In the meantime, in the central sub-basins, like the Bks Basin, subsidence continued on a higher rate. Therefore erosional
processes removed the majority of the HST deposits of the Pa-3 3rd-order cycle (except in the northernmost part), and much of the TST as well. The
whole process could have been enhanced by the oversteepening of the slope during the overriding of the Szeghalom- and the neighbouring basement
highs. After this point the whole system became rearranged and a major delta distributary appeared from a northerly direction.

segments of the thrust system (Windhoffer and Bada,
2005). A later phase of reactivation of the basement
structures was initiated by a strike-slip type stress
regime (Windhoffer and Bada, 2005). In the Derecske
Trough this might have started in this period, during the
TST of the Pa-3 3rd-order cycle (ca. between 78 Ma).
It might have caused the transpression of the Szegha-
lom-high (causing the onlap feature of the already-
deposited sediments) and the transtensional subsidence
of the Derecske trough. Earlier studies proved the en-
 Gy. Juhsz et al. / Sedimentary Geology 202 (2007) 7295
echelon folds and transpressive feature of the Bihar-
nagybajomSzeghalom basement high (Rdly, 1990).
At SB Pa-4, at about 6.8 Ma, a sharp change in
accommodation to sediment supply ratio can be observed.
The relative lake-level fall is estimated at greater than
200 m (perhaps enhanced by subsequent tectonic activity),
abrupt facies changes occurred and in the Pa-4 cycle
sedimentation continued at about this decreased lake-level
afterwards while the morphological shelf ceased to exist
(Figs. 10, 11). Most of the HSTof Pa-3 3rd-order sequence
was eroded after this event; the prograding HST delta can
be traced only in the northernmost part of the study area,
on profile I (Fig. 8). This large-scale base level fall was
most probably caused by interaction of structural changes,
including tilting, compressional features, and/or basin
inversion and the area of the Derecske trough was
emerged while subsidence of the central sub-basins, like
the Bks Basin, was enhanced.
Therefore one proposed solution is that inversion
began at about 6.8 Ma in this part of the basin and not at
4 Ma as was earlier generally accepted (Vakarcs et al.,
1994; Fodor et al., 1999). Another solution could be that
the phenomena and also the variations of the rate of
sediment supply discussed above could be also generated
by large-scale tectonic changes in the hinterland areas,
towards the Eastern Carpathians. Actually during the Pa-3
3rd-order cycle, the overall progradation was interrupted
by a relative lake-level rise, then a sudden relative lake-
level fall of a 200300 m magnitude. It could have been
caused by a tilting event to the east and then tilting back to
the west of the area, superimposed by some strike-slip
movements. However, the different solutions interacted
intimately and one can have been caused by the other.
In conclusion, the sedimentary evolution of the Late-
MiocenePliocene (Pannonian s.l.) sedimentary succes-
sion, defined through the integrated stratigraphical
investigations, at the same time proved to be a key to
interpret and date the structural evolution of the sub-
basin. During the investigations it was found that the Pa-
3 cycle forms a major tectonostratigraphic cycle and
highlights the onset of a new era of structural evolution
in the region.
7.2. Tectonic versus climatic controls?
Earlier sequence stratigraphical studies in the central
Pannonian Basin have led to three markedly different
models. The first model was based on seismic reflection
profiles in the central and northern part of the study area.
Several unconformities within the Upper Miocene
Pliocene succession were identified and traced in
seismic profiles (Pogcss et al., 1988; Vakarcs and
91
Vrnai, 1991; Csat, 1993; Pogcss et al., 1994;
Vakarcs et al., 1994). These surfaces were interpreted as
third-order sequence boundaries caused by basin-wide
relative lake-level changes. The surfaces were dated by
palaeomagnetic data from four boreholes (Elston et al.,
1990). Correlation between the seismic profiles and the
magnetostratigraphical interpretations, uncertain though
the latter were, led to the conclusion that some of the
water-level changes of Lake Pannon occurred in the
same phase as global eustasy (Pogcss et al., 1988,
1994). A significant Late-Miocene 3rd-order water-level
change had been dated using magnetostratigraphy
(Csat, 1993) and a very close match with the age of
the Mediterranean salinity crisis was found.
Twelve sequence boundaries were identified on
seismic and well-log profiles in the Neogene sedimentary
fill of the Pannonian Basin by Vakarcs et al. (1994). In the
case of four sequence boundaries (8.2 Ma, 6.3 Ma,
4.2 Ma, 3.8 Ma) excellent correlation was found between
the magnetostratigraphic age of the sequences and the
ages of the boundaries of the Haq et al. (1987) curve.
Correlation of the other eight sequence boundaries was
only approximate (Vakarcs et al., 1994), but it was
presumed that this number and the revealed magnetos-
tratigraphic age of these sequences correspond well to the
published global eustatic curve for that period of time
(Vakarcs et al., 1994; Vakarcs, 1997).
The second model was based on seismic studies in the
southern part of the study area (Bks basin). Mattick et al.
(1994) identified only local hiatuses, which extended for
distances of a maximum of 30 km along strike. This pattern
was interpreted to reflect lateral switching of delta lobes
rather than basin-wide lake-level changes, and it was
considered to represent fourth-order stratigraphical cycles.
Mattick et al. (1994) predicted that if some of the hiatuses
are still related to basin-wide lake-level changes, further
research may show that they were triggered by climatic
cycles. Indeed, in the following years several attempts were
made to demonstrate the link between lacustrine cycles in
Lake Pannon deposits (observed mostly within the jfalu
Formation) and climatic cycles within the Milankovitch
band (Juhsz et al., 1997, 1999; Sprovieri and Sacchi,
1999; Korps-Hdi et al., 2000; Sprovieri et al., 2003;
Harzhauser et al., 2004; Sacchi and Mller, 2004).
The third model was based on high-resolution
sedimentological analysis of 10 fully-cored boreholes
from the Pannonian Basin, 5 of which, including
Kaskantyu-2 and Tiszapalkonya-I, were located in the
western and northern margins of the central Pannonian
Basin. Juhsz E. et al. (1996) documented one regional
unconformity within the Upper MiocenePliocene sec-
tion of the boreholes. This occurs between subaerially
92 Gy. Juhsz et al. / Sedimentary Geology 202 (2007) 7295
exposed delta-plain sediments with palaeosols and root
casts, and a fluvial succession consisting of channel
sands and floodplain marls. Juhsz E. et al. (1996)
claimed that the unit below the unconformity is
Miocene, whereas the unit above the unconformity is
Pliocene in age, but they provided no data in support of
this dating. These interpretations, however, have not
been linked yet to a regional or a basin-scale sequence
stratigraphical network. Recently three of these bore-
holes were correlated with each other using seismic
profiles in the SW part of the Hungarian Plain, far from
our study area, with some sub-basins in between
(Tthn Makk, 2003).
At first sight, the three models mentioned above
seem to be mutually exclusive. However, the authors'
observations and analyses demonstrate that all three
models, or at least their elements, are reconcilable. In the
study area, 3rd-order sedimentary sequences are un-
doubtedly present in the succession (cf. model 1). The
most spectacular sequence boundary, Pa-4, presented
earlier, was clearly identified in fully-cored boreholes,
such as Kaskantyu-2 and Tiszapalkonya-I (cf. model 3).
In the southwestern part of the study area (Bks basin),
above the Pa-4 SB, however, only local unconformities
can be detected on the seismic lines. Fourth-order
sequences, within the third-order sequence Pa-4, are
observable here (cf. model 2).
The integrated stratigraphical results, presented here,
strongly suggest that depositional cycles of different
orders were driven by different factors. On the regional
dip profiles in the main route of the deltaic infill of the
basin both 3rd- and 4th-order cycles were identified. On
the basis of the stacking pattern of the cycles and the
sedimentological, as well as sequence stratigraphical
analysis, the driving forces of the 3rd-order cycles were
regional scale tectonic changes. This was most charac-
teristic in the 3rd-order Pa-3 cycle, where the huge
amount of sediment delivery into the basin, developing
extremely thick sedimentary succession, made clearly
visible the effects of the tectonic background through
the inner architecture of the cycle, the extremities of the
progradation of the clinoforms and finally the large-
scale facies changes through the course of relative lake-
level fall at the Pa-4 sequence boundary.
The 4th-order cycles are clearly visible and correlat-
able in the study area with a good areal extent until the Pa-
4 sequence boundary. After the major relative lake-level
fall, clinoforms became thinner while extensive shelfal
areas could no longer be seen. Therefore progradation was
much faster, forming individual autocyclic deltaic lobes,
as noted by Mattick et al. (1994), and 4th-order cycles can
be easily recognized but with more limited areal extent at
least at a seismic scale. Since there were no marine
connections eustasy could have no effect on Lake Pannon.
Therefore 4th-order cycles were most probably driven
climatically in the Milankovitch band, through river
discharge, sediment supply, changes in fluvial style in the
alluvial environments and other factors. An unambiguous
definition of the origin and nature of the 4th-order cycles
would require a more precise and reliable geochronolog-
ical framework.
8. Conclusions
Integrated stratigraphical investigations were carried
out in the eastern part of the Pannonian Basin, focused
on the palaeogeographical evolution and the response of
fluvio-deltaic systems to tectonic and climatic controls.
Depositional facies were interpreted in a sequence
stratigraphical framework in the area of sediment input
into the basin from the NE. As a consequence a refined
picture on the stratigraphical architecture and the
sedimentary evolution of the study area has been
achieved, both for the fluvial and the deltaic complexes.
The facies and stratigraphic architecture of the
identified 3rd- and 4th-order depositional cycles indi-
cate that the 3rd-order cycles were driven by regional
scale structural changes. The investigations strongly
suggest that 4th-order depositional cycles were most
probably driven by large-scale cycles within the
Milankovitch band and their multiples.
In the study area in the 3rd-order cycle interval Pa-3
(sensu Vakarcs, 1997), between 9.16.8 Ma, a major
progradationalretrogradational pattern can be ob-
served. This cycle exhibits enormous sediment yield
and extremes in facies architecture, driven by tectonic
changes. The Pa-3 cycle forms a major tectonostrati-
graphic cycle and highlights the onset of a new era of
structural evolution in the region.
Through integrated stratigraphical analysis the dating
of the renewal of strike-slip movements and tilting of the
study area may be possible. It is marked by a major
backstep and backwards thickening of the delta complex
at about 7.27.5 Ma, causing a widespread major
transgression onto fluvial floodplains inland from the
shoreline, above 100's of m thick fluvial deposits.
The SB Pa-4 (ca. 6.8 Ma), at the end of the 3rd-order
cycle Pa-3, indicates a very strong relative base level
lowering of greater than 200 m in the eastern part of the
basin, accompanied by large-scale abrupt facies
changes. It is a result of a complete change of structural
style marking the beginning of basin inversion in the
coeval basin margins in the east, like the area and
surroundings of the Derecske Trough.
 Gy. Juhsz et al. / Sedimentary Geology 202 (2007) 7295
From the beginning of the 3rd-order cycle Pa-4,
sedimentation continued in the subsiding central basins
(e.g. Bks Basin), characterized by decreased water
depth and without any morphological shelf which was
well-developed earlier. The main routes of sediment
input changed after SB Pa-4: earlier it was from NEE,
later from the NNE. The position of the channel belt of
the main axial river at the boundary of the subsiding and
emerging areas was most probably determined by
structural movements in this zone.
Acknowledgements
The work was supported financially by the Hungar-
ian National Research Fund (OTKA 035168), and
scientifically by the Geological Institute of Hungary.
We are grateful to the Hungarian Geological Survey for
the possibility to review the necessary datasets.
Support of the MOL Hungarian Oil and Gas Co. is
gratefully acknowledged. We thank the reviewers for
their useful comments and corrections on the paper and
Professor Dr. Phil Gibbard for careful revision of the
English text.
References
Bada, G., Horvth, F., Tth, L., Tth, T., 2000. Seismotectonic
problems of radioactive waste disposal. Fldtani Kzlny,
Hungary, Budapest 130/4, 585610 (In Hungarian).
Balla, Z., 1984. The Carpathian loop and the Pannonian Basin: a
kinematic analysis. Geophysical Transactions 30, 313353.
Balogh, K., rva-Ss, E., Pcskay, Z., 1986. K/Ar dating of post-
Sarmatian alkali basaltic rocks in Hungary. Acta Mineralogica-
Petrographica, Szeged 28, 7593.
Brczi, I., Phillips, R.L., 1985. Processes and depositional environ-
ments in Neogene deltaiclacustrine sediments, Pannonian Basin,
SE Hungary. Geophysical Transactions, Hungary, Budapest 31,
5574.
Brczi, I., Hmor, H., Jmbor, ., Szentgyrgyi, K., 1988. Neogene
sedimentation in Hungary. In: Royden, L.H., Horvth, F. (Eds.),
The Pannonian Basin A Study in Basin Evolution. AAPG
Memoir, vol. 45, pp. 5767.
Berggren, W.A., Kent, D.V., Swisher, C.C., Aubry, M.-P., 1995.
A revised Cenozoic geochronology and chronostratigraphy. In:
Berggren, W.A., Kent, D.V., Aubry, M.-P., Hardenbol, J. (Eds.),
Geochronology Time Scales and Global Stratigraphic Correlation.
SEPM Special Publication, vol. 54, pp. 129212.
Brezsnynszky, K., Haas, J., Kovcs, S., Szederknyi, T., 2000.
Geology of Hungary 2000 Basement Geology. Hungarian
Geological Institute, Budapest. Special Publication.
Csat, I., 1993. Neogene sequences in the Pannonian Basin, Hungary.
Tectonophysics 226, 377400.
Csontos, L., 1995. Tertiary tectonic evolution of the intra-Carpathian
area: a review. Acta Volcanologica 7/2, 113.
Detzky, G., Lrincz, K., Tevan, K., 2000. Seismic study of neotectonic
phenomena in the Szolnok Flysch Belt (in Hungarian). Magyar
Geofizika 41/1, 3341.
93
Elston, D.P., Lantos, M., Hmor, T., 1990. Az Alfld pannniai (s.l.)
kpzdmnyeinek magnetosztratigrfija. A MFI vi Jelentse
az 1988. vrl, I. Rsz. MFI, Hungary, Budapest, pp. 109134.
Elston, D.P., Lantos, M., Hmor, T., 1994. High resolution polarity
records and the stratigraphic and magnetostratigraphic correlation of
Late Miocene and Pliocene (Pannonian s.l.) deposits of Hungary. In:
Teleki, P.G., Mattick, R.E., Kkay, J. (Eds.), Basin Analysis in
Petroleum Exploration. A Case Study from the Bks Basin,
Hungary. Kluwer Academic Publishers, Dordrecht, pp. 111142.
Fodor, L., Csontos, L., Bada, G., Gyrfi, I., Benkovics, I., 1999.
Tertiary tectonic evolution of the Pannonian Basin system and
neighbouring orogens: a new syntesis of paleostress data. In:
Durand, B., Jolivet, L., Horvth, F., Seranne, M. (Eds.), The
Mediterranean Basins: Tertiary Extension within the Alpine
Orogen. Geological Society, London, Special Publications,
vol. 156, pp. 295334.
Gerner, P., Bada, G., Dvnyi, P., Mller, B., Oncescu, M., Cloetingh,
S., Horvth, F., 1999. Recent tectonic stress and crustal
deformation in and around the Pannonian Basin: data and models.
In: Durand, B., Jolivet, L., Horvth, F., Seranne, M. (Eds.), The
Mediterranean Basins: Tertiary Extension within the Alpine
Orogen. Geological Society, London, Special Publications,
vol. 156, pp. 269294.
Haas, J., Hmor, G., Korps, L., 1999. Geological setting and tectonic
evolution of Hungary. Geologica Hungarica Series Geologica,
Hungary, Budapest 24, 179196.
Haq, B.U., Hardenbol, J., Vail, P.R., 1987. Chronology of fluctuating
sea level since the Triassic. Science 235, 11561167.
Hmor, G., 1978. Die Orogenphasen des Badenian, Chronostratigra-
phia und Neostratotypen, vol. 6. Veda, Bratislava, pp. 109110.
Hmor, G., 1995. Miocene paleogeographic and facies map of the
Carpathian Basin (I-II-III). Geological Atlas of Hungary, vol. 19.
Hungarian Geological Institute, Budapest.
Hmor, G., Pogcss, Gy., Jmbor, ., 2001. Paleogeographic/
structural evolutionary stages and the related volcanism of the
CarpathianPannonian Region. Acta Geologica Hungarica, Hun-
gary, Budapest 44/23, 193222.
Harzhauser, M., Daxner-Hck, G., Piller, W.E., 2004. An integrated
stratigraphy of the Pannonian (Late Miocene) in the Vienna Basin.
Austrian Journal of Earth Sciences 95/96, 619.
Horvth, F., 1993. Towards a mechanical model for the formation of
the Pannonian Basin. Tectonophysics 226, 333357.
Horvth, F., Cloetingh, S., 1996. Stress-induced late-stage subsidence
anomalies in the Pannonian Basin. Tectonophysics 266, 287300.
Horvth, F., Tari, G., 1999. IBS Pannonian Basin Project: a review of
the main results and their bearing on hydrocarbon exploration. In:
Durand, B., Jolivet, L., Horvth, F., Seranne, M. (Eds.), The
Mediterranean Basins: Tertiary Extension within the Alpine
Orogen. Geological Society, London, Special Publications,
vol. 156, pp. 195214.
Hunt, D., Tucker, M.E., 1995. Stranded parasequences and forced
regressive wedge systems tract: deposition during base-level fall
reply. Sedimentary Geology 95, 147160.
Jiricek, R., 1979. Tektogeneticky vyvoj Karpatskeho oblouku behem
oligocenu a neogenu. In: Mahel, M. (Ed.), Tektonicke Profily
Zapadnych Karpat. Geol. Ustav. Dionys Stura, Bratislava, pp. 205215.
Jmbor, ., 1989. Review of the geology of the s. l. Pannonian
formations of Hungary. Acta Geologica Hungarica, Hungary,
Budapest 32, 269324.
Juhsz, Gy., 1991. Sedimentological and lithostratigraphical framework
of the Pannonian (s.l.) sequence in the Hungarian Plain, Eastern
Hungary. Acta Geologica Hungarica, Budapest 34/12, 5372.
94 Gy. Juhsz et al. / Sedimentary Geology 202 (2007) 7295
Juhsz, Gy., 1992. A pannniai s.l. formcik trkpezse az Alfldn:
elterjeds, fcies s ledkes krnyezetek- Pannonian s.l. forma-
tions in the Hungarian Plain: distribution, facies and sedimentary
environment. Fldtani Kzlny, Budapest 122, 133165 2-4.ktet,
(In Hungarian and English).
Juhsz, Gy., 1993. Relatv vzszintingadozsok rtegtani-szedimento-
lgiai bizonytkai az Alfld pannniai s.l. ledksszletben. Magyar, I., Geary, D.H., St-Szentai, M., Lantos, M., Mller, P.,
(Sedimentological and stratigraphical evidences of relative water-
level fluctuations in the Pannonian Lake.). Fldtani Kzlny,
Budapest 123/4, 379398 (with English abstract).
Juhsz, Gy., 1994. Magyarorszgi neogn medencerszek pannoniai s.l.
ledksornak sszehasonlt elemzse. (Comparison of the sedi-
mentary sequences in Late-Neogene sub-basinsub-basins in the
Pannonian Basin, Hungary). Fldtani Kzlny, Hungary, Budapest
124 (3), 341365 (with English abstract).
Juhsz, E., Mller, P., Ricketts, B.D., Tth-Makk, ., Hmor, T.,
Farkas-Bulla, J., Suto-Szentai, M., 1996. High-resolution sequence
stratigraphy and subsidence analysis of the Late Neogene in the
Pannonian Basin, Hungary. Acta Geologica Hungarica 39,
129152 Budapest.
Juhsz, E., Kovcs, L.., Mller, P., Tth-Makk, ., Phillips, L.,
Lantos, M., 1997. Climatically driven sedimentary cycles in the
Late Miocene sediments of the Pannonian Basin, Hungary.
Tectonophysics 282, 257276.
Juhsz, E., Phillips, L., Mller, P., Ricketts, B., Tth-Makk, M.,
Lantos, M., Kovcs, O.L., 1999. Late Neogene sedimentary facies
and sequences in the Pannonian Basin, Hungary. In: Durand, B.,
Jolivet, L., Horvth, F., Seranne, M. (Eds.), The Mediterranean
Basins: Tertiary Extension within the Alpine Orogen. Geological
Society, London, Special Publications, vol. 156, pp. 335356.
Juhsz, Gy., Mller, P., Tth-Makk, ., 2004a. Alluvial architec-
ture and fluvial cycles in Quaternary deposits in a continental
interior basin, E Hungary. Zagreb, Geologia Croatica 57/2,
171191.
Juhsz, Gy., Pogcss, Gy., Magyar, I., Vakarcs, G., 2004b. Late-
Neogene facies, sequences and basin subsidence in the eastern part of
the Pannonian Basin (Hungary). IAS 2004 Coimbra, Portugal, p. 155
(Abstracts).
Juhsz, Gy., Pogcss, Gy., Vakarcs, G., Magyar, I., 2005a.
Sedimentary evolution and integrated stratigraphy: a key to
structural evolution in the Eastern Pannonian Basin, Hungary.
European Geological Union (EGU) General Assembly, Wienna,
2429/04/2005. Geophysical Research Abstracts, vol. 7 (CD)
EGU05-A-09682.
Juhsz, Gy., Pogcss, Gy., Vakarcs, G., Magyar, I., 2005b. Evolu-
tion of fluvio-deltaic systems in the central part of the Pannonian
Basin. 8th International Conference on Fluvial Sedimentology,
Delft, 712/08/1995, Abstracts, p. 148.
Kkai, J., Pogcss, Gy., 1991. Hydrocarbon plays in Mesozoic
nappes, Tertiary wrench basins, and interior sags in the Pannonian
Basin. First Break 9/7, 315334.
Korps-Hdi, M., Nagy, E., Nagy-Bodor, E., Szkvlgyi, K., Kovcs,
L.., 2000. Late Miocene climatic cycles and their effect on
sedimentation (west Hungary). In: Hart, M.B. (Ed.), Climates: Past
and Present. Geological Society, London, Special Publications,
vol. 181, pp. 7988.
Lantos, M., Hmor, T., Pogcss, Gy., 1992. Magneto- and seismos-
tratigraphic correlations of Pannonian s.l. (Late Miocene and
Pliocene) deposits in Hungary. Paleontologia i Evoluci 2425,
3546 (Sabadell).
Lemberkovics, V., Brny, ., Gajdos, I., Vincze, M., 2005.
A szekvencia sztratigrfiai esemnyek s a tektonika kapcsolata
a Derecskei-rok pannoniai rtegsorban. Fldtani Kutats,
Budapest, Hungary XLII (1), 1624.
Lourens, L., Hilgen, F., Shackleton, N.J., Laskar, J., Wilson, D., 2004.
The Neogene period. In: Gradstein, F.M., Ogg, J.G., Smith, A.G.
(Eds.), A Geologic Time Scale 2004, vol. 409440. Cambridge
University Press, pp. 469484.
1999. Integrated biostratigraphic, magnetostratigraphic and chron-
ostratigraphic correlations of the Late Miocene Lake Pannon
deposits. Acta Geologica Hungarica 42 (1), 531.
Mattick, R.E., Rumpler, J., Ujfalusy, A., Szanyi, B., Nagy, I., 1994.
Sequence stratigraphy of the Bks basin. In: Teleki, P.G., Mattick,
R.E., Kkay, J. (Eds.), Basin Analysis in Petroleum Exploration. A
Case Study from the Bks Basin, Hungary. Kluwer Academic
Publishers, Dordrecht, pp. 99110.
Miall, A.D., 1996. The Geology of Fluvial Deposits. Springer-Verlag.
582 pp.
Molenaar, C.M., Rvsz, I., Brczi, I., Kovcs, A., Juhsz, Gy.K.,
Gajdos, I., Szanyi, B., 1994. Stratigraphic framework and
sandstone facies distribution of the Pannonian sequence in the
Bks Basin. In: Teleki, P.G., Mattick, R.E., Kkay, J. (Eds.),
Basin Analysis in Petroleum Exploration. A Case Study from the
Bks Basin, Hungary. Kluwer Academic Publishers, Dordrecht,
pp. 99110.
Nemcok, M.G., Pogcss, Gy., Pospisil, L., 2005. Activity timing of
the main tectonic systems in the CarpathianPannonian region in
relation to the rollback destruction of the lithosphere. In: Golonka,
J., Picha, J. (Eds.), The Carpathians and their Foreland.: Geology
and Hydrocarbon Resources. AAPG Memoir, vol. 84.
Papp, A., Jmbor, ., Steininger, F.F. (Eds.), 1985. Chronostratigra-
phie und Neostratotypen. . Miozn der Zentralen Paratethys, vol. 7.
Akadmiai Kiad, Budapest. Pannonien, 636 pp.
Phillips, R.L., Rvsz, I., Brczi, I., 1994. Lower Pannonian deltaic-
lacustrine processes and sedimentation, Bks Basin. In: Teleki, P.G.,
Mattick, R.E., Kkay, J. (Eds.), Basin Analysis in Petroleum
Exploration. A Case Study from the Bks Basin, Hungary. Kluwer
Academic Publishers, Dordrecht, pp. 99110.
Plint, A.G., Nummedal, D., 2000. The falling stage systems tract:
recognition and importance in sequence stratigraphic analysis. In:
Hunt, D., Gawthorpe, R.L. (Eds.), Sedimentary Responses to Forced
Regressions. Geological Society, London, Special Publications, 172,
pp. 117.
Pogcss, Gy., Lakatos, L., Rvsz, I., Ujszszi, K., Vakarcs, G.,
Vrkonyi, L., Vrnai, P., 1988. Seismic facies, electrofacies and
Neogene sequence chronology of the Pannonian Basin. Acta
Geologica Hungarica 31/34, 175207.
Pogcss, Gy., Lakatos, L., Barvitz, A., Vakarcs, G., Farkas, Cs.,
1989a. PliocenQuaternary strike-slipe faults is the Great
Hungarian Plain, Hungary (Pliocn kvarter oldaleltoldsok a
Nagyalfldn). ltalnos Fldtani Szemle 24, 149169.
Pogcss, Gy., Vakarcs, G., Barvitz, A., Lakatos, L., 1989b. Post-rift Strike-
slip Fault Sin the Pannon Basin and Their Role in the Hydrocarbon
Accumulation. Proceedings of the 34th International Geophysical
Symposium, Budapest, September 1989, vol. II, pp. 601611.
Pogcss, Gy., Simon, E., Vrnai, P., 1991. Hydrocarbon perspectives
of overthrusted nappes in the Preneogene basement of the
Pannonian Basin. 36th International Geophysical Symposium
Kiev. Proceedings, pp. 6575.
Pogcss, Gy., Mattick, R.E., Elston, D.P., Hmor, T., Jmbor, .,
Lakatos, L., Lantos, M., Simon, E., Vakarcs, G., Vrkonyi, L.,
Vrnai, P., 1994. Correlation of Seismo- and Magnetostratigraphy
in Southeastern Hungary. In: Teleki, P.G., Mattick, R.E., Kkay, J.
 Gy. Juhsz et al. / Sedimentary Geology 202 (2007) 7295
(Eds.), Basin Analysis in Petroleum Exploration. A Case Study
from the Bks Basin, Hungary. Kluwer Academic Publishers,
Dordrecht, pp. 143160.
Posamentier, H.W., Vail, P.R., 1988. Eustatic controls on clastic
deposition II. Sequence and system tracts models. In: Wilgus, C.K.,
Hastings, B.S., Kendall, C.G.St.C., Posamentier, H.V., Ross, C.A.,
Van Wagoner, J.C. (Eds.), Sea-level Changes: an Integrated
Approach. SEPM Spec.Publ., vol. 42, pp. 125154.
Rdly, P., 1990. Az oldaleltoldsos rendszerek kompresszis/
transzpresszis znihoz kapcsold sznhidrogn csapdzds
lehetsgei s korltai alfldi pldkon. Alkot Ifjsg Plyzat, Tthn Makk, . 2003: Szekvencia sztratigrfiai vizsglatok a Duna-
Budapest 68.
Rvsz, I., Brczi, I., Phillips, R.L., 1989. A Bksi-medence
alspannniai ledkkpzdse. (Lower-Pannonian sedimentation
in the Bks Basin.). Magyar Geofizika 30, 98113.
Royden, L.H., 1988. Late Cenozoic Tectonics of the Pannonian Basin
System. In: Royden, L., Horvth, F. (Eds.), The Pannonian Basin, a
Study in Basin Evolution. AAPG Memoir, vol. 45, pp. 2748.
Sacchi, M., Mller, 2004. Orbital cyclicity and astronomical calibration
of the Upper Miocene continental succession cored at the
Iharosberny-I well site, western Pannonian Basin, Hungary. In:
D'Argenio, B., Fischer, A.G., Premoli Silva, I., Weissert, H., Ferreri,
V. (Eds.), Cyclostratigraphy: Approaches and Case Histories. SEPM
Special Publication, vol. 81, pp. 275294.
Schumm, S.A., 1993. River response to base-level change: implications
to sequence stratigraphy. The Journal of Geology 101, 279294.
Serra, O., 1985. Sedimentary Environments from Wireline Logs.
Schlumberger, Montrouge, France, p. 211.
Shanley, K.W., McCabe, P.J., 1994. Perspectives on the sequence
stratigraphy of continental strata. AAPG Bulletin 78 (4), 544568.
Shanley, K.W., McCabe, P.J. (Eds.), 1998. Relative Role of Eustacy,
Climate and Tectonics in Continental Rocks. SEPM Spec.Publ.,
vol. 59. 234 pp.
Sprovieri, M., Sacchi, M., 1999. Correlation between Paratethys and
Mediterranean events during the Tortonian: a working hypothesis.
Neogene Newsletter 6, 6070.
Sprovieri, M., Sacchi, M., Rohling, 2003. Climatically influenced
interactions between the Mediterranean and the Paratethys during
the Tortonian. Paleoceanography 18/2, 1034.
Szabo, Cs., Harangi, Sz., Csontos, L., 1992. Review of Neogene and
Quaternary volcanism of the CarpathianPannonian region.
Tectonophysics 208, 243256.
Tari, G., Horvth, F., Rumpler, J., 1992. Styles of extension in the
Pannonian Basin. Tectonophysics 208, 203219.
95
Tari, G., Dvnyi, P., Dunkl, I., Horvth, F., Lenkey, L., Stefanescu,
M., Szafin, P., Tth, T., 1999. Lithospheric structure of the
Pannonian Basin derived from seismic, gravity and geothermal
data. In: Durand, B., Jolivet, L., Horvth, F., Sranne, M. (Eds.),
The Mediterranean Basins: Tertiary Extension within the
Alpine Orogen. Geological Society, London, Special Publications,
vol. 156, pp. 215251.
Teleki, P.G., Mattick, R.E., Kkay, J. (Eds.), 1994. Basin Analysis in
Petroleum Exploration. A Case Study from the Bks Basin,
Hungary. Kluwer Academic Publishers, Dordrecht, p. 330.
Tisza kze kzps s dli rsznek ks miocn rtegsorban
(Sequence stratigraphy of Late Miocene deposits in the southern
and central part of the Danube-Tisza interfluve). Unpubl. PhD
thesis, ELTE Budapest, 156.
Vakarcs, G., 1997. Sequence stratigraphy of the Cenozoic Pannonian
Basins, Hungary. PhD thesis. Rice University, Houston, Texas, 514.
Vakarcs, G., Vrnai, P., 1991. A Derecskei-rok krnyezetnek szeizmosz-
tratigrfiai modellje. (Seismostratigraphic model of the area of the
Derecske-trough.). Magyar Geofizika XXXII (12), 3850.
Vakarcs, G., Vail, P.R., Tari, G., Pogcss, Gy., Mattick, R.E., Szab,
A., 1994. Third-order Middle MioceneEarly Pliocene depositonal
sequences in the prograding delta complex of the Pannonian Basin.
Tectonophysics 240, 81106.
Van Wagoner, J.C., Posamentier, H.V., Mitchum, R.M., Vail, P.R., Sarg,
J.F., Loutit, T.S., Hardenbol, J., 1988. An overview of the
fundamentals of sequence stratigraphy and key definitions. In:
Wilgus, C.K., Hastings, B.S., Kendall, C.G.St.C., Posamentier, H.V.,
Ross, C.A., Van Wagoner, J.C. (Eds.), Sea-level Changes: an
Integrated Approach. SEPM Spec.Publ., vol. 42, pp. 3946.
Van Wagoner, J.C., Mitchum, R.M., Campion, K.M., Rahmanian, V.D.,
1990. Siliciclastic Sequence Stratigraphy in Well-logs, Cores and
Outcrops. AAPG Methods in Exploration Series, vol. 7. USA, Tulsa.
Weissmann, G.S., Mount, J.F., Fogg, G.E., 2002. Glacially driven
cycles in accumulation space and sequence stratigraphy of a
stream-dominated alluvial fan, San Joaquin Valley, California,
USA. Journal of Sedimentary Research 72/ 2, 240251.
Windhoffer, G., Bada, G., 2005. Formation and deformation of the
Derecske Trough, Pannonian Basin: insight from analog modeling.
Acta Geologica Hungarica 48/4, 351369.
Windhoffer, G., Bada, G., Nieuwland, D., Wrum, G., Horvth, F.,
Cloetingh, S., 2005. On the mechanics of basin formation in the
Pannonian Basin: inferences from analogue and numerical
modelling. Tectonophysics 410, 389415.
 Available online at www.sciencedirect.com

Sedimentary Geology 202 (2007) 96 123

www.elsevier.com/locate/sedgeo

Volcanic and climatic controls on fluvial style in a high-energy

system: The Lower Cretaceous Matasiete Formation,
Golfo San Jorge basin, Argentina
Jos Matildo Paredes a,, Nicols Foix a , Ferrn Colombo Piol b , Adriana Nillni a ,
Jos O. Allard a , Rosa A. Marquillas c
a
Dpto. de Geologa, F.C.N., Universidad Nacional de la Patagonia San Juan Bosco, Ruta Provincial N 1,
Km. 4, (9005) Comodoro Rivadavia, Chubut, Argentina
b
Dpto. de Estratigrafa, Paleontologa y Geociencias Marinas, Fac. de Geologa, Universidad de Barcelona, 08071, Barcelona, Spain
c
CONICET, Dpto. de Geologa, F.C.N., Universidad Nacional de Salta, Buenos Aires 177 (4400) Salta, Argentina

Abstract

The Cretaceous Chubut Group in the Golfo San Jorge Basin (Patagonia, Argentina) comprises up to 6000 m of continental
sediments. At the base of the succession, an extensive salinealkaline lacustrine unit (Pozo D-129 Fm) grades laterally towards the
basin margin into the fluvial Matasiete Fm (Hauterivian?Aptian). The Matasiete Fm comprises up to 650 m of siliciclastic and
pyroclastic deposits in its type area, where three Members have been distinguished. The following facies associations were
identified: 1) single fluvial channels; 2) multistorey fluvial channels; 3) proximal floodplain; 4) distal floodplain; 5) pyroclastic
(ash-fall and ground-surge) deposits. Additional observations in the time-equivalent Pozo D-129 Fm provide evidence for 6)
shallow lacustrine, and 7) deep-lacustrine facies associations. The floodplain association contains paleosols with abundant
carbonate concretions and fine pyroclastic tuffs. Paleosols are mostly vertisols, characterized by deep desiccation cracks,
slickensides and carbonate nodules. These observations, coupled with the occurrence of shallow-water oolitic grainstones in the
lacustrine Pozo D-129 Fm indicate a semiarid environment. Tree trunks up to 1 m in diameter and 15 m long, preferentially located
near channel margins, provide evidence for vegetated riverbanks. The exposures of the Matasiete formation in the San Bernardo
Fold Belt show individual channel belts containing straight, meandering and braided channels. Evidence of changes in fluvial style
attributable to base-level control has not been observed. The most dramatic changes in fluvial architecture, which were
comparatively short-lived, were provoked by intermittent pyroclastic deposition on the floodplain. The reduced infiltration capacity
of the floodplain and the concomitant increase in sediment load increased runoff, as well as volume and flashiness of the discharge,
which favoured the development of shallow, multichannel rivers. The fluvial system abruptly changed from single, ribbon channels
to a braided system, which notably increased the width/thickness ratio of sandbodies.
2007 Elsevier B.V. All rights reserved.

Keywords: High-energy fluvial systems; Fluvial architecture; Pyroclastic input; Climatic and volcanic controls; Cretaceous; Patagonia; Argentina

1. Introduction

Corresponding author. Tel.: +54 297 154 125 039; fax: +54 297
455 0339.
E-mail address: paredesj@unpata.edu.ar (J.M. Paredes).
0037-0738/$ - see front matter 2007 Elsevier B.V. All rights reserved.
doi:10.1016/j.sedgeo.2007.05.007
Understanding of the evolution of river systems re-
quires reconstruction of changes in depositional en-
vironments in space and time. Such reconstructions
 J.M. Paredes et al. / Sedimentary Geology 202 (2007) 96123
must be based on detailed measurements of parameters
related to the hydrological behaviour of channel belts and
the associated floodplain, as well as the correct recogni-
tion of changes in external controls on sedimentation (e.g.
base level, tectonics, climate, and volcanism). In recent
years, a number of seminal papers have addressed
changes in fluvial architecture produced under conditions
of tectonic forcing (e.g. Bridge and Leeder, 1979;
Alexander and Leeder, 1987; Schumm, 1993; Holbrook
and Schumm, 1999; Arche and Lopez-Gomez, 1999;
Gawthorpe and Leeder, 2000), base-level changes
(Legarreta et al., 1993; Wright and Marriot, 1993; Shanley
and McCabe, 1994; Emery and Myers, 1996; Leeder and
Steward, 1996; Legarreta and Uliana, 1998; Shanley and
McCabe, 1998; Kraus, 1999), and subtle changes in
climate of source areas and alluvial basins (Shanley and
McCabe, 1998; Blum and Tornqvist, 2000; Cecil, 2003).
The dramatic impact of pyroclastic input into the drainage
system and the floodplain, particularly relevant for basins
located near areas of volcanic activity, has also been
evaluated in the last decades (Cas and Wright, 1987;
Smith, 1987; Fisher and Smith, 1991).
This work seeks to contribute to the insights obtained
in the above studies, by unravelling the controls on
sedimentology, stratigraphy and depositional architec-
ture of the fluvial systems of the Matasiete Formation,
based on detailed field observations in the San Bernardo
Fold Belt. Particular attention has been given to the role
of pyroclastic input and its effects on fluvial architec-
ture. A regional framework for synthesizing the controls
on deposition has been constructed by means of ad-
ditional observations on the lacustrine Pozo D-129 Fm,
which comprises deposits of a lake that acted as a
regional base level for the fluvial systems.
1.1. Stratigraphic and tectonic setting
The Chubut Group of the Golfo San Jorge basin
provides a record of fluvial and lacustrine deposition
in an intracratonic setting. The deposits reach about
6000 m in thickness in the subsurface of the basin, and
comprise around 2000 m of epiclastic and pyroclastic
deposits along the San Bernardo Fold Belt (Fig. 1), a
NorthSouth trending belt in which extensional Meso-
zoic faults were inverted mostly during the Tertiary
(Homovc et al., 1995). In recent decades, detailed in-
vestigations of the Chubut Group have been carried out,
mostly due to their importance for the oil industry. The
Pozo D-129 Fm is considered to be the main source
rock of the basin, and the volcaniclastic sandstones
of the Castillo and Bajo Barreal Fms are the main oil
reservoir rocks.
97
The Neocomiano or Las Heras Group (Fig. 2), a
marine to transitional megasequence deposited under
synrift conditions (Figari et al., 1999) underlies the
Chubut Group. The Las Heras Group is not exposed in
the San Bernardo Fold Belt. The Tertiary sediments that
cover the Chubut Group consist of fluvial and aeolian
deposits preserved in local depressions, and scarce
marine sediments preserved in the southern part of the
Fold Belt. The lithostratigraphic arrangement of the
Chubut Group indicates the development of an ex-
tensive salinealkaline lacustrine unit at the base (Pozo
D-129 Fm), which is laterally equivalent to the fluvial
Matasiete Fm (Hauterivian? to Aptian). The Castillo
Fm, a volcaniclastic unit deposited in fluvial and very
shallow lacustrine environments during the Albian,
conformably covers both units. The Chubut Group also
contains younger fluvial deposits of the Bajo Barreal Fm
(Cenomanian to Campanian) and the Laguna Palacios
Fm (Santonian to Maastrichtian), the latter characterized
by stacked paleosols and minor channels.
Regional studies of the Golfo San Jorge basin based
on subsurface and outcrop data (Barcat et al., 1989;
Fitzgerald et al., 1990; Peroni et al., 1995; Figari et al.,
1999; Uliana and Legarreta, 1999; Hechem and
Strelkov, 2002) have depicted the general evolution of
the basin and its petroleum systems. Fitzgerald et al.
(1990) regarded the Pozo D-129 Fm as part of an early
sag stage in the evolution of the basin, whereas other
authors (Figari et al., 1999) proposed an extensional or
transtensional setting during deposition of the Pozo D-
129 Formation. Thermal subsidence has been the main
mechanism during deposition of the other units of the
Chubut Group (Figari et al., 1999; Rodriguez and Littke,
2001). Feruglio (1949), Lesta and Ferello (1972),
Sciutto (1981), Clavijo (1986) and Hechem et al.
(1990) formulated the main characteristics and strati-
graphic relationship of the Chubut Group. Hechem et al.
(1990) recognized six depositional sequences in the
Chubut Group. The lower three depositional sequences
include the Pozo D-129, Matasiete and Castillo Fms,
characterized by the development of a fluvial and lacus-
trine environment, in which transgressive to regressive
cycles are attributed to changes in tectonically and/or
climatically forced changes in lake size. The upper three
depositional sequences are represented by the Bajo
Barreal and Laguna Palacios Fms. Both consist of vol-
caniclastic beds and braided ephemeral fluvial deposits
(Hechem et al., 1990).
Detailed stratigraphic studies in selected outcrops of
the Castillo and Bajo Barreal Fms have been carried out
during the past two decades (Meconi, 1990; Rodriguez,
1992; Legarreta et al., 1993; Hechem, 1994, 1998; Bridge
98 J.M. Paredes et al. / Sedimentary Geology 202 (2007) 96123

Fig. 1. Location of the Golfo San Jorge basin in central Patagonia, and the study area. A) Radar image of Patagonia, the boxed area marks the location of the San
Bernardo Fold Belt. B) Sedimentary basins and uplifted areas in Patagonia. C) Geological map of the San Bernardo Fold Belt, with indication of the exposures
of the Matasiete and Pozo D-129 Fms. Both units are exposed in the cores of anticlines. 1) Codo del Senguerr Anticline, 2) Chenque (Silva) Hill, 3) Cachetamn
Hill, 4) Baya Peninsula, 5) Matasiete Canyon, 6) Sierra Nevada Anticline, 7) Tronador Canyon. Basaltic and intrusive rocks (Eocene-recent) cover most of the
San Bernardo Fold Belt area. Paleozoic and Jurassic rocks in the northern part of the fold belt represent deposits of a previous basin.
 J.M. Paredes et al. / Sedimentary Geology 202 (2007) 96123 99

Fig. 2. Synthetic lithostratigraphy and palaeoenvironments of the San Bernardo Fold Belt. Ages according to Fitzgerald et al. (1990), Bridge et al.
(2000) and Rodriguez and Littke (2001).

et al., 2000). The paper of Bridge et al. (2000) is, until
now, the most detailed work on fluvial architec-
ture developed in the Chubut Group. Their work was
focused on outcrops of the Bajo Barreal Fm and part of the
Castillo Fm, in exposures mostly southward of the study
area of the present paper. They found evidence of
perennial fluvial systems with straight channels (sinuosity
less than 1.23) and rare braided channels, and mean
thickness in the order of 2 to 5 m, but variable between
the studied areas (Bridge et al., 2000). Previously pub-
lished data from the Matasiete Fm in the Matasiete Can-
yon (Paredes et al., 2003) were used in the present study of
this unit over the entirety of the San Bernardo Fold Belt.
1.2. The Pozo D-129 and Matasiete depositional system
The Matasiete and Pozo D-129 units are exposed in
the San Bernardo Fold Belt as isolated outcrops in the
cores of asymmetrical anticlines (Fig. 1). The lacustrine
Pozo D-129 Fm is exposed in the Codo del Senguerr
anticline, Silva Hill and Baya Peninsula. The base of the
unit is not exposed in the basin, and the stratigraphic
interval exposed in individual outcrops does not exceed
100 m. However, in the subsurface of surrounding areas
(e.g. Well Ch-SS.es.1) its thickness exceeds 900 m. A
thickness of at least 1500 m commonly occurs in the
areas to the South and East of the San Bernardo Fold
100 J.M. Paredes et al. / Sedimentary Geology 202 (2007) 96123

Belt (Figari et al., 1999). Because of the scarcity of
exposures of lacustrine rocks, their subdivision is based
on subsurface data. Regional facies distribution in the
lacustrine Pozo D-129 Fm suggests a provenance of
clastics from the North Patagonian and Deseado regions
(Fig. 3, after Figari et al., 1999), in an EastWest
elongate basin with main depocenters parallel to major
normal faults. An increase in pyroclastic content to-
wards the western sector was attributed to early volcanic
activity in the Andean Ranges (Folguera and Ianniz-
zotto, 2004).
Clavijo (1986) informally identified four main units
(sections) in the Pozo D-129 Fm of the western part of
the basin. The lower section contains tuffs and tuffa-
ceous siltstones, grey-to-green laminated shales, and a
subordinate amount of sandstones deposited in deep-
lacustrine environments. This evolves upwards into pale
grey tuffs containing thin tuffaceous sandstones and
oolitic grainstones, which represent deposition in littoral
shoals. A section dominated by stacked green tuffs with
minor coarse-to-medium-grained tuffaceous sandstones,
considered to be shallow lacustrine deposits with high
pyroclastic content, occurs higher up in the unit. Finally,
an upper section composed of white tuffs containing
sandstones, minor oolitic grainstones and green tuffa-
ceous sandstones was recognized, inferred to have been
deposited in a marginal to fluvio-deltaic environment
(Clavijo, 1986).
Anoxic conditions have been inferred for different
sections of the Pozo D-129 Fm, based on the presence of

Fig. 3. Regional patterns of the Matasiete and Pozo D-129 Fms. A) Lateral relationship between Pozo D-129 and Matasiete Fms, showing thickness
variation from North to South and the main progradationalretrogradational cycles (modified from Hechem and Strelkov, 2002). B) Regional
distribution of the Pozo D-129 Fm and lithological character of the unit, mainly based on subsurface information. Coarse clastics were mostly
provided from the North Patagonian and Deseado regions. Pyroclastic material from the former Andes Ranges is most abundant in the western part of
the area. Local source areas are marked by the presence of coarse clastics (after Figari et al., 1999).
 J.M. Paredes et al. / Sedimentary Geology 202 (2007) 96123
laminated shales containing pyrite, dark colors and
absence of burrows (Van Nieuwenhuise and Ormiston,
1989). The dominance of Classopolis pollen and the
abundance of oolitic grainstones, as well as the presence
of Botryococcus-like algal forms, led these authors to
suggest that deposition occurred in a stratified saline
alkaline lacustrine system in a semiarid climate. The
Barremian to Aptian age assigned to the unit was based
on the presence of the Flabellochara harrisi association,
obtained from outcrops of the Chenque (Silva) Hill
(Hechem et al., 1987), but some studies support an
Hauterivian age for the beginning of lacustrine deposi-
tion (Archangelski et al., 1984; Fitzgerald et al., 1990).
The Matasiete Fm (Lesta and Ferello, 1972) was
originally named Reddish Sandstones and Tuffs by
Feruglio (1949). Lesta and Ferello (1972) identified three
Members (Lower, Middle and Upper) based on the ratio of
sandstone bodies to floodplain suites. Sciutto (1981)
provided details of the distribution of these units in
outcrops of the San Bernardo Fold Belt and commented on
their main characteristics. This author was the first to
recognize the lateral relation between the Matasiete Fm and
the lacustrine Pozo D-129 Fm, and interpreted the former
as a meandering high-energy fluvial system. Galeazzi
(1989) described the sedimentology of the Matasiete Fm in
the southern part of Matasiete Canyon, and proposed a
subdivision of this formation into two depositional
sequences (the Lower and Middle Members are included
in Sequence 1 and the Upper Member is equivalent to
Sequence 2). He attributed changes in channel geometry
within the Matasiete Fm to base-level variations. The
fluvial systems were considered to be braided in the lower
sequence and fully lacustrine to meandering in the upper
sequence with paleocurrents to the southeast (120)
(Galeazzi, 1989). Paredes et al. (2003, 2004) recognized
the large variability in fluvial deposits within individual
Members and between superimposed channel belts, and
highlighted the variation in paleocurrent directions, the
dominance of channel migration by avulsion, and the
presence of large tree trunks in some of the channel fills
(some 1 m in thick and 15 m long). Paredes et al. (2003,
2004), who analyzed the lithostratigraphy of the Matasiete
Formation in and around its type area, showed the Lower
Member to be composed of a 205 m thick section de-
posited by a low-sinuosity fluvial system with three main
channel belts, separated by floodplain fines. The Middle
Member is 215 m thick, and characterized by the
dominance of floodplain fines over channel fills. The
fluvial system provides evidence of meandering and
braided behaviour, but many sandbodies have no
diagnostic features. The Middle Member of the Matasiete
Fm grades laterally in a southward direction to lacustrine
101
facies of the Pozo D-129 Fm (Hechem et al., 1990;
Hechem and Strelkov, 2002; Paredes et al., 2003). The
Upper Member, which regionally covers the lacustrine
Pozo D-129 Fm, consists of a 230 m thick section of red-
coloured fluvial deposits comprising straight, meandering
and braided channel fills.
2. Materials and methods
This study comprises the description and interpreta-
tion of detailed stratigraphic sections of the Matasiete
Fm in seven areas of the San Bernardo Fold Belt (Figs. 4
and 5), and analysis of additional observations from
the lacustrine Pozo D-129 Fm. Three Members were
proposed in the Matasiete Canyon based on grain-size
analysis and facies assemblages (Lesta and Ferello,
1972), and all references in the present study to Lower,
Middle and Upper Members of the Matasiete Fm are
indexed to this scheme. All three Members are exposed
in the Matasiete Canyon (Fig. 5). In the other areas
there are only exposures of the Upper Member of the
Matasiete Fm or the lacustrine Pozo D-129 Fm (Fig. 4).
The term sandbody is used in a general sense and
includes single and multistorey channel fills. The ar-
chitectural elements of Miall (1996) and their bounding
surfaces were deduced from interpretation of outcrop
profiles and photomosaics of laterally continuous out-
crops. Paleocurrent directions were measured from
unidirectional sedimentary structures such as cross-
stratification, asymmetrical ripples or oriented tree trunks,
and were structurally corrected (Ramsay, 1961). The
apparent dimensions of channel fills were measured using
GPS point data along the edges of sandbodies, and
corrected with the mean of the paleocurrent data for each
channel fill. The lithofacies analysis largely follows the
classifications and facies definitions of Miall (1996),
Bridge (1993) and Talbot and Allen (1996), with
modifications for volcaniclastic environments suggested
by Mathisen and Vondra (1983), Smith (1987) and Cas
and Wright (1987). Thin sections were obtained from
most of the lithofacies and from the unit exposed in
the Matasiete Canyon. Paleosol analysis included a
description of color, grain size, nodule type, as well as
matrix and mottles. Determination of major oxide weight
percent was carried out by X-ray fluorescence of whole-
rock samples. Clay minerals were identified by X-ray
diffraction.
3. Sedimentology
Galeazzi (1989) and Paredes et al. (2003) proposed
lithofacies schemes for the Matasiete Fm based on
 102
J.M. Paredes et al. / Sedimentary Geology 202 (2007) 96123
Fig. 4. Stratigraphic sections of the Upper Member of the Matasiete Fm and lacustrine Pozo D-129 Fm. Location of the sections in Fig. 1C. The most complete section of the unit is exposed in the
Matasiete Canyon (Fig. 5).
 J.M. Paredes et al. / Sedimentary Geology 202 (2007) 96123 103

Fig. 5. Synthetic stratigraphic section of the Matasiete Fm at Matasiete Canyon. The Lower, Middle and Upper members are defined according to the
ratio of floodplain fines over channelized facies. The stratigraphic distribution of the sandbodies and their W/T ratio is indicated (modified from
Paredes et al., 2003).
104 J.M. Paredes et al. / Sedimentary Geology 202 (2007) 96123
exposures in the Matasiete Canyon. As new observa-
tions allowed us to identify new lithofacies and
depositional processes, short descriptions of the litho-
facies are presented in Table 1 and Fig. 6. Five
genetically significant lithofacies associations were
identified in the Matasiete Fm with the aid of textures,
sedimentary structures, geometry, paleocurrent indicators
and lateral and vertical arrangement of lithofacies: 1)
single fluvial channels; 2) multistorey fluvial channels; 3)
proximal floodplain; 4) distal floodplain; 5) pyroclastic
deposits. Two additional lithofacies associations were
distinguished in the time-equivalent lacustrine Pozo D-
129 Fm: 6) shallow lacustrine association and 7) deep-
lacustrine association.
3.1. Facies association of the Matasiete Fm
3.1.1. Facies association 1: single fluvial channels
3.1.1.1. Description. Facies association (FA) 1 con-
sists of lenticular beds of sandstone less than 3 m thick
and a few tens of meters in width, encased in red mud-
stone. The sandbodies are characterized by the absence
of lateral accretion surfaces. The channel fills have a low
relief basal erosion surface, do not fine upward sig-
nificantly, and are volumetrically dominated by matrix-
supported fine conglomerates and sandstones (Facies
LF1), and clast-supported conglomerates, either hori-
zontally bedded (Facies FL2) or containing trough
cross-bedding (Facies LF3). Paleocurrent directions ob-
tained from cross-strata, and tree trunks generally show
low dispersion (b50) around the mean. Occasionally,
consecutive channels in a single section show orthog-
onal palaeotransport directions.
3.1.1.2. Interpretation. Deposition in single-thread flu-
vial channels is indicated by a significant basal erosion
surface, lenticular geometry, coarse infilling and unidi-
rectional paleocurrents. Single episodes of deposition are
suggested by the lack of internal erosion surfaces and by
uniformity in thickness of cross-sets that fill the channels
(Lunt et al., 2004). Paleoflow variation between channels
is attributed to tectonic tilting or nodal avulsions. The
absence of lateral accretion surfaces implies that the
channels were laterally stable.
3.1.2. Facies association 2: multistorey fluvial channels
3.1.2.1. Description. Facies association (FA) 2 con-
sists of erosionally-based, lenticular and tabular sand-
bodies 3 to 20 m thick, with multiple internal erosion
surfaces. Most sandbodies lack lateral accretion sur-
faces, have ribbon geometries and internally comprise 3
to 4, slightly laterally shifted, vertically stacked stories
(e.g. Fig. 7). They mostly comprise clast-supported
conglomerate, either horizontally bedded (Facies LF2)
or containing trough cross-bedding (Facies LF3), with
matrix-supported conglomerate and sandstone (Facies
LF1), features shared with FA1. The upper parts of the
sandbodies are dominated by coarse-grained sandstones
with low-angle cross bedding, trough and planar cross-
bedding (Facies LF3 and LF4), and occasional medium-
grained sandstones with asymmetrical ripples (Facies
LF6). In some cases, sandstones with desiccation cracks
(Facies LF5) were observed on the upper surfaces of
fining-upward sandbodies (e.g. Fig. 7). The general
trend of the channel fills is fining-upward, but many
strata show no vertical grain-size trend. Where fining-
upward trend is preserved, there is also a decrease in
bedform height and bed thickness upward. Thicknesses
of cross-beds range from 0.3 to 1.8 m. They show
unimodal paleocurrent directions with generally low
dispersion. Internal erosional surfaces are gently con-
cave upwards, show scours in the scale of tens of cen-
timeters and are filled by a thick, poorly organized
lag and low-angle, horizontally bedded conglomerates
(Facies LF2). Occasional tree-trunk fragments, are 30 to
50 cm in diameter (maximum observed diameter was
1 m) and 4 to 15 m long. The larger trunks are oriented
parallel to the channel margins and slightly inclined
in opposed direction to the flow. Few multistorey sand-
bodies have several large-scale inclined surfaces con-
taining trough cross-bedding and asymmetrical ripples
(Facies LF7) on top. Single stories in the later sand-
bodies are separated by pale grey, laminated mudstones
(Fig. 8). The tabular sandbody emplaced on the regional
tuff bed of Fig. 8 preserves a convex-upward, large-
scale inclined surface in a direction normal to the mea-
sured paleocurrent.
3.1.2.2. Interpretation. The coarse-grain size, basal
and internal erosional surfaces, unimodal paleocurrents,
trough and planar cross-stratification in sandstone up
to massive conglomerate beds, compound cross-strata
and the absence of marine fossils suggest deposition in
multistorey fluvial channels (Bridge, 1993; Miall, 1996).
Some elongate sheets of conglomerate containing hor-
izontal bedding (Facies LF2) represent longitudinal bed-
forms. The trough-cross stratified conglomerates and
sandstones (Facies LF3) were deposited in 3D dunes and
bars. Planar cross-stratified sandstones (Facies LF 4)
were formed as 2-D dunes. The dominance of coarse and
ungraded beds, low paleocurrent variability and rarity
of lateral accretion surfaces indicates transportation by
 J.M. Paredes et al. / Sedimentary Geology 202 (2007) 96123 105

Table 1
Lithofacies of the Matasiete and Pozo D-129 Fms in the San Bernardo Fold Belt
Lithofacies and sedimentary features Interpretation
Fluvial facies
LF1. Matrix supported, cobble to pebble conglomerate: Poorly sorted, Mass flow deposited from highly concentrated
massive or crudely bedded, not imbricated, medium-to-fine gravel clasts in sediment flows.
a matrix rich in coarse-to-medium tuffaceous sand. Volcanic clasts, moderately
rounded, although occasional broken clasts are found. Maximum clast-size
observed was 12 cm, but most range from 2 to 5 cm. Thickness less than 2 m,
in gradational or sharp contact with other coarse-grained lithofacies.
LF2. Clast supported, cobble to pebble conglomerate with horizontal Lag deposits or sieve deposits, the thicker beds
bedding: Moderately sorted, crudely bedded with clast imbrication. Weak represent amalgamation of longitudinal bedforms.
grading, a few beds fine upward. Sub-angular to moderately rounded
clasts, volcanic and pyroclastic. Could contain large trunks and bone
remains. Maximum clast-size observed 10 cm, thickness 0.2 to 1.5 m.
Base frequently erosional, top locally gradational to finer lithologies.
LF3. Clast supported, trough cross-stratified, pebble conglomerate and Downstream migration of 3-D dunes or oblique
sandstone: Base commonly erosional, with fining-upward. Local mud migration of longitudinal bars in channels.
intraclasts and dominance of volcanic clasts. Oriented wood fragments.
Individual cross-sets are 0.21.8 m thick, separated by scour surfaces.
Amalgamated bodies up to 7 m thick. Unimodal paleocurrents with
generally low dispersion about the mean.
LF4. Clast supported planartabular stratified, pebble conglomerate and coarse Downstream migration of 2-D dunes or linguoid bars.
sandstone: Planar or scoured base, sets 0.3 to 0.9 m thick that fine upward and
are in many cases amalgamated. Rare mudstone intraclasts and scarce wood
fragments. Imbricated clasts and unimodal paleocurrent distributions.
LF5. Pebble conglomerate and coarse to medium sandstone with desiccation Abandoned channel with subaerial exposure.
cracks: This lithofacies occurs rarely on top of Facies LF3 or LF4. The
cracks are polygonal to irregular in plan view, 2 to 4 cm deep, and are filled
by dark, massive, more consolidated mudstone.
LF6: Fine to medium sandstone with asymmetrical ripples: Rare. Thin beds Downstream migration of ripples in lower
(b0.3 m) overlying flat-based conglomerate bodies, associated with small- flow regime.
scale planar cross-lamination. Ripple sets are 0.01 to 0.03 m thick and
0.1 m in wavelength. Unimodal paleocurrents.
LF7: Coarse to fine sandstone with inclined large-scale surfaces: Found in the Upper Lateral accretion deposits.
Member. Consists of bedsets 0.5 to 5 m thick, with low-angle basal surfaces
(b10), internally containing low-angle cross bedding and trough cross-bedding
that fines upward. Finer lithologies locally show asymmetrical ripples.
Palaeocurrents are near orthogonal to the strike of the inclined surface.
LF8: Medium to fine sandstone with horizontal lamination. Dm-scale Upper or lower flow regime plane beds.
packages of planar-laminated, fine-grained sandstones. Individual beds are
1 cm thick and can be traced for several m, internally ungraded or fine
upward. Some beds contain gravel.
LF9. Tabular or lobate fine to medium sandstone: Planar or erosional base, Overbank deposition into crevasse channels,
lenticular geometry; planar or convex-upward top, occasional pebbles and lobate deposits represents crevasse splay.
general fining-upward trend. Low-angle and trough cross-bedding;
commonly interbedded with beds of Facies LF9. Thickness ranges from
few tens of cm to 0.8 m thick.
LF10. Red to yellowbrown, massive to laminated siltstone and claystone: Subaerial, distal floodplain deposits, soils with
Tabular geometry, beds are fine-laminated or massive, occasionally chemical precipitation. Occasional isolated,
mottled. Locally contains abundant calcium carbonate nodules (caliche), as ephemeral, very shallow-water deposition.
isolated, concentric nodules or thin, tabular beds, sparsely root-bioturbated.
Occasionally pale-grey or white in color; contains small-scale symmetrical
ripples and Planolites, Palaeophycus and Taenidium fossil traces.

Pyroclastic facies
LF11: Tabular fine-grained tuff with horizontal bedding: Planar base, Pyroclastic deposits from volcanic ash fall-out.
ungraded beds 0.3 to 3.5 m thick. Massive or horizontal stratified. Incipient to full pedogenic modification.
Moderate to well sorted, subangular clasts, rare accidental lithic clasts.
Mottling (local to pervasive) and root traces (vertical to subvertical,
up to 10 cm long) on top of individual beds.
(continued on next page)
106 J.M. Paredes et al. / Sedimentary Geology 202 (2007) 96123

Table 1 (continued)
Lithofacies and sedimentary features Interpretation
Pyroclastic facies
LF12: Grey fine-grained tuff with low-angle cross-bedding and antidune Volcaniclastic flow (ground surge), upper and
bedding: Tabular geometry, thickness up to 1 m, individual sets are less lower flow regime conditions. Turbulent flow.
0.15 m thick and 1.5 m wavelength, unwelded, with undulated base,
pumice clasts, glass shards, and orientation of lithic clasts. Gradational
contact between coarse and fine strata, containing low-angle cross
stratification that laterally grades to plane-laminated beds. Dome
like structures with low-angle cross-laminated beds dipping
upstream and downstream.
LF13: Grey fine-grained tuff with accretionary lapilli: Plane or Volcaniclastic flow deposits. Deposition of
undulated base, ungraded beds 0.2 to 0.4 m thick. Massive, locally accretionary lapilli from condensation
contains parallel lamination and rarely low-angle cross-bedding. into the pyroclastic surge or from ash-fall deposition.
Accretionary lapilli are up to 0.7 cm in diameter and exhibit
concentric internal structure.
LF14: Grey horizontally laminated fine-grained tuffaceous sandstone. Hyperconcentrated flood flow, possibly formed
Tabular geometry, consists of laterally continuous laminae or very under rapidly aggrading conditions in super-critical
thin beds internally massive; often associated to LF12. Common flow conditions
superposed sets with different laminae thickness.

Lacustrine facies (Pozo D-129 Fm)

LF15: Grey or bluegrey, tabular, oolitic limestone: Planar or Fresh-water limestone
irregular base, amalgamated bodies up to 0.8 m in thickness, with
concave-up tops. Internally massive or contain small-scale cross-bedding.
Ooids form well sorted grainstones, their shape and size are
spherical and about 2 mm in diameter, with no dolomite in their
composition and diagenetic silicification.
LF16: Grey or pale blue siltstone and claystone: Planar base and tabular Alternating deposition from suspension and traction
geometry, internally massive. Some thicker beds contain sand to silt current in deep-lacustrine setting.
laminae a few cm in thickness. Locally interbedded with white, tabular,
fine-laminated or massive tuff. Scarce bioturbation and abundance
of elongate carbonate nodules.
LF17: Tabular, coarse to medium sandstone: Planar or erosional base, and Bars and dunes deposited in a shallow
frequent concave-up top. Individual bodies are 0.30.6 m thick, lacustrine setting
moderately to well sorted, ungraded or fine-upward, amalgamated into
packages c. 3.5 m thick. Tuffaceous clasts are dominant, but volcanic
and rare limestone clasts present. Contains low-angle cross-bedding
and occasional hummocky cross-bedding.

high-energy currents and suggests the presence of low
sinuosity, fixed channels (Bridge et al., 2000). The local
occurrence of desiccation cracks on top of fining-upward
in-channel strata records desiccation of surficial sedi-
ments between flood events (Collinson, 1996). Storeys
displaying vertical reduction in grain size and internal
structure, and preservation of rippled cross-strata (Facies
LF6) on top indicate progressive reduction in discharge
during filling, with slow and gradual abandonment of the
channel. Avulsion was an important process in the infill
of these channels. A few coarse-grained channels contain
lateral accretion surfaces, which suggest helicoidal flow
and lateral migration of these channels. Single storeys of
these meandering channels are separated by laminated
mudstones suggesting variable discharge and periods of
inactivity. Convex-upward, large scale, inclined surfaces
in tabular sandbodies suggest the diversion of the flow
and presence of braided rivers.
3.1.3. Facies association 3: proximal floodplain
3.1.3.1. Description. This Facies Association (FA3)
consists of small sandstone bodies less than 0.8 m thick
(Facies LF9) encased in sparsely bioturbated reddish
mudstones (Facies LF10) and occasional thin ash-fall
beds (Facies LF11) with variable terrigenous clastic
content, deposited adjacent to major sandbodies. The
sandstone beds are mostly tabular, planar or erosionally-
based and many contain red-coloured mudstone intra-
clasts. Internally they show low-angle and trough-cross
bedding (Facies LF3) in sets up to a few tens of cm thick.
Paleocurrent shows low dispersion in the measurements
 J.M. Paredes et al. / Sedimentary Geology 202 (2007) 96123 107

and is mostly oriented at high angle to the adjacent, upward trends are locally present in the interbedded
major sandbodies (Fig. 8). Many beds fine upward. Tops sandstone and mudstone associations. This succession is
of sandstone beds are planar or concave-up, and typically commonly truncated by the presence of channel sand-
gradational into Facies LF10. Coarsening or fining- bodies. Where present, paleosols are dominated by thin
108 J.M. Paredes et al. / Sedimentary Geology 202 (2007) 96123

Fig. 6. Selected photographs of lithofacies. A) Matrix supported, fine gravels and sandstones (Facies LF1) preserved at the base of multistorey
channels. B) Clast supported, fine gravel beds (Facies LF2). The gravels are moderately rounded, mostly dominated by acid volcanic clasts,
ignimbrites and tuffs. Pen is 15 cm long. C) Planartabular fine gravels and sandstones (Facies LF4) over horizontally bedded sandstones (Facies
LF2). D) Oriented trunks (voids) preserved in channel fill deposits. Lower Member of the Matasiete formation in the Matasiete Canyon. The arrow
points to a hammer 0.3 m long. Trunks shown are up to 0.5 m in diameter and 15 m long. E) Fine distal floodplain deposits (Facies LF10)
containing abundant, rounded, carbonaceous concretions (arrows). F) Coarse sandstone bed with planar cross-bedding (Facies LF4). G) Rooted
white tuff (Facies LF10). Bioturbation increases towards the top, while the base contains parallel lamination. Scale in centimetres. H) Ripple cross-
lamination (Facies LF6) preserved on top of a fining-upward bed. Sandstone over the erosive surface contains mud intraclasts (arrow). Coin is
15 mm in diameter. I) Amalgamated, meter-scale cross-bedded sandstones (Facies LF3). Paleoflow is to the upper right. Hammer is 0.3 m long. J)
Fine laminated tuff (Facies LF14) at the base of a volcaniclastic surge deposit. K) Massive tuff containing accretionary lapilli (Facies LF13). L)
Tuffs containing low-angle cross bedding (Facies LF12) are part of the surge deposits. Arrows mark post-event bioturbation tubes.

horizons of pedogenic carbonate and small rhizoliths coarsening upward facies trend suggests avulsion via
(sensu Kraus and Hasiotis, 2006) preserved in fine crevasse splay progradation of the main channel (Sling-
sandstones. erland and Smith, 2004).

3.1.3.2. Interpretation. Based in the small scale of their
features, the sandstones represent crevasse splays and
small crevasse channels, formed during high-discharge
flood events in areas proximal to LA2 (Miall, 1996),
with paleocurrents oriented at high angle to those of the
adjacent channel fills. Coarsening-upward trends are
indicative of successive flows arriving on a low-energy
floodplain (Smith et al., 1989). Fining-upward trends are
attributed to an increase in the distance to active channel
belts or to the gradual reduction of the discharge. The
return to suspension fallout is marked by deposition of
mudstones or tuffs with paleosol horizons. Truncation of
3.1.4. Facies association 4: distal floodplain
3.1.4.1. Description. This facies association (FA4) is
dominated by reddish, massive to laminated claystones
and siltstones (Facies LF10), and tabular beds of fine-
grained tuff (Facies LF11). The association reaches a
thickness of several tens of meters, and the main char-
acteristics are the abundant carbonate concretions in the
form of isolated, concentric nodules up to 0.4 m in
diameter, or occurring as tabular and more consolidated
beds. Pedogenic features are vertical root traces, mot-
tling, slickensides and coalesced carbonate nodules (see
 J.M. Paredes et al. / Sedimentary Geology 202 (2007) 96123 109

Fig. 7. Details of the multistorey channels. A) Photomosaic of a selected fixed, straight, multistorey sandbody, Lower Member in Matasiete Canyon
B) Interpretation and distribution of paleocurrent data. C) Measured section of the multistorey sandbody and lithofacies details.

below). Trace fossils are common on bed surfaces or with-
in beds, in the form of vertical, cylindrical, unlined and
passively filled tubes. Some tabular, laminated fine-
grained sandstone beds contain symmetrical ripples at
their tops, as well as horizontal, multidirectional, menis-
cate burrows. The bioturbation intensity of most beds in-
creases upwards, and pervasive bioturbation is common in
some localities (Bellosi et al., 2002). In the upper Member,
superimposed cycles grade upward from thick, yellow
brown coloured mudstone beds to red mudstone beds.
bonate nodules indicates an excess of alkaline solutes and
precipitation due to fluctuations in ground-water table in a
well-drained floodplain (Retallack, 1988). Mottling
results from periodic waterlogging (Wright, 1999). The
presence of beds containing wave ripples and burrows
(faunal feeding and/or dwelling activity) near the top
suggests that deposition took place in very shallow-water
or isolated ephemeral ponds (Abdul Aziz et al., 2003).
3.1.5. Facies association 5: pyroclastic deposits

3.1.4.2. Interpretation. The red colour of most of this 3.1.5.1. Description. This facies association (FA5) con-
LA indicates subaerial exposure. The abundance of car- sists of tabular, pale-grey or white, lithic-free tuffs (Facies
110 J.M. Paredes et al. / Sedimentary Geology 202 (2007) 96123

Fig. 8. Multistorey meandering fluvial channel from Upper Member in Matasiete Canyon (location in Fig. 5). A, B) Photomosaic and interpretation of
the sandbody. Facies LF7 represent large-scale inclined surfaces that separated single episodes of infill of the channel. A braided channel is present on
top of the white, tabular tuff shown in the lower part of the picture. C) Measured section and paleocurrent data. Pictures show details of sandstone
lobes preserved in the margin of the channel and root impressions at the base of the channel. Hammer is 0.3 m long.

LF11), preserved in the three Members of the Matasiete
Fm. The pyroclastic beds consist mainly of zeolite/
smectite-replaced cuspate pumice shards, with sparse
crystals of plagioclase and quartz. The tuffs are massive or
finely laminated and frequently show root traces and
vertical to horizontal burrows on top. Two beds identified
towards the top of the Middle Member in the Matasiete
Canyon (Fig. 9) are about 4.5 km in lateral extent and 1 to
3 m thick and show the vertical stacking of plane-based,
fine laminated tuffs (Facies LF14) containing accretionary
lapilli (Facies LF13), followed by several dm-thick beds of
low-angle cross-bedded and finely laminated tuffs, which
tend to thicken within gentle depressions and thin over
crests of undulations (Facies LF12). Accretionary lapilli
are in many places concentrated into thin strata. Near-
symmetrical bedforms are thin (b0.15 m thick) with
wavelengths of about 0.7 to 1.5 m, and internal lamina-
tions dipping downstream and upstream. Individual beds
are ungraded, but the general trend is coarsening-upward
owing to the presence of accidental tuffaceous clasts in the
upper sets. Occasional elongate biotite grains are displayed
parallel to the base of the bedforms.
3.1.5.2. Interpretation. The deposition of fine, mas-
sive tuffs with root traces or burrows was formed by
intermittent deposition of ash-fall as tabular beds in low-
gradient, subaerial and subaqueous areas of floodplains
(Cas and Wright, 1987). Tuff beds containing asym-
metrical ripples and small-scale cross-bedding indicate
transportation by turbulent flows in lower flow regime
 J.M. Paredes et al. / Sedimentary Geology 202 (2007) 96123 111

Fig. 9. A) Stratigraphic section of ground-surge deposits in uppermost part of Middle Member at Matasiete Canyon (location in Fig. 5). B)
Photomicrograph of finely laminated tuff (Facies LF12) with cuspate Y-shaped shards and plagioclase fragments (arrow). Scale bar is 0.5 mm. C)
Photomicrograph of accretionary lapilli, rim-type, from the basal section of the pyroclastic surge. The coarse-grained core is coated by at least seven
layers of progressively finer-grained, darker, ash. Scale bar is 1 mm. D, E) Outcrop view and sketch of dome-like forms interpreted as antidune
structures (Facies LF12) interbedded with low-angle cross-stratified tuffs. Pen is 14 cm long. Flow is from left to right.

condition. The dome-like structures with sets of laminae
dipping upstream and downstream are interpreted as
antidune structures (Alexander et al., 2001).
The extensive, tuffaceous beds containing low-angle
cross bedding and horizontal lamination and rare antidune
bedding indicates transportation by flows that are tran-
sitional from the upper plane bed to antidune stability
fields (Fielding, 2006). These beds indicate the presence
of ground-surge deposits associated with contemporary
volcanic events (Fisher and Schmincke, 1984; Carey,
1991; Schumacher and Schmincke, 1995).
3.2. Facies associations of the Pozo D-129 formation
3.2.1. Facies association 6: shallow lacustrine
environment
3.2.1.1. Description. The shallow lacustrine associa-
tion (FA6) consists of tabular oolitic limestone beds
(Facies LF15) and low-angle cross-bedded and hum-
mocky cross-stratified coarse to fine sandstone beds
(Facies LF17). Some clastic beds display erosional
bases and fining-upward trends, but others have a planar
112 J.M. Paredes et al. / Sedimentary Geology 202 (2007) 96123

base and concave-upward top, and show partial su-
perposition of lenses about 1 m thick and few tens of
meters width. Many beds contain oolitic limestone clasts
as part of a basal lag. The carbonate lithofacies comprise
oolitic grainstones without dolomite. The ooids are
regularly shaped, show several concentric laminae, and
do not contain detrital components (Fig. 10B). Their
nuclei often consist of microfossils.
3.2.1.2. Interpretation. The presence of basal erosion-
al surfaces, fining-upward trends and moderately-to-
well-sorted sandstone beds, plus the presence of sedi-
mentary structures generated in a lower flow regime,
suggest in-channel deposition from turbulent flows.
Hummocky cross-bedding is evidence of oscillatory
and multidirectional flows, related to storm waves. The
occasional presence of partially superposed, lobate beds
indicates unconfined deposition in bar complexes. Both
facies may be preserved in a low-gradient marginal
lacustrine setting (Link and Osborne, 1978). The oolitic
grainstone beds require a well-oxygenated shallow-
water body and excess of alkaline solutions for their
formation. Such conditions are met during intervals of
reduced sediment supply, and in areas sheltered from
clastic input (Talbot and Allen, 1996).
3.2.2. Facies association 7: deep-lacustrine environment
3.2.2.1. Description. This facies association (FA7)
is recorded in outcrops exposed to the south of the
Matasiete Canyon, Silva Hill and Codo del Senguerr
anticline (Locations 1, 2 and 4 in Fig. 1). It mostly
consists of grey or pale blue, poorly consolidated, lo-
cally bioturbated but more commonly fine-laminated
claystones and siltstones (Facies LF16) and thin tuff
beds (Facies LF11). Thin, lenticular, fine sandstone
beds containing ooids and abundant carbonate nodules
a few cm in diameter are occasionally intercalated in
these laminated beds. In southern areas (e.g. Codo del
Senguerr), FA7 is dominated by stacking of white,
massive or finely laminated tuffs.
3.2.2.2. Interpretation. A low-energy environment
is inferred from the dominance of extensive, finely
laminated deposits. The calcareous composition of the
nodules indicates salinealkaline waters and oxidizing
conditions (Platt and Wright, 1991). The absence of any
preserved evaporite facies suggests an abundant supply
of fresh water to the lake, which prevented evaporation
and concentration of salts (Talbot and Allen, 1996). The
associated tuff beds indicate spasmodic volcanic activity

Fig. 10. A) Stratigraphic section of the lacustrine Pozo D-129 Fm in Chenque (Silva) Hill. B) Photomicrograph of well-rounded oolitic grainstones
(Facies LF15). Scale bar is 1 mm. C) Grey mudstones (Facies LF16) and amalgamated sandstone beds (Facies LF17) preserved in a shallow lacustrine
environment.
 J.M. Paredes et al. / Sedimentary Geology 202 (2007) 96123 113

in the hinterland and slow deposition under subaqueous
conditions.
4. Fluvial architecture
Good exposures of the channelized deposits in the
Matasiete Canyon, architectural diagrams and detailed
measurements of stratigraphic profiles were used to
reconstruct the fluvial architecture. The coarse fill of
channels and the rarity of lateral accretion surfaces
(observed in only 2 out of 100 channels) permitted
the real width of each channel to be obtained from its
apparent width and the direction normal to paleoflow. A
compilation of the paleocurrent measurements obtained
from the unit is presented as Fig. 11. In this way,
measurements of channel width/thickness ratios (here-
after abbreviated as W/T ratios) were obtained (Fig. 5)
for each of the three Members in the Matasiete Canyon
and compared to the channel classifications developed
by Friend (1983) and Friend et al. (1979). Most of the
single or multistorey fluvial channels of the Matasiete
Fm could be described as sandstone ribbons using the
criterion (W/T b 15) of Friend et al. (1979). Despite the
large lateral extent of some of the sandbodies, no sheets
with W/T N 100 were identified (see Fig. 12).
W/T ratio b 10. In the uppermost sandbodies the W/T
ratio reaches 35.6 (W/T values range from 2.3 to 35.6).
Multistorey channels of the Lower Member have an
average thickness of 7.2 m (range 2.0 to 16.5 m) and an
average width of 47 m (range 20 to 162 m). The
proportion of channels in two measured sections at
Matasiete Canyon is 27% and 38%. Three main channel
belts, separated by floodplain deposits, are present in the
most sand-rich section of this Member. Paleocurrent
directions within each belt are variable (see Fig. 11), but
dispersion of measurements in single channels is low.
The basal channel belt flows to the northeast (vector
mean = 64, standard deviation = 21, n = 89), the middle
one to the southeast (mean = 174, SD = 21, n = 24), and
the upper belt to the east-southeast (mean = 110, SD =
31, n = 63). Channel sandbodies lack lateral accretion
surfaces, show multiple erosional surfaces, low vari-
ability in the paleoflow in single channels, and frequent
fining-upward trend. This evidence suggests that avul-
sion was an important mechanism for migration of these
fixed, coarse-grained channels. The presence of desic-
cation cracks on top of some amalgamated sandstone
beds suggests complete desiccation of the channel bed
after avulsion or abandonment, and exposure of in-
channel dunes and bars.

4.1. Lower Member 4.2. Middle Member

This Member is exposed in the Matasiete Canyon This Member is exposed in the Matasiete Canyon
only, where most of the multistorey sandbodies have a only. The deposits reflect a dominance of floodplain

Fig. 11. Rose diagrams showing the distribution of mean paleoflow in the three members of the Matasiete Formation. Data from the Lower Member at
the Matasiete Canyon are separated by channel belts. Data from the Middle and Upper Member in the Matasiete Canyon and Sierra Silva are mainly
parallel to the major fault of the basin. Data from the remainder areas (e.g. Sierra Nevada, Tronador and Cachetaman area) are scarce, but show a quite
similar paleoflow direction.
114 J.M. Paredes et al. / Sedimentary Geology 202 (2007) 96123

Fig. 12. A) Thickness variation, mean and standard deviation data from the Upper Member of the Matasiete Formation at several locations. B) W/T
ratio of fluvial sandbodies of the Matasiete and Castillo formations at Matasiete Canyon. The data reflect a reduction of sediment supply during the
deposition of the Castillo Formation and changes in the geometries of the sandbodies.

fines over channelized facies, the latter comprising 15%
and 24% of the total thickness of two measured sections.
Sandbodies in the Middle Member have an average
thickness of 6.5 m (range 2 to 11 m) and an average
width of 134 m (range 65 to 450 m). The channels are
not connected and reflect transportation to the south
(vector mean = 192, SD = 27, n = 27). The W/T ratio is
laterally variable (Fig. 5), but increases toward the top of
the Member (range 14.5 to 75). The largest W/T ratio
(75) is recorded on top of the extensive volcaniclastic
beds shown in Fig. 9.
4.3. Upper Member
The architecture of the Upper Member is similar to
that of the Middle Member, with isolated multistorey
sandbodies encased in thick floodplain deposits. Multi-
storey channels make up 21% and 26% in two measured
sections at Matasiete Canyon, 30% in the Chenque
(Silva) Hill, 28% in the Cachetamn area, 16% in the
Sierra Nevada anticline, and 11% in the Tronador
section. Channel thicknesses and their standard devia-
tions are presented in Fig. 12A. The average thickness of
the channels in the Upper Member at Matasiete Canyon
is 7.0 m (range from 2 to 14.4 m in Matasiete Canyon,
one sandbody of 20 m thick in Silva Hill) and the
average width equals 126 m (range from 50 to 440 m).
The mean paleotransport direction is south-southeast in
the Matasiete Canyon and Sierra Silva anticline (mean =
168, SD = 46, n = 111) but individual sandbodies show
variations of 80 to 100 around the mean. The W/T
ratio of the multistorey channels is less than 20 (see
Fig. 5), except in a unique sandbody characterized by a
W/T ratio of 90. This sandbody is a homogeneous,
coarse-grained bed, directly overlying a white, tabular,
3.5 m thick tuff bed. In the core of the Sierra Nevada
anticline the exposures of the Upper Member of the
Matasiete Fm (Sciutto and Martnez, 1996) show an
average channel thickness of 8.2 m (n = 4), comparable
to the sandbodies in the Matasiete Canyon. Exposures
 J.M. Paredes et al. / Sedimentary Geology 202 (2007) 96123
of the Upper Member in the Tronador Canyon and
Cachetamn Hill are characterized by channels with an
average thickness of 1.6 m (n = 10) and 2.5 m (n = 10),
respectively. The small size of these channels suggests
that these outcrops represent deposits of a less integrated
fluvial system or, alternatively, they could be part of a
tributary system draining to the southern areas (e.g.
Sierra Nevada, Matasiete and Sierra Silva), where the
scale of the sandbodies is about 4 times greater.
4.4. Architecture of the Matasiete and Castillo Fms
Differences in thickness and W/T ratio of the Matasiete
and Castillo Formation are shown in Fig. 12B, which is
based on measurements at Matasiete Canyon. The data
reflect a reduction in thickness of sandbodies in the
Castillo Formation relative to those below and a general
change in channel shape from ribbons (W/T b 15) in the
Matasiete Formation to sheets (W/T N 15) in the Castillo
Formation. These data support the hypothesis that
sediment supply decreased during deposition of the
Castillo Formation (Hechem and Strelkov, 2002), but
further studies are needed to clearly understand the role of
the ash-dominated floodplain and other controls (vegeta-
tion, discharge, changing paleogeography at the basin
margins) on the behavior of the fluvial systems of the
Castillo Formation, which is beyond the scope of this
contribution.
5. Paleosols
More than 70% of the sections of the Matasiete
Fm consist of mudstone and siltstone, and frequently
the quality of the exposures is poor. The Upper Member
was examined at Matasiete Canyon, where a detailed
analysis of a representative section (20 m thick) was
carried out. Yellowbrown and red siltstoneclaystone
beds are the dominant lithologies in the studied section
(Fig. 13). Smectite and analcime are the dominant
components in the b2 m size fraction, and low pro-
portions of illite are present in two samples. Red and
yellowbrown mudstones frequently have a blocky or
prismatic texture (N 20 mm), although the yellow
brown strata show a finer blocky texture (b 10 mm).
Carbonate concretions form discrete nodules ranging
from 5 to 40 cm in diameter, or rods of 10 to 25 cm in
diameter and 100 to 150 cm length. Locally, coalescent
rods and nodules are found (Stage I, II and III sensu
Machette, 1985). Yellowbrown mudstones contain
slickensides, which are found as clay-lined fractures
that intersect to form concave-up, dish-shaped struc-
tures. On the top of some soil profiles (A horizon)
115
there are desiccation cracks to 0.25 m deep, filled by a
mixture of unsorted sandstone and mudstone or lami-
nated beds. Rhizoliths (Klappa, 1980; Kraus and
Hasiotis, 2006) a few millimeters in diameter are
common both in the red and the yellowbrown beds.
Elongate, light-grey rims (mottling) with circular cross-
sections are common; they represent drab haloes
(locally reduced zones) around root traces (Retallack,
1990). XRD analyses showed that the red to yellow
brown colour reflects the relative proportions of hema-
tite (Fe2O3) and goethite (FeO(OH)), respectively
(Cornell and Schwertmann, 2003). One sample contains
jarosite (KFe3(SO4)2(OH)6). The paleosols record well-
developed stress features (e.g. pseudo-anticlines, well
developed blocky or prismatic ped structure and deep
desiccation cracks) typically associated with intense
expansion and contraction (FitzPatrick, 1980; Birke-
land, 1984), features frequently attributed to vertisols
(Marriot and Wright, 1993). The common occurrence of
well-developed carbonate rods and nodules also indi-
cates low sedimentation rates and seasonal oscillation
of the water table (Allen, 1986; Marriot and Wright,
1993). These observations support the semiarid cli-
mate conditions suggested by Hechem et al. (1987)
based on microfossils in the time-equivalent lacustrine
Pozo D-129 Fm. Paleosols developed on pyroclastic
strata only show vertical root traces and burrows at-
tributed to Planolites, Skolithos and rare Taenidium;
these incipient paleosols without horizons were inter-
preted as inceptisols.
6. Fluvial styles
The inferred depositional processes, facies and litho-
facies associations, as well as architecture and paleocur-
rents of the Matasiete Fm, have been used to interpret the
evolution of fluvial styles within the units. The appli-
cation of discrete classification schemes to fluvial sys-
tems is problematic because they are based on different
properties, and aim at simplifying the continuum of
natural processes (Miall, 1996). The well-known sub-
division into straight and meandering rivers is based on
sinuosity. Other classification schemes are based on
sinuosity and sediment type (Schumm, 1963, 1977),
floodplain behaviour (Nanson and Croke, 1992) or
channel style (Woolfe and Balzary, 1996). Our approach
to the analysis of fluvial architecture is based on channel
planform and rates of channel migration relative to
floodplain aggradation rate.
Interpretation of fluvial style was based on the pres-
ence of: (a) coarse-grained infill of many of the sand-
bodies; (b) ribbon geometries of most of the sandbodies;
116 J.M. Paredes et al. / Sedimentary Geology 202 (2007) 96123

Fig. 13. Stratigraphic and chemical data for fine-grained rocks from the Upper Member at Matasiete Canyon (location in Fig. 5).
 J.M. Paredes et al. / Sedimentary Geology 202 (2007) 96123
Fig. 14. Schematic environmental reconstruction for continental deposits of the Matasiete and Pozo D-129 formations. Vertical scale is greatly
exaggerated.
(c) low dispersion in paleocurrent vectors in single
channels; (d) stories displaying vertical reduction in grain
size and internal structure, and preservation of rippled
cross-strata in the upper parts; (e) scarce sandbodies
containing lateral accretion bedding; (f) laminated mud-
stones on top of fining-upward trends related to lateral
accretion bedding in multistorey meandering channels;
(g) rare desiccation cracks on top of single channels;
(h) oriented tree trunks of 1 m in diameter along channel
margins, likely incorporated as bedload; (i) high, but
variable, mudstone to sandstone ratio, and (j) vertisols in
the distal floodplain, which contain desiccation cracks,
slickensides and carbonate nodules.
Most of the above features indicate that transporta-
tion occurred primarily during high-discharge flood
events, with evidence of gradual abandonment in some
multistorey channels. The paucity of deep incisions and
the preservation of interchannel deposits indicate that
aggradation rates in the channel belts exceeded those of
the floodplain. Most of the studied sandbodies corre-
spond to Field 1 of Woolfe and Balzary (1996).
117
The proportion of sandbodies in several sections from
the three Members varies from 11% to 38%. Up to 80%
of the channels analyzed in the seven studied areas are
coarse-grained, lenticular, lack lateral accretion surfaces
and display low dispersion of paleoflow measurements
(b50). We consider these sandbodies to be straight
(Rust, 1978). The presence of large oriented trunks
located at the margins of these channels suggests that
riverbanks were vegetated. A few channels contain
epsilon cross-stratification and lateral accretion surfaces,
have a fining-upward trend and display a considerable
variation in paleoflow measurements ( 80100 around
the vector mean), indicating the presence of helicoidal
flow patterns and meandering behaviour (Miall, 1996). It
should be noted that such features could have been
produced in meandering or braided rivers (Allen, 1983;
Bristow, 1987). Single episodes of lateral migration are
separated from each other by pale-grey laminated
mudstones several tens of centimeters thick, suggesting
that channel beds were periodically inactive. A few
channels are filled by cobble to pebble conglomerates,
118 J.M. Paredes et al. / Sedimentary Geology 202 (2007) 96123
have high W/T ratio (N 50) and are volumetrically
dominated by the stacking of in-channel dunes. These
show very low dispersion in paleocurrents, and are
regarded as braided rivers. This pattern developed on top
of tabular, pyroclastic ash-fall deposits or pyroclastic
surge deposits, which suggests that the modification
from single channels to a multichannel system (braided)
was strongly controlled by substrate characteristics, and
bears no relation to base-level changes.
The fluvial systems of the Matasiete Fm were pre-
viously interpreted as meandering (Sciutto, 1981) and
braided to meandering (Galeazzi, 1989), while changes
in fluvial style were attributed to base-level changes.
Paredes et al. (2003, 2004) also favoured an interpre-
tation in terms of braided and meandering fluvial styles,
but pointed out that many multistorey channel fills did
not fit into this classification. The presence of straight,
meandering and braided channels in the Matasiete Fm
suggests that different reaches of the fluvial system have
had a different pattern along their course.
Both anastomosing rivers (Nadon, 1994; Makaske,
2001) and anabranching rivers (Nanson and Knighton,
1996; Tooth and Nanson, 1999) share many sedimen-
tological attributes with the Matasiete Fm, notably the
large proportion of overbank and floodplain deposits,
the presence of straight, meandering and braided chan-
nels, the predominance of ribbon-like geometries and
the poorly developed upward-fining trends in many
channels fills. We believe that a hybrid model better
represents the complex features observed in this gravel-
dominated fluvial system (Fig. 14), but the occurrence
of the Matasiete Fm in the core of anticlines, with few
three-dimensional exposures, and the impossibility of
proving that individual channel-fills were formed by
coeval active channels make doubtful the categorization
of the unit into one of these fluvial styles.
7. Discussion: external controls on sedimentation
7.1. Tectonism
The increase in accommodation during deposition of
the lacustrine facies of the Pozo D-129 Fm (Barre-
mian?Aptian) is considered to be related to extensional
or transtensional tectonics (Figari et al., 1999). Evidence
includes rapid thickness variation and facies changes
in the subsurface of the basin. The original tectonic
relationships are obscured in the outcrops of the Pozo
D-129-Matasiete depositional system along the San
Bernardo Fold Belt, because of the strong tectonic
inversion and extensive flood basalts covering the
landscape (see Fig. 1C). The only indirect evidence of
tectonic control on sedimentation is the near-orthogonal
pattern of paleocurrent directions recorded between
three consecutive channel belts of the Lower Member
(Fig. 11), which indicate an alternation of transporta-
tion to the NE (64), S (174) and E-SE (110). These
changes could be related to nodal avulsions, or represent
the entry of fluvial systems transversal and longitudinal
to the NNW-SSE oriented ancient normal faults. The
Middle and Upper Members show a consistent paleo-
flow direction to the south at Matasiete Canyon and
Sierra Silva (Middle Member = 192; Upper Member =
168), broadly parallel to the tectonic structure.
7.2. Base-level changes and depositional sequences
Galeazzi (1989) considered the Lower and Middle
Member as part of a unique depositional sequence, and
located the sequence boundary at the base of the chan-
nels that cover the volcaniclastic beds in the topmost
part of the Middle Member. The Upper Member was
considered to be part of a second depositional sequence.
By contrast, Hechem et al. (1990) considered the Lower
Member as a single depositional sequence, and included
the Middle and Upper Members in a second depositional
sequence. The Middle Member was interpreted as a
retrogradational system and the Upper Member as a
progradational system, both related to episodes of ex-
pansion and contraction of the lake. However, field
evidence did not allow the recognition of sequence
boundaries according to the criteria proposed by Leeder
and Steward (1996) and Shanley and McCabe (1994).
Paredes et al. (2003) rejected the concept of base-
level control on channel geometry, and instead noted
a relationship between sandbody geometry and the oc-
currence of tabular pyroclastic beds in the Middle and
Upper Members at Matasiete Canyon, which is unrelated
to base level. There are no substantial changes in fluvial
architecture (eg. mean thickness of the channels, W/T
ratio, and sandstone/mudstone ratio) between the Middle
Member (retrogradational sets) and Upper Member
(progradational sets). Moreover, most of the sandbodies
are located towards the upper part of the Upper Member
(Fig. 5, see Fig. 14 in Paredes et al., 2003), which reflects
a reduction in the rate of accommodation, or higher rates
of avulsion in the uppermost Upper Member.
7.3. Climate
The climate during deposition of the Matasiete-Pozo
D-129 system was deduced from the presence of Clas-
sopolis pollen, and the common occurrence of oolitic
grainstones and Botryococcus-like algal forms in the
 J.M. Paredes et al. / Sedimentary Geology 202 (2007) 96123 119

Fig. 15. Geometry of the fluvial sandbodies and their relation to the supply of pyroclastic material. A) Single or multistorey channels carved in an
epiclastic floodplain show straight and rare meandering patterns. B) Channels are wider and display a braided pattern after pyroclastic supply events.

lacustrine facies (Hechem et al., 1987; Van Nieuwen-
huise and Ormiston, 1989). These observations suggest
a perennial and stratified salinealkaline lacustrine sys-
tem in a semiarid climate. Paleosols indicate a seasonal
climate, as inferred from the presence of desiccation
cracks, blocky and prismatic structure, slickensides and
carbonate nodules, all of which characterize present-day
vertisols (FitzPatrick, 1980).
Fluvial channels of the Lower and Upper members
contain large tree-trunk fragments. Their sizes imply the
existence of vegetated channel margins, or high rates of
precipitation in the source area. We favour the former
interpretation, based on (a) the good preservation of
the larger trunks at the margins of the channels, (b) the
lacustrine facies interpretation, and (c) the common
occurrence of densely vegetated banks of present-day
rivers in semiarid and arid regions (e.g. Tooth and Nanson,
2000). The abundance of pedogenic carbonates in
floodplain deposits of the Upper Member reflects an excess
of alkaline solutions and oscillation of the water table,
which could constitute evidence of increasing aridity during
deposition of the unit. By contrast, there are no appreciable
changes in the size of the sandbodies within the unit, which
suggest no substantial changes in discharge. Fluctuations of
the water table and variable (seasonal?) runoff could
account for low organic-matter content and for the absence
of coal in the floodplain deposits (McCabe and Parrish,
1992), owing to degradation of organic-matter during dry
seasons. Meandering channels located in the Upper
Member show in-channel mud deposition, which reflects
occasional desiccation of the channel bed in response to a
seasonal or ephemeral supply of water and sediment.
7.4. Volcanism
Evidence of intermittent volcanic activity during de-
position of the Chubut Group comprises fine-grained
pyroclastic beds and locally ignimbrites (Sciutto, 1981;
120 J.M. Paredes et al. / Sedimentary Geology 202 (2007) 96123
Fitzgerald et al., 1990; Figari et al., 1999; Bridge et al.,
2000; Genise et al., 2002). Most of the volcanic deposits
in the Matasiete Fm reflect distal (pyroclastic) events
related to the early stages of uplift and deformation of the
Andean Chain, located 150200 km westward of the San
Bernardo Fold Belt. Of particular interest are the
pyroclastic surges identified in the upper part of the
Middle Member in view of the inferred location of the
Cretaceous volcanic centres. The geometry and lateral
extent of surges depend on the type of surge, the topo-
graphical control and post-depositional erosion (Cas and
Wright, 1987), but studies of recent eruptions indicate that
maintenance of turbulent suspension of clasts over
distances longer than 30 km is rare. A nearby source of
the surge deposits exposed at the Matasiete Canyon is
inferred from the tuff-dominated surge composition with
well-preserved glass shards, the dominance of dune forms
and wavy layering generated by turbulent flows, and the
presence of accidental clasts and accretionary lapilli.
The abrupt increase in the W/T ratio of sandbodies
deposited directly on top of pyroclastic beds in the Middle
and Upper Member is interpreted as a direct consequence
of the deposition of large quantities of pyroclastic material
on the floodplain and inside channels, which produced
several interrelated changes in the river system (Fig. 15).
Firstly, sediment load abruptly increased due to the supply
of large quantities of loosely consolidated material, either
by fluvial processes or by heavy rains. Additionally, the
low permeability of the fine-grained tephra relative to
soils (Swanson et al., 1982; Smith, 1991) reduced the
infiltration capacity of the floodplain and produced an
increase in the runoff. The high aggradation rate of the
floodplain and the increase of volume and variability of
the sediment load could have increased the avulsion
frequency and produced the observed widening of the
sandbodies (Vessell and Davies, 1981).
In contrast to the widely acknowledged importance
of tectonic and climate controls on the evolution of
continental systems, our field data indicate that episodic
input of pyroclastic material into the basin produced
the most dramatic changes in fluvial architecture of
the Matasiete Formation. However, these changes only
temporarily affected the fluvial system, as demonstrated
by the rapid return to previous geometries and sizes of
the channelized sandbodies.
8. Conclusions
Detailed studies of lithofacies and facies associa-
tions in the Pozo D-129-Matasiete depositional system
(HauterivianAptian) have been carried out in the San
Bernardo Fold Belt, where seventeen epiclastic, chem-
ical and pyroclastic lithofacies were identified. Two
facies associations were identified in the Pozo D-129
Fm, which represent shallow and deep-lacustrine en-
vironments. Five facies associations were identified in
the fluvial Matasiete Fm: single fluvial channels, multi-
storey fluvial channels, proximal floodplain, distal
floodplain, and pyroclastic deposits.
The lacustrine system of the Pozo D-129 Fm de-
veloped under semiarid climate conditions in a basin
without connection to the sea. The lacustrine rocks in
the San Bernardo Fold Belt indicate low gradient basin
margins, coarsening-upward and shallowing facies
trends and occasional excess of alkaline solutions, with
precipitation of oolitic grainstones.
Near the basin margin, the fluvial Matasiete Fm is
represented by single or multistorey channels encased in
red and yellowbrown coloured floodplain fines with
occasional pyroclastic deposits. The most complete sec-
tion of the unit is exposed in the Matasiete Canyon,
where it reaches a thickness of 650 m, and three mem-
bers (Lower, Middle and Upper Member) have been
distinguished based on stacking patterns and mudstone
to sandstone ratio.
The Lower Member consists of a 205 m thick section,
with three main channel belts separated by thick
floodplain deposits. Most channels of this Member are
coarse-grained (pebble conglomerates and coarse sand-
stones), with low dispersion (b50) of paleocurrents.
However, the channel belts show near-orthogonal
changes in paleoflow directions. In-channel dunes are
common in single and multistorey channels. They display
rare fining-upward trends and occasional sedimentary
structures generated under low-velocity conditions.
Desiccation cracks are preserved on top of some multi-
storey channels, suggesting variable runoff and complete
desiccation of the channel bed. Most channels are straight
and appear to have migrated by avulsion during floods.
The Middle Member is a 215 m thick section
dominated by floodplain deposition in laterally extensive,
very shallow-waters or in vegetated lowlands. Several
sandbodies are coarse-grained, lenticular, and have no
diagnostic features. The paleotransport direction was to
the south, and most of the bodies display low paleocurrent
variability. The W/T ratio is variable through the Member,
but the largest W/T ratio is recorded directly on top of
tabular, pyroclastic surges, which appear to have
provoked a modification of the channels from narrow
ribbons to wider and shallower braided rivers.
The Upper Member consists of a 230 m thick section
containing isolated sandbodies encased in thick red-
coloured floodplain deposits. Channels with straight,
meandering or braided behaviour are recorded in different
 J.M. Paredes et al. / Sedimentary Geology 202 (2007) 96123
sections of the Member at the San Bernardo Fold Belt.
Transportation was to the south with a considerable
dispersion in individual sandbodies. The scarce multistorey
channels with meandering behaviour reveal occasional
desiccation of the channel bed. Paleosols analyzed in the
floodplain of the Upper Member could be considered as
vertisols, due to the presence of deep desiccation cracks and
slickensides, as well as the abundance of carbonate con-
cretions and caliche horizons. These features reflect oscil-
lations of the water table, an excess of alkaline solutions
and a semiarid (seasonal) climate. The absence of coal and
the low organic-matter content suggest variable runoff and
degradation of organic-matter during dry seasons. Tree
trunks up to 1 m in diameter and 15 m long, preferentially
located near channel margins, provide evidence for vege-
tated riverbanks.
The most dramatic changes in fluvial architecture were
provoked by intermittent pyroclastic deposition on the
floodplain. The fluvial system changed from single,
ribbon channels to a short-lived braided river, which
notably increased the W/T ratio of sandbodies. The re-
duced infiltration capacity of the floodplain and the
concomitant increase in sediment load increased runoff, as
well as volume and flashiness of the discharge, which
favoured the development of shallow, multichannel rivers.
Acknowledgements
We are indebted to Jorge F. Rodriguez, Ren E.
Hudecek and Ral E. Giacosa for field assistance and
constructive discussions. We thank V. Funes for XRD
analysis, M. Barqun and A. Saiz (Chemical Department,
UNPSJB) for chemical determinations and XRF analysis
in paleosol samples. The Departamento de Geologa of
the UNPSJB is acknowledged for logistic support. This
paper is a contribution to the PI CIUNPAT 521. Partial
funding by the Research Project BTE2002-04316-C03-
01 and Accin Complementaria CGL-2004-20409 del
Ministerio de Educacin y Ciencia, Spain, and by the
Grup de Qualitat de la Generalitat de Catalunya, DURSI,
2001-SGR-00074 are gratefully acknowledged. The
comments of two anonymous reviewers contributed to
the clarity of this paper. Editors Gert Jan Weltje and
Christopher Fielding are thanked for their very con-
structive comments and editorial assistance.
References
Abdul Aziz, H., Sanz-Rubio, E., Calvo, J.P., Hilgen, F.J., Krijgsman,
W., 2003. Palaeoenvironmental reconstruction of a middle Miocene
alluvial fan to cyclic shallow lacustrine depositional system in the
Calatayud Basin (NE Spain). Sedimentology 50, 211236.
121
Alexander, J., Leeder, M.R., 1987. Active tectonic control on alluvial
architecture. In: Ethridge, F.G., Flores, R.M., Harvey, M.D. (Eds.),
Recent Developments in Fluvial Sedimentology. Special Publica-
tion, vol. 39. Society of Economic Paleontologists and Miner-
alogists, pp. 243259.
Alexander, J., Bridge, J.S., Cheel, R.J., LeClair, S.F., 2001. Bedforms
and associated sedimentary structures formed under supercritical
water flows over aggrading sand beds. Sedimentology 48,
133152.
Allen, J.R.L., 1983. Studies in fluviatile sedimentation: bars, bar-
complexes and sandstone sheets (low sinuosity braided streams) in
the Brownstones (L. Devonian), Welsh Borders. Sedimentary
Geology 33, 237293.
Allen, J.R.L., 1986. Pedogenic calcretes in the Old Red Sandstone
facies (Late SilurianEarly Carboniferous) of the Anglo-Welsh
area, southern Britain. In: Wright, V.P. (Ed.), Paleosols: Their
Recognition and Interpretation. Princeton Univ. Press, Princeton,
NJ, pp. 139179.
Archangelski, S., Baldn, A., Gamerro, J.C., Sller, J., 1984. Palinologa
estratigrfica del Cretcico de Argentina Austral. III. Distribucin de
las especies y conclusiones. Ameghiniana 21, 1533.
Arche, A., Lopez-Gomez, J., 1999. Tectonic and geomorphic controls
on the fluvial styles of the Eslida Formation, Middle Triassic,
Eastern Spain. Tectonophysics 315, 187207.
Barcat, C., Cortias, J., Nevistic, V., Zucchi, H., 1989. Cuenca del Golfo
San Jorge. In: Chebli, G., Spalletti, L. (Eds.), Cuencas Sedimentarias
Argentinas, pp. 319345.
Bellosi, E.S., Gonzalez, M., Genise, J., 2002. Paleosuelos y
sedimentacin cretcica de la cuenca San Jorge (Grupo Chubut)
en la Sierra de San Bernardo, Patagonia Central. 15 Congreso
Geolgico Argentino, vol. 2, 747753 (Calafate).
Birkeland, P.W., 1984. Soils and Geomorphology. Oxford Univ. Press,
Oxford. 372 pp.
Blum, M.D., Tornqvist, T.E., 2000. Fluvial response to sea-level
change: a review and look forward. Sedimentology 47, 248.
Bridge, J.S., 1993. Description and interpretation of fluvial deposits: a
critical perspective. Sedimentology 40, 801810.
Bridge, J.S., Leeder, M.R., 1979. A simulation model of alluvial
stratigraphy. Sedimentology 26, 617644.
Bridge, J.S., Jalfin, G.A., Georgieff, S.M., 2000. Geometry, lithofacies,
and spatial distribution of Cretaceous fluvial sandstone bodies, San
Jorge Basin, Argentine: outcrops analog for the hydrocarbon-bearing
Chubut Group. Journal of Sedimentary Research 70, 319337.
Bristow, C.S., 1987. Brahmaputra River: channel migration and
deposition. In: Ethridge, F.G., Flores, R.M., Harvey, M.D. (Eds.),
Recent Developments in Fluvial Sedimentology. Special Publica-
tion, vol. 39. Society of Economic Paleontologists and Miner-
alogists, pp. 6374.
Carey, S.N., 1991. Transport and deposition of thepra by pyroclastic
flows and surges. In: Fisher, R.V., Smith, G.A. (Eds.), Sedimen-
tation in Volcanic Settings. Special Publication, vol. 45. Society of
Economic Paleontologists and Mineralogists, pp. 3957.
Cas, R.A., Wright, J.V., 1987. Volcanic Successions, Modern and
Ancient. Allen and Unwin, London. 528 pp.
Cecil, N., 2003. The concept of autocyclic and allocyclic controls on
sedimentation and stratigraphy, emphasizing the climatic variable.
In: Blaine Cecil, N., Terence Edgar, N. (Eds.), Climate Controls on
Stratigraphy. Special Publication, vol. 77. Society of Economic
Paleontologists and Mineralogists, pp. 1320.
Clavijo, R., 1986. Estratigrafa del cretcico inferior en el sector
occidental de la Cuenca Golfo San Jorge. Boletn de Informaciones
Petroleras 9, 1532 (Buenos Aires).
122 J.M. Paredes et al. / Sedimentary Geology 202 (2007) 96123
Collinson, J.D., 1996. Alluvial sediments, In: Reading, H.G. (Ed.),
Sedimentary Environments: Processes, Facies and Stratigraphy,
3rd Ed. Blackwell, Oxford, pp. 3782.
Cornell, R.M., Schwertmann, U., 2003. The Iron Oxides: Structure,
Properties, Reactions, Occurrences, and Uses. Wiley-VCH, Wein-
heim. 664 pp.
Emery, D., Myers, K., 1996. Sequence Stratigraphy. Blackwell,
Oxford. 297 pp.
Feruglio, E., 1949. Descripcin Geolgica de la Patagonia. Yacimien-
tos Petrolferos Fiscales 1, 1334 (Buenos Aires).
Fielding, C.R., 2006. Upper flow regime sheets, lenses and scout-fills:
extending the range of architectural elements for fluvial sediment
bodies. Sedimentary Geology 190, 227240.
Figari, E., Strelkov, E., Laffife, G., Cid de la Paz, M., Courtade, S.,
Celaya, J., Vottero, A., Lafourcade, P., Martinez, R., Villar, H.,
1999. Los sistemas petroleros de la Cuenca del Golfo San Jorge:
Sntesis estructural, estratigrfica y geoqumica. 4 Congreso de
Exploracin y Desarrollo de Hidrocarburos, pp. 197237.
Fisher, R.V., Schmincke, H.U., 1984. Pyroclastic Rocks. Springer-
Verlag, Berln. 472 pp.
Fisher, R.V., Smith, G.A. (Eds.), 1991. Sedimentation in Volcanic Settings.
Special Publication, vol. 45. Society of Economic Paleontologists and
Mineralogists, pp. 1257.
Fitzgerald, M.G., Mitchum, R.M., Uliana, M.A., Biddle, K.T., 1990.
Evolution of the San Jorge Basin, Argentina. American Associ-
ation of Petroleum Geologists Bulletin 74, 879920.
FitzPatrick, E.A., 1980. Soils. Longman, New York. 353 pp.
Folguera, A., Iannizzotto, N., 2004. The Lagos La Plata and Fontana
fold-and-thrust belt: long-lived orogenesis at the edge of western
Patagonia. Journal of South American Earth Sciences 16,
541566.
Friend, P.F., 1983. Toward the field classification of alluvial architecture
or sequence. In: Collinson, J.D., Lewin, J. (Eds.), Modern and
Ancient Fluvial Systems. Special Publication, vol. 6. International
Association of Sedimentologists, pp. 345354.
Friend, P.F., Slater, M.J., Williams, R.C., 1979. Vertical and lateral
building of river sandstone bodies, Ebro basin, Spain. Journal of
the Geological Society (London) 136, 3946.
Galeazzi, J.S., 1989. Anlisis de facies y paleocorrientes de la
Formacin Matasiete en la seccin Sur del Caadn homnimo: Lesta, P., Ferello, R., 1972. Regin Extraandina del Chubut y norte de
Sierra de San Bernardo, Chubut, Argentina. Msc. Thesis, UBA.
Buenos Aires, Argentina, pp. 182.
Gawthorpe, R.L., Leeder, M., 2000. Tectono-sedimentary evolution of
active extensional basins. Basin Research 12fs, 195218.
Genise, J.F., Sciutto, J.C., Laza, J.H., Gonzalez, M.G., Bellosi, E.S.,
2002. Fossil bee nests, coleopteral pupal chambers and tuffaceous
paleosols from the Late Cretaceous Laguna Palacios Formation,
Central Patagonia (Argentina). Palaeogeography, Palaeoclimatol-
ogy, Palaeoecology 177, 215235.
Hechem, J.J., 1994. Modelo predictivo de reservorios en un sistema
fluvial efmero del Chubutiano de la cuenca del Golfo San Jorge,
Argentina. Asociacin Argentina de Sedimentologa Revista 1, 314.
Hechem, J.J., 1998. Arquitectura y paleodrenaje del sistema fluvial
efmero de la Formacin Bajo Barreal, cuenca del Golfo San Jorge,
Argentina. 1 Congreso Latinoamericano de Sedimentologa,
pp. 315323. Isla Margarita.
Hechem, J.J., Strelkov, E.E., 2002. Secuencia sedimentaria mesozoica
del Golfo San Jorge. In: Haller, J.M. (Ed.), Geologa y recursos
Naturales de Santa Cruz. 15 Congreso Geolgico Argentino,
vol. 1, pp. 129147. Buenos Aires.
Hechem, J.J., Figari, E.G., Musacchio, E.A., 1987. Hallazgo de la
Formacin Pozo D-129. Petrotecnia 28 (11), 1315.
Hechem, J.J., Homovc, J.F., Figari, E.G., 1990. Estratigrafa del
Chubutiano (Cretcico) en la Sierra de San Bernardo, cuenca del
Golfo San Jorge, Argentina. 11 Congreso Geolgico Argentino,
vol. 3, pp. 173176. San Juan.
Holbrook, J., Schumm, S.A., 1999. Geomorphic and sedimentary
response of rivers to tectonic deformation: a brief review and
critique of a tool for recognizing subtle epeirogenic deformation in
modern and ancient settings. Tectonophysics 305, 287306.
Homovc, J.F., Conforto, G.A., Lafourcade, P.A., Chelotti, L.A., 1995.
Fold belt in the San Jorge Basin, Argentine: an example of tectonic
inversion. In: Buchanan, J.G., Buchanan, P.G. (Eds.), Basin Inver-
sion. Special Publication, vol. 88. Geological Society of London,
pp. 235248.
Klappa, C.F., 1980. Rhizoliths in terrestrial carbonates: classification,
recognition, genesis and significance. Sedimentology 26, 613629.
Kraus, M.J., 1999. Paleosols in clastic sedimentary rocks: their
geological applications. Earth-Science Reviews 47, 4170.
Kraus, M.J., Hasiotis, S.T., 2006. Significance of different modes of
rhizolith preservation to interpreting paleoenvironmental and
paleohydrologic settings: examples from Paleogene paleosols,
Bighorn basin, Wyoming U.S.A. Journal of Sedimentary Research
76, 633646.
Leeder, M.R., Steward, M.D., 1996. Fluvial incision and sequence
stratigraphy: alluvial responses to relative sea-level fall and their
detection in the geological record. In: Hesselbo, S.P., Parkinson, D.N.
(Eds.), Sequence Stratigraphy in British Geology. Special Publica-
tion, vol. 103. Geological Society of London, pp. 2539.
Legarreta, L., Uliana, M., 1998. Anatomy of hinterland depositional
sequences: Upper Cretaceous fluvial strata, Neuqun basin, West-
central, Argentina. In: Shanley, K., McCabe, P. (Eds.), Relative
Role of Eustasy, Climate and Tectonism in Continental Rocks.
Special Publication, vol. 59. Society of Economic Paleontologists
and Mineralogists, pp. 8392.
Legarreta, L., Uliana, M., Larotonda, C.A., Meconi, G.R., 1993.
Approaches to nonmarine sequence stratigraphy theoretical
models and examples from Argentine basins. In: Eschard, R., Doliez,
B. (Eds.), Subsurface Reservoir Characterization from Outcrop
Observations. Collection Colloques et Sminaires, vol. 51. Editions
Technip, Paris, pp. 125145.
Santa Cruz. In: Leanza, A., Academia Nacional de Ciencias (Eds.)
Geologa Regional Argentina: 601654. Crdoba.
Link, M.H., Osborne, R.H., 1978. Lacustrine facies in the Pliocene Ridge
basin, California. In: Matter, A., Tucker, M.E. (Eds.), Modern and
Ancient Lake Sediments. Special Publication, vol. 2. International
Association of Sedimentologists, pp. 167187.
Lunt, I.A, Bridge, J.S., Tye, R.S., 2004. A quantitative, three dimensional
depositional model of gravely braided rivers. Sedimentology 51,
377414.
Machette, M.N., 1985. Calcic soils of the southwestern United States.
In: Weide, D.L. (Ed.), Soils and Quaternary Geology of the South
West United States. Special Paper, vol. 203. Geological Society of
America, pp. 121.
Makaske, B., 2001. Anastomosing rivers: a review of their classification,
origin and sedimentary products. Earth-Science Reviews 53, 149196.
Marriot, S.B., Wright, V.P., 1993. Paleosols as indicators of geomorphic
stability in two Old Red Sandstone alluvial suites, South Wales.
Journal of the Geological Society (London) 150, 11091120.
Mathisen, M.E., Vondra, C.F., 1983. The fluvial and pyroclastic
deposits of the Cagayan Basin, northern Luzon, Philippines an
example of nonmarine volcaniclastic sedimentation in an interarc
basin. Sedimentology 30, 369392.
 J.M. Paredes et al. / Sedimentary Geology 202 (2007) 96123
McCabe, P.J., Parrish, J.T., 1992. Tectonic and climatic controls on
the distribution of Cretaceous coals. In: McCabe, P.J., Parrish, J.T.
(Eds.), Controls on the Distribution and Quality of Cretaceous
Coals. Special Paper, vol. 267. Geological Society of America,
pp. 115.
Meconi, G., 1990. Facies, arquitectura fluvial y paleoambientes
del Grupo Chubut en el Codo del Ro Senguer, lmite provincial
de Chubut-Santa Cruz. 3 Reunin Argentina de Sedimentologa,
pp. 193196. San Juan.
Miall, A.D., 1996. The Geology of Fluvial Deposits: Sedimentary
Facies, Basin Analysis and Petroleum Geology. Springer-Verlag,
Berlin. 582 pp.
Nadon, G.C., 1994. The genesis and recognition of anastomosed fluvial
deposits: data from the St. Mary River formation, Southwestern
Alberta, Canada. Journal of Sedimentary Petrology 64, 451463.
Nanson, G.C., Croke, J.C., 1992. A genetic classification of
floodplains. Geomorphology 4, 459486.
Nanson, G.C., Knighton, A.D., 1996. Anabranching rivers: their cause,
character and classification. Earth Surface Processes and Land-
forms 21, 217239.
Paredes, J.M., Hudecek, R., Foix, N., Rodriguez, J.F., Nillni, A., 2003.
Anlisis paleoambiental de la Formacin Matasiete (Aptiano) en su
rea tipo, noroeste de la Cuenca del Golfo San Jorge, Argentina.
Asociacin Argentina de Sedimentologa Revista 10, 82101.
Paredes, J.M., Foix, N., Hudecek, R., Rodriguez, J.F.R., Nillni, A.,
2004. Arquitectura y estilos fluviales de la Formacin Matasiete,
Aptiano de la Cuenca del Golfo San Jorge, Argentina. 10 Reunin
Argentina de Sedimentologa, pp. 129130. San Luis.
Peroni, G., Hegedus, A., Cerdan, J., Legarreta, L., Uliana, M., Laffite,
G., 1995. Hydrocarbon accumulation in an inverted segment of the
Andean Foreland: San Bernardo Belt, Central Patagonia. In:
Tankard, A., Suarez, R., Welsink, H. (Eds.), Petroleum basins of
South America. American Association of Petroleum Geologists
Bulletin, vol. 62, pp. 403419.
Platt, N.H., Wright, V.P., 1991. Lacustrine carbonates: facies models,
facies distributions and hydrocarbon aspects. In: Anadn, P.,
Cabrera, L., Kelts, K. (Eds.), Lacustrine Facies Analysis. Special
Publication, vol. 13. International Association of Sedimentolo-
gists, pp. 5774.
Ramsay, J.G., 1961. The effects of folding upon the orientation of
sedimentary structures. Journal of Geology 69, 84100.
Retallack, G.J., 1988. Field recognition of paleosols. In: Reinhardt, J.,
Sigleo, W.R. (Eds.), Paleosols and Weathering Through Geologic
Time: Principles and Applications. Special Paper, vol. 216. Geological
Society of America, pp. 120.
Retallack, G.J., 1990. Soils of the Past. Unwin-Hyman, London. 520 pp.
Rodriguez, J.F., 1992. Interpretacin paleoambiental de la Formacin
Bajo Barreal (Cretcico tardo) en Estancia Ocho Hermanos, Chubut.
4 Reunin Argentina de Sedimentologa, vol. 2, pp. 8188. La Plata.
Rodriguez, J.F., Littke, R., 2001. Petroleum generation and accumu-
lation in the Golfo San Jorge Basin, Argentina: a basin modeling
study. Marine and Petroleum Geology 18, 9951028.
Rust, B.R., 1978. A classification of alluvial channels systems. In:
Miall, A.D. (Ed.), Fluvial Sedimentology. Canadian Society of
Petroleum Geology, Memoir, vol. 5, pp. 187198.
Schumacher, R., Schmincke, H.U., 1995. Models for the origin of
accretionary lapilli. Bulletin of Volcanology 56, 626639.
Schumm, S.A., 1963. A tentative classification of alluvial river
channels. US Geological Survey Circular 477.
Schumm, S.A., 1977. The Fluvial System. Wiley, New York. 338 pp.
123
Schumm, S.A., 1993. River response to baselevel change: implication
for basin stratigraphy. Journal of Geology 101, 279294.
Sciutto, J.C., 1981. Geologa del Codo del Ro Senguerr, Chubut,
Argentina. 8 Congreso Geolgico Argentino, vol. 3, pp. 203219.
San Luis.
Sciutto, J.C., Martnez, R.D., 1996. El Grupo Chubut en el Anticlinal
Sierra Nevada, Chubut, Argentina. 8 Congreso Geolgico Argen-
tino y 3 Congreso Geolgico de Exploracin de Hidrocarburos,
vol. 1, pp. 6775. Salta.
Shanley, K.W., McCabe, P.J., 1994. Perspectives on the sequence
stratigraphy of continental strata. American Association of Petroleum
Geologists Bulletin 78, 544568.
Shanley, K.W., McCabe, P., 1998. Relative role of eustasy. Climate and
Tectonism in Continental Rocks. Special Publication, vol. 59.
Society of Economic Paleontologists and Mineralogists, pp. 1234.
Slingerland, R., Smith, N.D., 2004. River avulsion and their deposits.
Annual Review of Earth and Planetary Sciences 32, 257285.
Smith, G.A., 1987. The influence of explosive volcanism on fluvial
sedimentation: the Deschutes Formation (Neogene) in central
Oregon. Journal of Sedimentary Petrology 57, 613629.
Smith, G.A., 1991. Facies sequences and geometries in continental
volcaniclastic sediments. In: Fisher, R.V., Smith, G.A. (Eds.),
Sedimentation in Volcanic Settings. Special Publication, vol. 45.
Society of Economic Paleontologists and Mineralogists, pp. 109121.
Smith, N.D., Cross, T.A., Dufficy, J.P., Clough, S.R., 1989. Anatomy
of an avulsion. Sedimentology 36, 124.
Swanson, F.J., Collins, B., Dunne, T., Wicherski, B.P., 1982. Erosion
of tephra from hillslopes near Mt. St. Helens and other volcanoes.
Proceedings of the Symposium on Erosion Control in Volcanic
Areas, Public Works Research Institute, Japan, Technical Mem-
orandum, vol. 1908, pp. 183221.
Talbot, M.R., Allen, P.A., 1996. Lakes, In: Reading, H.G. (Ed.),
Sedimentary Environments: Processes, Facies and Stratigraphy,
3rd Ed. Blackwell, Oxford, pp. 83124.
Tooth, S., Nanson, G.C., 1999. Anabranching rivers on the Northern
Plains of arid central Australia. Geomorphology 29, 211233.
Tooth, S., Nanson, G.C., 2000. The role of vegetation in the formation
of anabranching channels in an ephemeral river, Northern plains,
arid central Australia. Hydrological Processes 14, 30993117.
Uliana, M., Legarreta, L., 1999. Jursico y Cretcico de la Cuenca del
Golfo San Jorge. In: Caminos, R. (Ed.), Geologa Argentina.
Subsecretara de Minera de la Nacin, Buenos Aires, pp. 496510.
Van Nieuwenhuise, D.S., Ormiston, A.R., 1989. A model for the origin
of source-rich lacustrine facies, San Jorge Basin, Argentina. 1
Congreso Nacional de Exploracin de Hidrocarburos, vol. 2,
pp. 853883.
Vessell, R.K., Davies, D.K., 1981. Non-marine sedimentation in an
active forearc basin. In: Ethridge, F.G., Flores, R.M. (Eds.), Recent
and Ancient Nonmarine Depositional Environments: Model for
Exploration. Special Publication, vol. 31. Society of Economic
Paleontologists and Mineralogists, pp. 3145.
Woolfe, N.J., Balzary, J.R., 1996. Fields in the spectrum of channel
style. Sedimentology 43, 797805.
Wright, D., 1999. The role of sulphate-reducing bacteria and cyanobac-
teria in dolomite formation in distal ephemeral lakes of the Coorong
region, South Australia. Sedimentary Geology 126, 147157.
Wright, V.P., Marriot, S.B., 1993. The sequence stratigraphy of fluvial
depositional systems: the role of floodplain sediment storage.
Sedimentary Geology 86, 203210.
 Available online at www.sciencedirect.com

Sedimentary Geology 202 (2007) 124 137

www.elsevier.com/locate/sedgeo

Characterizing avulsion stratigraphy in ancient alluvial deposits

H.L. Jones , E.A. Hajek
Department of Geology and Geophysics, University of Wyoming, Laramie Wyoming 82071, USA

Abstract

Guidelines for identifying ancient avulsion deposits were set forth by Kraus and Wells [Kraus, M.J., Wells, T.M., 1999.
Recognizing avulsion deposits in the ancient stratigraphical record. In: Smith, N.D., Rogers, J. (Eds.), Fluvial Sedimentology VI,
Special Publication of the International Association of Sedimentologists, vol. 28, pp. 251268], building on the study by Smith
et al. [Smith, N.D., Cross, T.A., Dufficy, J.P., Clough, S.R., 1989. Anatomy of an avulsion. Sedimentology 36, 123] of the
modern Saskatchewan River system (Cumberland Marshes, central Canada), and serve to characterize avulsion depositional
sequences in the ancient Willwood and Fort Union Formations (Paleogene, Bighorn Basin, NW Wyoming, USA). We recognize,
however, that the model is not universally applicable to avulsion-dominated successions, specifically systems which lack defining
heterolithic avulsion deposits, set forth by Kraus and Wells [Kraus, M.J., Wells, T.M., 1999. Recognizing avulsion deposits in the
ancient stratigraphical record. In: Smith, N.D., Rogers, J. (Eds.), Fluvial Sedimentology VI, Special Publication of the International
Association of Sedimentologists, vol. 28, pp. 251268]. Observations in several fluvial intervals suggest that the avulsion
stratigraphy outlined by Kraus and Wells [Kraus, M.J., Wells, T.M., 1999. Recognizing avulsion deposits in the ancient
stratigraphical record. In: Smith, N.D., Rogers, J. (Eds.), Fluvial Sedimentology VI, Special Publication of the International
Association of Sedimentologists, vol. 28, pp. 251268] represents one category of avulsion stratigraphy found in the rock record,
but does not capture the nature of avulsion deposits everywhere. Based on observations (using measured sections, outcrop photo-
panels, and aerial photographs) in the Willwood Formation (Eocene, Wyoming) and Ferris Formation (Cretaceous/Paleogene,
Wyoming), we present two end-member categories of avulsion stratigraphy in ancient deposits; stratigraphically abrupt, when a
main paleochannel is stratigraphically juxtaposed directly atop floodplain/overbank deposits, and stratigraphically transitional,
where crevasse splays and other non-floodplain/-overbank deposits stratigraphically precede a main paleochannel. This
characterization provides a broader, more inclusive way to recognize and describe avulsion stratigraphy in ancient deposits and
may be an important factor to consider when modeling connectivity in fluvial reservoirs. Furthermore, our observations show that
one type of avulsion channel stratigraphy may prevail over another within an ancient basin, suggesting that system-wide factors
such as splay-proneness or avulsion style (i.e. aggradational, incisional, etc.; [Slingerland, R., Smith, N.D., 2004. River avulsions
and their deposits. Annual Review of Earth and Planetary Sciences 32, 257285]) may be primary controls on the type of avulsion
stratigraphy deposited and preserved in ancient basin-fills.
2007 Elsevier B.V. All rights reserved.

Keywords: Avulsion; Stratigraphy; Alluvial architecture; Willwood Formation; Ferris Formation

1. Introduction

Avulsion is a dominant depositional mechanism

Corresponding author. active during alluvial basin filling. Avulsions affect
E-mail address: hazellee@uwyo.edu (H.L. Jones). alluvial architecture, specifically depositional channel
0037-0738/$ - see front matter 2007 Elsevier B.V. All rights reserved.
doi:10.1016/j.sedgeo.2007.02.003
 H.L. Jones, E.A. Hajek / Sedimentary Geology 202 (2007) 124137
stacking patterns, and therefore can have great influence
on the character of the stratigraphy preserved. River
avulsions are rarely observed on human timescales,
however ancient alluvial basins contain thousands of
manifestations of channel avulsions and provide a rich
database with which to study this phenomenon. Existing
avulsion-related research includes studies of modern
and Holocene avulsions (e.g. Fisk, 1944; Smith et al.,
1989; Trnqvist, 1994; Jones and Harper, 1998; Sling-
erland and Smith, 1998; Aslan and Autin, 1999;
Ethridge et al., 1999; Morozova and Smith, 1999;
Stouthamer, 2001; Slingerland and Smith, 2004). These
studies provide details of how channels redistribute
water and sediment during an avulsion, however it is
unclear if and how these results can be extrapolated to
longer timescales and the rock record. Theoretical and
experimental modeling have been used to improve
understanding of how avulsions fill basins over longer
timescales (thousands to millions of years) (e.g. Allen,
1978; Leeder, 1978; Bridge and Leeder, 1979; Bryant
et al., 1995; Mackey and Bridge, 1995; Heller and
Paola, 1996; Slingerland and Smith, 2004); however,
these models do not constrain the actual character and
structure of preserved sediments associated with channel
avulsion, and the ancient rock record remains relatively
unexploited as a tool for studying avulsions at this scale.
Relatively few descriptions have been published of
ancient (pre-Quaternary) avulsed channels and/or their
stratigraphic context in the rock record (e.g., Kraus and
Gwinn, 1997; Kraus and Wells, 1999; Mohrig et al.,
2000). Guidelines for identifying ancient avulsion
deposits were proposed by Kraus and Wells (1999),
who characterize depositional sequences in the ancient
Willwood and Fort Union Formations and build upon the
avulsion model of Smith et al. (1989) that was derived
from detailed study of the modern Saskatchewan River
system. Kraus' and Wells' (1999) main criterion for
recognizing avulsion deposits is that heterolithic
avulsion deposits underlie trunk channels. Heterolithic
avulsion deposits are laterally extensive, consist of weak,
pedogenically modified, fine-grained deposits, and
encase ribbon (width:depth b 10) and thin sheet sand-
stones which show paleoflow parallel or subparallel to
the trunk channels. In addition, heterolithic avulsion
deposits are surrounded by coal or moderately well-
developed to well-developed paleosols. These observa-
tions match depositional successions documented in
modern avulsions, particularly the Saskatchewan River
(Smith et al., 1989). Because of this convincing
connection between ancient and modern, the Kraus and
Wells (1999) model has become a popular way to
interpret ancient alluvial stratigraphy. Alternatively,
125
Mohrig et al. (2000) studied the GuadalopeMatarranya
system (Oligocene, Ebro Basin, Spain) and Wasatch
Formation (Eocene, western Colorado), both of which
apparently lack heterolithic avulsion deposits and
instead contain channels incised directly into floodplain
deposits. These observations can be confusing because it
is unclear whether the incised channels represent: 1)
channels that have enlarged and reworked/eroded
associated heterolithic avulsion deposits; 2) ribbon
sandstones contained within heterolithic avulsion depos-
its; or 3) a different type of avulsion process (incisional
avulsion, Mohrig et al., 2000).
Part of the confusion of characterizing avulsion
stratigraphy in the ancient is one of defining avulsion.
Kraus and Wells (1999) define ancient avulsions by
recognizing heterolithic avulsion deposits in a strati-
graphic succession. This recognition criterion assumes
that all avulsions occur by successive sediment and water
diversion from the trunk channel via widespread
crevasse splay progradation onto the adjacent floodplain.
River avulsion however is more generally defined as an
established channel relocating to a new position on the
floodplain (e.g., Mohrig et al., 2000; Slingerland and
Smith, 2004), regardless of the specific processes by
which sediment and water are diverted during channel
relocation. We propose that this basic definition is an
appropriate starting point for interpreting ancient stra-
tigraphy and use this broader definition to recognize
ancient avulsions as the appearance of a main paleo-
channel (identified by having sedimentary structures
indicative of persistent flow) within a stratigraphic suc-
cession. In order to better document the range and
character of avulsion stratigraphy in ancient deposits, we
document the sediment surrounding a main paleo-
channel and the channel's stratigraphic context.
We compare avulsion stratigraphy in two well-exposed
ancient fluvial successions (the lower Eocene Willwood
Formation and the Maastrichtian/Paleocene Ferris For-
mation) using measured sections, outcrop photo-panels,
and aerial photographs. Two distinct manifestations of
avulsion stratigraphy were observed. Stratigraphically
transitional channels are directly preceded by non-
overbank deposits, including crevasse splays, crevasse
channels, distributary channels, distributary mouth bars,
etc. The non-overbank deposits preceding the channel
may contain the suite of heterolithic avulsion deposits
as defined by Kraus and Wells (1999). In contrast,
channels juxtaposed directly atop floodplain/overbank
deposits are classified as stratigraphically abrupt. There
is no evidence of heterolithic avulsion deposits or other
non-overbank deposits in proximity to channels within
the stratigraphically abrupt successions we observed.
126 H.L. Jones, E.A. Hajek / Sedimentary Geology 202 (2007) 124137
This categorization of avulsion stratigraphy is widely
applicable in ancient deposits allowing even those
successions that lack heterolithic avulsion deposits to be
recognized and described as avulsion-dominated inter-
vals. In this study, and perhaps that of Mohrig et al.
(2000), stratigraphically abrupt channels are not likely a
simple case of channel-reworking and removal of
underlying heterolithic avulsion deposits (e.g. Stoutha-
mer, 2001), but rather represent a distinct type of
avulsion stratigraphy. Stratigraphically abrupt fluvial
successions are probably more common in the rock
record than has been previously recognized and may
reflect differences in splay-proneness and avulsion
styles between systems exhibiting stratigraphically
transitional or stratigraphically abrupt avulsion stratig-
raphy. From an applied standpoint, these two styles
of avulsion stratigraphy have different implications
for predicting and modeling reservoir architecture and
connectivity.
2. Methods
In order to use the rock record to understand the
avulsion process over basin-filling timescales, it is
important to identify and compare main paleochannels
that represent probable avulsions of Stouthamer
(2001). Probable avulsions include full, partial, and
failed avulsions, which are indistinguishable in ancient
deposits because at long timescales, dating evidence
cannot be used to document the timing of parent channel
abandonment (Stouthamer, 2001). In ancient stratigra-
phy, where there is no evidence of allogenic cut and fill
cycles, isolated coarse-grained channel-belt deposits
encased in fine-grained material are considered the
result of the avulsion process (Kraus and Wells, 1999;
Mohrig et al., 2000). For this study, the stratigraphic
context of such paleochannels (i.e. probable paleoavul-
sions) was characterized using a combination of
measured sections, outcrop photo-panels, and aerial
photographs of the Willwood and Ferris Formations.
Main paleochannels/channel belts (i.e. trunk chan-
nels of Kraus and Wells, 1999) are identified as
relatively large, coarse-grained channel bodies with a
basal scour surface and sedimentary structures indicat-
ing persistent flow conditions; including clinoforms,
internal scour surfaces and stories, pervasive cross-beds,
and thalweg deposits (coarse-grained lags, trough cross-
beds, and/or soft-sediment deformation at the base of the
channel body). A probable avulsion is presumed to have
occurred where a single channel or channel-belt deposit
is separated laterally and vertically from other channel
deposits in a stratigraphic cross-section, and is sur-
rounded by mostly fine-grained deposits. The strati-
graphic context of such channels their avulsion
stratigraphy is described.
The deposits surrounding paleochannels are classified
as either overbank or non-overbank. Overbank deposits
include fine-grained material, commonly laminated
mudstones, associated with overbank flooding (Prez-
Arlucea and Smith, 1999; Mohrig et al., 2000). Non-
overbank deposits include crevasse splays (e.g., Jorgen-
sen and Fielding, 1996) and crevasse channels formed by
breaches in levees, ephemeral floodplain features formed
by overland flow (e.g. Taylor, 1999), and/or deposits
associated with avulsion; such as heterolithic deposits
(Kraus and Wells, 1999), distributary channels and bars
(Smith et al., 1989), or small deltas in floodplain lakes
(Tye and Coleman, 1989). These non-overbank deposits
range from mudstone to sandstone and typically contain
ripple- and dune-scale tabular cross-beds and planar
bedding, and show evidence of intermittent flow. Both
overbank and non-overbank deposits may exhibit mod-
ification by floodplain processes including bioturbation
and varying degrees of paleosol development.
The Willwood and Ferris Formations are well suited
for this study because of their extensive exposure. Fur-
thermore, throughout both study intervals there is
no evidence of major extrinsic controls on paleochan-
nel formation and distribution. For example, paleoflow
depths, grain-size, sediment composition, and paleochan-
nel dimensions are consistent within the Willwood and
Ferris Formations. Additionally, there are no major un-
conformities or changes in floodplain or paleosol char-
acter, suggesting that tectonic, climate, or base-level
changes were not primary controls on channel bodies.
This is consistent with conclusions drawn by Mohrig et al.
(2000) in the GuadalopeMatarranya system and the
Wasatch Formation.
3. Field areas
3.1. Willwood Formation
The lower Eocene Willwood Formation was depos-
ited during the tectonic development of the Bighorn
Basin (NW Wyoming, USA) in the latter part of the
Laramide Orogeny, and consists of northerly flowing
fluvial systems that drained the Bighorn, Beartooth, and
Owl Creek mountains (Bown, 1980; Brown, 1993)
(Fig. 1). During Willwood Formation deposition the
Laramide Orogeny was waning, and several large
climatic shifts occurred, including the Paleocene
Eocene Thermal Maximum event and the Eocene
Warm Interval (Bao et al., 1999; Koch et al., 2003).
 H.L. Jones, E.A. Hajek / Sedimentary Geology 202 (2007) 124137 127

Fig. 1. Location map of study areas. The two shaded regions in the main figure are the Bighorn Basin in northern Wyoming and the Hanna Basin in
southern Wyoming. Abbreviations are as follows: PPolecat Bench; MMcCullough Peaks; EElk Creek (Willwood Formation study areas), and
HHanna, Wyoming (Ferris Formation study area). Map modified from Fig. 2 of Dickinson et al. (1988).

The Willwood Formation consists of channel-belt
sandstones, crevasse splay deposits, proximal overbank
deposits, and floodplain mudstones with well-developed
paleosols. Channel-belt sandstones are typically fine-to
medium-grained, 10 to 25 m thick, 1 to 1.5 km wide and
consist of 2 or more channel storeys that are individually
46 m thick on average (Fig. 2AB). Internally,
channel-belt sandstones exhibit abundant soft-sediment
deformation, tabular and trough cross-bedding, massive
bedding, bioturbation, and occasional clinoforms
(Fig. 2CH). Typical paleoflow depths are estimated to
be 24 m based on clinoform and cross-bed heights (e.g.
Paola and Borgman, 1991; Leclair et al., 1997; Bridge and
Tye, 2000). The margins of the channel sandstones are
usually associated with well-developed levee deposits
consisting of fine-grained, thinly bedded, bioturbated,
ripple-dominated overbank deposits. Overall, channel
sandstones comprise only 20% of the Willwood
Formation (Kraus and Wells, 1999), the rest of which
consists of weakly developed paleosols, crevasse splay
deposits and well-developed paleosols. The weakly
developed paleosols are clay-rich, drab in color and
weakly weathered with remnants of planar bedding.
Abundant crevasse splay deposits are commonly found
within intervals associated with weakly developed
paleosols. Crevasse splay deposits are typically thin
sandstones that consist of climbing ripples, planar
bedding, contorted bedding and evidence of bioturbation
and rooting near the top. Well-developed paleosol
horizons in the Willwood Formation are typically red to
reddish-purple in color and contain slickensides, burrow-
ing, mottling, and rooting throughout their vertical
thickness and encase weakly developed paleosols and
crevasse splay deposits. The crevasse splay deposits and
associated weakly developed paleosols are the heterolithic
avulsion deposits of Kraus and Wells (1999). Field
locations of measured sections and outcrop photo panels
are shown in Fig. 1. The Willwood Formation outcrop
photos in Fig. 2 were taken in the Elk Creek study area
(E in Fig. 1) and the Willwood Formation stratigraphic
128 H.L. Jones, E.A. Hajek / Sedimentary Geology 202 (2007) 124137

Fig. 2. Willwood Formation. A) An outcrop photo panel of the Willwood Formation taken at McCullough Peaks study area (M in Fig. 1). Note well-
developed paleosols marked by banding and relatively few channel-belt sand bodies. Line denotes location of stratigraphic section presented in Fig. 5.
B) Outcrop photo panel of a typical Willwood Formation channel belt. Note repeated vertical succession of splays and channel-belt sand bodies. CH)
Internal sedimentary structures typical within the Willwood Formation: C) decimeter scale tabular cross-bedding coset; D) Small-scale climbing
ripples; E) vertical meniscate burrow; F) preserved leaf litter; G) cast of a vertical burrow; H) palaeosol carbonate nodules, maximum diameter 1 cm.
Photo scale is cm on left and inches on right.
 H.L. Jones, E.A. Hajek / Sedimentary Geology 202 (2007) 124137 129

section in Fig. 5 was measured in the McCullough Peaks
study area (M in Fig. 1).
3.2. Ferris Formation
The Ferris Formation was deposited in the Hanna
Basin (south-central Wyoming, USA) by fluvial sys-
tems draining late Sevier and Laramide highlands
southeast of the retreating western interior seaway
(Cannonball Sea) in the latest Cretaceous through
early Paleocene (Weimer, 1984) (Fig. 1). As a result
of late Laramide structural deformation (Lillegraven
et al., 2002), the Ferris Formation in the Hanna Basin
has been tilted such that stratigraphy is dipping nearly
vertically into the ground. Consequently this unit is
well exposed and can be observed in cross section
from air photos (Fig. 3AB). In the Hanna Basin, the
Ferris Formation ranges from marginal marine and
estuarine in the lower portion of the section to wholly
fluvial in the upper half (Lillegraven et al., 2002).
This study focuses on the fluvial portion in the upper
section.

Fig. 3. Ferris Formation. A) Aerial photograph of the fully fluvial section of the Ferris Formation (location H in Fig. 1). The sand bodies appear as
white and the dark hues are fine-grained floodplain deposits and vegetation. Due to post-depositional tectonic tilting of nearly 90, the deposits are
now vertical, with stratigraphic up towards the top of the photo, and paleoflow direction is into the ground. The thick black line shows the location of
the stratigraphic section presented in Fig. 5. B) View of the Ferris Formation looking east. Note the distinct separation of sand bodies by muddy
floodplain material in both A and B. C) The Ferris Formation is typified by channel-belt incision directly into floodplain material. Here, the hammer is
resting on dune-scale trough cross bedding at the base of a channel-belt. The feature cut directly into fine-grained floodplain material (contact
indicated by white line). Paleoflow direction is away from viewer. D) Typical Ferris Formation muddy floodplain material. It is slightly laminated and
shows little or no pedogenic modification.
130 H.L. Jones, E.A. Hajek / Sedimentary Geology 202 (2007) 124137
The fluvial interval is mudstone-dominated with
isolated channel-belt sandstones and conglomerates
distributed throughout the deposit (Heller et al., 2004,
2005). Channel sandstones (coarse-to very coarse-
grained) and conglomerates (clasts up to 9 cm in
diameter) are typically 1.54 m thick, several tens to
hundreds of meters in width (measured roughly
perpendicular to paleoflow measurements), and contain
abundant trough cross-beds and poorly developed
clinoforms. Channel deposits are commonly 12 storeys
and estimates from clinoform and cross-bed height
suggest average paleoflow depth was 12 m. Mudstones
separate individual sand bodies and conglomerates, and
crevasse splay deposits are rare and are never found
directly below a channel deposit (Fig. 3CD). The field
location of the measured section shown in Fig. 5 and
aerial photographs in Fig. 3 are located in the Hanna
Basin study area (indicated by H in Fig. 1).
4. Avulsion stratigraphy
4.1. Two generalized types
Based on observations in the Willwood and Ferris
Formations, we propose that avulsion stratigraphy can be
categorized into two generalized types: stratigraphically
transitional and stratigraphically abrupt. These terms are
chosen to describe the stratigraphy below avulsed paleo-
channels. It is possible that stratigraphically transitional and
stratigraphically abrupt channels reflect progradational/
aggradational and incisional styles of avulsion, respectively
(refer to discussion), but too little is known about avulsion
and its stratigraphic record to invoke process-interpretative
terminology. These observations were only applied to main
channel/channel-belt deposits encased in fine-grained
material (i.e. probable paleoavulsions).
Transitional avulsion stratigraphy is characterized by
a generally coarsening-upward sequence of non-over-
bank deposits preceding a main paleochannel (Fig. 4A
and B). Material below the main paleochannel is
dominated by non-overbank deposits and is relatively
fine-grained compared to the overlying channel. The
underlying succession may contain indicators of inter-
mittent flow conditions such as climbing ripples, clay-
rich intervals, bioturbation, rooting, and weak pedogenic
modification. In the specific case of the Willwood and
Fort Union Formations, crevasse splays and other non-
overbank deposits (i.e. heterolithic avulsion deposits)
underlie main/trunk channels (Kraus and Wells, 1999
and this study). The term stratigraphically transitional
refers to the stratigraphic transition, in a vertical
succession, from fine-grained overbank deposits to
coarse-grained main channel deposits, via non-overbank
deposits, which are sandwiched in between.
In contrast, stratigraphically abrupt avulsion stratig-
raphy has coarse-grained channel deposits juxtaposed
directly atop fine-grained floodplain material and lacks
non-overbank deposits underlying the main channel
(Fig. 4C and D). The Ferris Formation, documented in
this study, and the Wasatch Formation and Guadalope
Matarranya system described by Mohrig et al. (2000),
commonly show stratigraphically abrupt paleochannels.
4.2. Willwood Formation
Within the Willwood Formation, channel deposits are
stratigraphically associated with a series of thickening-
and coarsening-up crevasse splay deposits (Fig. 5). The
splay deposits are recognized by a vertical succession of
weakly developed drab colored paleosols alternating
with relatively thin (0.251 m), fine-grained, planar to
rippled to bioturbated sandstone. Occasionally there is
evidence that the main channel slightly incised into the
underlying non-overbank deposits (Figs. 4A and 5).
These observations were consistent spatially (through
the Elk Creek, McCullough Peaks, and Polecat Bench
sections, refer to Fig. 1) and temporally ( 56 Ma to
53 Ma) within the basin. Of twenty-two channels
documented (n = 7 in Elk Creek area, n = 10 in McCul-
lough Peaks area, and n = 5 in the Polecat Bench area)
three channels are characterized as stratigraphically
abrupt and nineteen channels are characterized as
stratigraphically transitional, preserving a dominance
of stratigraphically transitional avulsion stratigraphy
within the Bighorn Basin throughout deposition of the
Willwood Formation. A stratigraphic section from the
McCullough Peaks (M) area is shown in Fig. 5 to
exemplify the character of stratigraphically transitional
channels. A thicker line in the stratigraphic column
outlines each channel and its associated non-overbank
deposits. One or more splay deposits precede each
channel throughout the 130 m section and each main
paleochannel has evidence of persistent channelized
flow. This measured section is representative of Will-
wood Formation stratigraphy basin-wide.
4.3. Ferris Formation
Throughout the 10 km2 study area of the Ferris
Formation, main paleochannel deposits are found only to
occur directly atop floodplain deposits or unconformably
atop older channel-belt deposits. Most commonly,
channel-belt deposits are juxtaposed directly atop fine-
grained floodplain deposits (Figs. 4C and 5). Where
 H.L. Jones, E.A. Hajek / Sedimentary Geology 202 (2007) 124137 131

Fig. 4. Categories of avulsion stratigraphy: A) Willwood Formation sequence showing an upward thickening and coarsening of crevasse splay
deposits preceding the main avulsed channel above. Note slight scouring into preceding splay deposits at white line. B) Schematic stratigraphic
column of idealized stratigraphically transitional stratigraphy. Note a thickening and coarsening up of crevasse splay deposits in relation to the
overlying main avulsed channel. The main avulsed channel may or may not scour into the preceding deposits. C) Ferris Formation sequence
(upsection is to the right) showing an abrupt transition from fine-grained floodplain deposits with little or no pedogenic modification into a channel
sand. The contact is marked on the photo with a white line. D) Schematic stratigraphic column of idealized stratigraphically abrupt avulsion
stratigraphy. Note the abrupt transition from fine-grained deposit to a coarse-grained channel deposit.

channel-belt sandstones are found overlying older
channel-belt deposits, the base of the younger channel
deposit often truncates underlying channel deposits. The
juxtaposition of channel deposits atop fine-grained
floodplain material or older channel deposits is a
consistent observation throughout the Ferris Formation
in the Hanna Basin. Of thirty Ferris Formation channels
documented to date in the Hanna Basin, all thirty
channels are characterized as stratigraphically abrupt and
none of the Ferris Formation paleochannels are preceded
by non-overbank deposits. Therefore, stratigraphically
abrupt channels dominate the Ferris Formation stratig-
raphy in the Hanna Basin. Fig. 5 shows a 150 m section
measured in the Hanna Basin. It is clear that the channels
are not preceded by non-overbank deposition, which
exemplifies the character of stratigraphically abrupt
132 H.L. Jones, E.A. Hajek / Sedimentary Geology 202 (2007) 124137

Fig. 5. Two measured stratigraphic sections depicting the recognition and character of stratigraphically transitional and stratigraphically abrupt avulsion
stratigraphy. The Whistle Creek section (Willwood Formation) on the left is located in the McCullough Peaks area (M in Fig. 1) and the line of section is
shown in Fig. 2A. The location of the Hanna Basin section (Ferris Formation) on the right is shown by a line of section in Fig. 3A and H in Fig. 1.
Components of avulsion stratigraphy recognition criteria are outlined with thick black lines. Note the propensity of splay/non-overbank deposits and range
of grain sizes in the Whistle Creek section versus the lack of splay deposits and relatively bimodal distribution of grain sizes in the Hanna Basin section.

avulsion stratigraphy. This measured section is repre-
sentative of Ferris Formation stratigraphy in the Hanna
Basin.
5. Discussion
Recognizing two categories of avulsion channels
provides a tool for classifying and communicating about
ancient avulsion deposits and builds upon the common-
ly applied Kraus and Wells (1999) model. Furthermore,
categorizing probable paleoavulsions as either strati-
graphically transitional or stratigraphically abrupt seems
to capture two distinctly different types of avulsions
preserved in the rock record; the former characterized by
an obvious phase of sediment deposition prior to the
formation of the avulsed channel, and the latter
characterized simply by the appearance of the channel
directly onto the floodplain. It is our contention that this
 H.L. Jones, E.A. Hajek / Sedimentary Geology 202 (2007) 124137
difference is an important aspect of ancient avulsion
stratigraphy, and may have implications about the
avulsion process itself. In particular, other authors
have attributed a lack of crevasse splays below avulsed
channels to subsequent channel reworking and erosion
of splay deposits (e.g., Stouthamer, 2001). Our
observations in the Ferris Formation have yet to find
an example of a channel showing transitional stratigra-
phy in the study area. Unless all the channels observed
in the Ferris Formation completely reworked underlying
splay deposits associated with avulsion, this type of
stratigraphy reflects a type of avulsion not characterized
by an initial phase of progradational deposition (e.g.,
Mohrig et al., 2000; Slingerland and Smith, 2004). It is
possible that the Ferris Formation has captured this
incisional style of avulsion.
Classifying avulsion stratigraphy as stratigraphically
transitional or stratigraphically abrupt may provide
important insights into outstanding questions about the
behavior of fluvial systems on intermediate timescales
including: How common are regional versus local
avulsions? Does the propensity for levee crevassing
and splay deposition vary within and between river
systems? What are different mechanisms for river
avulsion and when/where are they likely to occur?
Additionally, stratigraphic avulsion style may have a
significant impact on reservoir connectivity within
fluvial deposits.
5.1. Avulsion distance
It is possible that the presence of stratigraphically
transitional or stratigraphically abrupt paleochannels is a
function of the proximity of the probable paleoavulsion
to the avulsion site. An avulsed channel exposed near
the avulsion site or formed relatively close to the parent
channel may be more likely to occur on top of or amid
non-overbank deposits than an avulsion relocated
further from the parent channel and potentially outside
the zone of splay deposition, or outside the range of the
prograding sediment wedge formed during avulsion.
The former might exhibit stratigraphically transitional
channels, whereas the latter could result in stratigraphi-
cally abrupt avulsion stratigraphy. If avulsion stratigra-
phy is controlled by the proximity of an avulsed channel
to its parent channel, it may indicate the difference
between relatively local avulsions and regional avul-
sions. If this is the case, the distribution and abundance
of channels that are part of stratigraphically transitional
and stratigraphically abrupt successions within a basin
might indicate the relative abundance of local and
regional ancient avulsions. Further research into the
133
spatial distribution of avulsed channel sands within a
basin is needed to test this hypothesis.
We believe our observations of the Willwood and
Ferris Formations are sufficient to determine that
classifying paleochannels as stratigraphically transition-
al or stratigraphically abrupt is not simply a result of
outcrop exposure showing only intervals proximal or
distal to avulsion sites. Rather, our field observations,
which span both the spatial and temporal scales of
deposition in each basin, suggest that avulsion channel
stratigraphy is dominantly stratigraphically transitional
in the Willwood Formation and stratigraphically abrupt
in the Ferris Formation. If avulsion stratigraphy reflects
avulsion distance, this might suggest that Willwood
Formation river systems were dominated by local
avulsions whereas regional avulsions were common in
the Ferris Formation river systems. However, there are
several other factors that could account for system-wide
avulsion stratigraphy and differences between fluvial
systems. These include river splay proneness and
avulsion mechanisms.
5.2. River splay proneness
Differences in avulsion stratigraphy may be a result
of the amount of splay deposits produced within a
fluvial system, or river splay proneness. Aslan and
Autin (1999) suggest that floodplains aggrade via two
processes: overbank flood sedimentation and crevasse
splay deposition. It is possible that systems that produce
stratigraphically transitional channels (e.g. Willwood
Formation fluvial systems) were innately prone to
crevasse splay deposition whereas other systems, such
as the Ferris Formation fluvial systems, are more heavily
dominated by overbank deposition.
Work in modern systems has noted a difference in
association of river systems and river avulsion with
crevasse splay deposition. Aslan and Autin (1999) sug-
gest that floodplain aggradation is dominated by crevasse
splay deposition during intervals of active avulsion of
the Holocene lower Mississippi River. Kraus and Wells
(1999) attribute up to 50% of floodplain aggradation
in the Willwood Formation to avulsion-related non-
overbank deposition. This is similar to the percentage
of crevasse splay deposition during the Holocene
lower Mississippi River avulsions. Such settings are
likely to produce stratigraphically transitional avulsion
stratigraphy.
In contrast, there are other modern river systems
known to be much less splay-prone. Studies in several
modern river systems, such as Narew River (NE
Poland), Okavango fan rivers (Botswana), Fitzroy
134 H.L. Jones, E.A. Hajek / Sedimentary Geology 202 (2007) 124137
River (Australia) and Cooper Creek (Australia) have
documented river avulsions that do not fit the Smith
et al. (1989) avulsion model of river avulsion (Rust and
Legun, 1983; McCarthy et al., 1992; Taylor, 1999;
Knighton and Nanson, 2000; Aslan et al., 2003;
Gradzinski et al., 2003; Fagan and Nanson, 2004).
Rather, these particular rivers have avulsed by scouring
and incising a new pathway via nick-point retreat and/or
overland flow on the floodplain. In particular, a study of
the modern Fitzroy River, northwest Australia (Taylor,
1999) demonstrates that channel avulsion is not
associated with crevasse splay deposition in that system.
Rather, channels relocate into topographic depressions
on the floodplain, which are generated by scouring
during overbank flow. Taylor (1999) suggests that
crevasse splay facies should be rare or absent in the
Fitzroy's stratigraphic record, in part because of levees
with very low topographic relief (800 m wide by b1 m
high), high-magnitude flooding events, and highly
cohesive floodplain material. Such a system could
dominantly produce stratigraphically abrupt channels.
In comparison, Ferris Formation channels lack well-
developed levee deposits and have muddy floodplains,
which might indicate low-relief levees and highly
cohesive banks, respectively.
It has been proposed that, generally, crevasse splay
production may be a function of grain size distribution,
riverbank cohesiveness, levee dimensions, elevation
differences between the channel water surface and
adjacent floodbasin, water table levels, or variable
hydrographs (e.g., Brakenridge, 1988; Nanson and
Croke, 1992; Nanson and Knighton, 1996; Taylor,
1999; Moody and Troutman, 2000), although the
specific relationships between these factors are not
straightforward (e.g., Nanson and Knighton, 1996). The
sedimentary record of ancient avulsion deposits may
help constrain the relative importance of some of these
parameters. For example, in this study, the Willwood
Formation, characterized by common non-overbank
deposits, is a dominantly sandy system, with maximum
grain-sizes of medium sand and abundant silt and mud
deposits. In contrast, the Ferris Formation, which lacks
non-overbank deposits, has an apparently bimodal
grain-size distribution where very coarse (conglomerat-
ic to very coarse sand) grain sizes are found within
channels, and overbank deposits are dominated by very
fine-grained (silt and clay) material (Fig. 5). Perhaps,
when levees were partially breached during flood
events, Ferris Formation rivers lacked sufficient sus-
pended silt and sandy material to deposit splays,
whereas the upper portions of the water column in
Willwood Formation rivers carried significant sus-
pended silt and fine sand which was readily deposited
upon flow expansion during splay events. It is also
possible that highly cohesive, muddy levees, similar to
those of Ferris Formation rivers, would resist crevas-
sing, whereas sandier levees, such as those of Willwood
Formation rivers, would be less cohesive and more
readily give way, spilling sediment and water onto the
floodplain, readily producing non-overbank deposits.
Studying the context and abundance of crevasse splay
deposits in additional ancient systems might help
elucidate which factors are dominant controls on their
deposition.
5.3. Avulsion mechanisms
Avulsion channel stratigraphy may also reflect
different avulsion mechanisms, or the actual process
by which main flow left the parent channel and formed a
new channel. Three avulsion styles are described in the
literature are related to the process of sediment and
water dispersal and distribution during an avulsion
(Slingerland and Smith, 2004). They include avulsion
by incision, avulsion by progradation or aggradation,
and avulsion by annexation (Slingerland and Smith,
2004). Incisional avulsion occurs when the avulsion
begins with incision and scouring into the floodplain
(e.g., Mohrig et al., 2000). This style of avulsion could
produce deposits consistent with stratigraphically abrupt
channels. In contrast, progradational or aggradational
avulsion is characterized by an early stage of deposition
wherein numerous crevasse splays and multi-channeled
distributary networks develop as the parent channel
loses an increasing amount of flow and sediment to a
newly developing avulsion channel (e.g., Smith et al.,
1989; Ethridge et al., 1999).
Stratigraphically transitional avulsion stratigraphy
may reflect the process of progradation and aggrada-
tion during avulsion captured in the rock record, where
transitional non-overbank deposits preserved below
main paleochannels represent the stage of avulsion
where sediment from the main channel progrades out
across and aggrades a portion of the surrounding
floodplain. Likewise, stratigraphically abrupt channels
may represent the incisional style of avulsion, where
new channels are incised directly into the floodplain.
This can occur in settings with erosion-resistant flood-
plains (which encourages significant overland flow
during flooding; Taylor, 1999), by the process of nick-
point retreat (Mohrig et al., 2000), or by an avulsing
channel annexing a channel that is proportionally
smaller (Slingerland and Smith, 2004). Furthermore, a
system showing stratigraphically abrupt channels may
 H.L. Jones, E.A. Hajek / Sedimentary Geology 202 (2007) 124137
not be prone to crevasse splay activity, as discussed
above. In contrast, stratigraphically transitional avul-
sion stratigraphy may represent progradational, or
aggradational, avulsion style which may occur in
splay-prone systems, such as the Willwood Formation.
Avulsion by annexation occurs when an avulsed
channel reoccupies an abandoned channel or appro-
priates an existing channel. There are several possible
stratigraphic manifestations of a channel that avulsed by
annexation if preserved. One scenario requires that the
newly avulsed channel is much smaller than the
previous channel and therefore is under-fit in relation
to the channel morphology of the previous channel.
Stratigraphically, the avulsed channel resides within the
previous channel's deposit and is recognized by a
through-going scour surface delineating the older from
the younger channel deposits and, sometimes, preserv-
ing either splays or fine-grained non-overbank material
below the channel scour. Conversely, a newly avulsing
channel may appropriate a pre-existing channel or
channel deposit that is smaller. The relatively larger
avulsing channel in this scenario might erode any
evidence of the previous channel and appear as
stratigraphically abrupt. It is possible that avulsion by
annexation could produce either stratigraphically tran-
sitional channels (where non-overbank deposits are
sandwiched between an older channel and the newly
avulsed channel) or stratigraphically abrupt channels
(where the avulsed channel incises directly into the
older, underlying channel).
5.4. Implications of avulsion stratigraphy
The style of avulsion as it occurs (i.e. aggradational,
incisional, annexation) and resultant avulsion channel
stratigraphy is dependent upon how the flow was
initially diverted from the parent channel during the
avulsion process. It is widely accepted that the diverted
flow will follow the greatest slope advantage, or the path
of greatest flow efficiency. Many factors contribute and
coalesce so that each avulsion is unique, including the
size and duration of the avulsion itself and how the
avulsion interacts with the character of the flood basin it
is occupying (Slingerland and Smith, 2004). Flood basin
characteristics and conditions include: topographic
relief, presence, absence, or type of vegetation, water-
table level, presence or absence of pre-existing chan-
nels, and substrate erodability. If we were to assume that
such flood-basin characteristics are relatively consistent
and persistent throughout a basin at a given time it is
within reason to hypothesize that one variation of
avulsion stratigraphy would be dominant, although not
135
exclusive, throughout a given basin during a particular
period of deposition. Field observations within the study
areas of the Bighorn Basin and Hanna Basin comply
with this hypothesis, although further research must be
performed in several more basins with avulsion-
dominated stratigraphy in order to assess the universal-
ity of our findings.
Avulsion is one of the main controls on the
distribution of coarse- and fine-grained deposits within
a basin, and therefore is an important control on
hydrocarbon reservoir quality and distribution. Fluvial
reservoirs can be difficult plays to predict for a number
of reasons including non-uniform sandy-body distribu-
tion and subtle lithologic heterogeneities which can
limit pressure communication between two neighboring
sand bodies, even within a predicted flow unit.
Predicting connectivity between channel-scale sand
bodies in fluvial reservoirs is critical to maximizing
sweep efficiency, but is difficult given the high degree of
variability within fluvial systems. Coarsening-upward
non-overbank deposits preceding stratigraphically tran-
sitional channel successions could permit enhanced
fluid flow and increased pressure communication
between fluvial sand bodies. Although we predict that
reservoirs dominated by transitional avulsion stratigra-
phy may be well connected, these underlying deposits
may also act as thief zones, and the channel sand bodies
may lack sufficient seals. Conversely, stratigraphically
abrupt channel sequences consist of channel sands
juxtaposed directly atop floodplain deposits and do not
show evidence of splay deposition prior to channel
avulsion. We therefore predict that the sharp juxtaposi-
tion of channel sands and floodplain fines will seal and
compartmentalize the channel sand bodies creating
potentially limiting connectivity between channels. If
continuing research further supports the hypothesis that
fluvial basin fills are dominated by a particular type of
avulsion channel stratigraphy, recognizing and charac-
terizing the distribution and prevalence of particular
avulsion channel stratigraphy will help provide a new
tool for predicting sand body interconnectedness within
fluvial reservoirs.
6. Conclusions
Two distinct stratigraphic contexts of paleochannels
are recognized in the ancient rock record: stratigraphically
transitional or stratigraphically abrupt. This classification
recognizes that some ancient avulsion stratigraphy
(stratigraphically transitional) contains crevasse splay
and other non-overbank deposits preceding the appear-
ance of the newly avulsed channel, whereas other
136 H.L. Jones, E.A. Hajek / Sedimentary Geology 202 (2007) 124137
avulsion stratigraphy (stratigraphically abrupt) lacks
underlying non-overbank deposits. Our observations
indicate that one type of stratigraphy can be dominant in
a sedimentary basin. Twenty-two Willwood Formation
probable paleoavulsions spatially distributed throughout
the Bighorn Basin (Wyoming, USA) were documented.
Nineteen of the paleochannels are stratigraphically
transitional and three are stratigraphically abrupt. This
dominance of stratigraphically transitional paleochannels
in the Willwood Formation might have been influenced
by splay proneness in Willwood Formation rivers and a
propensity towards an aggradational/progradational style
of avulsion. In contrast, all thirty of the Ferris Formation
paleochannels documented throughout the Hanna Basin,
Wyoming, USA are stratigraphically abrupt. Crevasse
splay and other non-overbank deposits are rare throughout
the basin and were never found directly below a channel
deposit. Ferris Formation fluvial systems may not have
been prone to crevasse splay and other non-overbank
activity and may have relocated dominantly through
incisional style avulsions.
There is a need for future work documenting the
extent to which this trend is apparent in other avulsion-
dominated fluvial successions. Further characterization
and quantitative documentation of the relative abun-
dance and distribution of each type of avulsion channel
stratigraphy in other systems will provide information
needed to understand fully the factors resulting in each
type of stratigraphy, and to constrain how various
intrinsic and extrinsic factors affect the avulsion process.
Furthermore, this work might also result in better
models for understanding and predicting sand body
connectivity of fluvial reservoirs.
Acknowledgements
The authors would like to thank the International
Association of Sedimentologists, the University of
Wyoming Graduate School, and the University of
Wyoming Department of Geology and Geophysics for
providing support for international travel to present this
research at ICFS 2005, Delft, Netherlands and to the
GSA and AAPG foundations for student research
funding that supported much of the field work required
for this research. Additionally, this work was partially
supported by Cabot Oil and Gas, Chevron Energy
Technology Company, ConocoPhillips Subsurface
Technology, and EOG Resources. We are grateful to
Paul Heller and Norm Smith for helpful discussions.
This paper has greatly benefited from constructive and
critical reviews by Susan Marriott, Christopher Fielding
and two anonymous reviewers.
References
Allen, J.R.L., 1978. Studies in fluviatile sedimentation: an exploratory
quantitative model for the architecture of avulsion-controlled
alluvial suites. Sedimentary Geology 21, 129147.
Aslan, A., Autin, W.J., 1999. Evolution of the Holocene Mississippi River
floodplain, Ferriday, Louisiana; insights on the origin of fine-grained
floodplains. Journal of Sedimentary Research 69 (4), 800815.
Aslan, A., White, W.A., Warne, A.G., Guevara, E.H., 2003. Holocene
evolution of the western Orinoco Delta. Venezuela. Geological
Society of American Bulletin 115 (4), 479498.
Bao, H., Koch, P.L., Rumble III, D., 1999. PaleoceneEocene climatic
variation in western North America: evidence from the 18 of
pedogenic hematite. Geological Society of America Bulletin 111,
14051415.
Bown, T.M., 1980. Summary of latest Cretaceous and Cenozoic
sedimentary, tectonic and erosional events, Bighorn Basin,
Wyoming. In: Gingerich, P.D. (Ed.), Early Cenozoic Paleontology
and Stratigraphy of the Bighorn Basin, Wyoming. University of
Michigan Papers on Paleontology, vol. 24, pp. 532.
Brakenridge, G.R., 1988. River flood regime and floodplain stratigra-
phy. In: Baker, V.R., Kochel, R.C., Patton, P.C. (Eds.), Flood
Geomorphology. John Wiley & Sons, New York, pp. 139156.
Bridge, J.S., Leeder, M.R., 1979. A simulation model of alluvial
stratigraphy. Sedimentology 26, 617644.
Bridge, J.S., Tye, R.S., 2000. Interpreting the dimensions of ancient
fluvial channel bars, channels, and channel belts from wireline-
logs and cores. AAPG Bulletin 84, 12051228.
Brown, W.G., 1993. Structural style of Laramide basement-cored
uplifts and associated folds. In: Snoke, A.W., Steidtman, J.R.,
Roberts, S.M. (Eds.), Geology of Wyoming. Mem. Geol. Surv.
Wyo., vol. 5, pp. 19.
Bryant, M., Falk, P., Paola, C., 1995. Experimental study of avulsion
frequency and rate of deposition. Geology 23 (4), 365368.
Dickinson, W.R., Klute, J.A., Hayes, M.J., Janecke, S.U., Lundin, E.R.,
McKittrick, M.A., Olivares, M.D., 1988. Paleogeographic and
paleotectonic setting of Laramide sedimentary basins in the central
Rocky Mountain region. Geological Society of America Bulletin
100, 10231039.
Ethridge, F.G., Skelly, R.L., Bristow, C.S., 1999. Avulsion and
crevassing in the sandy, braided Niobrara River: complex response
to base-level rise and aggradation. In: Smith, N.D., Rogers, J. (Eds.),
Fluvial Sedimentology VI. Special Publication of the International
Association of Sedimentologists, vol. 28, pp. 179192.
Fagan, S.D., Nanson, G.C., 2004. The morphology and formation of
floodplain surface channels, Cooper Creek, Australia. Geomor-
phology 60, 107126.
Fisk, H.N., 1944. Geological Investigations of the Alluvial Valley of
the Lower Mississippi River. U.S. Army Corps of Engineers,
Waterways Experimentation Station, Mississippi River Commis-
sion, Vicksburg, MS. 78 pp.
Gradzinski, R., Baryla, J., Doktor, M., Gmur, D., Gradzinski, M.,
Kedzior, A., Paszkowski, M., Soja, R., Zielinkski, T., Zurek, S.,
2003. Vegetation-controlled modern anastamosing system of the
upper Narew River (NE Poland) and its sediments. Sedimentary
Geology 157 (34), 253276.
Heller, P.L., Paola, C., 1996. Downstream changes in alluvial
architecture: an exploration of controls on channel-stacking
patterns. Journal of Sedimentary Research 66 (2), 297306.
Heller, P.L., Jones, H.L., Paola, C., 2004. Autocyclic avulsion clusters
in alluvial basins: examples from the Maastrictian/Paleocene of
Wyoming. GSA Abstracts with Programs 36 (5), 462.
 H.L. Jones, E.A. Hajek / Sedimentary Geology 202 (2007) 124137
Heller, P.L., Hajek, E.A., Jones, H.L., Paola, C., Sheets, B.A., 2005.
Avulsion Clusters in Alluvial Basins: Examples from Wyoming,
USA. 8th International Conference on Fluvial Sedimentology,
Delft, Netherlands, p. 127.
Jones, L.S., Harper, J., 1998. Channel avulsions and related processes,
and large-scale sedimentation patterns since 1875, Rio Grande,
San Luis Valley, Colorado. Geologic Society of America Bulletin
110, 411421.
Jorgensen, P.J., Fielding, C.R., 1996. Facies architecture of alluvial
floodbasin deposits: three-dimensional data from the Upper
Triassic Callide Coal Measures of east-central Queensland,
Australia. Sedimentology 43, 479497.
Knighton, A.D., Nanson, G.C., 2000. Waterhole form and process in
the anastomosing channel system of Cooper Creek, Australia.
Geomorphology 35, 101117.
Koch, P.L., Clyde, W.C., Hepple, R.P., Fogel, M.L., Wing, S.L.,
Zachos, J.C., 2003. Carbon and oxygen isotope records from
paleosols spanning the PaleoceneEocene boundary, Bighorn
Basin, Wyoming. In: Wing, S.L., Gingerich, P.D., Schmitz, B.,
Thomas, E. (Eds.), Causes and Consequences of Globally Warm
Climates in the Early Paleogene. GSA, Boulder, pp. 4963.
Kraus, M.J., Gwinn, B.M., 1997. Facies and facies architecture of
Paleogene floodplain deposits, Willwood Formation, Bighorn
Basin, Wyoming, USA. Sedimentary Geology 114 (14), 3354.
Kraus, M.J., Wells, T.M., 1999. Recognizing avulsion deposits in the
ancient stratigraphical record. In: Smith, N.D., Rogers, J. (Eds.),
Fluvial Sedimentology VI. Special Publication of the International
Association of Sedimentologists, vol. 28, pp. 251268.
Leeder, M.R., 1978. A quantitative stratigraphic model for alluvium,
with special reference to channel deposit density and interconnec-
tedness. Canadian Society of Petroleum Geologists, Memoir 5,
587597.
Leclair, S.F., Bridge, J.S., Wang, F., 1997. Preservation of cross-strata
due to migration of subaqueous dunes over aggrading and non-
aggrading beds; comparison of experimental data with theory.
Geoscience Canada 24, 5566.
Lillegraven, J.A., Snoke, A.W., McKenna, M.C., 2002. Tectonic and
paleogeographic implications of the late Laramide geologic history
in the northeastern corner of Wyoming's Hanna Basin. Rocky
Mountain Geology 39, 764.
Mackey, S.D., Bridge, J.S., 1995. Three-dimensional model of alluvial
stratigraphy: theory and application. Journal of Sedimentary
Research. Section B, Stratigraphy and Global Studies 65 (1), 731.
McCarthy, T.S., Ellery, W.N., Stanistreet, I.G., 1992. Avulsion
mechanism on the Okavango fan, Botswana: the control of a
fluvial system by vegetation. Sedimentology 39, 779795.
Mohrig, D., Heller, P.L., Paola, C., Lyons, W.J., 2000. Interpreting
avulsion process from ancient alluvial sequences: Guadalope
137
Matarranya system (northern Spain) and Wasatch Formation
(western Colorado). GSA Bulletin 112 (12), 17871803.
Moody, J.A., Troutman, B.M., 2000. Quantitative model of the growth
of floodplains by vertical accretion. Earth Surface Processes and
Landforms 25, 115133.
Morozova, G.S., Smith, N.D., 1999. Holocene avulsion history of the
lower Saskatchewan fluvial system, Cumberland Marshes, Sas-
katchewanManitoba, Canada. In: Smith, N.D., Rogers, J. (Eds.),
Fluvial Sedimentology VI. Special Publication of the International
Association of Sedimentologists, vol. 28, pp. 231250.
Nanson, G.C., Croke, J.C., 1992. A genetic classification of flood-
plains. Geomorphology 4, 459486.
Nanson, G.C., Knighton, A.D., 1996. Anabranching rivers: their
cause, character and classification. Earth Surface Processes and
Landforms 21, 217239.
Paola, C., Borgman, L., 1991. Reconstructing random topography
from preserved stratification. Sedimentology 38, 553565.
Prez-Arlucea, M., Smith, N.D., 1999. Depositional patterns following
the 1970s avulsion of the Saskatchewan River Cumberland
marshes, Saskatchewan, Canada. Journal of Sedimentary Research
69, 6273.
Slingerland, R., Smith, N.D., 1998. Necessary conditions for a
meandering-river avulsion. Geology 26 (5), 435438.
Slingerland, R., Smith, N.D., 2004. River avulsions and their deposits.
Annual Review of Earth and Planetary Sciences 32, 257285.
Smith, N.D., Cross, T.A., Dufficy, J.P., Clough, S.R., 1989. Anatomy
of an avulsion. Sedimentology 36, 123.
Stouthamer, E., 2001. Sedimentary products of avulsions in the
Holocene RhineMeuse Delta, the Netherlands. Sedimentary
Geology 145 (12), 7392.
Taylor, C.F.H., 1999. The role of overbank flow in governing the form
of an anabranching river: the Fitzroy River, northwestern
Australia. In: Smith, N.D., Rogers, J. (Eds.), Fluvial Sedimentol-
ogy VI. Special Publication of the International Association of
Sedimentologists, vol. 28, pp. 7792.
Trnqvist, T.E., 1994. Middle and late Holocene avulsion history of
the River Rhine (RhineMeuse Delta, The Netherlands). Geology
22, 711714.
Tye, R.S., Coleman, J.M., 1989. Evolution of Atchafalaya lacustrine
delta, south-central Louisiana. Sedimentary Geology 65 (12),
95112.
Weimer, R.J., 1984. Relation of unconformities, tectonics, and sea-level
changes, Cretaceous of Western Interior, U.S.A. In: Schlee, J.S. (Ed.),
Interregional unconformities and hydrocarbon accumulation. Amer-
ican Association of Petroleum Geologists Memoir, vol. 36, pp. 735.
 Available online at www.sciencedirect.com

Sedimentary Geology 202 (2007) 138 157

www.elsevier.com/locate/sedgeo

Provenance of fluvial sandstones at the start of late JurassicEarly

Cretaceous rifting in the Cameros Basin (N. Spain)
Laura Gonzlez-Acebrn a,, Jos Arribas b , Ramn Mas a
a
Dto. Estratigrafa, Facultad de Ciencias Geolgicas (UCM)-Instituto de Geologa Econmica (CSIC),
C/ Jose Antonio Novais 2, 28040 Madrid, Spain
b
Dto. Petrologa y Geoqumica, Facultad de Ciencias Geolgicas (UCM)-Instituto de Geologa Econmica (CSIC),
C/ Jose Antonio Novais 2, 28040 Madrid, Spain

Abstract

The Cameros Basin (Iberian Chain, Central Spain) developed during the latest JurassicEarly Cretaceous in an extensional
regime characterized by high rates of subsidence. Its sedimentary fill has been subdivided into eight depositional sequences (DS)
mainly composed of continental sediments. DS 1 and DS 2 represent the first rifting stage (Tera Group, Tithonian). The purpose of
this study is to characterize the Tera Group in the eastern part of the basin based on provenance criteria derived from fluvial
sandstones. In this area of the basin, the Tera Group can be subdivided into three formations: the greda Formation, the Magaa
Formation and the Sierra de Matute Formation. These formations are composed of alluvial-fan deposits, meandering fluvial
sediments and lacustrinepalustrine mudstones.
A quantitative petrographic study indicated the presence of three main petrofacies in the Tera Group. The close correlation
between petrofacies and lithostratigraphic units indicates that sandstone composition is a powerful tool for deciphering the tectonic
processes active during the initial rift stages of the Cameros Basin.
Petrofacies 1 is sedimentolithic (mean: Qm54F3Lt43) and represents erosion of the Jurassic marine pre-rift substratum (mainly
Kimmeridgian limestones) during deposition of the DS 1 alluvial fan deposits (greda Fm.). Petrofacies 2 is quartzofeldspathic, and can be
subdivided into Petrofacies 2A, with an average composition of Qm84F15Lt1 and Petrofacies 2B, whose average composition is Q71F23Lt6.
Petrofacies 2 was generated by the erosion of low to medium-grade metamorphic terranes and plutonic source rocks. It characterizes the
Magaa Fm. (DS 2). Petrofacies 3 is quartzolithic (mean: Qm67F16Lt17), and is attributed to tectonic reactivation of the basin. This
petrofacies characterizes the Sierra de Matute Fm. (DS 2). Thus, the provenance evolution of this basin is characterized by erosion of the
pre-rift sedimentary substratum, followed by unroofing of the basement, as recorded in other ancient and modern rifted basins.
2007 Elsevier B.V. All rights reserved.

Keywords: Sandstone provenance; Depositional sequences; Tithonian; Cameros Basin; Northern Spain

1. Introduction mining the origin of ancient terrigenous deposits (Blatt,

Sandstone petrography is widely considered to be a
powerful tool for tectonic reconstructions and for deter-
Corresponding author. Tel.: +34 91 3944785; fax: +34 91 394480.
E-mail address: lgcebron@geo.ucm.es (L. Gonzlez-Acebrn).
0037-0738/$ - see front matter 2007 Elsevier B.V. All rights reserved.
doi:10.1016/j.sedgeo.2007.05.008
1967; Dickinson, 1970; Pettijohn et al., 1972). Many
factors, such as source area lithology, relief, climate and
diagenesis affect the final sandstone composition. Several
authors have described a relationship between the detrital
composition of sandstones and the tectonic setting (e.g.
Ingersoll, 1978; Dickinson and Suczek, 1979; Dickinson
et al., 1983; Dickinson, 1985; Valloni, 1985).
 L. Gonzlez-Acebrn et al. / Sedimentary Geology 202 (2007) 138157
The provenance of foreland-basin arenites has been
extensively investigated (Schwab, 1986; DeCelles and
Hertel, 1989; Criteri and Le Pera, 1994; Critelli, 1999).
In this type of basin, large volumes of clastic sediments
are produced by erosion of thrust belts, and sandstone-
provenance analysis provides valuable insights into the
record of thrust movement. Relatively few studies,
however, have examined detailed processes and res-
ponses within intra-plate rift basins (Zuffa et al., 1980;
Evans, 1990; Garzanti et al., 2001; Arribas et al., 2003;
Garzanti et al., 2003). Multi-phase rifting and tilted
crustal blocks lead to erosion and sediment redistribution
within the basin, such that detrital modes of syn-rift
sandstones strongly vary in relation to their paleotectonic
position in the basin. The first stage of rifting usually
starts with erosion of the pre-rift sedimentary substra-
tum, followed by unroofing of the basement. Thus,
sandstone provenance studies are essential for recon-
structing eroded sediments and the tectonic evolution of
rift basins.
The principal aims of this paper are: (1) to describe
and discuss petrologic data derived from fluvial
sandstones (Tera Group) deposited at the beginning of
rifting of the Cameros basin during the late Jurassic
Fig. 1. Geological map of the Cameros Basin indicating the sections: MOV, Montenegro-Villoslada en Cameros (northwestern section) and ALM,
Almajano (southeastern section). Modified from Guimer et al. (1995).
139
Early Cretaceous; (2) to establish the development of
source areas and provide a paleotectonic reconstruction
of the sandstones of the Tera Group in the eastern sector
of the basin; (3) to compare our results with sandstones
signatures of the Tera Group in the western sector of the
basin (Arribas et al., 2003); (4) to discuss provenance
models for intraplate rift basins through a comparison
with modern rifted-margin sands of the Red Sea and
Gulf of Aden (Garzanti et al., 2001, 2003); and (5) to
contribute to a global data base of petrographic prove-
nance data for intra-plate rift basins.
2. Geological and stratigraphic setting
The Cameros Basin in the northern Iberian Range
(Fig. 1) forms part of the Mesozoic Iberian Rift System
(Mas et al., 1993; Guimer et al., 1995; Salas et al., 2001;
Mas et al., 2002, 2003). Intraplate rifting was a conse-
quence of the opening of the oceanic Bay of Biscay, which
separated Iberia from Europe. The Cameros Basin is the
westernmost basin of the Mesozoic Iberian Rift System.
It shows several atypical features including scarce ma-
rine influence and high subsidence and sediment
accumulation rates. Another main feature of this basin is
140 L. Gonzlez-Acebrn et al. / Sedimentary Geology 202 (2007) 138157

Fig. 2. Stratigraphy of the depositional sequences (DS) of the Cameros Basin. The stratigraphic interval examined is indicated (Tera Group, DS 1 and DS 2).
MOV, Montenegro-Villoslada en Cameros (northwestern section) and ALM, Almajano (southeastern section). Modified from Mas et al. (2004).

that a large proportion of the Upper JurassicLower (e.g. Guiraud and Seguret, 1985; Casquet et al., 1992;
Cretaceous deposits of the northwestern area are affected Alonso-Azcarate et al., 1995, 1999; Barrenechea et al.,
by low to very low-grade hydrothermal metamorphism 2000; Mas et al., 2003).
 L. Gonzlez-Acebrn et al. / Sedimentary Geology 202 (2007) 138157
The basin-fill succession of the Cameros Basin
embodies a large cycle or megasequence composed of up
to 5000 m of sediment deposited from the Tithonian to
early Albian. These deposits overlie late Jurassic marine
carbonates, commonly those of the Kimmeridgian Torre-
cilla en Cameros Formation, and are separated from them
by an erosional unconformity with associated paleosols
and/or paleokarst features (Alonso and Mas, 1990; Benito,
2001; Benito and Mas, 2002; Benito et al., 2005). This so-
called infra-Tithonian unconformity is present throughout
the basin and occurs at the base of the Tera Group, which is
the subject of this study. The stratigraphic gap represented
by this unconformity is more important in the northern part
of the basin than in the central and southern areas. In the
northern area of the basin, it extends from the Late
Kimmeridgian to part of the Barremian. Towards the center
and south, it generally extends from the Late Kimmer-
idgian to the Early Tithonian.
The sedimentary infill of the Cameros Basin has been
divided into eight depositional sequences (Mas et al., 2002,
2003) spanning the Tithonian to early Albian (Fig. 2). This
sedimentary record consists of continental sediments
corresponding to alluvial and lacustrine systems, with
very rare marine incursions (Mas et al., 1993; Gmez
Fernndez and Melndez, 1994). The Tera Group rep-
resents the first stage of rifting and is formed by two
depositional sequences (DS 1 and DS 2, Fig. 2), which are
Tithonian in age (Mas et al., 1993, Martn-Closas and
Alonso Milln, 1998; Mas et al., 2004). The thickness and
lateral continuity of DS 1 are highly variable. This depo-
sitional sequence is represented by clastic alluvial facies
and lacustrinepalustrine carbonate facies. In contrast, DS
2 is especially thick, up to 1500 m in the depocenter, and
consists of a clastic fluvial facies which grades upwards
and laterally to carbonate lacustrine facies.
Arribas et al. (2002, 2003) characterized the basin-fill
successions of the western sector of the Cameros Basin
in terms of their clastic constituents and sandstone
sources. These authors identified four main petrofacies
(A, B, C and D) from DS 1 to DS 7. Petrofacies A is
quartzo-sedimentolithic (mean: Qm85F2Lt13) and char-
acterizes DS 1. This petrofacies records the erosion of
Jurassic marine pre-rift cover. Petrofacies B is quartzo-
feldspathic (mean: Qm81F14Lt5) and is derived from the
erosion of metamorphic terranes of the West-Asturian
Leonese Zone during the deposition of DS-2 and DS-3.
The quartzarenitic Petrofacies C (mean: Qm95F3Lt2)
was generated by recycling of the sedimentary cover,
mainly Triassic arkoses and carbonate rocks. Finally,
Petrofacies D is quartzofeldspathic (mean: Qm88F11Lt1),
and attributed to erosion of coarse crystalline plutonic
rocks from the Central Iberian Zone.
141
Analysis of sandstone provenance in the eastern
sector of the Cameros Basin has received less attention.
The present paper focuses on the beginning of rifting
(Tera Group) in this eastern sector.
Two representative stratigraphic sections of the Tera
Group in the eastern sector of the basin were examined
(Fig. 1): one located in the northwest (Montenegro
Villoslada, MOV, 377 m thick) and the other in the
southeast of the study area (Almajano, ALM, 343 m
thick). Based on illite and chlorite crystallinity data
(Barrenechea et al., 2001), the Tera Group deposits of the
MontenegroVilloslada section attained low-anchizonal
conditions, whereas the Almajano section deposits only
suffered diagenetic conditions.
In both study areas, the Tera Group can be subdivided
into three formations: greda, Magaa and Sierra de
Matute (Mas et al., 1993; Gmez Fernndez and Melndez,
1994) (Figs. 2 and 3). The main sedimentological
characteristics of DS 1 and DS 2 in the sections examined
can be summarized as follows:
Depositional sequence 1
greda Fm. (DS 1). In the southeastern section
(ALM), the greda Fm. is composed of conglomerates
including abundant carbonate-rock fragments that pass
upwards into coarse-medium grained sandstones with
trough cross bedding, interbedded with reddish mud-
stones containing calcrete profiles. The greda Fm.
corresponds to a distal system of alluvial fans with sub-
stantial development of meandering channels and mud
flats. These alluvial fans are related to the onset of
rifting. A high rate of vertical accretion is indicated by
the high ratio of floodplain to channel facies. In the
northwestern section (MOV), this formation is com-
posed of nodular limestones, with root marks and a bed
of calcareous conglomerate. The greda Fm. was
deposited in a palustrine environment and has abundant
pedogenic features. Its thickness varies from 7 m in the
northwester section to 78 m in the southeastern section
(Fig. 3).
Depositional sequence 2
Magaa Fm (DS 2). This formation is composed of
channel-fill (sandy point-bars) and crevasse deposits,
interbedded with floodplain mudstones displaying abun-
dant paleosols. Sandstones appear generally as sheets less
than 5 m thick. The Magaa Fm. was deposited in a
meandering fluvial system. In the southeastern section
(ALM), it attains a thickness of 121 m and contains some
conglomeratic channel fills (Fig. 3). In the northwestern
section (MOV), the formation is thicker (370 m), and can
be divided into Magaa Fms. 1 and 2 (Fig. 3). Channel
facies are more common in Magaa Fm. 1 and, occa-
sionally, amalgamated channel-fill and crevasse deposits
 142
L. Gonzlez-Acebrn et al. / Sedimentary Geology 202 (2007) 138157
Fig. 3. Stratigraphic sections of Almajano (southeastern section) and Montenegro-Villoslada (northwestern section). The Qmr/Qmo, the Qp/Qm and the Ms/Qmr ratios are represented in each section.
Qmr: monocrystalline quartz, undulosity b5; Qmo: Monocrystalline quartz, undulosity N5; Qp: polycrystalline quartz; Qm: Monocrystalline quartz; Ms: muscovite.
Table 1
Petrographic data base of sandstones from Almajano section (ALM)
ALM-101 ALM-102 ALM-103 ALM-104 ALM-105 ALM-106 ALM-107 ALM-108 ALM-109 ALM-110 ALM-111 ALM-112 ALM-113
N % N % N % N % N % N % N % N % N % N % N % N % N %
NCE Q Qmr Monocrystalline quartz, 30 6.6 31 7.6 4 2.6 46 9.3 23 5.0 19 4.7 13 3.2 10 2.5 3 0.7 9 2.2 27 5.8 8 1.6 29 6.3
undulosity b 5
Qmo Monocrystalline quartz, 50 11.0 101 24.8 5 3.3 81 16.4 159 34.4 144 35.8 228 56.2 195 48.5 214 53.0 187 46.6 186 40.1 161 32.1 131 28.4
undulosity N 5

L. Gonzlez-Acebrn et al. / Sedimentary Geology 202 (2007) 138157

Qm[Q] Monocrystalline 20 4.4 10 2.5 0.0 1 0.2 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
quartz with inherited
sintaxial cement
Qp2-3 Polycrystalline quartz 5 1.1 2 0.5 0.0 25 5.1 16 3.5 23 5.7 19 4.7 16 4.0 8 2.0 25 6.2 17 3.7 22 4.4 20 4.3
with 2-3 subgrains
Qp N 3 Polycrystalline quartz 0.0 0.0 0.0 8 1.6 42 9.1 36 9.0 35 8.6 62 15.4 54 13.4 73 18.2 51 11.0 77 15.3 44 9.5
N 3 subgrains
Qfrg Quartz in plutonic 0.0 0.0 0.0 0.0 0.0 14 3.5 7 1.7 11 2.7 1 0.2 19 4.7 0.0 0.0 0.0
rock fragment
Cq Carbonate replacement 10 2.2 6 1.5 0.0 28 5.7 18 3.9 3 0.7 3 0.7 0.0 0.0 1 0.2 11 2.4 13 2.6 21 4.5
on quartz
K Ks K-feldspar, 1 0.2 5 1.2 0.0 6 1.2 9 1.9 13 3.2 26 6.4 19 4.7 34 8.4 12 3.0 8 1.7 8 1.6 9 1.9
single crystals
Kfrg K-feldspar in coarse- 0.0 0.0 0.0 0.0 0.0 0.0 0.0 5 1.2 2 0.5 7 1.7 0.0 0.0 0.0
grained rock fragment
CaoK Kaolinite replacement 0.0 0.0 0.0 0.0 0.0 6 1.5 9 2.2 25 6.2 7 1.7 21 5.2 0.0 0.0 0.0
K-feldspar
Cik Kaolinite plus illite 0.0 0.0 0.0 0.0 0.0 2 0.5 1 0.2 10 2.5 2 0.5 3 0.7 0.0 1 0.2 0.0
replacement K-feldspar
Kil Illite replacement 0.0 0.0 0.0 0.0 2 0.4 3 0.7 3 0.7 0.0 0.0 3 0.7 5 1.1 2 0.4 5 1.1
K-feldspar
Ck Carbonate replacement 2 0.4 1 0.2 0.0 0.0 1 0.2 2 0.5 5 1.2 0.0 19 4.7 1 0.2 9 1.9 1 0.2 4 0.9
on K-feldspar
P Ps Plagioclase, 0.0 0.0 0.0 0.0 0.0 0.0 0.0 2 0.5 0.0 0.0 2 0.4 1 0.2 1 0.2
single crystals
Ab Albite 1 0.2 7 1.7 0.0 3 0.6 3 0.6 20 5.0 2 0.5 0.0 0.0 0.0 21 4.5 55 11.0 43 9.3
Ail Illite replacement 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 4 0.9 3 0.6 0.0
on albite
Cab Carbonate replacement 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 5 1.1 5 1.0 5 1.1
on Plagioclase
L Ch Chert 0.0 0.0 0.0 0.0 8 1.7 1 0.2 2 0.5 2 0.5 5 1.2 2 0.5 8 1.7 6 1.2 4 0.9
Lm Schist-Slate 0.0 0.0 0.0 0.0 1 0.2 3 0.7 0.0 0.0 0.0 0.0 1 0.2 1 0.2 2 0.4
(continued on next page)

143
 144
Table 1 (continued)
ALM-101 ALM-102 ALM-103 ALM-104 ALM-105 ALM-106 ALM-107 ALM-108 ALM-109 ALM-110 ALM-111 ALM-112 ALM-113
N % N % N % N % N % N % N % N % N % N % N % N % N %
M Ms Muscovite 0.0 0.0 0.0 0.0 2 0.4 10 2.5 3 0.7 9 2.2 2 0.5 5 1.2 12 2.6 8 1.6 25 5.4
Mfrg Muscovite in coarse- 0.0 0.0 0.0 0.0 0.0 2 0.5 0.0 0.0 0.0 0.0 0.0 0.0 0.0
grained rock fragment
Tu Tourmaline 1 0.2 0.0 0.0 0.0 1 0.2 0.0 1 0.2 1 0.2 0.0 1 0.2 0.0 4 0.8 3 0.6

L. Gonzlez-Acebrn et al. / Sedimentary Geology 202 (2007) 138157

Op Opaque 0.0 14 3.4 0.0 3 0.6 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
CE Ml Micritic limestone 77 17.0 128 31.4 0.0 0.0 14 3.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
Sc Sparitic limestone 22 4.9 17 4.2 4 2.6 6 1.2 22 4.8 0.0 0.0 0.0 0.0 0.0 35 7.5 2 0.4 5 1.1
Md Dolomicrite 21 4.6 9 2.2 42 27.5 49 9.9 13 2.8 0.0 0.0 0.0 0.0 0.0 1 0.2 6 1.2 46 10.0
Sd Dolosparite 14 3.1 1 0.2 0.0 2 0.4 1 0.2 0.0 0.0 0.0 0.0 0.0 0.0 2 0.4 11 2.4
Afrc Ankerite replacement 10 2.2 0.0 0.0 0.0 8 1.7 0.0 0.0 0.0 0.0 0.0 1 0.2 55 11.0 18 3.9
on carbonate rock
fragment
Fo Fossils 10.2 1 0.2 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1 0.2
Ep Echinoderm plates 81.8 13 3.2 0.0 2 0.4 0.0 0.0 0.0 0.0 0.0 0.0 3 0.6 0.0 0.0
0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 7 1.4 1 0.2
NCIIn Intraclast 1 0.2 2 0.5 2 1.3 3 0.6 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
CI Mli Micritic limestone 0.0 0.0 83 54.2 140 28.3 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
TOTAL 274 60.5 348 85.5 140 91.5 403 81.4 343 74.2 301 74.9 357 87.9 367 91.3 351 86.9 369 92.0 407 87.7 448 89.2 428 92.6
FRAMEWORK
[Ca] Calcite cement 0.0 10 2.5 4 2.6 39 7.9 7 1.5 0.0 0.0 0.0 0.0 0.0 8 1.7 1 0.2 0.0
[Dol] Dolomite cement 0.0 4 1.0 0.0 12 2.4 6 1.3 0.0 0.0 0.0 0.0 0.0 18 3.9 34 6.8 3 0.6
C [Anq] Anquerite cement 0.0 19 4.7 0.0 9 1.8 15 3.2 10 2.5 6 1.5 4 1.0 6 1.5 0.0 4 0.9 0.0 10 2.2
E [Q] Quartz cement 0.0 18 4.4 0.0 26 5.3 38 8.2 26 6.5 20 4.9 10 2.5 33 8.2 23 5.7 9 1.9 14 2.8 17 3.7
M Ank[Q] Anquerite replacing 0.0 0.0 4 0.9 4 1.0 0.0 0.0 0.0 0.0 2 0.4 0.0 2 0.4
quartz cement
E [Cao] Kaolinite cement 0.0 0.0 0.0 0.0 27 5.8 17 4.2 16 3.9 5 1.2 3 0.7 8 2.0 10 2.2 0.0 0.0
N Anq[Cao] Anquerite replacing 0.0 0.0 0.0 0.0 4 0.9 3 0.7 0.0 0.0 0.0 0.0 0.0 0.0 0.0
kaolinite cement
T [Cao-il] Kaolinite-illite cement 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
S [il] Illite cement 0.0 0.0 0.0 0.0 0.0 1 0.2 0.0 0.0 0.0 0.0 0.0 0.0 0.0
[Fe] Fe-oxide cement 0.0 7 1.7 7 4.6 5 1.0 18 3.9 40 10.0 7 1.7 16 4.0 9 2.2 1 0.2 6 1.3 5 1.0 2 0.4
[K] K-feldspar cement 0.0 1 0.2 0.0 1 0.2 0.0 0.0 0.0 0.0 2 0.5 0.0 0.0 0.0 0.0
TOTAL 274 60.5 407 100.0 153 100.0 495 100.0 462 100.0 402 100.0 406 100.0 402 100.0 404 100.0 401 100.0 464 100.0 502 100.0 462 100.0
Quartz typologies have been classified following the criteria of Basu et al. (1975). Plutonic rock fragments have been counted as quartz, feldspar and muscovite in plutonic rock fragments (Qfrg, Kfrg, Mfrg).
Table 2
Petrographic data base of sandstones from Montenegro-Villoslada section (MOV)
MOV-100 MOV-101 MOV-102 MOV-103 MOV-104 MOV-105 MOV-106 MOV-107 MOV-108 MOV-109 MOV-110 MOV-111 MOV-112
N % N % N % N % N % N % N % N % N % N % N % N % N %
NCE Q Qmr Monocrystalline 45 11.2 24 5.9 27 4.4 23 5.9 51 12.7 24 6.0 8 2.0 26 6.1 14 3.4 11 2.9 3 0.7 10 2.3 9 2.2
quartz, undulosity b5

L. Gonzlez-Acebrn et al. / Sedimentary Geology 202 (2007) 138157

Qmo Monocrystalline 101 25.1 104 25.7 81 13.1 142 36.5 108 26.8 145 36.3 173 42.9 150 35.3 158 38.6 175 46.2 166 41.2 185 42.5 151 36.7
quartz, undulosity N5
Qm Monocrystalline quartz 0.0 8 2.0 4 0.6 2 0.5 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
[Q] with inherited
sintaxial cement
Qp2-3 Polycrystalline quartz 15 3.7 13 3.2 11 1.8 7 1.8 35 8.7 23 5.8 6 1.5 20 4.7 33 8.1 13 3.4 7 1.7 30 6.9 22 5.3
with 2-3 subgrains
Qp N 3 Polycrystalline quartz 46 11.4 12 3.0 14 2.3 39 10.0 54 13.4 25 6.3 18 4.5 27 6.4 26 6.4 0.0 46 11.4 30 6.9 31 7.5
N3 subgrains
Qfrg Quartz in plutonic 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 9 2.2 0.0 12 3.0 24 5.5 1 0.2
rock fragment
Cq Carbonate replacement 3 0.7 7 1.7 3 0.5 2 0.5 0.0 10 2.5 3 0.7 8 1.9 1 0.2 12 3.2 0.0 2 0.5 3 0.7
on quartz
K Ks K-feldspar, 32 8.0 6 1.5 0.0 24 6.2 44 10.9 48 12.0 50 12.4 64 15.1 65 15.9 16 4.2 61 15.1 43 9.9 67 16.3
single crystals
Kfrg K-feldspar in coarse- 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 6 1.5 0.0 2 0.5 0.0 1 0.2
grained rock fragment
CaoK Kaolinite replacement 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
K-feldspar
Cik Kaolinite plus illite 3 0.7 0.0 0.0 0.0 0.0 0.0 2 0.5 0.0 5 1.2 0.0 17 4.2 1 0.2 0.0
replacement K-feldspar
Kil Illite replacement 3 0.7 4 1.0 0.0 2 0.5 0.0 1 0.3 0.0 8 1.9 7 1.7 0.0 0.0 3 0.7 13 3.2
K-feldspar
Ck Carbonate replacement 1 0.2 2 0.5 9 1.5 3 0.8 2 0.5 12 3.0 0.0 7 1.6 2 0.5 1 0.3 1 0.2 0.0 7 1.7
on K-feldspar
P Ps Plagioclase, 5 1.2 3 0.7 0.0 5 1.3 9 2.2 1 0.3 5 1.2 5 1.2 6 1.5 0.0 3 0.7 5 1.1 3 0.7
single crystals
Ab Albite 23 5.7 35 8.7 30 4.8 22 5.7 18 4.5 6 1.5 12 3.0 12 2.8 10 2.4 21 5.5 9 2.2 6 1.4 18 4.4
Ail Illite replacement 0.0 0.0 6 1.0 0.0 2 0.5 1 0.3 0.0 0.0 0.0 0.0 0.0 0.0 0.0
on albite
Cab Carbonate replacement 1 0.2 4 1.0 2 0.3 1 0.3 1 0.2 3 0.8 0.0 0.0 0.0 0.0 0.0 0.0 0.0
on Plagioclase
(continued on next page)

145
 146
Table 2 (continued)
MOV-100 MOV-101 MOV-102 MOV-103 MOV-104 MOV-105 MOV-106 MOV-107 MOV-108 MOV-109 MOV-110 MOV-111 MOV-112
N % N % N % N % N % N % N % N % N % N % N % N % N %
L Ch Chert 11 2.7 8 2.0 3 0.5 5 1.3 7 1.7 8 2.0 11 2.7 8 1.9 10 2.4 2 0.5 5 1.2 5 1.1 6 1.5
Lm Schist-Slate 7 1.7 2 0.5 2 0.3 3 0.8 2 0.5 1 0.3 8 2.0 4 0.9 5 1.2 0.0 2 0.5 5 1.1 0.0
M Ms Muscovite 29 7.2 19 4.7 28 4.5 3 0.8 0.0 5 1.3 17 4.2 18 4.2 10 2.4 69 18.2 6 1.5 55 12.6 54 13.1
Mfrg Muscovite in coarse- 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 2 0.5 0.0 0.0 3 0.7 2 0.5

L. Gonzlez-Acebrn et al. / Sedimentary Geology 202 (2007) 138157

grained rock fragment
Tu Tourmaline 1 0.2 3 0.7 1 0.2 0.0 0.0 1 0.3 0.0 0.0 0.0 1 0.3 0.0 1 0.2 0.0
Op Opaque 0.0 0.0 137 22.1 15 3.9 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
CE Ml Micritic limestone 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
Sc Sparitic limestone 0.0 1 0.2 2 0.3 3 0.8 3 0.7 4 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
Md Dolomicrite 0.0 13 3.2 16 2.6 17 4.4 11 2.7 2 0.5 0.0 0.0 0.0 0.0 0.0 0.0 0.0
Sd Dolosparite 7 1.7 2 0.5 2 0.3 3 0.8 1 0.2 2 0.5 0.0 7 1.6 0.0 0.0 0.0 0.0 0.0
Afrc Ankerite replacement 0.0 4 1.0 2 0.3 11 2.8 1 0.2 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
on carbonate
rock fragment
Fo Fossils 15 3.7 24 5.9 26 4.2 0.0 5 1.2 2 0.5 0.0 0.0 0.0 0.0 0.0 0.0 0.0
Ep Echinoderm plates 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
0.0 0.0 1 0.2 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
NCIIn Intraclast 1 0.2 0.0 9 1.5 0.0 0.0 4 1.0 0.0 0.0 0.0 2 0.5 0.0 0.0 0.0
CI Mli Micritic limestone 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
TOTAL 349 86.8 298 73.8 416 67.2 332 85.3 354 87.8 328 82.0 313 77.7 364 85.6 369 90.2 323 85.2 340 84.4 408 93.8 388 94.2
FRAMEWORK
[Ca] Calcite cement 0.0 0.0 0.0 0.0 5 1.2 7 1.8 0.0 3 0.7 0.0 0.0 0.0 0.0 0.0
[Dol] Dolomite cement 10 2.5 7 1.7 3 0.5 15 3.9 4 1.0 4 1.0 0.0 11 2.6 0.0 1 0.3 0.0 0.0 0.0
C [Anq] Anquerite cement 8 2.0 34 8.4 75 12.1 9 2.3 3 0.7 8 2.0 0.0 4 0.9 2 0.5 11 2.9 0.0 0.0 0.0
E [Q] Quartz cement 10 2.5 11 2.7 13 2.1 25 6.4 28 6.9 44 11.0 41 10.2 26 6.1 29 7.1 7 1.8 47 11.7 22 5.1 2 0.5
M Ank[Q] Anquerite replacing 5 1.2 3 0.7 11 1.8 0.0 0.0 3 0.8 1 0.2 4 0.9 0.0 0.0 3 0.7 0.0 0.0
quartz cement
E [Cao] Kaolinite cement 0.0 0.0 0.0 1 0.3 1 0.2 0.0 0.0 0.0 0.0 0.0 7 1.7 0.0 0.0
N Anq[Cao] Anquerite replacing 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
kaolinite cement
T [Cao-il] Kaolinite-illite cement 1 0.2 0.0 0.0 0.0 0.0 0.0 0.0 0.0 2 0.5 0.0 0.0 0.0 0.0
S [il] Illite cement 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1 0.2 0.0 0.0
[Fe] Fe-oxide cement 16 4.0 51 12.6 101 16.3 4 1.0 6 1.5 1 0.3 48 11.9 12 2.8 6 1.5 37 9.8 2 0.5 5 1.1 22 5.3
[K] K-feldspar cement 3 0.7 0.0 0.0 3 0.8 2 0.5 5 1.3 0.0 1 0.2 1 0.2 0.0 3 0.7 0.0 0.0
TOTAL 402 100.0 404 100.0 619 100.0 389 100.0 403 100.0 400 100.0 403 100.0 425 100.0 409 100.0 379 100.0 403 100.0 435 100.0 412 100.0
Quartz typologies have been classified following the criteria of Basu et al. (1975). Plutonic rock fragments have been counted as quartz, feldspar and muscovite in plutonic rock fragments (Qfrg, Kfrg, Mfrg).
 L. Gonzlez-Acebrn et al. / Sedimentary Geology 202 (2007) 138157
can be observed. Magaa Fm. 2 represents more distal
deposits of the meandering fluvial system.
Sierra de Matute Fm (DS 2). This formation, 145 m
thick, is only represented in the southeastern section
(ALM) (Fig. 3). It is composed of tabular carbonate beds of
grey wackestone containing ostracods and charophytes,
subordinate gastropods, bivalve fragments and scarce
oncolites. Root marks and brecciation related to water-
table fluctuations are common. Limestone bodies are less
than 45 m thick and interbedded with channelized
sandstone bodies, showing trough cross-bedding and
carbonate-rock fragments at the base (Fig. 3). The Sierra
de Matute Fm. was deposited in a lacustrinepalustrine
environment.
3. Methods
Samples of medium-grained sandstones were collected
from two representative stratigraphic sections of the Tera
Group (see Fig. 1 for locations) and 26 selected for
analysis. Thin sections were etched and stained using HF
and sodium cobaltinitrite for potassium feldspar, and
alizarin-red and potassium ferrocyanide for carbonate
identification (Chayes, 1952; Lindholm and Finkelman,
1972, respectively). To characterize detrital modes, a
quantitative petrographic analysis was performed on thin
sections by the integrated GazziZuffa point counting
method (Gazzi, 1966; Zuffa, 1985; Weltje, 2002). This
procedure combines the GazziDickinson and tradi-
tional criteria (Ingersoll et al., 1984). Four hundred to four
hundred and fifty points were counted per slide. Post-
depositional modifications to the original framework (e.g.
feldspar replacement) were assessed to restore the original
composition of the sandstone framework. The petro-
graphic data (Tables 1 and 2) reveal the restored frame-
work compositions, and in each case the way in which
composition differs from the original framework is
indicated: carbonate replacement of quartz (Cq); kaolin-
ite, kaolinite plus illite, illite or carbonate replacement of
K-feldspar (CaoK, Cik, Kill, Ck); illite replacement of
albite (Ail); carbonate replacement of plagioclase (Cab);
ankerite replacement of carbonate rock fragments (Afrc).
Thirty-two detrital classes were considered and grouped
into four categories according to the criteria of Zuffa
(1980): non-carbonate extrabasinal (NCE), carbonate
extrabasinal (CE), non-carbonate intrabasinal (NCI) and
carbonate intrabasinal (CI) (Tables 1 and 2).
4. Results
Forty-three petrographic classes were recognized,
comprising detrital grains (32 classes) and cements (11
147
classes) (Tables 1 and 2). Depositional matrix composed
exclusively of diagenetic phases was not recognized (i.e.,
pseudomatrix, epimatrix and pore filling). The epimatrix
and pseudomatrix affecting the original framework grains
were considered, however, when reconstructing the ori-
ginal framework composition (Tables 1 and 2). Diagenetic
processes (mechanical and chemical compaction and
cementation) were intense and destroyed the porosity of
the sediment.
4.1. Grain types
4.1.1. Non-carbonate extrabasinal grains (NCE)
This category included quartz, K-feldspars, plagio-
clase, muscovite, dense minerals (pyrite, tourmaline and
other opaque minerals), chert, and rock fragments
(slateschist fragments and plutonic rock fragments).
Monocrystalline quartz with abraded overgrowths (Qm
[Q] in Tables 1 and 2 and Fig. 4B) was observed at the
base of both sections; its percentage decreases towards
the top of the sections.
Replacement of feldspar by clay minerals and carbo-
nate is common (kaolinite, illite or illite replacement of
kaolinite, CaoK, Kil, Cik in Tables 1 and 2). Albites (Ab)
(Fig. 4H) are commonly replaced by clay minerals (illite)
(Ail).
Although scarce, non-carbonate lithic fragments
provided important information on source rocks. These
included slate and fine-grained schist (Lm) (Fig. 4I),
chert grains (Ch) (Fig. 4I) and plutonic rock fragments
(Fig. 4F). Chert grains are scarce and either show tec-
tonic fabric with oriented crystal units (of metamorphic
origin) (Fig. 4I) or an isotropic texture.
4.1.2. Carbonate extrabasinal (CE)
Carbonate-rock fragments were classified according
to their composition and grain size as: micritic limestone
fragments (Ml) (Fig. 4C), sparitic rock fragments (Sc)
(Fig. 4E and H), dolomicritic fragments (Md) and
dolosparitic fragments (Sd). In addition, recycled fossils
of extrabasinal origin (Fo) were also included in this
group. Echinoderm plates (Ep) (Fig. 4A) with inherited
syntaxial overgrowths (rounded and discontinuous out-
lines) are common. All these grains were derived from
sedimentary units and identified according to the
petrographic criteria of Zuffa (1987).
4.1.3. Noncarbonate intrabasinal (NCI)
This category is not quantitatively significant and
consists of soft siltyclay grains, deformed by mechan-
ical compaction and commonly associated with quart-
zofeldspathic deposits.
148 L. Gonzlez-Acebrn et al. / Sedimentary Geology 202 (2007) 138157

4.1.4. Carbonate intrabasinal (CI) 4.1.5. Recalculated parameters

Composed of micritic limestones associated with For an accurate analysis of clastic modes, we
calcretes, these grains often show textures of septarian established several petrographic variables. Table 3
nodules (Fig. 4D). summarizes the criteria used to obtain these variables
 L. Gonzlez-Acebrn et al. / Sedimentary Geology 202 (2007) 138157 149

Table 3 criteria of several authors: Zuffa (1985) (NCE/CI/CE);

Recalculated parameters used in the ternary plots
Ternary plot Parmeters
NCE-CE-CI NCE = Qmr + Qmo + Qm[Q] + Qp2 3 + Qp N 3 + Qfrg
+ Cq + Ks + Kfrg + CaoK + Cik + Cil + Ck + Ps + Ab + Ail
+ Cab + Ch + Lm + Ms + Mfrg + Tu + Opp
CE = Ml + Sc + Md + Sd + Afrc + Fo + E
CI = In
QFR Q = Qmr + Qmo + Qm[Q] + Qp2 3 + QpN3 + Qfrg + Cq
F = Ks + Kfrg + CaoK + Cik + Kil + Ck + Ps + Ab + Ail
+ Cab
R = Qfrp + Kfrg + Mfrg + Lm + CE
QmFLt Qm = Qmr + Qmo + Qm[Q] + Qp2 3 + QpN3 + Qfrg
+ Cq
F = Ks + Kfrg + CaoK + Cik + Kil + Ck + Ps + Ab + Ail
+ Cab
Lt = Ch + Lm + Ml + Sc + Md + Sd + Afrc + Fo + Pe
QmKP Qm = Qmr + Qmo + Qm[Q] + Qp2 3 + Qp N 3 + Qfrg
+ Cq
K = Ks + Kfrg + CaoK + Cik + + Kil + Ck
P= Ps + Ab + Ail + Cab
QmrQmoQp Qmr = Qmr
Qmo = Qmo
Qp = Qp2 3 + Qp N 3
RgRsRm Rg = Qfrg + Kfrg + Mfrg
Rs = CE
Rm = Lm
LmLsmLse Lm = Lm
Lsm = Ml + Md + Afrc
Lse = Sc + Sd + Fo + Pe
See Tables 1 and 2 for abreviations.
and Table 4 provides the values for both sections. These
recalculated parameters were used to plot clastic modes
in ternary diagrams (Fig. 4). The Qmr/Qmo and Qp/Qm
ratios obtained (Fig. 3) proved to be sensitive indicators
of metamorphic source areas. To characterize the
depositional sequences we also used Ms/Qmr ratios
(Fig. 3).
4.2. Sandstone composition and petrofacies
The sandstone framework composition was repre-
sented in different ternary plots (Fig. 5) following the
Dickinson et al. (1983) (QmFLt); Dickinson (1985)
(QmKP); Basu et al. (1975) (Qmr/Qmo/Qp); Arribas
et al. (1990) and Criteri and Le Pera (1994) (RgRsRm);
Arribas et al. (2003) (LmLsmLse). Three different
petrofacies were distinguished on the basis of these
plots. Their signatures and main characteristics are
detailed in Table 5, in the form of arithmetic averages
and standard deviations for all recalculated parameters.
Rigorous statistical analyses were not attempted in our
reconnaissance study of the provenance of the Tera
Group in the Eastern part of Cameros Basin.
Petrofacies 1: This sedimentolithic petrofacies plots
in the recycled orogenic field of the QmFLt diagram
(Fig. 5B). It shows a high lithic fragment content,
mainly including micritic carbonate fragments
(Fig. 4C), and plots on the Rs pole of the RgRsRm
ternary diagram and on the Lsm pole of the LmLsmLse
diagram (Fig. 5E and F). Fragments from old carbonate
cements or recrystalized fossils ascribed to marine
carbonate facies were observed. Inherited echinoderm
plates (Fig. 5A) with syntaxial cements are also present,
and there is evidence of intrabasinal contributions from
calcretes (Fig. 4D). The Qmr/Qmo ratio is high (Table 4,
Fig. 3), and some quartz grains display abraded
overgrowths. Feldspars and polycrystalline quartz are
also rare (see Qp/Qm Table 4, Fig. 3). The greda Fm.
(DS 1) is characterized by Petrofacies 1.
Petrofacies 2: Petrofacies 2 is quarzofeldspathic and
plots within the craton interior field of the QmFLt
ternary plot (Fig. 5B). Plutonic rock fragments and
schistslate fragments are present. Polycrystalline
quartz, usually with a tectonic fabric (Fig. 4G) and
more than three crystal units, is common. Relatively
high K-feldspar and muscovite contents are character-
istic of this petrofacies, tending to increase towards the
top (Fig. 5C). Petrofacies 2 was formed during the
deposition of the Magaa Fm. (DS 2).
This petrofacies is not compositionally homogeneous
in the southeastern and northwestern sections of the

Fig. 4. Photomicrographs of the detrital components and diagenetic features of the Tera Group sandstones. A.Inherited echinoderm plate (Ep)
(ALM, Petrofacies 1, greda Fm.). Crossed nichols. B.Monocrystalline quartz grain with inherited syntaxial overgrowth (Qm[Q]). (ALM,
Petrofacies 1, greda Fm.). Crossed nichols. C.Litharenite from the greda Fm., Petrofacies 1 (ALM). Note the high content of monocrystalline
quartz (Qm) and micritic limestone fragments (Ml). Crossed nichols. D.Septarian nodule of a calcrete from the greda Fm. (Petrofacies 1). Crossed
nichols. E.Medium-grained arkose from the Magaa Fm. (MOV, Petrofacies 2B) showing a sparitic rock fragment (Sc). A ferroan calcite cement
([Ca]) and an ankerite cement ([Anq]) also appear. Parallel nichols. F.Medium-grained arkose from the Magaa Fm. (MOV, Petrofacies 2B) with a
plutonic rock fragment (Prf), moscovite (Ms) and single grains of K-feldspar (Ks). The framework exhibits dense packing generated by compaction.
Crossed nichols. G.Petrofacies 2A (ALM, greda Fm.) containing polycrystalline quartz (Qp) with tectonic fabric. Kaolinite pore filling ([Kao]) is
very common in this petrofacies. Crossed nichols. H.Subarkose from the Sierra de Matute Fm. (ALM, Petrofacies 3) showing a sparitic rock
fragment (Sc). Note the presence of moscovite (Ms) and albite (Ab). Crossed nichols. I.Subarkose from the Sierra de Matute Fm. (ALM,
Petrofacies 3) with slate (Sl) and chert (Ch) fragments. As diagenetic features, it shows quartz overgrowths ([Q]), ankerite overgrowths ([Anq]),
calcite overgrowths ([Ca]) and kaolinite pore fillings ([Kao]). Crossed nichols.
 150
Table 4
Values of recalculated parameters for sandstones of Almajano (ALM) and Montenegro-Villoslada (MOV) sections (Arithmetic averages)
Samples ALM- ALM- ALM- ALM- ALM- ALM- ALM- ALM- ALM- ALM- ALM- ALM- ALM- MOV- MOV- MOV- MOV- MOV- MOV- MOV- MOV- MOV- MOV- MOV- MOV- MOV-

L. Gonzlez-Acebrn et al. / Sedimentary Geology 202 (2007) 138157

101 102 103 104 105 106 107 108 109 110 111 112 113 100 101 102 103 104 105 106 107 108 109 110 111 112

NCE% 43.6 50.9 6.4 49.9 83.1 100.0 100.0 100.0 100.0 100.0 90.2 83.9 80.8 93.4 85.2 86.1 89.8 94.1 95.7 100.0 98.1 100.0 99.4 100.0 100.0 100.0
CE% 56.0 48.6 32.9 14.6 16.9 0.0 0.0 0.0 0.0 0.0 9.8 16.1 19.2 6.3 14.8 11.8 10.2 5.9 3.0 0.0 1.9 0.0 0.0 0.0 0.0 0.0
CI% 0.4 0.6 60.7 35.5 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.3 0.0 2.2 0.0 0.0 1.2 0.0 0.0 0.0 0.6 0.0 0.0 0.0
Qm% 43.9 45.2 16.4 73.5 77.7 82.7 86.4 82.4 80.2 86.5 74.1 64.4 61.4 69.3 66.7 65.1 68.5 71.1 71.8 70.3 66.8 67.5 84.1 70.1 80.4 65.0
F% 1.5 3.9 0.0 3.5 4.5 15.9 13.0 17.1 18.3 12.9 13.7 17.4 16.8 22.4 21.4 21.9 18.2 21.8 22.8 23.3 27.7 28.3 15.1 27.8 16.7 33.1
Lt% 54.6 50.9 83.6 23.0 17.8 1.4 0.6 0.6 1.4 0.6 12.2 18.1 21.8 8.3 11.9 13.0 13.4 7.2 5.4 6.4 5.5 4.2 0.8 2.1 2.9 1.8
Qm% 43.9 45.2 16.4 73.5 77.7 82.7 86.4 82.4 80.2 86.5 74.1 64.4 61.4 69.3 66.7 65.1 68.5 71.1 71.8 70.3 66.8 67.5 84.1 70.1 80.4 65.0
K% 1.5 3.9 0.0 3.5 4.5 15.9 13.0 17.1 18.3 12.9 13.7 17.4 16.8 22.4 21.4 21.9 18.2 21.8 22.8 23.3 27.7 28.3 15.1 27.8 16.7 33.1
P% 54.6 50.9 83.6 23.0 17.8 1.4 0.6 0.6 1.4 0.6 12.2 18.1 21.8 8.3 11.9 13.0 13.4 7.2 5.4 6.4 5.5 4.2 0.8 2.1 2.9 1.8
Qmr% 43.9 45.2 44.4 73.5 77.7 82.7 86.4 82.4 80.2 86.5 74.1 64.4 61.4 69.3 66.7 65.1 68.5 71.1 71.8 70.3 66.8 67.5 84.1 70.1 80.4 65.0
Qmo% 1.5 3.9 55.6 3.5 4.5 15.9 13.0 17.1 18.3 12.9 13.7 17.4 16.8 22.4 21.4 21.9 18.2 21.8 22.8 23.3 27.7 28.3 15.1 27.8 16.7 33.1
Qp% 54.6 50.9 0.0 23.0 17.8 1.4 0.6 0.6 1.4 0.6 12.2 18.1 21.8 8.3 11.9 13.0 13.4 7.2 5.4 6.4 5.5 4.2 0.8 2.1 2.9 1.8
Rg% 0.0 0.0 0.0 0.0 0.0 82.7 86.4 82.4 80.2 86.5 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 67.5 0.0 70.1 80.4 65.0
Rs% 1.5 3.9 55.6 3.5 4.5 0.0 0.0 0.0 0.0 0.0 13.7 17.4 16.8 22.4 21.4 21.9 18.2 21.8 22.8 0.0 27.7 0.0 0.0 0.0 0.0 0.0
Rm% 0.0 0.0 0.0 0.0 17.8 1.4 0.0 0.0 0.0 0.0 12.2 18.1 21.8 8.3 11.9 13.0 13.4 7.2 5.4 6.4 5.5 4.2 0.0 2.1 2.9 0.0
Lm% 0.0 0.0 0.0 0.0 0.0 100.0 0.0 0.0 0.0 0.0 2.5 1.4 2.4 50.0 9.1 8.0 8.1 11.1 11.1 100.0 36.4 100.0 0.0 100.0 100.0 0.0
Lsm% 68.5 81.1 91.3 83.1 52.9 0.0 0.0 0.0 0.0 0.0 2.5 5.5 19.3 50.0 13.6 16.0 16.2 22.2 66.7 0.0 63.6 0.0 0.0 0.0 0.0 0.0
Lse% 31.5 18.9 8.7 16.9 45.1 0.0 0.0 0.0 0.0 0.0 95.0 93.2 78.3 0.0 77.3 76.0 75.7 66.7 22.2 0.0 0.0 0.0 0.0 0.0 0.0 0.0
P/F 0.2 1.7 0.6 0.6 5.0 0.5 0.5 5.9 11.8 9.6 7.0 9.5 5.9 7.0 7.2 2.0 4.2 4.0 3.9 5.5 3.0 2.5 5.1
Qmr/ 0.6 0.3 0.8 0.6 0.1 0.1 0.1 0.1 0.0 0.0 0.1 0.0 0.2 0.4 0.2 0.3 0.2 0.5 0.2 0.0 0.2 0.1 0.1 0.0 0.1 0.1
Qmo
Qp/Qm 0.1 0.0 0.0 0.3 0.3 0.4 0.2 0.4 0.3 0.5 0.3 0.6 0.4 0.4 0.2 0.2 0.3 0.6 0.3 0.1 0.3 0.3 0.1 0.3 0.3 0.3
Mosc/ 0.0 0.0 0.0 0.0 0.1 0.5 0.2 0.9 0.7 0.6 0.4 1.0 0.9 0.6 0.8 1.0 0.1 0.0 0.2 2.1 0.7 0.7 6.3 2.0 5.5 6.0
Qmr
 L. Gonzlez-Acebrn et al. / Sedimentary Geology 202 (2007) 138157 151

Fig. 5. Ternary plots describing sandstone composition. Values for the northwestern and the southeastern sections are shown as squares and circles,
respectively (see legend). Development for the western section (MOV) is marked by broken arrows; development for the southeastern section (ALM)
is indicated by continuous arrows. Petrofacies are indicated with numbers when possible. See Table 3 for recalculated parameters and Tables 1 and 2
for numerical values. The ternary diagrams were prepared according to the criteria of several authors: A.Zuffa (1985) (NCE/CI/CE); B.
Dickinson et al. (1983) (QmFLt); C.Dickinson (1985) (QmKP); D.Basu et al. (1975) (Qmr/Qmo/Qp); E.Arribas et al. (1990) and Criteri and
Le Pera (1994) (RgRsRm); F.Arribas et al. (2003) (LmLsmLse).

basin. Thus, it is possible to distinguish Petrofacies 2A Petrofacies 2B evolves from the craton interior to the
from 2B. Petrofacies 2A is represented in the craton transitional continental field of the QmFLt ternary
interior field of the QmFLt ternary plot, whereas plot (Fig. 5B). Petrofacies 2A plots along the RgRm
 152
Table 5
Main characteristics of the three petrofacies. () means arithmetic averages and standard deviations
Petrofacies Samples Lithology QmFLt QmKP Qp/Qm Qmr/Qmo P/K Main characteristics and
Lithostratigraphic units

L. Gonzlez-Acebrn et al. / Sedimentary Geology 202 (2007) 138157

1: Sedimentolithic ALM-101 Sedarenites Qm54 23 Qm45 23 0.08 0.12 0.54 0.20 0.85 0.78 Lithic fragments
ALM-102 F3 2 K2 2 Micritic carbonate fragments
ALM-103 Lt43 25 P53 25 Inherited echinoderm plates
ALM-104 Intrabasinal contributions
from calcretes
greda Fm. (ALM)
2A: Quarzofeldspathic ALM-105 Mainly subarkoses Qm84 4 Qm85 3 0.35 0.09 0.07 0.20 1.65 2.22 Plutonic rock fragments
ALM-106 F15 5 K13 5 Squist-slate fragments
ALM-107 Lt1 7 P2 7 QpN 3 with tectonic fabric
ALM-108 K-feldspar
ALM-109 Muscovite
ALM-110 Magaa Fm. (ALM)
2B: Quarzofeldspathic MOV-100 MOV-107 Mainly subarkoses Qm71 6 Qm71 6 0.70 0.14 0.27 0.16 5.14 2.98 Carbonate rock
MOV-101 MOV-108 F23 5 K23 5 (Magaa Fm. 1) fragments at the botton
MOV-102 MOV-109 Lt6 4 P6 4 0.08 0.05 Plutonic rock fragments
MOV-103 MOV-110 (Magaa Fm. 2) at the top
MOV-104 MOV-111 Squist-slate fragments
MOV-105 MOV-112 Qp N 3 with tectonic fabric
MOV-106 MOV-113 K-feldspar
Muscovite
Magaa Fm. 1 and 2
(MOV)
3: Quarzo- ALM-111 Subarkoses and Qm67 7 Qm67 7 0.44 0.12 0.13 0.20 9.11 2.14 Sparitic rock fragments
sedimentolithic ALM-112 lithoarenites F16 2 K16 2 Inherited echinoderm plates
ALM-113 Lt17 5 P17 5 Scarce schist-slate fragments
QpN 3 with tectonic fabric
Sierra de Matute Fm. (ALM)
 L. Gonzlez-Acebrn et al. / Sedimentary Geology 202 (2007) 138157
edge of the RgRsRm diagram (Fig. 5E). Petrofacies 2B
records an evolution from the Rs pole to RgRm edge of
the ternary plot. Petrofacies 2A shows lower Qmr/Qmo,
Qp/Qm and P/K indices than Petrofacies 2B (Table 4,
Figs. 3 and 5C). Petrofacies 2A shows rare lithic frag-
ments (Fig. 5B). Carbonate-rock fragments (Fig. 4E) are
present at the base of Petrofacies 2B (Fig. 4E).
Petrofacies 2B also suggests compositional differ-
ences between Magaa Fm. 1 and Magaa Fm. 2, since
the Ms/Qmr ratio is higher for Magaa Fm. 2 than Fm. 1
(Fig. 3). In addition, the Qmr/Qmo ratio is higher for
Magaa Fm. 1 than for Magaa Fm. 2 (Table 4, Fig. 3).
Petrofacies 3: This quartzolithic petrofacies plots in
the recycled orogenic field of the QmFLt ternary plot,
as does Petrofacies 1 (Fig. 5B). The main petrographic
characteristic is the presence of sparitic rock fragments
(Fig. 4H) and inherited echinoderm plates. This
petrofacies plots close to the Rs pole of the RgRsRm
ternary diagram and close to the Lse pole of the
LmLsmLse ternary diagram (Fig. 5E and F). Scarce
slateschist fragments occur (Fig. 4I). Other petro-
graphic features are a low Qmr/Qmo ratio (Table 4,
Fig. 3), abundant polycrystalline oriented quartz with more
than three crystal units (see Qp/Qm in Table 4), and a P/K
index higher than that for Petrofacies 2 (Table 4).
Petrofacies 3 was generated during the deposition of the
Sierra de Matute Fm. (DS 2).
5. Discussion
The changes observed in the petrofacies examined in
this study indicate variations in source areas during
sedimentation of the Tera Group.
Most of the micritic carbonate fragments of Petrofacies
1 are intrabasinal, and are probably derived from coeval
calcretes of the greda Fm. Sparitic rock fragments in
Petrofacies 1 were supplied by the Jurassic sedimentary
substratum. Petrofacies 2 shows an increase in metamor-
phic and secondary plutonic supplies, recorded in the
RgRsRm and LmLsmLse ternary plots (Fig. 5E and F).
The Qmr/Qmo/Qp ternary plot and the Qp/Qm and
Qmr/Qmo ratios (Figs. 5D and 3, Table 4) indicate a
decrease in non-undulatory monocrystalline quartz and
increased polycrystalline typologies at the top of both
sections. Polycrystalline quartz with tectonic fabric,
indicative of metamorphic sources, is also common at
the top.
The three established petrofacies are clearly related to
the lithostratigraphic units studied and record the sedi-
mentary evolution of the Tera Group. The greda Fm.
(Petrofacies 1) is a recycled (sedimentoclastic) unit, as
revealed by its high content of carbonate fragments,
153
inherited echinoderm plates, its low feldspar content and
dominance of monocrystalline quartz types (Table 4) and
recycled patterns. Petrofacies 1 records the erosion of the
Jurassic marine pre-rift deposits, mainly the Kimmer-
idgian limestones of the underlying unit (Torrecilla en
Cameros Fm.). This conclusion is supported by the simi-
larity of the extrabasinal carbonate-rock fragments to
the Torrecilla en Cameros Fm. limestones (Alonso and
Mas, 1990; Benito, 2001; Benito and Mas, 2001, 2002).
However, the high Qmr/Qmo index and the presence
of quartz with inherited syntaxial overgrowths suggest
that Petrofacies 1 also records the erosion of pre-rift
Mesozoic siliciclastic units (Callovian quartz-arenites,
Mas et al., 2003).
Petrofacies 1 corresponds to the undissected-transi-
tional stage of the non-volcanic rifted margin of
Garzanti et al. (2001, 2003). This type of provenance
has been recognized in present-day sands of the Red Sea
and Gulf of Aden (Yemen). Petrofacies 1 shows many
features of this stage: it plots in the same compositional
area of the QmFLt ternary plot and shows sedimentary
detritus, including recycled monocrystalline quartz and
carbonate grains.
Petrofacies 1 is equivalent to Petrofacies A from the
western sector of the Cameros Basin (Arribas et al., 2003).
Two main differences exist between both petrofacies: (1)
Petrofacies 1 is less quartzose than Petrofacies A, which is
richer in carbonate lithic fragments (Petrofacies 1:
Qm54F3Lt43; Petrofacies A: Qm85F2Lt13); and (2):
Petrofacies 1 contains no metamorphic rock fragments,
unlike Petrofacies A. Thus, sedimentary sources seem to
be more important in Petrofacies 1, which does not record
erosion of the crystalline basement. Sandstones in the
eastern sector of Cameros Basin have a signature that is
closer to the undissected-transitional stage of Garzanti
et al. (2001, 2003).
A complex provenance from metamorphic and granite
terranes can be deduced for Petrofacies 2 (Magaa Fm.,
DS 2). Despite the low percentage of slateschist frag-
ments (b2%), the presence of these rock fragments is
highly significant. Several studies on modern sediments
(Palomares and Arribas, 1993; Arribas and Tortosa, 2003;
Le Pera and Arribas, 2004) have demonstrated that
sandstone composition does not allow direct quantitative
estimations of the lithologies at the source, owing to the
different capacities of rocks to generate sands (Palomares
and Arribas, 1993). Indeed, slates and schists show a very
low sand-generating capacity (Sand Generation Index,
Palomares and Arribas, 1993) and are often underrepre-
sented, because of their low rock-fragment content. Slate
schist fragments are therefore considered significant
despite their scarcity.
154 L. Gonzlez-Acebrn et al. / Sedimentary Geology 202 (2007) 138157

Fig. 6. Location of the West Asturian Leonese Zone (WALZ) and the Central Iberian Zone (CIZ) in the scheme of Julivert et al. (1972) for the Iberian
(Hesperian) Massif. The situation of both sections is indicated as: MOV, Montenegro-Villoslada en Cameros (northwestern section) and ALM,
Almajano (southeastern section). The arrow indicates the tectonic development of the western basin, from the late Jurassic to Middle Albian
(modified from Arribas et al., 2003).

Medium sand-sized polycrystalline quartz grains de- slateschists tend to contain more than three crystal units
rived from granitoid and gneissic sources are formed by per grain (Basu et al., 1975; Tortosa et al., 1991; Palomares
quartz grains with few crystal units, typically less than and Arribas, 1993). The polycrystalline quartz at the top of
three, whereas polycrystalline quartz grains derived from both sections shows a tectonic fabric and is commonly

Fig. 7. Tectonic development of petrofacies during the Tithonian. Note that reactivation of the basin took place during deposition of Petrofacies 3 due
to back faulting. The location of both sections is indicated as: MOV, Montenegro-Villoslada en Cameros (northwestern section) and ALM, Almajano
(southeastern section).
 L. Gonzlez-Acebrn et al. / Sedimentary Geology 202 (2007) 138157
composed of more than three crystal units. Accordingly, its
origin can be traced to metamorphic source areas.
In contrast, several factors point to plutonic sources for
Petrofacies 2: (1) the scarcity of plutonic rock fragments in
Petrofacies 2A and at the top of Petrofacies 2B. Granitoids
have a very high potential to generate sands (Palomares and
Arribas, 1993), but their mechanical isotropy produces
sands formed by monomineral grains. Thus, plutonic
sources generate scarce plutonic rock fragments; (2) high
K-feldspar levels; and (3) the fact that Petrofacies 2 plots on
the QmF edge of the QmFLt diagram (Fig. 5B).
The high muscovite content of Magaa Fm. 2 (see
Fig. 3, Ms/Qmr ratio) also indicates an increase in the
influence of both metamorphic and plutonic sources,
due to the gradual unroofing of the basement.
The inferred sources for Petrofacies 2 (Magaa Fm.,
DS 2) are low-grade to medium-grade metamorphic
terranes. These terranes probably correspond to the West
Asturian Leonese Zone in the Iberian Massif (Fig. 6).
However, the granite source rocks are probably from the
Central Iberian Zone (Fig. 6). These interpretations are
based on the main lithologies of both zones: the West-
Asturian Leonese Zone consists mainly of a thick lower
Paleozoic sequence of slates and quartzites of greens-
chist metamorphic facies and minor contents of
amphibolite facies (Julivert, 1983). The main bedrock
lithologies in the Central Iberian Zone are Hercynian
granites, graniodiorites, and gneisses, with minor
exposure of low-grade metamorphic rocks (Villaseca
et al., 1993).
Moreover, different sources can be recognized in
Petrofacies 2A and 2B. Establishing the provenance of
Petrofacies 2B is more complex than that of Petrofacies
2A. Petrofacies 2B contains carbonate-rock fragments in
Magaa Fm. 1 (Table 2 and Figs. 4E and 5). However, no
carbonate-rock fragments were observed in Petrofacies
2A. Moreover, the P/K ratio is higher for Petrofacies 2B
(5.14) than 2A (1.19). These observations suggest different
source rocks for the two petrofacies, and probably indicate
that the marine substrate was still being eroded during the
deposition of the lower part of Petrofacies 2B.
An increase of K-feldspar content is recorded from
Petrofacies 1 to Petrofacies 2, indicating an increasing
depth of erosion. This tendency points to a change from
undissected-transitional to transitional signatures (Garzanti
et al., 2001, 2003). During the transitional stage, sands are
derived from sedimentary successions and underlying
basement rocks in varying proportions, as a function of the
erosion level and type of rocks exposed (Garzanti et al.,
2001, 2003).
Petrofacies 2 is equivalent to Petrofacies B (Arribas
et al., 2003) from the western sector of the Cameros Basin.
155
The main difference between both petrofacies is the pres-
ence of plutonic rock fragments in Petrofacies 2, which are
absent from Petrofacies B. Thus, deeper erosion of the
basement was recorded in the eastern sector of the basin in
sediments of Tithonian age.
Sedimentary sources are inferred for Petrofacies 3. In
addition, a secondary influx of metamorphic sources is
recorded. Sedimentary sources were probably generated
by erosion of Jurassic marine sedimentary rocks, mainly
Kimmeridgian limestones of the Torrecilla and Cameros
Fm., as indicated by sparitic rock fragments and inherited
echinoderm plates. This finding suggests reactivation of
this part of the basin, probably due to back-faulting
(Fig. 7), as has been inferred for the western part of the
Cameros Basin (Arribas et al., 2003).
Petrofacies 3 is characteristic of the eastern sector of
the Cameros Basin. This petrofacies is absent in the
western sector. Moreover, Petrofacies 3 shows a signature
between the undissected-transitional and the transitional
stages of Garzanti et al. (2001, 2003), revealing the in-
creasing relevance of sedimentary sources and the reacti-
vation of this part of the basin.
Our provenance data broadly fit into the Garzanti et al.
(2001, 2003) model for a modern rifted basin, which we
consider a more appropriate conceptual framework than
the more general Dickinsonian model (Dickinson et al.,
1983). Furthermore, our study shows that the use of
RgRsRm and LmLsmLse diagrams, and Qmr/Qmo, Qp/
Qm and Ms/Qmr ratios are important for a comprehensive
provenance analysis in rifted basins.
6. Conclusions
The present study of sandstone provenance in the
initial rifting stage of the Cameros Basin illustrates that
petrofacies analysis is a powerful tool for paleogeo-
graphic and paleotectonic reconstructions in tectonically
controlled basins.
The Tera Group (Tithonian) records sediment accumu-
lation during the initial formation of the Cameros Basin
during the latest JurassicEarly Cretaceous rift stage. The
Tera Group is made up of alluvial-fan deposits, meandering
fluvial sediments and lacustrinepalustrine carbonates.
Detrital modes of sandstones reveal three different
petrofacies, which are closely related to lithostratigraphic
units. As in other rifted basins, petrofacies indicate erosion
of pre-rift sedimentary substratum at the beginning of the
rifting, followed by unroofing of the basement in later
stages. This progression was also recorded in the eastern
part of Cameros Basin (Arribas et al., 2003) and in the
modern Red Sea-Gulf of Aden system (Garzanti et al.,
2001, 2003).
156 L. Gonzlez-Acebrn et al. / Sedimentary Geology 202 (2007) 138157
Petrofacies 1 is sedimentolithic (mean: Qm54F3Lt43)
and developed during the deposition of the greda Fm.
(DS 1). This petrofacies mainly records erosion of the
Jurassic marine sedimentary substrate.
The sandstones of Petrofacies 2 are quartzofelds-
pathic. Petrofacies 2 can be subdivided into Petrofacies
2A (mean: Qm84F15Lt1) and Petrofacies 2B (mean:
Q71F23Lt6). This petrofacies records the sedimentation
of the Magaa Fm. (DS 2). The deduced provenance
terranes for the Magaa Fm. are low to medium grade
metamorphics from the West Asturian Leonese Zone of
the Iberian Massif, which represents the Paleozoic
basement of the Cameros Basin. A minor influence of
plutonic source areas from the Central Iberian Zone is
also deduced. Progression from sedimentary sources to
metamorphic sources from DS 1 to DS 2 is also
observed in the western part of the Cameros Basin
(Arribas et al., 2002). However, erosion of plutonic
rocks (Petrofacies B in Arribas et al., 2003) has not been
documented in the western part of the basin.
Petrofacies 3 (mean: Qm67F16Lt17) characterizes the
Sierra de Matute Fm. (DS 2). The sedimentary source for
this formation is the marine Jurassic sedimentary substra-
tum, as is the case for Petrofacies 1. This suggests an
episode of reactivation of the basin, due to back-faulting.
Thus, erosion of the pre-rift sedimentary substratum is
followed by unroofing of the basement, as documented in
other ancient and modern rifted basins.
Acknowledgements
Funding for this research was provided by the Spanish
DIGICYT projects BTE 2001-026 and CGL 2005-07445-
C03-02/BTE. The authors would like to thank M.I.
Benito, L. Domingo and A. Martn for their useful sug-
gestions and comments. This manuscript also benefited
from critical reviews by G.J. Weltje, C. Fielding and two
anonymous referees.
References
Alonso, A., Mas, R., 1990. El Jursico Superior marino en el sector de
la Demanda-Cameros (La Rioja-Soria), vol. 14, pp. 173198.
Alonso-Azcarate, J., Barrenechea, J.F., Rodas, M., Mas, R., 1995.
Comparative study of the transition between very low grade
metamorphism and low grade metamorphism in siliciclastic and
carbonate sediments. Early Cretaceous, Cameros Basin (North
Spain). Clay Minerals 30, 407419.
Alonso-Azcarate, J., Rodas, M., Bottrell, S.H., Raiswell, R., Velasco, F.,
Mas, R., 1999. Pathways and distances of fluid flow during low grade
metamorphism: evidence from pyrite deposits of the Cameros Basin,
Spain. Journal of Metamorphic Geology 17 (4), 339348.
Arribas, J., Tortosa, A., 2003. Detrital modes in sedimenticlastic sand
from low-order streams in the Iberian Range, Spain: the potential
for sand generation by different sedimentary rocks. Sedimentary
Geology 159, 275303.
Arribas, J., Gmez-Gras, D., Rosell, J., Tortosa, A., 1990. Estudio
comparativo entre las areniscas Paleozoicas y Trisicas de la isla de
Menorca: Evidencias de procesos de reciclado. Revista de la
Sociedad Geolgica de Espaa 3, 105116.
Arribas, A., Mas, R., Ochoa, M., Alonso, A., 2002. Composicin y
diagnesis del registro detrtico en el borde suroccidental de la cuenca
de Cameros. Zuba. Instituto de Estudios Riojanos 14, 99119.
Arribas, J., Alonso, A., Mas, R., Tortosa, A., Rodas, M., Barrenechea,
J.F., Alonso-Azcarate, J., Artigas, R., 2003. Sandstone petrography
of continental depositional sequences of a intraplate rift basin:
Western Cameros Basin (North Spain). Journal of Sedimentary
Research 73 (2), 309327.
Basu, A., Young, S.W., Suttner, L.J., James, C.W., Mark, G.H., 1975.
Re-evaluation of the use of undulatory extinction and polycrys-
tallinity in detrital quartz provenance interpretation. Journal of
Sedimentary Petrology 45, 873882.
Barrenechea, F.J., Rodas, M., Frey, M., Alonso-Azcarate, J., Mas, J.R.,
2000. Chlorite, Corrensite and Chlorite-Mica in late Jurassic
Fluvio-Lacustrine sediments of the Cameros Basin of Northeastern
Spain. Clay and Clay Minerals 48 (2), 256265.
Barrenechea, F.J., Rodas, M., Frey, M., Alonso-Azcarate, J., Mas, J.R.,
2001. Clay diagenesis and low-grade metamorphism of Tithonian
and Berriasian sediments in the Cameros basin. Clay Minerals 36
(3), 325333.
Benito, M.I., 2001. Estudio comparativo de la evolucin sedimentaria
y diagentica de los litosomas carbonatados arrecifales (pre-rifting)
de la Cuenca de Cameros. Kimmeridgiense. La RiojaSoria. Ph.D.
Thesis. Universidad Complutense de Madrid, 410 pp.
Benito, M.I., Mas, R., 2001. Diagnesis temprana meterica de la Fm.
arrecifal Torrecilla en Cameros (Kimmeridgiense Inferior: prerift) y
de los carbonatos de la base del Grupo Tera (Titnico: sinrift) en el
sector de Soria. Cuenca de Cameros. N de Espaa. Geotemas 3 (1),
8388.
Benito, M.I., Mas, R., 2002. Evolucin diagentica de los carbonatos
arrecifales de la Formacin Torrecilla en Cameros y de los
carbonatos continentales suprayacentes (Kimmderidgiense inferi-
or-Titnico) en el Sector de Soria. Cuenca de Cameros, N. de
Espaa. Journal of Iberian Geology 28, 6592.
Benito, M.I., Lohmann, K.C., Mas, R., 2005. Late Jurassic palaeogeo-
graphy and palaeoclimate in the Northern Iberian Basin of Spain:
Constraints from diagenetic records in reefal and continental
carbonates. Journal of Sedimentary Research 75, 8296.
Blatt, H., 1967. Provenance determination and the recycling of
sediments. Journal of Sedimentary Petrology 37, 10311044.
Casquet, C., Galindo, C., Gonzlez-Casado, J.M., Alonso, A., Mas, R.,
Rodas, M., Garca, E., Barrenechea, J.F., 1992. El metamorfismo
en la Cuenca de Los Cameros. Geocronologa e implicaciones
tectnicas. Geogaceta 11, 2225.
Chayes, F., 1952. Notes of the staining of potash feldspar with sodium
cobaltonitrite in thin section. American Mineralogist 37, 337340.
Critelli, S., 1999. The interplay of lithosferic flexure and thrust
accommodation in forming stratigraphic sequences in the southern
Apennines foreland basin system. Italy: Lincei Scienze Fische e
Naturali, Rendiconti, Series IX, 10, pp. 257326.
Criteri, S., Le Pera, E., 1994. Detrital modes and provenance of Miocene
sandstones and modern sands of the Southern Apennines thrust-top
basins, Italy. Journal of Sedimentary Research, A 64, 824835.
DeCelles, P.G., Hertel, F., 1989. Petrology of fluvial sands from the
Amazonian foreland basin, Peru and Bolivia. Geological Society
of America Bulletin 101, 15521562.
 L. Gonzlez-Acebrn et al. / Sedimentary Geology 202 (2007) 138157
Dickinson, W.R., 1970. Interpreting detrital modes of graywacke and
arkose. Journal of Sedimentary Petrology 40, 695707.
Dickinson, W.R., 1985. Interpreting provenance relations from detrital
modes of sandstones. In: Zuffa, G.G. (Ed.), Provenance of
Arenites. D. Reidel, Dordrecht, pp. 333361.
Dickinson, W.R., Suczek, C.A., 1979. Plate tectonics and sandstone
compositions. American Association of Petroleum Geologists 63,
21642182.
Dickinson, W.R., Beard, L.S., Barkenrigge, G.R., Erjavec, J.L., Ferguson,
R.C., Inman, K.F., Knepp, R.A., Lindberg, F.A., Ryberg, P.T., 1983.
Provenance of North America Phanerozoic sandstones in relation to
tectonic setting. Geological Society of America Bulletin 94, 222235.
Evans, A.L., 1990. Miocene sandstone provenance relations in the
Gulf of Suez: insights into Synrift Unroofing and Uplift History.
The American Association of Petroleum Geologists Bulletin 74,
13861400.
Garzanti, E., Vezzoli, G., And, S., Castiglioni, G., 2001. Petrology of
Rifted-Margin Sand (Red Sea and Gulf of Aden, Yemen). Journal
of Geology 109, 277297.
Garzanti, E., And, S., Vezzoli, G., Dell'Era, D., 2003. From rifted
margins to foreland basins: investigating provenance and sediment
dispersal across desert Arabia (Oman, U.A.E.). Journal of
Sedimentary Research 73 (4), 572588.
Gazzi, P., 1966. Le arenarie del flysch sopracretaceo dell'Appenino
modenese; correlazioni con il flysch di Monghidoro. Mineralogica
Petrografica Acta 12, 6997.
Gmez Fernndez, J.C., Melndez, N., 1994. Estratigrafa de la
Cuenca de los Cameros (Cordillera Ibrica Noroccidental, N de
Espaa) durante el trnsito JursicoCretcico. Revista Sociedad
Geolgica de Espaa 7 (12), 121139.
Guimer, J., Alonso, A., Mas, R., 1995. Inversion of an extensional-ramp
basin by a newly formed thrust: the Cameros Basin (N Spain). In:
Buchanan, J.G., Buchanan, P.G. (Eds.), Basin Inversion. Geological
Society Special Publication, vol. 88, pp. 433453.
Guiraud, M., Seguret, M., 1985. A realising solitary overstep model
for the late Jurassic-Early Cretaceous (Wealdian) Soria strike-slip
basin (Northern Spain). SEMP Special Publication 37, 159175.
Ingersoll, R.V., 1978. Petrofacies and petrologic evolution of the Late
Cretaceous forearc basin, northern and central California. Journal
of Geology 86, 335352.
Ingersoll, R.V., Bulard, T.F., Ford, R.L., Grimm, J.P., Pickle, J.D.,
Sares, S.W., 1984. The effect of grain size on detrital modes: a text
of the Gazzi-Dickinson point-counting method. Journal of
Sedimentary Petrology 54, 103116.
Julivert, M., 1983. La estructura de la Zona Asturoccidental-Leonesa.
In: Ros, J.M. (Ed.), Geologa de Espaa. Libro Jubilar. Instituto
Geolgico y Minero de Espaa, vol. T1, pp. 381407.
Julivert, M., Fontbot, J.M, Ribeiro, A., Navais Conde, L.E., 1972.
Mapa tectnico de la Pennsula Ibrica y Baleares, E. 1:1.000.000.
Instituto Geolgico y Minero de Espaa, Madrid.
Le Pera, E., Arribas, J., 2004. Sand composition in an Iberian passive-
margin fluvial course: the Tajo River. Sedimentary Geology 171,
261281.
Lindholm, R.C., Finkelman, R.B., 1972. Calcite staining: semiquan-
titative determination of ferrous iron. Journal of Sedimentary
Petrology 42, 239245.
Martn-Closas, M., Alonso Milln, A., 1998. Estratigrafa y Bioestra-
tigrafa (Charophyta) del Cretcico inferior en el sector occidental
de la Cuenca de Cameros (Cordillera Ibrica). Revista de la
Sociedad Geolgica de Espaa 11, 253269.
Mas, R., Alonso, A., Guimer, J., 1993. Evolucin tectonosedimen-
taria de una cuenca extensional intraplaca: la cuenca finijursica-
157
eocretcica de Los Cameros (La Rioja-Soria). Revista Sociedad
Geolgica de Espaa 6 (34), 129144.
Mas, R., Benito, M.I., Arribas, J., Serrano, A., Guimer, J., Alonso, A.,
Alonso-Azcarate, J., 2002. La Cuenca de Cameros: desde la
extensin finijursica-eocretcica a la inversin terciaria
implicaciones en la exploracin de hidrocarburos. Zuba. Instituto
de Estudios Riojanos 14, 964.
Mas, R., Benito, M.I., Arribas, J., Serrano, A., Alonso, A., Alonso-
Azcarate, J., 2003. The Cameros Basin: From late Jurassic-Early
Cretaceous Extension to Tertiary Contractional Inversion-Implications
of Hydrocarbon Exploration. In: AAPG International Conference and
Exhibition, Barcelona, Spain. Geological Field Trip 11 (52 pages).
Mas, R. (coord.), Garca, A (coord.), Salas, R., Melndez, A., Alonso,
A., Aurell, M., Bdenas, B., Benito, M.I., Carenas, B., Garca-
Hidalgo, J.F., Gil, J., Segura, M., 2004. 5.3.3. Segunda fase de
rifting: Jursico SuperiorCretcico Inferior. In: Geologa de
Espaa (J. Vera Ed.), Sociedad Geolgica de Espaa Instituto
Geolgico y Minero de Espaa, 503509.
Palomares, M., Arribas, J., 1993. Modern stream sands from
compound crystalline sources: composition and sand generation
index. In: Johnsson, M.J., Basu, A. (Eds.), Processes Controlling
the Composition of Clastic Sediments. Geological Society of
America, Special Paper, vol. 284, pp. 313322.
Pettijohn, F.J., Potter, P.E., Siever, R., 1972. Sand and Sandstones.
Springer Verlag, New York, p. 618.
Salas, R., Guimer, J., Mas, R., Martn-Closas, C., Melndez, A.,
Alonso, A., 2001. Evolution of the Mesozoic Central Iberian Rift
System and its Cainozoic Inversion (Iberian Chain). In: Cavazza,
W., Roberson, A.H.F.R., Ziegler, P. (Eds.), Peri-Tethyan Rift
Wrench Basins and Passive Margins. Mmoires du Musum
National d' Histoire Naturelle, vol. 186, pp. 145185.
Schwab, F., 1986. Sedimentary signatures of foreland basin
assemblages: real or counterfeit? In: Allen, P.A., Homewood, P.
(Eds.), Foreland Basins: International Association of Sedimentol-
ogists. Special Publication, vol. 8, pp. 395410.
Tortosa, A., Palomares, M., Arribas, J., 1991. Quartz grain types in
Holocene deposits from the Spanish Central System: some
problems in provenance analysis. Developments in Sedimentary
Provenance Studies. Geological Society Special Publication, vol.
57, pp. 4754.
Valloni, R., 1985. Reading provenance from modern marine sands. In:
Zuffa, G.G. (Ed.), Provenance of Arenites: The Netherlands,
Dordrecht, Reidel, pp. 309332.
Villaseca, C., Barbero, L., Huertas, M.J., Andonaegui, P., Bellido, F.,
1993. A cross-section though Hercynian granites of the Central
Iberian Zone, Excursion guide, Servicio de Publicaciones. Consejo
Superior de Investigaciones Cientficas, Madrid, p. 122.
Weltje, G.J., 2002. Quantitative analysis of detrital modes: statistically
rigorous confidence regions in ternary diagrams and their use in
sedimentary petrology. Earth-Science Reviews 57, 211253.
Zuffa, G.G., 1980. Hybrid arenites: their composition and classifica-
tion. Journal of Sedimentary Petrology 50, 2129.
Zuffa, G.G., 1985. Optical analyses of arenites: influence of methodol-
ogy on compositional results. In: Zuffa, G.G. (Ed.), Provenance of
arenites: The Netherlands, Dordrecht, Reidel, pp. 165189.
Zuffa, G.G., 1987. Unravelling hinterland and offshore palaeogeo-
graphy from deep-water arenites. Marine Clastic Sedimentology,
pp. 3961. Chapter 2.
Zuffa, G.G., Gaudio, W., Rovito, S., 1980. Detrital mode evolution of
the rifted continental-margin Longobucco sequence (Jurassic),
Calabrian Arc, Italy. Journal of Sedimentary Petrology 50 (1),
5161.
 Available online at www.sciencedirect.com

Sedimentary Geology 202 (2007) 158 173

www.elsevier.com/locate/sedgeo

Destruction of a fluvial reservoir by hydrothermal activity

(Cameros Basin, Spain)
M. Ochoa a,, J. Arribas a , R. Mas b , R.H. Goldstein c
a
Departamento de Petrologa y Geoqumica, Universidad Complutense de Madrid, Facultad de Ciencias Geolgicas,
C/ Jos Antonio Novais 2, 28040, Madrid, Spain
b
Departamento de Estratigrafa, Universidad Complutense de Madrid, Facultad de Ciencias Geolgicas,
C/ Jos Antonio Novais 2, 28040, Madrid, Spain
c
Department of Geology, University of Kansas 1475 Jayhawk Boulevard, 120 Lindley Hall, Lawrence, Kansas 66045-7613, USA

Abstract

This study provides an example of a high-quality fluvial hydrocarbon reservoir that was completely destroyed by hydrothermal
processes. The reservoir unit was deposited in the Cameros Basin, located in the NW sector of the Iberian Chain (Spain). The basin
was filled with clastic fluvial deposits (sandstones and conglomerates) between Late Berriasian and Early Aptian times. Provenance
of sands was mainly from coarse crystalline rocks. A humid tropical climate produced intense weathering of K-feldspar during
transport from source to basin. Thus, a mineralogically mature rigid framework with high porosity existed at the time of deposition,
which would have constituted a high-quality hydrocarbon reservoir. At present however, the porosity of the reservoir is negligible.
Porosity was reduced by a sequence of diagenetic processes: (1) mechanical compaction (i.e. crushing of metamorphic lithic grains)
and chemical compaction, (2) kaolinite and siderite cementation, and (3) early quartz cementation. Hydrocarbon emplacement
probably occurred between phases (2) and (3). A low-grade metamorphic (hydrothermal) event, reaching greenschist facies, took
place during the Late Cenomanian. It dramatically reduced the remaining porosity of the reservoir and destroyed the hydrocarbon
charge. Hydrothermal processes which affected the sandstones include (1) re-compaction; (2) late quartz cementation and
silicification of remaining feldspars; (3) carbonate cementation; (4) chloritization of feldspars, metamorphic lithic fragments and
intrabasinal argillaceous grains; and (5) growth of pyrite and chloritoid crystals on argillaceous material of intrabasinal,
extrabasinal or even diagenetic origin. Hydrocarbons that migrated to the margins of the basin escaped these hydrothermal
modifications and were preserved. The results of this study may be used to predict the diagenetic and hydrothermal evolution of
other potential reservoirs in similar tectonic settings.
2007 Elsevier B.V. All rights reserved.

Keywords: Sandstone diagenesis; Hydrothermal activity; Metamorphism; Porosity reduction; Cameros Basin; Spain

1. Introduction Tithonian to the Early Albian (Fig. 1). During this

Among the Mesozoic basins of the Iberian Chain
(Spain), the Cameros Basin preserves the most complete
sedimentary record of the time interval from the
Corresponding author.
E-mail address: mochoa@geo.ucm.es (M. Ochoa).
0037-0738/$ - see front matter 2007 Elsevier B.V. All rights reserved.
doi:10.1016/j.sedgeo.2007.05.017
period, the anomalously high subsidence and sedimen-
tation rates within the basin produced a sedimentary
record that reached up to 5 km in vertical thickness
(representing up to 9 km of stratigraphic record in the
direction of the northward migration of the successive
depositional sequences which filled the basin). Two
sectors with different features can be distinguished: the
 M. Ochoa et al. / Sedimentary Geology 202 (2007) 158173
Fig. 1. Simplified geologic map showing study area and location of the transect through the outcrops. The stratigraphic sections are located in Muriel (A), Cidones-Abejar (B), Golmayo (C), Yanguas (D),
San Pedro Manrique (E), Valdemadera (F), Trevijano (G), Jubera (H), Arnedillo (I) and Prjano (J).

159
160 M. Ochoa et al. / Sedimentary Geology 202 (2007) 158173
northeastern sector, characterized by the most rapid rates
of subsidence and affected by low-grade (hydrothermal)
metamorphism; and the southwestern sector, character-
ized by the presence of secondary depocenters and
unaffected by metamorphism.
The focus of this study is the northern and central part of
the eastern Cameros Basin (Fig. 1, outcrops from D to J). A
SWNE trending geological cross-section from this area
is shown in Fig. 2. In this area, Mas et al. (2003) described
eight depositional sequences (DS) ranging in age from
Tithonian to Early Albian (Fig. 3). This study is focused
on the maximum synrift filling stage (DS-4 to DS-7),
from Late Berriasian to Early Aptian (Urbin Group), in
which a succession of up to 2.2 km of clastic sediments
was formed.
Although there are many publications dealing with
stratigraphy, sedimentology and mineralogy of the Urbin
Group (e.g. Salinas and Mas, 1989; Casquet et al., 1992;
Mas et al., 1993; Alonso-Azcrate et al., 1995; Mantilla
et al., 1998; Alonso-Azcrate et al., 1999a,b; Barrenechea
et al., 2001; Mata et al., 2001; Mantilla et al., 2002; Mas
et al., 2003), little attention has been paid to reservoir
characterisation and evaluation of these clastic sediments
Fig. 2. Stratigraphic transect of Urbin Group through the southern area (S), central area (C) and northern area (N) of the basin. See Fig. 1 for location
of sections.
(Arribas et al., 2003). The petroleum systems of the basin
are discussed by Mas et al. (2002). The absence of hy-
drocarbons in the basin fill prompted the present analysis
of the metamorphic (hydrothermal) processes that caused
their destruction.
This paper documents the destructive effects of burial
diagenesis and low-grade hydrothermal metamorphism on
reservoir quality, and provides insights into the evolution
of hydrocarbon reservoirs in other basins affected by these
processes. Hydrothermal alteration of clastic reservoirs
has been frequently documented (i.e., Hoffman and
Hower, 1979; Larese, 1997). Common related processes
include mineral cementation and recrystallization, gener-
ated by acid hot water flows, which promotes porosity/
permeability reduction (i.e., Oelkers et al., 1996; De Ros
et al., 2000). Low-grade hydrothermal metamorphism
may produce a complete occlusion of fluid pathways in the
reservoir (Dutkiewicz et al., 1995; Buick et al., 1998). In
many basins hydrothermal metamorphism develops local-
ly, and preservation of economically exploitable reservoirs
may occur in areas of the basin unaffected by metamor-
phism (Sptl et al., 1994; Pittman and Splt, 1996; Sptl
et al., 2000). Thus, understanding the processes that
 M. Ochoa et al. / Sedimentary Geology 202 (2007) 158173
Fig. 3. Rift and post-rift stages in the evolution of the Iberian Basin (Modified from Salas et al., 2001) and idealised stratigraphic section showing the
latest Paleozoic and Mesozoic sratigraphic record in the Cameros area.
affected palaeo-reservoir quality is an important objec-
tive in predictive diagenetic reservoir modelling.
2. Geological setting
The Cameros Basin is an intracratonic basin, rep-
resenting a folded segment of the Alpine Chain that
developed as a consequence of the contractional inver-
sion of the Mesozoic Iberian rift basin. The development
of this basin took place in two phases: the first phase
lasted from Early Permian to the Late Triassic and the
second from Tithonian to Early Albian. These two
rifting phases were followed by post-rift thermal sub-
sidence, during which mainly shallow marine carbo-
nates were deposited (Guimer et al., 1995; Salas et al.,
2001) (Figs. 13).
During the second stage of rifting, a thick succession
(b5 km total thickness) of alluvial to lacustrine, depo-
sitional sequences was accumulated in the depocentre of
the Cameros Basin (Fig. 3). According to Mas et al.
(2003), the basin fill constitutes a large megasequence
bounded by two main unconformities (Fig. 3). The
megasequence can be subdivided into eight depositional
sequences (DS-1 to 8); this study focused on the interval
between DS-4 and DS-7, which constitutes the Urbin
Group.
The thickness of the Urbin Group varies dramati-
cally from 100 m in the marginal areas of the basin to
2200 m in the depocentres (Fig. 2). It comprises sand-
161
stones and conglomerates interbedded with shales,
which were deposited in settings ranging from braided
systems (Mas et al., 2003) in the southwestern area of
the basin (A to C in Fig. 1) to meandering fluvial and
lacustrine systems (Mas et al., 2003) in the central and
northeastern areas (D to F in Fig. 1); and alluvial-fan
facies (Mas et al., 2003) in the northern part of the basin
(G to J in Fig. 1).
3. Basin evolution, potential source rocks and
metamorphism
Mas et al. (2003) summarized the genesis and evo-
lution of the Cameros Basin (Fig. 4). Extensive car-
bonate platforms developed during Jurassic times, i.e.,
before the Tithonian to Early Albian rifting phase (1, in
Fig. 4). The Cameros Trough began to form as an
extensional ramp basin from Tithonian to Berriasian
time. Depositional Sequences 1 to 3 were deposited in
this time interval (2, in Fig. 4). The maximum thickness
of strata is reached in the DS-4 to DS-8 deposits of the
central zone, corresponding to the time interval of max-
imum extension from Berriasian to Early Albian. This
initiated organic matter maturation and hydrocarbon
migration towards permeable units in the basin (3a and
3b, in Fig. 4). Marine environments re-occupied the
Cameros area during the Late Cretaceous, constituting
the Post-Rift Stage 2. During Cenomanian times, hydro-
thermal metamorphism affected a significant volume of
162 M. Ochoa et al. / Sedimentary Geology 202 (2007) 158173

Fig. 4. Cross-sections illustrating the geological history of the Cameros Basin. See text for explanation (modified from Mas et al., 2003).

the basin fill in the Cameros Trough and hydrocarbon The potential petroleum systems of the Cameros area
reservoirs were destroyed (4a and 4b, in Fig. 4). Finally, have been reviewed by Mas et al. (2002, 2003) based on
Tertiary contractional inversion of the Cameros Basin an analysis of the Cameros Extensional Ramp Basin
took place (5, in Fig. 4). (Cameros Trough) depocentre and two surrounding
 M. Ochoa et al. / Sedimentary Geology 202 (2007) 158173
half-graben sub-basins (Rioja Trough to the North and
Bigornia Trough to the South). The Cameros Trough is
considered the central part of the Cameros Basin in
which hydrothermal activity took place (Mas et al.,
2003), and where early generation and migration of oil
from source rocks (today over-mature) to reservoirs took
place. Potential source rocks of the Cameros Trough are
Callovian organic marls and black shales (type II, Tissot
and Welte, 1978), Early Cretaceous lacustrine organic
marls, (type II/I, Tissot and Welte, 1978), and syn-rift
Early Cretaceous lignites and organic-rich shales (type
III, Tissot and Welte, 1978) (Mas et al., 2003). Among
the most important reservoirs were syn-rift, Early
Cretaceous channelized fluvial sandstones (DS 4 to 7)
(Mas et al., 2003). The rocks that represented the seals at
that time were the interbedded mudstones. The timing of
hydrocarbon generation can be considered as early and
produced in two events: (1) Albian, during which or-
ganic matter matured by burial and (2) Late Cretaceous,
during which maturation of organic matter was induced
by abnormal heat flow caused by the first hydrothermal
event (Mas et al., 2003). The early emplacement of
hydrocarbons is supported by petrography as they pre-
date quartz overgrowths.
The metamorphism of the basin is considered as
hydrothermal and allochemical (Casquet et al., 1992;
Barrenechea et al., 1995; Alonso-Azcrate et al., 1995,
2001; Mantilla et al., 2002). Two thermal events can be
recognized (e.g. Mantilla et al., 2002). The first one is the
most relevant here and took place after complete filling of
the basin. It has been dated at the post-rift age of 106 to
86 Ma BP (from Late Albian to Coniacian). Metamorphic
conditions ranged from low-grade to very low-grade, with
a maximum temperature of 340370 C and a maximum
pressure of 1 kbar (aprox. 3900 m). This hydrothermal
event coincides with other regional geotectonic events
such as the opening of the Gulf of Biscay (Olivet et al.,
1984), and is coincident with the age of the alkaline
magmatism in the Pyrenees (Montigny et al., 1986).
Another hydrothermal phase has been recorded at 40 Ma
(Paleogene). This phase was related to the beginning of
basin inversion during the Alpine contraction.
4. Methods
Sandstone sampling was carried out in several strat-
igraphic sections (Figs. 1, 2). In the southern part of the
basin, these sections are located at Muriel (A), Cidones-
Abejar (B) and Golmayo (C). In the central part, locations
of the sections are Yanguas (D), San Pedro Manrique (E)
and Valdemadera (F). In the north, locations of these
sections are Trevijano (G), Jubera (H), Arnedillo (I) and
163
Prjano (J). About 150 samples of sandstone were an-
alysed to determine clastic and authigenic products. An-
alytical methods included conventional petrography, fluid
inclusion microthermometry, cathodoluminescence, and
analyses with electron microprobe and scanning electron
microscope (SEM).
For petrographic analysis, standard doubled-polished
thin sections were etched and stained using HF and
sodium cobaltinitrite for potassium feldspar identifica-
tion, and alizarin red and potassium ferrocyanide to better
distinguish carbonate components (Chayes, 1952; Lind-
holm and Finkelman, 1972, respectively). Four hundred
points were counted per sample for quantitative petro-
graphic analyses. The Gazzi-Dickinson method (Inger-
soll et al., 1984) and petrographic groups defined by Zuffa
(1980) were used to classify points. Post-depositional
changes to the original framework were also considered
and evaluated. These analyses permit a reconstruction of
the original framework composition required for prove-
nance deductions and diagenetic inferences.
Electron microprobe analysis was used to character-
ise carbonates cements, feldspars, clay minerals and
solid fibres of hydrocarbons. Acceleration voltage and
sample current intensity were set at 15 Kv and 20 nA.
Standards used were TAP, LD2, PETJ, PETJH, LIF and
LIFH. Many images were taken under the microprobe in
backscattered light in order to recognize different bright-
ness levels.
Double-polished thin rock slices detached from their
glass mount were prepared from most samples for fluid
inclusion microthermometric analysis. The thickness of
the slices was between 30 and 50 m. In all cases, cold
preparation techniques were used to avoid re-equilibra-
tion of fluid inclusions (Goldstein and Reynolds, 1994).
The measurements were carried out on a petrographic
microscope equipped with a Linkam THMSG 600
heating and cooling stage which enables temperatures of
phase transitions in the range of 180 to 300 C. This
heating stage is mounted on an Olympus BX60 po-
larizing microscope.
Small fragments of double-polished rock slices were
also imaged under CL using an Oxford Instrument
photomultifilter-based CL detector installed on a JEOL
T330 SEM. Cathodoluminescence photography was
performed after homogenisation temperatures had been
determined in order to avoid any negative effects on data
acquisition.
SEM-CL textures were used to verify that the studied
fluid inclusions were located in quartz cement and not in
quartz clasts, to evaluate the possibility of finding some
zoning inside the quartz overgrowth itself, and also to
determine the timing of quartz cementation relative to
164 M. Ochoa et al. / Sedimentary Geology 202 (2007) 158173
other diagenetic processes. The latter contributed to es-
tablishing the diagenetic sequence on the basis of textural
relationships.
5. Results
5.1. Petrofacies
Following the criteria of Dickinson (1985), it is
possible to note a trend in framework composition from
South to North (Fig. 5). In proximal zones (southern
area), the sandstone composition is quartzofeldspathic
(mean Q81F18Lt1; Fig. 6A) and K-feldspar prevails over
plagioclase. Sandstones from the southern part corre-
spond to subarkoses instead of ideal arkoses due to
chemical weathering in a tropical climate, which pro-
moted a rapid decay of feldspars (Ochoa et al., 2004).
These rocks evolve to more mature quartzose sand-
stones in depocentres (mean Q96F3Lt1; Fig. 6B) suggest-
ing significant mineralogical and textural maturation
during transport over a distance of N 50 km. This sys-
tematic spatial variation can also be attributed to intense
weathering in a humid climate (Rat, 1982). In addition,
local sources of Triassic and Jurassic sedimentary rocks
(carbonate and clastics) produced quartzolithic sand-
stone petrofacies (mean Q93F1Lt6) (Fig. 6C), in alluvial-
fan environments in the northeastern part of the basin. In
spite of the different petrofacies, all sandstones are
considered to have an original quartz-rich rigid frame-
work. Their mean grain-sizes range from fine to coarse.
They are moderately to well sorted, which suggests that
the original porosity was between 34% and 40% (Beard
and Weyl, 1973). The high values of original porosity
and the relatively low clay content characterize these
deposits as very good reservoir rocks.
Fig. 5. Ternary plots showing sandstone petrofacies in the Urbin
Group (according to Dickinson, 1985) in the southern area (S), central
area (C) and northern area (N) of the basin. Qm: Monocrystalline
quartz; F: K-feldspar; Lt: total lithic population.
5.2. Diagenetic and hydrothermal processes and products
Urbin Group sandstones experienced intense dia-
genesis. The most relevant processes are indicated in
Fig. 8.
5.2.1. Compaction
Compaction has been evaluated following the criteria
of Houseknecht (1987) and Lundegard (1992) for
estimation of the intergranular volume of the original
sediment and its reduction during diagenesis. Porosity loss
by compaction (COPL, Lundegard, 1992) in the analyzed
sandstones ranges from 11.9% to 37.29% for sandstones
in the north; from 23.6% to 39.6% in the central area; and
from 15.4% to 39% in the south (Table 1). Note that these
values include both mechanical and chemical compaction.
According to Lundegard (1992), the ICOMPACT (Com-
pactional Index = COPL/(COPL + CEPL)) in the Urbin
Group varies between 0.6 and 0.9, indicating that com-
paction can be considered the main process responsible for
porosity loss (Figs. 7A and 9). Likewise, different dia-
genetic pathways of sandstones from the north, centre and
south indicate that compaction has most strongly affected
the central part of the basin, which has experienced the
most subsidence.
5.2.2. Cements and replacements
Several diagenetic precipitates occluded the inter-
granular pore space. Some of these replace framework
components or other cements.
5.2.2.1. Clay Minerals. Kaolinite occurs as pore-filling
cement in all areas, occupying intergranular pore space
and occasionally as replacement of K-feldspar, generat-
ing diagenetic matrix (epimatrix). Its occurrence is highly
variable (mean 0.7%, 1.6%, 8.9% in the northern, central
and southern areas respectively).
It is possible to distinguish two generations of pore
filling, an earlier generation included in the eodiagenesis
and a later generation that fills secondary pores. The first
pore filling appears to be enclosed by other cements and
deformed by mechanical compaction, which suggest an
early stage of generation. Eodiagenetic kaolinite is com-
mon in warm and wet continental environments (i.e.
Worden and Morad, 2003). The last phase shows vadose
features and is included in the telodiagenetic stage. It has
only been recognized in samples from the southern area of
the basin. Telodiagenetic kaolinite must be associated with
an influx of low-pH meteoric waters during uplift (Lanson
et al., 2002; Ketzer et al., 2003; Worden and Morad, 2003).
Illite appears as pore lining around detrital grains and
as replacements of K-feldspar (epimatrix, Fig. 6E). The
 M. Ochoa et al. / Sedimentary Geology 202 (2007) 158173 165

Fig. 6. Thin-section photomicrographs of detrital components and diagenetic features in Urbin Group sandstones. A) Medium-grained sandstone
from the southern area of Cameros Basin, showing a quartzose framework with K-feldspar (K) grains. Plane-polarized light. B) Medium-grained
quartzarenite from the central area of the basin. The long and concave-convex grain-to-grain contacts suggest that chemical compaction occurred.
Cross-polarized light. C) Coarse-grained sublitharenite from the northern area of the basin, exhibiting a well-sorted framework with carbonate grains
(echinoid fragment). Plane-polarized light. D) Illite pore lining in a quartzarenite from the northern area. E) K-feldspar cement ([K]) in a very coarse-
grained quartzarenite of the central area. Cross-polarized light. F) Fibres of hydrocarbons showing features of deformation. G) Hydrocarbon traces
surrounding quartz grains. H) Tiny siderite rhomb suggesting an early diagenetic origin. Note the diagenetic matrix (epimatrix) on the left side,
produced by the replacement of a K-feldspar by clay minerals. Plane-polarized light.

pore lining is inferred to be early and related to the centage 13.2%, 15.7% and 2.9%, mean 2.3%, 2.8% and
eodiagenetic stage. It is generally thin and discontinuous 0.5% in the northern, central and southern areas re-
(Fig. 6D). Its occurrence is variable (maximum per- spectively). The illite pore lining is tangential to grain
166 M. Ochoa et al. / Sedimentary Geology 202 (2007) 158173

Fig. 7. Thin-section photomicrographs of detrital components and diagenetic features in Urbin Group sandstones. A) Non-carbonate intrabasinal
grain (NCI, rip-up clast) deformed by mechanical compaction. Cross-polarized light. B) SEM cathodoluminescence image of a detrital quartz grain
with internal zoning in the overgrowth. Note the crushing and microfracturing of the grain before quartz precipitation. C) Quartz replacing a still
recognisable K-feldspar grain, cross-polarised light. D) Backscattered image of a plagioclase from a sample in the northern area with relicts of K-feldspar.
This could be evidence of albitization. E) Backscattered image of carbonate cement from the central area. F) prismatic chloritoid evidencing a low-grade
metamorphic process. Cross-polarized light. G) SEM cathodoluminescence image of detrital quartz grains from the northern area showing two phases of
quartz cement. The sketch on the right side better differentiates the two quartz cementation phases.

surfaces. Some illitic coatings are texturally similar to
smectite, which suggests a smectite precursor (Morad
et al., 1994) that was illitised during diagenesis. Re-
placement of K-feldspar by illite is considered to have
taken place during mesodiagenesis in a relatively closed
system at 120 C140 C, and at burial depths greater
than 3.7 km (Chuhan et al., 2000, 2001). Illitisation of
feldspar pre-dates quartz cementation. It is not clear
whether the process continued after precipitation of
quartz.
5.2.2.2. K-feldspar. This cement occurs as euhedral,
thin and discontinuous overgrowths and is commonly
corroded and replaced by illite and carbonate cements
(Fig. 6E). It attains 0.5% in all areas of the basin. It is a
very early phase, which predates quartz and carbonate
cements. This indicates early precipitation in eodiage-
netic environments (Fig. 8).
5.2.2.3. Injection of hydrocarbons. The first event of
hydrocarbon injection is inferred to have occurred early,
because some hydrocarbon fibres lie between quartz
grains, pre-dating quartz cements (Fig. 6F and G). The
fibres occur as very thin, brown coatings. The presence
of these hydrocarbon residues has been confirmed with
electron microprobe analysis and with organic solvents.
The residues are related to the migration of oil into the
reservoir, which took place in the Late AptianEarly
Albian, prior to the hydrothermal event. This migration
involved the Cameros Trough and its satellite basins,
located to the north and south. Furthermore, Mas et al.
(2003) suggest that the tectonic inversion of the basin
 M. Ochoa et al. / Sedimentary Geology 202 (2007) 158173
Fig. 8. Chronology of the main post-depositional processes and products ascribed to eodiagenesis, mesodiagenesis, telodiagenesis and to the
metamorphic event (shaded zone).
(EoceneEarly Miocene) produced another late migra-
tion of hydrocarbons towards satellite basins.
5.2.2.4. Siderite. Siderite occurs as isolated crystals
(Fig. 6H) and is considered an early phase, as it is often
appears embedded in quartz cement. Siderite precipita-
tion requires a reducing environment related to metha-
nogenesis (e. g. Mozley and Wersin, 1992).
5.2.2.5. Early and late quartz. Quartz is generally the
most abundant cement but is variable in abundance. It is
mainly developed in the northern and central areas,
reaching 9.9% and 12.9% of total rock volume (mean,
2% and 2.8%) respectively.
Diagenetic quartz cement appears as overgrowths
around detrital grains, and locally includes other earlier
diagenetic phases such as clay minerals (kaolinite and
illite) and siderite. Quartz cement typically appears to
have been corroded by other minerals (i.e. carbonates). It
postdates compaction, which indicates that it precipitated
during relatively deep burial (mesodiagenesis, Fig. 8) at
depths of more than 2.5 km and temperatures above 80 C
(e.g. Worden and Morad, 2000; Walderhaug et al., 2001).
In addition, quartz cement postdates hydrocarbon
emplacement. Although early hydrocarbon emplacement
has been cited as a barrier to authigenic quartz nucleation
(Worden et al., 1998; Bloch et al., 2002), quartz cemen-
tation is possible regardless of the timing of hydrocarbon
generation and migration (Reed et al., 2005). Pressure
solution, clay mineral reactions in intercalated mudstones,
167
and the dissolution of K-feldspars during burial diagenesis
(Bjrlykke and Egeberg, 1993; Barclay and Worden,
2000; Bloch et al., 2002) are likely sources for quartz
cement. Quartz cementation is facilitated in the presence of
micas, clay minerals and/or organic matter, which release
the water necessary to diffuse silica out of the contact zone
(Wilson, 1994). Chemical compaction generated quartz
cementation prior to the metamorphic peak event.
In addition, the fluid inclusion analysis discussed
below indicates the presence of a second quartz cement
phase (late quartz in Fig. 8) that is ascribed to the
metamorphic event. This phase of quartz cementation is
recognizable on cathodoluminescence images as a
zonation in the overgrowths (Fig. 7B). The zonations
in the quartz overgrowths are due to variations in alu-
minium and transition metals (Kraishan et al., 2000).
5.2.2.6. Secondary porosity. Secondary porosity is
high in the southern area (maximum value of 20.1%,
mean 9.2%), and decreases to lower values in the central
and northern areas (mean 2.5 and 7.9%, respectively). It is
manifested by the presence of corroded grains and partial
dissolution and originated by the dissolution of framework
grains, mainly K-feldspars and carbonates. Occasionally,
secondary pores have developed in intergranular cement.
5.2.2.7. Hematite. Hematite cement occurs as grain
coatings or as secondary pore filling. Its occurrence is
very variable (mean 1.2%, 1.1% and 3.4% in the northern,
central and southern areas respectively). Occurrences of
168 M. Ochoa et al. / Sedimentary Geology 202 (2007) 158173
this cement are typically accompanied by high values of
secondary porosity. Hematite cement precipitates under
oxidizing conditions and may appear as an early grain
coating or as a telodiagenetic product, resulting from de-
dolomitization of Fe-rich dolomite at shallow depths
(Morad et al., 1995) and pyrite oxidation.
5.2.3. Metamorphic processes and products
In addition to diagenetic processes and products, hydro-
thermal imprints on the sandstones have been observed in
samples from the northern and central areas of the basin.
5.2.3.1. Pyrite precipitation. The percentage of pyrite
considered of metamorphic origin ranges from 0.2% to
1.4% in the northern area (mean 0.1%) and from 0.2% to
8.2% (mean 0.9%) in the central area. These pyrite
crystals appear as euhedral coarse crystals that cement and
replace framework and previous diagenetic products in
the sandstones (Alonso-Azcrate et al., 2001). This pyrite
is attributed to highly saline waters that produced sulphur
mineralisation in the presence of hydrocarbons (Rowe and
Burley, 1997; Alonso-Azcrate et al., 2001). Tempera-
tures of 367 C have been estimated for the precipitation
of this mineral (Alonso-Azcrate et al., 1999a).
5.2.3.2. Replacement of K-feldspars by quartz. Re-
placement of K-feldspars by quartz is common in samples
from the central and northern areas (Fig. 7C). It occurs in
percentages between 0.2% and 1.6%, in the northern area
and between 0.2% and 4.2%, in the central area and is
manifested by the presence of illite inclusions on quartz
overgrowths that preserve the external shape of precursor
K-feldspar grains. This process requires high silica
concentrations and high temperatures. For this reason, it
must be related to hydrothermal fluids supersaturated in
silica.
5.2.3.3. Carbonate cements. Dolomite appears as
small rhombic crystals that locally occlude pores. Its
content varies from 0.2% to 36.1%. The mean values are
12.5% in the northern area and 1.5% in the central area.
Dolomite crystals typically have abundant fluid inclu-
sions that give the cements a turbid aspect in transmitted
light. Dolomite cement has a bright cathodolumines-
cence signal that is explained by anoxic conditions
during precipitation (Tucker, 1988). The origin of this
dolomite is presumably related to hydrothermal fluids,
and postdates the pyrite phase. Later carbonate cements,
especially calcite, frequently replaced dolomite cement
in the telodiagenetic stage.
Ankerite occurs as isolated patches and is replaced by
later calcite cement (Fig. 7E). Ankerite cement occupies
such a low percentage of intergranular volume that it has
not been quantified. In the central part of the basin,
ankerite has bright luminescence (Fig. 7E) and high values
of Mn (manganic ankerite). In many cases, the trace-
element content of late diagenetic carbonate cements is
depleted in iron and magnesium, and enriched in
manganese (Lynch and Land, 1996; Milliken, 1998).
Alternatively the high Mn values can be generated by
hydrothermal flow (Chow et al., 1996; Morad et al., 2000).
5.2.3.4. Albitization. The presence of twinned and
untwinned albite in the northern and central areas of the
basin suggests that albitization took place. The albites
appear as fresh idiomorphic and subidiomorphic grains,
and sometimes have textural evidence of K-feldspar
replacement (Fig. 7D). The composition of authigenic
albite is close to the Na end member and is largely non-
luminescent (Kastner and Siever, 1979). Albitization pre-
dates several carbonate cementation phases and is
common as a replacement in depths greater than 2500 m,
especially in sandstones rich in K-feldspars (Morad et al.,
1990). The Na required for albitization is probably derived
from dissolution of Triassic evaporites (Saigal et al., 1988).
5.2.3.5. Chlorite and chloritoid. A metamorphic
event is also inferred from the presence of two meta-
morphic minerals, chlorite and chloritoid, in samples
from the northern and central areas of the basin
(Fig. 7F). The content of chlorite varies from 0.2% to
4.2% (mean 1.5%) in the northern area and from 0.2% to
10% (mean 2%) in the central area.
Chlorite replaces quartz, K-feldspars and micas. It
appears on argillaceous framework grains (intrabasinal
grains and metamorphic rock fragments) and is asso-
ciated with clay minerals from the sandstone matrix.
Diagenetic chlorite is recognized as polycrystalline
micronodules or as radial aggregates. The content of
chloritoid reaches 2.4% (mean 0.4%). It appears as
tabular crystals or in aggregates. It is possible that these
crystals retrograde into chlorite whilst maintaining their
previous morphology (pseudomorphosis).
5.3. The imprint of hydrothermal processes on quartz
overgrowths
Under transmitted light, quartz overgrowths are
generally 20 to 40 m thick and display uniform extinc-
tion. Two types of quartz cement can be recognized
under SEM-CL, which will be referred to as cathodo-
luminescence zones Q1 and Q2. These two cements
occur together in different samples from the northern
and central areas of the basin.
 M. Ochoa et al. / Sedimentary Geology 202 (2007) 158173 169

Q1 (Fig. 7G) is dark under CL and it preserves

ghosts of fibrous structure that are perpendicular to
the detrital grain-overgrowth boundary. Q2 is bright and
has concentric growth zoning (Fig. 7B) or is iso-
pachous. Locally, this cement displays irregular bound-
aries inconsistent with growth and has a patchy
distribution. The dark Q1 cement has been fractured
and its spaces filled with bright Q2. Fragments of Q1 are
frequently included in the Q2 cement. The destruction
of the original texture of Q1 cements is probably due to
high-temperature recrystallization (Goldstein and
Rossi, 2002). Hence, the hot hydrothermal fluids that
generated the bright Q2 cement are likely to have
caused the recrystallization of Q1 as well. In contrast,
Q1 cement is less common in sandstones from the
central area and where it appears is thin and discontin-
uous. This could also be related to the predominance of
hydrothermal processes, which lowered the preserva-
tion potential of Q1 in the central area relative to the
northern area.
Homogenization temperatures were measured in
syntaxial quartz overgrowths in sandstones from the
northern and central areas. Two types of fluid inclusions
were observed in distinct parts of quartz overgrowths, as
shown by SEM-associated cathodoluminescence images.
One type of fluid inclusion (the first generation) occurs in
the dark, non-luminescent phase (Q1), closest to the
detrital grain-overgrowth boundary. It was formed at low
temperatures (from 100.5 C to 115 C, mean 110 C),
and is interpreted to have a diagenetic origin. The other Fig. 9. Diagenetic paths diagrams showing the loss of porosity by
type of fluid inclusion (the second generation) is included cementation processes (CEPL) versus compaction processes (COPL)
in bright and more luminescent cements (Q2, Fig. 7G), of the studied sandstones in the southern, central and northern areas of
the basin, according to Lundegard (1992).
further away from detrital grain-overgrowth boundaries. It
was formed at higher temperatures (from 95 C to 227 C,
mean 175 C), and is considered to be of hydrothermal The circulation of hydrothermal fluids took place
origin (Fig. 9). We propose that the high temperatures after the basin was filled. This suggests that hydrother-
reflect the injection of hydrothermal fluids whereas the mal activity was not restricted to the deepest part of the
low temperatures reflect maximum burial. The spatial basin (Barrenechea et al., 2001). The maximum
distribution of the two types of inclusions and their temperatures did not exceed 340 C (highest metamor-
temperatures of formation indicate that the Q2 phase is of phic conditions) and the pressure remained below 1.0 Kb
hydrothermal origin. (around 3.5 to 4 km deep). Fluid circulation could have
The first cement (Q1) precipitated from a single fluid been enhanced by extensional fractures, sedimentary
reservoir, because salinity values are homogeneous within discontinuities and the relatively high permeability of the
each sample, whereas the fluid inclusions with higher deposits (Alonso-Azcrate et al., 1999a).
temperatures are inferred to correspond to a mixture of
two fluid reservoirs, recognized by different salinity 6. Discussion
values for similar homogenization temperatures (Th) in
the same sample. For Th = 100 C, salinity values between Although petrofacies vary from south (quartzofelds-
3.4 and 6.2 were registered. These reservoirs could pathic) to north (quartzose and quartzolithic, respec-
have been affected by the hydrothermal metamorphism tively), all sandstones from the Urbin Group in the
related to the rifting evolution (Casquet et al., 1992; Cameros Basin had a rigid framework with high original
Barrenechea et al., 1995). porosity values at the time of deposition. Textural
170 M. Ochoa et al. / Sedimentary Geology 202 (2007) 158173
maturation of sediments during transport and intense
weathering produced moderately to well-sorted sand-
stones with a quartzose framework. The presence of
primary porosity in some samples from the northern and
southern areas indicates that conduits for fluid flow were
not completely eliminated during burial. Together, these
features suggest that the fluvial sandstones were
originally a high-quality reservoir. Unfortunately, dia-
genetic and metamorphic processes degraded these
original characteristics. Diagenetic processes reduced
original porosity by intense compaction (COPL) and
cementation (CEPL). Concavo-convex contacts, which
indicate pressure solution, are common. Mechanical and
chemical compaction was the most important mecha-
nism of porosity reduction in the basin.
Several cementation phases have been distinguished.
Pore-filling kaolinite, pore-lining illite and K-feldspar
cements were recognized. These cementation phases
started to fill the intergranular pores in the eodiagenetic
stage. Petrographic observations suggest that kaolinite
formed at shallow depth. An influx of fresh water is a
likely explanation for its origin (Bjrlykke, 1998). K-
feldspar overgrowths formed at relatively shallow burial
depths. K-feldspar precipitation requires high silica
activities and high K + /H + ratios (Morad et al., 2000)
and the occurrence of feldspar overgrowths may imply
that detrital feldspars were dissolving (De Ros et al.,
1994). Generation and migration of hydrocarbons
occurred during early mesodiagenesis. Today, this is
recognizable by the presence of carbon residues
appearing as thin fibres, locally showing deformation
features. These are the result of mixing of hydrocarbons
with subsequent hydrothermal fluids, which generated a
carbon residue and triggered water release.
In addition, in the mesodiagenetic realm, a re-com-
paction event took place that induced breakage of clasts
and previous cements by framework collapse, and
reduced the porosity (Figs. 7B and 10). Re-compaction
has been inferred only in samples from the northern and
southern areas of the basin, in which sufficient K-
feldspar and carbonate cement were present to promote
framework collapse by dissolution. Thus, porosity
reduction is proposed to have occurred in four stages
(Fig. 9): (1) Loss of primary porosity by early
mechanical compaction; (2) early cementation, mainly
by kaolinite and K-feldspar; (3) dissolution of frame-
work grains in the southern area and dissolution of
carbonates in the northern area; and (4) framework
collapse by re-compaction.
The diagenetic trend of sandstones from the central
area is different, for many reasons. Firstly, there are not
as many K-feldspars that are prone to dissolution, so re-
Fig. 10. Histograms of homogenisation temperatures in primary fluid
inclusions in quartz overgrowths. Notice the two populations of fluid
inclusion homogenisation temperatures.
compaction did not take place. Secondly, the central area
experienced the most subsidence, which means that
compaction has been greatest there. Also, quartz cement
seems to be the main cause of porosity reduction in the
central area relative to the northern and southern areas.
Carbonate cementation has been of minor importance for
porosity reduction. The diagenetic trend implies constant
compaction with minimal carbonate cementation.
In the northern and central areas of the basin, hydro-
thermal metamorphism destroyed reservoirs and caused
a second generation of quartz overgrowths and replace-
ment of K-feldspar by quartz, in addition to carbonate
cementation (including dolomite and ankerite), albitiza-
tion, and precipitation of pyrite and chlorite-chloritoid.
All these processes are consistent with the presence of
fluids supersaturated in silica, which are typical of hy-
drothermal systems (Fournier, 1985). Moreover, accord-
ing to Alonso-Azcrate et al. (1999a), the estimated
temperatures for pyrite precipitation are consistent with
the hydrothermal metamorphic realm.
Textural and fluid inclusion evidence indicates that
precipitation of silica of diagenetic and metamorphic
origin occurred throughout the burial history. Tempera-
tures deduced from fluid inclusion analysis in the second
generation of quartz overgrowths of metamorphic origin
range from 95 C to 227 C, which is compatible with the
maximum temperature of 340 C proposed by Casquet
et al. (1992) for the metamorphic peak. The hydrothermal
phase that affected the Cameros Trough also destroyed the
hydrocarbon charge and the porosity of the reservoirs.
Nevertheless, those hydrocarbons that migrated early
from the Cameros depocentre towards the margins of the
basin may have been preserved, as they were not affected
by hydrothermal metamorphism (Mas et al., 2003). Thus,
sufficient permeability and intergranular volume must
 M. Ochoa et al. / Sedimentary Geology 202 (2007) 158173
have been preserved in rigid-grain sandstones, despite
significant burial.
The final processes that affected the Urbin Group
rocks are related to telodiagenesis, and include the
generation of secondary porosity, cementation of
haematite, oxidation of pyrites, and replacement of
dolomite by calcite. All these processes were more
intense in the northern and southern areas than in the
central area, possibly because circulation of meteoric
waters was slower in the central area of the basin, due to
the lower porosity values of the hydrothermally altered
sandstones.
7. Conclusions
1. The dispersal systems that fed the Cameros Basin
produced large volumes of sands with quartz-rich
petrofacies during sedimentation of the Urbion Group.
Their rigid framework and low matrix content pro-
moted high primary porosity values favourable for a
high-quality reservoir.
2. The provenance of the sandstones has influenced
diagenesis mainly by controlling the spatial distribu-
tion of K-feldspars. The most important diagenetic
processes are compaction (mechanical and chemical)
and cementation by kaolinite, siderite and quartz.
Hydrocarbon emplacement occurred before quartz
cementation.
3. Diagenetic processes related to hydrothermal meta-
morphism significantly reduced porosity and destroyed
all reservoir potential in the Cameros Trough. Dom-
inant processes were re-compaction, late quartz and
carbonate cementation, and growth of chlorite, chlor-
itoid and pyrite.
4. The preservation of hydrocarbons in peripheral and
satellite basins supports the idea that the Urbin
Group is a good potential hydrocarbon reservoir in
locations not affected by hydrothermal processes.
5. This study in the Cameros Basin may serve to reassess
the possibility of finding potential clastic reservoirs in
intracratonic rift basins, despite the likely presence of
locally destructive effects on reservoir quality pro-
duced by clastic diagenesis and low-grade hydrother-
mal metamorphism.
Acknowledgements
Founding for this research was provided by the Spanish
DIGICIT projects BTE2001-026 and CGL2005-07445-
C03-02/BTE. This manuscript has been improved by
comments and suggestions from G.J. Weltje and four
anonymous referees.
171
Appendix A. Supplementary data
Supplementary data associated with this article can
be found, in the online version, at doi:10.1016/j.
sedgeo.2007.05.017.
References
Alonso-Azcrate, J., Barrenechea, J.F., Rodas, M., Mas, J.R., 1995.
Comparative study of the transition between very low-grade and
low-grade metamorphism in siliciclastic and carbonate sediments:
early Cretaceous, Cameros Basin (Northern Spain). Clay Minerals
30, 407419.
Alonso-Azcrate, J., Rodas, M., Botrell, S.H., Raiswell, R., Velasco, F.,
Mas, R., 1999a. Pathways and distances of fluid flow during low-
grade metamorphism: evidence from pyrite deposits of the Cameros
Basin, Spain. Journal of Metamorphic Geology 17, 339348.
Alonso-Azcrate, J., Rodas, M., Bottrell, S.H., Mas, J.R., Raiswell, R.,
1999b. Estudio textural e isotpico de los sulfuros diseminados en
los sedimentos de la cuenca de Cameros (La Rioja, Espaa).
Revista de la Sociedad Geolgica de Espaa 12, 241249.
Alonso-Azcrate, J., Rodas, M., Fernndez-Daz, L., Bottrell, S.H.,
Mas, R., Lpez-Andrs, S., 2001. Causes of variation in cristal
morphology in metamorphogenic pyrite deposits of the Cameros
Basin (N of Spain). Geological Journal 36, 159170.
Arribas, J., Mas, R., Alonso, A., Ochoa, M., 2003. Diagenetic
processes controlling the quality of potential clastic reservoirs in a
continental rift basin: western Cameros Basin, Spain. AAPG
International Conference and Exhibition, Barcelona, Spain, (Poster
presentation).
Barclay, S.A., Worden, R.H., 2000. Effects of reservoir wettability on
quartz cementation in oil fields. In: Worden, R.H., Morad (Eds.),
Quartz cementation in sandstone reservoirs. Special Publication
of the International Association of Sedimentologists, vol. 29,
pp. 103117.
Barrenechea, J.F., Rodas, M., Mas, J.R., 1995. Clay mineral variations
associated with diagenesis and low-grade metamorphism of early
Cretaceous sediments in the Cameros Basin, Spain. Clay Minerals
30, 119133.
Barrenechea, J.F., Rodas, M., Frey, M., Alonso-Azcrate, J., Mas, J.R.,
2001. Clay diagenesis and low grade metamorphism of Tithonian
and Berriasian sediments in the Cameros Basin (Spain). Clay
Minerals 36, 325333.
Beard, D.C., Weyl, P.K., 1973. Influence of texture on porosity and
permeability of unconsolidated sand. American Association of
Petroleum Geologists Bulletin 57, 349369.
Bjrlykke, K., 1998. Clay mineral diagenesis in sedimentary basins: a
key to the prediction of rock properties. Examples from the North
Sea Basin. Clay Minerals 33, 1534.
Bjrlykke, K., Egeberg, P.K., 1993. Quartz cementation in sedimentary
basins. American Association of Petroleum Geologists Bulletin 77,
15361548.
Bloch, S., Lander, R.H., Bonnell, L., 2002. Anomalously high porosity
and permeability in deeply buried sandstone reservoirs: origin and
predictability. American Association of Petroleum Geologists
Bulletin 86, 301328.
Buick, R., Rasmussen, B., Krapez, B., 1998. Archean oil: evidence for
extensive hydrocarbon generation and migration 2.5-3.5 Ga.
American Association of Petroleum Geologists Bulletin 82, 5069.
Casquet, C., Galindo, C., Gonzlez-Casado, J.M., Alonso, A., Mas, J.R.,
Rodas, M., Garca, E., Barrenechea, J.F., 1992. El metamorfismo en
172 M. Ochoa et al. / Sedimentary Geology 202 (2007) 158173
la cuenca de los Cameros. Geocronologa e implicaciones tectnicas.
Geogaceta 11, 2225.
Chayes, F., 1952. Notes of the staining of potash feldspar with sodium
cobaltinitrite in thin section. The American Mineralogy 37, 337340.
Chow, N., Morad, S., Al-Aasm, I.S., 1996. Origin of authigenic
carbonates in Eocene to Quaternary sediments from the Arctic and
Norwegian-Greenland Sea. In: Myhre, A., Thiede, J., Firth, J.,
Ruddiman, W.F., Johnsson, L. (Eds.), Proceedings of the Ocean
Drilling Program, Scientific Results, vol. 151, pp. 415434.
Chuhan, F.A., Bjorlykke, K., Lowrey, C., 2000. The role of
provenance in illitization of deeply buried reservoir sandstones
from Haltenbanken and North Viking Graben, offshore Norway.
Marine and Petroleum Geology 17, 673689.
Chuhan, F.A., Bjorlykke, K., Lowrey, C.J., 2001. Closed-system burial
diagenesis in reservoir sandstones; examples from the Garn
Formation at Haltenbanken area, offshore Mid-Norway. Journal
of Sedimentary Research 71, 1526.
De Ros, L.F., Sgarbi, G.N., Morad, S., 1994. Multiple authigenesis of
K-feldspar in sandstones; evidence from the Cretaceous Areado
Formation, Sao Francisco Basin, central Brazil. Journal of Sedimen-
tary Research 64, 778787.
De Ros, L.F., Morad, S., Broman, C., De Csereo, P., Gomez-Gras, D.,
2000. Influence of uplift and magmatism on distribution of quartz
and illite cementation: evidence from Siluro-Devonian sandstones
of the Paran Basin, Brazil. Special Publication of the International
Association of Sedimentologists 29, 231252.
Dickinson, W.R., 1985. Provenance relations from detrital modes of
sandstones. In: Zuffa, G.G. (Ed.), Provenance of arenites. NATO
ASI Series, vol. C-148, pp. 333362.
Dutkiewicz, A., Rasmussen, B., Buick, R., 1995. Oil preserved in fluid
inclusions in Archaean sandstones. Nature 395, 885888.
Fournier, R.L., 1985. The behaviour of silica in hydrothermal
solutions. In: Berger, B.R., Bethke, P.M. (Eds.), Geology and
Geochemistry of Epithermal Systems: Society of Economic
Geologists. Reviews in Economic Geology, vol. 2, pp. 4561.
Goldstein, R.H., Reynolds, T.J., 1994. Systematics of fluid inclusions
in diagenetic minerals. SEPM Short Course, vol. 31. SEPM, Tulsa.
199 pp.
Goldstein, R.H., Rossi, C., 2002. Recrystallization in quartz over-
growths. Journal of Sedimentary Research 72, 432440.
Guimer, J., Alonso, A., Mas, R., 1995. Inversion of an extensional ramp
basin by a newly formed thrust: the Cameros basin (N Spain). In:
Buchanan, J.G., Buchanan, P.G. (Eds.), Basin Inversion. Geological
Society of London Special Publication, vol. 88, pp. 433453.
Hoffman, J., Hower, J., 1979. Clay mineral assemblages as low grade
metamorphic geothermometers: Application to the thrust faulted
disturbed belt of Montana, U.S.A. SEPM Special Publication 26,
5579.
Houseknecht, D.W., 1987. Assessing the relative importance of
compaction processes and cementation to reduction of porosity in
sandstones. American Association of Petroleum Geologists Bulletin
71, 633642.
Ingersoll, R.W., Bullard, T.F., Ford, R.L., Grimm, J.P., Pickle, J.D.,
Sres, S.W., 1984. The effect of grain size on detrital modes: a test of
the Gazzi-Dickinson point counting method. Journal of Sedimen-
tary Petrology 54, 103116.
Kastner, M., Siever, R., 1979. Low temperature feldspars in sedimentary
rocks. American Science 279, 435479.
Ketzer, J.M., Morad, S., Amorosi, A., 2003. Predictive clay cementation
in a sequence stratigraphy framework. In: Worden, R.H., Morad, S.
(Eds.), Clay Cementation in sandstones. International Association of
Sedimentologists, Special Publication, vol. 34, pp. 4259.
Kraishan, G.M., Rezaee, M.R., Worden, R.H., 2000. Significance of
trace element composition of quartz cement as a key to reveal the
origin of silica in sandstones: an example from the Cretaceous of the
Barrow Sub-basin, Western Australia. In: Worden, R.H., Morad, S.
(Eds.), Quartz Cementation in sandstones. International Association
of Sedimentologists, Special Publication, vol. 29, pp. 317331.
Lanson, B., Beaufort, D., Berger, G., Bauer, A., Cassagnabre, A.,
Meunier, A., 2002. Authigenic kaolin and illite minerals during
burial diagenesis of sandstones: a review. Clay Minerals 37, 122.
Larese, R.E., 1997. Impact of diagenetic processes on sandstone
reservoir quality; controls, effects, and predictive evaluation using
data from natural and experimental systems. American Association
of Petroleum Geologists Bulletin 81, 1955.
Lindholm, R.C., Finkelman, R.B., 1972. Calcite staining; semiquan-
titative determination of ferrous iron. Journal of Sedimentary
Petrology 42, 239245.
Lundegard, P.D., 1992. Sandstone porosity loss a big picture
view of the importance of compaction. Journal of Sedimentary
Petrology 62, 250260.
Lynch, F.L., Land, L.S., 1996. Diagenesis of calcite cement in Frio
Formation sandstones and its relationship to formation water
chemistry. Journal of Sedimentary Research 66, 439446.
Mantilla, L.C., Casquet, C., Mas, R., 1998. Los paleofluidos en el
Grupo Oncala, Cuenca de Cameros (La Rioja, Espaa): datos de
inclusiones fluidas, istopos de oxgeno y SEM. Geogaceta 24,
207210.
Mantilla, L.C., Casquet, C., Galindo, C., Mas, R., 2002. El
metamorfismo hidrotermal cretcico y palegeno en la cuenca de
Cameros (Cordillera Ibrica, Espaa). Revista Zuba, Monogrfico
14, 143154.
Mas, R., Alonso, A., Guimer, J., 1993. Evolucin tectonosedimen-
taria de una cuenca extensional intraplaca: la cuenca finijursica-
eocretcica de Los Cameros (La RiojaSoria). Revista de la
Sociedad Geologica de Espaa 6 (34), 129144.
Mas, R., Benito, M.I., Arribas, J., Serrano, A., Guimer, J., Alonso, A.,
Alonso-Azcrate, J., 2002. La cuenca de Cameros: desde la extensin
finijursica-eocretcica a la inversin terciaria -implicaciones en
la exploracin de hidrocarburos-. Zuba Monogrfico 14, 964.
Logroo.
Mas, R., Benito, M.I., Arribas, J., Serrano, A., Guimer, J., Alonso, A.,
Alonso-Azcrate, J., 2003, The Cameros Basin: From Late Jurassic
Early Cretaceous Extensin to Tertiary Contractional Inversin
Implications of Hydrocarbon Exploration. Northwest Iberian Chain,
North Spain. Geol. Field Trip, 11, AAPG International Conference
and Exhibition. Barcelona, Ed. Cent. Recherches. Elf-Total-Fina.
56p.
Mata, M.P., Casas, A.M., Canals, A., Gil, A., Pocov, A., 2001. Termal
history during Mesozoic extendion and Tertiary uplift in the
Cameros Basin, northern Spain. Basin Research 13, 91111.
Milliken, K.L., 1998. Carbonate diagenesis in non-marine foreland
sandstones at the western edge of the Alleghanian overthrust belt,
southern Appalahians. In: Morad, S. (Ed.), Carbonate Cementation in
Sandstones. International Association of Sedimentologists Special
Publication, vol. 26, pp. 87105.
Montigny, R., Azambre, B., Rossy, M., Thuizat, R., 1986. K-Ar study
of Cretaceous magmatism and metamorphism in the Pyrenees; age
and length of rotation of the Iberian Peninsula. Tectonophysics
129, 257273.
Morad, S., Bergan, M., Knarud, R., Nystuen, J.P., 1990. Albitization of
detrital plagioclase in Triassic reservoir sandstones from the Snorre
Field, Norwegian North Sea. Journal of Sedimentary Petrology 60,
411425.
 M. Ochoa et al. / Sedimentary Geology 202 (2007) 158173
Morad, S., Ben Ismail, H., De Ros, L.F., Al-Aasm, I.S., Serrihini, N.E.,
1994. Diagenesis and formation water chemistry of Triassic
reservoir sandstones from southern Tunisia. Sedimentology 41,
12531272.
Morad, S., Al-Aasm, I.S., Longstaffe, F.J., Marfil, R., De Ros, L.F.,
Johansen, H., Marzo, M., 1995. Diagenesis of a mixed siliciclastic/
evaporitic sequence of the middle Muschelkalk (Middle Triassic),
the Catalan Coastal Range, NE Spain. Sedimentology 42, 749768.
Morad, S., Ketzer, J.M., De Ros, F., 2000. Spatial and temporal
distribution of diagenetic alterations in siliciclastic rocks: implica-
tions for mass transfer in sedimentary basins. Sedimentology 47,
95120.
Mozley, P.S., Wersin, P., 1992. Isotopic composition of siderite as an
indicator of depositional environment. Geology 20, 817820.
Ochoa, M., Arribas, J., Mas, R., 2004. Changes in sandstone
composition during Lower Cretaceous synrift fluvial sedimenta-
tion (Cameros Basin, Spain). 32nd International Geological
Congress, Florence (Italy). Abstract CD, Session 24234.
Oelkers, E.H., Bjrkum, P.A., Murphy, W.M., 1996. A petrographic
and computational investigation of quartz cementation and porosity
reduction in North Sea sandstones. American Journal of Science
296, 420452.
Olivet, J.L., Bonnin, J., Beuzart, P., Auzende, J.M., 1984. Cinmatique
de l'Atlantique nord et central. CNEXO. Rapport Scientifique et
Technique 54, 1108.
Pittman, J.K., Splt, Ch., 1996. Origin and timing of carbonate
cements in the St. Peter Sandstone, Illinois Basin: evidence for a
genetic link to Mississippi Valley-type mineralization. SEPM
Special Publication 55, 187203.
Rat, P., 1982. Factores condicionantes en el Cretcico de Espaa.
Cuadernos de Geologa Ibrica 8, 10591076.
Reed, J.S., Eriksson, K.A., Kowalewski, M., 2005. Climatic,
depositional and burial controls on diagenesis of Appalachian
Carboniferous sandstones: qualitative and quantitative methods.
Sedimentary Geology 176, 225246.
Rowe, J.E., Burley, S.D., 1997. Fault-related diagenetic cementation in
Triassic Sandstones at Alderley Edge, north-eastern Cheshire. In:
Meadows, N., Cowan, G. (Eds.), Petroleum Geology of the Irish
Sea and its Margins. Geological Society of London, Special
Publication, vol. 124, pp. 325352.
Saigal, G.C., Morad, S., Bjrlykke, K., Egeberg, P.K., Aagaard, P.,
1988. Diagenetic albitization of detrital K-feldspars in Jurassic,
173
Lower Cretaceous, and Tertiary clastic reservoir rocks from
offshore Norway, I. Textures and origin. Journal of Sedimentary
Petrology 58, 10031013.
Salas, R., Guimer, J., Mas, J.R., Martn-Closas, C., Melndez, A.,
Alonso, A., 2001. Evolution of the Mesozoic Central Iberian Rift
System and its Cenozoic inversin (Iberian Chain). In: Cavazza,
W., Robertson, A., Ziegler, P. (Eds.), Peri-Tethyan Rift/Wrench
Basins and Passive Margins. Mmoires du Musum National
d'Histoire Naturelle, vol. 186, pp. 145185.
Salinas, F.J., Mas, J.R., 1989. Individualizacin de la cubeta lacustre
de Cervera del Ro Alhama (la Rioja) durante la sedimentacin del
Grupo Urbin (Cretcico inferior). Comunicaciones, XII, Con-
greso Espaol de Sedimentologa, Bilbao, pp. 7982.
Sptl, C., Houseknecht, D.W., Longstaffe, F.J., 1994. Authigenic
chlorites in sandstones as indicators of high-temperature diagen-
esis, Arkoma foreland basin, USA. Journal of Sedimentary
Research 64, 553566.
Sptl, C., Houseknecht, D.W., Riciputi, L.R., 2000. High-temperature
quartz cement and the role of stylolites in a deep gas reservoir,
Spiro Sandstone, Arkoma Basin, USA. International Association
of Sedimentologists Special Publication 29, 281297.
Tissot, B.P., Welte, D.H., 1978. Petroleum formation and occurrence. A
New Approach to Oil and Gas Exploration. Springer, Berlin. 538 pp.
Tucker, M., 1988. Techniques in Sedimentology. Blackwell Scientific
Publications. 394 pp.
Walderhaug, O., Bjorkum, P.A., Nadeau, P.H., Langnes, O., 2001.
Quantitative modelling of basin subsidence caused by temperature-
driven silica dissolution and reprecipitation. Petroleum Geoscience
7, 107113.
Wilson, M.D., 1994. Reservoir quality assessment and prediction in
clastic rocks. SEPM Short Course 30. 432 pp.
Worden, R.H., Morad, S. (Eds.), 2000. Quartz cementation in
sandstone reservoirs. Special Publication of the International
Association of Sedimentologists, vol. 29, pp. 103117.
Worden, R.H., Morad, S., 2003. Clay mineral cements in sandstones.
International Association of Sedimentologists Special Publication
34. 509 pp.
Worden, R.H., Oxtoby, N.H., Smalley, P.C., 1998. Can oil emplace-
ment stop quartz cementation in sandstones? Petroleum Geosci-
ence 4, 129137.
Zuffa, G.G., 1980. Hybrid arenites: their composition and classifica-
tion. Journal of Sedimentary Petrology 50, 2129.
 Available online at www.sciencedirect.com

Sedimentary Geology 202 (2007) 174 192

www.elsevier.com/locate/sedgeo

Fluvial responses to tectonics and climate change during the Late

Weichselian in the eastern part of the Pannonian Basin (Hungary)
Annamria Ndor , Edit Tham-Bozs, rpd Magyari, Edit Babinszki
Geological Institute of Hungary, Stefnia ut 14, H-1143 Budapest, Hungary

Abstract

Fine-grained sandysilty channelbelt and floodplain deposits of the BerettyKrs Rivers, a main eastern transverse
tributary system of the modern Tisza River in the eastern part of the Pannonian Basin, were deposited during the Late Pleistocene
under net subsiding conditions. The palaeo-drainage network pattern of a 2500 km2 large part of the alluvial plain was
reconstructed based on interpretation of airborne photographs and analysis of 18th century topographic maps, which show the
natural river patterns that predate the introduction of river regulation schemes. The investigation showed that a large meandering
river system, with two main channel belts surrounding a floodbasin, entered the alluvial plain from the northeast, and a braided
river entered the alluvial plain from the southeast. Detailed sedimentary logs of seven continuous corings and several sand and
clay-pit sections were used to characterize different alluvial units. Optical luminescence dating (OSL) of 25 quartz samples and
four 14C datings showed that the sediments are of Late Pleniglacial to Late Glacial age. Transport directions inferred from heavy
mineral analyses combined with the OSL ages strongly suggest that the large meandering system represents the palaeo-Tisza River,
which supposedly flowed along the northeastsouthwest striking rmellk depression during the Late Pleniglacial. The braided
river can be regarded as a precursor to the Fekete and FehrKrs Rivers, which entered the alluvial plain from the southeast
during the Late Glacial. The interpretation of seismic profiles, field measurements of neotectonic activity, and variations in
thickness of sediments along the studied profile revealed that river development was largely controlled by subsidence along the
rmellk depression until 14 to 16 ky, and by uplift of the southeastern part of the catchment area. The studied fluvial successions
also document the response of the palaeo-Tisza and Krs system to the climate changes of the Weichselian Late PleniglacialLate
Glacial period. Much of the sand from the meandering zones was deposited during the BllingAllerd and SgvrLascaux
interstadials, whereas some dated sand units from the braided zone represent the Older and Younger Dryas. The error ranges of
OSL dates, which often exceed the duration of Weichselian substages and subdivisions, prevented an unambiguous correlation of
the studied sections with the millennial-scale climate changes of the last 25 ky. Meandering and braided river activity coexisted
under different climate conditions, whereas locations of the main channel belts are related to subsidence anomalies. The results of
our study thus clearly indicate that tectonics was the primary control on river development.
2007 Elsevier B.V. All rights reserved.

Keywords: BerettyKrs rivers; Sedimentology; Stratigraphy; OSL dating

1. Introduction

In the past decade a large number of studies,

especially on northwestern European rivers, have
Corresponding author. revealed that climate variations, tectonics and sea-level
E-mail address: nador@mafi.hu (A. Ndor). changes exerted a major control on river patterns and
0037-0738/$ - see front matter 2007 Elsevier B.V. All rights reserved.
doi:10.1016/j.sedgeo.2007.03.001
 A. Ndor et al. / Sedimentary Geology 202 (2007) 174192 175

thus depositional styles of fluvial sediments over
glacialinterglacial cycles. Most of the models discuss-
ing the role of climate control on river development are
based on observations in uplifted areas, in which
variations in fluvial discharge and sediment flux are
recorded in the form of riverterrace series (e.g.
Bridgland, 2000, Rhine: Boenigk, 2002, Thames:
Gibbard, 1985; Maddy and Bridgland, 2000, Meuse:
Vandenberghe, 1993; Van den Berg, 1994, Seine and
Somme: Antoine et al., 2000). However, totally different
processes can be envisaged for a subsiding setting, in
which topographic differences and stream gradients are
too low for river incision, and subsidence is balanced by
fluvial aggradation. One of the best studied areas for this
latter type is the RhineMeuse system (e.g. Kasse et al.,
1995; Trnquist, 1995; Berendsen and Stouthamer,
2000; Wallinga et al., 2004; Busschers et al., 2005 and
many references therein), whose development was
dominated by North Atlantic climate changes (Johnsen
et al., 1992), sea-level changes (Trnquist et al., 2000;
Waelbroeck et al., 2002) and invasions and retreats of
the Fennoscandian ice sheet (Ehlers and Gibbard, 2004).
This paper provides a case study from a subsiding
setting in the eastern part of the Pannonian Basin (Krs
basin) (Fig. 1), in which sedimentation was continuous
under net subsiding conditions during the Pleistocene
(Ndor et al., 2003). Controls on river development
were quite different from northwest Europe, and thus
can complement our understanding of the evolution of
fluvial systems during Late Pleniglacial and Late Glacial
times. The Pannonian Basin, surrounded by the Alps,
Carpathians and Dinarides is Europe's largest intra-
montane basin (Fig. 1). The Danube River is the major
outflow and is connected to the Black Sea only through
a narrow opening in Romania, some 1200 km down-
stream. Hence, sea-level changes can be excluded as an
allogenic control on fluvial development. However, the
complicated tectonic evolution of the Pannonian Basin,
especially its latest compressional phase with differen-
tial uplift and subsidence rates, exerted a strong control
on river patterns, which will be discussed in detail in this
paper. The extent to which fluvial evolution was
governed by climate is a more complex question. The
Pannonian Basin is a transitional area where three large

Fig. 1. Regional geotectonic setting of the Pannonian Basin within the Carpathian arc (after Bada and Horvth, 2001). Neotectonic basin inversion led
to the formation of several actively subsiding basins and uplift of the basin margin areas. VB = Vienna basin, DaB = Danube basin, DrB = Drava
basin, SB = Sava basin, GP = Great Plain. The DvavnyaSzarvas fault zone had a major control on river development in the Krs basin during the
Late Pleistocene, discussed in detail in the paper.
176 A. Ndor et al. / Sedimentary Geology 202 (2007) 174192

climate zones interact: oceanic, continental and also
sub-Mediterranean (Bacs, 1959), and the interaction of
these zones also played a crucial role during the
Quaternary.
The aim of this paper is to outline the Late
Pleistocene evolution of the BerettyKrs transverse
fluvial system, related to the avulsion history of the
Tisza River, which is the main modern axial river of the
Hungarian Great Plain. The BerettyKrs system is
situated at the eastern transition between the Carpathian
orogen and the Pannonian back-arc domain (Figs. 1, 2).
The rivers originate in the Apuseni Mountains, which
are currently undergoing uplift, and are now flowing
across one of the deepest sub-basins (Krs basin) of the
subsiding Great Plain. The modern Beretty, a fairly
small river, flows along the southwestern part of the
rmellk depression (Fig. 2), which was presumably
occupied by the much larger palaeo-Tisza River during
the Late Pleniglacial (Gbris, 1998, 2002). The main
objectives of the study were to reconstruct ancient
channel geometries in space and time along the lower
reach of the BerettyKrs system during the Late
Pleistocene by imaging the sedimentary architecture of
Late Pleistocene to recent fluvial sequences. An
essential point was the construction of a detailed
chronological framework by the use of optical dating
(Optically Stimulated Luminescence, OSL, Tham-
Bozs et al., 2007). Through this, reconstruction of
the palaeo-environments and the structural and climatic
control on river development was attempted.
2. Geological setting
The Pannonian Basin formed by back-arc style
rifting during the Early and Middle Miocene, coeval
with thrusting of the Carpathian belt (Royden and
Horvth, 1988) (Fig. 1). From the Late Miocene
onward, a brackish to freshwater lake, the so-called

Fig. 2. Index map of the Great Plain and surrounding orogens showing the topography and geographical names used in the text. Dotted areas represent
major Pleistocene alluvial fans: D alluvial fan of the palaeo-Danube, M alluvial fan of the palaeo-Maros, Ny Nyrsg alluvial fan, deposited
by rivers draining the Northeast Carpathians and the northern part of Transylvania. Smaller alluvial fans also developed in front of the North
Midmountains and along the western foothills of the Apuseni Mountains, deposited by the ancestral Krs Rivers.
 A. Ndor et al. / Sedimentary Geology 202 (2007) 174192
Lake Pannon occupied its depression, which was filled
by fluvial-dominated delta systems that prograded from
the northwest and northeast towards the basin interior
(Brczi and Phillips, 1985; Pogcss et al., 1988). From
the Late Miocene onward, the sediment accumulation
rate significantly exceeded the rate of thermal subsi-
dence, which led to the infilling of the entire basin and to
the formation of a paludal to terrestrial environment by
the Pliocene (Brczi and Phillips, 1985; Jmbor, 1989;
Juhsz, 1994; Magyar et al., 1999). Several compres-
sional events occurred since the early Late Miocene and
compression became continuous during the last 2 Ma
(Horvth and Cloetingh, 1996). This resulted in the
complete termination of extension throughout the whole
Pannonian Basin, which caused uplift of large areas
including the North Midmountains and almost the entire
Transdanubia (area west of the Danube). The uplifted
regions surrounded major plains, which were still
subsiding and became areas of continuous fluvial
sedimentation during the Quaternary (Great Plain,
Danube basin, Drava- and Sava basins, Vienna basin)
(Fig. 1). The spatial distribution of vertical movements
reflects lithospheric folding with a large wavelength
(Horvth and Cloetingh, 1996; Cloetingh et al., 1999;
Bada and Horvth, 2001).
The Krs basin with an average height of 84 to 85 m
above sea level represents one of the lowest regions in
the eastern part of the Pannonian Basin. It is a classic
example of a continuously subsiding sub-basin on the
Great Plain, in which late-stage subsidence has been
triggered by compressional basin inversion. The main
modern rivers of the area are the Beretty, the Sebes-,
Fekete- and FehrKrs Rivers, which display an
approximately eastwest oriented transverse pattern.
These rivers drain the Apuseni Mountains in the east,
and transport sediments to the Pannonian Basin. After
their confluence in the western part of the Krs basin,
they continue as the HrmasKrs River towards the
low gradient, southerly directed, axial Tisza River in the
central part of the Pannonian Basin (Fig. 2). The lengths
of these four transverse rivers are quite similar.
However, there are differences in the size, geology
and morphology of their catchment areas in the Apuseni
Mountains in Romania, which result in characteristic
differences in stream gradients (Fig. 3). Generally the
steep (10 to 25 m/km) stream gradients of the rivers
coming from the mountain belt are abruptly reduced at
the basin margin, close to the Hungarian border, and
the rivers continue with a 5 to 10 cm/km gradient
towards the axial Tisza River. Of these four rivers, the
Fekete- and FehrKrs coming from the southeast
have the highest gradients in their mountain sections
177
(Fig. 3). Today, all of these rivers carry mostly sus-
pended load material in the Hungarian parts of their
drainage areas.
3. Drainage pattern evolution of the Great Plain
during the Quaternary
Fluvial sediments transported by the Danube and
Tisza Rivers and their tributaries filled the Great Plain
during the Quaternary. Rivers discharging from the
uplifting marginal areas built large alluvial fans (Fig. 2)
and were drawn towards the subsiding depressions.
Changes in subsidence rates in space and time
throughout the Quaternary resulted in the evolution of
a complex drainage pattern, which has been recently
reviewed by Gbris and Ndor (in press). During Early
and Middle Pleistocene, the rapidly subsiding central
part of the Great Plain served as a local base level
(Fig. 4A) which forced the palaeo-Danube from
northwest to southeast, and contributed to the construc-
tion of a huge alluvial fan (Fig. 2). Another major
alluvial fan developed in the Nyrsg area (Fig. 2). It was
fed by the rivers which drained the Northeast
Carpathians and the northern part of Transylvania. The
palaeo-Tisza River coming from the Northeast Car-
pathians and the Apuseni Mountains flowed along the
southeastern margin of the Nyrsg alluvial fan
(Fig. 4A). During the Early Weichselian it shifted
further to the southeast and formed a 1015 km wide
erosional plain along the rmellk which was a
northeastsouthwest striking depression between the
Nyrsg and the alluvial fans of the palaeo-Szamos and
-Kraszna Rivers draining the Northeast Carpathians. A
third large alluvial fan was built in the southern part of
the Great Plain by the palaeo-Maros River draining the
East Carpathians and Apuseni Mountains (Borsy, 1990)
(Fig. 2).
The Krs basin, located between these large alluvial
fans, was a continuously subsiding area. The ancestral
Beretty and Krs Rivers coming from the east, as well
as the palaeo-Tisza from the northeast, flowed into this
basin (Figs. 2, 4A). This resulted in the deposition of a
nearly 500 m thick cyclic fluvial record spanning the
entire Pleistocene (Ndor et al., 2003).
During Late Pleniglacial and Late Glacial times, a
third major river (palaeo-Bodrog) could have appeared
(Fig. 4B) in addition to the palaeo-Danube and palaeo-
Tisza, which flowed diagonally towards the central
depression of the Great Plain (Gbris, 1998, 2002) The
appearance of this third main river could have been
initiated by renewed subsidence in the foreland of the
North Midmountains.
178 A. Ndor et al. / Sedimentary Geology 202 (2007) 174192

Fig. 3. Hydrological characteristics and stream gradients of the rivers of the Krs basin (Tham-Bozs et al., 2002).

At the end of the Pleistocene, subsidence of small
basins at the margins of the Great Plain caused a major
re-arrangement of the drainage network and gave rise
to the present-day hydrography. As a result of
subsidence along the southern reach of the present
Danube (Jask and Krolopp, 1991) and the decreasing
rate of subsidence of the central part of the Great Plain,
the palaeo-Danube shifted westwards from the axis of
its previous alluvial fan during the Late Pleistocene to
occupy its present course (Fig. 4B). This is shown by
the absence of terraces older than the Late Weichselian
(No II/a) in the present Danube valley (Pcsi, 1959).
As a result of a combined effect of subsidence of the
Bodrogkz basin on the northern margin of the Great
Plain and the uplift of the Nyrsg area (Fig. 2), the
Tisza moved away from the rmellk depression
during the Late Glacial (Gbris, 1998, 2002). After
entering the Great Plain, the Tisza first flowed to the
northwest towards the Bodrogkz basin, and abruptly
turned to the southwest to occupy the position of the
former palaeo-Bodrog (Fig. 4B).
4. Research methods
The drainage network and channel-pattern evolution
of the BerettyKrs system were reconstructed by
integrating various methods. Information on the stratig-
raphy of the area was based on the 1:100 000 scale
geological map series of the Geological Institute of
Hungary. The most important alluvial morphological
features (various channel patterns, marshes, flood-
basins) have been reconstructed and edited on a digital
 A. Ndor et al. / Sedimentary Geology 202 (2007) 174192 179

Fig. 4. A) Hypothetical river network during the Early- and Middle Pleistocene (after Gbris and Ndor, in press). The rapidly subsiding central part
of the Great Plain forced the palaeo-Danube from the northwest and the palaeo-Tisza from the northeast. B) Three-axis palaeo-drainage system of the
Great Plain during the Late Pleistocene (after Gbris and Ndor, in press). The appearance of the third axis resulted from rejuvenated subsidence in
front of the North Midmountains.

map, based on detailed analyses of airborne photographs belts. This was compared to the natural river network
at 1:60 000 scale, covering about 2500 km2. Special patterns, which existed until river regulation was
attention was paid to the analysis of intersecting channel introduced in the mid 19th century, and which were
180 A. Ndor et al. / Sedimentary Geology 202 (2007) 174192
reconstructed on the basis of topographic maps dated
17821785. According to these old maps drawn prior to
river regulation, most of the area was submerged, and
characterized by a swampy environment, in which rivers
frequently changed their courses through avulsion.
In order to acquire a better understanding of the
origin of the different planform geometries of rivers on
the alluvial plain, seven new continuous cores were
drilled into different types of palaeo-channels by an
Eijkelkamp percussion borer for detailed lithological,
sedimentological and geochronological (OSL, 14 C)
analyses. Undisturbed sediments were obtained in
dark, non-transparent plastic tubes of 1 m length and
5 cm diameter. Macroscopic description of the cores
included grain size, colour, organic and non-organic
admixtures and sedimentary structures. The seven cores
were correlated along a northsouth profile. Several
clay and sand pits were logged to complement the
borehole information.
Optical dating (OSL) of sand-sized quartz from 25
samples was carried out to obtain chronostratigraphic
information. Samples for OSL were taken by cutting
out 20 cm long sections of sediment in plastic tubes to
avoid exposure to light. The samples were sieved to
isolate the 80 to 200 m grain-size fraction. Sand was
then treated with hydrogen peroxide (30%) to remove
organic material, and hydrochloric acid (10%) to
dissolve carbonates. Density separation of the quartz-
rich fraction was carried out by an aqueous solution of
sodium polytungstate. This fraction was etched with
hydrofluoric acid (40%) for 60 min to remove any
remaining feldspar grains, as well as the outer ca.
10 m layer of the quartz grains. The etching is
intended to remove any quartz that may have absorbed
a dose from alpha radiation from the surrounding
sediments. OSL measurements were made using a
RIS TL/OSL automatic reader with a calibrated
90
Sr/ 90 Y beta source. The SAR (Single-Aliquot
Regenerative dose) protocol (Murray and Wintle,
2000, 2003) was used for equivalent dose determina-
tion. Dose rates were calculated based on the results
of high-resolution gamma spectrometry measure-
ments of the surrounding sediments, performed in
the laboratory of the Etvs Lornd Geophysical
Institute, Hungary. Tham-Bozs et al. (2007) dis-
cussed the measurement procedure and optical dating
results in detail. To cross-check OSL dates, four 14 C
ages were obtained from charcoal-rich layers of
floodplain deposits in selected borehole sections.
Radiocarbon analyses were performed by the Beta
Analytic Radiocarbon Dating Laboratory, Miami,
Florida.
Heavy-mineral composition and inferred transport
directions of the fluvial sediments have also been
integrated into our reconstruction. The same samples
that were dated by OSL were also analysed for heavy
mineral composition, in order to couple age and
provenance information. Heavy minerals were separated
from the OSL samples after sieving and before other
treatments, and were separated by bromoform using a
centrifuge. The heavy mineral grains were identified
under a polarising microscope and quantified by point
counting. The heavy mineral composition of the
samples was compared to the composition of modern
river sediments of the Great Plain by means of cluster
analysis. A hierarchical classification was applied to the
detrital heavy-mineral suite, based on Euclidean dis-
tance and the weight-paired average method (Davis,
1986; .Kovcs, 1986). Samples within a particular
cluster have similar detrital heavy mineral compositions,
which suggests they originated from the same source
area. We presume that in this region the Late Pleistocene
physiography and geology of the potential source areas
were comparable to the present-day situation. Therefore
we argue that transport paths of modern rivers,
originating from known source areas and characterized
by similar heavy mineral compositions can be used to
draw inference from older sediments. This method has
been successfully used to trace provenance and
Pleistocene palaeo-transport directions of fluvial sedi-
ments in the Krs basin (Tham-Bozs et al., 2002).
Tectonic reinterpretation of more than one hundred
industrial seismic profiles (Dudko, 2002) contributed to
the evaluation of structural control on river develop-
ment. Although these seismic profiles do not resolve the
uppermost ca. 400 m of the Pleistocene sedimentary
units, the structural character and spatial distribution of
the main fault zones which cut through the Mio
Pliocene strata permit reconstruction of the tectonic
evolution of the area.
5. Reconstruction of palaeo-drainage patterns
5.1. Results of airborne photographic interpretation
The palaeo-morphology of the studied alluvial plain
drawn on the basis of airborne photographic interpre-
tation shows that several major zones can be distin-
guished on the basis of characteristic planform
geometries (Fig. 5A). In the north, between the modern
Hortobgy and Beretty Rivers, there is a zone with
remnants of large meanders (km-size radius). The
orientations of the meander loops and their scroll bars
indicate that palaeo-flow was mainly from the northeast.
 A. Ndor et al. / Sedimentary Geology 202 (2007) 174192 181

Fig. 5. A) Palaeo-channel remnants of the Krs basin based on airborne photographic interpretation with the location of the studied boreholes and
outcrops. Remnants of meanders with different sizes and braided planform geometries were identified. The DvavnyaSzarvas sinistral strikeslip
zone exerted a strong neotectonic control on river development. For discussion see the text. B) Main zones of different palaeo-channel remnants in the
Krs basin based on the interpretation of historical maps, showing natural river patterns before river regulation. It clearly outlined two major
meandering rivers enclosing a swampy area, and remnants of a braided river on the south.
182 A. Ndor et al. / Sedimentary Geology 202 (2007) 174192

South of this zone of large meanders in the central part
of the study area, there is an area characterized by a
blurred pattern of small meanders of a few tens to
hundred meters amplitude, north of the modern
HrmasKrs Rivers. A northeastsouthwest oriented
sharp line marks the contact between these two zones
(Fig. 5A). The spatial distribution of different planforms
and their intersections shows that medium-size meander
remnants of a few hundred metres amplitude are also
found along the southern boundary of the small meander
zone, south of the present HrmasKrs River
(Fig. 5A). All meander remnants are found at a
topographic height of 8485 masl, irrespective of their
size. In the southern part of the studied area there is a
zone with a braided pattern (Fig. 5A) at 8790 m asl.
Based on the orientation of the planforms, palaeo-flow
was from the southeast. The ancient braided channel
remnants are sub-parallel with the modern Ketts
Krs River (confluence of the Fekete and FehrKrs
Rivers).
5.2. Interpretation of historical maps
Three main channel generations were distinguished
based on the comparison of historic and present-day
topographic maps: (1) presently active river channels
which coincide with natural channels and artificial
canals (mostly cut-offs of meander loops) which were
constructed during river regulation; (2) channels active
in the mid 18th century, which represent the natural river
pattern before regulation; (3) channel remnants which
were no longer active in the mid 18th century: these are
fossil channel remnants which can be also seen on
airborne photographs.
The active and inactive channels of the natural river
pattern network clearly outline remnants of two major
meandering river systems (roughly along the modern
Beretty SebesKrs and Hortobgy Rivers) which
flowed from the northeast. They enclosed a large
swampy area into which many smaller tributaries
flowed, and created several small meanders. The area

Fig. 6. Cross section along the boreholes and outcrops. OSL and 14C dates are shown in Table 1. Transport directions are inferred from cluster analysis
of heavy minerals, summarized in Fig. 7. Sedimentary pattern of the sections closely mirror the river pattern of the different zones, shown on Fig. 5.
Upward fining successions with notable differences in their thickness characterize the large- and medium-size meandering zones. Section is the
swampy area consists of clay and silt, while the braided zone logs are characterized by trough cross-stratified sands, lacking upward-fining units.
Interpreted profile is shown on Fig. 8.
 A. Ndor et al. / Sedimentary Geology 202 (2007) 174192 183

further to the south is characterized by braided channel
remnants, which are parallel to each other and lack
natural smaller tributaries/creeks. No swampy environ-
ment could be seen in this area. The distinctive large- to
medium-size meanders (km to few hundred meters
radius) of the two main former channel systems can be
easily traced, while the small meanders in the central
part and the braided channels in the south show a blurred
pattern (Fig. 5B).
6. Sedimentary facies of boreholes and outcrops
Sediments from the zone of large meanders in the
north (Dvs-1 and Tr-1 sections) exhibit sandysilty
upward fining successions 5 to 6 m thick (Fig. 6). Their
lower part consists of fine- to medium-grained, cross-
stratified sand. The 0.1 to 0.5 m thick sets consist of
planar and trough-cross-bedded laminae, and beds
passing upwards into silty units. At least three different
generations of sandy units occur in the shallow
subsurface in the area of the medium-sized meanders
(Gyoma, Gyomaendrd, Gye-4 and Gye-2 sections)
(Fig. 6). Each unit consists of 2 to 3 m thick fining-
upward sequences (Gye-2, Gye-4) with cross- to
horizontal stratified fine-grained sand, capped by silt.
Cross-bedded or massive fine- to medium-grained sand
units crop out in the Gyoma and Gyomaendrd sections.
The Gye-3 borehole was drilled in the swampy area
surrounded by the two main meandering channel
systems (Fig. 5). This section consists almost exclu-
sively of alternating clayey silt and clay in 1 to 2 m thick
cycles (Fig. 6).
The uppermost sediments (upper 3 to 4 m below the
surface) from the braided zone (Hu-1, Kondoros and Ks-1
sections) are mostly fine- to medium-grained, (locally
coarse-grained) trough cross-stratified sands (Fig. 6). The
0.1 to 0.5 m thick sets consist of planar and trough cross-
stratified laminae and beds. Upward fining units cannot
be distinguished in these sections, which may confirm the
braided (or low-sinuosity) origin suggested by the
planform geometry.
7. Dating results
The OSL ages of the fluvial samples from 1.5 to
7.7 m below the modern surface vary between about 10

Table 1
OSL and 14C ages. Location of the samples is shown in Fig. 6
Sample OSL age (ky) Chronostratigraphy Transport directions
Large meanders 1. DVS-1 3.503.70 m 14.1 1.2 Blling NE
2. DVS-1 4.504.70 m 14.2 1.4 Blling E, NE, SE
3. DVS-1 5.505.70 m 14.4 1.1 Blling NE, N
4. DVS-1 6.506.70 m 14.4 1.4 Blling E, NE, SE
5. TR-1 5.505.70 m 16.2 1.7 Older Dryas NE
6. TR-1 6.506.70 m 17.9 2.4 Late Pleniglacial NE
7. TR-1 7.507.70 m 22.8 2.0 Late Pleniglacial E, NE, SE
Medium-size meanders 8. Gyoma 3.00 m 15.9 1.5 Older Dryas NE
9. GYE-4 2.602.80 m 17.9 1.6 Late Pleniglacial NE
10. GYE-4 3.603.80 m 18.0 2.1 Late Pleniglacial NE
11. Gyomaendrd 1.70 m 12.2 1.3 Younger Dryas SE
12. Gyomaendrd 2.50 m 16.0 1.2 Older Dryas NE, N
13. Gyomaendrd 3.20 m 19.0 1.7 Late Pleniglacial NE, N
14. GYE-2 5.605.90 m 25.9 2.3 Late Pleniglacial NE, N
15. GYE-2 6456.65 m 26.5 2.4 Late Pleniglacial NE, N
16. HU-1 1.501.70 m 12.4 1.2 Younger Dryas E, NE, SE
Braided 17. HU-1 2.502.70 m 13.3 1.1 BllingAllered NE
18. Kondoros hb. 3.40 m 11.9 1.1 Younger Dryas SE
19 KS-1 1.501.70 m 10.4 1.2 Preboreal NE, N
20. KS-1 2.502.70 m 13.7 1.1 BllingAllered NE, N
21. KS-1 3.503.70 m 15.4 1.3 Older Dryas SE

14 14
Sample C age BP (ky) Chronostratigraphy C age cal BP (y)
I. TR-1 1.801.85 m 12.95 0.06 Allerad 15 89014 840
II. TR-1 5.90 m 13.18 0.06 Blling 16 26015 430
III. Gyoma 3.80 m 29.17 0.35 Late Pleniglacial
IV. HU-1 6.50 m 27.86 0.30 Late Pleniglacial
184 A. Ndor et al. / Sedimentary Geology 202 (2007) 174192
and 27 ky, representing sediments from the Late
PleniglacialLate Glacial period (Table 1, Fig. 6).
These OSL ages are stratigraphically consistent with
the 14C age data. (Table 1, Fig. 6).
OSL ages of cross-stratified channel sands from the
northern part of the large meander zone cluster around
14 ky (Blling) in the Dvs-1 borehole and 16 to 18 ky
(SgvrLascaux interstadial, cf. Smegi and Krolopp,
2002) in the Tr-1 borehole (Table 1), in which
overlying charcoal-rich layers of floodplain deposits
have 14C ages of 15.89 to 14.84 ky cal BP. The
radiocarbon age of the lower sample in the Tr-1
borehole (number II: 16.26 to 15.43 ky cal BP) closely
matches the OSL ages of over- and underlying samples
(samples 5 and 6 with ages of 16.2 1.7 ky and 17.9
2.4 ky, respectively.) The oldest sample in the Tr-1
borehole, which is located at a depth of 7.6 m, is 22.8 ky
old. It is represented by horizontally stratified sand, and
appears to belong to another unit than the overlying
cross-stratified sand which forms the basal member of
an upward fining cycle.
A 29 ky old charcoal-rich clay located at a depth of
3.8 m is the oldest dated deposit at Gyoma, in the area of
the medium-size meanders. The ages of the different
sand units of upward fining pointbar sequences,
present at depths of 6.6 to 1.6 m below the surface,
range from 26 to 12 ky (Table 1, Fig. 6). Some sands
with similar ages are present at horizons which can be
correlated, e.g. ca. 18 ky old sands at Gyomaendrd and
Gye-4, and 16 ky old sands at Gyoma and Gyomaend-
rd (Fig. 6).
Sands in the area with braided planforms are younger
than sands from the meandering areas. Samples taken at
1.5 to 3.7 m depths are 10 to 15 ky old, and represent
deposition in Preboreal, Younger Dryas, Blling
Allerd interstadial and Older Dryas periods (Table 1).
No braided fluvial activity was previously known from
the early Holocene of Hungary.
Fig. 7. Composition of detrital heavy mineral clusters and sediments of modern rivers from the Great Plain and their transport directions, and studied
samples with similar mineralogical composition. For discussion see the text. cpx: clino-pyroxene, gr: garnet, amh: hornblende, px: pyroxene, m:
magnetite, chl: chlorite, amo: oxy-hornblende.
OSL ages also showed that only 1 to 2 m of fine-
grained sediment, mostly affected by pedogenesis,
accumulated in the entire area during the last 10 ky.
8. Transport directions inferred from heavy mineral
analyses
Heavy mineral analyses showed that amphibole,
pyroxene, garnet, magnetite and chlorite are the main
detrital heavy minerals in the studied samples (Fig. 7).
Practically the same species are found in all samples.
Sediments from the different modern rivers have only
very rare characteristic provenance-indicator minerals.
This is due to the similar source rocks of the catchment
areas (Fig. 1). Quantitative evaluation of the composi-
tional variation of the samples by means of cluster
analysis yielded the dendrogram shown in Fig. 7. It
indicates that the heavy mineral composition of most
samples is similar to that of the recent sediments of the
Tisza River and its northern tributaries, with high
pyroxenehornblendegarnet- (magnetitechlorite)
content, which strongly suggests derivation from the
northeast. Some sands have compositions similar to the
sediments of the modern Beretty and Krs Rivers,
with garnethornblendepyroxene, indicative of supply
from an easterly direction. Only a few samples are
compositionally similar to modern sediments of the
FehrKrs and Maros Rivers, with clinopyroxene
garnethornblende, which suggests derivation from the
southeast. Fig. 6 shows the inferred transport directions
of the OSL-dated samples. These transport directions
represent the flow directions of recent rivers and the
location of their source areas.
9. Neotectonic structures of the area
The analysis of neotectonics in the study area was
based on re-interpretation of seismic profiles by Dudko
 A. Ndor et al. / Sedimentary Geology 202 (2007) 174192
(2002). The sections analysed were acquired by the
Hungarian Oil Company. Seismic interpretation was
performed on paper prints of migrated sections.
Previous seismic interpretations by Pogcss et al.
(1989) led to the identification of a northeastsouthwest
striking PliocenePleistocene sinistral strikeslip zone
north of the area.
Generally, the basement of the study area is
characterized by a northeastsouthwest striking basin-
and-range geometry which gradually shallows towards
the northeast. Lower and Middle Miocene sediments are
typically missing above basement highs. Upper Mio-
cene, Pliocene and Pleistocene strata are mainly
horizontal, but show doming at certain locations.
Three zones consisting of faults which cut through
Upper MiocenePliocene strata were recognized by
Dudko (2002). These fault zones have both extensional
and compressional elements. One of these, the so-called
DvavnyaSzarvas zone, is particularly relevant to
the study area (Fig. 5A). It is a northeastsouthwest
oriented strikeslip zone, the southeastern side of which
is downfaulted and can be followed into the basement,
as confirmed by numerous well logs. Vertical offset of
Upper MiocenePliocene strata increases from north to
south from 70 to 150 m (Szentgyrgyi, 1988; Pstynyi,
1993). The studied sections show mainly negative
flower structures along cross-sections, but positive
flower structures are present at some localities.
10. Discussion
The most important issue to be addressed is the
spatial and temporal link between the zones with
different planforms (meanders with different sizes and
braided). Do they represent different coexisting rivers,
or successive stages in the evolution of one or more river
system(s)? Another important question is to what extent
river development was controlled by tectonics and/or
climate changes.
To a certain extent, the various planform geometries
were known from earlier literature (e.g. Papp, 1960;
Mike, 1975). The large meanders in the north of the
studied area were considered as the latest remnants of
the palaeo-Tisza, which flowed along the rmellk
depression during the Late Weichselian. The small
meanders in the central zone have been attributed to
ancestors of the Krs River (Papp, 1960), which
flowed across a flat marshy area. Different interpreta-
tions of the braided planforms have been proposed in
earlier literature. Mike (1975) outlined some linear
patterns for this area, but he called them floodmarks.
Somogyi (1961) and Borsy (1990) related them to the
185
alluvial fan of the Maros River, south of the studied
area.
The arrangement of the large meanders reconstructed
on the basis of airborne photographic interpretation and
historical maps clearly suggests a palaeo-flow direction
from the northeast, where two major branches of a large
meandering river can be outlined, surrounding a flood-
basin, which is characterized by remnants of small
meanders (Fig. 5). Borehole logs from the northern area
of large meanders represent the fill of a major channel
with a bankful depth of 5 to 6 m (Tr-1, Dvs-1).
Floodplain deposits are less common in borehole Dvs-1,
in which stacked cross-stratified sands are attributed to
superimposed channels and/or braided bars (Figs. 6, 8).
The thickness of the pointbar sequences is less in the
southern zone of medium-size meanders (Gye-4, Gye-2
sections). This pattern suggests that the main meander-
ing channel system was to the north, while in the
southern part somewhat smaller meandering channels
co-existed (Figs. 6, 8).
The size of the palaeo-meanders in the northern part
of the study area is similar to those of the modern Tisza
River. Furthermore, the similarity of heavy mineral
assemblages of most of the sands from the large- and
medium-size meanders with the sediments of the
modern Tisza River and its northern tributaries suggests
that this meandering river system can be identified as the
palaeo-Tisza, which flowed along the rmellk depres-
sion during the Late Pleistocene. The major northward
avulsion of the palaeo-Tisza, triggered by the subsi-
dence of the Bodrogkz basin, occurred in the Late
Glacial (Gbris, 1998, 2002). This is consistent with our
data, which show that the channel sands from the
meanders exhibiting northeast transport directions are at
least 14 ky old. Transport directions also indicate
occasional sediment supply from the east (Fig. 6). This
suggests that precursors of the Krs Rivers coming
from the east also contributed to alluvial sedimentation
in this area.
The Gye-3 log from the area of the small meanders
revealed that the entire section consists of clayeysilty
sediments (Fig. 6), and historical maps show that this
area was a large marsh before river regulation. This
suggests that the area of the small meanders can be
interpreted as a floodbasin surrounded by the two main
branches of the meandering palaeo-Tisza.
The spatial distribution of the braided channels
outlines an approximately 10 km wide southeast
northwest striking zone with a palaeo-flow direction
from the southeast. The hypothesis that the southern part
of the former alluvial plain of a meandering river system
is partly overlain by a braided sand ridge is supported by
186 A. Ndor et al. / Sedimentary Geology 202 (2007) 174192

Fig. 8. Interpretation of the cross-section shown in Fig. 6. The meandering branches of the palaeo-Tisza flowed along subsiding depressions
controlled by the fault zones of the DvavnyaSzarvas sinistral strikeslip fault zone.

the OSL ages of the sediments: sands from the braided
zone are mostly 10 to 15 ky old, whereas most sands
from the meandering zone are at least 14 to 18 ky old
(Table 1). Furthermore, the elevation of the sand ridge of
the braided zone is 2 to 3 m higher than the meandering
zones. This suggests that a major change occurred when
a braided river entered the alluvial plain of the
meandering palaeo-Tisza and Krs Rivers from the
southeast, sometime during the Late Glacial. The
appearance of this braided river was more or less coeval
with the abandonment of the Tisza course along the
rmellk depression (Gbris, 1998, 2002). Borehole
logs and outcrop sections from the braided zone (Fig. 6)
are interpreted as deposits of a sandy braided channel
belt system. Sands with similar ages are at a lower
position at Hu-1 than at the Kondoros and Ks-1 sections
(Fig. 6), which may reflect the ancient slope morphol-
ogy and palaeo-flow direction from east-southeast to
west-northwest, as also suggested by the orientation of
planform channel geometries. Modern stream gradients
of the Fekete and FehrKrs Rivers draining from the
southeast are greater than those of the Beretty and
SebesKrs Rivers (Fig. 3), which could also explain
the braided pattern. The appearance of this braided river
system may have been related to accelerated uplift of the
southern part of the Apuseni Mountains source area.
Based on the heavy mineral assemblage, this river
system can be regarded as a precursor of the Fekete and
FehrKrs Rivers, which carry sediments mostly
from the southeast. However, the alternation of
northeast and southeast transport directions suggests
reworking of sediments that were previously deposited
by rivers draining northeastern source areas.
Supply of sediment by the Maros River is suggested
by the southeast transport direction and the composition
of the heavy-mineral assemblage (Fig. 7), and by
previous workers (Somogyi, 1961; Borsy, 1990). The
Maros alluvial fan complex (Fig. 2) includes three
distinct gravel horizons, of which the middle one is
directed towards the northwest, indicating that subsi-
dence of the Krs basin attracted the palaeo-Maros
River sometime during the MiddleLate Pleistocene
(Urbancsek, 1960, 1965). The uppermost (Late Pleisto-
cene?) gravel units of the Maros alluvial fan are east
west oriented. This indicates that the braided planform
geometries in the southern part of the Krs basin are
not linked to the Maros alluvial fan.
10.1. Tectonic control on river development
OSL ages of the sediments provided the possibility to
outline changes in sediment accumulation rate through
 A. Ndor et al. / Sedimentary Geology 202 (2007) 174192
time. The geological setting of the study area indicates
that accommodation for fluvial sediments was mainly
created by subsidence. This implies that changes in
sediment accumulation rates are a direct reflection of
changes in subsidence rates.
Several chronostratigraphic horizons were marked
along the studied profile (at 12, 14, 16, 18 and 20 ky
respectively) (Fig. 6), based on ages from the boreholes.
The difference between a faster sedimentation rate for
channel sands (e.g. the 3 m thick sand body in Dvs-1
borehole was deposited within an approximately
1000 year period according to OSL ages) and the
relatively slower deposition of floodplain deposits was
taken into account when interpolating dated levels
between boreholes and outcrops. The spatial arrange-
ment of selected chronostratigraphic horizons in the
boreholes and outcrops shows that the sedimentation
rate was highest in the northern meandering zone until
14 ky, as shown by the branching patterns of the time
lines. Between 18 (20) and 14 ky, the sedimentation rate
was also high in the southern branch of the meandering
zone, as shown by the thickening section in the Gye-4
borehole. These data support the idea that the two
branches of the meandering palaeo-Tisza must have
coexisted for a relatively long time (several thousand
years), and flowed along subsiding depressions which
channelled the main flow directions. From 14 (16) ky
onwards, fluvial sands were mainly deposited to the
south, coeval with the appearance of the braided palaeo-
Fekete and FehrKrs Rivers. The parallel time lines
in the south indicate a spatially uniform sedimentation
rate of approximately 0.51.0 m/ky. Channels often
changed their course and deposited a sand sheet over a
large area.
The alternations in sediment thickness are related to
subsidence anomalies, which may be linked to varia-
tions in the activity of structural zones inferred from
seismic interpretations (Figs. 5, 8). The linear contact of
the northern meandering zone and the adjoining flood-
basin area coincides with one of the faults within the
northeastsouthwest striking DvavnyaSzarvas
zone. One of the northward dipping elements of this
zone, which supposedly shows a negative flower
structure, was interpreted as a sinistral strikeslip zone
(Szentgyrgyi, 1988; Pstynyi, 1993). It joins a larger
Pleistocene left lateral strikeslip zone towards the
northeast, identified by Pogcss et al. (1989). This fault
zone continues towards the northeast at the edge of the
Derecske trough, where Late Pleistocene sinistral
strikeslip movements were proved by Pogcss et al.
(1989) and also by the re-interpretation of seismic
profiles (Dudko, 2002). These interpretations are in
187
good agreement with 3-D numerical modelling of
Windhoffer et al. (2004), who interpreted the Derecske
trough as a recently reactivated negative flower
structure. Further to the northeast, this zone continues
into the rmellk region, an area in which sinistral
strikeslip has been identified as the main control on
landform evolution during the Late Pleistocene, based
on neotectonic field measurements and morphotectonic
analyses (Magyari et al., 2006).
10.2. Climate control on river development
The next question to be discussed is whether there is
any correlation between the dated sediments from the
different zones, and millennial scale climate oscillations
of the Late PleniglacialLate Glacial periods. Many
studies of northwest European rivers showed that rivers
generally developed a braided pattern during glacial/
stadial periods due to the large amount and availability
of coarse-grained sediments and ephemeral discharges,
while during milder periods they tend to have a
meandering pattern due to well-established vegetation,
more balanced discharges and limited sediment load
(e.g. Kozarski, 1991; Starkel, 1991; Vandenberghe
et al., 1994; Bohncke et al., 1995; Kasse et al., 1995;
Mol et al., 2000).
Fig. 9 summarizes the palaeoclimatic and palaeoen-
vironmental changes in the Pannonian Basin over the
last 25 ky, with emphasis on the controls on fluvial
activity. During glacial/stadial periods, conditions in
the Pannonian Basin were similar to those in Northwest
Europe to a certain extent (dry periods with ephemeral
discharges, strong aeolian activity), but there are also
striking differences. The Pannonian Basin, as an
unglaciated region during the last Glacial period
(Ehlers and Gibbard, 2004) also belonged to the
periglacial zone (Vandenberghe and Pissart, 1993),
where permafrost phenomena have been known for a
long time (Pcsi, 1961; Maruszczak, 1987). However,
recent studies showed that some of the sediment
deformation features, which were originally ascribed
to cryoturbation, are in fact related to palaeo-seismic
activity (Van-Vliet Lanoe et al., 2004). Reconstructions
of palaeo-vegetation (Willis et al., 1995, 2000) showed
that during the Late Weichselian the Hungarian
landscape was characterized by closed mixed taiga
spots with coniferous trees and pockets of deciduous
trees, associated with treeless steppe elements. This
suggests that sediment supply to the Hungarian rivers
could have been lower than in Northwest Europe.
Interstadials with a well-established vegetation cover
in combination with higher average discharge and
188 A. Ndor et al. / Sedimentary Geology 202 (2007) 174192

Fig. 9. Palaeoenvironmental changes in the Pannonian Basin during the last 25 ky (after Gbris and Ndor, in press). Mean July temperature and
humidity is based on Kordos, 1977; Szr et al., 1991; Smegi et al., 1991; Smegi and Hertelendi, 1998; Smegi and Krolopp, 2002. Vegetation
based on Willis et al., 1995, 2000; Smegi and Krolopp, 2002. Aeolian processes based on Gbris et al., 2002; Gbris, 2003; Ujhzy et al., 2003.

limited sediment supply could have been similar in
both areas.
Comparison with northwest European sections was
further impeded by the fact that the classical Late
Weichselian chronostratigraphy based on radiocarbon
data was originally designed for Northwest European
sites (Mangerud et al., 1974), and derived largely from
early palaeobotanical studies referring to vegetational
stages of Scandinavia. Recognition of comparable
pollen zones throughout Northwest Europe led to the
wide application of the same pollen zones, which were
considered to be broadly synchronous and presumed to
reflect particular climate events. However, a rigorous
chronostratigraphic interpretation of pollen records is
 A. Ndor et al. / Sedimentary Geology 202 (2007) 174192
untenable, because of the diachronous nature of
biozones, and regional climatostratigraphic signals
cannot be neglected (Bjrck et al., 1998). Terrestrial
data from other parts of Europe, like the Pannonian
basin, also document significant climatic shifts during
the Last Termination, but these are diachronous and
modified by local thresholds.
In our attempt to compare dated fluvial deposits of
the studied sections in the Krs basin with climate
oscillations of the Late PleniglacialLate Glacial
periods, we faced the problem that the error range of
OSL dates often exceeds the duration of the Weichselian
substages. Therefore, the correspondence in ages of
certain units is somewhat arbitrary. However, a large
number of the samples showed a good correspondence
with the millennial-scale climate changes of the Late
PleniglacialLate Glacial periods (Table 1).
Much of the sand from the zone of the meanders
represents deposition during the BllingAllerd and
SgvrLascaux interstadials, both wet and mild
periods (Fig. 9), in which meandering rivers were
inferred to have been active in the Middle Tisza region
as well (Vandenberghe et al., 2003; Gbris and Nagy,
2005). In the braided zone some dated sand units
represent the Older and Younger Dryas, which were
relatively cold and dry periods (Fig. 9), during which
braided fluvial activity also characterized the Middle
Tisza region (Vandenberghe et al., 2003; Gbris and
Nagy, 2005). There are some samples from the braided
zone the ages of which are consistent with the Blling
Allerd interstadial. Although this period is regarded as
the beginning of the Late Glacial warming, its middle
part was characterized by intensive aeolian activity in
Hungary (Gbris, 2003), which may be linked to the
braided fluvial style in the southern part of the Krs
basin. Despite these climate-induced secular trends in
river patterns, meandering and braided rivers appear to
have coexisted under different climate conditions.
11. Summary and conclusions
Various planform geometries of meandering and
braided rivers were reconstructed on the basis of
airborne photographic interpretation combined with
analysis of 18th century topographic maps, that show
natural river patterns in the Krs basin. The arrange-
ment of the meanders suggests a palaeo-flow direction
from the northeast, where two major branches of a large
meandering river can be outlined, surrounding a
floodbasin.
The similarity of heavy mineral assemblages of most
of the sands from the meanders with the sediments of the
189
modern Tisza River and its northern tributaries suggests
that this meandering river system can be identified as the
palaeo-Tisza. Transport directions also indicate occa-
sional sediment supply from the east by precursors of
the Krs Rivers. Most of the sands from this
meandering zone are at least 14 to 18 ky old.
The spatial distribution of the braided channels
indicates palaeo-flow from the southeast. Sands from
the braided zone are mostly 10 to 15 ky old, which
suggests that a braided river entered the alluvial plain of
the meandering palaeo-Tisza and Krs Rivers during
the Late Glacial. Based on the heavy mineral assem-
blage, this river system can be regarded as a precursor of
the Fekete and FehrKrs Rivers, which carry
sediments mostly from the southeast. The appearance
of this sandy braided river was more or less coeval with
the northward avulsion of the palaeo-Tisza (Gbris,
1998, 2002).
Late-stage (PlioPleistocene) subsidence anomalies
in the Pannonian Basin are linked to compressional
basin inversion (Horvth and Cloetingh, 1996; Bada and
Horvth, 2001). Hence, changes in subsidence rate in
various parts of the Krs basin allow us to infer a major
tectonic control on river development. Interpretation of
seismic profiles of the area, studies of the recent stress
field, and neotectonic measurements all indicate a
sinistral strikeslip zone (DvavnyaSzarvas zone)
along the northwestern margin of the Krs basin,
which manifests itself as a range of fault patterns
(mainly flower structures). Borehole logs revealed
intensive subsidence along the northern part of the
Krs basin (rmellk depression) until 14 to 16 ky, as
shown by the increased thickness of sediments. This
syn-sedimentary subsidence controlled by the sinistral
DvavnyaSzarvas strikeslip zone forced the
meandering palaeo-Tisza to flow in the rmellk
depression during the Late Pleniglacial. The waning
subsidence of the rmellk depression after 14 ky, and
the intensive subsidence of the Bodrogkz basin in the
north led to avulsion of the palaeo-Tisza away from the
rmellk depression, which was then occupied by the
much smaller Beretty River. Another important event
was the appearance of the braided palaeo-Fekete and
FehrKrs Rivers during the Late Glacial in the
southern part of the Krs basin, which deposited a sand
sheet on top of the meanders left by the palaeo-Tisza and
Krs Rivers. The appearance of the ancestral Fekete
and FehrKrs Rivers may be related to accelerated
uplift of the southern part of the Apuseni Mountains,
their source area.
An unambiguous relationship between millennial-
scale climate changes of the Late PleniglacialLate
190 A. Ndor et al. / Sedimentary Geology 202 (2007) 174192
Glacial periods and reconstructed fluvial activity in the
Krs basin could not be demonstrated. Much of the sand
from the zone of the meanders represents deposition
during the BllingAllerd and SgvrLascaux (post-
LGM) interstadials. Some dated sand units from the
braided zone represent the Older and Younger Dryas. Our
data clearly show that braided and meandering rivers
coexisted under different climate conditions.
Acknowledgements
The research was supported by the Hungarian
National research Fund (OTKA) T. 46307. Thorough
reviews of the manuscript by Phil Gibbard, Orsolya
Sztan and Gert Jan Weltje, as well as discussions with
Rajiv Sinha and Sampat Tandon were particularly
helpful in improving this paper.
References
Antoine, P., Lautridou, J.P., Laurent, M., 2000. Long-term archives in
NW France: response of the Seine and Somme rivers to tectonic
movements, climatic variations and sea-level changes. Geomor-
phology 2000, 183207.
Bacs, N., 1959. Magyarorszg ghajlata. (Climate of Hungary).
Akadmiai Kiad, Budapest. 302 pp.
Bada, G., Horvth, F., 2001. On the structure and tectonic evolution of
the Pannonian basin and surrounding orogens. Acta Geologica
Hungarica 44, 301327.
Brczi, I., Phillips, R.L., 1985. Processes and depositional environ-
ments within deltaic-lacustrine sediments, Pannonian Basin,
Southeast Hungary. Geophysical Transactions 31, 5574.
Berendsen, H.J.A., Stouthamer, E., 2000. Late Weichselian and Holo-
cene palaeogeography of the RhineMeuse delta, The Netherlands.
Palaeogeography, Palaeoclimatology, Palaeoecology 161, 311335.
Bjrck, S., Walker, M.J.C., Cwynar, les C., Johnsen, S., Knudsen, K.L.,
Lowe, J.J., Wohlfarth, B., INTIMATE Members, 1998. An event
stratigraphy for the Last Termination in the North Atlantic region
based on the Greenland ice-core record: a proposal by the
INTIMATE group. Journal of Quaternary Science 13, 283292.
Boenigk, W., 2002. The Pleistocene drainage pattern in the Lower
Rhine Basin. Netherlands Journal of Geosciences 81, 202209.
Bohncke, S., Kasse, C., Vandenberghe, J., 1995. Climate induced envi-
ronmental changes during the Vistulian Lateglacial at Zabinko,
Poland. Quaestiones Geographicae. Special Issue 4, 4364.
Borsy, Z., 1990. Evolution of the alluvial fans of the Alfld. In: Rachocki,
A.H., Church, M. (Eds.), Alluvial Fans, a Field Approach. John
Wiley and Sons, pp. 229248.
Bridgland, D.R., 2000. River terrace systems in north-west Europe: an
archive of environmental change, uplift and early human
occupation. Quaternary Science Reviews 19, 12931303.
Busschers, F.S., Weerts, H.J.T., Wallinga, J., Cleveringa, P., Kasse, C.,
De Wolf, H., 2005. Sedimentary architecture and optical dating of
Middle and Late Pleistocene RhineMeuse deposits fluvial
response to climate change, sea-level fluctuation and glaciation.
Geologie en Mijnbouw 84, 2541.
Cloetingh, S., Burov, E., Poliakov, A., 1999. Lithosphere folding:
primary response to compression? (from central Asia to Paris basin).
Tectonics 18, 10641083.
Davis, J.C., 1986. Statistics and Data Analysis in Geology. Wiley, New
York. 646 pp.
Dudko, A., 2002. Szeizmikus szelvnyek neotektonikai rtkelse a
Krs medence terletn. Kzirat. Orszgos Fldtani s Geofizikai
Adattr 26 p.
Ehlers, J., Gibbard, P.L. (Eds.), 2004. Quaternary Glaciations extent
and chronology, Part I: Europe. Developments in Quaternary
Science, vol. 2a. Elsevier, Amsterdam.
Gbris, Gy., 1998. Late Glacial and Post Glacial development of drainage
network and the paleohydrology in the Great Hungarian Plain. In:
Bassa, L., Kertsz, . (Eds.), Windows on Hungarian Geography.
Akadmiai Kiad, Budapest, pp. 2336.
Gbris, Gy., 2002. A Tisza helyvltozsai. (The shifting Tisza River).
In: Mszros, R., Schweitzer, F., Tth, J. (Eds.), Jakucs Lszl, a
tuds, az ismeretterjeszt s a mvsz. Magyar Tudomnyos
Akadmia Fldrajzi Kutatintzet, Pcsi Tudomnyegyetem
kiadsa, Pcs, pp. 91105.
Gbris, Gy., 2003. A fldtrtnet utols 30 ezer vnek szakaszai s a
futhomok mozgsnak fbb peridusai Magyarorszgon. (The
periods of the history of the Earth for the last 30 thousand years and
the periods of the movement of aeolian sand). Fldrajzi Kzlem-
nyek 127, 114.
Gbris, Gy., Nagy, B., 2005. Climate and tectonically controlled river
style changes on the SajHernd alluvial fan (Hungary). In:
Harvey, A.M., Mather, A.E., Stokes, M. (Eds.), Alluvial Fans:
Geomorphology, Sedimentology, Dynamics. Geological Society,
London, Special Publications, vol. 251, pp. 6167.
Gbris,Gy., Ndor, A., in press. Long-term fluvial archives in Hungary:
response of the Danube and Tisza rivers to tectonic movements
and climatic changes during the Quaternary. Quaternary Science
Reviews.
Gbris, Gy., Horvth, E., Novothny, ., Ujhzy, K., 2002. History of
environmental changes from the Last Glacial period in Hungary.
Praehistoria 3, 922.
Gibbard, P.L., 1985. The Pleistocene History of the Middle Thames
Valley. Cambridge University Press, Cambridge. 155 pp.
Horvth, F., Cloetingh, S., 1996. Stress induced late-stage subsidence
anomalies in the Pannonian Basin. In: Cloetingh, S., Ben Avraham,
Z., Sassi, W., Horvth, F. (Eds.), Dynamics of Extensional Basins
and Inversion Tectonics. Tectonophysics, vol. 266, pp. 287300.
Jask, S., Krolopp, E., 1991. Negyedidszaki kregmozgsok s
folyvzi ledkfelhalmozds a Duna-vlgyben Paks s Mohcs
kztt. (Quaternary crustal movements and fluvial sedimentation
in the Danube valley between Paks and Mohcs). Magyar llami
Fldtani Intzet vi Jelentse 1989, 6583.
Johnsen, S.J., Clausen, H.B., Dansgaard, W., Fuhrer, K., Gundestrup,
N., Hammer, C.U., Iversen, P., Jouzel, J., Stauffer, B., Steffensen,
J.P., 1992. Irregular interstadials recorded in a new Greenland ice
core. Nature 359, 311313.
Jmbor, ., 1989. Review of the geology of the s.l. Pannonian formations
of Hungary. Acta Geologica Hungarica 32, 269324.
Juhsz, Gy., 1994. Magyarorszgi neogn medencerszek pannniai s.
l. ledksornak sszehasonlt elemzse. Fldtani Kzlny 124,
341365.
Kasse, C., Bohncke, S., Vandenberghe, J., 1995. Fluvial periglacial
environments, climate and vegetation during the middle Weichse-
lian with special reference to the Hengelo Interstadial. Mededlin-
gen Rijks Geologische Dienst 52, 387413.
Kordos, L., 1977. Holocn klmavltozsok kimutatsa Magyarorsz-
gon a pocok-hmr segtsgvel. (Holocene climate changes in
Hungary based on the vole-thermometer method). Fldrajzi
Kzlemnyek 25, 222229.
 A. Ndor et al. / Sedimentary Geology 202 (2007) 174192 191

Kozarski, S., 1991. Warhta a case study of a lowland river. In: Starkel, L., 1991. Long-distance correlation of fluvial events in the
Starkel, L., Gregory, K., Thornes, J. (Eds.), Temperate Palaeohy- temperate zone. In: Starkel, L., Gregory, K.J., Thornes, H.J. (Eds.),
drology. John Wiley and Sons, pp. 189215. Temperate Palaeohydrology. John Wiley and Sons, pp. 473495.
Maddy, D., Bridgland, D.R., 2000. Accelerated uplift resulting from Smegi, P., Hertelendi, E., 1998. Reconstruction of microenviron-
Anglian glacioisostatic rebound in the Middle Thames Valley, mental changes in Kopasz Hill loess area at Tokaj (Hungary)
UK?: evidence from the river terrace record. Quaternary Science between 15,00070,000 BP years. Radiocarbon 40, 855863.
Reviews 18, 245260. Smegi, P., Krolopp, E., 2002. Quatermalacological analyses for
Magyar, I., Geary, D., Mller, P., 1999. Paleogeographic evolution of modelling of the Upper Weichselian paleoenvironmental changes
the Late Miocene Lake Pannon in Central Europe. Palaeogeo- in the Carpathian Basin. Quaternary International 91, 5363.
graphy, Palaeoclimatology, Palaeoecology 147, 151167. Smegi, P., Szr, Gy., Hertelendi, E., 1991. Paleoenvironmental
Magyari, ., Kercsmr, Zs., Unger, Z., Tham-Bozs, E., 2006. reconstruction of the last period of the Upper Wrm in Hungary,
Neotectonic investigations of the rmellk region (NE Pannonian based on malacological and radiocarbon data. Soosiana 19,
Basin, NW Transylvania). GeoLines 20, 8788. 512.
Mangerud, J., Andersen, S.T., Berglund, B.E., Donner, J.J., 1974. Qua- Szentgyrgyi, K-n, 1988. Szarvas (Endrd-II) lehatrol kutatsi
ternary stratigraphy of Norden, a proposal for terminology and fzis fldtani zrjelentse. Kzirat. Orszgos Fldtani s
classification. Boreas 4, 109128. Geofizikai Adattr T. 14798.
Maruszczak, H., 1987. Problems of paleogeographic interpretation of Szr, Gy., Smegi, P., Hertelend, E., 1991. Malacological and isotope
ice-wedge casts in European loess: SEM characterisation micro- geochemical methods for tracing Upper Quaternary climatic
features on frost shattered quartz grains. In: Pcsi, M., French, H.M. changes. In: Pcsi, M., Schweitzer, F. (Eds.), Quaternary
(Eds.), Loess and the Periglacial Phenomena. Akadmiai Kiad, Environment in Hungary. Akadmiai Kiad, Budapest, pp. 6173.
Budapest, pp. 285302. Tham-Bozs, E., Kercsmr, Zs., Ndor, A., 2002. Tectonic control on
Mike, K., 1975. Utilisation of the analysis of ancient river beds for the changes in sediment supply on Quaternary alluvial systems, Krs
detection of Holocene crustal movements. Tectonophysics 29, sub-basin, SE Hungary. In: Jones, S.J., Frostick, L.E. (Eds.), Sediment
359368. Flux to Basins: Causes, Controls and Consequences. Geological
Mol, J., Vandenberghe, J., Kasse, C., 2000. River response to variation Society, London, Special Publications, vol. 191, pp. 3753.
of periglacial climate in mid latitude Europe. Geomorphology 33, Tham-Bozs, E., Murray, A.S., Ndor, A., Magyari, ., Babinszki, E.,
132148. 2007. Investigation of river network evolution using luminescence
Murray, A.S., Wintle, A.G., 2000. Luminescence dating of quartz dating and heavy mineral analysis of Late-Quaternary fluvial sands
using an improved single-aliquot regenerative-dose protocol. from the Great Hungarian Plain. Quaternary Geochronology 2,
Radiation Measurements 32, 5773. 168173.
Murray, A.S., Wintle, A.G., 2003. The single aliquot regenerative dose Trnquist, T.E., 1995. Alluvial architecture of the Quaternary Rhine
protocol: potential for improvements in reliability. Radiation Meuse system in The Netherlands discussion. Geologie &
Measurements 37, 377381. Mijnbouw 74, 183186.
Ndor, A., Lantos, M., Tth-Makk, ., Tham-Bozs, E., 2003. Milan- Trnquist, T.E., Wallinga, J., Murray, A.S., De Wolf, H., Cleveringa, P.,
kovitch-scale multi-proxy records for the fluvial sediments of the last De Gans, W., 2000. Response of the RhineMeuse system (west-
2.6 Ma from the Pannonian Basin, Hungary. Quaternary Science central Netherlands) to the last Quaternary glacio-eustatic cycles: a
Reviews 22, 21572175. first assessment. Global and Planetary Change 27, 89111.
.Kovcs, L., 1986. Cluster-analzis eljrsok TPA/L szmtgpen. Ujhzy, K., Gbris, Gy., Frechen, M., 2003. Ages of periods of sand
Magyar llami Fldtani Intzet vi Jelentse 1985-rl, 571582. movement in Hungary determined through luminescence measure-
Papp, A., 1960. Fiatalkori vzrajzi vltozsok a Tiszntl kzps ments. Quaternary International 111, 91100.
rszben trtnelmi adatok alapjn. Fldrajzi Kzlemnyek 8, Urbancsek, J., 1960. Az alfldi rtzi kutak fajlagos vzhozama s
7784. abbl levonhat vzfldtani s sfldrajzi kvetkeztetsek.
Pcsi, M., 1959. A magyarorszgi Duna-vlgy fejldstrtnete. Hidrolgiai Kzlny 1960/5, 398403.
(Entwicklung und Morphologie des Donautales in Ungarn). Urbancsek, J., 1965. Az Alfld negyedkori fldtani kpzdmnyeinek
Akadmiai Kiad, Budapest. 346 pp. mlyszerkezete. Hidrolgiai Kzlny 1965/3, 111124.
Pcsi, M., 1961. A periglacilis talajfagyjelensgek fbb tpusai Vandenberghe, J., 1993. River terrace development and its relation to
Magyarorszgon. Fldrajzi Kzlemnyek 9, 124. climate: the Saalian Caberg terrace of the Maas river near Maastricht
Pogcss, Gy., Lakatos, L., Ujszszi, K., Vakarcs, G., Vrkonyi, L., (The Netherlands). Mededelingen - Rijks Geologische Dienst 47,
Vrnai, P., Rvsz, I., 1988. Seismic facies, electro facies and 1924.
Neogene sequence chronology of the Pannonian basin. Acta Vandenberghe, J., Pissart, A., 1993. Permafrost changes in Europe during
Geologica Hungarica 31, 175207. the last glacial. Permafrost and Periglacial Processes 4, 121135.
Pogcss, Gy., Lakatos, L., Barvitz, A., Vakarcs, G., Farkas, Cs., 1989. Vandenberghe, J., Kasse, C., Bohncke, S., Kozarski, S., 1994. Climate-
PliocnQuarter oldaleltoldsok a Nagyalfldn (Pliocene related river activity at the WechselianHolocene transition: a
Quaternary transcurrent fault in the Great Hungarian Plain. comparative study of the Wartha and Maas rivers. Terra Nova 6,
ltalnos Fldtani Szemle 24, 149169. 476485.
Pstynyi, F., 1993. Szeizmikus feldolgozsi jelents, Dvavnya- Vandenberghe, J., Kasse, K., Gbris, Gy., Bohncke, S., Van Huissteden,
Kelet. Kzirat. Orszgos Fldtani s Geofizikai Adattr T. 19875. K., 2003. Fluvial style changes during the last 35.000 years in the
Royden, L.H., Horvth, F. (Eds.), 1988. The Pannonian Basin. A Case Tisza valley. XVI. INQUA Congress, 2330 July 2003, Reno,
Study in Basin Evolution. AAPG Memoirs, vol. 45. 394 pp. Nevada, USA, Abstracts, p. 68.
Somogyi, S., 1961. Haznk folyhlzatnak fejldstrtneti vzlata. Van den Berg, M.W., 1994. Fluvial sequences of the Maas: a 10 Ma
(Entwicklungsgeschichtige Skizze der Wassernetzes von Ungarn). record of neotectonics and climatic change at various time-scales.
Fldrajzi Kzlemnyek 9, 2550. PhD Thesis, Universal Press, Veenendaal, 181 pp.
192 A. Ndor et al. / Sedimentary Geology 202 (2007) 174192
Van-Vliet Lanoe, B., Magyari, ., Melliez, F., 2004. Neotectonic and
seismic activity in Europe: the differentiation between co-seismic
and periglacial deformations. Global and Planetary Change 43,
103127.
Waelbroeck, C., Labeyrie, L., Michel, E., Duplessy, J.C., McManus, J.F.,
Lambeck, K., Balbon, E., Labracherie, M., 2002. Sea-level and deep
water temperature changes derived from benthic foraminifera
isotopic records. Quaternary Science Reviews 21, 295305.
Wallinga, F.J., Trnquist, T.E., Busschers, F.S., Weerts, H.J.T., 2004.
Allogenic forcing of the late Quaternary RhineMeuse fluvial
record: the interplay of sea-level change, climate change and
crustal movements. Basin Research 16, 535547.
Willis, K.J., Smegi, P., Braun, M., Tth, A., 1995. The Late Qua-
ternary environmental history of Btorliget, N.E. Hungary.
Palaeogeography, Palaeoclimatology, Palaeoecology 118, 2547.
Willis, K.J., Rudner, E., Smegi, P., 2000. The full-glacial forest of
Central and Southeastern Europe. Quaternary Research 53,
203213.
Windhoffer, G., Bada, G., Nieuwland, D., Wrum, G., Horvth, F.,
Cloething, S., 2004. On the mechanics of basin formation in the
Pannonian basin: inferences from analogue and numerical
modeling. Tectonophysics 410, 389415.
 Available online at www.sciencedirect.com

Sedimentary Geology 202 (2007) 193 210

www.elsevier.com/locate/sedgeo

Fluvial response to rapid climate change during the

Devensian (Weichselian) Lateglacial in the River Great Ouse,
southern England, UK
C. Gao a,, S. Boreham b , R.C. Preece c , P.L. Gibbard b , R.M. Briant d
a
Ontario Geological Survey, 933 Ramsey Lake Road, Sudbury, Ontario P3E 6B5, Canada
b
Cambridge Quaternary, Department of Geography, University of Cambridge, Cambridge CB2 3EN, England, United Kingdom
c
Cambridge Quaternary, Department of Zoology, University of Cambridge, Cambridge CB2 3RS, England, United Kingdom
d
Department of Geography, Queen Mary, University of London, Mile End Road, London, E1 4NS, England, United Kingdom

Abstract

Few Lateglacial (15,20011,500 Cal BP) fluvial deposits are known from southern England, UK. This paper outlines a
sedimentological investigation of a Lateglacial site near the village of Roxton in the middle reach of the River Great Ouse. Below
the Holocene argillaceous alluvium, several large gravelly lithofacies units representing long-term fluvial processes are recognised
based on their lower bounding surfaces and sedimentary features. In line with palaeontological evidence and radiocarbon dating,
detailed facies analysis indicates that the river changed from a braided to a wandering/meandering regime in response to the
warming-induced reduction in nival peak discharge and increase in vegetation cover consisting of birch woodland and grassland
during the Lateglacial Interstadial (15,20012,700 Cal BP). This was followed by a change back to a braided river in response to
climatic deterioration during the Younger Dryas Chronozone (12,70011,500 Cal BP). The interstadial floodplain consisted of both
overbank fines and channel lateral or point-bar sands and gravels. The two suites of contemporaneous sediments in the geological
succession may be easily mis-interpreted as independent climato-stratigraphical units due to their cross-cutting relationship and
marked difference in facies assemblage. Lateral channel movement in such a system also implied that the floodplain underwent
constant erosion, resulting in underdeveloped overbank sequences, insufficient to withstand post-depositional removal by erosion.
This probably accounts for the rare preservation of fine-textured overbank deposits of the Lateglacial Interstadial in southern
England.
2007 Elsevier B.V. All rights reserved.

Keywords: Facies models; Wandering/meandering river; Climate change; Lateglacial; Great Ouse; Southern England

1. Introduction continental sedimentation and geomorphological pro-

The Lateglacial climate (ca. 15,200 to 11,500 Cal
year BP, Table 1) underwent drastic change over the
North Atlantic region, imposing strong impacts on
Corresponding author. Fax: +1 705 670 5905.
E-mail address: george.gao@ndm.gov.on.ca (C. Gao).
0037-0738/$ - see front matter 2007 Elsevier B.V. All rights reserved.
doi:10.1016/j.sedgeo.2007.02.004
cesses (Bond et al., 1993; Dansgaard et al., 1993; Lowe
et al., 1999). In response to climatic oscillations and
associated environmental changes, many lowland rivers
in the northwestern and central European mainland
experienced changes in behaviour and style (Vanden-
berghe et al., 1994; Starkel, 1995; Mol et al., 2000). In
contrast, few Lateglacial fluvial deposits containing
similar evidence are known from the British Isles (Rose
194 C. Gao et al. / Sedimentary Geology 202 (2007) 193210

Table 1
British Late Devensian stratigraphy based on Rose (1989), Bjrck et al. (1998) and Lowe et al. (1999)

et al., 1980; Collins et al., 1996; Lewis et al., 2001). This British lowland rivers characteristically emplaced
paucity has been attributed to a high process threshold gravelly deposits in a braided regime in non-glaciated
for the rivers to cross or the removal of the deposits by southern England during the last glacial Devensian
erosion (Van Huissteden et al., 2001). (Weichselian) Stage (Fig. 1), whereas Holocene

Fig. 1. Location map of the study area. Star indicates the study site shown in Fig. 2. The two capital letters such as TL, TM and SU are British
Ordinance Survey grid references. The grey area in the upper left inset map indicates the Late Devensian ice advance (Bowen et al., 1986).
 C. Gao et al. / Sedimentary Geology 202 (2007) 193210
deposition is characterised by widespread accretion of
argillaceous alluvium from a meandering regime due to
changing vegetational and hydrological conditions as
well as the subsequent change in sediment supply
(Gibbard and Lewin, 2002). Although rare, available
sedimentary evidence from the Late Devensian succes-
sion suggests changing river styles in response to the
climatic variability (Rose, 1995). While sedimentary
records alone cannot be over-emphasised for environ-
mental implications, recorded fossiliferous materials in
the sediments provide independent palaeoenvironmental
evidence and chronological control by dating them. This
allows the sedimentary records to be studied under the
regional stratigraphical framework and compared
directly with the climatic events.
The aims and objectives of this study are, through a
case study area, to construct facies models through
detailed sedimentological investigations, and to under-
stand fluvial response to rapid climate change during the
Lateglacial period of 15,20011,500 Cal BP in southern
UK. This paper presents our observations and research
results at a Late Devensian site near the village of
Roxton in the middle reach of the River Great Ouse
where evidence exists for marked fluvial change in
response to Lateglacial climatic oscillations (Fig. 1).
2. Methods
The recognition of fluvial system response to climate
change is made possible through a thorough analysis of
the sedimentological and palaeontological evidence, and
a well constrained geochronology. Through detailed
analysis of internal and external geometries of sedi-
mentary deposits such as structures, grain size and
bounding surfaces, facies models can be constructed, as
Miall (1996) summarised, to characterise the change of
fluvial planform in a geological succession. Palaeonto-
Table 2
Lithofacies classification (modified from Miall, 1996)
Code Lithofacies
Gh Glast-supported gavel, massive or with crude horizontal bedding
Gp Gravel, planar cross-stratified
Gt Gravel, trough cross-stratified
Sh Sand, horizontally bedded
Sl Sand, low-angle (<15) cross beds
Sp Planar cross-bedded sand, solitary or grouped
Sr Sand with ripple cross lamination
Ss Cross-bedded sand in broad, shallow scours
St Trough cross-bedded sand, solitary or grouped
Fl Sand, silt and mud with fine lamination and small ripple marks
Fm Massive mud and silt, occasionally with desiccation crakes
Fsm Massive silt and mud, frequently with freshwater shells
195
logical evidence provides a proxy indicator for
palaeoenvironment and, therefore, establishes the pos-
sible link between changing channel pattern and
changing climate. As macroscopic plant remains were
available, radiocarbon dating was used to provide
chronological control for regional correlation in this
study.
In the field, the available sections were recorded
using a combination of vertical sections, drawings and
photographs. For convenience and simplification, codes
were used in lithofacies description (Tables 2 and 3).
Important characteristics such as lithology, sedimentary
structures, bounding surfaces, external features such as
shape, thickness and lateral extent were recorded.
Palaeocurrent directions were measured on 3rd-order
bounding surfaces and, where these were not available,
on cross-bedding planes. Key facies and prominent,
large-scale bounding surfaces were identified and
classified into lithofacies units. Bounding surface
terminology follows the classification of Miall (1993,
1996) based on cross-cutting relationships, in which 1st
and 2nd orders refer to cross-bed set and coset bounding
surfaces, the smallest-scale features, 3rd and 4th orders
define internal erosion (e.g., reactivation surfaces of a
point bar) and upper bounding surfaces of macroforms,
respectively, 5th order corresponds to large lithosome
boundaries, and 6th order, the largest scale encountered
in this study, refers to major lithosome boundaries that
define groups of large lithofacies units outlined by 5th-
order boundaries.
Over 45 borehole records were acquired from the
local aggregates producer the Redland, allowing insight
into the geometry of the sedimentary and lithofacies
units, making reliable correlation possible across the
study area. As discussed later, the boreholes have played
a key role in the delineation of an abandoned palaeo-
meander in this study. The study site was also levelled in
Depositional process
Longitudinal bedforms, lag deposits, sieve deposits
Transverse bedform or bars, or deltaic growth
Minor channel fills
Plane-bed flow, critical flow
Scour fills, antidunes
Linguoid and transverse bedforms
Ripples
Scour fills
Sinuous-crested and linguoid sand dunes
Overbank, abandoned channel, or waning flood deposits
Overbank or drape deposits
Backswamp or abandoned channel deposits
196 C. Gao et al. / Sedimentary Geology 202 (2007) 193210

Table 3
Architectural elements in fluvial deposits (modified from Miall, 1993, 1996)
Element Symbol Lithofacies Geometry
Longitudinal gravel bedforms GB Gh, Gt, Gp with minor Gravelly tabular bodies, often fining upwards with sand or pebbly
Sr, Sh, Sl, Ss sand beds at top
Lateral accretion macroform LA Sp, Gp, Gm, Sl, St, Fl Point bars attached to accretionary banks. Wedge or lobe with
3rd-order erosion surfaces dipping obliquely to stream
Overbank fines FF Fl, Fm, Fsm Silty blankets or lenses in abandoned channels
Downstream accretion macroform LA(DA) Sp, St, Gp Bars attached to banks (alternate bars) with 3rd-order
internal erosion surfaces
Channels CH Gh, Gt, Fl, Fm, Fsm, Sl Concave-up erosional base, filled by any combination of above
elements. Internal concave-up 3rd-order erosion surface common

order to obtain the elevations of the sedimentary and
lithofacies units, and their bounding surfaces, using
Ordnance Survey bench marks.
Samples were collected as necessary for laboratory
analyses. 20 to 30 kg of gravelly sediment was obtained
for pebble-lithology determination from the lithofacies
units identified in the field. Samples weighing 2 to 5 kg
were collected from silty sand and organic mud layers
for analysis of plant macroscopic remains and Mollusca.
For pollen analysis, organic sandy silt was sampled in
small tubes (Gao, 1997).
3. Field observations
The present-day River Great Ouse has a single-thread
meandering morphology, confined in a valley incised
into the chalky Lowestoft Formation till that was
emplaced during the Anglian ice advance at ca. 450 ka
BP (Bowen et al., 1989). The river has a mean annual
discharge of 14 m3 s 1 near St. Ives (Fig. 1). Four
topographical levels exist in the river valley from the
bottom to the valley shoulder, namely, the floodplain
and terraces T1, T2 and T3. The Holocene argillaceous
alluvium that encroaches onto the low terraces is
widespread on the floodplain.
The gravel pits for study are located in a low terrace
and the floodplain on the western bank of the river near
Roxton (TL158 538) (Fig. 2). The low terrace that
stands at about 16 m OD (above sea level) has been
mapped as the combined 1st2nd terrace in the Great
Ouse (Edmonds and Dinham, 1965). The 3rd terrace
occurs only locally at Roxton. The gravelly deposits

Fig. 2. Pits and sections at Roxton.

 C. Gao et al. / Sedimentary Geology 202 (2007) 193210
beneath the 3rd terrace are not exposed here but are
believed to have formed in the early Devensian (Gao,
1997). The chalky Lowestoft Formation till occurs on
valley shoulders and in deeply buried glacial channels
with its base at 2 m OD below the present valley in this
area (Horton, 1970; Gao, 1997).
The pits have revealed 3 to 4 m of gravelly fluvial
deposits beneath the Holocene alluvial silty clay, which
rarely exceeds 1 m thick but thickens towards the river
(Fig. 2). Boreholes provided by the Redland have
uncovered silt and sand deposits filling an abandoned,
northsouth aligned meander bend (Fig. 2). This
meander bend infill has been subsequently examined
in pit sections and will be described later in detail.
The fluvial deposit is dominated by gravels with clast-
supported, massive to crude horizontal bedding struc-
tures (Gh) (Fig. 3). Clean, medium to coarse sand fills the
framework space or voids. Cross-bedded sand layers
occur locally (Sr, Sl, Ss). In places, silt beds or lenses
containing molluscan shells and plant debris occur (Fl)
(Fig. 3). The sub-rounded to sub-angular framework
clasts ranging from 1 to 6 cm diameter are dominated by
flint chert and include chalk, limestone and quartzite.
This range of lithology is identical to the Devensian
fluvial gravels in the Great Ouse that result from multiple
Fig. 3. Column sections, showing lithofacies units L-1 to L-5 with characteristic facies codes of Fl, Gh and Sp etc. Numbers at the top indicate pit
sections shown in Fig. 2. Elevations are metres (m) above sea level (OD).
197
phases of river reworking of the Lowestoft Formation till
(Rogerson, 1987; Gao, 1997). The fluvial gravel
sequence is underlain unconformably by the Lowestoft
Formation till and slightly lithified, bluish grey Jurassic
Oxford Clay (Fig. 3). In the till, glaciofluvial gravel
deposits occur, characterised by abundant round chalk
pebbles and cobbles. Several ice-wedge casts or
pseudomorphs occur in the pit sections (Fig. 2).
3.1. Lithofacies units
Based on the lower bounding surfaces and sedimen-
tary features, several lithofacies units have been
recognised in the fluvial succession above the Lowestoft
Formation till and Oxford Clay bedrock. The facies
assemblages of lithofacies units can be similar or
different depending on their depositional conditions
(Table 4). The lower bounding surfaces are easier to
define in the field where clear contrast in facies
assemblage exists between neighbouring lithofacies
units. The uppermost lithofacies unit L-5 consists
entirely of Holocene argillaceous alluvium with a 6th-
order lower bounding surface. Its depositional environ-
ment has been summarised in Table 4. For this reason,
this unit will not be described in detail.
198 C. Gao et al. / Sedimentary Geology 202 (2007) 193210

Table 4
Lithofacies units based on their lower bounding surfaces and sedimentary features
Litho-units Architectural Lower bounding Geometry and relationships Depositional environment
elements surface (and order)
L-5 FF Flat, unconformable Silty to clayey blankets on floodplain, Floodplain alluvium by a single-thread
(6th-order) encroaching to the 1st terrace. Small meandering river in middle and late Holocene.
channels occasionally seen at base High sinuosity
L-4 GB, CH Erosional, tabular Gravel beds with crude horizontal Gravel-bed braided river with well-defined
to concave-up bedding to trough cross-bedding, topographic levels. Low sinuosity a
(5th-order) often filling channels
L-3 GB, FF, LA Erosional, tabular to Elements separated by 4th-order erosion Gravel to sand-bed wandering/meandering
irregular (5th-order) surfaces. LA overlying directly on FF river with intermediate sinuosity b. Meandering
with erosional surfaces. GB commonly channels existed with attached point bars and
present well-built floodplain
L-2 GB Erosional, planar Gravels with massive to crude horizontal Gravel-bed braided river with development
concave-up (5th-order) bedding, minor sand beds as scour infill of topographic levels. Low sinuosity
and bar top sheets
L-1 GB Erosional, planar Same as above Similar to above
(6th-order)
a
Suggested sinuosity (P) values: low <1.2, intermediate 1.21.5, high >1.5 (after Miall, 1996).
b
Gravel-bed wandering rivers stand between braided and meandering rivers in planform and sinuosity (cf. Miall, 1996).

3.1.1. Lithofacies unit L-1 bedding (Sp, Gp) in massive to crudely bedded gravels
This is the basal gravel that directly overlies the (Gh) (Fig. 5). The abandoned meander bend occurs in
valley floor and has a tabular 6th-order lower bounding the upper part of L-3, extending about 350 m long and
surface that truncates both the Lowestoft Formation till
and Jurassic Oxford Clay. Best exposed in Section 6
(Fig. 3), it contains up to 0.5 m of light orange brown,
coarse gravel clasts and thins to the east. Although
crudely bedded, this gravel appears more massive in
texture than the overlying L-2. L-1 corresponds to the
lower gravel defined by Gao (1997).

3.1.2. Lithofacies unit L-2

This gravel unit has a 5th-order tabular lower
bounding surface that truncates L-1 and, in places, the
Lowestoft Formation till and Oxford Clay (Fig. 3). It
contains light brownish gravel with massive to crude
horizontal bedding structures (Gh). Lag cobbles
typically line the base of this gravel, outlining the
lower bounding surface. Where similar gravelly facies
and colour exist, the contact between L-2 and L-1
becomes obscured. Palaeocurrent directions were mea-
sured in the field on bedding planes from sand lenses,
which indicate a mean value of 341 (n = 10) (Fig. 4a).
L-2 was previously referred to as the upper gravel
(Gao, 1997).
Fig. 4. Palaeocurrent data shown as equal-area projection and rose
3.1.3. Lithofacies unit L-3 diagram. (a) Lithofacies unit L-2. Mean linear vector (MLV) in dip/
Lithofacies unit L-3 is defined by an irregular to inclination. (b) Lithofacies unit 3 is indicated by small triangles and
tabular, 5th-order lower bounding surface (Fig. 3). In black dots in equal-area projection, and marked by I and II in rose
diagram for the lower and upper parts of Section 3, respectively. Small
contrast to the underlying L-2, this unit frequently open circles and the marker III indicate lithofacies unit L-4. Mean
contains silt and silty sand beds (Fl, Fm, Fsm) and values and sample numbers are given in parentheses in the rose
pebbly sand and fine gravel sets with tabular cross- diagram.
 C. Gao et al. / Sedimentary Geology 202 (2007) 193210 199

Fig. 5. Schematic cross-section based on field observations, showing the lithofacies units and their lower bounding surfaces with highlighted
characteristic architectural elements, i.e., GB, FF and CH. Not to scale.

25 m wide at the top (Figs. 2 and 5). It contains 1.6 m of
laminated bluish grey, shelly silt and clay (Fl) in the
lower part and stratified to massive sand and pebbly
sand (Fl, St) in the upper.
In Section 1, a channel-like shallow depression filled
by 0.5 m of fossiliferous silt and sand occurs (Fig. 6).
This sequence, which has been radiocarbon-dated and
systematically sampled, consists of, from the base to the
top, a massive, shelly silt and sand (Fl), laminated shelly
silt (Fsm) containing abundant plant macrofossils, and a
massive to stratified shelly silty sand (Fl). This sequence
is truncated by the overlying L-4 (Fig. 6).
The pebbly sand and fine gravel sets with tabular
cross-bedding (Sp, Gp), ranging from 1.3 to 1.5 m
thick, are recorded at Sections 3 and 5 (Fig. 2). They
exhibit a complex internal geometry characterised by

Fig. 6. Main section, showing the silty channel infill at Section 1. Circled number indicates the rank of lower bounding surface.
200 C. Gao et al. / Sedimentary Geology 202 (2007) 193210

prominent (4th-order) intra-formational bounding sur-
faces (Fig. 7). These cross-sets show in places a
lower part consisting of sand, pebbly sand and fine
gravels with steep tabular foresets (Sp, Gp), and an
upper section of gently dipping gravels (Gh) and thin
silty sand layers (Fl) (Fig. 7). The bedding planes in
the lower part exhibit a consistent northeastward dip
with inclination of 26 (n = 2) and the upper bedding
planes or internal reactivation surfaces (3rd-order)
show an eastward dip with gentler inclination (12,
n = 5) (Figs. 4b, 7).
Although juxtaposed against massive to horizontal-
bedded gravels (Gh), the cross-stratified sand and
gravel sets are also in contact with silt beds (Figs. 5
and 7). In Section 3 for example, the cross-stratified
sand and gravel truncates a flat-based, massive to
laminated, bluish sandy silt layer (Fl), 30 to 40 cm in
thickness and 10 m in lateral extent, containing
molluscan shells (Fig. 7).
3.1.4. Lithofacies unit L-4
Lithofacies unit L-4 is outlined by an irregular to
tabular, 5th-order lower bounding surface truncating L-3
(Figs. 3 and 5). It is overlain unconformably by
Holocene argillaceous alluvium (L-5, Table 4).
Although dominated by massive to crudely horizontally
stratified gravels (Gh), L-4 contains several gravel sets
with trough cross-bedding (Gt) in cross-sections, filling
shallow channels with a dominantly NNE trend (Fig. 5).
The channels are <1.5 m deep, ranging from a few
metres to 10 m wide at the top. This unit becomes
thicker towards the river, reaching 2 to 3 m east of
Section 8 (Fig. 1). Bedding planes were measured from
sand lenses for palaeocurrent current analysis, indicating
a mean dip of 347 (n = 11) (Fig. 4b).
Several ice-wedge casts or pseudomorphs are
recorded in the pit sections (Fig. 2). They taper
downwards to form a V- or funnel-shaped morphology,
extending from the upper part of L-4 down into L-1, with
depth of over 1.8 m and width ranging from 0.8 to 1.1 m
at the top (Figs. 3 and 5). The enclosing strata are
disturbed and downturned by small faults (Fig. 8). The
alignment of the surface depressions of these ice-wedge
casts indicates a cross-cutting relationship (Fig. 2).
3.2. Interpretation
Lithofacies units L-1 and L-2 are interpreted as
gravelly bar and bedform deposits (GB) formed in a
braided river with low sinuosity (Table 4) (cf. Rust,
1972; Bryant, 1983; Gibbard, 1994; Gao et al., 2000;
Gao, 2004). Scour-infill sand beds and lenses are minor
components. Sands with ripple cross-lamination, prob-
ably common at the time of deposition, are infrequently
preserved as sand sheets on bar tops. Palaeocurrent
measurement from lithofacies unit L-2 indicates a

Fig. 7. Floodplain sequence at Section 3, containing an overbank alluvium (FF) of laminated shelly, bluish grey silt and silty sand in the lower part and
a lateral accretion complex (LA, LA(DA)) in the upper. The latter comprises a lower alternate bar (LA(DA)), consisting of fine to medium sand,
pebbly sand and fine gravel with steep, northeast-dipping planar cross-bedding, and an upper true point-bar (LA) of gravels and laminated silty sand
layers with gentle, east-dipping reactivation surfaces. 89/12: dip/inclination.
 C. Gao et al. / Sedimentary Geology 202 (2007) 193210
Fig. 8. Ice-wedge cast (upper part) with an estimated total depth of
over 2 m in Section 7. The double-head arrow at the top indicates the
surface depression. Dashed lines are small tension faults. The slumped
silty sand layers within the wedge structure are indicated by white
arrows. Shovel is 1 m long.
northward-flowing direction, similar to that of the
present-day river (Figs. 2 and 4a).
Lithofacies unit L-3 consists of overbank fines (FF)
and lateral accretion or point-bar deposits (LA) as well
as gravelly bar and bedform deposits (GB) (Table 4).
The frequency of overbank fines suggests floodplain
development along the river, which implies that, unlike
braided rivers with multiple channels that are mobile
and undergo frequent shifts, this river probably had
fewer but more stable channels at this time. These
channels would have been defined in the valley by
stable areas where frequent overbank flooding took
place, leading to the accretion of argillaceous floodplain
alluvium.
The overbank fines include overbank flood deposits
consisting of predominantly inorganic silty fines,
abandoned channel infills and backswamp organic
deposits (Figs. 57). As illustrated by Fig. 6, the
sedimentary features of the channel infill sequence
suggest an initial emplacement of sand and silt after the
abandonment of the channel probably with persistent
flow, and subsequent accumulation of organic-enriched
201
fines resulting from the established vegetation in the
vicinity, followed by later-phase deposition of silt and
sand. The increased grain size towards the upper
sequence is also seen in the abandoned meander bend,
suggesting increased flooding activities. The floods not
only washed coarse-grained sands but also molluscan
shells into the ponds.
The cross-stratified pebbly sand and fine gravel sets
are interpreted in the context of lateral accretion or
point-bar deposition (LA) (Figs. 5 and 7). Point bars,
the characteristic macroform for a meandering river,
generally develop on the accretionary bank of a
sinuous river and their internal reactivation surfaces
dip obliquely to the stream, forming epsilon cross-
bedding (Allen, 1963). Because of the coarse nature of
the sediments, small sedimentary structures suitable for
flow direction measurement were not available. The
abandoned meander bend, indicative of the presence of
a former sinuous river at the study site, runs to the
north at its apex (Fig. 2). For a river with intermediate
to high sinuosity, the channel stream, as exemplified
by the present-day Great Ouse, tends to flow parallel
or sub-parallel to the river direction at the apexes of
meanders and threads away from there (Fig. 2).
Because of the southnorth aligned valley, the river
that left the meander bend should have flowed
northwards as evidenced by both the present river
and rivers that emplaced the adjacent lithofacies units
L-2 and L-4 in a persistent northwestnorth direction
(Figs. 2 and 4). Based on these observations, a
northward-flowing direction is estimated for the
channel flow of the river at the apexes of the
meanders. In such a river, the reactivation surfaces of
point bars in areas near these locations are expected to
dip at a high angle to the north. The east dip of the
macroform bedding planes (89, n = 5) is transverse to
this direction, thus providing palaeocurrent evidence
for a lateral accretion interpretation.
The change of dip direction from northeast to east
upwards through the sequence was probably related to
gradual development of channel sinuosity at the study
site (Fig. 7). Thus, the lower cross-bedded set with steep
and simple bedding planes can be interpreted as
downstream accretion of an alternate bar (LA(DA)) in
a low- to medium-sinuosity river (e.g., McCabe, 1977;
Crowley, 1983; Miall, 1985, 1996). With increased
channel sinuosity, true point bars (LA) developed with
internal 3rd-order reactivation surfaces indicated by the
partially eroded overbank silty sand layers (Fig. 7).
It is not uncommon for a braided river to consist of a
large number of small sinuous-flow threads, separated
by mid-channel bars, within a large channel. Because
202 C. Gao et al. / Sedimentary Geology 202 (2007) 193210

the channels and bars are mobile and frequently shift to it are interpreted here as derived from former
during annual, seasonal or even diurnal flooding (e.g., epigenetic ice wedges (Figs. 3, 5, 8).
Vandenberghe and Woo, 2002), development of large
point bars as macroforms is present but atypical, and the 4. Palaeontology
common macroforms in a gravel-bed braided river are,
as discussed earlier, longitudinal gravel bars and sheets, The fossiliferous L-3 has been sampled for macro-
characterised by clast-supported gravels with massive floral, pollen and molluscan analyses (Figs. 3 and 6).
or crude horizontal stratification (Rust, 1972; Bryant, With the exception of the Holocene alluvium L-5, all
1983; Miall, 1985, 1996; Gibbard, 1994; Lewis et al., other lithofacies units are barren of fossils.
2001). Recent observations confirmed the presence of
lateral accretion deposits in large braided rivers with 4.1. Plant macroscopic fossils
sand or sandy-gravel bedload (Best et al., 2003; Lunt
and Bridge, 2004). This is further complicated by the One bulk sample M1 of 5 kg from the central organic
fact that mid-channel bars may develop from chute- silt of the abandoned channel infill at Section 1 was
cutoff point bars (Ashmore, 1991, Lunt and Bridge, washed through a sieve with a mesh size of 0.5 mm for
2004). As such, caution must be taken when facies plant macroscopic analysis (Figs. 3 and 6). The sample
models are constructed. In our study, multiple lines of contains the following plant remains: Menyanthes
evidence were used in facies analysis. For example, trifoliata, Filipendula ulmaria, Hippuris vulgaris,
interpretation for a floodplain environment was based Chara spp., Apiaceae, Potamogeton spp., Thalictrum
on not only characteristic sedimentary features such as flavum, Ranunculus subg. Ranunculus and Carex spp.
overbank and lateral deposits and channels but also, as Plant remains are good indicators of the local
discussed later, palaeontological evidence. environment because of their close connection to site-
Taken together, all the evidence available for specific conditions (Godwin, 1975). The macroscopic
lithofacies unit L-3 points to a gravel to sand-bed plant fossils as a whole suggest a damp, floodplain
meandering river with point bars that migrated across environment with standing water, as indicated by marsh
the valley floor, developing a large floodplain. Direct and aquatic plant species such as Thalictrum vulgaris,
evidence for river sinuosity comes from the preserved Ranunculus subg. Ranunculus, Potamogeton, M. trifo-
meander bend illustrated in Fig. 2. As the river migrated liata, H. vulgaris and Chara. The climate was probably
or shifted laterally, some meander bends were aban-
doned on the floodplain and gradually silted up. On the Table 5
other hand, the frequent longitudinal gravel bars and Results (%) from pollen analysis of the silty channel infill in lithofacies
bedforms (GB) in the sequence, typical of braided river unit L-3 at Section 1
deposition, imply that a wandering river, classified Sample No. P1 P2
between a single-channelled meandering and a multiple- Depth 910 cm 1112 cm
channelled braided river, could be responsible for the Betula 35.7 35.6
deposition of L-3 (Church, 1983; Brierley and Hickin, Pinus 3.6 2.6
Poaceae 35.7 33.3
1991; Miall, 1996).
Cyperaceae 5.9 7.6
Similar to L-1 and L-2, lithofacies unit L-4 is Compositae Tub. 0.7 0.7
dominated by gravelly bar and bedform deposits (GB) Caryophyllaceae 1.0
formed in a braided river with low sinuosity (Table 4). Chenopodiceae 0.3 0.3
However, the presence of many gravel-filled channels Filipendula 6.6 10.2
Ranunculus 0.3
suggests that the river was restrained to some degree
Rumex 0.7
probably by established topographic levels (cf. Wil- Thalictrum 0.3
liams and Rust, 1969; Rust, 1972). Some of the Umbelliferae 0.7
topographically raised areas stood for a prolonged Filicales 7.9 8.9
period before reworking by the stream. In these newly Pteridium 1.3
Myriophyllum 1.3
emerged areas, minimal vegetation and hence a limited
Nymphaea 0.3
capacity for trapping drifting snow would have Potamogeton 1.0
provided favourable conditions for ice-wedge growth Typha latifolia 1.0
under permafrost conditions (Mackay, 1993; Mackay Sum trees 39.3 38.3
and Burn, 2002; Gao, 2005). The ice-wedge casts that Sum herbs 51.5 52.8
Sum lower plants 9.2 8.9
all terminate at the ground surface and hence are related
 C. Gao et al. / Sedimentary Geology 202 (2007) 193210 203

Table 6
Molluscan analysis of overbank fines in lithofacies unit L-3
Sample No. S1 S2 S3 S4 S5 S6 S7 S8
(Original No) RX020101-I(1) RX020101-I(2) RX020101-II RX020101-III RX020203 RX020202 RX020201 RX030101
Sample weights (kg) 2.5 2 3 2.5 2 2 2 2
Freshwater species
Valvata cristata Mler 169 210 107 36 37
Valvata piscinalis (Mler) 67 99 95 8 1 1 2 119
Bithynia tentaculata (L.) 116 243 113 29 142
Shells opercula 136 184 167 68
Physa fontinalis (L.) 1 1
Lymnaea truncatula (Mler) 3 4 1 1 3 11 4 8
Lymnaea palustris (Mler) 4 5 1 1 2 1 3
Lymnaea stagnalis (L.) 1 6 1 2 7
Lymnaea peregra (Mler) 4 4 3 2 21 35 50 54
Anisus leucostoma (Millet) 1 1 1 5 2 5 1
Bathyomphalus 9 12 7 2 1 1 2
contortus (L.)
Gyraulus laevis (Alder) 8 7 2 2 5 59
Armiger crista (L.) 32 74 32 5 15 22 14 91
Hippeutis complanatus (L.) 6 7 3 1 6 3 4 9
Ancylus fluviatilis Mler 1
Sphaerium corneum (L.) 3 1 2 2
Sphaerium lacustre (Mler) 4 3
Pisidium amnicum (Mler) 1
Pisidium casertanum (Poli) 8 34 2 8 9 27
Pisidium obtusale (Lamarck) 10 27 3 4 1 3 4 30
Pisidium milium Held 38 118 37 8 16 4 7 130
Pisidium subtruncatum 36 125 53 11 7 6 25 150
Malm
Pisidium nitidum Jenyns 92 366 113 19 32 27 69 456
Pisidium spp. 1 1
Subtotal 741 1527 741 196 118 127 210 1324
Land snails
Carychium minimum Mler 6 6 2 9
Succinea oblonga 4 6 1
Draparnaud
cf. Oxyloma pfeifferi 5 7 5 2 111 20 21 20
(Rossmsler)
Cochlicopa lubrica (Mler) 2 20 2 2 3
Vertigo pygmaea 1
(Draparnaud)
Pupilla muscorum (L.) 1 2 1 33 4 12 5
Vallonia costata (Mler) 2 12
Vallonia pulchella (Mler) 1 5 7 5 18 20
Vallonia excentrica Sterki 1
Vallonia pulchella/excentrica 3 3 1 3 61 28 32 17
Punctum pygmaeum 1 55 1 4
Draparnaud
Vitrina pellucida (Mler) 1
Deroceras/Limax 1
Euconulus fulvus (Mler) 1 1
Helicopsis striata (Mler) 1
Trichia hispida (L.) 1 2
Subtotal 10 31 12 6 291 63 93 95
Total 751 1558 753 202 409 190 303 1419
Freshwater species 99% 98% 98% 97% 29% 67% 69% 93%
Land snails 1% 2% 2% 3% 71% 33% 31% 7%
204 C. Gao et al. / Sedimentary Geology 202 (2007) 193210
not a stadial type because of the presence of the warmth-
demanding plant Filipendula (cf. Lewis et al., 2001).
4.2. Pollen
Two samples from the central organic silt of the
abandoned channel infill at Section 1 contain
sufficient pollen grains for analysis (Figs. 3 and 6,
Table 5). The pollen assemblage, dominated by Betula
(36%) and Poaceae (3336%), indicates a mixed
vegetation of birch woodland and grassland in the
surrounding area. The presence of birch woodland in
the region suggests typical interstadial conditions that
prevailed in southern England in the early part of the
Devensian Lateglacial substage (Godwin, 1975; Hunt
and Birks, 1982).
4.3. Mollusca
Multiple samples (S1 to S8) have been taken for
molluscan analysis (Figs. 3 and 6). The samples were
washed through 1 and 0.5 mm mesh sieves and the
shells extracted with the help of a desk magnifier.
Except for S5, which contains primarily land snails,
the samples are dominated by freshwater species
(Table 6). In accordance with their ecological
characteristics and tolerance (Sparks, 1961), the
freshwater species are classified into slum, catholic,
ditch and moving-water groups (Fig. 9). The species in
the slum group, Lymnaea truncatula, Anisus leucos-
toma, Sphaerium lacustre, Pisidium casertanum and
Pisidium obtusale, can tolerate poorly oxygenated
conditions. In contrast, the moving-water taxa Valvata
piscinalis, Bithynia tentaculata, Physa fontinalis,
Lymnaea stagnalis, Ancylus fluviatilis and Pisidium
amnicum require well-oxygenated habitats, such as
found in streams and large ponds in southern England
Fig. 9. Percentages of aquatic molluscan taxa with different habitats.
today. Valvata cristata is the only representative of the
ditch group that requires habitats in clean, slowly
moving water. The remaining species are catholic,
living in a wide range of habitats, except the slum
water conditions.
The abundance of moving-water taxa in samples S1
to S4 at Section 1 is striking, suggesting habitats in a
large pond (Fig. 9). The moving-water species decline
slightly in frequency in S2 but increase in S3, reaching
a maximum in S4 (Fig. 9). The frequency of the slum
shells, on the other hand, shows the opposite trend.
This probably reflects the evolution of the habitat from
an initial oxygenated phase, towards impoverished
oxygen in the pond during the interim phase while the
later phase saw improved conditions. A. fluviatilis is
only present in the basal sample S1, which suggests
that in-channel flow was persistent immediately
following the abandonment of the channel, since this
species normally lives in a habitat of freely moving
water with strong currents. Such a change was
probably a response to local factors such as the
gradual establishment of vegetation along the aban-
doned channel and the later-phase reduction in
vegetation resulting from increased flood erosion,
consistent with the sedimentary evidence discussed
previously.
Samples from other localities show increased fre-
quency of land snails and reduced moving-water species
(Table 6, Fig. 9). In sample S5, the fauna is dominated by
terrestrial taxa, suggesting a distal floodplain environ-
ment. Among the terrestrial snails, Vertigo pygmaea,
Pupilla muscorum and Vallonia costata are xerophilous
(arid environment) species living typically in open
grassland. The other terrestrial snails are hydrophilous
species such as would be expected at the margins of a
river floodplain in the British lowland today. Among
them, Oxyloma pfeifferi and Carychium minimum are
typical species of marshy habitats. The shells from these
various habitats were washed into the ponds and swamps
on the floodplain during flood events.
The climate indicated by the molluscan fauna is an
interstadial type. Many taxa in the fauna are frequently
found in the British Isles in interstadial records, but are
unknown from the stadials of the Devensian (Holyoak,
1982). These include V. cristata, B. tentaculata, C.
minimum, P. fontinalis, Bathyomphalus contortus, Hip-
peutis complanatus, Cochlicopa lubrica, Punctum
pygmaeum and Euconulus fulvus. Stadial molluscan
faunal assemblages recorded in southern England often
contain Columella columella, a species living today
above the tree-limit in the Alps and the semi-tundra of
northern Europe (Gao et al., 1998; Briant et al., 2004).
 C. Gao et al. / Sedimentary Geology 202 (2007) 193210 205

5. Dating the Lateglacial Interstadial in line with lithostratigra-

A bulk sample R1 from the central organic silt of the
abandoned-channel infill at Section 1 (Figs. 3 and 6)
was washed through a 0.5 mm mesh sieve. Collected
plant twigs greater than 0.5 mm were radiocarbon-dated
at Godwin Laboratory, University of Cambridge,
yielding an uncalibrated age of 14,050 80 14C BP
(laboratory code: Q-2937). The age is based on the
Libby half-life for 14C isotope of 5568 years and the
zero year of 1950 AD. The calibrated age ranges from
16,320 to 17,130 Cal BP with a medium probable age of
16,750 Cal BP using the computer program CALIB Rev
5.01 (Stuiver and Reimer, 1993; Reimer et al., 2004).
The dating suggests an age slightly earlier than the
Lateglacial Interstadial beginning at ca. 15,200 Cal
BP (Table 1). However, the presence of aquatic plants
in the sample (see above) may have caused hard-
water contamination, resulting in an older date than
the true age of the sample (Day, 1996). As will be
discussed below, the sequence can be correlated with
phy and palaeontological analyses.
6. Discussion
The 5th-order lower bounding surfaces that underlie
the lithofacies units L-2 to L-4 indicate long-term fluvial
processes (Miall, 1993). The sedimentary features of
these units suggest changing river style from braided to
wandering/meandering and back to braided in planform
(Fig. 10). There is no evidence for a significant drop in
base level in the Fenland basin, triggering basin-wide
erosion during the Late Devensian (Gao, 1997). On the
contrary, sea level was slowly rising to reach the present-
day level in the Holocene. In light of the palaeontolo-
gical evidence and radiocarbon dating, the changing
river style and erosional lower bounding surfaces are
interpreted as fluvial response to climate change during
the Late Devensian substage of Lateglacial.
The change of river planform from braided to
wandering/meandering coincided with, as indicated by

Fig. 10. Cartoon diagram showing the proposed facies models. Late Dimlington Stadial: braided river with developed topographic levels (lithofacies
unit L-2). Ice wedges developed, as recorded at Buckden, about 15 km downstream from Roxton (Gao, 1997). Windermere Interstadial: wandering/
meandering river characterised by channel lateral accretion and an established floodplain (L-3). Younger Dryas Chronozone: braided river with well-
defined topographic levels (L-4) at which ice wedges developed. Early Holocene: anastomosing river at ca. 11,5009500 Cal BP.
206 C. Gao et al. / Sedimentary Geology 202 (2007) 193210
the palaeontological analyses, temperature amelioration
during the time span represented by L-3. Improved
vegetation and subdued climatic severity probably
curtailed surface erosion and reduced nival discharge
in spring, forcing the river to adjust its planform in
response to the new conditions. Based on the palaeo-
botanical analysis, lithofacies unit L-3 is correlated to
the Lateglacial Windermere Interstadial at ca. 15,200 to
12,700 Cal BP when birch began to colonise the British
Isles (Table 1) (Godwin, 1975; Hunt and Birks, 1982).
The Lateglacial Interstadial is thought to have had a
similar thermal maximum to the present-day summer
temperature on the basis of beetle fauna in the British
Isles (Coope, 1977; Atkinson et al., 1987). As discussed
previously, the radiocarbon date of 14,050 80 14C BP
(16,750 Cal BP) is marginally too old, resulting
probably from hard-water error from the aquatic plants
in the sample dated. For instance, early Holocene
samples frequently containing the aquatic plant Pota-
mogeton have yielded radiocarbon dates 1500 to
2000 years older than the true age (Day, 1996).
The erosional lower bounding surface of lithofacies
unit 3 can be explained by lateral migration of the
channel typical of a meandering river. In this process,
lateral channel movement or shift frequently eroded the
existing floodplain, forming an irregular surface con-
sisting of the former channel floors (Gao, 2004). Since
the wandering/meandering river was a result of
Lateglacial warming, such an erosional lower bounding
surface can still be interpreted in the context of regional
climate change.
Palaeontological information is unavailable from the
non-fossiliferous lithofacies units L-1 and L-2. How-
ever, palaeontological analyses of organic silt and sand
lenses, recorded elsewhere in southern England, gen-
erally indicate treeless, open conditions in Devensian
fluvial gravels with facies assemblages (GB) character-
istic of L-1 and L-2 (cf. Bryant, 1983; Gibbard, 1994;
Gao, 1997; Gao et al., 1998; Briant et al., 2004). Based
on the sequence events of the Late Devensian, L-2 is
correlated with the Devensian Dimlington Stadial
beginning around 26,000 14C BP (Table 1). Although
formed in an earlier time, the age of L-1 is difficult to
determine due to substantial erosion after its deposition.
Fossiliferous deposits dating from Dimlington Stadial
have been reported from the Great Ouse at Buckden,
about 15 km downstream from Roxton (Fig. 1) (Gao,
1997). There, beneath the lowest terrace, an organic
sandy silt bed occurs in a gravelly deposit similar to L-2
in facies assemblage and stratigraphical occurrence. The
organic bed, which is radiocarbon-dated at 16,200 530
14
C BP or 19,410 Cal BP (Q-2957), contains a tundra-
type vegetation characterised by shrub willows Salix
herbacea and Salix polaris under cool to cold condi-
tions. Climatic severity is also indicated by the presence
of ice-wedge casts in the deposits at that site (Gao,
1997).
The change back from a wandering/meandering to a
braided river in lithofacies unit L-4 is probably related to
climate deterioration and a reduction in vegetation cover
during the subsequent Loch Lomond Stadial or Younger
Dryas Chronozone at 12,700 to 11,500 Cal BP (Fig. 10,
Table 1). Strong incision occurred before the emplace-
ment of L-4, consistent with observations elsewhere in
southern England (Rose et al., 1980; Collins et al., 1996;
Lewis et al., 2001). This incision probably resulted from
increased nival peak discharge whereas sediment supply
was limited due to the existence of resistant river banks,
still covered by cold-resistant vegetation of birch and
grass that required time to respond to the changing
climate.
The existence of ice-wedge casts in L-4 indicates
permafrost development under severe conditions in this
area (Fig. 8). Using ice-wedge casts to estimate palaeo-
temperatures is difficult because of many uncertainties.
They include the duration of cold phases, history of ice-
wedge growth and decay, and the local conditions of
winter temperature, vegetation cover and snow depth,
all of which are critical to ice-wedge growth (Mackay,
1993; Mackay and Burn, 2002; Murton and Kolstrup,
2003; Gao, 2005). A mean annual air temperature
(MAAT) no higher than 6 to 8 C is often considered
the prerequisite for ice-wedge growth in sands and
gravels (Pw, 1966a, b; Romanovskij, 1985; Van
Huissteden et al., 2003). However, active ice wedges
may occur at a MAAT of 3 to 4 C (Burn, 1990). As
such, the palaeo-temperature can probably be estimated
no higher than 3 to 4 C, consistent with the estimate
from Coleoptera assemblages from southern England
(Atkinson et al., 1987).
The end of the Younger Dryas Chronozone saw a
rapid thermal amelioration in climate and the Great
Ouse became stabilised in the early Holocene (Gao,
1997). There is evidence to suggest that the Great Ouse
developed an anastomosing planform prior to a single-
thread meandering pattern (Fig. 10). At Buckden, a
stabilised area is indicated by many alder (Alnus sp.) tree
roots dated at 8645 110 14C BP (9660 Cal BP) (Q-
2916); further downstream at St. Ives, palaeosol
horizons occur in stabilised areas between small
channels filled by early Holocene deposits dated at
9970 55 14C BP (11,420 Cal BP) (Q-2939) (Gao,
1997). It is probable that the river adopted the relict
channels from the braided river of the previous cold
 C. Gao et al. / Sedimentary Geology 202 (2007) 193210
phase, forming an anastomosing morphology, charac-
terised by interconnected streams separated by small
islands (Fig. 10) (Brown et al., 1994; Gibbard and
Lewin, 2002). The fine-grained deposits of lithofacies
unit L-5 are typical of the middle to late Holocene
sequence, resulting from intense vertical accretion of
argillaceous alluvium on floodplains beginning at ca.
5000 BP (Brown et al., 1994; Brown, 1995). The lack of
lateral or point-bar accretion in the sequence strongly
suggests the river had a single-thread meandering
planform, similar to the present-day River Great Ouse.
Regional fluvial responses to Lateglacial climatic
variability can hardly be established from a single-site
investigation such as this study. However, field
observations in the Gipping, Kennet and upper Thames
valleys all suggest changing fluvial styles, comparable
to this site (Rose et al., 1980; Collins et al., 1996; Lewis
et al., 2001). There is evidence that meandering
channels also developed in the Nene valley during this
episode (Brown et al., 1994). This, as a whole, suggests
that many lowland gravel-bed rivers responded to the
Lateglacial climatic oscillations in a similar way to the
sand-bed dominated rivers in northwestern and central
mainland Europe (Vandenberghe et al., 1994; Starkel,
1995; Mol et al., 2000). The rare reports of Lateglacial
fluvial deposits in southern England can thus be
explained favourably in the context of post-depositional
erosion and infrequent preservation of the records.
It is noticeable that facies models for sinuous fluvial
regimes during the Lateglacial Interstadial from else-
where in southern England are based on the existence
of increased accumulation of overbank fines and hence
established floodplain complexes (cf. Rose, 1995).
Sedimentary details of the channels are lacking. While
our observations have confirmed the development of
overbank fines, the recognition of lateral or point-bar
deposition indicates active lateral channel migration
within the valley. This suggests that the floodplain
contained deposits from both overbank accretion and
lateral aggradation. Such a floodplain reflects a normal,
actively meandering river, quite different from the
present-day Great Ouse and other lowland rivers,
which are inactive and lacking, in general, point-bar
forming processes. This interpretation is consistent
with the frequent gravel to sand macroform elements in
the sediments spanning the Windermere Interstadial
(Fig. 10).
The existence of a gravel to sand-bed wandering/
meandering river during the Windermere Interstadial
has two important implications for interpretation of the
fluvial architecture. First, the erosive power of flood-
plain flow during floods was small due to a low water
207
depth. Consequently, the accretion of argillaceous
alluvium on the floodplain could reach a considerable
thickness (up to 5 m or more for inactive meandering
rivers, such as the present-day Great Ouse) without
lateral channel erosion (Gao, 1997). In an active
meandering system, however, lateral channel move-
ments often cause strong erosion of existing floodplains,
as shown by dislocated silt or peat clasts present in
point-bar gravels (Lewis et al., 2001; Gao, 2004),
resulting in underdeveloped overbank sequences. Tak-
ing into consideration the relatively short duration of the
Lateglacial Interstadial, it is expected that only limited
overbank fines accumulated, insufficient to withstand
post-depositional removal by erosion. The opportunities
for preservation also may be related to the lateral extent
of deposited overbank fines, i.e., higher preservation
potential for a wider floodplain and lower for a smaller
one. Because the Great Ouse is confined in a valley
incised into the Lowestoft Formation till, the palaeo-
floodplain is unlikely to have been any larger in size
than the present one. The medium sinuosity for the
Lateglacial Great Ouse, as Fig. 10 depicts, strongly
suggests a limited floodplain in the form of a narrow,
near-channel zone of sandy argillaceous alluvium,
unfavourable for preservation. These factors together
probably account for the rare preservation of the fine-
textured overbank alluvium of this episode.
Second, the fine-grained overbank deposits often
contain plant debris that can be radiocarbon-dated. In
contrast, the point-bar sands and gravels are normally
barren of organics and their ages are often inferred.
Because of their erosional contact and different facies
characteristics (see Fig. 7), these two suites of
contemporaneous sediments easily may be mis-inter-
preted as independent climato-stratigraphical units. This
is probably true of some reported Lateglacial sites, for
example, at Raunds and Ashton Keynes in the Nene and
upper Thames valleys, respectively (Brown et al., 1994;
Lewis et al., 2001).
7. Conclusions
The lithofacies units defined by thoroughgoing
lower bounding surfaces in the Lateglacial succession
at Roxton represent long-term fluvial processes. Based
on facies analysis and supported by palaeontology and
radiocarbon dating, local facies models show a
changing fluvial style from a braided to a wandering/
meandering river in response to climatic warming
during the Lateglacial Interstadial, followed by a
change back to a braided river in response to climatic
deterioration during the Younger Dryas Chronozone.
208 C. Gao et al. / Sedimentary Geology 202 (2007) 193210
Although rare, the evidence available elsewhere in
southern England indicates a similar fluvial pattern,
suggesting a regional fluvial response to the Lateglacial
climate variability.
The interstadial floodplain consisted of both over-
bank silty fines and channel lateral or point-bar sands
and gravels. The two suites of contemporaneous
sediments in the geological succession easily may be
mis-interpreted, however, as independent climato-strati-
graphical units due to their cross-cutting relationship
and marked difference in facies assemblage. Lateral
channel movement in such a system also indicates that
the floodplain underwent constant erosion, resulting in
underdeveloped overbank sequences, insufficient to
withstand post-depositional removal by erosion. This
probably accounts for the rare preservation of fine-
textured overbank deposits of the Lateglacial Intersta-
dial in southern England.
Acknowledgements
We thank Dr. R. Switsur for undertaking the
radiocarbon dating and Dr. M. Field for assistance in
analysis of the plant macroscopic remains. Redland not
only gave permission for access to its gravel pits but also
provided borehole data. One of the authors (CG)
received a studentship (19921996) from the Sino-
British Friendship Scholarship Scheme (SBFSS). Addi-
tional financial assistance was provided by Clare Hall
college of the University of Cambridge, and Cambridge
Philosophical Society. Comments from two anonymous
reviewers, Drs. S.B. Marriott and C.R. Fielding, the
Guest Editor and the Editor greatly improved the quality
of the manuscript. Dr. Marriott also provided additional
editorial assistance. For their support and assistance, we
are very grateful.
References
Allen, J.R.L., 1963. The classification of cross-stratified units, with
notes on their origin. Sedimentology 2, 93114.
Ashmore, P.E., 1991. How do gravel-bed rivers braid? Canadian
Journal of Earth Sciences 28, 326341.
Atkinson, T.C., Briffa, K.R., Coope, G.R., 1987. Seasonal tempera-
tures in Britain during the past 22 000 years, reconstructed using
beetle remains. Nature 325, 587592.
Best, J.L., Ashworth, P.J., Bristow, C., Roden, J., 2003. Three-
dimensional sedimentary architecture of a large, midchannel sand
braid bar, Jamuna River, Bangladesh. Journal of Sedimentary
Research 73, 516530.
Bjrck, S., Walker, M.J.C., Cwynar, L.C., Johnsen, S., Knudsen,
K.-L., Lowe, J.J., Wohlfarth, B., INTIMATE Members, 1998.
An event stratigraphy for the Last Termination in the North
Atlantic region based on the Greenland Ice-core record: a
proposal by the INTIMATE group. Journal of Quaternary
Science 13, 283292.
Bond, G., Broecker, W., Johnsen, S., McManus, J., Labeyrie, L.,
Jouzel, J., Bonani, G., 1993. Correlations between climate records
from North Atlantic sediments and Greenland ice. Nature 365,
143147.
Bowen, D.Q., Rose, J., McCabe, A.M., Sutherland, D.G., 1986.
Correlation of Quaternary glaciations in England, Ireland, Scotland
and Wales. Quaternary Science Reviews 5, 299340.
Bowen, D.Q., Hughes, S., Sykes, G.A., Miller, G.H., 1989. Landsea
correlations in the Pleistocene based on isoleucine epimerization in
non-marine molluscs. Nature 340, 4951.
Briant, R.M., Coope, G.R., Preece, R.C., Keen, D.H., Boreham, S.,
Griffiths, H.I., Seddon, M.B., Gibbard, P.L., 2004. Fluvial system
response to Late Devensian (Weichselian) aridity, Baston,
Lincolnshire, England. Journal of Quaternary Science 19,
479495.
Brierley, G.J., Hickin, E.J., 1991. Channel planform as a non-
controlling factor in fluvial sedimentology: the case of the
Squamish River floodplain, British Columbia. Sedimentary
Geology 75, 6783.
Brown, A.G., 1995. Holocene channel and floodplain change: a UK
perspective. In: Gurnell, A., Petts, G. (Eds.), Changing River
Channels. Wiley, Chichester, pp. 4364.
Brown, A.G., Keough, M.K., Rice, R.J., 1994. Floodplain evolution
in the East Midlands, United Kingdom: the Lateglacial and
Flandrian alluvial record from the Soar and Nene valleys.
Philosophical Transactions of the Royal Society of London. A
348, 261293.
Bryant, I.D., 1983. Facies sequences associated with some braided
river deposits of Late Pleistocene age from southern Britain. In:
Collinson, J.D., Lewin, J. (Eds.), Modern and Ancient Fluvial
Systems. Special Publication of the International Association of
Sedimentologists, vol. 6, pp. 267275.
Burn, C.R., 1990. Implications for palaeoenvironmental reconstruction
of recent ice-wedge development at Mayo, Yukon Territory.
Permafrost and Periglacial Processes 1, 314.
Church, M., 1983. Pattern of instability in a wandering gravel bed
channel. In: Collinson, J.D., Lewin, J. (Eds.), Modern and Ancient
Fluvial Systems. Special Publication of the International Associa-
tion of Sedimentologists, vol. 6, pp. 169180.
Collins, P.E.F., Fenwick, I.M., Keith-Lucas, D.M., Worsley, P., 1996.
Late Devensian river and floodplain dynamics and related
environmental change in northwest Europe, with particular
reference to a site at Woolhampton, Berkshire, England. Journal
of Quaternary Science 11, 357375.
Coope, G.R., 1977. Fossil coleoptera assemblages as sensitive
indicators of climatic changes during the Devensian (last) cold
stage. Philosophical Transactions of the Royal Society of London.
B 280, 313340.
Crowley, K.D., 1983. Large-scale bed configurations (macroforms),
Platte River basin, Colorado and Nebraska: primary structures and
formative processes. Geological Society of America Bulletin 94,
117133.
Dansgaard, W., Johnsen, S.J., Clausen, H.B., Dahl-Jensen, D.,
Gundestrup, N.S., Hammer, C.U., Hvidberg, C.S., Steffensen, J.P.,
Sveinbjornsdttir, S.E., Jouzel, J., Bond, G., 1993. Evidence for
general instability of past climate from a 250-kyr ice-core record.
Nature 364, 218220.
Day, P., 1996. Devensian Late-glacial and early Flandrian environ-
mental history of the Vale of Pickering, Yorkshire, England.
Journal of Quaternary Science 11, 924.
 C. Gao et al. / Sedimentary Geology 202 (2007) 193210
Edmonds, E.A., Dinham, C.H., 1965. The geology of the country
around Huntingdon and Biggleswade. Memoir of Geological
Survey, Sheet 187 and 204. Her Majesty's Stationary Office,
London.
Gao, C., 1997. Sedimentology and stratigraphy of the River Great
Ouse, southern England. Ph.D thesis. University of Cambridge,
Cambridge.
Gao, C., 2004. Sedimentary facies changes and climatictectonic
controls in a foreland basin, the Urumqi River, Tian Shan,
Northwest China. Sedimentary Geology 169, 2946.
Gao, C., 2005. Ice-wedge casts in Late Wisconsinan glaciofluvial
deposits, southern Ontario, Canada. Canadian Journal of Earth
Sciences 42, 21172126.
Gao, C., Coope, G.R., Keen, D.H., Pettit, M.E., 1998. Middle Devensian
deposits of the Ivel valley at Sandy, Bedfordshire, England.
Proceedings of the Geologists' Association 109, 127137.
Gao, C., Keen, D.H., Boreham, S., Coope, G.R., Pettit, M.E.,
Stuart, A.J., Gibbard, P.L., 2000. Last Interglacial and
Devensian deposits of the River Great Ouse at Woolpack
Farm, Fenstanton, Cambridgeshire, UK. Quaternary Science
Reviews 19, 787810.
Gibbard, P.L., 1994. Pleistocene History of the Lower Thames valley.
Cambridge University Press, Cambridge. 229 pp.
Gibbard, P.L., Lewin, J., 2002. Climate and related controls on
interglacial fluvial sedimentation in lowland Britain. Sedimentary
Geology 151, 187210.
Godwin, H., 1975. History of British Flora, 2nd edition. Cambridge
University Press, Cambridge. 541 pp.
Holyoak, D.T., 1982. Non-marine mollusca of the Last Glacial period
(Devensian) in Britain. Malacologia 22, 727730.
Horton, A., 1970. The Drift Sequence and Sub-glacial Topography in
Parts of the Ouse and Nene Basins. Institute of Geological
Sciences, Report No. 70/9.
Hunt, T.G., Birks, J.B., 1982. Devensian late-glacial vegetation
history at Sea Mere. Norfolk. Journal of Biogeography 9,
517538.
Lewis, S.G., Maddy, D., Scaife, R.G., 2001. The fluvial system
response to abrupt climate change during the last cold stage: the
Upper Pleistocene River Thames fluvial succession at Ashton
Keynes, UK. Global and Planetary Change 28, 341459.
Lowe, J.J., Birks, H.H., Brooks, S.J., Coope, G.R., Harkness, D.D.,
Mayle, F.E., Sheldrick, C., Turney, C.S.M., Walker, M.J.C., 1999.
The chronology of palaeoenvironmental changes during the Last
GlacialHolocene transition: towards an event stratigraphy for the
British Isles. Journal of the Geological Society (London) 156,
397410.
Lunt, I.A., Bridge, J.S., 2004. Evolution and deposits of a gravelly
braid bar, Sagavanirktok River, Alaska. Sedimentology 51,
415432.
McCabe, P.J., 1977. Deep distributary channels and giant bedforms in
the Upper Carboniferous of the central Pennines, northern
England. Sedimentology 24, 271291.
Mackay, J.R., 1993. Air temperature, snow cover, creep of frozen
ground, and the time of ice-wedge cracking, western Arctic coast.
Canadian Journal of Earth Sciences 30, 17201729.
Mackay, J.R., Burn, C.R., 2002. The first 20 years (19781979 to
19981999) of ice-wedge growth at the Illisarvik experimental
drained lake site, western Arctic coast. Canadian Journal of Earth
Sciences 39, 95111.
Miall, A.D., 1985. Architectural-element analysis: a new method of
facies analysis applied to fluvial deposits. Earth-Science Reviews
22, 261308.
209
Miall, A.D., 1993. The architecture of fluvialdeltaic sequences in the
Upper Mesaverde Group (Upper Cretaceous), Book Cliffs, Utah.
In: Best, J.L., Bristow, C.S. (Eds.), Braided Rivers. Geological
Society Special Publication, vol. 75, pp. 305332.
Miall, A.D., 1996. The Geology of Fluvial Deposits: Sedimentary
Facies, Basin Analysis and Petroleum Geology. Springer, Berlin.
582 pp.
Mol, J., Vandenberghe, J., Kasse, C., 2000. River response to
variations of periglacial climate in mid-latitude Europe. Geomor-
phology 33, 131148.
Murton, J.B., Kolstrup, E., 2003. Ice-wedge casts as indicator of
palaeotemperatures: precise proxy or wishful thinking? Progress in
Physical Geography 27, 155170.
Pw, T.L., 1966a. Ice wedges in Alaskaclassification, distribu-
tion and climatic significance. Proceedings of the 1st Interna-
tional Permafrost Conference, National Academy of Science-
National Research Council of Canada. Publication, vol. 1287,
pp. 7681.
Pw, T.L., 1966b. Palaeoclimatic significance of fossil ice wedges.
Biuletyn Peryglacjalny 15, 6573.
Reimer, P.J., Baillie, M.G.L., Bard, E., Bayliss, A., Beck, J.W., Bertrand,
C.J.H., Blackwell, P.G., Buck, C.E., Burr, G.S., Cutler, K.B., Damon,
P.E., Edwards, R.L., Fairbanks, R.G., Friedrich, M., Guilderson, T.P.,
Hogg, A.G., Hughen, K.A., Kromer, B., McCormac, F.G., Manning,
S.W., Ramsey, C.B., Reimer, R.W., Remmele, S., Southon, J.R.,
Stuiver, M., Talamo, S., Taylor, F.W., van der Plicht, J., Weyhen-
meyer, C.E., 2004. IntCal04 terrestrial radiocarbon age calibration,
260 ka BP. Radiocarbon 46, 10291058.
Rogerson, R.J., 1987. The terrace of the River Great Ouse. Ph.D thesis.
University of London, Surrey.
Romanovskij, N.N., 1985. Distribution of recently active ice and soil
wedges in the USSR. In: Church, M., Slaymaker, O. (Eds.), Field
and theory: Lectures in Geocryology. University of British
Columbia Press, Vancouver, pp. 154165.
Rose, J., 1989. Stadial type sections in the British Quaternary. In:
Rose, J., Schlchter, C. (Eds.), Quaternary Type Sections;
Imagination or Reality? A.A. Balkema, Rotterdam, pp. 4567.
Rose, J., 1995. Lateglacial and early Holocene river activity in lowland
Britain. In: Frenzel, B., Vandenberghe, J., Kasse, K., Bohncke, S.,
Glser, B. (Eds.), European River Activity and Climatic Change
during the Lateglacial and Early Holocene. G. Fischer, Stuttgart,
pp. 5174.
Rose, J., Turner, C., Coope, G.R., Bryan, M.D., 1980. Channel
changes in a lowland river catchment over the past 13,000 years.
In: Cullingford, R.A., Davidson, D.A., Lewin, J. (Eds.), Time-
scales in Geomorphology. Wiley, Chichester, pp. 159175.
Rust, B.R., 1972. Structure and process in a braided river.
Sedimentology 18, 221246.
Sparks, B.W., 1961. The ecological interpretation of Quaternary non-
marine Mollusca. Proceedings of the Linnological Society of
London 172, 7180.
Starkel, L., 1995. Changes of river channels in Europe during the
Holocene. In: Gurnell, A., Petts, G. (Eds.), Changing River
Channels. Wiley, Chichester, pp. 2742.
Stuiver, M., Reimer, P.J., 1993. Extended 14C database and revised
CALIB radiocarbon calibration program. Radiocarbon 35,
215230.
Van Huissteden, J., Gibbard, P.L., Briant, R.M., 2001. Periglacial
fluvial systems in northwest Europe during marine isotope stages 4
and 3. Quaternary International 79, 7588.
Van Huissteden, J., Vandenberghe, J., Pollard, D., 2003. Palaeotem-
perature reconstructions of the European permafrost zone during
210 C. Gao et al. / Sedimentary Geology 202 (2007) 193210

marine oxygen isotope stage 3 compared with climate model
results. Journal of Quaternary Science 18, 453464.
Vandenberghe, J., Kasse, C., Bohncke, C., Kozarski, S., 1994.
Climate-related river activity at the WeichselianHolocene transi-
tion: a comparative study of the Warta and Maas rivers. Terra Nova
6, 476485.
Vandenberghe, J., Woo, M.K., 2002. Modern and ancient periglacial
river types. Progress in Physical Geography 26, 479506.
Williams, P.F., Rust, B.R., 1969. The sedimentology of a braided river.
Journal of Sedimentary Petrology 39, 649679.
 Available online at www.sciencedirect.com

Sedimentary Geology 202 (2007) 211 229

www.elsevier.com/locate/sedgeo

Late Quaternary climatic evolution of the Arno coastal plain

(Western Tuscany, Italy) from subsurface data
Margherita Aguzzi a , Alessandro Amorosi a,, Maria Luisa Colalongo a ,
Marianna Ricci Lucchi a , Veronica Rossi a , Giovanni Sarti b , Stefano Claudio Vaiani a
a
Dipartimento di Scienze della Terra e Geologico-Ambientali, Universit di Bologna, Via Zamboni 67, 40126 Bologna, Italy
b
Dipartimento di Scienze della Terra, Universit di Pisa, Via Santa Maria 53, 56126 Pisa, Italy

Abstract

A multidisciplinary study of a 105-m-long core was carried out on the Tyrrhenian coast of Tuscany, Western Italy. Detailed
description of sedimentary facies, foraminifer and ostracod assemblages, pollen, and 14C ages is presented in this paper.
Identification in core of two transgressive surfaces (TSs) as the most prominent stratigraphic markers allows subdivision of the
Late Quaternary stratigraphic succession into two transgressiveregressive (TR) sequences, attributed to the last 150 kyr BP.
Sequence boundaries have no unequivocal physical expression in the core.
Detailed pollen analysis documents a direct relationship between vertical facies evolution and climate fluctuations. Coastal to
shallow-marine sediments in the lower part of TR sequences were deposited during the last two interglacial periods (OIS 1 and
5e), under rising sea-level conditions and during the following sea-level highstands. By contrast, alluvial sedimentation (upper part
of TR sequences) took place during periods of sea-level fall and subsequent sea-level lowstands, and was invariably linked to the
onset of glacial periods (OIS 42 and 6, respectively).
This paper presents the first detailed facies documentation of a Late Quaternary incised-valley fill sequence from Italy. About
51 m of Holocene sediments are recorded beneath the present Arno River valley. Early transgression is documented in the lower
part of the incised-valley fill by wave-dominated estuarine facies overlying lowstand fluvial deposits. Late transgression records the
rapid landward migration of a beach-barrier system, followed by the establishment of an open-marine environment. Highstand
sedimentation is represented by a shallowing-upward succession, which reflects progradation of the modern delta/strandplain.
2007 Elsevier B.V. All rights reserved.

Keywords: Late Quaternary stratigraphy; Sea-level change; Incised-valley system; Pollen; Micropalaeontology; Arno River

1. Introduction

Stratigraphic investigation of Late Quaternary depos-

Corresponding author. Tel.: +39 51 2094586; fax: +39 051
2094522.
E-mail addresses: MAguzzi@arpa.emr.it (M. Aguzzi),
alessandro.amorosi@unibo.it (A. Amorosi),
marianna.riccilucchi@unibo.it (M.R. Lucchi),
veronica.rossi4@unibo.it (V. Rossi), sarti@dst.unipi.it (G. Sarti),
stefano.vaiani@unibo.it (S.C. Vaiani).
0037-0738/$ - see front matter 2007 Elsevier B.V. All rights reserved.
doi:10.1016/j.sedgeo.2007.03.004
its has become increasingly popular in the last decade,
because of the unique opportunity to use tectonically
undisturbed and easily datable sediments as a basis for
the definition of realistic and accurate depositional
models (Blum and Trnqvist, 2000; Cattaneo and Steel,
2003). These models may be successfully used for the
prediction of possible scenarios of environmental
212 M. Aguzzi et al. / Sedimentary Geology 202 (2007) 211229
change under rising sea-level conditions, and for the
development of new conceptual models in the field of
sequence stratigraphy (e.g. Saito, 1994; Blum and
Trnqvist, 2000; Plint and Nummedal, 2000; Amorosi
and Colalongo, 2005). Such models can also be applied
to the interpretation of older successions and may serve
to elucidate late Quaternary tectonic activity.
Facies analysis and sequence stratigraphy of Late
Quaternary deposits traditionally deals with detailed
reconstructions of Holocene successions buried in the
uppermost few tens of metres beneath the modern
alluvial and coastal plains (Lowrie and Hamiter, 1995;
Somoza et al., 1998; Amorosi et al., 1999a; Hori et al.,
2002; Amorosi et al., 2003; Overeem et al., 2003; Tanabe
et al., 2003; Hori et al., 2004; Amorosi et al., 2005;
Storms et al., 2005). Although a cyclic stacking pattern of
facies in response to Milankovich-scale cycles has been
observed within pre-Holocene deposits (Massari et al.,
2004), there have been few attempts to depict the detailed
facies architecture of the last two glacialinterglacial
cycles (from OIS 6 onwards). This is mostly due to
problems in the recovery of pre-Holocene data within
Fig. 1. Map of the study area with location of the major coastal plains along the Tyrrhenian coast of Italy. AA: section trace of Fig. 5.
buried successions. Long-cored pollen series from
continental successions in Europe have documented the
climatic evolution of the last 150 ky (Wijmstra, 1969;
Wijmstra and Smit, 1976; Woillard, 1978; Follieri et al.,
1988; De Beaulieu and Reille, 1992; Tzedakis, 1993;
Reille et al., 1998; Tzedakis, 1999), and have been shown
to correlate with oxygen-isotope stratigraphy (see
Tzedakis et al., 1997; Caspers and Freund, 2001; Guiter
et al., 2003; Preusser, 2004). These studies, however,
generally lack detailed facies analysis. Few studies have
investigated in detail the response of coastal and deltaic
depositional systems to the last two Quaternary glacio-
eustatic cycles on the basis of multiple datasets (Amorosi
et al., 1999b; Trnqvist et al., 2000, 2003; Lim and Park,
2003; Amorosi et al., 2004; Konradi et al., 2005;
Hanebuth et al., 2006).
The aim of this paper is to present the first detailed
example of an integrated sedimentological and micro-
palaeontological study of late Quaternary deposits in
the Tyrrhenian area, based on analysis of a 105 m-long
core that was drilled in 2003 in western Tuscany (Italy),
only 200 m from the present shoreline (Fig. 1). Specific
 M. Aguzzi et al. / Sedimentary Geology 202 (2007) 211229
objectives of this paper are: i) to reconstruct the depo-
sitional history of the Tuscan coastal plain in terms of
its response to sea-level fluctuations during the past
150 ky, ii) to document the vertical stacking pattern of
facies within the Holocene incised-valley fill beneath
the Arno River, and iii) to establish a high-resolution
chronology for the studied succession on the basis of
pollen data.
2. OIS 5e deposits along the Tyrrhenian coast
of Italy
The Quaternary evolution of the Tyrrhenian coast of
Italy is difficult to unravel, due to complex superposition
of tectonic uplift, subsidence and eustatic effects
on sedimentation. A review of Late Quaternary tecton-
ics along the Tyrrhenian coast and its influence on
stratigraphic architecture was recently presented by
Ferranti et al. (2006). Apart from a few detailed studies,
however, stratigraphic information on the Late Quater-
nary succession is scarce in comparison with the well-
documented Adriatic coast (Amorosi et al., 1999b;
Cattaneo and Trincardi, 1999; Ridente and Trincardi,
2002; Amorosi et al., 2004). This applies especially to
deposits of the last interglacial.
Early investigations of Late Quaternary deposits
along the Tyrrhenian coast have been carried out in the
Tiber Plain, around the city of Rome, where deposits
attributed to the last interglacial (OIS 5e) have been
encountered at depths ranging between 30 and 80 m
below surface (Bellotti et al., 1994b, 1995; Milli, 1997).
South of Rome, tidal notches referred to OIS 5e have
been recognized at + 7.3 m a.s.l. in outcrops bordering
the Fondi Plain (Fig. 1), whereas lagoonal OIS 5e
deposits have been encountered in subsurface at 6.5 m
(Antonioli et al., 1988).
Close to Naples (Fig. 1), OIS 5e marine deposits have
been identified at 50 m core depth in the Volturno plain
(Romano et al., 1994), on the basis of 230 Th/234U dating
of coral Cladocora coespitosa. Further south, in the
Sarno Plain (Fig. 1), marine deposits attributed to OIS 5
have been recognised at a depth of 25 m below surface
(Barra et al., 1991).
In Tuscany, subsurface investigations carried out on
the Ombrone coastal plain have been restricted to the post-
OIS 5 succession (Carboni et al., 2002; Bellotti et al.,
2004; Biserni et al., 2004). To the north, in the Piombino
alluvial plain, lagoonal deposits assigned to OIS 5e on the
basis of pollen data have been identified at 20 m core
depth (Amorosi et al., 2004). Marine deposits attributed to
OIS 5 on the basis of thermo-luminescence dates (Mauz,
1999) and amino-acid racemization (Hearty and Dai Pra,
213
1987) crop out patchily along the coast between Piombino
and Leghorn, and are overlain by reddish weathered
alluvial deposits (Cortemiglia et al., 1983; Hearty and Dai
Pra, 1987; Mauz, 1999; Sarti et al., 2005).
The Arno coastal plain is located north of Livorno
(Leghorn), in the southern part of the subsiding
Viareggio Basin (Fig. 1). The Viareggio Basin includes
the Versilia Plain and is bounded to the NE by the Pisa
Mts. Its southern boundary, marked by the Livorno
Sillaro tectonic line (Bortolotti, 1966), is coincident with
the foothills of the Leghorn Mountains. The depocentre
of this basin, which is located close to the present
coastline, was filled by up to 2500 m of upper Miocene
Present deposits (Pascucci, 2005). The Viareggio Basin
is the northernmost of a series of extensional basins that
developed on the south-western side of the Northern
Apennines (Tuscany) during the early Messinian. This
basin evolved as a half-graben, with NWSE orienta-
tion, during a rifting phase in the Early Pliocene (Mariani
and Prato, 1988; Argnani et al., 1997).
Aguzzi et al. (2005) established a stratigraphic
framework for the Late Quaternary deposits beneath the
lower Arno River valley. Approximately 20 km north of
the study area, in the Versilia Plain (Fig. 1), 230Th/234U
dating of a 90 m-long core has suggested the attribution
of marine deposits to OIS 5e, at a depth of approxi-
mately 70 m (Antonioli et al., 1999). Unfortunately, no
facies analysis has been performed on this core.
3. Materials and methods
The 105 m-long core (labelled M1) was drilled close to
the village of Tirrenia, along the Tyrrhenian coast of
Tuscany (Fig. 1). Drilling system consisted of a double
corer, with 12 cm core diameter. Core recovery was 95%.
The core was split lengthwise and carefully described in
terms of colour, lithology, grain size, sedimentary struc-
tures and accessory materials (roots, wood fragments,
organic material, peat, mollusc shells). Five AMS 14C
uncalibrated dates (performed at CEDAD, University of
Lecce, Italy) were obtained from wood, organic-rich layers
and mollusc shells, collected in the upper 50 m (Fig. 2).
A total of 179 samples of approximately 150 g were
collected for micropalaeontological analyses. All samples
were i) dried at 60 C, ii) washed with H2O through sieves
of 63 m (240-mesh), and iii) analyzed for ostracods and
foraminifera. Qualitative analyses were performed for all
samples on the N 63 m size fraction. Sixty-six samples
with perfectly preserved foraminiferal specimens within
the N 125 m size fraction were used for counting. These
samples were split into small portions, which included at
least 300 foraminifera. Identification of foraminifera and
214 M. Aguzzi et al. / Sedimentary Geology 202 (2007) 211229

Fig. 2. Stratigraphic column of Core M1, inferred depositional environments, and general micropalaeontological divisions. TS: transgressive surface;
TRS: tidal ravinement surface; WRS: wave ravinement surface; MFS: maximum flooding surface; SB: sequence boundary.
 M. Aguzzi et al. / Sedimentary Geology 202 (2007) 211229
ostracods was based on the original microfossil descrip-
tions, and a series of key papers (Bonaduce et al., 1975;
Jorissen, 1988; Athersuch et al., 1989; Albani and
Serandrei Barbero, 1990; Henderson, 1990; Cimerman
and Langer, 1991; Sgarrella and Moncharmont Zei,
1993; Fiorini and Vaiani, 2001). Ecological information
on species and palaeoenvironmental significance of as-
semblages are mainly provided by Blanc-Vernet (1969),
Bonaduce et al. (1975), Jorissen (1987), Murray (1991),
Pugliese and Stanley (1991), Yassini and Jones (1995),
Debenay et al. (2000), Ruiz et al. (2000), Donnici and
Serandrei Barbero (2002), Smith and Horne (2002),
Amorosi et al. (2004) and Carboni et al. (2004).
Eighty-nine samples were analyzed for pollen.
Sediment samples were dried and treated with HCl
(20%), HF (40%) and NaOH (10%). Pollen concentra-
tion was calculated by addition of tablets with a known
number of Lycopodium spores to a specific weight of
sediment (Stockmarr, 1971). An average of 300 pollen
grains per sample was counted. Only the synthetic
pollen diagrams are shown, in which the pollen sum
includes all pollens with the exception of aquatics.
Arboreal pollen types (AP) are split into four classes:
i) Pinus; ii) mixed deciduous oak-wood, including
mesophilous and thermophilous trees living in warm
and humid climates, such as Quercus, Corylus, Carpi-
nus betulus, Ostrya, Fraxinus, Hedera, Tilia, Ulmus,
Acer, Betula, and Alnus; iii) mountain taxa (Abies, Picea
and Fagus), living at present at altitudes N1500 m, and iv)
Mediterranean taxa, living in Mediterranean climates
characterized by high summer drought and including
sclerophyllous and evergreen trees/shrubs, such as
Quercus ilex, Erica arborea, Phillyrea type, Pistacia
and Olea. Similarly, non-arboreal types (NAP) are
subdivided in three classes: 1) steppic shrubs and herbs
(Artemisia, Chenopodiaceae, Ephedra, Hippophae) with-
standing cold and dry conditions; 2) Poaceae and
3) ubiquists, represented by all other herbs without any
particular ecological demand. The aquatics are repre-
sented by hygro-hydrophyte plants growing in humid
environments, such as Sparganium type, Typha, Myrio-
phyllum and Cyperaceae.
4. Sedimentology and micropalaeontology of core M1
Core M1 (Fig. 2) was drilled in the car park of a
bathing establishment close to the present shoreline. The
uppermost 2 m are of anthropogenic origin. The
succession can be divided into a series of twenty-six
depositional units, the sedimentological and micropa-
laeontological (foraminifers, ostracods, pollen) charac-
teristics of which are described in descending order.
215
4.1. 2.1013.15 m: delta front facies association
The upper part of this facies association, between
2.10 and 6.00 m core depth, consists of well sorted,
medium to coarse sand, with scattered pebbles. This
shows a gradation into the underlying fine to medium
sand, with local silty sand intercalations. Wood frag-
ments and plant debris are abundant, and shell material
is widespread throughout this unit. Poorly preserved
foraminifers (Ammonia, Elphidium and Miliolidae)
and ostracods (Pontocythere turbida and Loxoconcha
turbida) are locally observed within this facies associ-
ation. No pollen were found in this interval.
On the basis of its stratigraphic position and morpho-
logical features of the present coastal plain, this facies
association is attributed to the delta front of a prograding
wave-dominated delta system. The coarser part is inter-
preted to reflect upper shoreface deposits, whereas the
underlying finer-grained deposits are attributed to the
lower shoreface. The presence of few microfossils within
this facies association, characteristic of littoral environ-
ments, is consistent with this interpretation, and suggests
transport under high-energy conditions.
4.2. 13.1519.15: prodeltadelta front transition
facies association
This facies association includes a rhythmic alternation
of silty clay and very fine, sharp-based sand layers. A 2 m
thick sand body is recorded between 17 and 19 m core
depth. Shell material and wood fragments are abundant.
The silty clays include a foraminiferal assemblage
dominated by Ammonia tepida and A. parkinsoniana
(4673%), with common Cribroelphidium spp. (mainly
C. granosum and C. poeyanum). L. turbida with
subordinate P. turbida, Carinocythereis whitei and Lep-
tocythere spp. characterize the ostracod fauna (Fig. 3).
Microfauna is scarce within the sandy deposits, with rare
and poorly preserved foraminifers, similar to the fauna
recorded in the overlying unit. A 14C age of 3224 35 yr
BP was obtained at 15.75 m core depth (Fig. 2).
High percentages of arboreal pollen (AP about 65%),
with mixed deciduous oak-wood, Pinus and Mediterra-
nean taxa, characterize this interval (Fig. 4). Also
significant is the relative abundance of cultivated trees
(such as Olea and Castanea).
This facies association is interpreted to reflect the
gradual transition from a prodelta to a delta front, occur-
ring during a period of intense human activity. The micro-
fauna is diagnostic of a shallow-marine environment with
strong influence by fluvial discharge. The sharp-based
sand layers are interpreted as the result of flood events.
 216
M. Aguzzi et al. / Sedimentary Geology 202 (2007) 211229
Fig. 3. Distribution of benthic foraminifers and ostracods in the upper 50 m of core M1. TS: transgressive surface; TRS: tidal ravinement surface; WRS: wave ravinement surface; MFS: maximum
flooding surface; SB: sequence boundary.
 M. Aguzzi et al. / Sedimentary Geology 202 (2007) 211229
4.3. 19.1528.90 m: prodelta facies association
This facies association differs from the overlying
deposits by the generally lower sand/mud ratio, which
increases from bottom to top of the interval, and the
lower amount of shells and wood fragments. The
foraminiferal assemblage is characterized by the dom-
inance of A. tepida and A. parkinsoniana (3070%),
with subordinate Cribroelphidium and Quinqueloculina
seminulum (Fig. 3). Relatively high frequencies of
Adelosina spp., Ammonia beccarii, Ammonia inflata,
Textularia spp. and Valvulineria bradyana are observed
in the lower part of this interval, below 24 m core depth.
Quantitative palaeontological data are not available
for several sand-rich samples from the upper part
of this facies association, for which less than 300
foraminiferal tests were observed. The ostracod assem-
blage includes mainly Cytheridea neapolitana and
L. turbida, with local high concentrations of P. turbida
and C. whitei.
Pollen data record a progressive downward increase in
AP percentages, reaching about 85% (Fig. 4). In the
uppermost part, forests were characterized by a mixed
sclerophyllous (Mediterranean taxa) and deciduous
broadleaved vegetation typical of a subMediterranean
climate, with irregular and short summer drought
(Jalut et al., 2000). Starting at 2526 m depth, partic-
ularly significant is the relative abundance (about 20%
on average) of mountain taxa, whose components
(mainly Abies and Fagus) grow only under high humidity
conditions.
This facies association represents a prodelta environ-
ment developed during an interglacial period. The
microfauna consists of shallow-marine foraminifers
and ostracods commonly observed in recent Mediterra-
nean deltas (e.g. Colalongo, 1969; Jorissen, 1988;
Bellotti et al., 1994a); the abundance of the euryhaline
foraminifers A. tepida and A. parkinsoniana reflects a
significant fluvial influence. Upward increase in sand/
mud ratio and decreasing amounts of Adelosina, Valvu-
lineria and Textularia suggest increasing freshwater
influx with the progressive approach to the fluvial
mouth.
4.4. 28.9030.50 m: inner-shelf facies association
This facies association is characterized by a homo-
geneous succession of bioturbated silty clay. Apart from
a general abundance of Ammonia, this interval records
the highest specific diversity of both foraminifers and
ostracods, and is characterized by the highest percen-
tages of Adelosina, Triloculina and Textularia (Fig. 3).
217
These are coupled with low frequencies of Q.
seminulum and Cribroelphidium spp. The ostracod
assemblage is dominated by Semicytherura spp. and
Carinocytherei whitei, with subordinate Leptocythere
spp., L. tumida, Cytheroma spp., and Pterygocythereis
ceratoptera.
This interval records the highest percentages of AP
(about 90%) and a significant increase in total pollen
concentration (Fig. 4).
The microfauna observed in this facies association is
typical of a relatively open-marine environment, with
scarce fluvial influence, and is here interpreted as an
inner-shelf, slightly below the wave base level. Pollen
data point to a phase of maximum forest cover under
interglacial climate conditions.
4.5. 30.5032.95 m: transgressive barrier
facies association
This facies association displays a typical fining-
upward tendency, from fine sand to silty sand, with
local intercalations of silty clays. This unit displays an
erosional lower boundary, marked by a veneer of mollusc
shells. The microfauna consists in the lower part of poorly
preserved ostracods (Aurila and Loxoconcha) and
foraminifers (Elphidium and Ammonia), passing upwards
to a foraminiferal association similar to that described for
the inner-shelf facies association, except for the lower
abundance of Textularia (Fig. 3). The ostracods include
L. turbida, P. turbida and Semicytherura spp. In the
uppermost part of the unit Hiltermannycythere turbida
and Pseudocytherura spp. have been observed.
Pollen samples from this facies association, collected
within two fine-grained intercalations, show a similar
proportion of AP groups relative to the overlying unit
(Fig. 4). A remarkable peak of Poaceae is recorded in the
lowermost sample, paralleled by high values of aquatics
and total pollen concentration. A silt intercalation at
32.30 m core depth yielded a 14C age of 7769 45 yr BP
(Fig. 2).
This facies association formed in a high-energy,
coastal environment. The basal erosion surface is inter-
preted as a wave ravinement surface (WRS in Fig. 2)
(Swift, 1968; Nummedal and Swift, 1987). This unit, as
a whole, has been interpreted as a transgressive sand
sheet, reflecting the landward migration of a barrier
during the Holocene transgression. The fining-upward
trend and the microfauna suggest rapid transition to
relatively deeper water environments. Pollen character-
istics of the lowest sample suggest an environmental
control on the pollen association, possibly related to
proximity to marsh environments.
 218
M. Aguzzi et al. / Sedimentary Geology 202 (2007) 211229
Fig. 4. Synthetic pollen and total concentration diagrams of core M1. Facies associations, oxygen-isotope stages and 14C ages are indicated. Dotted lines mark minor divisions based on pollen spectra.
AP, arboreal pollen; NAP, non-arboreal pollen. For main composition of pollen groups, see text.
 M. Aguzzi et al. / Sedimentary Geology 202 (2007) 211229
4.6. 32.9536.10 m: flood tidal delta facies association
This facies association consists of alternating silty
sand and fine-grained to medium sand layers, a few dm
thick. Small fragments of mollusc shells are present. A
rare, poorly preserved microfauna, consisting of
A. tepida, A. parkinsoniana, Brizalina spp., Bulimina
spp. and Cyprideis torosa, characterizes the sand layers
(Fig. 3); a well preserved microfauna composed almost
entirely of A. tepida and A. parkinsoniana (95% of the
foraminiferal assemblage), in association with brackish
ostracods C. torosa and Loxoconcha elliptica, was
observed in a silt intercalation at 33.85 m core depth.
Very high pollen concentration values and maximum
percentages (85%) of mixed deciduous oak-wood are
recorded within this interval (Fig. 4).
The diagnostic microfauna of this facies association,
showing the co-existence of brackish/euryhaline
(A. tepida, A. parkinsoniana, C. torosa, L. elliptica) and
open-marine (Brizalina spp., Bulimina spp.) taxa, suggest
that deposition of this unit took place in a brackish-water
environment, with local input of sediment material of
marine origin. In view of the textural features and the
stratigraphic relationships with the overlying and under-
lying units, this facies association is interpreted as a flood
tidal delta, i.e. a sedimentary body located in the back-
barrier area behind a tidal inlet. Maximum development of
the mesophilous forest suggests a stable warm-temperate
climate.
4.7. 36.1050.90 m: estuary facies association
This facies association includes a thick, homoge-
neous succession of dark grey silty clay. Sand layers,
displaying both fining-upward and coarsening-upward
tendencies and wave ripples at their tops, are commonly
encountered between about 39 and 47 m core depth.
Wood and shell fragments are also scattered throughout
this interval. A 5 cm-thick layer, entirely made up of
mollusc shells, with abundant Cerastoderma glaucum,
is present at the base of this facies association.
The foraminiferal assemblage consists almost entire-
ly of A. tepida and A. parkinsoniana (about 80100%),
with local occurrences (up to 20%) of Haynesina
depressula (Fig. 3). Among the ostracods C. torosa is
the dominant species, with subordinate L. elliptica.
Changes in the proportion of foraminiferal and ostracod
specimens allow subdivision of this facies association
into three intervals (Fig. 2): ostracods are more abundant
than foraminifers between 39 and 45 m core depth,
where mesohaline to freshwater ostracod species
(Candona lactea, C. neglecta, Cypris pubera and Ilyo-
219
cypris gibba) are recorded. A reverse proportion of these
taxa is recorded above and below this interval.
With regard to pollen data, this interval can be
subdivided into two parts (Fig. 4): similar to the over-
lying facies association, the upper interval is character-
ized down to the depth of about 45 m by a dominance of
mixed deciduous oak-wood and low values of the other
arboreal groups. In the lower interval, the relative
proportion of Pinus increases significantly, while total
pollen concentration decreases. A sample at 39.40 m
core depth yielded a 14C age of 8840 75 yr BP (Fig. 2).
The sedimentological features, combined with the very
low microfaunal diversity and the presence of dominant
brackish assemblages suggest that deposition of this
facies association took place within a wave-dominated
estuary (central basin deposits of Dalrymple et al.,
1992), in which two distinct subenvironments can be
identified. The lower and upper intervals, showing maxi-
mum foraminiferal abundance, are interpreted to reflect an
outer estuary environment. By contrast, the ostracod-rich
stratigraphic interval including mesohaline to freshwater
species, which displays a significantly stronger fluvial
influence, is interpreted as an inner estuary deposit. The
basal shell layer, with a characteristic brackish fauna, is
interpreted as a transgressive lag, representing the initial
stage of the estuarine transgression, and corresponds to
the tidal ravinement surface (TRS in Fig. 2) of Allen and
Posamentier (1993) and Zaitlin et al. (1994), or the bay
ravinement surface of Nummedal and Swift (1987) and
Foyle and Oertel (1997).
Pollen data from the lower part of this facies
association indicate a scattered warm-temperate forest
vegetation, with significant occurrence of drought-
tolerant trees (Pinus), growing under early interglacial
climate conditions. Subsequently, a mixed deciduous
mesophilous forest was established.
4.8. 50.9051.20 m: swamp facies association
This facies association is composed of a few dm-
thick grey clay beds with abundant organic material. No
microfauna was observed within this interval.
A 14C age of 10,004 50 yr BP was obtained close to
the base of this facies association, at 51.15 m core depth
(Fig. 2).
The pollen spectrum available for this interval
displays a remarkable peak in total pollen concentration
and high values of AP (80%), mainly represented by
mixed deciduous oak-wood taxa (Fig. 4).
This facies association is interpreted as a freshwater
swamp deposit that formed at the onset of the Holocene
transgression, and could represent the initial expression
220 M. Aguzzi et al. / Sedimentary Geology 202 (2007) 211229
of a transgressive phase (TS in Fig. 2). Pollen charac-
teristics support this interpretation, indicating well-
developed forest vegetation typical of interglacial cli-
mate conditions.
4.9. 51.2054.60 m: fluvial-channel facies association
This facies association consists of gravelly sand, sand
and dark brown sandy silt, with a distinctive fining-
upward trend and a sharp erosional lower boundary. An
overconsolidated, dark horizon, approximately 50 cm
thick, caps the succession. No microfossils are present in
this interval.
Pollen are very scarce or absent throughout this unit.
Pollen analysis of two samples collected within the stiff
deposits at top of this facies association, between 51.40
and 51.20 m core depth, showed dominance of a single
herbaceous taxon (Asteraceae, belonging to ubiquists),
with parallel very low AP values (Fig. 4).
This facies association is interpreted as fluvial-channel
sediments, discharged by the ancient Arno River, and
formed during a cold climatic period. The fining-upward
trend is interpreted to mark the progressive channel aban-
donment, with subsequent soil development on top. The
unusual pollen spectrum recorded close to the top of the
facies association, including grains that are very resistant
to oxidation and corrosion (Bottema, 1975), is consistent
with its interpretation as a palaeosol. In this case, the high
dominance of Asteraceae should be interpreted as due to
differential pollen preservation and has no value as open
vegetation indicator (Havinga, 1984).
4.10. 54.6056.70 m: floodplain facies association
This facies association consists of a monotonous
succession of silty clay. No microfossils are present, with
the exception of scattered, poorly preserved ostracods.
Despite the general low pollen content, a vegetation
association dominated by NAP, with scattered trees of
mixed deciduous oak-wood and Pinus, is observed
(Fig. 4). Of note, are the very high values of steppics.
A 14C age of 23,462 125 yr BP was obtained from a
sample at 54.95 m core depth (Fig. 2).
This facies association has been attributed to a
floodplain environment. The characteristic pollen spec-
trum indicates the development of open vegetation
typical of cold and arid climate conditions.
4.11. 56.7061.40 m: fluvial-channel facies association
This facies association shows a distinctive fining-
upward trend from coarse to fine sand, with an erosional
lower boundary. No microfauna was observed within
this interval.
Samples from this interval are very poor in pollen
and display an association similar to that recorded in the
overlying deposits (Fig. 4).
Similarly to the stratigraphic interval between
50.90 m and 54.60 m core depth, this facies association
is interpreted to represent a fluvial-channel deposit.
4.12. 61.4063.30 m: marsh facies association
This facies association is composed of grey clay,
with a few silt intercalations in the uppermost 60 cm.
The microfauna records the occurrence throughout this
interval of an ostracod assemblage dominated by
C. torosa, with rare Leptocythere lagunae (Fig. 2).
The pollen spectra show relatively high values of
mixed deciduous oak-wood (about 40% in average), and
a consequent reduction of NAP (Fig. 4).
These deposits, characterized by a typical brackish
microfauna, are inferred to have formed in a brackish-
water environment, such as a marsh, in close proximity
to a marine environment (bay or estuary). Pollen data
suggest a moderate expansion of mesophilous forests
within an open vegetation landscape, indicating a slight-
ly warmer and more humid climate.
4.13. 63.3072.40 m: lake facies association
This facies association includes a thick, homogeneous
succession of clays. Horizontal lamination with laminae
of different colours (from yellowish grey to light grey,
dark grey and black) is common. The darker colours are
due to the abundance of organic material. A considerable
concentration of freshwater molluscs was observed. The
microfauna of this interval consists uniquely of rare
oligohaline to freshwater ostracods, such as C. lactea,
C. neglecta and Cyclocypris spp. (Fig. 2).
Pollen data display very high values of Poaceae and
moderate percentages of mixed deciduous oak-wood,
decreasing downward (Fig. 4). A significant occurrence
of aquatics is recorded throughout this interval.
The laminated clays are believed to reflect a lake
deposit. Lack of connection with the open sea is
supported by the diagnostic fossil content. The anom-
alously high values of Poaceae and aquatics are an
additional, obvious indication of humid environment.
4.14. 72.4074.80 m: subdelta facies association
This facies associations consists of silty clay and fine
to medium sand. No microfossils are present in this
 M. Aguzzi et al. / Sedimentary Geology 202 (2007) 211229
interval and no samples were found suitable for pollen
analysis.
In view of the stratigraphic relationships with the
overlying and underlying units (see next section),
this facies association is presumed to reflect a sub-
delta deposit, formed at the boundary between a wet
coastal plain and an alluvial plain. This sand body
could be linked to overflow events of either a river or a
distributary channel into an enclosed freshwater terres-
trial area.
4.15. 74.8076.20 m: swamp facies association
This facies association is composed of grey and
dark grey clay, with rare silt layers. Planorbis molluscs
and carbonized wood fragments are abundant. Dark
laminae, due to concentration of organic material, are
common.
The pollen association is characterized by high
values of NAP, mainly steppics and Poaceae, and low
values of AP (Fig. 4). Pinus is the most significant
tree and the aquatics show a remarkable peak (about
40%).
The sedimentological characteristics suggest that this
facies association formed in a quiet, freshwater paludal
environment. This interpretation is strongly supported
by the anomalously high values of aquatics.
4.16. 76.2077.70 m: floodplain facies association
This facies association is made up of grey silty clay,
with plant debris and rare terrestrial gastropods (Helix).
No microfossils were found within this interval.
Pollen analysis shows considerably high percentages
of Pinus and lack of aquatics (Fig. 4).
This facies association is interpreted to represent the
finest portion of deposits which settle in the alluvial
plain and are afterwards exposed to subaerial processes.
Lack of aquatic plants indicates the occurrence of dry
paleoenvironmental conditions.
4.17. 77.7083.90 m: fluvial-channel facies association
This facies association consists of sandy gravel,
grey sand and silty sand, with a distinctive fining-upward
trend and a sharp erosional lower boundary. No
microfossils are present in this interval. This facies
association yielded no samples suitable for pollen
analysis.
These deposits are interpreted as fluvial-channel
sediments, discharged by the palaeo-Arno river possibly
during a cold period.
221
4.18. 83.9084.50 m: swamp facies association
These sediments consist of brown peaty clay with
abundant wood fragments and plant debris. Small
freshwater gastropods (Bithynia, Planorbis) are present.
No microfauna was found in this interval.
This unit displays a high content of pollen with
dominant Poaceae and significant percentages of aqua-
tics (Fig. 4).
This association appears to have been deposited in a
humid, terrestrial environment, such as a swamp.
4.19. 84.5087.70 m: floodplain facies association
This facies association includes grey silty clay, with
plant debris and rare terrestrial gastropods. No micro-
fauna was found in this interval.
Pollen of NAP dominate this interval, while AP are
represented by remarkable values of mixed deciduous
oak-wood (reaching a maximum of 22%) and Pinus
(Fig. 4).
This facies association reflects a floodplain environ-
ment developed during a moderately warm and humid
period.
4.20. 87.7089.40 m: levee facies association
This facies association shows a sharp lower boundary
with the underlying facies association, and consists of a
rhythmic alternation of silty clay and fine to medium
sand layers. Plant debris is present. The clay intercala-
tions contain calcareous nodules. No microfauna was
found in this interval.
Pollen samples are characterized by high percentages
of Pinus, an abundance of NAP, and very scarce
occurrence of mixed deciduous oak-wood and low total
pollen concentration (Fig. 4).
This facies association formed in a terrestrial envi-
ronment, probably as an overbank deposit on a levee.
The pollen spectrum suggests the development of open
vegetation with scattered trees and shrubs, typical of
cold climate conditions.
4.21. 89.4097.90 m: delta front facies association
This facies association is dominated by coarse to
medium sand, with abundant shell material, grading
upward into gravelly sand. Wood fragments are present
throughout this interval. Rare and poorly preserved
foraminifers (Ammonia, Elphidium, Miliolidae) and
ostracods (Aurila, Loxoconcha and P. turbida) have
been observed (Fig. 2).
222 M. Aguzzi et al. / Sedimentary Geology 202 (2007) 211229
The available pollen spectra show very high mean
values of Pinus (60%), a considerable amount of mixed
deciduous oak-wood (30% in the uppermost sample),
mountain and Mediterranean taxa, and very low NAP
values (Fig. 4).
This facies association is interpreted to reflect a
shallowing-upward succession within a high-energy,
delta-front environment, with sandy lower shoreface
deposits grading upward into gravel rich, upper shore-
face deposits.
The pollen association and the high total pollen
concentration are indicative of a dense, mixed evergreen
and deciduous forest vegetation, indicating a cool-
temperate climate. A superposed environmental control
on pollen association is suggested by the extremely high
percentages of Pinus, as the result of selective preser-
vation of pollen grains within these beach deposits.
4.22. 97.90100.00 m: transgressive barrier
facies association
This facies association includes fine to medium sand
with an erosional lower boundary and a remarkable
abundance of shell materials. The microfauna is very
similar to that from the above interval. No pollen have
been found in this interval.
This unit is interpreted to reflect a transgressive sand
sheet due to the landward migration of a transgressive
barrier. The erosional lower boundary corresponds to
the wave ravinement surface (WRS in Fig. 2).
4.23. 100.00100.85 m: washover facies association
This facies association consists of fine to medium
sands, with rare small fragments of mollusc shells. No
microfauna was found in this interval.
Total pollen concentration is low (Fig. 4). Although
NAP are predominant, moderate amounts of mixed
deciduous oak-wood and Mediterranean taxa also
characterize this interval.
Given the stratigraphic relationships with the adjacent
units, this facies association is interpreted to have formed
in a back-barrier environment. Particularly, this deposit
could have accumulated on the landward side of a beach-
barrier, due to barrier overstepping by storm processes
(washover sand). The low total pollen concentration is
consistent with such a peculiar depositional environment.
4.24. 100.85101.70 m: lagoon facies association
This facies association includes a sandy layer with
a characteristic brackish-water mollusc fauna (C. glau-
cum); this is overlain by peaty silty clay with shell material
and carbonized wood fragments. A typical brackish
microfauna is recorded in the fine-grained deposits of this
interval (Fig. 2). The foraminiferal assemblage is
dominated by A. tepida and A. parkinsoniana, while the
ostracod association is mainly composed of C. torosa.
The pollen association is dominated by NAP (mostly
Poaceae and steppics) and shows increasing-upward
values of mixed deciduous oak-wood (Fig. 4). The
abundance of aquatics is significantly high.
The lower boundary of this facies association rep-
resents the initial phase of a marine ingression and corre-
sponds to the transgressive surface (TS in Fig. 2). The
sedimentological and micropalaeontological features of
this facies association are typical of a brackish lagoon
environment. Pollen spectra suggest that sedimentation
took place during a transitional phase of reforestation
under warming climate conditions. The abundance of
aquatics indicates the local presence of stagnant waters.
4.25. 101.70102.20 m: swamp facies association
This facies association consists of dark sandy clay,
with abundant wood fragments and plant debris. No
microfossils are present in this interval.
Pollen data record very high amounts of NAP (90%),
while AP display low values, with very low quantities of
Pinus and abundant aquatics (Fig. 4).
This facies association is typical of an enclosed wet
terrestrial environment, such as a swamp. The characteristic
pollen association is indicative of a steppe environment
developed under very cold and arid climate conditions.
4.26. 102.20105.00 m: fluvial-channel facies association
This facies association consists of fine to medium
sands, with a distinctive fining-upward trend. No micro-
fauna was recorded in this interval.
The pollen spectrum available records the highest
values of NAP (98%) of the whole succession and a very
low pollen concentration (Fig. 4).
This facies association is interpreted as fluvial-
channel sediments discharged by the palaeo-Arno river
during a cold and arid phase. This is documented by
pollen data, suggesting the presence of scattered vege-
tation, characterized mainly by steppic herbs.
5. Palaeoenvironmental and climatic evolution
The vertical stacking pattern of facies observed in
core M1 shows that the Late Quaternary succession of
the Arno coastal plain consists of a repeated alternation
 M. Aguzzi et al. / Sedimentary Geology 202 (2007) 211229
of coastal and alluvial deposits (Fig. 2). The two major
transgressive surfaces (TSs) detected at about 51 and
102 m core depth, respectively, appear as the two most
valuable stratigraphic markers in the core. These may
enable a pragmatic subdivision of the stratigraphic
succession into two vertically stacked transgressive
regressive (TR) sequences (sensu Embry, 1993, 1995),
of Holocene and Late Pleistocene age, respectively. A
minor transgression surface could be represented by the
lower boundary of swamp deposits resting on floodplain
deposits at 76 m core depth.
Sediments deposited in brackish-water environments
directly overlie the TSs, which are generally developed
on top of alluvial sediments. These early transgressive
deposits are overlain by nearshore and shallow-marine
sediments (late transgressive + highstand). The upper
parts of TR sequences show an upward transition to
continental deposits, formed during periods of sea-level
fall and subsequent sea-level lowstand.
Similar to recent observations by Amorosi and
Colalongo (2005) from the Late Quaternary deposits of
the Po Plain, in Northern Italy, transgressive sediments
above the TSs carry a distinctive pollen signature, cor-
responding to: i) maximum expansion of forest vegeta-
tion (AP), ii) abrupt decrease in non-arboreal pollen
(NAP), and iii) major peaks in total pollen concentration
(Fig. 4). These characteristics point to an obvious
relationship between transgressive episodes and the
onset of warm (interglacial) phases. By contrast, alluvial
plain deposits beneath the TSs show pollen assemblages
dominated by open vegetation communities (NAP) and
significantly lower concentration values, which are
typical of glacial periods.
Given the uncertain age attribution of the lower TR
sequence, for which no absolute age data are available,
we will discuss our core data from top to bottom.
5.1. The upper TR sequence (OIS 1)
The upper TS identified at about 51 m core depth
records the onset of the Holocene transgression (Fig. 2),
as inferred from five radiocarbon dates. Of particular
interest is the remarkable thickness (about 20 m) of early
transgressive, brackish-water facies between the tidal
ravinement surface (TRS) and the wave ravinement
surface (WRS), suggesting that sedimentation between
10 and 7.8 kyr BP took place within a wave-dominated
estuary (Dalrymple et al., 1992; Nichol et al., 1994).
This interpretation is consistent with recent work from a
more landward location in the Arno coastal plain,
showing the presence of an incised-valley fill, about
25 m thick, roughly perpendicular to the present shore-
223
line and corresponding to the modern Arno River course
(Fig. 5). Away from the valley axis, the Holocene suc-
cession is less than 20 m thick and overlies an indurated
and weathered palaeosurface. This surface, correspond-
ing to an interfluve sequence boundary (Van Wagoner
et al., 1990; Aitken and Flint, 1996; McCarthy and Plint,
1998), was formed during the last glacial period and was
transgressed by brackish waters around 7.8 kyr BP (see
core M3 in Fig. 5).
The thick estuarine complex observed in core M1
thus appears to constitute the SW prolongation of the
incised-valley (see Fig. 1). A similar estuary infilling by
clayey sediments during the last transgression has been
reported from several incised-valley systems (Hori et al.,
2001; Ta et al., 2001; Boski et al., 2002; Li et al., 2002).
The transgressive barrier sands dated to about
7800 yr BP document the rapid landward migration of
a beach-barrier system during the late stage of trans-
gression. These sands, in turn, are overlain by thin inner-
shelf clays, indicating rapidly increasing water depth.
The maximum flooding surface, identified at 29.85 m
core depth within lithologically homogeneous inner-
shelf deposits (Fig. 2), was recognized on the basis of
subtle micropalaeontological evidence (Fig. 3), such as
the highest percentages of Textularia (17%) and Ade-
losina (18%), associated with relatively low contents of
A. tepida and A. parkinsoniana (see also Fiorini and
Vaiani, 2001; Amorosi and Colalongo, 2005).
Late transgressive marine sediments are significantly
thinner (3 m) than the overlying highstand succession of
prograding deltaic deposits, which attains a total
thickness of 30 m. This asymmetry is characteristic of
the Holocene transgressiveregressive sedimentary
wedges described from the major coastal plains all
around the world (see review in Amorosi and Milli,
2001).
Detailed pollen analysis enables the reconstruction
of a three-fold palaeoclimatic evolution for the Holo-
cene of core M1, documenting a direct relationship
between vertical facies patterns and vegetation history
(Fig. 4). The first phase, marking the transition from
glacial to interglacial conditions, was characterized by
the abrupt spread of a mixed deciduous oak-wood,
which took place in response to a pronounced phase
of climate warming at the beginning of the Holocene.
This remarkable change in climatic conditions was
paralleled by the onset of transgressive sedimentation
within the incised-valley, showing transition from
fluvial to paludal, and then outer-estuarine deposits
(see also Zaitlin et al., 1994; Hori et al., 2002). Similar to
other pollen sequences of the Mediterranean area
(Prez-Obiol and Juli, 1994; Nicol-Pichard and
224 M. Aguzzi et al. / Sedimentary Geology 202 (2007) 211229
Fig. 5. Cross-section showing the geometry of the Holocene Arno incised-valley fill, 10 km NE of core M1. For detailed facies description of core
M3, see Aguzzi (2006). For section trace, see Fig. 1.
Dubar, 1998; Carrin and Van Geel, 1999), early
Holocene vegetation is associated with the presence of
drought-tolerant Pinus woodland (Fig. 4). Although
masked by a peak in Poaceae, which probably reflects
proximity to a paludal environment, the first vegetation
phase terminates at 45 m core depth with a small, but
significant decline of temperate trees. This is accompa-
nied by a slight regressive tendency within the incised-
valley, recorded in the core by rapid transition from an
outer-to inner-estuarine environment.
During the second phase, extending up to about
7800 yrs BP, Pinus woodland was largely replaced by the
oak-wood, which shows a considerable expansion and
culmination. This vegetation phase, suggesting a progres-
sive increase in temperature and humidity, matches the
progressive landward shift of the shoreline, shown by
upward transition from inner estuary to outer estuary, and
then flood tidal-delta deposits.
Maximum AP values, corresponding to the maxi-
mum development of trees and thus suggesting the most
favourable temperature and humidity conditions, are
recorded at the onset of marine sedimentation, within
transgressive-barrier and inner-shelf deposits. After
peak transgression, forest composition underwent an
important change: the mixed deciduous oak-wood
declined and the Mediterranean and Pinus woodland
spread. The development of the Mediterranean forest,
mainly characterized by Q. ilex, represents a significant
palaeovegetation event recorded in many Italian sites,
such as Valle di Castiglione (Follieri et al., 1988), Lago
di Vico (Magri and Sadori, 1999), Lagaccione (Magri,
1999), Ripasottile (Ricci Lucchi et al., 2000), and the
Tyrrhenian Sea (Russo Ermolli and di Pasquale, 2002).
This event suggests a change toward a more Mediter-
ranean climate that persisted and probably increased its
characteristic summer drought during the second half of
the Holocene, when the Mediterranean vegetation be-
came dominant and a progressive reduction in other
forest components was recorded (mixed deciduous oak-
wood and mountain taxa), parallel to an expansion of
open vegetation.
5.2. The lower TR sequence (OIS 5e-2)
No direct age attribution is available for the nearshore
sands overlying the lower TS, at about 102 m core depth.
Although additional chronological control must be
obtained to constrain the exact timing of deposition of
this unit, we are inclined to assign these sands to the last
interglacial period, corresponding to the Tyrrhenian or
 M. Aguzzi et al. / Sedimentary Geology 202 (2007) 211229
Eemian (Oxygen-Isotope Substage 5e 128116 kyr
BP), in view of the absence of marine deposits between
this interval and the Holocene sediments (Fig. 2). This
attribution is consistent with data from the Enea core in
the adjacent Versilia Plain (Fig. 1), which have provided
an age of 129.2 and 132.8 kyr BP at about 69 m and 72 m
core depth, respectively (Antonioli et al., 1999; Ferranti
et al., 2006).
Pollen from the swamp and the fluvial deposits
below the TS (102105 m core depth) provide strong
indication of glacial climate conditions (dominance of
NAP and particularly steppics) corresponding to OIS 6
(Fig. 4). Pollen data within the lagoonal sediments just
above the TS record a progressive increase of arboreal
taxa, mainly deciduous oak-wood and Mediterranean,
which can be related to the onset of the last interglacial
(OIS 5e). This is followed by a no-pollen zone, coin-
cident with the deposition of washovers and transgres-
sive barrier sands, which likely records the maximum
warming phase of the Eemian. Scattered pollen samples
within the overlying, regressive lower-shoreface sedi-
ments indicate the presence of a dominant coniferous
forest (characterized by Pinus and mountain taxa), with
scarcity of mesophilous trees belonging to the oak-
wood. This pollen association could be the result of the
progressive cooling typical of the latest phase of the
Eemian (Guiot et al., 1989; De Beaulieu and Reille,
1992; Zagwijn, 1996; Tzedakis et al., 2003; Preusser,
2004).
An abrupt facies change separates the Eemian marine
deposits from the overlying alluvial sediments, suggest-
ing the presence of a significant stratigraphic hiatus.
Pollen data available for the levee deposits capping the
delta front succession indicate open vegetation, charac-
terized mainly by steppics and Pinus, which can be
related to an early Weichselian stadial phase (possibly
corresponding to OIS 5b). The pollen association re-
corded within the overlying floodplain and swamp
deposits represents a significant expansion of mesophi-
lous forest (mixed deciduous oak-wood) that can be
associated with an interstadial phase (OIS 5a).
The alluvial (fluvial-channel and floodplain) deposits
between about 76 and 84 m core depth have been
assigned to OIS 4, in view of their stratigraphic position,
sedimentological characteristics and pollen assemblage
dominated by Pinus, which indicate the onset of glacial
climate conditions. Accordingly, we assume that the
overlying swamp, subdelta, lake and marsh deposits
(6176 m) were primarily formed during OIS 3, which
implies that the minor transgressive surface at 76 m
core depth marks the OIS 4/3 transition. This is, again,
suggested by pollen data, which indicate the persistence
225
of a considerable mesophilous forest typical of intersta-
dial climate conditions within an open vegetation
landscape.
The two erosional surfaces marking the base of
fluvial-channel deposits at 61.40 and 54.60 m core
depth, respectively, are both potential candidates for the
sequence boundary (SB), corresponding to the OIS 3/2
transition. On the basis of the radiocarbon date of
23.4 ka at about 55 m core depth, which would translate
after calibration (Stuiver et al., 1998) into about 27.5 ka
(late OIS 3), we are inclined to place the SB at the upper
unconformable surface.
6. Conclusions
Through integrated sedimentological and micropa-
laeontological (foraminifer, ostracod and pollen) anal-
yses of a 105-m-long core from western Tuscany, new
insights were obtained into the response of this coastal
system to combined sea-level and climate fluctuations
during the last 150 kyr.
Two prominent intervals with marine sedimentation
permit us to define two distinct transgressiveregressive
(TR) sequences, of Holocene and Late Pleistocene age,
respectively. Stratigraphic correlation with the marine
oxygen-isotope record on the basis of pollen data
documents strict relationships between TR sequences
and interglacial/glacial cycles, which shows that trans-
gressive surfaces correlate invariably with the onset of
forested conditions during interglacials, whereas the
return to alluvial sedimentation correlates with an abrupt
change to open vegetation conditions during glacials.
An attempt to delineate the palaeoenvironmental
evolution of the study area in response to sea-level
fluctuations and climate forcing resulted in the estab-
lishment of a well-resolved chronology for the studied
succession, from OIS 6 to Present. The transgressive
surfaces are much more readily identifiable surfaces
than sequence boundaries.
This paper presents the first detailed facies documen-
tation of an incised-valley sequence from the Late
Quaternary of Italy. The Holocene succession is 51 m
thick and includes about 20 m of early transgressive
estuarine deposits above a tidal ravinement surface.
These are capped by a wave ravinement surface, overlain
by a thin succession of late transgressive barrier-beach
and inner-shelf deposits. Highstand deposits above the
maximum flooding surface are represented by a
shallowing-upward succession of prodelta to delta-front
sediments.
The detailed reconstruction of palaeoenvironmental
evolution and vegetation changes in the Tuscan area
226 M. Aguzzi et al. / Sedimentary Geology 202 (2007) 211229
during the last 150 kyr may help significantly in refining
our knowledge of sea-level fluctuations and climate
changes in the Mediterranean area during the Late
Quaternary.
Acknowledgements
This study was funded by the Bologna University as
part of Progetto Strategico d'Ateneo (co-ordinator:
A. Amorosi). We are grateful to M. Gibling and M.G.
Carboni for their accurate reviews, and to C. Fielding and
G.J. Weltje for the editorial advice.
References
Aguzzi, M., 2006. Stratigrafia di sottosuolo di depositi tardoquaternari
del Valdarno inferiore (Toscana). PhD Thesis. University of
Bologna.
Aguzzi, M., Amorosi, A., Sarti, G., 2005. Stratigraphic architecture of
Late Quaternary deposits in the lower Arno Plain (Tuscany, Italy).
Geol. Rom. 38, 110.
Aitken, J.F., Flint, S.S., 1996. Variable expressions of interfluvial
sequence boundaries in the Breathitt Group (Pennsylvanian),
eastern Kentucky, USA. In: Howell, J.A., Aitken, J.F. (Eds.), High
Resolution Sequence Stratigraphy: Innovations and Applications.
Geol. Soc. London Spec. Publ., vol. 104, pp. 193206.
Albani, A., Serandrei Barbero, R., 1990. I Foraminiferi della Laguna e
del Golfo di Venezia. Mem. Sci. Geol. 42, 271341.
Allen, G.P., Posamentier, H.W., 1993. Sequence stratigraphy and
facies model of an incised valley fill: the Gironde Estuary, France.
J. Sediment. Petrol. 63, 378392.
Amorosi, A., Colalongo, M.L., 2005. The linkage between alluvial and
coeval nearshore marine successions: evidence from the Late
Quaternary record of the Po River Plain, Italy. In: Blum, M.D.,
Marriott, S.B., Leclair, S.F. (Eds.), Fluvial Sedimentology VII.
International Association of Sedimentologists Special Publ., vol. 35,
pp. 257275.
Amorosi, A., Milli, S., 2001. Late Quaternary depositional architecture of
Po and Tevere river deltas (Italy) and worldwide comparison with
coeval deltaic successions. Sediment. Geol. 144, 357375.
Amorosi, A., Colalongo, M.L., Pasini, G., Preti, D., 1999a. Sedimentary
response to Late Quaternary sea-level changes in the Romagna
coastal plain (northern Italy). Sedimentology 46, 99121.
Amorosi, A., Colalongo, M.L., Fusco, F., Pasini, G., Fiorini, F., 1999b.
Glacio-eustatic control of continental-shallow marine cyclicity
from Late Quaternary deposits of the south-eastern Po Plain
(Northern Italy). Quat. Res. 52, 113.
Amorosi, A., Centineo, M.C., Colalongo, M.L., Pasini, G., Sarti, G.,
Vaiani, S.C., 2003. Facies architecture and Latest Pleistocene
Holocene depositional history of the Po Delta (Comacchio area).
Italy. J. Geol. 111, 3956.
Amorosi, A., Colalongo, M.L., Fiorini, F., Fusco, F., Pasini, G., Vaiani,
S.C., Sarti, G., 2004. Palaeogeographic and palaeoclimatic
evolution of the Po Plain from 150-ky core records. Glob. Planet.
Change 40, 5578.
Amorosi, A., Centineo, M.C., Colalongo, M.L., Fiorini, F., 2005.
Millennial-scale depositional cycles from the Holocene of the Po
Plain, Italy. Mar. Geol. 222223, 718.
Antonioli, F., Dai Pra, G., Hearty, P.J., 1988. I sedimenti quaternari
nella fascia costiera della Piana di Fondi (Lazio meridionale). Boll.
Soc. Geol. Ital. 107, 491501.
Antonioli, F., Girotti, O., Impronta, S., Nisi, M.F., Puglisi, C., Verrubbi, V.,
1999. Nuovi dati sulla trasgressione marina olocenica e sulla
subsidenza della Pianura Versiliese attraverso in sondaggio di
90 metri. Atti del Convegno Le Pianure. Conoscenza e salvaguardia,
Regione Emilia Romagna, Bologna, Italy, pp. 214218.
Argnani, A., Bernini, M., Di Dio, G.M., Papani, G., Rogledi, S., 1997.
Stratigraphic record of crustalscale tectonics in the Quaternary of
the Northern Apennines (Italy). Quaternario 10, 595602.
Athersuch, J., Horne, D.J., Whittaker, J.E., 1989. Marine and brackish
water ostracods. In: Kermack, D.M., Barnes, R.S.K. (Eds.), Synopses
of the British Fauna (New Series), vol. 43, Brill E.J, Leiden.
Barra, D., Cinque, A., Gedwelt, M., Hurtgen, C., 1991. L'ospite caldo
Sylvestra seminis (Bunaduce, Masoli and Pugliese, 1976)
(Crustacea, Ostracoda): un possibile marker dell'ultimo intergla-
ciale dell'area mediterranea. Quaternario 4, 327332.
Bellotti, P., Carboni, M.G., Di Bella, L., Palagi, I., Valeri, P., 1994a.
Benthic foraminiferal assemblages in the depositional sequence of
the Tiber Delta. Boll. Soc. Paleontol. Ital., Spec. Vol. 2, 2940.
Bellotti, P., Chiocci, F.L., Milli, S., Tortora, P., Valeri, P., 1994b.
Sequence stratigraphy and depositional setting of the Tiber delta:
integration of high-resolution seismics, well logs, and archeological
data. J. Sediment. Res., Sect. B Stratigr. Glob. Stud. 64, 416432.
Bellotti, P., Milli, S., Tortora, P., Valeri, P., 1995. Physical stratigraphy
and sedimentology of the Late PleistoceneHolocene Tiber Delta
depositional sequence. Sedimentology 42, 617634.
Bellotti, P., Caputo, C., Davoli, L., Evangelista, S., Garzanti, E.,
Pugliese, F., Valeri, P., 2004. Morpho-sedimentary characteristics
and Holocene evolution of the emergent part of the Ombrone River
delta (southern Tuscany). Geomorphology 61, 7190.
Biserni, G., Berendsen, H.J.A., Sandrelli, F., 2004. Shape reconstruc-
tion of the Pleistocene/Holocene unconformity in the Grosseto
Plain (Tuscany, Italy). Quaternario 17, 443451.
Blanc-Vernet, L., 1969. Contribution l'tude des foraminifres de
Mditerrane. Travaux de la Station Marine d'Endoume, Marseille.
Blum, M.D., Trnqvist, T.E., 2000. Fluvial response to climate and
sea-level change: a review and look forward. Sedimentology 47
(Suppl. 1), 248.
Bonaduce, G., Ciampo, G., Masoli, M., 1975. Distribution of
Ostracoda in the Adriatic Sea. Pubbl. Stn. Zool. Napoli 40, 1304.
Bortolotti, V., 1966. La tettonica trasversale dell'Appennino 1. La linea
LivornoSillaro. Boll. Soc. Geol. Ital. 85, 529540.
Boski, T., Moura, D., Veiga-Pires, C., Camacho, S., Duarte, D., Scott,
D.B., Fernandes, S.G., 2002. Postglacial sea-level rise and
sedimentary response in the Guadiana Estuary, Portugal/Spain
border. Sediment. Geol. 150, 103122.
Bottema, S., 1975. The interpretation of pollen spectra from prehistoric
settlements (with special attention to Liguliflorae). Palaeohistoria
17, 1735.
Carboni, M.G., Bergamin, L., Di Bella, L., Iamundo, F., Pugliese, N.,
2002. Late Quaternary sea-level changes along the Tyrrhenian Sea.
Paleoecological evidences from foraminiferal and ostracod
assemblages. Geobios 35, 4050.
Carboni, M.G., Frezza, V., Bergamin, L., 2004. Distribution of Recent
foraminifers in the Ombrone River Basin (Tuscany Continental Shelf,
Tyrrhenian Sea, Italy): relations with fluvial run-off. In: Coccioni, R.,
Galeotti, S., Lirer, F. (Eds.), Proc. First Ital. Meeting Environ. Micropal.,
Grzybowski Found. Spec. Publ., vol. 9, Krakow, Poland, pp. 716.
Carrin, J.S., Van Geel, B., 1999. Fine-resolution Upper Weichselian
and Holocene palynological record from Navarrs (Valencia,
 M. Aguzzi et al. / Sedimentary Geology 202 (2007) 211229
Spain) and a discussion about factors of Mediterranean forest
succession. Rev. Palaeobot. Palynol. 106, 209236.
Caspers, G., Freund, H., 2001. Vegetation and climate in the Early-Pleni-
Weichselian in northern central Europe. J. Quat. Sci. 16, 3148.
Cattaneo, A., Steel, R.J., 2003. Transgressive deposits: a review of
their variability. Earth-Sci. Rev. 62, 187228.
Cattaneo, A., Trincardi, F., 1999. The Late Quaternary transgressive
record in the Adriatic epicontinental sea: basin widening and facies
partitioning. In: Bergman, K.M., Snedden, J.W. (Eds.), Isolated
Shallow Marine Sand Bodies: Sequence Stratigraphic Analysis
and Sedimentologic Interpretation. Soc. Econ. Paleontol. Mineral.
Spec. Publ., vol. 64, pp. 127146.
Cimerman, F., Langer, M.R., 1991. Mediterranean Foraminifera. Acade-
mia Scientiarum et Artium Slovenica, Ljubljana, p. 30. Classis IV.
Colalongo, M.L., 1969. Ricerche sugli Ostracodi nei fondali antistanti
il delta del Po. G. Geol. 36, 335362.
Cortemiglia, G.C., Mazzanti, R., Parea, G.C., 1983. Geomorfologia
della Baia Baratti (LivornoToscana) e della sua spiaggia. Geogr.
Fis. Din. Quat. 6, 148173.
Dalrymple, R.W., Zaitlin, B.A., Boyd, R., 1992. Estuarine facies models:
conceptual basis and stratigraphic implications. J. Sediment. Petrol.
62 (6), 11301146.
De Beaulieu, J.-L., Reille, M., 1992. The last climatic cycle at La
Grande Pile (Vosges, France). A new pollen profile. Quat. Sci. Rev.
11, 431438.
Debenay, J.-P., Guillou, J.-J., Redois, F., Geslin, E., 2000.
Distribution trends of foraminiferal assemblages in paralic
environments. A base for using foraminifera as bioindicators.
In: Martin, E.R. (Ed.), Enviromental Micropaleontology. Topics
in Geobiology, vol. 15, Kluwer Academic/Plenum Publishers,
New York, pp. 3967.
Donnici, S., Serandrei Barbero, R., 2002. The benthic foraminiferal
communities of the northern Adriatic continental shelf. Mar.
Micropaleontol. 44, 93123.
Embry, A.F., 1993. Transgressiveregressive (TR) sequence analysis
of the Jurassic succession of the Sverdrup Basin, Canadian Arctic
Archipelago. Can. J. Earth Sci. 30, 301320.
Embry, A.F., 1995. Sequence boundaries and sequence hierarchies:
problems and proposals. In: Steel, R.J., Felt, V.L., Johannesses,
E.P., Mathieu, C. (Eds.), Sequence stratigraphy on the Northwest
European Margin. Norwegian Petr. Soc. Spec. Publ., vol. 5,
pp. 111.
Ferranti, L., Antonioli, F., Mauz, B., Amorosi, A., Dai Pra, G.,
Mastronuzzi, G., Monaco, C., Orr, P., Pappalardo, M., Radtke, U.,
Renda, P., Romano, P., Sans, P., Verrubbi, V., 2006. Markers of the
last interglacial sea-level high stand along the coast of Italy: tectonic
implications. Quat. Int. 145146, 3054.
Fiorini, F., Vaiani, S.C., 2001. Benthic foraminifers and transgressive
regressive cycles in the Late Quaternary subsurface sediments of
the Po Plain near Ravenna (Northern Italy). Boll. Soc. Paleontol.
Ital. 40, 357403.
Follieri, M., Magri, D., Sadori, L., 1988. 250,000-year pollen record
from Valle di Castiglione (Roma). Pollen Spores 30, 329356.
Foyle, A.M., Oertel, G.F., 1997. Transgressive systems tract de-
velopment and incised-valley fills within a Quaternary estuary-
shelf system: Virginia inner shelf, USA. Mar. Geol. 137,
227249.
Guiter, F., Andrieu, V., de Beaulieu, J.-L., Cheddadi, R., Calvez, M.,
Ponel, P., Reille, M., Keller, T., Goeury, C., 2003. The last climatic
cycles in Western Europe: a comparison between long continuous
lacustrine sequences from France and other terrestrial records.
Quat. Int. 111, 5974.
227
Guiot, J., Pons, A., de Beaulieu, J.L., Reille, M., 1989. A
140,000 years continental climate reconstruction from two
European pollen records. Nature 338, 309313.
Hanebuth, T.J.J., Saito, Y., Tanabe, S., Vu, Q.L., Ngo, Q.T., 2006. Sea
levels during late marine isotope stage 3 (or older?) reported from
the Red River delta (northern Vietnam) and adjacent regions. Quat.
Int. 145146, 119134.
Havinga, A.J., 1984. A 20-year experimental investigation into the
differential corrosion susceptibility of pollen and spores in various
soil types. Pollen Spores 26, 541558.
Hearty, P.J., Dai Pra, G., 1987. Ricostruzione paleogeografica degli
ambienti litoranei quaternari della Toscana e del Lazio setten-
trionale con l'impiego dell'aminostratigrafia. Boll. Serv. Geol. Ital.
106, 189224.
Henderson, P.A., 1990. Freshwater ostracods. In: Kermack, D.M.,
Barnes, R.S.K. (Eds.), Synopses of the British Fauna (New Series),
vol. 42, Brill E.J., Leiden.
Hori, K., Saito, Y., Zhao, Q., Cheng, X., Wang, P., Sato, Y., Li, C.,
2001. Sedimentary facies of the tide-dominated paleo-Changjiang
(Yangtze) estuary during the last transgression. Mar. Geol. 177,
331351.
Hori, K., Saito, Y., Zhao, Q., Wang, P., 2002. Evolution of the coastal
depositional systems of the Changjiang (Yangtze) River in response
to Late PleistoceneHolocene sea-level changes. J. Sediment. Res.
72, 884897.
Hori, K., Tanabe, S., Saito, Y., Haruyama, S., Nguyen, V., Kitamura, A.,
2004. Delta initiation and Holocene sea-level change: example
from the Song Hong (Red River) delta, Vietnam. Sediment. Geol.
164, 237249.
Jalut, G., Amat, A.E., Bonnet, L., Gauquelin, T., Fontugne, M., 2000.
Holocene climatic changes in the Western Mediterranean, from
south-east France to south-east Spain. Palaeogeogr. Palaeoclimatol.
Palaeoecol. 160, 255290.
Jorissen, F.J., 1987. The distribution of Benthic Foraminifera in the
Adriatic Sea. Mar. Micropaleontol. 12, 2148.
Jorissen, F.J., 1988. Benthic Foraminifera from the Adriatic Sea;
principles of phenotypic variation. Utrecht Micropaleontol. Bull.
37, 1176.
Konradi, P.B., Larsen, B., Srensen, A.B., 2005. Marine Eemian in the
Danish eastern North Sea. Quat. Int. 133134, 2131.
Li, C., Wang, P., Sun, H., Zhang, J., Fan, D., Den, GB., 2002. Late
Quaternary incised-valley fill of the Yangtze delta (China): its
startigraphic framework and evolution. Sediment. Geol. 152, 133158.
Lim, D.I., Park, Y.A., 2003. Late Quaternary stratigraphy and
evolution of a Korean tidal flat, Haenam Bay, Southeastern Yellow
Sea, Korea. Mar. Geol. 193, 177194.
Lowrie, A., Hamiter, R., 1995. Fifth and sixth order eustatic events
during Holocene, fourth order highstand influencing Mississippi
delta-lobe switching. In: Finkl Jr., C.W. (Ed.), Holocene Cycles:
Climate, Sea Levels and Sedimentation. J. Coastal Res., vol. 17,
pp. 225229.
Magri, D., 1999. Late Quaternary vegetation history at Lagaccione near
Lago di Bolsena (central Italy). Rev. Palaeobot. Palynol. 106, 171208.
Magri, D., Sadori, S., 1999. Late Pleistocene and Holocene pollen
stratigraphy at lago di Vico, central Italy. Veget. Hist. Archaeobot.
8, 247260.
Mariani, M., Prato, R., 1988. I bacini neogenici costieri del margine
tirrenico: approccio sismico stratigrafico. Mem. Soc. Geol. Ital. 41,
519531.
Mauz, B., 1999. Late Pleistocene records of littoral processes at the
Tyrrheniana Coast (Central Italy): depositional environments and
luminescence chronology. Quat. Sci. Rev. 18, 11731184.
228 M. Aguzzi et al. / Sedimentary Geology 202 (2007) 211229
Massari, F., Rio, D., Serandrei Barbero, R., Asioli, A., Capraro, L.,
Fornaciari, E., Vergerio, P.P., 2004. The environment of Venice
area in the past two million years. Palaeogeogr. Palaeoclimatol.
Palaeoecol. 202, 273308.
McCarthy, P.J., Plint, A.G., 1998. Recognition of interfluve sequence
boundaries: integrating paleopedology and sequence stratigraphy.
Geology 26, 387390.
Milli, S., 1997. Depositional setting and high-frequency sequence
stratigraphy of the Middle-Upper Pleistocene to Holocene deposits
of the Roman Basin. Geol. Rom. 33, 99136.
Murray, J.W., 1991. Ecology and Palaeoecology of Benthic Forami-
nifera. Longman Scientific & Technical, London.
Nichol, S.L., Boyd, R., Penland, S., 1994. Stratigraphic response of
wave-dominated estuaries to different relative sea-level changes
and sediment supply histories: Quaternary case studies from Nova
Scozia, Louisiana and Eastern Australia. In: Dalrymple, R.W.,
Boyd, R., Zaitlin, B.A. (Eds.), Incised-Valley Systems: Origin and
Sedimentary Sequences. Soc. Econ. Paleontol. Mineral. Spec.
Spec. Publ., vol. 51, pp. 266283.
Nicol-Pichard, S., Dubar, M., 1998. Reconstruction of Late-glacial and
Holocene environments in south-east France based on the study of
a 66-m long core from Biot, Alpes Maritimes. Veget. Hist.
Archaeobot. 7, 1115.
Nummedal, D., Swift, D.J.P., 1987. Transgressive stratigraphy at
sequence-bounding unconformities: some principles derived
from Holocene and Cretaceous examples. In: Nummedal, D., Pilkey,
O.H., Howard, J.D. (Eds.), Sea-Level Fluctuation and Coastal
Evolution. Soc. Econ. Paleontol. Mineral. Spec. Spec. Publ., vol. 41,
pp. 241260.
Overeem, I., Kroonenberg, S.B., Veldkamp, A., Groenesteijn, K.,
Rusakov, G.V., Svitoch, A.A., 2003. Small-scale stratigraphy in a
large ramp delta: recent and Holocene sedimentation in the Volga
delta, Caspian Sea. Sediment. Geol. 159, 133157.
Pascucci, V., 2005. Neogene evolution of the Viareggio Basin
Northern Tuscany (Italy). Geoacta 4.
Prez-Obiol, R., Juli, R., 1994. Climatic change on the Iberian
Peninsula recorded in a 30,000-yr pollen record from Lake
Banyoles. Quat. Res. 41, 9198.
Plint, A.G., Nummedal, D., 2000. The falling stage systems tract:
recognition and importance in sequence stratigraphic analysis. In:
Hunt, D., Gawthorpe, R.L. (Eds.), Sedimentary Response to Forced
Regressions. Geol. Soc. London, Spec. Publ., vol. 172, pp. 117.
Preusser, F., 2004. Towards a chronology of the Late Pleistocene in the
northern Alpine Foreland. Boreas 33, 195210.
Pugliese, N., Stanley, D.J., 1991. Ostracoda, depositional environ-
ments and Late Quaternary evolution of the eastern Nile delta,
Egypt. Quaternario 4, 275302.
Reille, M., Andrieu, V., de Beaulieu, J.-L., Guenet, P., Goeury, C.,
1998. A long pollen record from Lac du Bouchet, Massif Central,
France, for the period ca. 325 to 100 ka (OIS 9c to OIS 5e). Quat.
Sci. Rev. 17, 11071123.
Ricci Lucchi, M., Calderoni, G., Carrara, C., Cipriani, N., Esu, D.,
Ferreli, L., Girotti, O., Gliozzi, E., Lombardo, M., Longinelli, A.,
Magri, D., Nebbiai, M., Ricci Lucchi, F., Vigliotti, L., 2000. Late
Quaternary record of the Rieti Basin, central Italy: paleoenviron-
mental and paleoclimatic evolution. G. Geol. 62, 105135.
Ridente, D., Trincardi, F., 2002. Eustatic and tectonic control on
deposition and lateral variability of Quaternary regressive
sequences in the Adriatic basin (Italy). Mar. Geol. 184, 273293.
Romano, P., Santo, A., Voltaggio, M., 1994. L'evoluzione morfologica
della pianura del fiume Volturno (Campania) durante il tardo
Quaternario. Quaternario 7, 4156.
Ruiz, F., Gonzalez-Regalado, M.L., Baceta, J.I., Menegazzo-Vitturi, L.,
Pistolato, M., Rampazzo, G., Molinaroli, E., 2000. Los ostrcodos
actuales de la laguna de Venecia (NE de Italia). Geobios 33,
447454.
Russo Ermolli, E., di Pasquale, G., 2002. Vegetation dynamics of
south-western Italy in the last 28 kyr inferred from pollen analysis
of a Tyrrhenian Sea core. Veget. Hist. Archaeobot. 11, 211219.
Saito, Y., 1994. Shelf sequence and characteristic bounding surfaces in
a wave-dominated setting: Latest PleistoceneHolocene examples
from Northeast Japan. Mar. Geol. 120, 105127.
Sarti, G., Zanchetta, G., Ciulli, L., Colonese, A., 2005. Late Quaternary
oligotypical non-marine mollusc fauna from southern Tuscany:
climatic and stratigraphical implications. Geoacta 4.
Sgarrella, F., Moncharmont Zei, M., 1993. Benthic Foraminifera of the
Gulf of Naples (Italy): systematics and autoecology. Boll. Soc.
Paleontol. Ital. 32, 145264.
Smith, A.J., Horne, D.J., 2002. Ecology of marine, marginal marine
and nonmarine ostracodes. In: Holmes, J.A., Chivas, A.R. (Eds.),
The Ostracoda Applications in Quaternary Research. Geophysical
Monograph, vol. 131, AGU, Washington DC, pp. 3764.
Somoza, L., Barnolas, A., Arasa, A., Maestro, A., Rees, J.G.,
Hernandez-Molina, F.J., 1998. Architectural stacking patterns of
the Ebro delta controlled by Holocene high-frequency eustatic
fluctuations, delta-lobe switching and subsidence processes.
Sediment. Geol. 117, 1132.
Stockmarr, J., 1971. Tablets with spores in absolute pollen analysis.
Pollen Spores 13, 615621.
Storms, J.E.A., Hoogendoorn, R.M., Dam, R.A.C., Hoitink, A.J.F.,
Kroonenberg, S.B., 2005. Late-Holocene evolution of the Mahakam
delta, East Kalimantan, Indonesia. Sediment. Geol. 180, 149166.
Stuiver, M., Reimer, P.J., Braziunas, T.F., 1998. High-precision
radiocarbon age calibration for terrestrial and marine samples.
Radiocarbon 40 (3), 11271151.
Swift, D.J.P., 1968. Coastal erosion and transgressive stratigraphy.
J. Geol. 76, 444456.
Ta, T.K.O., Nguyen, V.L., Tateishi, M., Kobayashi, I., Saito, Y., 2001.
Sedimentary facies, diatom and foraminifer assemblages in a late
PleistoceneHolocene incised-valley sequence from the Mekong
River Delta, Bentre Province, Southern Vietnam: the BT2 core.
J. Asian Earth Sci. 20, 8394.
Tanabe, S., Saito, Y., Sato, Y., Suzuki, Y., Sinsakul, S., Tiyapairach, S.,
Chaimanee, N., 2003. Stratigraphy and Holocene evolution of the mud-
dominated Chao Phraya delta, Thailand. Quat. Sci. Rev. 22, 789807.
Trnqvist, T.E., Wallinga, J., Murray, A.S., De Wolf, H., Cleveringa, P.,
De Gans, W., 2000. Response of the RhineMeuse system (west-
central Netherlands) to the last Quaternary glacio-eustatic cycles: a
first assessment. Glob. Planet. Change 27, 89111.
Trnqvist, T.E., Wallinga, J., Busschers, F.S., 2003. Timing of the
last sequence boundary in a fluvial setting near the highstand
shoreline insights from optical dating. Geology 31, 279282.
Tzedakis, P.C., 1993. Long term tree populations in northwest Greece
through multiple Quaternary climatic cycles. Nature 364, 437440.
Tzedakis, P.C., 1999. The last climatic cycle at Kopais, central Greece.
J. Geol. Soc. 156, 425434.
Tzedakis, P.C., Andrieu, V., de Beaulieu, J.-L., Crowhurst, S., Follieri, M.,
Hooghiemstra, H., Magri, D., Reille, M., Sadori, L., Shackleton, N.J.,
Wijmstra, T.A., 1997. Comparison of terrestrial and marine records of
changing climate of the last 500,000 years. Earth Planet. Sci. Lett.
150, 171176.
Tzedakis, P.C., Frogley, M.R., Heaton, T.H.E., 2003. Last Interglacial
conditions in southern Europe: evidence from Ioannina, northwest
Greece. Glob. Planet. Change 36, 157170.
 M. Aguzzi et al. / Sedimentary Geology 202 (2007) 211229
Van Wagoner, J.C., Mitchum, R.M., Campion, K.M., Rahmanian, V.D.
(Eds.), 1990. Siliciclastic Sequence Stratigraphy in Well Logs,
Cores, and Outcrops: Concepts for High-Resolution Correlation of
Time and Facies. Am. Assoc. Petrol. Geol. Methods in Exploration
Series, vol. 7, AAPG, Tulsa (55 pp.).
Wijmstra, T.A., 1969. Palynology of the first 30 m of a 120 m deep
section in Northern Greece. Acta Bot. Neerl. 18, 511527.
Wijmstra, T.A., Smit, A., 1976. Palynology of the middle part
(3078 meters) of the 120 m deep section in Northern Greece.
Acta Bot. Neerl. 25, 297312.
Woillard, G.M., 1978. Grand Pile peat bog: a continuous pollen record
for the last 140,000 years. Quat. Res. 9, 121.
229
Yassini, I., Jones, B.G., 1995. Foraminiferida and Ostracoda from
Estuarine and Shelf Environments on the Southeastern Coast of
Australia. University of Wollongong Press, Wollongong.
Zaitlin, B.A., Dalrymple, R.W., Boyd, R., 1994. The stratigraphic
organization of incised-valley systems associated with relative sea-
level change. In: Dalrymple, R.W., Boyd, R., Zaitlin, B.A. (Eds.),
Incised-Valley Systems: Origin and Sedimentary Sequences. Soc.
Econ. Paleontol. Mineral. Spec. Spec. Publ., vol. 51, pp. 4560.
Zagwijn, W.H., 1996. An analysis of Eemian climate in western and
central Europe. Quat. Sci. Rev. 15, 451469.
 Available online at www.sciencedirect.com

Sedimentary Geology 202 (2007) 230 248

www.elsevier.com/locate/sedgeo

Combining sedimentological and geophysical data for

high-resolution 3-D mapping of fluvial architectural
elements in the Quaternary Po plain (Italy)
R. Bersezio a,c,, M. Giudici b , M. Mele a
a
Earth Science Department, University of Milan, via Mangiagalli 34, I-20133 Milano, Italy
b
Earth Science Department, University of Milan, via Cicognara 7, I-20139 Milano, Italy
c
CNR IDPA, Via Mangiagalli 34, I-20133 Milano, Italy

Abstract

Current approaches to the reconstruction of the geometry of fluvial sediments of Quaternary alluvial plains and the
characterization of their internal architecture are strongly dependent on core data (1-D). Accurate 2-D and 3-D reconstructions and
maps of the subsurface are needed in hydrostratigraphy, hydrogeology and geotechnical studies. The present study aims to: 1)
improve current methods for geophysical imaging of the subsurface by means of VES, ERGI and GPR data, and calibration with
geomorphological and geological reconstructions, 2) optimize the horizontal and vertical resolution of subsurface imaging in order
to resolve sedimentary heterogeneity, and 3) check the reliability/uncertainty of the results (maps and architectural reconstructions)
by comparison with exposed analogues. The method was applied to shallow (0 to 15 m) aquifers of the fluvial plain of southern
Lombardy (Northern Italy). At two sites we studied fluvial sediments of meandering systems of the Last Glacial Maximum and
post-glacial historical age. These sediments comprise juxtaposed and superimposed gravelsand units with fining-upward
sequences (channel-bar depositional elements), which are separated by thin and laterally discontinuous silty and sandy clay units
(overbank and flood plain deposits). The sedimentary architecture has been studied at different scales in the two areas.
At the scale of the depositional system, we reconstructed the subsurface over an area of 4 km2 to a depth of 18 m (study site 1).
Reconstructed sequences based on 10 boreholes and water-well stratigraphic logs were integrated with the interpretation of 10
vertical electrical soundings (VES) with Schlumberger arrays and 1570 m long dipoledipole electrical resistivity ground imaging
profiles (ERGI). In unsaturated sediments, vertical and horizontal transitions between gravelsand units and fine-grained sediments
could be mapped respectively at the meter- to decameter scale after calibration of the VES with borehole data. Similar information
could be obtained in waterlogged sediments, in which the largest units could be portrayed and the lateral continuity of major
hydrostratigraphic units could be assessed. Maps of apparent resistivity were combined with sand-to-clay ratio maps obtained from
stratigraphic data, which substantially increased their quality. ERGI profiles added substantial information about the horizontal
transitions between fine- and coarse-grained units. At the scale of depositional elements (channel-bar systems) we studied quarry
exposures, over an area of about 4000 m2, down to 8 m below ground level (study site 2). In this case, facies analysis was
performed on progressing quarry faces and integrated with a network of 165 m long ERGI profiles and 1100 m long ground-
penetrating radar (GPR) profiles. Channel boundaries and accretion surfaces of point bars were resolved by both GPR and ERGI,
which permitted 3-D mapping of these surfaces.
Comparison between the results obtained for the two study sites demonstrates that integration of sedimentological data with
geophysical imaging (ERGI and VES) enables the identification of stratigraphic units at the scale of depositional elements.
Moreover, fining-upward trends and other internal features of the deposits, such as the transitions from coarse to fine-grained

Corresponding author. Earth Science Department, University of Milan, via Mangiagalli 34, I-20133 Milano, Italy.
E-mail address: riccardo.bersezio@unimi.it (R. Bersezio).

0037-0738/$ - see front matter 2007 Elsevier B.V. All rights reserved.
doi:10.1016/j.sedgeo.2007.05.002
 R. Bersezio et al. / Sedimentary Geology 202 (2007) 230248
sediments within channel-bar complexes, could be resolved. Hence, the combination of sedimentological and geophysical methods
provides a more accurate 3-D reconstruction of hydrostratigraphically significant sedimentary units compared to reconstructions
based solely on borehole/point data.
2007 Elsevier B.V. All rights reserved.
Keywords: Aquifer characterization; Electrical resistivity ground imaging; Fluvial sediments; Ground-penetrating radar; Hydrogeophysics;
Hydrostratigraphy
1. Introduction
Knowledge of the detailed geometry and architecture
of fluvial sequences is crucial for hydrogeological
studies, subsurface granular and aggregate mapping
and for the development of quantitative models.
Unfortunately, alluvial basin fills are extremely complex
and reconstructions are often oversimplified models
based on borehole data. In the field of hydrogeological
applications, general agreement exists about the ideal
procedure for aquifer characterization, as recently
summarized for example by Bridge and Hyndman
(2004). According to these authors, the first stage
includes a) stratigraphic correlation of borehole and core
data, b) analysis of the sedimentological and physical
properties of sedimentary units, and assessment of their
shapes, and c) determination of horizontal and vertical
continuity of sediment packages between borehole data,
assisted by geophysical exploration tools. The next
stage comprises 3-D modeling of (hydro)-facies distri-
bution, porosity and permeability by geostatistical
methods. North (2005), who summarized some pitfalls
of the most popular stochastic methods and geophysical
techniques used to interpolate between boreholes,
concluded that a good understanding of fluvial sedi-
mentology and stratigraphic evolution is required for
spatial modeling of aquifers and reservoirs. In this paper
we focus on assessment of the spatial continuity of
sedimentary units and their hydrological properties
between boreholes. We attempt to improve upon
existing methods by combining geological, geomorpho-
logical, stratigraphical and sedimentological data with
common, quickly employable and inexpensive geo-
physical techniques. In our case study, the integration of
different data sets yields good proxies of the location of
features relevant to hydrogeology at different scales,
such as the horizontal transitions and interfaces between
sandgravel bodies and fine-grained, less permeable
units.
Several geophysical techniques are available to
explore the stratigraphy of alluvial basins. High-
resolution shallow seismic (HRSS) is a powerful tool
231
for obtaining high-quality stratigraphic images in the
presence of a reasonably high P-wave velocity contrast
between lithologies. Such contrasts are generally present
if clayshale units alternate with coarse-grained units
(Birkelo et al., 1987; Geissler, 1989; Bruno and Godio,
1997; Bradford et al., 1998; Liberty, 1998; Whiteley
et al., 1998; Juhlin et al., 2002; Bradford, 2002a,b;
Francese et al., 2002, 2005). Unfortunately, HRSS is an
expensive and time-consuming method that may entail
environmental and security risks.
Ground-penetrating radar (GPR) is at present very
popular, and many aquifer analogue studies have been
published so far (e.g. Jol and Smith, 1991; Stephens,
1994; Asprion and Aigner, 1999; Bristow and Jol, 2003;
Heinz and Aigner, 2003). This technique is fast,
inexpensive and relatively easy to use. Moreover, it
can provide very high-resolution images if a small
penetration depth is sufficient. Unfortunately, this
technique cannot be readily applied in the presence of
even very thin soil or shallow clay layers, and
significant images of saturated sediments below the
water table are difficult to obtain.
The inversion of seismic, electric, and electromag-
netic data collected with cross-borehole acquisition
arrays is another powerful geophysical tool, which
requires appropriate wells. This technique cannot be
applied with existing water wells, so that it tends to
become very expensive (e.g., Tronicke et al., 2004;
Binley et al., 2004). Also direct-push electrical conduc-
tivity logging has been used to characterize unconsol-
idated sediments, providing vertical logs with an
extremely high resolution (Schulmeister et al., 2004).
However, horizontal correlations between measure-
ments are hampered by the uncertainties associated
with the use of 1-D data.
The most common geoelectrical techniques are
vertical electrical soundings (VES) and electrical
resistivity ground imaging (ERGI) (Telford, 1990;
Reynolds, 1996; Dahlin, 2001). These techniques have
been traditionally used in hydrogeology and other
environmental applications (Reynolds, 1996; Daily
and Ramirez, 2000) but are rarely applied for
232 R. Bersezio et al. / Sedimentary Geology 202 (2007) 230248

characterizing aquifer sediments at different scales
(Zalasiewicz et al., 1985; Alfano and Mancuso, 1996;
Baines et al., 2002; Giudici et al., 2003; Gourry et al.,
2003; Lapenna and Rizzo, 2003; Zaleha et al., 2005).
These techniques can be employed in many different
geological and environmental settings but they yield
mostly average images of the investigated volumes of
sediments so that the lithological boundaries are not
accurately resolved. In addition, resolution decreases
with increasing investigation depth. As a rule of thumb,
the spatial resolution of ERGI surveys is approximately
one-half of the electrode spacing (Baines et al., 2002),
whereas the maximum depth of investigation is about
one-fifth of the total length of the survey line (Barker,
1989; Ward, 1990). Using both VES and ERGI is
worthwhile, because they are well-established techni-
ques, user-friendly, inexpensive, rapid and not harmful
to the environment.
Baines et al. (2002) demonstrated the usefulness of
ERGI for mapping the geometry of channel-belts and
valley fills, which were independently constrained by
the properties of pore waters. Such applications require
that calibration with local borehole data is possible and
sufficient lithological contrasts are present. Gourry et al.
(2003) obtained a similar result from ERGI in a fluvial
plain with incised meandering palaeo-channels (Val de
Loire), that is, in a situation comparable to our local-
scale case (site 1, palaeo-Sillaro fluvial plain and aban-
doned meandering river; see the following Section 3).
Furthermore, Zaleha et al. (2005) were able to resolve
fluvial and glacio-fluvial depositional elements with
ERGI. These promising results obtained by ERGI for

Fig. 1. Sketch map of the study area with location of study site 1, the subsurface study area, and study site 2, the quarry aquifer analogue.
 R. Bersezio et al. / Sedimentary Geology 202 (2007) 230248
different scales of alluvial aquifer units (Baines et al.,
2002; Gourry et al., 2003; Lapenna and Rizzo, 2003
among others) encouraged us to combine VES and
ERGI data with geological, stratigraphical and sedi-
mentological data. Our purpose is to develop a
stratigraphic model in which detailed vertical resolution
of boreholes (centimeters to decimeters) can be
combined with the horizontal resolution provided by
electrical resistivity soundings (meters to decameters).
This integrated method, which represents an improve-
ment over the conventional approach based on correla-
tion of borehole and logging data (1-D), will be used to
map buried fluvial deposits and to obtain a fully 3-D
characterization of hydrostratigraphic units. We applied
this integrated approach to the Upper Pleistocene
Holocene alluvial succession of two study areas in the
Po plain of Lombardy, south of Milan, between the
Lambro and Adda Rivers (Fig. 1). The two sites have
been selected for two different-scale analyses: a local-
scale study of a real aquifer complex in the subsurface of
Fig. 2. Flow chart of the methodology for 3-D mapping of fluvial architectural elements at sites 1 and 2, in the Quaternary Po plain.
233
study site 1, and a detailed-scale study of an aquifer
analogue exposed at site 2 (Fig. 1).
2. Methods and techniques
A flow chart of the studies carried out at sites 1 and 2
is presented in Fig. 2. At both sites a preliminary
geological and stratigraphic reconstruction was obtained
by geological and geomorphological surface mapping
(1:10,000) and correlation of borehole logs (site 1) or
sedimentological logs measured along vertical excava-
tion fronts (site 2).
At site 1, after the elaboration of the qualitative
geological model, we performed 10 VES with the
Schlumberger array (maximum AB/2 150 m) (Figs. 3,
4 and 5). The resistivity profiles were calibrated to the
nearest available stratigraphic logs. The interpretation
was made by trial-and-error fitting of field curves with
those computed for a layered structure with the linear
digital filter method (Koefoed, 1979). Interpolation
234 R. Bersezio et al. / Sedimentary Geology 202 (2007) 230248
Fig. 3. VES (Schlumberger array) and ERGI (dipoledipole array) techniques for acquisition of subsurface electrical resistivity images. The general
equations for calculating apparent resistivity (A) are also indicated.
of the apparent resistivities measured for different
values of AB/2 (Fig. 3) yielded a set of maps of
apparent resistivity (Fig. 6). The integration of
geological and resistivity models allowed us to locate
the boundaries of the major sedimentary units and the
interfaces between coarse permeable sediments (sand
and gravel) and fine-grained units (very fine sand, silt
and clay). These results were used to guide the
collection of 10 ERGI profiles with dipoledipole
array (1570 m total length; dipole spacing 5 m), whose
aim was to identify the precise position and shape of
these interfaces (Fig. 7).
Site 2 is about 6 km south of site 1 (Fig. 1). During
2004 and 2005, Holocene to recent sediments of the
Lambro River were exposed in a gravel quarry at this
site, which allowed us to construct a 3-D map of
30 40 6 m3 before and during excavation (Fig. 8).
Three orthogonal faces of the quarry were studied in
detail. Individual beds and facies association were
mapped in the field as well as on orthonormalized
photos. Twenty-seven sedimentological logs were
obtained (with centimetre resolution), correlated directly
in the field and displayed on vertical facies maps
(1:100). In addition, an intermediate face of excavation
was mapped based on a line drawing of a photomosaic
and three sedimentological logs (Fig. 8). The same
volume was studied by a geophysical survey that was
carried out on the top surface of the quarry, a near-
horizontal surface from which the shallow soil had been
removed. This preliminary GPR survey (400 MHz
monostatic antenna, grid spacing 1 m, total profile
length 1100 m; Fig. 9) permitted us to map the major
boundary between the two highest-order stratigraphic
units of the excavated sediments in 3-D. Furthermore,
five intersecting ERGI profiles, 165 m long, with
dipoledipole array (Fig. 3) and dipole spacing of 1 to
3 m, were acquired (Figs. 8 and 10). The results of the
ERGI were directly compared to reality and to the map
drawn on the basis of observations and the GPR survey.
The sedimentary maps of the northern and eastern
quarry faces and the corresponding 20 sedimentological
logs (Fig. 8) were used to calibrate the interpretation of
the ERGI profiles. After the geophysical survey, maps of
the intermediate and southernmost exposures supported
 R. Bersezio et al. / Sedimentary Geology 202 (2007) 230248 235

Fig. 4. Geomorphological sketch of study site 1 (framed area), with location of boreholes (triangles), VES (dots, label S) and ERGI profiles (straight
solid lines, label L). Geomorphological symbols: solid lines = boundaries of the oldest terraces of the Sillaro palaeo-river valley (I) and of the Lambro
River valley (IV), ornamentation towards the lower surface; dashed lines (unornamented) = boundaries of minor terraces; dotted lines = interpreted
crest-lines of scroll bars; roman numbers = relative chronology of cut-offs among terrace scarps. Satellite view was derived from the Google Maps site
(http://maps.google.com/), on October, 10, 2004.

by 10 logs (Fig. 8) were obtained during short time
intervals between successive excavation phases. We
used this data set for the final verification of the
geophysical image. Therefore, at this site, integration of
sedimentological and geophysical data yielded a
reconstruction of the 3-D architecture of the aquifer
analogue, with a vertical resolution of 0.1 to 1 m and a
horizontal resolution of 1 to 10 m. The sandy and
gravelly beds exposed on the quarry faces represent
superimposed channel-bar depositional elements, with
rare intercalations of silt and clay lenses. At site 2, we
evaluated the internal architecture of depositional
elements comparable to those at site 1. The study of
this analogue was used to refine the interpretation of the
subsurface study, and helped us to reduce the gap
between the vertical and horizontal resolution for the
final map and the aquifer reconstruction.
3. Site 1 Upper Pleistocene succession of the
palaeo-Sillaro meandering River: depositional system
scale
At site 1 we investigated a buried succession of
Upper Pleistocene alluvial sediments of the palaeo-
valley of an abandoned river, the palaeo-Sillaro River
(Figs. 1 and 4). This area belongs to the Neogene Po
plain foredeep basin which developed between the
frontal thrusts of two diachronous orogens, the Southern
Alps to the north and the external arcs of the Northern
Apennines to the south. The area of Southern Lombardy
236 R. Bersezio et al. / Sedimentary Geology 202 (2007) 230248

represents the external ramp of the Apenninic foredeep
basin that was filled during Early to Middle Pleistocene
by at least three regressive cycles of marine to coastal
deposits. Deposition of alluvial and glacio-fluvial
sequences, fed from the alpine side, started after
regression, under the control of Middle to Late
Pleistocene glacial cycles (Bini, 1997b; ENI- Regione
Lombardia, 2002; Muttoni et al., 2003; Bersezio et al.,
2004).
The meandering valley of the Sillaro River devel-
oped during the last glacial maximum (LGM) as an
outflow from the Lecco glacier (Bini, 1997b) and
progressively entrenched into the pre-LGM plain, the
so-called Livello Fondamentale della Pianura Auct.
(Castiglioni and Pellegrini, 2001, and references
therein). Traces of the meandering river are preserved
in an abandoned valley bounded by 2 to 5 m high terrace
scarps (I in Fig. 4) with an average width of 3 km. The
well-preserved meander loops have a mean radius of
curvature of about 1 km, one order of magnitude larger
than the meanders of the Lambro River valley to the
West (Fig. 1). The post-glacial terrace scarps of the
Lambro River intersect the boundaries of the palaeo-
Sillaro (scarp IV in Fig. 4).

Fig. 5. Data from borehole stratigraphic logs (a) and VES (b) at study site 1. Legend and location of data are represented on the right-hand frames
(compare with Fig. 4). AB and CD are the traces of the cross-sections portrayed in Fig. 7.
 R. Bersezio et al. / Sedimentary Geology 202 (2007) 230248 237

Fig. 6. Apparent resistivity maps (a, b) and coarse/fine sediments ratio maps (c, d) at different depths, obtained from interpolation of geoelectrical and
stratigraphic data respectively. The AB/2 distances of the maps (a) and (b) represent investigation depths comparable to the depth at which are drawn
the maps in (c) and (d).

The fluvial sediments of the palaeo-Sillaro are coarse
to fine sands with gravels, grading upwards into silts and
clays. Individual upward-fining sequences are up to 6 m
thick. These sequences grade vertically and laterally to
siltclay units. The 14C results obtained from this
succession indicate ages between 22000 300 y B.P.
(topmost part of underlying sediments) and 19909 y cal
BC, which is consistent with an LGM age (Bersezio
et al., 2004).
Stratigraphic correlations between boreholes at the
study site (Figs. 4 and 5) show that the LGM fluvial
sediments of the palaeo-Sillaro are bounded below by an
erosional surface that truncates the Upper Pleistocene,
pre-LGM coarse-grained alluvial units attributed to the
Upper Pleistocene glacio-fluvial Besnate Allogroup (Da
Rold, 1990; Bini, 1997a,b; Bersezio et al., 2004).
The study site, which covers an area of about 4 km2,
is located close to the western boundary of the palaeo-
valley, where it is intersected by the most external scarp
of the Lambro system (Figs. 1 and 4). At this site we
investigated the entire succession of the palaeo-Sillaro
depositional system to a maximum depth of 18 m.
Water-well data show minor fluctuations of the water
table around a depth of 6 m below the ground surface.
Therefore the study includes both the non-saturated
zone and the phreatic aquifer below, that is, a volume of
sediments of primary interest for the definition of
surface infiltration and protection of groundwater.
3.1. Geology and stratigraphy
At study site 1 palaeo-Sillaro channel-bar deposi-
tional elements have been reconstructed (Fig. 4) from
field and photo-geomorphological surveys. The spatial
distribution of gravel, sand and siltclay surface sedi-
ments was mapped based on shallow natural exposures,
excavations and hand-drilled sampling (1 m deep), and
integrated with the morphological map. The final result
of Fig. 4 shows three abandoned meandering traces,
represented by channel-bar elements, with cross-cutting
238 R. Bersezio et al. / Sedimentary Geology 202 (2007) 230248
relationships. The oldest meanders are confined to the
eastern area (I in Fig. 4) and are cut by two more recent
meander traces (II and III in Fig. 4). The most recent one
(III in Fig. 4) is very well preserved and is bounded by a
3 m high terrace scarp. It truncates trace II which
occupies most of the study area and is cross-cut in the
west by the most outward situated terrace scarp of the
Holocene Lambro River (IV in Fig. 4).
A preliminary subsurface reconstruction based on
correlation of lithological data from 10 boreholes and
water wells shows 1) the Upper Pleistocene sands and
gravels of the Besnate Allogroup (unit 2 in Bersezio
et al., 2004; Figs. 5 and 7), 2) the lower erosional
boundary of the palaeo-Sillaro sediments (unit 3 after
Bersezio et al., 2004), and 3) unit 3 which consists of
two superimposed minor fining-upwards sequences.
The uppermost minor sequence of unit 3 represents
the deposits of meander belts II and III (Figs. 4 and 7).
Basal gravelly to sandy bedsets are interpreted as point-
bar deposits that accreted in EW direction and
migrated down-stream to the south-east. The gravelly
sandy bedsets are overlain by thick silt and sandysilt
units and are locally capped by laminated to massive
organic-rich clays, interpreted as fills of abandoned
channels. These are laterally equivalent to the fine-
grained flood plain deposits of the Sillaro palaeo-valley.
The lowermost minor sequence does not show relations
with the present-day topography in the study area. It
consists of gravel-to-sand and sand-to-clay upward-
fining sequences that are interpreted as channel-point
bar and flood plain deposits, respectively.
At site 1, deposition of unit 3 started at the end of the
Late Pleistocene Besnate glacial cycle (Bini, 1997b) as
shown by erosion at the top of these sediments.
Radiocarbon dating indicates that deposition of the
lowermost minor sequence of unit 3 started during the
LGM (Cant glacial cycle in the regional literature, Bini
et al., 1997b; Bersezio et al., 2004) and continued until
late- to post-glacial times, in a moderately entrenched
meander belt with confined flood plains. Later, the
palaeo-Sillaro River was abandoned after cut-off at the
former LGM position of the glacier mouth which had
retreated by that time (Bini et al., 2004). The Sillaro
depositional system was deactivated before the Holo-
cene, as testified by the clear preservation of its final
trace and by the cross-cutting relations with Holocene
rivers (Lambro to the west and Adda to the east).
3.2. Results of VES and ERGI profiles
Ten VES were carried out, eight of which were used
to develop the model and two to verify the results (VES
S9 and S10). The first step for the interpretation of the
field resistivity profiles was calibration of VES S3 and
S4 with wells AS004 and AS105 (Fig. 5). The
interpretation of the VES is shown in Figs. 5 and 7 in
the form of vertical logs. It provided the following
information:
i) The water table is mapped from any VES at a
depth close to that measured in the adjacent water
wells (67 m below the ground surface). Below
this depth, resistivity contrasts are attenuated but
still recognizable, thus allowing for lithological
interpretation by comparison with calibrated VES.
However, spatial variations of pore-water compo-
sition could affect the electrical resistivity and
prevent interpretation of the variations of bulk
electrical resistivity in terms of lithological
variations. Unfortunately we could not perform
direct measurements of the electrical conductivity
of pore waters, but the results of monitoring
surveys over a large area (including this study
site) do not show significant variations. Hence, we
assumed that spatial variations of pore-water
characteristics do not affect the electrical resistiv-
ity of the sediments.
ii) Shallow, highly-resistive electrical units in the
unsaturated zone can be confidently associated
with the uppermost gravelsand fining-upwards
sequence that characterizes unit 3 (Figs. 5 and 7).
The stratigraphic model showed a south-east-
wards fining of the shallowest sediments (transi-
tion from point-bar to channel and overbank
deposits) and the development of a siltclay
interface between the two minor sequences of unit
3. This is clearly illustrated by the resistivity
profiles (Figs. 5 and 7), which allowed us to widen
the area where this transition has been recognized
and to refine its position at the studied site.
Nevertheless, the geometry of this transition and
the relations between the sedimentary bodies are
still poorly resolved. This warranted the position-
ing of NWSE oriented ERGI profiles across this
transition (Figs. 5 and 7).
iii) In the saturated sediments, the lowermost minor
sequence of Unit 3 (gravel to sand) could be
locally separated from the underlying sequence
that is formed by gravelly sand grading upwards to
silty clays (Besnate Allogroup, Fig. 7). The
resistivity contrasts are weak but still interpretable.
Based on the VES, 10 ERGI profiles were obtained
to refine the geometry of the stratigraphic model
 R. Bersezio et al. / Sedimentary Geology 202 (2007) 230248
Fig. 7. Comparison between a stratigraphic cross-section (a), a VES correlation panel (b) and the ERGI profiles (b, c). The location of sections AB and CD is in Fig. 5. The blue dashed line represents

239
the average position of the water table.
240 R. Bersezio et al. / Sedimentary Geology 202 (2007) 230248

Fig. 8. Study site 2, quarry aquifer analogue. a) Aerial view of location of the quarry before excavation; b) the top of the quarry, where GPR and ERGI
were surveyed (white dashed lines: ERGI profiles); c) location of the stratigraphic logs and of the GPR-ERGI grids. The image presented in Fig. 10a
corresponds to the intermediate front of excavation (compare the position of the stratigraphic logs labeled with M).

(location in Figs. 4 and 5, results in Fig. 7). Calibration
was performed by comparison with the closest bore-
holes (AS105, S1, AS005 and AS106). All ERGI
profiles portray the water table at about 6 m below the
ground surface. However, the resistivity image of the
waterlogged sediments is clearly interpretable in terms
of lithostratigraphy (Figs. 5 and 7). The ERGI profile
AB (Figs. 4 and 7) illustrates the continuity and shape
of the shallowest electrical units, which are character-
ized by intermediate resistivity (200500 m) and can
be satisfactorily superimposed on the stratigraphic
scheme. In this way, the 2-D shape of the lowermost
gravelsand body of unit 3 is obtained along this
transect. The fine sediments that separate this body from
the uppermost sequence represent a conductive electri-
cal unit (b 50 m) that is clearly visible in profile L8
(Fig. 5). The lateral transition into fine-grained deposits
is quite clearly visible south-east of profile L10. Fine
sand to clay-grade channel fills are depicted by the thick
conductive lense in profile L9 (Fig. 7).
Three additional ERGI profiles, acquired in the area
between VES S7 and S8, some decameters south-west
of profile AB (ERGI L1, L2 and L3, Figs. 5 and 7),
were used to resolve the lateral transition of coarse to
fine sediments. This group of geophysical images
shows: i) the southeastern termination of the resistive
body corresponding to the point bar deposits of the
shallowest sequence of unit 3, ii) the concave-up
geometry of the adjacent fine-grained channel fill and
iii) the conductive (fine-grained) sediments that separate
the two superposed minor sequences of unit 3. The
shallow investigation depth of this detailed ERGI does
not reach the sediments of the Besnate Allogroup below
unit 3.
 R. Bersezio et al. / Sedimentary Geology 202 (2007) 230248 241

Fig. 9. a) Block diagram of sediments at study site 2, obtained from the GPR survey. Note the evidence of the erosion surface (thick black dashed line
labeled with ) at the base of the Renaissance age barchannel depositional element and the dipping sets within the same (thin black lines); b) GPR
profiles AB and BC (location is shown in panel a).

3.3. Data integration and discussion of results at site 1
The geomorphological, stratigraphic and geophysical
data provide consistent pictures of the sedimentary
architecture and can be integrated into a coherent model
of a part of the depositional system of the palaeo-Sillaro
meandering River, corresponding to unit 3. The model
shown in Figs. 47 allows for a reconstruction of the
evolution of the fluvial system that is consistent with the
regional geomorphological evolution, the history of LGM
onset and decay, the palaeo-Sillaro meandering River
dynamics, and the late switch-off of river discharge. The
geophysical data show a good match with the geological
description at the control sites where two VES were
surveyed after compilation of the model (S9 and S10).
The data set allows for mapping of relevant hydrostrati-
graphic properties of the non-saturated zone and of the
phreatic aquifer beneath it, within unit 3. These maps
include i) coarse-to-fine sediment ratio maps, (C/F: gravel
and sand vs. silt and mud) that are directly related to
porosity and permeability of sediments and can be derived
for the different sequences from the stratigraphic data, and
ii) apparent resistivity maps for different spacing of the
electrodes and different investigation depths. Some
examples of these maps, which refer to different depth
ranges of about 0 to 2.5 m and 7.5 to 10 m are shown in
Fig. 6. The C/F maps show a good match with the spatial
distribution of coarse point-bar sediments and fine
channel-fill and overbank sediments. The agreement of
apparent resistivity maps with the C/F maps is good for
shallow depths (Fig. 6a and c) and satisfactory for the
deepest sediments (Fig. 6b and d). Hence, it seems
possible to introduce the use of the apparent resistivity
maps as a complement to, or as a proxy for the sand/mud
ratio maps that are usually drawn to characterize the
hydrostratigraphic properties of aquifer complexes.
242 R. Bersezio et al. / Sedimentary Geology 202 (2007) 230248

Fig. 10. Comparison of a quarry exposure (a) with the relative stratigraphic logs (M1, M2, M3) and ERGI image (b: T2 section), based on
composition of two consecutive segments of the NS eastern face and of the intermediate EW face (location in Fig. 8c). The thick black dashed line
represents the erosion surface at the base of the Renaissance age barchannel depositional element, whereas the thin black lines represent the
dipping sandgravel sets.

A comparison of our case with recent literature (see
the brief summary in the introduction) suggests that
the methodology here adopted is effective in resolving
the external geometry of the bar, channel and over-
bank sedimentary bodies, and may be used to reduce
the gap between horizontal and vertical resolution of
borehole data. However, the obtained reconstruction
does not include the internal architecture of the
sedimentary units, because bedding and textural changes
(that relate to porosity and permeability) could not be
resolved in most of the investigated volume, far away
from the boreholes.
To test the effectiveness of the methodology at the fine
scale of the internal architecture of the barchannel
complexes, we applied a similar procedure to an exposed
succession that provides an analogue of the buried
depositional elements and aquifer units studied at site 1.
4. Site 2 Historical point-bar complexes of the
Lambro River: depositional elements scale
4.1. Geology and stratigraphy at site 2
Two superposed point-barchannel-fill depositional
elements belonging to the Holocene sediments of the
meandering Lambro River are exposed at site 2. The
depositional elements are largely composed of sands
and gravels. Very rarely silt and clay drapes and lenses
are present at the top of bar or channel sequences. In
other cases these represent the fill of abandoned chute
channels. Mud clasts are commonly present above the
erosion surface at the base of the bar and channel
sequences.
At the study site the present-day Lambro valley is
entrenched 15 m below the average elevation of the
 R. Bersezio et al. / Sedimentary Geology 202 (2007) 230248
regional flat top surface of the pre-LGM fluvial
sediments (Livello Fondamentale della Pianura
Auct.; Fig. 1). The studied succession belongs to the
lowermost terrace in the Lambro valley, which is
situated between the most external and elevated terrace
and the flood plain of the present-day river (Fig. 1). The
lowermost terrace is separated from the present-day
flood plain by a 7 m high scarp. The terrace is formed by
two superposed barchannel depositional elements that
are separated by a sub-horizontal erosion surface
(Figs. 9 and 10). The upper depositional element is of
MedievalRenaissance age because it contains typical
ceramic and brick artifacts (Fig. 11). It is truncated and
covered by a shallow sandgravel unit with a modern
soil profile, which was not considered because it was
removed at the start of excavation. The oldest and
lowermost depositional element contains artifacts that
are presumably of Roman age (M. Cremaschi, pers.
com., 2005). The archaeological materials are still under
study to derive a more precise age determination. These
sediments deposited in historical times directly overlie
the truncated top of the Late Pleistocene Besnate
Allogroup (Bersezio et al., 2004). This contact is not
exposed at present but can be easily recognized based on
lithological criteria in boreholes. This is due to the fact
that the drainage area of the HoloceneRecent Lambro
River is limited to the foothills of the Lombardian
Southern Alps, which provide sediments belonging to a
distinct petrofacies (mostly sedimentary lithoclasts of
JurassicCretaceous formations). The drainage area did
Fig. 11. Renaissance age point-bar at study site 2. Bricks (indicated by arrows) are imbricated together with lithic clasts (eastern NS striking quarry
face, close to the northwestern edge of the studied volume; view from the west).
243
not reach the main Alpine range that fed the Besnate
Pleistocene depositional systems with different clast
types (crystalline basement rocks and Palaeozoic
Triassic formations).
The uppermost barchannel depositional element is
a gravelsand lithosome, with a maximum thickness of
about 3 m that shows a well-developed lateral accretion
structure dipping on average to the south and south-west
(Figs. 9, 10 and 11). The gravelsand bedsets fine
upwards into trough cross-bedded sands interpreted as
dune deposits, and into fine-sand rippled units. A
transition to channel-fill deposits occurs to the SW,
where medium to fine sand concave-up bedsets drape
the basal erosion surface and are covered by fining-
upwards large-scale trough cross-bedded sands, and
finally by rippled bedsets of very fine sands. The
margins of the channel are not exposed. The upper part
of the barchannel depositional element is truncated by
chute channels, up to 10 m wide and 2 m deep, that are
filled by planar to trough cross-stratified gravels and
laminated sands, which form bars and dunes that are
abruptly capped by horizontal, planar laminated to
massive silty clay, clay and peat. The basal surface of the
uppermost barchannel depositional element is almost
flat at the centre of the quarry exposure and shows
concave-up scours that are some decimeters deep and
several meters wide, in the west.
The Roman age lowermost depositional element
below this surface consists of a more than 4 m thick
gravelsand succession, the base of which is not
244 R. Bersezio et al. / Sedimentary Geology 202 (2007) 230248

exposed (Fig. 12). It is formed by a SSW-dipping lateral
accretion complex up to 4 m in height, with gravelsand
inclined bedsets. Beds are typically decimeter-thick and
normally graded (Fig. 12a). Open-framework, well-
sorted fine gravels occur at the base of all bedsets. These
fine upwards, passing into stratified/laminated or
massive gravelly sands and sands that develop laterally
and upwards into sets of trough cross-bedded sands,
interpreted as dune deposits, and then to meter-thick,
fine to very fine rippled sand beds. The lateral accretion
complex progressively fines westwards and southwards
into sand with minor gravel, that is mostly restricted to
the base of cross-bedded sets (Fig. 12b). Several dis-
crete accretion phases are documented by erosion sur-
faces that obliquely cut the laterally accreted unit and
are covered by inclined sets with a slightly different dip
(Fig. 12b). The average direction of accretion varies
from SSE to SSW. The westernmost part of the expo-
sure (Fig. 12c and d) shows a transition into channel-
fill sand sediments, with large-scale concave-up sand
layers, that develop above a concave-up lower scour
about 15 m wide and 3 m deep, with a basal lag rich in
pebbles, mud cobbles and slide blocks of laminated
silty clay. The large-scale curved sets develop upwards
and laterally into massive sands and then to cross-
bedded cosets of well-sorted, medium to fine sand with
large-scale trough cross-bedding, which display an
upward reduction in set thickness (Fig. 12c and d). The
basal surface of the entire barchannel depositional
element emerges from the ground floor close to the
eastern corner of the exposure and rises to the east
where the bar terminates, presumably at the location
where it was attached to the bench.
The position of the piezometric surface at site 2 is just
below the lowermost quarry floor, at the base of the
exposed sediments. During our study it fluctuated in
response to the discharge of the nearby Lambro River, as
observed from temporary occurrences of springs at the
base of the quarry. Hence, the studied sediments were
not saturated, but characterized by variable pore
moisture content due to infiltration, evaporation and
capillarity.

Fig. 12. Views of the northernmost quarry face (location in Fig. 8c). This almost EW exposure shows: i) the two superimposed barchannel
depositional elements, ii) the bounding erosion surface between the Renaissance age and the Roman age point barchannel depositional elements,
and iii) the transition from point bar (pictures a and b) to channel-fill deposits (picture c and line drawing of the same photo in d) occurring in the
lowermost (Roman age) depositional element, from E to W (view of photos from the North; see main text for description). Location of photos is
indicated by numbers of the measured sections (S2 to S10) that are indicated also in the index map of Fig. 8c.
 R. Bersezio et al. / Sedimentary Geology 202 (2007) 230248 245

4.2. Geophysical imaging
The relevant features of the aquifer analogue archi-
tecture that we expected to resolve with geophysical
imaging at site 2 were: i) the surface bounding the two
superimposed point barchannel depositional ele-
ments, ii) the transition from the gravel core of the
two laterally accreted units to the correlative sandy
channel fills, and iii) the position of the water table.
The GPR survey allowed us to map in detail the
shape of the base of the Roman age unit (Fig. 9, surface
labeled with ). It also showed the inclined geometry of
the gravelsand accretion units and the transition
towards the almost transparent channel-fill sand
unit. Below this surface the GPR resolution decayed
rapidly, probably due to the strong contrast between the
dry gravels (above) and the wet sands below. Presum-
ably also the large number of concave-up features, at
different scales, disturbed the GPR image that was
almost completely transparent at the base.
The results of ERGI profiles are shown in Figs. 10
and 13. Fig. 10 provides a comparison between the
intermediate stratigraphic cross-section (M1M4 in
Fig. 8c) and the closest ERGI image (T2 in Fig. 8c);
Fig. 13 presents a block diagram of the four ERGI
profiles with a few stratigraphic logs for calibration.
Some encouraging results can be underlined:
1) The Roman and the Renaissance age depositional
elements are clearly separated in the ERGI image.
The geometry of the bounding surface () is
accurately depicted and coincides satisfactorily both
with the GPR image and the measured stratigraphic
logs; vertical and horizontal resolution in this case
are at decimeter to centimeter scale;
2) The position of the water table is correctly identified;
3) The eastern gravel core of the two lateral accretion
complexes has been clearly recognized and can be
distinguished from the western and southern sands
that fringe the bars and fill the lateral channels. The

Fig. 13. Block diagram obtained from the ERGI survey at study site 2. Three stratigraphic logs (M2, M3, M4) are shown for comparison. Location of ERGI
panels and stratigraphic logs is given in Fig. 8c. Note the good representation of the fining-upward trend of the Roman-age sediments, of the erosion surface
at its top and the sharp contrast with the overlying gravelsand sediments of the Renaissance age point-barchannel depositional element.
246 R. Bersezio et al. / Sedimentary Geology 202 (2007) 230248
south-westward dipping shape of these cores has
been observed also in the ERGI image, in this case
showing a meter-scale resolution;
4) Some minor features, for instance the sand lenses that
fill some of the chute-channels, have been recog-
nized, but their dimensions and positions could not
be accurately resolved;
5) Comparison of the ERGI model of the sediment
volume with the vertical maps and logs of the
southern and intermediate quarry faces provided a
satisfactory verification;
6) The horizontal resolution of the ERGI image, obtained
with 13 m electrode spacing and an investigation
depth of about 8 m, is much closer to the vertical
resolution of the stratigraphic logs than images
obtained with standard survey configurations. It lies
in the decimeter range, only one order of magnitude
below that of the stratigraphic logs. Obviously this is
due to the specific acquisition setup. Nevertheless, this
result encourages the study of analogues to improve
the interpretation of geophysical images of buried
sediments. The integration of different geophysical
tools (like VES, ERGI and GPR) with the geological
analysis seems to provide a relatively fast, efficient and
accurate methodology for our purpose.
5. Conclusions
1) This work aims to solve some problems related to
characterization of alluvial aquifers and mapping of
subsurface fluvial sediments, such as the definition and
quantification of thickness and continuity of imper-
meable fine-grained units, and the distribution, geo-
metry and connectivity of the most permeable gravel
sand bodies. These features are not satisfactorily
resolved by current subsurface reconstructions based
on correlation of 1-D stratigraphic data. Crucial
problems are: i) locations of lateral boundaries bet-
ween fine and coarse sediments are not obtained with
sufficient accuracy, ii) stratigraphic models and maps
are affected by unknown levels of uncertainty in the
correlations and iii) there is a discrepancy between
vertical and horizontal resolution of the reconstruc-
tions (tens of centimetres vs. tens to hundreds of
meters, respectively). Our approach is based on integ-
ration of current geological methods with different
geophysical techniques and on the comparative study
of subsurface sediments with exposed analogues. It
comprises the following steps: i) define the geological
model of the subsurface at the scale of the depositional
system, by application of current geomorphological
and stratigraphic methods to surface and subsurface
reconstruction; ii) use VES to refine the rough
stratigraphic model and to select sites for 2-D ex-
ploration; iii) obtain 2-D subsurface images along key
transects with ERGI; iv) redraw the stratigraphic
model; v) verify the new model by execution of some
new VES at selected sites; vi) compare the results with
the ERGI image of the analogue case to check re-
solution and significance of resistivity images at the
fine scale of architectural elements (sensu Miall, 1985).
2) The method permits 3-D reconstruction and charac-
terization of shallow aquifers, with a comparable
accuracy (decimeters to meters) in vertical and
horizontal dimensions, which permits a reconstruction
of the external geometry of architectural elements
(gravelsand bars, channel fills; Miall, 1985) and of
their bounding surfaces. Comparable conclusions have
been drawn by other authors in different basins, at the
larger scale of depositional systems or depositional
elements. The present study demonstrates that the
accuracy of this approach can be improved and adds
new evidence to support the general validity and
relevance of the methodology. Moreover we demon-
strate that a comparison with exposed analogues helps
to improve the interpretation of the geophysical image
and allows us to add information on the internal
architecture of these depositional elements. In fact, the
qualitative comparison between the geophysical
images that we obtained at the subsurface study site
1 and at the analogue site 2 (Figs. 5, 7 and 13)
respectively) provided an indirect verification of the
geological interpretation given at site 1. The compar-
ison showed that: i) the individual depositional units
(beds or bedsets, at the physical scale of dm to m) can
be distinguished if either gravel/sand on silt/clay or
gravel on sand contacts are present; ii) the boundaries
of channel-bar depositional elements can be recog-
nized and mapped. The geophysical images of these
features at the subsurface and analogue study sites are
clearly comparable, which provides strong support for
interpretation of the subsurface; iii) fining-upward
sequences can be recognized by resistivity imaging,
both in unsaturated and in waterlogged sediments, at
different scales (decimeters to meters) and hierarchic
ranks at the two study sites.
3) The ERGI and VES techniques also yield useful results
below the water table, although resolution in depth is
attenuated and resistivity contrasts are smaller than in
unsaturated sediments. Uncertainties associated with
imaging of saturated sediments (aquifers) are attribut-
able to the possibly low resistivity of high-permeable
clean gravels, and to pore-water chemistry. Neverthe-
less, calibration with borehole data allowed us to
 R. Bersezio et al. / Sedimentary Geology 202 (2007) 230248
associate the upward and lateral decrease of resistivity
with gravel-to-sand sequences capped by fine-grained
layers. This suggests that although water saturation
attenuates resistivity contrasts, it does not prevent the
recognition of gravel/sand and silt/clay layers.
4) The maps of apparent resistivity we obtained from
VES and ERGI for different values of the Schlumber-
ger array length (i.e., at different depths of investiga-
tion) represent a proxy of the maps of the Coarse/Fine
sediment ratio, which is used to characterize aquitard
efficiency and aquifer storage capacity.
5) Integration of geological, sedimentological, and geo-
morphological studies with VES and ERGI imaging
of the subsurface is an inexpensive, environmentally
friendly and fast methodology for reliable recon-
struction, characterization and mapping of fluvial
aquifers/aquitard/aquiclude complexes. The method
provides improvements upon the traditional point
data approach to aquifer reconstruction and charac-
terization, and may also be applied in geotechnical
mapping and in studies aimed at investigation of the
relation between fluvial evolution and environmental
change. This method of reconstruction can also be
used to condition or verify 3-D geostatistical simu-
lations of subsurface alluvial architecture.
Acknowledgements
Financial support to the research was provided by
PRIN 20052006 funds, CNR-IDPA funds to RB and by
FIRST (University of Milan) funds to MG. We
acknowledge the cooperation during field-work and the
suggestions of Diana Dell'Arciprete (study site 2),
Beniamino Calella, Giuseppe Malgesini, Silvia Otto-
mano, Fabrizio Pavia, Fabrizio Felletti, Alfredo Bini and
Mariangelo Baio, as well as the kindness of the proprietor
of the Ca' de' Geri quarry, who permitted unlimited
access to study site 2. Careful reviews by K.F. Rijsdijk, an
anonymous Referee and the Editors substantially im-
proved the first version of the manuscript.
References
Alfano, L., Mancuso, M., 1996. Sull'applicabilit del metodo dipolo-
dipolo continuo nelle ricerche idriche a media profondit in aree di
pianura. Acque Sotterranee XIII 2 (50), 6171.
Asprion, U., Aigner, T., 1999. Towards realistic aquifer models: a three-
dimensional georadar case-study of Quaternary gravel deltas (Singen
Basin, SW Germany). Sedimentary Geology 129, 281297.
Baines, D., Smith, D.G., Froese, D.G., Bauman, P., Nimeck, G., 2002.
Electrical resistivity ground imaging (ERGI): a new tool for
mapping the lithology and geometry of channel-belts and valley-
fills. Sedimentology 49, 441449.
247
Barker, R.D., 1989. Depth of investigation of collinear symmetrical
four electrode arrays. Geophysics 54, 10311037.
Bersezio, R., Pavia, F., Baio, M., Bini, A., Felletti, F., Rodondi, C.,
2004. Aquifer architecture of the Quaternary alluvial succession of
the southern Lambro basin (Lombardy, Italy). Il Quaternario 17 (2/1),
361378.
Bini, A., 1997a. Problems and methodologies in the study of
Quaternary deposits of the southern side of the Alps. Geological
Institute 2 (2), 1120.
Bini, A., 1997b. Stratigraphy, chronology and palaeogeography of
Quaternary deposits of the area between the Ticino and Olona
Rivers (ItalySwitzerland). Geological Institute 2 (2), 2146.
Bini, A., Strini, A., Violanti, D., Zuccoli, L., 2004. Geologia del
sottosuolo dell'alta pianura a NE di Milano. Il Quaternario 17 (2/1),
343354.
Binley, A., Cassiani, G., Winship, P., 2004. Characterization of
heterogeneity in unsaturated sandstone using borehole logs and
cross-borehole tomography. In: Bridge, J.S., Hyndman, D.W. (Eds.),
Aquifer Characterization. SEPM Spec. Publ., vol. 80, pp. 129138.
Birkelo, B.A., Steeples, D.W., Miller, R.D., Sophocleous, M., 1987.
Seismic reflection study of a shallow aquifer during a pumping
test. Ground Water 25, 703709.
Bradford, J.H., 2002a. Depth characterisation of shallow aquifers with
seismic reflection, part I the failure of NMO velocity analysis
and quantitative error prediction. Geophysics 67, 8997.
Bradford, J.H., 2002b. Depth characterisation of shallow aquifers with
seismic reflection, part II Prestack depth migration and field
examples. Geophysics 67, 98109.
Bradford, J.H., Sawyer, D.S., Zelt, C.A., Oldow, J.S., 1998. Imaging a
shallow aquifer in temperate glacial sediments using seismic reflection
profiling with DMO processing. Geophysics 63, 12481256.
Bridge, J.S., Hyndman, D.S., 2004. Aquifer characterization. In:
Bridge, J.S., Hyndman, D.W. (Eds.), Aquifer Characterization.
SEPM Special Publication, vol. 80, pp. 12.
Bristow, C.S., Jol, H.M. (Eds.), 2003. Ground Penetrating Radar in
Sediments. Geological Society, Spec. Publ., vol. 211. 197 pp.
Bruno, P.P.G., Godio, A., 1997. Environmental risk assessment of a
shallow aquifer in Piana Campana (Italy): a field comparison
between seismic refraction and reflection methods. European
Journal of Environmental and Engineering Geophysics 2, 6176.
Castiglioni, G.B., Pellegrini, G.B., 2001. Note illustrative della Carta
Geomorfologica della Pianura Padana illustrative notes of the
geomorphological map of Po Plain (Italy). Supplementi di
Geografia Fisica e Dinamica Quaternaria IV (207 pp, Torino).
Dahlin, T., 2001. The development of DC resistivity images techniques.
Computer and Geosciences 27, 10191029.
Daily, W., Ramirez, A.L., 2000. Electrical imaging of engineered
hydraulic barriers. Geophysics 65 (1), 8394.
Da Rold, O., 1990. L'apparato glaciale del Lago maggiore, settore
orientale. Ph.D. Thesis, University of Milan, 200 pp.
ENI- Regione Lombardia, 2002. Geologia degli acquiferi padani della
Regione Lombardia. Relazione Tecnica. S.EL.CA, Firenze. 130 pp.
Francese, R., Hajnal, Z., Prugger, A., 2002. High-resolution images of
shallow aquifers. A challenge in near surface seismology.
Geophysics 67, 177187.
Francese, R., Giudici, M., Schmytt, D.R., Zaja, A., 2005. Mapping the
geometry of an aquifer system with a high resolution reflection
seismic profile. Geophysical Prospecting 53, 817828.
Geissler, P.E., 1989. Seismic profiling for groundwater studies in
Victoria, Australia. Geophysics 54, 3137.
Giudici, M., Manera, M., Romano, E., 2003. The use of hydrological
and geoelectrical data to fix the boundary conditions of a ground
248 R. Bersezio et al. / Sedimentary Geology 202 (2007) 230248
water flow model: a case study. Hydrology and Earth System
Sciences 7 (3), 297303.
Gourry, J.C., Vermeersch, F., Garcin, M., Giot, D., 2003. Contribution of
geophysics to the study of the alluvial deposits: a case study in the
Val d'Avaray area of the River Loire, France. Journal of Applied
Geophysics 54, 3549.
Heinz, J., Aigner, T., 2003. Three dimensional GPR analysis of various
Quaternary gravel-bed braided river deposits (southwestern Ger-
many). In: Bristow, C.S., Jol, H.M. (Eds.), Ground Penetrating Radar
in Sediments. Geological Society Spec. Publ., vol. 211, pp. 99110.
Jol, H.M., Smith, D.G., 1991. Ground penetrating radar of northern
lacustrune deltas. Canadian Journal of Earth Sciences 28, 19391947.
Juhlin, C., Palm, H., Mullern, C.F., Wallberg, B., 2002. Imaging of
groundwater resources in glacial deposits using high-resolution
reflection seismic. Journal of Applied Geophysics 51, 107120.
Koefoed, O., 1979. Geosounding Principles. Elsevier, Amsterdam.
Lapenna, V., Rizzo, E., 2003. Tomografie geoelettriche dipolari del
Bacino dell'Alta Val d'Agri. In: Colella, A. (Ed.), Le risorse idriche
sotterranee dell'Alta Val d'Agri. Collana Editoriale di Studi e
Ricerche, vol. 3. Autorit Interregionale di Bacino della Basilicata,
Potenza, pp. 87112.
Liberty, L., 1998. Seismic reflection imaging of a geothermal aquifer
in an urban setting. Geophysics 63, 12851294.
Miall, A.D., 1985. Architectural-element analysis: a new method of
facies analysis applied to fluvial deposits. Earth-Science Reviews
22, 261308.
Muttoni, G., Carcano, C., Garzanti, E., Ghielmi, M., Piccin, A., Pini, R.,
Rogledi, S., Sciunnach, D., 2003. Onset of major Pleistocene
glaciations in the Alps. Geology 31 (11), 989992.
North, C.P., 2005. Subsurface fluvial modelling: fashions, fads, and
classic black numbers. 8th International Conference on Fluvial
Sedimentology Delft, Book of Abstracts. TU Delft, pp. 1821.
Reynolds, J.M., 1996. An Introduction to Applied and Environmental
Geophysics. Wiley, New York. 796 pp.
Schulmeister, M.K., Butler, J.J., Franseen, E.K., Wysocki, D.A.,
Doolittle, J.A., 2004. High-resolution stratigraphic characterization
of unconsolidated deposits using direct-push electrical conductivity
logging: a floodplain-margin example. In: Bridge, J.S., Hyndman,
D.W. (Eds.), Aquifer Characterization. SEPM Spec. Publ., vol. 80,
pp. 6778.
Stephens, M., 1994. Architectural element analysis within the Kayenta
Formation (Lower Jurassic) using ground-probing radar and
sedimentological profiling, southwestern Colorado. Sedimentary
Geology 113, 195209.
Telford, W.M., 1990. Applied Geophysics, second edition. Cambridge
University Press, Cambridge. 770 pp.
Tronicke, J., Holliger, K., Barrash, W., Knoll, M.D., 2004.
Multivariate analysis of cross-hole georadar velocity and attenu-
ation tomograms for aquifer zonation. Water Resources Research
40 (W01519) 14 pp.
Ward, S.H., 1990. Resistivity and induced polarization methods. In:
Ward, S.H. (Ed.), Geotechnical and Environmental Geophysics,
vol. 1, pp. 147190.
Whiteley, R.J., Hunter, J.A., Pullan, S.A., Nutalaya, P., 1998.
Optimum offset seismic reflection mapping of shallow aquifers
near Bangkok, Thailand. Geophysics 63, 13851394.
Zalasiewicz, J.A., Mathers, S.J., Cornwell, J.D., 1985. The
application of ground conductivity measurements to geological
mapping. Quarterly Journal of Engineering Geology 18, 139148
(London).
Zaleha, M.J., Ritter, J.B., Rumschlag, H.J., 2005. Resolving fluvial,
glaciofluvial and glacial deposits using electrical resistivity ground
imaging (ERGI). 8th International Conference on Fluvial Sedi-
mentology Delft, Book of Abstracts, vol. 324. TU Delft.
 Available online at www.sciencedirect.com

Sedimentary Geology 202 (2007) 249 268

www.elsevier.com/locate/sedgeo

The development of fluvial stochastic modelling in the Norwegian

oil industry: A historical review, subsurface implementation
and future directions
Kevin Joseph Keogh a,, Allard Willem Martinius b , Rune Osland a
a
Statoil ASA, Technology and Projects, 4035 Stavanger, Norway
b
Statoil ASA, Research Centre, 7005 Trondheim, Norway

Abstract

Fluvial sandstones are an important reservoir type for the petroleum industry. In the late 1970's and early 1980's, large
hydrocarbon discoveries in the Norwegian North Sea in fluvial strata prompted the need for generating geologically meaningful,
stochastic, object-based models of fluvial deposits. The aim of this focus was to allow the geologist to provide the reservoir
engineers with a more realistic representation of permeability contrasts within channelised, fluvial deposits by being able to use
appropriate measurements from outcrop analogues as direct input data into the modelling software. This initiative resulted in the
development of a suite of geologically driven, stochastic modelling algorithms supported by an extensive fieldwork program aimed
at collecting stratigraphic and quantitative data from ancient outcrop analogues to support enhanced reservoir characterisation and
geological modelling. Today, these reservoirs are still important hydrocarbon producing fields with accurate reservoir description
and 3D modelling capabilities playing a vital role in targeting remaining oil, especially now that many of the fields on the
Norwegian continental shelf are past peak production and are in a decline phase.
As both computing capabilities and quantitative outcrop analogue studies have increased the understanding of, and the ability to
model fluvial reservoirs, so have stochastic modelling techniques continued to provide the most suitable and robust means of
building geologically realistic 3D reservoir models that incorporate increased geological understanding and heterogeneity
complexity. In the recent past, a multitude of data, such as seismic and production data have been used to condition the stochastic
algorithms.
This review paper aims to outline the role of stochastic algorithms in building geologically-realistic, 3D fluvial reservoir models
and highlight the success of these developments with case studies from both producing fields and ancient outcrop analogue studies.
Finally, the paper will allude to possible improvements in stochastic fluvial modelling and future directions in the modelling of
fluvial petroleum reservoirs. These include the use of physical or process-based models, high-resolution near wellbore models, and
multi-point statistics, that allow for more realistic representations of heterogeneities of fluvial deposits at a variety of scales and by
a variety of methods.
2007 Elsevier B.V. All rights reserved.

Keywords: Fluvial depositional systems; Reservoir characterisation; Stochastic modelling; Norwegian petroleum industry

Corresponding author. Tel.: +47 90983823.

E-mail address: keke@statoil.com (K.J. Keogh).

0037-0738/$ - see front matter 2007 Elsevier B.V. All rights reserved.
doi:10.1016/j.sedgeo.2007.05.009
250 K.J. Keogh et al. / Sedimentary Geology 202 (2007) 249268
1. Introduction
Reservoirs of fluvial sandstones account for more than
20% of the worlds remaining reserves of hydrocarbons,
and are indeed an important reservoir type for the
petroleum industry. In petroleum reservoir description it
has long been recognised that modelling sedimentary
deposits as totally homogeneous bodies, both with regards
to sedimentological and structural heterogeneities, is a
gross simplification of their potential flow behaviour.
Ultimately, reservoir simulation models aim to integrate
data from multiple scales of measurement to capture a
geologically realistic range and spatial variability in
petrophysical properties, so as to be able to realistically
simulate flow through them. The models should also be
able to predict future oil and gas production of alternative
recovery scenarios and to optimize field development.
Simple trend mapping and interpolation techniques that
were standard in the late 1970's and early 1980's failed to
give reliable predictions for creating geologically realistic
models of fluvial deposits. A wealth of sedimentary
studies on outcrops and modern systems had already
identified that fluvial deposits are heterogeneous on a
variety of scales, from the microscopic to the megascopic
scale (Fig. 1). The understanding of the occurrence and
variability of these various scales of heterogeneities was
deemed important in order to better characterise these
reservoir types and produce hydrocarbons from them.
Predicting and representing heterogeneities caused by
faulting within a reservoir model is an equally important
task in fluvial reservoir characterisation. A wealth of
research and publications exists on this topic (Yielding
et al., 1997; Knai and Knipe, 1998; Manzocchi et al.,
1999; Lecour et al., 2001; Thore et al., 2002; Sperrevik
et al., 2002; Bailey et al., 2002; Holden et al., 2003;
Sverdrup et al., 2003; Ottesen et al., 2005; Hoffman et al.,
2006) and the ability to more accurately model seismic to
sub-seismic fault populations, predict fault seal capacities
and model transmissibilities across faults has also
enhanced hydrocarbon recovery in fluvial reservoirs as
much as increased depositional understanding and
stochastic facies modelling. The subject of fault model-
ling is however beyond the scope of this paper and only
the sedimentological and stratigraphical aspects of
reservoir modelling will be addressed herein.
The discovery of potentially large hydrocarbon
accumulations within fluvial deposits in the Norwegian
North Sea in the late 1970's to early 1980's led to the
realisation that to produce these deposits economically
and effectively, better prediction of their connectivity
and flow behaviour needed development of specialist
modelling algorithms. The heterogeneities identified
from sedimentological studies also needed to be captured
in a geological modelling environment to quantitatively
represent their occurrence and variability and to aid in
transferring geologically realistic models into the
reservoir simulation domain. With the current initiatives
of late-life IOR (Increased Oil Recovery) projects on the
Norwegian continental shelf, particularly on fields
within the Tampen Spur area (Statfjord, Snorre, Gullfaks
Fields and their satellites) and Haltenbanken area there is
currently a strong focus on enhanced reservoir char-
acterisation and modelling of the fluvial reservoirs of the
Upper Triassic Lunde, Lower Jurassic Statfjord (Tam-
pen) and Lower Jurassic re (Haltenbanken) Formations
in the Norwegian continental shelf. The aim of this paper
is to highlight how fluvial stochastic modelling began
in the Norwegian oil industry and how it has been
successfully applied to producing fields, particularly in
the Norwegian sector of the North Sea. Some remarks on
what could be done in the future to better our under-
standing and modelling of fluvial petroleum systems are
also presented.
2. What is stochastic reservoir modelling and why
use it?
This paper is not meant as a review of stochastic
modelling techniques and their uses in the oil industry, as
much has been previously published on specific algo-
rithms and their application (Haldorsen et al., 1988;
Haldorsen and Damsleth, 1990; Srivastava, 1994; North,
1996; Damsleth and Omre, 1997; Dubrule, 1998),
however, it is useful to review certain elements of
stochastic modelling that are relevant to the theme of this
paper.
Reservoirs have potentially measurable, determinis-
tic properties at all scales that arise from the interaction
of many, complex processes and are therefore intrinsi-
cally deterministic in their nature. However, the
information we gather of a reservoir is incomplete and
of differing length scales (seismic scale vs well log scale
vs core plug scale; Fig. 2). To be able to best predict oil-
in-place and recovery, a method must therefore be used
whereby the gaps in data can be filled and the natural
variability of the system at a certain volume scale can be
captured. Previously, interpolation, based on the idea
that there was little or no variability between data points,
was used as the method for computing maps honouring
the data for the parameter of interest. However, the
smoothness of these interpolations did not realistically
describe the spatial variability of the parameters
modelled in fluvial reservoirs. Lake and Carroll (1986)
defined reservoir characterisation as a process for
 K.J. Keogh et al. / Sedimentary Geology 202 (2007) 249268 251

Fig. 1. Schematic view of the different scale and type of heterogeneities that can affect fluid flow behaviour in fluvial depositional systems. Diagram
modified from Weber (1986), Haldorsen (1986) and Dreyer et al. (1990a).

quantitatively assigning reservoir properties, recognis-
ing geological information and uncertainties in spatial
variability. They stated that geostatistics and stochastic
modelling provide the methods and tools to integrate the
available information from different sources and scales
in a unified and consistent framework and to generate
multiple realisations that can account for the uncertainty
and spatial variability of the key reservoir parameters.
The practice of stochastic modelling is a relatively
recently established discipline within geosciences with
252 K.J. Keogh et al. / Sedimentary Geology 202 (2007) 249268
Fig. 2. Chart showing the various subsurface data types that sample the reservoir volume plotted against the corresponding vertical resolution of the
data type. Both outcrop analogue studies and physical-based modelling studies can help bridge the gaps between the different subsurface data types.
some of the earlier work dating back to only the early/
mid-1980s. It is a common misconception that stochas-
tic modelling involves a purely random process of
distributing properties away from the data points. The
origin of the word stochastic from the Greek adjective
stokastikos means skilful at aiming or guessing and
Haldorsen and Damsleth (1990) in their review of
stochastic modelling refer to the stochastic component
of reservoir description as formalised guessing. The
aim of stochastic modelling (either by discrete,
continuous or hybrid methods) is to generate synthetic
but statistically realistic realisations from limited data
for the reservoir under study (Haldorsen and Damsleth,
1990) by honouring the data available at known points/
locations in the reservoir and then simulating data
between by conforming to a set of statistical rules and
user-defined conditions. Those conditions can include
the geometry and distribution of facies objects, the
distribution, orientation and lengths of faults, the
correlation lengths of continuous parameters and the
covariance of dependent parameters. Conditional sto-
chastic simulation aims to achieve more realistic
variability by systematically adding noise to the
modelling parameters according to the conditions
defined. The noise is zero at the data points and the
resulting simulated parameter gives a more realistic
outcome because the observations are honoured and the
statistical properties of the parameter are better mim-
icked than the interpolated parameter (Haldorsen and
Damsleth, 1990; their Fig. 15). Initially, in stochastic
reservoir modelling, these a priori conditions were not
necessarily geologically related, but computational
advancements and increased geological understanding
during the development of these algorithms led grad-
ually to increasing geological control on the input to the
models. Object-based techniques, also referred to as
Boolean or Marked point process models, (Matheron
et al., 1987; Dubrule, 1989; Haldorsen and Damsleth,
1990; MacDonald and Halland, 1993; Hatly, 1994;
Srivastava, 1994; Deutsch and Wang, 1996; Deutsch
and Tran, 2002) became the preferred stochastic method
of representing channelised fluvial reservoirs. The
ability to model channel belts and individual channels
as discrete objects within which appropriate petrophy-
sical distributions could be modelled, often using pixel-
based techniques, was a major step forward (Journel
et al., 1998). This advance led to a more accurate
representation of the subsurface fluvial reservoir, giving
realistic input for dynamic reservoir flow simulation
studies.
It has been shown through various publications
(Henriques et al., 1990; Tyler et al., 1994) and internal
company research that in subsurface reservoir modelling,
stochastic techniques are well suited to producing
multiple, quantified, statistically regular approximations
of the geoscientists' formalised and educated impression,
conditioned to the field data available and other soft
knowledge. A study by Clemetsen et al. (1989)
highlighted this finding by comparing various methods
using oilfield data. During the appraisal stages of the
Snorre field, Norwegian North Sea, the process-driven
model approach of Bridge and Leeder (1979) was
evaluated along with a stochastic modelling approach.
The process-driven model of Bridge and Leeder (1979)
was found to be unsuitable as the knowledge of the
sedimentary and tectonic processes responsible for
sediment deposition needed to define the model were,
and often still are in mature fields, poorly understood. The
published model of Johnson and Krol (1984) was also
tested (Henriques et al., 1990). The input data describing
 K.J. Keogh et al. / Sedimentary Geology 202 (2007) 249268
fluvial channel sandbody stacking and connectivity
within the reservoir were geologically interpreted 2D
log correlation panels. The large well spacing in the
Snorre Field, Norwegian North Sea where this method
was tested gave large uncertainties in the correlation of log
data, a problem not encountered by Johnson and Krol
(1984) in the Brent Field, UK North Sea where there was
much closer spacing between wells at that time and thus
less correlation uncertainty. Given also that different
geologists would create a different log correlations with
the same data, this was thought to be a cumbersome
method to use for reservoir simulation as it was found to
be too subjective for representing the data and also
difficult to update. In this and most other cases a stochastic
model has shown to produce the most robust and
predictive model results.
More advanced stochastic modelling algorithms
allow surfaces of sequence stratigraphic significance
to be used as zone boundaries in the model building.
This allows more appropriate use of input geometrical/
dimensional data relevant to the sequence stratigraphic
interval in question and various studies have high-
lighted the advantage of using a sequence strati-
graphic zonation principle within stochastic modelling
(MacDonald et al., 1992; Ritchie et al., 1998; Ains-
worth et al., 1999; Howell and Flint, 2002; Keogh,
2002). Despite the relative immaturity of sequence
stratigraphic principles within fluvial environments,
the flexibility of stochastic modelling techniques
allows for uncertainty arising from alternative inter-
pretations or zonations to be tested within the model-
ling process.
3. The development of fluvial stochastic modelling in
the Norwegian oil industry
In the Norwegian North Sea, the discovery of the
giant Snorre Field, then operated by Saga Petroleum,
and the decision to develop the field meant that a
computer program also had to be developed to model
and simulate fluvial sandbody architecture. In the mid
1980's an initiative in the Norwegian oil industry was
undertaken to create stochastic modelling algorithms to
better model these deposits and also to collect geological
data from well-exposed outcrop analogues world-wide
to give geological input to the modelling algorithms.
3.1. SISABOSA
The first generation of three-dimensional fluvial
modelling applications named SISABOSA (SImulation
of SAnd BOdies for SAga) was developed primarily for
253
use in field appraisal and development stages. The
development of these stochastic algorithms was a
breakthrough in reservoir characterisation for the Norwe-
gian oil industry as the geology of fluvial deposits could
be more realistically defined. The model assumes that the
reservoir matrix (non-channel deposits) is non-permeable
and is penetrated by permeable bodies (channel deposits)
as shown in Fig. 3a. The channel deposits are modelled as
a rectangular parellepiped shape and the dimensions of
these shapes are user-defined and constant. The channel
deposits are located randomly within the permeable
matrix whilst the orientation of the channel bodies is
specified by the user with a degree of freedom around a
mean value. The channel volume fraction of the fluvial
system is also user-defined and is used by the model as a
stop criterion for the model run. Augedal et al. (1986)
document in detail the model algorithms, input specifi-
cation and corresponding output.
The resulting reservoir geometry output from this
first generation 3D fluvial modelling algorithm held
many assumptions concerning the nature of fluvial
deposits and as such was an over-simplification of the
geology. However, Stanley et al. (1989) published
results comparing the flow prediction from a fluvial
reservoir defined using the SISABOSA model com-
pared to the flow prediction from using an infinite,
homogeneous reservoir definition. These predictions
were plotted against the flow results obtained from a
well flow test and show that the definition of the
reservoir using the SISABOSA model gave much better
predictions compared to the well test observations that
could be achieved using a homogeneous reservoir
definition (Fig. 4). These results mirror the heteroge-
neous nature of fluvial deposits that have been
documented in the various sedimentological studies
and emphasise the importance of representing these
heterogeneities if realistic predictions of the subsurface
reservoir are to be obtained.
3.2. FLUREMO
The development of FLUREMO (FLUvial REser-
voir MOdelling) was to supersede SISABOSA with the
ability to be able to give more geologically realistic
input to the algorithms. The improvements in FLUR-
EMO allowed for the ability to define avulsion and
clustering of channels and channel belts (Fig. 3b;
Clemetsen et al., 1989). This geological principle was
mimicking the process-based rules employed in the
model of Bridge and Leeder (1979) via a mathematical
function and was a great improvement from the random
placement of channels as employed in the SISABOSA
254 K.J. Keogh et al. / Sedimentary Geology 202 (2007) 249268

Fig. 3. 3D conceptual diagrams representing the output from the development of the three main fluvial stochastic algorithms used in the Norwegian
oil industry. (a) SISABOSA algorithm producing rectangular parallelepiped shaped channels with uniform thickness. (b) FLOREMO algorithm
producing clustering or avulsing, sinuous channels that can show a range of thicknesses. (c) SISA algorithm producing multiple facies, with variable
channel dimensions along individual channels, channel belt interactions, multi-well correlations and non-homogeneous permeability value
assignment to individual facies.

model. The channel bodies of the SISABOSA model
were essentially rectangular shaped, inferring that the
bodies had no deviation from their axial orientation.
This again was a geological simplification and an
improvement in the FLUREMO model was the ability to
define a sinuosity to the channels and channel belts,
again giving more realistic geometrical form to the
channel and channel belts (Fig. 3b). The third
development in the FLUREMO model was the ability
to vary the width and height of both individual channels
 K.J. Keogh et al. / Sedimentary Geology 202 (2007) 249268 255

Fig. 4. Comparison of reservoir flow performance of two reservoir models together with the field well test observation. The reservoir model defined as
an infinite homogeneous reservoir over-estimates the reservoir flow performance significantly, whereas the reservoir model built using the
SISABOSA fluvial stochastic modelling algorithms reproduced the test observation reasonably well. Diagram modified from Stanley et al. (1989).

and channel belts. It was already known from studies of
modern and ancient fluvial systems that individual
channels and channel belts did not have exactly the
same height and width, but rather always varied from
one channel to the next. The constant width and
thickness imposed in the SISABOSA model was again
a geological simplification of a known component of the
nature of fluvial systems, and the inclusion of this
improvement in the FLUREMO model (Fig. 3b) was
again an attempt to create models guided by actual
geological principles and observations. Clemetsen et al.
(1989) details the model and its improvements over
SISABOSA.
3.3. SISA
With some years' production experience from fields
on the Norwegian shelf and more field data available
from production wells, the complexity of heterogene-
ities in these fluvial reservoirs that was affecting the
production behaviour became more apparent and
quantifiable. The geological simplifications of the
SISABOSA and FLUREMO modelling algorithms
became insufficient to honour the geological complexity
now seen in the field data and that was already docu-
mented in the numerous published studies of ancient and
modern systems (Fig. 1). Egeland et al. (1993) and
Georgsen et al. (1994) published a revised fluvial
stochastic modelling algorithm (SISA SImulation of
Sedimentary Architectures) and its application to sub-
surface data, that was better able to handle both the
increased amount of subsurface data and the implemen-
tation of the increased geological understanding of the
reservoir and its flow behaviour complexities gained
during the production phase of field life.
The revised algorithms implemented in SISA could
now handle the explicit modelling of multiple facies in
the form of background, channel sand, sheet-splay sands
and internal channel barriers (Fig. 3c). Production data
from wells also revealed a new source of data to the
geologist whereby communication from injector to
producer well pairs gave geological information regard-
ing the connectivity of the fluvial system. The use of this
information was implemented in the SISA model by
allowing the geologist to be able to correlate the same
facies between multiple wells. An updated conditioning
algorithm for the treatment of deviated wells was
published by Skorstad et al. (1999) in response to the
challenge of conditioning object-based stochastic fluvial
models to geological information from these well types.
A feature of the previous stochastic fluvial modelling
algorithms was the geologically poor definition of the
dimensions of the channels and channel belts. In SISA,
an improvement was made to better define geologically
256 K.J. Keogh et al. / Sedimentary Geology 202 (2007) 249268
realistic channel dimensions by allowing them to vary
along the length of a single channel or channel belt
(Fig. 3c).
Data collection from ancient fluvial analogues had
recognised that the stratigraphic stacking of fluvial
channels and channel belts was often of an ordered
nature and in some studies this ordering is considered a
consequence of base level fluctuations and therefore
systems tract-dependent (Legarreta et al., 1993; Shanley
and McCabe, 1993; Olsen, 1995; Curie, 1997; Martin-
sen et al., 1999). This phenomenon had also been
recognised in modern fluvial systems, had been
quantified as a process and was an integral part of the
process-based fluvial modelling algorithm of Bridge and
Leeder (1979). The specification of channel and channel
belt interactions into the stochastic modelling algo-
rithms (Egeland et al., 1993; Georgsen et al., 1994) was
again an attempt to create geologically realistic fluvial
channel models, and also an attempt to represent known
fluvial processes in a non-process-based mathematical
formula.
Another major improvement of the SISA algorithms
from the previous models was the ability to assign
permeability values on a per-facies basis and also to
specify a range around the mean permeability of each
facies. This allows for the modelling of non-homoge-
neous facies, giving a more realistic outcome from the
modelling that matches more the natural variation seen
in core plug measurements from the same facies type.
These improvements represented the efforts of many
geologists within the Norwegian oil community in
addressing the challenge of producing geologically
realistic mathematical models from stochastic modelling
algorithms.
3.4. Moheres
Moheres (Modelling of Heterogeneous Reservoirs)
contained two modules that were developed within the
Norwegian oil community, specifically by the Norwe-
gian Computing Centre, ODIN and Geomatic in the
early 1990's for the modelling of fluvial depositional
systems (Bratvold et al., 1994). These modules,
Fettuccini and Sevfac, were a two-step modelling
process. In Fettuccini (as an analogy to pasta shapes),
the channel belts and channels were modelled. Sevfac
(Several Facies) was used to model within-channel
heterogeneities in the form of lithofacies variations. The
modelled lithofacies distribution represented the sys-
tematic change in lithofacies within a channel as the
channel and channel belt developed through time with
changing sediment volume and type input and water
velocity changes, as observed from ancient analogues,
modern systems and subsurface core data. The devel-
opment of these algorithms was again in response to a
drive by geologists in the Norwegian oil community for
geologically guided stochastic models.
3.5. Statoil and BP Alliance
For a four year period during the mid-1990s, a group
of researchers from Statoil and BP oil companies
worked together on a co-operative research program
aimed at identifying which heterogeneity types, both
sedimentological and structural, had the most effect on
static and dynamic fluid flow in shallow marine and
fluvial reservoirs. Whilst the Statoil and BP Alliance
(19921996) did not have the explicit goal of improving
the geological modelling algorithms of fluvial systems,
two important modelling developments were made
during this project. Jones et al. (1995) detail the
geological findings of this project, but in addition to
these, it was through this project that the workers
recognised the need to represent both small scale
heterogeneities and sub-seismic resolution faults in
reservoir models (Fig. 6) to be able to better predict flow
behaviour of fluvial depositional systems. These find-
ings ultimately led to the development of the now
commercialised SBED (Sedimentary (BEDding) soft-
ware (Wen et al., 1998; Elfenbein et al., 2005; Ringrose
et al., 2005) and HAVANA (sub-seismic and stochastic
fault modelling) software (Holden et al., 2003; Ottesen
et al., 2005).
3.6. SAFARI Project
Within the Norwegian oil community, the SAFARI
(Sedimentary Architecture of Field Analogues for Re-
servoir Information) project, a co-operation between
three Norwegian oil companies (Norsk Hydro, Saga and
Statoil) and the Norwegian Petroleum Directorate
(NPD), was an innovative fieldwork programme. The
aim of the project was to collect field analogue data from
a host of localities world-wide considered to be ancient
analogues to the reservoirs on the Norwegian continen-
tal shelf. Specifically, the project would deliver a
database of stratigraphic criteria and quantitative mea-
surements (Fig. 5) for the purpose of providing input to
geologically-guided stochastic models. Within fluvial
depositional systems, a total of seven different field
analogues were studied and data collected for quan-
titatively characterising the depositional geometry and
internal heterogeneities of fluvial deposits analogous to
Lunde and Statfjord Formations in the North Sea.
 K.J. Keogh et al. / Sedimentary Geology 202 (2007) 249268 257

Fig. 5. Examples of the types of quantitative data collected during the SAFARI field analogue project to give dimensional input to reservoir models on
architectural elements constrained to stratigraphic intervals with the aim of creating geologically-guided stochastic models. (a) 2.5D architectural
fence diagram describing the stratigraphic evolution of individual channel and channel belts; (b) Histogram plot of thickness distribution of channel
belt bodies; (c) Width distributions of architectural elements; (d) Thickness vs width plot for channel belt and channel sandbodies. Diagram modified
from Dreyer et al. (1993).

Dreyer et al. (1993) published a case study from the
Escanilla Formation of the Ainsa Basin in Spain as an
example of how quantitative data from the SAFARI
dataset could be implemented in stochastic reservoir
modelling studies. These quantitative data (Fig. 5bd)
are constrained from analysis in the field of sedimentary
bodies and from 3D fence diagrams constructed from
2D cross sections (Fig. 5a). Such quantified data is key
input to object-based stochastic modelling but extraction
of such data from 1D core from the subsurface is often
limited. It is here that such outcrop measurements can
help bridge the gap between scales of data resolution in
the subsurface (Fig. 2) to give geologically constrained
input to the stochastic techniques.
A 3D facies model was built of the studied outcrops of
the Escanilla Formation (Nss et al., 1997) using the
SAFARI dataset and the channel connectivity patterns
subsequently analysed using streamline simulation meth-
ods (Blunt et al., 1996; King and Datta-Gupta, 1998;
Sharif et al., 2000). The study by Nss et al. (1997) used a
stochastic modelling tool conditioned to outcrop data
(digitised outcrop logs, geometrical parameters of channel
cross-section and planform parameters) to produce 3D,
geologically realistic representations of the fluvial
channels identified during the SAFARI fieldwork pro-
gram. The stochastic channels were subsequently com-
pared to the actual outcrop interpretations as a reality
check and the models were found to be consistent with
258 K.J. Keogh et al. / Sedimentary Geology 202 (2007) 249268
field correlations. Nss et al. (1997) claim that the
generation of such channel geometries observed in the
Escanilla Fm using conventional mapping techniques
would have required the use of at least several hundred
control points. By utilising a stochastic modelling tech-
nique, Nss et al. (1997) produced geologically realistic
channel geometries with a much more limited condition-
ing dataset. The resulting 3D geological model could then
be quantitatively investigated to better understand the
dynamic impact of channel geometry and connectivity in
fluvial systems.
It is the nature and quality of the outcrop exposures
that is key to being able to properly reconstruct channel
geometries in three-dimensions with high confidence.
The importance of understanding the three-dimensional
nature of low (b 35%) net-to-gross fluvial depositional
systems with regards to reservoir connectivity potential
is highlighted in Fig. 6. By analysing a simple 2D
section of the outcrop (Fig. 6a), it is apparent that the
lower interval consists of isolated, discontinuous
channels, whereas the uppermost interval consists of a
broad, amalgamated channel belt. Four pseudo wells
were placed in the corners of the 3D model, two
assigned as producer wells and two as injector wells. A
simple streamline simulation was then performed with
the aim of testing how connected the outcrop is in three-
dimensions. Where channel facies are encountered in
the wells, these intervals are set as perforated intervals.
Despite its apparent non-connectedness, flow from the
injector well was able to reach the production well in the
lower interval of the model (Fig. 6b). In 3D, these
seemingly isolated channels (Fig. 6a) are locally
connected, due to their sinuous nature, thus allowing
for a tortuous flow path to be found between the injector
and producer pair.
An important consideration, despite static connected-
ness in this lower interval, is the dynamic aspect of the
connectivity. Injector- and producer-well flow rates must
be maintained at certain levels to meet production
demands and to uphold pressure stability in a subsurface
reservoir. A very tortuous flow path between two wells
(Fig. 6b) may result in poor well economics with regard to
daily hydrocarbon production rate. Infill drilling can
increase dynamic reservoir connectivity and the cost of
drilling more wells can be offset by increasing daily
production and hence increased supply to the market. The
ability to model these outcrop analogues in a geologically
realistic way, such that drilling scenarios can be
quantitatively tested, allows better evaluation of reservoir
engineering solutions and hence better predictions of the
economical viability of such solutions for subsurface
fields. In the uppermost interval, the streamline pattern
produced shows that the broad, amalgamated channel belt
allows for good dynamic communication between the
injector and producer pair (Fig. 6c).
3.7. Developments in the last decade
It is noticeable from the literature and from sustained
use of existing software tools that from the mid-1990s to
the present day there have been no major advances or
developments in fluvial stochastic modelling algorithms.
The question remains then: what has happened in that
time? During this period, there was a conscious shift of
focus in the oil industry. The mid-late 1990's saw a
paradigm shift from algorithm development to integrated
workflows and ease of use in an attempt to make
available the modelling algorithms developed from the
academic and company research arena to the desktop of
the asset teams. To complement this, the 3D reservoir
modelling tools became a platform for visualisation of
data and the modelling algorithms needed to take into
account the possibilities for conditioning on different
data types, often of differing quality and resolution.
Dubrule and Damsleth (2001), in their review of the
achievements and challenges in petroleum geostatistics,
regarded the future of geostatistics as not of developing
more algorithms but in integrating these tools and
training the various reservoir subsurface personnel to use
these tools. Their conclusions highlight that the change
in focus in the Norwegian oil industry from algorithm
development still has some way to go before petroleum
geostatistics and stochastic modelling techniques is a
toolbox that is integrated and utilised by all geoscientists,
not just the specialists.
The advancement of seismic processing and inter-
pretation techniques has allowed seismic data to be used
directly in conditioning fluvial reservoir models (Skare
et al., 1996; Hauge et al., 1998; Svanes et al., 2004; their
Fig. 9). The study by Svanes et al. (2004) also highlights
the flexibility of stochastic fluvial modelling algorithms
and how successful they were in being able to assess the
feasibility of proposed well placements in the Sincor
Field, Venezuela. Svanes et al. (2004) were able to use
seismic data in three alternative ways for conditioning
the stochastic model, each alternative depending on
seismic data quality and stochastic model purpose.
Firstly, 2D trend maps derived from seismic attribute
maps were used to condition trends in the stratigraphic
units represented in the stochastic model. Secondly, the
seismic data in certain areas and stratigraphic intervals
were of adequate quality to allow interpretion of facies
that would be modelled as stochastic objects. From the
seismic interpretation, it was possible to determine
 K.J. Keogh et al. / Sedimentary Geology 202 (2007) 249268 259

Fig. 6. Streamline simulation of the 3D deterministic model from Escanilla Formation, Spain, built using the SAFARI dataset. Pseudo wells have been
created to act as injector and producer pairs for simulation purposes. Dynamic simulation of such outcrop models allows testing of production
scenarios to find optimum well numbers and placements for economic field development purposes of analogous subsurface fields. (a) Cross-section
showing the fluvial architecture in 2D. The lower interval appears to contain poorly connected channels, whereas the uppermost interval shows a
broad, amalgamated, well connected channel belt. Net-to-gross is 29% in this section. (b) Streamline simulation of the lower interval showing that
despite the apparent non-connectedness of the channels in the 2d section, their 3D geometry connects them up allowing a tortuous path to be found
between the injector and producer well. (c) Streamline simulation of the uppermost interval showing the simple flow path from injector to producer
well due to the well connected nature of the channel belt.
260 K.J. Keogh et al. / Sedimentary Geology 202 (2007) 249268
ranges in geometries, dimensions and paleoflow direc-
tions of these objects and thus tune the input parameters
to the stochastic model from these data. Thirdly, Svanes
et al. (2004) used the 3D seismic (inverted impedance)
cube directly in conditioning the stochastic object-based
algorithms for controlling the spatial distribution of the
various modelled facies. The resultant 3D stochastic
models represented the subsurface geology such that
evaluation of planned horizontal wells and in place and
recoverable volumes could be performed directly in the
geological model.
Additionally, on fields with many years production,
geological information from production data has to now
be taken into account as a conditioning parameter and
some interesting methods and results from the integration
of production (dynamic) data into reservoir models (Yoon
et al., 2001; Kashib and Srinivasan, 2006) have recently
been published. However, these areas still require
research and development before they become more
routine within stochastic reservoir modelling studies.
Despite the ability of stochastic algorithms to
integrate multiple data types and to generate multiple
realisations capturing uncertainty and spatial variability
of reservoir parameters, often the uncertainty of a
reservoir parameter is not necessarily connected to the
variation around the mean and standard deviation of the
same distribution but rather the mean value of the
distribution itself is uncertain. Uncertainty studies were
thus undertaken to establish the significance in the
uncertainty of both outcrop (Martinius and Nss, 2005)
and reservoir parameters (Tyler et al., 1993; Sandsdalen
et al., 1996; Sharif et al., 2000), and a key focus here
was how to best integrate these workflows and findings
with the stochastic reservoir model. This is a challenge
that today has more focus, and reservoir modelling
software vendors are investing in integrating uncertainty
modules within the stochastic modelling environment.
A recent geological uncertainty study (Osland et al.,
2006), based on the Snorre Field in the Norwegian North
Sea was undertaken with the aim of evaluating the effect
of using different stochastic modelling algorithms for
representing the high, commonly up to 80%, net sand-to-
gross fluvial system (Mjs et al., 2005). A series of
object-based, stochastic facies modelling scenarios were
run and the differences between resulting models
quantitatively evaluated through use of a reservoir
simulator. The two object-based algorithms used both
utilise a marked point process model, where one is
adapted specifically for representing fluvial-style hetero-
geneities (Holden et al., 1998) and the other, a general
marked point process, for producing objects that
describe heterogeneities found in a range of depositional
environments (Lia et al., 1996). The study was
performed on a sector area of the full field and the
results provided input to determine and optimize the
modelling strategy employed for the full field modelling.
The sector model area contains nineteen wells in total,
and while all well data were used to generate input
statistics to the stochastic models together with geolog-
ical knowledge gained from an ancient outcrop ana-
logue, only four wells were conditioned in the model
runs to allow the input distribution around the condi-
tioning parameters to vary according to the stochastic
model used and to speed up the modelling process. The
various facies modelling scenarios were to test the effect
of using different modelling approaches (i.e. using a
fluvial specific object-based algorithm or not) to geo-
logically represent channel, crevasse and background
(non-reservoir) facies associations and the effect of
uncertainty on geological conditioning parameters to
the stochastic models (i.e. channel direction, channel
volume fraction).
Fig. 7 shows examples of the various stochastic
model scenarios tested. Fig. 7af all use the object-
based modelling algorithm specific for modelling fluvial
environments, such that crevasse splay deposits can be
explicitly attached as objects to the channel objects.
Fig. 7a is treated as the reference case model and
evaluation of other model scenarios refer back to the
base case results. Fig. 7b is a different realisation from
the same base case model setup where the differences in
object spatial distribution are a function of the random
element of the modelling algorithm. Fig. 7c has another
channel object directional trend with respect to the base
case. An alternative algorithm is used to model the
crevasse objects (Fig. 7d). Fig. 7ef vary from the base
case in that the input channel and crevasse volume
fractions are 20% lower and higher than the base case,
respectively. An alternative algorithm is used to model
both crevasse and non-reservoir objects with the channel
volume fraction being represented by the remainder of
the grid that is not populated by these objects (Fig. 7g).
Within the sector area there are currently two producer
wells and one injector well and these actual wells were
used in the dynamic simulation model setup.
The simulation of oil recovery (Fig. 8) is used to
evaluate the effect of the various modelling algorithms
and the effect of changing the channel and crevasse
volume fractions. To be able to compare with a model
based on interpolation of well values, a model based on
kriging of petrophysical properties is included. As can
be observed in Fig. 8, the case based on kriging has
unrealistically high recoveries, compared with the
reference case model 10a + b. This response is similar
 K.J. Keogh et al. / Sedimentary Geology 202 (2007) 249268 261

Fig. 7. Horizontal slices of various stochastic, object-based facies modelling scenarios of the Snorre field. Input statistics are taken from surrounding
wells and supplemented with geological knowledge gained from an ancient outcrop analogue. Reference case: (a) using the Channel module in Irap
RMS (b) another realisation. Alternative scenarios: (c) with alternative channel orientation; (d) crevasses are modelled using a different object-based
stochastic algorithm (general marked point process); (e) as (a) but with a lower channel system volume fraction; (f) as (a) but with a higher channel
system volume fraction; (g) both crevasse and floodplain facies associations are modelled using a general marked point process with the channel
facies association defined as the background facies.

to that shown in Fig. 4. The effect of the random seed is
illustrated by comparing the two realisations of the
reference case model (Fig. 7; Fig 8 curves a and b). The
difference between them is small, which should be
expected as the model objects are small compared to the
volume of the model. The evaluation of the alternative
modelling algorithms suggests that the methods chosen
for representing crevasses is of little importance.
By modelling both crevasses and non-reservoir facies
as objects (Fig. 7g), recovery is significantly reduced in
comparison to the base case (Fig. 8 curves a and g).
These results are in agreement to the findings of
MacDonald and Halland (1993) as they suggest that
explicit shale modelling is a relevant procedure for
reservoirs with less than 30% non-reservoir. Evaluation
of the simulation results addressing the effect of channel
262 K.J. Keogh et al. / Sedimentary Geology 202 (2007) 249268

Fig. 8. Graph comparing recovery factor as a function of pore volume injected to quantitatively evaluate the different modelling scenarios shown in
Fig. 7ag. The results showing the largest deviation from the reference case are the scenario with reduced channel volume fraction (Fig. 7e), the
scenario using an alternative modelling approach (Fig. 7g) and the kriged model. These results suggest that even in high net-to-gross scenarios it is
still important to represent the geology being modelled as connectivity can still be an important issue for flow. Similarly to Fig. 4, the kriged model
over-estimates recovery from the reservoir using this modelling technique.

and crevasse volume fraction shows that increasing
volume fraction (Fig. 7f) gives a better recovery.
However, decreasing volume fraction (Fig. 7e) gives a
much greater downside in recovery (Fig. 8 curves e and
f). This highlights that connectivity can still be an issue
even in high net sand-to-gross reservoirs. By employing
stochastic modelling algorithms, the sensitiveness of
recovery factor to changes in various geological input
parameters have been evaluated and recommendations
for the full field stochastic modelling task have been
given based on results from this study.
The advent of marine sequence stratigraphy as a tool
for assisting the reservoir geologist in zoning the
reservoir was being gradually extended to the fluvial
depositional domain in the early to mid-1990s (Shanley
and McCabe, 1993; Wright and Marriott, 1993) and the
Norwegian oil industry also began to focus on increasing
their understanding of the development of fluvial
systems within a sequence stratigraphic framework
(Olsen, 1995; Ryseth et al., 1998; Martinsen et al.,
1999). The aim was that geometrical parameters and
spatial relationships between channels and channel belts
could be assigned to stochastic modelling of fluvial
objects that were appropriate to the systems tract the
fluvial interval was interpreted to reside in (Dreyer,
1990; Dreyer et al., 1990b; Reynolds, 1999; Dalrymple,
2001). This remains a challenge today in both academia
and the oil industry as the acceptance of sequence
stratigraphic principles in fluvial systems and the role of
base level and eustatic controls versus local, autocyclic
controls in fluvial depositional development are still
under strong debate.
4. Future directions of fluvial reservoir modelling
Object-based stochastic algorithms for modelling
fluvial systems will not be replaced in the foreseeable
future. This means that the general philosophy of fluvial
reservoir modelling needs to be changed. The way
channels and their associated facies are modelled needs
to be revisited to improve the algorithms so as to model,
for example, more complex channel patterns, upstream
downstream changes in channel style and channel
branching, and more geologically realistic relationships
to the associated marginal reservoir facies in the
overbank/floodplain areas.
The philosophy of Pore-to-Field modelling is
becoming a research area of great interest in the
Norwegian petroleum community (ren et al., 1998)
including initial attempts within fluvial depositional
systems (Theting et al., 2005). The philosophy behind
pore-to-field modelling is that the various scales of
heterogeneities within the reservoir are represented in
scale-specific models (Fig. 9) built for the Representa-
tive Element Volume (REV; Fig. 9a; Nordahl et al.,
2005), where both subsurface and outcrop analogue data
are used to condition these models (Fig. 9b). At all
scales, the models can be flow simulated in order to
produce type curves (i.e. relative permeability) for the
heterogeneity scale modelled. For modelling the next
 K.J. Keogh et al. / Sedimentary Geology 202 (2007) 249268 263

Fig. 9. Idealised workflow illustrating the philosophy of the pore-to-field scale modelling concept. Recent developments in modelling algorithms
and software capabilities are allowing the small scale heterogeneities of fluvial depositional systems to be geologically modelled at a representative
scale (a) in 3D and flow simulated to analyse their effects on flow behaviour in these reservoir types. These results (c) can then be used to geologically
define the input to the next scale of modelling such that the various scales of heterogeneity in fluvial systems are better preserved during upscaling and
are not misrepresented as was traditionally the case when modelling directly at the geological model scale. The use of ancient field analogues will
continue to be a key arena for increasing our understanding of fluvial depositional systems and for collecting multi-scale data (b) key to input for
stochastic, or otherwise, reservoir models. Ancient and modern analogues can also help in bridging the gap between scale and resolution of data
collected from the subsurface. The outcrop pictures are from the Lourinha Formation, Portugal which is currently being studied as an outcrop
analogue to the Statfjord Formation on the Norwegian continental shelf.

scale up of heterogeneity it is not the model itself but the
model properties (type curves) that are used to bridge
the missing scale gap (Fig. 9c). By employing such an
upscaling philosophy, this results in a seamless
integration of heterogeneities that honour all scale-
related data present in fluvial rocks that are of relevance
for dynamic behaviour. SBED, in particular, is fast
developing into a toolkit for bridging the missing scale
gap and for the development of modelling algorithms to
capture lithofacies to facies association scale hetero-
geneities of fluvial systems (Fig. 9c). This can only help
to achieve better modelling and understanding of the
effect of small scale heterogeneities on fluid flow
behaviour. The integration of these modelling toolkits
into the conventional geological- and reservoir-scale
modelling workflow is also a key issue in successful
implementation of such research-developed toolkits into
subsurface reservoir production operations.
Karssenberg et al. (2001) showed it was possible to
create process-driven models of fluvial channel belts
conditioned to well observation data from a simplified
form of the process model first published by Bridge and
Leeder (1979). Whilst the process-driven model algo-
rithms needed to be simplified in order to account for
well conditioning and run times in particular, the results
were shown to be geologically meaningful in terms of
channel belt evolution and resulting vertical and
planform patterns (Fig. 10). The development of these
264 K.J. Keogh et al. / Sedimentary Geology 202 (2007) 249268
Fig. 10. Process-based modelling of fluvial depositional systems has recently been addressed again by Karssenberg et al. (2001). ac show some of
the input required for the modelling algorithm and (d) is an example of the resulting output. A major advancement in the work of Karssenberg et al.
(2001) was the ability to condition the model to well data.
algorithms needs to be continued to be able to produce
conditioned geological models based on more complex
physical process interactions before it could be
considered an alternative to stochastic modelling
techniques. However this data type could potentially
be considered as another form of geologically-guided
conditioning data for the stochastic algorithms. These
hybrid models could combine the best and essential
parts of process-based techniques and stochastic tech-
niques, training stochastic, object-based models with
process-based model results.
For stochastic reservoir modelling to be fully utilised
further in the value chain, the challenge of history-
matching multiple realisations needs to be addressed.
Treating multiple stochastic realisations in a reservoir
simulator is still in a maturing phase, but with the
expansion of cost-effective high-power computing this
can hopefully be achieved in the foreseeable future. This
is a view also expressed by Dubrule and Damsleth (2001)
where they refer to history matching under uncertainty.
The development of multi-point statistics (MPS) has
shown promising results from published studies,
including channelised fluvial reservoir facies (Caers
et al., 2000), and has the potential for use as an alter-
native method to object-based stochastic techniques,
particularly when speed of results is a key factor. As
input to conditioning the MPS model, the algorithms
require a training image (a 3D realistic representation of
the subsurface geology, but not constrained by any hard
data), which often has been in the form of a stochastic
realisation. With the improvement of process-driven
models, this could also provide an alternative as a
training image for multi-point conditioning. Caers and
Zhang (2004) present an excellent overview of MPS
techniques and training image construction.
It is important to mention that studying both ancient
and modern field analogues should still be at the
forefront in gaining a better understanding of different
fluvial systems and their stratigraphic development as
well as providing quantitative data for stochastic, or
other, modelling input. A well-exposed, coastal fluvial
outcrop in Portugal is being studied as an ancient
analogue to the Statfjord Formation in the Norwegian
North Sea. The work being carried out here involves
understanding the sedimentological facies that make up
the architectural elements important for fluvial reservoir
description and the stratigraphic ordering of such
elements for predicting facies and architectural styles
away from subsurface well locations. Quantitative data
(dimensions, geometries, interactions) are being collect-
ed on the various observed sandbody types at a variety
of scales (Fig. 9b) to be used as input to condition the
multi-scale outcrop models and to act as a database for
subsurface modelling input in analogous subsurface
fields. Here, the philosophy of pore-to-field modelling is
being employed to catalogue a library of quantified
multi-scale heterogeneities, the models generated and
their dynamic impact on flow in fluvial reservoirs.
Dynamic simulation of all scale-related models popu-
lated with appropriate petrophysical properties and
coupled with experimental design will quantify and
rank the relative importance of the various scales of
heterogeneities under various drainage strategies. This
systematic and quantitative approach to gathering and
modelling data from outcrop analogues will allow
personnel to better define (1) the scale of modelling
important for their particular fluvial reservoir, and
(2) the appropriate drainage strategy given the reservoir
rock properties. Another important aspect of this study
is to understand the nature of the paleosols and the
stratigraphic information that can be recorded within
them. By understanding the spatial and stratigraphic
development of the paleosols (McCarthy and Plint,
2003) and their relationships with associated sand
bodies, this information can be compared to similar
developments observed in paleosols in core data from
the Statfjord Formation in the Norwegian North Sea and
could be a further tool for helping with reservoir
zonation and sandbody prediction in the subsurface
environment. It is the integration of such stratigraphic
 K.J. Keogh et al. / Sedimentary Geology 202 (2007) 249268
studies that allows for more meaningful application of
geometrical data on sandbodies to be used as modelling
input within appropriate stratigraphic zones. It is
important to work more on the specifics of the different
types of fluvial systems (e.g. ephemeral or perennial,
fine-grained versus coarse-grained sediment supply,
specific dry land river systems, tide influenced systems).
In addition, most quantified outcrop work published, for
example, is on lower delta plain fluvial systems whereas
there are more types of systems that form equally
important fluvial reservoirs. The utilisation of sophisti-
cated stochastic modelling techniques is not a replace-
ment for geological understanding. It is rather that a
better understanding of fluvial systems should lead to
more geologically realistic stochastic modelling of the
flow patterns and behaviour of these systems. This
should in turn lead to enhanced well placement and
drainage strategies over field life scales and ultimately
production of more oil, more efficiently.
5. Conclusions
The discovery of the Snorre field on the Norwegian
continental shelf in 1979 prompted the Norwegian oil
industry into developing object-based stochastic mod-
elling algorithms to better model the complex architec-
tural styles in fluvial depositional systems. The current
focus on advanced reservoir description of fluvial
reservoirs on the Norwegian continental shelf for aiding
increased oil recovery includes revising reservoir
models of fields under evaluation. This article is
meant to provide an overview of the development of
stochastic fluvial modelling algorithms in the Norwe-
gian oil industry over the past few decades and to
address how these algorithms have enhanced reservoir
description of fields during appraisal, development and
production phases. The paper has addressed potential
and ongoing developments that can improve the
geological modelling of fluvial systems across multiple
scales of heterogeneities known to affect fluid flow in
these types of petroleum reservoirs.
In brief, this article has highlighted how early models
of fluvial reservoirs using interpolation techniques failed
to produce geologically correct models that give correct
predictions and can improve recovery within fluvial
reservoirs. Currently, stochastic modelling techniques
provide the only tools in the industry for modelling fluvial
systems that integrate all data types whilst being able to
represent the uncertainty and variability of key reservoir
parameters along with appropriate input to dynamic flow
simulations. The process phenomena controlling the
geometry, relative position and distribution of channels
265
have steadily been incorporated into stochastic algorithms
allowing more realistic fluvial geometries and stacking
patterns to be modelled. Many published and unpublished
data have highlighted that the characterisation of fluvial
reservoirs is difficult, given that heterogeneities and data
are on multiple scales. Therefore, the reservoir modelling
approach must be aligned to the problems as general-
isations are difficult.
Analogue studies can help to bridge the sampling gap
inherent in subsurface data collection. These types of
studies have provided understanding for building a more
deterministic framework within which appropriate
dimensional and relational data can be used as modelling
input to geologically-guided scale-related stochastic
models. Multi-scale scenario modelling and the resultant
outputs can help to better define the important hetero-
geneities that affect flow in fluvial reservoirs given a
specific drainage strategy.
The Norwegian oil industry is currently very active
in modelling fluvial reservoirs, and by addressing the
system on a basin-wide scale individual fields are bene-
fiting from a regional understanding of the evolution of
the system. Case studies highlighted in this article have
shown that a thorough appreciation and representation of
the variability and uncertainty of the system is required if a
full understanding of the connectivity and flow behaviour
of the reservoir is to be captured. Efforts are currently
being focused to address the philosophy of modelling
fluvial systems across a variety of scales to better predict
fluid flow behaviour and optimise field production.
Acknowledgements
The authors would like to acknowledge Statoil ASA
for the permission and financial support to publish this
article and their many colleagues who have contributed
to the discussions regarding the historical perspective,
scientific content and scope of this paper. Snorre staff
working with updating the reservoir model are thanked
for use of their data and results. The Graphics
Department of Statoil ASA are thanked for redrafting
the published figures. Two anonymous reviewers,
together with Bob Hoogendoorn and Chris Fielding are
acknowledged for their comments that have improved
the focus and clarity of this paper. Opinions stated in this
article are those of the authors and not of Statoil ASA.
References
Ainsworth, R.B., Sanlung, M., Duivenvoorden, S.T.C., 1999.
Correlation techniques, perforation strategies and recover factors:
an integrated 3D reservoir modeling study, Sirikit Field, Thailand.
AAPG Bulletin 83 (10), 15351551.
266 K.J. Keogh et al. / Sedimentary Geology 202 (2007) 249268
Augedal, H.O., Omre, H., Stanley, K.O., 1986. SISABOSA a
program for stochastic modelling and evaluation of reservoir
geology. In: Kjeller (Ed.), Conference Proceedings: Reservoir
Description and Simulation with Emphasis on EOR, pp. 118.
Bailey, W.R., Manzocchi, T., Walsh, J.J., Keogh, K.J., Hodgetts, D.,
Rippon, J.H., Nell, P.A.R., Flint, S.S., 2002. The effect of faults on
the 3-D connectivity of reservoir bodies: a case study from the East
Pennines Coalfield, U.K. Petroleum Geoscience 8 (3), 263278.
Blunt, M.J., Liu, K., Thiele, M.R., 1996. A generalized streamline method
to predict reservoir flow. Petroleum Geoscience 2 (3), 259269.
Bratvold, R., Holden, L., Svanes, T., Tyler, K., 1994. STORM:
integrated 3D stochastic reservoir modeling tool for geologists and
reservoir engineers. SPE 27563.
Bridge, J.S., Leeder, M.R., 1979. A simulation model of alluvial
stratigraphy. Sedimentology 26 (5), 617644.
Caers, J., Zhang, T., 2004. Multi-point geostatistics: a quantitative
vehicle for integrating geologic analogs into multiple reservoir
models. In: Grammer, G.M., Harris, P.M.M., Eberli, G.P. (Eds.),
Integration of Outcrop and Modern Analogs in Reservoir Modeling.
AAPG Memoir, vol. 80, pp. 383394.
Caers, J.K., Srinivasan, S., Journel, A.G., 2000. Geostatistical quantifi-
cation of geological information for a fluvial-type North Sea
reservoir. SPE Reservoir Evaluation & Engineering 3 (5), 457467.
Clemetsen, R., Hurst, A.R., Knarud, R., Omre, K.H., 1989. A computer
program for evaluation of fluvial reservoirs. In: Buller, A.T., et al.
(Ed.), North Sea Oil and Gas Reservoirs; II, Proceedings of the
North Sea Oil and Gas Reservoirs Conference. Graham & Trotman,
pp. 373385.
Curie, B.S., 1997. Sequence stratigraphy of nonmarine Jurassic
Cretaceous rocks, central Cordilleran foreland-basin system.
Bulletin of the Geological Society of America 109, 12061222.
Dalrymple, M., 2001. Fluvial reservoir architecture in the Statfjord
Formation (northern North Sea) augmented by outcrop analogue
statistics. Petroleum Geoscience 7 (2), 115122.
Damsleth, E., Omre, H., 1997. Geostatistical approaches in reservoir
evaluation. Journal of Petroleum Technology 37681, 498501.
Deutsch, C.V., Tran, T.T., 2002. FLUVSIM: a program for object-
based stochastic modeling of fluvial depositional systems.
Computers & Geosciences 28 (4), 525535.
Deutsch, C.V., Wang, L., 1996. Hierarchical object-based stochastic
modeling of fluvial reservoirs. Mathematical Geology 28 (7), 857880.
Dreyer, T., 1990. Sand body dimensions and infill sequences of stable,
humid-climate delta plain channels. In: Buller, A.T., Berg, E.,
Hjelmeland, O., Kleppe, J., Torster, O., Aasen, J.O. (Eds.), North
Sea Oil and Gas Reservoirs, vol. 2. Graham & Trotman, pp. 337351.
Dreyer, T., Scheie, A., Walderhaug, O., 1990a. Minipermeameter-
based study of permeability trends in channel sand bodies. AAPG
Bulletin 74 (4), 359374.
Dreyer, T., Falt, L.-M., Hy, T., Knarud, R., 1990b. Quantified fluvial
outcrop data for reservoir modelling purposes from the Escanilla
Fm, Sobrarbe sub-basin, Spanish Pyrenees. 13th International
Sedimentological Congress. IAS, Nottingham, United Kingdom,
p. 141.
Dreyer, T., Falt, L.-M., Hoy, T., Knarud, R., Steel, R., Cuevas, J.-L.,
1993. Sedimentary architecture of field analogues for reservoir
information (SAFARI): a case study of the fluvial Escanilla
formation, Spanish pyrenees. In: Flint, S.S., Bryant, I.D. (Eds.),
The Geological Modelling of Hydrocarbon Reservoirs and
Outcrop Analogues. Blackwell Scientific Publications, pp. 5780.
Dubrule, O., 1989. A review of stochastic models for petroleum
reservoirs. In: Armstrong, M. (Ed.), Geostatistics. Kluwer Publishers,
pp. 493506.
Dubrule, O., 1998. Geostatistics in petroleum geology AAPG
Continuing Education Course Note Series 38. American Associ-
ation of Petroleum Geologists, p. 250.
Dubrule, O., Damsleth, E., 2001. Achievements and challenges in
petroleum geostatistics. Petroleum Geoscience 7, S1S7 (supplement).
Egeland, T., Georgsen, F., Knarud, R., Omre, H., 1993. Multifacies
modelling of fluvial reservoirs. SPE 26502, 863872.
Elfenbein, C., Ringrose, P., Christie, M., 2005. Small-scale reservoir
modeling tool optimizes recovery offshore Norway. World Oil 10,
4550.
Georgsen, F., Egeland, T., Knarud, R., Omre, H., 1994. Conditional
simulation of facies architecture in fluvial reservoirs. In:
Armstrong, C., Dowd, P.A. (Eds.), Geostatistical Simulations.
Kluwer Academic Publishers, pp. 235250.
Haldorsen, H.H., 1986. Simulator parameter assignment and the
problem of scale in reservoir engineering. In: Lake, L.W., Carroll
Jr., H.B. (Eds.), Reservoir Characterization. Academic Press, Inc.,
pp. 293340.
Haldorsen, H.H., Damsleth, E., 1990. Stochastic modeling. Journal of
Petroleum Technology 20321, 404412.
Haldorsen, H.H., Brand, P.J., MacDonald, C.J., 1988. Review of the
stochastic nature of reservoirs. In: Edwards, S., King, P.R. (Eds.),
Mathematics in Oil Production. Oxford Science Publications,
Oxford, pp. 109210.
Hatly, A., 1994. Numerical facies modelling combining deterministic
and stochastic methods. In: Yarus, J.M., Chambers, R.L. (Eds.),
Stochastic Modelling and Geostatistics: Principles, Methods and
Case Studies. AAPG Computer Applications in Geology, vol. 3.
AAPG, pp. 109120.
Hauge, R., Skorstad, A., Holden, L., 1998. Conditioning a fluvial
reservoir on stacked seismic amplitudes. In: Buccianti, A., Nardi, G.,
Potenza, R. (Eds.), Proceedings of the Fourth Annual Conference of
the International Association for Mathematical Geology, Isola
d'Ischia, Italy, pp. 800805.
Henriques, A., Tyler, K., Hurst, A., 1990. Characterisation of fluvial
sedimentology for reservoir simulation modelling. SPE 18323.
Hoffman, K.S., Neave, J.W., Nilsen, E.H., Sarkisov, G., 2006.
Application of the fused fault block technique to fault network
modeling. SPE, paper 102375-MS.
Holden, L., Hauge, R., Skare, O., Skorstad, A., 1998. Modeling of
fluvial reservoirs with object models. Mathematical Geology 30
(5), 473496.
Holden, L., Mostad, P., Nielsen, B.F., Gjerde, J., Townsend, C.,
Ottesen, S., 2003. Stochastic structural modelling. Mathematical
Geology 35 (8), 899914.
Howell, J.A., Flint, S.S., 2002. Application of sequence stratigraphy
in petroleum geology and 3-D reservoir modelling. In: Armentrout,
J.M. (Ed.), Sequence Stratigraphic Models for Exploration and
Production: Evolving Methodology, Emerging Models and
Application Histories. 22nd Annual Gulf Coast Section SEPM
Foundation Bob F. Perkins research conference, Houston, Texas,
pp. 633644.
Johnson, H.D., Krol, D.E., 1984. Geological modelling of a heteroge-
neous sandstone reservoir: Lower Jurassic Statfjord Formation, Brent
Field. SPE 13050.
Jones, A., Doyle, J., Jacobsen, T., Kjonsvik, D., 1995. Which sub-
seismic heterogeneities influence waterflood performance? A case
study of a low net-to-gross fluvial reservoir. In: de Haan, H.J.
(Ed.), New Developments in Improved Oil Recovery. Geological
Society of London, pp. 518.
Journel, A.G., Gundeso, R., Gringarten, E., Yao, T., 1998. Stochastic
modelling of a fluvial reservoir: a comparative review of
 K.J. Keogh et al. / Sedimentary Geology 202 (2007) 249268
algorithms. Journal of Petroleum Science and Engineering 21 (12),
95121.
Karssenberg, D., Tornqvist, T.E., Bridge, J.S., 2001. Conditioning a
process-based model of sedimentary architecture to well data.
Journal of Sedimentary Research. Section B, Stratigraphy and
Global Studies 71 (6), 868879.
Kashib, T., Srinivasan, S., 2006. A probabilistic approach to
integrating dynamic data in reservoir models. Journal of Petroleum
Science and Engineering 50 (34), 241.
Keogh, K.J., 2002. Sequence stratigraphy and 3-D modelling of the
East Pennine Coalfield, UK: a deterministic and stochastic
approach. Doctor of Philosophy Thesis, University of Liverpool,
Liverpool, UK, 281 pp.
King, M.J., Datta-Gupta, A., 1998. Streamline simulation: a current
perspective. In Situ 22 (1), 91140.
Knai, T.A., Knipe, R.J., 1998. The impact of faults on fluid flow in the
Heidrun Field. In: Jones, G., Fisher, Q.J., Knipe, R.J. (Eds.),
Faulting, Fault Sealing and Fluid Flow in Hydrocarbon Reservoirs.
Geological Society of London Spec Pub, vol. 147, pp. 269282.
Lake, L.W., Carroll, H.B., 1986. Reservoir Characterization. Reservoir
Characterization Technical Conference. Academic Press. 650 pp.
Lecour, M., Cognot, R., Duvinage, I., Thore, P., Dulac, J.-C., 2001.
Modelling of stochastic faults and fault networks in a structural
uncertainty study. Petroleum Geoscience 7, S31S42 (Special Issue).
Legarreta, L., Uliana, M.A., Larotonda, C.A., Meconi, G.R., 1993.
Approaches to nonmarine sequence stratigraphy theoretical
models and examples from Argentine basins. In: Eschard, R.,
Doligez, B. (Eds.), Subsurface Reservoir Characterization from
Outcrop Observations. ditions Technip, Paris, pp. 125143.
Lia, O., Tjelmeland, H., Kjellesvik, L., 1996. Modelling of facies
architecture by marked-point models. In: Baafi, E.Y., Schofield, N.A.
(Eds.), Geostatistics Wollongong '96, Proceedings of the Fifth interna-
tional Geostatistics Congress, Wollongong, Australia, pp. 386397.
MacDonald, A.C., Halland, E.K., 1993. Sedimentology and shale
modeling of a sandstone-rich fluvial reservoir; upper Statfjord
Formation, Statfjord Field, Northern North Sea. AAPG Bulletin 77
(6), 10161040.
MacDonald, A.C., Hoye, T.H., Lowry, P., Jacobsen, T., Aasen, J.O.,
Gringheim, A.O., 1992. Stochastic flow unit modelling of a North
Sea coastal-deltaic reservoir. First Break 10 (4), 124133.
Manzocchi, T., Walsh, J.J., Nell, P., Yielding, G., 1999. Fault
transmissibility multipliers for flow simulation models. Petroleum
Geoscience 5 (1), 5363.
Martinius, A.W., Nss, A., 2005. Uncertainty analysis of fluvial outcrop
data for stochastic reservoir modelling. Petroleum Geoscience 11 (3),
203214.
Martinsen, O.J., Ryseth, A., Helland-Hansen, W., Flesche, H.,
Torkildsen, G., Idil, S., 1999. Stratigraphic base level and fluvial
architecture: Ericson Sandstone (Campanian), Rock Springs
Uplift, SW Wyoming, USA. Sedimentology 46 (2), 235259.
Matheron, G., Beucher, H., de Fouquet, C., Galli, A., Guerillot, D.,
Ravenne, C., 1987. Conditional simulation of the geometry of
fluvio-deltaic reservoirs. SPE 16753.
McCarthy, P.J., Plint, A.G., 2003. Spatial variability of palaeosols across
Cretaceous interfluves in the Dunvegan Formation, NE British
Columbia, Canada: palaeohydrological, palaeogeomorphological
and stratigraphic implications. Sedimentology 50 (6), 11871220.
Mjs, R., Nilsson, I., Wathne, E.I., Strmbck, A., Kjemperud, A.V.,
Schomacker, E.R., Cross, T.A., 2005. Stratigraphically controlled
depositional changes and reservoir characterisation of the Snorre
Field, North Sea. In: Weltje, G.J., de Boer, P., Vanden Berghe, J.,
van der Zwan, K., Stouthamer, E., Wolfert, H. (Eds.), 8th
267
International Conference on Fluvial Sedimentology. IAS, Delft
University, Netherlands, p. 206.
Nss, A., Falt, L.-M., Corregidor, J., Laegreid, U., 1997. 3D-models
from outcrop analogues; fluvial 3D architecture of the Escanilla
Fm., Spanish Pyrenees, American Association of Petroleum
Geologists 1997 annual convention. Annual Meeting Abstracts.
American Association of Petroleum Geologists and Society of
Economic Paleontologists and Mineralogists, Dallas, TX, United
States, pp. 8687.
Nordahl, K., Ringrose, P.S., Wen, R., 2005. Petrophysical character-
ization of a heterolithic tidal reservoir interval using a process-
based modelling tool. Petroleum Geoscience 11 (1), 1728.
North, C.P., 1996. The prediction and modelling of subsurface fluvial
stratigraphy. In: Carling, P.A., Dawson, M.R. (Eds.), Advances
in fluvial dynamics and stratigraphy. John Wiley & Sons Ltd.,
pp. 395508.
Olsen, T.R., 1995. Sequence stratigraphy, alluvial architecture and
potential reservoir heterogeneities of fluvial deposits: evidence from
outcrop studies in Price Canyon, Utah (Upper Cretaceous and Lower
Tertiary). In: Steel, R.J., Felt, V.L., Johannessen, E.P., Mathieu, C.
(Eds.), Sequence Stratigraphy on the Northwest European Margin.
Elsevier for Norwegian Petroleum Society, pp. 7596.
ren, P.-E., Bakke, S., Arntzen, O.J., 1998. Extending predictive
capabilities to network models. SPE Journal 3 (4), 324336.
Osland, R., Strmbck, A., Omdal, S., Krogh Berg, F., Seldal, M.,
2006. Reservoir Model of the Snorre Field 2006. RA-SNORRE-
00262, Statoil ASA (Internal Report).
Ottesen, S., Townsend, C., verland, K.M., 2005. Investigating the effect
of varying fault geometry and transmissibility on recovery: using a
new workflow for structural uncertainty modeling in a clastic
reservoir. In: Boult, P., Kaldi, J. (Eds.), Evaluating Fault and Cap
Rock Seal. AAPG Hedberg Series, vol. 2. AAPG, pp. 125136.
Reynolds, A.D., 1999. Dimensions of paralic sandstone bodies. AAPG
Bulletin 83 (2), 211229.
Ringrose, P.S., Nordahl, K., Wen, R., 2005. Vertical permeability
estimation in heterolithic tidal deltaic sandstones. Petroleum
Geoscience 11 (1), 2936.
Ritchie, J.S., Pilling, D., Hayes, S., 1998. Reservoir development,
sequence stratigraphy and geological modelling of Westphalian
fluvial reservoirs of the Caister C Field, UK Southern North Sea.
Petroleum Geoscience 4 (3), 203211.
Ryseth, A., Fjellbirkeland, H., Kloster Osmundsen, I., Skalnes, A.,
Zachariassen, E., 1998. High-resolution stratigraphy and seismic
attribute mapping of a fluvial reservoir: Middle Jurassic Ness
Formation, Oseberg Field. AAPG Bulletin 82 (9), 16271651.
Sandsdalen, C., Barbieri, M., Tyler, K., Aasen, J., 1996. Applied
uncertainty analysis using stochastic modelling. SPE 35533.
Shanley, K.W., McCabe, P.J., 1993. Alluvial Architecture in a
sequence stratigraphic framework a case history from the
upper Cretaceous of southern Utah, U.S.A. In: Flint, S.S., Bryant,
I.D. (Eds.), The geological modelling of hydrocarbon reservoirs.
IAS Special Publication, vol. 15. Blackwell Scientific Publica-
tions, pp. 2156.
Sharif, A., MacDonald, A., Kjorholt, O., Gilbrandsen, S., 2000.
Integrated reservoir characterisaion and uncertainty analysis using
stochastic methods and streamline simulations. SPE 65178.
Skare, ., Skorstad, A., Hauge, R., Holden, L., 1996. Conditioning a
fluvial model on seismic data. In: Baafi, E.Y., Schofield, N.A. (Eds.),
Geostatistics Wollongong '96, Proceedings of the Fifth International
Geostatistics Congress, Wollongong, Australia, pp. 465476.
Skorstad, A., Hauge, R., Holden, L., 1999. Well conditioning in a
fluvial reservoir model. Mathematical Geology 31 (7), 857872.
268 K.J. Keogh et al. / Sedimentary Geology 202 (2007) 249268
Sperrevik, S., Gillespie, P.A., Fisher, Q.J., Halvorsen, T., Knipe, R.J.,
2002. Empirical estimation of fault rock properties. In: Koestler,
A.G., Hunsdale, R. (Eds.), Hydrocarbon Seal Quantification. NPF
Special Publication, vol. 11. Elsevier Science B.V., Amsterdam,
pp. 109125.
Srivastava, R.M., 1994. An overview of stochastic methods for
reservoir characterisation. In: Yarus, J., Chambers, R. (Eds.),
Stochastic Modeling and Geostatistics: Principles, Methods and
Case Studies. AAPG Computer Applications in Geology, pp. 316.
Stanley, K.O., Jorde, K., Raestad, N., Stockbridge, C.P., 1989.
Stochastic modelling of reservoir sand bodies for input to reservoir
simulation, Snorre Field, northern North Sea, Norway. In: Buller,
A.T., et al. (Ed.), North Sea Oil and Gas Reservoirs; II, Proceedings
of the North Sea Oil and Gas Reservoirs Conference. Graham &
Trotman, pp. 91101.
Svanes, T., Martinius, A.W., Hegre, J., Maret, J.-P., Mjs, R., Molina,
J.C.U., 2004. Integration of subsurface applications to develop a
dynamic stochastic modelling workflow. AAPG Bulletin 88 (10),
13691390.
Sverdrup, E., Helgesen, J., Vold, J., 2003. Sealing properties of faults
and their influence on water-alternating-gas injection efficiency in
the Snorre Field, northern North Sea. AAPG Bulletin 87 (9),
14371458.
Theting, T.G., Rustad, A.B., Lerdahl, T., Stensen, J.., T., B., ren, P.-E.,
Bakke, S., Ringrose, P., 2005. Pore-to-field multi-phase upscaling for
depressurization. 13th European Symposium on Improved Oil
Recovery, Budapest.
Thore, P., Shtuka, A., Lecour, M., Ait-Ettajer, T., Cognot, R., 2002.
Structural uncertainties; determination, management, and applica-
tions. Geophysics 67 (3), 840852.
Tyler, K., Svanes, T., Omdal, S., 1993. Faster history matching and
uncertainty in predicted production profiles with stochastic
modeling. SPE 26420.
Tyler, K., Henriquez, A., Svanes, T., 1994. Modelling heterogeneities
in fluvial domains: a review of the influence on production
profiles. In: Yarus, J.M., Chambers, R.L. (Eds.), Stochastic
Modeling and Geostatistics: Principles, Methods and Case Studies.
AAPG, pp. 7789.
Weber, K.J., 1986. How heterogeneity affects oil recovery. In: Lake, L.W.,
Carroll Jr., H.B. (Eds.), Reservoir Characterization. Academic Press,
Inc., pp. 487544.
Wen, R., Martinius, A.W., Nss, A., Ringrose, P., 1998. Three-
dimensional simulation of small-scale heterogeneity in tidal deposits.
In: Buccianti, A., Nardi, G., Potenza, R. (Eds.), Proceedings of the
4th Annual Conference of the International Association of
Mathematical Geology (IAMG). Ischia, pp. 129134.
Wright, V.P., Marriott, S.B., 1993. The sequence stratigraphy of fluvial
depositional systems: the role of floodplain sediment storage.
Sedimentary Geology 86 (34), 203210.
Yielding, G., Freeman, B., Needham, D.T., 1997. Quantitative fault
seal prediction. AAPG Bulletin 81 (6), 897917.
Yoon, S., Malallah, A., Datta-Gupta, A., Vasco, D., Behrens, R., 2001.
A multiscale approach to production-data integration using
streamline models. SPE 56653.
 Available online at www.sciencedirect.com

Sedimentary Geology 202 (2007) 269 280

www.elsevier.com/locate/sedgeo

The stochastic river: The use of budget-capacity modelling as a

basis for predicting long-term properties of stratigraphic successions
John C. Tipper
Geologisches Institut, Albert-Ludwigs-Universitt Freiburg, Albertstrasse 23B, D-79104, Freiburg im Breisgau, Germany

Abstract

Sediment budget and sedimentation capacity are basic variables that can be used to describe the operation of sedimentation
systems of all types; they are generalisations of sediment supply and accommodation. The sediment budget at a site is defined as
the net flux of sediment across the site's boundaries; the sedimentation capacity is a measure of the deviation from equilibrium of
the sedimentation system concerned. The operation of a sedimentation system is described by first specifying the system's budget-
capacity relationship; this is most conveniently done by providing a probabilistic means of predicting budget from capacity. Once
the budget-capacity relationship has been specified, a model of the system is constructed; this model follows the general rules put
forward in the budget-capacity theory of sedimentation. The model enables the prediction of the patterns of deposition, stasis and
erosion that the system will produce in time; from these patterns can then be determined the properties of the stratigraphic
successions the system is expected to leave behind.
There are many kinds of budget-capacity model. One of these, the stochastic river, is particularly useful in predicting long-
term regularities in the stratigraphic thickness-time relationship. The stochastic river is a two-dimensional budget-capacity model
that mimics an idealised single-channel river. It is composed of a large number of similar sedimentation systems linked together in
a chain, with each system fed exclusively from its upstream neighbour. The systems are continually perturbed by having their
capacities altered by various external factors; these may be long-term regional-scale factors such as tectonically-induced basement
subsidence and isostatic adjustment, as well as short-term local-scale factors such as changes in river discharge. The capacity
perturbations are modelled as being driven by a random process; the perturbations of adjacent systems are spatially and temporally
coupled. This stochastic river reproduces many of the patterns of transport, deposition and erosion found in modern river systems.
In particular, it reproduces the irregularly alternating, downstream-migrating patterns of deposition and erosion that result, for
instance, in the development, migration and amalgamation of bars in gravel-bed rivers.
2007 Elsevier B.V. All rights reserved.

Keywords: Rivers; Sediment transport; Sedimentation; Sequence stratigraphy; Stratigraphic columns; Theoretical models

1. Introduction variability of the systems concerned affects substantially

Every stratigraphic succession is the record of the
operation of one or more sedimentation systems; from this
record, the stratigrapher attempts to work out how and
when those systems operated. The number and the
Tel.: +49 761 203 6476; fax: +49 761 203 6496.
E-mail address: john.tipper@geologie.uni-freiburg.de.
0037-0738/$ - see front matter 2007 Elsevier B.V. All rights reserved.
doi:10.1016/j.sedgeo.2007.01.026
the properties of the succession that finally results. If
several systems are involved in its production, a
succession is likely to be heterogeneous, especially if
those systems are highly variable ones. In contrast, a
succession produced by a single system is likely to be
relatively homogeneous, especially if that system operates
in much the same way for the whole of the time
concerned. Within these more homogeneous successions
270 J.C. Tipper / Sedimentary Geology 202 (2007) 269280
there commonly will be regularities to be found, e.g.,
patterns of facies repetition or patterns of bed thickness
variation. Such regularities can be expected to be
indicative of the system that produced them, therefore it
is important that they be recognised wherever they exist.
The regularities within stratigraphic successions are
often far from obvious, for two reasons. Firstly, they tend
to be regular only in a statistical sense; secondly, they
tend to be recognisable only over the long term. That
these regularities should have these characteristics is
hardly surprising, for most natural sedimentation sys-
tems tend both to fluctuate in the short term and to have
considerable short-range spatial variability. Perhaps the
only exceptions are sedimentation systems operating in
water that effectively is still, where sediment is supplied
exclusively in suspension. Other than that, it is hard to
think of any natural sedimentation system that is likely to
leave behind stratigraphic successions that are truly
regular.
Three questions arise naturally in this context, and it is
with these questions that this present paper is concerned.
Firstly, what should the stratigraphic successions pro-
duced by a given sedimentation system look like on
average, over the long term? Secondly, what sorts of
regularities should be expected within those successions?
Thirdly, how can these regularities be predicted? The
paper starts with a brief discussion of two contrasting
approaches that can be used to answer these questions;
then it presents a theory of sedimentation that can act as
the basis for suitable sedimentation models; finally it
introduces and illustrates the use of some of these models.
The paper deals exclusively with regularities in the
stratigraphic thickness-time relationship.
2. Two contrasting approaches
In principle, there are two approaches that may be
used to answer the questions posed. Either the records of
the past operation of the given sedimentation system are
analysed and used directly as a guide for prediction; or
that system is studied as it is operating today, and models
of it are built that then serve as predictors. The first
approach is an inductive one; the second is essentially
deductive. Both approaches are valuable, but each has its
difficulties. The chief difficulty with the inductive
approach is the requirement that records of the past
operation of the system have been preserved and can be
identified, i.e., that there do exist stratigraphic succes-
sions that are known to have been laid down as the result
of that system's operation. That must always be doubtful,
for how can it ever be possible to know not simply to
assume that particular stratigraphic successions were
laid down by the system of interest? The danger is that
regularities detected in particular successions will be
interpreted as being typical for that system, and that these
interpretations will in turn be used to erect elaborate yet
ultimately unrealistic models of how the system oper-
ates. Stratigraphers who wish to take the inductive
approach would do well to look back at the history of the
rise and fall of the Bouma sequence model for turbidite
sedimentation, described so eloquently by Shanmugam
(1997).
The chief difficulty with the deductive approach is
that models built of sedimentation systems operating
today cannot be tested over timespans that are relevant
stratigraphically. No natural sedimentation system has
ever been monitored over a timespan of more than a few
hundred years, even rudimentarily. Systematic records of
natural sedimentation systems for instance, records of
river discharges and flood levels rarely go back more
than a few decades. Detailed scientific monitoring of
the type necessary to test sedimentation models effec-
tively has been carried only over very short timespans
(typically weeks or months), and then only for a very few
systems. Perhaps the most impressive example of this
latter type of monitoring is the use of arrays of high-
resolution sensors to record sedimentation patterns in
estuarine and fluvial systems (Lawler, 2005).
This impossibility of testing sedimentation models
over stratigraphically relevant timespans is certainly a
problem, but it need not be an insuperable one. If a
model has a sufficiently strong theoretical basis, and if
the effects of its assumptions being violated are spelled
out clearly, then often it can still be used as a predictor,
even though most of its predictions are untestable. Three
points are critical to this argument: (1) the strength of the
theory underpinning the model, (2) the spelling out of
the effects of assumptions being violated, and (3) the
nature of the predictions that the model makes. The
strength of the theory refers to the degree to which that
theory can be deduced from first principles: a strong
theory is one that follows directly from first principles, a
weak theory is one that is largely empirical. The
assumptions referred to are those incorporated into a
model in order that it can be made to work; they form
what are termed the model's auxiliary hypotheses
(Oreskes et al., 1994; Tipper, in press). All predictions
made using a model are conditional on its auxiliary
hypotheses being true; therefore, strictly, no model
should ever be used when any of its assumptions are in
any way violated. In practice, it is rare for there not to be
violations of at least some of a model's assumptions,
especially when natural systems are being modelled.
The effects of such violations must then, however,
 J.C. Tipper / Sedimentary Geology 202 (2007) 269280 271

always be made absolutely clear. This can be done most
easily by formulating model predictions probabilistical-
ly, i.e., as expected values with associated confidence
bounds.
Of the two approaches, it is the essentially deductive
one that generally is the better provided always that
appropriate models are used. In studying the thickness-
time relationship in stratigraphic successions, one set of
models is particularly appropriate; these are models
based on the budget-capacity theory of sedimentation
(Tipper, 2000). Budget-capacity modelling offers a
straightforward yet rigorous way of predicting the likely
long-term properties of stratigraphic successions.
3. The budget-capacity theory of sedimentation
The budget-capacity theory of sedimentation stems
from the premise that the ways in which sedimentation
systems behave can be described formally in terms of the
interaction of two fundamental controls sediment
budget and sedimentation capacity. (These are general-
isations of the conventionally used ideas of sediment
supply and accommodation, which strictly are applicable
only in sedimentation environments that are entirely
erosion-free.) The sediment budget at a site (here
denoted by b) is defined as the net flux of sediment
across the site's boundaries (import flux minus export
flux). The sedimentation capacity of a site (denoted by c)
is defined as either the amount of space at the site in
which sediment can be deposited or the amount of space
at the site from which sediment can be released into
transport by erosion. Sedimentation capacity is measured
by the difference in elevation between the baselevel at
the site and the lithic surface (baselevel elevation minus
lithic surface elevation); baselevel is taken to be defined
as the elevation corresponding to the locally stable equi-
librium of the sedimentation system concerned. Sedi-
mentation capacity is effectively a measure of the
deviation from equilibrium of the sedimentation system
operating at the site.
Three types of sedimentation site can be recognised
(Fig. 1). If the budget and capacity values at a site are
both positive, the site is depositional; if the values are
both negative, the site is erosional; if the budget is zero,
there will be stasis neither deposition nor erosion
irrespective of the site's capacity. The sedimentation
systems at sites for which the capacity is non-zero are out
of equilibrium, by definition; any deposition or erosion
that occurs at these sites acts to move the systems back
towards their equilibria. The sedimentation system at a
site for which the capacity is zero is in equilibrium, by
definition; the budget at this site is necessarily zero, and
the system therefore remains in equilibrium.
It is not unreasonable to treat sediment budget and
sedimentation capacity as being essentially independent
variables, ones whose effects can in principle be sepa-
rated. However, it is hard to conceive of any type of
natural sedimentation system for which budget and
capacity would be entirely unrelated variables, if only
because each is ultimately a reflection of larger-scale
environmental factors such as climate change and
tectonics; see Tipper (2000) for discussion of this
point. The strength and the nature of the relationship
between budget and capacity will undoubtedly always

Fig. 1. The budget-capacity theory of sedimentation. (A) General contingency table. (B, C) Definition diagrams for depositional and erosional sites
respectively, for a given time interval. Solid arrows indicate import and export fluxes; size of arrow gives relative magnitude. c is capacity at start of
interval (baselevel elevation minus lithic surface elevation); b is budget during interval (import flux minus export flux); c is resulting capacity
change, manifest as change in lithic surface elevation (starred arrow).
272 J.C. Tipper / Sedimentary Geology 202 (2007) 269280

vary from site to site, but the variation between sites
within the same general sedimentation environment can
probably be assumed always to be less than the variation
between sites in environments that are radically different.
Thus the budget-capacity relationships at two sites on a
carbonate platform can probably be expected to be more
similar to each other than either would be to the budget-
capacity relationship at a site on a river floodplain. This
assumption that there is likely to be some more or less
consistent relationship between the budget and the
capacity values at sites at which similar sedimentation
systems are operating is used in the budget-capacity
theory to justify characterising sedimentation systems
directly in terms of the forms of their budget-capacity
relationships.
4. Describing the form of the budget-capacity
relationship
Most naturally occurring sedimentation systems can
be expected to have budget-capacity relationships that
are both temporally and spatially variable. It will
therefore probably be better to describe budget-capacity
relationships in a probabilistic rather than a deterministic
way, for instance by using some framework within which
the likely budget value at a site can be predicted for any
given value of capacity (Fig. 2). The construction of such
a framework for a given sedimentation system is in
principle straightforward: all that has to be done is to
specify the set of conditional frequency distributions for
the budget values, one distribution for each of the values
of capacity the system can have. (The frequency func-
tions for these distributions are denoted here by f (b|c).)
In practice, there will be considerable difficulties in
specifying these distributions, for any natural sedimen-
tation system. Firstly, it is unlikely that they will ever be
known exactly; secondly, their forms can be expected to
be different for different values of capacity, even within
the same system. Given these difficulties, there is only
one reasonable course of action to take: the distributions
should be modelled using some suitable general-purpose
frequency distribution, the parameters of which can be
adjusted as required.
One general-purpose frequency distribution that can
conveniently be used as a model in this context is part
of a set of distributions developed by Tipper (manu-
script submitted to Computers & Geosciences); it has
a unimodal frequency function, it is defined over a
limited range, and it has either three or four adjustable
parameters. The frequency function, f (x), is taken to
cover the interval 0 x a; it is constructed in two
sections, which join smoothly at the mode, m. For
0 x m, f (x) = fl (x)/R, where f l (x) = w l g(x) + (1
wl) g(m); for m x a, f(x) = f u(x)/R, where f u(x) =
w u g (x) + (1 w u) g(m). wl and wu are blending coef-
ficients; 0 (wl, wu) 1. R, a scaling constant, is given

Fig. 2. Probabilistic prediction of budget from capacity, using conditional frequency functions f (b|c). (A) Budget-capacity plane; for simplicity, only
positive quadrant is shown. |b| |c|, therefore f (b|c) exists only over unshaded part of quadrant. (B) Examples of f (b|c), along transects identified on
budget-capacity plane; see text for corresponding system behaviour. (C) Interpolation scheme used for determining shape parameters ( p, q) used in
constructing f( b|c) for experiments described in text. Blending coefficients calculated as follows: wl = min (2m/c, 1.0), wu = (2 2m/c, 1.0), where m
is the beta distribution mode (see text). Distribution 11: p = 1.25, q = 6.0, wl = 0.1, wu = 1.0. Distribution 22: p = 6.0, q = 6.0, wl = 1.0, wu = 1.0.
Distribution 33: p = 6.0, q = 1.75, wl = 1.0, wu = 0.26. Distribution 44: p = 1.6, q = 1.01, wl = 1.0, wu = 0.03. Distribution 55: p = 1.01, q = 1.01,
wl = 1.0, wu = 1.0.
 J.C. Tipper / Sedimentary Geology 202 (2007) 269280
by w u + (1 w u) a g(m) +(w u w l ) (m g(m) G pattern to be expected in many systems is what com-
(m)). g() and G() are the frequency and distribution
functions of a beta distribution defined

p1 overx
the interval
q1
0 a. g(x) is given by B p;qa
1
 ax 1 a , where
p and q are shape parameters (these have to be restricted
to values 1, in order that the frequency function be
unimodal), and B(p,q) is the beta function. G(x) is equal to
Bx/a( p,q), the incomplete beta function. The mode m is
given by a ( p 1) / ( p + q 2).
It is important to stress that there are no geological
reasons for selecting this particular general-purpose
frequency distribution for use in modelling budget-
capacity relationships. The reasons for selecting it are
instead entirely pragmatic: it provides an efficient
and flexible representation of frequency functions such
as f(b|c) that (1) are defined over a fixed interval, and (2)
can reasonably be expected always to be unimodal. The
definition of f (b|c) over a fixed interval is a direct
consequence of the budget-capacity theory of sedimen-
tation (which requires that b c 0 and 0 |b| |c|).
The expectation that f (b|c) will always be unimodal is
justified on the ground that the systems being modelled
are assumed all to have single equilibria; this is because
the budget-capacity theory expresses how sediment is
redistributed in systems with single locally stable
equilibria.
It is important also to be clear about the status and the
function of the parameters for which values have to be
supplied, i.e., the shape parameters ( p,q) and the
blending coefficients (wl, wu). These parameters have
no physical meaning whatsoever for any type of
sedimentation system; their function is solely to describe
how the system being studied is likely to behave when it
finds itself in a particular non-equilibrium state. Three
examples will serve to show how the values of these
parameters are chosen (Fig. 2B). The first example is a
system that is required to respond hardly at all to being
out of equilibrium; the appropriate combination of
parameter values is one that produces a function f (b|c)
that is skewed towards zero (Fig. 2B, distribution 11).
The second example is a system that is required always to
attempt to return to equilibrium as quickly and directly as
possible; the combination of parameter values has now to
produce a function f (b|c) that is skewed towards the
value c (Fig. 2B, distribution 33). The third example is
a system whose response is required to be essentially
independent of its deviation from equilibrium; the com-
bination of parameter values has now to produce a
function f (b|c) that is substantially flat over the entire
range 0c (Fig. 2B, distribution 55).
Most sedimentation systems will surely behave
differently at different values of capacity. A behaviour
273
monly is referred to as hill-climbing behaviour; this is
found in natural systems of very many types (see
Dawkins, 1997, for biological examples). A sedimen-
tation system that displays hill-climbing behaviour will
(1) respond independently of capacity when remote
from its equilibrium (i.e., at points far out in the plains
surrounding the hill to be climbed), (2) show strong
back-to-equilibrium behaviour when closer to its
equilibrium (i.e., at points on the hill's main slopes),
and (3) respond hardly at all when in its equilibrium's
immediate neighbourhood (i.e., at points close to the
hill's summit, where the ground surface flattens off).
This hill-climbing behaviour is produced using the set of
budget frequency functions shown in Fig. 2.
5. Some examples of budget-capacity modelling
Budget-capacity modelling is illustrated most easily
in one dimension, i.e., for a single sedimentation site.
The elaboration to two dimensions follows naturally.
5.1. 1D modelling, using a probabilistic budget-capacity
relationship
For simplicity, the site concerned is assumed to be
subsiding and non-erosional; the baselevel is assumed to
stay constant through time. The budget-capacity rela-
tionship of the sedimentation system at the site is
specified probabilistically, as described earlier; the site
is non-erosional, therefore budget frequency functions
have only to be provided for non-negative capacity
values. The lithic surface at the site starts at baselevel
and subsides at a specified rate, thereby changing the
site's capacity; the implied capacity change rate can be
held constant, or it can be allowed to vary in time
according to some given frequency distribution. The site
is supplied with sediment at a specified rate; this rate of
sediment supply the import flux can either be held
constant, or it can be allowed to vary in time according
to some given frequency distribution.
This one-dimensional budget-capacity model is imple-
mented most conveniently as a discrete-time simulation.
At each time step, a value for the capacity change is
obtained and added to the existing capacity value; the new
capacity is then used to determine the parameters of the
appropriate budget distribution. A value for the budget is
drawn at random from this distribution and compared
with a value obtained for the import flux; an amount of
sediment equal to the lesser of these two values is
deposited at the site. The site's capacity is adjusted to
reflect this deposition, then the next time step begins.
274 J.C. Tipper / Sedimentary Geology 202 (2007) 269280

Some results produced by this model are illustrated in
Fig. 3. They relate to three experiments: the first to
investigate the effects of changes in a system's ability to
respond to being out of equilibrium, the second to
investigate the effects of variability in the capacity change
rate, the third to investigate the effects of variability in the
capacity change rate and in the import flux. In the first
experiment, the import flux and the capacity change rate
were both held constant; the import flux was made very
slightly greater than the capacity change rate this
ensured that the system did not stray irrevocably from its
equilibrium. Seen clearly in the results of this experiment
(Fig. 3A) are what would automatically be identified by
sequence stratigraphers as parasequences. However, these
parasequences are not produced in the usually accepted
way, i.e., as the result of externally-driven capacity
increases (e.g., Coe, 2003). Instead, they result from the
system being unable to respond immediately to being out
of equilibrium. The quicker the system responds, the more
likely it is that the parasequences it produces will be
individually distinct, more-or-less uniform in size, and
developed close to baselevel (Fig. 3A, run 1); the slower
the system responds, the more likely it is that the
parasequences will be stacked together into irregular,
shallowing-upward sets (Fig. 3A, run 3). Parasequences
produced in this manner are essentially autogenic; they
are independent of externally-driven capacity changes
and therefore cannot be expected to be in any way
correlatable in age.
In the second experiment, the capacity change rate
was allowed to vary; it was given a gamma distribution
with a mean value very slightly less than the import flux.
Individually distinguishable parasequences are now not
produced, but the shallowing-upward motif is still
clearly recognisable (Fig. 3B; cf. Fig. 3A, run 3). In
the third experiment, the import flux and the capacity
change rate were both gamma distributed; the mean
import flux was again made very slightly greater than the
mean capacity change rate. Shallowing-upward sequences
are now recognisable only with difficulty (Fig. 3C). What
are seen instead are relatively long periods of relative
stability alternating with short intervals in which there are

Fig. 3. Capacity-time plots for experiments with 1D budget-capacity model with probabilistic budget-capacity relationship; see text for detailed
description. (A) Constant import flux; constant capacity change rate. Note more-or-less uniform parasequences in run 1, and irregular shallowing-
upward parasequence sets in run 3; example of each is marked with half-arrow. (B) Constant import flux; gamma distributed capacity change rate.
Examples of shallowing-upward sequences marked with half-arrows. (C) Gamma distributed import flux; gamma distributed capacity change rate.
Short intervals of rapid capacity change marked with pairs of solid arrows.
 J.C. Tipper / Sedimentary Geology 202 (2007) 269280 275

large capacity changes. This pattern of capacity fluctuation

is attributable to the fact that the import flux and the
capacity change rate vary independently. It is therefore
likely that there will be chance co-occurrences both of low
import flux and high capacity change rate and of high
import flux and low capacity change rate. The first of these
co-occurrences will result in the system becoming starved
of sediment, and hence in the system's capacity increasing
greatly; the system will then be far from its equilibrium
and will behave accordingly (for instance with a func-
tion f(b|c) similar to distribution 55 in Fig. 2B). The
co-occurrence of high import flux and low capacity
change will result in the system's capacity decreasing
greatly, sometimes to zero. The system will then behave
in its close-to-equilibrium manner, for instance with a
function f(b|c) similar to distribution 11 in Fig. 2B.

5.2. 1D modelling, using prescribed budget and

capacity variations

It may sometimes be necessary to predict the pro-

perties of stratigraphic successions produced at sites at
which the variations in time of budget and capacity are
prescribed. The budget-capacity relationship of the
sedimentation system concerned does then not need to
be specified probabilistically. An example of a budget-
capacity model developed for use in this situation is the
one-dimensional model for cyclic sedimentation de-
scribed by Tipper (2000). This model assumes (1) that
there is a constant rate of tectonic subsidence at the site in
question, (2) that baselevel varies cyclically in time, and
(3) that the sedimentation system is capable always of
keeping the lithic surface at baselevel, except when
either of two fixed budget limits are reached. (The effect
of these budget limits is to place upper bounds on the
sediment transport rate into and out of the site.) This
model can be formulated in terms of three dimensionless
variables, each of which has a limited range of values.
The form of the stable sedimentation cycle can be
obtained for any of the combinations of values these
variables can have, and from this cycle can then be
predicted the resulting stratigraphic succession (Fig. 4).
The complete set of possible forms of cyclic succession
is given in Tipper (2000).
5.3. 2D Modelling, using probabilistic budget-capacity
relationships
The simplest type of budget-capacity model in two
dimensions corresponds to what can be thought of as an
idealised single-channel river. It is made up of a large
number of essentially similar sedimentation systems
Fig. 4. 1D budget-capacity model for cyclic sedimentation; see Tipper
(2000). Elevation-time plot (top): heavy line is track of lithic surface,
light line is baselevel sinewave, open circle marks start of sequence,
filled circle marks start of sequence-bounding unconformity. Subsi-
dence rate = 45%; maximum deposition rate = 85%; maximum erosion
rate = 11%. Capacity-time plot (bottom): shaded area marks time when
system is out of equilibrium. Preserved stratigraphic succession (right):
BL is baselevel facies, DU is shallowing-upward facies, SU is
deepening-upward facies.
linked together in a chain, with each system fed
exclusively from its upstream neighbour. The systems
are continually perturbed by having their capacities
altered by various external factors; these may be long-
term regional-scale factors such as tectonically-induced
basement subsidence and isostatic adjustment, as well as
short-term local-scale factors such as changes in river
discharge. The capacity changes are driven by a random
process, and the changes at adjacent sites are spatially
and temporally coupled (Fig. 5). Sediment is supplied to
the most upstream site, the import flux values there being
drawn from some known frequency distribution. This
two-dimensional stochastic river is implemented by
repeatedly applying the discrete-time algorithm used in
the one-dimensional simulation described earlier. The
sites are treated in turn at each time step, with the import
flux to site i at time step j being taken to be the export
flux from site i 1 the site immediately upstream of i
at time step j 1.
Results from one run of this simulation are given in
Fig. 6. They illustrate some of the behaviour this model
276 J.C. Tipper / Sedimentary Geology 202 (2007) 269280
Fig. 5. The stochastic river. Signal model for externally-driven capacity changes (c). Capacity change at site i at time j + 1 is weighted average of
(1) capacity change at site i at time j, and (2) capacity changes at sites i 1 and i + 1 at time j + 1, and (3) independent random noise component ().
ws is space-weighting factor; wt is time-weighting factor.
commonly displays, most noticeably the quasi-periodicity
seen both in the capacity and in the downstream flux.
(This quasi-periodic behaviour is not unexpected: it is due
to the coupling of the capacity perturbations, which makes
the model behave as a spatio-temporal Markov process.)
The quasi-periodicity is both temporal and spatial:
(1) each system in the chain tends to alternate in time
between periods in which it is principally depositional and
periods in which it is principally erosional; (2) at any one
time there tend to be sets of principally depositional
systems alternating with sets of principally erosional
systems. The quasi-periodicity is irregular, and the loci of
deposition and erosion migrate unsteadily downstream.
As they do so, they leave behind a stratigraphic record
built of sets of diachronous sequences and unconformities.
These have a scaling that reflects (1) the budget-capacity
relationships of the individual sedimentation systems, (2)
the relative strength of the spatial and temporal coupling of
the capacity changes, and (3) the variability of the external
factors producing those changes.
Patterns of downstream-migrating areas of alternating
deposition and erosion are found both in meandering and
in braided river systems. They are found at many scales
(Frostick and Jones, 2002). The most obvious examples
involve simple bedforms such as dunes and ripples; other
examples, at a spatially larger scale, involve single
and composite bars and bar complexes. It is tempting
to suggest that the two-dimensional budget-capacity
model described here could be applied to all of these
examples effectively to say that there may be one
single scale-independent framework within which all
such patterns of alternating deposition and erosion
should be seen. That, however, would not be correct. It
is important to realise that the model described, in the
implementation given above, is only applicable when all
of the sedimentation systems are of substantially the
same type, i.e., when they share the same general budget-
capacity relationship. Whenever there is more than one
type of system, more complicated models are needed. As
an example, consider a budget-capacity model made for
a field of migrating dunes. In that model there would be
two very distinct types of sedimentation system to be
considered: those operating at sites where the flow is
attached, and those operating at sites between a flow
detachment point and its immediately-downstream flow
reattachment point. These two types of system can be
expected to have greatly different budget-capacity rela-
tionships. The boundaries of the areas of the bed over
which the two types of systems exist will necessarily
change in time as the dunes responsible for the flow
detachment and reattachment migrate downstream.
Therefore the system at any given site will switch
periodically from being of one type to being of the other
type, and then back again. Such switching behaviour
would have to be incorporated into the model being used,
and it would greatly increase that model's complexity.
It would seem to be at the reach scale that the
stochastic river is probably best applied, at least
initially. A particularly promising application would
seem to be to help explain the riffle-and-pool morphol-
ogy so characteristic of gravel-bedded meandering rivers
(Leopold et al., 1964). The irregularly alternating,
downstream-migrating patterns of deposition and ero-
sion seen in the example illustrated (Fig. 6B) are imme-
diately reminiscent of the patterns of bar development,
migration and amalgamation that occur in such rivers,
 J.C. Tipper / Sedimentary Geology 202 (2007) 269280 277

Fig. 6. The stochastic river. Three snapshots (1, 2, 3) from one experiment. Transport path is composed of chain of 2000 identical systems; each
system has budget-capacity relationship shown in Fig. 2, with the same relationship for negative as for positive capacities. (A) Capacity (heavy line)
and sediment flux (shaded area) plotted against distance down transport path. (B) Net change in lithic surface elevation. Note irregular quasi-
periodicity in capacity and flux, with downstream-migrating loci of deposition and erosion. PP marks the part of the transport path used to
synthesise the budget transition probability distribution.

and which are the means by which the riffle-and-pool
morphology is maintained. The characteristically fluc-
tuating transport flux patterns found in gravel-bed rivers
(e.g., Brewer and Passmore, 2002; Fuller et al., 2002;
Milan et al., 2002) are also reproduced by the model
(Fig. 6A).
6. Two strategies for predicting long-term patterns
in thicknesstime relationships
The thickness-time relationship in a stratigraphic
succession is conventionally represented using the
standard Barrell diagram of net thickness against time
(Fig. 7A; Barrell, 1917, Fig. 5; Tipper, 1998, Figs. 3, 4).
This diagram can be constructed for any sedimentation
site for which a time-series of budget values is given.
Budget-capacity models such as the ones described
above produce such time-series automatically as part of
their output; therefore the Barrell diagram for strati-
graphic successions left behind by a modelled system
can always be constructed.
The simplest strategy to use in predicting long-term
patterns in the stratigraphic thickness-time relationship
is as follows: (1) a one-dimensional budget-capacity
278 J.C. Tipper / Sedimentary Geology 202 (2007) 269280

Fig. 7. (A) The Barrell diagram. Main plot shows net thickness as a function of time. Black-shaded intervals on time plot (bottom) correspond to beds
(,, ) preserved in resulting stratigraphic succession (right). Inset (left) shows construction of diagram using succession of budget values obtained
from an appropriate transition probability distribution. (B) The budget transition probability distribution, for timestep i to timestep i + 1. Left-hand
diagram shows transition frequency function, f (bi + 1|bi), plotted on bi + 1 bi plane; darker greys indicate higher function values. Right-hand diagrams
show transition frequency function in cross-section, along transects 11 and 22 identified on bi + 1 bi plane. Succession of budget values is
obtained as follows: (1) a value is chosen at random from the function along the current transect (e.g. white filled circle on transect 11); (2) this value
sets the location of the next transect (e.g. 22), which lets the next value be chosen, etc.

model is set up for the system in question and run a large
number of times, (2) a Barrell diagram is constructed
from each of the calculated budget time-series, (3) the
set of Barrell diagrams is analysed in an appropriate
way. This one-dimensional modelling strategy certainly
will work, but its results will probably be highly sen-
sitive to the boundary conditions used in running the
model, in particular to the distribution assumed for the
import flux. This problem can be avoided by using the
following two-dimensional modelling strategy: (1) a
version of the stochastic river model is constructed for
a transport path long enough to remove any dependen-
cies on the flux conditions at the upstream and
downstream path ends (Fig. 6), (2) the model is run
first until the effects of whatever initial conditions were
chosen have worn off, (3) the model is then run further
and a record made of all of the budget time-series in the
central section of the transport path, (4) from this record
is synthesised the one-step transition probability distri-
bution for the budget values (Fig. 7B). This transition
probability distribution is a probabilistic encapsulation
of the stratigraphic results that the system being
modelled will be expected to produce over the long
term; from it can be constructed as many synthetic
stratigraphic successions as are required, each with its
associated Barrell diagram.
 J.C. Tipper / Sedimentary Geology 202 (2007) 269280
7. Some brief final comments: on data, on estimation,
and on modelling
As was stated initially, the aim of this present paper is
only to introduce and illustrate the use of budget-
capacity models. The systematic application of these
models to particular types of sedimentation system, the
analysis of the resulting Barrell diagrams, and the
recognition in those diagrams of significant stratigraphic
regularities, must wait for later work.
A clear prerequisite for this later work will be that the
budget-capacity relationships be known for the sedimen-
tation systems being modelled. This will require (1) that
appropriate data be available from which budget-capacity
relationships in natural sedimentation systems can be
estimated, and (2) that procedures exist by which the
estimation can be carried out. Substantial problems can be
foreseen with both of these requirements, especially with
the availability of the appropriate data. These data will
certainly best be obtained from arrays of high-resolution
sedimentation sensors (Lawler, 2005, and references
therein), from regularly re-recorded bathymetric surveys
(e.g. Degrendele et al., 2004), or from other continuously
recording remote-sensing techniques. Data obtained
using such techniques are commonly expensive, howev-
er; therefore it seems highly unlikely that they will be
available in sufficient quantity in the near future, except
for some rather specialised sedimentation systems.
The estimation problem is also a substantial one, for
there are no obvious procedures by which budget-capac-
ity relationships can be estimated directly from data for
modern sedimentation systems even assuming that
those data are available in sufficient quantity and with
sufficient quality. What will have to be done therefore is to
develop procedures for estimating budget-capacity rela-
tionships indirectly. One possible approach is via the
budget transition probability distribution, which can be
calculated from any time-series of measured sediment
flux or measured change in lithic surface elevation. A
suggested estimation procedure is as follows. Firstly,
some rough ideas are developed of how a system of the
type being considered is likely to respond when out of
equilibrium; these ideas are formalised probabilistically
using sets of budget frequency functions. A suitable
budget-capacity model is next constructed, for instance
the stochastic river. This model is run and the budget
transition probability distribution determined, then this
model-derived distribution is compared with the distri-
bution calculated from the supplied time-series data; the
forms of the budget frequency functions are modified in
the light of this comparison, and the run-compare-and-
modify steps of the procedure are repeated. The procedure
279
terminates once a result has been obtained that is deemed
satisfactory, on whatever grounds.
The estimate of the budget-capacity relationship
obtained by using this procedure should not be expected
to produce a model-derived budget transition probability
distribution that is a perfect fit to the distribution
calculated from the supplied time-series data. But that is
not necessarily a bad thing; it certainly does not mean that
the estimate of the budget-capacity relationship should be
rejected. To appreciate this point, it is necessary only to
remember that the budget-capacity relationship being
sought the true budget-capacity relationship is one
that applies to a potentially infinite set of sedimentation
systems of the same type, all of which are broadly the
same yet all of which are in various ways still different.
Therefore the behaviour of any one system from within
this set for instance, the system from which the data
used in the estimation were collected is properly seen
only as one example of how systems in the set can behave.
It must be stressed that the appropriate budget-capacity
relationship to use in modelling sedimentation systems of
a particular type is not necessarily the relationship that
provides the best fit to data obtained from any single
system. On the contrary, the appropriate budget-capacity
relationship to use is the one that best describes the ways
in which systems of the type in question usually respond
to being out of equilibrium.
These problems of data availability and estimation
might seem to be the final confirmation of what many
sedimentary geologists will already believe that there
is no place in stratigraphy for modelling that is as
theoretically oriented as that presented here. Modelling (it
will be argued) certainly has its part to play in stratigraphy,
but only when the models concerned are straightforward,
conventional, deterministic things the sort that have an
obvious relationship to sedimentation systems seen
operating today, the sort that are grounded in the equations
of basic physics, the sort that can be calibrated directly
with measurements made in the field. That argument is
superficially persuasive, but it will founder always on the
rock of time. For those conventional models impressive
though they undoubtedly are at reproducing the behaviour
of modern sedimentation systems are powerless
whenever the system being modelled is one that no
longer exists. In that situation when the structure of the
system cannot be observed, when the appropriate values
for the parameters cannot be measured, when the initial
conditions are inherently unknowable then it is models
of a very different sort that are needed. These models
appear at first glance to have little relevance to real-world
sedimentation systems; they seem abstract and incapable
of being implemented. Yet they are built on appropriate
280 J.C. Tipper / Sedimentary Geology 202 (2007) 269280
system-based theory, from which they derive their ability
to handle the unknowable, the unmeasureable and the
unobservable. That ability is what stratigraphy needs.
Acknowledgment
I thank the journal reviewers for their helpful
comments.
References
Barrell, J., 1917. Rhythms and the measurements of geologic time.
Geol. Soc. Amer. Bull. 28, 745904.
Brewer, P.A., Passmore, D.G., 2002. Sediment budgeting techniques
in gravel-bed rivers. In: Jones, S.J., Frostick, L.E (Eds.), Sediment
Flux to Basins: Causes, Controls and Consequences. Geol. Soc.
Lond. Spec. Publ., vol. 191, pp. 97113.
Coe, A.L. (Ed.), 2003. The Sedimentary Record of Sea-Level Change.
Cambridge University Press, Cambridge. 288 pp.
Dawkins, R., 1997. Climbing Mount Improbable. Penguin Books,
London. 308 pp.
Degrendele, K., Roche, M., Schotte, P., 2004. Three years of bedform
evolution on the Flemish banks, based on successive multibeam
surveying of two monitoring areas. In: Hulscher, S.J.M.H., Garlan,
T., Idier, D. (Eds.), Marine Sandwave and River Dune Dynamics
II, Workshop Proceedings. University of Twente, Netherlands,
pp. 4854.
Frostick, L.E., Jones, S.J., 2002. Impact of periodicity on sediment
flux in alluvial systems: grain to basin scale. In: Jones, S.J.,
Frostick, L.E. (Eds.), Sediment Flux to Basins: Causes, Controls
and Consequences. Geol. Soc. Lond. Spec. Publ., vol. 191,
pp. 8195.
Fuller, I.C., Passmore, D.G., Heritage, G.L., Large, A.R.G., Milan, D.J.,
Brewer, P.A., 2002. Annual sediment budgets in an unstable gravel-
bed river: the River Coquet, northern England. In: Jones, S.J., Frostick,
L.E. (Eds.), Sediment Flux to Basins: Causes, Controls and
Consequences. Geol. Soc. Lond. Spec. Publ., vol. 191, pp. 115131.
Lawler, D.M., 2005. The importance of high-resolution monitoring in
erosion and deposition dynamics studies: examples from estuarine
and fluvial systems. Geomorphology 64, 123.
Leopold, L.B., Wolman, M.G., Miller, J.P., 1964. Fluvial Processes in
Geomorphology. Freeman, San Francisco. 522 pp.
Milan, D.J., Heritage, G.L., Large, A.R.G., 2002. Tracer pebble
entrainment and deposition loci: influence of flow character and
implications for riffle-pool maintenance. In: Jones, S.J., Frostick, L.E
(Eds.), Sediment Flux to Basins: Causes, Controls and Conse-
quences. Geol. Soc. Lond. Spec. Publ., vol. 191, pp. 133148.
Oreskes, N., Shrader-Frechette, K., Belitz, K., 1994. Verification,
validation, and confirmation of numerical models in the earth
sciences. Science 263, 641646.
Shanmugam, G., 1997. The Bouma Sequence and the turbidite mind
set. Earth-Sci. Rev. 42, 201229.
Tipper, J.C., 1998. The influence of field sampling area on estimates of
stratigraphic completeness. J. Geol. 106, 727739.
Tipper, J.C., 2000. Patterns of stratigraphic cyclicity. J. Sediment. Res.
70, 12621279.
Tipper, J.C., in press. Models that talk back. In: de Boer, P.L., Postma,
G., van der Zwan, C.J., Burgess, P.M., Kukla, P. (Eds.), Analogue
and Numerical Forward Modelling of Sedimentary Systems; from
Understanding to Prediction. Int. Assoc. Sedimentologists, Spec.
Publ., 39.
 Available online at www.sciencedirect.com

Sedimentary Geology 202 (2007) 281 296

www.elsevier.com/locate/sedgeo

Influence of superimposed bedforms and flow unsteadiness on

formation of cross strata in dunes and unit bars
A.J.H. Reesink , J.S. Bridge
Department of Geological Sciences, Binghamton University, P.O. Box 6000, Binghamton, NY 13902-6000, USA

Abstract

Formation of angle-of-repose cross strata on the lee side of dunes and unit bars is dependent on three main grain-sorting
mechanisms: (1) presorting of sediment that arrives at the lee side of the bedform, related to superimposed bedforms and longer-
term variations in water flow and sediment transport; (2) sorting due to differential deposition of sediment on the lee side and
associated grain flows, and; (3) movement of sediment on the lee side by the water currents in the lee side flow-separation zone.
Although most emphasis has been put on mechanism (2), recent field and experimental studies of the dynamics of sandy and
gravelly dunes and unit bars show that mechanism (1) is at least as important.
Bedforms superimposed on the backs of dunes and unit bars are ubiquitous, and are composed of size sorted sediment. These
superimposed bedforms travel faster than the host bedform and overtake it. If the overtaking bedform has a relatively large height, the lee
side of the host bedform becomes reduced in slope and decelerates as the superimposed bedforms overtakes. This results in a
reactivation surface that is commonly lined with relatively fine-grained sediment. However, if an overtaking superimposed bedforms
has a relatively small height (e.g., a ripple or bedload sheet), the lee side of the host bedform remains at the angle-of-repose. The sediment
within the overtaking bedform forms a relatively thick cross stratum on the lee side of the host bedform. Arrival of the trough of the next
superimposed bedform at the crest of the host bedform results in deposition of a relatively thin stratum. The thickness of the cross stratum
couplet formed in this way is related to the height and length of the superimposed bedform and the height of the host bedform, which are
related to flow depth and dimensionless bed shear stress on the back of the host bedform. The grain size of the sediment in the
superimposed bedforms varies with their height, and this is reflected in the grain size of successive cross strata. The formation of cross-
stratified open-framework gravel is associated with size sorting of sand and gravel in superimposed bedforms, and further size sorting
associated with flow-separation on the lee side of large host bedforms.
Published by Elsevier B.V.

Keywords: Cross-stratification; Cross strata; Dunes; Unit bars; Superimposition; Bedforms; Flow unsteadiness

1. Introduction widely used to interpret conditions of flow and sediment

Cross-stratification formed by dunes and unit bars is
the most common sedimentary structure in river-channel
deposits, and is common in many other depositional
environments. Ancient cross-stratified deposits are
Corresponding author.
E-mail address: areesin1@binghamton.edu (A.J.H. Reesink).
0037-0738/$ - see front matter. Published by Elsevier B.V.
doi:10.1016/j.sedgeo.2007.02.005
transport, such as water depth, flow velocity magnitude
and direction during deposition (review in Bridge,
2003). Cross-strata are defined primarily by spatial var-
iation in grain size, which in turn affects spatial variation
in permeability, and the movement of fluids through
cross-stratified deposits. Grain size sorting in cross strata
is determined by three main factors (Fig. 1): (1) presorted
sediment that arrives at the brink point, related to
282 A.J.H. Reesink, J.S. Bridge / Sedimentary Geology 202 (2007) 281296

Fig. 1. Controls on the grain size sorting in cross strata formed by dunes.

superimposed bedforms and longer-term unsteadiness in
water flow and sediment transport; (2) sorting due to
differential deposition of sediment on the lee slope and
associated grain flows, and; (3) movement of sediment
on the lee side by water currents in the lee side flow-
separation zone. Factor (2) is commonly taken as dom-
inant. However, recent experimental and field studies
indicate that the role of bedform superimposition (e.g.,
ripples or bedload sheets on dunes; ripples or dunes on
unit bars) in the formation of cross-stratification is much
more important than commonly appreciated.
This paper will start with a short outline of our
understanding of the formation of cross strata, identify
gaps in our knowledge, and address the common mis-
conception that grain flows cause the formation of most,
if not all, cross-stratification. Then, new experimental
and field data will be presented that demonstrate the
importance of presorting of sediment, specifically super-
imposition of bedforms, in determining the grain size
sorting in cross strata. Finally, further research on this
topic is suggested in order to allow for more accurate
prediction of spatial variation in porosity and perme-
ability (of concern to hydrogeologists and petroleum
geologists), and to improve quantitative interpretation of
ancient flow and sediment transport conditions.
2. Controls on the grain size sorting in cross strata
formed by dunes and unit bars
2.1. Presorting of sediment due to superimposed bed-
forms and longer-term unsteadiness in water flow and
sediment transport
2.1.1. Long-term flow unsteadiness
Changes in the rate and grain size of sediment supply
to the crests of bedforms may be related to relatively
long-term flow unsteadiness associated with weather-
related floods, bank-slumps, and cyclical variations in
discharge such as caused by snowmelt and freezing in
cold regions, or tides. The grain size of bedload is ex-
pected to decrease as flow stage and bed shear stress
decrease (review in Bridge, 2003). It has been suggested
that relatively coarse-grained cross strata are deposited at
high flow stage, and that fine-grained cross strata
are deposited at low flow stage (Smith, 1972, 1974;
Steel and Thompson, 1983; Smith, 1985). Alternations
of sandy cross strata and muddy drapes in tidal dunes
(called tidal bundles: Fig. 2) are generally taken to be
formed by alternating periods of tidal current flow and
tidal slack water (e.g. Visser, 1980; Allen, 1982; Nio
et al., 1983). Systematic variation in the thickness of tidal

Fig. 2. Tidal bundles in an excavation in the Oosterschelde (courtesy of Dr. Van den Berg). Height is approximately 2 m.
 A.J.H. Reesink, J.S. Bridge / Sedimentary Geology 202 (2007) 281296 283

bundles has been related to variations in tidal range
associated with neap-spring cycles and with the diurnal
inequality of the tide. However, the exact nature of
the effects of long-term flow unsteadiness on sorting in
cross strata has never been examined systematically
using flume experiments.
2.1.2. Superimposed bedforms
Related to flow unsteadiness is temporal change in
the geometry of dunes, and the likelihood that dunes of
different sizes will be superimposed (reviews in Allen,
1982; Bridge, 2003). However, experimental and field
data show that ripples and bedload sheets occur on the
backs of dunes, and bedload sheets, ripples and dunes
occur on the backs of unit bars, under both steady and
unsteady flow conditions (Allen, 1982; Bridge, 2003;
Venditti et al., 2005; Lunt and Bridge, 2007: Fig. 3).
Furthermore, it has been recognized that many bedforms
grow by amalgamation of smaller bedforms, and that
small, embryonic bedforms are continually forming on
the backs of larger ones (Mantz, 1978; Allen, 1982;
Raudkivi and Witte, 1990; Ditchfield and Best, 1992;
Gabel, 1993; Coleman and Melville, 1994, 1996;
Raudkivi, 1997; Leclair, 2002). The influence of these
ubiquitous superimposed bedforms on the nature of
cross strata has never been investigated systematically.
The influence of superimposed bedforms on the
development of cross strata in unit bars was investigated
early on by Hooke (1968), Smith (1972), and Rust (1984).
Hooke (1968) and Smith (1972) observed how granules
accumulate in the troughs of sandy superimposed bed-
forms, such that granule cross strata in unit bars is
associated with the passing of the troughs of the super-
imposed bedforms (Fig. 4). Rust (1984) argued that open-
framework gravel occurred within isolated dunes super-
imposed on unit bars, whereas bedload sediment in the
absence of superimposed dunes was a mixture of sand and
gravel (Fig. 4). Therefore, cross strata formed of open-
framework gravel, bounded by cross strata formed of sandy
gravel, were attributed to deposition from superimposed
dunes alternating with deposition from non-dune regions.
Experiments with sandy gravels by Lunt and Bridge (2007)
indicated that bedload sheets and dunes superimposed on
unit bars are composed mainly of open-framework gravels,

Fig. 3. Bedforms superimposed on dunes and unit bars. (A) Bedload sheets on the back of sand dune. Dark areas are troughs with relatively fine-
grained sediment. Flume is 0.6 m wide. (B) Ripples on sand dunes. Flume is 0.6 m wide. (C) Ripples on a unit bar in sand. Flume is 0.6 m wide.
(D) Dunes and people on crest of a sandy unit bar, Congaree River. Trough of unit bar is to left.
284 A.J.H. Reesink, J.S. Bridge / Sedimentary Geology 202 (2007) 281296

Fig. 4. Grain size sorting in unit bars resulting from size sorting in superimposed dunes, according to (A) Hooke (1968) and Smith (1972), and
(B) Rust (1984). s is sand, g is gravel, sg is sandy gravel and ofg is open-framework gravel.

as suggested by Rust (1984). The highest superimposed
bedforms have the best developed open-framework
gravels, but the smaller superimposed bedforms are
composed of sandy gravel. Therefore, cross strata in unit
bars composed of alternations of open-framework gravel
and sandy gravel (Fig. 5) are related to variation in the
height of superimposed bedforms (c.f. Rust, 1984). These
experiments also indicate that a veneer of suspended fine-
grained sand is deposited in the trough of bedload sheets
and dunes, and partly infiltrates the coarse-grained
sediment in this region. Such co-accumulation of small
and large grains in bedform troughs results in highly
bimodal sediments. As a superimposed bedform trough
arrives at the brink point of a host bedform, fine-grained
suspended load and bimodal trough sediment are deposited
on the lee side of the host. As the superimposed bedform
arrives, dominantly coarse-grained sediment is deposited
on the lee side of the host (Fig. 6).
If superimposed bedforms are high relative to the
host bedform (Hs/H N 0.2 where Hs and H are heights
of superimposed and host bedforms, respectively), the
lee side slope of the host bedform becomes reduced as
the superimposed bedforms approaches its crest, due
to erosion associated with the flow-separation zone of
the superimposed bedforms. This overtaking process
produces a low-angle reactivation surface rather
than an angle-of-repose cross stratum (Fig. 6), and the
internal structure of the host bedform is composed of
cross-sets made by the superimposed bedforms with
set boundaries dipping downstream at less than the
angle-of-repose (Fig. 6). However, an identical
structure (called compound cross-stratification by
Allen, 1982) can be formed by superimposed bed-
forms migrating over a larger host bedform that does
not have a lee face at the angle-of-repose, because it
is either developing or dying out (Fig. 7). This is

Fig. 5. Experimental data from Lunt and Bridge (2007) showing cross strata of varying thickness and grain size within a unit bar, associated with
deposition from superimposed bedforms of varying height. Peel is 0.5 m across and 0.2 m high.
 A.J.H. Reesink, J.S. Bridge / Sedimentary Geology 202 (2007) 281296 285

Fig. 6. (A) Formation of cross strata by migration of small bedforms superimposed on a dune back as observed in recent experiments (see also Figs. 4
and 5). (B) Formation of a reactivation surface by migration of a large bedform over a dune.

commonly produced as ripples or dunes migrate over the wedge of sediment that eventually flows down the lee
developing unit bars (e.g., Jackson, 1976; Levey, side depends on the spatial variation of deposition rate on
1978; Bridge et al., 1986; Best et al., 2003; Lunt et al., the lee slope. For coarse-grained bedload sediment
2004).

2.2. Sorting due to differential deposition of sediment

on the lee side and associated grain flows

Bedload sediment and near-bed suspended sediment

that arrives at the brink point of the bedform is deposited
on its lee side at a distance downstream from the brink
point that is controlled by: (1) grain velocity; (2) height
above the bed; (3) grain settling velocity, and; (4) tur-
bulence in the separated flow (Fig. 8; Jopling, 1965;
Allen, 1963, 1968, 1982; Hunter, 1985b). Grains travel
further downstream from the brink point as their velocity
and height above the bed increase, and as their settling
velocity (hence grain size) decreases. Therefore, the rate
of deposition and the grain size of deposited sediment
decrease down the lee slope, resulting in building of a
down slope thinning wedge of sediment that fines down
slope (i.e. tangentially).
The slope of the outer edge of the wedge of sediment
that builds up on the lee side increases to the angle of
initial yield, at which point a grain flow occurs and
comes to rest at the so-called residual angle. The angle of
initial yield is commonly around 35, and up to 10
larger than the residual angle after grain flow. In reality,
the angle of yield may not reach its maximum value prior
to grain flow, because turbulence in the flow-separation
zone can cause the sediment wedge to fail prematurely. Fig. 7. (A) Compound cross strata formed by migration of dunes over
The lee side wedge builds quickly if sediment transport unit bar, Congaree River. Angle-of-repose cross strata associated with
rate is high, and if most of the sediment load is near the the avalanche face of the unit bar only occur at the downstream edge of
the unit bar. (B) Close-up of outer edge of unit bar deposits showing
bed. Thus, grain flow frequency increases with sediment grain size sorting due to dune migration. The finest layers are due to
transport rate, provided that most material is deposited passing of the dune troughs. Note internal lamination and distinct
on the upper lee slope. The size (cross-sectional area) of reactivation surface below trowel.
286 A.J.H. Reesink, J.S. Bridge / Sedimentary Geology 202 (2007) 281296
Fig. 8. Idealized paths of grain fall on bedform lee side for grains of
different size and distance from the bed.
moving relatively slowly, most sediment is deposited
close to the brink point, and the failing wedge of sedi-
ment is relatively small. The lack of settling of suspended
sediment on the lower lee side results in an angular
contact between the lee side slope and the bedform
trough (Fig. 9). For fine-grained near-bed sediment
moving relatively fast, much of the sediment is
transported well past the brink point and even into the
trough, and the wedge would be larger. The build up of
fine-grained sediment at the base of the lee side reduces
the angle of the lower slope (tangential bases: Fig. 9).
Small sediment wedges at the point of failure result in
small grain flows that may not involve enough material
to reach the base of the lee slope. However, subsequent
addition of sediment from other small grain flows can
produce a large composite wedge that can reach the base
of the lee slope (Htu et al., 1995). There are no realistic
models for the size of the avalanching wedges relative to
bedform size, nor for the thickness of grain flow
deposits. Grain size sorting within grain flows originat-
ing from composite wedges is also poorly known.
Differential rates of deposition of different sediment
sizes on the angle-of-repose lee side of bedforms, and
associated grain flows, are commonly taken as the most
important controls of grain size sorting in cross strata
(Allen, 1963; Jopling, 1965; Allen, 1968, 1970, 1982;
Hunter, 1985a,b; Hunter and Kocurek, 1986; Carling and
Glaister, 1987; Makse, 1997; Makse et al., 1997a,b;
Koeppe et al., 1997; Kleinhans, 2004, 2005). Grain size
sorting during discrete grain flows has been related to
Fig. 9. Grain size sorting by grain flow and grain fall on bedform lee side, and geometry of cross strata, for low and high sediment transport rate.
Arrows point to direction of decreasing grain size.
kinetic sieving (Middleton and Southard, 1984) and to
dispersive grain stress (Bagnold, 1954; Sallenger, 1979),
although the latter mechanism has been substantially
discredited (Legros, 2002). With kinetic sieving, small
grains settle between the spaces created between the
larger grains during grain shearing and dilation. The
large grains move towards the faster-moving outer edges
of the grain flow, where they can move down slope more
quickly than smaller grains. The largest grains on top of
the grain flow roll down the slope relatively easily be-
cause they have small friction angles. Therefore, the
grain flow deposit becomes coarser grained down slope
and towards its outer edge (Fig. 9). The bedform must be
high enough (lee side must be long enough) for this
process to be effective. At relatively low sediment
transport rates, grain flows are intermittent, and it may be
possible to distinguish relatively coarse-grained grain
flow strata from relatively fine-grained grain-fall stra-
ta (Jopling, 1965; Hunter, 1985a). At high sediment
transport rate, when grain flows are very frequent, much
fine-grained suspended sediment is incorporated during
avalanching and also deposited on the lower lee slope
(Allen, 1970, 1982; Hunter, 1985a,b; Carling and
Glaister, 1987; Kleinhans, 2004). This may counteract
the sorting caused by kinetic sieving, such that grain size
sorting may not be obvious.
Avalanching of sediments with a large range of
grain size, such as sandy gravels, may lead to complete
segregation of different grain sizes along the angle-of-
repose slope, and to absence of cross strata (e.g. Koeppe
et al., 1997; Makse, 1997; Makse et al., 1997a,b, 1998;
Fig. 10). This is due to the large differences in pivoting
(friction) angles between the different grain sizes in a
mixture. The largest grains moving over smaller grains
have the smallest pivoting angles and readily move to
the base of the lee slope. This means that coarse-grained
cross strata in many sandy gravels cannot be related to
sorting in grain flows and instead must be related to
superimposed bedforms or longer-term flow unsteadi-
ness. Unfortunately, most of the experimental studies
(cited above) of the nature of sediment deposition and
 A.J.H. Reesink, J.S. Bridge / Sedimentary Geology 202 (2007) 281296
Fig. 10. Experimental grain flow of sandy gravel produces grain size separation but no cross strata.
grain flow on lee slopes did not involve serious
consideration of variations in rate and size of sediment
arriving at the brink point. It is clear, however, that grain
flow cannot explain many types of cross strata.
2.3. Movement of sediment on the lee side by the water
currents in the flow-separation zone
The separated (reverse) flow over the bedform lee
side can reduce the velocity of the grain flows and move
sediment up the lower lee slope (in some cases as rip-
ples: Boersma et al., 1968), causing a slope decrease
there. It is readily observed in flume experiments that
grain flows do not always reach the base of the lee slope.
However, cross strata commonly continue into the
trough (Fig. 9). This tangential cross-stratification can-
not be formed by grain flows, and can only be associated
with deposition of suspended sediment and/or action
of the reverse-flow. These processes also need to be
investigated systematically.
3. Recent experimental studies
Experiments were initiated in the Binghamton flume
in order to: (1) generate cross strata within dunes and unit
bars formed in sands and sandy gravels under steady and
unsteady flow conditions; (2) describe the geometry,
grain size sorting, porosity and permeability of cross
strata formed under different conditions; (3) evaluate
the controls on the formation of the cross strata, and;
(4) examine the possibility of using the cross strata to
interpret formative flow conditions in much more detail
than possible now. A prime objective is to be able to
distinguish the effects of sorting from grain flows and
presorting on the back of the bedforms. In reality, grain
287
size sorting probably results from a combination of
presorting, lee side grain flows, and reworking by the
lee side flow vortex. Preliminary results are presented
here for the cases of cross strata formed by ripples and
bedload sheets superimposed on dunes, and ripples and
dunes superimposed on unit bars. The results are com-
pared with cross strata formed by subaqueous grain
flows developed in a Hele-Shaw cell, in which presorting
can be assumed negligible.
3.1. Binghamton flume and experimental procedure
The laboratory flume in the Department of Geological
Sciences, Binghamton University, is 0.61 m wide, 0.4 m
deep, and 7.6 m long. It recirculates water and sediment
up to about 50 mm in diameter, using a combined
propeller pump (for water and suspended sediment) and
diaphragm pump (for water and gravelly bedload
sediment). Sand and water can be recirculated using
only the propeller pump. A conveyor belt at the upstream
end of the flume allows introduction of sediment across
the width of the flume at a constant rate, for use in
experiments with aggradation. An ultrasonic depth
profiler is used for measuring bed elevation to within
0.1 mm, and can easily resolve mm-high bedforms. A
computerized instrument carriage can be used to move
the depth profiler along the flume at a fixed rate. A
Marsh-McBirney electromagnetic flow meter with a
12 mm diameter sensing head is available for measuring
profiles of flow velocity.
Two sediment mixes were used in the experi-
ments (Fig. 11): moderately well-sorted medium
sand (D50 = 0.43 mm), and poorly-sorted sandy gravel
(D50 = 2 mm). Bedforms were developed from an ini-
tially flat bed. The flume was run for at least the time it
288 A.J.H. Reesink, J.S. Bridge / Sedimentary Geology 202 (2007) 281296

Fig. 11. Grain size distribution of the sand and sandy gravel used in the experiments.

takes one dune to migrate through the entire length of
the flume before measurements were made. Unit bars
were developed in the flume by overloading the
upstream end of the flume with sediment. Super-
imposed bedforms of specific types and sizes were
controlled by varying the sediment feed, flow velocity
and water depth (Table 1). Watersurface elevations
were measured throughout the experiments through
the sidewalls along the length of the flume. Water
surface slopes in the downstream half of the flume
were derived from regression analysis of repeated
measurements of watersurface elevation. The experi-
ments were recorded on videotape, which facilitated a
detailed analysis of the depositional processes. Fol-
lowing each experiment, the flume was drained,
photographs were taken of the bed and the deposits
seen through the flume sidewalls, and cores were taken
in various orientations. Epoxy-resin peels were made
of some of the cores, and others were used for grain
size analysis and determination of permeability and
porosity.
The experiments are true scale, meaning that the
flow, sediment transport and bedforms are representa-
tive of natural rivers with relatively small flow depths
and subcritical flow. The results could be upscaled to
deeper flows insofar as the heights and lengths of dunes
are proportional to flow depth, the heights of unit bars
are proportional to flow depth, and their lengths are
proportional to flow width. However, upscaling must
also involve other parameters, such as grain size, which
must be scaled up in the same ratio as flow depth.
3.2. Grain flow apparatus and experimental procedure
The Hele-Shaw cell is a metal box (0.04 m wide,
0.5 m long and 0.2 m high) with a removable side plate

Table 1
Experimental conditions used in the flume
Sediment Depth average Average water Average water Froude Host bedform Superimposed bedforms Aggradation
type velocity (m/s) a depth (m) a surface slope () number rate (mm/s)
Sandy gravel 0.47 (0.87) 0.20 (0.11) 0.007 0.34 (0.84) Unit bar Bedload sheets, dunes 0.019
Sandy gravel 0.84 0.16 0.004 0.67 Bedload sheets, dunes 0.0075
Sandy gravel 0.34 () 0.22 (0.09) 0.004 0.23 () Unit bar Bedload sheets, dunes 0.018
Sandy gravel 0.20 (0.75) 0.20 (0.06) 0.005 0.14 (0.98) Unit bar Bedload sheets, dunes Unsteady
Sand 0.55 0.14 0.004 0.47 Dune Ripples 0
Sand 0.65 0.14 0.003 0.55 Dune Ripples, bedload sheets 0
Sand 0.75 0.14 0.005 0.64 Dune Bedload sheets, ripples 0
Sand 0.85 0.14 0.006 0.73 Dune Bedload sheets 0
Sand-test run 0.35 (0.10) Unit bar Ripples 0.02
Sand-test run 0.20 (0.15) Unit bar Ripples 0.02
Sand 0.19 (0.6) 0.25 (0.08) 0.13 (0.68) Unit bar ripples 0.12
Sand 0.19 (0.6) 0.25 (0.09) 0.13 (0.68) Unit bar Ripples, dunes with ripples 0.08
Numbers not available or not significant.
a
First number is measured in trough of unit bar (number in brackets measured on back of unit bar).
 A.J.H. Reesink, J.S. Bridge / Sedimentary Geology 202 (2007) 281296 289

and grain size generally increases upward within the

cross stratum to medium sand. The true thickness of
grain flow deposits (measured normal to the slope)
ranges from 4.5 to 8 mm, with a mean of 6.2 mm. Grain
flow deposits that are 0.18 m and 0.07 m high have cross
strata with the same thickness (Table 2). The angle of the
cross strata varies from 34 at the top to about 29 at the
bottom. The contact between the cross strata and the
bottom of the Hele-Shaw cell is angular. Grain flows
that reached the bottom of the Hele-Shaw cell were built
up from several smaller flows that did not contain
enough sediment to reach the bottom. The flow me-
chanics were the same for the small and large grain
flows. The relative uniformity of grain flow thickness is
due to the limiting volume (weight) of sediment required
to allow the flow to reach the bottom.

4.2. Flume experiments with dunes in sand

Sand dunes formed in water depth of 0.2 m and

0.15 m, with flow velocity ranging from about 0.55 m/s
to 0.85 m/s (Table 1). The dunes were about 0.05 m high
Fig. 12. (A) Peel of subaqueous grain flows in sand. Peel is 0.3 m wide
and 1 m long, and had superimposed ripples and
and 0.18 m high. (B) Section of the upper part of the peel. Scale on the bedload sheets, with heights of 0.003 m to 0.03 m and
right side is in centimeters. lengths of 0.1 m to 0.2 m (Fig. 3). Ripples and bedload
sheets are differentiated by an arbitrary cutoff height of
that allows epoxy peels to be made. Sediment was fed to 3 mm based on flume observation and existing literature
the Hele-Shaw cell through a 2D funnel at one end of the (Best and Bridge, 1992; Bennett and Bridge, 1995;
box initially. Once the sediment reaches the angle-of- Bridge, 1997; Bridge and Best, 1997). There is a con-
repose, the sediment slope builds to the top of the box by tinuum between bedload sheets and ripples because
repeated grain flows. The slope is then allowed to build bedload sheets on the backs of dunes can amalgamate
out along the length of the box at constant height by and grow into ripples. Ripples were superimposed on
moving the funnel slowly across the top of the box. dunes at relatively low flow velocities (b 0.65 m/s),
Subaqueous grain flows were made by placing the entire whereas primarily bedload sheets were superimposed on
apparatus under water. Grain flow deposits of height less dunes at higher flow velocities.
than the box height was made by partially filling the Both ripples and bedload sheets arriving at the crest of
Hele-Shaw cell with sediment. In order for grain flow a dune affected the shape and migration rate of the dune
deposits to be compared with the cross strata developed lee side, and influenced the nature of deposition there
in the flume, their heights were 0.07 m and 0.18 m. In (Fig. 13). As ripples have a large relative height (Hs/H
addition, observations were made in a Hele-Shaw cell of greater than about 0.30.4), the lee side of the dune is
similar dimensions made out of Plexiglas (after Makse
et al., 1998).
Table 2
4. Experimental results Details of subaqueous cross strata produced in the Hele-Shaw cell,
using different grain flow heights and sediments
4.1. Hele-Shaw cell experiments with grain flows in sand Sediment type Total Average Standard Number of
height thickness deviation of measurable
The cross strata formed in the Hele-Shaw cell by (m) (mm) thickness (mm) strata
subaqueous grain flows are picked out clearly on the Sand 0.18 0.65 0.09 33
peels by variations in grain size (and permeability), Sand 0.12 0.62 0.10 11
particularly in the upper parts of the peels (Fig. 12). A Sand 0.07 0.58 0.08 24
Sandy gravel 0.18 0
veneer of fine sand occurs at the base of a cross stratum,
290 A.J.H. Reesink, J.S. Bridge / Sedimentary Geology 202 (2007) 281296

Fig. 13. Timing of the occurrence of: (1) superimposed bedforms

(ripples of about 0.01 m height) arriving at the brink point of a sand
dune; (2) grain flows that reach the base of the dune lee slope, and;
(3) reduction of the lee side slope of the dune below the angle-of-
repose (note lack of grain flows when slope angles are reduced).

commonly reduced in slope when the trough of the ripple

approaches the dune crest. Bedload sheets do not tend to
cause such slope reduction because of their limited height.
The arrival of the body of the superimposed bedform re-
establishes the angle-of-repose lee slope, and is associated Fig. 14. (A) Peel (0.5 m wide) of deposits of sand dunes with super-
with rapid advance of the dune lee slope and frequent imposed ripples. Note the compound cross-stratification with re-
grain flows (Fig. 13). Most grain flows were associated activation surfaces. (B) Close-up of the same peel: the reactivation
with the passage of the main body of superimposed surfaces are lined with finer material.
bedforms. However, some grain flows were initiated by
turbulent eddies impinging on the dune lee slope, and 4.3. Flume experiments with solitary unit bars in sand
these tend to be thinner than those initiated by
depositional oversteepening. The toes of grain flows Solitary unit bars in sand varied from 0.08 to 0.18 m
were in some cases slowed down and even lifted up into in height (Table 1). Bedforms superimposed on sandy
the flow by the reverse-flow of the separation zone vortex. unit bars were ripples with heights up to 0.02 m and
The increased dilation associated with the introduction of lengths up to 0.20 m (Hs/H of about 0.10.2) In view of
water into the grain flow possibly decreases the the relatively small Hs/H, the arrival of ripples at the
effectiveness of kinetic sieving while increasing the effect crest of the unit bar did not cause reduction of its lee
of differential settling. Grain size sorting associated with slope. Most grain flows are associated with passage of
these effects requires further study.
As the superimposed ripples normally caused reduc-
tion in slope of the dune lee side, the internal structure of
the dunes commonly contained compound cross strata
(Fig. 14). However, angle-of-repose cross strata are
formed on dune lee slopes in association with the arrival
of bedload sheets. Fig. 15A and B clearly show discrete
cross strata on the order of millimeters thick being
formed by migration of successive bedload sheets. The
cross strata are recognized by virtue of changes in grain
size from fine to medium sand, as in the case of the Hele-
Shaw experiments. The cross strata are tangential be-
cause of the effects of grain fall and reworking of
sediment by the reversed flow. The cross strata have a
maximum angle of about 32 at the top and can be traced
into very-low-angle strata and regressive ripples at the
base. The cross strata are distinctive from the top to the
base of the set, in contrast to the cross strata formed in the Fig. 15. (A) Low bedforms, classified as bedload sheets because of their
limited height, (1,2) superimposed on dune. Bedform 1 is forming a
Hele-Shaw cell. The average thickness of the cross strata cross stratum. Note cross strata of variable thickness that is highlighted
is about 8 mm, and there is more variability in thickness by dark, fine-grained layers. (B) Peel (0.5 m wide) of cross strata
than in the Hele-Shaw grain flow experiments. formed in sand dune with superimposed-bedload sheets.
 A.J.H. Reesink, J.S. Bridge / Sedimentary Geology 202 (2007) 281296 291

the unit bar trough, and formation of bottomsets that are

Fig. 16. Timing of arrival of ripples at the brink of a sandy unit bar, and
grain flows down the lee side. Most grain flows are associated with the
passage of the main body of superimposed bedforms, some are
triggered by turbulence associated with their troughs.
the main body of the ripples, and several large grain
flows occur as each ripple passes (Fig. 16). The rela-
tively larger height/flow depth of unit bars compared to
dunes ( 2/3 for unit bars and 1/4 for dunes in these
experiments) results in greater deceleration on lee sides
of unit bars relative to dunes (flow velocity in the trough
of the unit bar 5025% of flow velocity over its top),
enhanced settling of fine-grained suspended sediment in
distinctly finer-grained (very fine sand) than the cross
strata (medium to fine sand) (Fig. 17). Co-flow (e.g.
Fig. 3C) and reverse-flow ripples (e.g. Fig. 18B) are
present in the unit bar trough depending on the flow
characteristics in the trough.
The cross strata have an average thickness of
12.5 mm and can be traced into the strata of the topsets
on the unit bar (Fig. 17B). The cross strata formed by
migration of the unit bar are two or three times thicker
than those formed by grain flows of the same height in
the Hele-Shaw cell. The cross strata are recognized due
to internal grain size variations from medium to fine
sand. The number of cross strata preserved corresponds
to the number of ripples arriving at the bar crest and not
with the number of grain flows that reached the bottom
of the lee slope. In a few examples, cross strata formed
by migration of ripples over unit bars show internal
lamination with the same thickness and grain size
sorting as expected from grain flows (Fig. 7B). Some
cross strata have small reactivation surfaces at the top,

Fig. 17. (A) Sandy unit bar with superimposed ripples. Scale is a meter rule. (B) Peel (0.5 m wide) of sandy unit bar deposits with cross strata formed
by superimposed ripples. Strata delineated by fine-grained drapes that stick out because of capillary effects. Note faint internal laminations caused by
grain flows. (C) Peel (0.5 m wide) of across-flume view of the cross strata associated with peel B. (D) Cut section of a unit bar peel that shows the
grain size variation of peel B. The dimensions of the section are 26 by 60 mm. Resin percolated from the right to the left. The light parts on the left side
are partially impregnated sediment, the right side is fully impregnated sediment. The bottomsets in the lower right are composed of dark, very fine
sand with internal laminations (partially impregnated). The foresets in the middle and upper right show fully impregnated medium to fine sand,
indicating the higher permeability.
292 A.J.H. Reesink, J.S. Bridge / Sedimentary Geology 202 (2007) 281296

Fig. 18. (A) Unit bar in sandy gravel with superimposed-bedload sheets and dunes (B) lee side of unit bar showing down slope increase in grain size
and sharp boundary with rippled fine-grained sand in the trough. Flume is 0.6 m wide. (C) Peel (0.5 m wide) of deposits of unit bar showing cross-
strata of varying thickness and grain size, associated with deposition from superimposed bedforms of varying height.

which allows them to be directly associated with the
passing of a superimposed bedform (Figs. 7B, 17B).
The contact between the cross-set (foreset) and the
bottomset is angular (Fig. 17B), indicating that grain
flow rather than grain fall dominates the deposition on
the lee slope.
The cross strata within the unit bar are laterally con-
tinuous and planar in the across-flume view (Fig. 17C),
even though most of the superimposed ripples have
sinuous crest-lines. This is because ripple crest-lines
were mostly laterally continuous (see Fig. 3C). How-
ever, in some experiments (Table 1), the superimposed
ripples occasionally amalgamated into dunes, which
caused the formation of large reactivation surfaces and
a sinuous crest-line of the unit bar. This resulted in
curved cross strata in across-stream views. The
geometry of the crest-line and lee side of the unit bar
thus controls the shape of strata viewed normal to the
flow direction.
4.4. Flume experiments with solitary unit bars in sandy
gravel
Solitary unit bars in sandy gravel were 0.1 to 0.15 m
high, and water depth over the trough and back were
0.2 m and 0.05 to 0.1 m, respectively. Bedforms super-
imposed on unit bars were dunes and bedload sheets that
were 0.3 to 0.6 m long and up to 0.02 m high (Fig. 18A).
The cross strata are on the order of tens of millimeters
thick (up to 40 mm), and their thickness and grain size
vary considerably (Fig. 18C). Coarsest grains accumulate
at base of the cross-set, and rest sharply on bottomsets
made of fine-grained sand (Fig. 18B). Relatively thick
cross strata formed of coarse-grained open-framework
gravel are due to passage of the highest superimposed
bedforms, whereas thinner cross strata composed of
sandy gravel are due to the passage of lower super-
imposed bedforms. Grain flows in sandy gravels did not
form cross strata, only grain size segregation (Fig. 10).
 A.J.H. Reesink, J.S. Bridge / Sedimentary Geology 202 (2007) 281296
5. Discussion
5.1. Relationship between cross-sectional areas of cross
strata and superimposed bedforms
A simple model for cross-stratal area and thickness
associated with superimposed bedforms, as seen in a
vertical cross-section parallel to flow direction, is:
A Hs Ls =2 H th=sina 1a
th Hs Ls sin a=2H 1b
where A is cross-sectional area of the superimposed
bedform (assuming triangular cross-section) and the
cross stratum, Hs and Ls are the height and length of the
superimposed bedforms respectively, H is the height of
the host bedform, th is thickness of the cross stratum,
and is the slope of the lee side of the host bedform
(residual angle-of-repose). Eqs. (1a), (1b) does not apply
if the Hs/H is greater than about 0.2, because such
relatively large superimposed bedforms would cause
reduction in slope of the lee side of the host bedform.
Eqs. (1a), (1b) also does not apply if the cross-sectional
area of superimposed bedforms is less than that of the
wedge of sediment that fails and becomes a grain flow.
In this case, the presorting in the superimposed bedform
would be lost in the accumulating wedge. If the cross-
sectional area of the superimposed bedform is greater
than that of the failing wedge, as in these experiments, it
may be possible to recognize internal cross strata related
to individual grain flows.
It is important to recognize that the cross-sectional
area of cross strata in preserved sets of cross strata is
much less than that in fully preserved bedforms, as
considered in these flume experiments. According to the
theory of Leclair and Bridge (2001), approximately two-
thirds of the height of a bedform, on average, is removed
by erosion of the upper part. However, in these
experiments, only a small amount of the upper part of
the host bedform and cross-set is removed by super-
imposed bedforms.
Fig. 19 shows that there is close correspondence
between the cross-sectional areas of superimposed
bedforms and cross strata for the case of bedload sheets
on dunes, and bedload sheets, ripples and dunes on unit
bars. The data used (Table 3) include not only these
experiments, but our field data from the South
Saskatchewan River in Canada and the Congaree River
in South Carolina (plus extra data from Levey, 1978).
In the case of fully developed dunes, it is likely that
Ls and H are linearly related to flow depth, and Hs may
293
be related to flow depth and/or grain size. Therefore,
cross-stratal thickness should be related to flow depth
and/or grain size. In the case of bedload sheets or dunes
superimposed on unit bars, it is possible that Hs, Ls and
H are linearly related to depth, such that cross-stratal
thickness would also be related to depth. These ideas
will be tested with data that we are collecting. If a
relationship can be developed between thickness of
cross strata, thickness of cross-sets, the geometry of
superimposed and host bedforms, and hydraulic para-
meters such as flow depth and mean flow velocity, it
would be very useful for reconstructing flow conditions
in ancient deposits.
5.2. Effects of long-term flow unsteadiness
It is commonly thought that long-term changes in
water discharge, associated for example with weather-
related floods or diurnal melting and freezing, can result
in variation in the grain size of adjacent cross strata. The
grain size is normally assumed to increase as discharge
increases. However, preliminary experiments show that
this is a rather simplistic view. For example, if flow
velocity and bed shear stress are increased in time, the
top of a bar may be eroded if the increased sediment
transport rate exceeds the upstream supply. If the finest-
grained sediment occurs at the top of the bar deposit
(coarsest at the base), erosion of the bar top will cause a
decrease in grain size in sediment delivered to the lee
side of the bar (Kleinhans, 2002; Lunt and Bridge,
2007). If erosion proceeds, the sediment delivered to the
Fig. 19. Comparison of cross-sectional areas (along-stream view) of
cross strata and superimposed bedforms for bedload sheets super-
imposed on sand dunes, and ripples, bedload sheets, and dunes
superimposed on sandy and gravelly unit bars. Data sources shown
in Table 3.
294 A.J.H. Reesink, J.S. Bridge / Sedimentary Geology 202 (2007) 281296

Table 3 bars proposed here should lead to much more reliable

Data sources for Fig. 19 predictions of spatial variation in porosity and perme-
Source of data Ripple Strata Host superimposed ability, and thus to characterization of aquifers and hy-
area area bedform bedform type drocarbon reservoirs. Numerical models that predict the
(m2) (m2) type
effect of sedimentary structures on porosity and per-
Experiment 1 0.0050 0.0058 Unit bar Ripples meability (e.g. Kortekaas, 1985; Kelkar and Gupta,
Experiment 2 0.0027 0.0027 Unit bar Ripples
1988; Ringrose et al., 1992; Scheibe and Freyberg,
Experiment 3 0.0013 0.0015 Unit bar Ripples
Experiment 4 0.0020 0.0023 Unit bar Ripples 1995, Scheibe and Murray, 1998) require realistic rep-
Experiment 5 0.0025 0.0025 Dune Bedload sheets resentation of the geometry and grain size sorting of
Sandy gravel 0.0075 0.0062 Unit bar Dunes, different scales of cross strata. This is yet to be attained.
experiments 2005 bedload sheets
Congaree 2006 0.1875 0.1920 Unit bar Dunes
6. Conclusions
Congaree 2004 0.2000 0.2400 Unit bar Dunes
South Saskatchewan 0.0016 0.0014 Unit bar Ripples
2006 Grain size sorting in cross strata formed by dunes and
Levey (1978) 0.2520 0.2400 Unit bar Dunes unit bars is controlled by presorting of sediment
supplied to their lee sides as well as by sorting by
lee side of the bar will become increasingly coarse. grain fall and grain flow on their lee sides. Presorting of
Another, related way of changing the sediment supply to sediment is associated with superimposed bedforms
the front of a bar is for a bank-slump to deliver a large (e.g., bedload sheets, ripples, small dunes) and with
slug of sediment into the river, possibly resulting in a longer-term flow unsteadiness. Variations in the size and
migrating unit bar. The finest grain fractions outrun the grain size sorting of superimposed bedforms result in
coarser fractions, and this fine material can be deposited variations in the thickness and grain size of the cross
on the lee side of previously existing unit bars before the strata. The thickest and coarsest cross strata are
coarser sediment arrived. Therefore, a fine-coarse associated with the largest superimposed bedforms.
couplet in unit bar cross strata does not necessarily Long-term flow unsteadiness is related to factors such as
mean changing flow stage. cyclic floods, associated with tides and melting-freezing
cycles, erosion and deposition of upstream bars, and
5.3. Further research and its implications upstream bank-slumps, each of which can give rise to
alternations of coarse- and fine-grained cross strata.
The full implications of linking the formation of cross Detailed experimental studies are required to fully
strata with superimposed bedforms and longer-term flow explore the relationship between presorting of sediment
unsteadiness will only be realized after more detailed arriving at the lee sides of dunes and unit bars, and
experimental studies. It is first necessary to undertake variations in the geometry and grain size sorting of their
comprehensive studies of the 3D geometry and grain size cross strata. Such studies will allow much more refined
sorting of superimposed bedforms, the relationship quantitative interpretation of ancient flow and sediment
between the passage of superimposed bedforms and transport conditions (e.g., flow depth and velocity) from
the nature of grain flow on the lee side of the host cross strata, and more accurate prediction of spatial
bedform, and how these relate to the 3D geometry and variation in porosity and permeability, a major concern
grain size sorting of cross strata. The next step is relating to hydrogeologists and petroleum geologists.
the geometry, migration rate, and grain size sorting of
both the host bedforms and the superimposed bedforms
References
to flow and sediment transport conditions. Then the
geometry and grain size sorting of cross strata can also be
related to flow and sediment transport conditions. Allen, J.R.L., 1963. Sedimentation to the lee of small underwater
sandwaves: an experimental study. Journal of Geology 75,
However, the preservation ratio (host bedform height/ 95116.
cross-set thickness) of cross strata must be known in Allen, J.R.L., 1968. The accumulation of sediment in the lee of ripples
order to do this (Leclair and Bridge, 2001). and dunes, and the thickness of bottomset deposits. Geological
Textural variations in cross strata can be related to Magazine 105, 166176.
spatial variations in porosity and permeability (reviews Allen, J.R.L., 1970. The avalanching of granular solids on dune and
similar slopes. Journal of Geology 78, 326351.
in Brayshaw et al., 1996; Tidwell and Wilson, 2000; Allen, J.R.L., 1982. Sedimentary Structures, their Character and
Bridge, 2003). The improved descriptions and inter- Physical Basis, vol. 2. Developments in Sedimentology 30 b, 2.
pretations of cross strata associated with dunes and unit Elsevier Scientific Publishing Company, Amsterdam.
 A.J.H. Reesink, J.S. Bridge / Sedimentary Geology 202 (2007) 281296
Bagnold, R.A., 1954. Experiments on the gravity-free dispersion of
large solid spheres in a Newtonian fluid under shear. Proceedings
of the Royal Society of London 187A, 4963.
Bennett, S.J., Bridge, J.S., 1995. The geometry and dynamics of low
relief bed forms in heterogeneous sediment in a laboratory channel,
and their relationship to water and sediment transport. Journal of
Sedimentary Research A65, 2939.
Best, J.L., Bridge, J.S., 1992. The morphology and dynamics of low-
amplitude bedwaves upon upper stage plane beds and the
preservation of planar laminae. Sedimentology 39, 737752.
Best, J.L., Ashworth, P.J., Bristow, C.S., Roden, J.E., 2003. Three
dimensional sedimentary architecture of a large, mid-channel sand
braid bar, Jamuna River, Bangladesh. Journal of Sedimentary
Research 73 (4), 516530.
Boersma, J.R., Van de Meene, E.A., Tjalsma, R.C., 1968. Intricated
cross stratification due to interaction of a mega ripple with its
leeside of backflow ripples (upper point bar deposits, Lower
Rhine). Sedimentology 11, 147162.
Brayshaw, A.C., Davies, G.W., Corbett, P.W.M., 1996. Depositional
controls on primary permeability and porosity at the bedforms scale
in fluvial reservoir sandstones. In: Carling, P.A., Dawson, M.R.
(Eds.), Advances in Fluvial Dynamics and Stratigraphy. John Wiley
& Sons Ltd, pp. 373394.
Bridge, J.S., 1997. Thickness of sets of cross strata and planar strata
as a function of formative bedwave geometry and migration, and
aggradation rate. Geology 25, 971974.
Bridge, J.S., 2003. Rivers and Floodplains; Forms, Processes and
Sedimentary Record. Blackwell Publishing, Oxford. 491 pp.
Bridge, J.S., Best, J.L., 1997. Preservation of planar laminae due to the
migration of low relief bed waves over aggrading upper stage plane
beds: comparison of experimental data with theory. Sedimentology
44, 253262.
Bridge, J.S., Smith, N.D., Trent, F., Gabel, S.L., Bernstein, P., 1986.
Sedimentology and morphology of a low sinuosity river: Calamus
River, Nebraska Sand Hills. Sedimentology 33, 851870.
Carling, P.A., Glaister, M.S., 1987. Rapid deposition of sand and gravel
mixtures downstream of a negative step: the role of matrix-infilling
and particle-overpassing in the process of bar front accretion.
Journal of the Geological Society of London 144, 543551.
Coleman, S.E., Melville, B.W., 1994. Bed form development. Journal
of Hydraulic Engineering 120, 544560.
Coleman, S.E., Melville, B.W., 1996. Initiation of bedforms on a flat
sand bed. Journal of Hydraulic Engineering 122, 301310.
Ditchfield, R., Best, J.L., 1992. Development of bed features;
discussion. Journal of Hydraulic Engineering 118, 647650.
Gabel, S.L., 1993. Geometry and kinematics of dunes during steady
and unsteady flows in the Calamus River, Nebraska, USA.
Sedimentology 40, 237269.
Htu, B., Van Steijn, H., Bertran, P., 1995. Le role des coulees de
pierres seches dans la genese d'un certain type d'eboulis stratifies/
the role of dry grain flow in the genesis of a type of stratified scree.
Permafrost and Periglacial Processes 6 (2), 173194.
Hooke, R.L., 1968. Laboratory study of the influence of granules on
flow over a sand bed. Geological Society of America Bulletin 79,
495500.
Hunter, R.E., 1985a. A kinematic model for the structure of lee side
deposits. Sedimentology 32, 409422.
Hunter, R.E., 1985b. Subaqueous sand flow cross strata. Journal of
Sedimentary Petrology 55, 886894.
Hunter, R.E., Kocurek, G., 1986. An experimental study of
subaqueous slipface deposition. Journal of Sedimentary Petrology
56, 387394.
295
Jackson, R.G., 1976. Large scale ripples of the Lower Wabash River.
Sedimentology 23, 593623.
Jopling, A.V., 1965. Laboratory study of the distribution of grain sizes
in cross bedded deposits. In: Middleton, G.V. (Ed.), Primary
Sedimentary Structures and Their Hydrodynamic Interpretation.
SEPM Special Publication, vol. 12. SEPM, pp. 5365.
Kelkar, B.G., Gupta, S.P., 1988. The effects of small scale heterogeneities
on the effective dispersivity of porous medium. SPE/DOE 17339,
215226.
Kleinhans, M.G., 2002. Sorting Out Sand and Gravel; Sediment Transport
and Deposition in Sand-Gravelbed Rivers. Utrecht University, Utrecht.
293 pp.
Kleinhans, M.G., 2004. Sorting in grain flows at the lee side of dunes.
Earth Science Reviews 65, 75102.
Kleinhans, M.G., 2005. Grain size sorting in grainflows at the lee side
of deltas. Sedimentology 52, 291311.
Koeppe, J.P., Enz, M., Kakalios, J., 1997. Phase diagram for avalanche
stratification of granular media. Physical Review E 58 (4),
R4104R4107.
Kortekaas, T.F.M., 1985. Water/oil displacement characteristics in
cross bedded reservoir zones. Society of Petroleum Engineers
Journal 25, 917926.
Leclair, S.F., 2002. Preservation of cross strata due to migration of
subaqueous dunes: an experimental investigation. Sedimentology
49, 11571180.
Leclair, S.F., Bridge, J.S., 2001. Quantative interpretation of sedimentary
structures formed by river dunes. Journal of Sedimentary Research
71, 713716.
Legros, F., 2002. Can dispersive pressure cause inverse grading in
grain flows? Journal of Sedimentary Research 72, 166170.
Levey, R.A., 1978. Bed form distribution and internal stratification of
coarse grained point bars upper Congaree River, S.C. In: Miall, A.D.
(Ed.), Fluvial Sedimentology. Canadian Society of Petroleum
Geologists, Calgary, pp. 105127.
Lunt, I.A., Bridge, J.S., 2007. Formation and preservation of open-
framework gravel strata in unindirectional water flows. Sedimen-
tology 54, 7187.
Lunt, I.A., Bridge, J.S., Tye, R.S., 2004. A quantative, three dimensional
depositional model of gravelly braided rivers. Sedimentology 51,
377414.
Makse, H.A., 1997. Stratification instability in granular flows. Physical
Review E 56 (6), 70087018.
Makse, H.A., Cizeau, P., Stanley, H.E., 1997a. Possible stratification
mechanism in granular mixtures. Physical Review Letters 78 (17),
32983301.
Makse, H.A., Havlin, S., King, P.R., Stanley, H.E., 1997b. Spontaneous
stratification in granular mixtures. Nature 386, 379382.
Makse, H.A., Ball, R.C., Stanley, H.E., Warr, S., 1998. Dynamics of
granular stratification. Physical Review E 58 (3), 33573367.
Mantz, P.A., 1978. Bedfroms produced by fine cohesionless granular
and flakey sediments under subcritical water flows. Sedimentology
25, 83103.
Middleton, G.V., Southard, J.B., 1984. Mechanics of Sediment Move-
ment. Lecture Notes for Short Course, vol. 3. SEPM, Oklahoma.
Nio, S.D., Siegenthaler, C., Yang, C.S., 1983. Megaripple cross
bedding as a tool for the reconstruction of the paleo-hydraulics in a
Holocene subtidal environment, S.W. Netherlands. Geologie en
Mijnbouw 62 (2), 499510.
Raudkivi, A.J., 1997. Ripples on stream bed. Journal of Hydraulic
Engineering 116, 5864.
Raudkivi, A.J., Witte, H.-H., 1990. Development of bed features.
Journal of Hydraulic Engineering 116, 10631079.
296 A.J.H. Reesink, J.S. Bridge / Sedimentary Geology 202 (2007) 281296
Ringrose, P.S., Sorbie, K.S., Corbett, P.W.M., Jensen, J.L., 1992.
Immiscible flow behaviour in laminated and cross bedded sand-
stones. Journal of Petroleum Science and Engineering 9, 103124.
Rust, B.R., 1984. Proximal braidplain deposits in the Middle Devonian
Malbaie Formation of Eastern Gaspe, Quebec, Canada. Sedimen-
tology 31, 675696.
Sallenger, A., 1979. Inverse grading and hydraulic equivalence in grain
flow deposits. Journal of Sedimentary Petrology 49, 553562.
Scheibe, T.D., Freyberg, D.L., 1995. Use of sedimentological infor-
mation for geometric simulation of natural porous media structure.
Water Resources Research 31 (12), 32593270.
Scheibe, T.D., Murray, C.J., 1998. Simulation of geological patterns: a
comparison of stochastic simulation techniques for groundwater
transport modeling. In: Fraser, G.S., Davis, J.M. (Eds.), Hydro-
geologic Models of Sedimentary Aquifers. SEPM Concepts in
Hydrogeology and Environmental Geology, vol. 1. SEPM, Tulsa,
OK.
Smith, N.D., 1972. Some sedimentological aspects of planar cross-
stratification in a sandy braided river. Journal of Sedimentary
Petrology 42, 624634.
Smith, N.D., 1974. Sedimentology and bar formation in the upper
Kicking Horse River: a braided outwash stream. Journal of
Geology 82, 205223.
Smith, N.D., 1985. Proglacial fluvial environment. In: Ashley, G.M.,
Shaw, J., Smith, N.D. (Eds.), Glacial Sedimentary Environments.
SEPM Short Course, vol. 16. SEPM.
Steel, R.J., Thompson, D.B., 1983. Structures and textures in Triassic
braided stream conglomerates (Bunter pebble beds) in the
Sherwood Sandstone Group, North Staffordshire, England. Sedi-
mentology 30, 341367.
Tidwell, V.C., Wilson, J.L., 2000. Heterogeneity, permeability patterns
and permeability upscaling: physical characterization of a block of
Massillon sandstone exhibiting nested scales of heterogeneity. SPE
Reservoir Evaluation and Engineering 3 (4), 283291.
Venditti, J.G., Church, M., Bennett, S.J., 2005. Morphodynamics of
small-scale superimposed sandwaves over migrating dune bed-
forms. Journal of Geophysical Research 110, F01009.
Visser, M.J., 1980. Neap-spring cycles reflected in Holocene subtidal
large-scale bedform deposits: a preliminary note. Geologie en
Mijnbouw 8, 543546.
 Available online at www.sciencedirect.com

Sedimentary Geology 202 (2007) 297 316

www.elsevier.com/locate/sedgeo

Grain size and sediment concentration effects on

channel patterns of silt-laden rivers
D.S. van Maren
Delft University of Technology, Faculty of Civil Engineering and Geosciences, section of Hydraulic Engineering,
P.O. Box 5048, 2600 GA Delft, The Netherlands
WL | Delft Hydraulics, P.O. Box 177, 2600 MH Delft, The Netherlands

Abstract

The Yellow River in China is characterized by channel erosion and a meandering channel pattern at both low sediment
concentrations and at hyperconcentrations, whereas deposition and braiding tends to dominate at intermediate sediment
concentrations. A 3D numerical model (Delft3D) is used to analyze the effect of sediment-induced density effects on the formation
of channel patterns using a highly simplified river geometry. Results show that the effect of vertical sediment density stratification
and hindered settling significantly influence river channel patterns. At some critical concentration (the saturation concentration),
vertical mixing is totally damped by the density gradients at the lutocline (a strong vertical gradient in sediment concentration).
This increases the secondary circulation in the river channel significantly, which in turn promotes meander development. When the
flow is not close to saturation, the main channel-forming process seems to be channel avulsion through blocking of old channels
and subsequent channel cut-off. When the concentrations are so high that sediment is primarily held in suspension by hindered
settling, sedimentation is so much reduced that channels are no longer blocked and therefore the river course becomes confined to a
single meandering channel.
2007 Elsevier B.V. All rights reserved.

Keywords: Yellow River; Hyperconcentrated flow; Sediment transport; River morphology; Braiding; Meandering; Morphological modelling; Delft3D

1. Introduction land. The river dikes have to be raised constantly to

The rivers that originate from the Chinese Loess
Plateau carry large amounts of suspended sediment (up
to 1600 kg m 3 ; Wan and Wang (1994)), of which a
large part is deposited in the downstream reaches,
resulting in a rapid bed level rise. The best known
example is the Yellow River, where the diked section of
the floodplain is up to 10 m higher than the adjacent
Delft University of Technology, Faculty of Civil Engineering and
Geosciences, section of Hydraulic Engineering, P.O. Box 5048, 2600 GA
Delft, The Netherlands. Tel.: +31 15 2785438; fax: +31 30 2785124.
E-mail address: d.s.vanmaren@citg.tudelft.nl.
0037-0738/$ - see front matter 2007 Elsevier B.V. All rights reserved.
doi:10.1016/j.sedgeo.2007.04.001
prevent flooding, but it is obvious that this cannot
continue indefinitely and more sustainable solutions
have to be found. One of these solutions is to reduce
siltation by strategic sediment flushing from upstream
reservoirs. Reservoirs in China's silt-laden rivers allow
flushing of sediment as well as water, and increasing
attention is paid to the optimal sediment concentration
of water released from the reservoirs. Of particular
interest is the modification of the sediment transport
capacity of the river by its sediment load, which is at a
maximum at either very high, or low sediment
concentrations (e.g. Wan and Wang, 1994; Chien and
Wan, 1998; Xu, 2004). The river channel deepens
298 D.S. van Maren / Sedimentary Geology 202 (2007) 297316

during high transport capacity conditions (low or hyper-
concentrations exceeding 100 kg m 3), whereas sedi-
ment is deposited at intermediate concentrations (see
Xu, 2004). River channel deepening during hypercon-
centrated flow is especially significant. Locally erosion
may be up to 9 m during the passage of a hyperconcen-
trated flood (Wan and Wang, 1994). A numeric model
that can predict channel adjustment to sediment flushing
is essential for efficient management of these upstream
dams. Therefore a research project was initiated within
the KNAW-MOST Programme of Scientific Alliances
between China and The Netherlands, which aims to
model sediment transport and morphology in the Lower
Yellow River (de Vriend et al., 2005).
The lower Yellow River (LYR) is characterized by
braiding and meandering sections (Fig. 1). A morpho-
dynamic model should therefore be able to reproduce
the tendency of a river section to either meander or
braid. Under which conditions rivers meander or braid is
well known from natural and laboratory data (Leopold
and Wolman, 1957; Schumm, 1977; Ferguson, 1987;
Van den Berg, 1995) and from theoretical models
(Engelund and Skovgaard, 1973; Parker, 1976; Fredse,
1978; Blondeaux and Seminara, 1985). There is general
consensus that channel geometry is mainly controlled by
discharge, sediment diameter and valley slope (Bridge,
1993): meandering predominates in fine grained, low-
slope constant-discharge rivers whereas braiding is
favoured by a steep valley slope, a coarse grain size
and perennial flow. However, that meandering is more
common in fine-grained sediments is related to the low
gradient typical for meandering rivers and not to the
grain size itself. Even more, the critical river slope for
braiding is higher for gravel bed rivers than for fine-

Fig. 1. Yellow river basin (a) and the lower Yellow River (b) showing the braided and meandering section.
 D.S. van Maren / Sedimentary Geology 202 (2007) 297316 299

grained rivers (Ferguson, 1987). The tendency to braid
may also decrease with increasing bank erodibility by
affecting the width/depth ratio and secondary flow
(Ferguson, 1987). However, the channel patterns in silt-
laden rivers on the Chinese Loess Plateau probably also
depend on the sediment concentration (Xu, 2002):
meandering dominates at low and high sediment
concentrations, whereas braiding prevails at intermedi-
ate concentrations. Xu related this transitional behaviour
to the occurrence of hyperconcentrated floods, though
the physical mechanisms regulating this transition are
not yet known. Previous work on stratification effects in
highly concentrated sediment-laden flow by Winterwerp
(2001) and Winterwerp (2006) revealed the existence of
three flow regimes controlled by the sediment concen-
tration: subsaturated low concentration flow, supersat-
urated flow, and subsaturated hyperconcentrated flow.
Concentrations over a few g/l may already invoke
vertical density gradients that suppress turbulence to
such a degree that suspended sediment rapidly settles
from suspension (supersaturated flow), whereas the
water column becomes fully mixed again at even higher
concentrations due to hindered settling effects (subsat-
urated hyperconcentrated flow). These flow regimes
may be related to the previously mentioned concentra-
tion-dependence of channel patterns in silt-laden rivers.
The aim of this paper is to determine and numerically
reproduce the influence of sediment-induced density
effects on river morphology.
First, sediment transport patterns in the Yellow River
are introduced (Section 2), followed by an analysis of
sediment-induced density stratification (Section 3). The
relation between sediment density stratification and
morphology is then explored with a 3D morphodynamic
model, which is introduced in Section 4, and results are
presented in Section 5. The discussion follows in
Section 6, and conclusions are drawn in Section 7.
2. The Yellow River
The Yellow River has a length of 5464 km and a basin
area of 795.000 km2. From 1940 to 1969, the average
discharge in the lower Yellow River was 1630 m3 s 1 and
1560 million tons of sediment were transported annually
(Wang and Xu, 1999). Since 1969 the discharge and
sediment load have decreased significantly because of
increased human water demand. The upper part of the
lower Yellow River has a braiding channel pattern (Fig. 1)
with a slope of 22.5 10 4 and a main channel width
between 1 and 3.5 km. The downstream part of the lower
Yellow River is meandering with a slope of 1 10 4 and a
channel width between 300 and 1000 m (Long et al., 2002).

Fig. 2. Bed level change z (in m) during a flood event in the Yellow River upstream of Sanmenxia, in relation to the maximum sediment
concentration C observed during that flood (a; data modified from Xu (2004), with 2nd order polynomial fit), erosion and sedimentation per flood
event in the Huangpuchuan River as a function of discharge Q and sediment concentration C (b; modified from Xu (2004)), and channel sinuosity
(defined as the ratio of the channel length to the valley length) in a number of Chinese rivers in relation to the mean annual sediment concentration Ca
(c; modified from Xu (2002).
300 D.S. van Maren / Sedimentary Geology 202 (2007) 297316
The average suspended sediment concentration in the
Yellow River is 30 kg m 3, but during floods the
concentration can exceed 100 kg m 3. Although these
high-concentration floods are often referred to as
hyperconcentrated by various authors, there is no clear
definition of hyperconcentrated flow. Elaborating the
effect of sediment on the vertical structure of the water
column through the Richardson number, van Maren et al.
(2006) defined hyperconcentrated flow (HCF) as a
sediment-laden flow in which the vertically homogeneous
concentration profile is strongly affected by hindered
settling, and can be maintained at low turbulence rates
when the sediment concentration exceeds a critical
sediment concentration. This critical sediment concentra-
tion increases with the water depth, the grain size and the
space-filling concentration, and decreases with the cubed
flow velocity. HCF is common in the tributaries of the
Yellow River but less frequent in the lower reaches: due to
hydroelectric dams they are now mainly artificially
created during reservoir flushing.
Although the sediment transport processes at very
high concentrations are still poorly understood, it seems
that the transport capacity of the flow is so high that
sediment concentration and river bed erosion is restricted
by the scour rate of the bed (Wang, 1999), rather than by
the spatial gradients in the sediment transport. The
transition from a normal flow to HCF is associated with a
transition from deposition to scouring in the main
channel (Wang et al., 1997; Wang, 1999; Wang et al.,
2002; Xu, 2004), and therefore plays an important role in
channel morphology. Between Aishan and Lijin, for
instance, scouring begins when Q N 1800 m3 s 1 whereas
sediment is deposited when Q b 1800 m3 s 1 (Wang and
Liang, 2000). The transition occurs during higher
discharges according to Pu (1997): Q = 3000 m3 s 1 in
the meandering reach (corresponding to a flow velocity
u of 1.9 m s 1) and Q = 4000 m3 s 1 in the braided
reach. Xu (2004) distinguished three erosion-sedimen-
tation regimes: erosion occurs at high (N 200 kg m 3) and
low (b 80 kg m 3) concentrations whereas sedimentation
rates are at a maximum at intermediate concentrations (C
between 80 and 200 kg m 3). Fig. 2a suggests that
erosion/sedimentation correlates with the sediment
concentration, especially at higher concentrations.
However, this relation is obscured by the temporal
variation in grain size transported in the Yellow River. A
clearer example is therefore the erosion/sedimentation
pattern in a tributary of the Yellow River with a more
uniform sediment type, such as the Huangpuchuan River
(Fig. 2b). In this river, sediment is deposited between
approximately 100 to 600 kg m 3 whereas the channel is
eroded at lower and higher concentrations. This
alternation of erosion and deposition significantly affects
river channel patterns because predominant scouring
decreases the width-to-depth ratio and increases mean-
dering, whereas deposition increases the width-to-depth
ratio and promotes braiding, in the Yellow River (Xu,
2004), but also elsewhere (Germanoski and Schumm,
1993). Therefore the channel sinuosity of silt-laden
rivers on the Chinese Loess plateau is highest at
intermediate concentrations (Fig. 2c). It therefore
seems that the rivers on the Chinese Loess Plateau
have three different sediment transport regimes: at both
low and very high concentrations the sediment carrying
capacity is high, resulting in predominantly erosion and
meandering, and at intermediate concentrations deposi-
tion and braiding prevails. Understanding the morpho-
logic behaviour of the Lower Yellow River therefore
requires a detailed analysis of the relation between
sediment concentration and morphology.
3. Sediment density effects in silt-laden flow
The high sediment concentrations typical for floods
in the Yellow River (several 100 g/l) increase the density
of the watersediment mixture significantly. Because
concentrations of several g/l are already sufficient to
significantly dampen turbulence (Winterwerp, 2001),
the turbulence structure in the Yellow River can be
expected to be heavily influenced by its suspended
sediment load. This has two major consequences:
Sediment is held in suspension by turbulence.
Vertical sediment-induced density gradients increase
with the sediment concentration, thereby effectively
damping turbulence, and therefore reducing the
capacity of the flow to keep particles in suspension.
This, in turn, increases the downward sediment flux,
thereby further increasing the vertical density
gradients. When the flow is close to saturation, this
positive feedback mechanism may result in a total
damping of turbulence, and therefore sediment can
no longer be held in suspension, leading to a total
collapse of the vertical profile of turbulence and
sediment concentration (see Winterwerp, 2001;
Winterwerp, 2006).
Suppression of turbulence leads to a decreasing
vertical velocity gradient near-bottom. The velocity
profile consequently changes from a logarithmic into
a parabolic-like profile, resulting in decreasing
hydraulic resistance and therefore higher depth-
averaged flow velocities. However, at even higher
concentrations, the yield stress of the fluid sediment-
mixture may exceed the velocity shear stress. In that
 D.S. van Maren / Sedimentary Geology 202 (2007) 297316 301

case plug flow may develop, characterized by a high
velocity gradient near-bottom.
In addition to these turbulence-related effects, high
sediment concentrations also affect sediment transport
by:
An increased viscosity, and at very high concentra-
tion, also yield stress. The flow in some tributaries of
the Yellow River may become non-Newtonian
during floods, characterized by laminar flow (high
viscosity) or plug flow (cross-sectionally and verti-
cally uniform flow velocities due to high yield
stresses) (Wan and Wang, 1994).
Hindered erosion and hindered settling. Sediment
settling rates decrease with increasing sediment
concentrations due a number of processes, including
the particle-induced return flow and wake formation,

Fig. 3. Cross-sectional distribution of flow velocity (a), sediment concentration (b) and median grain size (c) during a flood event in the Yellow River
near Huayuankou in 1959, with vertical profiles in the channel (marked with an arrow) to the right of each cross-section (d, e, and f) (Yellow River
Conservancy Comission, 1959).
302 D.S. van Maren / Sedimentary Geology 202 (2007) 297316
Fig. 4. Relation between the stream power u3/ghw0 and the sediment
concentration C under equilibrium conditions (i.e. no erosion or
sedimentation) obtained from streams on the Loess Plateau and the
main stem of the Yellow River, with a grain size D50 between 20 and
80 m (Xu, 1999). The solid line is a least-squares fit of Eq. (1).
particleparticle collision and interaction, an increas-
ing viscosity, and reduced gravity (see Winterwerp
and van Kesteren (2004) for an overview). Similarly,
erosion of sediment from the bed will be reduced at
increasing sediment concentrations because there is
less water available to displace sediment particles
from the bed.
Soil mechanics. Sediment deposition rates may be so
high during some HCF that water cannot easily
escape from the depositing bed, making it susceptible
to liquefaction. Additionally, high flow velocities
(velocities during hyperconcentrated floods may
exceed 5 m/s), inhibit water from penetrating the
bed and displace single particles. Undrained erosion
may then occur, during which blocks of sediment are
eroded rather than single particles (Winterwerp and
van Kesteren, 2004).
The concentration of sediment transported in sus-
pension during floods in the Yellow River is not
vertically uniform, at least not during hyperconcentrated
conditions. This is illustrated by the cross-sectional
distribution of sediment concentration and grain size in
Fig. 3 measured during a moderate flood, showing
vertical concentration gradients exceeding 10 kg/m3/m.
Consequently, the sediment-induced density gradients
during a moderate flood are as large as those encoun-
tered in highly stratified estuarine environments, and
possibly much higher during large floods when the
sediment concentration exceeds several 100 g/l. How-
ever, accurate data with vertical distribution of concen-
tration and velocity during such events are very scarce.
Some further information can be obtained from depth-
averaged data relating the flow strength (u3/ghw0) to
depth-average sediment concentration (Fig. 4). At
C b 100 kg m 3 the sediment concentration increases
with flow strength u3/ghw0, but at higher concentra-
tions the flow strength required to hold sediment in
suspension decreases again. This may be attributed to
hindered settling effects: at high sediment concentra-
tions less energy is required to keep sediment in
suspension because the sediment settling velocity is
reduced by hindered settling. The role of hindered
settling can be estimated by fitting a formula for the
sediment transport capacity to the data in Fig. 4. For that
purpose the saturation concentration Cs, based on the
flux Richardson number, is used (Winterwerp, 2001):
qm qs u3T
Cs K1 1
gqs  qw hw0 1  /5
where K1 is a coefficient, m is the density of the water
sediment mixture, s the sediment density, w the water
density, g is the gravitational acceleration, u is the bed
shear velocity, h is the water depth, w0 is the particle fall
velocity in clear water and is the volumetric sediment
concentration given by = C/Cg; C is the concentration
in kg m 3 and Cg is the concentration at which a space-
filling network forms. The most pronounced feature of
the figure, the low flow strength needed to keep
sediment in suspension at high concentrations, is related
to the hindered settling part (1 )5 of Eq. (1). Hindered
settling is consequently also expected to play a major
role in river morphology of silt-laden rivers. Therefore
both the effects of hindered settling and sediment
density stratification on river morphology will be
analyzed using a 3D numerical model.
4. 3D modelling
4.1. Processes
Theoretical studies on channel patterns have mainly
been done using analytical models. Linear stability
analyses revealed that the channel pattern is primarily
controlled by the formation of alternate bars which in
turn is determined by the channel width-to-depth ratio
(Engelund and Skovgaard, 1973; Fredse, 1978) and
channel slope (Parker, 1976). Instability of alternate bars
in straight channels was considered to be the driving
 D.S. van Maren / Sedimentary Geology 202 (2007) 297316
mechanism behind the initial formation of meanders
until the work of Ikeda et al. (1981), Struiksma et al.
(1985), and Blondeaux and Seminara (1985), who
extended the alternate bar theory with resonance and
overshoot effects. Non-linear stability analyses (e.g.
Tubino and Seminara, 1990; Seminara and Tubino,
1992; Tubino et al., 1999) focussed on the behaviour of
these alternate bars, whereas the effect of bar formation
by cross-sectional width variations was analyzed by
Repetto et al. (2002).
Numerical models may be applied to simulate channel
pattern evolution as long as the bed forms that determine
channel pattern formation are larger than the model grid.
In order to simulate meander formation, the model should
be able to reproduce alternate bars which have a
magnitude in the order of the channel width. Initially
straight channels evolve into braiding channels by (1)
deposition and accumulation of a central bar, (2) chute
cutoff of point bars, (3), conversion of a transverse bar into
braid bars, and (4) multiple bar braiding (Ashmore, 1991).
The spatial scales associated with these processes are
several times smaller than the channel width. A numerical
model with 1020 grid cells in the cross-sectional
direction should reproduce these processes as long as
the hydrodynamics are accurately solved. Recent numer-
ical modelling studies include the simulation of the initial
formation and evolution of meanders (Olsen, 2003;
Ruether and Olsen, 2003), braided channels (Enggrob
and Tjerry, 1999) and estuarine channel-shoal patterns
(Hibma et al., 2003, 2004; Marciano et al., 2005).
In this study, the state-of-the-art sediment online
version of the Delft3D model is used to simulate
planform morphology development. Delft3D solves the
unsteady, shallow water equations in 2 or 3 dimensions
under the hydrostatic pressure assumption, and com-
putes sediment transport and updates morphology
simultaneously with the flow (see Lesser et al. (2004)
for a description and validation). An older version of this
model has already been used to simulate channel
avulsion processes (Jagers, 1999) and also estuarine
channel-shoal patterns (Hibma et al., 2003, 2004;
Marciano et al., 2005). In addition, most of the
sediment-induced density effects described in Section
3 have been implemented in the Delft3D model
(Winterwerp and van Kessel, 2003). Simulation of
these sediment density effects requires 3D simulations,
and therefore the model is applied in 3D mode.
4.2. Model set-up
The Yellow River is highly simplified with a length
of 50 km and a uniform slope i of 2.2 10 4 in the
303
downstream direction, with a parabolically shaped
channel cross-section several km wide (see Fig. 5).
This is the typical gradient and channel width of the
braided section of the lower Yellow River (Fig. 1). A
spatially random perturbation with a maximum ampli-
tude of 5 cm is used to invoke initial sediment transport
gradients. The grid cells are 50 by 125 m in the cross-
channel and along-channel direction, respectively. Ten
vertical layers are used, with a thickness increasing
logarithmically from bed to surface. A discharge Q of
5000 m3 s1 is prescribed at the upstream boundary
(approximately the yearly annual flood in the Yellow
River), and an equilibrium water level computed from i
and Q is prescribed at the downstream boundary. A
constant Chezy value of 60 m1/2/s is used for the bed
roughness. River bank erosion is parameterized by dry
cell erosion (Roelvink et al., 2003). Active (wet) cells
are defined as having a water depth exceeding 0.1 m and
are used for hydrodynamic computations; cells with a
lower water depth are inactive (dry) and not active in
hydrodynamic computations. However, the erosion in
active (wet) cells adjacent to dry cells can be transferred
to these dry cells. In this way, not the wet cell but the dry
cell (representing the river bank) is eroded. As a result of
this erosion the water depth in the dry cell increases in
time until the water depth exceeds 0.1 m and becomes
an active cell, resulting in a migration of the channel.
Yellow River sediment is mainly fine to coarse silt
and transported in suspension as either bed material or
washload. Washload is very fine sediment which is
advected from upstream and almost uniformly distrib-
uted over the water column at a concentration depending
mainly on supply and not on the flow velocity. The
washload fraction, mainly composed of very fine silt, is
eroded from the Chinese Loess Plateau and mostly
flushed seaward (Zhao et al., 1989). Bed material load is
coarser than washload (mostly coarse silt and fine sand),
it is in local equilibrium with the bed, and therefore the
concentration depends on the local flow velocity. The
Fig. 5. Cross-channel initial bed level and water depth of the idealized
Yellow River section, consisting of a wide and shallow channel.
304 D.S. van Maren / Sedimentary Geology 202 (2007) 297316
median grain size D50 of channel bed sediment is
approximately 65 m whereas the D50 of suspended
load is approximately 40 m during normal conditions,
but much finer during some flood conditions. The yearly
averaged D50 of the suspended load is 19 m (Fig. 6).
Therefore two sediment fractions are used: one coarse
fraction of 65 m, representing the bed material load,
and one fine sediment fraction representing the wash-
load. The coarse sediment fraction (65 m) is in local
equilibrium with the bed through erosion and deposi-
tion, and computed with the Van Rijn formulations for
bedload and suspended load transport (van Rijn, 2000).
The upstream sediment concentration of this fraction is
in equilibrium with flow velocities at the upstream
boundary.
The washload fraction, on the other hand, does not
interact with the bed (i.e. no erosion or deposition, all
sediment transported as washload is transported down-
stream). Both the upstream prescribed sediment con-
centration C and median grain size D50 of this washload
fraction are varied. Numerical experiments are done
with two washload grain sizes because different grain
sizes result in different vertical density gradients. The
effect of hindered settling and turbulence damping on
morphologic evolution are difficult to separate because
they occur simultaneously. However, for very fine
sediment the vertical density gradients are low and their
effect on the vertical flow velocity profile is only
marginal. Therefore very fine washload simulations
(D50 = 4 m) are used to analyze the effect of hindered
settling, and coarse washload (D50 = 40 m) simulations
Fig. 6. Cumulative fraction f of the diameter D of suspended load and
bed material in the braided reach of the Yellow River near Huayuankou
from 1957 to 1959 (Yellow River Conservancy Comission, 1959).
for the combined effect of hindered settling and
turbulence damping. Both are done for a range of
washload sediment concentrations, results of which will
be shown here for 0, 100, 200, and 400 kg m 3. The
space-filling concentration Cg of the washload is based
on the data shown in Fig. 4: Eq. (1) fits the data best
when Cg = 700 kg m 3.
Idealized morphological modelling may provide
insights into certain mechanisms that regulate the
morphologic behaviour, but the necessary simplifica-
tions also have some drawbacks. In natural river
systems, sediment carried in suspension and in bed
load is graded. Coarse sediment prevails in the main
channel, where flow velocities are generally higher, and
fine sediment occurs where flow velocities are lower.
This segregation reduces sediment transport in the
deeper parts relative to the shallower parts. In single-
fraction simulations such as done here, this negative
feedback mechanism is not accounted for, and conse-
quently the model tends to compute unrealistically deep
channels. This is further aggravated by diffusion of
sediment from high-concentration areas (deep channel)
to lower concentration areas (shallow areas). Both
effects are compensated by increasing the gravity-
driven part of the bed load transport component
(directed down-slope, which mainly is in the cross-
channel direction and consequently perpendicular to the
main flow) with a factor . Because the transport by
suspended load is three orders higher than by bed load,
was set to 1000.
5. Results
5.1. Morphodynamics
The initial phase of planform morphology develop-
ment is characterized by downstream migrating alternate
bars (Fig. 7ab). The propagation speed gradually
decreases and a more asymmetric pattern emerges with
subaereal shoals or bars, intersected by migrating
channels (Fig. 7c and d). These channels diverge,
converge, and are abandoned in a manner typical for
braided river systems, resulting in a planform morphol-
ogy such as Fig. 7d.
In nature, new braiding channels are mainly initiated
by mid-channel bar formation or channel cut-off
(Section 4.1). Analysis of the simulated channel
development revealed that, although braid bar formation
is important at the initial stage of development (similar
to results of Repetto et al. (2002)), channel cut-off seems
to be the dominant braiding mechanism when a distinct
pattern with channels and dry areas has formed. This
 D.S. van Maren / Sedimentary Geology 202 (2007) 297316 305

Fig. 7. River bed development at 6-week intervals with a constant discharge of 5000 m3/s and without a washload fraction (simulation (a) in Fig. 9).
The water depth ranges between 0 (white) and 5 m (black shading).

channel diversion is exemplified by Fig. 8. Sediment off gradually deepens and forms a new channel, while
carried in some active braiding channels is deposited the old channel is abandoned (Fig. 8bd).
resulting in the formation of a cut-off in the outer bend The planform morphologies that emerge using fine-
of the channel by headwards erosion (Fig. 8a). This cut- grained washload (4 m) concentrations of 0, 100 and
306 D.S. van Maren / Sedimentary Geology 202 (2007) 297316

Fig. 8. Detailed computed channel pattern development using a constant discharge of 5000 m3/s and no washload fraction, at 20-day intervals. Black
shading indicates sediment deposition, whereas gray shading indicates erosion. Channel cut-off is initiated by sediment deposition in the main
channel (black arrow in (a)) which results in new channel formation (gray arrow in (a)). This channel gradually incises in the downstream direction
until a new channel has formed (gray arrows in (b) and (c) while simultaneously the channel downstream of the cut-off is abandoned (black arrows in
(b) to (d).

200 kg m 3 (Fig. 9ac) are all very similar and
characterized by a braiding pattern, although simula-
tions without a washload fraction (Fig. 9a) have more
channels compared to those with an intermediate
washload fraction. Using a concentration of 400 kg
m 3 (Fig. 9d), however, a slightly meandering single-
channel patterns evolves.
Simulations with coarse-grained washload (40 m,
resulting in larger vertical density gradients) are much
more sensitive to the sediment concentration (Fig. 10).
Channel patterns that evolve vary with the washload
sediment concentrations. A meandering channel pattern
evolves at C = 100 kg m 3 and 400 kg m 3 whereas a
braiding pattern develops at C = 0 kg m 3 and 200 kg
m 3. The main difference with the fine-grained wash-
load simulations is the meandering pattern that evolves
at C = 100 kg m 3. The reasons for this will be explored
in more detail in Section 5.2.
It is noted that one of the morphological timescales is
determined by the settling velocity of the sediment. Due
to hindered settling, the settling velocity decreases (and
hence this morphological timescale increases) with a
factor 2.2, 5.4 and 69.2 for C = 100, 200, and 400 kg m 3
respectively. However, the current model is too short for
simulating the longer timescales because the results
become affected by the boundary conditions. A longer
model though is prohibited from a computation time
point of view. It is therefore not possible to obtain
dynamic equilibrium conditions for all simulations.
However, the results presented here are indicative for
the patterns that develop.
5.2. Hydrodynamics
The previous section shows that the observed
channel patterns become increasingly meandering with
(a) an increasing sediment concentration, and (b) an
increasing vertical sediment concentration gradient. In
the suspended-load dominated system analyzed here,
the morphodynamic behaviour is determined by the
vertical sediment fluxes between the bed and the water
column. Note that in these simulations, the silt fraction
Csilt does not interact with the bed and therefore
morphodynamic behaviour only depends on the sand
fraction Csand. The vertical sediment fluxes are deter-
mined by the uplift of sediment by bed-generated
 D.S. van Maren / Sedimentary Geology 202 (2007) 297316 307

Fig. 9. River bed development with a constant discharge of 5000 m3/s using a washload with a concentration of 0, 100, 200, and 400 kg m 3 and a
grain size of 4 m. The water depth ranges between 0 (white) and 5 m (black shading). Note that a, b, and c are half year simulations whereas d is a 1
year simulation.

turbulent kinetic energy, and by the downward motion
of sediment through the sediment fall velocity ws. Bed
generated turbulent kinetic energy is proportional to the
cubed depth-averaged velocity, but modified by vertical
density gradients. However, the most pronounced effect
of sediment-induced density gradients is damping of the
bed-generated vertical mixing (i.e. the eddy viscosity ),
leading to a decreasing capacity of the flow to keep
sediment in suspension. This can be illustrated through
the vertical profiles of the computed flow velocity U,
308 D.S. van Maren / Sedimentary Geology 202 (2007) 297316

Fig. 10. River bed development with a constant discharge of 5000 m3/s using a washload with a concentration of 0, 100, 200, and 400 kg m 3 and a
grain size of 40 m. The water depth ranges between 0 (white) and 5 m (black shading). Note that a, b, and c are half year simulations whereas d is a 1
year simulation.

eddy viscosity , and sediment concentration C. 5.2.1. Fine-grained washload

Therefore vertical profiles of U, and C, computed in For fine-grained washload simulations (D50 = 4 m)
the deepest point of the initial river cross-section 5 km the washload distribution is almost vertically uniform
downstream of the upstream boundary before beginning (Fig. 11c; note the detailed horizontal scale). Therefore
of morphologic bed update, are analyzed in more detail. the sediment-induced density gradients d/dz are low
 D.S. van Maren / Sedimentary Geology 202 (2007) 297316 309

and consequently the vertical profiles of U and
(Fig. 11a and b) are logarithmic and parabolical,
respectively, as in clear water, and are only slightly
varying with the sediment concentration (note that any
increase of kinematic viscosity with sediment concen-
tration is not included in the numerical model). The
vertical distribution of sand (Fig. 11d), however,
strongly varies with the sediment concentration. For
low washload concentrations the sand concentration
profile follows a Rouse profile, but at high washload
concentrations the sand is vertically uniform due to
hindered settling. In Section 5.1 it was concluded that
the most important braiding process in the computations
was channel blocking and channel cut-off. Channel
blocking, however, is less frequent at high concentra-
tions because the sedimentation rates are much lower
due to reduced fall velocity. As a result, the channel
pattern becomes more meandering at high concentra-
tions. It is therefore concluded that the computed
meandering channel pattern in Fig. 9d is caused by
hindered settling effects.
5.2.2. Coarse-grained washload
The vertical distribution of the washload fraction is
not uniform at all sediment concentrations for the coarse
(D50 = 40 m) washload simulations (Fig. 12c), which
has some important implications. When sediment is not
well vertically mixed, the vertical sediment-induced

Fig. 11. Vertical profile of flow velocity (a), eddy viscosity (b), relative washload concentration (c), and sand concentration (d) for a washload D50 of
4 m at various concentrations.
310 D.S. van Maren / Sedimentary Geology 202 (2007) 297316

density gradients increase with sediment concentration.
As a result, vertical mixing is increasingly damped,
which in turn further increases the vertical density
gradients. At some critical concentration the vertical
mixing becomes so low that the turbulence structure and
the sediment concentration profile collapse (Winter-
werp, 2001, 2006). At still higher concentrations,
however, sediment settling rates become so low due to
hindered settling that the concentration profile is re-
established (Winterwerp, 2006). In the coarse washload
simulations, the largest vertical gradients in silt
concentration occur at C = 100 kg m 3, for which the
concentration increases from several kg m 3 near-
surface to several hundreds of kg m 3 near-bottom over
a depth of 2 m. The resulting sediment-induced density
gradients are very large, significantly damping vertical
mixing (Fig. 12b). The result of this dampened mixing is
a strongly modified velocity profile with increased near-
surface flow velocities (Fig. 12a) and a more or less
decoupling of the near-surface and the near-bed water
flow. This is compensated for by a reduced water depth
(not shown in Fig. 12 because the water depth is made
dimensionless). It therefore seems that the meandering
channel pattern in Fig. 10b is caused by these sediment-
induced density effects. The question then arises how to
explain this relationship in physical terms.
Comparing the computed hydrodynamics in a mean-
dering channel and a braiding channel pattern reveals

Fig. 12. Vertical profile of flow velocity (a), eddy viscosity (b), relative washload concentration (c), and sand concentration (d) for a washload D50 of
40 m at various concentrations.
 D.S. van Maren / Sedimentary Geology 202 (2007) 297316 311

some further differences such as a more pronounced
secondary circulation in meandering streams (not shown).
These higher secondary flow velocities may be related to
the more strongly curved bends in the meandering
channel. Therefore the hydrodynamics of the various
model simulations are analyzed in more detail during the
initial phase of channel deformation (characterized by a
more or less symmetric pattern such as in Fig. 7b). This
reveals a fundamentally different secondary flow pattern
between simulations with C = 100 kg m 3 and
D50 = 40 m (Fig. 13) and simulations without washload
(Fig. 14) or with C = 100 kg m 3 and D50 = 4 m
(Fig. 15). It should be noted that Fig. 13 shows model
results after 2 weeks whereas the other two are after
4 weeks, because the morphologic changes in Fig. 13 are
faster. The magnitude of the secondary flow (defined as
the velocity normal to the depth-averaged flow velocity)
is several orders of magnitude larger for C = 100 kg m 3
and D50 = 40 m. This is caused by a strongly reduced
exchange of turbulent kinetic energy at the density
interface (luto cline) around C = 100 kg m 3 (Fig. 13c),
which decouples the flow dynamics in the surface layer
from the bottom layer (Fig. 13b). The centripetal
acceleration needed to keep the flow in a circular path
increases with the flow velocity whereas the lateral
pressure gradient is constant over depth. Therefore the
surface layer (with high flow velocities) is deflected
towards the outer bend while the bottom layers are
deflected towards the inner bend (e.g. Jansen et al., 1979).
A high vertical velocity gradient, caused by stratification
(such as Fig. 12a), will therefore lead to a stronger
secondary circulation. This stronger secondary circu-
lation leads to increased lateral migration of the channel,
and therefore an increased curvature of the channel

Fig. 13. Cross-sectional distribution at X = 20 km of flow velocity in depth-averaged flow direction (a), flow velocity normal to depth-averaged flow
direction (b), and total sediment concentration (c) after 2 weeks simulation, using a washload concentration of 100 kg m 3 and a grain size of 40 m.
312 D.S. van Maren / Sedimentary Geology 202 (2007) 297316

Fig. 14. Cross-sectional distribution at X = 20 km of flow velocity in depth-averaged flow direction (a), flow velocity normal to depth-averaged flow
direction (b), and total sediment concentration (c) after 4 weeks simulation, without washload.

bend, which in turn further strengthens the secondary
circulation.
The concentration gradient is over 250 kg m 3/m,
which equals a density gradient f /z of 155 kg m 3/m
(where f is the density of the watersediment mixture).
Turbulence can be maintained as long as the flux
Richardson number Rif is less than 0.2 (Tennekes and
Lumley, 1994). However, this number is not be easily
obtained because it requires knowledge of the eddy
viscosity and the eddy diffusivity. Alternatively, assum-
ing that the eddy viscosity is approximately equal to
eddy diffusivity (see Tennekes and Lumley, 1994) the
gradient Richardson number Rig can be used:
@qf
 qg @z
Rig f

2
@u 2
@z
with u/z at the density interface being approximate-
ly 1 m/s/m, Rig is approximately 1.3, which means that
vertical mixing is suppressed. Figs. 1315 show that
this damping yields larger secondary currents, which
promotes meandering because (1) secondary flow is an
essential process for meander growth, and (2)
complete channel blocking (which was concluded to
be the main process responsible for the channel
braiding) is prevented, because when large amounts
of sediment settle from suspension, the secondary
currents increase, and redistribute sediment to one side
of the channel, leading to a sideways migration of the
channel rather than to abandonment. Computing Rig
numbers for the measurements in Fig. 3 yields Rig
numbers exceeding 0.2, which shows that dampened
vertical mixing does indeed play a role in the Yellow
River.
 D.S. van Maren / Sedimentary Geology 202 (2007) 297316 313

Fig. 15. Cross-sectional distribution at X = 20 km of flow velocity in depth-averaged flow direction (a), flow velocity normal to depth-averaged flow
direction (b), and total sediment concentration (c) after 4 weeks simulation, using a washload concentration of 100 kg m 3 and a grain size of 4 m.

At C = 400 kg m 3, hindered settling reduces the fall
velocity of the washload fraction to such a degree that
Csilt is vertically uniform. As a result, sediment-induced
density gradients are low which is reflected in a
parabolic eddy viscosity (Fig. 12b) and a logarithmic
flow velocity profile (Fig. 12a). Therefore the mean-
dering pattern at C = 400 kg m 3 (Fig. 10d) can be
attributed to hindered settling, similar to the fine
washload simulations (Fig. 9d).
6. Discussion
6.1. Model assumptions
An important question is to what degree the applied
model used here can reproduce the complicated
processes associated with channel pattern formation,
considering its simplifications and limited number of
processes included. Processes on a subgrid scale (here
50125 m) are not simulated, because most of the
relevant morphological units (alternate bars, river
channels) are much larger than this grid size. Further-
more, the factor used to compensate for the diffusion
of sediment towards the shallow areas of the river
(Section 4.2) tends to widen and shallow the river
channel. Since it is understood from both linear stability
analyses (Engelund and Skovgaard, 1973; Parker, 1976;
Fredse, 1978) and geomorphic studies (Leopold and
Wolman, 1957) that braiding occurs at high width-to-
depth ratios, this means that the factor promotes
braiding relative to meandering. The correction may
be needed because of the use of a single grain-size
fraction for the morphodynamics and an oversimplifi-
cation of the bed roughness.
314 D.S. van Maren / Sedimentary Geology 202 (2007) 297316
An assumption within the k model is that the
PrandtlSchmidt number t (see Section 5.2) equals 0.7.
However, Cellino and Graf (1999) observed in labora-
tory experiments that t should be closer to 2 in nearly
saturated flows (note that these authors use a factor,
where = 1/t). This was confirmed by recent numerical
experiments of sediment-laden flow (Winterwerp,
2006). A higher PrandtlSchmidt number results in a
reduced sediment diffusivity ( = /t), thereby
increasing the vertical concentration gradients (and
therefore sediment-induced density gradients) in the
concentration profiles such as those in Fig. 12c.
Following the reasoning in Section 5.2, this would result
in an increased tendency to meander. Indeed, repeating
the simulation with D50 = 40 m and Csilt = 100 kg m 3
(Fig. 10b) and t = 2 results in a much more pronounced
tendency to meander. However, these simulations are
morphologically unstable and therefore not suitable for
long computations.
Inevitably, there are a number of model simplifica-
tions regarding the modelling of highly concentrated
flows. Although very rare in the lower Yellow River,
hyperconcentrated floods are known to erode bed
sediment in blocks by mass erosion, scouring the river
bed up to 9 m. (Wan and Wang, 1994; Wang et al., 1997,
2002). Mass erosion is not included in the model used
here. There are also indications that sediment erosion
rates may be reduced during hyperconcentrated condi-
tions because of hindered erosion effects (Winterwerp
and van Kesteren, 2004), a process very similar to
hindered settling. Finally, some hyperconcentrated flows
become non-Newtonian, which may result in laminar
plug flow. Of particular interest may be the transitional
regime between laminar and turbulent flow during which
large-scale KelvinHelmholtz instabilities are generated
that dominate both turbulence production and fluid
mixing (Baas and Best, 2002). These are presently not
included in the model, but may be important for the
previously mentioned mass erosion processes.
6.2. Channel pattern formation
The channel pattern in Fig. 7 is initiated by alternate
bar instability, followed by enlargement of the bars and
channels. This pattern occurred in all simulations, albeit
with different wavelengths and amplitudes, even in
simulations where the final pattern was meandering. The
classic braiding process is channel bifurcation resulting
from streamline convergence around a steady braid bar
(Leopold and Wolman, 1957), and a braid bar forms
downstream of a widening river (Mosley, 1976). A
recent stability analysis (Repetto et al., 2002) repro-
duced this braid bar formation as a result of channel
widening and narrowing, which is regulated by
transverse flow components. However, there are many
additional mechanisms that initiate braid development
(see e.g. Ashmore, 1991; Ferguson, 1993; Bridge, 1993;
Bridge, 2003). In the numerical simulations presented
here, it seems that channel blocking and channel bend
cut-off dominate the braiding channel formation. With
the given discharge (Q = 5000 m3/s) and the slope
(i = 2.2 10 4) one would expect a braiding pattern
using the Leopold and Wolman (1957) discrimination of
i = 0.0125Q0.44. Although this expression is an outdat-
ed oversimplification because channel pattern is not
determined by discharge and channel slope alone (and,
even more, by valley slope rather than channel slope
(Ferguson, 1987; Van den Berg, 1995)), it does indicate
that under normal flow conditions, a braiding channel
pattern is to be expected. However, in silt-laden rivers
there seems to be an additional role of sediment in
channel pattern formation. A meandering channel
pattern may develop instead of a braiding channel
through: (a) increasing secondary currents due to
turbulence suppression at the lutocline, and (b) hindered
settling effects that delay or prevent channel blocking.
The degree of turbulence suppression depends on the
velocity shear and the density gradient (see Eq. (2)). The
density gradient increases with grain size and sediment
concentration (until hindered settling effects become too
important), which implies that in highly concentrated
suspended load-dominated rivers the tendency to
meander increases with grain size and concentration.
This is in agreement with previous relations between
grain size and channel pattern (Ferguson, 1987; Van den
Berg, 1995; Dade, 2000), even though these findings are
based mostly on rivers dominated by bedload transport
or mixed load transport.
The concentration-dependence of the channel pattern
was already described by Xu (2002) and Xu (2004), but
this paper provides a physical explanation for channel
pattern variations with sediment concentration in silt-
laden flows. However, the exact relation between
concentration and channel pattern such as presented by
Xu (meandering at low concentrations and high
concentrations, braiding in between) is not exactly
reproduced since meandering was mainly simulated at
concentrations close to 100 kg m 3 instead of the several
10s of kg m 3 observed by Xu (2002, 2004). This may
be caused by the relation between channel slope i and
sediment concentration in natural rivers. Rivers with low
sediment concentrations probably have a lower slope
than rivers with a high sediment load, and therefore tend
to meander more strongly. The numerical simulations,
 D.S. van Maren / Sedimentary Geology 202 (2007) 297316
however, have a constant slope. Additionally, the low-
concentration meandering required a vertical sediment-
induced density gradient. This vertical density gradient,
however, increases significantly with an increasing
spread in grain size. Therefore the use of more sediment
fractions would lead to a stronger tendency to meander,
also at lower sediment concentrations.
7. Conclusions
Sediment concentration significantly influences the
channel pattern in sediment-laden flows, such as found
in the silt-laden rivers on the Chinese Loess plateau.
Previous studies revealed that the channel pattern of
these rivers becomes meandering at both low and high
concentrations, and braided in between. Based on a
numerical modelling study, it is concluded that the main
mechanism for meandering at relatively low concentra-
tions is damping of vertical mixing by sediment-induced
density gradients, which strengthen secondary currents
significantly. The coarser the grain size, the higher the
sediment density gradients, and therefore the more
pronounced this effect is. Braiding is mainly caused by
channel blocking due to sedimentation in the main
channel, ponding and ultimately channel avulsion. At
very high concentrations, hindered settling becomes so
dominant that channels are no longer blocked and a
single meandering channel evolves.
Acknowledgements
This work is part of the project Predictive Morpho-
logical Modelling of the Lower Yellow River, financed
by the Dutch Royal Academy of Sciences (KNAW) and
the Chinese Ministry of Science and Technology within
the framework of the Programme of Scientific Alliances
between China and the Netherlands. B. S. Wu, Z-Y
Wang, J. J. Zhou, and J. Q. Xia from Tsinghua
University are acknowledged for discussions and the
use of data, and J. C. Winterwerp and two anonymous
reviewers for their useful comments and suggestions.
References
Ashmore, P.E., 1991. How do gravel-bed rivers braid? Canadian
Journal of Earth Sciences 28, 326341.
Baas, J.H., Best, J.L., 2002. Turbulence modulation in clay-rich
sediment-laden flows and some implications for sediment
deposition. Journal of Sedimentary Research 72 (3), 336340.
Blondeaux, P., Seminara, G., 1985. A unified bar-bend theory of river
meanders. Journal of Fluid Mechanics 157, 449470.
Bridge, J.S., 1993. The interaction between channel geometry, water
flow, sediment transport and deposition in braided rivers. In: Best,
315
J.L., Bristow, C.S. (Eds.), Braided rivers. Geol. Soc. Spec. Publ.,
vol. 75, pp. 1371.
Bridge, J.S., 2003. Rivers and floodplains: forms, processes and
sedimentary record. Blackwell Science, Oxford, p. 489.
Cellino, M., Graf, W.H., 1999. Sediment-laden flow in open-channels
under noncapacity and capacity conditions. Journal of Hydraulic
Engineering 125 (5), 456562.
Chien, N., Wan, Z.H., 1998. Mechanics of sediment transport. ASCE
press, Reston, Virginia. 913 pp.
Dade, W.B., 2000. Grain-size, sediment-transport, and alluvial channel
pattern. Geomorphology 35, 119126.
de Vriend, H.J., Zhou, J.J., Wang, Z.B., Wang, Z.Y., Wu, B.S.,
Winterwerp, J.C., van Maren, D.S., 2005. Predictive modelling of
sediment transport and morphology in the lower Yellow River.
Proc. of the 2nd Int. Yellow River Forum on River Basin
Management, Zhengzhou, China, vol. 3, pp. 128134.
Engelund, F., Skovgaard, O., 1973. On the origin of meandering and
braiding in alluvial streams. Journal of Fluid Mechanics 57, 289302.
Enggrob, H.G., Tjerry, S., 1999. Simulation of morphological
characteristics of a braided river. Proc. of the 1st IAHR Symposium
on River, Coastal, and Estuarine Morphodynamics, Genova, Italy,
1999, pp. 585594.
Ferguson, R.I., 1987. Hydraulic and sedimentary controls of channel
pattern. In: Richards, K.S. (Ed.), River Channels: Environment and
process. Blackwell, Oxford, pp. 129158.
Ferguson, R.I., 1993. Understanding braiding processes in gravel-bed
rivers: progress and unsolved problems. In: Best, J.L., Bristow, C.S.
(Eds.), Braided rivers. Geol. Soc. Spec. Publ., vol. 75, pp. 7387.
Fredse, J., 1978. Meandering and braiding of rivers. Journal of Fluid
Mechanics 84 (4), 609624.
Germanoski, D., Schumm, S.A., 1993. Changes in braided river
morphology resulting from aggradation and degradation. Journal
of Geology 101, 451466.
Hibma, A., de Vriend, H.J., Stive, M.J.F., 2003. Numerical modelling of
shoal pattern formation in well-mixed elongated estuaries 57, 981991.
Hibma, A., Schuttelaars, H.M., de Vriend, H.J., 2004. Initial formation
and long-term evolution of channel-shoal patterns. Continental
Shelf Research 24 (15), 16371650.
Ikeda, S., Parker, G., Sawai, K., 1981. Bend theory of meanders. Part
1: Linear development. Journal of Fluid Mechanics 112, 363377.
Jagers, H.R.A., 1999. Numerical analysis of cutoff development. Proc.
of the 1st IAHR Symposium on River, Coastal, and Estuarine
Morphodynamics, Genova, Italy, 1999, pp. 553562.
Jansen, P.P., van Bendegom, L., van den Berg, J., de Vries, M., Zanen,
A., 1979. Principles of River Engineering, the non-tidal alluvial
river. Pitman, London.
Leopold, L.B., Wolman, M.G., 1957. River channel patterns: braided,
meandering and straight. Us. S. G. S. Prof. Paper, vol. 282-B.
Lesser, G.R., Roelvink, J.A., van Kester, J.A.T.M., Stelling, G.S.,
2004. Development and validation of a three-dimensional
morphological model. Coastal Engineering 51, 883915.
Long, Y., Liang, G., Zhang, Y., Shen, G., Zhang, L., Cheng, L., 2002.
Range survey of deposition in the lower Yellow River. Interna-
tional Journal of Sediment Research 17 (2), 91101.
Marciano, R., Wang, Z.B., Hibma, A., de Vriend, H.J., Defina, A.,
2005. Modeling of channel patterns in short tidal basins. Journal of
Geophysical Research 110 (F01001). doi:10.1029/2003JF000092.
Mosley, M.O., 1976. An experimental study of channel confluences.
Journal of Geology 74, 535562.
Olsen, N.R.B., 2003. Three-dimensional CFD modeling of self-
forming meandering channel. Journal of Hydraulic Engineering
129 (5), 366372.
316 D.S. van Maren / Sedimentary Geology 202 (2007) 297316
Parker, G., 1976. On the cause and characteristic scales of meandering
and braiding in rivers. Journal of Fluid Mechanics 76, 457480.
Pu, Q., 1997. Effect of perennial sediment regulation in Xiaolangdi
reservoir on reduction of deposition in the lower Yellow River.
International Journal of Sediment Research 12 (2), 5867.
Repetto, R., Tubino, M., Paola, C., 2002. Planimetric instability of channels
with variable width. Journal of Fluid Mechanics 457, 79109.
Roelvink, J.A., van Kessel, T., Alfageme, S., Canizares, R., 2003.
Modelling of barrier island response to storms. Proceedings of
Coastal Sediments '03, Clearwater, Florida, pp. 111.
Ruether, N., Olsen, N.R.B., 2003. CFD modeling of alluvial channel
instabilities. Proc. of the 3rd IAHR Symposium on River, Coastal, and
Estuarine Morphodynamics, Barcelona, Spain, 2003, pp. 636642.
Schumm, S.A., 1977. The fluvial system. John Wiley & Sons, New
York. 338 pp.
Seminara, G., Tubino, M., 1992. Weakly nonlinear theory of regular
meanders. Journal of Fluid Mechanics 244, 257288.
Struiksma, N., Olesen, K.W., Flokstra, C., de Vriend, H.J., 1985. Bed
deformation in curved alluvial channels. Journal of Hydraulic
Research 23 (1), 5779.
Tennekes, H., Lumley, J.L., 1994. A first course in turbulence. Proc.
Estuarine and Coastal Modelling. MIT press, Cambridge,
Massachusetts, p. 300.
Tubino, M., Seminara, G., 1990. Free-forced interactions in develop-
ing meanders and suppression of free bars. Journal of Fluid
Mechanics 214, 131159.
Tubino, M., Repetto, R., Zolezzi, G., 1999. Free bars in rivers. Journal
of Hydraulic Research 37, 759775.
Van den Berg, J.H., 1995. Prediction of alluvial channel pattern of
perennial rivers. Geomorphology 12, 259279.
van Maren, D.S., Winterwerp, J.C., Wu, B.S., Zhou, J.J., 2006.
Morphological modelling of hyperconcentrated floods in the
Yellow River. Proc. of the River Flow conference, Lissabon,
Portugal, 2006, vol. II, pp. 13831392.
van Rijn, L.C., 2000. General view on sand transport by currents and
waves: data analysis and engineering modelling for uniform and
graded sand (TRANSPOR 2000 and CROSMOR 2000 models).
Tech. Rep. Z2899.20. Delft Hydraulics.
Wan, Z., Wang, Z., 1994. Hyperconcentrated flow. IAHR monograph
series. Balkema, Rotterdam, The Netherlands. 230 pp.
Wang, G., Xu, M., 1999. Analysis of peculiarities of 92.8 flood event
in the lower Yellow River. International Journal of Sediment
Research 14 (2), 205213.
Wang, G., Xia, J., Zhang, H., 2002. Theory and practice of
hyperconcentrated sediment-laden flow in China. Proc. of the
13th Congress of the Asia and Pacific Division, IAHR, Singapore,
pp. 10551071.
Wang, Z.Y., 1999. Experimental study on scour rate and river bed
inertia. Journal of Hydraulic Research 37 (1), 1737.
Wang, Z.Y., Liang, Z.Y., 2000. Dynamic characteristics of the Yellow
River mouth. Earth Surface Processes and Landforms 25, 765782.
Wang, Z.Y., Huang, J., Su, D., 1997. Scour rate formula. International
Journal of Sediment Research 13 (3), 1120.
Winterwerp, J.C., 2001. Stratification effects by cohesive and non-
cohesive sediment. Journal of Geophysical Research 106 (C10),
22.55922.574.
Winterwerp, J.C., 2006. Stratification effects by fine suspended
sediment at low, medium and very high concentrations. Journal of
Geophysical Research 111 (C5). doi:10.1029/2005JC003019.
Winterwerp, J.C., van Kessel, T., 2003. Siltation by sediment-induced
density currents. Ocean Dynamics 53, 186196.
Winterwerp, J.C., van Kesteren, W.G.M., 2004. Introduction to the
physics of cohesive sediment in the marine environment.
Developments in Sedimentology, vol. 56. Elsevier, Amsterdam,
the Netherlands. 466 pp.
Xu, J., 1999. Erosion caused by hyperconcentrated flow on the loess
plateau of China. Catena 36, 119.
Xu, J., 2002. Complex behaviour of natural sediment-carrying
streamflows and the geomorphological implications. Earth Surface
Processes and Landforms 27, 749758.
Xu, J., 2004. Double thresholds in scour-fill processes and some
implications in channel adjustment. Geomorphology 57, 321330.
Yellow River Conservancy Commission, 1959. Hydrological data
from the Yellow River. Red Books. (in Chinese).
Zhao, Y., Pan, X., Fan, Z., Han, S., 1989. Sedimentation in the lower
reaches of the Yellow River and its basic laws. In: Brush, L.M.,
Wolman, M.G., Huang, B.W. (Eds.), Taming the Yellow River: silt
and floods. Kluwer Academic Publishers, Dordrecht, pp. 477516.
 Available online at www.sciencedirect.com

Sedimentary Geology 202 (2007) 317 332

www.elsevier.com/locate/sedgeo

Estimating and modeling the annual contribution of authigenic

sediment to the total suspended sediment load in the
Kleine Nete Basin, Belgium
Elin Vanlierde a,, Jan De Schutter b , Patric Jacobs a , Frank Mostaert b
a
Department of Geology and Soil Science, Research unit for Sedimentary Geology and Engineering Geology, Ghent University,
Krijgslaan 281 S8, B-9000 Ghent, Belgium
b
Flanders Hydraulics Research, Ministry of the Flemish Community, Berchemlei 115, B-2140 Borgerhout (Antwerp), Belgium

Abstract

In the Nete basin (Flanders, Belgium) chemical precipitates derived from groundwater-associated Fe2+ seeping into the
overlying surface water significantly contribute to the composition, concentration, and fluxes of suspended sediment.
The degree of contribution of these precipitates for the period of 1999 up to 2004 was determined by comparing measured
annual suspended sediment loads with theoretical calculations as well as values generated with MARS (Model for Authigenic
River Sediment) (1.0), a numerical model newly developed for predicting authigenic sediment flux. Theoretical calculations, based
on various estimates of the concentration of Fe2+ in groundwater, after stoichiometric and sorption correction, place the averaged
annual contribution of authigenic sediment between 43% and 100% for the period-of-record. MARS 1.0 predicts annual contributions
of authigenic suspended sediment of 58% up to 96%.
2007 Elsevier B.V. All rights reserved.

Keywords: Sediment transport; Suspension; Authigenesis; Ferruginous composition; Mathematical models

1. Introduction anthropogenic (e.g. sewage treatment discharges and

Sediments in rivers can pose problems for policy-
makers and river managers, who need to address issues
such as dredging, flood control, erosion and water qual-
ity. Determining the sources and their relative contribu-
tion to the total sediment flux is of primary importance
if these problems are to be addressed. When locating
potential sources of fluvial sediment, different types of
sources should be taken into account such as detrital
(e.g. land erosion), biological (e.g. algae and diatoms),
Corresponding author. Fax: +32 9 264 49 43.
E-mail address: Elin.Vanlierde@ugent.be (E. Vanlierde).
0037-0738/$ - see front matter 2007 Elsevier B.V. All rights reserved.
doi:10.1016/j.sedgeo.2007.05.003
industrial discharges) and authigenic sources. Authi-
genic sediments are formed when compounds in solution
precipitate due to changing environmental conditions,
thus forming sediment. This can for example happen
when groundwater seeps into surface water.
On an international level, virtually nothing is known
about the contribution of authigenic sediment to the total
sediment load. Generally, articles and reports addressing
sediment sources (Collins et al., 1997; Walling et al., 1999;
AMINAL, 2004) and models calculating the amount of
sediment entering the river (Morgan et al., 1998; Van
Rompaey et al., 2003), do not take authigenic sediment
into account. This is probably due to the fact that the
regions of interest in these studies are often characterized
318 E. Vanlierde et al. / Sedimentary Geology 202 (2007) 317332

by high land erosion, which will decrease the contribution
of authigenic sediments to the total sediment load
transported by a river, or that in these regions, due to the
lack of geological conditions favorable to the production
of authigenic sediment, little is produced.
As an illustration, in Flanders (Belgium), the geolog-
ical and hydrogeological settings are extensively studied
(e.g. Schiltz et al., 1993; Lebbe, 1999; Jacobs et al.,
2000; DOV, 2004), but little is known about the fluvial
sediment transport. Sediment studies in the Nete basin
(Soresma, 2002; Librecht and Van Orshoven, 2003),
estimate the total suspended sediment load (the sum of
soil erosion and discharges from industry and house-
holds) for the entire basin at 7000 t/year. However,
measurements by Flanders Hydraulics Research (FHR,
2006) indicate that the Kleine Nete, the largest tributary
of the basin, transports on its own on average more than
12 000 t/year. This clearly indicates that an important
source of sediment has been overlooked, i.e. the authi-
genic fraction.
This article addresses the issue of authigenic material
as a potential source of fluvial sediment, gives a short
overview of known processes involved in authigenic
sedimentation and gives estimates of the annual contri-
bution of authigenic sediment to the total suspended
sediment load for the Kleine Nete in Belgium. These
estimates are generated through theoretical deductions
as well as through modeling with MARS (Model for
Authigenic River Sediment), a model further developed
from previous research (Vanlierde et al., 2005).
2. Area description
2.1. Local setting
The Nete basin, selected as study area for this
research, is located in the North-East of Belgium and
contains the Grote Nete and the Kleine Nete and their
tributaries (see Fig. 1). In total, the basin drains about
1670 km2 of which approximately 590 km2 lie upstream

Fig. 1. Localization of the Nete basin in Belgium, with its two major tributaries, the Kleine Nete and the Grote Nete, and the monitoring station at
Grobbendonk. The location of the NS cross-section, which is shown in Fig. 2, is also indicated.
 E. Vanlierde et al. / Sedimentary Geology 202 (2007) 317332
Fig. 2. Hydrogeological profile of the NS cross-section through the Nete basin, modified from Envico (2001). Values for horizontal permeability (Kh) of the aquifers and hydraulic resistance (c) are
provided by Lebbe (1999) and Envico (2001).

319
320 E. Vanlierde et al. / Sedimentary Geology 202 (2007) 317332
Table 1
Values of annual measured surface water discharge (Q), numerically
filtered annual baseflow (Bf) and interflow (If) for the period 19992004
1999 2000 2001 2002 2003 2004 1999
2004
Q (106 m3) 226.5 254.3 281.1 251.1 169.7 195.5 1378.2
Bf (106 m3) 150.4 173.4 183.5 170.6 116.2 129.7 923.8
If (106 m3) 44.9 49.0 57.1 48.4 32.7 38.3 270.5
Contribution of 86.2 87.5 85.6 87.3 87.8 85.9 86.7
groundwater
to total Q (%)
of Grobbendonk, on the Kleine Nete, where a monitor-
ing station of Flanders Hydraulics Research is located
(Fig. 1).
The specific settings of the Nete Basin ensure a
significant contribution of authigenic sediment to the
total sediment load of the basin. Its flat topography
ensures a low run-off erosion rate, which, combined with
the presence of ditches and trenches adjacent to the
fields, minimizes the detrital sediment source. Further-
more, the geological structure of the Nete basin ensures a
high production of authigenic sediment. Fig. 2 illustrates
the presence of thick Tertiary deposits in the subsurface,
overlain by a thin layer of Quaternary sediment. The
Boom Clay Formation serves as an aquitard that divides
the Nete groundwater system into two separate hydro-
geological units (De Schutter et al., 1995; Envico, 2001).
As a result, it can be assumed that all groundwater
seeping into the surface water system originates from the
upper unit. This unit consists of three Neogene sandy
aquifers: (1) the Merksplas, Brasschaat, and Mol For-
mations; (2) the Lillo, Poederlee, and Kasterlee Forma-
Fig. 3. Schematic overview of processes involved in the production of authigenic sediment in the Nete basin.
tions; and (3) the Berchem and Diest Formations.
Aquifers 2 and 3 contain high quantities of iron minerals
(e.g. glauconite) and aquifer 3, due to its high perme-
ability and thickness (see Fig. 2), is responsible for most
of the groundwater seepage into the basin.
Several attempts to estimate the contribution of
groundwater seepage to the total water discharge in the
Nete basin have yielded a variety of results. Van Der
Beken and Huybrechts (1990) estimated that between
1901 and 1986, groundwater contributed on average
62% to surface-water discharge of the Nete basin.
Batelaan (2006) compared Sloto and Crouse's (1996)
baseflow separation method with Wittenberg's (1999)
for different sites, including some in the Nete basin. The
method by Sloto and Crouse (1996) yielded baseflow
contributions in the Nete basin between 72 and 81% of
the total discharge, whereas Wittenberg's (1999) placed
the contribution of the baseflow between 73 and 84%.
Groundwater contributions for this study were
separated from the total hydrograph using WETSPRO
(Willems, 2000), which is based on the numeric filtering
procedure by Nathan and McMahon (1990). Hence,
average groundwater contributions were estimated to
account for 87% of the surface water discharge for the
period 1999 to 2004, shown in Table 1. The variability
in these results can be explained by the fact that the
flexibility of the non-linear baseflow separation techni-
ques yields a computed baseflow closer to the total
discharge hydrograph than can be achieved with clas-
sical linear approaches, as noted by Wittenberg (1999).
Subsequently, it can be concluded that the contribution
of seepage in the Nete basin is important in relation to
the total water discharge.
 E. Vanlierde et al. / Sedimentary Geology 202 (2007) 317332 321

Water discharge and sediment transport have been
measured continuously in Grobbendonk, respectively
from 1983 onwards and from 1999 until present. These
data will be presented in Section 5.
2.2. Authigenic sedimentation processes
Fig. 3 schematically depicts the processes in the
study area from the moment precipitation (e.g. rain and
snow) enters the soil until the water seeps into the river,
causing formation of authigenic sediment. The precip-
itation, only being slightly buffered in the upper meters
of the deposits, has to percolate through a thick layer of
glauconite rich sands. The low pH buffering capacity in
combination with the reigning redox conditions, as can
be seen in Table 2, creates a reducing environment, thus
ensuring the presence of substantial amounts of Fe2+ in
the groundwater (AMINAL, 1995; Blommaert et al.,
1988; VLM, 1997). The more oxidizing conditions, as
seen in Table 2, of the Formation of Mol, and the sandy
part of the Formation of Lillo are due to their position
near the surface, in the Nete basin. As the level of the
groundwater table drops in dry periods, oxidation can
take place. However, Swings and Boeye (1997) found
that deeper groundwater currents contributed to seepage
in the Kleine Nete, which ensures the supply of reducing
groundwater.
Once the groundwater seeps into the river, it encoun-
ters completely different environmental conditions, such
as higher pH and redox potential (VMM, 2006; FHR,
2006), which will cause oxidation of the Fe2+, followed
by the hydrolysis to hydrous ferric oxide (Fe2O3XH2O)
as indicated by Stumm and Morgan (1970). These
authors also indicate that the attachment of a hydroxy
group to ferric ions is a relatively fast process. Amor-
phous hydro-ferric oxide is an insoluble compound,
which can sorb contaminants, nutrients a.o. onto its
surface and which can easily flocculate, thus forming
authigenic sediment. Therefore, due to its high seepage
levels of Fe2+-enriched groundwater, the Nete Basin is
an ideal location to study contributions of authigenic
precipitates to the total flux.
2.3. Sediment characterization
It has been established that flocculation has a direct
effect on the settling velocity of fine particles (Li and
Ganczarczyk, 1987; Droppo et al., 2005 a.o.) and that
flocculation is a function of biological agents (e.g. bacteria
and polysaccharides etc.), organic material, clay minerals,
meteorological factors (e.g. rainfall related water discharge),
hydrological conditions (e.g. turbulence, tide) and chemical
components (e.g. complex iron-oxides and -hydroxides)
(Van Leussen, 1994; Droppo et al., 1997; Droppo et al.,
2000; Mikes et al., 2004; Govoreanu, 2004; Nopens, 2005;
Chen et al., 2005).
As complex iron-oxides and hydroxides are abun-
dantly present in the Nete, sampling campaigns were

Table 2
Values for redox potential (Eh) (DOV, 2006), pH and concentrations of FeTot, (VMM, 2006) in the different aquiferal systems in the Kleine Nete
basin as well as values for redox potential (Eh) (FHR, 2006), pH, concentrations of FeTot, and the percentages of FeTot and PTot present in the total
suspended sediment (VMM, 2006)
Surface water Groundwater
Kleine Nete Neogene Sand Neogene Sand Neogene Sand
Grobbendonk aquifer 1 aquifer 2 aquifer 3
Merksplas Brasschaat Mol Lillo Poederlee Kasterlee Diest Berchem
Eh (mV) Min 181 36 97 121 2 101 64
Max 362 690 580 420 551 280 240
Median 129 174 354 256.5 260.5 148.5 113
pH Min 6.30 4.69 5.05 6.37 4.46 5.28 5.48
Max 8.21 6.21 6.97 7.53 7.39 8.04 7.44
Median 7.30 5.96 6.54 7.09 5.83 6.26 6.89
[FeTot] mg/l) Min 0.70 0.20 0.00 0.03 0.05 0.05 0.07
Max 11.40 68.10 10.46 26.04 98.73 148.01 42.80
Median 4.29 4.77 0.43 2.36 9.10 15.68 1.56
FeTot  Min 6.45
% Max 75.20
SSC
Median 27.07
PTot  Min 0.81
% Max 43.33
SSC
Median 3.60
322 E. Vanlierde et al. / Sedimentary Geology 202 (2007) 317332

The flocculated state of the sediments in the Kleine

Nete will alter the sediment transport dynamics such as
settling velocities. Therefore, flocculation will have to
be taken into account in the process of modeling authi-
genic sediment transport, as is addressed in Section 3.3.

3. Methods

3.1. Determining total suspended sediment concentra-

tions and fluxes

To estimate the contribution of authigenic sediment,

Fig. 4. Microscopic image of suspended sediment, sampled in the Kleine
Nete, at vertical 4 at 1 m below surface level, on 08 March 2006, and
filtered on a 0.45 m Millipore nitro-cellulose membrane filter. Image
obtained with an Olympus BH2 microscope with transmitted light.
executed at Grobbendonk on 7 and 8 March and 9 May
2006 to obtain insight into the nature of the suspended
sediment in the Kleine Nete. On different locations in
the cross-section, samples were taken and then filtered
on 0.45 m Millipore nitro-cellulose membrane filters.
The filters were then put under a microscope and
photographed, after which the images were statistically
analyzed. Fig. 4 shows an example of such an image.
The analyses confirm that material transported in
suspension in the Kleine Nete is highly flocculated, as
reported by Belien (2006) and Belien et al. (2006). Most
floc length values, observed by image analysis, varied
between 100 and 300 m, while the largest floc length
observed was 363 m.
Furthermore, Belien (2006) also showed that flocs in
the Kleine Nete have dimensions that vary as a function
of water depth. Close to the water surface, the flocs are
large. At around 60% of the total water depth, flocs are
significantly smaller (see Table 3). The largest flocs
were observed at approximately 90% of the total water
depth, close to the river bed.
one needs to know the authigenic load and the total
sediment load in the basin. However, only suspended
sediment fluxes (SSFes) have been measured on the
Kleine Nete at the monitoring station in Grobbendonk.
Between 3 February 1999 and 31 January 2000 and
again from 29 July 2002 up to present, the site was
equipped with an automatic sampler, which collects
water samples every 7 h. These samples are collected
weekly and processed in the laboratory where, after
filtration of 100 ml through a 0.45 m nitrocellulose
membrane Millipore filter followed by air drying, sus-
pended sediment concentrations (SSCs) are determined.
Consequently, daily averages are calculated, as well as
annual suspended sediment fluxes, through multiplica-
tion by daily discharge values. Discharge values are
obtained from on-site gauge height measurements and
from a site-specific, validated gauge heightdischarge
relationship.
Between 1 January 2002 and 28 July 2002 the sam-
plers were removed from the site and no samples were
taken, so SSCs for this period could only be indirectly
determined by developing a site-specific regression
model. Empirical relations such as regression models
and rating curves are frequently used to estimate SSCs
in the absence of actual concentration data (e.g. Walling
and Webb, 1981, De Vries and Klavers, 1994; Philips
et al., 1999; Asselman, 2000; Horowitz et al., 2001).

Table 3
Average values for shape and size characteristics, obtained through image analysis, of flocculated material, sampled on different water depths, on two
different verticals in the cross-section of the Kleine Nete in Grobbendonk
Vertical 8 7/3/06 Vertical 6 9/5/06
Sampling depth under the water surface (m) 0.34 m 1.02 m 1.65 m 0.07 m 0.27 m 0.81 m 1.31 m
Area (m2) 152.32 117.02 159.3 500 528.57 371.43 533.33
Perimeter (m) 68.9 64.04 68.71 116.29 117.43 88.43 116.89
Major axis length (m) 16.73 15.66 16.75 31 30.14 26.57 30.89
Minor axis length (m) 8.97 8.47 9.06 15.43 16.29 15 15.56
Elongation (m) 2.49 2.57 2.57 2.11 2.02 1.83 2.23
Roundness () 0.35 0.35 0.38 0.43 0.41 0.6 0.44
Feret diameter (m) 10.03 9.38 10.06 18.71 18.86 17.86 18.67
Compactness () 0.66 0.66 0.67 0.69 0.7 0.77 0.69
 E. Vanlierde et al. / Sedimentary Geology 202 (2007) 317332
The regression model was also used to complete the
existing dataset, in which values were missing due to
instrument failure.
Before determining correlations and developing the
regression model, skewness in the datasets was reme-
died by logarithmic transformation. Where necessary,
the initial values were increased by 1, to avoid transfor-
ming values of zero. Subsequently, correlations between
these transformed parameters were investigated. Para-
meters of which correlation with log (SSC + 1) was
significant at a level of 0.01 were put into a stepwise
linear regression (SPSS, 2003). The final model, with an
R2 of 0.612, took the form of:
logSSC 1 1:183logQ 0:616logQBf 1
0:580logQIf 1
 0:235logQRo 1
 0:830logQp 0:501
in which SSC is the suspended sediment concentration
(in mg/l), and Q is the daily average discharge (in m3/s).
QBf, QIf and QRo are respectively the average daily
baseflow, interflow and run-off (in m3/s), all of which
are explained in Section 3.2.1. Finally, Qp is the daily
average discharge of the previous day.
All parameters, save the two latter, have a positive
correlation with log (SSC + 1); indicating that SSC will
increase with higher discharge values (total discharge as
well as groundwater contributions). The negative corre-
Fig. 5. Factors and fluxes involved in the genesis of authigenic sediment; values for all of these need to be obtained, for estimation of the authigenic
suspended flux.
323
lation of log (QRo + 1), however, indicates that run-off
does not contribute significantly to the delivery of sus-
pended sediment, while the negative correlation of log
(Qp) indicates that high flow conditions the previous day,
will lead to decreasing delivery of sediment the day after.
3.2. Determining authigenic sediment fluxes by
theoretical deduction
Fig. 5 shows the variety of chemical, mineralogical
and hydrological data needed for determination of the
authigenic sediment flux. To estimate the flux of authi-
genic sediment in the river one needs to multiply the
flux of authigenic Fe2+ entering the river with factors
accounting for stoichiometric and sorption corrections.
The authigenic Fe2+ flux can be obtained by multiplying
Fe2+ concentrations in the groundwater with their res-
pective discharge. How estimates of these components
1 were obtained are explained in detail below.
3.2.1. Estimating the annual authigenic Fe2+flux
The Fe2+-flux entering the Kleine Nete upstream of
Grobbendonk can be calculated by multiplying ground-
water discharge by the concentration of Fe2+ therein
present. The groundwater contribution has been deter-
mined by WETSPRO (Willems, 2000), which divides
the measured discharge into a contribution of baseflow,
interflow and run-off. Baseflow is herein defined as the
groundwater flow in the saturated zone, while interflow
represents the groundwater flow in the unsaturated zone
324 E. Vanlierde et al. / Sedimentary Geology 202 (2007) 317332
and run-off is the overland flow (DHI, 2000). This
procedure was previously calibrated for the Nete basin
as part of the project Actualisation of the Sigmaplan
(Swings et al., 2003). In this study, the calibrated
procedure was used to expand the existing time series of
daily base-and interflow data to include the entire period
from 1999 to 2004. By addition of these daily discharge
values, annual values for these parameters were
obtained (Table 1).
Next, Fe2+ concentrations need to be determined for
baseflow and interflow. Vanlierde et al. (2005), based on
various sources and databases (the Belgian Nuclear
Research Centre (Blommaert et al., 1988), PIDPA
(public water supply) (DOV, 2004), AMINAL-Division
Water (DOV, 2004), Institute of Nature Conservation
(INBO, 2004)), showed that total Fe concentrations in
the groundwater vary strongly throughout the region
and display even local differences. Table 2 gives an
overview of this variability as well as values for pH and
redox. To account for this variation, concentrations for
baseflow and interflow were obtained from median val-
ues of respectively the Formation of Diest (15.68 mg/l)
and the upper phreatic layer (15 to 25 mg/l). The values
of total Fe concentration can be used to estimate Fe2+
concentrations; because of the reigning pH and redox
conditions, all iron present will be in its reduced state.
These median concentration values were then multiplied
by their respective flow components and results were
added to obtain the total Fe2+ flux entering the river.
3.2.2. Calculating the authigenic sediment flux
As noted above, the Fe2+ entering the river will be-
come chemically unstable and will oxidize due to chang-
ing environmental conditions (pH and redox conditions).
However, the mass of authigenic iron compounds
generated by the same amount of dissolved Fe2+ can
vary significantly, depending on the wide variety of
mineralogical forms that can occur in freshwater (Stumm
and Morgan, 1970) and the sorption of the precipitate
formed. As Vance (2002) stated, ferric hydroxide (Fe
(OH)3) is the direct result of ferrous iron oxidation and
precipitation. In time, ferric hydroxide is mineralized.
The principal forms of mineralized ferric iron found in
soils are amorphous hydrous ferric oxide (Fe2O3XH2O),
maghemite (-Fe2O3), lepidocrocite (-FeOOH), hema-
tite (-Fe2O3), and goethite (-FeOOH), of which the
latter two are definitely not present in the Kleine Nete, as
determined by Mssbauer spectroscopy. Iron phosphates
however, are abundantly present in the Nete basin.
Measurements by the Flemish Environment Agency in
the period 19992004 (VMM, 2006) in the Kleine Nete
at Grobbendonk demonstrated that phosphorus contents
reached high values (see Table 2). However, Fe concen-
trations were even more abundantly present, indicating
that most available phosphate is bound to iron, thus
explaining the relative low concentrations of soluble
phosphates in the surface water (b 0.3 mg P/l) as men-
tioned in AMINAL (2003). The high phosphate contents
measured in suspended sediment are caused by intensive
agricultural activity as well as by the high population
density in the basin. Therefore, to obtain insight into
the authigenic suspended sediment flux (ASSF), the
Fe2+-flux needs to be corrected.
Stoichiometrically, 1 g of Fe2+ can form iron com-
pounds with a mass ranging from 1.5 (ferrihydrite) up to
3.8 g (mitridatite). This range will be used as the range
for the stoichiometric correction factor (FSt).
Besides the stoichiometric factor, another correction
is needed to obtain the authigenic SSF. Colloidal and
flocculated iron compounds will sorb cations, which
will add to the mass of the authigenic precipitates; how-
ever, this addition in mass is likely to be substantially
smaller than the increase due to stoichiometric factors.
As a rule of thumb, Vance (2002) indicates that amor-
phous hydro ferric oxide is capable of adsorbing
0.5 mmol of ionic material per gram. Clarck (2005),
however, states that the cation exchange capacity of
ferric iron oxyhydroxides (ferrihydrite) is even greater
than that of smectite clays, with values often exceeding 5
to 7 mmol/g. However, this is the case only at neutral to
high pH. Measured pH values in the Kleine Nete fall
within this range.
Depending on which ion is being adsorbed and
which compound is being formed, the weight of the
sediment will thus increase in mass respectively by 0.1
up to 1 g per gram sediment. Therefore, a theoretical
range of 1.05 up to 2 was considered as sorption correc-
tion factor (FSo).
To obtain a total correction factor (FT), both correc-
tion factors (FSo and FSt) must be multiplied. However,
to obtain the theoretical interval of FT, one should take
notice that high sorption capability is mostly confined to
hydrous ferric oxide, and much less to compounds with
higher stoichiometric values, as the former is usually
present in colloidal form and therefore has greater
specific surface area, leading to a higher sorption coef-
ficient (Stumm and Morgan, 1970; Lijklema, 1979).
Therefore, the upper boundary FT-value of 4.00 is
obtained by multiplying the highest FSt-value (3.8) with
the lowest FSo-value (1.05), while the lower boundary
value of FT, 1.57, is obtained by multiplying the lowest
FSt-value (1.5) also with the lowest FSo-value (1.05).
This does, however, not include any increase in mass
due to microbiological activity.
 E. Vanlierde et al. / Sedimentary Geology 202 (2007) 317332 325

To procure the broadest theoretical range in annually
produced authigenic sediment, the lowest value of total
Fe2+ flux is multiplied with the lowest value of FT, and
the highest value of total Fe2+ flux is multiplied with the
highest value of FT. Results can be seen in Table 4 and
are explained in Section 5.
3.3. Determining authigenic suspended sediment fluxes
by MARS (1.0)
MARS (Model for Authigenic River Sediment)
version 1.0 is a numerical model, developed by Flanders
Hydraulics Research in a Microsoft Excel environment
in which it also generates its output. MARS predicts the
flux of authigenic sediment, produced by groundwater
seepage. MARS takes into account site-specific settling,
erosion and accumulation processes by using algorithms
that intend to approximate the observed complexity in
hydraulic, physical, chemical and biological processes
in the river. These algorithms are empirically derived
and optimized by fitting the modeled ASSF's onto the
total SSF's.
MARS starts by calculating the authigenic sediment
flux entering the river through groundwater seepage
(ASFSeep) by multiplying the flux of Fe2+ entering the
river with factors accounting for stoichiometric and sorp-
tion corrections, as explained in Section 3.2.2. Mathe-
matically this is expressed as:
ASFSeep QBf d Fe2
Bf QIf d FeIf d FSt d FSo
2
2 idation algorithm (5) has an exponential form, derived
2+
in which FeBf and FeIf2+ respectively represent the Fe2+
concentrations in the baseflow and interflow.
Additionally, MARS uses three separate algorithms to
approximate the observed sediment transport patterns in
the river. The first algorithm, the settling algorithm,
determines the percentage of authigenic suspended sedi-
ment, delivered by groundwater (ASFSeep) that will settle
on the river bed due to low flow conditions, flocculation,
presence of rooted water plants and other factors capable
of influencing the settling behaviour of sediment. It also
determines how much of the ASFSeep will remain in
suspension. These fluxes are calculated as follows:
ASFSus ASFSeep d ad emQ  3
and
X
t
ASFSet ASFSeep d ad emQ  4
i1
in which ASFSus is the part of the ASFSeep that remains in
suspension and ASFSet is the flux that will settle on the
riverbed. Coefficients a and m are dimensionless con-
stants and t is the time period of interest. The expression
ad emQ represents the fraction of the authigenic sedi-
ment that settles out.
The second algorithm models the effect of consol-
idation processes and the residence time on the settled
sediment. The longer sediment remains on the riverbed,
the more it will consolidate, which will, in turn, demand
higher flow conditions for remobilization. The consol-
from the exponential functions describing the consoli-
dation process as noted by De Smedt et al. (1989). It
determines how much of the settled authigenic material

Table 4
Overview of the total SSFes, at Grobbendonk on the Kleine Nete, as well as the deduced and modeled ASSFes and their relative contribution to the
total SSF
Measured SSF ASSFes by theoretical deduction ASSFes by modeling with MARS
Total SSF (t) Potential range of Potential ASSF (t) Contribution of
ASSF (t) contribution ASSF to the total
of ASSF to the SSF (%)
total SSF (%)
Min Max Min Max
1999 14 775 4 774 13 922 36 100 a 10 061 68
2000 11 220 5 440 15 777 46 100 a 10 763 96
2001 15 624 5 882 17 223 45 100 a 14 148 91
2002 15 482 5 358 15 544 41 100 a 11 660 75
2003 7274 3 643 10 561 41 100 a 4 230 58
2004 6314 4 109 11 968 52 100 a 3 934 62
Average 11 782 4 868 14 166 43 100 a 9 133 75
a
The maximum values of authigenic sediment fluxes are higher than the measured SSFes, which leads to values of contribution over 100%. As this
is physically impossible, the maximum authigenic contribution is considered to be 100%.
326 E. Vanlierde et al. / Sedimentary Geology 202 (2007) 317332
can be resuspended, given sufficient flow conditions,
based on the amount of accumulated sediment.
ASFrCons bd enASFSet 5 1999
in which ASFrCons represents the dimensionless fraction
of accumulated authigenic sediment that can be resus-
pended given sufficient flow conditions, as determined
in the resuspension algorithm (6) and b and n are
dimensionless constants.
The final algorithm, the resuspension algorithm (6),
calculates the percentage of the available settled and
accumulated authigenic sediment (ASFCons) that could
be eroded and resuspended. It takes both the available
amount of sediment into account and the discharge-
dependent specific erodibility of the river. Again, it
takes the form of an exponential function of the mea-
sured discharge, which is consistent with the observed
exponential relation between water discharge and sedi-
ment flux at the Grobbendonk site (FHR, 2006).
ASFrResus cd erQ d ASFrCons 6 due to the fact that 2003 and 2004 were less humid than
in which ASFrResus is the dimensionless fraction of
authigenic settled sediment which is being resuspended.
Coefficient c and r are dimensionless constants and the
expression cderQ represents the dimensionless erodibil-
ity of the river as a function of discharge.
The actual flux of resuspended authigenic sediment
(ASFResus) can be obtained by multiplying the fraction
(ASFrResus) by the total accumulated authigenic sedi-
ment available (ASFSet):
ASFResus ASFrResus d ASFSet 7 authigenic flux values were higher than the measured
Finally, MARS calculates the total authigenic sus-
pended sediment flux (ASSF):
ASSF ASFSus ASFResus 8 that the contribution of authigenic sediment is substan-
In all three algorithms the coefficients and constants
were obtained by modeling and fitting the ASSF to the
observed total suspended sediment flux.
4. Results and discussion
In Table 5 the annual total measured SSF's for the
period 19992004 are summarized, ranging between
6300 and 15,600 t/year, which would correspond to 0.11
and 0.26 t/year/ha. The average value for this period
accounts to 11800 t/year, or 0.2 t/year/ha, while the
average soil loss in the Nete Basin is estimated to be
Table 5
Total SSF's determined by combination of measurements and
regression model values and the number of values of both sources
2000 2001 2002 2003 2004
Total SSF (t/year) 14 775 11 220 15 624 15 482 7274 6314
Number of daily SSC 324 62 0 142 306 360
values measured
Number of daily SSC 41 304 365 223 59 6
values by regression
model
0.12 t/year/ha (CIW, 2006). Table 5 also shows how
many of the SSC values, used to determine this flux,
were actual measurements, and how many of them were
regression model estimates (as explained in Section
3.1). Although total SSF's of 2000, 2001 and even 2002
should be interpreted cautiously, because of their
extensive use of regression model derivations and
therefore greater uncertainty, it is still obvious that the
2003 and 2004 SSF's are less than half of what was
annually transported in 1999 to 2002. This is most likely
the previous years; 2003 was even a dry year, as Van den
Hurk et al. (2005) confirm. This resulted in low surface
flow discharge values as can be seen in Table 1. This has
caused a lower production of authigenic sediment, and
also decreased the supply of detrital sediment.
In Table 4, the results of the theoretically calculated
annual authigenic fluxes and their relative contribution
to the total SSF are summarized. The average authigenic
flux contribution is predicted to range between 43% and
100% of the total SSF for the 19992004 period in the
Kleine Nete at Grobbendonk. However, the maximum
SSFes, which indicates that possibly not all authigeni-
cally produced sediment is annually transported, but can
also be partially accumulated on the river bed or be
transported as bedload. Nonetheless, it can be concluded
tial in this basin.
Table 5 also summarizes MARS estimates for the
contribution of authigenic SSF to the total suspended
load. These range from 58% to 96%, and thus all lie
within the theoretically determined range. Values for
daily discharges and total mass accumulated on the river
bed (ASFSet) for the period of 19992004 are
presented in Fig. 8. The accumulated sediment is stored
on the riverbed and can be remobilized, given suffi-
ciently high flow conditions. These conditions tend to
be met in winter, as is shown by the decreases in
accumulation during the winters of the first three years.
However, from spring 2002 onward, the accumulation,
 E. Vanlierde et al. / Sedimentary Geology 202 (2007) 317332
Fig. 6. Daily discharge values in m3/s and total accumulated authigenic sediment (ASFSet) data in 106 t for the period 19992005 at Grobbendonk,
Kleine Nete basin.
as predicted by MARS, has been continuously rising
with no substantial decrease up to 2004, which suggests
sediment consolidation as a reason for lower SSF's in
2003 and 2004 (Fig. 6).
Consolidation would also explain the seeming incon-
sistency of the lower SSF in 2004 compared to that
observed in 2003, even though precipitation and
discharge values were higher in 2004. During the very
dry summer of 2003, sediment was consolidated on the
riverbed, possibly fixed by water plants, which disabled
remobilization, except in the case of high discharge
conditions. These conditions were not met during the
winter period of 2003, nor during 2004, which is
explained in Fig. 7. This figure shows the daily water
discharges measured at Grobbendonk for the years 1999,
2000, 2003 and 2004. Additionally, the minimal,
maximum and average values of daily discharge
measured over the time period 19832004, were also
plotted. It is important to note that the highest discharge
values for the years 2000 and 2004 (Fig. 7a) were almost
of the same magnitude, but that, nevertheless, signifi-
cantly less sediment was transported in 2004 compared to
2000 (see Table 5). This can be explained by looking at
the discharge conditions of the preceding years (Fig. 7b).
2003 was a dry year, with discharge values 84% of the
time lower than the average discharge value that day for
the period 19832004, which allowed consolidation
processes to help immobilize accumulated sediment on
the river bed, thus prohibiting resuspension in 2004.
1999, however, was characterized by more frequent
and higher discharge peaks, one of which occurred after
327
the dry summer period. This ensured that in 1999 little
sediment was left to accumulate on the river bed,
therefore being available for transportation in 2000;
whereas in 2003 consolidation processes helped immo-
bilize accumulated sediment on the river bed, thus
prohibiting resuspension in 2004.
Fig. 8 shows that for the time period investigated, the
increase in accumulation does not seem to significantly
influence the predictive power of the model. Fig. 8a
depicts total and authigenic SSF's during a period in
which accumulation increase is followed by a distinct
decrease, while Fig. 8b shows a period characterized by
a fairly constant increase, with only a slight decrease
during the highest discharge peaks. Both figures show
that the ASSF peaks coincide with the total SSF peaks
and tend to underestimate the total SSF. This is very
likely, as the authigenic contribution should not be
100% or more. Although the predictive power of MARS
1.0 is not significantly affected by the constant accu-
mulation of sediment on the riverbed, this might cause
problems for longer time series.
Earlier authigenic modeling attempts (Vanlierde et al.,
2005), were limited to the period February 1999 up to
January 2000, for which an average authigenic contri-
bution of 70% was predicted. This corroborates predic-
tions of MARS (1.0) for the same period (i.e. 66%).
Authigenic contributions have a strong positive
correlation with annual discharge. Whether this is due
to the remobilization of the previously deposited
authigenic sediment during high flow conditions, or to
the fact that, because of the low erosion rate in the basin,
328 E. Vanlierde et al. / Sedimentary Geology 202 (2007) 317332

Fig. 7. Daily water discharges measured at Grobbendonk (Kleine Nete) for the years 2000 and 2004 (a) and 1999 and 2003 (b), plotted against the
minimal, maximal and average daily discharges for the period 19832004.

detrital sources relatively contribute less during these
conditions, or is caused by other phenomena is currently
being researched.
Although the results from MARS (1.0) seem to
confirm the assumption that Fe2+ in groundwater sig-
nificantly contributes to the suspended sediment flux in
the Kleine Nete, as measured in Grobbendonk, a vali-
dation of the model remains necessary. This validation
process is under development and will be based on
following principles: (1) determining Fe2+ and Fe3+
concentrations in representative suspended sediment
samples as well as bed material, in different tributaries
of the Kleine Nete, while also determining organic
matter concentrations as well as percentages of the
detrital fraction. Samples will be taken under different
discharge conditions. This will give an indication of
the variability of the Fe-content in the sediment as
well as of the contribution of detrital sediment to the
total flux. (2) As the flocs found in the Kleine Nete
are amorphic, classical methods of determining miner-
alogical nature of the sediment, such as XRD analysis
are not applicable. Therefore, the mineralogical struc-
ture has to be derived from chemical analyses of the
material, which is not a straightforward issue in
 E. Vanlierde et al. / Sedimentary Geology 202 (2007) 317332 329

Fig. 8. Total suspended sediment fluxes (SSFes), authigenic suspended sediment fluxes (ASSFes), and total accumulated authigenic sediment
(ASFSet), at Grobbendonk, Kleine Nete basin, for (a) the period 01 January 199931 March 2000 and (b) the period 1 August 200231 July 2003.

combination with electron microscopy. As XRF or ICP
analyses of sediments demand an extensive calibration,
preliminary tests are being conducted with spectropho-
tometry to determine iron concentrations (Golterman,
2004). The next step is to perform sequential extraction
tests, to regain insight in the mineralogical structure
of the iron flocs. (3) A sensitivity analysis will be
performed to estimate the impact of varying the different
parameters.
Consequently, MARS will be subjected to validation
by means of comparing the Fe values measured in the
sediment and by the percentages of total suspended
sediment it represents.
The question remains whether authigenic sedimen-
tation is an important process in basins which attract
their sediment from upstream erosion. This case study
in the Kleine Nete seems to indicate that authigenic
sedimentation produces substantial amounts of sediment
in basins characterized by certain geological and
hydrogeological conditions, independent from whatever
other sources of sediment are present in the basin.
Furthermore, even when quantitatively the authigenic
sediment contribution might not be significant in a
basin, it can have a substantial impact on the sediment
and water quality, as iron colloids, hydroxides and flocs
have sorption capacities of the same magnitude as clays
or higher (Stumm and Morgan, 1970). Therefore, it can
be concluded that when regional settings allow for the
formation of authigenic sediment, this process should be
taken into account.
330 E. Vanlierde et al. / Sedimentary Geology 202 (2007) 317332

5. Conclusions Appendix A

In the Kleine Nete basin, the annual suspended Abbreviations

sediment flux at the Grobbendonk monitoring station ASFResus Resuspended Authigenic Sediment Flux
was estimated to range between 6300 and 15,600 t/year, ASFSeep Authigenic Sediment Flux entering the river
with an average of 11,800 t/year for the period 1999 through groundwater Seepage
2004. These results are based on actual measurements ASFSet Part of the ASFSeep that settles on the river bed
and a site-specific regression model. ASFSus Part of the ASFSeep that remains in suspension
Groundwater inflow in the Nete basin substantially ASFrCons Fraction of accumulated and Consolidated
contributes to the surface water discharge. Based Authigenic Sediment that can resuspend given
on groundwater separation techniques in WETSPRO sufficient flow conditions
(Willems, 2000), groundwater accounts on average for ASFrResus Fraction of Authigenic settled Sediment
87% of the surface water flow upstream Grobbendonk, which is being resuspended
for the period-of-record. ASSF Authigenic Suspended Sediment Flux
Furthermore, this groundwater discharge at Grob- FEr Erosion Factor
bendonk is rich in Fe2+ due to lowered pHs and reducing FRes Residence Factor
conditions within the aquifer; when the reduced FSo Sorption Correction Factor, parameters used in
groundwater enters the surface water system, substantial MARS
quantities of authigenic iron minerals precipitate and FSet Settling Factor
flocculate. Theoretical calculations indicate that the FSt Stoichiometric Correction Factor, parameters
authigenic precipitates could, on average, account for 43 used in MARS
up to 100% of the suspended sediment fluxes, for the FT Total Correction Factor = combination of FSt
19992004 period. The best current estimate for annual and FSo
2+
authigenic mineral contributions to the suspended FeBf Fe2+ concentrations present in the baseflow
2+
sediment flux at the site for the period-of-record is FeIf Fe2+ concentrations present in the interflow
modeled by a new model called MARS (Model for mAcc total mass of authigenic sediment accumulated
Authigenic River Sediment), version 1.0. Its predictions on the river bed (ton)
of contribution range from 58 until 96%, with an MARS Model for Authigenic River Sediment
average of 75% for all 5 years. SSC Suspended Sediment Concentration
Also, the positive correlation between annual dis- SSF Suspended Sediment Flux
charge values and authigenic suspended sediment fluxes Q Daily average water discharge
indicates that due to an increase of surface water QBf Baseflow; Groundwater flow in the saturated
discharge, sources other than authigenic precipitates will zone, contributing to the total water discharge
relatively decrease their contribution. QIf Interflow; Groundwater flow in the unsaturated
Finally, the decrease of total suspended sediment zone, contributing to the total water discharge
flux in 2003 and 2004, equaling a decrease of more Qp Daily average water discharge value of the
than 50% compared to the previous years, is most previous day
likely due to consolidation of sediment deposited on QRo Run-off; Overland run-off, contributing to the
the riverbed. total water discharge

Acknowledgements References

We would like to thank Jan Ronsyn and Maarten
Deschamps for their help in filtering the discharge data
and Prof. Dr. Peter Van den Haute for the use of his
microscopes. Also, we would like to thank Art Horowitz
for his continuous support and knowledge. Furthermore,
we would like to thank the field personnel of Flanders
Hydraulics Research for their continuous efforts to
collect quality data. Finally, our thanks goes to Petra
Vijncke and Bart Pannemans. Without them, this article
would never have existed in this form.
AMINAL Division Water, 1995. Grondwaterkwaliteit in de provincie
Antwerpen. Report for the Flemish Government, Belgium.
AMINAL Division Water, 2003. Bekkenbeheerplan van de Nete:
Samenvatting omgevingsanalyse Ontwerpvisie. Report for the
Flemish Government, p. 53. Online available at: https://circa.
vlaanderen.be/Public/irc/mmis/bekken/library?l=/netebekken/
bekkenbeheerplan/omgevingsanalyse&vm=detailed&sb=Title
(accessed 15 August 2006).
AMINAL Division Water, 2004. Omgevingsanalyse in het kader
van de opmaak van de bekkenbeheerplannen, Bekken van de
Demer, Perceel 7. Thema Sedimenttransport report, for the Flemish
Government, p. 67. Online available at: https://circa.vlaanderen.
 E. Vanlierde et al. / Sedimentary Geology 202 (2007) 317332
be/Public/irc/mmis/bekken/library?l=/demerbekken/bekkenbe-
heerplan/omgevingsanalyse/samenvatting&vm=detailed&sb=Title
(accessed on 25 May 2004).
Asselman, N.E.M., 2000. Fitting and interpretation of rating curves.
Journal of Hydrology 234, 228248.
Batelaan, H., 2006. Phreatology, Characterizing groundwater recharge
and discharge using remote sensing, GIS, ecology, hydrochemistry
and groundwater modelling, PhD Thesis, Vrije Universiteit
Brussel, Belgium.
Belien, H., 2006. Studie van geflocculeerd ijzerhoudend sediment in
de Kleine Nete, unpublished Msc thesis, Ghent University,
Belgium, pp. 103.
Belien, H., Vanlierde, E., Mostaert, F., Jacobs, P., 2006. A preliminary
study of the flocculation of iron-bound sediment in a Belgian river.
Geophysical Research Abstracts, vol. 8, 04478. EGU General
Assembly, Vienna, Austria. 02-07 April.
Blommaert, W., de Marmol, P., Patyn, J., 1988. Hydrochemical
Research within the Hades Project (Disposal of high radioactive
waste into the Boom clay), Studiecentrum voor Kernenergie.
Flemish Government, Belgium.
CIW, 2006. Het Bekkenbeheerplan van het Netebekken. Flemish
Government, p. 202. Online available at:http://www.volvanwater.
be/deel-bekkenbeheerplannen/bekkens.html/nete.html/deel-bekken-
beheerplannen/BBP_Nete_ontwerpvOO_03_11_06.pdf, (accessed
30 November 2006).
Chen, M.S., Wartel, S., van Eck, B., van Maldegem, D.C., 2005.
Suspended matter in the Scheldt estuary. Hydrobiologia 540,
79104.
Clarck, I., 2005. Chapter 8: Sorption and Ion Exchange Reactions, in
course: Groundwater Geochemistry. University of Ottawa,
Canada. Online available at: http://www.science.uottawa.ca/est/
eng/prof/clark/GEO4342/Ion%20Exchange%20Reactions.pdf,
(accessed 16 December 2005).
Collins, A.L., Walling, D.E., Leeks, G.J.L., 1997. Source type
ascription for fluvial suspended sediment based on a quantitative
composite fingerprinting technique. Catena 29, 127.
DOV, 2004. Databank Ondergrond Vlaanderen. Database with
geological, geotechnical and hydrogeological data Online avail-
able at: http://dov.vlaanderen.be, (accessed on 15 August 2004).
De Schutter, G., Matheeussen, C., Tyberghein, J., Alwan, A., 1995.
Grondwaterkwaliteit in de Provincie Antwerpen. AMINAL in
samenwerking met Provinciebestuur Antwerpen, report for the
Flemish Government.
De Smedt, F., Muls, J., Sas, M., Smits, J., 1989. Een computermodel voor
de berekening van de consolidatie van slib. Water 47, 140143.
De Vries, A., Klavers, H.C., 1994. Riverine fluxes of pollutants:
monitoring strategy first, calculation method second. European
Journal of Water Pollution Control 4, 1217.
DHI, 2000. A Modelling System for Rivers and Channels, Reference
Manual. DHI Water & Environment, Hrsholm, Danmark, p. 514.
Droppo, I.G., Leppard, G.G., Flannigan, D.T., Liss, S.N., 1997. The
freshwater floc: a functional relationship of water and organic and
inorganic floc constituents affecting suspended matter properties.
Water, Air and Soil Pollution 99, 4354.
Droppo, I.G., Walling, D.E., Ongley, E.D., 2000. The influence of floc
size, density and porosity on sediment and contaminant transport.
In: Stone, M. (Ed.), The role of erosion and sediment transport in
nutrient and contaminant transfer. Publication, vol. 263. IAHS,
Waterloo, Canada, pp. 141147.
Droppo, I.G., Leppard, G.G., Liss, S.N., Milligan, T.G., 2005.
Flocculation in Natural and Engineered Environmental Systems.
CRC Press, p. 188.
331
Envico, 2001. Uitwerking van thema's ter invulling van de omgevingsa-
nalyse in het kader van de opmaak van waterhuishoudingsplannen en
bekkenbeheerplannen voor een aantal pilootprojecten- Bekken van
de Nete, Report ed. by W. Mennen. AMINAL- Division Water.
Flemish Government, Belgium.
FHR, 2006. HYDRA Database, online partly available at: http://www.
lin.vlaanderen.be/awz/waterstanden/hydra, and available on re-
quest hic@vlaanderen.be, Flanders Hydraulics Research, Depart-
ment of Mobility and Public Works, Flemish Government.
Golterman, H.L., 2004. The Chemistry of Phosphate and Nitrogen
Compounds in Sediments. Kluwer Academic Press, Dordrecht,
p. 251.
Govoreanu, R., 2004. Activated sludge flocculation dynamics: on-line
measurement methodology and modelling. Doctoral Thesis, Faculty
of Bioscience Engineering. Ghent University, pp. 290.
Horowitz, A.J., Elrick, K.A., Smith, J.J., 2001. Estimating suspended
sediment and trace element fluxes in large river basins: methodological
considerations as applied to the NASQAN programme. Hydrological
Processes 15, 11071132.
INBO (Instituut voor Natuurbehoud), 2004. WATINA. Available on
request at: http://www.instnat.be, (accessed 6 July 2004).
Jacobs, P., Louwye, S., Polfliet, T., Adams, R., Vermeire, S., De Moor,
G., 2001. Quartairgeologische Kaart van Belgi, Vlaams Gewest.
Kaart en Toelichtingen bij het Kaartblad Antwerpen (15). Afdeling
Natuurlijke Rijkdommen en Energie, pp. 163.
Lebbe, L.C., 1999. Hydraulic Parameter Identification: Method for
Single and Multiple Pumping Tests. Springer, Berlin, p. 359.
Li, D.H., Ganczarczyk, J.J., 1987. Stroboscopic determination of
settling velocity, size and porosity of activated sludge flocs. Water
Research 21, 257262.
Librecht, I., Van Orshoven, J., 2003. Aandachtszones voor erosiebes-
trijding: de bodemkundige erosiegevoeligheidskaart. Report to the
Land Division of the Flemish Government.
Lijklema, L., 1979. Binding van o-fosfaat door ijzer (III)-en aluminium-
hydroxiden; theorie en praktische betekenis. H2O 12 (23), 511513.
Mikes, D., Verney, R., Lafite, R., Belorgey, M., 2004. Controlling
factors in estuarine flocculation processes, experimental results
with material from the Seine estuary. Journal of Coastal Research
SI (41), 8289.
Morgan, R.P.C., Quinton, J.N., Smith, R.E., Govers, G., Poesen, J.W.A.,
Auerswald, K., Chisci, G., Torri, D., Styczen, M.E., 1998. The
European Soil Erosion Model (EUROSEM): a dynamic approach for
predicting sediment transport from fields and small catchments. Earth
Surface Processes and Landforms, vol. 23, Issue. 6. John Wiley &
Sons, Ltd., pp. 27544.
Nathan, R.J., McMahon, T.A., 1990. Estimating low-flow character-
istics in ungauged catchments. Water Research Management 6,
85100.
Nopens, I., 2005. Modelling the activated sludge flocculation process:
a population balance approach. Doctoral Thesis. Faculty of
Bioscience Engineering. Ghent University, pp. 293.
Philips, J.M., Webb, B.W., Walling, D.E., Leeks, G.J.L., 1999.
Estimating the suspended sediment loads of rivers in the LOIS
study area using infrequent samples. Hydrological Processes 13,
10351050.
Schiltz, M., Vandenberghe, N., Gullentops, F., 1993. Toelichting bij de
Geologische kaart van Belgie, Vlaams Gewest, Kaartblad (16),
Lier, Schaal 1:50000. Technical Report, Belgische Geologische
Dienst, Bestuur Natuurlijke Rijkdommen en Energie.
Sloto, R.A., Crouse, M.Y., 1996. HYSEP: A computer program for
streamflow hydrograph separartion and analysis. Tech. Rep. Water-
Resources Investigations report 96-4040. U.S. Geological Survey.
332 E. Vanlierde et al. / Sedimentary Geology 202 (2007) 317332
Soresma, 2002. Eindrapport: Invulling van de thema's voor de
uitvoering van de omgevindgsanalyse in het kader van de opmaak
van het bekkenbeheerplan van de Nete en het aterhuishoudingsplan
van de Grote Nete, Perceel 5. Report for the Flemish Government.
SPSS, 2003. SPSS 12.0 for Windows, Copyright SPSS Inc., 19892003.
Stumm, W., Morgan, J.J., 1970. Aquatic Chemistry, An Introduction
Emphasizing Chemical Equilibria in Natural Waters. Wiley-
Interscience, New York, USA.
Swings, J., Boeye, D., 1997. Ecohydraulische studie van het Olens
Broek. Universiteit Antwerpen, studie uitgevoerd in opdracht van
AMINAL Afdeling Natuur (AMINAL/NAT 1996/nr13), p. 72.
Swings, J., Vanderkimpen, P., Smets, S., Blankaert, J., 2003. Actualisatie
Sigmaplan-Deelopdracht 3: Hydrologische en hydraulische mod-
ellen Volume 1b: Statistiek/Hydrologie Rupelbekken, Version 2.1,
AWZ - Sea Scheldt Division. Flemish Government, Belgium.
Vance, D.B., 2002. Iron The environmental impact of a universal
element. Online available at: http://2the4.net/iron.htm, (accessed 4
December 2005).
Van den Hurk, B., Beersla, J., Keller, F., 2005. Klimaatscenario's van
watertekorten. In course: Inzicht in watertekorten. KNMI. November.
Van Der Beken, A., Huybrechts, W., 1990. De Waterbalans van het
Vlaams Gewest, Een rationeel Waterbeheer via kennis van de
Waterbalans. Water 50, 8892.
Van Leussen, W., 1994. Estuarine macroflocs, PhD Thesis, University
of Utrecht.
Vanlierde, E., de Schutter, J., Meys, J.F.A., Mostaert, F., Jacobs, P.,
2005. Contributions of authigenic iron compounds to fluvial
suspended sediment concentrations and fluxes in the Nete sub-
basin, Belgium. In: Walling, D.E., Horowitz, A.J. (Eds.), IAHS
publication, vol. 291, pp. 5463. Sediment budgets 1.
Van Rompaey, A., Krasa, J., Dostal, T., Govers, G., 2003. Modelling
sediment supply to rivers and reservoirs in Eastern Europe during
and after the collectivisation period. Hydrobiologia: The Interac-
tions between Sediments and Water 494 (13), 169176.
VLM, 1997. Fosfaatverzadiging van zandige bodems in Vlaanderen.
Flemish Land Agency Report. Ministry of the Flemish Community,
Belgium.
VLM, 2006. Database physico-chemical measuring network. Online
available at: http://www.vmm.be, (accessed 15 August 2006).
Walling, D.E., Webb, B.W., 1981. The reliability of suspended sediment
load data. Erosion and Sediment Transport Measurement. IAHS
Publication, vol. 133. IAHS Press, Wallingford, UK, pp. 177194.
Walling, D.E., Owens, P.N., Leeks, G.J.L., 1999. Fingerprinting
suspended sediment sources in the catchment of the River Ouse,
Yorkshire, UK. Hydrological Processes 13, 955975.
Willems, P., 2000, Probabilistic immission modeling of receiving
surface waters, PhD thesis, Katholieke Universiteit Leuven,
Faculty of Engineering, Leuven, Belgium.
Wittenberg, H., 1999. Baseflow recession and recharge as nonlinear
storage processes. Hydrological Processes 13, 715726.