The effects of changes to estuarine hydrology on system phosphorous

retention capacity: The Mondego estuary, Portugal

A.I. Lilleb
a,
*
, M. Otero
b
, J.P. Coelho
b
, E.T. Rodrigues
c
, M.E. Pereira
b
, A.C. Duarte
b
,
M.A. Pardal
d
, M.R. Flindt
e
a
CESAM - Centre for Environment and Marine Studies, Department of Biology, University of Aveiro, Campus Universitrio de Santiago, 3810-193 Aveiro, Portugal
b
CESAM - Centre for Environment and Marine Studies, Department of Chemistry, University of Aveiro, Campus Universitrio de Santiago, 3810-193 Aveiro, Portugal
c
IMAR - CMA (Marine Environmental Research Centre), Department of Life Sciences, University of Coimbra, Apartado 3046, 3001-401 Coimbra, Portugal
d
CFE - Centre for Functional Ecology, Department of Life Sciences, University of Coimbra, Apartado 3046, 3001-401 Coimbra, Portugal
e
Centre of Environmental Technology, Institute of Biology, University of Southern Denmark, Campusvej 55, 5230 Odense M, Denmark
a r t i c l e i n f o
Article history:
Received 7 July 2011
Accepted 13 December 2011
Available online 8 January 2012
Keywords:
phosphorus
sorption
estuary management
soils
sediments
eutrophication
a b s t r a c t
The Mondego estuary is a mainly polyhaline estuary in central Portugal in which eutrophication
increased during the last decades of the 20th century. In 1998 the system hydrology was changed, aiming
to reverse the eutrophication process. A long environmental monitoring database showed that the mean
concentrations of dissolved inorganic phosphorus (DIP) increased by 50%, due to the increase in winter
concentrations, i.e., in periods characterized by lower temperature and lower salinity. Given existing
background knowledge, especially on the system hydrodynamics, phosphorus dynamics and phosphorus
speciation in the estuary, we framed the hypothesis that the signicant increase of DIP could be related
to different sorption capacities of the sediments. The results highlighted two scenarios: i) Before 1998 the
nutrient-rich freshwater input from the upstream cultivated lands entered the system through the area
with the highest PO
4
eP adsorption capacity (Q
*
657 mg P g
1
wwt), thus the PO
4
eP availability in the
water column was mostly dependent on the mineralization processes, which is in agreement with
previous ndings; ii) After 1998, the water residence time diminished from moderate (weeks) to short
(days). This change coincided with a diversion of the water to an area with a much lower PO
4
eP
adsorption capacity (Q
*
410 mg P g
1
wwt), which represented a decrease in 7.3% of the system
adsorption capacity. This means that sediments were not able to adsorb much of the PO
4
eP and a higher
equilibrium concentration occurred in the water column. The sorption study proved to be a valuable tool
in testing our hypothesis and provided essential information on the mobility of PO
4
eP from soils/sedi-
ments to the water column.
2011 Elsevier Ltd. All rights reserved.
1. Introduction
Eutrophication is one of the major water quality problems that
must be faced by water authorities worldwide, and it can be highly
dependent on dissolved inorganic phosphorus (DIP, o-phosphate,
PO
4
eP) dynamics and uxes, both in fresh water systems
(Bechmann et al., 2005; Carpenter, 2005; Garnier et al., 2005;
Heredia and Fernandez-Cirelli, 2007) and in estuaries (Pinckney
et al., 2001; de Jonge et al., 2002; Coelho et al., 2004). In estua-
rine systems DIP can increase considerably during spring/summer,
while spatially DIP may be highly variable and dependent on the
river loads, the physicochemical and the biological conditions in
the water and bottom sediments (de Jonge et al., 2002). The excess
of o-phosphate in aquatic systems is mainly due to exogenous
sources of anthropogenic origin, namely from urban wastewater
treatment plants, from where P has not been removed to be re-
used, and agricultural practices, as a result of the continuous
application of fertilizers (e.g. de Jonge et al., 2002; Carpenter, 2005;
Garnier et al., 2005). Most of the phosphorus added to soil as
fertilizer and manure is rapidly bound by the soil. Sediments
produced by soil erosion may carry a phosphorus load through
rivers, representing also a source of o-phosphate to coastal water
bodies (e.g. Sharpley et al., 1992; van Beusekomand de Jonge, 1998).
Knowing the processes that intensify the transport of phosphorus
to water bodies is of great importance due to the accumulation of o-
phosphate (the bioavailable P) in surface water, resulting in water
quality deterioration (Heredia and Fernandez-Cirelli, 2007).
* Corresponding author.
E-mail addresses: lillebo@ua.pt, lillebo@ci.uc.pt (A.I. Lilleb).
Contents lists available at SciVerse ScienceDirect
Estuarine, Coastal and Shelf Science
j ournal homepage: www. el sevi er. com/ l ocat e/ ecss
0272-7714/$ e see front matter 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.ecss.2011.12.016
Estuarine, Coastal and Shelf Science 99 (2012) 85e94
Sorption studies are a valuable way of obtaining essential
information on the mobility of chemicals either fromwater to soils/
sediments or from soils/sediments to water, namely concerning
phosphorus availability (e.g. Pomeroy et al., 1965; de Jonge and
Engelkes, 1993; de Jonge et al., 1993; Pant and Reddy, 2001; Cyra
et al., 2009). The determination of sorption isotherms by labora-
tory experiments can be a valuable approach to compare the
binding efciency of different matrices (i.e. soils and sediments)
(Barone et al., 1992).
The Mondego estuary (Fig. 1) is a well-documented system
located on the North Atlantic Ocean Ecoregion, where eutrophica-
tion has increased during the last decades of the 20th century
(Lilleb et al., 2005, 2007; Leston et al., 2008). This system is rela-
tively small (3.4 Km
2
) with a catchment area of 6670 km
2
(Flindt
et al., 1997). The estuary receives intensely fertilized agricultural
runoff from about 150 km
2
of local upstream cultivated land,
mainly dominated by corn and rice elds (Lilleb et al., 2005). In
1998, a management programme was carried out and the system
hydrodynamics was altered in order to reverse the eutrophication
process in the most affected area of the estuary. However, over the
last 10 years, and based on the classication proposed by Wasmund
et al. (2001), the estuary has became polytrophic concerning the
concentration of DIP (Lilleb et al., 2007), with the annual mean
concentration increasing from 1.0 mmol P L
1
to 1.5 mmol P L
1
between 1998 and 2008. A previous study by Coelho et al. (2004)
showed that there is not a clear seasonal trend in the total
amounts of exchangeable P or P speciation in supercial sediments,
and that pools of exchangeable P were z30 times higher than the
annual net-export from the estuary (14 t P year
1
), concluding that
the annual P release from the sediment appears small compared to
the pool-size. This background information, namely on the system
hydrodynamics, phosphorus dynamics and phosphorus speciation,
leads us to question if endogenous mineralization processes are still
the main phosphate source to the water column, and if this
signicant increase in the concentration of DIP, namely during
winter, could be related to the shifting of the Mondego estuary
hydrology and to P adsorption processes.
In this study it was hypothesised that following the changes in
the system hydrology, the signicant increase of DIP was related to
differences between sorption capacities of upstream agricultural
soils and downstream estuarine sediments. Thus, sorption experi-
ments were performed in the laboratory to determine and compare
the phosphorus isotherms corresponding to several upstream
agricultural soils and downstream estuarine sediments. The main
aim of this approach was to estimate the Mondego estuary ef-
ciency in adsorbing or releasing phosphorus, taking into account
background information on DIP concentrations and the imple-
mented management actions to reduce eutrophication. The study
took place in late winter (March) for three main reasons: i)
temperature is not sufciently high to enhance the phosphorus
internal loading; ii) DIP concentrations increased for the past 10
years during the winter period, and iii) in the agricultural soils the
growth of the new crops always start in the following spring. The
Mondego estuary was therefore used as a case study to evaluate the
impact of drainage basin land use on the phosphorus dynamics of
transitional waters.
1.1. Background information about the Mondego estuary to frame
the hypothesis
In the estuarine area, the Mondego river divides into a northern
and southern arm, separated by the Murraceira island (Fig. 1). Until
1998, the Mondego river outow occurred mainly via the northern
arm, which is the main navigation channel. The southern arm was
almost silted up in the innermost areas, being characterised by
large areas of intertidal ats exposed during low tide. Water
circulation in the south arm was, until 1998 (Fig. 1A), mainly
dependent on tides and on the freshwater input from the Pranto
river articially controlled by a sluice. At this time the system
showed clear eutrophication symptoms. In 1998, the rst mitiga-
tion measures were implemented, via changes in hydrology. Flow
between the north and the south arm was intensied to increase
circulation and diversion of nutrient-rich freshwater inow, trying
to reverse the eutrophication process in the most affected area of
the estuary e its south arm. After 1998, the Pranto river sluice
opening was effectively minimised to reduce the nutrient loading,
and the water residence time was reduced due to engineering work
in the upstream areas (Fig. 1B). A rice-farmer association controlled
both sluice operations and the local users forumcarried out the rst
mitigationphase in 1998. An evaluation of the systemin 2005 led to
a larger scale intervention under the National Institute for Water
(INAG) and local authorities supervision. Thus, in May 2006 the
upstream connection between the two arms of the estuary was re-
established (Fig. 1C).
A long-term monitoring programme in the Mondego estuary
has been carried out since the early 1990s, within the scope of
European and national projects. The South arm of the Mondego
estuary is characterised in Table 1, including the main changes due
to the implementation of the management programmes.
In order to assess the water quality in the south arm of the
estuary, three distinct areas have been monitored (circle in Fig. 1A)
since 1993. Most of the results from this monitoring have been
published elsewhere (Lilleb et al., 2005, 2007; Leston et al., 2008).
Fig. 2A shows the relation of o-phosphate with water temperature.
For the 1993e1997 period o-phosphate concentrations showed
a highly signicant (P < 0.001) positive linear relationship with
water temperature (R
2
0.44, N 40), suggesting a clear depen-
dency on mineralization processes (Lilleb et al., 2004, 2005). Yet,
after 1998, there is an absence of a signicant positive linear rela-
tion of o-phosphate with temperature (Fig. 2B). DIP can be strongly
dependent on mineralization process (Coelho et al., 2004), and has
been previously simulated by applying a simple dynamic temper-
ature model (Fig. 2C) (Lilleb et al., 2004). This last study also
showed that 43% of dissolved inorganic phosphorus efuxes to the
main water body occurred during low tide and that mud ats
represent 79% of the net internal loading. The south arm of the
Mondego estuary is z2.1 km
2
, with 63% corresponding to intertidal
areas, of which 60% are mud ats. From Fig. 2C data
(mmol P m
2
d
1
) we can calculate the sum of daily uxes for one
year in the considered area as z4.6 t P y
1
. Recalculating for
a complete tidal cycle and for the entire area, the net internal
loading becomes z13.5 t P y
1
. This value corresponds to the
annual net-export from the estuary (14 t P year
1
) calculated by
Flindt et al. (1997). However, after the interventions in the system
hydrodynamics, DIP concentrations became signicantly higher
(Leston et al., 2008). In addition, since 1998, high o-phosphate
concentrations, which used to occur only during spring and
summer, also persist during autumn and winter, under relatively
low temperatures (between 6

C and 12

C), suggesting the exis-
tence of additional sources associated to the freshwater inputs,
through surface and subsurface runoff.
2. Material and methods
2.1. Sample collection and characterization
Three composite agricultural soil samples (each N 7) (A, B and
C) and three composite sediment samples (each N 7) (D, E and F)
were collected in March 2007 at the sites indicated in Fig. 1C. Soil
surface samples were taken from agricultural soils upstream the
A.I. Lilleb et al. / Estuarine, Coastal and Shelf Science 99 (2012) 85e94 86
Fig. 1. A - The Mondego estuary, with the location of the long-term monitoring area (A, B and C) and the management areas; B - Freshwater main point sources before and after
managements; C - Sites where soils (A, B and C) and sediments (D, E and F) were sampled.
A.I. Lilleb et al. / Estuarine, Coastal and Shelf Science 99 (2012) 85e94 87
system: A - from soil used for rice cultivation; B e from unculti-
vated soil; and C - from soil used for maize cultivation. Estuarine
surface sediments were collected during low tide at three sites,
from the innermost part to the outermost part of the system D e
Pranto river sluice; E e Gala bridge; and F e Southern arm. At each
site approximately 500 g of sediments/soil were collected from the
surface (depth 10 cm), and all stones/shells, fauna or vegetation
discarded. Samples were transported in refrigerated boxes to the
laboratory. This depth was chosen based on previous research
(Coelho et al., 2004), which showed that, in agreement with
observations in other estuarine systems, the maximum values for
extractable P pools occurred at surface layers and declined with
depth to stable values. That study indicated that P labile fractions
seemed to be responsible for these surface maxima, while more
refractory pools dominated in deeper sediments. In the laboratory,
soils and sediments were air-dried at room temperature
(20e25

C). Disaggregation was then manually performed before
grinding and sieving the samples to a particle size 1 mm and
homogenised. Soil and sediment sub-samples were analysed for
moisture content (WwteDwt, wet weight minus dry weight, 120

C
for 24 h) and organic matter content as percentage of loss on
ignition (%LOI) (DwteAwt, dry weight minus ashes weight, 500

C
for 4 h) and for ne particles (less than 63 mm) after wet sieving of
the dried sample through a 63 mm mesh size sieve. The soil and
sediment pH was measured in water (pH(H
2
O)) and in calcium
chloride 0.01 mol CaCl
2
L
1
(pH(CaCl
2
)) according with ISO (1994).
Total P content of soils and sediments, were analysed in dried and
homogenised samples after combustion (500

C for 4 h). The ashes
were digested in concentrated hydrochloric acid, then ltered and
total P quantied in the form of PO
4
eP using a standardised
colorimetric ascorbic acid method (Limnologisk Metodik, 1992).
Total Fe and Mn concentrations in soils and sediments were
determined as described by Monterroso et al. (2007) while Fe and
Mn oxyhydroxides were evaluated after a partial extraction using
a hydroxylammonium chloride-acetic acid mixture (HAA),
following the method described in Ramalhosa et al. (2006) and
Monterroso et al. (2007). For the procedure 20 ml of 0.04 mol L
1
hydroxylamine hydrochloride (NH
2
OHCl) in 25% glacial acetic acid
was added to 1 g of dried sediment. The extractions with HAA were
conducted in centrifuge tubes, which were subjected to continuous
stirring for 6 h. After extraction solid-solution separation was made
by centrifuging (B. Braun, model Sigma 4e10) at 6000 rpm for
30 min. The supernatant was ltered through 0.45 mm pore size
Millipore lters and analysed by Flame Atomic Absorption Spec-
trophotometry (FAAS) in a Perkin Elmer model AAnalyst 100. All
the determinations were made on three analytical replicates of
each sample.
2.2. Statistical analysis
A principal components analysis (PCA) was performed using the
PRIMER Version 5 software. Projections considered the principal
components 1 and 2 for environmental variables vectors (sedi-
ments/soils physicochemical properties, Table 2) and sampling sites
(A, B, C, D, E and F). Concentration data were log(x a) transformed.
In the case of total P and Mn (total and oxides) a 0.01 due to the
comparative low values, whilst for all other concentrations a 1.
2.3. Sorption study
A sorption isotherm represents the quantity of a substance
sorbed on the surface of a solid as a function of the equilibrium
concentration of that substance in the solution that the solid is in
contact with, at a constant temperature (Abou et al., 2007).
Adsorption experiments were then carried out using homogenised
samples (particle size 1 mm). The adsorption study carried out
was based on the OECDTG106 (2000), for the Determination of Soil
Adsorption/Desorption, using a Batch Equilibrium Method. This
consisted of a preliminary study to nd out the appropriate
experimental conditions to study the PO
4
eP sorption and of an
equilibrium study to determine the PO
4
eP sorption capacity of
soils/sediments and the corresponding isotherms.
2.4. Preliminary study
This consisted in the determination of appropriate solid-
mass
:solution
volume
ratio (R) to be used for subsequent sorption
experiments. R depends on the specic solid (in this study, soils or
sediments) adsorption capacity and on the limit of the analytical
methodology for the specic target solute (in this study, PO
4
eP).
For preliminary studies it is recommended to test a few xed R
and to chose one for which the percentage of the target solute
adsorbed is >20%, and preferably > 50% (Boesten, 1990), while the
target solute concentration is high enough to be measured accu-
rately. Furthermore, for adsorption/desorption studies into soils it
is advised to use at least 1 g of solid, in order to get reliable results
(OECD TG 106, 2000).
Different masses (1, 2 and 5 g, dwt) of soil/sediment were
exposed under stirring with a xed volume (20 mL) of a PO
4
eP
solution so to determine which of the resulting R (0.05, 0.1, 0.25)
was the most suitable one for the following equilibrium study to be
carried out.
Table 1
Summary characterization of the South arm of the Mondego estuary.
Characteristics
Geographic location 40

08
0
N, 8

50
0
W
System intertidal
area (km
2
)
1.75
System subtidal
area (km
2
)
0.96
System volume
(10
6
m
3
)
5
Mean depth (m) 2e4 high tide
Tidal range (m) 0.35e3.3
Mixing characteristics Well-mixed with irregular
river discharges
Mean substratum
composition
Silt, clay and sand
Annual insolation
of PAR (400e700 nm)
(mol phot. m
2
y
1
)
3200e32,000
Characteristics Before management After management
1993e1997 1999e2006 (May) 2006e2008
(March)
Salinity
range (psu)
1.9e33.1 0.2e35.8 17.4e33.8
Salinity
(mean std)
(22.0 7.88) (26.2 7.42) (29.4 4.28)
Mean water
temperature
range (

C)
8.0e23.7 7.6e29.0 9.3e27.1
Residence time Moderate (weeks) Short (days)
*
Current velocity Low and dependent
on the Pranto
river sluice
Higher and not
dependent on the
Pranto river sluice
(Not dependent
on the Pranto
river sluice)
*
Turbidity High Lower
*
DIP(maximum)
(mmol L
1
)
2.02 3.06 2.79
DIP (Minimun)
(mmol L
1
)
0.44 0.20 0.63
DIP (mean std)
(mmol L
1
) (n)
1.01 0.07 (40) 1.54 0.07 (77) 1.45 0.13 (20)
* Information not available.
A.I. Lilleb et al. / Estuarine, Coastal and Shelf Science 99 (2012) 85e94 88
First, a stock solution of PO
4
eP (20,000 mg P L
1
) was prepared
by dissolving KH
2
PO
4
in water 24 h before use and kept closed in
the dark at 4

C. Then the corresponding mass (1, 2 or 5 g) of each
solid was placed in a glass ask and shaken (150 r.p.m.) overnight
with 18 mL of CaCl
2
(0.01 mol L
1
) at controlled room temperature
(25

C) in a GFL 3017 Shaker. Then, shaking was stopped and 2 mL of

PO
4
eP (20,000 mg L
1
) were added to each ask so to have 20 mL of
PO
4
eP 2000 mg L
1
.
Since the moment at which the solid starts to be shaken
together with the PO
4
eP solution, adsorption (and desorption) of
PO
4
eP into the solid starts and the PO
4
eP concentration in solution
decreases until the equilibrium, from which PO
4
eP concentration
remains constant. In order to assess the time necessary to attain
adsorption equilibrium, different shaking periods (t 4, 8, 24, 48
and 60 h) were tested. For each solid and R, as many asks as
different t were prepared in triplicate and run as described above.
After each t, the corresponding three replicates were removed from
the shaker, the supernatant solution was ltered (1.2 mm Whatman
GF/C) and analysed for PO
4
eP quantication using the ascorbic acid
method, following the Limnologisk Metodik (1992) methodology.
For each solid, R and t, one blank was run in triplicate (the
corresponding mass of each sediment/soil and 20 mL of
Fig. 2. A - The variation of mean temperature (

C) in the water and the concentration of phosphate (PO

4
eP, mean SE) between January 1993 and March 2008; B - Relation
between mean water temperature and the concentration of phosphate before (1993e1997) and after (1999e2008) the managements; C - Simulation of the yearly variation in
PO
4
eP uxes in the mudats, using real weekly temperatures from 1999 to 2001 as forcing function (white circles), and real uxes measured seasonally in the eld (black dia-
monds) (from Lilleb et al., 2004).
A.I. Lilleb et al. / Estuarine, Coastal and Shelf Science 99 (2012) 85e94 89
0.01 mol L
1
CaCl
2
, without PO
4
eP). Triplicate controls (20 mL of
PO
4
eP in 0.01 mol L
1
CaCl
2
without soil or sediment) were also
run in parallel.
2.5. Equilibrium study
After having determined the most suitable R for each solid and
the time necessary to attain the equilibrium, adsorption equilib-
rium experiments were carried out and the adsorbed PO
4
eP onto
the soils/sediments at different initial PO
4
eP concentrations
(between 2000 and 100,000 mg L
1
) was quantied. For each solid,
as many asks as the different concentrations considered were
prepared in triplicate. Taking into account the results from the
preliminary study, the most appropriate R was 0.05 thus, 1 g of each
soil/sediment was placed in each glass ask together with 18 mL of
0.01 mol L
1
CaCl
2
and shaken (150 r.p.m.) overnight at controlled
room temperature (25

C). Then 2 mL of stock solution with the
appropriate PO
4
eP concentration (20,000, 30,000, 40,000, 50,000,
60,000 or 80,000 mg L
1
) were added to each ask so to have asks
containing 20 mL of solution with different initial PO
4
eP concen-
trations (between 2000 and 100,000 mg L
1
). Flasks were shaken
(150 r.p.m. at 25

C) for 48 h, which was determined in the
preliminary study as the time required to reach adsorption equi-
librium (t
e
). Afterwards, the necessary volume of supernatant was
withdrawn from each ask, ltered (Whatman GF/C) and analysed
for equilibrium PO
4
eP concentration.
For each experiment, one triplicate blank (one sediment/soil
and 0.01 mol L
1
CaCl
2
, without PO
4
eP) was run in parallel. Also,
triplicate controls (PO
4
eP in 0.01 mol L
1
CaCl
2
at the different
initial concentrations, without soil/sediment) were subjected to the
same experimental procedure.
2.6. Calculations and equilibrium model
At equilibrium, the amount of PO
4
eP adsorbed onto the
different solids, q
e
(mg P g
1
dwt), was calculated by a mass balance
relationship (eq. (1)).
q
e
C
0
C
e

V
W
(1)
where C
0
(mg P L
1
) is the initial PO
4
eP concentration in solution, C
e
(mg P L
1
) is the blank corrected equilibrium concentration of
PO
4
eP, V (L) is the volume of the solution and W(g dwt) is the mass
of soil/sediment.
After q
e
calculation, the corresponding distribution coefcient
(K
d
), which represents the ratio between PO
4
eP in the solid and in
the solution at the equilibrium, was calculated:
K
d

q
e
C
e

L g
1

(2)
The percentage of adsorption (A%) at the equilibrium was
determined as:
A%
C
0
C
e
C
0
100 (3)
To describe the adsorption equilibrium, the Langmuir isotherm
model (eq. (4)) was used.
q
e

QK
L
C
e
1 K
L
C
e
(4)
where K
L
(L mmol
1
) and Q (mmol mg
1
dwt) are the Langmuir
sorption equilibriumconstant and maximumcapacity, respectively.
3. Results
3.1. Physicochemical properties of sediments and agricultural soils
Physicochemical characterization of sediment/soil samples is
shown in Table 2. The 2-dimensional PCA conguration (Fig. 3)
displays the relationship between sampling sites (A, B, C, D, E and F)
and the environmental variables vectors (sediments/soils physi-
cochemical properties), explaining 81.0% of the variance. Fig. 3
shows that the percentage of moisture and of LOI (loss on igni-
tion), and the amount of P
total
, Fe
total
, total Fe
oxides
and Mn
total,
have
positive values in the rst axis (PCA
1
), explaining 63.3% of the
variance, separating soil A and sediments F and D. The environ-
mental variables Mn
oxides
and pH have negative values in the
second axis (PCA
2
), which explains 17.7% of the variance, clearly
separating sediment D, Thus, the analysis clearly shows that the
three estuarine sediments and the three upstreamagricultural soils
can be separated into three groups according to the similarities of
their physicochemical properties: a) sediment D, b) soils B and C,
and sediment E, and c) soil A and sediment F.
3.2. Preliminary study
The preliminary study veried the fact that, for most of the
solids used, the adsorption equilibrium was attained within 24 h.
However, in the case of soil B, equilibrium was reached only after
48 h. So, in the subsequent adsorption equilibrium experiments
shaking was maintained for 48 h for all the solids under study.
Selection of the appropriate ratio was made on the basis of the plot
of solid
mass
:solution
volume
ratios (R) versus K
d
and percentage of
adsorption (A%). These plots are shown in Fig. 4. Differences
between soils and sediments are not noticeable although disparity
among soils, both on A (%) and on K
d
(L g
1
), is slightly higher than
Table 2
Physicochemical properties of soils and sediments used for the study of P adsorption (mean value standard deviation).
Property (unit) Soils Sediments
A B C D E F
Moisture (%) 47.7 9.4 15.2 1.1 21.7 0.8 44.2 3.3 33.6 3.1 48.5 3.0
LOI (%) 6.6 0.5 2.6 1.6 4.0 0.05 6.6 0.8 4.0 0.4 6.8 1.5
Fine fraction (% <63 mm) 46.5 3.3 17.6 0.5 24.4 1.2 49.7 4.1 36.2 1.8 32.8 0.3
Bulk density (kg L
1
Kg L
1
) 0.69 0.18 1.64 0.04 1.37 0.03 0.75 0.07 1.01 0.08 0.67 0.06
pH [H
2
O] 5.7 0.1 7.5 0.1 7.4 0.0 7.1 0.0 7.1 0.1 6.8 0.0
pH [CaCl
2
] 5.3 0.1 6.9 0.0 7.1 0.1 7.0 0.0 6.9 0.1 6.7 0.1
Total P (mg g
1
) 0.66 0.02 0.30 0.05 0.64 0.12 0.78 0.04 0.37 0.04 0.48 0.09
Total Fe (mg g
1
) 34.77 0.79 17.36 0.22 11.00 0.43 39.65 0.39 22.03 0.99 30.23 0.45
Fe oxides (mg g
1
) 3.39 0.16 0.67 0.04 0.80 0.04 3.64 0.04 1.47 0.05 2.14 0.04
Total Mn (mg g
1
) 0.19 0.00 0.24 0.01 0.19 0.01 0.29 0.01 0.21 0.03 0.21 0.01
Mn oxides (mg g
1
) 0.05 0.01 0.11 0.00 0.14 0.00 0.09 0.00 0.02 0.00 0.04 0.00
A.I. Lilleb et al. / Estuarine, Coastal and Shelf Science 99 (2012) 85e94 90
among sediments. For the three R considered here, either for soils
or for sediments, A was >50%. Furthermore, for all of them, the
remaining PO
4
eP in solution after equilibrium was over the
detection limit (0.2 mg P L
1
; y y
blank
3 std
blank
(Miller and
Miller, 1984)) and over the accurate quantication limit
(1 mg P L
1
; y y
blank
10 std
blank
(Miller and Miller, 1984)).
According to error analysis studies (Boesten, 1990) K
d
values below
0.0003 L g
1
cannot be estimated accurately from a decrease in
concentration in the aqueous phase. Thus, for having the maximum
K
d
, the R 0.05, which corresponds to 1 g of soil or sediment in
20 mL of PO
4
eP solution, was chosen for the equilibrium study.
3.3. Equilibrium study
We tted a Langmuir model to the PO
4
eP adsorption isotherm
data (Fig. 5). This gure shows that for all soils and sediments used
under the experimental conditions, q
e
increases with the C
e
of
remaining PO
4
eP in solution and the Langmuir model ts experi-
mental points. However, there are differences in the adsorption
into the different solid matrixes. It can be seen in Fig. 5 that soil A
and especially sediment D adsorbed more PO
4
eP than the rest of
solids and even at low C
e
, comparatively high q
e
were reached in
these solids. Sediments F and E showed a slightly lower adsorption
of PO
4
eP, but higher than that of soils B and C. The Langmuir
maximum capacities (Q, mg P g
1
dwt) in Table 3 reect differences
between the adsorption of PO
4
eP onto the solids under study. It is
emphasized that the results in Table 3 were obtained using the
dried soils and sediments and that, (see Table 2) they had
a different % of moisture. Thus, for eld inferring of these results,
PO
4
eP adsorption capacity of each material was corrected for % of
moisture. Table 4 shows the results on adsorption capacity ob-
tained using the soils and sediments after correction for moisture
content (Q
*
, mg g
1
wwt). The corrected results clearly show that
sediment D has a comparatively higher adsorption capacity
(Q
*
657 mg P g
1
wwt), when compared to the rest of sediments
(Q
*
515 and 410 mg P g
1
wwt, for sediments E and F,
respectively), and soils (Q
*
472, 469 and 526 mg P g
1
wwt, for
soils A, B and C, respectively).
4. Discussion
In aquatic systems, phosphorus can be present in the form of
iron, aluminium and calcium oxides or adsorbed on the surface of
minerals and organic materials (e.g. Bostrmet al., 1988; Morel and
Hering, 1993; Coelho et al., 2004). In this last case, the environ-
mental conditions such as salinity, pH or redox potential will
determine the degree of PO
4
eP adsorption. In freshwater, phos-
phorus is transported mainly adsorbed to Fe and Al oxides (Lebo,
1991; Zwolsman, 1994) but, in marine areas, the increase in pH
with salinity may prevent phosphate adsorption onto Fe oxides, by
changing both the speciation of phosphate from H
2
PO

4
to HPO
2
4
,
and the surface charge on the Fe oxides (Lebo, 1991; Zwolsman,
1994). Solid calcium carbonates are also inuenced by salinity
and pH. At high salinity conditions, both in estuarine and marine
environments, calcite is produced through precipitation reactions
and biological activity, co-precipitating phosphate and forming an
adsorption substrate for dissolved phosphate. Calcite transported
upstream through tidal currents may dissolve because of the lower
pH, releasing phosphate to the water column (de Jonge and
Villerius, 1989).
Fig. 3. Results from the PCA analysis, showing the principal components 1 and 2 for
soils and sediment physicochemical properties vectors and soils (A, B and C) and
sediments (D, E and F).
Fig. 4. A - Plot of soil to solution ratio (R) versus the distribution coefcient (K
d
)
(isolated symbols) and the percentage of adsorption (A (%)) (symbols with lines); B -
Plot of sediment to solution ratio (R) versus the distribution coefcient (K
d
) (isolated
symbols) and the percentage of adsorption (A (%)) (linked symbols).
A.I. Lilleb et al. / Estuarine, Coastal and Shelf Science 99 (2012) 85e94 91
The organic matter content and total Fe and, specically, Fe
oxides content are the properties that most signicantly affect the
adsorption capacity of soils (McGechan, 2002) and sediments (Pant
and Reddy, 2001). In the present work, comparing results in
Tables 2 and 3 shows a clear relationship between the Q (mg P g
1
dwt) determined for the adsorption of PO
4
eP onto the studied soils
and sediments and these properties.
Except for soil A, which has a relatively high afnity for PO
4
eP,
the adsorption extent of PO
4
eP into the soils studied is in the range
of results found for soils from cultivated elds (Zhou and Li, 2001;
Heredia and Fernandez-Cirelli, 2007). Also, the PO
4
eP adsorption
capacity of the sediments estimated here, especially that of D,
was higher than most of estuarine sediments in the study by
Pant and Reddy (2001), in which Sorption maxima under anaer-
obic conditions ranged between 31 and 500 mg P g
1
wwt,
N 9) or in oodplain sediments (Tian and Zhou, 2008;
Q
max
44e188 mg P g
1
dwt, N 7). The relatively higher
adsorption capacity of PO
4
eP by sediment D may be explained by
the comparatively higher organic matter content (expressed as %
LOI), higher percent of sediment ne fraction (<63 mm) and the
higher percentages of metals, especially iron in the oxidised form,
as was highlighted by the principal components analysis (Fig. 3),
explaining 81.0% of the variance. Taking into account the results
from the sorption study and the existing background information,
three major areas with distinct characteristics have been consid-
ered in order to scale up to the systemlevel, i.e., to the south arm of
the Mondego estuary (z2.1 km
2
). The considered areas are: i) E e
the downstream area with z1.2 km
2
; ii) D e the downstream area
of the Pranto river, including the mixing zone with z0.4 km
2
, and
iii) F e the upstream area of the South arm with z0.5 km
2
.
However, intertidal areas represent 63% of the total area and mud
ats account for 60% of the intertidal areas, so the area to be
considered should be, respectively: i) E z 0.45 km
2
, ii)
D z 0.15 km
2
and iii) F z 0.19 km
2
. The Langmuir maximum
adsorption capacity of sediments at each area corrected for mois-
ture is i) E e the downstream area with Q* 515 mg P g
1
; ii) D e
the downstream area of the Pranto river with Q* 657 mg P g
1
;
and iii) F e the upstream area of the South arm with
Q* 410 mg P g
1
. Each calculated Q* corresponds to a sediment
surface area of 7.1 cm
2
g
1
sediment (batch experiment) thus
extrapolating to the system, each area has potentially the following
maximum adsorption capacity corrected for moisture (C*): E
C* 3.3 ton P; D C* 1.4 ton P; F C* 1.1 ton P. Considering the
mud ats areas of the south arm (E D F) the maximum
adsorption capacity corrected for moisture C*z 5.9 t P. This value
corresponds to the 1993e1997 period. For the 1999e2008 situa-
tion, a 70% reduction in the D area corresponding to the down-
stream area of the Pranto river should be considered. Thus, the area
corresponding to Dfor the 1999e2008 situation is 0.11 ha. Then, for
this situation D C* 1.0 ton P. So, considering the mud ats areas of
the south arm(E D F) C*z5.4 t P, meaning a decrease in 7.3% of
the maximum adsorption capacity of the system. Therefore, at the
system level two scenarios should be compared: i) Before 1998 the
nutrient-rich freshwater from the upstream cultivated lands
entered the system through the sediment D area. This area has the
highest PO
4
eP adsorption capacity (Q
*
657 mg P g
1
wwt), thus
the PO
4
eP availability in the water column was mostly dependent
on the mineralization processes. However, during this period, the
estuary was mesotrophic concerning the concentration of DIP
(Lilleb et al., 2007). ii) After 1998, the nutrient-rich freshwater
from the upstream cultivated lands entered the system in the
sediment F area with a much lower PO
4
eP adsorption capacity
(Q
*
410 mg P g
1
wwt). These sediments were not able to adsorb
as much as PO
4
eP and higher equilibrium concentration occurs in
the water column. Simultaneously, the water residence time
diminished from moderate (weeks) before 1997, to short (days)
after 1998. The PO
4
eP availability in the water column is no longer
solely dependent on the mineralization processes, and high o-
phosphate concentrations persist during autumn and winter.
Additionally, and in accordance with the same environmental
classication, the estuary became polytrophic (Lilleb et al., 2007).
During this period, salinity also increased signicantly
(ManneWhitney U Statistic; P <0.001) from a median value of
24.3e28.8 (z18% increase) but still within the polyhaline
denition (salinity range between 18 and 30). It is known that
salinity may also affect P adsorption, since a salinity increase
involves an increase ionic competition for adsorption sites.
Namely, when riverine particles reach the estuarine mixing zones
Fig. 5. The adsorption isotherms of PO
4
eP onto the soils (A, B and C) and sediments (D,
E and F) and the corresponding ttings of the experimental points to the Langmuir
model.
Table 3
Langmuir parameters and standard deviation of residuals (s
y.x
) corresponding to the
ttings of the equilibriumisotherm(25

C) data on the adsorption of PO

4
eP onto the
different soils (A, B and C) and sediments (D, E and F).
Property A B C D E F
Q, mg g
1
dwt 902 554 672 1177 775 796
K
L
, L g
1
1.E-04 4.E-05 3.E-05 6.E-05 5.E-05 9.E-05
s
y.x
26.98 34.16 20.50 25.44 22.76 32.05
Table 4
Langmuir maximum adsorption capacity of PO
4
eP onto the different soils (A, B and
C) and sediments (D, E and F) after correction for moisture (%).
Property A B C D E F
Moisture% 47.7 15.2 21.7 44.2 33.6 48.5
Q,
*
mg g
1
wwt 472 469 526 657 515 410
*
Langmuir maximum adsorption capacity corrected for moisture.
A.I. Lilleb et al. / Estuarine, Coastal and Shelf Science 99 (2012) 85e94 92
the particle-bound P is released to seawater. But it has been
demonstrated that the inuence of salinity on Kd is greatest in
sediments containing low exchangeable P, diminishing in sedi-
ments with high exchangeable P (Zhang and Huang, 2011). Coelho
et al. (2004) previously showed that when compared to sediment E,
sediment F had a higher amount of exchangeable P (z42.8 g P m
2
)
and a decrease was observed from the inner mud ats to the outer
mud ats of the estuary (z26.0 g P m
2
), which is also in agree-
ment with the opposite pattern found here for the PO
4
-P adsorp-
tion capacity corrected for moisture (Q
*
410 mg P g
1
for sediment
F and Q
*
515 mg P g
1
for sediment E). However, the higher Q
*
by
sediment E cannot be explained by the content in organic matter
content nor by the percentages of metals, namely oxidised iron, as
expressed by the principal components analysis. In fact, in the inner
mud ats (sediment F), the percentages of phosphorus present in
the form of iron, aluminium and calcium oxides were Al/clay-
P 18% CaeP 24% and Fe/Mn-P 32%, whilst in the outer mud
ats of the estuary (sediment E) the percentages were Al/clay-
P 11% CaeP 43% and Fe/MneP 17% (Coelho et al., 2004).
Hence, for sediment E, from the higher salinity area, co-
precipitation of PO
4
eP with calcium carbonate seems to be
responsible for the comparatively higher PO
4
eP retention capacity
(this study) and the lower amount of exchangeable P (Coelho et al.,
2004). Another experimental study has shown that in addition to
sediment composition, temperature and salinity are the most
common environmental variables that inuence the phosphate
uxes in sediment water interface (Zhang and Huang, 2011). In our
study, P adsorption onto the different agricultural soils and estua-
rine sediments was studied under identical conditions and P
solutions for experimentation were prepared in CaCl
2
(0.01 mol
L
1
), as described in the M&M section. Thus, possible salinity and
temperature effects on P adsorption were not considered in the
present work. Furthermore, DIP concentrations increased for the
past 10 years during the winter period, when an increase in
freshwater inputs due to precipitation lowers the salinity. Another
possible effect could be related with Rochfords suggestion, so
called by Pomeroy et al. (1965), suggesting that phosphate efux
fromthe sediment only occurs when the concentration in the water
decreases below 1.0 mmol L
1
, and sediment takes up phosphate
only when the concentration in water is above 3 or 4 mmol L
1
. This
is a valuable area of future study.
Assuming that the amount of fertilizers applied in the cultivated
elds was not increased during the last ve years, the changes in
the system hydrodynamics have degraded the water quality, as far
as the concentration of dissolved inorganic phosphorus (DIP, o-
phosphate, PO
4
eP) is concerned.
Studies based on time series data fromthe Wadden Sea and Ems
estuary have shown that P-reduction measures have resulted in an
effective decrease of DIP (de Jonge et al., 2002), indicating that P-
reduction measures should be implemented in the Mondego
estuary. The sorption study that was carried out has proven to be
a valuable tool and a valuable method to obtain essential infor-
mation on the mobility of PO
4
-P from soils/sediments to the water
column.
5. Conclusions
The change in the Mondego estuary hydrology has conditioned
the spatial and temporal variability in the systemcapacity to adsorb
P. Results obtained reinforce the idea that management at the
system level should also be accomplished through the reduction of
nutrient loads, particularly of phosphorus, which will constitute
a longer term solution. On a broader scale, results may be seen as
a case study for the effect of land uses on estuarine eutrophication,
highlighting the role of soils and sediments PO
4
eP sorption to the
system P sorption capacity.
Acknowledgements
The authors acknowledge funding from Portuguese FCT (Foun-
dation for Science and Technology) to CESAM (Centre for Envi-
ronmental and Marine Studies). The Post-Doc grant SFRH/BPD/
48449/2008 (P. Coelho) funded by FCT is also acknowledged.
Special thanks are given to Dr Jos Raposo for drawing Fig. 1C.
References
Abou, N.J.S., Madramootoo, C.A., Hendershot, W.H., 2007. Modelling phosphate
adsorption to the soil: application of the non-ideal competitive adsorption
model. Environmental Pollution 149, 1e9.
Barone, F.S., Rowe, R.K., Quigley, R.M., 1992. A laboratory estimation of diffusion and
adsorption coefcients for several volatile organics in a natural clayey soil.
Journal of Contaminant Hydrology 10, 225e250.
Bechmann, M.E., Bergeb, D., Eggestada, H.O., Vandsemb, S.M., 2005. Phosphorus
transfer from agricultural areas and its impact on the eutrophication of lake-
sdtwo long-term integrated studies from Norway. Journal of Hydrology 304,
238e250.
Boesten, J.J.T.I., 1990. Inuence of soil/liquid ratio on the experimental error of
sorption coefcients in pesticide/soil system. Journal of Pest Science 30, 31e41.
Bostrm, B., Andersen, J.M., Fleischer, S., Jansson, M., 1988. Exchange of phosphorus
across the sediment-water interface. Hydrobiologia 170, 229e244.
Carpenter, S.R., 2005. Eutrophication of aquatic ecosystems: bistability and soil
phosphorus. PNAS- Proceedings of the National Academy of Sciences of the
United States of America 102, 10002e10005.
Coelho, J.P., Flindt, M.R., Jensen, H.S., Lilleb, A.I., Pardal, M.A., 2004. Phosphorus
speciation and availability in intertidal sediments of a temperate estuary:
relation to eutrophication and annual P-uxes. Estuarine Coastal and Shelf
Science 61, 583e590.
Cyra, H., McCabe, S.K., Nrnberg, G.K., 2009. Phosphorus sorption experiments and
the potential for internal phosphorus loading in littoral areas of a stratied lake.
Water Research 43, 1654e1666.
de Jonge, V.N., Villerius, L.A., 1989. Possible role of carbonate dissolution in estua-
rine phosphate dynamics. Limnology and Oceanography 34, 332e340.
de Jonge, V.N., Engelkes, M.M., 1993. The role of mineral compounds and chemical
conditions in the binding of phosphate in the Ems estuary. Netherlands Journal
of Aquatic Ecology 27, 227e236.
de Jonge, V.N., Engelkes, M.M., Bakker, J.F., 1993. Bio-availability of phosphorus
compounds in sediments of the western Dutch Wadden Sea. Hydrobiologia 253,
151e163.
de Jonge, V.N., Elliot, M., Orive, E., 2002. Causes, historical development, effects and
future challenges of a common environmental problem: eutrophication.
Hydrobiologia 475/476, 1e19.
Flindt, M.R., Kamp-Nielsen, L., Marques, J.C., Pardal, M.A., Bocci, M., Bendoricho, G.,
Nielsen, S.N., Jorgensen, S.E., 1997. Description of the three shallow estuaries:
Mondego river (Portugal), Roskilde Fjord (Denmark) and the Lagoon of Venice
(Italy). Ecological Modelling 102, 17e31.
Garnier, J., Nmery, J., Billen, G., Thry, S., 2005. Nutrient dynamics and control of
eutrophication in the Marne River system: modelling the role of exchangeable
phosphorus. Journal of Hydrology 304, 397e412.
Heredia, O.S., Fernandez-Cirelli, A., 2007. Environmental risks of increasing phos-
phorus addition in relation to soil sorption capacity. Geoderma 137, 426e431.
Lebo, M.E., 1991. Particle-bound phosphorus along an urbanized coastal plain
estuary. Marine Chemistry 34, 225e246.
Leston, S., Lilleb, A.I., Pardal, M.A., 2008. The response of primary producers
assemblages to the mitigation measures to reduce eutrophication symptoms in
a temperate estuary. Estuarine Coastal and Shelf Science 77, 688e696.
Lilleb, A.I., Neto, J.M., Flindt, M.R., Marques, J.C., Pardal, M.A., 2004. Phosphorous
dynamics in a temperate intertidal estuary. Estuarine Coastal and Shelf Science
61, 101e109.
Lilleb, A.I., Neto, J.M., Martins, I., Verdelhos, T., Leston, S., Cardoso, P.G.,
Ferreira, S.M., Marques, J.C., Pardal, M.A., 2005. Management of a shallow
temperate estuary to control eutrophication: the effect of hydrodynamics on
the system nutrient loading. Estuarine Coastal and Shelf Science 65, 697e707.
Lilleb, A.I., Teixeira, H., Pardal, M.A., Marques, J.C., 2007. Applying water quality
status criteria to a temperate estuary (Mondego, Portugal) before and after the
mitigation measures to reduce eutrophication symptoms. Estuarine Coastal and
Shelf Science 72, 177e187.
Limnologisk Metodik, 1992. In: Kbenhavns Universitet (Ed.), Ferskvandsbiologisk
Laboratorium. Akadmisk Forlag, Kbenhavns.
McGechan, M.B., 2002. SW-soil and water - sorption of phosphorus by soil, part 2:
measurement methods, results and model parameter values. Biosystems
Engineering 82, 115e130.
Miller, J.C., Miller, J.N., 1984. Statistics for Analytical Chemistry. Ellis Horwood
Limited, Chischester.
A.I. Lilleb et al. / Estuarine, Coastal and Shelf Science 99 (2012) 85e94 93
Monterroso, P., Pato, P., Pereira, M.E., Millward, G.E., Vale, C., Duarte, A., 2007. Metal-
contaminated sediments in a semi-closed basin: implications for recovery.
Estuarine Coastal and Shelf Science 71, 148e158.
Morel, F.M.M., Hering, J.G., 1993. Principles and Applications of Aquatic Chemistry.
Wiley, New York, pp. 320e570.
OECD TG 106, 2000. Determination of Soil Adsorption/Desorption, Using a Batch
Equilibrium Method. OECD Guidelines for the Testing of Chemicals. Organi-
sation for Economic Co-operation and Development. Environment
Directorate.
Pant, H.K., Reddy, K.R., 2001. Phosphorus sorption characteristics of estuarine
sediments under different redox conditions. Journal of Environmental Quality
30, 1474e1480.
Pinckney, J.L., Paerl, H.W., Tester, P., Richardson, T.L., 2001. The role of nutrient
loading and eutrophication in estuarine ecology. Environment Health
Perspective 109 (Suppl. 5), 699e706.
Pomeroy, L.R., Smith, E.E., Grant, C.M., 1965. The exchange of phosphate between
estuarine water and sediments. Limnology and Oceanography 2, 167e172.
Ramalhosa, E., Segade, S.R., Pereira, E., Vale, C., Duarte, A., 2006. Mercury cycling
between the water column and surface sediments in a contaminated area.
Water Research 40, 2893e2900.
Sharpley, A.N., Smith, S.J., Jones, O.R., Berg, W.A., Coleman, G.A., 1992. The transport
of bioavailable phosphorus in agricultural runoff. Journal of Environmental
Quality 21, 30e35.
Tian, J., Zhou, P., 2008. Phosphorus fractions and adsorption characteristics of
oodplain sediments in the lower reaches of the Hanjiang River, China. Envi-
ronmental Monitoring and Assessment 137, 233e241.
van Beusekom, J.E.E., de Jonge, V.N., 1998. Retention of phosphorus and nitrogen in
the Ems estuary. Estuaries 21, 527e539.
Wasmund, N., Andrushaitis, A., Lysiak-Pastuzak, E., Muller-Karulis, B., Nausch, G.,
Neumann, T., Ojaveer, H., Olenina, I., Postel, L., Witek, Z., 2001. Trophic status of
the south-eastern Baltic sea: a comparison of coastal and open areas. Estuarine
Coastal and Shelf Science 53, 849e864.
Zhang, J.Z., Huang, X.L., 2011. Effect of temperatureandsalinityonphosphatesorption
on marine sediments. Environmental Science and Technology 45, 6831e6837.
Zhou, M., Li, Y., 2001. Phosphorus-Sorption characteristics of Calcareous soils and
Limestone from the southern Everglades and Adjacent Farmlands. Soil Science
Society of America Journal 65, 1404e1412.
Zwolsman, J.J.G., 1994. Seasonal variability and biogeochemistry of phosphorus in
the scheldt estuary, South-west Netherlands. Estuarine Coastal and Shelf
Science 39, 227e248.
A.I. Lilleb et al. / Estuarine, Coastal and Shelf Science 99 (2012) 85e94 94