REVIEWS IN

ECONOMIC GEOLOGY
Volume 14

STRUCTURAL CONTROLS ON
ORE GENESIS
Editors
Jeremy P. Richards and Richard M. Tosdal

SOCIETY OF ECONOMIC GEOLOGISTS, INC.

 REVIEWS IN ECONOMIC GEOLOGY
(ISSN 0741-0123)
Published by the
SOCIETY OF ECONOMIC GEOLOGISTS, INC.
Graphic Design by Johnson Printing, 1880 S. 57th Court, Boulder, CO 80301
Printed by Johnson Printing, 1880 S. 57th Court, Boulder, CO 80301
Additional copies o f this volume may be obtained from:
Society o f Economic Geologists, Inc.
7811 Shaffer Parkway
Litdeton, CO 80127 USA
Tel. 1.720.981.7882
Fax 1.720.981.7874
e-mail: seg@segweb.org

Vol. 1: FLUID-MINERAL EQUILIBRIA IN HYDROTHERMAL SYSTEMS (1 9 8 4 ) ISBN 0-9613074-0-4

Vol. 2: GEOLOGY AND GEOCHEMISTRY OF EPITHERMAL SYSTEMS (1 9 8 5 ) ISBN 0-9613074-1-2
Vol. 3: EXPLORATION GEOCHEMISTRY: DESIGN AND
INTERPRETATION OF SOIL SURVEYS (1 9 8 6 ) ISBN 0-9613074-2-0
Vol. 4: ORE DEPOSITION ASSOCIATED WITH MAGMAS (1 9 8 9 ) ISBN 0-9613074-3-9
Vol. 5: SEDIMENTARY AND DIAGENETIC MINERAL DEPOSITS:
A BASIN ANALYSIS APPROACH TO EXPLORATION (1 9 9 1 ) ISBN 0 -9 6 1 3 0 7 4 4 -7
Vol. 6a THE ENVIRONMENTAL GEOCHEMISTRY OF MINERAL
DEPOSITS PART I: PROCESSES, TECHNIQUES,
AND HEALTH ISSUES (1 9 9 9 ) IS B N 1-887483-50-0
Vol. 6b THE ENVIRONMENTAL GEOCHEMISTRY OF MINERAL
DEPOSITS PART II: CASE STUDIES AND RESEARCH TOPICS (1 9 9 9 ) IS B N 1-887483-504)
Vol. 7: APPLICATIONS OF MICROANALYTICAL TECHNIQUES
TO UNDERSTANDING MINERALIZING PROCESSES (1 9 9 8 ) IS B N 1-887483-51-9
Vol. 8: VOLCANIC-ASSOCIATED MASSIVE SULFIDE DEPOSITS:
PROCESSES AND EXAMPLES IN MODERN AND
ANCIENT SETTINGS (1 9 9 9 ) IS B N 1-887483-52-7
Vol. 9: ORE GENESIS AND EXPLORATION:
THE ROLES OF ORGANIC MATTER (2 0 0 0 ) IS B N 1 -8 8 7 4 8 3 -5 3 5
Vol. 10 TECHNIQUES IN HYDROTHERMAL ORE DEPOSITS GEOLOGY (1 9 9 8 ) ISB N 1 -8 8 7 4 8 3 5 4 -3
Vol. 11 METAMORPHOSED AND METAMORPHOGENIC ORE DEPOSITS (2 0 0 0 ) ISBN 1 -8 8 7 4 8 3 5 3 1
Vol. 12 APPLICATION OF RADIOGENIC ISOTOPES TO ORE
DEPOSITS RESEARCH AND EXPLORATION (1 9 9 9 ) ISB N 1 -8 8 7 4 8 3 5 6 -X
Vol. 13: GOLD IN 2000 (2 0 0 0 ) ISB N 1 -8 8 7 4 8 3 5 7 -8
Vol. 14: STRUCTURAL CONTROLS ON ORE GENESIS (2 0 0 1 ) ISBN 1 -8 8 7 4 8 3 5 8 -6

Reviews in Economic Geology is a publication o f the Society o f Economic Geologists designed to accompany the Society's
Short Course series. Like the Short Courses, each volum e provides comprehensive updates on various applied and acad­
em ic topics for practicing econom ic geologists and geochem ists in exploration, developm ent, research, and teaching. Vol­
umes arc produced in conjunction with each new Short Course, first serving as a textbook for that course, and
subsequently made available to SEG members and others at modest cost.
© Copyright 2001, Society o f Economic Geologists. Inc.
Perm ission is granted to individuals to make single copies o f chapters for personal use in research, study, and teach­
ing, a n d to use short quotations, illustrations, and tables from Rn/ietm in Economic Geology for publication in scientific
wotks. Such uses must be appropriately credited. Copying for general distribution, for promotion and advertising, for ere-
aung new collective works, or for other commercial purposes is not permitted without the specific written permission o f
the Society o f Econom ic Geologists, Inc.
 REVIEWS IN
ECONOMIC GEOLOGY
(issn 0741-0123) Volume 14

STRUCTURAL CONTROLS ON
ORE GENESIS

ISBN 1-887485-58-6

Volume Editors
J erem y P. R i c h a r d s Ric h a r d M . T o sd a l
Department o f Earth and Atmospheric Sciences Mineral Deposit Research Unit
University of Alberta Department o f Earth and Ocean Sciences
Edmonton, Alberta T 6G 2E3 University of British Columbia
Canada Vancouver, British Columbia V6T 1Z4
Canada

SOCIETY OF ECONOMIC GEOLOGISTS, INC.

The Authors:
Byron Berger Steve Ludington Richard H. Sibson
U.S. Geological Survey U.S. Geological Survey University of Otago
MS 964, Federal Center MS 989, 345 Middlefield Rd. Departm ent of Geology
Denver, CO 80225-0046 Menlo Park, CA 94025 PO Box 56
USA USA Dunedin
Tel. +$03,256-5533 Tel. +650.329.5371 New Zealand
Fax: +303.236-3200 Fax: +650.329.5374 Tel. +64.3.479.7506
E-mail: bberger@usgs.gov E-mail: slud@usgs.gov Fax: +64.3.479.7527
E-mail:
Jean Braun Nick Oliver rick.sibson@stonebow.otago.ac.nz
Research School of Earth Sciences Economic Geology Research Unit
The Australian National University School o f Earth Sciences Richard M. Tosdal
Canberra, ACT 0200 James Cook University Mineral Deposit Research Unit
Australia Townsville, QLD 4811 University o f British Columbia
Tel. +61.2.6125.5512 Australia Vancouver, BC V6T 1Z4
Fax: +61.2.6125.5443 Tel. +61.7.07.4781.5049 Canada
E-mail: Jean.Braun@anu.edu.au Fax: +61.7.07.4725.1501 Tel. +604.822.3149
E-mail: Nick.OHver@jcu.edu.au Fax: +604.822.6088
Stephen F. Cox E-mail: rtosdal@cos.ub.ca
Departm ent o f Geology and Alison Ord
Research School o f Earth Sciences CSIRO Division of Phaedra Upton
The Australian National University Exploration and Mining Department of Geology
Canberra, ACT 0200 PO Box 437 University o f Otago
Australia Nedlands, WA 6009 PO Box 56
Tel. +61.2.6125.0045 Australia Dunedin
Fax: +61.2.6125.5544 Tel. +61.8.9389.8421 New Zealand
E-mail: sfcox@geology.anu.edu.au Fax: +61.8.9389.1906 Telephone: +64.3.479.7519
E-mail: a.ord@dem.csiro.au Fax: +64.3.479.7527
Thomas G. Hildenbrand E-mail:
U.S. Geological Survey K Howard Poulsen phaedra.upton@stonebow.otago.ac.nz
MS 989, 345 Middlefield Rd. 34 Walford Way
Menlo Park, CA 94025 Nepean, ON K2E 6B6 Richard K. Valenta
USA Canada MJM Exploration
Tel. +650.329.5303 Tel. +613.723.1766 GPO Box 1042
Fax: +650.329.5313 Fax: +613.723.7192 Brisbane, QLD 4001
E-mail: tom@usgs.gov E-mail: hpoulsen@attcanada.ca Australia
Tel. +61-7-3214.9245
Robert C. Jachens Jeremy P. Richards Fax: +61-7-3214.9133
U.S. Geological Survey Department of Earth E-mail: rkvalem@mim.com.au
MS 989, 345 Middlefield Rd. and Atmospheric Sciences
Menlo Park, CA 94025 University o f Alberta
USA Edmonton, AB T6G 2E3
Tel. +650.329.5300 Canada
Fax: +650.329.5313 Tel. +780.492.3430
E-mail: jachens@usgs.gov Fax: +780.492.2030
E-mail: Jcremy.Richards@ualberta.ca
Mark A. Knackstedt
Research School of Francois Robert
Physical Sciences and Engineering Barrick Gold Corporation
The Australian National University 7257 Dunver
Canberra, ACT 0200 Verdun, QC H4H 2H6
Australia Canada
Tel. +61.2.6125.2495 Tel. +514.732.0021
Fax: +61.2.6125.1884 Fax: +514.732.0021
E-mail: makllO@ rsphyl.anu.edu.au E-mail: frobert@bai rick.eom

u
BYRON R. BERGER graduated from Occidental College, Test Site, with the purpose of examining environmental
Los Angeles, with a B.A. degree in economics, and from die restoration issues, (3) study of crustal structure and funda­
University of California at Los Angeles with an M.S. degree in mental geologic processes related to earthquake and vol­
geology. From 1971 to 1977, he worked in the Minerals canic hazards in die central and western United States (pri­
Exploration Department and Minerals Exploration Research marily the New M adrid seismic zone and rift systems on
Division, Research and Development Departm ent, of the volcanic islands such as Hawaii), and (4) determination of
Continental Oil Company. In 1977 he joined the U.S. Geo­ die location, attitude, and displacement history of potentially
logical Survey, where he has been a research geologist and dangerous, often concealed faults in the San Francisco and
has served as Chief, Branch o f Geochemistry (1983-1988) Los Angeles areas.
and Deputy Chief, Office of Mineral Resources (1988-1992).
He is presendy a member of die Crustal Imaging and Char­ ROBERT JACHENS received his B.S. degree in geology
acterization Team and conducts research on (1) thermal, from San Jose State University and his M.S. and Ph.D.
mechanical, and hydraulic coupling in ore formation, (2) degrees in geophysics from Columbia University. He was a
the environmental geology and geochemistry of mineralized research associate, Lamont-Doherty Geological Observa­
areas, and (3) the development of tectonic and structural tory, from 1972 tol976, and has been a geophysicist with die
models of epizonal hydrothermal mineral deposits for min­ U.S. Geological Survey from 1976 to the present. Jachen's
eral resource assessment applications. primary interest is in the application of gravity and magnetic
techniques to die solution o f geologic problems. Work has
JEAN BRAUN is a fellow in the Geodynamics Group at die included studies of tidal gravity, temporal gravity changes
Research School of Earth Sciencesat the Australian National associated with tectonic and volcanic deformation, earth As­
University. His research specializes in computational model­ suring due to groundwater extraction, regional geophysics
ing o f geodynamic processes, including fluid transport in in mineral resource assessments, plate tectonics of western
permeable media. He holds a B.Sc. degree in physics from North America, continent-scale relationships between grav­
die University of Liege, and a Ph.D. degree in oceanography ity and upper-crustal geology, and die tectonic evolution of
from Dalhousic University. the California coast ranges. His m ost recent work has
focused on die use of regional gravity and magnetic anom­
STEPHEN COX is professor of structural and economic alies to understand the structure o f and offsets on faults of
geology, jointly in the D epartm ent of Geology and the the San Andreas system, applications o f geology and geo­
Research School of Earth Sciences at die Australian National physics to defining the hydrogeologic framework of aquifer
University. His research interests are primarily in the cou­ systems, and the creation and visualization o f 3-D geologic
pling between deform ation processes and fluid flow in maps.
crustal regimes, widi applications to ore genesis and crustal
mechanics. His research is pursued via field-based studies, MARKKNACKSTEDT is an Australian Research Council
microstructural and microchemical analyses, and high-pres- Queen Elizabeth II Fellow in the D epartm ent o f Applied
sure/high-tem perature experim ental studies. He holds a Mathematics a t the Research School o f Physical Sciences
B.Sc. (Hons) degree from the University of Tasmania and a and Engineering at the Ausu alian National University. His
Ph.D. degree from Monash University. research interests include multiphase flow in porous media,
application o f percolation theory to the geological sciences,
THOMAS G. HILDENBRAND is a geophysicist with the U.S. and characterization o f disordered materials. He holds a
Geological Survey in Menlo Park, California. He received his bachelor’s degree from Columbia University and a Ph D.
B.S. degree in engineering physics and his M.S. and Ph.D. degree from Rice University, boil) in chemical engineering.
degrees in engineering geoscience from the University of
California at Berkeley. He began his career at the U.S. Geo­ STEVE LUDINGTON received his B.S. degree in geology
logical Survey as a postdoctoral fellow from 1975 to 1977. from Stanford University in 1967, and a Ph.D. degree in geol­
From 1984 to 1990, he served as Section Chief and then as ogy from the University of Colorado in 1974. He worked for
Chief of the Branch of Geophysics, an interdisciplinary team the mineral indusu y from 1969 to 1974, doing exploration
devoted to die theoretical development and applicadon of in the central and southern Rocky Mountains. In 1974, he
potendal-field, elecuical, and remote sensing techniques. As joined the U.S. Geological Survey. He has worked for the
a research scientist, Hildenbrand is interested in interpreting Survey in Denver, Colorado, and Reston, Virginia, and since
potential-field data by developing geologic models address­ 1987, he has been in Menlo Park, California. Until the mid-
ing the origin and evolution of crust in various tectonic set­ 1980s, Ludington was a specialist, interested in Climax-type
tings. Recent research activities have been directed toward molybdenite deposits and other mineral deposits associated
extracting geologic information from potential-field data in with evolved granites. More recently, he has participated in
the ( 1 ) assessment of mineral resources on local and regional mineral resource studies in Costa Rica, Venezuela,
regional scales in western United States, (2) investigation of and Bolivia, and in Nevada. He coordinated the National
local structures affecting groundwater flow at the Nevada Mineral Resource Assessment o f the U.S. Geological Survey
for precious and base metals. At present, he is engaged in vey of Canada, he is currently a consulting economic geolo­
using regional geochemisu'y to elucidate the tectonic frame­ gist who specializes in structural geology of ore deposits. His
work of the northern Great Basin. research has focused mainly on problems of significance to
the global mineral exploration and he has authored or co-
NICK OLIVER is professor of economic geology and Direc­ authored 55 journal papers and reports. He served as an
tor of the Economic Geology Research Unit at James Cook associate editor for Economic Geology from 1993 to 1998 and is
University, a position he has held for four years. Previously he a fellow of the Society of Economic Geologists as well as a
held academic posiuons at Curdn University, Monash Uni­ member o f the Association of Geoscientists o f Ontario. He
versity, and research posiuons at CSIRO and the Carnegie currently teaches a graduate-level course in structural geol­
Institute Geophysical Laboratory, where his research ogy and ore deposits for the mineral exploration program at
involved analysis o f large hydrotherm al systems through Queen’s University.
combinations o f petrology, geochemistry, and structural
geology. He has a B.Sc. (Hons) degree from the University of JEREMY P. RICHARDS received his B.A. degree (1983) in
Queensland and a Ph.D. degree from Monash University. geology from the University o f Cambridge, UK, his M.Sc.
His research experience covers a broad range o f deposit degree (1986) in economic geology from the University of
types and geological settings, from banded iron formations Toronto, Canada, and a Ph.D. (1990) in economic geology
through various epigenetic base metal and gold deposits, from die Australian National University. After holding a two-
regional alteration patterns in the Mt. Isa block, Yukon gran­ year postdoctoral position at die University of Saskatchewan,
ite-related gold, and greenstones and metamorphosed mas­ Canada (1990-1992), he joined the faculty at the University
sive sulfides in Finland. With collaborators in CSIRO and die of Leicester, UK, and was Course Director of the Mineral
minerals industry, he has applied numerical modeling to Exploration M.Sc. program from 1994 to 1997. He joined
assist with interpretation of field and mine patterns for most die University of Alberta in 1997, and is currently associate
of these districts. He is a fellow of the Society of Economic professor of economic geology. Research interests have
Geologists and was awarded the 1999 Mineralium Deposit# spanned a wide range in both pure and economic geology,
Reviewer of die Year. and include investigadons o f Keweenawan and central
African copperbelt geology, alkalic-type epitherm al Au
ALISON ORD received a B.Sc. (Honours) degree from the deposits in Papua New Guinea and British Columbia, por­
University of Edinburgh in 1977, and a Ph.D. degree from phyry Cu-Mo deposits in Ontario, and in Chile, Argentina,
the University of California at Los Angeles in 1981. From and Ecuador, and volcanology and structural geolog)' of the
1981 to 1984, she held a postdoctoral research fellowship at Central Andes. In the face of the current downturn in metal
Mon ash University, studying the effect of defect chemistry on prices, his most recent foray is into sand and gravel resources
the strength of quartz at high pressures and temperatures. in Alberta. Actually, there is a lot o f gold in those gravels!
She joined the Commonwealth Scientific and Industrial
Research Organisation o f Australia (CSIRO) in 1984 as a FRANCOIS ROBERT is a graduate o f Ecole Polytcchnique
research sciendst in the division of Geomechanics. She is in Montreal, where he received his B.Sc. (1978), M.Sc.
now a chief research scientist in the division of Exploration (1980) and Ph.D. (1983) degrees. His doctoral work focused
and Mining, leading the research group, Structural Controls on the geology, structure, and hydrothermal characteristics
on Mineralisation. H er principal areas o f research are explo­ of the Sigma gold deposit in Abitibi. Following one-year post­
ration o f the fundamental controls o f mechanics, solid and doctoral studies at the Univeraty o f Michigan, where he stud­
fluid, on the form ation o f giant ore deposits, including ied fluid inclusions o f the Sigma deposit, he joined the Geo­
numerical modeling of deformation-fluid flow-thermal trans­ logical Survey of Canada, in 1985, as a research scientist with
port-chemical reaction feedback systems. She has interests the mineral deposits group. His research focused largely on
also in 3-D vision systems for the mining and mineral pro­ the geology and structure o f gold deposits and districts in
cessing industries, particularly remote and automatic systems metamorphic terranes, in Canada and abroad. He worked
for determ ining joint structures in pit walls and fragment closely with industry and participated in a num ber o f mis­
size distributions on muck piles and conveyor belts. She is on sions for international agencies and governments. He
the editorial board o f the AustralianJournal ofEarth Sciences received the Society o f Economic Geologists WaldemarLind-
and is a member of the Geological Society of Australia. gren Award in 1990 for his work on gold deposits. In 1997,
he joined Barrick Gold Corporation as Senior Research
HOWARD POULSEN was bom and raised at Thunder Bay, Geologist. In addition to his other duties, he serves as in-
Ontario, and attended the University o f Waterloo (Honors housc technical consultant on different projects worldwide.
B.Sc. degree, physics, 1970), Lakehead University (M.Sc.
degree, structural geology, 1980) and Q ueen’s University RICHARD (RICK) SIBSON graduated with a B.Sc. (Hons)
(Ph.D. degree, economic geology, 1984). With more than 25 degree in geology from the Univeraty of Auckland, and M.Sc.
years o f cumulative research experience at Lakehead Unver- and Ph.D. degrees from Imperial College, London. He
sity, the Ontario Geological Survey, and the Geological Sur­ taught at Imperial College (1973-1982) and at the University
of California at Sania Barbara (1982-1990) before returning PHAEDRA UPTON received a B.Sc. (Hons) degree in
to New Zealand as professor of geology at the University of chemistry from Canterbury University in 1989, and a Ph.D.
Otago (1990-2001). His research focuses on the structure of degree in geology from the University of Otago in 1995. She
crustal fault zones and the mechanics of shallow crustal earth­ was a postdoctoral fellow with CSIRO from 1996 to 1998.
quakes. Increasing recognition of fluid involvement in fault­ Since 1998 she has been a Foundation o f Research, Science
ing has led to an interest in mechanisms for fluid redistribu­ and Technology postdoctoral fellow at the University o f
tion around seismicaily active faults, and structural controls Otago. H er research is focused on the geodynamics of, and
on the development o f fault-hosted hydrothermal mineral­ the fluid flow systems associated with, convergent plate
ization. He is author o r co-author o f more than 60 research boundary settings. The majority of her work has been based
papers and is an elected fellow o f the Geological Society of around the Southern Alps of New Zealand. She uses geody­
London, the Geological Society o f America, the American namic modeling, fieldwork, and stable isotope geochemistry
Geophysical Union, and the Royal Society o f New Zealand. in her studies of coupled deformation and fluid flow. She is
a m em ber o f the Geological and Geophysical Societies of
RICHARD TOSDAL received a B.A. degree from the Uni­ New Zealand and of the American Geophysical Union.
versity of California at Santa Barbara, an M.Sc. degree from
Queen’s University in Kingston, Ontario, and a Ph.D. degree RICK VALENTA is Chief Geologist of M ount Isa Mines
from the University of California at Santa Barbara. Tosdal Exploration and has been with MIM for over six yews. Prior
has worked for the U.S. Geological Survey and as an inde­ to that he lectured in geophysics, image interpretation, and
pendent geologist for the mining industry. He currently is structural geology at Monash University and also consulted
the Director of the Mineral Deposit Research Unit at the for MIM, World Geoscience, and Acrodata. He has a B.Sc.
University of British Columbia in Vancouver, where he super­ (Hons) degree from the University of New Brunswick and a
vises research groups invesdgadng intrusion-related Au Ph.D. degree from Monash University (specializing in defor­
deposits, Carlin-type sedimentary rock-hosted deposits, mation and ore genesis at the Hilton Pb-Zn mine near Mt.
epithermal systems, and pericratonal VMS deposits, and is Isa). Rick has been instrumental in regional and local area
developing projects in die diamond-related fields and in car­ selection and geological synthesis of MIM projects in Aus­
bonate terranes. He has worked on geologic, tectonic, and tralia, Africa, and South America. He is well known across
metallogenic problems in the cordilleras and cratons of Ausualia as an expert in integrated geological and geophysi­
North and South America. His personal research interests cal approaches to project generation and terrain interpreta­
include the role o f structure and tectonics in ore genesis and tion. H e is involved with several collaborative research pro­
magmatism, the role of climate in ore genesis, and the jects, including one on geophysical interpretation, structure,
source of metals in ore deposits. and ore genesis in the Ernest Henry Cu-Au district
 PREFACE
The spark for this volume arose am id the euphoria of For those readers who seek more discussion o f this aspect
completing Techniques in Hydrothermal Ore Deposits Geology o f the discipline, we provide below a Table o f Events for the
(Reviews in Economic Geology, v. 10), and the accompany^ construction o f a Reviews in Economic Geology volume:
ing Society of Economic Geologists-sponsored short course
at the 1998 Geological Society of America Annu;il Meeting in Event Reality
Toronto. Dick Tosdal noted the lack o f any structural geol­
Protolith formation Volume editors dream up concept;
ogy in the Techniques volume, and I glibly replied that this
inrite authors to submit chapter
could be the subject o f a whole Reviews volume on its own.
outlines
Several espressos later we shook hands on a new project, this
one also to be presented in Toronto, bu t as part o f the D, Editors tell authors what they
Prospectors and Developers Association of Canada Conven­ actually wanted
tion in March 2001. Unconformity Time passes...
The reason that Dick's comment triggered such precipi­ Sj, M, Manuscripts prepared by authors
tate action was that we recognized, both in ourselves and in regardless, and sent ou t for review
the wider economic geology community, a lack o f clear
understanding about this fundamental aspect of our disci­ Unconformity Time passes...
pline. T h e m antra o f hydrogcologists and hydrotherm al Reviews received
ore deposits geologists alike is “permeability, permeability, S, Reviews considered
permeability," and one o f the best ways to enhance perme­ Unconformity Time passes...
ability is by fracturing rock. This we all know, but for many
o f us, formal training in structural geology ceased in our M, Some changes made and revised
undergraduate years. What this volume aims to do is to pro- versions submitted
ride a refresher on basic aspects o f structural geology and Mineralization Revised manuscripts accepted by
its control on fluid flow, and also to review some o f the editors
m ore sophisticated models that can be used to predict DS,M , Copyediting and production
paleo fluid flow paths—and therefore, zones o f potential
Uplift Publication of volume
mineral deposition—from structural and rheological data.
The closing chapter attempts to place deposit-scale struc­ Erosion No! Volume withstands the test of
tural controls on ore deposition in the context o f regional time...
crustal stress fields and plate tectonic models, using por­
phyry deposits as an example. No volume such as this can be prepared without the ded­
The volume is intended for economic geologists whose icated efforts of the various authors, and Dick and I thank
specialization is not structural geology, and should appeal the 14 contributors who stayed the course. We also thank
to a wide audience from academia and industry, including David John, who guest-edited our own contribution, and all
students. It is not intended to supplant formal instruction the reviewers who provided timely and incisive critiques
in this field, but instead aims to renew interest in structural that have, we hope, kept us honest—their names are listed
approaches to ore deposits geology, and to showcase some below. Production of this volume would not have been pos­
of the developments that allow the formulation of predic­ sible w ithout the untiring efforts of Lisa Laird and AJice
tive models for deposit localization. Bouley at SEG, and Lee Ewert, who expertly handled the
Perhaps surprisingly—and this may be a relief to some copyediting.
people—only one chapter alludes to multiple deformation
events, the staple of many classic structural geology courses. Jerem y P. Richards
Richard M. Tosdal

Reviewers
Jennifer Adams Michel Jebrak
Anne-Marie Boullier David John
Ron Bruhn Robert Kerrich
Mark Duffett John Ridley
David Hill Tom Schroeter
Jay Hodgson Stephen Temperlcy
Carol Finn Ted Theodore
 CONTENTS
Chapter 1—Principles o f Structural Control
on Permeability and Fluid Flow in
Hydrothermal Systems
S. F. Cox, M. A. Knackstedt, andJ. Braun
Absu-act............................................................................. 1
In tro d u c tio n .................................................................... 1
Principles of Fluid Flow in Porous and
Fractured Rocks ........................................................... 2
Principles o f Macroscopic Fracture F orm ation......... 10
Evolution of Flow Pathways
during D e fo rm a d o n ...................................................19
Conclusions....................................................................22
Acknowledgm ents......................................................... 22
R eferences...................................................................... 22
Chapter 2—Seismogenic Framework for
Hydrothermal Transport and
O re Deposition
Richard H. Sibson
A b stract...........................................................................25
Introduction ..................................................................26
Seismogenic C ru st......................................................... 26
Tectonic Stress and Fluid Pressure R egim es...............31
Faults, Fractures, and Fluid Pressure L evels..............35
Dynamic Structural Permeability ........................38
Fluid Redisuibution Mechanisms
Linked to the Earthquake Stress C ycle.................... 41
Summary D iscussion.....................................................47
Acknowledgm ents......................................................... 47
R eferences...................................................................... 47
Chapter 3—Deformation, Fluid Flow,
and Ore Genesis in Heterogeneous Rocks,
with Examples and Numerical Models
from the Mount Isa District, Australia
Nicholas H. S. Oliver, Alison Old,
Richard K. Valenla, and Phaedra Upton
A b stract.......................................................................... 51
Introduction ..................................................................51
Principles o f Fluid Flow .............................................. 52
Deformation and Fluid Flow ........................................54
Heterogeneous Deformadon and Fluid F low ........... 54
Numerical M odels......................................................... 56
Exam ples........................................................................ 57
D iscussion...................................................................... 71
Acknowledgments......................................................... 72
References...................................................................... 72
vu
Chapter 4—Utility o f Magnetic and
Gravity Data in Evaluating Regional
Controls on Mineralization:
Examples from the Western United States
T. G. Hildenbrand, Byron Bergen
R. C. Jachens, and Steve Ludington
A b stra c t...........................................................................75
Introduction ..................................................................75
Potendal-Fieid Data and M ethods............................... 76
Mineralized Areas in the Western United States . . .85
Relationships of Geophysical Structures
and Deposit P attern s...................................................90
Conclusions and Summary ......................................104
Acknowledgm ents....................................................... 106
R eferences.................................................................... 106
Chapter 5—Vein Formation and
Deformation in Greenstone Gold Deposits
Fmnfois Robert and K. Howard Poulsen
A b stra c t.........................................................................I l l
Introduction and Scope ............................................ I l l
Structure of Greenstone Gold Districts .................. 112
Veins and Their A nalysis............................................ 118
Classification of V e in s .................................................124
Vein Deformation and Superimposed Strain ......... 138
Structural Timing o f V e in s ........................................ 146
Discussion o f Concluding Rem arks........................... 150
Acknow ledgm ents....................................................... 152
R eferences.................................................................... 152
Chapter 6—Magmatic and Structural
Controls on the Development of
Porphyry Cu ± Mo ± Au Deposits
R. M. Tosdal and J. P. Richards
A b stract.........................................................................157
Inuoducuon ................................................................ 157
General Porphyry Model .......................................... 158
Convergent Plate Margin M agmadsm ...................... 165
Volcanoes and Shallow-Level
Intrusive Complexes: The
Environment o f Porphyry Cu D eposits.................. 166
Structural Sequence in Porphyry Systems ..............171
Origin of Breccia Pipes in Porphyry Cu Deposits . .173
Tectonic Setting o f Porphyry Cu Deposits ..............174
Concluding R em arks...................................................175
Acknowledgm ents....................................................... 177
R eferences.................................................................... 177
Society of Economic Geologists
Reviews v. 14, 2001, p. 1-24

Chapter 1

Principles o f Structural Control on Permeability and

Fluid Flow in Hydrothermal Systems
S. F. Cox,+
Centre fo r Advanced Studies o f Ore Systems, Department o f Geology a n d Research School o f Earth Sciences,
The Australian N ational University, Canberra, A C T 0200, Australia

M . A . Kn ack sted t,

Research School o f Physical Sciences and Engineering, The Australian N ational University, Canberra, A C T 0200, Australia

a n d J. B raun
Research School o f Earth Sciences, The Australian N ational University, Canberra, A C T 0200, Australia

Abstract

Fluid pathways between metal sources and sites of ore deposition in hydrothermal systems are gov­
erned by fluid pressure gradients, buoyancy effects, and the permeability distribution, Structural controls
on ore formation in many epigenetic systems derive largely from the role that deformation processes and
fluid pressures play in generating and maintaining permeability within active faults, shear zones, associ­
ated fracture networks, and various other structures at all crustal levels.
In hydrothermal systems with low intergranular porosity, pore connectivity is low, and fluid flow is typ­
ically controlled by fracture permeability. Deformation-incluced fractures develop on scales from microns
to greater than hundreds of meters. Because mineral scaling o f fractures can be rapid relative to the life­
times o f hydrothermal systems, sustained fluid flow occurs only in active structures where permeability is
repeatedly renewed.
In the brittle upper crust, deformation-induced permeability is associated with macroscopic fracture ar­
rays and damage products produced in episodically slipping (seismogenic) and aseismically creeping faults,
growing folds, and related structures. In the more ductile mid- to lower crust, permeability enhancement is
associated with grain-scale dilaiancy (especially in active shear zones), as well as with macroscopic hydraulic
fracture arrays. Belotv the seismic-aseisntic transition, steady state creep leads to steady state permeability
and continuous fluid flow in actively deforming structures. In contrast, in the seismogenic regime, large
cyclic changes in permeability lead to episodic fluid flow in faults and associated fractures.
The geometry and distribution o f fracture permeability' is controlled fundamentally by stress and fluid
pressure states, but may also be influenced by preexisting mechanical anisotropies in the rock mass. Frac­
ture growth is favored in high pore fluid factor regimes, which develop especially where fluids discharge
from faults o r shear zones beneath low-permeability flow barriers. Flow localization within faults and shear
zones occurs in areas o f highest fracture aperture and fracture density, such as damage zones associated
with fault jogs, bends, and splay’s. Positive feedback between deformation, fluid flow, and fluid pressure
promotes fluid-driven growth o f hydraulically linked networks o f faults, fractures, and shear zones.
Evolution o f fluid pathways on scales linking fluid reservoirs and ore deposits is influenced by the rel­
ative proportions o f backbone, dangling, and isolated structures in the network. Modeling o f the growth
o f networks indicates that fracture systems reach the percolation threshold at low bulk strains. Just above
the percolation threshold, flow is concentrated along a small proportion of the total fracture population,
and favors localized ore deposition. At higher strains, flowr is distributed more widely throughout the frac­
ture population and, accordingly, the potential for localized, high-grade ore deposition may be reduced.

Introduction monly localized wi thin faults, shear zones, o r associated frac­

T he Formation of many types of epigenetic ore deposits
involves some form of control by structures produced dur­
ing crustal deformation. In particular, ore deposition is com-
- ’Corresponding au th or e-mail, sfcox@gcology.ami.cdu.au
ture systems. Additionally, mineralization can be restricted
to particular parts of faults o r shear zones such as jogs or
bends (Hulin, 1929; Newhouse, 1942; McKinstry, 1948; Sib-
son, 1987), to fold-related structures such as saddle reefs
(Hulin, 1929; Chace, 1949; Cox et ah, 1991), o r related to
deform ation around heterogeneities such as com petent
rock masses (Oliver etal., 2001). Epigenetic mineralization
also is typically localized along low displacement faults and
fracture systems (e.g., Robert et al., 1995; Cox, 1999).
Hydrothermal mineral deposits are produced by focused
flow o f large volumes o f fluids (Henley et al., 1985; Fyfe,
1987; Cox, 1999). Mass balance calculations, based for
example on silica and metal solubility, indicate that time-
integrated fluid volumes in many types of hydrothermal ore
systems are typically greater than several cubic kilometers.
Crustal deform ation processes influence fluid flow in
hydrothermal systems via controls on rock penneability and
the driving forces for fluid flow. Undei-slanding how defor­
mation processes and structures influence permeability evo­
lution is a key aspect to understanding the architecture of
fluid pathways in hydrothermal systems, and die factors con­
trolling where ore deposits form along these pathways. This
review explores how active deformation controls the permc-
ability of rock masses through the formation of macroscopic
fracture systems, damage zones, and grain-scale microcrack
permeability within actively deforming structures. The role
of fluid pressures and stress regimes in driving permeability
enhancem ent and growth o f perm eable fau lt/fractu re/
shear networks is highlighted. We outline how stress regimes
and variations in fluid pressure in hydrothermal systems gov­
ern the location and geometry o f fluid pathways between
metal sources, fluid sources, and sites o f ore deposition. We
also examine how linkages among structures develop during
progressive deformation, and how they control the architec­
ture o f fluid pathways from deposit to crustal scales.
Principles of Fluid Flow in Porous and Fractured Rocks
Fluid m igration through the E arth’s crust occurs in
response to various driving forces. The fluid flux and geom­
etry of flow is fundam entally controlled by permeability
variations in the crust. In this section, we outline the forces
driving crustal fluid flow and examine the dependence of
fluid flux on rock permeability. We also discuss how the evo­
lution of grain-scale to macroscopic permeability in
hydrothermal systems is influenced by stress regimes and
fluid pressure regimes during both brittle and ductile styles
of deformation. Two central points are highlighted: firstly,
the geometry of fluid pathways is strongly dependent on
the geometry and style of deformation; and secondly, per­
meability is a transient rock property that can be rapidly
reduced by porosity-destruction processes such as com­
paction and sealing of intergranular pores and fractures.
Maintenance of permeability and fluid flow in hydrother­
mal systems is, therefore, dependent upon active deforma­
tion repeatedly regenerating permeability. Accordingly,
fluid flow is localized within structures that were active dur­
ing the operation of hydrothermal systems.
D riv in g forces f o r f l u i d flo w
Fluid flow occurs in response to two major classes of dri­
ving forces: (1 ) pressure-driven flow, and (2) buoyancy-
driven flow.
Pressure-driven flow arises from a num ber of causes. Par­
ticularly at crustal depths less than several kilometers, rock
permeability is generally high enough that pore spaces are
highly interconnected and fluid pressures are close to
hydrostatic values (i.e., owing to the weight of the fluid col­
um n). In this regime, topographic relief is a m ajor factor
establishing hydraulic gradients that drive subsurface flow
downward from regions o f high topography towards
regions o f lower topography. Depths o f topographically
driven fluid penetration are typically up to several kilome­
ters (Forster and Smith, 1990).
Transient, deformation-induced pore pressure changes
that set up vertical and lateral hydraulic gradients can be
im portant at all crustal depths. These changes arise from
compaction o f intergranular pore spaces during burial or
regional deformation, grain-scale dilatation associated with
granular flow, or from microscopic to macroscopic crack
growth and closure during deform ation. Such deform a­
tion-induced pore pressure gradients are associated with
both small elastic responses (poroelastic effects) or larger
inelastic deform ations o f intergranular pore spaces and
fractures owing to stress changes associated with the seismic
cycle and heterogeneous deform ation (Muir-Wood and
King, 1993). In particular, macroscopic dilatancy associated
with episodic fault slip exerts a powerful control on fluid
migration around active faults in the crustal scismogenic
regime (Sibson, 1987,1993, 2001). Transient fluid migra­
tion in response to pressure gradients due to deformation-
induced grain-scale dilatancy during regional deformation
has been discussed by Cox and Etheridge (1989) and
McCaig and Knipe (1990), and modeled by O rd and Oliver
(1997) and Oliver e ta l. (2001).
Driving pressure-gradients within the crust are also gen­
erated by development o f suprahydrostatic-pressurcd fluid
reservoirs. For example, suprahydrostatic fluid pressuriza­
tion can be associated with emplacement of magmas into a
cooler, porous, fluid-saturated rock-mass. Transient driving
pressure-gradients are also associated with m etam orphic
fluid-procluction and fluid-absorption reactions (Walther,
1990; Rumble, 1994), as well as by fluid expulsion associ­
ated with crystallization of hydrous silicate melts. Modeling
of two-dimensional advective flow regimes associated with
drainage of suprahydrostatic reservoirs has been conducted
by Matthai and Roberts (1997). In these flow regimes, spa­
tial variations in permeability (for example, owing to the
presence of active and perm eable faults or shear zones)
establish hydraulic gradients that focus fluid flow and gov­
ern flow pathways (Fig. 1). Temporal and spatial variations
in permeability induced by episodic fault rupture particu­
larly influence the dynamics of fluid flow in pressure-driven
flow regimes in the seismogenic upper crust (Sibson et al.,
1988; Cox, 1999; Sibson, 2001).
Buoyancy drive for crustal fluid flow arises from vertical
variations in fluid density, either through effects of tempera­
ture gradients or variable concentrations of dissolved species
(e.g., salinity) in pore fluids. Decreasing density of hydro-
thermal fluid with increasing tem perature and depth for
many pore fluid compositions leads to gravitational instabil­
ity and convection of fluid through porous media in near-
hydrostatic fluid pressure regimes. An excellent example of

Fig. 1. Numerical (finite element) models simulating steady state, pres­
sure-driven fluid flow patterns around permeable faults or shear zones
embedded in a less permeable medium. A vertical lithostntic fluid pressure
gradient is maintained in the medium away from the fault. Length o f flow
vectors corresponds to flow velocity; flow vectors within faults not shown,
Contours indicate departures o f fluid pressure from liihostatic values: light
areas arc below and dark areas arc above liihostatic pressure, a. Simple pla­
nar fault or shear zone with a permeability 10s limes that o f the surround­
ing host rock matrix. Note fluid focusing at the upstream (lower) part o f
the structure, and fluid discharge around the downstream (upper) part of
the structure, b. Fluid flow patterns associated with the presence o f a fault
stepover region. Permeability and imposed fluid pressure gradients arc
the same as in (a).
this is provided by thermally driven convection around hot
intrusive complexes, as m odeled by Norton and Knight
(1977), Forster and Smith (1990), and Cathles et al (1997).
Flow in porous media
In addition to a driving force, crustal fluid flow also
requires a rock mass with pore spaces to contain fluid, as
well as connectivity between pore spaces. The pore spaces
can be intergranular pores o r grain-scale to macroscopic
fractures. The one-dimensional macroscopic flux o f a sin­
gle-phase fluid in a horizontal plane through an isotropic,
porous medium is described by a simplified form o f Darcy’s
law, which states that
q = Q /At = k / fj(dP/dx), (1)
where
q is the fluid flux (i.e., the volume Q o f fluid traversing
cross-sectional area A perpendicular to the x axis, per
unit time t);
k is the permeability o f the porous medium;
t) is the kinematic viscosity o f the pore fluid; and
d P /d x is the horizontal fluid pressure gradient that drives
flow (Bear, 1972; Fig. 2).
T he volume flux has dimensions of velocity, and is some­
times referred to as the Darcian velocity. Note, however,
that the Darcian velocity is not the actual velocity at which
the fluid moves through the pore space. Rather, it is a mea­
sure o f the fluid volume per unit time moving through unit
cross-sectional area o f the rock. It is related to the real aver­
age fluid velocity, v, by the relationship,
q = v-Q. (2)
Particularly in low-porosity rocks, the real fluid velocity at
the pore scale can be much larger than the Darcian veloc­
ity. Darcy’s law assumes laminar flow, and applies in porous
rocks at flow rates up to about 1 m /s. At higher flow rates,
turbulence and high inertial forces lead to breakdown of
Darcy’s law (Gueguen and Palciauskas, 1994).
Permeability has units m2 (1 darcy = 10-12 m2), and is an
intrinsic rock property quantifying the capacity of fluids to
pass through rock. Permeability is influenced by the con­
nectivity between intergranular pore spaces o r fractures in
a rock, and is particularly sensitive to the minimum throat
size between connecting pores. Natural rocks have perme­
abilities that range over more than ten orders of magnitude
(Brace, 1990; Manning and Ingebritsen, 1999). For exam­
ple, porous sandstone (porosity, <f>>15%) and poorly com­
pacted tuffs can have perm eabilities as high as 10"12 m2,
whereas unfractured, “tight” metamorphic or igneous rocks
can have permeabilities less than 10-22 m2. Fluid fluxes and
the geometry of flow are, therefore, particularly dependent
on spatial and time variations in permeability within the
crust. We examine below the critical effect o f deformation
processes in generating large, although commonly tran­
sient, changes in rock permeability.

DARCY’S LAW
Q/At = k/r|(dP/dx)
Fic. 2. Darcy’s law and control o f horizontal fluid flow rate by driving
pressure gradient, rock permeability, and fluid viscosity.
The Darcy equation also illustrates how fluid flux is
inversely proportional to the kinematic viscosity of the pore
fluid. At temperatures between 100° and 800°C, and pres­
sures between 50 and 300 MPa, the viscosity of water ranges
over about one o rder o f m agnitude (40-400 ji/Pa-s),
decreasing with increasing tem perature and decreasing
pressure (H aaretal., 1984). Viscosity increases with increas­
ing salinity (Garven and Freeze, 1984). The extreme range
in permeability in rocks compared with the much smaller
range in viscosity in typical ore fluids, means that variations
in permeability usually have a greater impact on fluid flux
than variations in viscosity in simple water-dominated
hydrothermal systems. The dynamics of two-phase flow in
porous media can lead to complex effects that will not be
treated in this review (Sahimi, 1994).
For fluid flow with a vertical com ponent o f m otion (z
axis), the driving force for flow is no t the absolute fluid
pressure gradient, but the nonhydrostatic com ponent of
the gradient. An expanded expression for Darcy’s law is
given by the relationship,
where
qi is the fluid flux in die coordinate direction i;
ky is die permeability tensor (allowing for anisotropy of per­
meability);
p is the fluid density; and
g is the gravitational acceleration (Forster and Smith,
1990);
dP/dxj represents die driving force due to gradients in fluid
pressure, whereas the p%(di/<kj) term accounts for gravity
and also buoyancy effects associated with spadal variations
in fluid density. This full Darcian relationship illustrates
that the driving force for fluid flow is the difference
between the fluid pressure gradient and the hydrostatic
com ponent o f the fluid pressure gradient. This result is
illustrated by considering fluid pressure variations in a ver­
tical colum n o f fluid in a porous medium with fully con­
nected porosity. The depth dependence of fluid pressure Pr
in a column o f fluid at rest is given by
Pf^Pgz. (4)
where z is the depth below the surface.
In this case, the vertical fluid pressure gradient, pg,
owing to the weight o f the fluid column docs not drive flow,
and is known as the hydrostatic fluid pressure gradient.
Deviations in d P /d z from pg are required to drive vertical
flow. Accordingly, for vertical flow o f a uniform density
fluid, Darcy’s law reduces to
q = ( k /7j)* (d P /d z-p g ). (5)
Where d P /d z is greater than pg, upwards flow occurs,
whereas if dP /dz is less than pg, downwards flow occurs in a
fluid of uniform density. For general pressure-driven flow in
rocks with isotropic permeability, flow is parallel to the non-
hydrostatic com ponent o f the fluid pressure gradient. Per­
meability anisou opy may result in flow that is not parallel to
the driving pressure gradient
Types ofporosity
Porosity, is the ratio o f the volume of void space to the
bulk volume o f a porous medium,
^ = Vpora/(V pores + VslJids). (6)
Permeability is not always simply related to porosity. It is
dependent on various factors, including pore diam eters
and shapes, diameters o f pore throats (which connect large
intergranular pores), and connecdvities between pores.
From the standpoint o f flow through a porous medium, the
interconnected pore volume is o f interest.
Porosity usually comprises intergranular porosity and
fracture porosity. Intergranular porosity, such as primary
pore spaces in poorly cemented o r weakly compacted clas­
tic sediments and pyroclasdc materials, can control fluid
flow in shallow, low-tem perature hydrotherm al systems.
Examples include some carbonate-hosted Pb-Zn systems
(Garven, 1985) and some parts o f shallow, volcanic-related
geothermal systems. Intergranular porosity, although usu­
ally very low, may also be present in metamorphic regimes
(Holness, 1997). Reacdon-enhanced porosity is associated
with volume changes during metamorphic reacdons at ele­
vated tem peratures and pressures (Rumble e t al., 1982;
Zhang et al., 2000). This effect is particularly important in
controlling grain-scale fluid infiltration in skarn environ-
 O

ments, and pervasive hydrotherm al alteration associated

with fluid discharge from fracture-controlled hydrothermal
systems into relatively low-permeability wall-rocks. Volume
a
change between the solid reactants and reaction products
tends to produce transient changes in intergranular poros­
ity and permeability in these regimes. Macroscopically duc­
tile deformation in shear zones can be associated with grain
translation and grain-scale cracking, which also enhances
intergranular porosity. Similarly, in low-temperature brittle
deform ation regimes, cataclasis in active fault zones may
modify the intergranular porosity of the host rocks.
In rocks with low, or poorly connected intergranular
porosity, fluid transport is controlled by fracture porosity.
Fractures can be present at scales ranging from grain-scale
to macroscopic (> tens o f meters long). Fracture formation
is usually controlled by the stress and fluid pressure states
during deform ation o f a rock mass (see “Principles of
Macroscopic Fracture Formation” below), but may also be
influenced by thermal shocking.
Powsity-permeabttUy relationships
Low temperature regimes— Clastic rocks and fault rocks: In low-
tem perature crustal regimes where primary intergranular
porosity can survive for geologically long periods, perm e­
ability exhibits a simple relationship with porosity, espe­
cially at high porosities. For porosity greater than about 6 to
10 percent, pores are generally fully interconnected and
porosity is related to permeability by a relationship of the
form

koc 0n, (7)

where n is approximately 3. In this regim e, permeability

decrease with decreasing porosity is controlled by pore
shrinkage. At lower porosities, progressive loss of connectiv­
ity between pores occurs through closure o f narrow throats
between the larger intergranular pores. Pores that form part
o f a connected network that spans the sample/systcm com­
prise what is known as the backbone porosity (Fig. 3a). All
flow is localized along this backbone pore network; accord­
ingly, the backbone porosity controls the permeability. The
rem aining porosity comprises two elements: (1 ) isolated
pores, which do not connect with the backbone porosity,
even' though they may form localized clusters o f pores that
are connected to each other; and (2) dangling, or dead-end Fi<;. 3. a. Schematic illustration o f classes o f intergranular pores. Flow
pores, which connect to the backbone porosity from one rate is controlled by the backbone porosity, rather than by total porosity or
side only. Although they are p art o f the interconnected total con n ected (i.e., backbone + dangling) porosity, b. Relationship
porosity and contain pore fluid, they do not contribute to between total porosity and permeability in Fontainebleau samistone (after
Bourbic and Zinszner, 1985). The percolation threshold porosity and the
the flow. At low porosities, permeability changes are direedy relationship k« arc indicated.
related to progressive changes in the reladve proportions of
backbone, dangling, and isolated pores, as well as the total
porosity o f a rock. threshold. W here permeability is controlled by approxi­
T he relationship, k « 0s, breaks down where the back­ mately equant intergranular pores, the percolation thresh­
bone porosity becomes less than the total porosity. At lower old typically occurs at total porosities in the range 3 to 6
porosity, permeability typically becomes m ore sensitive to percent in many types of clastic rocks.
small changes in porosity (Fig. 3b). The porosity at which Above the percolation threshold, permeability associated
all pore connectivity is lost (i.e., ^ ckb(,nc = 0, £lso!;ilcd = 1 ) with intergranular porosity is influenced only moderately
and perm eability vanishes, is known as th e percolation by confining pressure (David and Darot, 1989). Permeabil-
ity decreases with increasing confining pressure (Fig. 4) in
response to elastic and inelastic deformation, causing pore
shrinkage and collapse. Note that at constant confining
pressure, changes in pore fluid pressure (i.e., changes in
effective confining pressure) lead to relatively small
changes in permeability. Poroelastic effects on total inter­
granular pore volume, driven by changes in stress states
associated with fault rupture events and interseismic stress
recovery, may play a role in episodic fluid redistribution
around aedve fault zones (Muir-Wood and King, 1993; Sib-
son, 1993). Such effects are likely to be a small part o f the
overall fluid budget in a high-flux hydrothermal system.
The relationships between porosity, permeability, and
effective stresses, found for clastic sedimentary rocks, will
apply also to granular damage products (gouge, cataclasite,
and breccia) produced in fault zones.
High temperature regimes— metamorphic rocks: In high-
tem perature m etam orphic regimes, atomic diffusion,
creep, grain boundary migration, and chemical reactions
are fast on geologic timescales. In this case, intergranular
porosity and pore connectivity in rock-fluid systems is con­
trolled principally by interfacial surface energy effects, pro­
vided that the polycrystalline aggregate is not deforming.
Where pore shapes are controlled by surface energy mini­
mization, there is a balance o f surface tension forces along
the surfaces where two solid grains and a fluid phase meet,
and we can define a dihedral wetting angle, 0 , which is the
angle between the two solid-fluid interfaces (Fig. 5a). The
magnitude of die wetting angle is controlled by the balance
of interfacial forces expressed by the relation,
0 = 2 arcos(ys_s/ 2ys_fl) , (8)
where y ^ and y ^ are the solid-solid and solid-fluid sur­
face energies per unit area, respectively (Smith, 1964).
FlG. 4. R elationship betw een perm eability and confining pressure, at
various pore fluid pressures, in Fontainebleau sandstone with porosity of
0.06 (after David and Darot, 1989),
b
Fig. 5. Schematic illustration o f pore geometries in polycrystalline grain
aggregates at elevated temperatures where pore shapes are controlled by
surface energy minimization, a. Cross section through a gran-edge chan­
nel; dihedral wetting angle, S . is indicated, b. For dihedral wetting angles
S60“. fluid forms isolated pockets o n twograin interfaces, and fluid chan­
nels along three-grain edges may connect fluid pockets at four-grain cor­
ners. c. For dihedral wetting angles >60°, fluid o n two-grain interfaces,
three-grain edges, and four-grain com ers occurs as isolated pockets (after •
Watson and Brenan, 1987).
The three-dim ensional connectivity o f intergranular
pores in isotropic mineral aggregates is dictated by interfa-
cial wetting angles and total porosity. We distinguish
between pores that form at two-grain interfaces, along
three-grain edges, and at grain corners (Fig. 5). For dihe­
dral welting angles greater than 60*, pores at two grain
interfaces tend to form isolated pockets. At low porosity,
grain-edge channels pinch off to form discontinuous beads
along grain edges, and pores at grain corners tend to be
isolated (Fig. 5c). In contrast, at dihedral wetting angles less
than or equal to 60°, connectivity between pores at grain
corners may be provided to low porosities by continuous
channels along grain edges (Fig. 5b). For ©greater than 0°,
the equilibrium fluid distribution on two-grain interfaces is
disconnected bubbles. Only in die extreme case of 0 equal
to 0° can a continuous fluid film exist stably on two-grain
interfaces. Experim ental studies o f interfacial wetting
angles for common minerals and pore fluids indicate that
for some common mineral-fluid systems, dihedral wetting
angles are greater than 60* (Holness, 1997). For example,
in the quartz-H20 - C 0 2 system at elevated pressures and
tem peratures in the range 950® to 1,150°C, dihedral wet­
ting angles are typically greater than 60® unless high solute
concentrations are present in the pore fluid (Watson and
Brenan, 1987).
The evolution of permeability with decreasing porosity at
elevated temperatures, where surface energy effects control
 m X C IPU -S OF STRUCTURAL COXTROl. IN HYDROTHERMAL SYSTEMS
pore geometry, is illustrated by high-temperature isostatic
compaction experiments on calcite aggregates (Zhang et
al., 1994a). The overall porosity-permeability relationship is
similar to that for clastic sedimentary rocks (Fig. 6a). For
the calcite-argon system, the dihedral wetting angle is
greater than 60°, so the situation is analogous to many min­
eral-fluid systems at elevated temperatures. The porosity-
permeability relationship approximately follows the cube
law (i.e., n = 3 in equation 7) to porosities down to about
0.06, where full connectivity between pores starts to be lost
(Fig. 6b). Below this porosity, permeability is much more
strongly dependent on total porosity. The critical porosity,
0C, at which pores completely lose connectivity (i.e., the per­
colation threshold) occurs at a porosity of 0.04. For min­
eral-fluid systems with wetting angles less than 60®, the per­
colation threshold will occur at porosities less than 0.04.
Near (f>c, the permeability of the intergranular pore network
obeys the scaling law (Knackstedt and Cox, 1995),
k ~ ( 0 - 0 c)2- (9)
For high-temperature, isostatically stressed (i.e., all prin­
cipal stresses equal) m ineral-fluid systems, where pore
geometry is controlled by minimization of interfacial sur­
face energy, dihedral wetting angles have a profound effect
on fluid transport. For wetting angles greater than about
60®, pore connectivity is lost at porosities of several per­
cent. Accordingly, many metamorphic fluid-rock systems,
which have porosities less than one percent, will be below
the percolation threshold and effectively im perm eable,
unless deform ation processes actively generate fracture
networks.
Fracture-controlledfluid flout
In hydrothermal systems where compaction, pore sealing
processes, o r diffusion-controlled surface energy effects
have driven intergranular porosity below the percolation
threshold, fluid migration is dependent on the generation
of fracture porosity.
Macroscopic to microscopic fractures at low temperatures: For
steady state, laminar, incompressible flow in an ideal, hori­
zontal, parallel-sided plane fracture, the volume flow rate
(m3/s) is given by the relationship,
Q,= wb3/ 12 tj*(dP /dx), (10)
where
b is the fracture aperture;
w is the width o f the fracture (m easured in the fracture
plane); and
d P /d x is the fluid pressure gradient along the fracture
(Gueguen and Dienes, 1989).
By analogy with Darcy’s law, for a fracture with cross-
sectional area wb, the equivalent fracture permeability is
given by -
k = b2/ 12 . (H )
7
h 0.16 |-----------------1---------------- T1-----------------1--------------
~\-----------------
a experimentally measured
□
. ✓ ti•e
/tl “
P • model backbone porosity A
GO 0.12 -
£ ✓ O
Hi 0 .0 8 - , d'a ° ♦* #
m rf' ♦
sf. a **
co
•*
8o 0 .0 4 -
//
/
e •
.
< S/ O *
/
0 * ----------- ^ ----------- 1------------- 1------------- 1
0 0 .0 4 0 .0 8 0 .1 2 0 .1 6
TOTAL POROSITY
Fig. 6. a. Porosity-permeability relationships for calcitc grain aggre­
gates, isostatically hot-pressed at temperatures between 360° and 560*0.
Confining pressures ranged from 200 to 300 MPa and argon pore fluid
pressures ranged from 100 to 250 MPa (after Zhang c ia l., 1994a). b. Rela­
tionship between total porosity and connected (or accessible) porosity
experimentally determ ined during isostatic hot-pressing o f calcite aggre­
gates. Also shown are numerically modeled values o f (lie backbone poros­
ity. M odified after Zhang c ta l. (1994a) and Knackstedt and Cox (1995).
In reality, relationships between fracture apertures and
flow rates are more complicated owing to the effects of frac­
ture roughness leading to aperture changes and tortuous
flow paths. Experimental studies provide clear evidence of
surface roughness causing departures from the cubic law in
natural rough fractures (Witherspoon et al., 1980; Tsang
and W itherspoon, 1981; Brown, 1987). Numerical model­
ing o f flow through rough fractures also confirms that frac­
ture roughness and tortuosity lead to substantial departures
from the cubic law (Waite et al., 1999).
Because cracks have high aspect ratios, elastic opening or
closure o f cracks in response to changes in effective stress
states has im portant effects on crack aperture and perme­
ability (Walsh, 1981). The dependence o f fracture perme­
ability on effective confining pressure (i.e., confining pres­
s u re -flu id pressure), illustrated in Figure 7, shows that the

EFFECTIVE CONFINING PRESSURE, MPa
Fig. 7. Dependence o f crack permeability on effective confining pres­
sure {after Guegucn and Palciauskas, 1994).
pressure dependence is substantially greater than for mate­
rials where perm eability is controlled by approximately
equant intergranular pores (Fig. 4). Accordingly, low poro-
eiastic and inelastic strains in cracked rocks will be more
effective than deformation of intergranular pores in chang­
ing permeability, and driving fluid redistribution in
response to stress changes (e.g., around active faults).
For a cracked rock with fracture porosity <p(, and where
penny-shaped cracks are fully interconnected and ran­
domly oriented, permeability k is given by
k=aV 3, U 2)
where a is the average crack half-aperture. For an average
crack radius, r, and average crack sparing, 1,
0c = 2m-L,a /I3, (13)
and so (Gudguen and Palciauskas, 1994),
k = 2m -V /3F. (14)
This relationship holds true for fractured media where
fractures are developed at microscopic to macroscopic
scales. It illustrates how both fracture density and fracture
connectivity play a critical role in controlling permeability.
For randomly distributed fractures that are not fully con­
nected, permeability is given by
k= (4 7 T /1 5 H aV /l:\ (13)
where connectivity, f, is 0 < f < 1 (Gueguen and Dienes,
1989). As in the case for a single fracture, this relationship
illustrates how the permeability of a cracked medium is very
sensitive to average fracture aperture. The evolution of con­
nectivity among elements of networks of fractures, faults,
and shear zones, and its influence on flow architecture, is
an important aspect of the evolution of hydrothermal sys­
tems, and is discussed below in “Evolution o f Flow Pathways
during Deformation.”
In summary, for flow in fracture-controlled hydrothermal
systems, the highest flux will occur where and when ( l)
fracture apertures are highest, (2) fracture density Is high­
est, and (3) fracture connectivity is highest.
Permeability during macroscopically ductile deformation:
Experimental results, together with field observations o f
fluid-rock interaction in exhum ed shear zones, indicate
that perm eability enhancem ent is comm only associated
with macroscopically ductile deform ation. Experimental
studies indicate that macroscopically ductile deformation at
elevated temperatures and confining pressures is only asso­
ciated with permeability enhancem ent when a com ponent
o f strain (albeit small) is accommodated by microfractur-
ing, which generates grain-scale fluid pathways.
T he evolution o f perm eability during deform ation
involving both intragranular plastic deformation and grain-
scale crack growth is illustrated by experimental studies on
calcite rocks (Zhang e ta h , 1994b). This work demonstrates
that for low effective confining pressures, permeability
increase with increasing strain can be very rapid and large
(Fig. 8a, b ). For example, a t an effective confining pressure
of 30 MPa, permeability increases by two orders o f magni­
tude with three percent shortening, and increases by a fur­
ther order of magnitude after ten percent shortening. Only
small increases in permeability occur with higher strains,
however.
Major increase in permeability, at strains as low as a few
percent, is associated with growth o f m icrocracks and
rapid development of connectivity in grain-scale crack net­
works (Fig. 9a). Such behavior persists well into the domi­
nantly crystal plastic deform ation regime, provided that
fluid pressures are high enough to facilitate some defor­
m ation by m icrocracking (Fischer and Paterson, 1992;
Storm ont and Daemen, 1992; Zhang e t al., 1994b; Peach
and Spiers, 1996). As both tem perature and effective con­
fining pressure increase, brittle/frictional processes are
impeded and intracrystalline plasticity is favored. So, crack
growth rates reduce, and the critical strain required to
develop a well-connected, high-permeability crack network
increases.
The experimental studies demonstrate that where high
fluid pressures produce low effective confining pressures,
grain-scale crack growth significantly increases the perme­
ability o f active shear zones relative to their host-rocks,
even though most displacement may occur by microscop­
ically ductile deformation mechanisms such as dislocation
flow and dissolution-precipitation creep. A significant
aspect of the experimental work is that fracture networks
can develop high crack connectivity and high permeabil­
ity at very low strains. One implication of this result is that
low strain deform ation, especially when localized in net­
works of faults, shear zones, and associated fracture arrays,
may have a big im pact on the localization o f fluid flow
(see “Evolution of Flow Pathways d uring D eform ation”
below).

Fig. 8. Relationship between (a) porosity an<l strain, and (l>) perme­
ability and strain, as a function o f effective con fin in g pressure during
deformation o f Carrara marble at room temperature and in the presence
o f argon pore fluid. Relationships at cflcctivc confining pressures o f SO.
50. and 100 MPa arc illustrated. Confining pressure « 300 MPa; nominal
strain ia te= 1 .2 x 1(H s-1 (after Zhang eta)., 100-lb).
Competition between deformation-induced permeability
enhancement and permeability reduction processes
;: At depth in the Earth’s crust, and especially at elevated
tem peratures in active hydrotherm al systems, porosity-
destruction processes such intergranular cem entation,
compaction, and healing and sealing o f fractures can cause
permeability to decrease on timescales that are short rela­
tive to the lifetimes o f hydrothermal systems. Penneability
evolution is, therefore, controlled directly by competition
between deformation-induced porosity-creation processes
. and various porosity-destruction processes.
The structure o f veins in ore systems provide spectacular
evidence o f repeated fracturing and fracture sealing in
hydrothermal systems. For example, crack-seal microstruc­
tures in veins (Ramsay, 1980; Cox, 1995) indicate that
, macroscopic fractures in some deep hydrothermal regimes
/: can open and seal up to several thousand times in the life-
time o f one active vein (Fig. 9b). Similarly, overprinting
relationships in veins in shallow epitherm al systems and
deeper mesothermal systems commonly indicate muldple
episodes o f fracturing and vein sealing (Fig. 9c). Experi­
m ental studies o f fracture healing (Hickman and Evans,
1987; Brantley e t al., 1990; Zhang e t al., in press), also
dem onstrate that, at tem peratures greater than approxi­
mately 300*C, diffusional crack healing and associated loss
o f crack connectivity in minerals such as quartz and calcite
can be fast, even on laboratory' timescales.
Permeability evolution in faults in the seismogenic upper
crust can be influenced strongly by porosity changes associ­
ated with defoimation during repeated, episodic slip events
(over several seconds), or during aseismic creep. In rocks
with high initial intergranular porosity, fault slip is associated
with reduction in porosity and permeability owing to the
production of fine-grained, compact, low-porosity gouges
(Knipe, 1998; Zhang et al., 1999). In this case, faults become
seals or aquitards in hydrothermal systems. However, in
intrinsically low-porosity, tight rocks (e.g., metamorphic and
igneous rocks), faulting leads to porosity and permeability
increase during slip, owing to fracture growth and cataclasis.
In contrast, porosity evolution during interseismic intervals is
governed by progressive porosity destruction by pore col­
lapse and hydrodiermal sealing (Angevine et al., 1982; Cox
and Paterson 1991; Cox, 1995). Episodic slip and interseis-
mic sealing may result in large, time-dependent changes in
penneability in fault zones (Fig. 10). Particularly in fiacture-
controlled hydrotherm al systems, flow within perm eable
faults and associated fracture arrays (Fig. 1) promotes rapid
pore sealing by mineral deposition. As soon as porosity drops
below the percolation threshold, fluid flow will shut off. An
im portant conclusion then, is that in the immediate post­
rupture phase, faults in low-permeability host rocks are
highly permeable structures that act as fluid conduits. How­
ever, with progressive interseismic sealing and loss of perme­
ability, faults can become aquitards until a later slip event
Accordingly, in active hydrothermal regimes, permeability
is rapidly destroyed unless ongoing deformation regenerates
permeability. Additionally, for permeable fluid pathways to
develop, the rate o f deformation-induced penneability
enhancement must be greater than the rate o f permeability
reduction owing to closure and sealing of pores and fractures.
E pisodic versus c o n tin u o u s flo w
Com petition between porosity-creation processes and
porosity-destruction processes in actively deforming rocks
results in a contrast between flow regimes in the upper
crustal seismogenic regime and the lower crustal aseismic
regime. At depths below the seismic-aseismic transition
(typically 10-20 km deep), where steady state creep
processes usually dom inate, a balance between rates of
porosity destruction a n d porosity creation in creeping
shear zones is expected to generate quasi-steady state per­
meabilities that are higher than those o f the surrounding,
less rapidly deforming rock mass. This leads to essentially
continuous fluid flow along actively creeping elements of
shear networks. Creeping faults and shear zones probably
play a key role in focusing the migration of deeply sourced
 Fig. 9. a. Grain-scale crack networks produced in Carrara marble during macroscopically ductile deformation at room
temperature and an effective pressure o f 100 MPa. The shortening direction is paraltet to the micrograph long axis. Inci­
dent light micrograph, b. Quartz-rich extension vein with dark, crack-seal inclusion bands composed o f tourmaline. This
texture indicates multiple episodes o f extcnsional fracturing and fracture-scaling associated with gold mineralization,
Sigma Mine, Val d ’Or, Quebec. Transmitted light micrograph, c. Laminated fault-fill veins indicating multiple episodes
o f fault slip and dilation during gold mineralization. Revenge gold deposit, St Ives goldfield. Eastern Goldfields Province,
Western Australia.

fluids upwards lo the base o f the seismogenic regime. In
contrast, above the seismic-aseismic transition, potentially
large, cyclic changes in fault permeability cause episodic
fluid flow. Episodic flow is associated with fluid pressure
cycling and episodic fluid redistribution around active
faults, and has important implications for the dynamics of
flow and reaction in hydrothermal systems in the seismo­
genic regime (Cox, 1999; Sibson, 2001).
Principles o f Macroscopic Fracture Formation
The development of faults, shear zones, and associated
fracture arrays plays a key role in controlling the local per­
meability distributions and the macroscopic architecture of
fluid pathways in hydrothermal systems. The geometry of
these structures is inherently influenced by the orientations
and relative magnitudes o f stresses in hydrothermal systems.
The distribution o f these structures is also controlled by fluid
pressure regimes. This section outlines the critical influence
o f both fluid pressures and stress states in controlling macro­
scopic fracture growth and associated permeability genera­
tion in hydrothermal systems. Importantly, macroscopic frac­
ture systems can develop not only in the brittle upper crust,
but also in the more ductile deeper crust, provided fluid
pressures are high enough.
Types offractures and orientation relationships
with stress fields
Three classes of macroscopic fractures may form during
 a regimes during crustal deformadon produce distinctive ori­
TIME
b and 14a). High fracture densities and apertures can also
Fig. 10. Schematic illustration o f liinc-cicpcndetu changes in (a) poros­
ity and (b) permeability in fault rocks during the seismic cycle. Sudden
permeability enhan cem ent is associated with episodic fault slip (EQ).
Interscismic pore sealing and compaction reduces permeability between
slip events. After Cox (1995).
brittle deformation: (1 ) pure extension fractures; (2) shear
fractures; and (3) hybrid extensional-shear fractures. Mineral
filling in each of these fracture types produces veins (Fig. 11).
Any stress field can be resolved into three mutually per­
pendicular components, which are known as the maximum
principal stress (Oj), the intermediate principal stress (o2),
and the m inim um principal stress (o3). Pure extension
fractures form perpendicular to 0:<and open perpendicu­
lar to die fracture wall (Fig. 12a). In intact, isotropic rock,
shear fractures (i.e., faults) tend to form at angles typically
between 20® and 35® to the orientation o f a ,, such that the
fault plane contains the orientation o f <r2 (Fig. 12b). Two
conjugate sets o f shear fractures can form, and where devel­
oped, their intersecdon is subparallel to o2. Hybrid extern
sionai-shear fractures have com ponents o f displacem ent
both parallel and perpendicular to the fracture plane. They
form a t angles between 0° and approximately 25° to the <Tt
direction.
The orientauons o f the maximum principal stresses asso­
ciated with contractional, extensional, and wrench tectonic
entauons of shear fractures (faults) and spatially associated
extension fractures (and veins; Fig. 13). The relative orien­
tations o f faults and extension veins provide one o f the
most powerful ways of determ ining shear sense on struc­
tures that were active during hydrothermal mineralization
(Robertand Poulsen, 2001).
Faults and associated fracture arrays exhibit heteroge­
neous distributions o f fracture apertures and fracture den­
sities at mesoscopic to macroscopic scales. Areas with the
highest apertures a n d /o r fracture densities produce the
highest permeability sites, which localize most fluid migra­
tion if they connect to fluid reservoirs. High fracture aper­
tures a n d densities are typically associated with dilatant
fault bends, stepover regions, or jogs that link approxi-
mately planar segments of faults and shear zones (Figs, l i d
develop in areas o f com petence contrast (Oliver e t al.,
2001) and during dilation a t fold hinges during flexural
slip folding (Cox e t al., 1991). High fracture permeability
can also localize aro u n d fault term ination zones, where
fault splays, wing cracks, or brecciated regions develop (Fig.
14b). Contractional jogs, as well as dilationa! jogs, can be
sites o f high fracture density, which localize fluid (low and
ore deposition.
T he orientation relationships between shear fractures,
jogs, and slip directions result in the long axis o f jogs devel-
oping approxim ately parallel to the <%orientation (Fig.
14c). This produces a permeability anisotropy favoring flow
parallel to the jo g axis. For reverse and norm al faulting
regimes, extensional and contractional jogs have subhort-
zontal plunges (Fig. 14d). This produces good horizontal
fracture connectivity within faults, and may cause ore shoots
to have gentle plunges. In contrast, strike-slip regimes pro­
duce jogs with good vertical connectivity that can control
the geometry of steeply plunging ore shoots (Fig. 14d).
Stress magnitudes and fracture formation
The types o f fractures that control permeability in hydro-
thermal systems in the brittle regime are governed by the mag­
nitudes of stress differences (<T( - <J3). The stress states in rocks,
and the relationships between stress magnitudes, stress differ­
ences, fracture type, and fracuire orientation can be illustrated
by two-dimensional Mohr circle constructions.
The stresses acting on a plane inclined to O, at an angle
a can be resolved into a normal stress (o„) acting perpen­
dicular to the plane, and a shear suess (fs) acting parallel to
the plane (Fig. 15a). The magnitudes o f the normal stress
and shear stress are given by the relationships,
on = l/2(o, + o3) + l/2(o, - a,)-cos 2a (16)
and
r,= 1/2(0, -O t) sin 2a. (17)
A M ohr diagram plots on against Ts, and has the geomet­
ric property that the normal and shear stresses acting on a
plane inclined at angle a to O, are given by the coordinates
12 COXKTAL

Fig. 1 i. Examples o f vein types classified according to fracture modes, a. Pure extension veins; opening direction is
po rjoenclicu lnr to the fracture walls. Mt Lyell. Tasmania. Coin is 2 cm in diameter. I>. Hybrid cxtcnsional-shear vein, with
opening direction inclined to the fracture wall indicated by orientation o f displacement-controlled crack-seal quartz
fibers. This vein is associated with smaller, snbhorizontal extension veins. Mt Lyell. Tasmania c. Gendy left-dipping exten­
sion veins spatially associated with gendy right-slipping fault-fill veins, within dilatant segments o f two faults associated
with gold mineralization. The orientation relationship between the subhorizontal extension veins and the fault veins
indicates that mineralization formed during reverse faulting. Victory gold m ine. St Ives goldfield. Western Australia.
Field o f view is 6 m wide. d. Calcite-Filled dilacintjogs developed on stepovers Iretwcen small, steeply dipping, sinisual
wrench faults. Lcs Matches, Languedoc, France.

of the point P on the perimeter of the circle haring diame­
ter (<Tj - <T3) and center at (<7, + <%)/2 on the normal stress
axis (Fig. 15a). This circle is cailed the stress circle (or
Mohr circle). Note that P is on a radius inclined at 2 a to the
normal stress axis.
Brittle shearfailure: U nder fluid-absent conditions, the
stress state that causes compressional brittle shear failure
(i.e., faulting) in an intact, isotropic medium is approxi­
mated by the relationship,
T>= C + P ni
where
C is the cohesive strength of the medium; and
u is die coefficient o f friction. .....................
This Coulomb shear failure critenon is indicated by the
line BC in Figure 15a. This line specifies the brittle shear ■'M
failure strength of the m edium as a function o f norm al 'M
stress. So, in dry, intact rock, brittle shear failure occurs
where and when the M ohr circle (describing the shear
stress and normal stress state in a rock) contacts the failure
envelope BC. Shear failure may, therefore, be induced by
decreasing <73 a n d /o r increasing the value o f <7j (stress cir­
cle A in Fig. i5b). Because the angle between die Mohr cir­
cle radius (AP) and the normal stress axis is 2a, the angle a
between the shear failure plane and the orientation o f the
maximum principal stress is given by the relationship,
(1 ®)
a= (90° - arctan fl)/2. (19)
For typical friction coefficients of approximately 0.75,
shear fractures are inclined at approxim ately 27° to (7j.
Shear fractures form ed in intact rock and obeying rela-
itf
 PRIM PIJES OF STRUCTURAL CONTROL IN HYDROTHERMAL SYSTEMS

a EXTENSION FRACTURES a R E V E R S E FAULT

b NORMAL FAULT

subvertical extension
veins

Fig. 12. Orientations of fracture types with respect to principal stress C WRENCH FAULT
directions, CT>, and o3. a. Pure extension fractures, b. Conjugate shear
fractures, with sense of shear indicated.

tlonship (19) a re term ed “optimally oriented” faults (Sib-

son, 1985). Activation, o r re-activation, o f misorienecd
structures may also be im portant in controlling fracture
permeability in hydrothermal systems. For example, mts-
priemed faults may develop by rotation away from optimal
orientations with respect to the stress field, due either to
rotation o f stress fields a n d /o r rock masses d uring pro­
gressive deformation. Additionally, preexisting faults, or
o th er mechanical anisotropies such as bedding o r folia­
tion, can also be (re-)activated. Shear (re a c tiv a tio n
occurs when the shear stress on the misoriented plane sat­ Fig. 13. Orientations of faults and associated extension veins relative to
isfies the Mohr-Coulomb failure criterion (Fig. I5c). Fig­ the principal stress directions for contraction.!/ (n), extensions! (b), and
u re 15c has been constructed with the cohesive strength, strike-slip (c) tectonic regimes.
C, o f the preexisting fault or anisotropy less than that of
intact rock. Note, however, that rapid sealing o r com­ and shear stress state (Mohr circle) contacts die nonlinear
paction o f gouge on fault slip patches in hydrothermal sys­ part o f the failure envelope at negative (i.c., tensile) nor­
tems may cause some faults to quickly regain cohesive mal stresses. Pitre extension failure occurs where the nor­
strength (Kanagawa et al., 2000). mal stress equals the tensile strength (T) o f the rock (stress
circle C in Fig. I5b). Note that the angle 2 a is zero, as
Extension and hybrid extensional-shearfailure: Extension fail­ expected for extension failure a long the Oj-dg plane, per­
ure a n d extensional shear failure occur if the normal stress pendicular to Gy Hybrid extensional shear failure occurs
 COX ETA/.

Fig. U . a. Geometry of coturactional and dUautnt jogs. b. Wing cracks and coniractional splays developed around
fault icmmauons. c. Geometry of a tUlaiant jog relaiive lo fault slip direction, d. Orientations ofjogs in coniractional,
extension*!, and wrench tectonic refines. High fracture density, aperture, and connectivity along jogs favors high fluid
flux along the jog axis.

w here the stress circle contacts the failure envelope
between -T and C (stress circle B in Fig. 15b).
Bute extension failure only occurs a t relatively small
stress differences, typically less than approximately 4T. For
4T < (CFj - o-\) < 6T, failure occurs in extensional-shear
m ode. Shear failure occurs a t stress differences greater
than approximately 6T. Because rock tensile strengths are
typically less (sometimes substantially less) than about 10
MPa, the occurrence of pure extension veins in hydrother-
ma\ systems indicates stress differences less than 40 MPa
during vein opening (Etheridge, 1083).
T he role o f fluids
Fluid pressure modifies stress stales at depth in the
Earth’s crust. The effect of pore fluid pressure (Pr) is lo
reduce the effective normal stress (<Jn') according to the
relationship,
< =
in term s of the M ohr circle representation o f stress
states, the role of fluid pressure is to move the stress circle
to the left (Fig. 16). Note that although fluid pressure mod­
ifies normal stress, it does not influence shear stress.
In hydrotherm al regimes, changes in stresses a n d /o r
fluid pressures can induce brittle failure. For example, at
low stress differences, pure extension failure is induced by
increase in fluid pressure (Fig. 16a), provided the effective
minimum principal stress becomes negative and equal in
magnitude to the tensile strength of the rock. This fluid-dri­
ven extension fracturing is known as hydraulic extension
fracture. The hydraulic fracture criterion is, thus,
Pr=<r3 + T. (21)
The abundance of mineral-filled extension fractures in
many epigenetic ore deposits indicates that tensile effective
stress states, and, thus, fluid pressures greater than <r3, are
(20) common in hydrothermal systems. Importantly, equation
 a
Fig. 15. a. Mohr diagram and inset illustrating relationship between
principal stresses, shear stress, and normal stress. The normal stress (<7n)
and shear stress ( t5) on a plane inclined at angle a to the orientation o f
the maximum principal stress are given by the coordinates o f the point P
on a Mohr circle having diameter (or, - er;1) and center at (o’, + a3) / 2 on
the norma! stress axis. The curved line -TCB is a typical brittle failure
envelope (modified Griffith-Mohr-Coulomb failure criterion). T is the ten*
silc strength o f the rock; the shear failure envelope (CB) is approximated
by the relationship t ® C + pr where r is the shear strength, C is the cohc-
' sive strength, jA is the coefficient o f friction, and Cn is the normal stress.
Note that shear strength is dependent on normal stress. Stress states to the
right o f the failure envelope do not result in brittle failure, b. Mohr dia­
gram illustrating Mohr circles for (A) shear failure (faulting), (B) hybrid
cxtcnsional shear, and (C) pure extension failure o f intact rock. c. Mohr
diagram illustrating stress conditions for reactivation o f a nonoplimally
oriented, Iqw cohesion, shear failure plane (c.g., preexisting fault, or foil*
ation/bedding anisotropy) inclined at angle (i to the maximum principal
stress: Note that when the shear failure envelope (AB) for a preexisting
: mechanical weakness lies below the failure envelope for intact rock, the
preexisting structure may be preferentially reactivated.
Fig. 16. Mohr circte constructions illustrating the effects o f increasing
fluid pressure, by an am ount P(, o n effective stress states and conditions
for failure in intact rock. a. Increasing fluid pressure at low stress differ­
ences results in extension failure, b. Increasing fluid pressure at high stress
differences induces shear failure.
(2 1 ) illustrates how the opening o f hydraulic extension
fractures limits the maximum fluid pressures that may
develop in hydrothermal systems.
An im portant aspect of the role o f fluid pressure in frac­
ture processes is that brittle failure may be induced at any
depth, provided fluid pressures are high enough. Perme­
ability enhancem ent by fracture growth can occur in
response to changes in stress difference (i.e., through
changes to Oj a n d /o r <j3), and also by changes in fluid
pressure. This concept is illustrated in Figure 17, which
plots failure modes as a function o f stress difference and
fluid pressure. H ere, fluid pressure is expressed as the
pore fluid factor (Av), the ratio o f fluid pressure to over­
burden stress (<rv). The pore fluid factors and stress differ­
ences leading to failure in extensional, extensional-shear,
and shear modes are plotted for a strike-slip regim e at a
depth of 3 km. From some ambient stress state, failure can
be induced either by increasing the pore fluid factor along
the trajectories such as Q a n d S (Fig. 17) without changing
the absolute stress state (an example o f purely fluid-driven
failure), or by maintaining a constant pore fluid factor and
increasing the stress difference (e.g., trajectory P), or some
 strike-slip faulting, depth * 3 km (Nguyen et al., 1998) that cannot be ascribed to repeated
fluctuations in tem perature and depth. They are instead
related to fluctuations in shear stress and fluid pressure asso­
ciated with episodic fault slip and fault-valve behavior in
shear zones (Sibson et al., 1988; Sibson, 2001). Changes in
deformation style, caused by changes in fluid pressure states,
are a key factor influencing deformation style, the nature of
permeability, and pervasiveness of flow and alteration.
For example, in die shcan-hostcd Revenge gold deposit in
the St. Ives Goldfield (Eastern Goldfields Province, Yilgarn
craton, Western Australia), Au-mineralized shear zones
exhibit initial brittle shear failure, subsequent mixed brit­
tle-ductile behavior, and finally, fully brittle behavior during
progressive hydrothermal alteration (Nguyen etal., 1998).
Initial fluid flow was controlled by macroscopic shear frac­
ture at pore factors near one, that is, near-lithostadc fluid
pressures (Fig. 18a). Progressive potassic alteration o f
Fig. 17. Failure mode diagram illustrating how changes in stress differ­ metabasalt and m etadolerite host-rocks produced weaker,
ence and fluid pressure (expressed as pore fluid factor; sec text for expla­ biotite-rich alteration assemblages. Reduced plastic shear
nation) lead to brittle failure. Reids for pure extension, extcnsional-shcar.
and shear failure are indicated for a strike-slip regime at a depth o f 3 km. strength, due to reaction-weakening, induced a transition to
A failure envelope (AB) is indicated for intact rock with cohesive strength ductile deformation (Fig. 18b). In this deformation regime,
o f 10 MPa. Failure occurs when the pore fluid factor an d /or stress differ­ fluid flow was controlled by permeability enhancem ent asso­
ence increase to touch the failure envelope AB. Trajectory P leads to shear ciated with grain-scale dilatancy and reaction-enhanced
failure at constant fluid pressure with increasing shear stress. Trajectory R
leads to shear failure by increase in both fluid pressure and stress differ­
porosity. Repeated brittle slip events occurred during ongo­
en ce. Trajectories S and Q lead to shear failure and extension failure, ing ductile deformation, and were associated with formation
respectively, by pore fluid pressure increase alone. of breccias and laminated fault-fill veins in dilatant bends
and jogs. Episodic fast seismic (?) slip events are interpreted
to have been driven by increasing pore fluid factors during
com bination of both stress and fluid pressure increase individual fault-valve cycles. Each brittle shear failure event
(e.g., trajectory R). was preceded by an interval of interseismic ductile shear at
Fluid-driven failure is extremely important in generating lower pore fluid factors (Fig. 18b). Sodic metasomatism late
fauk/fracturc networks in hydrothermal systems in both in the hydrothermal history produced localized albite-rich
the upper and lower crust, provided fluid pressures are assemblages, which resulted in reaction-strengthening and a
high enough (Sibson, 1996; Cox, 1999). A significant late transition back to fully brittle behavior and macroscopic
aspect of fluid-driven fracturing is that, even without fracture-controlled fluid flow (Fig. 18c).
changes in the stresses acting on a previously nondeform­
ing rock medium, infiltration o f high-pressure fluids can Implications of stress and fluid pressure regimes for
localization offracturing, fluid flow, and mineralization
drive the spontaneous growth of fractures. The orientations
of these fractures are controlled by the orientation of the Fluid pressure regimes in the crust are controlled to a
principal stresses, magnitudes of the stress differences, and first o rder by permeability and fluid driving pressures. In
geometry of any preexisting mechanical anisotropies in the the upper few kilometers o f the continental crust, where
rock mass. fracture and pore connectivity is high, pore fluid pressures
are generally close to hydrostatic. At deeper levels, where
Effects of stress and fluid pressure on transitions pore sealing and collapse are more rapid, loss of long-term
between failure modes pore connectivity results in fluid pressures increasing
Particularly in hydrothermal systems form ed at mid- towards Uthostatic levels. Deeper level fluid reservoirs are
crustal depths (e.g., some mesothermal gold systems), fluid accordingly suprahydrostatically pressured. Tapping of
flow, hydrothermal alteration, and mineralization are con­ these reservoirs by active and permeable faults, fractures,
trolled by faults and shear zones that exhibit mixed brit­ and shear zones provides transient pathways for rapid
tle-plastic behavior. First order transitions between brittle upwards migration of fluids under high driving pressure
and plastic behavior in the crust are commonly ascribed to gradients. Fluid discharge from the upper levels of these
the effects of increasing tem perature and confining pres­ structures can generate overpressured dom ains at these
sure (with increasing depth) inhibiting brittle deformation sites (Fig. 1), especially if die fluid discharge zone is capped
and promoting the operation of thermally activated defor­ by a low-permeability dom ain. Fluid-driven growth of
m ation processes such as dislocation creep and diffu- fault/fracture systems is more favored as pore fluid factors
sional/advcctive mass transfer processes. However, shear progressively increase above hydrostatic levels (Fig. 19).
zones in some mesothermal gold systems exhibit evidence An example o f localized fluid pressurization and associ­
for repeated transitions between brittle and plastic behavior ated mineralization is provided by the development of the

STRESS DIFFERENCE, MPa
F ig. 18. Failure m ode diagrams illustrating brittle and plastic failure
envelopes as a function o f pore fluid factor and stress difference. The dia­
gram is constructed for optimally oriented reverse faulting at a depth o f 12
km in rock with a cohesive strength o f 10 MPa, tensile strength o f 5 MPa.
and, friction coefficient 0.73. Fluid pressure and stress changes associated
with fnult-valve behavior may produce episodic transitions between no
deformation, ductile shear failure, brittle shear failure, and brittle exten­
sion failure at various phases o f the seismic cycle. The failure envelope is
indicated by the curve PQRST. Ductile shear failure occurs for fluid pres­
sure and stress states in die interval ST; brittle shear failure occurs between
S and R; cxtcnsional-shcar occurs between R and Q; and extensionn! fail­
ure occurs in the interval PQ. a. Brittle shear failure at point B induced by-
increase in stress difference and fluid pressure, b. Reaction-weakening
decreases the ductile shear strength (ST), so that increase in stress differ­
en ce and fluid pressure leads firstly to ductile shear failure (ascismic
creep) at point B, followed by brittle shear failure at point S, in response to
progressive increase in pore fluid factor, c. Latc-stagc reaction-hardening
increases ductile shear strength, so that failure occurs by brittle shear fail­
ure at high stress differences, or extension failure occurs nt low stress dif­
ferences and high pore fluid factors.
swarm o f subhorizontal extension veins which hosts tung­
sten mineralization at Panasqueira in Portugal (Foxford et
aL, 2000). The vein swarm in this deposit developed above a
granitoid stock in a subhorizontal, lenticular domain which
is less than 200 m thick. The veins are interpreted to have
grown where fluid m igrating up from the stock became
trapped below a low permeability zone and developed a
supralithostatic-pressured domain of hydraulic fracturing
(i.e., Pf = c%+ T). Because extension veining at Panasqueira
is no t associated with faulting, stress differences are inferred
to have been low. The thickness o f the veined domain is con­
trolled by the limited depth interval over which supralitho-
stadc fluid pressures developed above die stock.
The Bendigo goldfield in central Victoria, Australia, pro­
vides an example o f where the structure of low penneabil-
ity stratigraphic units may have localized the distribution of
high pore fluid factor sites and associated vein gold
deposits. The Bendigo goldfield occurs in the regional cul­
m ination zone o f a series o f doubly-plunging folds. Addi­
tionally, most fault-related and vein-related gold deposits in
the area are located in and near saddle reefs in the hinges
of anticlines. High pore fluid factors, associated hydraulic
fracturing, and veining at Bendigo are interpreted to have
been controlled by fluid migration to the structurally high­
est levels beneath folded, low-permeability stratigraphic
seals (C oxetal., 1991).
At the Porgera gold deposit in Papua New Guinea, early
hydrotherm al alteration, fluid flow, and low-grade, dis­
persed mineralization was associated with pervasive grain-
scale and fracture-controlled flow around a mafic intrusive
complex (Munroe, 1995). Early in the development o f the
hydrotherm al system, stress differences were not high
enough to generate faults, but high pore fluid factore pro­
duced extension veins. Subsequent growth o f a fault across
the active, intrusive-related hydrothermal system was asso­
ciated with a rotation o f the stress Held and increase in
stress difference (Cox and Munroe, 1999). The increase in
stress difference, together with high pore fluid factors,
drove the growth of a fault and its associated high-perme­
ability damage zone, which tapped deeper level hydrother­
mal fluids. Localization of fluid flow within the fault and its
damage zone led to the formation of a very high grade gold
deposit around the fault. This contrasts with the early min­
eralization which is more dispersed and not spatially associ­
ated with faulting.
Competence contrasts within rock masses also control
localization of fracturing (see also Oliver et al., 2001). For
example, where strong rock masses occur within a weaker
matrix, with boundaries oriented at a high angle to the short­
ening direction, viscous or frictional drag along the contacts
reduces the minimum principal stress in the competent rock
unit. As fluid pressures build-up, hydraulic extension frac­
ture or shear failure (depending on stress difference) occurs
first in the zone of decreased minimum effective principal
stress in the com petent rock unit. Examples include the
development of mesothermal Au-mineralized extension vein
arrays in dolerite in the Mt Charlotte orebodies at Kalgoorlie
(Clout et al., 1990); ladder vein systems in dikes, such as at
 a surface
b

near-hydrostatic

Fig. 19, Localization o f fluid-driven failure controlled by depth-dependence o f stresses and fluid pressure, a. Upwards,
head-driven fluid flow in a fault, which taps an overpiessurcd fluid reservoir. Fluid discharge at the downstream end o f
the fault drives fluid pressure build-up beneath a low permeability leaky seal zone. W here stress differences are low,
hydraulic extension fracture arrays can form in the domain where Pf > a 3 + T. For higher stress differences, britde shear
failure can occur where Pf < + T. b. Schematic profile o f fluid pressure as a function o f depth (solid line XY) for the
fault-controlled flow system illustrated in (a). Depth dependence o f fluid pressures equivalent to hydrostatic and (crs +
T) levels shown for comparison. For low stress differences (<J( - <4T ), hydraulic extension fracturing occurs where P(
■ cr3 + T. At higher stress differences, shear failure may occur at lower fluid pressures.

Wood’s Reef, Victoria (Clappison, 1954; Edwards, 1954), and For T{greater than or equal to 0, shear failure will occur;
at Lamaque, Quebec (Robert, 1990); and gold-bearing vein for rf less than 0, shear failure will not occur. This parame­
systems hosted within dikc-like, competent felsic intrusions ter is the same as the Coulomb stress param eter used to
in the St Ives goldfield in Western Australia. predict locations o f aftershocks triggered by stress transfer
In the seismogenic regime, fluid pressure and shear after major earthquakes (King et al., 1994; Stein, 1999).
stress cycling associated with fault-valve behavior potentially Similarly, the growth o f hydraulic extension fractures will
lead to complex, but cyclically repeated changes to pore occur where and when
fluid factors and stresses (Sibson, 2001). Time ancl space
variations in pore factors, shear stress, and normal stress are Pf > ff3 + T. (24)
the main parameters influencing the timing and location of
episodic fracture, and distribution of high-permeability T hat is, where and when fluid pressure is highest, c;* is
zones within hydrothermal systems. least, or tensile strength is least (Cox et ah, 1987).
According to Coulomb failure criteria, brittle shear fail­ Understanding how stress and fluid pressure states vary in
ure occurs where and when shear stress rs is greater than a rock mass has important implications for predicting local­
the rock shear strength; that is, where ization o f fluid flow and ore deposition in hydrothermal sys­
tems. The use of Coulomb failure criteria to predict where
rs > C + M ^ - P t). (22) deformation-induced flow paths can form requires a knowl­
edge of the variation of shear stress, normal stress, fluid pres­
Proximity to shear failure can, thus, be described by a sure, friction coefficients, and cohesive strengths in rock
tim e-dependent param eter, Tp which is the difference masses in three dimensions, over time. Numerical modeling
between the shear stress and the shear strength. Thus, approaches such as that by Holyland and Ojala (1997) used
two-dimensional and three-dimensional linear elastic defor­
Tf=Ts_ C -/i(< rn - P f). (23) mation modeling to predict where minimum mean stresses
 or effective minimum principal stresses occur. However, cient connectivity to create networks that link fluid source
equation (23) illustrates that these approaches are highly rocks and ore deposition sites. Percolation theory (Sahimi,
simplified and might not adequately treat all the important 1994) provides useful insights about the evolution of
parameters controlling “proximity to failure.” In pardcular, hydraulic connectivity and the partitioning of fluid flow
potential spatial variations in fluid pressures need to be mod­ amongst elem ents of a network of relatively perm eable
eled. Modeling strategies, used to assist prediction of where structures in an otherwise low-permeability medium
fluid flow and mineralization may be localized, need to treat (Berkowitz, 1995; Cox and Knackstcdt, 1999).
coupling between stress and fluid pressure, as well as poten­ In the same way as for grain-scale pore networks, macro­
tial time and space variations in stress states associated with scopic percolation networks, comprising faults, fractures,
stress relief, stress transfer, and fluid pressure fluctuations and shear zones, can be described in terms of three types of
during the seismic cycle. This is a fertile field for further elements: backbone, dangling, and isolated elements (Fig.
development. 20a). Backbone elements provide a direct connection from
An example of potendal feedbacks between deformation, one side of the system to the other (e.g., metal source reser­
effective stress states, and permeability in controlling ore voir to ore deposit) and carry the bulk of the fluid flux. Dan­
localization is provided by the distribution of some gling elements, also known as dead-end elements, branch
fault/shear-hosted mesothermal gold deposits around shear from the flow backbone and act as fluid feeders to the back­
systems. Archean mesothermal deposits are typically located bone in the upstream part of the system, or as distributary or
within low displacement faults, shear zones, and associated discharge structures in the downstream part o f the system.
fracture systems adjacent to larger, crustal-scale fault systems Isolated elements are disconnected from both the backbone
(Robert et al., 1995; Cox, 1999). Timing, spatial, and kine­ and dangling elements in the network, and are, therefore,
matic relationships between low-clisplacement and high-dis- low-flux sunctures not connected to fluid reservoirs.
placement structures indicate that many of the low-displace­ At very low bulk crustal strains, most of the faults or shears
ment faults, shear zones, and fracture networks that host Au in a deforming domain will be short, isolated structures. With
deposits probably formed as aftershock structures in response increasing deformation, active faults and shears increase in
to slip events on major faults. Co-seismic stress transfer is a length and surface area; new structures also nucleate and
major factor controlling the locadon of aftershock activity grow so that fault connectivity increases with strain. A critical
(Stein, 1999). Accordingly, fluid invasion of slipped fault seg­ point, known as the percolation threshold (Sahimi, 1994), is
ments, the consequent decrease in effeedve stresses, and reached when enough elements connect to allow fluid flow
fluid-driven failure after major slip events, may act together across the entire width o f the network. Reaching die percola­
with mainshock-related stress transfer in localizing growth of tion threshold corresponds to the sudden onset o f flow. The
low-displaccment structures, high fluid flux, and mineraliza­ point at which the percolation threshold is reached is depen­
tion that produces mesothermal gold systems. d en t on several factors including strain accommodated by
growth o f permeable faults, fractures, and shear zones, as well
Evolution o f Flow Pathways during Deformation
as by fracture geometries and relative rates o f fracture growth
and nucleation (Zhang and Sanderson, 1994; An and Sam-
Localization o f deformation and flow
mis, 1996; Roberts et al., 1998).
The strain distribution associated with crustal deformation Partitioning o f flow among elements o f a fracture/shear
is typically very heterogeneous, with higher strains and asso­ network is dependent on the relative proportions o f back­
ciated permeability enhancem ent being localized along bone, dangling, and isolated elements (Cox and Knackst-
structures such as faults and shear zones. The significance of edt, 1999). Just above the percolation threshold, the flow
active fracture networks in controlling fluid migration is illus­ backbone is a very small fraction o f the total network, but
trated by analytical solutions and numerical modeling stud­ most of the flow is localized along this part o f the system
ies of two-dimensional flow patterns around high-permeabil­ (Fig. 20b). At higher strains the proportion o f elem ents
ity zones in a less permeable matrix (Phillips, 1991; Matthai that are p art o f the backbone progressively increases, so
and Roberts, 1997; Taylor et al., 1999). For permeable struc­ that flow becomes progressively m ore evenly distributed
tures inclined at low angles to the regional gradient of across the system as m ore faults or shears becom e con­
hydraulic head, fluid focusing occurs around the higher nected to each other and to fluid reservoirs.
pressure (upstream) levels of shear zones, whereas fluid dis­ At low displacements, all fracture segments are isolated
charge occurs in the lower pressure (downstream) levels o f from one another and the fluid reservoir. The percolation
shear zones (Fig. l).T h e dimensions of fluid charge regions threshold is reached when approximately 30 percent o f the
are comparable to the dimensions o f the high-permeability sites are occupied. For the three-dimensional fault configu­
segments o f faults and shear zones. Importantly, shears ration illustrated in Figure 20b, and using typical
inclined at high angles to the regional hydraulic head gradi­ length/displacem ent scaling relationships for faults
en t are much less effective at focusing fluid flow. (Scholz, 1990), the percolation threshold is reached at bulk
strains o f only a few percent. For three-dimensional perco­
Development ofpercolation networks in hydrothermal systems
lation, flow pathways can be very tortuous and produce
; Fracture/shear-controlled hydrothermal systems develop point-like distribution o f high fluid flux sites along
where and when linked permeable structures develop suffi­ restricted segments of faults and shear zones (Fig. 21). This
a jfj
§cc
0 mal systems well above the percolation threshold, fluid flow
3UJ
1 lation. This more dispersed flow is likely to produce more
o distributed, potentially lower grade mineral deposits.
I clusters o f deposits, along a small proportion of the total
b systems, where flow occurs through dense and highly inter­
FRACTION OF TOTAL SITES
OCCUPIED
Fig. 20. a. Schematic two dimensional representation o f a fault/fracturc
network consisting o f isolated dem ents, dangling dem ents, and the back­
bone structure. Most flow is localized along the flow backbone. Dangling
elements in the upstream (lower, in this ease) part o f the system feed fluid
to the backbone o f the system, whereas dangling elements o f the network
in the downstream (upper) pait o f the system act as fluid discharge sites, b.
Fraction o f isolated, dangling, and backbone sites as a function o f total
number o f sites for the three-dimensional ease o f conjugate fractures or
faults inclined at -15s to die bulk flow direction (after Cox and Knackstcdt.
1999).
factor will act together with variations in fracture densities
and apertures to control flow distribution.
Several points emerge directly from percolation model­
ing. Firstly, th e result that the percolation threshold is
reached at bulk strains o f only a few percent indicates that
fracture-controlled hydrothermal systems and associated
ore deposits can develop at very low strains. High-displace-
m eiu structures and large regional strains are not necessary
to facilitate high fluid fluxes. In this regard, small strains
during the initial stages of changes in plate kinematics may
be im portant in developing new fracture systems that tap
into previously inaccessible fluid reservoirs.
Secondly, for networks just above the percolation thresh­
old, where the backbone is a very small proportion o f the
total fault population, flow is localized on relatively few struc­
tures, which link fluid sources and sinks. This situation max­
imizes fluid/rock ratios in die fluid-accessible parts of perco­
lation networks, and therefore maximizes the potential for
generation of “giant” ore deposits. In conuast, for hydrother­
is distributed over a larger proportion of die fracture popu­
An example of this control could be provided by the typ­
ical distribution o f m ajor mesothermal gold deposits, or
population of kinematically related (and thereby broadly
co-active) faults and shear zones (Cox, 1999). This distribu­
tion indicates dial most fluid flow is localized along a small
num ber o f active structures, and, therefore, that the
fau lt/sh e ar networks that successfully generate m ajor
mesothermal lode gold systems have operated near die per­
colation threshold during ore genesis. Some poi*phyiy-type
connected fracture networks (e.g., Haynes and Titley,
1980), may be deposit-scale examples o f flow systems well
above the percolation threshold.
At any instant, hydraulic connecdvity in a network is
dependent upon which parts of the system are actively frac­
turing (and hence permeable). For example, as deforma-
don migrates through a shear system, not all parts o f a geo­
metrically connected and kinematically linked network of
faults and shear zones are active. So, the locus of fluid flow
will be transient and m igrates with active deform ation
through the system. In thrust systems, where new thrusts
commonly nucleate progressively deeper into the footwall
o f previously active thrust sheets, the locus o f fluid flow will
migrate towards the foreland widi time.
Particularly in the middle to upper crustal seismogenic
regime, rapid changes in permeability in fault networks on
timescales o f seismic slip recurrence can also lead to sud­
den changes in the location and architecture o f flow back­
bones, unless m ost displacem ent is also localized on the
fluid flow backbone.
Self-organization o f flow networks
It was noted earlier that the distribution o f mesothermal
gold deposits along a small proportion of die total popular
don o f kinematically related and co-active faults or shear
zones may indicate that these systems develop m ost effec­
tively in percolation networks that operated close to their
percolation threshold. This raises a question: do some
processes maintain fault/fracture/shear networks near the
percolation threshold? Such behavior is a manifestation o f
"self-organized criticality,” a term used to describe processes
which hold complex o r disordered systems near their critical
point (e.g., percolation threshold). Many natural processes
such as slope failure in sandpiles and earthquake rupturing
events exhibit self-organized behavior (Bak, 1997).
In the seismogenic regime, interseismic fault scaling can
repeatedly drive networks back below the percolation
threshold after rupture events, so that on the lifetime o f the
hydrothermal system the network may oscillate about the
 FLUID OUT A

Fig. 21. Three-dimensional flow paths associated with flow through part o f a network of intersecting fractures or faults
ju s t above th e percolation threshold. Fluid enters the network at the bottom and follows a tortuous path (shown in
black) to higher structural levels. Three-dimensional flow near the percolation threshold produces a point-like distrib­
ution o f high flux in map view.

percolation threshold. However, other factors may also lead
to self-organization in hydrothermal systems.
The network m odeling described above assumed that,
although elements o f a percolation network nucleate ran­
domly, all elements grow at the same rate. In cases where
fracture systems connect to an overpressured fluid reser­
voir, invasion of high-pressure fluids along fluid-accessing
elements of the network may preferentially weaken these
elements and dramatically enhance their growth rates rela­
tive to elements isolated from fluid reservoirs. This occurs
through the effects o f high fluid pressures promoting fail­
ure and reaction-softening processes. Accordingly, con­
nected elements can slip and grow at much faster rates than
isolated elements. The fluid-driven growth, or “self-genera­
tion” (Sibson, 1996), o f percolation networks in response
to invasion o f overpressurized fluid, therefore, provides a
positive feedback between fluid access and fracture growth
rate. This may localize repeated slip and fluid flow on the
flow backbone that form s at the percolation threshold.
Fluid-driven network growth may well lead some hydrother­
mal systems to self-organize near the percolation threshold.
Such extreme fluid focusing generates high potential for
formation o f giant ore deposits.
Ore deposition in percolation networks
In many cases, ore deposition is controlled by processes
such as (1) fluid mixing, (2) fluid-rock reaction, (3) phase
separation driven by sudden fluid pressure drop, and (4)
gradients in pressure and tem perature (Skinner, 1997).
O re deposition is dep en d en t on structurally controlled
delivery o f fluids to appropriate reaction sites. These
processes operate with different effectiveness in different
parts of percolation networks.
Isolated elements of percolation networks have the low­
est fluid flux and, therefore, have minimal ore potential in
hydrothermal systems. The ore potential of backbone and
dangling elements depends on the nature o f the ore depo­
sition reactions, and whether the site is in the upstream o r
downstream part o f the system.
Most flow occurs along the network backbone. Even
though they are high-flux sites, flow backbones are unlikely
to be sites o f substantial fluid-rock interaction o r fluid mix-
mg away from their upstream o r downstream terminations.
However, backbone sites do provide potential for ore depo­
sition in response to tem perature and pressure gradients.
Backbone elements also have the potential for sudden fluid
pressure drops (and associated phase separation) caused by
co-seismic dilatancy associated with large slip on these struc­
tures in the seismogenic regime. The potential for ore
deposition involving fluid-rock reaction is dependent on
the grain-scale pervasiveness of flow. Especially during flow
along macroscopic fractures, fluid migrating along the flow
backbone has limited interaction with the fracture walls.
However, where flow is controlled by grain-scale dilatancy,
such as in cataclasite o r ductile shear zone materials, fluid-
rock reaction can occur along the flow path.
High potential for fluid-rock reaction is provided in the
downstream parts of percolation networks, where fluids dis­
charge from active structures into die surrounding, perme­
able rock medium. Fluid discharge from the downstream
termination of the flow backbone (Fig. 1), or from down­
stream dangling members o f the network, promotes fluid-
rock interaction. Similarly, this part o f the flow network
generates potential for fluid mixing if deeply derived fluids
discharging from the network interact with shallow-level
fluid reservoirs already present in the discharge regime, as
may be the case for many types of porphyry-type and
epithermal systems.
Potential for ore deposition by fluid m ixing reactions
also occurs in the upstream parts o f networks. T he
upstream regions o f flow backbones are dom inated by
focusing o f fluids from dispersed reservoirs and host rocks,
into the perm eable backbone network. Mixing of fluids
derived from different reservoirs will be particularly effec­
tive where dangling (fluid feeder) elem ents of networks
intersect the upstream segments o f backbone structures.
Conclusions
Active deformation is necessary to generate and maintain
permeability, and to sustain large-scale fluid flow in many
hydrothermal systems. At upper crustal levels, and in rocks
with low intergranular porosity, flow is controlled by per­
meability associated with grain-scale cataclasis and macro­
scopic fracture growth. At deeper crustal levels, fluid flow in
active shear zones typically is controlled by deformation-
induced, grain-scale m icrocrack permeability. However,
macroscopic fracture perm eability still develops in high
pore fluid factor regimes at elevated temperatures.
The distribution and geom etry of fracture-controlled
fluid pathways is governed by fluid pressure and stress state.
High pore fluid factors are important in driving growth of
fracture permeability at all scales. High pore fluid factor
regimes tend to develop above leaking, suprahydrostatic-
pressured fluid reservoirs, and below low-permeability
domains. Faults and shear zones play a key role in tapping
fluids from deeper level fluid reservoirs.
The highest fluid flux in individual faults or shear zones
is localized in sites with the highest fracture apertures
a n d /o r fracture density. Thus, both dilational and contrac-
tional bends and jogs on faults and shear zones, as well as
fault termination zones (wing cracks and splays) and fault
intersections, form structural sites with high potential for
localizing fluid flow.
T he macroscopic fluid pathways in fracture-controlled
hydrothermal systems are influenced by the evolving con­
nectivity among elements of fracture/fault/shear networks
during progressive deform ation. The most extreme flow
localization and greatest potential for ore localization
occurs at the onset of system-wide fluid flow, when the sys­
tem reaches the percolation threshold. Percolation thresh­
olds may be reached at bulk strains as low as a few percent.
Systems well above die percolation threshold are associated
with m ore dispersed fluid flow, and may provide lower
potential for ore formation.
Acknowledgments
This work was funded partly by a g ran t from the Aus­
tralian Research Council, as well as by support from the
Research School o f Earth Sciences and WMC Resources
Ltd. R. Henley, S. Munroe, P. Nguyen, K- Ruming, R. Sib-
son, W. Stone, J. Streit, and S. Zhang all provided valuable
discussion about ore systems and fluid flow in rock media.
KA. Foxford is thanked for providing a preprint. The con­
structive review comments o f M. Jebrak and S. Temperley,
as well as the editorial contributions o f J. Richards, are also
greatly appreciated.
REFERENCES
An, L.-J., and Samtnis, C.G., 1996, A cellular automaton for the develop­
m ent o f crustal sltcar zones: Tcctonophysics, v. 253. p. 247-270.
Angcvinc, C.L., Turcouc, D .l_ and Furnish. M.D., 1932, Pressure solution
liihification as a mechanism for stick-slip behavior o f faults: Tectonics,
v. 1, p. 151-160.
Bak, P., 1997, How nature works— the science o f sclforganizcd criticality:
Oxford, Oxford University Press, 212 p.
Bear, J., 1972, Dynamics o f fluids in porous media: New York, Elsevier, 764 p.
Berkowitz, B., 1995, Analysis o f fracture network connectivity using perco­
lation theory: Mathematical Geology, v. 27, p. 467—183.
Bourbic, T., and Zinszncr, B., 1985, Hydraulic and acoustic properties as a
function o f porosity in Fontainebleau sandstone: Journal o f Geophysical
Research, v. B13, p. 11,52-1-11,532.
Brace, W.F., 1990, Permeability o f crystalline and argillaceous rocks—sta­
tus and problem: International Journal o f Rock Mechanics, Mining Sci­
ence and Gcomcchanics Abstracts, v. 17, p. 876-893.
Brantley, S.L., Evans, B„ Hickman, S.H., and Crcrar, D A ., 1990, Healing
o f microcracks in quartz— implications for fluid flow: Geology, v. 18,
p. 136-139.
Brown, S.R.. 1987, Fluid flow through rock joints— the effect o f surface
roughness: journal o f Geophysical Research, v. 92, p. 1337-1347.
Cnthles, L.M., Ercndi, A.H.J., and Barrie, T„ 1997, How long can a
hydrothermal system be sustained by a single intrusive event?: Economic
Geology, v. 92, p. 766-771.
Chacc, F.M., 1949, Origin o f the Bendigo saddle reefs, with comments on
the formation o f ribbon quartz: Economic Geology, v. 44, p. 561-597.
Clappison, R.J.S., 1954, The M orning Star M ine, W ood’s Point, in
Edwartls, A.B., ed.. Geology o f Australian ore deposits: Melbourne, Aus-
tralasian Institute o f Mining and Metallurgy, p 1077-1082.
Clout, J.M.F., C lcghom , J.H., and Eaton, P.C., 1990, Geology o f the Kalgo-
orlie goldfield, in Hughes, F.E., ed„ Geology o f the mineral deposits o f
Australia and Papua New Guinea: Australasian Institute o f Mining and
Metallurgy Monograph Scries, v. I-J. p. 411—431.
Cox, S.F.. 1995, Faulting processes at high fluid pressures An example o f
fault-valve behavior from the Wattle Gully Fault, Victoria, Australia:Jour­
nal o f Geophysical Research, v. 100, p. 841-859.
 ------ 1999, Deformaticmal controls on the dynamics o f fluid flow in
mcsothcrmal gold systems, in McCaffrey, K., Loncrgan. L., and Wilkin­
son,.!.. cds.. Fractures, fluid flow, and mineralization: Geological Soeieiy
o f London Special Publication, v. 155, p. 123-140.
Cox, S.F., and Etheridge. M A , 1989, Coupled grain-scale dilaiancy and
mass transfer during deformation at high fluid pressures, Mt Lycll area,
Tasmania: Journal o f Structural Geology, v. l l , p . 147-162.
Cox, S.F., and Knackstedt, M A ., 1999, Ore generis in fracture-controlled
hydrothermal systems: Percolation theory approaches: PACRIM *99,
Bali, Indonesia, October 10-13,1999, Proceedings, p. 639-6-12.
Cox, S.F., and Munroe, S.M., 1999, Fluid-driven faulting processes in an
intrusive-related hydrothermal system, Porgera, Papua New G uinea
(abs.): Eos. v. 80 (46), p. F728-F729.
Cox. S.F.. and Paterson, M.S.. 1991, Experimental dissolution-precipitation
creep in quartz aggregates at high temperatures: Geophysical Research
Letters, v. 18. p. 1401-1404.
Cox, S.F., Etheridge, M A ., and Wall, VJ., 1987, The role o f fluids in syn-
tectonic mass transport, and die localization o f mctamorphic vein-type
ore deposits: Ore Geology Reviews, v. 2. p. 63-86.
Cox, S.F, Wall, VJ., Etheridge, MA ., and Potter, T.F.. 1991, Dcformational
and m ctam orphic processes in the form ation o f mcsotherm al vein-
hosted gold deposits— examples from the Lachlan Fold Belt in central
Victoria, Australia: Ore Geology Reviews, v. 6 . p. 391-423.
David. C., and Darot, M., 1989, Permeability and conductivity o f sand­
stone. in Maury. V., and Fourmaintreaux, D.. cds., Rocks at great depth:
Amsterdam, Dalkcma, p. 203-210.
Edwards, A.B., 1954, Mines o f the Waihalla-Wood's Point auriferous belt,
in Edwards. A.B.. cd.. G eology o f Australian ore deposits: Melbourne,
Australasian Institute o f Mining and Metallurgy, p. 1061-1076.
Etheridge, M.A., 1983, Differential stress m agnitudes during regional
deformation and metamorphism: Upper bound imposed by tensile frac­
turing: Geology, v. 11, p. 231—234.
Fischer, GJ.. and Paterson, M.S., 1992, Measurements o f permeability and
storage capacity in rocks during deformation at high temperature and
pressure, in Evans, B., and Wong. T.-f„ cds., Fault mechanics and trans­
port properties o f rocks: San Diego, Academic Press, p. 213-252.
Forster, C„ and Smith, L,, 1990, Fluid flow in tectonic regimes: Mineralog-
ica! Association o f Canada Short Course Handbook, v. 18, p. 1—17.
Foxford, KA., Nicholson, R., Polya, D A ., and Hebblethwaite, R.P.B., 2000,
Extcnsiona! failure and hydraulic salving at Minas da Panasqucim, Por­
tugal—evidence from vein spatial distributions, displacem ents, and
geometries: Journal o f Structural Geology, v. 22, p. 1065-1086.
Fyfe, W.S., 1987, Tectonics, fluids, and ore deposits— mobilization and
rcmobilization: Ore Geology Reviews, v. 2, p. 21-36.
Garvcn, G., 1985. The role o f regional fluid flow in the genesis o f the Pine
Point deposit, Western Canada sedimentary basin: Economic Geology,
v. 80, p. 307-324.
Garven, G., and Freeze, R.A., 1984, Theoretical analysis o f the role o f
groundwater flow in the genesis o f stratabound ore deposits. 1 . Mathe­
matical and numerical m odel: American Journal o f Science, v. 284,
p. 1085-1124.
Gu4gucn, Y., anti Dienes, J., 1989, Transport properties o f rocks from sta­
tistics and percolation: Mathematical Geology, v. 21, p. 1-13.
G uegucn, Y., and Palciauskas. V., 1994, Introduction to the physics o f
rocks: Princeton, Princeton University Press, 294 p.
Haar, L., Gallagher, J.S., and Kell, G.S., 1984, Steam tables: Washington
D.C., Hampshire Publishing Corporation, 319 p.
Haynes, F.M., and Tilley, S.R., 1980, The evolution o f fracture-related per­
meability within the Ruby Star granodioritc, Slcrrita porphyry copper
deposit, Pima County, Arizona: Economic Geology, v, 75, p. 673-683.
Henley, R., Trucsdell, A.H., and Barton, P.B., Jr., 1985, Fluid-mineral equi-
libria in hydrothermal systems: Reviews in Econom ic Geology, v. 1 ,
p. 1-267.
Hickman, S.H., and Evans, B., 1987, Influence o f geom etry upon crack
• healing in calcitc: Physics ami Chemistry o f Minerals, v. 15, p. 91-102.
Holness, M.B., 1997, The permeability o f non-deforming rock, in Hoi ness,
• M.B., ed„ Deformation-enhanced fluid transport in the Earth's emst and
. mantle: London, Chapman and Hall, p . 9-39.
Holyland, P.W., and Ojala, VJ., 1997, Computer-aided structural targeting
. in mineral exploration: Australian Journal o f Earth Sciences, v. 44,
p. 421-432.
Hulin, C.D., 1929, Structural control o f ore deposition: Economic Geol­
ogy, v. 2-1, p. 15-49.
Kanagawa, K., Cox, S.F., and Zhang, S., 2000, Eflccts o f dissolution-precip­
itation processes o n the strength and m echanical behavior o f quartz
gouge at high-temperature hydrothermal conditions: Journal o f Gco-
physical Research, v. 105, p. 11,115-11,126.
King. G.C.P., Stein. R.S., and Lin. J., 199-1, Static stress changes and the
triggering o f earthquakes: Bulletin o f the Scismological Society o f Amer­
ica, v. 84, p. 935-953.
Knackstedt, M., and Cox. S.F., 1995, Percolation and pore geom etry o f
crustal rocks: Physical Review E, v. 51, p. R5181-R5184.
Knipe. R.J., 1998, Faulting, fault sealing and fluid flow in hydrocarbon
reservoirs—an introduedon, i«Jones, G., Fisher. Q J., and Knipe, RJ.,
etls., Faulting, fault sealing, and fluid flow in hydrocarbon reservoirs:
Geological Society o f London Special Publications, v. 147, p. vii-xxi.
Manning, C.E., and Ingebritscn, S.E., 1999, Permeability o f the continen­
tal crust—th e implications o f geotherm al data and m ctam orphic sys­
tems: Reviews in Geophysics, v. 37, p. 127-150.
Matthai, S.K., and Roberts, S.G.. 1997, Transient versus continuous fluid
flowin scismically active faults: An investigadon by electric analogue and
numerical m odelling, in Jamtveit, B., and Yardlcy, B.W.D., eds., Fluid
flow and transport in rocks— mechanisms and effects: London, Chap­
man and Hall, p. 263-292.
McCaig. A.M ., and Knipe, R.J.. 1990, Mass transport m echanism s in
deforming rocks—recognition using mtcrostructural and microchemical
criteria: Geology, v. 18, p . 82*1-827.
McKinstry, H.E., 1948, Mining geology: Englewood Cliffs, N.J., Prentice-
Hall Inc., 680 p.
Muir-Wood, R., and King, G.C.P., 1993, Hydrological signatures o f earth­
quake strain: Journal o f Geophysical Research, v. 98, p. 22,035-22,068.
Munroe, $31., 1995, T h e Porgera gold deposit. Papua New Guinea: The
influence o f structure and tectonic setting on hydrothermal fluid flow
and m ineralisation a t a convergent margin: PACRIM *95, Auckland,
N.Z., November 19-22,1995 Proceedings, p. 413—>116.
N cwhousc. W.H., 1942, O re deposits as related to structural features:
Princeton, N.J., Princeton University Press, 280 p.
Nguyen. P.T., Cox, S.F., Powell, C.M., and Harris, L., 1998, Fault-valve
behaviour in optim ally-oriented shear zon es at Revenge gold m ine,
Kambalda, Western Australia: Journal o f Structural G eology, v. 20,
p. 1625-16*10.
Norton. D., and Knight, J., 1977, Transport phenomena in hydrothermal sys­
tems: Cooling plutons: American Journal o f Science, v. 277, p. 937-981.
Ord, A . and Oliver, N.H.S., 1997, Mechanical controls on fluid flow dur­
ing regional metamorphism: som e numerical models: Journal o f Mcta­
morphic Geology, v. 15, p. 345-359.
Oliver, N.H.S., Ord, A., Vnlentn, R.K., and Upton, P., 2001, Deformation,
fluid flow, and ore genesis in heterogeneous rocks, with examples and
numerical models from the Mount Isa district, Australia: Reviews in Eco­
nomic Geology, v. 14, p. 51-74.
Peach, C.J., and Spiers, C.J., 1996, Influence o f crystal plastic deformation
on dilatancy and permeability developm ent in synthetic salt rock:
Tectonophysics, v. 256, p. 101-128.
Phillips, O.M., 1991, Flow and reactions in permeable rocks: Cambridge,
Cambridge University Pr ess, 285 p.
Ramsay, J.G., 1980, T he crack-scal mechanism o f rock deformation:
Nature; v. 284, p. 135-139.
Robert, F., 1990, Structural setting and control o f go!d<]uartz veins o f the
Val d'Or area, southeastern Abitibi Subprovince, in Ho, S.E., Robert, F..
and Groves, D.I., cds.. Gold and base metal mineralisation in the Abitibi
Subprovince, Canada, with emphasis on the Quebec segment: Nedlands.
University o f Western Australia Publication 24, p. 167-209.
Robert. F., and Poulscn, K.H., 2001, Vein formation and deformation in
green ston e gold deposits: Reviews in E conom ic G eology, v. 14,
p. 111-155.
Robert. F., Boullier, A.-M., and Firdaous, K., 1995. Gold<|uartz veins in
mctamorphic icrranesand their bearing on the role o f fluids in faulting:
Journal o f Geophysical Research, v. 100, p. 12,861-12.881.
Roberts, S., Sanderson, D.J., and Gumiel, P„ 1998, Fractal analysis o f Sn-W
mineralization from central Iberia— insights into the role o f fracture
connectivity in die formation o f an ore deposit: Economic Geology, v. 93,
p . 360-365.
Rumble, D., 1994, Water circulation in meuimorphisnt: Journal o f Geo­
physical Research, v. 99, p. 15,499-15,502.
Rumble, D., Ferry, J.M., Hoering. T.C., and Boucoi, A.J., 1982, Fluid fiow
during metamorphism at the Beaver Brook fossil locality, New Hamp­
shire: American journal o f Science, v. 282, p. 886-919.
Sahitni. M., 1994, Applications o f percolation 1 11 0 0 1 7 : London, Taylor and
Francis, 258 p.
Scholz. C.H., 1990, The m echanics o f earthquakes and faulting: Cam­
bridge, Cambridge University Press, 439 p.
Sibson, R.H., 1985, A note on fault reactivation: journal o f Structural Geol­
ogy, v, 7, p. 751-754.
------ 1987, Earthquake rupturing as a mineralizing agent in hydrothermal
systems: Geology, v. 15, p. 701-704.
—— 1993, Load-strengthening versus load-weakening faulting: journal o f
Structural Geology, v. 15, p. 123-128.
------ 1996, Structural permeability o f fluid-driven fauit-fracturc meshes:
journal o f Structural Geology, v. 18, p. 1031-1042,
— 2001, Scism ogcnic framework for hydrothermal transport and ore
deposition: Reviews in Economic Geology, v, 14, p. 25-50.
Sibson, R.H., Robert, F., and P oulsen, K.H., 1988, High-anglc reverse
faults, fluid-pressure cycling, and nicsoihcrmal gold deposits: Geology,
v. 16, p. 551-555.
Skinner, B.J., 1997, Hydrothermal mineral deposits—what we d o and
don't know, in Barnes, H.L., ec!., G eochemistry o f hydrothermal ore
deposits. 3rd ed.: New York, Wiley, p. 1-26.
Smith, C.S., 1964, Som e elem entary principles o f polycrystallinc
microstmcturcs: Metallurgical Reviews, v, 9, p. 1-47.
Stein, R.S., 1999, The role o f stress transfer in earthquake occurrence:
Nature, v. 402, p. 605-609.
Stormont, J.C., and Daemon, J.K., 1992, Laboratory study o f gas perme­
ability changes in rock salt during deformation: International Journal o f
Rock M echanics and M ining Science G com cchanics Abstracts, v. 29,
p . 325-342.
Taylor, W.I., Pollard, D.D., and Aydin, A., 1999, Fluid flow in discrete joint
sets—field observations and numerical simulations: journal o f Geophys­
ical Research, v. 104, p. 28,983-29,006.
Tsang. Y.W.. and Witherspoon, PA., 1981, Hydromecitanical behavior o f a
deformable rock fracture subject to normal stress: Journal o f Geophysi­
cal Research, v. 8 6 , p. 9287-9298.
Waite, M.E., G e, S., and Spettier, H., 1999, A new conceptual m odel for
fluid flow in discrete fractures—an experimental and numerical study:
Journal o f Geophysical Research, v. 104, p. 13,049-13,059.
Walsh, j.B ., 1981, Effect o f pore pressure and confining pressure on per­
meability: International Journal o f Rock Mechanics, v. 18, p. 429-435.
Walther,J.V., 1990. Fluid dynamics during progressive regional inctamor-
phism, in Brcdehoft,J.D., and Norton, D.L., cds.. The role o f fluids in
crustal processes: W ashington, D.C.. National Academy o f Sciences,
P. 6-1-71.
Watson, E.B., and Brenan,J.M., 1987, Fluids in the lithosphere. I. Expert*
menially-determined wetting characteristics o f CCVH.,0 fluids and their
implications for fluid transport, host rock physical properties, and fluid
inclusion formation: Earth and Planetary Science Letters, v. 85, p. 497-515.
W itherspoon, PA.. Wang, J.S.Y., Iwai, K., and G ale.j.E .. 1980. Validity o f
the cubic law in a deformable rock fracture: Water Resources Research,
v. 19, p. 1016-1021.
Zhang, S., Paterson, M.S., and Cox, S.F., 1994a, Porosity and permeability
evolution during hot isostatic pressing o f calcitc aggregates: Journal o f
Geophysical Research, v. 99, p. 15,741-15,760.
Zhang, S., Cox, S.F., and Paterson, M.S., 1994b, T he influence o f room
temperature defoim aiion o n porosity and permeability in calcitc aggre­
gates: Journal o f Geophysical Research, v. 99, p. 15,761-15,775.
Zhang. S-. Tullis, T.E., and Scruggs, VJ., 1999, Permeability anisotropy and
pressure dependence o f permeability in experimentally sheared gouge
materials: Journal o f Structural Geology, v. 21, p. 795-806.
Zhang. S., FitzGerald, J.G., and Cox. S.F., 2000, Reaction-enhanced per­
meability during decarbonation o f caicitc + quartz -» wollastonile + car­
bon dioxide. Geology, v. 28, p. 911-914.
Zhang, S.. Cox. S.F., and Paterson, M25., in press. Microcrack growth and
healing in deformed calcite aggregates: Tectonophysics.
Zhang, X., and Sanderson, D.J., 1994, Fractal structures and deformation
o f fractured rock masses, in Kruhl, J.H., cd., Fractals and dynamic sys­
tems in geoscience: Berlin, Springer-Vcrlag, p. 37-52.
Society of Economic Geologists
Reviews v. H , 2001, p. 25— 50

Chapter 2

Seism ogenic Framework for Hydrothermal Transport and Ore Deposition

Richard H . S ibson *
D epartm ent o f Geology, U niversity o f O tago, P.O. B o x 5 6 , N ew Z ea la n d

A further result attending the disturbances o f the surface of the Earth has been to produce rents
or Fissures in the rocks which have been subjected to these violent movements, and to convert
them into receptacles o f metallic ores accessible by the labors of man. The greater part of metal­
liferous veins o r ig in a te d in enormous cracks and crevices, penetrating irregularly and obliquely
d o w n w a rd s to a n u n k n o w n depth, and resembling the rents and chasms which are produced by
m o d e r n e a rth q u a k e s .
— VV! B u ckla n d , 1 8 3 6 (w ritin g o n th e “a d va n ta g eo u s effect o f
distuvbingforces in g iv in g origin to m ineral v e in s’")

Abstract
Fault motion in th e upper c o n tin e n ta l crust is accommodated principally by earthquake rupturing
within a seismogenic z o n e w h o se base, depending on composition, generally lies in the 300° to 450°C
tem perature range. R u p tu r e in itia tio n , propagation, and termination within this zone arc affected by
structural and rh e o lo g ic a l irregularities. Sawtooth accumulation and release of shear stress on the seis­
mogenic structures leads to cy clin g o f both shear and mean stress (affecting fluid content) throughout
the surrounding rock m ass, w ith significant fluid redistribution throughout the aftershock phase follow­
ing large earthquakes. S tr u c tu r a l permeability in such regions is intrinsically dynamic: episodic creation
of permeability a c c o m p a n y in g seismic slip and fracturing is counteracted by the development o f low-per­
meability fault gouge and hydrothermal cementation, so that flow systems are modulated by intercoupled
stress and permeability cycling. Because criteria for all modes o f brittle failure and fault reshcar depend
o n fluid pressure as well as tectonic stress, a variety o f mechanisms may link fluid redistribution to
episodic faulting and fracturing. Stress changes accompanying large-scale rupturing on established faults
redistribute fluids through sul>$idiary fracture networks during aftershock periods, but packages of over-
pressured fluid migrating through stressed crust may also create new structural permeability by distrib­
uted brittle failure, generating earthquake swarms.
T he fluid pressure state at different crustal levels is critical to the formation and preservation o f void
space. Fluid overpressuring above hydrostatic values is generally easier to sustain in compressional tec­
tonic regimes, but maximum sustainable overpressure in any particular setting depends not only on the
intrinsic permeability o f the rock mass but also on the tectonic stress state and existing fault architecture.
Large-scale hydrothermal flow through low-permeability rocks is often channeled within dilatant mesh
structures o f interlinked shear and extension fractures. These fault-fracture meshes can form and reacti­
vate only u n d er low effective stress ( o f < 0, o rP f > o f ) in the absence o f throughgoing low-cohesion faults
that are well oriented for frictional reactivation. High-flux flow o f this kind can, therefore, occur only
u nder special structural circumstances. In extensional-transtensional tectonic regimes, dilatant meshes
can be m aintained under hydrostatic fluid pressures in the shallow crust to depths dependent on rock
tensile strength, defining the cpizonal environment for mineralization. However, at all depths within
compressional-transpressional regimes, development o f fault-fracture meshes involves hydrothermal flu­
ids overpressured to ncar-lithostatic values. In particular, mesozonal lode mineralization requires the ac­
cumulation and interm ittent high-flux discharge of strongly overpressured fluids in the midcrust, most
commonly around the base o f the continental seismogenic zone.
Im portant precipitation mechanisms linked to intermittent seismic slip include the suction-pump mech­
anism arising from rapid slip transfer across dilational faultjogs and bends, and various forms of fault-valve
action where ruptures transect boundaries to overpressured portions of the crust. These mechanisms in­
duce abrupt localized reduction in fluid pressure at specific structural sites, triggering phase s e rra tio n
and hydrothermal precipitation throughout the postseismic period o f readjustment (aftershock phase).
However, renewal o f fault-fracture permeability may also lead to episodic mixing of fluids derived from dif­
ferent sources. For example, each fault-valve discharge may promote precipitation through the mixing of

♦E-mail, Rick.Sib$on@stonebow.otago.ac.nz
 originally deep, hot, overpressured fluids of mctamorphic a n d /o r magmatic origin with colder fluids cir­
culating in the near-surface hydrostatic regime.
Regional episodes o f fluid redistribution are likely to accompany major tectonic transitions (c.g., tec­
tonic inversion) because o f changes in the stress state and sustainable levels of fluid overpressure, and an
inherited architecture o f faults poorly oriented for slip in the new stress field. Evidence o f structural
channeling in such settings reveals interesting comparisons between the flow paths of hydrothermal and
hydrocarbon fluids.

Introduction tectonic stress field (Barton et al., 1995). Thus, processes of

fluid redistribution in the vicinity of faults, if not directly dri­
St u d ie s o f exhumed fault zones provide abundant evidence ven, seem likely to be strongly modulated by stress and per­
for their role as fluid conduits at all crustal levels (McCaig, meability cycling tied to the earthquake stress cycle (Fig. 1).
1997; Parry, 1998), but from the viewpoint of mineralization Different fluid flow regimes are, therefore, expected to
the key problem is to identify the specific conditions that prevail in intraplate regions, where seismic activity is low, in
promote focused, large-flux flow at specific sites (Cox et al., comparison with seismically active plate boundaries. In (he
2001). A large proportion of hydrothermal mineral deposits former, steady flow systems may develop under a relatively
are associated with faults and allied fracture systems that constant intraplate stress field. In the latter, seismic failure
were active in the upper, seismogcnic crust, although it is along m ajor fault systems, with episodic regeneration of
notable that the hosting faults are often comparatively low- structural permeability coupled to stress cycling in the sur­
displacement features (Knopf, 1929; McKinstry, 1948). Solu­ rounding crust, leads to cyclical modulation o f flow systems
bility criteria, incremental vein textures, and the need for (Sibson, 1994). This contribution discusses dynamic
rapid transport between different P-T-X environments to processes of fluid redistribution coupled to the earthquake
promote instability and localized precipitation suggest that stress cycle that may induce mineral precipitation at specific
mineralization generally results from multiple transient structural sites through dramatic reductions in fluid pressure
episodes of rapid flow (Henley and Berger, 2000). This, cou­ a n d /o r fluid mixing. In terms o f basic exploration practice,
pled with structural evidence that mineralization was synk- it emphasizes the importance o f identifying (1 ) the tectonic
inematic, developing in the same tectonic stress field as the stress regime and mode o f faulting, (2) the structural geom­
hosting faults, has led many workers over the years to postu­ etry related to fault irregularities, (3) the level in the crust,
late a relationship between episodes o f hydrothermal pre­ and (4) the fluid pressure state at the time of mineralization.
cipitation and seismic slip increments (e.g., Buckland, 1836;
Seismogenic Crust
Hulin, 1925; Knopf, 1929; Newhouse, 1942; Robert and
Brown, 1986; Cox, 1995). Earthquakes in the shallow crust arise from frictional
O ur understanding of the relationship between fault instability with ruptures propagating over p a rt o r all o f a
structure and earthquake activity hits improved greatly over fault surface (Scholz, 1998). Earthquake m agnitude (M)
the past three decades through the development o f high- increases with rupture dimension (L) and rupture area (A =
density seismograph networks and improved location tech­ L2), and the am ount o f slip (u) averaged over the rupture
niques (e.g., Hill etal., 1990). Paramount is the recognition surface. Scaling relationships have been established on the
that away from subduction interfaces, earthquakes are gen­ basis o f the near-constancy of earthquake stress drops allow­
erally confined to the upper half of deforming continental ing approximate correlations to be made between M, L, and
crust and that earthquake rupturing, involving frictional A and Cl (Table 1). In any active region, the statistical distri­
reactivation of existing faults, is die predominant means of bution o f different sized earthquakes follows a power law
accom m odating fault slip within the upper seismogenic relationship such that earthquake frequency increases by
crust. Additionally, modem seismotcctonic analyses employ­ about a factor of 10 for every unit decrease in magnitude.
ing focal mechanism studies have given us a far better pic­ Also shown in Table 1 is the relative seismic wave energy
ture of the styles and patterns o f fauldng that occur in dif­ release with respect to an Ml earthquake, estimated from
ferent tectonic settings. Direct evidence of fluid involvement the Giuenberg-Richtcr relationship (wave energy increases
with shallow crustal earthquakes comes from observations of by a factor o f -32 for every unit increase in magnitude). It
substantial postseismic discharge in the vicinity o f some follows that, although small earthquakes are highly numer­
active faults (Muir-Wood and King, 1993), and recent geo­ ous com pared with large ones, the energy released as a
chemical studies suggest a very large CO., flux through the whole by small events tends to be insignificant compared
San Andreas fault system (Kennedy et al., 1997). In addi­ with that released by die largest events in a region.
tion, theoretical and laboratory studies emphasize the cou­
pling of stress and fluid pressure conditions in fault and Continental seismogenic zone
fracture formation and in the reactivation of existing struc­ Away from areas o f active subduction, where thermal and
tures, and there is accumulating evidence that active struc­ fluid pressure regimes are strongly perturbed (e.g., Hase-
tural pemieability in the crust is controlled by the prevailing gawa e t al., 1994), m icroearthquake activity (M < 3) is

Fit;. 1. "Delta" diagram illustrating how cyclic.il earthquake (EQ) rup­
turing leads to illtcrcoupied changes in tectonic stress, permeability, and
fluid pressure, thereby effecting fluid redistribution and hydrothermal
precipitation.
largely confined to the top 10 to 20 km of deforming conti­
nental crust (Fig. 2) in a r e a s of moderate to high heat flow
(60-100 m W /m 2; Sibson, 1983; Hill et al., 1990). Seismicity
deepens to -25 km in colder cratonic crust, a s in eastern
North America, but also shallows to depths o f only a few
kilometers in areas of intense geothermal activity. Predom­
inant factors affecting seisrnogcnic depth are inferred to be
heat flow and crustal composition: temperatures at the con­
tinental seismic-aseismic transition have been inferred to
vary from ~300° to -450*0 for, respectively, quartz-domi­
nant and feldspar-dom inant rheologies (Sibson, 1984;
Scholz, 1988). Similarly, Ito (1999) infers a temperature of
350* ± 50®C for the base o f the seismogenic zone in south­
westJapan. In the Kakkonda geothermal field o f northeast
Japan, drilling through a contact aureole into a late Qua­
ternary granitic pluton (T > 500*C) has allowed direct cor­
relation o f the local microseismicity cut-out at -2 km depth
to the 350*C isotherm (Muraoka ct al., 1998).
Variations in seismic style
Mainshock-aftershock sequences: Moderate to large tectonic
earthquakes generally occur as mainshock-aftershock
sequences, sometimes with premonitory foreshocks. Recur­
rence intervals between successive mainshocks typically
range from decades to many thousands of years, depending
on the level o f fault activity. These larger earthquakes tend
to nucleate towards the bottom o f the microseismically
defined seismogenic zone (Fig. 2) with the ruptures propa­
gating either unilaterally o r bilaterally as well as upwards,
typically at around the shear wave velocity of ~3 km /s. As a
general rule, only M > 6 ruptures break through to the
Earth’s surface to produce a surface rupture trace. Seismo-
logical and geodetic studies coupled with analyses of sur­
face fault-breaks and precision aftershock locations suggest
that the bulk o f slip during rupture is confined to a thin
tabular zone o f indeterm inate thickness. However, after­
shock activity may be distributed through substantial vol­
umes o f rock, especially around the perim eter of the main-
shock rupture o r near rupture irregularities (e.g.,
Reasenberg and Ellsworth, 1982). Aftershock sequences
extend for time periods generally ranging between months
and years for moderate to large ruptures, the level of activ­
ity decaying inversely with time from the mainshock.
Scholz (1982) argued that the typical 10 to 20 km thick­
ness o f the seismogenic zone imposes a fundamental length
scale on crustal rupturing (Fig. 2). Small ruptures (M < 6)
confined within the seismogenic layer follow somewhat dif­
ferent scaling laws from large earthquakes (M > 6), which
rupture the full depth o f the seismogenic zone and then
extend along strike to total rupture lengths that are many
times the seismogenic thickness (Pacheco e t al., 1992).
Major fault zones such as the San Andreas fault system also
exhibit significant along-strike variations in seismic style
(Hill et al., 1990). Background micioearthquake activity is
extremely low along the two segments o f the San Andreas
Fault, which ruptured historically in great (M - 8) earth­
quakes in 1857 and 1906, b u t along segments o f the San
Andreas and Calaveras faults that arc actively creeping, the
fault zone is defined by clustered microseismicity as a sub-
vertical tabular structure extending through m ost o f the
seismogenic zone.
Swann activity: Earthquake swarms are a variant o f seismic
activity' where a large num ber o f small earthquakes occur
without a distinct principal shock, the activity waxing and
waning through time and sometimes migrating. Swarms
tend to be distributed throughout a substantial rock vol­
ume—a key distinction from standard tectonic earth-

TAIILF. 1. Approximate Correlations between Earthquake Magnitude, Rupture Parameters,

Relative Wave Energy Release with Respect to an M l Earthquake, and Relative Frequency o f
Different Magnitude Ruptures in Relation to the Average Occurrence per Year o f M7 Earthquakes
Magnitude u L A Energy release w.r.t. Ml Relative frequency
M7 -1 m -3 0 km - 1 0 s km 2 - 10 ” N /yr
Mo - 1 0 cm - 3 km - 1 0 km 2 - 10 R -1 0 0 N /yr
M3 - 1 cm -3 0 0 m ~ l 0 ->in2 - 10 * -10,000 N /yr
Ml - 1 mm ~30 m - 1 0 * m2 1 -1.000,000 N /yr
Abbreviations: M * earthquake magnitude, u= average slip, L = rupture dimension, A = L2 = area
Note: assumes constant shear stress drop i » 3 MPa across circular ruptures (estim ated from relationships in
Kanamori and Anderson, 1975, and Hanks and Kanamori. 1979)
 Fic. 2. Strike-parallel longitudinal profile illustrating varying seismic style along die transition from locked to creep­
ing segments o f a major strike-slip fault zone (m odeled in part on the San Andreas fault) and the relationship o f small
and large ruptures (perimeters defined by aftershock concentrations—filled circles) lo die continental seismogenic zone
defined by background microseismicity (open circles). Stars represent hypocenters (points o f rupture initiation) for
laiger ruptures.
quakes. They are frequently associated with areas of recent
volcanic o r geotherm al activity within extensional and
transtensional tectonic regimes. Within the latter, swarms
are often localized within dilaiional stepovers. Hill (1977)
proposed a mechanical model accounting for swarm char­
acteristics that involves m igration o f magmatic a n d /o r
hydrotherm al fluids through a distributed honeycomb
mesh of interlinked m inor shear and extension fractures
(see below).
Correlation tofault structure and rheology
Studies of exhumed faults suggest that major transcrustal
fault zones arc predom inantly tabular features in upper
brittle and lower ductile regions of the crust (Fig. 3), flaring
out somewhat with increasing depth in the middle to lower
crust (Hanmer, 1988). Their internal structure is often het­
erogeneous with slip a n d /o r strain localized on principal
displacement zones that may continue as planar features
over considerable distances o r anastomoze within a mesh­
like structure (Wallace and Morris, 1986). In the brittle
upper crust, principal displacem ent zones are localized
within or, commonly, at one or other o f the margins to a
fault zone o f varying thickness (o -l km for large displace­
m ent strike-slip faults) comprising variably fractured and
cataclastically deform ed material. At deeper ductile levels,
high-strain shear zones anastomoze around lozenges of
comparatively low-su'ain material. A generalized depth pro­
gression o f dom inant fault rock types—g o u g e /b re ed a
(often day-rich) —» cataclasite —» mylonite —> mylonitic
gneiss (Sibson, 1983; Passchier and Troutv, 1996; Snoke et
al., 1998)—provides the basis for a simple rheological
model that accounts for the depth o f the seismic-aseismic
transition in quartzo-feldspathic continental crust.
T he predom inance o f pressure-dependent cataclastic
processes in the u p p e r crust defines a frictional (FR)
regim e where shearing resistance increases with depth
unless increasing overburden pressure is counteracted by
fluid overpressure (Fig. 3). A fundam ental change in
shearing processes occurs within continental fault zones
at the onset of greenschist fades metamorphic conditions
(T > 300°C), m arked by the appearance o f mylonite series
fault rocks with penetrative L-S fabrics (foliation plus
stretch Uneation) developed by temperature-sensitive crys­
tal plastic flow. Flow by dislocation creep sets in for quartz
a t tem peratures h ig h er than 300° to 350°C, a n d for
feldspar at tem peratures higher than 450°C (Carter and
Tsenn, 1987). Greenschist mylonitization, however, gener­
ally involves a m ixture o f deform ation mechanisms with
plastic flow of quartz accompanied by feldspar cataclasis,
perhaps aided by diffustonal mechanisms as grain size is
reduced by dynamic recrystallization a n d /o r cataclasis
(White e t al., 1980). The quasi-plastic (QP) flow regime
(Sibson, 1983) embraces all these flow mechanisms. How­
ever, flow shear resistance throughout the quasi-plastic
regim e is likely bounded by a tem perature-dependent
power law appropriate to dislocation creep, decreasing
with increasing tem perature and d e p th , so th a t peak
sh ear resistance occurs in the neighborhood o f the
FR /Q P transition. U nder higher greenschist to amphibo­
lite facies m etamorphic conditions in the lower crust (T >
500*0), textures o f mylonitic gneisses record m ore con­
tinuous and presum ably aseismic shearing deform ation
within ductile sh ear zones that com m only range from
hundreds o f m eters to kilometers in thickness (Hanmer,
1988) (N.B. T he F R /Q P transition correlated with the
base o f the seismogenic zone can only loosely be equated
with the ill-defined “brittle-ductile transition,” which has
different connotations for field geologists, seismologists,
and experimentalists in rock deformation; Rutter, 1986).
Fault rock textures in the vicinity of the FR/Q P transition
record a m ixture o f continuous and discontinuous shear-
ing (e.g., mylonitic fabrics interlaced with deform ed
pseudotachylytes o r hydrothermal fault-veins). Commonly
inferred temperatures at the base o f the seismogenic zone
 SEISMOGENIC FRAMEWORK FOR ORE DEPOSITION

SHEAR RESISTANCE (MPa)

Fig. 3. General rheological model and schematic strength profile for a transcntscal strike-slip fault zone (SS) in typi­
cally heterogeneous crust (horizontal scale = 2 x vertical sente). Base o f scismogcnic zone is defined by the transition
from unstable frictional (FR) sliding on principal displacement zones (PDZ) to distributed quasi-plastic (QP) shearing
flow. Frictional fault strength has been calculated assuming optimal fault orientation for a friction coefficient p%= 0.75
under hydrostatic fluid pressure O.'l; comparative strengths o f optimal normal (NJ and reverse faults [R] also
shown). Average frictional strength through the scismogcnic zone imposed by heat flow constraint on the San Andreas
fault shown for comparison.

in continental crust (350° < T < 450°C) accord well with the
recognition o f a fundam ental change from frictional to
quasi-plastic deform ation in fault zones occurring some­
where within the greenschist facies environment, depend­
ing on the quartz/feldspar ratio. Crustal heterogeneity con­
tributes to die smearing-out of the FR/QP transition zone,
which involves a complex m ixture o f discontinuous and
continuous shearing processes.
The earthquake shear stress cycle
Faulting occurs to relieve accumulated shear stress (T) on
faults. Seismological and geodetic studies demonstrate that
shear stress adjacent to a major seismically active fault fol­
lows an approximate sawtooth oscilladon (Fig. 4a). Periods
of steady linear accumulation of shear stress (and elastic
strain) over interseismic periods ranging from decades to
30 RICHARD H. SIBSON

O ther important stress parameters influencing fluid flow

(a) such as m ean stress, o, are tied to this basic shear stress
cycle, although the nature of the coupling depends on the
mode of faulting (Fig. 4b, c). The aftershock phase, conse­
q u en t on ab ru p t localized stress change and creation o f
fresh fracture permeability (Fig. 1), seems likely to repre­
log k sent the most intense period of fluid redistribution around
a seismically active structure.
“Characteristic ”earthquakes and structural
controls on rupturing
(b ) »
REVERSE FAULT Paleoscismic studies along fault zones have demonstrated
th at activity along individual fault segments commonly
involves “characteristic" ruptures o f about the same dimen­
sion and slip, rather than a mixture o f different sized earth­
quakes (Schwartz and Coppersmith, 1984). Repeating char­
acteristic earthquakes have also been observed down to the
(0 s microearthquake (M < 3) level in detailed studies of the San
Andreas and other major active faults (Bakun and McEvilly,
1984; Nadeau and McEvilly, 1997). Such behavior implies
the existence o f structural controls that determ ine where
characteristic ruptures initiate and term inate. A range o f
geometrical and rheological irregularities have been pro­
posed as structural controls accounting for characteristic
earthquake behavior, including fault bends (King, 1986), en
echelon segment stepovers (Segall and Pollard, 1980; Sib-
son, 1986), and rheological heterogeneity at the base of the
seismogenic zone (Sibson, 1984). As yet, no consensus exists
on the relative importance o f these different controls.
Along strike-slip faults, where microearthquake epicen­
ters can be (Erectly correlated to geometrical irregularities
seen in map view, there are clear correlations between after­
shock concentrations and major fault irregularities. Partic­
FlC». -1. Fluid-pressure cycling in relation to the earthquake (EQ) stress ularly noticeable is the tendency for intense clustering o f
cycle, a. Sawtooth oscillation o f shear stress, T, along a seism ogenic fault aftershocks within dilational sites, especially dilational
with inferred fluctuations in fault zone permeability, k. b. Coupled varia­ stepovers (jogs) linking en echelon fault segments (Fig. 5).
tion in mean stress, 8 , along a reverse fault, c. Coupled variation in mean By contrast, in the case o f contractional stepovers, after­
stress, a, along a normal fault, d. Fluid-pressure cycling accompanying
fault-valve action in overpressured crust, and suction-pump effect within a shocks often extend well into the surrounding crust adja­
dilational fault jo g (Ph and P, equal hydrostatic anti lithostatic fluid pres­ cent to the stepover (Sibson, 1985a; 1986). Strike-slip rup­
sure levels, respectively). tures often term inate, o r are severely perturbed, in the
neighborhood o f larger dilational stepovers with subse­
quent tim e-dependent slip transfer across the jo g (Peltzer
many thousands o f years are punctuated by rapid stress e t al., 1996). Recent advances in Coulom b interaction
drops during rupturing, with evidence o f nonlinear relax­ analysis (see review by Harris, 1998), involving modeling of
ation during the postseismic aftershock phase, which for changes in shear and normal stress throughout the region
larger earthquakes may last from m onths to years. A key around a complex earthquake rupture, have demonstrated
point, therefore, is that the stress state around a fault post- good correlations between areas of enhanced failure poten­
failure may differ substantially from the prefailure state, tial (raised shear stress and lowered norm al stress) with
especially if the degree o f shear stress relief is high. observed aftershock distributions (e.g., King e t al., 1994).
Whereas the absolute level o f shear stress driving faulting at This technique holds great promise for identifying likely
depth remains contentious, the average shear stress drop mineralizing sites around complex fault traces.
accompanying rupture generally lies in the range 1 < A < In a global study of the distribution of major irregularities
10 MPa (Kanamori and Anderson, 1975). Am plitude o f on strike-slip faults, Wesnousky (1988) demonstrated a gen­
stress cycling diminishes with distance from the causative eral decrease in the frequency o f fault irregularities with
fault: for larger earthquakes that rupture across the seismo­ increasing finite displacement, and a corresponding increase
genic zone, stress cycling will extend through the adjacent in the maximum size of earthquake ruptures. This recogni­
crust to distances comparable with the depth o f die zone tion o f fault-smoothing with increasing displacement has
(say, ±15 km). important implications for mineralization: the high degree

FlO. 5. Contrasting aftershock distributions around dilational and con-
cractional jogs on dextral strike-slip fault ruptures in California (stars indi­
cate mainshock epicenters, arrowheads indicate direction o f rupture prop­
agation, stippling intensity reflects aftershock concentrations), a. 1979
M5.9 Coyote Lake earthquake (after Reascnbcrg and Ellsworth, 1982).
b. 1968 M6.4 Borrego M ountain earthquake (after H am ilton, 1972).
c. 1992 M7.3 Landers earthquake (after Hauksson et a!., 1993).
o f complexity along low-displacement structures renders
them much more favorable as mineralizing sites.
Tectonic Stress and Fluid Pressure Regimes
State o f stress tn fluid-saturated crust
The stress state at a point in a continuum under load may
be represented by an axial cross comprising three orthogo­
nal principal compressive stresses:
<7, > ff2 > (!)
(Fig. 6a). No shear stress occurs on the principal planes
containing any two o f the principal stresses; planes a t all
(b)
(c)
(0 )
I
F i g . 6 . Basic properties o f stress, a. Axial cross o f the principal com*
pressive stresses (stress ellipsoid) defining the state o f stress at a point,
b. Components o f shear ( f) and normal stress (<7n) on a plane containing
a , and lying at an angle $ to <T,. c . Plot illustrating the satiation o f rand cr„
as a function o f 0 for the case when = 5 and Oj = 1 stress unit, rcspcc*
lively, d. \foriation o f local stress field within a budding layer with respect
to far-field applied stress (n.s. = neutral surface).
o th er orientations have resolved com ponents o f shear
stress, t, as well as normal stress (ojj < <7n < <7|). The differ­
ential stress (<T| —a3) is a measure o f the “devtatoric" char­
acter o f the stress field. Maximum shear stress (rni;tx = 0.5
(<Ti —0*3 ) ) always lies along planes containing the <X2 axis
oriented a t ±45° to the <JXdirection. Mean stress (which.
 w^mim

because it determines volumetric strain, can be thought of Compressional

as the “grasp on the sponge”) is Thrust Fault

_ = (gi_+ y gO (2)

In fluid-saturated crust with internal pore-fluid pressure,

Pf, however, the mechanical response of a rock mass is gov­
erned by the effective principal compressive stresses:

<y{ = (dj - Pf) > = (tr2 - Pf) > cr3' = {<%- Pf) (3)

(H ubbertand Rubey, 1959).

For the family o f planes containing the 0>2 axis» shear
stress is given by

t - & + 2 !L sinae (4)

and the effective nonnal stress by

{0\ + 0*) (0\ + 0$) cos20, (5)

<7n'a <t n - P f)
2 2

where 0 is the angle between <Ti and the plane (Fig. 6b, c).
Effective mean stress is then

{01+02+03)
5' (ff-P f) (6)
3

"A n d e rs o n ia n " stress ( a n d f a u lt ) regimes

Anderson (1951) recognized the existence of three basic Fig. 7. The three Andersonian stress regimes, showing the expected
stress regimes within the crust. Given that the Earth’s sur­ orientations o f faults and associated extension fractures at initiation.
face is approxim ately horizontal in most cases, with the
rock-air interface incapable o f supporting shear stress, it provinces, corresponding to one o r o th e r o f the three
follows that stress uajectories near the surface are generally A ndersonian regim es, extending over large portions o f
either vertical o r horizontal with the vertical stress, 0 . * <Tj, the Earth’s crust (Zoback, 1992).
<72, o r o j (Fig. 7). Recognizing also that brittle faults gener­
ally form through Coulomb failure along planes containing Correlation of stress regimes to tectonic setting
the CTj, axis at 6 25° to 30° to <7, (see below), Anderson Andersonian extensional regimes incorporating normal
(1951) defined three fundamental modes o f faulting in the fault systems are dominant along divergent plate boundaries
different stress regimes: steeply dipping normal faults form­ (Yeats et al., 1997). Sets of parallel, predominandy inward­
ing in extensional stress regimes (Oy = Cf(); gendy dipping facing norm al faults tend to be symmetrically disposed
thrust faults forming in compressiona! stress regimes (cfy = about die rift axes along midocean ridges and some conti­
a3); and subvertical strike-slip faults form ing in wrench nental rifts (Cowie, 1998), where they are frequently associ­
regimes (cfy *=<r2) (Fig. 7). Note, however, that Anderson’s ated with magmatic intrusion o f parallel-striking dikes. In
classification is based on fault orientation with respect to other regions of distributed continental extension, such as
the stress field at the time o f fault inception. the Basin and Range province o f western North America,
At d e p th , stress trajectories may curve as a conse­ sets o f normal faults with consistent feeing direction bound
quence o f horizontal shear tractions developed, for sets o f tilting fault blocks undergoing domino rotation, sep­
exam ple, a t th e base o f th e b rittle u p p e r crust o r the arated by half-graben sedimentary basins (Fig. 8; Roberts
lithosphere (Westaway, 1998). C onsiderable stress h e t­ and Yielding, 1994). In such regions, the fault sets remain
erogeneity may also result from m aterial o r structural seismicalty active down to dips of 30° to 35°, at which they
heterogeneity within the Earth (e.g., Segall and Pollard, apparently lock up (Jackson and White, 1989). Formation of
1980). Indeed, situations may arise where the local stress a new generation o f steep normal faults and their subse­
field is very different from the far-field stress (Fig. 6d). quent dominoing may then allow the early inactive fault set
Nonetheless, regional studies employing a combination to rotate down to still lower dips (Proffett, 1977). Arc-paral­
o f focal m echanism and b orehole stress analyses have lel normal faults may, however, also develop in island arc set­
dem onstrated the existence o f rath er uniform stress tings, often in association with magmatic activity in back-arc
 (a) EXTENSIONS. DOMINO LOCK-UP
(b) COMPRESSIONAL DOMINO LOCK-UP
Fig. 8 . Scu o f initially Andersonian normal and reverse faults undergo­
ing domino rotation to frictional lock up at 55° to 60® to cr, during progres­
sive regional extension and contraction, respectively. Light dashed Hues
indicate likely orientations o f secondary fault sets forming after lock up.
basins and, in the vicinity o f the outer rise in subduction sys­
tems, through bending o f the subducting slab. Normal fault­
ing also develops orthogonally to thrusting in active colli­
sion orogens such as the Himalaya to allow lateral distension
of the uplifted welt (Armijo et al., 1986).
Andersonian congressional regimes and active reverse
faulting are principally associated with convergent plate
boundaries (Yeats et a t, 1997). In island and mountain arc
systems, imcrplatc motion is primarily accommodated on
the m egathrust at the subduction interface; subsidiary
thrusting may develop in the fore-arc basin on the subduc­
tion hanging wall, and on opposite-vergent thrusts in the
back-arc. Thrust faulting likewise dominates in the foreland
and hinterland flanks o f collisional orogens. Local areas of
thrusting are also associated with contractional irregulari­
ties on continental transform faults. However, isolated
reverse fault ruptures also occur in intraplate regions such
as the cratonic crust o f eastern N orth America and Aus­
tralia. Steeper reverse faulting tends to characterize for­
merly rifted crust undergoing shortening (positive inver­
sion tectonics), but may also develop in collision-accretion
complexes where, as a consequence o f progressive hori­
zontal contraction, sets o f early shallow-dipping thrusts
undergo dom ino steepening to frictional lock up at dips
approaching 60° (Fig. 8).
Comprehensive reviews o f strike-slip fault geometry, prin­
cipally in continental and island arc environments, are pro­
vided by Yeats e t al. (1997) and Woodcock and Schubert
(1994). Continental strike-slip fault systems occur in a
range of patterns and tectonic settings: (1 ) as braided sys­
tems o f subparallel strike-slip faults defining continental
transforms; (2) as trench-parallel strike-slip faults con­
tributing to strain partitioning along island and mountain
arcs; (3) as indent-related strike-slip faults in collisional oro­
gens; (4) as local transfer structures in predominandy dip-
slip fault systems; and (5) as distributed strike-slip faults,
developed in either parallel o r conjugate sets, accommo­
dating crustal deformauon over broad areas, often with sig­
nificant rotations o f crustal blocks about vertical axes (e.g.,
Luyendyk, 1991). In many cases it seems that the orienta­
tion o f large-displacement strike-slip faults no longer bears
any simple Andersonian relationship to regional Oj stress
trajectories (e.g., M ount and Suppe, 1992). Major irregu­
larities induce subsidiary deformation. Thrusting develops
in the vicinity of restraining bends and within contractional
jogs, and normal faulting occurs near releasing bends and
within dilational jogs. Such irregularities serve as local
amplifiers o f vertical m otion within the strike-slip system
but many appear to be comparatively short-lived (Brown et
al., 1991).
Fluid pressure regimes in the crust
The pressure, Pf, exerted by fluids saturating pore and
fracture space within the rock mass is o f huge importance
to mineralization. Through the principle o f effective stress
(H ubbert and Rubey, 1959), fluid pressure may reduce
rock strength a n d induce brittle failure, and also plays a
critical role in the creation o f void space. At depth, z, within
the Earth, the level o f fluid pressure is conveniently defined
with reference to the vertical stress (overburden pressure)
by means o f the pore-fluid factor,
A y) (Pf)
P rg * • (7)
where pR is the average rock density and g is gravitational
acceleration. Effective overburden pressure may then be
written,
<V#= (Ov- pf) = p Rgz • <1 - Ay). (8)
Where pore and fracture space is saturated with aqueous
fluid and freely interconnected up to the water table (usu­
ally assumed coincident with the Earth’s surface), the fluid
pressure state is said to be hydrostatic with Ay = pu.„vr/ p R *
0.4 along a linear hydrostatic gradient of approximately 10
MPa/km (Fig. 9). In sedimentary basins with impermeable
horizons (e.g., shale o r evaporite formations), however, it is
not uncom m on for fluid pressures to exceed hydrostatic
rallies (Ay > 0.4) a t depths o f m ore than a few kilometers
and to approach lithostatic values (Ay —» 1 .0 ) at greater
 FLUID PRESSURE (MPa)
100 200 300 400

Fits. 9. Hypothetical fluid pressure profiles and associated frictional shear resistance (normalized to hydrostatic Tfr)
through the FR regime o f an active fault zone. Thick dashed lines (a, b) represent two o f a range o f smoodt progressions
from hydrostatic to neai -lithostatic fluid pressures towards the base o f d ie seisinogenic zone; line c represents a profile
through a compartmentalized fault zone with hydrostatic gradients between each scaling horizon (stippled). Thin
dashed lines define overpressures at constant pore-fluid factor (Ay).

depths. Underpressured zones (Ay < 0.4) also exist in some
sedimentary basins and in vapor-dominated geothermal sys­
tems (Allis, 2000).
More detailed recent studies o f fluid pressure distribu­
tions in sedimentary basins have revealed evidence o f over-
pressured compartments bounded by sealing surfaces that
may include stratigraphic horizons, tabular zones o f
hydrothermal cementation in some cases following isother­
mal surfaces, and im perm eable, generally clay-rich fault
seals that may be steep o r shallow (H unt, 1990; Powley,
1990). Though overpressured to suprahydrostatic values,
vertical pressure gradients inside such sealed com part­
ments generally rem ain hydrostatic. Hydrocarbon prove­
nance studies coupled with the presence o f localized
hydrothermal plumes in the high levels o f some overpres­
sured basins (Roberts et a!., 1996; McKenna and Sharp,
1997), and abnorm al diagenede assemblages localized in
die vicinity of faults (Burley et al., 1989; Bruhn et al., 2000)
provide evidence that bounding seals are breached from
time to time by faulting and fracturing (fault-valve action).
Fluid overpressuring may arise through the following:
( 1 ) compaction, especially in areas o f rapid sedimentadon
perhaps aided by compressionat deformadon as in fore-arc
prisms (Davis c ta l., 1983); (2) diageneuc and metamor-
phic dehyclradon processes aided by organic maturation;
(3) igneous intrusion into fluid saturated crust (Fournier,
1999); and, perhaps (4) direct linkage to zones o f mantle
degassing, as recendy postulated for the San Andreas fault
system (Kennedy e t al., 1997). Fluid pressure levels are usu­
ally reckoned to be near-lithostatic (Ay ~ 1.0 with Pf» crv) in
regions undergoing prograde m etam orphism a t green-
schist and higher grades, where m ineral deform ation
mechanisms contribute to rapid sealing, infilling, and col­
lapse o f pore space (Edieridge et al., 1984). As a general
rule, fluid overpressures are easier to sustain in congres­
sional as opposed to extensional regimes because o f the
tendency for fault and fracture systems to be shallow dip­
ping in the former, but steeply inclined in the latter (Sibson
and Scott, 1998).
Various dynamic forms o f overpressuring have also been
postulated as mechanisms accounting for the apparent
weakness o f major seismically aedve plate boundary faults
such as the San Andreas fault. Rice (1992) suggested a
model where overpressuring is sustained by a continuous
upwards fluid flux through the fault zone, whereas Byerlee
(1993) proposed that fault zones are subdivided into over­
pressured com partm ents separated by seals, in each o f
which the pressure gradient remains hydrostatic, with local
fluid redistribution occurring at the time of earthquakes.
Sleep and Blanpicd (1994) advanced a model in which
fluid overpressure is regenerated in each seismic cycle
through compaction o f fault gouge. Figure 9 illustrates a
range o f possible fluid pressure profiles through the seis-
m ogenic zone, and their effect in reducing frictional
strength below that expected for a fault that is uniformly
hydrostatically pressured (see eq. 9). Profile (a) in Figure 9
represents an initial low degree of overprcssuring with a
sharp transition to near-lithostatic pressures under green-
schist conditions near the base of the scismogenic zone;
profile (b) a more progressive transition to near-lithostatic
pressures; and profile (c) a compartmentalized fault zone
with the pressure gradient reverting to hydrostadc within
each compartment.
Faults, Fractures, and Fluid Pressure Levels
Three modes o f macroscopic brittle failure—shear frac­
tures (faults), pure extension fractures, and hybrid exten-
sional-shcar fractures—may develop in intact, isotropic
rock with predictable orientations to the principal stress
axes (Fig. 10). In a fluid-saturated rock mass, the type of
failure that occurs is greatly affected by fluid pressure
through the law of effective stress (eq. 3). Relevant failure
criteria are listed in Table 2, both in their standard forms in
t / o f space and, to emphasize their dependence, as func­
tions of fluid pressure, P,-. For a particular rock type, brittle
failure may be induced by increasing either differential
stress or fluid pressure with the o th er held fixed, o r
through some combination o f these changes.
Criteria for theformation offaults and fractures
Stress/fluid pressure conditions under which the differ­
ent failure modes develop are illustrated by a composite
Griffith-Coulomb failure envelope for intact, hom oge­
neous, and isotropic rock plotted on a M ohr diagram of
shear stress, r, versus effective normal stress, o f = (<7n - P,),
normalized to rock tensile strength, T (Fig. 10). The condi­
tion for frictional reactivation (reshear) of a cohesionless
existing fault (equivalent to that for a cohesionless aggre­
gate) is also illustrated. For simplicity, the slopes of the lin­
ear portions of the intact failure envelope in die compres-
sional field and the reshear criterion are both constructed
with slopes of 0.75, corresponding to the average of the
experim entally determ ined range for the coefficient of
internal friction (0.5 <p%< 1.0;Jaeger and Cook, 1979) and
near the middle o f Byerlec's (1978) range for static friction
(0.6 </is < 0.85).
Within a particular rock unit, the mode o f brittle failure
depends on the balance between the differential stress (CTj
- c%) and rock tensile strength, T (Secor, 1965). When (cTj
- Oj) < 4T, extension fractures form in accordance with the
hydraulic fracture criterion along planes perpendicular to
Oil (Fig. 10). When (tjj —<T3) > 5.66T, shear fractures (faults)

Fic. 10. Composite failure envelope for intact rock (bold line) plus the reshear condition for a cohesionlcss fault
(dashed) plotted o n a Mohr diagram o f shear stress, r, against effective normal stress, o f . normalized to rock tensile
strength, T. Slopes o f the linear portions o f the failure envelopes for intact rock and reshear arc both taken as 0.75. cor­
responding respectively 1 0 the coefficients o f internal friction and static friction ( /t j . Critical stress circles arc shown
for the three m acroscopic m odes o f brittle failure, and for the reshear o f an optim ally oriented cohesionlcss fault
(dashed). Expected orientations with respect to the principal stress axes o f new-formed shear (sh.). cxtcnsional-shcar
(e-s) and extension fractures (ext.) arc shown in the attached cartoon.
 Note: T * tensile strength, T. C 2T * cohesive strength, and cocfficcni o f internal friction, f t = 0.75;
reshear condition for a cohesionlcss fault calculated for static friction, p , = 0.75
form in accordance with the Coulomb criterion along
planes containing the cr2 axis at initiation angles 9{ = 27° ±
5* to (allowing for the 0.5 < fij < 1.0 experim ental
range). When 4T < (Cj - of) < 5.66T, extensional-shear
fractures form along planes also containing the a2 axis but
at lower angles (0< Q{) to Ov Note that only Coulomb shear
discontinuities may develop in a cohesionless aggregate, at
similar orientations to faults forming in intact rock.
Thus, for the standard Andersonian stress states, normal
faults should form with initial dips of 58° to 68° in exten-
sional tectonic regimes (Oy = <7,), striking parallel to verti­
cal extension fractures; thrust faults should form in com-
pressional tectonic regimes (crv = o f at initial clips of 22° to
32°, possibly in association with subhorizontal extension
fractures; and strike-slip faults should form with subvertical
dips in wrench regimes (<JV « <t2), striking at 22° to 32° to
sets o f subvertical extension fractures (Fig. 7).
Frictional reactivation of existingfaults
Reshear o f existing cohesionless faults depends on their
orientation as well as differential stress and fluid pressure lev­
els. The general criterion for frictional reshear o f an existing
cohesionlcss fault with static coefficient o f friction, /is, is
T= J M ,'= /is(o; , - p r>.
which, for faults containing the o., axis and oriented at a
reactivation angle 9r to o v may be rewritten in terms of the
ratio of effective principal stresses as
G\ _<X i - P f ) _0 +/i»ccfl r)
( 10)
(03- Pf) - (1 -/is tafl r)
(Sibson, 1985b). Figure 11 illustrates how this stress ratio
varies with 9r for p s = 0.75 (cf. Fig. 10), reflecting the rela­
tive ease of reshear at different reactivation angles. The
optimal angle for frictional reactivation (at which the stress
ratio for reactivation reaches a positive m inim um ) then
occurs when 9 * = 0.5tan- , ( l / ^ s) ~ 27°. U nder this optimal
reshear condition, equation ( 10) reduces to
(Cj - 0-3) = 3Oy' or (0! - 03) = o.750v' (11)
for compressional and extensional regimes, respectively.
These optimal conditions for reshear have special signifi­
cance in that the tectonic stress state in many settings
appears to be governed by the critical stress required for
reactivation o f favorably oriented faults (Townend and
Zoback, 2000).
As the orientation o f an existing fault becomes progres­
sively less favorably oriented for reactivation (either greater
<fl) o r less than optimal) the ratio o f effective stresses required
for reactivation increases, reaching infinity when Br = 20r*
(-54° for = 0.75), the angle of frictional lock up (Sibson,

Fit;. 11. Ratio o f greatest to least effective principal stresses (tTj'/Oj')
required for reshcar o f a coliestonless fault plotted against the reactivation
angle (0r). 9 * is the optimum angle for reactivation at which the stress
ratio reaches a positive minimum. Frictional lock up occurs at 29*.
1985b). This accounts nicely for the commonly observed
lock up o f normal and reverse faults within dominoing sets
at 50° to 60° to vertical and horizontal Oj directions, respec­
tively (Fig. 8). Reshear o f severely misoriented faults with $t
> 29 * is possible only under special circumstances when
the tensile overpressure condition, <7S' < 0, is met.
Extremely im portant from the viewpoint o f mineraliza­
tion is that a throughgoing cohesionless fault, favorably ori­
ented for frictional reactivation within die prevailing stress
field, inhibits any o th er form o f brittle failure within the
adjoining intact rock mass (Sibson, 2000a). T he tensile
overpressure condidon ( o f < 0) needed for extensional-
shear o r pure extension fracturing cannot be attained in
the presence o f an existing fault oriented at 0r < 20r*,
because reshear will always occur first (fig. 10). Cohesion-
less faults other than those containing the a , direction may
likewise inhibit brittle failure o f intact rock, but the critical
fault orientations in three dimensions then depend also on
the value o f the intermediate principal stress.
Brittlefailure mode plot
Stress conditions for the three modes o f brittle failure
and for fault reshear (as in Fig. 10) may be transcribed by
m eans o f the in d e p e n d e n t forms o f the failure criteria
listed in Table 2 to plots of differentia! stress (<Tj - <%) ver­
sus effective vertical stress, Oy' = (<fy- Pf), for different
Andersonian stress regimes (Sibson, 1998). Failure curves
for intact rock are constructed as a function o f intact tensile
strength, T, which, for m ost sedimentary rocks lies in the
range 1 to 10 MPa, but may reach 20 MPa o r more for crys­
talline rocks (Lockner, 1995). Values o f <7V' may then be
converted to equivalent depths for different values o f the
pore-fluid factor, Ay.
In Figure 12, brittle failure conditions for optimal res­
hear and for intact rock have been constructed for com­
pressions! (<7V= ojj) and extensional (C7V = tectonic
regimes with T = 5,10,15, and 20 MPa. The two sets of fail­
ure conditions also represent end-member cases for strike-
slip faulting where <r3 £ <7V = a 2 ^ <7,. Each failure curve
defines the transition with increasing cfv' from purely
extensional fracturing through extensional-shear to com-
pressional shear failure, the equivalent depths for the tran­
sitions increasing with rock tensile strength. For example,
in an extensional regime under hydrostatic fluid pressures
(Ay » 0.4), pure extension fracturing for T * 10 MPa
extends to depths of -2 km and the transition from exten­
sional-shear to compressional-shear failure occurs at depths
o f - 3 km. These brittle failure m ode plots have a broad
range of implications for the development of hydrothermal
mineralization around faults:
1. It is much easier, in terms of differential stress a n d /o r
fluid pressure levels, to develop brittle faults and fractures
within an intact rock mass in an extensional tectonic
regime than within a compressional regime.
2. For positive Oy' (Ay; < 1.0), only compressional shear
failure can occur in contractional tectonic regimes (except
in cases of extrem e stress heterogeneity), whereas exten-
sional and extensional-shear fractures may form in exten-
sional regim es to depths dep en d en t on rock tensile
strength.
3. Processes such as hydrotherm al silicification, by
increasing the tensile strength o f the rock mass, may
increase the depth range over which extension fracturing
may occur.
4. At high values o f o f , brittle structures can only form
by com pressional shear failure, but when o f is low, the
mode of failure becomes extremely sensitive to variations in
tensile strength reflecting changes in rock competence.
Mixed m ode failure then becomes likely within a heteroge­
neous rock mass.
5. Again, it is apparent that the presence of an through-
going cohesionless fault that is favorably oriented for fric­
tional reactivation within the prevailing stress field prevents
attainm ent of the tensile overpressure condition (Pf > <r3)
needed for hydraulic extension fracturing and extensional-
shear failure.
Maximum sustainable overpressure
In low-permeability rocks the degree o f overpressuring is
limited by the formation (or reactivation) of fractures and
faults that may act as flow conduits. Critically im portant is
the lowest o f (corresponding to the highest Pf and Ay val­
ues) that can be sustained in a particular tectonic environ­
m ent. Consider two end-m em ber situations. Hydraulic
 ,=1.2
6
4
2 \
0 i—
4

8
12

16

20
24

28

32

36

40
0.8

to „ . . r .. . . ,c ,^ ,h differential stress (cr, - crn) plotted against effective vertical stress, o v'. for
I-io. 1_. Buttle failure inode p > (cctonjc regimes. Effective vertical stress can be equated to depth
extension^ (a v - cr,) and comiJreMi ^ Dashed lines represent the optimal reshcar conditions forcohe-
(ng ,t) fo r ■.H e ro n * a lu c,o f l i r e i n t a c t r o c k v m h thedifferentfailu refieldsoudin ed (shadedpor-
!!o" 'defin e ti ” ex.e"is!o,I’a l l e a r (<*) iransiuon regions between purely extensional fntm tring and compressional
shear failure). Insels shore expected orientations o f the resulung bntde structures (abbrevtauons as in Ftg. 10).

extension fracturing with o f = - T occurs at the lowest pos­
sible Gf value for a given tensile strength (Fig-10) but can
only occur under low differential stress with (<7, - <%) < 4T.
Alternatively, if fluid overpressure is limited by the presence
of throughgoing cohesionless faults that are optimally ori­
ented for reactivation, then maximum sustainable over­
pressure is governed by equation ( 1 1 ), and diminishes with
increasing differential stress. Relevant here are the obser­
vations of Barton e t al. (1995), who found that the most
perm eable fracture systems in a crystalline assemblage
u n d er triaxial stress were those that were optimally ori­
ented for reshear.
In Figure 13, maximum sustainable fluid pressure. Ay, is
plotted against depth for a normal-fault regime (Gy= cr,)
and for a thrust-fault regime (Gy = <73), again representing
end-m em ber cases for strike-slip faulting. Following the
procedure of Secor (1965), Ay values for hydraulic exten­
sion fracturing at maximum permissible levels o f differen­
tial stress, (cr, - <73) = 4T, are plotted against depth for dif­
ferent values of T in the two regimes. Also shown (from eq.
1 1 ) are the Ay values required at different levels of differ­
ential stress for reshear of optimally oriented cohesionless
faults in each regime. Interm ediate situations between
these end-members arise when the faults retain some
degree of cohesive strength. Clearly, the presence of
throughgoing cohesionless faults drastically lowers the
maximum fluid overpressure that can be sustained at a par­
ticular depth and the higher the differential stress, the
lower die sustainable overpressure.
Dynamic Structural Permeability
Components of stress-controlled structural permeability
Stress-controlled structures affecting rock permeability
include brittle mtcrocracks, extensional and extensional-
shear fractures, faults, and stylolitic solution scams. Their
characteristic orientations at initiation with respect to a tri­
axial stress state are illustrated in Figure 14, along with some
o f the interlinkages that are frequently observed such as
dilational fault jogs, en echelon tension gash shear zones,
and fault-fracture meshes o f the kind proposed by Hill
(1977). Permeability o f individual components is, however,
not simply predictable and is also generally time dependent
(Cox et al., 2001; Sibson, 1994). This is especially the case
for faults, where fault permeability relative to the country
rock depends critically on the nature of that rock, and may
also vary with displacem ent activity, gouge development,
and hydrothermal cementation.
M inor faults developing in initially high porosity sedi­
mentary or volcanic rock may, through grain comminution,
porosity collapse, and alteration to clay-rich assemblages,
form deform ation bands that are relatively impermeable
with respect to the wall-rock (Antonellini and Aydin, 1994).
Thus, some inactive faults in sedimentary basins form seal-
 EXTENSIONAL NORMAL FAULT REGIME COMPRESSIONAL THRUST FAULT REGIME
(<*v = Oi)
0 0.5 1.0

Fig. 13. Maximum sustainable fluid pressures. Ay. plotted against depth in compresstonai and cxtensional tectonic
regimes. Bold lines for T = 5 and 10 MPa assume that fluid overpressure is limited by the formation o f hydraulic exten­
sion fractures under maximum allowable differential stress, (a ( -cr3) = 4T. Thin lines arc the maximum sustainable over­
pressures in the presence o f existing, optimally oriented, cohcsioulcss faults under various levels o f differential stress.
Double-headed arrows link sustainable fluid pressure curves for the different failure modes at the same differential stress
level.

ing barriers to fluid m igration but others clearly act as
cross-stratal conduits for fluid flow (Burley et al.( 1989). In
contrast, faults developed in strong, initially low-porosity
rocks tend to enhance local permeability through the mis­
match o f opposing walls, by cataclastic brecciation, and by
the developm ent o f subsidiary fractures (Brown and
Bruhn, 1996). Studies of exhumed fault zones in crystalline
rock suggest that a low-permeability core of extremely fine­
grained ultracataclasite is often flanked by a comparatively
high-permeability damage zone o f distributed fracturing
(Evans etal., 1997).
Directional permeability
■ Both macroscopic and microscopic extensional fractur­
ing enhance permeability in the cfi/cr2 plane provided the
cracks rem ain unccm ented (Fig. 14); the effect becomes
more pronounced as Pf $7 3, with large apertures when Pf >
Pg. U nder these conditions, extensional shears may likewise
form gaping conduits for high-flux flow. O n the other
.hand, flow perpendicular to the <j 2/ cfh plane may be
restricted by stylolidc seams o f insoluble clay material in
planes perpendicular to o l (Groshong, 1988). Note, how­
ever, that the common intersections o f all stress-controlled
components o f structural permeability lie parallel to the <f2
axis, as do tubular components of permeability such as dila-
tional jogs and fault-fracture intersections. Strong direc­
tional perm eability in the <r2 direction is, therefore,
expected provided there is adequate three-dim ensional
continuity and interconnectedness between the various
components of structural permeability. Moreover, regard­
less of w hether faults are low-permeability structures in
highly porous and permeable host-rocks, or comparatively
high-permeability structures disrupting low-porosity rocks,
developm ent o f conjugate fault sets inevitably leads to
strong <72 directional permeability.
Fault permeability in relation to the earthquake stress cycle
Evidence from geotherm al fields and experim ental
deformation under hydrothermal conditions suggests that
fault permeability in active hydrothermal environments can
be destroyed over time periods that are short com pared
with earthquake recurrence intervals (Fournier, 1991,
1996; Moore etal., 1994; Hickman et al.f 1995). Permeabil­
ity within seismically active fault zones is, thus, inherently

gaa&esss
 Self-generation of structural permeability by migratingfluids

Because the failure criteria for different modes o f britde

failure (Table 2) all depend strongly on fluid pressure, infil­
tration o f pressurized fluids into stressed low-permeability
crust may selfgenerate structural permeability throughout
the rock mass, the components including reactivated exist­
ing surfaces (bedding, cleavage, fractures, dike contacts,
etc.) and newly form ed extension fractures, extensional-
shcai'S, and m inor faults. As each individual structure forms
or reactivates, the stress field is locally perturbed, adding to
stress heterogeneity within the evolving fault-fracture sys­
tem. A range o f fault-fracture meshes comprising various
combinadons o f structural permeability components (Fig.
14) may then develop (Sibson, 1996). At high o f , only shear
fractures can form o r reactivate, but wherever the tensile
overpressure condition, o f < 0, is achieved, developing
fault-fracture meshes may also incorporate gaping exten-
sional and extensional-shear fractures as in Figure 15. In
general, fault-fracture meshes tend to be highly irregular as
a consequence o f heterogeneity o f m aterial properties,
stress, and fluid pressure within die rock mass, but system­
atic mesh structures may develop where the stress field is
symmetrically disposed with respect to competence layering.
Hill (1977) proposed that mesh structures of interlinked
shear and extensional fractures driven by infiltrating
m agma o r hydrotherm al fluid could account for earth­
quake swarm aedvity distributed throughout substandal
rock volumes. Field studies suggest th at Hill-type mesh
structures comprising extensional vein systems interlinked
through m inor shears (Fig. 16) are common over a broad
range o f scales, often serving as im portant conduits for
hydrothermal a n d /o r hydrocarbon flow (Sibson, 1996; Sib­
son and Scott, 1998).

Fic;. 14. Stress-controlled com ponents of Structural permeability in rela­

tion to a trinxial stress field (shear-sense indicators define faults [Coulom b
shears]; extension fractures and extcnsional-shears: cross-hatched; stylo-
lites: "squiggly” lines). As drawn, the diagram illustrates possible structural
permeability in vertical sections through an extensional regim e ( ctv = 0 ]);
or, w hen ro tated 90°, for a com pressional regim e (<?v = cr:i); or, w hen
treated as a m ap view, for a strike-slip regim e (0y ■ &'])■ Note the com m on
intersection o f all stress-controlled com ponents in the a> direction.
dynamic, with faulting, fracturing, micro-cracking, and
brecciation com peting with processes of permeability
destruction that include hydrothermal alteration accompa­
nying gouge formation, microcrack healing, hydrothermal
cem entation of fractures, and solution-precipitation
(Angcvine et al., 1982; Cox et al., 2001). In particular,
hydrotherm al precipitation is generally prom oted by
decreases in fluid pressure; the more abrupt the pressure
drop, the faster such self-sealing will tend to occur (Parry,
1998; Fournier, 1999). Permeability in the vicinity of the
fault zone is, thus, likely to be highest immediately postrup­
ture (Brown and Bruhn, 1996), diminishing through the
aftershock phase to comparatively low values through the
interseismic period (Fig. 4).
High-flux flora th rough fault-fracture meshes
In general, rock permeability increases as P| and o f
decreases (Seront et al., 1998; Cox et al., 2001), increasing
flow rate for a given head gradient. However, in low-perme­
ability rock masses, bulk permeability is dominated by frac­
ture flow. Flow rate is then highly sensitive to fracture aper­
ture, varying as the cube of aperture for a set o f parallel
planar Fractures (Snow, 1968). Although faults may signifi­
cantly enhance bulk permeability, especially postfailure, max­
imum flow rates under a particular head gradient will be
achieved at the highest levels of fluid overpressure, with gap­
ing extensional and extensional-shear fractures distributed
throughout a rock volume in some form of fault-fracture
mesh (Figs. 15 and 16). This requires the tensile overpressure
condition, o f < 0, to be maintained, at least locally, providing
a general condition for high-flux flow in low-permeability
rocks. The higher the tensile strength, the greater the range
of differential stress levels under which extensional and
extensional-shear failure may occur in preference to faulting
(Fig. 12). High-flux flow through fault-fracture meshes is
therefore favored by the presence of high tensile strength
material within the rock mass, provided that the hydrofrac­
ture condition (Pf * <TS + T) is met for the stronger material.

FlO. 15. Cross sections o f Hill-type fault-fracture m eshes developed in
extensional and com pressional stress regimes: a. in the absence o f existing
faults (b o ld er lines indicate eventual am algam ation into thro u g h g o in g
fault); b. in co rp o ra tin g existing faults th at arc severely misoriented for
reactivation in th e prevailing stress field (a fte r Sibson and Scott, 1998).
Attainment of the tensile overpressure condition requires
either absolute tensile values o f cr3 (i.e., cr3 < 0) or the con­
dition Pf > 0$ to obtain locally. Local reductions in <73 may be
brought about by stress field heterogeneity, for example at
fault tips or in dilational jogs between en echelon fault seg­
m ents (Segal! and Pollard, 1980; O hlm acher and Aydin,
1997). Favored localities for mesh development, therefore,
include linkage structures along large-displacement fault
zones such as dilational jogs, lateral ramps, and transfer
faults. However, in general the condition Pf > cr3 must be
maintained for meshes to continue deforming and acting as
high-permeability structural conduits. In extensional-
transtensional regimes, this condition can be achieved
under near-hydrostatic fluid pressure levels at shallow
depths (Figs. 12 and 13), contributing to the formation of
epizonal m ineral deposits in the boiling environm ent of
ascending hydrotherm al plumes (Henley, 1985; Sillitoe,
1993). At other than shallow depths in extensional-transten-
sional settings, significant fluid overpressuring is required
for this condidon for gaping extensional and extensional-
shear fractures to be maintained.
A critical point also apparent in Figures 12 and 13 is that
the tensile overpressure condition allowing high-flux flow
through dilatant fault-fracture meshes can only be achieved
in; the absence of throughgoing low-cohesion faults diat are
well oriented for frictional reactivation in the prevailing
stress field. High-flux flow through fault-fracture meshes, as
illustrated in Figure 15, is therefore, to be expected only in
three circumstances: (1) in effectively intact low-permeability
crust devoid o f throughgoing favorably oriented faults, for
example in crust recently reconstituted through meuunor-
phism (Sibson and Scott, 1998); (2) where existing faults
have become severely m isoriented in the prevailing stress
field, perhaps through progressive dominoing of an existing
fault set (Fig. 8); and (3) where existing faults regain cohe­
sive strength between successive slip increments through
hydrothermal cementation (e.g., Nguyen etaL, 1998).
In fact, field evidence suggests that distributed fault-frac­
ture meshes are in many cases precursors to the develop­
m ent o f throughgoing faults which form by the progressive
amalgamation o f mesh com ponents into a favorably ori­
ented structure, after which the rest o f the mesh system
becomes inactive (Sibson and Scott, 1998). Conditions
favoring high-flux flow at maximum sustainable overpres­
sure are, therefore, likely to be short-lived.
Concentrated flow along individual faults
High-flux flow along an individual fault may arise when it
serves as the principal high-level discharge route during
fault-valve action (see below), following the postfailurc col­
lapse o f a dilatant overpressured fault-fracture mesh at
depth . Such behavior is especially likely on load-strength­
ening reverse o r reverse-oblique slip faults in compres-
sional-transpressional settings, where fault-normal stress is
reduced postfailure. However, the progressive amalgama­
tion o f any kind o f fault-fracture mesh into a dom inant
throughgoing fault (Fig. 15) may also provide transient
conditions for high-flux flow along individual faults. In
such circumstances, local flow focusing is likely at dilational
sites arising from m inor fault curvature (Brown and Bruhn,
1996) and in areas o f stress heterogeneity near fault tips,
especially as they approach dilational jogs or isolated fault
bends (Segalland Pollard, 1980).
Fluid Redistribution Mechanisms Linked
to the Earthquake Stress Cycle
Interlinkages between changes in tectonic stress, fault-
fracture permeability, and fluid pressure (Fig. 1) give rise
to a range o f potential mechanisms coupling fluid redis­
tribution to the earthquake stress cycle. More than one
mechanism may operate at any time and place, and their
relative im portance in different tectonic regimes and at
different crustal levels is far from fully established. Taking
special account o f the conditions for high-flux flow estab­
lished above, this review places emphasis on redistribution
mechanisms capable o f giving rise to the focused large-
volume flow needed for m ineralization. Note, however,
that much of the thinking is based on simple two-dimen­
sional geom etrical and stress analysis of fault environ­
ments. The more complex natural environment inevitably
allows for a greater range of three-dimensional structural
controls with high mineralization potential (e.g., Bruhn et
al., 1990).

Ftc. 16. Cross sections through mineralized hull-fracture meshes, a. Extension*! cpizona! mesh. Martha mine. Coro­
mandel, New Zealand (after Morgan, 1924). b Comprcssional mesozona) mesh. Perron m ine (now Pascalis N ord), Val
d ’Or, Abiribi belt, Quebec (after Ames. 1948).
Cyclical dilatancy pumping
Various forms of shear-stress-dependent grain-scale,
microcrack, and fracture dilatancy directly related to the
earthquake stress cycle (Fig. 4a) have been proposed, both
for the crust adjacent to seismogenic faults, and for mater­
ial within the fault zones themselves (reviewed by Sibson,
1994). Early suggestions of regional microcrack dilatancy
operating at high (>100 MPa) levels of shear stress in the
crust adjacent to active fault zones and capable of large-
scale fluid redistribution (Scholz et al., 1973; Sibson ct al.,
1975) have not been substantiated, although the search for
other forms of stress-dependent dilatancy continues
(Crampin, 1994). Until the stress levels driving faulting and
appropriate constitutive dilatancy laws are more fully estab­
lished, it is not possible to evaluate the contributions of
these various mechanisms to crustal fluid redistribution.
Coupled mean stress (6~)-shear stress ) cycling
In general, tectonic shear stress on faults cannot change
without fault-normal stress also changing, thereby altering
the level of mean stress and the fault’s frictional strength
(eq. 9; Sibson, 1991). For dip-slip faults, changes in mean
stress (d~) are roughly comparable to the shear stress drop
(generally 1 < A < 1 0 MPa, equivalent to changes in
hydraulic head o f 0.1 to 1.0 km ), b u t the coupling is dia­
metrically opposite for reverse and normal faults (Figs. 4b,
c). Neglecting local stress heterogeneity at fault tips, normal
N
Ct
faults are load-weakening with average mean stress decreas­
ing during loading only to increase abruptly postfailure,
whereas reveree faults are load-strengthening, with average
mean stress increasing during loading only to decrease at
failure. Mean stress cycling around strike-slip faults may lie
anywhere between these two end-member cases, approach­
ing the normal fault and reverse fault situations in transten-
sional and U'anspressional settings, respectively.
Because o f strong directional permeability along
strike, m ean stress cycling in dip-slip fault systems may
induce significant strike-parallel redistribution postfailure,
with lateral movement of fluids away from a failed normal
fault segment but inwards towards a failed reverse fault seg­
m ent (Fig. 17). Effects are likely to be most noticeable in
the vicinity of fault tips and overlaps where stress hetero­
geneity allows the condition for high-flux flow to be main­
tained. Such redisu'ibution may be especially important in
the case of norm al faults, which are likely to have a high
concentration o f associated subvertical fractures in the
near-surface (Sibson, 2000b). Closure of such subvertical
cracks from increased horizontal stress postfailure has been
postulated as the cause of the large surface effusions some­
times observed following rupture of normal faults (Muir-
Wood and King, 1993).
Postseismic redistribution around rupture irregularities
Rupturing and slip transfer across fault irregularities leads
to abrupt postfailure changes in m ean stress localized
 NORMAL FAULT SYSTEM REVERSE FAULT SYSTEM

Flo. 17. Map views o f strike-parallel fluid redistribution around failed segm ents o f normal and reverse dip-slip faults
as a consequence o f the change in mean stress accompanying failure ( d - ).

around the irregularities (Segall and Pollard, 1980), with a where reverse faults refract through fold hinges under con­
tendency for fluids to be redistributed from areas o f raised ditions o f strong fluid overpressuring (e.g., Cox, 1995).
to areas of reduced mean stress (Nur and Booker, 1972). In The internal structure o f large dilationaljogs is generally
strike-slip fault systems where epicenU'a! concentrations can a fault-fracture mesh comprising various combinations o f
be directly correlated in map view to structural irregularities subsidiary faults, sheeted arrays o f extension veins, and
on subvertical faults (jogs, isolated bends, and fault inter­ multiply recem ented wall-rock breccias developed by
sections; Figs. 5 and 18), there is a notable tendency for hydraulic implosion (Sibson, 1985a). In mechanical terms,
aftershocks to cluster in areas of inferred mean stress reduc­ it appears that dilationaljogs and bends act as suction
tion at fault tips and, especially, in dilational jogs and bends pumps (Fig. 4d). Rapid slip transfer during rupture propa­
(Sibson, 1986). Slip transfer across dilational jogs reduces gation leads to abrupt localized reductions in fluid pressure
mean stress in the stepover region with compensating lobes below ambient (hydrostatic?) levels, triggering brecciation
of elevated mean stress in the flanking areas, in conuust to by hydraulic implosion accompanied by episodes of boiling
contractional jogs where mean stress is increased in the and mineral precipitation at high levels in geothermal sys­
stepover region but lowered in the flanking regions (Fig. tems (Sibson, 1987). Induced suctions contribute to rup­
18). Structural response at isolated fault bends is affected by ture perturbation or arrest, promoting the time-dependent
the direction of rupture propagation, which determ ines slip transfer across dilationaljogs that has been observed
whether mean stress on the across-bend segment decreases during the aftershock phase as fluid pressures restore to
(promoting dilation) or increases as a rupture propagates background levels (Peltzer et ah, 1996).
into the bend. It is in such dilational sites, through local From the viewpoint o f the explorationist, the three-
stress heterogeneity, that the high-flux flow condition (<7/ < dimensional geometry of dilational structures is paramount.
0) is most likely to be achieved. Simple dilationaljogs usually develop approximately per­
Jog suiictures linking en echelon strike-slip fault segments pendicular to fault slip vectors (Fig. 14), acting as pipe stinc­
have cross-strike dimensions ranging up to several kilome­ tures of locally enhanced permeability within fault zones. In
ters. Detailed aftershock studies show that, in at least some suike-slip systems, larger dilationaljogs also serve to localize
instances, such structural irregularities extend through the magmatic intrusions, including porphyries (Tosdal and
full depth o f the seismogenic zone (e.g., Reasenberg and Richards, 2001). This, coupled with strong localized
Ellsworth, 1982). However, dilational and contractional jogs enhancem ent of vertical permeability, may set up powerful
also develop on dip-slip faults, though generally on a lesser hydrothermal systems its in the Salton Sea geothermal fields
scale. Dilationaljogs (sometimes referred to as cymoid loops; of the southern San Andreas fault (Sibson, 1987). Dilational
M[cKinstry,1948) are often well developed on normal faults jogs in dip-slip systems tend to enhance along-strike perme­
at shallow depth, reflecting the comparative ease of exten- ability; and upwards flow of buoyant hydrothermal plumes
sional fracturing under hydrostatic fluid pressures in the epi- may be significantly deflected by gently rakingjogs, which
zonal environment (Fig. 12), especially where competence outcrop as fault overlaps (Sibson, 2000b). Explorationists
layering promotes mixed-mode brittle failure and fault should note, however, that complexities in rupture geome­
refraction (Sibson, 1998). Dilationaljogs may also develop try coupled with three-dimensional material heterogeneity

DILATIONAL FAULT JOG CONTRACTIONAL FAULT JOG
tfot8i<8 0< intomal structure
(b) Oitational Response at Isolated Fault Bend
rupture nuctealion
Contractions! Response at isolated Fault Bend
rupture nucieation
Complex Rupture Trace
(c)
rupture nudeation
contractional
jog. tfHationai
bend
contractional
bend
Fir.. 18- Rupture interaction with smictural irregularities on faults, a.
Dilations*! and contractional jogs (schematic indicates interna! structure of
dilational jog) with associated areas of mean stress decrease and increase,
b, Dilational and contractional response of isolated fault bends, c. Zones
of contraction and dilatation along a complex rupture trace.
in th e rock m ass may lead to m u ch m ore co m p lex dilational
structures in th ree d im e n sio n s th an th e stand ard two-
d im en sion al d ilational j o g m o d el. B ruhn e t a!. (1 9 9 0 ). for
ex a m p le , have c o n sid e re d th e sp a ce p ro b lem s created by
ru p tu rin g w ith in a n orm al fault system m ad e up o f inter­
linked nonparallel fault segm ents.
F ault-valve action
V alving a c tio n , in v o lv in g th e p o stfa ilu re d isch a rg e o f
overpressured fluids th rou gh fauit-fracture systems, occurs
w herever ruptures breach im p erm eab le barriers b ou n d in g
overp ressu red p o rtio n s o f th e crust. T h ese barriers n e e d
n o t be su b h o riz o n ta l b u t m ay a lso in volve ste e p sea lin g
faults b o u n d in g flu id pressure com partm ents (H u n t, 1990;
Powiey, 1 9 9 0 ). O v erp ressu rin g m ay b e restricted to th e
vicinity o f th e fault zo n es th em selves o r e x ten d in to broad
areas o f th e su rrou n d in g crust (Fig. 19). Settings for differ­
e n t form s o f fault-valve action in clu d e overpressured sed i­
m entary basins (R oberts e t a!., 1996; M cKenna an d Sharp,
1997), areas o f active m agm atic intrusion (Fournier, 1999),
a n d th e brittle carap ace to reg io n s u n d e r g o in g prograde
m etam orphism (Sibson a n d Scott, 1998). Valve action may
involve localized discharge alon g established faults e x ten d ­
in g through th e aftershock period, o r distributed discharge
in volvin g self-g en era tio n o f fault-fracture m e sh e s d u rin g
earthquake swarm activity.
F o llo w in g ru p tu re, flu id d isch arge a lo n g th e tran sien t
p erm ea b ility o f th e rup tu re z o n e o r z o n e s ca u ses lo ca l
reversion towards a hydrostatic fluid pressure gradient. T h e
accom panyin g pressure d rop p rom otes hydrotherm al self-
sea lin g , a llo w in g flu id o verp ressu res to reb u ild at d ep th
(Figs. 4d and 19a). T im in g o f successive failure ep isod es is,
thus, co n tro lled by th e cycling o f tectonic shear stress, fluid
pressure, and frictional fault strength th rou gh ou t th e inter-
seism ic p e r io d . W id esp read d e v e lo p m e n t o f sy n te c to n ic
fault-veins in e x h u m e d fau lt zo n es d e v e lo p e d w ithin both
sed im en ta ry and crystalline h o st rocks, c o u p le d w ith evi­
d en ce for fluid pressure cycling from fluid inclu sion studies
o n th e vein m aterial (e.g ., Robert e t al., 1995; Parry, 1998),
su ggest th at m in o r valving action involving sm all fluid vol­
u m es is w id esp rea d in cru stal fa u lt z o n es. H ow ever, th e
p o ten tia l for ex trem e valving activity clearly relates to th e
m axim um overpressure that is sustainable in d ifferen t tec­
tonic settings (Fig. 13).
C onditions fo r extreme valve action: For significan t m ineral­
iza tio n to d e v e lo p th ro u g h fault-valve a c tio n , e a c h cycle
sh o u ld id ea lly in volve rapid d isch a rg e o f large flu id vol­
um es co u p le d to a large-am plitude d rop in flu id pressure.
dilational B ecause high-flux flow at d ep th in th e crust req uires dila-
Jog tan t fau lt-fracture m e sh e s, a first r e q u ir e m e n t is th e
absen ce o f th rou gh goin g, low -cohesion faults that are w ell
o r ie n te d for reactivation (Fig. 15). T h u s, th e greater the
d eg r ee o f m iso riem a tio n o f existin g faults in th e tecton ic
stress field , fo r e x a m p le by progressive d o m in o rotation
(F ig. 8 ) , th e g r ea ter th e lik e lih o o d o f sig n ifica n t valve-
a ctio n an d flu id pressure cycling. C om pressional regim es
ca p a b le o f su sta in in g th e h ig h e st lev els o f overp ressu re
(F ig. 13) h ave th e h ig h e s t p o ten tia l fo r large-am p litu d e
flu id p ressu re c y c lin g (S ib son e t a!., 1988; C ox, 1995;
R obert e t al., 1995). T hus, alth ough in esozon al gold-quartz
v ein s o c c u p y in g fau lt-fractu re m e sh e s m ay d e v e lo p in a
ran ge o f tecto n ic en v iro n m en ts (Sibson an d Scott, 1 9 9 8 ),
th e m o st in ten se form s o f su ch m in eralization ten d to b e
h o sted in th e sem i-d u ctile roots o f steep reverse fault sys­
tem s in c o llisio n -a c c r e tio n settin g s (e .g ., th e M eso zo ic
M o th er L o d e v ein system in th e fo o th ills o f th e Sierra
N evada, C aliforn ia). S u ch vein system s co m m o n ly ex ten d
over 1 to 2 km vertically, com prising fault-veins d evelop ed
o n reverse or reverse-oblique faults in term esh ed in m utual
cross-cutti n g relationships with arrays o f flat-lying exten sion
veins (Figs. 15 an d 16). M ineralization generally occu rred
u n d e r sub -grecn sch ist to green sch ist m ctam orp h ic c o n d i­
tio n s (1 0 ± 5 km d e p th ) in com pressional-transpressionai
tecto n ic regim es. T h e com pressional stress field served to
co n ta in d eep -so u rced , lithostaticafly overp ressu red H 20 -
COy flu ids a ro u n d d ie base o f th e seism o g en ic z o n e until
(a) R isin g r In c rea sin g Pt

FAULT W S T A S IU n
•c*C*j»s(0n*Pf__ — Pf CYCLE
RU PTU RE .

F luid D ischarge
‘ D ecrea sin g P ,

(b) EXTENSIONAL TECTONIC REGIME

VALVING THROUGH FORMATION EXTREME VALVE-ACTION ON

OF A FAULT-FRACTURE MESH SEVERELY MISORIENTED FAULTS

v ///////////////////^ ^ ^ ^

X ^ °3

..............
/ w/ / w/ /...s/ \s/ \/w./

(C) COMPRESSIONAL TECTONIC REGIME

Fig. 19. \farieucs o f fault-valve behavior, a. Schematic o f general fault-valve cycle, illustrating coupled cycling o f shear
stress ( t), fluid pressure (Pr),a n d frictional fault strength (ff). b. Distributed and localized valving in an cxtensional tec­
tonic regime, c. Distributed and localized valving in a compressions! tectonic regime (diagrammatic and not to scale).
rupture, followed by focused upwards discharge along the (a)
reverse faults. S N W E
Valving action in sedimentary basins: There is accumulating
evidence that varieties of fault-valve activity may also occur at
comparatively high levels in overpressured sedimentary
basins, promoting oil and gas as well as hydrothermal fluid
migration. The presence o f localized hydrothermal plumes
above overpressured portions o f the Gulf Coast basin
(Roberts et al., 1996; McKenna and Sharp, 1997) is sugges­
tive of valving action along normal faults, as are abnormal
diagenctic assemblages localized around faults in the North
Sea basin (Burley e t al., 1989). A likely candidate for an
active fault-valve province in an area of compressional tec­
tonics is the east Santa Barbara Channel-Ventura Basin
region in the western Transverse Ranges of California, an
actively contracting and overpressured fold-and-thrust belt
characterized by steep, seismically active reverse faults and
rapidly amplifying folds (Yeats et al., 1988; Shaw and Suppe,
1994). One such fold structure is the Vemura-Rincon antl-
clinorium (Fig. 20), host to a super-giant oilfield. Growth of
this structure has occurred within the last 200,000 years, and
the anticlinal crest is currently uplifting differentially at up
to 10 mm/yr. Oil production from strata as young as Pleis­ Fki. 20. a. Reverse faulting, often rooted in bedding shears, defined by
tocene suggests that hydrocarbon migration is still ongoing. drilling in the core o f the growing, fluid-overpressured Ventura anticline
This growing su*ucture is flanked by steep, outward-dipping o f southern California (after Yeats. 1983). b. Gold-bearing quartz veins
hosted in saddle reef structures and reverse faults rooted in bedding
reverse faults that remain microseismically active with focal
shears within the Garden Gully Anticline in deform ed Ordovician tur­
mechanisms defining almost pure reverse dip-slip (Yeats et bidites, Confidence Extended mine, Bendigo, Australia (after Cox et al.,
al., 1987). The core of the Ventura Anticline, which is 1991; Whitelaw, 1918).
strongly overpressured with Ay values approaching 0.9 at
depths of 3 km or so, is likewise disrupted by a mesh of bed­
ding-parallel and cross-stratal reverse-slip faults, m ost of associated with major episodes of fluid redistribution within
which dip at moderate to steep angles (Yeats, 1983). the crust. Particularly interesting situations arise during
Figure 20 draws attention to the remarkable similarities positive tectonic inversion, where there is an approximately
between the internal structure o f the Ventura Anticline and coaxial change from regional extension to compression,
the system o f gold-quartz veins hosted by a folded and and during negative inversion where the reverse occurs.
cleaved sequence of Ordovician turbidites within the Gar­ During positive inversion, the mean stress increases from
den Gully Anticline (Whitelaw, 1918) near Bendigo in the a value less titan the vertical stress in the extensional regime
State of Victoria, Australia. Incremental development of the to a value greater than the vertical stress in the compres­
vein system provides evidence for episodic large-volume sional regime, thereby boosting fluid pressure levels (Sib-
flow along reverse-slip discontinuities, some of which paral­son, 1995). Maximum sustainable overpressure is also
lel, and some of which crosscut bedding (C oxet al., 1991). higher in the compressional regime because inherited nor­
A reasonable inference, given the similar structural perme­ mal faults are poorly oriented for reactivation in the new
stress Field. Such scenarios favor the developm ent o f
ability in the two anticlinal structures, is that fluid migration
within the actively growing Ventura Anticline is likewise regional fault-valve provinces where compressional reacti­
episodic and concentrated along active reverse-slip sur­ vation o f inherited normal faults leads to extreme valving
faces. Migration o f overpressured fluids through compara­ acuon on steep reverse faults. Extreme fault-valve behavior
ble reverse-fault-coied anticlines in Cook Inlet, Alaska, hasin such settings will eventually be terminated by the devel­
also been recognized by Bruhn e t al. (2000). Despite the opm ent o f new, favorably oriented thrusts in the compres­
considerable variations in scale and structural level within sional stress field, and is, thus, likely to be comparatively
the crust, redistribution o f overpressured fluids in all o fshort lived. T h e eastern Santa Barbara Channel-Ventura
these cases seems attributable to valving action on moder- Basin region discussed above is one such area o f active tec­
ately-to-steeply dipping reverse-slip discontinuities. tonic inversion where inherited normal a n d /o r strike-slip
faults developed during Miocene extension have, since the
Regional valve-action from changes in stress regime Pliocene, been reactivated within an actively contracting
The dependence of maximum sustainable overpressure fold-and-thrust belt as steep reverse faults. The changeover
on the stress state and fault architecture (Fig. 13) suggests between active extension and shortening took place within
that regional changes in die tectonic regime are likely to be as little as 2 million years (Yeats et al., 1988).
 D uring negative inversion, high fluid overpressures
developed and m aintained during progressive regional
contraction cannot be sustained in the extensional stress
field because o f the decrease in maximum sustainable over­
pressure (Fig. 13). Rapid release o f overpressured fluids
may then take place through self-generauon o f extensional
fault-fracture meshes (Fig. 15), as appears to have occurred
during formation of the Alaska-Juneau gold-bearing quarts
vein system in southeast Alaska (Miller et al., 1992; Sibson
and Scott, 1998).
During both positive and negadve inversion, the speed of
the tectonic transition is likely to be a critical factor in
determining the extent o f fluid redistribuuon and its effec-
dveness in prom oting mineralizadon. The same phenom ­
ena may also occur on rather short dme-scales in strike-slip
fault systems, where poruons of the crust move rapidly from
areas of local transtension to transpression, or vice versa.
Summary Discussion
Dynamic fluid redistribution within the seismogenic
upper crust involves intercoupled stress and permeability
cycling. Large earthquakes on established faults induce
fluid redistribution into diladonal sites during aftershock
periods, but packages o f overpressured fluids m igrating
through stressed crust also self-generate structural perme­
ability by earthquake swarm activity involving distributed
brittle failure and reshear. Intermittent seismic slip in fault-
fracture systems causes localized hydrothermal precipita­
tion through a variety of mechanisms. Abrupt reducuons in
fluid pressure at specific structural sites may arise from suc­
tion-pump acdon involving rapid slip transfer across dila­
donal irregularities, o r through various forms of fault-valve
behavior where ruptures transect boundaries to overpres­
sured pordons o f the crust (Fig. 4d). Such pressure reduc-
dons may induce phase separadon and hydrothermal pre-
cipitaiion throughout the postseismic (aftershock) period
of readjustment. Episodic renewal o f fault-fracture perme­
ability may also, however, promote mixing of fluids derived
from different sources, pardcularly during valving acdon.
Originally deep, hot, and overpressured fluids o f diage-
nedc, metamorphic, or magmadc origin may be discharged
episodically into colder fluids circulating within the near­
surface, hydrostatically pressured regime.
High-flux flow through low-permeability rocks requires
dilatant fault-fracture meshes, which can form and reacti­
vate only under low effective stress (a3' < 0, o r Pf > a 3) in
the absence o f throughgoing low-cohesion faults that are
well oriented for frictional reactivation. W hen present,
such, throughgoing faults limit the degree o f overpressur­
ing and the amplitude of any fluid-pressure drops accom­
panying fault-valve action, lowering the potential for
hydrothermal precipitation from pressure cycling. High-
flyk flow accom panied by large am plitude fluid-pressure
cycling is, therefore, achievable only under special struc-
; tural circumstances, for example around existing fault sets
that have undergone domino-rotation to frictional lock up.
These conditions for pulsed high-flux flow do much to
explain observed relationships between styles of mineral­
ization and tectonic setting. Within extensional-transten-
sional tectonic regimes, conditions for high-flux flow lead­
ing to epizonal mineralization can be m aintained under
hydrostatic fluid pressures in the shallow crust to depths
dependent on rock tensile strength, but in compressional-
transpressional regimes, fluids overpressured to around
lithostatic values are needed for mesozonal mineralization
to develop through fault-valve action in the midcrust.
Extreme valve-action on steep reverse faults is most likely to
develop in areas of positive tectonic inversion, or in colli­
sion-accretion complexes where initially shallow-dipping
thrusts have steepened to lock up as a consequence of pro­
gressive horizontal contraction.
Acknowledgments
Thanks to G. Broadbent, S. Cox, L, Miller, H. Poulsen,
and F. Robert, and to the earthquake researchers of the
U.S. Geological Survey, Menlo Park, especially D. Hill, S.
Hickman, B. Ellsworth, and T. Hanks, who over many years
have focused my attention on issues o f faulting and fluid
flow, and to the course organizers for the opportunity to
participate. R. Bruhn, D. Hill, andj. Richards provided con­
structive advice on the manuscript and S. Read helped with
the figures. This work was funded by the NZ Public Good
Science Fund through FRST Contract #C05611.
REFERENCES
Allis, R.G., 2000, Insights on the formation o f vapor-dominated geother­
mal systems: World Geothermal Congress 2000. Kyushu-Tohuku. Japan.
28 May-10 June, 2000, Proceedings, p. 2489-2196.
Am es. H.G ., 1948, The Perron m ine, in Structural geology o f Canadian
ore deposits: Canadian Institute o f Mining and Metallurgy, p. 893-898.
Anderson. E.M., 1951. The dynamics o f faulting and dyke formation with
application to Britain, 2nd cd.: Edinburgh, Oliver and Boyd. 206 p.
Angevine. C L ., Turcotte, D.L., and Furnish, M.D., 1982, Pressure solution
lithificadon as a mechanism for the stick-slip behavior o f faults: Tecton­
ics, v. l , p . 151-160.
A m oneliini, M„ and Aydin, A., 199-1, Effect o f faulting on fluid flow in
porous sandstones: Petrophysical properties: American Association o f
Petroleum Gcotogists Bulletin, v. 78, p. 355-377.
Armijo, R-, Tapponnicr, P„ Mercicr, J.L., and Han, T-L.. 1986, Quaternary
extension in southern Tibet: Field observations and tectonic implica­
tions:Journal o f Geophysical Research, v, 91, p. 13.803-13,872.
Bakun, W.H., and McEvitly, T.V., 1984, Recurrence models and ParkHcld,
California, earthquakes: (ournal o f Geophysical Research, v. 89,
p. 3051-3058.
Barton, G , Zoback, M.D., and Moos, D„ 1995, Fluid flow along potentially
active faults in crystalline rock: Geology, v. 23, p. 683-686.
Brown, N., Fuller, M., and Sibson, R„ 1991, Palcom agnctism o f the
Ocottlio Badlands, southern California, and implications for slip trans­
fer through an antidilational faultjog: Earth and Planetary Science Let­
ters. v. 102, p. 277-288.
Brown, SJL, and Bruhn, R.L-. 1996, Formation o f voids and veins during
faulting:Journal o f Structural Geology, v. 18, p. 657-671.
Bruhn, R.L., Yonkce. W„-Y, and Parry, W.T., 1990, Structural and fluid-
chcm ical properties o f seism ogenic normal faults: Tcctonophysics,
v. 175, p. 139-157.
Bruhn, R.L., Parry, W.T., and Bunds, M.P., 2000,Tectonics, fluid migration,
and fluid pressure in a deform ed fore-arc basin, C ook Inlet, Alaska:
Geological Society o f America Bulletin, v. 112, p. 550-563.
Buckland, W., 1836, Geology and mineralogy considered with reference to
natural theology: London, William Pickering, 5*18 p.
Burley, S.D., Mullis, J., and Matter, N-, 1989, Tim ing diagenesis in the Tar­
tan reservoir (U.K. North Sea): Constraints from combined cathodolu-
 mincsccncc spectroscopy and fluid inclusion studies: Marine and Petro­
leum Geology, v. 6, p. 08-120.
Byerlee, J.D., 1978. Friction o f rocks: Pure and Applied Geophysics, v. 116,
p .615-626.
------ 1093, A model for episodic flow o f high-pressure water in fault zones
before earthquakes: Geology, v. 21, p. 303-306.
Garter, N.L., and Tscnn, M .C, 1987, Flow properties o f continental lithos­
phere: Tectonophysics, v. 136, p, 27-63.
Cowic, P A , 1998, Normal fault growth in 3 dimensions in continental and
oceanic crust, i»i Buck, R„ Delaney, P.T., Karson.JA. and Lagabriellc, Y.,
cds„ Faulting and magmatism at mid-ocean ridges: American Geophys­
ical Union Monograph, v. 106, p. 325-348.
Cox, S.F., 1995, Faulting processes at high fluid pressures: An example o f
fault-valve behaviour from the Wattle Gully Fault, Victoria, Australia:
journal o f Geophysical Research, v. 100, p. 12,8*11-12,860.
Cox, S.F., Wall, V.J., Etheridge, M A , and Potter, T.F., 1991, Deformation
and m etam orphic processes in the formation o f m csothcnnal vein-
hosted gold deposits— examples from the Lachlan fold belt in central
Victoria, Australia: Ore Geology Reviews, v. 6, p. 391—123.
Cox, S.F., Knackstedt, M.A., and Braun, J., 2001, Principles o f structural
control on permeability and fluid flow in hydrothermal systems: Reviews
in Economic Geology, v. 14, p. 1-2-1.
Cram pin, S., 199-1, T he fracture criticality o f rock: Geophysical Journal
International, v. 118, p. -128—*138.
Davis, D.J., Suppe, J.. and Dahien, F A , 1983, The mechanics o f fold-and-
thrust belts and accrctionary wedges: journal o f Geophysical Research,
v. 88, p. 1153-1172.
Etheridge, M A , Wall, VJ„ Cox, S.F., and Vernon, R.H., 198*1, High fluid
pressures during regional metamorphism and deformation: Implica­
tions for mass transport and deformation mechanisms: Journal o f G e o
physical Research, v, 89, p. 434*1-1358.
Evans, J.P., Forster, C.B., and Goddard, J.V., 1997, Permeability o f fault-
related rocks, and implications for hydraulic structure o f fault zones:
Journal o f Structural Geology, v. 19, p. 1393-1404.
Fournier, R.O., 1991, T he transition from hydrostatic to greater than
hydrostatic fluid pressures in presendy active hydrothermal systems in
crystalline rock: Geophysical Research Letters, v. 18, p. 955-958.
— 1996, Compressive and tensile failure at high fluid pressure where pre­
existing fractures have cohesive strength, with application to the San
Andreas fault: Journal o f Geophysical Research, v. tOI, p. 25,499-25.509.
— 1999, Hydrothermal processes related to m ovement o f fluid from
plastic into brittle rock in the magmatic-epithermal environment: Eco­
nomic Geology, v. 94, p. 1193-1211.
Groshong, R.H., 1988. Low-temperature deformation mechanisms and their
iniciprcintion: Geological Society o f America Bulletin, v. 100. p. 1329-1360.
Hamilton, R.M., 1972, Aftershocks o f the Borrego Mountain earthquake
from April 12 to June 12, 1968: U.S. G eological Survey Professional
Paper 787. p. 31-5*1.
Hanks, T.C., and Kanamori, I F.. 1979, A moment-magnitude scale:Journal
o f Geophysical Research, v. 81, p. 2848-2350.
Hanmcr, S., 1988, Great Slave Lake Shear Zone, Canadian Shield: Recon­
structed vertical profile o f a crustal-scale fault zone: Tectonophysics,
v. 149. p.245-264,
Harris, R A , 1998, Introduction to special session: Stress triggers, stress
shadows, and implications for seismic hazard: journal o f Geophysical
Research, v. 103, p. 3*17-358.
Hasegawa, A., Horiuchi, S., and Umino, N., 1994, Seismic structure o f the
northeastern Japan convergent margin: A synthesis: journal o f Geo­
physical Research, v. 99, p. 22,295-22,311.
Hauksson, E„ Jones. L.M., Hutton, K., and Ebcrhart-Phillips, D., 1993, The
1992 Landers earthquake sequence: Scismological observations: Journal
o f Geophysical Research, v. 98, p, 19,835-19,858.
Henley, R.W., 1985, The geothermal framework for epithermal deposits.
in Berger, B.R., and Bcthke, P.M., cds., Geology and geochemistry o f
epitherm al systems: Society o f Econom ic Geologists. Reviews in Eco­
nomic Geology, v. 2, p. 1-24.
Henley, R.W., and Bciger, B.R., 2000. Self-ordering and complexity in opt­
ional mineral deposits: Annual Reviews o f Earth and Planetary Sciences,
v. 28. p. 669-719.
Hickman, S., Sibson, R., and Bnihn, R., 1995, Introduction to special sec­
tion: "Mechanical involvement o f fluids in faulung”: Journal o f Geo­
physical Research, v. 100. p. 12,831-12,840.
Hill, D.P., 1977, A m odel for earthquake swarms: Journal o f Geophysical
Research, v. 82. 1347-1352.
Hill, D.P., Eaton, J.P., and Jones, L.M., 1990, Seismicity, 1980-86, in Wal­
lace. R.E., cd., The San Andreas fault system. California: U.S. Geological
Survey Professional Paper 1515, p. 115-151.
Hubbert, M.K., and Rubey, W.W., 1959, Role o f fluid pressure in mechan­
ics o f overthrust faulting: Geological Society o f America Bulletin, v. 70,
p. 115-205.
Hulln, C.D., 1925, Structural control o f ore deposition: Economic Geol­
ogy. v. 24, p. 15—19.
Hunt.J.M ., 1990, Generation and migration o f petroleum from abnor­
mally prcssurccl fluid compartments: American Association o f Petro­
leum Geologists Bulletin, v. 74, p. 1-12.
Ito, K., 1999, Seismogenic layer, reflective lower crust, surface heat flow,
and large inland earthquakes: Tectonophysics, v. 306, p. 423-433.
jackson.JA ., and White. N.J., 1989, Normal faulting in the upper conti­
nental crust: Observations from regions o f active extension: Journal o f
Structural Geology, v. 11, p, 15—36.
Jaeger. J.C., and Cook, N.G.W., 1979, Fundamentals o f rock mechanics,
3rd cd: London, Methuen, 593 p.
Kanamori, II., and Anderson, D.L., 1975, Theoretical basis o f some empir­
ical relations in seismology: Scismological Society o f America Bulletin,
v. 65, p, 1073-1096.
Kennedy, B.M., Kharnka, YK., Evans, W.C., Ellwood, A., DePaolo. D.J.,
Thordsen, J„ Ambats, G., and Mariner, R.H., 1997, Mantle fluids in the
San Andreas fault system, California: Science, v. 278, p. 1278-1281.
King, G.C.P., 1986, Speculations on the geom euy o f the initiation and ter­
mination processes o f earthquake rupture and its relation to morphol­
ogy and geological structure: Pure and Applied Geophysics, v. 124,
p. 567-585.
King, G.C.P., Stein, R.S„ and Lin, J„ 1994, Static stress changes and the
triggering o f earthquakes: Scismological Society o f America Bulletin,
v. 8*1, p. 935-953.
Knopf, A., 1929, The Mother Lode system o f California: U.S. Geological
Survey Professional Paper 157,88 p.
Lockncr, D A , 1995, Rock failure: American Geophysical U nion Refer­
ence Shelf, v. 3, p. 127-147.
Luvcndyk, B.P„ 1991, A model for Ncogenc crustal rotations, transtension,
and iranspression in soutlrem California: Geological Society o f America
Bulletin, v. 103, p. 1528-1536.
McCaig, A M ., 1997, The geochemistry o f volatile fluid flow in shear zones,
in H olness, M.B., ed ., Deform ation-enhanced fluid transport in lire
Earth’s crust and mantle: London, Chapman and Hall. p. 227-266.
McKenna. T.E., and Sharp, J.M., 1997, Subsurface temperatures, fluid
pressures, and salinities in tire Rio Grande embayment. Gulf o f Mexico
basin. U .S A : International G eological Congress, 30th, Proceedings,
v. 8, p. 263-274.
McKinstry, H.E., 1948, M ining geology: Englewood Cliffs, New Jersey.
Prentice-Hall, 677 p.
Miller, L.D., Barton, C.C., Frcdcrickscn, R.S., and Bressler, J.R., 1992,
Structural evolution o f the Alaska Juneau gold deposit, southeastern
Alaska: Canadian Journal o f Earth Science, v. 29, p. 865-878.
Moore, D.E., Lockncr, D., and Byerlee, J.D., 1994, Reduction o f perme­
ability in granite at elevated temperatures: Science, v. 265, p. 1588-1561.
Morgan, P.G., 192-1. Tire geology and mines o f tire Waihi District, Hauraki
Goldfield: New Zealand Geological Survey Bulletin, v. 26, 218 p.
Mount, V.S., and Suppe, J., 1992, Present-day stress orientations adjacent
to active strike-slip faults: Journal o f Geophysical Research, v. 97,
p. 11,995-12,013.
Muir-Wood, R-, and King. G.C.P., 1993, Hydrological signatures associated
with earthquake strain: jou rn al o f Geophysical Research, v. 98,
p. 22,035-22,068.
Muraoka, H ., U chida, T., Sasada, M„ Yagi, M„ Akaku, K., Sasaki, M.,
Yasukawn. K_. Miyazaki, S-L, Doi, N.. Saito. S.. Sato, K.. and Tanaka. S.,
1998, Deep geothermal resources survey program: Igneous, metamor­
phic. and hydrotlicnnal processes in a well encountering 500°C at 3,729
m depth. Kakkonda, Japan: Ccothennics. v. 27. p. 507-533-1.
Nadeau, R.M., and McEvilly, T.V., 1997, Scismological studies at Parkfield
V: Characteristic microearthquakc sequences as fault-zone drilling tar­
gets: Seismological Society o f America Bulletin, v. 87, p. 1463-1472.
N cwhouse, W.H., 1942, Ore deposits as related to structural features:
Princeton, New Jersey, Princeton University Press, 280 p.
Nguyen, P.T., Cox. S.F., Harris, L.B.. and Powell. C M ., 1998. Fault-valve
behaviour in optim ally oriented shear zones: An exam ple at the
Revenge gold mine, Kambalda, Western Australia: Journal o f Structural
Geology, v. 20. p. 1623-1640.
Nur, A., and Booker, J.R., 1972, Aftershocks caused by pore-fluid flow?: Sci­
ence, v. 173, p. 883-887.
Ohlmacher, G.C., and Aydin, A., 1997, Mechanics o f vein, fault, and solu­
tion surface formation in the Appalachian Valley and Ridge, northeast­
ern Tennessee. U.S.A.: Implications for fault friction, state o f stress, and
fluid pressure: Journal o f Structural Geology, v. 19, p. 927-94-1.
Pacheco, J.F., Scholz. C.H., and Sykes. L.R.. 1992, Changes in frequency-
size relationship from small to large earthquakes: Nature, v. 355,
p. 71-73.
Party, W.T., 1998, Fault-fluid com positions from fluid inchision observa­
tions and solubilities o f fracture-sealing minerals: Tcctonophysics, v. 290,
p. 1-26.
Passchlcr, C.W., and Trouw, RA J., 1996, Microtcctonics: Berlin, Springer-
Verlag, 289 p.
Peltzcr, C»., Rosen, P., Rogez, F., and Mudnut, K.. 1996, Postseismic
rebound in fault step-overs caused by pore fluid flow: Science, v. 273,
p . 1202-1204.
Powley, D.E., 1990, Pressures and hydrogeology in petroleum basins: Earth
Science Reviews, v. 29, p. 215-226,
Proffett, J.M., 1977, Ccnozoic geology o f the Ycrington District, Nevada,
and implications for the nature and origin o f Basin and Range faulting:
Geological Society o f America Bulletin, v. 88, p. 247-266.
Reasenberg, P., and Ellsworth, W.L., 1982, Aftershocks o f the Coyote Lake,
California, earthquake o f August 6, 1979: A detailed study: Journal o f
Geophysical Research, v. 87, p. 10,637-10,655.
Rice, J.R., 1992, Fault stress states, pore pressure distributions, and the
weakness of the San Andreas fault, in Evans, B„ and Wong, T-F., eds.,
Fault mechanics and transport properties o f rocks: San Diego, Acade­
mic Press, p. -175-503.
Robert, F., and Brown, A.C., 1986, Archcan gold-bearing quartz veins at
the Sigma mine, Abitibi greenstone belt, Quebec: Part I—G eologic rela­
tions and formation o f the vein system: Econom ic Geology, v. 81,
p. 578-592.
Robert, F., Boullier, A-M., and Firdaous, K., 1995, Gold-quartz veins in
mcLamorphic terrnnes and their bearing on the role o f fluids in fault­
ing: Journal o f Geophysical Research, v. 100, p. 12,861-12,879.
Roberts, A.M., and Yielding, G., 1994, Continental cxtcnsional tectonics,
in Hancock, P.L., cd., Continental deformation: O xford, Pergamon
Press, p. 223-250.
Roberts, S.J., N u r m J A , Cathles, L., and Cipriani, F-D„ 1996, Expulsion of
abnormally pressured fluids along faults: Journal o f Geophysical
Research, v. 101. p. 28,231-28,252.
Rutter, E.H., 1986, On the nomenclature o f m ode o f failure transitions in
locks: Tcctonophysics. v. 122, p. 381-387.
Scholz, C.H., 1982, Scaling taws for large earthquakes: Consequences for
physical m odels: Scistnological Society o f Am erica Bulletin, v. 72,
p. 1-14.
------ 1988, The britde-plnstic transition and the depth o f seismic faulting:
Geologische Rundschau, v. 77, p. 319-328.
— 1998, Earthquakes and friction laws: Nature, v. 391, p. 37-42.
Scholz, C.H., Sykes, L.R., and Aggnrwat, Y.P., 1973, Earthquake prediction:
A physical basis: Science, v. i l l , p. 803-810.
Schwartz, D.P., and Coppersmith, K.J., 1984, Fault behavior and charac­
teristic earthquakes: Examples from the Wasatch and San Andreas
faults: Journal o f Geophysical Research, v. 89, p. 5681-5698.
Sccor, D.T., 1965, Role o f fluid pressure in jointing: American journal o f
Science, v. 263, p. 633-646.
; Segall, P„ and Pollard, D.D., 1980, Mechanics o f discontinuous faults: Jour-
nal o f Geophysical Research, v. 85, p. 4337-4350.
Seront, B., Wong,T-F., Caine,J.S., Forster, C.B., Brohn, R.L, and Frcdrich,
J.T., 1998, Laboratory characterization o f hydromechanical properties
••• o f a seism ogcnic normal fault system: Journal o f Structural Geology,
y. 20, p. 865-882.
Shaw,J.H., and SuppeJ., 1994, Active faulting and fold growth in the east­
ern Santa Barbara Channel, California: Geological Society o f America
Bulletin, v. 106, p. 607-626.
Sisson, R H ., 1983, Continental fault structure and the shallow earthquake
source:Journal o f the Geological Society, London, v. 140, p. 741-767.
------ 1984, Roughness at the base o f die seismogcnic zone: Contributing
factors: Journal o f Geophysical Research, v. 89, p. 5791-5799.
------ 1985a, Stopping o f earthquake ruptures at dilational fault jogs:
Nature, v. 316, p. 2-18-251.
1985b, A note on fault reactivation: Journal o f Structural Geology,
v. 27, p. 751-75-1.
------ 1986, Earthquakes and lineam ent infrastructure: Royal Society o f
London Philosophical Transactions, Earth Sciences, v. A317, p. 63-79.
------ 1987, Earthquake rupturing as a mineralizing agent in hydrothermal
systems: Geolog)', v. 15, p. 701-704.
------ 1991, Loading o f faults to failure: Scismological Society o f America
Bulletin, v. 81. p. 2493-2497.
------ 199*1, Crustal stress, faulting, and fluid flow, in Parnell, J., ed.. Geoflu­
ids: Origin, migration and evolution o f fluids in sedimentary basins:
Geological Society o f London Special Publication, v. 78, p. 69-84.
. 1995, Selective fault reactivation during basin inversion: Potential for
fluid redistribution through fault-valve action, in Buchanan, J.G., and
Buchanan, P.G., eds.. Basin inversion: Geological Society o f London
Special Publication, v. 88, p. 3-19.
------ 1996, Structural permeability o f fluid-driven fault-fracture meshes:
journal o f Structural Geology, v. 18. p. 1031-1042.
------ 1998, Brittle failure m ode plots for compressional and extensional
tectonic regimes: Journal o f Structural Geology, v. 20. p. 655—660.
------ 2000a, A brittle failure m ode plot defining conditions for high-flux
flow: Economic Geology, v. 95, p. 41—18.
------ 2000b, Fluid involvement in normal faulting: Journal o f Geodynam­
ics, v. 29, p. 469—199.
Sibson, ILH., and Scou,J.. 1998, Stress/fauh controls o n the containment
and release o f overpressured fluids: Examples from gold-quartz vein sys­
tems in Juneau, Alaska; Victoria, Australia; and O tago. New Zealand:
Ore Geology Reviews, v. 13, p. 293-306.
Sibson, R.H., Moore. J.M., and Rankin. A.H ., 1975, Seismic pumping: A
hydrothermal fluid transport mechanism: Journal o f the Geological
Society, London, v. 131, p. 653-659.
Sibson, R.H ., Robert. F., and Poulscn, K.H., 1988, H igh-angte reverse
faults, fluid pressure cycling, and mesochermal gold-quartz deposits:
Geology, v. 16, p. 551-555.
Sillitoc. R.H., 1993, Epithermal models: Genetic types, geometrical con­
trols. and shallow features, in Kirkham. R.V.. Sinclair. W.D., Thorpe, R.I.,
and Duke. J.M.. eds.. Mineral deposit modeling: Geological Association
o f Canada Special Paper *10. p. 403-417.
Sleep, N.H., and Blanpicd, M.L., 1994, Ductile creep and compaction: A
mechanism for transiendy increasing fluid pressure in mostly scaled
fault zones: Pure and Applied Geophysics, v. 1-13, p. 9—10.
Snokc, A.W., Tullis, J„ and Todd, V.R., eds., 1998, Fault-related rocks: A
photographic atlas: Princeton, New Jersey, Princeton University Press,
617 p.
Snow, D.T., 1968, Rock fracture spacings, openings, and porosities: Jour­
nal o f Soil Mechanics and Foundations Division, Proceedings American
Society Civil Engineers, v. 94, p. 73-91.
Tosdal, R.M., and Richarris,J.P., 2001, Magmatic and structural controls on
the developm ent o f porphyry Cu±Mo±Au deposits: Reviews in Eco­
nomic Geology, v. 14, p. 157-181.
Townend.J., and Zoback, M.D., 2000, How faulting keeps the crust strong:
Geology, v. 28, p. 399—402.
Wallace. R.E., and Morris, H.T., 1986, Characteristics o f faults and shear
zones in deep mines: Pure and Applied Geophysics, v. 124, p. 107-126.
Wesnousky, S.G.. 1988, Scismological andstniccutal evolution o f strike-slip
faults: Nature, v. 335, p. 340-342.
Westaway, R., 1998, D ep en d en ce o f active normal fault dips on lower
crustal flow regimes: Journal o f the Geological Society, London, v. 155,
p . 233-254.
White, S.H., Burrows, S.E., Carreras, J„ Shaw, N.D., and Humphreys, FJ.,
1980, On mylonites in ductile shear zones: Journal o f Structural Geol­
ogy, v. 2, p . 175-187.
Whitclaw, H.S., 1918, The Confidence group o f mines: Geological Survey
o f Victoria Bulletin, v. 3 0 ,3 2 p.
Woodcock, N.H., and Schubert, C., 1994, Continental strike-slip tectonics,
in Hancock, P.L., cd., Continental deformation: O xford, Pergamon
Press, p. 251-263.
Yeats, R.S., 1983, Large-scale Quaternary detachments in Ventura basin,
southern California:Journal o f Geophysical Research, v. 88, p. 569-583.

Yeats, R.S., Lee, W.H.K., and Yerkes, R.F.. 1987, Geology and seismicity of
the eastern Red Mountain fault, Ventura County: U.S. Geological Sur­
vey Professional Paper, v. 1339, p. 161-167.
Yeats, R.S., Hufulc, G.J., and Grigsby, F.B.. 1988, Oak Ridge fault, Ventura
fold belt, and the Sisardecollemeni: Geology, v. 16, p. 1112-1116.
Yeats, R.S., Sieh, K.. and Allen, C.R., 1997, The geology o f earthquakes:
Oxford, Oxford University Press, 568 p. .
Zoback, M.L., 1992, Stress field constraints on intraplate seismicity in east­
ern North America: Journal o f Geophysical Research, v. 97,
p. 11,761-11,782.
Society o f Economic CrttlogiUs
Reviews v. 14, 2001, p . 5 1 -7 4

Chapter S

Deformation, Fluid Flow, and Ore Genesis in H eterogeneous Rocks, with

Examples and Numerical Models from the M ount Isa District, Australia
N icholas H. $. O liver*
Econom ic Geology Research U n it, School o f E a rth Sciences, Jam es Cook U niversity, Tow nsville Q L D 4 8 1 I, A u stra lia

Alison O rd,
C S fR O D ivisio n o f E xploration a n d M in in g , P O B ox 4 3 7 , N ed la n d s, VVA 6 0 0 9 , A u stralia

Richard K. Valenta,
M o u n t Is a A lin es E xploration, B risbane, A u stra lia

and P haedra U pton **

C SIR O D ivisio n o f E xp loration a n d M in in g , A 'edtands, W A 6 0 0 9 , A u stra lia

Abstract

Consideration o f the role o f rock property variations is crucial in any analysis o f the effects of defor­
mation o n fluid flow and mineralization. An empirical analysis of any mineralized terrain should consider
this factor, in addition to those used in any other analysis o f geometry and kinematics, such as orienta­
tion, evolution o f the stress and strain Acids, and the known distribution o f veins, shear zones, breccias,
a n d a lte ra tio n . T h e conceptual models that arise from such an analysis can be enhanced by computer
m o d e ls . T h e m o d e ls sh o w n here are finite difference models that simulate fluid flow in deforming rock
masses, one fo r flu id flow along predefined rock boundaries (Universal Distinct Element Code, UDEC),
a n d another for fluid flow through d e f o r m i n g porous media (Fast Lagrangian Analysis o f Continua,
FLAC). UDEC modeling of the p e r tu r b e d stre ss field a r o u n d stronger, lower permeability meta-intnisive
rocks in the Mary Kathleen district a n d th e H ilto n m in e o f th e M o u n t Isa district, northwest Queensland,
reproduces the observed location of th e m o s t in te n s e v c in in g a n d a lte ra tio n . FLAC modeling o f the Mary
Kathleen U-REE orebody reproduces th e lo c a tio n a n d g e o m e try o f o r e s h o o ts and provides an explana­
tion for focusing of regional fluid to w a rd s th e o r e d e p o s itio n sites. FL A C m o d e ls of the giant Mount Isa
copper deposit reveal that the effect o f th e rh e o lo g ic a l h e te r o g e n e ity o n flu id flow and solute transport
is amplified if consideration is made o f w h e th e r o r n o t th e ro c k s a r e c o n tr a c t a n i or dilatant, with in ­
creasing strain. Multiple working h y p o th e s e s c a n b e e v a lu a te d q u ick ly by s u c h modeling; therefore, the
models can be used in exploration a n d o re b o d y e x te n s io n stu d ie s . F u r th e r m o r e , it is suggested from our
work that the size and spacing of e p ig e n e tic , s tru c tu ra lly c o n tr o lle d o r e d e p o s its is related in a fairly sys­
tematic way to the scale and degree o f ro c k p r o p e rty v a ria tio n s , a t le a st f o r a g iv en strain history. For giant
ore deposits to form, it requires th a t th e g r a d ie n ts in p o r e p r e s s u r e g e n e r a te d a t local scales by hetero­
geneous rock packages must be s u b o r d i n a te to th o s e o p e r a tin g a t b r o a d e r scales.

Introduction netic ore deposits on major structural boundaries is well

T his Paper is primarily concerned with understanding the
dynamics of fluid flow and rock deformation assuming het­
erogeneous rock properties, in order to predict better the
spadal distribution of epigenetic ore deposits in exploration
and expansion of existing operations. Fluid flow controls
the dissolution, transport, and concentration of metals into
epigenetic ore deposits. The common occurrence of epige-
♦Corrcsponding author: e-mail, nick.oliver@ jcu.edu.au
*Present address: MIM Exploration, GPO Box 1042, Brisbane, QLD
4001, Australia.
’‘“"Present address: D epartm ent o f Geology, University o f Otago,
D unedin, New Zealand.
known, and such localization is a complex interplay between
rock permeability, rock rheology', fluid properties, and the
applied stress suite. Deformation of rock in the presence of
fluid can cause a change in the stress state, which in turn
can affect the relationships between permeability and pore
fluid pressure, causing fluid advection. Likewise, fluid move­
m ent can affect the bulk stress state of the rock, which in
turn can cause deformation (Ord and Oliver, 1997).
This paper also presents some principles and examples
of com puter m ethods used in hydrodynamic analysis of
deformed terrains. Com puter simulations, constrained as
much as possible by field observations, provide an iterative
means to test different models of deformation, fluid flow,
and alteration, and to assess the relative importance o f var­ gradients that affect fluid flow. More comprehensive details
ious structural features. Realistic simulations can now be of these principles are provided by Cox et al. (2001) and
run on m icrocom puters with short run times, allowing Robert and Poulsen (2001). Although the numerical mod­
consideration o f num erous input conditions and iterative els we present are capable of including the effects of topog­
comparison with reality. Detailed cause and effect scenarios raphy, buoyancy, compaction, etc., on fluid advection, we
for interrelationships between fluid flow and deformation focus attention on the interplay o f deformation, stress, per­
cannot be readily defined, and for this reason numerical meability, and rock properties, such as may occur in the
modeling is playing an increasingly im portant role in the active core o f deforming m ountain belts.
developm ent o f new research and practical exploration Channeled fluid flow through fractures, and flow via a
strategies. Questions that can be addressed include: continuum of grain-scale permeability (porous media flow),
are convenient end-m ember styles o f fluid flow, although
1. What were the relative magnitudes and orientations there is a com plete spectrum between these two types of
o f the principal stresses during deformation? flow for most geological scenarios. The governing equations
2. Where were the zones o f deformation-enhanced per­ for fracture flow as opposed to porous media flow are dif­
meability relative to the large scale geometry? ferent, although they converge; nevertheless, as yet, no sin­
3. What were the resultant patterns of fluid flow? gle num erical code has been able to address the conver­
4. What was the likely location of mineralization? gence realistically for deform ing geological materials. In
both cases, fluid flow is strongly affected by heterogeneous
Commonly, die structural geology in a given area is well rocks with different mechanical properties. It is also perti­
understood geometrically and temporally. By combining nent to note at this stage that the two numerical codes we
conventional structural analysis with an understanding of present in this paper, UDEC and FLAC, simulate fracture
the spatial and temporal variation o f rock and fluid prop­ flow (by discrete elem ent modeling) and porous media flow
erties, an analysis o f the effects of deform ation and fluid (by continuum modeling), respectively. Each type o f model,
flow on mineralization can be performed. Here, we explore and each way of thinking, has advantages and disadvantages.
the range of relauonships in which the distribution o f rocks Porous m edia flow in which fluid flow occurs through
o f varying rheological properties influences the develop­ interconnected pore space (or theoretically through inter­
m ent and evolution o f fluid pathways, and ultimately, epi­ connected fracture networks) occurs in response to
genetic syntectonic orebodies. The work has implications changes in hydraulic head as expressed by Darcy’s law
for all o f the categories o f ore deposits for which syn- or (Etheridge etal., 1984):
late-tectonic origins have been proposed, including
Archean greenstone-hosted gold deposits, orogenic gold J^ K - \ r \.V P , (1)
deposits, and many Proterozoic and Phanerozoic Cu and
Cu-Au deposits. Numerical models can assist this process by where J A is the volumetric fluid flux (m V m -/s),
attempting to determine the likely size and spacing o f ore Kis the permeability (m2),
deposits form ed by epigenetic processes in deform ing
v the viscosity (k g /m /s ), and
crust. Models for deformation of homogeneous materials
generally fail to predict the location of epigenetic ore VP the gradient in hydraulic head (M Pa/m).
deposits, and do not readily describe the redistribution of
ore-forming components in rocks already containing min­ The units for flux are given in this form because they can
eralization. Deposits in which sulfides are localized in struc­ be equated with estimates of time-integrated fluid fluxes
tural sites, but arc regarded as being deformed and remo­ determined petrologically, using the principle by which the
bilized, may also be strongly influenced by the distribution progress of a reaction as measured by the proportions of
o f rocks o f varying strength around them. reactant and p ro d u ct m inerals is a m easure o f the total
fluid flux (e.g., Ferry and Dipple, 1991). For a given fluid
Principles o f Fluid Flow viscosity fluid flow is controlled by permeability and gradi­
Porous media flow and fracture flow are terms used to ents in hydraulic head. In deforming rocks with aqueous
describe fluid m igration in relatively shallow subsurface fluids, the head gradients may be largely a function of
environments involving, for example, basinal fluid flow and deform ation and mineral reactions, whereas in shallower
hydrocarbon and contam inant transport (e.g., Domenico scenarios this VP term may be dominated by the effects of
and Schwartz, 1998), and the formation of sediment-hosted buoyancy, topographic, o r thermal gradients.
ore deposits (e.g., Garven and Freeze, 1984a, b). This sec­ There are several important differences between shallow
tion briefly examines to what degree these same principles groundwater flow and that inferred for deeper crustal flu­
can be applied to the understanding o f transport and accu­ ids. Primary porosity is largely irrelevant at depth, because
m ulation o f metals in deeper crustal environments. The it is vanishingly small for most crystalline rocks. In crys­
geological environments we are considering in this paper talline rocks, the creation and destruction o f secondary
are ones in which rock perm eabilities are generally low, porosity, particularly effective porosity (that available to
fluid is most commonly overpressured, and the role o f rock contribute to permeability), is dom inated by combinations
deformation is prominent or predom inant in the hydraulic of: ( 1 ) mineral reactions that involve a change in volume
 o f the solid phase a n d /o r liberation o f volatiles previously
bound up in the minerals (Rumble and Spear, 1983; Ferry
and Dipple, 1991); (2) dynamic crack propagation during
deformation (Cox e t al., 2001); and (3) pressure solution
and dissolution/precipitation creep resulting from load or
tectonic stress (Durney, 1976; Beach, 1977).
Many models of fluid flow in deform ing rocks use
Darcy’s law as the basic principle, with the assumption that
the behavior of fluids in rocks at depth can be described
by porous m edia flow (O rd and Oliver, 1997). This
assumption requires that the permeability o f rocks can be
represented by realistic averages o f all of the above com­
plex processes. In comparison to hydrogeological studies
in which perm eability can be m easured and related to
grain size, grain size distribution, pore shapes, and wetting
angles, our qualitative to semiquantitadve estimates of per­
meability in rocks at depth are frustratingly inadequate,
and are the primary limitation on quantification of fluid
flow models. Despite this limitation, more realistic under­
standing of permeabilities at depth has been achieved in
recent years through recognition o f the complex arrays of
processes that affect permeability, a few o f which we pre­
sent here.
At high tem peratures (>200*0, m ineral reactions can
make a substantial contribution to permeability and fluid
flow, because most metamorphic devolatilizadon reactions
liberate fluid that may result in changes in V? and therefore
fluid flux (Etheridge et al., 1984). The effect of mineral
reactions on hydraulic conductivity is difficult to quantify
because wetting angles and percolation thresholds for rocks
undergoing mineral reactions are different to those pre­
dicted o r measured by compaction experiments in the pres­
ence o f fluid (e.g., Zhang et al., 1994). Typical prograde
mineral reactions involve a decrease in the volume o f the
solid products and an increase in the total volume o f all the
products (solids plus released volatiles), and, thus, have con­
siderable potential to create permeability (Rumble and
Spear, 1983). However, such reaction-enhanced permeabil­
ity will most likely only predominate at relatively low strain
rates; otherwise, intragranular plastic deformation a n d /o r
formation of fluid-filled fractures will predominate (Oliver,
1996; Cox et al., 2001). For example, in a ductile shear zone,
mineral reactions that create fluid overpressure will be
important, but permeability enhancem ent is unlikely to be
dominated by porosity related to the AV o f reaction; rather,
it will be dominated by dynamic processes o f microcracking
and intracrystalline deformation.
Abundan t veins in many crustal rocks attest to the flow of
fluid along fractures, and in low-permeability rocks at low to
moderate temperatures (<200°C), it is well understood that
fracture-flow dom inates the permeability, for example in
limestones (e g., Lattman and Parizek, 1964) o r fractured
granites (Brace, 1968). However, to solve explicitly for the
fracture flow requires information on fracture orientation,
density, connectivity, aperture, hydraulic conductivity, and
smoothness, a very challenging list o f requirem ents. Flow
through fractures is simulated by various rock failure theories
which specify the relationships between fractures, fluid
pressure, and stress (e.g., Griffith, Anderson, Riedel, Mohr-
Coulomb theories), and also by considering fluid flow solu­
tions in parallel-walled o r cylindrical channels (much engi­
neering literature), especially laminar flow. The important
equation here is the condition for laminar flow between par­
allel-sided smooth plates:
(Pa-Pb)W 2
Q= 12 Lv » (2)
where Q is the instantaneous fluid flux (m /s),
L is the distance between the levels o f pressure Pa and Pb,
W is the channel width, and
vth e viscosity (modified from Phillips, 1972).
Equation (2) contains the basic information needed to
consider the specific geometry of flow channels, although
equations for rough (e.g., Brown, 1987) or highly tortuous
channelways (e.g., Helgeson and Lichtner, 1987) are con­
siderably more complex. The degree of intcrconnectivity of
individual fractures (see Sibson, 1996; Cox et al., 2001) is
also particularly important. By assessing fracture connectiv­
ity and percolation thresholds for a specific, known fracture
array (Cox e t al., 2001), flow m odeling can proceed to
demonstrate the localization, for example, of high fluid flux
sites and, hence, potential sites for mineralization.
The m odeling conundrum and its link to exploration
models is now a p p a re n t-sh o u ld we include specific frac­
tures in models, or try to represent rocks by averages in
which dynamic fracture permeabilities are implicit o r vary
throughout rock bodies in a systematic way? Equation (2)
can be extended to a volume o f rock having several or
many fractures (Dom enico and Schwartz, 1998). If an
appropriate average can be selected for a representative
rock volume, the permeability o f the fractured rock body
can then be approximated by an average, and Darcy’s law
may be applied so that the rock can be modeled in a con­
tinuum sense. In o rd er for an approxim ation o f porous
media flow to hold in fractured rocks, the representative
volume of rock under consideration must contain an aver­
age num ber o f fractures with respect to their abundance,
size, and hydraulic conductivity, an approximation that is
difficult to demonstrate practically. In our porous medium
analogy for fluid flow through fractured rocks in deep-
seated crustal environm ents presented below, we m ust
assume that the representative volume of fractured rock is
smaller than the size of the finite difference grid elements
(typically 50 x 50 m ), that the models, therefore, present
realistic averages o f fluid flow and rock properties, and that
Darcy’s law, therefore, applies. This assum ption also
requires, for example, that the average permeability o f frac­
tured rocks is characteristically up to Five o r six orders of
magnitude greater than revealed by in situ testing o f intact
rock (see synopsis by Domenico and Schwartz, 1998). This
approximation has been presented, but is still difficult to
constrain, in many m odeling studies (cf. Holyland and
Ojala, 1997; Ord and Oliver, 1997). For example, a 1-mm-
wide open fracture running across an essentially imperme­
able 1 m2-rock face will conduct as much fluid as a porous
m aterial with a perm eability o f 84 Darcys (c.g., a coarse
unconsolidated sand body near the E arth’s surface;
Domenico and Schwartz, 1998).
A difficulty remains with assessing specific fracture arrays
in gathering sufficient information to adequately account
for the variability of fracture properties in time and space.
This difficulty explains the preference o f many authors,
both in petrological and geom echanical simulations, to
attempt to model fluid flow using continuous porous media
sim ulations in which the details o f the fracturing are
em bedded within assumptions concerning average rock
properties and permeability (c.g., Dipplc and Feriy, 1992;
Hanson, 1992; O rd and Oliver 1997).
A nother m odeling approach is the discrete elem ent
m ethod in which fluid flow is confined to rock boundaries
or specified joints and fractures (e.g., Holyland and Ojaia,
1997; Jiang e t a!., 1997; some examples below). This
approach focuses on the stress conditions accompanying
deformation, and numerically treats the rock bodies them­
selves as impermeable. T he distribution o f areas o f low
mean and differential stress are then used as a proxy to pre­
dict fluid pathways, rather than the continuum methods in
which it is possible to fully couple deform ation and fluid
flow. Although this type o f modeling has drawbacks (e.g.,
Darcy’s law cannot fully be incorporated), it is one o f few
approaches to date, other than percolation theory, that
allows sim ulation o f fluid flow along specified fracture
arrays. To make a decision concerning the use o f any par­
ticular practical or conceptual m odel of fluid flow in
deforming, mineralized rocks, the following questions must
first be answered:
1. Which type o f process dominates the flow system: flow
along specific meso- o r macroscale fractures, o r flow
through permeable (porous or microfractured) bulk rock?
2. Can the permeability of different fractured rocks in
the area be averaged over a suitably small representative
volume?
3. Are sufficient data available on properties and con­
nectivity o f the fractures to warrant a simulation that spec­
ifies the properties o f individual fractures?
Deformation and Fluid Flow
The general principles o f the interrelationship between
deformation, fluid pressure, and fluid flow have been dealt
with at length in the literature, and are summarized in this
volume by Cox et al. (2001) and Stbson (2001). Pertinent
to this discussion are (a) the law of effective stress:
<Vir = <7„-pf. (3)
where C7vi} is the effective normal stress acting on a plane,
(Tn is the normal stress, and
Pf is the fluid pressure (e.g.,Jaeger, 1969),
as depictable on Mohr circle diagrams; and (b) the condi­
tion for tensile failure at high fluid pressures:
Pf >ff3 + T, (4)
where a 3 is the minimum compressive stress, and
T is the tensile strength of the rock (Secor, 1965;Jaeger, 1969).
These equations describe the simplest possible relation­
ships between an applied stress and the fluid pressure, in
which high fluid pressures reduce effective stresses and may
lead to rock failure, and, conversely, high applied stresses
may increase the fluid pressure. Although these relation­
ships are superficially simple, they become very complex
when different rock types are included in the analysis, and
when complex deform ation histories and geometries are
considered. For example, deformation o f several different
rock types, each with different tensile strengths, will result
in variations of + T throughout, which will cause local­
ization o f the potential conditions for tensile failure at any
given time. Continued deformation, which would result in
another distribution of <73, would further modify the area
where tensile failure is possible (Holyland and Ojaia, 1997).
Once failure occurs, the fluid pressure changes, causing
effective stress changes, which may lead to further defor­
mation. Some numerical models can simulate this complex
coupling between fluid flow and deformation (see below).
Homogeneous deformation
Although a hom ogeneous stress field can be imagined
leading to evenly distributed fluid flow, plastic deformation
in frictional m aterials generally leads to localization o f
strain, and hence o f fluid flow, owing to liny perturbations
of the stress field emanating from instabilities (Ord, 1990).
In reality, small irregularities in homogeneous rock bodies
can lead to localization of deformation and of fluid flow.
However, such irregularities will generally be far subordi­
nate to those occurring along the boundaries between rock
bodies, because o f the rheological contrasts.
An approach to homogeneous flow in duciilely deform­
ing rocks may occur if the fluid flux is very high relative to
die strain rate (Ord and Oliver, 1997). In such cases the
rock is swamped with fluid and the deformation conditions
are o f secondary im portance (see below). In regional
m ctam orphic rocks, this situation would be relatively
unusual, but may apply in some cases of contact m eta­
m orphism , near-surface flow, o r intrusion-related fluid
expulsion where fluid fluxes are very high relative to strain
rates. A m ore comm on scenario, however, would be the
situation where stress perturbations arising because o f the
juxtaposition o f two (or more) unalike rocks extend out­
wards into otherwise hom ogeneous rocks, causing local­
ization of fluid flow (see below). Again, such localization
will most likely be subordinate in comparison to that devel­
oped immediately around the rock interface that caused
the perturbation.
Heterogeneous Deformation and Fluid Flow
Three main factors apply in the control o f fluid flow dur­
ing deform ation of heterogeneous rock suites (see also
Oliver etal., 1990): (1) the pore pressure, Pf; (2) gradients
in Pf; and (3) the orientadon and magnitude o f the applied
stresses.
The effects of pore pressure and pore pressure gradients
on fluid flow in deform ing rocks are addressed by Cox et
al. (2001) andSibson (2001), building on a long history of
study in the field including seminal papers by Secor (1965),
Kerrich and Allison (1978), Etheridge etal. (1984),Sibson
e ta l, (1988), Boullier et al. (1994), Valenta e ta l. (1994),
and Cox (1999). In this section, and for the rest of the
paper, we focus particular attention on the role of variable
rock properties and boundary conditions in influencing
the above factors.
Rock boundaries and other planes of xueakness
Figure 1 shows how failure on an anisotropy is favored at
Pf lower than that required to fracture the bulk rock. Con­
sider a body o f rock with a plane o f weakness oriented at
60° to <7,. The cohesion and friction angle will, for most
cases, be considerably lower along the plane of weakness
than in the rock itself. The Mohr-Coulomb failure criterion
for this plane is different from that of die rock, as shown in
Figure 1. Even though a fracture set may have low perme­
ability according to the law of effective stress (eq. 3), if this
is the only anisotropy in the rock, then failure will occur on
this plane before failure occurs on the bulk rock. For exam­
ple, this process could result in the formation o f bedding- Fig. 1. T he effect o f planes o f weakness on failure criteria, a. Schematic
or foliation-parallel veins in unusual orientations relative to rock body with planes o f weakness, b. Mohr circle representation. Mohr
the inferred stress field (Reynolds and Lister, 1987; Valenta circles are show n with a secon d failure envelop e that represents the
etal., 1994). strength o f the rock along a plane o f weakness. Failure o n an anisotropy
(a) is favored a t Pr lower than that required to fracture the bulk rock. Note
Orientat ion of rock boundaries that for the fracture set a, although it has low permeability according to
the law o f effective stress (eq. 3), if this is the only anisotropy in the rock,
Localization of fluid flow and o f orebodies commonly then failure will occur on this set before failure occurs o n the bulk rock.
occurs along misoriented structures (see also Holyland and Modified from Oliver e ta l. (1990).
Ojala, 1997; Cox et al., 2001). Such misorientation (relative
to the plane o f shear for failure in homogenous rock) nor­
mally implies that the favorable structure is aligned at a rel­ may arise as a consequence o f deform ation o f highly het­
atively low angle to the m aximum compressive stress. erogeneous rocks, for example when space problems occur
Hence, the structure will be in a dilatant orientation in resulting from folding at different wavelengths in adjacent
which perm eability is enhanced according to the law o f layers (Cox et al., 2001), or when rigid bodies are deformed
effective stress (eq. 3), and in which tensile failure may be in a soft matrix, such as metagranite in pelites or calc-silicate
favored in contrast to shear failure on o th er parts o f the rocks (Oliver et al., 1990; see below). Large fluid-filled cracks
structure. Misorientation of this type may occur in response may be generated along rock boundaries, giving rise to large
to fault bends, fault-tip bends, fault linking, and fault refrac­ variations in Pf and, hence, a strong hydraulic driving force.
tion through different rocks (Holyland and Ojala, 1997; The large fluid pressure gradient may allow spalling o f the
Cox e t al., 2001). T h e m ore general case is one in which stronger rock into the dilatant zone created by the strain vari­
deform ation o f rock packages containing layers a n d /o r ation, thus, triggering brecciation and providing potential
blobs o f strong rocks in weaker matrix rocks causes local­ sites for mineralization (e.g., Phillips, 1972). The required
ization o f stress and strain, with rock boundaries in orien­ amounts o f overpressuring leading to fracture, and conse­
tations a t a high angle to the m aximum principal com­ quent underpressure leading to spalling and brecciation, will
pressive stress being the first to show particularly dilatant be far more readily attained along such boundaries than in
behavior (see below). more homogeneous rock bodies.
In general, it is difficult to separate an observed bulk
Perturbation ofstress and strainfields across rock boundaries strain gradient from an inferred stress gradient, even
/ Regions o f substantial dilatancy are required for the devel­ though the latter is harder to quantify. Strong bodies within
opm ent o f very large fractures in deep-seated lithostatically weaker materials tend to act as stress guides. W ithin the
pressured rocks. In rocks otherwise deforming in a largely stronger body, and in specific areas around it, both the ori­
ductile manner, dilation and fracturing may occur across entation and magnitude o f the principal stresses change rel­
areas o f very large strain gradients. Strain incompatibilities ative to the far-field o r applied stresses (Edwards, 1951;
Stromgard, 1973; Hobbs et al., 1976, figs. 1.6,1.7; jaeger and
Cook, 1979; Oliver et al., 1990; Ojala et al., 1993; Ord and
Oliver, 1997). Similarly, local stress varies significantly
around openings o r planes of weakness, such as cracks,
faults, tunnels, etc., an effect that has been well documented
in the mining engineering literature (see Jaeger and Cook,
1979 and references therein), and has equal application in
geology (e.g., Etheridge et al., 1984). These local variations
in stress can lead to failure in orientations and magnitudes
not predicted by simple analysis o f the average far-field
stresses away from the perturbation (Jiang et al., 1997).
Numerical Models
In recent years, numerical models for deformation and
fluid flow have been used to retrospectively demonstrate
reasons for fluid flow localization around rigid bodies and
related ore deposits (Oliver et al., 1990, 1999), to explain
broad variations in fluid flow in m etam orphic environ­
m ents (O rd and Oliver, 1997), and to take results from
known geometric relationships around ore deposits and
apply these to exploration in the surrounding areas (Ojala
et al., 1993; Holyland and Ojala, 1997; Mair et al., 2000).
The concepts are based primarily on analogue (e.g.,
Stromgard, 1973), finite elem ent, and finite difference
models of the state of stress within and around strong bod­
ies in a weak mauix, which allow for the prediction of areas
that may undergo tensile or shear failure. Such modeling,
however, is not a requirem ent for understanding the prin­
ciples of fluid flow in heterogeneous rocks. The models are
an aid to the logical process of identifying and characteriz­
ing key controls on fluid flow, and the process o f con­
structing and modifying a series o f num erical m odels is
akin to a risk analysis and refinem ent o f exploration crite­
ria. Knowledge o f the principles alone can lead to greatly
improved strategies for extension o f operating mines and
exploration for structurally controlled ore deposits.
UDEC
Two commercially available numerical codes have been
used by us and by colleagues in recent years for modeling
o f fluid flow in deform ing rock masses, and examples of
both are presented here. T he first o f these, UDEC (Uni­
versal Distinct Elem ent Code; Lemos et al., 1985; Starficld
and Cundall, 1988), is a two-dimensional, finite difference
m odel that does n o t fully couple deform ation and fluid
flow throughout the rock mass. A two-dimensional body of
rock is treated as a group o f separate, fully deform able
blocks that interact a t edge and co m er contacts. Internal
deform ation o f blocks is sim ulated using time-stepping
finite difference calculations, with specified block internal
properties. Constitutive properties o f contacts can also be
specified. Fluid flow through block interiors is not simu­
lated. Rather, UDEC m odels stress distributions ("stress
m apping”) in rock masses deform ed a t low bulk strains
(<5% bulk shortening), and emphasizes fluid flow along
predefined fractures, faults, jo in ts, and rock interfaces
(Holyland and Ojala, 1997). A three-dimensional version is
available (3DEC), b u t is n o t presented here.
UDEC is ideal for m odeling rocks that show predom i­
nantly brittle behavior and fluid flow channelized along
near-planar fractures. These fractures or interfaces between
rocks have their properties specifically defined in the mod­
els, so the fracture properties and the input geometry, thus,
control the fluid flow to a large degree. UDEC uses an
explicit calculation procedure. By this method, each calcu­
lation cycle covers a small enough time step that it is not
necessary to consider the spatial and temporal interdepen­
dence of variables in the model. This removes the need for
m anipulation o f large matrices, and allows simulation of
modest strains. However, small timesteps are required to
m aintain model stability, and UDEC has problem s with
complex grids and block overlap. Hence, the models can
only be run to relatively small total strains, because there is
no capacity for re-gridding during model inns which would
alleviate this problem. Its utility is, thus, restricted unless it
can be demonstrated from the geology that the configura­
tion o f rocks and deform ation fields at the time of fluid
flow did not change greatly (i.e., did not accumulate large
strains), and details of these assumptions for some scenar­
ios have been presented by Holyland and Ojala (1997).
Within these provisos, models of stress distribution, derived
from UDEC modeling, have been used to constrain paleo-
orefield fluid flow, even in rocks showing brittle-ductile
behavior. It is particularly useful if it can be demonstrated
that tire bulk o f fluid flow was along meso- to macro-scale
fractures, because UDEC treats the intervening rock mass
between the fractures as essentially impermeable. In this
regard it is not fully coupled even though the fluid obeys
Darcy’s law in the fractures, and this is the m odel’s main
limitation. However, it has some utility in rock masses that
contain micro-, meso-, and macro-scale fractures, because
the stress distribution in the rock bodies between the spec­
ified fractures can also be viewed as having the potential to
develop fracture-related permeability if the stress condi­
tions are suitable (see below). Jiang etal. (1997) use UDEC
in scenarios in which the geometry and evolution o f chan-
nelways are specified and varied, and this analysis includes
appraisal o f fluid flow rates and conduit width, critical in
specific localization o f m ineral deposits (Helgeson and
Lichtner, 1987).
FLAC
Fl-AC (Fast Lagrangian Analysis o f Continua; Cundall and
Board, 1988) is a finite difference code for m odeling
deform ing porous media; that is, it treats rock masses as
though they are continua represented by average values o f
mechanical, fluid flow, and heat transport properties. Fluid
flow in FLAC is more completely coupled with the deform­
ing rock mass than is the case with UDEC. Unlike UDEC,
FLAC can handle large strains, rotations, and displacements
both in two o r three dimensions, although complex simula­
tions using FLAC30 require long run times. We have favored
FLAC in recent years because most o f the deposit and
regional scenarios we have been considering involve fluid
flow during ductile and brittle-ductile deformation in which
flow occurs in rock bodies, as well as along rock boundaries.
 In this model, the fluid obeys Darcy’s law such that fluid
flow is a function primarily of gradients in hydraulic head
and permeability (eq. 1). Permeability can be fixed for
given rock types or allowed to change as a function of the
deform ation. In hydrogeological studies o f clastic sedi­
m ents dom inated by prim ary porosity (space between
clasts), porosity-permeability relationships are dominated
by expressions that focus on grain shapes, size distributions,
and pore space shapes (Domenico and Schwartz, 1998).
Such approximations are inappropriate for rocks in which
pore space is dom inated by secondary porosity, and in par­
ticular, dynamic porosity that arises from deformation. In
FLAC, the porosity has two components, one being a FLAC-
defined porosity that is rock-type dep en d en t b u t strain
independent; the second is deform ation related, being
related directly to the volume change occurring during
deform adon (O rd and Oliver, 1997). The volume change
during deform ation (and the linked porosity change) is
conceptualized by the diladon angle (Fig. 2). The increase
in pore volume during plasdc deform auon arises from the
sliding o f irregular surfaces past each other, such as along
grain boundaries and fractures surfaces (Brace, 1968; Ord,
1991). The assumpdon is then made diat this deformation-
induced dilatancy is duplicated at broader scales within
specified volumes o f rock, generally corresponding to dif­
ferent rock types. Common rocks such as sandstones and
marbles have dilation angles on the order o f +5° to +10®,
whereas some strong rocks (e.g., gabbro, skarn) may have
larger angles, and, less commonly, highly porous limestone
and similar rocks may have negadve diladon angles. Dila­
don angles may change during deformation; for example,
decreases may occur due to gradual sealing o f deformadon-
induced porosity by the reaction products o f protracted
fluid/rock interaction, such as vein infill or altcradon zones
: (see below).
Volume changes due to deform auon o f dilatant rocks
result in pore pressure changes, causing changes in the
hydraulic head and, hence, fluid flow according to Darcy’s
law. Changes in effective stress owing to volume change can
also lead to plastic deform ation, which leads back to
changes in volume. Feedback between fluid flow and defor­
m adon, thus, continues. These fundamental principles are
easiest to appreciate in the case o f intensely altered ductile
: shear zones, where micro-scale dilatancy, coupled with
localized shear, causes repeated fluid pressure changes
leading to concentrated ingress and flow o f fluid along the
shear zone (Cox e t al., 2001; Robert and Poulsen, 2001).
In the rem aining sections, we use com binations of
mapped field o r m ine relations, and either UDEC o r FLAC
models, to dem onstrate the principles o f fluid flow that
have led to ore accumulation.
UDEC stress modeling of the Mary Kathleen fold belt
Geologic background: O ur first model is based on the geol­
ogy o f the Mary Kathleen fold belt in the central Mount Isa
block (Oliver and Wall, 1987; Oliver et al., 1990; Oliver,
Fk;. 2. Simplified microstructural interpretations o f the dilation angle
o f a material, for granular porous media (a), and for material with per­
meability dominated by microcracks (b). showing how shearing can result
in increases in volume and potential for fluid inflow. For granular mate­
rials, the porosity generated by deformation may dominate over the pri­
mary intrinsic porosity, as depicted. In the FLAC models presented in the
text, the deformation-related porosity term is characteristically predomi­
nant over intrinsic porosity. Adapted from Ord and Oliver (1997).
1995). H ere, com petent, low perm eability m etadolerite
and metagranite bodies were surrounded by weak calc-sili­
cate rocks during syn-metamorphic fluid flow. Fluid flow in
those rocks o f the ~1760-Ma-old Corella Form ation
occurred at 1550 to 1500 Ma during the Isan orogeny, and
was strongly concentrated around the metadolerite-calc-sil-
icate interfaces. T he distribution o f alteration and veins
suggests that the m etadolerites acted as stress guides for
deformation and fluid flow (Fig. 3a, b). T he abundant brit­
tle and brittle-ductile structures within these alteration
zones are atypical o f normal amphibolite-facies rocks, and
are m ore characteristic o f deformation a t much shallower
levels in the crusL However, vein crosscutting relationships,
textural observations, and P-T data confirm that the laqje
majority o f alteration products and veins developed near
the peak o f regional amphibolite-facies m etam orphism
(Oliver, 1995). Mineralization (Cu, Au, U-REE) in the belt
is widespread and spatially related to the alteration. How­
ever, it is fairly consistently o f low tonnage, and with a few
exceptions, uneconom ic for metals. Links have recently
been drawn between this belt and the Q oncurry Belt some
40 to 60 km further east, where much richer Cu-Au miner­
alization occurs (Oliver et al., 1999).
Alteration and m inor mineralization is found near the
boundaries o f the meta-imrusive rocks and the metasedi­
ments, characteristically form ing calcite vein networks (±
chalcopyrite) surrounded by irregular or shear-related
albitic alteration zones (Fig. 3b). Large vein crystals of
clinopyroxene, amphibole, and calcite, locally up to meter
scales, suggest that there were substantial dilations during
vein formation.
 approx, scale boudinaged
altered beds

t N
i
TvT;
r [y \v v 0
«,/ S/ S/ N/
z p
\
/
\ X\
/ / / 30 M Pa
0 s
/
X
/
X
/
X
/
D
A W
0 K A V D

c D

200m T5 M Pa

200m
Fig . 3. Geology and UDEC m odeling o f the Mary Kathleen fold belt. a. Schem atic three-dim ensional geology o f the
Mary Kathleen fold belt show ing how distribution o f infiltrated, altered rocks (shaded) was controlled by deform ation
(E-W sh o rten in g ) o f co m p eten t m etadoierite bodies in calc-silicate host rocks (adapted from Oliver, 1995). b. Som e
detail in m ap view o f the distribution o f alteration products and the inferred m axim um principal stress direction d u r­
ing sh o rte n in g deform ation (m odified from O liver e t ai., 1990). c. C onceptual in p u t for two-dimensional, plan-view
UDEC m odel showing applied stresses on a co m p eten t m etadoierite body within w eaker calc-silicate rock (rock p rop­
erties from O liver c t al., 1990); lateral boundaries rem ain p lan a r d u rin g m odel deform ation but are free to change
length and move E-W; no fluid flow occurs, d. C ontoured effective m ean stress (MPa) g enerated after -5 p e rce n t E-W
shortening. N ote correspondence o f zones of low effective m ean stress (shaded) with field occurrences o f intense alter­
ation a n d veining (i.e., a “stress guide" effect).
 Model construction and parameters: Although several geo­
metric situations can be applied to specific m apped alter­
ation zones (Fig. 3a), the overall patterns of fluid flow may
be m odeled by deform ation of a tabular meta-intrusive
body surrounded by metasediments, and subject to short­
ening orthogonal to the elongate boundaries of the tabu­
lar body (Oliver et al., 1990). O ur first approach is to use
UDEC to determ ine w hether there is a good correlation
between the distribution o f the altered rocks and the stress
field operating over small deformation increments (Fig. 3
c, d). Physical properties o f rocks were obtained from the
tabulations o f Birch (1966), and all rocks were modeled as
Coulom b materials. Rock properties are presented in
Oliver et al. (1990) and are similar to those given in Table
1; data for the m odel m etadolerite are those for Sudbury
diabase, and the m etasedim ent is based on a m ixture o f
Framingham schist, Idaho Springs Formation schist, and
Cockseyville m arble (Oliver e ta l., 1990). The calculated
Young’s modulus ratio o f Sudbury diabase to this metased-
imentary aggregate is 2.4, consistent with the strain shadow
patterns observed around the meta-intrusive bodies in the
Mary Kathleen fold belt. Deformation in both bodies was
taken as elastic-plastic, and the applied stresses in the X and
Y directions (Oj and <r2) were 30 MPa and 5 MPa (300 bars
and 50 bars), respectively. The total area is divided into 200
finite-difference triangles (sub-blocks), with the appropri­
ate properties applied to the m etadolerite, the metasedi-
ments, and the boundary between the two rocks (Oliver et
al., 1990). The finite-difference triangles interact during
the simulation, from which an array of mean and differen­
tial stresses is com puted at intervals during the run. The
methods are generally the same as those presented by Holy-
land and Ojala (1997).
Model results, implications f o r exploration, and limitations:
The model shows good correlation with the observed dis­
tribution of alteration products, and suggests that low mean
(and differential) stresses were important for focusing flu­
ids. However, we cannot assess the realism o f the magnitude
o f the generated stresses because of limitations o f model
inputs. The model shown was run without fluids, and there­
fore the effective stress shown has arisen by our application
of low confining pressures on the block boundaries (Fig.
3c), to simulate artificially a total mean confining pressure
of 350 MPa (low pressure am phibolite facies) and an
applied differential stress (assumed) in the horizontal
dimension of 25 MPa. The m odel is, thus, not fully cou­
pled, nor three dimensional, so we cannot portray realistic
flow vectors because Darcy’s law (eq. 1) and the law of effec­
tive stress in the presence of fluid (eq. 3), are not incorpo­
rated. Despite these model restrictions, the extreme varia­
tion in rheology o f the m apped rock types produces (a)
large gradients in differential and mean stress (Oliver et al.,
1990), and (b) zones of high inferred fracture permeabil­
ity in orientations that would normally not be considered
suitable for dilation, particularly along the walls of the
intrusiye bodies aligned north-south, at right angles to the
maximum compressive far-field stress.
Matrix Block
Shear modulus (CPa) 1.0 5.0
Bulk modulus (CPa) 1.67 8.33
Density (k g /m 3) 2,700 2,700
Cohesion (MPa) 10.0 10.0
Friction angle 30® 30®
Dilation angle 10® 10®
Permeability (m-) lO e ll 5cI3
In an applied sense, the results o f the modeling allow us
to equate a regional deform ation with the localization of
alteration products, rather than requiring a specific defor­
mation phase o r change in the stress field, o r other defor­
m ation conditions. Secondly, the models show th at the
stress guide effect is sufficiently well correlated with the
shape and size o f the more competent, less permeable bod­
ies, that exploration for large deposits is m ore likely to be
successful when the stress guides arc large. In the Mary
Kathleen fold belt, this situation corresponds to 10-km-scale
metagranite + skam bodies, rather than the smaller meta­
dolerite bodies depicted in Figure 3. The Mary Kathleen U-
REE deposit, for example, is localized on such an interface
(see below). Finally, the m odels suggest that all o f the
observed vein and pod geometries (Fig. 3a) can be ratio­
nalized as being a product o f east-west shortening, includ­
ing steeply dipping, northerly oriented altered shear zones
along interfaces between the two rock types. The effect of
the different rock properties may be sufficiently strong to
perm it favorable low stress conditions for rock failure and
fluid flow in locations and orientations that would not read­
ily be predicted from an inspection o f the geometric con­
figuration relative to the shortening field.
A saturated, continuum modelfor the Mary Kathleenfold belt
The limitations o f UDEC do not readily allow us to test
conceptual ideas regarding fluid flow in general. In this sec­
tion, we briefly present numerical models that assist with an
understanding o f how variable strains and fluid fluxes
might influence the patterns predicted in Figure 3. UDEC
cannot fully couple deform ation with fluid flow through
the rock masses, so what is the effect o f fluid flow on the
deformation patterns? And what is the effect of having ver­
tical gradients in initial fluid pressure distribution, as would
be expected for deforming, metamorphic rocks?
Model construction and parameters: Figure 4 shows some
simple two-dimensional cross-section models, which ini­
tially are 5 km wide and 5 km deep, and contain a rectan­
gular block, 2 km wide, 1 km deep, and centrally placed.
Because o f o u r fluid pressure param eters and boundary
conditions (see below), similar results would pertain if the
larger block were made 20 by 20 km, so that we could sim­
ulate the amphibolite facies depths pertaining in the Mary
Kathleen fold belt. Elastic constants and plastic Mohr-
Coulomb properties have been selected for representative
 i i i i
0 1 2 3km
30% e, 30% e,
10% e, low fluid flux low fluid flux high fluid flux
Fig, <1. FLAC m odel o f ih e effects o f different fluid flux nucs and total strains o n die pattern o f fluid flow and defor­
mation in a two-dimensional cross-section o f heterogeneous rocks, a. T he conceptual m odel for input into FLAC is
based loosely on the geometry o f rocks in the Mary Kathleen fold belt (Fig. 3 ), but could easily be applied to any situa­
tion o f com petent rocks em bedded within weaker surrounds. Initial conditions (Table 1) arc: (1) two rock types, the
central block hating low permeability and moderate strength, and the surrounding rock hating low strength and m od­
erate permeability; (2) a lithostauc fluid pressure giadicnt (giving overall upwards flow; (3) the lithostatic fluid pres­
sure gradient is always enforced along the latcial boundaries and at the base o f the m odel to ensure fluid flow for the
duration o f the model; (4) lateral boundaries are free to move horizontally and vertically but must remain planar, sim­
ulating a general shortening strain; (5) the base can shorten but must remain planar, and the free top surface is open
to potential hydrostatic conditions. Rock property data are given in Orel and Oliver (1997). b. FLAC finite difference
m odel in two dimensions o f instantaneous fluid flow vectors after approximately 10 percent shortening (e) o f the orig­
inal block, showing the subtle effects o f the central block on the fluid flow vectors body, for a situation with 1 fluid flow
step per deformation step (equating to “typical" regional mctamorphic fluid fluxes), c. Same parameters except at 30
percent shortening, and in comparison to (b), note the strong fluid focusing and the developm ent o f a region o f sub-
lithostatic fluid pressure conditions corresponding to the draw-down vectors in the top center o f the model, d. Same
conditions as for (c), but with 100 fluid flow steps per deformation step, equating to fluid fluxes similar to those in con­
tact aureoles and metasomatic shear zones, showing how the high fluid fluxes can dampen the effect o f the deforma­
tion. Note that die draw-down effect shown in (c) is also swamped by the very high upward fluid fluxes. Adapted from
O rd and Oliver (1997).

rock types from Birch (1966; Table 1). The bulk modulus
for the elastic fluid in all models is 2 GPa and the fluid den­
sity is 1,000 kg/m 3 for fully saturated fluid flow. The mate­
rials follow a non-associated Coulomb constitutive law (Ord
and Oliver, 1997) with individual rock bodies having spec­
ified values of cohesion, friction angle, and dilation angle
(Table 1). In this example all properties for each rock stay
constant throughout the experiment; some properties are
allowed to vary in the M ount Isa model presented below.
The finite difference sectors within the models are initially
100 m across, and as little as a third o f this after deforma­
tion. Thirty to one hundred meters is, thus, our represen­
tative volume across which we assume average fracture-
related perm eability and rock properties. Hence, these
models are only pertinent to the flow o f fluid at broader
scales than this, although the same principles would apply
to local, more specific situations.
Each model is brought to equilibrium under gravity prior
to horizontal compression under plane strain conditions.
The out-of-plane stress is incorporated in the calculations.
In order to attain these conditions, the base of the model
is fixed vertically but is free to deform in a horizontal direc­
tion, the sides of the model are constrained to be planar
during deformation but are free to move horizontally and

■ ■
to extend vertically, and the top surface is free to deform in
any m anner (Fig. 4a). The fluid pressure throughout each
model is initialized to lithostatic conditions prior to begin­
ning the experiment, giving an overall upward flow due to
buoyancy, but is constrained to remain at these conditions
only at the edges o f the m odel for the duration o f the
experim ent. The vertical stress is also initialized to litho­
static conditions, and with Poisson’s Ratio approximately
one quarter, the horizontal stresses are initialized to one
third the vertical principal stress, prior to shortening.
For this model and the M ount Isa model presented
below, gravity (10 m /s 2) is applied, and the m odel is
brought to an initial force-equilibrium state before being
compressed horizontally a t a rate of 0.2 length units per
time step, which for a model 5,000 units wide (500 sectors)
results in a horizontal shortening of 4 percent every 1,000
steps. Fluid flow histories can be tabulated for individual
points, or plotted for the entire grid, but here we have used
vectors of instantaneous fluid flow to provide the best visu­
alization (Fig. 4).
Results and implications: Comparison between Figure 4b
and 4c shows that increasing strain at a constant fluid flux
rate has the effect of focusing fluid flow more strongly. This
scenario is not particularly applicable to predom inantly
brittle situations, where faulting can focus fluid a t relatively
low strains (Cox e t al., 2001), but does appear to be impor­
tant in m ore ductile rocks such as those m odeled here.
Comparison of Figure 4c and 4d shows that fluid pays little
attention to rock heterogeneities if fluxes are very high rel­
ative to deform ation rates. Such a scenario would not be
particularly useful for concentration o f mass into an ore
accumulation, but it may be useful in the dissolution and
transport phases o f a hydrothermal cycle during which it
would be desirable for fluid to scour as much rock as pos­
sible. By comparison with the real geology (Fig. 3a, b ), we
can, therefore, surmise that at the present level o f expo­
sure, the stress guide effect, was particularly important, and
the volumetric fluid flux was sufficiently high to cause sub­
stantial alteration, but not so high as to swamp the effects
of the heterogeneous deform ation. Differences between
the shapes o f the m etadolerite block in Figure 4c and 4d
testify as to how the fluid-coupling via Darcy’s law can feed
back to the stress state sufficiently that the strain pattern is
changed; however, this appears to be a second-order effect
for the materials modeled.
FLAC modeling of the Mary Kathleen U-REE ore deposit
Geological background: Modeling of a specific ore geome­
try was applied in numerical studies of the Mary Kathleen
ore deposit (Oliver, 1995; Oliver et al., 1999). This unusual
deposit (12 million tonnes at 0.12% U3O s, 6-7% REE),
hosted by a garnet-pyroxene skarn, was mined from 1956
until 1982, and is located 55 km east o f M ount Isa and
approximately 65 km west o f Ernest Henry. The host gar­
net-pyroxene skarn form ed during an early phase o f gran­
ite intrusion (1740 Ma) into the Corella Formation (Page,
1983). T he orebody was form ed late during the Isan
orogeny at 1550 to 1500 Ma, during D2 a n d /o r D3 (Page,
1983). One model appeals to protore preconcentration of
U-REE in the skarns during granite intrusion (Maas et al.,
1988), whereas another proposes more distal mass sources
(Oliver et al., 1999). In both cases, the focusing mechanism
for the present ore distribution is thought to be related to
fluid focusing during the Isan orogeny.
Allanite-uraninite-garnet ore shoots form an irregular
vein array that steepens westwards as the large Mary Kath­
leen shear zone is approached (Fig. 5). The shear zone has
a clear west-block-up m ovement sense defined by strati­
graphic markers and shear-sense indicators. The veins,
thus, do not represent deform ed planar markers, because
their overall curvature would otherwise give an east-block-
up m ovement sense. Rather, it would appear that the
change in orientation is a function of distance away from
the shear zone, and so may be related to ihe coupling o f
shear with brittle fracturing in the adjacent rocks, because
field relationships also point towards a likely synchronicity
of mineralization and shearing (Oliver etal., 1999).
Model construction, parameters, limitations: To test the
extent to which vein localization and geometry was related
to adjacent shearing, numerical modeling was applied to a
geometry in which a hard skarn body was coupled with a
soft shear zone at die time of ore genesis (Fig. 5). The start­
ing conditions we chose were based on an understanding
of the regional and local kinematics (Fig. 5a), such that a
broader east-west shortening was locally modified by west-
block-up shearing in the late stages of the D2 deformation.
Rock property data are presented in Tabic 2. The bound­
ary conditions were sim ilar to those for Figure 4, except
that both the bottom and top boundaries were allowed to
change shape to accom m odate the m ore complex kine­
matics. A lithostatic fluid pressure condition was imposed
for the duration o f the runs, based on the assumption that
sufficient fluid was available from the progress o f regional
metamorphic reactions a n d /o r late syntectonic granites.
We did not run models with different fluid pressure condi­
tions, nor did we change the rock properties during the
runs. Also, the total strains at the end of the run (Fig. 5c)
were considerably less than inferred from the geolog)’—the
200 m west-block-up model displacement is much less than
the calculated >2 km displacem ent from the geology
(Oliver et al., 1999).
Results and exploration implications: Both UDEC (Oliver
1995) and FLAC models (Fig. 5d, e) produce a heteroge­
neous stress field in vertical section in which the orienta­
tion and magnitude o f the stresses changes systematically
across the host skarn body, mimicking to some degree the
measured orientations of most ore veins (Fig. 5b). To a first
approxim ation, this suggests that deform ation-coupling
between the shear zone and the adjacent strong skarn body
was the prim ary mechanical control on ore deposition,
both spatially and with respect to ore shoot orientation.
The response o f the skam to the immediately adjacent high
strain zone was to deform via tensile and shear failure on a
 M ary

14.0

14.4
Depth
(km)

14.8

____ ] At yield in shear (plastic)

FlG. 5. Geology and FLAC model o f the Mar)’ Kathleen U-REE
deposit, adapted from Oliver et al. (1999). a. Regional cross section H H H Elastic, yield in shear in past
showing buttress effect of Wonga Belt against the Mary Kathleen fold
belt, based in part on Holcombe et al. (1992). Because o f this 10-km- At yield in tension
scale buttressing, the broader Isan orogeny D.j E-\V shortening is per­
turbed in the vicinity of the Maty Kathleen Shear Zone such that the
local strain field is heterogeneous and includes a significant compo­
nent of west-block-up shear, b. Cross section of geology showing ore
shoots in host skarn, progressively steepening down and to the west as
the Mary Kathleen Shear Zone is approached, c. Conceptual model
used as geom etric input into the FLAC modeling, showing the 200 m
resolved components of the regional shortening as an imposed hori­
zontal east to west velocity on the eastern boundary, and an inclined
west to east and up velocity for the western boundary, based on our
inferences from (a) and (b). The Figure shows the geometric starting
condition and 2 snapshots of the numerical modeling (middle and
end of the run), showing weaker shear zone rocks to the west, hard
skarn in the middle (hosting later ore veins), and moderate-strength
and -permeability calc-silicate rocks and amphibolite to the east (rock
properties given in Table 2). d. Summary o f results for enlargement
of the final run, showing curved lines of the trajectory of <7,, varying in
a fashion similar to the variation in the geometry of the ore shoots,
and variation in the failure mode, which is a rough approximation of
the ore grades at broad scale. These correlations vindicate our inter­
pretation that ore shoots formed by coupling o f the shear zone with
the hard skarn body during deformation, e. Instantaneous fluid flow
vectors at the final model stages, showing that fluid may have been
drawn from west o f the skarns towards the areas o f tensile failure
within the skarns, and that the calc-silicate rocks to the east may not
have allowed easy fluid access into the skarns.

Marble H ornfcls/
(west o f MKSZ) amphibolite Skarn
Shear modulus (GPa) 1.0 4.0 14.9
Bulk modulus (GPa) 1.67 6.67 2-1.8
Density (k g /m 3) 2.700 2.700 2,700
Cohesion (MPa) 10 50 500
Friction angle 15® 20® 30®
Dilation angle 10® 5®
25®
Permeability (m2) 2c-16 lc -1 6 lc-1 6
MKSZ = Mar>' Kathleen shear zone
set of planes systematically related to the stress field gener­
ated by this coupling o f hard rock against soft rock. This
effect has enabled fluid to be drawn in towards the skarn
from the west, and to a lesser extent from the east (Fig. 5e).
At the bottom western side o f the skarn body, the correla­
tion between the stress field and the observed vein orien­
tation is not so good, and this could be a function o f the rel­
atively low shear strains we have imposed for the modeling,
in comparison to reality. Alternatively, ore shoots near the
shear zone may have been deformed and rotated towards
m ore vertical orientations by the shearing, possibly by
expansion of the shear zone with time.
At ore-field scale, these model predictions would be use­
ful in controlling orebody extension planning, because the
parts o f the skarn away from the m ain shear zone are
unlikely to bear good grades unless other shear zones were
found. The m odeling also confirms that the ore shoots
were not preexisting veins that were deform ed by the Mary
Kathleen shear zone. In local exploration, large skarn bod­
ies adjacent to large shear zones would clearly be possible
targets, and in regional exploration, the results confirm the
propensity for localization o f ore deposits along boundaries
between the hardest and softest rocks available in the
regional setting (see below). Although this last observation
is hardly surprising, the models also provide a mechanism
for drawing fluid from the surrounding rocks into the vicin­
ity of the skarns. The protore model for orebody formation
(Page, 1983a; Maas e ta l., 1988) appeals to preconcentra­
tion o f U-REE in the skarns, but the model does not clearly
explain the low eNd signature and LREE-enrichment of the
orebody (Maas et al., 1988; Oliver et al., 1999). The alter­
native m odel presented here proposes external fluid
ingress from sources west o f the Mary Kathleen shear zone,
which would help to explain these geochemical signatures.
UDEC stress mapping of the Hilton Pb-Zn-(Cu) ore deposit
Geological setting: The Hilton Pb-Zn ore deposit is one of
a family o f well-known giant sediment-hosted base metal
sulfide deposits in the western Mount Isa district. Situated
some 40 km north o f M ount Isa, it differs from that deposit
m having far less copper. However, the lower grades and
i esuicted distribution of copper has allowed Valen ta (1994)
and Valenta et al. (1994) to define and refine a structural
m odel for syntectonic copper m ineralization which also
applies at Mount Isa (see below). Rocks in the Hilton area
are part of the - 1 670-Ma-old Mount Isa Group of the West­
ern Succession o f the M ount Isa block. M ineralization
occurs within impure dolomitic siltstones o f the Urquhart
Shale. T he structural history o f the area comprises early
thrusting (D{), regional north-trending folding (D2), and
late faulting and development of fold zones in various ori­
entations (D3-D 5). Rocks in the Hilton area lie on the west
limb of a large scale D2 anticline, contain a weak but pene­
trative $2 cleavage, and have undergone intense fauldng
and block rotation associated with the D«, to D- deforma­
tion episodes (Valenta, 1994).
Mineralization in the Hilton area can be divided into two
broad types: (1) stratiform lead-zinc mineralization occurring
in numerous lenses associated with strongly pyritic shales; and
(2) transgressive copper mineralizadon occurring within
lead-zinc orebodies and crosscutdng dolerite dikes in the
Hilton area, generally associated with deformed zones that
clearly postdate lead-zinc mineralizadon. Relationships
between copper mineralizadon and structural features have
shown that copper mineralizadon occurred mainly during
the D2 and D3 deform adon episodes, associated with late-
stage fluid circulation on the major fault zones in the area
(Valenta, 1994). In the Hilton mine, the copper-rich zone is
associated with an area of high strain within the three hang­
ing-wall lead-zinc orebodies where they are cut by a vertical
extensional fault adjacent to a precopper dike (Fig. 6a).
Model construction and parameters: The model used in this
study simulates behavior of rocks in the Hilton mine block
during syntectonic copper mineralizadon (i.e., late D2-earIy
D3). Modeling is performed on an east-west vertical section,
in order to test the relative contributions o f fracture and
boundary geometry and sulfide-doleritc-shale rheological
contrast to fluid flow and copper mineralization patterns
inferred from field observations.
Shear zones separate the rocks along most contacts. The
two largest bounding shear zones are steeply west-dipping,
and represent the Paroo fault zone on the west side and the
Barkly shear zone on the east. Six m inor shear zones, with
slightly su onger mechanical properties, occur between the
bounding shear zones. These represent m inor bedding-
parallel shear zones, and also serve as contacts between
shale and lead-zinc ore. Shearing also occurs on the hang­
ing-wall of the dike (Fig. 6a). Values of bulk modulus, shear
modulus, density, cohesion, friction angle, and tensile
strength have been estimated based on data reported in
Birch (1966), and are shown in Table 3. Dike rocks are con­
sidered to be the strongest, followed by barren shales. Lead-
zinc orebodies are the weakest model lithology. Estimates
o f normal stiffness, shear stiffness, cohesion, friction angle,
and tensile strength for m ajor and m inor faults are also
shown in Table 3. Low shear stiffnesses have been estimated
for faults in the m odel, based on the fact that faults at
Hilton are commonly wide zones o f mylonitization, poly-
clinal folding, and cataclasis (Valenta, 1994) rather than
simple rough joints. Values o f cohesion and tensile strength
have been set at zero for the same reason.
The constitutive model for internal deform ation of
blocks is elastic/piastic, with Mohr-Coulomb failure. Vein
 Lead-zinc Major Minor
Shale sulfides Dike faults faults

Dcnsilv (fcg/m*) 2,850 3.500 2,850

Bulk modulus (Fa) 40e9 30c9 5Ge9
Shear modulus (Fa) 20e9 I5c9 25e9
Tensile strength (Fa) 3e6 I.5c6 4e6 0 0
Cohesion (Pa) 4e6 2c6 4e6 0 0
Friction angle (can 0 0.6 0.4774 0.7 0.4 ; 0.5
Normal stiffness (Fa) 2.5c9 3.5e9
Shear stiffness (Fa) Ic2 2e2
Initial aperture (m) 0.05 0.03
Residua! aperture (m) 0.005 0.003

and shear zone orientations (Valenta, 1994) show that <7, dipping fault/dike system. The models have shown that <7,
was roughly subhorizontal and east-west at this time, and Oj rotated into a steep east-plunging orientation, parallel to
was subvertical. A horizontal compressive stress o f 20 MPa the dike margins (Figs. 6c, d), inferred to be a function of
was applied on the side boundaries, whereas stress on the the greater strength o f the dike. The resultant low normal
top and bottom boundaries was set a t 0 MPa. As for the stresses on jo in ts in this zone will cause dilatancy and an
Mary Kathleen fold belt UDEC model (Fig. 3), stresses this increase in jo in t aperture, leading in turn to fluid focusing
low were used to simulate the effect o f high pore pressures in the dike area.
within the blocks, according to the Law o f Effective Stress 2. Fluid m ust be focused into the intersection between
(eq. 3). Each timestep in the model corresponds to approx­ the Paroo fault zone and the Dyke Trace fault zone. The
imately .0025 percen t bulk shortening; 3,000 timesteps, models show that this is a zone o f large variation in m ean
thus, equates to only 7.5 percent shortening. and differential stress. Zones o f high and low stress migrate
and sometimes reverse during the evolution o f individual
Results: During model runs, the maximum principal stress model runs (R.K. Valenta, unpub. data). This would favor
is rotated into perturbed orientations around the dike, and alternate fluid inflow and outflow, accelerating the alter­
a zone of generally low to moderate differential stress and ation and m ineralization process by repeated fluid infil­
high m ean stress develops adjacent to the dike a t early tration.
stages (Fig. 6c, e). This zone developed as a result of the rhe­ 3. Fluid m ust m igrate into the orebody horizons. The
ological contrast between dike rocks, and the shales and Pl> area around the dike displays high mean stresses and low
Zn orebodies. It is likely that increased fracture permeabil­ differential stresses relative to the rest o f the m odel. A
ity and the low differential stress may have drawn fluid into result o f this will be that fluid migration outside the mzyor
the dike, even though the mean stress was higher than that shear zones should be predom inantly into the orebody
in die surrounding rocks. At later stages in the model (Fig. zones. Permeability gradients are provided by the rheolog­
6f), mean stress drops but remains relatively high at the dike ical contrast between the orebodies and interlayered shales,
tips, whereas the differential stress is lowered even further which causes brittle behavior in the shales.
(Fig. 6d). These model outputs are all consistent with the
intensity of alteration observed in the dike, and support the The models of both Hilton and the Mary Kathleen fold
inference that the perturbed stress field within and around belt indicate that variadons in mean and differendal stress
the dike has allowed high permeability even though the of up to an o rd er o f m agnitude can be expected in a
dike has accumulated high mean stresses. faulted, lithologically heterogeneous pile. This has m«yor
Copper mineralization is also strongly spatially associated implications for syndeformational fluid circuladon
with the intersection between the Paroo fault zone and the processes at various scales. Principal stresses can be reori­
dike-east-dipping fault system (i.e., the upper dike dp, Fig. ented by up to 90° from a given far-field stress. This obser­
6f). In the mine, there is a major zone of intense veining in vation also has implications for fluid flow, and places a cau­
this area. This observation requires a combination of high tionary note on acceptance o f regional stress orientadons
fluid pressure and low differential stress, or a local reori­ based on m easurem ent of local vein geometries.
entation of the principal stresses. The UDEC model shows
this is a zone of high mean stress but very low differential Model limitations: There are a num ber o f problems asso­
stress (Fig. 6e). If fluid pressures were high enough, it ciated with blind acceptance o f models o f the type pre­
would be possible to form veins perpendicular to the max­ sented in this paper. These include edge effects, zone gen-
imum principal stress in this zone. eradon effects, and problems with input parameters. We
are not trying to recreate the whole deform ation history,
Implications: The implications are as follows: but only the stress patterns associated with a short ume slice
1. Large volumes of fluid must be focused into the east­ during Cu mincralizadon. In this regard, we are uncertain
if the faulting along the contacts resulted in sufficient geo­
metric changes that the stress field was strongly perturbed.
Minimal offsets are shown in Figure 6c to 6f because large
offsets and high strains cause block overlap and crashing of
the program.
It can be seen in all stress plots that the least principal
stress decreases dramatically near the upper and lower
boundaries of the model, and that the corners bend in
towards the end of the run (Fig. 6d). Both these effects
arise because o f the low confining pressure on the top and
bottom boundaries which were implemented to artificially
simulate high fluid pressures (and low effective stresses).
The best way to avoid these problems is to make the dimen­
sions o f the block much greater than the area o f interest.
In most of the plots presented in Figure 6, it can be seen
that edge effects do not extend into the area of interest. In
some cases (e.g., Fig. 6f) it can be seen that model-gener­
ated stress build-ups o r shear zones appear outside the
bounding faults. These initiate on discontinuities in the
finite difference zones, produced by irregular block geome­
tries. It is difficult to avoid this problem, and the effects are
generally minor.
The fluid flow code in UDEC is limited by the fact that
fluid flow can only occur along block contacts, and flow is
not allowed through block contacts with a normal stress of
zero. The latter constraint severely limits the usefulness of
the code for simulation of fluid flow in models involving
localized dilation. Nevertheless, an attempt has been made
to infer fluid flow patterns by combining model-generated
flow with inferences from stress and permeability patterns.
FLAC modeling of the regional context of the
Mount Isa copper deposits
Geological setting: Perkins (1984) and Swager (1985)
caused controversy when they published data and models
for syntectonic origins for the sediment-hosted M ount Isa
copper deposits, but these models arc now widely accepted.
In detail, as at Hilton, the local structural controls are very
complex, and although the general paragenesis o f copper
overprinting preexisting lead-zinc orebodies is apparent,
the relationships are obscured by the size of die copper ore-
bodies, the intensity o f the surrounding silica-dolomite
alteration, and a spatial offset from the lead-zinc ores.
A nother controversy concerns the source o f the copper,
with the most recent model (Heinrich et al., 1995; Waring
et al., 1998) preferring regional sources rather than remo-
bilization from preexisting stradform Pb-Zn-Cu ore. At 1 to
50 km scale, the deposit is situated at die complex interface
between the Mount Isa Group shales, o f low metamorphic
grade, and a complex basem ent com prising the Eastern
Creek Volcanics (upper greenschist to amphibolite facies)
and the older Kalkadoon-Leichhardt block, exposed to the
east (Fig. 7a). Chalcopyrite mineralization and related sil­
ica-dolomite alteration form ed during the Isan orogeny,
under greenschist facies conditions, and involved m eta­
morphic, basinal, a n d /o r meteoric fluid sources (Heinrich
et al., 1995). The copper orebodies are hosted predom i­
nantly by the U rquhart Shales, and lie adjacent to but
deeper than the lead-zinc orebodies (Fig. 7a). The copper
deposits sit adjacent to the Paroo fault, the same fault sys­
tem that extends northwards to the Hilton mine. Copper
deposits do n o t extend westwards o f the fault.
Model construction and parameters: Unlike at the Hilton
mine, we are not testing specific local controls on copper
mineralization; rather, we are concerned with the broad
regional controls that may have led to focusing of copper
into the present site. The conceptual model we are propos­
ing involves east-west shortening o f the host sequence dur­
ing the Isan orogeny, and fluid flow via regional deforma­
tion-enhanced permeability. In this regard our models are
a test o f the physical aspects o f the Heinrich et a!. (1995)
model, which involves syntectonic extraction and concen­
tration of fluid and mass from external sources into the pre­
sent location of the orebodies in the Mount Isa Group.
A simplified geometric model was developed for a pre­
deformation half-graben geometry with the Eastern Creek
Volcanics sitting within Knlkadoon Basement rocks (O’Dea
et al., 1996; Betts and Lister, 2001), all o f which are overlain
by the M ount Isa G roup and a cover sequence o f sand­
stones, which have since been eroded (Fig. 7b). The
bounding fault o f the half-graben is the Paroo fault, which
inverts during shortening. In reality, the fault is a complex
fault zone with a protracted movement history (Fig. 7a). In
models presented here, the fault acts purely as a bounding
fault rather than a conduit or fluid aquitard. The models
are pushed with a constant velocity from die left (west) with
this boundary able to move vertically b u t constrained to
remain vertical. T he base is fixed vertically but is allowed to
move horizontally. T he eastern boundary is fixed in the
horizontal direction but allowed to move vertically. The top
boundary is allowed to move in any direction. These model
constraints correlate with the inference from the geology
that the Kalkadoon-Leichhardt basement below and to the
east acted as a rigid buttress d u rin g the Isan orogeny
(O ’Dea et al., 1996; Betts and Lister, 2001).
An elastic-plastic rheology, based on a non-associated
Mohr-Coulomb yield criterion and flow law, was used to
describe the mechanical behavior o f all rock types in the
models, and fluid flow obeys Darcy’s law, as for die earlier
FLAC models described. Rock properties are presented in
Table 4. Fluid pressures in the model were allowed to vary
from hydrostatic to greater than lithostatic, with an initial
hydrostatic fluid pressure regime. Unlike in the Mary Kath­
leen models, all boundaries except the top are impermeable
to fluid, and fluid is permitted to recharge from the surface.
This aspect o f the model corresponds to the shallow geo­
logic setting, the likely low permeability and fluid produc­
tivity of the basement rocks, and the likelihood that shallow
fluid sources (basinal, meteoric) were present during defor­
mation in addition to those produced during low rade
regional metamorphism of the Eastern Creek Volcanics and
the Mount Isa Group (Waring et al., 1998). The FLAC mod­
els calculate pore pressures and the instantaneous fluid flow
vectors. We have also calculated the total am ount o f fluid
that goes into or out of a cell during the run time o f the
 Paroo-
Mt Isa

A A 7\
10 km 1 0 km

FlCi. 7. Geology and FLAC simulations o f the district around the M ount Isa copper deposit, a. Simplified cross sec­
tion o f the western Mount Isa terrain, showing position o f copper orebodies in thickened and faulted western rift mar­
gin, silting above basement blocks. Approximate exposure level shown by horizontal line; ECV = Eastern Creek Vol­
canics. Adapted from G olcbycta). (1996) and Betts and Lister (2001). b. Starting geometry o f the models, and boundary
conditions for model runs (see text). c, d. Deformed grid and instantaneous fluid flow vectors after 25 percent short­
ening for a m odel with n o change in material properties during model run (Tabic 4). e. Corresponding zonaiion o f
areas o f inflow and outflow relate to differences in degree o f volumetric strain (dilation) within the deform ing rock
bodies. N ote how shear zone deformation occurs in several locations in the M ount Isa Group, which would not (under
these m odel conditions) be favorable for localization o f copper ore in only the present location (see text), f, g.
Deformed grid and instantaneous fluid flow vectors after 25 percent shortening for a model in which the dilation angle
o f the Eastern Creek Volcanics decreases from +5 to - 5 s during the model run (see text), h. Corresponding localiza­
tion o f outflow from the Eastern Creek Volcanics, and inflow into the Mount Isa Group, in a scenario that simulates the
localization leading to the present orebody disposition. Fluid movement is concentrated in the center o f the model,
where the half-grabcn has thickened and inverted. Fluid has moved out o f the Eastern Creek Volcanics, where they arc
m ost deform ed, and into the Mount Isa Group directly above the region o f the half grnben which is most thickened, i.
Pore pressure profiles at different degrees o f shortening (without erosion) through the western edge o f the half-graben
in the second m odel. As the Eastern Creek Volcanics in the half-graben contract, due to its increasingly negative dila­
tion angle, the pore pressure increases to be nearly lithostatic near the base, strongly driving fluid upwards into the
overlying rocks. Downwards flow is inhibited by the low penneability o f the basement rocks (Tabic 1).
 10 k m

f) g)

10 km 10 km
h) Copper
Pore Pressure (10s Ma)
orebodies
0 1 2 3 4

10 km
 T abu•: -1. Material Properties for the FLAG Model: Presented in Figure 7
Bulk Shear Tensile
Density modulus modulus strength Cohesion Friction Dilation Permeability
(k g /m :*) (Pa) (Pa) (Pa) (Pa) angle angle (m-)

Cover 2.700 2.7cl0 7e9 1.2c6 2.7c7 28= 5s 1o—l 3

MIG 2.700 2.3c 10 l.ielO 1.6c6 6.7c6 42° 10" Ic-12
ECV 2.700 2.7eI0 7e9 1.2e(5 2.7e7 28° 5®1 ic -1 3
Basement 2,700 4.4c 10 3c 10 i.3e7 G.Ge7 31° 5" lc - I 6

1 In the second m odel, the dilation angle is a linear function o f tlte amount o f plastic strain, ranging from*+»® initially to -5 ° after 25% shorten­
ing deformation (see text for details)
ECV = Eastern Creek Volcanics; MIG = Mount Isa Group

model. In these two-dimensional models, a net inflow into zones where the largest change in volume takes place (Fig.
a cell means that more fluid is moving into that cell from 7d, e). W ithin the M ount Isa Group, the shears are well
the surrounding cells, and a net outflow means that fluid is defined, whereas within the half-graben, the material is
moving from that cell into the surrounding cells (there is thickened rather than forming discrete shear zones, tins pat­
no consideration o f fluid m ovement in the third dim en­ tern developing as a function o f the different rock proper­
sion). This calculation provides a useful visualization o f the ties and their geometric arrangem ent during shortening.
two-dimensional time-integrated fluid flow, to complement Most o f the fluid movem ent occurs within the M ount Isa
the images of instantaneous fluid flow vectors. Group, the unit with the highest permeability and dilation
angle. Generally, flow is into the shear zones where the
Results and implications: It is outside the scope of this largest volume increase has taken place. Calculation o f net
paper to describe all of the 20 m odel perm utations that fluid flow within the grid show that fluid is moving from the
were run. However, o f key significance in comparison to the surrounding m aterial into the shear zones (Fig. 7c).
other models is that wc have allowed certain rock properties Because o f die dilatant nature o f the Mount Isa Group and
to vary during deform ation, in particular to simulate the the localization o f deformation within the conjugate shears
effects of reaction softening or hardening, and permeabil­ in this rock type, volume increase occurs in these shears.
ity changes accompanying alteration and syn-tectonic fluid- The increase in volume pulls in fluid from the surrounding
rock interaction. The model presented above for fluid flow material. However, the model fails to explain the localiza­
at the Mary Kathleen U-REE deposit contained no provision tion o f fluid flow in the area corresponding to the copper
for change of rock properties as a function of lime/strain. deposits; that is, there is insufficient focusing to explain the
The first model for Mount Isa follows the same principle, its very high tonnage and grade o f the ore bodies at these locar
a benchmark for the subsequent models (Fig. 7c-e). In the tions. Also, the model does not explain the geochemical rig-
second model, the dilation angle of the Eastern Creek Vol­ nature of the ore-forming fluids which indicates interaction
canics was made a function of the amount of plastic strain with the Eastern Creek Volcanics (Waring et al., 1998).
(Fig. 7f-h). As deformation takes place, the dilation angle
decreases until it becomes negative and the material is then Model 2 (dilation angle in Eastern Creek Volcanics decreases as
contractant rather than dilatant. We have inserted this func­ a function ofstrain): Having scavenged the necessary ore com­
tion to simulate the effect o f pore space collapse and fluid ponents from one rock unit (e.g., copper from the Eastern
pressure build up in the Eastern Creek Volcanics as plastic Creek Volcanics), fluid must then be focused in some way as
deform ation and mineral reactions progressively occlude it moves into a different environment where chemical, pore
the porosity. This effect will be greater in these rocks as they pressure, or thermal condidons cause it to precipitate m e t
are buried deeper in the pile, where secondary porosity als. As deformauon in this model proceeds, pore space in the
reduction by deformation and mineral reactions would have plasucally deforming regions of the Eastern Creek Volcanics
proceeded to a greater degree than in overlying rocks, is progressively destroyed (the diladon angle decreases lin­
enhanced also by the reactivity of the mafic rocks in contact early with percentage strain), the pore fluid becomes over­
with basinal fluids (Heinrich et a!., 1995). pressured, and fluid moves upward into the deform ing
regions o f the Mount Isa Group where volume increase with
Model l (no change in parameters during run): When the deform ation is occurring. Because o f the inverted basin
model geometry is shortened by 25 percent using FI AC, the geometry’, a large am ount o f fluid from the Eastern Creek
response of the basin is to invert and thicken with deforma­ Volcanics, thus, flows into a small region o f the Mount Isa
tion concentrated into conjugate shear zones whose loca­ Group (Fig. 7g, h), simulating the desired conditions for
tion is controlled by the geometry o f the basin within the generation of the copper ore deposits. Significant fluid flow
stronger basement rocks (Fig. 7c). The boundary between occurred from the Eastern Creek Volcanics into the Mount
the half-graben and the basem ent steepens as the basin Isa Group, near die Paroo fault (Fig. 7a), mimicking die pat­
inverts. Fluid flow is coupled to the deform ation as tern determined from the geochemical and isotopic studies
described earlier and fluid moves into o r out o f the shear (Heinrich et al-, 1995; Waring et al., 1998).
 The geometry of die focusing structures was provided by
the interaction between the deformation and the preexist­
ing configuration o f the different rock bodies, which any
stress modeling would provide, but the effectiveness o f the
focusing (and, hence, the ability to produce a giant ore-
body) was simulated in the model by our interpretation of
the relationship between dilation and deform ation in the
Eastern Creek Volcanics in comparison to the M ount Isa
Group. T he evolution o f pore pressure with tim e in this
model (Fig. 7i) shows the type o f pattern one would expect
in a compacting sedimentary basin affected by an initial vari­
able distribution o f rocks with different rheology and per­
meability. Upflow is favored, particularly at higher strains
when the fluid pressure In the Eastern Creek Volcanics
approaches lithostatic, whereas the overlying M ount Isa
Group continues dilating and maintains near hydrostatic
conditions due to hydrodynamic interconnection with over-
lying units and the surface. Although downflow would be
possible from the Eastern Creek Volcanics into the lower
basement because of the reverse fluid pressure gradient, this
is not favored because of the very low permeability and
greater cohesion of these underlying rocks.
The modeling provides an explanation for how syntec-
tonic copper orebodies could be localized in the vicinity of
former basin-bounding faults, thus, explaining superposi­
tion of copper on syn-sedimentary or diagenetic lead-zinc
orebodies. In the case o f the lead-zinc deposits, basinal
architecture would have been param ount in providing the
appropriate conditions for ore deposition (Betts and Lis­
ter, 2001). For the copper deposits, the geometric require­
ments would seem to be a localization in the area of great­
est thickening, adjacent to a former half-graben bounding
fault. O ur first model (Fig. 7d, e) provides for regional fluid
movement within the M ount Isa G roup, which may have
been a necessary com ponent o f the dissolution and trans­
port phase o f copper mineralization. However, the neces­
sary focusing mechanism is n o t provided by the real o r
FLAC-simulatcd geometry' alone, n o r would it be by a sim­
ple stress m apping approach (i.e., with UDEC modeling).
An additional necessary criterion is one in which some of
the properties o f the Eastern Creek Volcanics (such as dila­
tion angle) change during deformation so dial fluid is both
driven upwards and focused into the present location of
the ores (Fig. 7g, h). In combination, these models would
satisfy the regional geochemical m odel which appeals to
interaction o f evolved basinal brines with M ount Isa Group
sediments, the Eastern Creek Volcanics, and a metamor-
phic fluid source (Waring etal., 1998).
Discussion
An empirical approach may be used by exploration geol­
ogists when dealing with target selection in areas of strongly
heterogeneous rocks, and it also forms the basis for selec­
tion o f parameters and geometry in any numerical simula­
tion. Such an approach would concentrate on the follow­
ing: ( l) determining the likely orientation of the su*ess field
during fluid flow and mineralization; (2) identifying rock
boundaries and ranking these according to perceived dif­
ference in rheology between adjacent rocks; (3) ranking
rock boundaries according to orientation relative to the
inferred stress field; and (4) conducting an analysis of the
degree of interconnectivity of the indicated and prioritized
rock boundaries according to the principles outlined in
this volume by Sibson (1996, 2001) and Cox e t al. (2001).
Once the geometry and inferred kinematics o f the study
area arc established, assumptions m ust be m ade regarding
boundary conditions for input into the numerical models.
We have found this process o f boundary condition defini­
tion to be the m ost difficult, b u t scientifically the m ost
rewarding, o f all o f the processes o f m odel construction,
because It forces us to face the fundamental conditions of
the geology that will n o t generally be revealed by an empir­
ical, map-based analysis. Critical factors are the assumptions
regarding fluid pressure boundary conditions, and the pos­
sible shape changes and movements perm itted for the
model boundaries. Generally, it will only be possible to con­
strain these factors through a good knowledge of the geo­
logical problem. For example, the Mount Isa model (Fig.
7) has assumed initial hydrostatic conditions with hydraulic
interconnection to the surface, based on the known pre-
deformational history of the Mount Isa basin. In contrast,
the Mary Kathleen orebody formed during the regional
metamorphism, at depths of around 15 km, so it would not
be reasonable to assume initial hydrostatic conditions
because the metamorphism was likely to have elevated fluid
pressures to near-lithostatic values, and the possibility of ini­
tial hydraulic interconnection to the surface is remote. In
most cases, we have considered the local kinematic frame­
work and geometry in relation to the broader pattern (e.g.,
Fig. 5a), crucial in any analysis that moves from regional to
local scales. Testing these different scenarios is a vital part
o f the modeling, because it stimulates and focuses thought
processes towards realistic outcom es. Multiple working
hypotheses can be constructed that equate to different,
testable exploration models.
An understanding o f the scale and m agnitude o f rheo­
logical variability also provides a way o f determ ining the
likely spacing, and to some extent the size, o f ore deposits
whose localization is primarily controlled by the action o f
fluid flow during deformation (i.e., syntectonic epigenetic
ores). For a given strain, th e scales over which zones o f
anom alous stress develop around rheological hetero­
geneities are related mainly to the difference in material
properties o f the different rocks involved, as well as the spe­
cific geometry o f the rocks. For simple models, the scale o f
fluid flow perturbation induced by deform ation is on the
o rd er o f the size o f the rheological a n d /o r permeability
heterogeneity. If m ore com plicated situations are intro­
duced into the models, the fluid flow' pattern will be corre­
spondingly complicated (O rd and Oliver, 1997). In Figures
4,5, and 6, perturbation in the vicinity of the stronger rock
bodies gives rise to lateral and local downwards fluid flow,
because the lateral gradient in mean stresses and, hence,
fluid pressure, is similar in magnitude to vertical buoyancy-
related fluid pressure gradients; that is, the deformation-
induced com ponent of flow is at least as im portant as the
fluid buoyancy. The scale o f the fluid flow perturbation is tion. An improved understanding o f how reaction- and
related to the scale of the rock heterogeneity in all models. deform ation-enhanced permeability and softening/hard-
For structurally controlled ores, this implies that the spac­ ening relationships affect fluid flow, will lead to great
ing and abundance o f small, moderate, and large mineral­ improvements in the utility of the num erical m odeling
ization targets could be predicted by an analysis o f the geo­ approach. The overall aim of adding quantitative numeri­
metric configuration of rock bodies whose rheological cal models to the science of geology will ultimately benefit
properties can be adequately surmised. all who are concerned with the ways in which large ore
Orebodies form ed in deform ing rocks are commonly deposits were constructed.
located at sites of the greatest rheological and permeability
contrast in the regional system, and in many cases the size Acknowledgments
of the deposit reflects the scale of operation of the fluid We would like to thank G. Lister and R Betts for infor­
flow system. A requirem ent for large ore deposit genesis is, m ation pertaining to the M ount Isa reconstructions, R.
thus, that there are mechanisms or parameters of the sys­ Holcombe, R Pearson, and Z. Jiang for assistance with the
tem whereby local fluid flow perturbations are subordinate Mary Kathleen work, B. Mutton at the Hilton mine, and B.
to an overall flow pattern. Otherwise, insufficient source Hobbs, Y. Zhang, P. Gow, and M. Couthard for assistance
rocks may be accessed for metal dissolution, or there may with the modeling aspects. The field and mine com ponent
be insufficient focusing to concentrate much ore precipi­ of the research was supported and funded by M ount Isa
tation (Fig. 7d, e). Complex rock geometry at all scales will Mines Ltd, Monash University, AGSO (then BMR), Uni­
tend towards forcing a highly tortuous fluid flow with mul­ versity of Queensland, and the Australiari Research Coun­
tiple focii. Rocks that show strain-dependent permeability cil through its Large Grant scheme. We would also like to
changes will act as good sources or sinks for fluid flow. thank A. Boullier, an anonymous reviewer, and the superb
Broad gradients produced by buoyancy, topography, or editorial skills of J. Richards, for greatly contributing to
fluid overpressure will enhance long distance fluid flow as improvement of the manuscript.
long as the local perturbations occur over lesser magni­
tudes. Small fluid fluxes enhance the tendency for flow REFERENCES
localization, but conversely, very large fluid fluxes will Beach, A., 1977, Vein arrays, hydraulic fractures and pressure-solution
swamp the effect of rock heterogeneity, and the latter cir­ structures in a deform ed flysch sequence, SAV. England: Tectono-
physics, v. *10, p. 201-225.
cumstance is ideal for metal dissolution from large rock vol­
Betts, P.G., and Lister, G.S., 2001, A comparison o f the “strike-slip" versus
umes. To emphasize these points, the giant M ount Isa cop­ the "episodic rift-sag" hypotheses for the origin o f the Isa Superbasin:
per orebody lies on the boundary between upper Australian Journal o f Earth Sciences, v. 47, in press.
amphibolite facies and lower greenschist facies rocks (Fig. Birch, F., 1966, Elasticity, in Birch, S.P., Jr., ed., Handbook o f physical con­
7), and the Paroo/M ount Isa fault system almost certainly stants: Geological Society o f America, Memoir 97, p. 63-86.
Boullier, A.-M., Charoy, B., and Pollard, P.J., 1994, Fluctuation in porosity
represents both a major rheological boundary and one of and fluid pressure during hydrothermal events: Textural evidence from
the major features controlling the regional fluid flow pat­ the Emuford district, Australia: Journal o f Structural Ccology, v. 16,
tern, with rheological/perm eability contrast occurring at p. 1417-1429.
the 10-km scale. In contrast, fluid flow in the similar-aged Brace, W.F., 1968, T he m echanical effects o f pore pressure on the frac­
turing o f rocks, i« Baer, A.J., and Norris, D.K., cds., Research in tecton­
Mary Kathleen fold belt 50 km to the east (Oliver e t al., ics: Geological Survey o f Canada, Paper 68-52, p. 113-124.
1990) involved channelized fluid flow on the 100 m to 1 km Brown, S.R., 1987, Fluid flow through rock joints: T he effect o f surface
scale, focused around structural heterogeneities o f the roughness: Journal o f Geophysical Research, v. 92, p. 1537-1347.
same scale. As a consequence, mineralization in the latter Cox, S.F., 1999, Deformational controls on the dynamics o f fluid flow in
mcsothermal gold systems, in McCaffrey, K., Loncrgan, L., and Wilkin­
belt occurs at hundreds o f locations bu t does not form
son, J., cds., Fractures, fluid flow, and mineralization: Geological Soci­
large orebodies. The most notable exception, the modest­ ety o f London Special Publication, v. 155, p. 125-140.
sized Mary Kathleen U-REE syn-m etam orphic orebody, Cox, S.F., Knackstedi, M A , and Braun, J„ 2001, Principles o f structural
occurs on the largest rheological boundary in the entire control on permeability and fluid flow in hydrothermal systems: Reviews
belt, the Mary Kathleen shear zone (Oliver et al., 1999). In in Economic Geology, v. 14, p. 1-24.
the Cloncurry district even further east, syntectonic iron­ Cundall, PA., and Board, M., 1988, A microcomputer program for mod­
elling large-strain plasticity problems, in Swoboda, C., cd.. Numerical
stone-hosted Cu-Au deposits such as Ernest Henry, Starra- m ethods in geom echanics: International Conference on Numerical
Sehvyn, and Osborne, all show strong structural controls Methods in Gcomechanics, 6th, Rotterdam, Proceedings, p. 2101-2108.
(Laing, 1993), although fluid sources are inferred to be Dipple, G.M., and Ferry, J.M., 1992, Metasomatism and fluid flow in duc­
tile fault zones: C ontributions to M ineralogy and Petrology, v. i 12,
mostly from crystallizing magmas (Williams, 1998; Mark,
p . 149-164.
1999). Although data are incom plete and the degree of Dom enico, PA., and Schwartz, FAV., 1998, Physical and chemical hydro­
direct control of ore localization by proximity to specific geology: New York, Wiley, 506 p.
granite types is uncertain, the size and spacing o f these Durncy, D.W., 1976, Pressure-solution and crystallization deformation:
deposits may also be a function o f the spacing and rheo­ Royal Society o f London Philosophical Transactions, Earth Sciences,
v. A283, p. 229-240,
logical variability of rock types involved in the deformation. Edwards, R.H., 1951, Stress concentration around spheroidal inclusions
O ngoing m odeling work is directed towards gaining a and cavities: Journal o f Applied Mechanics, v. 18, p. 19-30.
greater understanding, largely from geological inferences, Etheridge. M A .. Wall, VJ„ Cox, S.F., and Vernon. R.H., 1984, High fluid
o f the ways in which rock properties change with deforma­ pressures during regional metamorphism and deformation: Implica-
 tions for mass transport and deformation mechanisms: Journal o f Geo­
physical Research, v. 89, p. 4344—4358.
Ferry, J.M., and D ipplc, G.M., 1991, Fluid flow, mineral reactions, and
metasomatism: Geology, v. 19, p. 211-214.
Garvcn, G., and Freeze, R A , 1984a, Theoretical analysis o f the role o f ground-
water flow in die genesis o f stratabound ore deposits. 1. Mathematical and
numerical model: American Journal o f Science, v. 284, p. 1085-1124.
------ 1984b, Theoretical analysis o f the role o f groundwater flow in the
genesis o f stratabound ore deposits. 2. Quantitative results: American
Journal o f Science, v. 284, p. 1125-1174.
Golcby, B.R., Drummond, B.J., and MacCrcady, T., 1996, T he deep seis­
mic reflection profile south o f M ount Isa and Cloncurry: Australian
Geological Survey Organisation Research Newsletter, v. 24, p. 6-8.
Hanson, R.B., 1992, Effects o f fluid production on fluid flow during
regional and contact metamorphism:Journal o f Meiamorphic Geology,
v. 10, p. 87-98.
Heinrich, C A ., Bain.J.H.C., Mernagh, T.P., Wybom, L A I., Andrew, A S.,
and Waring, C.L., 1995, Fluid and mass transfer during mctabasalt alter­
ation and copper m ineralization at M ount Isa, Australia: Econom ic
Geology, v. 90, p. 705-730.
Helgeson, H.C., and Lichtncr, P.C., 1987, Fluid flow and mineral reactions
at high temperatures and pressures: Geological Society o f London Jour­
nal, v. 144, p. 313-326.
Hobbs, B.E., Means, W.D., and Williams, P.F., 1976, An outline o f struc­
tural geology: New York, Wiley, 571 p.
Holcom be, RJ., Pearson, PJ., and Oliver, N.H.S., 1992, Structure o f the
Mary Kathleen fold belt, in Stewart, AJ„ and Blake, D.H., c<ls.. Detailed
studies o f the Mount Isa Inlicn Australian Geological Survey Organisa­
tion, Bulletin 243, p. 257-287.
Holyland, P.W., and Ojala, VJ., 1997, Computer-aided structural targeting
in m ineral exp lo ratio n : Two-and three-dimensional stress mapping:
Australian Journal o f Earth Sciences, v. 44, p. 421—132.
Jaeger, J.C., 1969, Elasticity, fracture, and flow: London, Chapman and
Hall,*268 p.
Jaeger, J.C., and Cook, N.G.W., 1979, Fundamentals o f rock mechanics:
London, Chapman and Hall, 593 p.
Jiang, Z., Oliver, N .H .S., Barr, T.D., Power, W.L., and O rd, A., 1997.
Numerical m odelling o f fault-controlled fluid flow in the generis o f tin
deposits o f the Maiage Orcficld, Gcjiu mining district, China: Economic
Geology, v. 92. p. 228-247.
Kerrich, R.. and A lison , I., 1978, Vein geometry and hydrostatics during
Yellowknife mineralization: Canadian Journal o f Earth Sciences, v. 15,
p . 1653-1660.
Laing, W.P.. 1993, Smictural/mctasomatic controls on ore deposits in the
east Mount Isa block: The key to tonnes and grade: Australian Institute
o f Geoscicinists Bulletin, v. 13, p. 17-24.
Lattman, L.A., and Parizck, R.R., 1964, Relationship between fracture
traces and the occurrence o f groundwater in carbonate rocks: Journal
o f Hydrology, v. 2, p. 7 3 -9 1.
Lemos, J.V., Hart, R.D., and Cundall, P A . 1985, A generalised program
for m odellingjointed rock masses—A keynote lecture, in Stcpiiansson,
O.. ed., Proceedings o f the International Symposium ou Fundamentals
o f Rockjoints: Centec, Lolea, Sweden, p. 335-344.
Maas. R., McCulloch, M.T., and Campbell, I.H., 1988, Sm-Nd isotope sys-
ternadcs in uranium rare-earth clement mineralization at Maty Kathleen
uranium mine, Queensland: Economic Geology, v. 82, p. 1805-1826.
Mair.J.L., Ojala. VJ., Salier, B.P., Groves, D.I., and Brown, S.( 2000, Appli­
cation o f stress mapping in cross section to understanding ore geom e­
try, predicting ore zones, and developm ent o f drilling strategics: Aus­
tralian Journal o f Earth Sciences, v. 47, p. 895-912.
Mark. G., 1999, Pctrogenesis o f Mesoprotcrozoic K-rich granitoids, south­
ern Mt Angclay igneous complex, Cloncurry district, northwest Queens*
■ land: Australian Journal o f Earth Sciences, v. 46, p. 933-950.
O'Dca, M.G., Ustcr, G.S., MacCrcady, T., Betts, P.G., Oliver, N.H.S., Pound, ICS.,
: Huang, W., and Valenta, R.K., 1996, Geodynamic evolution o f the Protero-
: zoic Mount Isa terrain, in Burg, J.-P., anti Ford, M., eds.. Orogeny through
time: Geological Society o f London Special Publication 121, p. 99-122.
Ojala, V.J., Ridley,J.R., Groves, D.I., and Hall, G.C., 1993, T he Granny
Smith gold deposit: T he role o f heterogeneous stress distribution at an
irregular granitoid contact in a greenstone facies tcrranc: Mincralium
Dcposita, v. 28, p. 409-419.
Oliver, N.H.S., 1995, The hydrothermal history o f the Mary Kathleen fold
belt, M ount Isa block, Q ueensland, Australia: Australian journal o f
Earth Sciences, v. 42, p. 267-280.
------ 1996, Review and classification o f structural controls on fluid flow
during regional mciamorphism:Journal o f Metamorphic Geology, v. 14,
p. 477—492.
Oliver, N.H.S., anti Wall, VJ., 1987, Metamorphic plumbing system in Pro­
terozoic calc-silicates, Queensland, Australia: Geology, v. 15. p. 793-796.
Oliver, N.H.S., Valenta, R.K., and Wall, VJ., 1990, The effect o f heteroge­
neous stress and strain on metamorphic fluid flow, Mary Kathleen, Aus­
tralia, and a m odel for large-scale fluid circulation: Journal o f Meta­
morphic Geology, v. 8, p. 311—331.
Oliver, N.H .S., Pearson, P.J., H olcom be, R.J., and Ord, A., 1999, Mary
Kathleen mctamorphic-hydrothcrmal uranium-rare-earth deposit: Ore
genesis and a numerical model o f coupled deformation and fluid flow:
Australian Journal o f Earth Sciences, v. 46, p. 467-484.
Ord. A , 1990, Mechanical controls on dilatant shear zones, in Knipc, R.J.,
and Rutter, E.H., cds.. Deformation mechanisms, rheology, and tecton­
ics: Geological Society o f London Special Publication 54, p. 183-192.
------ 1991, Deformation o f rock: A pressure-sensitive, dilatant material:
Pure and Applied Geophysics (Pagcoph), v. 137, p. 337-366.
Ord, A., and Oliver, N.H.S., 1997, Mechanical controls on fluid flow dur­
ing regional metamorphism: Som e numerical models: Journal o f Meta­
morphic Geology, v. 15, p. 345-360.
Page, R.W., 1983, Chronology o f magmatism, skarn formation, and ura­
nium mineralization, Mary Kathleen, Queensland, Australia: Economic
Geology, v. 78, p. 838-853.
Perkins. W.G., 1984, Mount Isa silica-dolomite and copper orebodies: The
rcsuliofasyntccionic hydrothermal alteration system: Economic Gcol-
ogy, v. 79, p. 601-637.
Phillips, WJ., 1972, Hydraulic fracturing and mineralisation: Geological
Society o f London Journal, v. 128, p. 337-359.
Reynolds, S.J., and Lister. G S ., 1987, Structural aspects o f fluid-rock inter­
actions in detachm ent zones: Geology, v. 15, p. 362-366.
Robert, F., and P oulsen, H ., 2001, Vein formation anti deform ation in
greenston e gold deposits: Reviews in Econom ic G eology, v. 14,
p. 111-155.
Rumble. D.I.. and Spear. F.S.. 1983. Oxygen-isotope equilibration and per­
meability enhancem ent during regional metamorphism:Journal o f the
Geological Society o f London, v. 1-10. p. 619-628.
Sccor, D.T.. 1965. Role o f fluid pressure in jointing: American Journal o f
Science, v. 265. p. 633-646.
Sibson, R.H., 1996, Structural permeability o f fluid-driven fauit-fracturc
meshes: Journal o f Structural Geology, v. 18, p. 1031-1043.
------ 2001, Seism ogcnic framework for hydrothermal transport and ore
deposition: Reviews in Economic G eology v. 14, p. 25-50.
Sibson, R.H., Robert, F., and Poulsen. K.H-, 1988. H igh angle reverse
faults, fluid pressure cycling, and mcsothermal gold-quartz deposits:
Geology, v. 16, p. 551-555.
Starfield, A M ., and Cundall. PA., 1988, Toward a m ethodology for rock
m echanics m odeling: International Journal o f Rock M echanics and
Minerals Society o f Ccom cchanics Abstracts, v. 25, p. 99-106.
Stromgard, K.-E., 1973, Stress distribution during formation ofboudinage
and pressure shadows: Tcciouophysics, v. 16, p. 215-248.
Swager, C P., 1985, Syndcformaiional carbonatc-rcpIacemcnt m odel for
the cop p er m ineralization at M ount Isa, Northwest Q ueensland: A
microstructural study: Economic Geology, v. 80, p. 107-125.
Valenta, R.K., 199-1, Syntcctonic discordant copper mineralization in the
Hilton m ine. Mount Isa: Economic Geology, v. 89, p. 1031-1052.
Valenta, R.K., Cartwright, I., and Oliver. N.H.S., 1994. Structurally-con­
trolled fluid flow associated with breccia vein formation: Journal o f
Metamorphic Geology, v. 12, p. 197-206.
Waring. C L .. Heinrich. C A , and W ill. VJ.. 1998, Proterozoic mctainor-
phic copper deposits: AGSO Journal o f Australian G eology and G eo­
physics, v. 17, p. 239-246.
Williams, PJ-. 1998, Metalliferous econom ic geology o f the Mt Isa Eastern
Succession, Queensland: Australian Journal o f Earth Sciences, v. -15.
p .329-341.
Zhang. S., Cox. S.F., and Paterson, M.S., 1991. The influence o f room tem­
perature deform ation on porosity and permeability in calcite aggre­
gates: journal o f Geophysical Research, v. 99, p. 15,761-15.778.
<

.
■

'
.

"
-
Society of Economic Geologists
Reviews v. 14, 2001, p. 75-109

Chapter 4

Utility o f Magnetic and Gravity Data in Evaluating Regional Controls on

Mineralization: Examples from the Western United States
T. G. H ildenbrand,1
U .S. G eological Survey, M a il Stop 989, 5 4 5 M id d lejield R oad, M enlo P ark, C alifo rn ia 9 4 0 2 5

Byron Berger,
U .S. Geological Survey, M a il Stop 964, D enver Federal Center, D enver, Colorado 8 0 2 2 5

R. C. J achens, and Steve L udington

U .S. G eological Survey, M a il Stop 9 8 9 , 3 4 5 M id d lefteld R oad, M enlo P ark, C alifo rn ia 9 4 0 2 5

Abstract

In te r a c tin g fra c tu re s enhance and localize permeability in the Earth’s crust and are, therefore, impor­
ta n t p h e n o m e n a in lo c a liz in g magmatic and hydrothermal systems. The ability to identify where such in­
te ra c tio n s a re p r e s e n t is useful in evaluating likely areas of mineralized rock, particularly in covered ter­
ra in s. R e g a rd le s s o f m a p scale, the interpretation of gravity and magnetic data can define deep-seated
c ru sta l f ra c tu re s a n d faults that may have guided emplacement of igneous rocks and large ore deposits.
H e r e w e e m p h a s iz e recurring regional-scale structural relationships mainly from the western United
S ta te s b a s e d o n th e interpretation of potential-field data, which can elucidate areas of past and present
fluid flow in th e crust.
In particular, we explore the utility of regional gravity and magnetic data to aid in understanding the
distribution of large Mesozoic and Cenozoic ore deposits (primarily epithermal and pluton-related pre­
c io u s and base metal deposits, a n d sediment-hosted g o ld deposits) in the western United States
c o rd ille ra . On the broadest scale, m o s t o r e deposits lie within areas characterized by low magnetization.
T h e Mesozoic Mother Lode gold belt displays characteristic geophysical signatures (regional gravity high,
re g io n a l low-to-moderate background magnetic field anomaly, long curvilinear magnetic highs) that
m ig h t serve as an exploration guide. Geophysical lineaments characterize the Idaho-Montana porphyry
belt and the La Caridad-Mincral Park belt (from northern Mexico to western Arizona) and, thus, indi­
cate deep-seated control for these mineral belts. At a more local scale, in Nevada, geophysical data de­
fine deep-rooted faults and magmatic zones that correspond to patterns of epithermal prccious-mctal de­
posits, and that may relate to the Carlin gold trend and the Battle Mountain-Eureka mineral belt. One
recurring structural model evolving from this study is that mineralization in the western United States
may be localized along strike-slip fault zones where pull-apart basins or releasing bends provided the in­
creased fracture permeability for the migrating ore-forming fluids (e.g., die Butte, Tombstone, Bagdad,
and Battle Mountain districts).
Many deposits discussed in the paper appear, at least in part, to be associated with reactivated older
faults as well as with faulting contemporaneous with ore deposition. We conclude that at a local scale,
structural elements work together to localize mineral deposits within regional zones or belts. Perhaps the
greatest utility of regional geophysical data is the identification of structural relationships that help nar­
row the study area, where more intensive multidisciplinary team studies can be carried out in a concerted
effort to evaluate the mineral potential.

Introduction ing primary structures related to mineralization (e.g., inter­

At a l l map scales, th e forem ost role o f the geophysicist in
mineral exploration is to provide geological inform ation in
three dim ensions, particularly in areas covered by younger
rocks. At local scales (deposit and district studies), high-res­
olution magnetic an d gravity data are clearly useful in defin-
^Corresponding author: e-mail, tonj@usgs.gov
secting faults, kimberlite dikes, and folds). At regional
scales, geophysical data help elucidate the geological frame­
work, mainly by defining major compositional boundaries
or structural zones that may be favorable environs for min­
eralization. Here we review and investigate the utility of grav­
ity and magnetic data in understanding the distribution of
large ore deposits primarily at regional scales, concentrating
on epitherm al and porphyry precious- and base-meta!
deposits, and sediment-hosted gold deposits in the western
United States.
The view that geophysical data provide structural and
lithological information essential in mineral investigations is
evidenced in the allocation o f m onetary resources by gov­
ernments to carry out national data collection programs to
encourage exploration. For example, mineral exploration
in Australia over the past four decades has relied heavily on
national geophysical databases (primarily magnetic data col­
lected at a 1.6-km flight-line spacing, and gravity data col­
lected at an average spacing o f l 1 km; Jaques et al., 195)7).
This data collection period resulted in the discovery of a
num ber of major ore deposits, some world class, such as
Olympic Dam (copper-gold-uranium), Kambalda (nickel),
Argyle (diamonds), and Century (zinc-lead). The success of
this collection of geological and geophysical data spawned
the 1990 National Geoscience Mapping Accord which incor­
porated the acquisition o f aeromagnetic data at a line spac­
ing o f 0.4 km in selected areas o f Australia. The high num­
ber o f new exploration licenses in surveyed areas
demonstrates the utility o f detailed geophysical data in min­
eral exploration, particularly in covered areas (Robson and
Lewis, 1997). In Canada, aerom agnetic and gravity data
were acquired mostly by the Geological Survey o f Canada
(CSC) through a 50-year-long national program, in which
consistent and high-quality survey specifications were main­
tained (Teskey et al., 1993). Canadian aeromagnetic anom­
aly maps are used extensively in exploration and geological
mapping by private, academic, and government agencies.
Other countries have observed the exemplary, cost-effective
geophysical programs of Australia and Canada and have ini­
tiated their own surveys (e.g., Botswana, Finland, Green­
land, India, Malaysia, Mexico, Namibia, Norway, Sweden,
Thailand, and Zimbabwe).
In order to investigate the usefulness of geophysical darn
to determine regional structures and their relationship to
the dismibution of ore deposits in the western United States,
we adopt several working hypotheses, ideas that have
appeared frequently in the literature. Implicit throughout
our study arc the following assumptions: (1) deep-seated
fault systems at both regional and local scales can localize
igneous activity, major ore deposits, ore districts, and min­
eral belts by providing plumbing systems for intrusions, ore-
forming fluids, and heat entering the crust; (2) these fault
systems are susceptible to repeated reactivation through
geological time; and (3) intersecting fault systems are espe­
cially favorable for ore deposits. In particular, we examine
the main utility of regional geophysical data in aiding the
definition of zones or lineaments representing these reacti­
vated fracture or fault systems. Examples of structures char­
acterized by fracture systems that penetrate the crust and
possibly the upper mantle include major fault zones (e.g.,
transform a n d /o r shear zones), intrusive zones, failed rifts,
and relicts of tectonic plate margins.
Geophysical studies at scales 1:24,000 o r greater are not
emphasized here. Rather, we focus on geophysical studies
leading to generalized ore deposit relationships or empiri­
cal information useful in evaluating m ineral potential at
regional scales (1:50,000 or less). As we explore the utility o f
potential-field data from regional to local scales
(-1:5,000,000-1:50,000; i.e., from lower to greater data res­
olution), the am ount of geological inform ation derived
from the potential-field data clearly increases, and, thus, the
benefits o f using potential-field data in mineral investigation
studies significantly increase.
We emphasize recurring geological structural relation­
ships that may apply to many classes of deposits. One com­
m on relationship evolving from the present study is that
many ore deposit types are present along major strike-slip
fault zones within magmatic arc environm ents (Fig. 1).
O ther empirical relationships such as ore deposit districts
lying at die intersection of major crustal boundaries are also
discussed.
Potential-Field Data and Methods
Magnetic data and method
The western United States aeromagnetic data set is based
on a patchwork o f over 300 surveys (line spacings ranging
from 0.4-8 km; Fig. 2 and inset), acquired to address a wide
variety o f objectives over a period of nearly 40 years (Hildcn-
brand e t al., 2000). During the past twelve years, the U.S.
Geological Survey recompiled these individual data sets by
continuing all the data to a common reference level o f 0.305
km above the ground, and by carefully merging the 1-km-
gridded data from each aeromagnetic survey with adjacent
data. Some surveys were made with line spacings o r altitudes
that were far too large to properly resolve shallow magnetic
sources. For example, over some parts o f the area, the only
existing data are those collected for the National Uranium
Resource Evaluation (NURE) Program at flightline spacings
of 5 km or greater, which are generally too wide to properly
characterize isolated sources lying at depths much shallower
than 5 km. Nevertheless, the data in Figure 2 are shown
below to be of adequate quality to provide new insights on
structures important to mineral investigations.
Several excellent reviews of the magnetic m ethod have
been published by Reford (1980), Grant (1985a, b), Hinze
(1985), Blakely (1995), Gunn and Dentith (1997), and
Reeves et al. (1997). We give a brief outline of the method
with which the reader can understand the basics o f aero­
magnetic interpretations. The magnetic method, one of the
oldest geophysical exploration techniques, is based on map­
ping the subsurface distribution of magnetic minerals con­
taining iron oxide minerals (e.g., magnetite) and iron sul­
fide minerals (e.g., pyrrhotite). The principle of magnetism
of rocks has been used in mineral exploration since 1640,
when a mining compass located iron ore bodies in Sweden.
Because oxides and sulfides of iron are associated with many
different types of mineral deposits, the magnetic method
continues to be a viable approach to directly focus mineral
exploration efforts (Grant, 1985a, b). Magnetic measure­
ments also contribute indirectly to unraveling the geological
setting of ore deposits.
The most effective platform from which to measure the
magnetic field is an aircraft, making the magnetic method
 one o f the most cost-effective approaches in 3D mineral
exploration studies. Although airborne geophysics has held
a prom inent position in mineral applications over the past
50 years, many striking technological advances have taken
place during the past 10 years, as evidenced by the substan­
tial increase in survey coverage during the 1990s (e.g., Den­
ham, 1997). Furthermore, advances in spatial control using
the Global Positioning System (GPS) have reduced survey
costs and increased data accuracy (Reeves et al., 1997).
Moreover, new visualization software enhances the expres­
sion o f subtle geological features with unprecedented clar­
ity, such that they are intuitively identifiable by nonspecial­
ists. The increased effective usage o f aeromagneuc data can
also be attributed to a combination o f faster computers,
improved inversion software, and geographic information
systems (GIS),
Knowledge of magnetic properties of minerals is critical
in understanding die magnetic response ofcrusuil sources.
The volume percentage, geometry, and history of magnetite
grains are generally of greatest importance in crustal mag­
netic studies. Although pyrite is nonm agnetic, hematite,
pyrrhodte, and alloy's of iron can also be important in some
mineral studies (e.g., pyrrhotile in massive sulfide deposits;
pyrrhotite can be roughly one-tenth as magnetic as mag­
netite). Individual dipoles representing magnetic grains can
coalesce into a distribution of dipole moment per unit vol­
“Q P iu to n - ' ume defining magnetization. Total magnetization of rocks is
related the vector sum of two components: induced magnetization
El Epithermal (which is proportional in magnitude and generally parallel
A Sediment- T.;$ to the Earth’s ambient field), and remanent magnetization
hosted gold yjf: <i (which has a direction and intensity dependent on the ori­
tirLow sulfide gin and geological history o f the rock). The intensity o f
gold quartz > ' 1 induced m agnetization is related to the intensity o f the
+ Distal disseminated ambient field through a constant o f proportionality known
Ag Au as the magnetic susceptibility, a unitlcss constant that is
t~30° 0 100 200 km determined by the physical properties o f the magnetic rock.
Rocks have widely varying magnetic properties. Typical
EXPLANATION
values o f susceptibility and rem anent magnetization for
Thnm fault— Sawteeth on common rock types are provided by Clark (1966,1997a, b)
Sedimentary basin upper plate and Carmichael (1982). At regional scales, magnetic highs
PY*.. I] Magmatic center are commonly associated with large masses o f igneous rock
pv7*T1 Zone of inctainotphism and with crystalline basement. Magnetic lows often occur
over areas of thick sedimentary basins and altered basement
Fit;. I. Map o f the western U nited States cordillera, showing ore rock (where magnetite is altered to hematite, for example).
deposits Mi|K*rimposcd on major tectonic elem ents and Laramidc igneous Mafic igneous rocks can be very magnetic, whereas fclsic
zones, sedimentary basins, and mctamorphic belts. T h e western United
States is divided into four generalized geological provinces (boundaries igneous, mctamorphic, and particularly sedimentary' rocks
shown as heavy solid and dashed lines): Pacific margin. Magmatic arc, generally have weaker magnetic expressions. Intrusive rocks
Cordillcran Fold and Thrust belt, and the Rocky Mountain foreland. The commonly have a lower rem anent magnetization and
smallest and largest post-Jurassic deposit symbols represent gross tallies o f higher susceptibility than extrusive rocks with the same
about $20 million and $60 billion, respectively. Intermediate sizes o f sym-
l>ol$ arc based linearly on deposit gross values lying between tlicsc extreme
chemical composition. These statements on magnetic prop­
values. I he short dashed line in northern Utah shows a segm ent o f the erties should be viewed as a general guideline with many
boundary between the Archcan basem ent to the north and Proterozoic exceptions.
l>ascmcni to die south. It should be noted that although Jurassic accretion T he magnetic properties o f rocks arc not only lied to
attd maginatism resulted in complex geological lerrancs along the Pacific their original chemistry but also to many factors related to
coastal states, during dtc l-aiamidc these regions experienced downwarp­
ing and basin developm ent. D = Butte and C - Cannivan Gulch deposits
their history (Grant, 1985a, b). For example, the magnetic
(M ontana); BD = Bagdad and TS = Tom bstone deposits (Arizona); T = response observed in magnetic anomaly maps is influenced
Thompson Creek deposit (Idaho). (Modified from M illerct al.. 1992.) by the geometry and depth o f the magnetic sources, their
orientation with respect to magnetic north, and the indina-
 1*10.2. Reduced-to-pole (KTP) ncromagncttc anomaly map o f the western United States. M any M csowmc- O uozcmc plu-
tonic and volcanic rocks are sufTicicnlly magnetic to be expressed in this map, Assuming induced magnetization is more
dominant than remanent magnetization, the RTP transformation reduces anomaly asymmetry (due to the inclined mag­
netic field) by roughly centering the anomalies over their sources (Blakely. 1995). Inset shows the highly variable fiightlm'e
spacings o f the many surveys forming the patchwork coverage, and highlights the overall poor data resolution. Black lines
show selected major magnetization boundaries (sec Pig. 7) within structural zones discussed in the text. Symbols as in Fig­
ure 1. BD ®Bagdad deposit, Arizona.

lion o f the am bient field at the latitude o f their location.
The degree of oxidation, alteration, weathering, and meta­
morphism also affect the rock’s magnetic properties. All
these factors combine to generate a complex picture of
crustal magnetic properties, particularly near ore deposits
where highly variable mineralogies and thermal histories of
associated rocks are frequently obseivcd.
Gravity data and method
Simpson et al. (1986), Jachens etal. (1989), Simpson and
Jachcns (1989), and Blakely (1995) provide excellent reviews
of the gravity m ethod and discuss m ajor density-defined
domains in the western United States. The gravity method
characterizes lateral density variations. It depends on reduc­
ing gravity measurements made at the Earth’s surface to
enhance die expression of geological targets by removal of
predictable effects. For example, to derive the complete
Bouguer gravity anomaly, corrections are made that relate to
the total mass, rotation, and ellipsoidal shape of the Earth, to
the elevation of the gravity station, and to die attraction of
nearby topographic or bathymetric relief. Bouguer gravity
anomaly maps traditionally have been used to provide a geo­
logical picture of the subsurface overland areas (Simpson et
al., 1986). At regional scales, however, the Bouguer gravity
map displays broad anomalies inversely correlated with
regional topography. These anomalies are caused by deep-
seated masses that isostaucally support the topographic
loads. To remove the effects of diese loads, the isostadc resid­
ual gravity anomaly is determined by subtracting an isostadc
regional field from the Bouguer gravity anomaly. This
regional field is commonly calculated using an Airy-Heiska-
nen model for local compensation (Simpson et al., 1986).
Using the topographic data to define the surface load and
assuming that a 30-km-thick crust has a 300 kg/m 3 density
conmist with die mande lead to the residual isostadc anom­
aly field shown in Figure 3. From diese data, die gravity lows
associated with the low-density sedimentary and volcanic
basins in the Basin and Range Province have been removed
using a modeling procedure described by Jachens and Mor-
ing (1990) and Saltus and Jachens (1995).
The gravity data (402,384 stadons) were taken from the
Department of Defence (DOD) database (Fig. 4). All data
were ued to the IGSN-71 gravity datum and reduced to com­
plete Bouguer-anomaly values using a reducuon density of
2.670 kg/m 3 and the 1967 formula for the theoretical grav­
ity. The spacing of gravity stations is variable within die study
area. For example, in areas of poor coverage (e.g., eastern
Oregon or northern Arizona), data resoludon is only one
station per 40 km2 or poorer, but in areas of dense coverage
(e.g., southern Nevada and California) one station per 1 to
2 km2 is commonly observed. Within the study area, the
average spacing is 1 station per 6 km2.
Density is a bulk property that varies considerably less
than rock magnetization. Sedimentary rocks have low den­
sities (generally <2,600 k g /m 3) and, thus, result in pro­
nounced gravity lows. Crystalline rocks with typical average
densities of 2,750 kg /m 3 generally produce gravity highs,
except when intruded by felsic plutons (densities of about
2.670 kg/m 3). Prominent gravity highs are generally attrib­
uted to mafic crystalline rocks.
Filtering techniques
An arsenal of interpretive methods exists for potential-
field enhancem ent and modeling (Blakely, 1995; Leaman,
1998). Modeling studies help determine the subsurface con­
figuration in the distribution of density and magnetization.
The interpreter applies all other available geoscience data
(e.g., geology, geochemical data, other geophysical data,
and physical property data) in order to narrow the possible
quantitative solutions characterizing the modeled sources.
Although m odeling sources is beyond the scope of this
review, it is the critical last phase in any derailed mineral
study. Here we emphasize data enhancem ent techniques.
Filter’s are applied to potential-field data to enhance partic­
ular anomaly characteristics, such as wavelength or trend.
For example, we apply wavelength filters to enhance the
expression of major crustal structures, shaded-relief maps to
highlight the trend of shallow sources, and gradient meth­
ods to define magnetization and density boundaries.
Long-wavelength structures: Part of the geophysicist’s goal in
mineral investigations is to characterize major geological
terranes or boundaries that extend deep into the crust. A
simple transformation of the reduced-to-pole (RTP) mag­
netic field (see Fig. 2 caption) to the magnetic potential
(also known as the pseudogravity transformation; Baranov,
1957; Fig. 5) helps alleviate the difficulty of identifying and
isolating broad regional magnetic features partially masked
by high-amplitude, shallow-sourced anomalies and by mag­
netic dipole edge effects. Thus, short-wavelength'magnetic
anomalies in Figure 2 related to sources with shallow roots
are largely absent in Figure 5.
The magnetic potential and residual isostatic gravity
anomaly maps are composed of anomalies from a variety of
sources that range in size and depth. This superposition o f
anomalies can result in interpretational ambiguities. For
example, broad wavelength anomalies (roughly >400 km;
Figs. 3 and 5) overwhelm the contributions o f some impor­
tant short-wavelength features. In o rd er to de-emphasize
both long wavelength (related to sources extending laterally
for several hundred kilometers) and short wavelength
anomalies (related to shallow, relatively thin sources resid­
ing in the upper crust), a match filter (Syberg, 1972) was
applied to highlight intermediate wavelength features (Figs.
6 and 7). The responses from middle-crustal sources lying at
depths roughly greater than 5 km are enhanced. Note that
the separation of anomalies based on rough depth estimates
is not complete because, for example, broad sources within
the shallow crust associated with intense anomalies will be
represented in Figures 6 and 7. The regional anomalies of
interest here appear on the intermediate-wavelength maps
(Figs. 6 and 7) and are observable in the unfiltered data
(Figs. 2 and 3). However, enhancing these interm ediate
wavelengths isolates the magnetic and gravity anomalies
most likely related to major crustal sources (i.e., those that
are relatively thick, deep-rooted, and limited to widths of
less than a few hundred kilometers). This enhancem ent,
thus, focuses and greatly simplifies our study o f regional
controls on mineral deposits.
Magnetization and density boundaries: Three general types of
geophysically defined boundaries are commonly considered:
1. An edge of a large crustal block with relatively uniform
density or magnetization properties that may be expressed
as a large isolated potential-field anomaly (e.g., flanks of a
linear zone of plutons).
2. A linear source with a depth extent much greater than
its width, expressed essentially as a narrow band on a geo­
physical map (e.g., a dike).
3. An edge formed by numerous sources and expressed
as an alignment o f anomalies o f varying intensity and size
 124° 120° 116° 112°
FlG. 3. Isostatic residual gravity anomalies in the western United States. In the Basin and Range Province, the anom­
alies primarily reflect density variations in prc-Cenozoic basement. The gravity effects o f the sedimentary and volcanic
cover deposits in the Basin and Range Province have been removed using tire procedure described byjachens and Mol­
ing (1990). These sedimentary and volcanic cover deposits obscure the gravity effects o f the basement and are related to
tectonic and volcanic processes accompanying widespread Ccnozoic extension. Symbols as in Figure 1.

(e.g., a major fault abruptly terminating various geophysical
terranes).
To emphasize magnetization boundaries, the magnetic
potential is shown as a shaded relief map in Figure 8. The
magnetic potential coupled with the shaded-relief filter pro­
vides a wealth o f inform ation on the regional structural
grain of the western United States.
An area in southern Arizona is used to illustrate two other
approaches to define boundaries. Maxima in the horizontal
gradient o f the gravity field o r the magnetic potential occur
near steep boundaries separating contrasting densities o r
magnetizations, respectively (Fig. 9). The locations of high
horizontal gradients are automatically determ ined and
drawn with the aid of a com puter in the First approach
(Blakely and Simpson, 1986). Such gradient maxima are
 Fig. 4. Gravity stations in western United States.
shown on Figures 6 and 9a (alignment o f very small black tri­
angles and crosses, respectively). The terrace m ethod, the
second boundary enhancem ent technique, is a lithological
or structural mapping tool that defines interpreted rock-unit
boundaries on the basis of local curvature of the gravity field
or the magnetic potential (Grauch and Cordell, 1987;
Cordell and McCafferty, 1989). Terracing utilizes the loca­
tions o f horizontal gradient maxima by simply steepening
the anomaly curve at maxima locations to resemble a step
function. In other words, a terraced m ap contains mathe­
matically derived terranes (areas o f constant density o r mag­
netization) separated by hard-edged boundaries that resem­
ble lithological or structural boundaries on a geological map.
Figure 9b shows a terraced gravity map o f southern Arizona.
Short-wavelength structures: The expression o f near-surface
sources can be enhanced using measured (McMullan and
McLellan, 1997) or calculated gradients (Blakely, 1995; e.g.,
vertical and horizontal gradient maps or shaded-relief maps;
Fig. 8). Here we also use another approach to sharpen the
effects o f near-surface sources. In a small area in southeast
Arizona (Fig. 9a), the magnetic data analytically are upward
continued a small interval (0.4 km) to generate a regional
field (i.e., calculating the magnetic field as if it was mea­
sured 0.4 km above the actual measurement surface). This
regional field o r low-frequency-passed field (because the
m easurem ent surface is farther from the sources) is then
subtracted from the unfiltered data set to derive a residual
or high-frequency-passed field. The unfiltered and residual
 Flo. 5. Magnetic potential map. The transformation o f the magnetic fietd to the magnetic potential emphasizes sources
o f large vertical extent by eliminating many o f the effects o f shallow sources. Units are arbitrary magnetic potential units
(mpu), dimensionally amperes. Symbols as in Figure 1.

magnetic data (Fig. 10a, b) illustrate the effectiveness o f this
approach in highlighting subtle geological information.
These magnetic data and the terraced gravity data (Fig.
lOd) reveal many linear features probably representing
faults. From a close inspection of Figure 10, one can infer
new faults, intersecting faults, and caldera boundaries, and
extend mapped faults.
Bmad geophysical patterns: Simplistic inteipretadonal meth­
ods can invert regional geophysical data to define a single sur­
face separating major changes in density or magnetization.
Although the isostatic gravity correction used to generate
Figure 3 accounts for locally com pensated surface loads
owing to topogtaphy, it fails to consider the possibility of
subcmstal masses that compensate for large geological bod­
ies residing in the crust (e.g., sedimentary basins and intru­
sive zones generally wider than 100 km). Cordell e t al.
(1991) considered the decompensative anomaly to account
for these large geological bodies with no surface topo­
graphic manifestations. By isolating the gravity effects of
sources within different zones o f the lithosphere, Cordel! et
 124° 120° 116° 112°
FlG. 6. Imcrmcdiatc-wavelengih isosialic residua) gra\ity anomaly map. A match filter was applied lo (he isostatic resid­
ual gravity anomaly shown in Figure S to remove the cflccis o f both very broad regional (shallow anti deep) and local
(shallow) gravity sources. The enhanced anomalies in this Figure are interpreted as expressing major regional masses
extending to substantial crustal depths. A heavy- while line shows the frontal thrust licit (Fig. I), because this thrust licit
separates thin-skinned thrust faults to the cast from deep-seated thrust faults to the west, the dec|>*rooicd structures
related to the Rocky Mountain Foreland probably lie west o f the thrust belt and may be related to mincm) deposit clus­
ters in western Montana and northern Utah (Fig. I). To emphasize density boundaries, locations o f horizontal gradient
maxima o f the intermediate-wavelength isostatic residual field are highlighted by very small black uianglcs. Symbols as in
Figure I. BH = Bingham deposit in Utah.

al, (1991) developed a scheme to approximate lithospheric
thickness. Following their interpretational scheme, Hilden-
brand ct al. (2000) calculated the lithospheric thickness in
the western United States (Fig. 11a), assuming a density con­
trast o f -5 0 k g /m 3 across the lithosphere-asthenosphere
boundary. Because many simplifying assumptions are neces­
sary, the absolute values of the calculated thicknesses are
highly unconstrained. Nevertheless, our estimated asthenos-
phere depths compare reasonably well with values based on
seismic velocities (Iyer and Hitchcock, 1989). Within the
 124° 120° 116° 112°
F iu 7. Intermediate-wavelength magnetic potential map. A match filter was applied to the magnetic potential data
shown in Figure 5 to remove the effects of both very broad regional (shallow and deep) and local (shallow) crustal mag­
netic sources. Thus, the enhanced anomalies in this Figure arc interpreted as expressing major regional magnetic sources
extending to substantial crustal depths. 'Id emphasize magnetization boundaries (abrupt lateral changes in magnetic
properties), locations o f selected horizontal-gradient maxima o f the intermediate-wavelength magnetic potential arc high­
lighted by the white dashed lines {black lines in Fig. 2) in areas o f structural zones discussed in the text. BD » Bagdad
deposit, Arizona, Symbols as in Figure 1.

Pacific coastal states, die overestimates in lithospheric thick­
ness based on gravity data are probably related to lower-
crustal dense sources, which are not compensated for in the
calculations. Elsewhere, the results shown in Figure l ia
seem useful in understanding relative changes in lithos­
pheric thickness.
To calculate changes in the thickness of the magnetic
lithosphere, we assume uniform lithospheric magnetic
properties using the interpretive scheme of Cordell and
Henderson (1968). The bottom of the magnetic lithosphere
is usually assumed to occur at the depth of the Curie point
isotherm, the temperature level at which minerals in rocks
lose their magnetic properties (roughly f>50°C for mag­
netite). In the inversion process, the thickness o f a single
magnetic-potential layer is calculated assuming die proper­
ties of the layer and the depth o f either the top o r bottom of
 124° 120° 116° 112°
Fk.. 8. Shadcd-rdicf map o f the magnetic potential shown in Figure 5. Such a map highlights magnetic lineaments.
Illumination direction is from the north. Symbols as in Figure I.

the layer. We assume that the magnetic lithosphere is uni­
formly magnetized and iis top lies at an average depth of0.7
km. Clearly, rock, magnetization varies considerably in the
western United States, where rocks range from highly mag­
netic ophiolttc and serpcntiniie to weakly magnetized meta­
morphosed and altered rocks. Thus, die assumption o f uni­
form magnetic properties is expected to lead to poorly
constrained results. Because die calculated values arc poorly
constrained (Fig. 1lb), only relative changes in the thick­
ness should be considered and interpreted as major
changes in the bulk magnetic properties of die lithosphere.
Nevertheless, remarkable correlations exist between gravity-
defined lithospheric thickness and the calculated thickness
of the magnedc portion of lithosphere (Fig. 11). Such thick­
ness calculations primarily highlight regions where major
lateral changes in the bulk density and magnetic properties
of the lidiosphcre are likely to be present.
Mineralized Areas in the Western United States
D eposit types
The ore deposits and districts portrayed in Figure I are
slighdy modified and updated from the database of known
deposits reported in Long et al. (1998). T he gross values
 (a) Magnetic-anomaly map of southern Arizona

-20 -10 0 10 20 30 40 mGal

(b) Terraced isostatic gravity map of southern Arizona
Fig. 9. a. Magnetic-anomaly map o f southern Arizona. To emphasize magnetization boundaries, locations o f horizon­
tal gradient maxima o f the RTF magnetic field (Fig. 2) arc highlighted by the small black crosses. I>. Terraced residual iso­
static gravity anomaly map o f southern Arizona. Terracing enhances density l>oundaries. In both maps, the northwest and
northeast structural grains are apparent. White dashed rectangles define areas shown in Figures 10 and 16. Orc-deposil
svmltols are circles for pluton-rclatcd anti squares for epithermal deposit types. BD = Bagdad dc|K>sit; TS * Tombstone
deposit. Arrows in the Tombstone anti Bagdad districts highlight northeast and northwest geophysical features shown at
larger scales in Figures 10 and 16, respectively.

mrxmrmrw
 a. Magnetic field b. Residual magnetic field
MOW 110*20* no* 110*40' 110*20' 110s

FlO. 10. a, Magnetic-anomaly map of the Tombstone area in southeast Arizona (location shown in Fig. 9; 1.6 km flight­
line spacing flown at 300 in above terrain), b. Residual magnetic-anomaly map created by subtracting an upward contin­
ued (0.4 km) regional field from the RTP magnetic field (Fig. 2). Small black dots in a and b highlight magnetization
boundaries based on horizontal gradient analysis, c. Isostatic gravity anomaly map showing basement features. Black dots
show gravity stations, d. Terraced isostatic map. Caldera boundaries (heavy dashed lines) taken from Lipman and
Hngstnim (1992) and Moore (1993). Most deposits (c.g., Tombstone deposit, TS) appear to be located near geophysical
anomalies reflecting inferred fault zones, intersecting faults zones, and calderas. These structures may have interacted to
increase permeability allowing magmas to be channeled into the shallow cnist. Ore deposit symbols arc circles for pluton-
relatcd and squares for epithermal deposit types. Two inferred fault zones are highlighted with arrows. DC = Dragoon
caldera; PTC = Parker and Turkey Canyon calderas, TC * Tombstone caldera. Solid lines depict mapped faults (taken
from Wilson etal., 1977).

SI

mum
 (a) Estimated thickness of lithosphere (b) Estimated thickness of the magnetic
based on the inversion of gravity data part of lithosphere
Fig, 11. a. Lithospheric thickness obtained by inversion o f a derived gravity field (following Cordell c t al., 1991)
assumed to represent mass effects deeper than the M oho. Symbols as in Figure 1. Seismically determined estimates o f
lithospheric thickness arc shown with numbers from P-wave models and with numbers within ellipses from Swave models
(Iyer and Hitchcock, 1989). b. Thickness o f the magnetic portion o f the lithosphere is based on the inversion o f magnetic
data. Because many simplifying assumptions arc necessary (c.g., uniform magnetization) in making this inversion, the
poorly constrained data shown here are only useful in understanding relative changes in lithospheric thickness or bulk
magnetic properties. Symbols as in Figure 1.

used to determine the size o f the symbols in Figure 1 and of deposits of this size among all five classes of deposit con­
subsequent figures are based on a compilation of past pro­ sidered in this study.
duction and remaining reserves for those deposits and dis­ For this analysis, we have grouped the deposits into five
tricts (Long c t al., 1998). T he largest symbols represent classes: epitherm al, ptuton-related, sediment-hosted gold,
gross values, based on current metal prices, o f roughly $60 low-sulfide gold-quartz vein, and distal-disseminated silver-
billion; an example is the porphyry copper deposit at Bing­ gold deposits. Deposits belonging to the first three classes
ham Canyon, Utah. Porphyry copper deposits arc the are primarily of Laramide to Holocene age, along with a few
largest, most valuable type o f metallic mineral deposit in the Jurassic porphyry-related deposits. The age o f sediment-
American cordillera. The smallest symbols represent gross hosted (Carlin-type) deposits in the Great Basin is a matter
values of about $20 million; there are numerous examples of continuing debate; although there is a growing consensus
that most deposits are middle Tertiary (Hofstra et al., 1999),
some may be as old as Jurassic. The low-sulfide gold-quartz
vein deposits, primarily in California’s Mother Lode district,
are Cretaceous in age. Deposit types o f the epithermal class
include hot-spring gold-silver, quartz-alunite vein, and
quartz-adularia vein (including Comstock, Creede, and
Sado subtypes; Cox and Singer, 1986). These deposits are
commonly hosted in volcanic rocks and are generally asso­
ciated with subvolcanic intrusive activity; they are most abun­
dant in the Great Basin. Deposit types in the pluton-related
class include porphyry copper (including copper-molybde­
num and copper-gold, as well as skarn-related copper),
skarn gold, skarn lead-zinc, polymetallic replacement, and
polymetallic vein deposits. These deposit types are com­
monly associated with mesozonal-to-epizonal, intermediate-
to-felsic plutons, found throughout the cordillera. The dis­
tal-disseminated silver-gold deposits include both those
defined originally by Cox (1992) and the gold-rich deposits
discussed by Theodore (1998). Although in specific
instances these deposits may be difficult to distinguish from
sediment-hosted gold deposits, they are believed to be fun­
damentally a product of igneous activity. Sediment-hosted
gold deposits lack a dem onstrated magmatic affiliation,
although opinion is far from unanimous on this point. In
the United States, they are geographically limited to a small
area in northeast Nevada and adjacent parts of Utah and
Idaho, and are unknown in areas with many porphyry-
related deposits such as southern Arizona. Low-sulfide gold-
quartz veins are included in this study because of their clear
association with prominent geophysical features. Unlike the
other deposit types, the veins in the Mother Lode occur in a
long linear belt that is closely coincident in time and space
with Mesozoic accretion along the western cordllleran mar­
gin. The veins, mainly in shear zones in volcanic and vol­
cano-sedimentary rocks, form ed during subsequent
regional metamorphism that accompanied the intrusion of
the major part o f the Sierra Nevada batholith exposed a few
tens o f kilometers to the east o f the deposits.
Deposit distribution
The complex geological evolution o f the western United
States provided favorable environm ents for the episodic
concentration o f metals over the past 3 b.y., but here we
focus on Mesozoic and Cenozoic deposits and their rela­
tionship to regional geophysical features. A fundamental
problem encountered in conducting any geophysical-ore
deposit study in the western United States is defining a rea­
sonable representation o f the distribution o f mineral
deposits. T he geophysical data sets, in general, present a
continuous image of the distribution o f density and magne­
tization. In contrast, most known ore deposits are discrete
entities generally confined to areas o f exposed bedrock.
Because most o f the western United States is covered by sed­
imentary and volcanic rocks, many covered areas must con­
ceal undiscovered mineral deposits. Therefore, the popula­
tion o f known deposits provides an incomplete information
base with which to interpret the interrelationships between
geophysical features and deposits. Nevertheless, we assume
here that the present distribution o f known ore deposits is
representative o f the actual distribution o f both known and
undiscovered deposits on a regional scale.
Another problem in understanding the regional controls
for mineralization is that the Mesozoic and Cenozoic metal-
logeny o f the western United States is related to continu­
ously evolving magmatic, stratigraphic, and tectonic systems.
We limit our discussion to particular deposit types, primarily
those form ed from Laramide time (80-40 Ma) to the
Holocene (Fig. 1). Even for this limited set o f deposit types
and ages, our discussion will not be comprehensive owing to
the large volume o f deposit data and to the expanse of the
study area. These data nevertheless reveal some interesting
examples of relationships between mineral deposits and
inferred regional crustal features in the western United
States.
During Laramide time, a calc-alkaline magmatic arc
related to subduction along the Pacific Coast shifted east­
ward owing to an acceleration and change in direction of
plate motion. The subducting plate apparently flattened
(dip -20°), resulting in uplift and magmatism as far inboard
as 1,500 km from the trench (Lipman e t al., 1972; Coney,
1978). Associated northeast-southwest-dirccted regional
compression accom panied em placem ent o f plutons and
porphyry and vein deposits throughout the region from
Idaho to Arizona (Hutchinson and Albers, 1992).
From 80 to 55 Ma, calc-alkaline magmatism and related
m ineralization occurred primarily in the northern and
southern cordillera (e.g., porphyry copper deposits near
Butte, M ontana, and porphyry copper and polymetallic
deposits o f Arizona, respectively), although deposits are
also scattered throughout the southwestern United States
to the Rocky M ountain front (Fig. 1). From 55 to 40 Ma,
additional magmatic activity occurred in the northern
Cordillera (e.g., the T hunder Mountain district in Idaho).
Although the shallowing subduction plate hingeline ini­
tially moved eastward, the dip o f the plate later steepened,
resulting in a southwestward retreat o f the hingeline
(-40-22 Ma; Coney, 1978). O ne o f the world’s largest cop­
per deposits, the 36 Ma Bingham porphyry copper deposit
in Utah, formed during this period o f steepening subduc­
tion (James, 1978). From late Tertiary to Quaternary (22-0
Ma) most epithermal districts formed during and after the
transition from a com pressional to extensionai tectonic
regime, which led to die formation o f the Great Basin. Dur­
ing this transition, the most important epithermal deposits
in western Nevada were form ed in near-surface volcanic
rocks (Ludington et al., 1993).
The absence o f deposits in certain geological terranes is
apparent. For example, the Colorado Plateau lacks impor­
tant metallic hydrothermal mineral deposits, perhaps due to
ovcrthickcncd crust that im peded magmatism (Coney,
1976). In post-Laramide time, the Columbia River flood
basalts have likely concealed older mineral deposits in Ore­
gon and Washington. In southern Nevada in an area
roughly between 36* and 37*N known as the “amagmatic
corridor,” mineralized rocks, Cretaceous plutons, and Ter­
tiary volcanic rocks are rare (see Stewart, 1980).
 Relationships of Geophysical Structures
and Deposit Patterns
Comparisons of the spatial distribution o f mineralized
systems and the locations of geophysically defined lithotec-
tonic crustal features have contributed to a better under­
standing of deep-seated regional controls on mineral con­
centrations (e.g., Gunn et al., 1997a, b; Jaques et al., 1997;
Leclair et al., 1997; Moore et al., 1998). For example,
Kutina and H ildenbrand (1987) noted that some ore
deposits in the western United States conspicuously cluster
within pronounced gravity lows or near the flanks of promi­
nent gravity highs. Their regional study concluded that the
interpretation of gravity data to characterize the regional
mass distribution in the crust constitutes a useful explo­
ration tool for defining low-density felsic intrusions and tec­
tonic zones possibly related to the regional controls on
mineralization. Grauch et al. (1995) identified a relation­
ship between a basement gravity lineament and die Battle
Mountain-Eureka mineral belt in Nevada. They suggested
that the gravity lineament expresses a boundary separating
major crustal blocks, possibly reflecting a Jurassic or pre-
Jurassic fault. This boundary would have subsequently influ­
enced the localization of hydrothermal activity. Blakely et
al. (1991) carried out a statistical proximity study of sedi­
m ent-hosted gold and distal-disseminated silver-gold
deposits in Nevada to gravity-defined boundaries, and pro­
posed that these deposits are preferentially located near
prom inent crustal boundaries. Hildenbrand et al. (2000)
expanded on these earlier regional geophysical studies uti­
lizing upgraded m agnetic and gravity databases and fil-
tered-anomaly maps of the western United States. Here we
review the results of these earlier studies and provide some
additional insights.
To investigate the regional controls on mineral deposits,
we first discuss the broad relationships between geophysical
features and the structural geology important to the local­
ization of some deposits. Then we turn to specific linear
arrays and clusters of deposits to discuss possible deep-
seated, regional controls on ore deposition. Here, die geo­
physical data indicate that the pull-apart structural model
may be related to the localization of many deposits regard­
less of their type or age. Lastly, the utility of high-resolution
potential-field data is highlighted in detailed ore-deposit
studies. In particular, we demonstrate the ability of geophys­
ical data to help focus exploration in complex geological
regimes by defining deep-crustal fracture systems or intru­
sive zones, which represent potentially favorable environs
for mineralization.
O v e ra ll deposit p atte rn s in. the western U nited States
(scale ~ 1 :5 ,0 0 0 ,0 0 0 )
R egional geophysical patterns: O n the broadest scale, the cal­
culated thicknesses of gravity-defined lithosphere and mag­
netic lithosphere are thinnest in regions of high heat flow,
such as the Great Basin and Yellowstone National Park
(Morgan and Gosnold, 1989). In contrast, a zone of low
heat flow, extending from eastern Utah northward into
Wyoming, is characterized by thick gravity-defined lithos­
phere and magnetic lithosphere.
The apparent coincidence of thin gravity-defined lithos­
phere (Fig. 11a) and magnetic lithosphere (Fig. lib ) may
simply result from shallow asthenosphere leading to higher
heat flow and, thus, a shallow Curie point isotherm. Thin
magnetic basement produces low regional magnetic inten­
sities (Blakely, 1988). Thin lithosphere might also be associ­
ated with voluminous magmatism and consequent
hydrothermal alteration, possibly leading to the destruction
of magnetic properties (e.g., alteration of m agnetite to
hematite or pyrite). Interestingly, mineral deposits tend to
be located in areas of thinner magnetic lithosphere.
Although correlation of regionally thin gravity-defined
lithosphere and magnetic lithosphere may be reasonable,
their correlation with post-Jurassic mineral deposits is enig­
matic, especially because the lithospheric thickness at vari­
ous times of prior mineralization may have been signifi­
cantly different from its thickness today. For example, many
metallic deposits were formed during the Mesozoic to late
Tertiary subduction o f the Farallon plate beneath N orth
America. By the Late Cretaceous, back-arc tectonism had
resulted in a thick entst through overthrusting over a large
region (50-60-km thickness; Miller et al., 1992). Beginning
in the Eocene in eastern Nevada, the ductile lower crust
began to extend through delam tnation (Zoback e t al.,
1981), and the crust had been thinned 30 to 50 percent by
the early late Miocene. From the Eocene to the Miocene,
rapid delam ination in the ductile lithosphere was appar­
ently manifested in the brittle crust by strike-slip faulting
(Berger et a l, 1998; Bourne et al., 1998; B. Berger, U.S. Geo­
logical Survey, unpub. data, 2000; Drew and Berger, in
press). This style o f faulting, for example along the Carlin
and Getchell mineral u ends, accom panied delam ination
due to traction forces across the brittle-ductile transition.
Extensional faulting did not dom inate the surficial land­
scape of the region until approximately 20 to 18 Ma in the
southern Basin and Range, and as late as 11 to 10 Ma in the
northern Basin and Range.
Mineral deposits were formed during all stages of this tec­
tonic evolution. During the Late Cretaceous to early Eocene
period of overthickened crust in the western United States,
a num ber of world-class mining districts were formed. Dur­
ing the uppermost Late Cretaceous and earliest Tertiary, the
polymetallic veins and porphyry copper deposits at Butte,
Montana, were emplaced (Snee et al., 1999), whereas most
porphyry copper deposits of Arizona and New Mexico were
formed in the Paleocene to earliest Eocene (Titley, 1995).
As the crust began to thin, the giant Carlin-stylc gold
deposits in north-central Nevada probably formed during
the Eocene to early 01igocene(?) (e.g., Hall et al., 1997;
Hofstra et al., 1999). By the middle Miocene the crust was
considerably thinner and the world-class Comstock Lode
and Goldfield epitherm al deposits form ed in western
Nevada (cf. SeedorfF, 1991). The net result is that most west­
ern United States ore deposits (Fig. 1) tend to be now
located within broad regions characterized by inferred thin­
ner magnetic lithosphere (Fig. l ib ) and low magnetic
 potential (Fig. 7). However, like the crust, the lithosphere
thickness changed during die course of die Cenozoic; con­
sequently, proposing a general genetic relationship between
mineralization and thin lithosphere is problematic- Perhaps
in som e local areas w here gravity-defined lithosphere and
m agnetic lithosphere are currently characterized by small-
to-im erm ediate thicknesses, deep faults related to m ineral­
ization w ere reactivated d u rin g e x te n sio n (i.e., existing
faults related to m ineralization possibly facilitated a n d local­
ized la te r e x te n sio n ). Clearly, areas largely u naffected by
m agm atism since Precam brian tim e a n d presently charac­
terized by thick lithosphere (Fig. 11a) are nearly devoid of
post-Jurassic m in eralized system s, su ch as th e C o lo rad o
Plateau, c entral W yoming, a n d central M ontana.
Such speculations at continental scales based on the inter­
pretation of magnetic and gravitydata may be useful in lim-
itingmincrai exploration to very large regions (e.g., possibly
restricting exploration to southern Arizona and eliminating
large regions in northern Arizona). For the type of deposits
being considered here, more fruitful studies to limit the
area in evaluating mineral potential are likely when a closer
inspection of the regional geophysical data is carried out,
even at scales o f 1:5,000,000. For example, the magnetic
data show that 90 percent of the deposits considered here
correlate with low intensities (<0.0 mpu) in the intermedi­
ate-wavelength magnetic potential (Fig. 7), although the
ratio of the area of negative magnetic potential values to
total area i s about 67 percent (Hildenbrand et al., 2000).
Fta. 12. Bouguci gravity map of south-central Australia, Note that the Olympic Dam deposit is located a u h e intersec­
tion o f 2 major gravity lineaments. (Taken from O'Driscoll, 1990; reprinted with permission of the Australasian Institute
o f Mining and Metallurgy)
Thus, regional magnetic lows may be a useful indicator of
favorable geological environments for mineral occurrence.
This tendency for deposits to b e associated with m agnetic
potential lows in th e w estern U n ited States w arrants addi­
tional study. In C anada, regional m agnetic lows a re also
im portant in m ineral exploration because greenstone belts,
the location o f m any ore deposits, are deficient m m agnetite
(Grant, 1985a, b).
Intersection of regional structures: Geophysical lineaments
defined at regional scales are also useful in mineral explo­
ration. Billingsley and Locke (1941), Lowell (1974), Sillitoe
(1975), and Kurina and Hildenbrand (1987) proposed an
influence of crustal lineament intersections in the distribu­
tion of mineral deposits. A classical example in southern Aus­
tralia is the giant Olympic Dam deposit (Fig. 12) located at
the intersection of major gravity discontinuities (O’DriscoH,
1990). In reviewing the geological framework of the Olympic
Dam deposit, Woodall (1998, p . 9) states, “We need to b e
more aware of the effects of d e e p basem ent shears. Deep
structures penetrate the Earth's crust and upper mantle and
may act as the plumbing system for fluid and heat entering
the lithosphere. Em pirical evidence indicates th a t such
su’uctures control many of the giant m ineral deposits..."
An analog in northern Utah may be deep-basement frac­
tures associated with the cluster of ore deposits at Bingham,
one of the world’s largest copper districts. East-west gravity
and magnetic trends east of Bingham (Erickson, 1976) on
Figures 6 to 8 delineate features paralleling the east-west-
 Relationships o f Geophysical Structures
and Deposit Patterns
Comparisons o f the spatial distribution o f mineralized
systems and the locations of geophysically defined lithoiec-
tonic crustal features have contributed to a better under*
standing o f deep-seated regional controls on mineral con­
centrations (e.g., Gunn et al., 1997a, b; Jaques et al., 1997;
Leclair e t al., 1997; M oore e t al., 1998). For example,
Kutina and H ildenbrand (1987) noted that some ore
deposits in the western United States conspicuously cluster
within pronounced gravity Iowa o r near the flanks of promi­
nent gravity highs. Their regional study concluded that the
interpretation o f gravity data to characterize the regional
mass distribution in the crust constitutes a useful explo-
radon tool for defining low-density felsic intrusions and tec­
tonic zones possibly related to the regional controls on
mineralization. Grauch e t al. (1995) identified a relation­
ship between a basement gravity lineament and the Battle
Mountain-Eureka mineral belt in Nevada. They suggested
that the gravity lineament expresses a boundary separating
m ajor crustal blocks, possibly reflecting a Jurassic o r pre-
Jurassic fault This boundary would have subsequently influ­
enced the localization of hydrothermal activity. Blakely et
al. (1991) carried out a statistical proximity study o f sedi­
m ent-hosted gold and distal-disseminated silver-gold
deposits in Nevada to gravity-defined boundaries, and pro­
posed that these deposits are preferentially located near
prom inent crustal boundaries. Hildenbrand et al. (2000)
expanded on these earlier regional geophysical studies uti­
lizing upgraded m agnetic and gravity databases and fil-
tcred-anomaly maps of the western United States. Here we
review the results of these earlier studies and provide some
additional insights.
To investigate the regional controls on mineral deposits,
we first discuss the broad relationships between geophysical
features and the structural geology important to the local­
ization of some deposits. Then we turn to specific linear
arrays and clusters of deposits to discuss possible deep-
seated, regional controls on ore deposition. Here, the geo­
physical data indicate that the pull-apart structural model
may be related to the localization of many deposits regard­
less of their type or age. Lastly, the utility of high-resolution
potential-field data is highlighted in detailed ore-deposit
studies. In particular, we demonstrate the ability of geophys­
ical data to help focus exploration in complex geological
regimes by defining deep-crustal fracture systems or intru­
sive zones, which represent potentially favorable environs
for mineralization.
O v e ra ll deposit p atte rn s in the western U nited States
(scale ~ 1 :5 ,0 0 0 ,0 0 0 )
R egional geophysical patterns: On the broadest scale, the cal­
culated thicknesses of gravity-defined lithosphere and mag­
netic lithosphere are thinnest in regions of high heat flow,
such as the Great Basin and Yellowstone National Park
(Morgan and Gosnold, 1989). In contrast, a zone of low
heat flow, extending from eastern U tah northw ard into
Wyoming, is characterized by thick gravity-defined lithos­
phere and magnetic lithosphere.
The apparent coincidence o f thin gravity-defined lithos­
phere (Fig. 11a) and magnetic lithosphere (Fig. lib ) may
simply result from shallow asthenosphere leading to higher
heat flow and, thus, a shallow Curie point isotherm. Thin
magnetic basement produces low regional magnetic inten­
sities (Blakely, 1988). Thin lithosphere might also be associ­
ated with voluminous magmausm and consequent
hydrothermal alteration, possibly leading to the destruction
o f m agnetic properties (e.g., alteration o f m agnetite to
hematite o r pyrite). Interestingly, mineral deposits tend to
be located in areas of thinner magnetic lithosphere.
Although correlation of regionally thin gravity-defined
lithosphere and magnetic lithosphere may be reasonable,
their correlation with post-Jurassic mineral deposits is enig­
matic, especially because the lithospheric thickness a t vari­
ous times o f p rio r m ineralization may have been signifi­
cantly different from its thickness today. For example, many
metallic deposits were formed during the Mesozoic to late
Tertiary subduction o f the Farallon plate beneath N orth
America. By the Late Cretaceous, back-arc tectonism had
resulted in a thick crust through overthrusting over a large
region (50-60-km thickness; Miller etal., 1992). Beginning
in the Eocene in eastern Nevada, the ductile lower crust
began to extend through detam inaiion (Zoback e t al.,
1981), and the crust had been thinned 30 to 50 percent by
the early late Miocene. From the Eocene to the Miocene,
rapid delam ination in the ductile lithosphere was appar­
ently manifested in the brittle crust by strike-slip faulting
(Berger et al., 1998; Bourne e t al., 1998; B. Berger, U.S. Geo­
logical Survey, unpub. data, 2000; Drew and Berger, in
press). This style o f faulting, for example along the Carlin
and Getchell mineral trends, accom panied delam ination
due to traction forces across the brittle-ductile transition.
Extensional faulting did not dom inate the surficial land­
scape of the region until approximately 20 to 18 Ma in the
southern Basin and Range, and as late as 11 to 10 Ma in the
northern Basin and Range.
Mineral deposits were formed during all stages of this tec­
tonic evolution. During the Late Cretaceous to early Eocene
period of overthickened crust in die western United States,
a num ber of world-class mining disu'icts were formed. Dur­
ing the uppermost Late Cretaceous and earliest Tertiary, the
polymetallic veins and porphyry copper deposits at Butte,
Montana, were emplaced (Snee et al., 1999), whereas most
porphyry copper deposits of Arizona and New Mexico were
formed in the Paleocene to earliest Eocene (Titley, 1995).
As the crust began to thin, the giant Carlin-style gold
deposits in north-central Nevada probably formed during
the Eocene to early 01igocene(?) (e.g., Hall et al., 1997;
Hofstra et al., 1999). By the middle Miocene the crust was
considerably thinner and the world-class Comstock Lode
and Goldfield epitherm al deposits form ed in western
Nevada (cf. Seedorff, 1991). The net result is that most west­
ern United States ore deposits (Fig. 1) tend to be now
located within broad regions characterized by inferred thin­
n er m agnetic lithosphere (Fig. l ib ) and low magnetic
potential (Fig. 7). However, like the crust, the lithosphere
thickness changed during the course o f the Cenozoic; con­
sequently, proposing a general genetic relationship between
mineralization and thin lithosphere is problematic. Perhaps
in some local areas where gravity-defined lithosphere and
magnetic lithosphere are currently characterized by small-
to-intermediate thicknesses, deep faults related to mineral­
ization were reactivated during extension (i.e., existing
faults related to mineralization possibly facilitated and local­
ized later extension). Clearly, areas largely unaffected by
magmatism since Precambrian time and presently charac­
terized by thick lithosphere (Fig. 1 la) are nearly devoid of
post-Jurassic mineralized systems, such as the Colorado
Plateau, cenu-al Wyoming, and central Montana.
Such speculations at continental scales based on the inter­
pretation of magnetic and gravity data may be useful in lim­
iting mineral exploration to very large regions (e.g., possibly
restricting exploration to southern Arizona and eliminating
large regions in northern Arizona). For the type of deposits
being considered here, m ore fruitful studies to limit the
area in evaluating mineral potential are likely when a closer
inspection of the regional geophysical data is carried out,
even at scales o f 1:5,000,000. For example, the magnetic
data show that 90 percent of the deposits considered here
correlate with low intensities (<0.0 m pu) in the intermedi­
ate-wavelength magnetic potential (Fig. 7), although the
ratio o f the area o f negative magnetic potential values to
total area is about 67 percent (Hildenbrand et al., 2000).
Fic. 12. Bougiicr gravity map o f south-central Australia. Note that the Olympic Dam deposit is located at the intersec­
tion o f 2 major gravity lineaments. (Taken from O ’DriscoII, 1990: reprinted with permission o f the Australasian Insdtutc
o f Mining and Metallurgy)
Thus, regional magnetic lows may be a useful indicator of
favorable geological environments for mineral occurrence.
This tendency for deposits to be associated with magnetic
potential lows in the western United States warrants addi­
tional study. In Canada, regional magnetic lows are also
important in mineral exploration because greenstone bells,
the location of many ore deposits, are deficient in magnetite
(Grant, 1985a, b).
Intersection ofregiotial structures: Geophysical lineaments
defined at regional scales are also useful in mineral explo­
ration. Billingsley and Locke (1941), Lowell (1974), Sillitoe
(1975), and Kutina and Hildenbrand (1987) proposed an
influence of crustal lineament intersections in the distribu­
tion o f mineral deposits. A classical example in southern Aus­
tralia is the giant Olympic Dam deposit (Fig. 12) located at
the intersection of major gravity discontinuities (O’Driscoll,
1990). In reviewing the geological framework of die Olympic
Dam deposit, Woodall (1993, p. 9) states, “Wc need to be
more aware of the effects o f deep basement shears. Deep
structures penetrate the Earth’s crust and upper mantle and
may act as the plumbing system for fluid and heat entering
the lithosphere. Empirical evidence indicates that such
structures control many o f the giant mineral deposits..
An analog in northern Utah may be deep-basement frac­
tures associated with the cluster o f ore deposits at Bingham,
one o f the world's largest copper districts. East-west gravity
and magnetic trends east o f Bingham (Erickson, 1976) on
Figures 6 to 8 delineate features paralleling the east-west-
uendingArchean-Proterozoic boundary (Fig. 1). Near Bing­
ham, north-northeast-trending gravity anomalies roughly
follow the boundary (frontal thrust belt, Fig. 6) between car­
bonate platformal rocks to the east (characterized by thin-
skinned thrust faults) and Paleozoic miogcoclinal slope-
facies rocks to the west (characterized by deep-seated thrust
faults; Miller e t a t , 1992). This region of intersecting geo­
physical features also includes nonhwest-trending gravity
and m agnetic lineam ents o f unknown geological origin
(Figs. 6 and 7). Billingsley and Locke (1941) nam ed this
area the “Salt Lake crossroads” (Fig. 1), where ancient inter­
secting fracture systems may have cut the crust and acted as
channelways for plutons and metal-bearing fluids. Whatever
the specific origins o f m ineral clusters, their correspon­
dence with linear magnetic and gravity features (Figs. 6 and
7) supports an origin related to deep-seated structural dis­
continuities. Below, we review how die diverse structural ele­
ments work together to localize mineral deposits.
Mineral bells (scale -1:500,000)
Linear arrays o f deposits, generally called “belts" or
“trends,” have long been the focus o f study by economic
geologists. For example. Card and Poulsen (1998), in study­
ing the m ineral deposits o f the Superior Province o f the
Canadian Shield, describe num erous clearly defined belts
characterized by deposits o f diverse metal associations and
genetic types. Many Superior Province gold deposits tend to
cluster near large faults and shear zones, sometimes
referred to as deformation corridors o r tectonic zones due
to their relationship to intensely deform ed rock in zones
several kilometers wide (see figs. 2.25, 2.50, and 2.52 in Card
and Poulsen, 1998). Also, most die gold deposits ofVictoria,
Australia, lie along linear structural zones within a 70-km-
wide belt in the Lachlan fold belt (Maher et al., 2000; Van-
derB ergetal., 2000).
Several m ineral belts in the western U nited States are
explored here with respect to their relationships to geo­
physical lineaments at scales of roughly 1:500,000.
Mother Lode gold belt: Brittle-ductile mesozonal hydrother-
mal gold deposits are localized in major fault zones near
sutured plate margins, and regional geophysical data are
helpful in defining such favorable areas. For example, a
group o f low-sulfide gold-quartz vein deposits, commonly
term ed “M other Lode” deposits (star symbols in NE Cali­
fornia, Fig. 1), arc hosted mostly by metamorphic rocks in
a series o f accreted volcanic arc and other oceanic terranes
in the western foothills o f the Sierra Nevada and the adja­
cent Klamath M ountains, California (Landefeld, 1990;
H utchinson and Albers, 1992). A few districts occur in
Mesozoic plutons that intrude the metamorphic rocks and
are part of the Sierra Nevada batholith, primarily exposed
a few tens o f kilom eters to the east. T he M other Lode
deposits in the Sierra Nevada are closely associated with
regional-scale, curvilinear fault zones (Melones, Bear
Mountains, and related fault zones tens to hundreds o f kilo­
meters long; Landefeld, 1990; Ashley, 1991),some of which
may be reactivated suture zones first developed during
accretion o f the volcanic arc terranes. Gold may have been
leached and transported during high-tem perature Meso­
zoic m etam orphism , and precipitated in quartz veins in
greenschist-grade metamorphic rocks and syndeformation
intrusions.
The Mother Lode districts correlate with geophysical fea­
tures at both regional and sub-regional scales. They lie along
a well-defined north-northwest-trending gravity high (Figs.
3,6, and 13b) caused by mafic metavolcanic and dense Cre­
taceous plutons o f the western Sierra Nevada batholith.
These igneous rocks likely intruded slabs o f dense accreted
oceanic crust, from which they may have inherited their
dense mafic character (Jachens and Griscom, 1985). Mag­
netically, the districts generally lie within a region o f low-to-
m oderate m agnetic field strength typical o f only slightly
m agnetic rocks, bu t n ear long, narrow, high-am plitude
curvilinear magnetic highs (Figs. 2 and 13a). The long, nar­
row magnetic highs are caused by tabular bodies o f serpen-
tinite which lie along and define many o f the regional-scale
LateJurassic fault zones that are associated with the deposits
(H enderson e t al., 1966; U.S. Geological Survey, 1969,
1976). T he low-to-moderate background magnetic field
strength over die metamorphic terranes hosting the ore dis­
tricts is consistent with the metamorphic grade o f the host
rocks because greenschist metamorphism normally destroys
the magnetite in these rocks (Grant, 1985a; Clark, 1997b).
The predom inant lithologies o f the host rock include: (1)
harzburgite, pyroxenite, and dunite; (2) basaltic pillow lava
and breccia; (3) basaltic-andesitic lava; (4) debris How and
turbidite deposits derived from the basaltic-andesitic lava;
(5) epiclastic conglomerate, mixed volcaniclastic-epiclastic
wacke, sandstone, and siltstone; (6) hemipelagic organic,
carbon-rich argillite; and (7) mafic dikes (Landefeld, 1990).
Greenschist facies metamorphism is characteristic of low-sul­
fide gold deposit environments worldwide (e.g., Kerrich,
1999), thus, providing a direct logical link between the
M other Lode deposits and their associated regional mag­
netic signature. In the Canadian Superior Province,
Archean greenstone belts have characteristic magnetic sig­
natures (Grant, 1985b) similar to those over the M other
Lode belt—namely, regional lows interrupted with linear,
local, high-amplitude anomalies. Grant (1985b) pointed out
that favorable ore environs within these greenstone belts are
related to permeable tectonic zones sometimes character­
ized by linear magnetic features, where the quantity of mag­
netite has been enhanced (e.g., through serpentinization)
o r diminished (e.g., through hydrothermal alteration). An
example is the group of gold and base metal deposits along
the 400-km-long Porcupine-Destor-Qucuco fault system in
Ontario and Quebec.
As described above, the M other Lode districts have a
characteristic three-part geophysical signature consisting of
a regional gravity high, a regional low-to-moderate back­
ground magnetic field anomaly, and long curvilinear mag­
netic highs. Hildenbrand et al. (2000) suggested that these
characteristics may serve as an exploradon guide for similar
deposits in nearby areas, especially areas where oceanic
crust is likely to exist o r m ight have existed in the past. Of
particular interest in Figure 13b are the gravity highs imme-
 122® 120° 1 2 2 ° 120®

a. RTP magnetic field b. Filtered isostatic gravity

Fici. 13. RTP magnetic-anomaly field (a) and intermediate-wavelength, rr^irfual isnstaiie graviiy anomaly PHd (b) over
the M other Lode gold belt in the western foothills o f the Sierra Nevada, California. T he Mesozoic Mother Lode gold belt
displays characteristic geophysical signatures (regional gravity high, regional low-co-modcratc background magn«»iir field
anomaly, long curvilinear magnetic highs) that might s e n e as an exploration guide. The Mother Lode deposits are closely
associated with regional-scale, curvilinear fault zones, som e o f which may b e reactivated suture zones first developed dur­
ing accretion o f the volcanic arc ternm es. Black lines show faults (taken from Jennings e l al„ 1977). Stars locate deposits
o f the M other Lode.

diately west of the Mother Lode gold belt (also west of lesser-
valued gold deposits not shown in Fig. 13). Here, accreted
volcanic arc terranes similar to those hosting the Mother
Lode deposits descend beneath younger sedimentary rocks
of the Great Valley and are likely intruded by Cretaceous
plutons. Detailed examination of the local magnetic field
over the eastern margin o f the Great Valley for areas of low
background magnetic field interrupted by linear magnedc
highs could be an effective means o f focusing a search for
buried Mother Lode type deposits.
Magmatic-hydrothermal mineral (porphyry) deposit bends: Mag­
matic-hydrothermal mineral deposits (e.g., porphyry copper
deposits) are frequently localized along m ajor strike-slip
fault zones within magmatic-arc environments (Nishiwaki,
1981, Drew and Berger, in press). An explanation is that
strike-slip fault zones provide localized zones of high per­
meability, thereby allowing the necessary focusing o f mag-
matism and fluid flow to form hydrothermal ore deposits.
Releasing bends o r extensional stepovers along strike-slip
fault zones (Dooley and McCIay, 1998) commonly localize
magmas in the shallow crust. A schematic model illustrating
the possible relationships between mineral deposits, faults,
grabens, and intrusions is shown in Figure 14. The exten-
sional stepover model may be important in understanding
the structural setting o f many deposits in western United
States (Berger et al., 1998; B. Berger, U.S. Geological Survey,
unpub. data, 2000.
In applying such a model, geophysical data are useful for
defining the deep-seated regional fracture zones important
to the intrusion o f magmas in the crust and the delineation
of near-surface volcanic features such as dikes or calderas.
O f particular importance are intersecting fracture trends
where permeability' may be enhanced. For example, Henley
and Adams (1992), in investigating linear trends of epither­
mal gold deposits in the Drummond Basin, Australia, sug­
gest that mineralization is related to local pull-apart basins
94 H lW E N liliA M ) E T A L .

b Intrusion in hinge C Intrusion within

of releasing bend pull-apart basin Extension

Fid. 14. a. Schematic model o f a duplex o f faults that make up a pull-apart giabcn or basin (releasing bend) between
two interacting cn Echelon right-lateral strike-slip faults; cross-basin strike-slip faults accommodate strain across the bend
when extension terminates (after Dooley and McQay, 1998). b. Form o f igneous intrusions in a hinge zone into the pull-
apart basin. The composite intrusive bodies tend to take the form o f a vertically elongated cylinder (B. Berger, U.S- Geo­
logical Survey, unpub. data, 2000), c. Form o f igneous intrusions within the center o f die pull-apart structure (B. Berger,
U.S. Geological Survey, unpub. data, 2000).

along linear reactivated strike-slip fault zones (identified in 15; O ’Neill and Lopez, 1985). Schmidt e t al. (1990) pro­
high-resolution aeromagnetic data). To illustrate the utility posed that the Boulder Batholith was em placed within a
of the pull-apart basin model and o f geophysical data in the local pull-apart basin between two east-westerly transfer
application o f this model in studying magmatic-hydrother­ faults active as the regional thrust front pushed east (Fig.
mal mineral deposits in die western United States, we pro­ 15a). When extension within the pull-apart could no longer
vide three examples related to two porphyry-copper deposit accommodate strain in response to the regional east-west
trends: the zone of porphyry deposits in the Idaho-Montana maximum principal stresses, northeast-striking, right-lateral
porphyry belt, and in southern Arizona (Fig. 1). strike-slip faults developed. Applying the structural model of
Example 1: A zone o f northeast-trending magnetic fea­ Figure 14, we propose that an extensional stepover between
tures from east-central Idaho across Montana to the Cana­ two e n echelon strike-slip faults controlled the emplace­
dian border (Figs. 7 and 8) contains a num ber of porphyry m ent of m ineralization-related intrusions at Butte (Fig.
molybdenum deposits including Cannivan Gulch and 15b). T he Butte district stepover appears to be located
Thompson Creek (Fig. 1). This alignm ent o f deposits where a reactivated, northwest-striking, Proterozoic fault sys­
defines the Idaho-Montana porphyry belt o f Rostad (1978), tem (Schmidt et al., 1990) caused the primary cn echelon
itself within the much longer, 1,500-km-long, Great Falls tec­ northeast-striking faults to elastically interact and thereby
tonic zone (O’Neill and Lopez, 1985; Fig. 1). form linking faults (Fig. 15b; Fig 14). Ore-bearing veins at
The zone of northeast-trending magnetic lineaments is Butte occur along the northeasi-suiking, northwest-striking,
coincident with a zone o f strike-slip faults, including the set and stepover faults.
that may have localized the Late Cretaceous Boulder Example 2: In southern Arizona, several porphyry copper
Batholith and the early Tertiary Butte mining district (Fig. deposits are localized along northwest-striking, right-lateral
 FlG. IS. a. Residua) Ki r magnetic anomaly map o f the Boulder batholiih area, Montana. These residual data were gen*
crated by subtracting an upward continued (1.0 km) regional field from the data in Figure 2. Pinion-related and cpilhcr*
mal deposits are shown as white circles and squares, respectively. Faults (narrow black lines) and the boundary o f the Butte
quartz monzonitc pluton (magenta dashed line), largely representing the boundary o f the Boulder batholiih, arc taken
from Smcdcs ct al. (1988). Heavy black lines show the proposed model o f emplacement o f the batholiih (located near B
in Fig. 1) in the pull-apart region during compressional tectonics due to cast-northeast segmentation o f the advancing
thrust sheet (Schmidt et a!.. 1990). b. Pattern o f faults in the Butte mining district (after Sales. 1913) within tltc Boulder
batholiih demonstrating (he localization o f porphyry-style copper deposits in a releasing bend along a right-lateral strike-
slip fault (highlighted with arrows).
fault /.ones where they step right due to interactions with
reactivated, older northeast-striking fault zones (B. Berger,
U.S. Geological Survey, unpub. data, 2000; Drew and
Berger, in press). These im portant northwest structural
zones, which are reflected in the magnetic and gravity data
(Figs. 6-10), may delineate the primary fault zones along
which porphyry copper deposits can be localized. The
northwest-trending geophysical anomalies, locally inter­ v
rupted by northeast-trending features, coincide with the
650-km-long zone of copper-bearing plutons (Fig. 1)
extending from northern Mexico to west-central Arizona,
sometimes called the “La Caridad-Mineral Park belt"
(Evcrnden and Kistler, 1970).
Another example of the control of stepovers on the local­
ization of magmatic activity (Fig. 14) may be the Tombstone w.w
volcanic center. We suggest that the pattern of faults and veins
in the Tombstone mining disu ict (Butler e t al., 1938) imply
that it is localized at the intersection of northwest- and north-
northeast-trending fault zones expressed in the geophysical
data jus lineaments adjacent to a circular magnetic feature
defining the Tombstone caldera (Figs. 9 and 10; cf. Moore, (a) Magnetic field
1993). Fold, fault, and polymetallic vein and replacem ent
mineralization patterns within the Tombstone district indi­
cate north-south compression at the time o f Late Cretaceous
mineralization, which implies that the geophysical features
reflect fractures with favorable orientations for reactivation.
Example 3: The Bagdad region underscores the impor­
tance of antecedent structures to the localization of magmas
and related ore deposits, and how a close inspection o f geo­
physical data can help in predicting such situations. Promi­
nent regional northeast-trending magnetic boundaries cross
this i cgion (Figs. 2 and 7). One such boundary (Figs. 9 and
16) separates a region of magnetic highs to the southeast
from magnetic lows to the northwest (e.g., the low associ­
ated with the Proterozoic Lawler Peak granite; cf. Anderson
ct al., 1956). Tile Bagdad deposit lies in the region of inter­
secting northwest- and northeast-trending magnetic bound­
aries or zones, coinciding in part with features in the poorly
resolved gravity data. As at Tombstone, the orientation of
Laramide regional stresses (N-S to NW-SE) favored princi­
pal displacem ent along northwest-striking, right-lateral
faults and the reactivation of northeast-striking faults such as 0 _ __ _J _ ___ 10 KM
the Proterozoic faulting observed in the geophysical data (b) Terraced gravity
(cf. Davis, 1981). Dike swarms (Anderson etal., 1956) delin­
eate the surface expression of the fault zones (Fig. 16). Fit;. 16- a- Reduccd-to-pole magnetic-anomaly map in the region o f the
Applying the structural model in Figure 14, we propose that Bagdad mineral district. The site o f Utc old Bagdad shaft is indicated by the
the mechanical interaction o f the two fault trends resulted letters Bl). Survey specifications are i.6 km fiighdinc spacing? and 150 tn
in the development of a releasing or extensional bend along above terrain. Small while circles highlight m agnetization boundaries
kised on horizontal gradient analysis. Black solid lines are interpreted mag­
die nordiwesi-striking fault zone. For fractures that intersect
netization boundaries (the two heavier lines are inferred major tectonic
at high angles, tensile forces are greatest at the intersections. zones highlighted in Fig. 9 ). b. Terraced isostatie basem ent gravity field
Thus permeability is also higher, and such intersections help over the Bagdad deposit (white ring). Although the gravity coverage is poor
to localize the em placem ent o f magmas such as the (stations located with white circles), several interpreted northeast and
Laramide stock that hosts the 72.6 ± 2.3 Ma Bagdad por­ northwest gravity lineaments (black lines) correlate with magnetic bound­
phyry copper deposit and related polymetallic veins. Analo­ aries. O f particular interest are the intersections o f magnetic and gravity
boundaries. Black dashed lines and the area highlighted with horizontal
gous intersections of geophysical features may be seen at
lines locate mapped faults and the Proterozoic Lawler Peak granite, respec­
several other localities in Figure 16. tively (taken from Wilson etal., 1977). White lines depict dike swarms near
Epithermal deposit trends: In Figures 17 to 20, the interme­ the Bagdad deposit (Anderson et al., 1956).
diate wavelength magnetic potential and gravity data and
 120° 118° 116® 114°

Fic.. 17. Intermediate-wavelength isostatic residual gravity anomaly map in (lie region ofim portant ore deposit trends
in Nevada. The blocky appearance o f die anomalies is due to die application o f a terrace filter, which enhances lithologic
or structural boundaries across which density varies (as explained in the text). Heavy black lines show interpreted major
density boundaries based on maxima o f horizontal gradient magnitudes shown in Figure 18. These boundaries define
zones of dense crustal blocks labeled A to C. Other gravity lineaments arc shown as light-weight black lines. The Protero­
zoic edge o f the craton (B?S r /w'Sr ■ 0.706) and the boundaries o f the Walker Lane tectonic zone are shown as dashed
white and black lines, respectively. Dotted line shows the Pine Nut (PN) Fault (Oldow et al., 1989). BT * Big Ten Peak
caldera; MJ = Ml. Jefferson caldera; RM = Round Mountain; SC> = Shoshone granite. Symbols as in Figure 1.

their horizontal gradient magnitudes are shown together
with the ore deposits in parts of Nevada and surrounding
regions. Using maxima of horizontal gradient magnitudes
(Figs. 17 and 20), Hildcnbrand etal. (2000) defined crustal
blocks characterized by major changes in density (Fig. 17)
and magnetization (Fig. 19). O ther secondary lineaments
are shown in these Figures. The attempt of Hildenbrand et
al; (2000) to define primary and secondary linear geophys­
ical features is subjective and not well constrained. For
example, although dense regional bodies labeled A, B, and
G in Figure 17 are shown as separate identities, these bod­
ies m ight be interconnected a t depth along secondary
boundaries.
Ludington et al. (1993) described the distribution of
epithermal deposits as displaying a crescent-shaped pattern
covering western and northern Nevada (black squares, Fig.
17). There appears to be a rough correlation of these deposits
with regional gravity highs (features A and B, Fig. 17). Iri the
southwestern arm of die coincident epithermal-deposit cres­
cent and gravity high (feature A), a prominent magnetic-high
zone (feature D, Fig. 19) suggests that some of the sources of
die gravity anomalies are dense and magnetic igneous intru­
sions, probably of Mesozoic age (Stewart, 1980). Of particular
interest are die northwest-striking magnetic and gravity fea­
tures (Figs. 8,18, and 20) aligned within the Walker Lane tec­
tonic zone (Figs. 17 and 19), a broad northwest-trending zone
 Fig, 18. HoriJtont.il gradient magnitudes of the intermediate-wavelength isostauc residual gravity anomaly data shown
in Figure 17. Maximum magnitudes (light gray shades) define boundaries separating regions o f contrasting densities.
Symbols as in Figure i.
of diverse topography and strike-slip faults (Stewart, 1988).
The northwest strike of die Pine Nut fault (exposed along die
eastern flank of die Wassuk Range near Walker Lake; Fig. 17),
Cretaceous dikes, and shapes of some plutons in the Walker
Lane imply that deep-seated nordiwest-striking and related
faults controlled die emplacement of pillions in the Mesozoic
back-arc. Blakely (1988) noted that the width of the north­
west-trending zone of magnetic highs in western Nevada
extends farther northeast dian die generally accepted struc­
tural boundary of the Walker Lane. Hilclenbrand et al. (2000)
pointed out that a related gravity high also extends nordieast
of die Walker Lane.
The geology (e.g., Shawe, 1995, 1999) along the north­
eastern edge of these northwest-trending regional geophys­
ical features indicates that Late Cretaceous intrusions (e.g.,
the Shoshone Granite in the Toquima Range) and mineral­
ization were localized in a region of northwest-trending
strike-slip faults (Fig. 21). Similarly striking lateral fault
zones, reflected in the magnetic data, bound the late
Oligocene to early Miocene Big Ten Peak and MountJeffer­
son calderas. Hilclenbrand et a!. (2000) proposed that the
calderas may be accommodating extensional strain between
these en echelon, strike-slip fault zones analogous to the
emplacement of granites into pull-apart structures (Fig. 14).
A nother im portant relationship is the one between vein-
bearing fractures and the regional principal stresses.
Caldera-associated mineralization as at Round Mountain is
often localized along fractures related to the principal
regional faults, in this case parallel to northwest-striking,
strike-slip faults (Shawe, 1995,1999), indicating the impor­
tance of die regional suess field to ore deposition. Hilclen-
brand et al. (2000) also pointed out that the correspon­
dence of the northeastern extent o f Tertiary epitherm al
mineralization and the boundaries o f geophysically inferred
 120° 118° 116° 114®

Fic. 19. Intermediate-wavelength magnetic potential map in the region o f important ore deposit trends in Nevada.
Heavy black lines show interpreted major magnetization boundaries based on maxima o f horizontal-gradient magnitudes
shown in Figure 20. These boundaries define zones o f magnetic crustal blocks labeled D anti E. Other magnetic linea­
ments are shown as light-weight black lines. The Proterozoic edge o f the craton (fi7S r /wi5r » 0.706) and the boundaries o f
the Walker Lane belt are shown as dashed white and black lines, respectively. Dotted line shows the Pine Nut (PN) fault
(Oldlow e t al., 1989). B T = Big Ten Peak caldera; MJ * Mt.Jefferson caldera; RM = Round Mountain; SC = Shoshone gran­
ite. Small squares define the map area shown in Figures 21 and 22. Symbols as in Figure 1.

Mesozoic intrusions (feature A in Fig. 17) supports a
hypothesis that there is a relationship between Tertiary
epithermal deposits and reactivated antecedent structures,
and further emphasizes the importance of deep-seated frac-
.lure systems to the localization of mineral deposits.
The northern part o f the epithermal deposit crescent is
characterized by low magnetic potential intensities occur­
ring in a northeast-trending, 500-km-long zone (Fig. 19),
and an abrupt change from northwest to a more northerly
direction of mountain ranges (see fig. 3 in Stewart, 1980).
Hildenbrand et al. (2000) proposed that the northeast­
trending magnetic low is a region characterized primarily by
upper-crustal, unaltered igneous rocks with magnetizations
of low amplitude and normal polarity. They suggested that
this northeast-trending zone o f magnetic lows reflects a
m ajor lithospheric irregularity, possibly the rem nant o f a
segment in the subducting plate that was propagated into
the overriding plate. Recent interpretation o f seismic data
(Van der Lee and Nolet, 1997) indicates that the subducted
Farallon plate is segmented and that one o f the boundaries
is roughly coincident with this trend o f magnetic lows.
There are numerous northeast-trending faults, dikes, and
mineralized veins within this zone that underscore its struc­
tural importance (Mackenzie and Bookstrom, 1976; Henry
and Bodcn, 1998).
Magnetic feature E strikes obliquely to the Carlin and Bat­
tle Moumain-Eureka mineral trends (Fig. 19), although its
western margin closely coincides with the western boundary
 120° 118° 116° H4°
F ig . 20. Horizontal-gradient magnitudes o f the intermediate-wavelength magnetic potential shown in Figure 19. Max­
imum magnitudes (light gray shades) define boundaries separating regions o f contiasting magnetization. Symbols as in
Figure 1.

of gravity feature C (Fig. 17). This north-northwest-trending belts, numerous investigations have focused on understand­
positive magnedc anomaly is coincident with highly mag­ ing the geological controls that led to these linear trends.
netic mafic rocks in the northern Nevada rift (Blakely and The Battle Mountain-Eureka trend, for example, lies near
Jachens, 1991; Zoback et al., 1994). Middle Miocene dikes the boundary separating less dense crustal rocks to the
and flows within the northern Nevada rift indicate that the southwest from dense rocks to the northeast (Figs. 17 and
rift post-dates the Eocene mineral deposits within the Battle 18; Blakely et al., 1991; Grauch et al., 1995,1998). Although
Mountain-Eureka and Carlin trends (discussed below). the trend o f the southern part of this mineral belt generally
However, the young rift zone parallels the western boundary follows inferred geophysical boundaries (Figs. 17 and 19),
of the dense crustal feature C and, thus, appears to have its n orthern p a rt appears to be oblique (~10°) to the
form ed along reactivated older structures. Epithermal regional geophysical trend.
deposits (15.5-15.0 Ma) follow the northern Nevada rift sys­ In contrast, the well-defined Carlin trend o f sediment-
tem extending roughly 300 km from the Oregon border to hosted gold deposits, paralleling and approximately 75 km
central Nevada (Fig. 19). John and Wallace (2000) suggest cast o f the Battle Mountain-Eureka trend, does not possess a
that faulting, mafic magmatism, and high heat flow related clear gravity or magnetic expression. O n the intermediate
to rifting led to precious metal mineralization by enhancing wavelength gravity anomaly map (Fig. 17), the combination
permeability and fluid flow along the lift o f removing the gravity effects of basins and de-emphasizing
Battle Mountain-Eureka and Carlin deposit trends: Because the effects of upper crustal and broad regional mass sources
many im portant gold deposits in Nevada lie along linear highlights a density boundary that roughly parallels and lies
 Fig. 21. Magnetic-anomaly field over the Round Mountain and Manhattan mineral districts, Nevada (area located in
Fig. 19). Four northwest-striking faults (black or white dotted lines) mapped by Shawe (1999) and Keith (1987) Hank the
Mt. Jefferson, Manhattan, and Big Ten Peak calderas (long dashed lines), suggesting a structural relationship. Black lines
arc inferred magnetization boundaries (based on interpretations o f derivative maps), indicating a northwest- and north­
east-trending structural grain. Epithermal deposits shown as squares. The Round Mountain and Manhattan mineral dis­
tricts spatially correlate with north west-trending magnetic lineaments. In the Round Mountain mineral district, local
northwest-striking faults have l>ccn mapped by Shawe (1999). The Mesozoic Baxter Springs intrusion supports the con­
clusion that the northwest-trending zones o f gravity atul magnetic highs (Figs. 17 and 20) reflect the extension o f the
Mesozoic granitoid rocks northeast o f the Walker Lane tectonic zone.

east of the Carlin trend. This gravity boundary continues far­
ther southeast to approximately latitude N39e,40', where it
abruptly changes direction to the southwest, intersecting die
Battle Mountain-Eureka trend near latitude N39°05'. A few
Carlin-style gold deposits and o th er deposit types occur
along this proposed gravity lineament representing the east­
ern boundary of feature C in Figure 17.
- The correspondence between the elliptical pattern of
Carlin-style gold deposits and the interpreted boundaries
defining a dense crustal zone (feature C, Fig. 17) is too close
to be dismissed as wholly fortuitous. In some locadons, off­
set of the boundaries defining the dense crustal block (fea­
ture C) from the alignm ent of deposits along the Battle
Mountain-Eureka and Carlin trends may be due to poor
data control in removing the gravity effects o f basins from
the gravity field (errors of several kilometers are possible;
see Jachens and Moring, 1990). The offsets associated with
these two ore deposit trends are troublesome but explain­
able. Three explanations are feasible if one accepts that the
intermediate-wavelength geophysical anomalies enhance
the effects o f deep crustal mass sources, whereas deposits
may be localized in shallow geological features which can be
laterally offset from their deeper roots. First, the crustal
block boundaries may have a variable dip and thus only
appear to be offset from the mineral trends. Second, north-
west-suiking strike-slip faults (e.g., Tosdal and Nutt, 1999) in
the overlying brittle crust formed due to extension during
delamination of the overthickened, early Cenozoic crust by
traction forces within the transition zone between the strong
brittle crust and the underlying weaker ductile zone (Berger
et al., 1998; Bourne et al., 1998; B. Berger, U.S. Geological
Survey, unpub. data, 2000; Drew and Berger, in press). How­
ever, over long time scales the slip rate in the brittle zone is
equivalent to the deformation rate (cf. Bourne et al., 1998),
so the differential strain may result in a structural offset
between shallow crustal regions and their deeper roots.
Third, rigid local bodies in the upper crust may act as obsta­
cles to upward fracture propagation and divert the strain
features around them , analogous to a knot in a piece o f
wood under stress.
The source producing the positive gravity anomaly
defined as feature C in Figure 17 may be related to both
dense carbonate rock and igneous intrusions (Grauch,
1998). Grauch et al. (1998) suggested that the density and
resistivity boundaries associated with the Carlin and Battle
M ountain-Eureka trends represent deep-penetrating
crustal faults, possibly related to the rifted Precam brian
continental margin, and that these faults later focused the
circulation o f hydrothermal fluids derived during increased
Tertiary heat uansfer. A recent study of feldspar Pb isotope
compositions in Jurassic, Cretaceous, and Ternary granitic
rocks supports this interpretation. Along the Carlin trend,
Wooden et al. (1998) define a northwest-trending Pb iso­
tope province boundary separating high ^ P b / ^ P b on
the northeast from lower values on the southwest. Another
Pb-province boundary is located near the 87S r/86Sr = 0.706
line (Figs. 17 and 19) and may reflect the Proterozoic era-
ton margin in Nevada. Wooden etal. (1998) proposed that
the Pb-isotope boundaries define Proterozoic crustal-scale
faults. Subsequent tectonism may have reactivated these old
northwest-trending structures. The episodic reaedvation of
deep-rooted faults as expressed in the gravity data (Figs. 17
and 18) can explain the alignment of ore deposits of vary­
ing ages and origins. Thus, die Carlin and Battle Mountain-
Eureka trends probably formed along antecedent Protero­
zoic faults.
Battle Mountain mining district: In the northern area of
the Battle M ountain-Eureka trend, Cretaceous and Ter­
tiary porphyry deposits of the Battle Mountain Mining Dis­
trict (BMMD) are generally associated with plutons
reflected as magnetic highs (Fig. 22). On the basis of gra­
dient analyses o f the m agnetic and gravity data, three
northwest- to north-northwest-trending zones or bound­
aries are defined. O ne geophysical zone represents the
western boundary o f the northern Nevada rift. In the Bat­
tle M ountain m ining district, Doebrich and T heodore
(1996) also defined two similar northwest-trending struc­
tural zones on the basis of topographic and magnetic lin­
ear features and on the elongations and alignm ents of
m ineral systems and dikes. The western two northwest-
striking geophysically defined zones bound Battle Moun­
tain and Buffalo M ountain. Because these m ountain
ranges have northeastern and southwestern topographic
expressions following the geophysical zones, the sources of
the geophysical zones appear to have influenced mountain
building. Q uaternary and late Tertiary faults identified
from a reconnaissance photogeologic m ap o f Nevada
(Dohrenwend et al., 1996) generally terminate or abruptly
change strike at all three o f these northwest-trending struc­
tural zones. O f particular interest is that most deposits lie
in the area between the two western northw est-trending
structural zones. The age o f these two zones may be Juras­
sic or older because dikes and faults along the southern
boundary o f a Jurassic pluton at Buffalo M ountain (Neff,
1966; Fig. 22) follow one o f these zones. Doebrich and
T heodore (1996) previously suggested that these north­
west-trending structural zones may be related to the shat­
tered hinges of broad, regional-scale Jurassic folds.
We propose that there was northwest-trending strike-slip
faulting in the Battle M ountain region during the late
Eocene, possibly related to forces imposed o n the brittle
crust by rapid extension in the lower crust. The two inter­
preted northw est-trending structures bounding Battle
M ountain were probably reactivated with right-lateral
m otion leading to roughly east-west extension (Fig. 22).
Apparently, the northw est faults interacted elastically
because, based on fault patterns and dike trends, the area
between them displays interlinked diladonal structures and
shears typical of pull-apart structures. Dikes tend to follow
the northwest-striking structural zones b u t strike largely
north-south in the region between them . Faults trend
north-south from the southern northwest-striking strike-slip
fault, and then bend to the northwest to merge with o r run
parallel to the northern northwest-striking strike-slip fault.
That the linking faults make up an interactive system o f ten­
sile and shear faults is known also from the m apping of
veins and faults in underground m ines (cf. Roberts and
Arnold, 1965). We propose that the Battle Mountain min­
eral deposits form ed in an extensional regim e between
northwest-trending faults with right-lateral motion. Thus,
the fracture permeability for the ore-forming fluids could
have developed in an extensional stepover (Fig. 14) with a
structural pattern similar to the dike and fault patterns at
Battle Mountain.
Mineral deposits (scale -1:50,000)
Compared to the regional geophysical data discussed thus
far, high-resolution magnetic and gravity data provide,
unquestionably, the greatest information on primary struc­
tures needed in evaluating mineral potential. Consequently,
national programs to systematically collect detailed or high-
resolution potential-field data over large regions have
evolved in several countries. At deposit-scales, Grant (1985a,
b), Hoover et al. (1992), and Gunn and Dentith (1997)
have provided discussions o f characteristic geophysical
expressions of several types of deposits.
Numerous excellent examples exist in the literature that
highlight the utility of geophysical studies based on high-
resolution potential-field data (see several examples in
Gubins, 1997). A notable exam ple relates to the gold
deposits of the Tennant Creek gold field, Northern Terri­
tory, Australia, which are scattered throughout an area of
about 2,000 mi2. Ivanac (1954) showed the importance of
m agnetic data in locating these hidden gold-copper-bis*
 117°20' 117° 116*40'

FlG. 22. (Top) Magnetic anomaly field over the region of the Battle Mountain mining district (area located in Fig. 19).
Battle Mountain mining district is highlighted with heavy dash line. The three northwest-trending stippled areas are
inferred structural zones based on gradient analyses of the magnetic and gravity data. Black square defines area shown in
Figure 24; symbols ;is in Figure f. Black lines are Quaternary and late Tertiary faults taken from a reconnaissance photo-
geologic map of Nevada (Dohrenwend ct ah, 1990). (Bottom) Red and black lines show dikes and faults, respectively, tit
Buffalo Mountain (Neff, 1900) and Battle Mountain (Roberts and Arnold, 1965). Cyan lines are interpreted magnetiza­
tion boundaries based on gradient analyses.
m uih deposits (which grade from 3-1,000 g A u /t) in a
magnetite* and hematite-rich ironstone. Many other Aus­
tralian studies a t district and ore-deposit scales dem on­
strate the utility o f geophysical interpretations to describe
deposits in a structural-stratigraphic context (e.g., Whiting,
1986; Isles et al., 1989; Edwards e t al., 1990; G unn and
Dentith, 1997; Gunn etal., 1997a, b; Willocks et al., 1999).
We discuss two examples: one from Australia and one from
the western United States.
Amnia block. Northern Territoiy: Whiting's (1986) classical
mineral study o f the Arunta block clearly demonstrates the
utility o f quality magnetic data (250 m line spacing and 80
m above ground) in compiling a geological map for a cov­
ered area with a precision that compares to that of maps of
exposed areas produced by conventional m apping tech­
niques. The ability to map accurately a concealed geologi­
cal unit o r param eter uniformly over a large area makes
aeromagneiic mapping a viable interpretive tool to charac­
terize structures and lithology over a vide range o f scales.
W hiting’s study benefited from a voluminous am ount o f
detailed physical property data and nearby limited expo­
sures o f targeted geological units. The measured magnetic
property data provided invaluable insights on the relation­
ship o f anomaly patterns and specific rock types, which
were used to generate the detailed structural and lithologi­
cal map shown in Figure 23. Economic concentrations o f
Ag-Pb-Zn and Cu-Au occur along prom inent elongate mag­
netic anomalies expressing hidden parasitic folds control­
ling mineralization.
Elder Creek porphyry copper deposit: In the Battle Mountain
m ining district (Fig. 22), the 37-Ma porphyritic monzo-
granitc intrusion related to the Elder Creek porphyry cop­
per system is clearly expressed in the magnetic data (Fig.
24). The central magnetic lows coincide with the heavily
potassium-silicatc-altered quartz stockworks. The abun­
dance of pyrrhotite in the surrounding homfels results in
prominent magnetic highs. Perhaps nearby magnetic highs
express concealed skarns or dikes of economic significance.
Conclusions and Summary
Several recurring structural relationships have been
highlighted throughout this review of the utility of geo­
physics in defining structures important in mineral studies.
The main relationships Include: (1) deep-seated faults, at
both regional and local scales reflected as geophysical lin­
eaments, can localize igneous activity, major ore deposits
and districts, and mineral belts; (2) these fault systems may
be reactivated episodically throughout geological time; and
(3) intersecting crustal fault systems are especially favorable
for ore deposition. Thus, magnetic and gravity data help
elucidate the regional geological framework or structural
setting and can contribute in locating favorable environ­
ments for hydrothermal mineralization.
The amount of geological information derived from the
potential-field data clearly increases from regional to local
scales, and thus the benefits of using potential-field data in
mineral investigation studies significantly increases. At local
scales, structural elements defined by high-rcsolulion geo­
physical data appear to work together to localize mineral
deposits, for example, along folds (e.g., Arunta Inlicr, Whit­
ing, 1986). Each mineralized zone or district appears to be
associated with a different geological history ancl with a dif­
ferent structural fabric associated with m ineralization.
Thus, the greatest utility o f geophysical data at either local
or regional scales is in the interpretation of structural rela­
tionships tliat help narrow an exploration area, particularly
in covered terrains, so that more intensive interdisciplinary
studies can be carried out in the effective evaluation of min­
eral potential.
In the present study, the broad application of potential-
field geophysics has led to new insights and reconfirm ed
previous interpretations o f the origin o f ore deposits in the
western conterminous United States. O ne recurring struc­
tural m odel evolving from this geophysical study is that
mineralization in the western United States is commonly
localized along strike-slip fault zones where pull-apart
basins o r releasing bends provided the increased fracture
permeability for the migration o f ore-forming fluids. The
Butte, Tombstone, Bagdad, and Battle M ountain districts
may be related to igneous activity and mineralization within
releasing bends in strike-slip faults (Fig. 14). O ther struc­
tural relationships have been highlighted:
1. Many deposits lie within regions characterized by low
total magnetization (H ildenbrand et al., 2000). This rela­
tionship warrants further study.
2. T he M other Lode gold belt closely follows a well-
defined norih-northwest-trending gravity high reflecting
dense mafic metavolcanic and piutonic rocks (H ilden­
brand et al., 2000).
3. Gravity and m agnetic maps em phasizing regional
crustal structures In Arizona, Montana, and Utah support
previous observations that linear geophysical zones define
deep structures apparently related to alignments of ore
deposits (e.g., Idftho-Montana porphyry belt) and ore-
deposit clusters (e.g., near Bingham, Utah).
4. In southwestern Nevada, the pattern of epitherm al
deposits (generally younger than 20 Ma) and inferred plu-
tons extends northeastward beyond the border of the north­
west-trending Walker Lane tectonic zone and mapped Meso­
zoic pillions, The correspondence between the northeast
edge of the pattern of epithermal deposits and geophysically
inferred Mesozoic plutons supports the hypothesis that a
relationship exists between Tertiary epithermal activity and
antecedent, reactivated structures. Further, it emphasizes the
importance of deep-seated fracture systems to the localiza­
tion of mineral deposits (Hildenbrand etal., 2000).
5. In north-central Nevada, an interpreted dense crustal
block has a western boundary' that correlates witli the north­
west-trending Battle Mountain-Eurcka trend (supporting
previous interpretations by Grauch etal., 1998), and an east­
ern boundary that roughly parallels the Carlin gold trend.
Both the Carlin and Battle Mountain-Eureka belts may have
formed along Proterozoic faults episodically reactivated dur­
ing active subduction of the Farallon plate beneath North
America. These episodic reactivations of old, deep-rooted
 105

k - 1 AdetaUeon [ | 3 Gnoissic Suito Basement depth contours

JERVOIS SCHISTS E 3 Magnetic metasedimonts Fault, Interpreted horizontal movement
Magnetite quartzite (NTRUSIVES _! Fault Interpreted voiticai movement
m C I 3 Granite.grenediorite oa» INTERPRETATION MAP
( y ^ Quartz saririte achltt Gabbro, amphbolite ,v— r Adelaidoan unconformity JERVOIS, N T
UndiffereritEated Uftramallc (ft) Mate or prospect
O motasodimonts Axial plane of fold
' ' i.j .j yAntndrm Synform

Fid. 23. Geological map o f thejervois area, Am nia Inlier. Northern Territory, Australia, based on the interpretation o f
acromagnctic data with the assistance o f geological and rock property data (taken from Whiting, 1986; published with the
permission o f the Geological Society o f Australia). The magnetic signatures o f scattered outcrops o f magnetite-quartz
rocks (associated with the lode horizon) are continued into coveted regions as intense linear magnetic anomalies. The
rcsuldng geological map reveals parasitic folds stmcturally controlling mineralization, highlighting the importance o f geo­
logical mapping based on the synthesis o f geophysical, geological, and rock-property data.

faults, as expressed in the gravity data, help explain the States should n o t be viewed as defining specific targets.
alignment of ore deposits of varying age and origins. Rather they provide a regional geological framework within
6. The Battle Mountain mineral district may be related which to help predict generally favorable environments for
to an extensional stepover between northwest-striking, mineralization.
right-lateral faults. Siting new exploration targets requires a careful analysis
o f geological, geochemical, and geophysical data at local
■In other regions, where the geology and origin o f ore scales in favorable areas defined by regional studies. A pre­
deposits may be less well known, mineral studies leading to dicted growth in the consumption of many commodities
similar geophysical relationships would be a useful early leads to the conclusion that greater success rates by
phase in exploration planning. The above geophysical rela­ prospectors in the exploration of covered terranes are
tionships with ore deposit patterns in the western United needed (Skinner, 1993). Skinner states (1993, p. 5), ‘T h e
 117°08' 117°06' 117°04'

Fig. 2*1. Residual magnetic field over the Elder Creek poiphyty copper system, Battle Mountain mining district (area
located in Figure 22). "Hie donation related to the pluton is clearly expressed in the magnetic data. The boundaries o f
exposed porphyritic monzograniic (white lines) and the outer limits o f the quartz stockworks (long dashed line) and
biotitc hornfcls (solid black lines) are taken from Gostyayeva ct al. (1996). A short-dashed line shows the approximate
boundary o f Quaternary sediment covering the eastern part o f the map area (Raines et al., 1996). These residual mag­
netic data were generated by subtracting an upward continued (50 in) regional field from the data in Figure 22.

greatest challenge lies in the finding o f buried ore bodies
and the solution to the challenge is vested in our scientific
understanding o f the way the crust is put together and our
ability to overcome the technological difficulties of map­
ping and sampling the crust in the third dimension."
High-resolution geophysical maps reduce the uncer­
tainty in identifying favorable environs for m ineral
deposits. In m ineral investigations at all m ap scales, the
interpretation of potenual-field data clearly will continue to
play a major role by defining geology in three dimensions,
if magnetic and gravity data of adequate resolution are
available to meet the challenge.
Acknowledgments
This study has greatly benefited from reviews by T.
Theodore, M. DufFett, C. Finn, andJ.P. Richards. D. Moore,
P. Gunn, and M. Pilkington kindly provided very useful ref­
erences discussing the utility o f geophysical data in mineral
studies outside the United States. Communications with R.
Ashley, D. Cox, T. Grauch, and D. John conu ibuted impor­
tant information on various aspects of the tectonic history,
mineral deposits, or structural development of the western
United States. Battle Mountain Gold Company kindly fur­
nished the high-resolution magnetic data over the Elder
Creek deposit (Fig. 24).
REFERENCES
Anderson, C A , Scholz, E.A., and Strobell, J.D., Jr., 1956, Geology nnd ore
deposits o f the Bagdad area, Yavapai County, Arizona: U.S. Geological
Survey Professional Paper 278,103 p.
Ashley, R.P.. 1991, Gold and silver deposits o f the United States: Geological
Society o f America, The Decade o f North American Geology (DNAG),
v. P-2, p. 3-22.
Baranov, V., 1957, A new m ethod for interpretation o f acromagnctic
maps— pscudogravity anomalies: Geophysics, v. 22, p. 359-383.
 Berger, B.R., and Drew, L.J., 1997, Role o f strike-slip duplexes in localiza­
tion o f volcanoes, related intrusions, and epizonal ore deposits [abs,):
Geological Society o f America Abstracts with Programs, v. 29, p. 359-360.
Berger, B.R., Goldhaber, M., Hildenbrand, T.G., and Wanly, R., 1998, Ori­
gin ofCarlin-style deposits, Nevada: Coupled regional fluid flow, core­
complex related extension, sirike-slip faults, and magmatism [abs.]: Geo­
logical Society o f America Abstracts with Programs, v. 30, p. 568.
Billingsley, P.R., and Locke, A., 1911, Structure o f ore deposits in the con­
tinental framework: Transactions o f the American Institute o f Mining
Engineers, v. 13-1, p. 9-64.
Blakely, R.J., 1988, Curie-temperature isotherm analysis and tectonic impli­
cations o f aerom agnetic data from Nevada: Journal o f Geophysical
Research, v. 93, p. 11,817-11,832.
—— 1995, Potential theory in gravity and m agnetic applications: Cam­
bridge University Press, 441 p.
Blakely, R.J., and Jachens,R.C., 1991, Regional study o f mineral resources in
Nevada: Insights from three-dimensional analysis o f gravity and magnede
anomalies: Geological Society of America Bulletin, v. 103, p. 795-803.
Blakely, R.J., and Simpson, R.W., 1986, Locating edges o f source bodies
from magnetic and gravity anomalies: Geophysics, v. 51, p, 1394-1396.
Blakely, RJ„ Jachens, R.C., and Sherlock, M.G., 1991, Bulk mineable pre­
cious metal deposits and regional crustal structure o f Nevada—A statisti­
cal comparison, in Dean, D A , Benedetto, K.M.F., and Dui'gin, D.C., cds,
Southern extension o f the Carlin Trend, influence o f structure and
stratigraphy on gold deposition. Post-meeting Field Trip no. 4, Great
Basin Symposium: Geological Society o f Nevada, Reno, Nevada, p. 30-38.
Bourne, S.J., England, P.C., and Parsons, B„ 1998, The motion o f crustal
blocks driven by flow o f the lower lithosphere and implications for slip
rates o f continental strike-slip faults: Nature, v. 391, p. 655-659.
Butler, B.S., Wilson, E.D„ and Rasor, C A , 1938, Geology and ore deposits
o f the Tombstone district, Arizona: Arizona Bureau o f Mines, Geological
Scries No. 10, Bulletin no. 133,113 p.
Card, K.D., and Poulscn, K.H., 1998, Geology and mineral deposits o f the
Superior Province o f the Canadian Shield, in Lucas, S., cd.. Geology o f
die Prccambrian Superior and Grenville Provinces and Prccambrian fos­
sils in North America: Geological Society o f America, The G eology o f
North America, v. C-l, p. 13-94.
Carmichael, R.S., 1982, Handbook o f Physical Properties o f Rocks: Boca
Raton, Florida, CRC Press, v. 2 ,3 4 5 p.
Clark, D A , 1997a, Magnetic properties o f rocks and minerals: Journal o f
Australian Geology and Geophysics, v. 17, p. 5-10.
------ 1997b, Magnetic petrophysics and magnetic petrology—aids to g e o
logical interpretation o f magnetic surveys: AGSO Journal o f Australian
Geology and Geophysics, v. 17, p. 83-103.
Clark, S.P., 1966, Handbook o f physical constants: Geological Society o f
America Memoir 9 7 ,5 8 7 p.
Coney. PJ.. 1976, Plate tectonics and the Laramidc orogeny: New Mexico
Geological Society Publication 6, p. 5-10.
------ 1978. Mesozoic-Ccnozoic Cordilleran plate tectonics: Geological Soci­
ety o f America Memoir 152, p. 33-50.
Cordell. L.. and Henderson, R.G., 1968, Iterative three-dimensional solu­
tion o f gravity anomaly data using a digital com puter Geophysics, v. 33,
;V.: p. 596-601.
: Cordell, L., and McCafTcrty, A E ., 1989, A terracing operator for physical
property mapping with potential-field data: Geophysics, v. 54, p. 621-634.
Cordell, L., Zorin, Y A , and Keller. C.R., 1991, The dccompcnsativc gravity
?.'/> anomaly and deep structure o f the region o f the Rio Grande rift:Journal
A / o f Geophysical Research, v. 96, p. 6557-6568.
Cox, D.P., 1992, Descriptive model o f distal disseminated Ag-Au: U.S. Geo-
V• logical Survey Bulletin 2004, p. 19.
Cox, D.P., and Singer, D.S., 1986, Mineral deposit models: U.S. Geological
Survey Bulletin 1693,379 p.
Davis, G.H., 1981, Regional strain analysis o f the superposed deformations
in southeastern Arizona and the eastern Great Basin: Arizona Geological
Society Digest, v. 13, p. 155-172.
Denham, D., 1997, Airborne geophysics in Australia— the governm ent
contribution: Journal o f Australian Geology and Geophysics, v. 17, no. 2.
p. 3-9.
Docbrich.J.L., and Theodore, T.G., 1996, Geologic history o f the Batde
Mountain mining district, Nevada, and regional controls on the distrib­
ution o f mineral systems, in Coyncr, A.R., and Fahey, P.L., cds., Geology
and ore deposits o f the American Cordillera: G eological Society o f
Nevada Symposium, Rcno/Sparks, Nevada, April 1995, Proceedings,
p. 453-483.
Dohremvend, J.C., Schell, BA., Mcngcs, C.M., Moring, B.C., and McKit-
trick, M.A., 1996, Reconnaissance phoiogcologic map o f young (Qua-
ternary and late Tertiary) faults in Nevada: Nevada Bureau o f Mines and
Geology, Open-File Report, 96-2, scale 1:1,000,000, p. 9.1-9.12.
Dooley, T., and McCiay, K., 1998, Analog modeling o f pull-apart basins: Amer­
ican Association o f Petroleum Geologists Bulletin, v. 81, p. 1804-1826.
Drew, L.J., and Berger, B.R., in press, Application o f the porphyry cop-
per/polymctallic vein kin-deposit system to mineral resource assessment
In the Maim Mountains, northern Hungary, in Andrea, F„ cd., Geoenvi-
ronmcntal models for resource exploitation and environmental security:
Dordrecht, Khmer Academic Publishers, 19 p.
Edwards, G.C., Booth, S.A., and Cozens, G.J., 1990, W hite Devil gold
deposit, in Hughes, F.E., cd., Geology o f the mineral deposits o f Ausualia
and Papua New Guinea: Australasian Institute o f Mining and Metallurgy,
v. 1, p. 849-855.
Erickson, A.J.,Jr., 1976, The Unita-Gold Hill trend—an economically impor­
tant lineament: Utah Geological Associadon Publication, no. 5, p. 126-137.
Evcrndcn.J.F., and Kistler, R.W., 1970, Geochronology o f emplacement o f
Mesozoic batholith com plexes in California and Western Nevada: U.S.
Geological Survey Professional Paper 6 23,42 p.
Gostyaycva, N., Theodore, T.G., and Lowcnstcm, J.B., 1996, Implications
o f fluid-inclusion relations in the Elder Creek porphyry copper system,
Battle Mountain mining district, Nevada: U.S. Geological Survey Open-
File Report 96-268,53 p.
Grant, F.S., 1985a, Acromagnetics, geology, and ore environments, I. Mag­
netite in igneous, sedimentary, and mctamorphic rocks— an overview:
Gcocxploration, v. 23, p. 303-333.
------ 1985b, Acromagnetics, geology, and ore environments, II. Magncdtc
and ore environments: Geocxplorauon, v. 23, p. 335-362.
Grauch, V.J.S., 1998, Crustal structure and its relation to gold belts in
north-central Nevada— overview and progress report: U.S. Geological
Survey Open-File Report-338, p. 34-37.
Grauch, VJ.S., and Coidcll, L , 1987, lim itations o f determining density
and magnetic boundaries from die horizontal gradient and pscudograv-
«y data: Geophysics, v. 52, p. 118-121.
Grauch, VJ.S.,Jachcns. R.C., and Blakely, R.J., 1995. Evidence for a basement
feature (elated to rite Cortez disseminated gold trend and implications for
regional exploration in Nevada: Economic Geology, v. 90, p. 203-207.
Grauch, VJ.S., Klein, D.P., and Rodriquez. B.. 1998, Progress on under­
standing the crustal structure near die Battlc-Mountain-Eurcka mineral
trend from geophysical constraints: U.S. Geological Survey Open-File
Report 98-338, p. 8-M .
Gubins, A G , cd., 1997, Geophysics and Geochemistry at the Millenium:
Prospectors and D evelopers Association o f Canada, Exploration 97,
Fourth D ecennial C onference o n Mineral Exploration, Toronto,
Ontario, Proceedings, 1068 p.
Gunn, PJ., and Dcntith, M., 1997, Magnetic responses o f mineral deposits:
AGSO Journal o f Australian Geology and Geophysics, v. 17, p. 135-158.
Gunn, PJ.. Maidmcnt, D., and Milligan. P.R., 1997a, Interpreting aero-
magnetic data in areas o f limited outcrop: AGSO Journal o f Australian
Geology and Geophysics, v. 17, p. 175-186.
Gunn, PJ., Milligan, P.. Mackey. T„ Liu, S., Murray, A . Maidmcnt, D., and
Harcn, R., 1997b. Geological mapping using the national airborne and
gravity datasets: An example focussing on Broken Hill: AGSO Journal o f
Australian Geology and Geophysics, v. 17, p. 127-136.
Hall, C.M., Simon. G.. and Kcslcr, S.E., 1997, Age o f mineralization at the
Twin Creeks SHMG deposit, Nevada: Society o f Econom ic Geologists
Guidebook 28. p. 151-154.
Henderson, J.R., Stroinquist, AA ., and Jerperscn, A.. 1966, Aeromagnetic
map o f parts o f the Mother Lode goltl and Sierra foothills copper min­
ing districts, California, and its geological interpretation: U.S. Geologi­
cal Survey Geophysical Investigations Map GP-661, scale 1:62,500.
Henley. R.W., and Adams, D.P.M., 1992. Strike-slip fault reactivation as a
control o n epitherm al vein-style gold mineralization: Geology, v. 20,
p. 443-446.
Henry, C.D., and Bodcn, D.R., 1998, Geology o f the Mount Blitzcn quad­
rangle, Elko County, Nevada: Nevada Bureau o f Mines and Geology Map
110, scale 1:24,000, 20 p.
Hilclenbrand, T.G., Berger, B.R.,Jachens, R.C., and Ludington, S.D., 2000.
Regional crustal structures and their relation to the distribution o f ore
deposits in western U nited States based on magnetic and gravity data:
Economic Geology, v. 95. in press.
Minze, W.J., ed„ 1985, The utility o f regional gravity and magnetic anom-
aly maps: Society of Exploration Geophysicists, 454 p.
Hofstra, A.H., Snee, LAV., Rye, R.O., Folger, H.W., Phinisey.J.D., Lorangcr,
R.J., Dahl, A.R., Naeser, C.W., Stein, H.J., and Lewchuk, M., 1999, Age
concentrations on Jerrit Canyon and other Carlin-type deposits in west-
ern United States— relationship to mid-Tertiary extension and magma-
tisnv. Economic Geology, v. 94, p. 769-802.
Hoover, D.B., Heran, W.D., and Hill, P.L., 1992, The geophysical expres-
sion o f selected mineral deposit models: U.S. Geological Survey Open-
File Report 92-557, 129 p,
Hutchinson, R.W., and Albers, J.P., 1992, M ctallogcnic evolution o f the
Cordillcran region o f the western United States: Geological Society o f
America, T he D ecade o f North American G eology (DNAG ), v. G-3,
p. 629-652.
Isles, D.J., Harman, P.G., and G unneen, J.P., 1989, The contributions o f
high-resolution aerom agnetics to Archcan gold exploration in the Maher, S., W illman, C.E., Taylor, D.I-I., Cayley, R A ., and M oore, D,H,,
region, Western AustniUa: Economic Geology Monograph 6, p. 389-397.
!vanac,J.F., 1954, The geology and mineral deposits o f the Tennant Creek
Gold-field, Northern Territory: Bureau o f Mineral Resources, Geology
and Geophysics, no. 22.16*1 p.
Iyer, H.M, and Hitchcock, T., 1989, Uppcr-mamle velocity structure in the
continental U.S. and Canada: G eological Society o f America Memoir
172, p. 681-710.
Jachens, R.C.. and Griscom, A , 1985, An isostatic residual gravity map o f
California— a residual map for interpretation o f anom alies from
imracrusial sources, in H inze, W.J., cd.. The utility o f regional gravity
and m agnetic anomaly maps: Tulsa, Society o f Exploration Geophysi-
cists, p. 347-360.
Jachens, R.C., and Moring, B.C., 1990, Maps o f the thickness o f Cenozoic
deposits and the isostatic residual gravity over basement for Nevada: U.S.
Geological Survey Open-File Report 90-404, scale 1:1,000,000,15 p.
Jachens, R.C., Simpson. RAV.. Blakely, R.J„ and Saints, R.W., 1989, Isostatic
residual gravity and cinsiai geology o f the United States: Geological Sod -
ety o f America Memoir 172, p, 405-424.
James, L.P., 1978, The Bingham copper deposits, Utah, as an exploration
target-history and pre-excavation geology: Econom ic Geology, v. 73,
p. 1218-1227.
Jaques, A.L., Wellman, P., Whitaker, A., and Wybom, D., 1997, Higli-reso-
lution geophysics in m odern geologic mapping: Journal o f Australian
Gcologynnd Geophysics, v. 17, p. 159-174.
Jennings. C.W.. Strand, R.G., and Rogers, T.H., 1977, Geologic map o f Gd- Nishiwaki, C., 1981, Tectonic control o f porphyiy copper genesis in the
ifornin: California Division o f M ines and Geolog>', G eologic Data Map
Series no. 2. scale 1:750,000.
John. D.A., and Wallace, A R „ 2000. Epithermal gold-silver dc|>osits
rclatcd to the northern Nevada rift, in Chier.JJL, Price. J.C ., Struh-
sackcr, E.M.. Hardyman, R.F., and Morris, C.L.. eds.. Geology and ore
deposits 2000-thc Great Basin and beyond: Geological Society o f Nevada
Symposium, Proceedings. 15-18 May 2000 (in press).
Keith, W.J., 1987, Preliminary geologic map o f the Big Ten Peak Quadran-
gle, Nye County, Nevada: U.S. Geological Survey Open-File Report 87-7, O ’Neill,J.M., and Lopez, D A , 1985. Character and regional significance
scale 1:62,500.
Kerrich. R.,’ 1999, Nature's gold factory: Science, v. 284, p. 2101-2102.
Kutina. J .a n d Hildcnbrand.T.G.. 1987, Ore deposits o f the western U.S. Raines. C.L., Sawatzky. D.L.. and Connors. K A , 1996. Great Basin geo-
in relation to mass distribution in the crust and mantle: Ceological Soci-
ety o f America Bulletin, v. 99, p. 30-11.
Landefeld, L.A., 1990, T h e geology o f th e M other Lode gold belt.
Foothills Mciamorphic belt. Sierra Nevada. California: American Associ-
ation o f Petroleum G eologists, Pacific Section, Guidebook GB68.
p. 117-124.
Lenman, D.E., 1998, Structure, contents, and setting o f Pb-Zn mincraliza- Refold, M.S., 1980, Magnetic method: Geophysics, v. 45, p. 1640-1658.
lion in the McArthur Basin, northern Territory Australia: Australian Roberts, R.J., and Arnold, D.C., 1965, Ore deposits o f the Antler Peak
Journal o f Earth Sciences, v. 45, p. 3-20.
Lcclair, A.D., Lucas, S.B., Broome, H.J.. Viljocn, D.W., and Weber. W. 1997,
Regional mapping o f Precambrian basement beneath Phancrozoic cover
in southeastern Trans-Hudson O rogen. Manitoba and Saskatchewan:
Canadian Journal o f Earth Sciences, v. 31. p. 618-634.
Lipman, P,W„ and Hagstmm, J.T., 1992, Jurassic ash-flow sheets, calderas,
and related intrusions o f the Cordillcran volcanic arc in southeastern
r
Arizona—implications for regional tectonics and ore deposits: Gcologi-
cal Society o f America Bulletin, v. 104, p. 32-39.
Liptnan, P.W., Prostka, H.J., and Christianson, R.L., 1972, Cenozoic vol-
canism and plate tectonic evolution o f western U.S.— I, Early and Mid-
d ie C enozoic: Royal Society o f London Philosophical Transactions,
Series A, v. 271, p .217-248.
Long, K.R., DeYoung. J.H .,Jr., and Ludington, S.D., 1998, Significant
deposits o f gold, silver, copper, lead, and zinc in the United Suites: U.S.
Geological Survey O pen-File Report 90-206A, 33 p. 98-206B. pm c 3.5
inch diskette.
Lowell, J.D., 1974, Regional characteristics o f porphyry copper deposits in
the soutlnvcsi: Economic Geology, v. 69, p. 601-617.
Ludington, S.D., Cox, D.P., Singer, D A , Sherlock, M.G.. Berger, B.R., and
Tingley, J.V., 1993, Spatial and temporal analysis o f precious-metal
deposits for a mineral resource assessment o f Nevada: Geological Asso-
ciation o f Canada, Special Paper 40, p. 31—10.
MacKcnzie, W.B., and Bookstrom, A A , 1976. Geology o f the Majuba Hill
area, Pershing County, Nevada: Nevada Bureau o f Mines and Geology
Bulletin 86, 23 p.
2000, Rethinking oragenic gold in the southern Lachlan fold belt labs.):
Australian G eological C onvention, 15 th, Sydney, Abstracts with Pro-
grams, v. 59.
McMullan, S.R., and McLellan, W.H., 1997, Measured is better, in Gubins,
A G „ cd., Geophysics and Geochemistry at the Millenium: Prospectors
and Developers Association o f Canada, Exploration 97, Fourth Deccn-
nial Conference on Mineral Exploration, Toronto, Ontario, GEO F /X ,
Proceedings, p. 87 ^ 8 7 6 .
Miller, D.M., Nilson, T.H., and Bilodeau, W.L., 1992, Late Cretaceous to
early Eocene geologic evolution o f the U.S. Cordillera: Geological Soci-
ety o f America, T he Decade o f North American Geology (DNAG), v. G-
3, p. 205-260.
M oore, D .H ., VanderBerg, A.H.M ., William, C.E., and Magart, A.P.M.,
1998, Paleozoic geology and resources o f Victoria: Journal o f Australian
Geology and Geophysics, v. 13, p. 107-122.
M oore, R.B., 1993, G eologic m ap o f the Tom bstone volcanic center,
Cochise County, Arizona: U.S. Ceological Survey Miscellaneous Invest!-
gallons Map, 1-2420, scale 1:50,000.
Morgan, P., and Gosnold, W.D„ 1 9 8 9 ,1-ieatflow and thermal regimes in
the continental United States: Geological Society o f America Memoir,
v. 172, p. 493-522.
Neff, T.R., 1966, Petrology and structure o f the Buffalo Mountain pluton,
Humbolt County, Nevada: Unpublished Ph.D. thesis, Stanford Univer-
sity, 120 p.
southwestern Pacific island arc region: Mining Geology, v. 31, p. 131-136
(inJapanese).
O ’Driscoll, 1990, Lineam ent tectonics o f Australian ore deposits: Aus-
tralasian Institute o f Mining and Metallurgy Monograph 14, p. 33-41.
Oldow. J.S., Bally, AAV , Ave Lallemant, H.G ., and Leem an, W.P.. 1989,
Phancrozoic evolution o f the North American Cordillera—United States
and Canada: G eological Society o f America, T he G eology o f North
America, v. A, p. 139-232.
o f the Great Falls tectonic zone, cast-central Idaho and west-central Mon-
tana: American Association o f Petroleum Geologists, v. 69, p. 437-177.
science database: U.S. Geological Survey, Digital Data Series DDS-41.
Reeves. C.V., Bedford. SAV., and M illigan. P.R., 1997. Airborne geo-
physics— old methods, new images, in Gubins, A.G., ed., Geophysics and
Geochemistry at the Millenium: Prospectors and Developers Association
o f Canada, Exploration 97, Fourth Decennial C onference on Mineral
Exploration, Toronto, Ontario, GEO F /X , Proceedings, p. 13-30.
Quadrangle, Humboldt and Landers Counties, Nevada: U.S. Geological
Survey Professional Paper 459-B, 94 p.
Robson, D.E., and Lewis, P.C., 1997, Regional geophysics and geochemistry
as an exploration initiative in New South Wales, Australia, in Gubins,
AG ., ed.. Geophysics and geochemistry at the Millenium: Exploration 97.
Fourth Decennial International Congress on Mineral Exploration, GEO
F /X , Proceedings, p. 793-804.
 UTILITY OF MAGNETIC AND GRAVITY DATA: EXAMPLES FROM WFJi'lERN U.S.
Rostad, O.H., 1978, K-Ar dates for mineralization in the White Cloud-Can-
nivart porphyry molybdenum belt o f Idaho-Montana: A Discussion: Eco­
nomic Geology, v. 73, p. 1366-1367.
Sales, R.H., 1913, Ore deposits at Butte, Montana: Transactions o f the
American Institute o f Mining Engineers, v. 46, p. 3-109.
Saltus, R.W., and Jachcns, R.C., 1995, Gravity and basin-depth maps o f the
Basin and Range Province, western United States: U.S. Geological Survey
Geophysical Investigations Map GP-1012, scale 1:2,500,00.
Schmidt. C.J., Smedcs, H.W., and O 'N eill, M., 1990, Syncomprcssional
emplacement o f the Boulder and Tobacco Root batholiths by pull-apart
along old faults: Wiley, Geological Journal, v. 25, p. 305-318.
Sccdorff, E., 1991, Magmatism, extension, and ore deposits o f Eocene to
H olocene age in the Great Basin—mutual effects and preliminary pro­
posed genetic relationships: Geological Society o f Nevada, Geology and ore
deposits o f the Great Basin, Reno, Symposium Proceedings, p. 133-178.
Shavvc, D.R., 1995, Geologic map o f the Round Mountain Quadrangle,
Nye County, Nevada: U.S. Geological Survey, scale 1:24,000.
------ 1999, G eologic map o f the Jefferson quadrangle, Nye County,
Nevada: U.S. Geological Survey Geologic Investigations Series 1-2670,
l:24,000-scale.
Simpson, R.W., and Jachcns, R.C., 1989, Gravity methods in regional stud­
ies: Geological Society o f America Memoir 172, p. 35-60.
Simpson, R.W., Jachens, R.C., Blakely, R.J., and Saltus, R.W., 1986, A new
isostatic residual gravity o f the conterminous United States, with a dis­
cussion o f the significance o f the isostatic residual anomalies: Journal o f
Geophysical Research, v. 91, p. 8348-8372.
Sillitoc, R.H., 1975, Subduction and porphyry copper deposits in south­
western North America (A reply to recent objections): Economic Geol­
ogy. v. 70, p. 1374-1377.
Skinner, B.J., 1993, Finding mineral resources and the consequences o f
using them—major challenges in the 21st century: Australasian Institute
o f Mining and Metallurgy Publication 93, v. 2, p. 1-8.
Smedcs, H.W., Montis, R.K., and Tilling, R.I., 1988, Preliminary map o f
piutonic units o f the Boulder batholith, southwestern Montana: U.S.
Geological Survey Open-File Report 88-283, scale 1:200,000.
Snee, L„ Miggins, D„ Geisman.J., Reed, M., Dilles, J., and Zhang, L., 1999,
Thermal history o f the Butte porphyry system, Montana [abs.j: Geologi­
cal Society o f America Abstracts with Programs, v, 31, p. A-S80.
Stewart, J.H., 1988, Tectonics o f the Walker Lane, western Great Basin:
Mesozoic and Cenozoic deformation in a shear zone, in Ernst, W.G., ed..
Metamorphism and crustal evolution o f the Western U.S. (Rubey Vol­
ume VII): Englewood Cliffs, N.J., Premicc-Mall, p. 683-713.
Stewart, R.B., 1980, Geology o f Nevada: Nevada Bureau o f Mines Special
Publication 4, 136 p.
Syberg, FJ.R., 1972, A Fourier method for the regional-residual problem
o f potential fields: Geophysical Prospecting, v. 20, p. 47-75.
Teskey, D.J., Hood, PJ., Morlcy, L.W.,Gibb, RA,Sawatzky, P„ Bower, M.,and
Ready, E.E,, 1993, The acromagnctic survey program o f the Geological
Survey o f Canada—contribution to regional geological mapping and min­
eral exploration: Canadian Journal o f Earth Sciences, v. 30, p. 243-2G0.
Theodore, T.G., 1998, Large distal-disseminated precious-metal deposits,
Battle Mountain mining district, Nevada: U.S. Geological Survey Open-
File Report 98-0338-B, p. 253-258.
Tidcy, S.R., 1995, Geological summary and perspective o f porphyty copper
deposits in southwestern North America: Arizona Geological Society
Digest 20, p. 6-20.
Tosdal, R.M., and Nutt, C.J., 1999, Late Eocene and Oligoccne tectonic set­
ting o f Carlin-type Au deposits, Carlin trend, Nevada, U .S A , in Stanley,
CJ. ct al, cds.. Mineral deposits— Processes to processing: Biennial SGA
1999 Meeting, 5th, London. A A Balkema, Proceedings, p. 905-908.
U.S. Geological Survey, 1969, Aeromagnedc map o f the northern Mother
Lode area, California: U.S. Geological Survey Gcophvsical Investigations
Map GP-671, scale 1:62,500.
------ 1976, Acromagnctic maps o f Nevada City and Alleghany 15 minute
quadrangles, California: U.S. Geological Survey Open-File Report 76-
274, scale 1:62,500.
VanderBerg, AH.M ., Willman, C.E., Maher, S., Simons, B A , Cayley, R A ,
Morand, V.J., Taylor, D.H., Moore, D., and Radojkovic, A , 2000, The Tas­
man fold bell system in Victoria: Geological Survey o f Victoria Special
Publication, scale 1:1,000,000.
Van der Lee, S., and Nolet, G., 1997, Seismic image o f the subducted trail­
ing fragments o f the Farallon plate: Nature, v. 386, p. 266-269.
Whidng, T.H., 1986, Acromagnctic data as an aid to geological mapping—
a case history from the Arunia Inlicr, Northern Territory: Australian
Journal o f Earth Sciences, v. 33, p. 271-286.
Willocks, A.J., Haydon, S.J., A te n , M.W., and Moore, D.H., eds., 1999,
Geophysical signatures o f base metal deposits o f Victoria: Geologic Sur­
vey o f Victoria Report 119, and Australian Society o f Exploration Geo­
physicists Special Publicadon 11,162 p.
Wilson, E. D., Moore, R. T., and Cooper, J. R., 1977, G eologic map o f Ari­
zona: U.S. G eological Survey and Arizona Bureau o f M ines, scale
1:500,000.
Woodall, R., 1993, The mulddisciplinary team approach to successful min­
eral exploration: Society o f Econom ic G eologists Newsletter, no. 13,
p. 6-11.
W ooden.J.L, Tosdal, k M ., and Kisder. R.W., 1998, Pb isotopic mapping o f
crustal structure in the northern Great Basin and relationships to Au
deposits trends: U.S. Geological Survey, Open-File Report98-338, p. 20-33.
Zoback, M.L., Anderson, R.E., and Thom pson, G A , 1981, Cenozoic evo­
lution o f the state o f stress and style o f tectonism in the western U S .:
Royal Society o f London Philosophical Transactions, series A v. 300,
p. 407-134.
Zoback, M.L., McKee, E.H., Blakely, R.J., and Thom pson, G A , 1994, The
northern Nevada rift—regional lecionic-magmadc rcladons and middle
Miocene stress dircedons: Geological Society o f America Bulletin, v. 106,
p. 371-382.
:
 Society o f Economic Geologists
Reviews v. H, 300/, />. 111—155

Chapter 5

Vein Formation and Deform ation in G reenstone G old D eposits

F rancois R o b e r t !
B arrick G old C orporation, 7 2 5 7 D unver, V erdun, Quebec, C a n a d a H 4 H 2 H 6

and K. H ow ard P o u l sen

5 4 W atford W as, N epean, O ntario, C a n a d a K 2 E 6 B 6

Abstract

Veins are common com ponents o f greenstone gold deposits. T heir analysis is one key aspect in un­
derstanding the sequence o f events leading to the formation o r deformation o f gold deposits. This analy­
sis is essential for the determ ination o f controls on mineralization and ore-forming processes, and for the
prediction o f the geometry and plunges o f deposits and orebodies. Many greenstone gold districts have
experienced a common structural evolution: D, thin skin-style shortening and D2 thick skin-style short­
ening are largely responsible for the structural trend and penetrative fabrics in a district, whereas Ds and
D4 transcurrent deformation are largely focused along preexisting major fault zones. A majority of green­
stone gold deposits consists o f quartz-carbonate veins in o r adjacent to high-angle reverse, and loss com­
monly transcurrent, shear zones, viewed as splays or subsidiaries of major, complex, belt-scale fault zones.
In Other deposits, veins simply overprint gold mineralization and provide im portant information about
the postore deformation history
T hree main types o f veins occur in greenstone gold deposits and each records small increments of bulk
strain. Laminated fault-fill veins form by slip along the central parts o f active shear zones in low-angle di-
lational bends, or less commonly by extensional opening o f foliation planes. Extensional and oblique-ex­
tension veins form within or adjacent to shear zones, at high angles to foliation and elongation lincation.
They represent opening and filling of extensional and hybrid extcnsional-shear fractures, respectively: In
more competent host rocks, extensional veins can form arrays of en echelon planar o r sigmoidal veins,
or of stacked planar veins, and can also combine into multiple sets to form stockwork and breccia bod­
ies. Multiple types and sets of auriferous veins commonly combine to form variably complex vein net­
works, especially in large deposits. These vein networks record deposit-scale bulk incremental strain, with
axes of elongation and shortening that can be compared with those of the main deformation increments
in the district as a further way of constraining their timing of formation. The formation of vein networks
in many districts is compatible with D2, and in a num ber of others with D3, reflecting their formation in
contractional or transcurrent deformation regimes, likely involving subhorizontal compressional stress
under high fluid pressures.
Veins in many districts also systematically display evidence of overprinting deformation, in the form of
folds, boudins, striated vein margins, and a number of internal vein textures such as recrystallized quartz
and stylolites. Overprinting deformation is a natural consequence of vein formation in active shear zones,
but it can also result from overprinting of veins by a younger increment of regional deformation. This can
lead to local shear zone reactivation or wholesale folding or boudinage of a deposit. The confident deter­
mination of the structural timing of veins in deposits is critical but challenging, and is at the center of di­
vergences of interpretation of the origin of many greenstone gold deposits. A num ber of guidelines are
offered to help distinguish pre-orogenic veins and deposits from those with syn- to postorogenic timing.

Introduction and Scope
G eologists have long noted the close spatial association
between gold deposits in greenstone belts and structures
such as faults, shear zones, and folds, as reflected by many of
the contributions in Newhouse (1042) and Canadian Insti­
tute o f Mining and Metallurgy (1048). A lthough there is
; wide acceptance that structures like shear zones, folds, and
% ^Corresponding author: e-mail, frobert@barrick.com
boudins are important loci of greenstone gold deposits at
some scale, there is no strong consensus on the geological
significance of this association (e.g., Mason and Helmstaedt,
1992; Hutchinson, 1993). Many of the conflicting genetic
interpretations o f gold deposits in greenstone belts, espe­
cially large complex deposits, stem in part from diverging
interpretations of the structural timing o f mineralization
(i.e., timing relative to their host structures). The contrast­
ing interpretations o f the Hemlo deposit by Pan and Fleet
(1992) and by Kuhns et al. (1994), and of the Campbell Red

111
Lake deposit by Andrews et al. (1986) and Penczak and
Mason (1997),serve to illustrate this poin t
Veins are common in greenstone gold deposits, in many
cases forming a major portion o f the ore, but overprinting
it in others (e.g., Tourigny et al., 1993; Robert, 1997). Veins,
like unconformities and igneous intrusions, arc also useful
in separating deformation predating their formation from
subsequent increments o f strain. T he analysis o f veins in
multiply deform ed metamorphic rocks can be valuable in
deciphering the sequence o f events leading to the forma­
tion o r deformation o f gold deposits. Furtherm ore, their
position, orientation, and timing relative to other strain fea­
tures such as shear zones, folds, and boudinaged compe­
ten t rock units, are factors o f direct interest in m ineral
exploration. In cases where veins and deposits can be
linked dynamically to an evolving structure, there are many
practical consequences concerning the predictability o f
their locations and orientations. In cases where deposits
have simply been overprinted by regional deformation giv­
ing the guise o f a relationship to major structures, the task
is to look through the deformation to deduce other causes
for gold localization.
T here are several m odem reviews of structural features
of gold deposits and districts (Bursnall, 1989; Eiscntohr et
al., 1989; Vearncombe et al., 1989; Mason and Helmstaedt,
1992; McCuaigand Kerrich, 1998). Similarly, the geometric
and kinematic features o f shear zones, so commonly associ­
ated with gold deposits, have also been extensively reviewed
(e.g., Ramsay, 1980a; Poulsen, 1986; Bursnall, 1989; Han-
m er and Passchier, 1991). However, there are com para­
tively few recent reviews emphasizing the structural charac­
teristics, significance, and controls on formation of veins in
greenstone gold deposits (e.g., Hodgson, 1989; Poulsen
and Robert, 1989;J6brak, 1992; Vearncombe, 1993).
The main objective of this contribution is to examine
structural aspects of veins in greenstone gold deposits, with
focus on the controls on vein formation, and on how veins
can be used in assessing the timing of gold mineralization
with respect to other geological events. The review also
serves as a basis for deciding am ong alternative processes
responsible for vein formation. In the absence o f conclusive
geochronological constraints, this interpretation is critically
dependent on the field criteria used to decide whether one
is exam ining features associated with vein form adon or
with vein deformation.
This paper combines aspects o f a review with those of a
discussion, and includes application o f the structural con­
cepts based on the authors’ experience. Illustrations largely
come from examples with which the authors are familiar,
and a number of deposits and districts are used repeatedly
throughout the text as examples. After a brief review of key
structural features of gold districts and deposits, the paper
is organized in three main sections. The first section covers
vein formation and analysis; it introduces selected relevant
theoretical concepts of vein formadon, followed by a review
of the different types o f veins, their structural characteris­
tics, and how they combine to form deposit-scale networks.
This section largely builds on previous work by the authors,
particularly Poulsen and Robert (1989) and Robert et al.
(1994). The next section addresses the effects of deforma­
tion o f veins in relation to geological structures such as
folds, shear zones, and mega-boudins. This leads to a third
section discussing the problem o f distinguishing syn-tec-
tonic veins from pre- or posttectonic veins. Key controls on
localization, shapes, and orientation o f orebodies are also
considered throughout the. text. The subject is approached
from the perspective o f the questions and problems facing
geologists exploring for, o r studying, greenstone gold
deposits. Although features discussed range from the scale
o f the district to that o f the thin section, outcrop-scale fea­
tures are emphasized for the most part.
Structure o f Greenstone Gold Districts
R ecurring lithologic and structural features o f gold
deposits a n d districts, and their evolution have been
reviewed by Card e t al. (1989), Vearncombe e ta l. (1989),
Groves and Foster (1991), Card (1992), Hodgson (1993),
Kerrich and Cassidy (1994), Groves et al. (1995), Groves et
al. (1998), McCuaig and Kerrich (1998), and Hagem ann
and Cassidy (2000). This section summarizes key aspects of
the geology o f greenstone gold disuicts to provide a frame­
work for discussing and analyzing veins in a m ore regional
context.
Lithologic composition o f districts
Greenstone belts consist of elongate units o f metavolcanic
rocks alternating with generally subordinate metasedimen-
tary rocks (Fig. 1), all m etam orphosed at greenschist to
lower amphibolite facies (e.g., Card, 1992). Mafic metavol­
canic rocks dominate most districts, with locally significant
ultramafic rocks and m inor volumes of felsic metavolcanic
rocks. Synvolcanic sills, dikes, and subvolcanic plutons
inuiidc metavolcanic rocks.
Metasedimentary successions also commonly occur in
greenstone gold disuicts (Fig. 1; e.g., Card etal., 1989). These
successions comprise older “flysh-like” graywacke-mudstone
turbidite sequences (Fig. 2), and younger “molassc-like,” allu­
vial-fluvial accumulations of polymictic conglomerate, aren-
ite, and mudstone (Figs. 3 and 4). Alluvial-fluvial sedimentaiy
successions unconformably overly volcanic rocks and the
older turbidite sequences, and represent an important time
marker in the evolution of the greenstone belts (Card, 1992).
Alluvial-fluvial sedimentary rocks are disuibuted along major
fault zones, where they occur as tectonic slices or as preserved
synclinal keels. Examples include the Timiskaming-type sedi­
mentary units of the Superior province (Card, 1992) and the
Missi type of the Trans-Hudson orogen (Stauffer, 1990) in
Canada, the Tarkwaian Group in the West-African greenstone
belts (Milesi et al., 1992), and the Merougil Formation and
Jones Creek Conglomerate in the Eastern Goldfields of West­
ern Australia (Krapez et al., 1997; Nelson, 1997a). In the
southern Superior Province, Timiskaming-type sedimentary
units are temporally associated with alkalic magmatisrn in the
form of quartz monzonite to syenite stocks and dikes, and
their locally preserved extrusive equivalents (Corfu et al.,
1991; Corfu, 1993).
 FlG. 1. Simplified geological map of the Abitibi greenstone belt showing the distribution o f major fault zones and sig­
nificant gold districts and deposits. Modified from Robert and Pouben (1997).

GREYWACKE, MUDSTONE t #• GOLD OEPOSIT, OCCURRENCE

MINOR CONGLOMERATE, B1F
greywacke. mudstone
[ '~ VOLCANIC ROCK
SHEAR ZONE
a QUARTZ-TOURMALINE-CARBONATE
F61SIC SYN- TO LATE-TECTONIC w VEIN TYPE
INTRUSION O QUARTZ-CARBONATE VEIN TYPE
[771 FELSICSYN-VOLCANIC INTRUSION A DISSEMINATED-STOCKWORK TYPE

Fig. 2. Geological map o f the Val d'Or district. Abitibi, showing the distribution of different t)pcs o f gold deposits in
the district. Modified from Robert ct al. {1995).
 Fig. 3. Simplified geological map o f die Rice Lake district, Manitoba,
showing the distribution o f gold deposits, faults, Timiskaming-type sedi­
mentary rocks, and intrusions. Modified from Poulscn et al. (1996).

Multiple generations of mafic to felsic intrusions, ranging

from batholiths to dikes, were emplaced throughout the
evolution of the greenstone belts (see Card, 1992; Kerrich
and Cassidy, 1994; Cassidy et al., 1998). O f these, felsic por­
phyry and lamprophyre dikes are particularly common fea­
tures of greenstone gold deposits and districts (Wyman and
Kerrich, 1989; Hodgson, 1993).
S tru c tu ra l evo lu tio n o f districts
Greenstone gold districts and deposits occur along
crustal-scale, first-order fault zones at boundaries between
contrasting lithological domains (e.g., Etsenlohr et al., 1989;
Hodgson, 1993). Examples include the Boulder-Lefroy fault
in Western Australia, the Porcupine-Destor and Larder
Lake-Cadillac fault zones in the Abitibi belt (Figs. 1 and 2),
the Wanipigow and Manigotogan fault zones in the Rice
Lake district, Manitoba (Fig. 3), and the McLennan tectonic
zone in die La Ronge district, Saskatchewan (Fig. 4). These
m ajor high-angle faults are commonly characterized by
anastomosing shear zones, intense L$ fabric (i.e., strong lin-
eation an foliation) developm ent and doubly plunging
folds, overprinting crenulation cleavages, and asymmetric
folds. They display evidence of complex, long-lived histories,
involving significant shortening accompanied by some com­
ponent o f high-angle reverse displacement, and overprint­
ing by strike-slip displacem ents (McCuaig and Kerrich,
1998). Along these faults, many districts are located at sites
o f structural complications such as bends and duplexes
(McCuaig and Kerrich, 1998) and intersections with cross
faults (Hodgson and Troop, 1988; Vcamcombe, 1998).
The dom inant structural grain o f many districts is subpar­
allel to the first-order faults and is defined by moderately to
steeply dipping supracrustal units, regional folds, and
numerous high-angle shear zones (Fig. 2). In detail, how­
ever, the structure of most gold districts is characterized by
the existence of multiple generations o f metamorphic fab­
rics, folds, and shear zones, indicating the presence o f sev­
eral recognizable increm ents of deform ation. Structural
analysis of gold districts is complicated by the highly hetero­
geneous nature o f the deform ation and by the domainal
developm ent o f structural fabrics (e.g., Hodgson and
Hamilton, 1989). However, three recurring principal incre­
ments of penetrative deformation can commonly be identi­
fied (Tfcble 1), even though the specific terminology used by
different authors varies from district to disuict (e.g., Hubert,
1990; Card, 1992; Milesi et al., 1992; Abraham and Spooner,
1995; Blceker and Parrish, 1996; Nelson, 1997b; Swager,
1997; H eather, 1998). Despite variations in their relative
importance from place to place, these three overprinting
increments of deformation account for the bulk of the sunc-
tural geometry o f most gold districts.
D, structural features include isolated, upright to recum­
bent folds with only local axial-plane cleavage (Fig. 5), and
localized layer-parallel penetrative foliation (Hubert 1990;
Lewry etal., 1990; Milesi etal., 1992; Swager, 1997; Heather,
1998). Low-angle thrusts are also well-documented effects of
D, in some districts, for example in the Kalgoorlie-Kam-

■
 T asil.F, 1, Common Deformation Increments in Gold Districts and Related Structures

Increment Structures Regime

Local evidence o f early sets o f folds or o f Unclear— possible extension?

lilting o f volcanic units
D, Localized foliation and cleavages “Thin-skinned" shortening
Tight to isoclinal, recumbent folds
Local thrust faults
Unconformity Deposition o f alluvial-nuviai sedimentary rocks Uplift and erosion

d2 Penetrative subvertical foliation and mostly "Thick-skinned” shortening

down-dip elongation lincations
Upright, tight to isoclinal folds
Moderately to steeply dipping reverse shear zones
D;i Localized crenulation cleavages Transcurrent deformadon
Asymmetric folds with steep plunges Reactivation o f preexisting structures
Tnmscurrcnt shear zones

Fie.. 5. Schematic representation o f trajectories o f the main structural fabrics and folds across the Val d ’O r district.
Modified from Robert (1990).

balda area in Western Australia (Swager and Griffin, 1990)
and in the Barberton Greenstone belt in South Africa
(Veamcombe et al., 1989). However, it is commonly the case
that the full significance o f the D| increm ent of deforma­
tion is difficult to establish because of overprinting by
younger structures. In a num ber of districts, the major D,
structures (folds and thrusts faults) trend at m oderate to
high angles to the dominant structural grain in the district
(e.g., the Rice Lake, Kalgoorlie-Kambalda, and N oranda
areas; Hubert, 1990; Swager and Griffin, 1990; Poulsen et
al., 1996). D, can be regarded as a period o f thin-skinned
thrusting and related isoclinal folding. In a few districts, D]
folds are also documented to overprint preexisting folds or
previously tilted volcanic units, reflecting the presence of an
earlier deformation event lacking associated penetrative fab­
rics and labeled D0 in Table 1 (Hodgson and Hamilton,
1989; Bleeker and Parrish, 1996; Swager, 1997).
In many districts, the unconformities at the base o f allu­
vial-fluvial (Timiskaming-type) sedimentary units truncate D,
folds (Hubert, 1990; Bleeker and Parrish, 1996) and thrusts
(Swager, 1997), indicating that sedimentation postdates the
D, increm ent of deformation (Card, 1992; Table 1). These
alluvial-fluvial sedimentary units are commonly overprinted
by tight to isoclinal upright folds and by a regional penetra­
tive foliation (Hodgson and Hamilton, 1989; Hubert, 1990;
Swager, 1997) commonly ascribed to D2.
D2 accounts for the most prominent regional penetrative
foliation, S2, parallel to the structural trend of the districts
(Table 1; Fig. 5). Sg foliation is typically subvertical and axial
planar to upright, tight to isoclinal, belt-parallel F2 folds
(e.g., Swager, 1997; Heather, 1998). It commonly contains
an elongation lineation, L2, with variable but most com­
monly down-dip plunges (e.g., Robert, 1990). In most cases,
S0 records horizontal shortening across the district with
accompanying subvertical to subhorizontal elongation in
the S2 plane. Moderate- to high-angle, reverse to reverse-
oblique shear zones are common in gold districts (Hodg­
son, 1993). They are products of D2, either as newly formed
structures o r as reactivated earlier structures, such as D }
thrusts steepened on the limbs o f F2 folds (Swager, 1997). Da
 ROBERT AND POULSEN

reverse to reverse-oblique shear zones are subparailcl to the 1990; Robert, 1990). This observation suggests the bulk o f
regional structural trends and in some cases form conjugate the deformation in these shear zones took place during D2.
pairs, recording deform ation consistent with D2 (e.g., The common overprinting of first-and second-order shear
Robert, 1990), which can be regarded as an increm ent of zones by crenulation cleavages and asymmetric folds
thick-skinned shortening across the greenstone b elt reflects their reactivation during D3 strike-slip deformation.
is characterized by steeply plunging asymmetric folds Third-order shear zones may form conjugate sets o f reverse
and Vertical axial planar crenulation cleavages overprinting or strike-slip shear zones (e.g., Robert, 1990; Poulsen et al.,
S,, and D2 shear zones. It may involve strike-slip reactivation 1996; Nguyen et al, 1998), consistent with their formation
of preexisting shear zones and, in some cases, formation o f during either D<_, o r Ds.
new strike-slip shear zones (Table 1; Hodgson and Hamilton,
1989; Hubert, 1990; Robert, 1990; Poulsen et al., 1996; Swa- Ool(l deposits
ger, 1997). En echelon folds, present in the Kalgooiiie-Kam- Where gold districts are distributed near first-order shear
balda area, have also been ascribed to D3 (Swager, 1997). In zones, it is well established that the majority of die deposits
some districts, D3 structures are best developed in the vidn- are actually hosted in higher-order shear structures
ity o f first-order shear zones, and present only in specific (McCuaig and Kerrich, 1998). This pattern o f shear zone
areas, reflecting a domainal distribution (Fig. 5). This third and deposit occurrence is well illustrated in the Val d ’O r
increment of deformation generally records an episode o f and La Ronge districts (Figs. 2 and 4). The reasons for this
late U'anscurrent deformation, resulting largely in suike-slip distribution o f mineralization remain unclear, but the pres-
reactivation of preexisting major shear zones (Card, 1992). ence o f gold-related alteration and some gold mincraliza-
In several districts, subsequent deformation increments tion in shear zones o f all orders suggest that they were inter-
have also been documented, involving development o f dis- connected in three dim ensions a t some stage in the
crete faults and kink bands oblique to the structural trend evolution o f the districts.
of the district, as well as further strike-slip reactivation o f In a large proportion o f gold deposits in greenstone
earlier structures (McRitchie, 1971; Hubert, 1990; Swager, belts, ore consists o f quartz-carbonate veins with a variable
1997; Heather, 1998). These features commonly have little com ponent of wall-rock sulfides. Single veins to complex
effect on the geometry and the structure o f the gold dis- vein networks commonly arc found in and adjacent to brit-
tricts and are not considered further here. tie-ductile to ductile shear zones in a variety o f host rocks,
In summary, it appears that most gold districts have expe- and deposits are vertically extensive (often over 2 km; e.g.,
rienced a similar structural evolution (Table 1), involving Hodgson, 1993). T he Sigma-Lamaque deposit at Val d O r
early, thin-skinned low-angle thrusting (Dt), followed by (Figs. 6 and 7), the San Antonio deposit at Rice Lake (Fig.
thick-skinned shortening (D2), evolving into transcurrent 8), and the deposits of the Star Lake district (Fig. 9) illus-
deformation (DH). In some districts at least, major D , struc- trate the range of complexity and nature o f deposit-scale
tures lie at a m oderate to high angle to the regional D2
structural trend, indicating significant differences in direc­
tions of crustal shortening between D t and D<, (e.g.,
Poulsen et al., 1996; Swager, 1997).
High-angle shear zones are an integral structural compo­
nent of gold districts. They can be grouped into crustal-scale
first-order structures, and smaller higher-order shear zones,
which form three-dimensional arrays of intersecting and
splaying structures (e.g., Eisenlohretal., 1989; McCuaig and
Kerrich, 1998; Nguyen et al., 1998). First-order crustal-scale
shear zones attain lengths of hundreds of kilometers and
are up to a kilometer wide. They are typically parallel to the
structural grain of the district. They are probably the
longest-lived structures and have undergone complex defor­
mation histories (e.g., Hodgson and Hamilton, 1989). Sec­
ond-order shear zones extend for a few to several tens of
kilometers and reach widths o f several tens o f meters,
whereas the more abundant third-order shear zones are less
than a few kilometers long and up to several meters wide.
Second-order shear zones are generally subparailcl to the
structural grain of the districts and third-order shear zones
are more typically oblique (in strike or dip).
The internal structure o f first- and second-order shear
zones is dom inated by an intense foliation and variably
developed elongation lineation parallel to the regional pic;. 6. Simplified geologicali map of iheSjgina-Umaquc deposit, Val
penetrative S2 and L., (e.g., Daigneault and Archambault, d’Or district. Modified from Rol>crt' ''
 Fig. 7. Cross sections through clifTerciU parts o f the Sigma-Lamaque deposit, Val d ’Or district; the cross sections are
located in Figure 6. Modified from Robert (1990). A. Section through the Sigma mine. B. Section through the Lamaquc
Main m ine. C. Section through the Lamaquc No. 2 mine.

vein and shear zone networks. Mineralized structures are
•.•.dominantly m oderate- to high-angle reverse to reverse-
oblique shear zones (Fig. 7), although strike-slip shear
zones host a small num ber o f deposits (Veamcombe et al.,
,1989; McCuaig and Kerrich, 1998). Quartz-carbonate vein
gold deposits are commonly interpreted to have formed in
compressional settings, at depths corresponding to the brit-
tle*ductile transition in the crust (Fig. 10; Sibson e ta l.,
1988). In some cases, however, rather than occurring in
association with through-going shear zones, mineralization
is hosted in fractured com petent units that have acted as
fluid conduits (Groves et al., 1995). This is the case at the
M ount Charlotte deposit in Kalgoorlie (Ridley and Men-
gier, 2000) and in part at the San Antonio deposit (Fig. 8),
where m ineralization is hosted in discrete structures o r
stockworks confined to com petent units within differenti-
 Fig. 8. Geology o f the San Antonio deposit. Rice Lake district. Modified From Poulsen et al. (1996.2000). A. Surface
map o f the San Antonio mine area. B. Cross section through the San Antonio deposit; line o f section shown in A. C. G eo­
logic plan o f level 7.
ated sills. In sedimentary rocks, the gold deposits can also
be hosted in folds, typically in association with reverse struc­
tures (Hodgson, 1993), as illustrated in Figure 10.
There are several structural styles of quartz vein orebod-
ies, including breccias and vein stockworks predominantly
within competent host rocks, continuous laminated quartz
veins in brittle-ductile shear zones, and thin, discontinuous,
highly attenuated veins in ductile shear zones (McCuaig
and Kerrich, 1998). These variations in structural styles of
quartz vein orebodies have been interpreted to represent
different erosion levels o f vertically extensive, crustal-scale
“orogcnic” systems, in which these different vein styles
formed synchronously (Colvine, 1989; Groves et al., 1995;
Gioves et al., 1998; McCuaig and Kerrich, 1998). However,
the existence o f deposits containing more than one age of
quartz veins within the sam e district, for exam ple a l Val
d ’O r (Robert, 1994) and a t Kalgoorlie (Clout etal., 1990),
complicates such an interpretation.
Additional ore styles, also located in o r near shear zones,
are present in gold deposits in greenstone belts. These
include dissem inated sulfide replacem ent zones, veinlet
stockwork zones, sulfide-rich veins, and massive sulfide
lenses (e.g., Robert and Poulsen, 1997). Some o f diese ore
styles arc unlikely to be related genetically to quartz veins
on the basis o f their timing, metal suites, and associated
alteration (e.g., massive sulfide lenses and sulfide-rich
veins). However, some deposits of disseminated-replace­
m ent and stockwork veinlet styles have more uncertain ori­
gins. In some cases they have been ascribed to the deeper
parts o f orogenic systems (e.g., Groves et al., 1995) and in
others to totally different genetic types o f deposits (Robert
and Poulsen, 1997). These complications and divergences
o f opinion reflect different interpretations o f the timing of
mineralization relative to its host structure. Nonetheless,
veins are present in nearly all variations in greenstone gold
deposits, and can be used as structural m arkers to place
constraints o n the tim ing o f ore relative to its host struc­
ture, and to place it in the context o f the structural evolu­
tion o f its host district.
Veins and Their Analysis
This section examines the characteristics o f veins com­
monly found in greenstone gold deposits and provides a
basis for their structural analysis. After a review o f theoreti­
cal concepts o f fracture dilation and vein formation, con­
sideration is given to vein classification. This is followed by
a review o f structural and textural characteristics o f differ­
ent types o f veins and vein arrays, leading to a discussion o f
 NW

KAHN 18
OREBODY

Fk;. 9. Star Lake district, La Rongc belt. Modified from Poulscn ct a). ( 1986b. 2000). A. Simplified geological map o f
the district. B. Geological map o f the 21 Zone (Star Lake) deposit. C. Cross section through die 2 1 Zone deposit; line o f
section shown in B.

structural analysis in gold deposits and districts. Although

mineralized veins are emphasized, much of the material
presented applies equally to barren veins that overprint
mineralization.
Theoretical concepts o f ve in fo rm a tio n a n d analysis

Veins and fracture dilation: The term “vein" is used here in

its strict sense to designate the hydrothermal material fill­
ing the space between separated fracture walls (e.g., Ram­
say and Huber, 1987). In a num ber o f cases, especially his­
torically, a distinction has been made between fissure veins,
those filling the space between separated fracture walls,
and replacem ent veins, those resulting from metasomatic
wall-rock replacem ent along fractures without dilation
(e.g., Cooke, 1946; Hodgson, 1989). Figure 11A illustrates
such a replacement vein, across which there is no dilation Fic. 10. Schem atic cross-sectional representation o f the com m on
and no offset o f preexisting markers. Replacem ent veins geom etry and structural setting o f quartz vein networks in greenstone
are not considered further in this paper because they do belts. Modified from Poulsen ct al. (2000).
 Fig. II. Geometric aspects o f fracture dilation leading to vein formation, applicable both to cases o f dilation o f pre­
existing fractures and o f newly formed fractures. A. Replacement vein developed along a fracture. B. Orthogonal exten­
sion o f a fracture. C. Oblique extension o f a fracture. D. Oblique extension resulting from shear displacement along a
non-planar fracture. See text for discussion.
noi convey any specific structural inform ation, but the
reader is referred to Hodgson (1989) fo ra treatment of this
topic.
From a conceptual point o f view, different schemes of
fracture dilation can lead to form ation of veins (Fig.
11B-D). Dilation can take place by orthogonal opening of
die fracture walls, as marked by the lack of shear displace­
m ent o f wall-rock markers and the alignm ent of mineral
fibers perpendicular to the fracture walls (Fig. 11B). Dila­
tion can also take place by oblique opening, indicated
both by the offset of the markers and by oblique alignment
o f fibrous minerals (Fig. 11C). An identical offset o f wall-
rock markers could result from an appropriate am ount of
shear displacement along the fracture followed by orthog­
onal opening. In such a case, however, fibrous minerals
would be found to be perpendicular to the vein as in Fig­
ure 1IB. Shear displacement along a planar fracture does
not normally lead to its dilation (Fig. 1ID). However, it is
well known that dilation o f fractures with this movement is
theoretically possible where there are favorably oriented
curves o r jogs. In geometric terms, dilation in such a jo g is
actually a special case o f oblique extension with only a
small angle between the vein and the direction o f opening
(F ig .llD ).
The dilation schemes shown in Figure 11 apply equally to
cases where a preexisting fracture, even a lithologic contact
o r a foliation plane, has been subsequently dilated, and to
cases where both the formation and dilation of a fracture
are part of a continuous process. Geologists have tradition­
ally dealt with the interpretations o f veins formed by dila­
tion o f fractures with two complementary approaches, dis­
cussed in the next two sections. O ne involves a dynamic
analysis, linking the dilation o f fractures to the causative
stresses that create them (e.g., Moritz and Crocket, 1990;
Cox, 1995; Nguyen e t al, 1998; Ridley and Mengler, 2000).
The o th er approach, adopted in this paper, considers the
dilation of veins as a contribution to the overall strain in the
host rocks (e.g., Ramsay, 1967; Robert and Brown, 1986;
Dube et al., 1989; Tourigny et al., 1989; Robert et al., 1994).
The predictive capacity o f dynamic analysis of veins is well
illustrated by the contributions o f Cox et al. (2001) and Sib-
son (2001) in this volume. In this particular paper, however,
we emphasize the relationships between veins and strain, as
well as stress. O ne advantage of this approach to green­
stone gold deposits is that they occur in environm ents
where, in addition to veins, one commonly encounters
many other manifestations o f strain in rocks including foli­
ation, lineation, shear zones, folds, and boudinaged com­
petent units. Linking veins to such structures provides a
broader framework for their analysis.
Relationshipsbetween tockfracturing, stressaxes, fluid pressure,
and veinformation: In numerous cases, the development of
fractures, their opening, and their filling (to form veins) can
be demonstrated to be part o f a continuous process. This is
well established in the case o f extensional fiber veins (Ram­
say and Huber, 1987), for laminated veins associated with
small shear fractures (Labaume et al., 1991; Petit etal., 1999;
Koehn and Passchier, 2000), and for a num ber of auriferous
veins (e.g., Boullierand Robert, 1992). Principles governing
rock fracturing, therefore, provide an appropriate frame­
work for discussing fracture and vein formation, their rela­
tionship to stress axes, and the important role o f fluid pres­
 sure. Only key concepts are reviewed here, because the
topic is covered more extensively in other chapters o f this
volume (Cox et al., 2001; Sibson, 2001). The reader can also
refer to the reviews of Cox etal. (1987), Kerrich (1989),and
Sibson (1990a), among others.
The Mohr-Coulomb theory o f rock failure predicts the
form ation of extensional, extensional-shcar, and shear
fractures in m eaningful directions relative to stress axes,
depending on the magnitudes o f stress and fluid pressure
(Fig. 12A). Because it acts equally in all directions, fluid
pressure (Pr) counteracts norm al stresses such that the
effective normal stress acting on a plane is defined as o f
= o n - Pf. Fracturing o f a rock mass is governed by the
effective principal stresses, defined as (Xj 2 f ■ 0\ a ^ - Pf
(Fig. ISA). ..................
Shear fractures develop at ~30° to CTj (i.e., along planes
offering minimal shear resistance; Fig. 12B) when the accu­
m ulating shear stress acting on a plane (r) exceeds the
shear resistance of the rock, at a minimum value of differ­
ential stress. Conditions for shear fracturing are expressed
by the equations r> C + p f and - o f > ~6T (where C
= cohesive strength, fi = coefficient of friction, and T * ten­
sile strength of the rock; see Sibson, 1990a). An increase in
fluid pressure reduces the effective normal stress acting on
the plane, but not the shear stress, thereby facilitating shear
fracturing of the rock. Departures from the predicted ori­
entations of shear fractures are expected in cases where the
host rock has a marked layer anisotropy, such as bedding or
penetrative foliation, depending on the orientation of the
layer relative to the principal stress axes (Peacock and
Sanderson, 1992).
Extensional fractures typically develop parallel to cr, (Fig.
12B when o f equals or exceeds the tensile strength of the
rock. This situation is only possible under conditions of low
differential stress, where o f t - T and o f - a f< 4T. Exten­
sional fracturing can, thus, only occur at negative ralues of
o f (Fig. 12A), a condition only possible through derated
fluid pressure in the inferred compressional environments
of formation o f greenstone gold deposits (see Sibson et al.,
1988; McCuaig and Kcrrich, 1998). Extensional failure by
hydraulic fracturing takes place when Pf > <r3 + T.
Extensional-shear fractures are hybrid between exten­
sional and shear fractures. They develop at an angle of less
than 30° to <7, (Fig. 12B), at negative values o f o f (Fig.
12A), and under the following conditions of differential
stress (Sibson, 1990a): 4T £ (<y,' - o f ) £ ~6T.
Based on these three modes of fracturing, one can pre­
dict the formation o f three corresponding types o f veins
.• (Fig. 12C). These veins are here designated extensional,
.oblique-extension, and fault-fill veins, and can be catego-
; rized by comparing the direction o f their inferred opening
vectors to the orientation o f their walls (Fig. 12C). The
result is a continuous spectrum in which opening or dis­
placement vectors range from perpendicular to subparallel
to the vein walls (Fig. 12C). By definition, the opening vec­
tor of extensional veins is perpendicular to the walls of the
veins, whereas in the case o f oblique-extension veins, the
opening vector is oblique to the walls. In shear veins, the
opening vector corresponds to the relative displacement of
the walls and is subparallel to the vein. A num ber o f studies
of gold-quartz vein deposits in deform ed terranes have
shown the existence of veins that correspond to these three
different types (e.g., Robert, 1990; Cox, 1995; Nguyen et al.,
1998). T he orientations o f the veins produced will be
highly dependent on the orientation of the principal stress
axes and on the prevailing stress field. Extensional veins are
predicted to be subhorizontal in compressional stress fields
(Fig. 12C), but subvertical in transcurrent and extensional
regimes (e.g., Sibson, 1990a). Similarly, fault-fill veins are
predicted to form at dips o f -30° in compressional stress
regimes, -60° in extensional regimes, and subvertical in
transcurrent regimes.
Gold-quartz veins in greenstone belts are interpreted to
have form ed dom inantly u n d er compressional regimes,
with a significant num ber o f fault-fill veins associated with
high-angle reverse structures (e.g., McCuaig and Kerrich,
1998). U nder such conditions, subhorizontal extensional
veins likely form as hydraulic fractures at mid-crustal levels
only where fluid pressure equals o r slightly exceeds litho-
static pressures, represented by <J3 in this case (Etheridge,
1983; Sibson et al., 1988; Cox et al., 2001). The association
of fault-fill veins with high-angle reverse shear zones departs
from their predicted shallow dips in compressional
regimes. This can be explained by the fact that, under con­
ditions o f high fluid pressures, preexisting high-angle struc­
tures and lithologic contacts are reactivated preferentially
to the formation o f new appropriately oriented, low-angle
reverse faults (Sibson e t al., 1988). Finally, under lithostatic
fluid pressure and very low differential stresses, orthogonal
dilation of planes o f weakness o th er than those oriented
perpendicular to is theoretically possible, including sub-
vertical foliation in high-angle shear zones, as proposed by
Kerrich and Allison (1978), and Kerrich (1989).
Consideration of veins as a component o f bulk strain: Another
approach to analysis o f veins in greenstone terranes is to
consider that the developm ent o f extensional, oblique-
extension, and fault-fill veins constitutes one o r more incre­
ments o f bulk strain in a rock mass and, therefore, can be
analyzed with reference to external axes o f strain (Fig. 13A;
e.g., Ramsay, 1982).
This approach allows for consideration o f the dilation o f
preexisting fractures, as well as o f new fractures formed by
failure while the bulk strain accumulated, and it requires a
less genetic terminology for describing the veins, as used in
figure 12C. It also allows for a comparison o f axes of bulk
strain indicated by die veins with those indicated by other
manifestations o f strain in their host rocks, such as folia­
tion, elongation lineations, shear zones, folds, and boudi-
naged com petent units. This approach provides a useful
framework for the interpretation of the evolution, timing,
and possible controls of the veins, as well as for considering
the effect of superimposed deformation.
The opening of veins o f diverse orientation can result in
n et dilation of the host rock in one or more directions (Fig.
13A). Unless counterbalanced by mctasomatic volume loss
 Failure envelope
Shear fracturing for intact rocks

E x ten sio n s!
fracturing

Extensions!
fracturing

O ^ C T j- P f cr,' = o , - pf

<*3
4

^1 G1 CT1

Extensions! Extensional-shear Shear fracturing

B fracturing fracturing (faulting)

MINERAL FIBERS

I dZ dZ

Extension Oblique-extension F au lt-fill

vein vein vein

Fl(i. 12. Relationships between stress and fluid pressure conditions, rocks fracturing, and vein formation. A. Mohr
representation of stress suite conditions required for cxtcnsional, oblique-extension, and shear fracturing of intact rocks,
and the effects of fluid pressure on effective stresses and fracturing (adapted from Sibson, 1990). The inset shows the
angular relationships between the effective principal stress axes (<jj\ <%'), effective normal stress (o’/ ) , and shear stress
(r). B. Illustration of the angular relationships of the three npcs of rock fractures and the principal stress directions. C.
liJustnuion of three types of veins expected to result from formation and opening of fractures in intact rocks, and of the
axes of incremental strain represented by the formation of such veins.
 Fig. 13. A. Isometric block diagram illustrating how dilation o f extension, oblique-extension, and fault-fill w ins might
contribute to the bulk strain o f a rock mass. As illustrated, vein dilation contributes to extension only in the dX direc­
tion but is compensated by unidirectional volume loss along dZ. B. Isometric block diagram o f an ideal mesh o f fault-fill,
oblique-extension, and extension veins. The shear sense on the shear zones and the opening vectors in the veins arc
compatible with a single increment o f plane, coaxial strain with no net volume change.

in the surrounding rock, this might result in a net volume Veins, however, may be only one manifestation of discon­
increase for the bulk rock (Ramsay and Wood, 1973). For tinuous strain in the bulk rock, and may be accompanied
example, the Sigma deposit (Fig. 7) contains approxi­ by the developm ent o f shear zones (Fig. 13B) which,
mately 100 significant subhorizontal quartz veins, on aver­ depending on orientation, need not have dilated (Hodg­
age 25 cm thick, plus more abundant but less well-recorded son, 1989). Such a mesh o f shear zones can, independent
thinner veins, distributed over a vertical interval o f 2,000 m
(Robert and Brown, 1986; M. Crevier, pers. com m un.,
2000). The veins are, thus, estimated to contribute to a ver­
dX
tical dilation of the rock mass on the order o f 1 to 2 per­
cent, even though they are discontinuously distributed.
Nevertheless, considerations o f volume change aside, the
vein-related dilation involved in many gold deposits is small
compared to that indicated by observation o f other features
in surrounding rocks. Shape fabrics in deformed conglom­
erate o r pillow basalt in greenstone belts commonly indi­
cate shortening o r extension exceeding 100 percent. This is
one o f the reasons why dilation o f veins can be treated
effectively as an increment of strain (Etheridge et al., 1984).
Some of the preexisting fractures and planes of weakness
in a deforming rock mass, such as lithologic contacts and
foliation planes, are m ore favorably oriented for dilation
than others. For a given two-dimensional strain (Fig. 14),
fractures with poles in the field of predicted elongation are
expected to dilate, unlike those with poles in the field of
shortening. For fractures with poles in the shortening sec­
tor, dilation is only possible at favorably oriented bends or
jogs along the fracture plane if it is activated as a shear,
according to the case illustrated in Figure 1 ID. The three-
dimensional distribution of fracture orientations favorable
for dilation is determ ined by the shape o f the strain ellip­
soid and is governed by the same principles that apply to FlC. II. Strain ellipse showing the expected sectors o f lengthening and
shortening during incremental horizontal shortening and vertical clonga-
the development o f boudins and folds during progressive don, as well as die orientation o f preexisting fractures for which dilation is
deformation (Flinn, 1962; Chapter 4 in Ramsay, 1967). expected.
o f contributions from vein dilation, contribute discontinue
ously to bulk suain (Ramsay and Huber, 1987), as can other
more familiar and more continuous geological features like
penetrative rock foliation and lineation.
There are several advantages to analyzing veins in terms
o f incremental strain axes in greenstone belts, because sev­
eral factors complicate the reconstruction o f their causative
stresses. The stress approach is only viable where there has
been very limited post-vein strain and tilting, something not
always possible to ascertain in multiply deformed terrancs
such as greenstone belts. Layer anisotropy, omnipresent in
gold districts and deposits, imparts significant refraction of
strain axes across layers o f contrasting competencies (e.g.,
Treagus, 1985; Hodgson, 1989). This effect will result, for
exam ple, in significant changes in orientation o f extern
sional veins across different layers (Fig. 15), which poses a
significant challenge to the reconstruction o f external
stress axes responsible for vein formation. Finally, the strain
approach allows for the analysis o f veins fonned by dilation
o f preexisting fractures, whose orientation may bear no
genetic relationships to the stress field at the time o f vein
formation (Ramsay, 1967).
Classification of Veins
A classification scheme for veins and vein arrays provides
a useful framework for their p roper identification, for
assessment o f their structural significance, and for struc­
tural analysis. In the context o f greenstone gold deposits,
veins have been classified on the basis of cither the nature
of their host fracture (e.g., McKinstry, 1948; Ramsay, 1967;
Hodgson, 1989; Poulsen and Robert, 1989; Cox, 1991) or
of their internal structure and texture (e.g., Hodgson,
1989;Jdbrak, 1992; Vearncombe, 1993). The internal struc­
ture and texture of the veins generally reflects the nature of
their host fracture or structure (e.g., shear versus exten-
sional fractures) and the mechanisms responsible for their
dilation.
The practical classificadon of veins adopted here (Table
2) is based on the host structure, geometry, and internal
features of the veins, all aspects that are generally observ­
able in the field. It is a hybrid classification in the sense that
it considers both individual veins and sets of intimately asso­
ciated veins, such as vein arrays and stockworks. Building
on the classifications of Hodgson (1989), Poulsen and
Robert (1989), Robert (1990), Sibson (1990a), and Cox
(1991), veins and vein groups are distinguished in this
paper as fault-fill veins, extensional (and oblique-exten­
sion) veins, extensional vein arrays, vein stockworks, and
breccia veins (Table 2).
Fault-fill veins, as their name indicates, typically occupy
faults and the central parts of shear zones (Fig. 12C)> where
they may be parallel or slightly oblique to their host struc­
ture (Fig. 16; Hodgson, 1989). T he term fault-fill vein is
preferred over shear vein l>ecause it is commonly unclear if
these veins actually formed as a result of movement along a
shear fracture (Fig. 11D ),or asaresult o f extensional open­
ing o f preexisting shear fractures o r foliation planes (Ker-
rich, 1989; Cox, 1991). Extensional veins occur as planar
Fig. 15. Refraction o f extensional veins across the contact between a
feldspar porphyry dike (left) and folded banded iron formation (right).
Note the difference in orientation o f die opening vectors (white arrows)
between the two rock types. Section view. Victory mine. Western Australia.
veins in low-strain rocks outside shear zones especially in
com petent units, as arrays o f en echelon sigmoidal o r pla­
nar veins within shear zones where they commonly fringe
fault-fill veins, o r as arrays o f stacked planar veins in com­
petent layers. Extensional veins typically lie at a high angle
to both local foliation and lineation (Fig. 16). Oblique-
extension veins are grouped with extensional veins in this
paper, in the light o f close similarities in their internal fea­
tures and form adon mechanisms. Vein stockworks consist
of multiple three-dimensional sets o f veins, most typically of
extensional or oblique-extension types, and are best devel­
oped in com petent lithologic units. Breccia veins (breccias
with a hydrothermal matrix) are commonly associated with
faults and shear zones, where they cut competent lithologic
units. Intense stockwork developm ent can also lead to
developm ent o f breccia zones. The setting, geom etric
arrangements, internal features, and formation of these dif­
ferent types o f veins are examined in detail in the following
sections.
Fault-fill veins
Fault-Fill veins are by far the most common type in gold
deposits and are the source o f most o f the vein-type ore
extracted from these deposits. Fault-fill veins are charac­
terized by the presence o f lam inated quartz, of slicken-
sides a n d siickenlines, and o f foliated wall-rock slivers.
They form elongate lenses within th eir host structures
and, thus, define oreshoots with plunges corresponding
to their long axes. Fault-fill veins are commonly hosted by
brittle-ductile to ductile sh ear zones, an d the rea d e r is
referred to Ramsay (1980a), Poulsen (1986), Bursnall
(1989), and H anm er and Passchier (1991) for reviews of
the geom etry, internal characteristics, a n d kinem atic
interpretation o f such structures.
 T a b i.f. 2. Classification and Main Characteristics of Different Types ofVeinsand Vein Arrays
Vein type Internal features Structural site Geometry
Fault-fill vein Laminated structure Shear zone or fault, especially Parallel or slightly oblique
Foliated wall-rock slivers at bends and dilation a! jogs to host structure
Slip surfaces Fold limbs, thrusts Lenticular veins common
Fibers at low angle to vein walls
Extcnsional vein Internal layering Outside shear zones Planar veins at moderate
(oblique- Open-space filling ACjoints in folds angle to shear zone
extension vein) Mineral fibers at high (moderate) At high (moderate) angles Perpendicular (oblique)
angle to vein walls to competent units to fold hinge
Rock bridges
Massive quartz
Extcnsional Internal layering: multiple Shear zones En echelon planar to
vein array openings Competent layers sigmoidal veins
Stacked planar veins
At a high angle to foliation
or competent layer
Vein stockwork Two or more oblique to Non-specific but common at Tabular to cigar-shapcd
orthogonal vein sets of vein intersections zones
extcnsional or oblique Preferentially developed in
extension veins competent lithologic units
Breccia texture developed in
intense stockworks
Breccia vein:
1.Jigsaw puzzle 1. Angular wall-rock clasts, no 1. /Xiong faults Parallel to host structure
(implosion) rotation, hydrothermal matrix
breccia
2. Fault breccia 2. Vein and wall-rock clasts with 2. Fault or shear zone;
rotation and abrasion, component o f fault-fill veins
hydrothermal matrix

F ig . 16. Common geometric arrangements o f fault-nil and extcnsional veins in shear zones, and their relationship to
incremental axes o f shortening (dZ) and elongation (<IX). A. Fault-fill veins in the central pan o f a reverse shear zone
showing conflicting crosscutting relationships with planar extcnsional veins extending outside the shear zone. B. Arrays
o f en echelon sigmoidal extcnsional veins within shear zones. C Arrays o f stacked planar extcnsional veins within shear
zones; these arrays can also develop in competent layers such as dikes.
 Geometric arrangements: Most fault-fill veins form mineral­ ated wall rock (Fig. 18B, C) o r by millimeter-thick septa likely
ized lenticular bodies in discrete faults or, more commonly, derived from the wall rocks. They can also be separated by
in the central parts o f shear zones. Several fault-fill veins, slip surfaces enhanced by the presence o f hydrothermal min­
separated by barren segments, may occur along strike or erals such as tourmaline, sericite, or chlorite (Fig. 18D, E). In
down dip within a single structure (Fig. 16A). Within shear o th er cases, the laminated character o f the veins is only
zones, fault-fill veins are either parallel or at low angles to defined by subtle differences in the color and texture o f
the shear zone boundaries, respectively distinguished as quartz. Individual laminae within fault-fill veins are generally
central and oblique types by Hodgson (1989); they are also parallel or subparallel to the vein margins.
at a low angle to the shear zone foliation. Fault-fill veins can Laminated fault-fill veins range from sheeted veinlet
also be associated with folds (in dilational jogs along limb- zones, in which the proportion of the wall-rock component
thrusts crossing bedding and along bedding on fold limbs, is approximately equal to or greater than that of the vein
corresponding to neck and leg reefs, respectively; Hodg­ com ponent, to book-textured or ribbon-textured veins in
son, 1989), as well as saddle reefs in fold hinges (Hodgson, which die vein component dominates (Hodgson, 1989). In a
1989; Cox etal, 1991). num ber o f fault-fill veins hosted by shear zones, observed
A majority o f fault-fill veins occur in m oderately to variations in vein to wall-rock ratio and in vein morphology
steeply dipping structures (>45*; Figs. 7-9), as reflected by define a lateral o r vertical zonation (Fig. 16). Fault-fill veins,
their dips in cross section (e.g., Colvine, 1989; Foster, 1989; near their lateral and vertical terminations, grade into iso­
Groves et al., 1995). A relatively small num ber o f fault-fill lated veinlets separated from each other by significant thick­
veins also occur in shallowly to moderately dipping suoic- nesses of wall rock. Such isolated veinlets are relatively com­
tures, such as those at Grass Valley, California (Sibson, mon in barren segments o f shear zones between fully
1990b), and Victory, Australia (Clark et al., 1986). developed fault-fill veins. Moving laterally towards die cenu-al
part of the fault-fill veins, individual veinlets increase in abun­
Internal features: Internally, fault-fill veins commonly con­ dance and are more closely spaced, defining what can be
sist o f laminated quartz (Figs. 17 and 18A-C; Hodgson, 1989; regarded as a sheeted veinlet zone (Fig. 17). These veinlets
Poulsen and Robert, 1989;Jebrak, 1992). Individual quartz are parallel or subparallel to die foliadon in die host shear
bands o r laminae are commonly separated by slivers of foli­ zone, and their dimensions are similar to those of individual

FlG. 17. Characteristics o f fault-fill veins in shear zones. A. Schematic representation o f tatcral zoning in the propor­
tion o f vein to wait rock along a shear zone. B. Line drawing, plan view, o f a fault-fill vein illustrating how individual vein-
lets amalgamate to form larger laminated quartz lenses {Sigma mine, Vnl d'Or).
 Fig. 18. Characteristics o f fault-fill veins. A. Laminated fault-fill vein, plan view. Internal quartz laminae arc separated
by dark (tourmaline-rich) semi<ontinuous to discontinuous wall-rock septa (1) and slip surfaces (2). Some o f the indi­
vidual quartz laminae have a breccia texture (3), whereas others have a massive texture (-1). Note also how some of the
quartz laminae have developed by reopening o f an existing vein (4). Sigma deposit, Val d’Or, plan view. Vein is 1 m thick.
B. Laminated fault-fill vein containing strongly foliated slivers o f wall rock (dark). Pamour deposit. Timmins, down-dip
view. Hammer for scale. C. Laminated fault-fill vein. Rod main zonc.Jolu deposit, La Rongc, cross section view. Hammer
for scale. D. Laminated fault-fill vein in which individual quartz laminae arc separated by discrete slip surfaces, lined with
flne-grained tourmaline (1). Note also the sigmoidal nature o f the lateral terminations o f individual quartz laminae (2).
Sigma dc|>osit, Val d’Or. oblique view. Wood blocks are -25 cm long. Courtesy o f the Geological Suivcy o f Canada, photo
no. I999-015S. E. Laminated fault-fill vein in which individual quartz laminae arc separated by discrete slip surfaces
coated with tourmaline (1). Note how a subhorizontal cxtcnsional vein in the hanging wall truncates the external pan
o f the fault-fill vein (2), to merge with slip surfaces well within the vein. Lucicu Beliveau deposit, Val d ’Or. cross section
view. Hammer at bottom right for scale. F. Stepped quartz and tourmaline fibers (slickcnlincs) along a slip surface widtin
a fault-fill vein, clearly indicating diat mineral precipitation accompanied reverse slip along the fault-fill vein. Lucien
Beliveau deposit, Val d'Or. Reprinted with permission o f Economic Grolngf, Robert, 1996, fig. 3. G. Fault breccia on the
margin o f a laminated fault-fill vein. The breccia consists o f angular clasts o f altered wall rocks (dark) in a hydrothermal
matrix o f quartz (white). Panunaque deposit, Val d ’Or. Plan view. Scale bar in centimeters.
128

laminae comprising fault-fill veins. The central parts o f fault-

fill veins are dominated by quartz laminae with subordinate
proportions o f wall-rock slivers, which commonly become
thoroughly altered and completely replaced by hydrother­
mal minerals (e.g., Robert and Brown, 1986). It is important,
in a drilling program, to recognize whether such lateral zon­
ing is present in fault-fill veins because relatively abundant
foliation-parallel veinlets may indicate the off-hole proximity
o f thicker fault-fill veins. The central o r thicker parts of fault-
fill veins can commonly result from the merging and juxta­
position of numerous quartz veinlets (Fig. 17B) due either to
reopening of existingveins (Fig. 18A), or to the formation of
new veinlets immediately adjacent to or overlapping with
existing ones. Accordingly, the incorporation of wall-rock
slivers within the veins is generally considered to be a natural
consequence of their incremental growth (e.g., Robert and
Brown, 1986; Moritz and Crocket, 1990).
Striated fault surfaces, or slickensides, are also commonly
observed along wall-rock slivers o r along the interfaces
between individual quartz laminae within fault-fill veins (Fig.
18F; Hodgson, 1989). In some cases, hydrothermal minerals
such as quartz, tourmaline, or calcitc define the striae, or
slickentines. Such striae are term ed slickenfibers and they
clearly indicate that vein development was accompanied by
slip along the veins. In some other instances, steps defined by
the slickenfibers further indicate the sense o f movement
(Fig. 18F). Such vein structures are exactly those expected to
form in fault-fill veins as defined above. Finally, in other rela­ Fig. 19. Schematic representation o f the geometric relationships
tively rare instances, dilational jogs are also observed along among structural elements of shear zones, the w ins they contain, and the
fault-fill veins, consistent with reverse movements (fig. 16e in corresponding incremental strain xxes. Lineations and suiations on slip
Robert, 1990; fig. 3c in Nguyen et al., 1998). Such dilational surfaces w ithin fault-fil! w in s a re subparallcl to the movement direction
(w hite arrow ); lo n g axes o f fault-fill veins a n d lines o f in te rse ctio n of
jogs not only indicate that vein development is accompanied
extensional veins are p erpendicular to the m ovem ent direction. See text
by slip, but they also indicate the sense of movement. It is not for discussion. A dapted from Pouiscn (1986).
dear at present if such dilational jogs along fault-fill veins are,
truly rare features, or if they are common but not recognized
because they occur at scales exceeding those of vein expo­ cases foliated) and vein material, enclosed in a hydrothermal
sures in outcrops or within stopes. matrix dom inated by quartz (Fig. 18G). The existence of
Where present within laminated fault-fill veins, slicken- fault breccia within fault-fill veins indicates active slip during
lines or slickenfibers commonly show consistent rakes within mineralization. It further indicates seismic slip during vein
single veins or sets of veins within a deposit (e.g., Robert and formation, in contrast to the more ductile flow commonly
Brown, 1986). Furthermore, where complete datasets exist, recorded in the host shear zone (e.g., Robert et al., 1995).
slickenlines indicate the same direction of movement as the
elongation lineation in the host shear zone (e.g., Robert, Dilation mechanisms: Several detailed studies provide good
1990; Nguyen et al., 1998). Where stepped slickenfibers or evidence for formation of the fault-fill veins during active
dilational jogs are observed, they indicate the same sense of slip along their host structure, based on structural features
movement as that deduced for the host shear zones (Fig. 19). described above (e.g., Robert and Brown, 1986; Dube et al.,
There seems to be good compatibility in direction and sense 1989; Nguyen e t a l, 1998). It is also well documented that
of shear between discrete slip events within the veins and a majority o f these veins have formed during active reverse
ductile flow in the host shear zone, indicating that both are slip (e.g., Vearncom be e t al., 1989; Hodgson, 1993;
part o f the same progressive deformation event (e.g., Robert McCuaigand Kerrich, 1998). However, the mechanisms for
and Brown, 1986; Nguyen et al., 1998). However, such com­ dilation and opening o f these veins are no t firmly estab­
patibility is not expected in cases where slickenlines are pro­ lished. Two mechanisms are most commonly proposed to
duced during postore reactivation of fault-fill veins and their explain the formation of fault-fill veins. In the first, dilation
host shear zones. occurs at releasing bends and jogs at a low angle to a shear
Some fault-fill veins, o r individual laminae within them, fracture or fault during slip (Fig. 11D; McKinstry, 1948; Sib-
consist largely of fault breccia and cataclasite (e.g., Nguyen et son, 1990a). This process may be recorded by crack-seal
al., 1998). These breccias are composed o f rotated angular textures parallel to the jo g and inclusion trails and mineral
clasts of variable sizes, made up of altered wall rocks (in some fibers parallel to the slip direction (Ramsay and Huber,
198S; Labaume e t al, 1991; Koehn and Passchier, 2000).
The second mechanism, designated crack-seal-slip, involves
dilation resulting from repeated episodes o f orthogonal or
oblique extension o f a shear fracture, alternating with
episodes of slip along the fracture (Jebrak, 1992; Petit et al.,
1999). The presence o f dilationaljogs (fig. 3c in Nguyen et
al., 1998) and o f foliation-parallel veins with orthogonal
mineral fibers (fig. 8 in Robert and Brown, 1986) indicates
that both mechanisms can operate to form fault-fill veins.
Other, less commonly invoked dilation mechanisms are
reviewed in Hodgson (1989).
Associated ore shoots: Ore shoots are common characteris­
tics of fault-fill veins in lode gold deposits in shear zones,
and determ ination of their plunge is a daily concern of
exploration and m ine geologists. O re shoots can be
defined by the elongate shape of fault-fill veins, by the
increased grade a n d /o r vein thickness at shear zone bends
or vein and shear zone intersections (Fig. 19), and by inter­
section of shear zones with chemically (e.g., banded iron
formation) or theologically (e.g., com petent dike) favor­
able layers (Poulsen and Robert, 1989). The plunges of ore
shoots in fault-fill veins are generally predictable, because
they bear specific geometric relationships to some struc­
tural elements of the vein, host shear zone, or entire vein
network. Ore shoots at splays and shear zone intersections
are perpendicular to the movement direction, which is also
parallel to the intersection of extensional veins with the
shear zones (Fig. 19). Ore shoots resulting from repeated
slip at dilational bends or jogs also tend to be perpendicu­
lar to the m ovement direction and their orientation is
clearly related to shear zone kinematics. O re shoots can
also be parallel to the slip direction along the host shear
zone, as pointed out by Veamcombe et al. (1989), particu­
larly if the veins are significantly overprinted by strain dur-
. ing continued, postdepositional shear movement.
O ther types o f ore shoots are controlled by the orienta­
tion o f the shear zones relative to intersected layers rather
. than by kinematics (Poulsen and Robert, 1989). This is the
case o f ore shoots a t the intersection between the shear
zones and chemically or Theologically favorable layers: the
plunge o f the resulting ore shoots is parallel to the line of
intersection. In addition, shear zones are commonly
deflected where they crosscut contrasting lithologies; the
resulting bends and jogs may become the sites o f ore shoots
whose plunge will again be parallel to that o f the intersec­
tion between the shear zone and the lithologic contacts.
Extensional veins and vein arrays
Extensional and oblique-extension veins, singly o r in
(arrays, are relatively common in gold deposits, where they
(preferentially, but not exclusively, develop in com petent
rocks. In a large num ber o f deposits, they are auriferous
and can be shown to be of the same age as the more signif­
icant fault-fill veins (see below). However, extensional veins
have limited economic significance: they form orebodies
only in rare cases where their grades and abundances justify
bulk mining o r where their grades and dimensions justify
selective mining, such as at the Sigma-Lamaque deposit at
Val d ’O r (R obert and Brown, 1986). In o th e r deposits,
extensional veins overprint the ore and may be barren o r
mineralized (Hodgson, 1989). Extensional veins can occur
within brittle-ductile shear zones, where they form arrays of
sigmoidal o r planar veins (Fig. 16B, C) and where they
commonly fringe fault-fill veins (see below). They also com­
monly occur in low-strain rocks outside shear zones, where
they are planar and m ore extensive (Fig. 16A). Finally,
extensional veins also occur as arrays o f stacked planar
veins in extended competent layers (Figs. 15 and 16C) and
in association with folds (Hodgson, 1989). Oblique-exten­
sion veins are generally planar and occur as links between
extensional or fault-fill veins or as conjugate sets in compe­
tent layers (see below).
Geometric arrangements: Extensional veins can occur in
three main geometric setdngs and arrangements relative to
shear zones (Table 2; Hodgson, 1989). They can occur as
arrays of en echelon veins, planar or sigmoidal in shape
(Figs. 15B and 20A, B), which define brittle-ductile shear
zones in their own right, even in absence o f well-developed
foliation and lineadon (Beach, 1975; Rickard and Rixon,
1983; Ramsay and Huber, 1987). In such arrays, the indi­
vidual veins are generally small, less than 10 cm thick, and
lie at a high angle to the foliadon and lineation in the host
shear zone if present (Figs. 16B). E n echelon arrays o f sig­
moidal extensional veins can occur as the single vein type
within the shear zone (Fig. 20B). They can also fringe fault-
fill veins present in the shear zone (fig. 7i in Burrows e t al.,
1989; Moritz and Crocket, 1990; fig. 7 in Cox ct al., 1995),
leading to form auon o f so-called centipede veins (Hodg­
son, 1989). T he sigmoidal shapes and the orientation o f
vein tips with respect to array boundaries are indicative o f
the sense o f shear in this case (Fig. 16B; Beach, 1975; Ram­
say and Huber, 1983).
Extensional veins also form arrays o f stacked planar veins
in com petent layers such as dikes (Fig. 15), preexisting
veins, altered portions of shear zones, or specific lithologic
units (Hodgson, 1989). These veins, commonly termed lad­
d er veins (Hodgson, 1989), are typically short and occur at
high angles to the host layer and to the foliadon if present
(Figs. 15,16C, and 20C), reflecdng the predicted refraction
o f the strain axes within the com petent layer (Treagus,
1983). Such arrays o f stacked planar veins record extension
o f the layer.
Extensional veins may occur as isolated tabular veins
extending away from shear zones and fault-fill veins in less
deform ed rocks (Figs. 16A and 20D). These are not com­
monplace but are well developed in a few districts around
the world, for example at Val d ’O r (Robert, 1990) and Tim­
mins (Moritz and Crocket, 1990). These tabular veins range
in thickness from a few cendmeters up to a few meters, and
they extend for several lens to a few hundreds o f meters
away from shear zones and fault-fill veins (Fig. 20D). For
example, the no. 1 extensional vein of the Donalda deposit,
in the Noranda disuict in die Abidbi belt, covers an area in
excess of 0.5 km2 (Riverin et al., 1990).
 Fig. 20. Characteristics o f extensions! veins and vein arrays. A. Sigmoidal extension veins surrounded by bleached
alteration haloes. Note the presence o f tourmaline fibers (black) within the vein. Perron deposit, Val d Or, cross section
view. Pen for scale. B. Array o f sigmoidal extension w in s in granodioritc. The shape o f die veins and the orientation o f
the vein tips relative to the left-dipping envelope o f the array indicate a reverse component o f vertical movement. Perron
deposit, Val d ’Or, cross section view. C. Array o f small planar extcnsional veins (also termed ladder veins) developed
within a competent, tourmatinized shear zone (black) in mafic volcanic rocks. Val d'Or. Compass for scale. D. Small hor­
izontal extcnsional vein, showing regular planar wall, as well as open-space filling internal textures defined by small cal-
cite crystals (white) attached to both walls. Sigma deposit, Val d ’Or, crass section view. Hammer for scale. E. Subhori­
zontal extcnsional shear vein. Note how the opening vector (arrows) is indicated by matching bends along the vein walls.
Internal tourmaline fibers (black) also track a similar opening vector. Sigma deposit, Val d ’Or, cross section view. Ham­
mer for scale. F. Subhorlzontnl extcnsional vein with w elkleveloped subvcriical tourmaline fibers (black). The vein
shows evidence o f multiple opening and filling episodes, including a posttourmaline fiber reopening in the center o f
the vein. The opening vector indicated by matching irregularities along the vein walls (arrows) is parallel to that indi­
cated by tourmaline fibers. Sigma deposit, Val d'Or, cross section view. Hammer for scale. G. Subhorizomal oblique-
extension vein with well-developed subvertical tourmaline fibers (black). Sigma deposit. Val d ’Or, cross section view.
Hammer for scale. Courtesy o f the Geological Survey o f Canada, photo no. 1999-015V. H. Close-up view o f crack-seal
bands (arrow) highlighted by black tourmaline, near the hanging wall contact o f a subhorizomal extcnsional vein in gra-
nodioritc. Perron deposit, Val d ’Or, cross section view. Pen for scale.
 Extensional veins associated with folds can form perpen­ Open-space filling textures are characterized by euhcdral
dicular to nearby fold axes, occupying AC jo in t positions o r radiating aggregates o f crystals o f hydrothermal minerals
(i.e., joints perpendicular to fold axis; Hobbs et al., 1976), such as carbonate, pyrite, tourmaline, and scheelite, attached
or parallel to axial plane foliation in fold hinges, forming to the walls o f the veins (Fig. 20D) o r to individual layers.
sheeted veinlct zones similar to those associated with fault- The precipitation o f continuous m onomineralic layers in
fill veins (see above). In both cases, the veins form arrays open spaces produces crustiform banding, present in a small
having long axes parallel to fold axes. num ber o f gold deposits in greenstone belts (Hodgson,
A majority o f extensional veins in greenstone gold 1989). Mineral fibers are n o t uncom m on in extensional
deposits, be they isolated o r part o f en echelon arrays, have veins: they consist o f highly elongate minerals showing a con­
shallow dips (Fig. 20B, D), as illustrated by the Sigma- stant preferred orientation within a vein o r within a layer
Lamaquc gold deposit (Fig. 7). This orientation is consistent (Fig. 20F, G). They represent crystals that grew progressively
with their formation in a compressional tectonic regime as the fracture opened and that generally track the direction
(Figs. 10 and 12C; Sibson et al., 1988; McCuaig and Kerrich, o f opening o f the veins (Dum eyand Ramsay, 1973; Cox and
1998). Examples o f both shallowly dipping and subvertical Etheridge, 1983). Potential mechanisms o f fiber growth o f
(Fig. 20C) extensional veins are known in stacked planar this type are reviewed by Ramsay and H uber (1983). T he
arrays, reflecting their formation in both compressional and extensional veins illustrated in Figure 20E and F proride
in transcurrent (or extensional) regimes, respectively. good examples o f m ineral fibers that indicate an opening
Oblique-extension veins (Fig. 20E) typically do not form direction consistent with that indicated by m atching vein
vein arrays. In some cases, they occur as link structures walls. Crack-seal textures (Ramsay 1980b) can be observed at
between segments o f fault-fill veins; i.e., in dilational jogs the microscopic to mesoscopic scale in many extensional
(Fig. 1ID) or extensional veins (Nguyen e t al., 1998). In veins (Fig. 20H). Such textures record multiple episodes o f
other cases, they form conjugate sets in com petent layers, fracturing along the vein walls, incorporating thin bands o f
recording layer-parallel extension (Robert, 1990; see wall-rock inclusions. The rate o f opening o f an extensional
below). Sets of oblique-extension veins are also expected to fracture relative to the rate o f mineral precipitation is a key
form by dilation o f preexisting fractures along a direction factor in determining the internal structure o f individual lay­
not perpendicular to them (Fig. 11C). ers within extensional veins (Cox, 1991;Foxfordetal., 1991).
According to Foxford et al. (1991), crack-seal textures form
Internal features: Extensional and oblique-extension veins when the rate of mineral precipitation exceeds that of frac­
are relatively easy to identify because they have many diag­ ture separation, whereas mineral fibers form when it equals
nostic attributes, including internal structures and textures the rate of fracture separation, and open-space textures form
indicative of their opening vector. In general, extensional when it is exceeded by the rate o f fracture separation. All
(and oblique-extension) veins have parallel and planar walls, three types o f structures can form in a single layer, reflecting
at least at the mesoscopic scale (Fig. 20C, D), in contrast with variations in rates o f dilation versus rates of mineral precipi­
the irregular outlines of fault-fill veins (Figs.lGB and 18E). tation (Boullier and Robert, 1992).
Irregularities or deflections of the vein walls are relatively Rock bridges (Fig. 21; Nicholson and Pollard, 1985; Fox-
common and can be matched across the vein, giving a good ford e t al., 1991,2000) are relatively common features in
indication of the opening vector. This is well illustrated in extensional veins, but not in oblique-extension veins. They
Figure 20E and F, where matching walls of the vein indicate consist o f planar to sigmoidal slabs o f wall rock partially or
moderately plunging and subvertical opening vectors, completely enclosed within extensional veins, o r separating
respectively. In addition, the opening vector of extensional individual, overlapping, en echelon extensional veins. Rock
veins is also indicated by offset of geological markers such as bridges arc generally oriented at low to moderate angles to
lithological contacts or preexisting veinlets (Fig. 1 IB). the vein walls. Rock bridges and en echelon extensional
Common features o f extensional and oblique-extension veins are natural consequences o f the propagation and dila­
veins include internal compositional layering, massive tion o f extensional fractures and can be regarded as diag­
quartz, open-space filling textures, mineral fibers, crack- nostic of extensional veins (Nicholson and Pollard, 1985).
seal textures, and rock bridges (Table 2). Internal layering Propagation o f an extensional fracture induces rotation of
parallel to vein walls is a common characteristic o f exten- the local principal stress axes a t die fracture tips, resulting
siona! and oblique-extension veins (Hodgson, 1989). The in the development of smaller en echelon extensional veins
presence o f internal layering indicates that the veins result (Fig. 21A, B). As the en echelon extensional veins dilate,
from multiple episodes o f opening and mineral precipita­ the rock segments separating them , o r rock bridges, will
tion (Fig. 20F, G). Mesoscopic internal textures and struc­ break and becom e slabs o f wall rocks partially enclosed
tures commonly observed within individual layers in a vein within the veins (Fig. 21C).
include massive homogeneous filling (bull quartz o r buck
quartz), open-space filling and crustiform textures, mineral Associated ore shoots: Diverse types o f ore shoots are associ­
fibers, and crack-seal textures (Hodgson, 1989; Vearn- ated with extensional veins and vein arrays. Intersections of
combe, 1993). Individual layers comprising a vein may dif­ extensional and fault-fill veins arc commonly the sites o f
fer from one another in m ineral proportions and in tex­ elevated gold concentrations (e g., Robert and Brown,
tures and structures (Fig. 20F). 1986). Such sites define high-grade ore shoots within fault-

Fig. 21. Rock bridges in cxtcnsional veins. A. Block diagram showing
how rock bridges between en ech elon segments o f an cxtcnsional vein
become incorporated within it as a result of growth o f the segments (mod­
ified from Foxfordct ah, 1991), B, Overlapping subhorizontal cxtcnsional
veins separated by rock bridges. Sigma deposit, Val d ’Or, cross section
view. Hammer for scale. C. Large subhorizontal cxtcnsional vein contain­
ing broken rock bridges (arrow). Perron deposit, Val d ’Or, cross section
view. Hammer for scale.
fill veins, the plunges o f which are parallel to the lines o f
intersection between the two vein types. In the Val cTOr dis­
trict, extensive subhorizontal extensional veins form elon­
gate tabular bodies in plan view (Robert, 1990). Their long
axes are parallel to the fault-fill veins and shear zones with
which they are associated, defining an o th er type o f ore
shoot. Extensional veins associated with fold hinges, as sets
o f parallel veins in either AC or axial planar orientations,
tend to be restricted to specific folded layers and will form
bodies with long axes parallel to those of the folds.
Stockwork zones and breccia bodies
A num ber o f gold orebodies in greenstone belts consist
o f stockwork zones and breccia bodies (Groves e t al., 1995;
McCuaig and Kerrich, 1998). Stockworks are defined here
to consist o f two or more intersecting vein sets. They are
best developed in competent host rocks and they are highly
variable in their morphology and internal complexities.
Hydrothermal breccias and breccia veins are typically devel­
oped within shear zones (especially brittle ones; Hagemann
et al., 1992) or where they intersect competent host rocks,
in common spatial association with fault-fill veins.
Geometry and internalfeatures: Vein stockworks range from
being internally well organized, with two or three well-
defined vein orientations (Fig. 22A, B), to complex net­
works o f random ly oriented veins and veinlets (Fig. 22C;
Table 2). The M ount Charlotte deposit in Western Australia
is an example o f an organized stockwork (Ridley and Men-
gler, 2000), whereas the stockwork orebodies o f the San
Antonio deposit in the Canadian Superior province repre­
sent a more complex example (Poulsen et al., 1986a; Lau,
1988). Intense stockwork veining results in disaggregation
of the host rocks into angular blocks showing no evidence
of rotation (Fig. 23C), cem ented by quartz and o th er
hydrothermal minerals. Such intense stockwork develop­
m ent produces zones with the appearance of hydrothermal
breccias (Fig. 23C). Organized stockworks can consists of
nearly orthogonal sets o f extensional veins, as a t M ount
Charlotte (Ridley and Mengier, 20(H)), o r o f conjugate sets
of oblique-extension veins, as a t the Louvicourt Goldfield
deposit at Val d ’O r (Fig. 24; R obert, 1990). Conflicting
crosscutting relationships are commonly observed between
veins of the different sets indicating their contem porane­
ous development (Ridley and Mengier, 2000). The overall
morphology of stockwork zones is variable and ranges from
cigar-shaped (c.g., at Louvicourt Goldfield) to elongate tab­
ular bodies (e.g., at M ount Charlotte). The following exam­
ples illustrate the diversity in m orphology and internal
geometry of stockwork orebodies.
Cigar-shaped stockworks o f the A and B ore zones at the
Louvicourt Goldfield deposit at Val d ’O r are confined to a
subvertical, east-west sulking gabbroic sill, cut by a feldspar
porphyry dike (Fig. 23A). The stockwork bodies plunge 25°
to 30° W; they extend down-plunge as much as 200 m but
are only 15 to 20 m high and approxim ately 10 m wide.
They consist o f two conjugate sets o f oblique-extension
veins (Figs. 22A, 23C) intersecting in a line parallel to their
 Fig. 22. Characteristics or stockworks and breccia zones. A. Stockwork zone comprised o f two oblique sets o f shallowly
dipping extcnsional siiear veins developed in a gabbro sill; the photo covers the upper left half o f the cigar-shape ore-
body shown in Figure 23C. Louvicourt Goldfield deposit, Val d ’Or, cross section view. Field o f view —1 m. B. Stockwork
com posed o f (wo orthogonal sets ofextensional veins grading into a breccia texture in the upper left part o f the photo­
graph. San Antonio deposit, Rice Lake, section view. Field o f view - 3 m. C. Breccia zone in the central part o f a stock-
work, with clasts o f altered wall rocks in a quartz-carbonate matrix, grading outward in two somewhat irregular sets o f
orthogonal veins. San Antonio deposit, Rice Lake, cross section view. Hammer for scale. Courtesy o f the n<»ningi«-ai Sur­
vey o f Canada, photo no. 1999-016F. D. Extcnsional quartz veins defining a slab-like brcccia/stockwork body o f o n e dom­
inant set o f veins (right-dipping), parallel to die mylonitic foliation o f die host shear zone, and a less well-developed set
o f left-dipping veins (arrow), cutting across the foliation. Star Lake deposit. La Rouge, cross section view. Hammers for
scale. The inset shows a sketch from a polished section ofaquartz-pyritc vein cutting at a high angle to the foliation. E.
Jigsaw puzzle breccia in a feldspar porphyry dike, consisting o f bleached angular wall-rock fragments in a fine-grained,
black tourmaline matrix. Sigma deposit, Val d'Or, cross section view. Hammer for scale. Reprinted with permission o f
Economic Geology, Robert and Brown, 1986, fig. 7a. F. Well-developed breccia vein fringed by two oblique sets o f extcn­
sional veins. Victory deposit. Western Australia, cross section view. Field notebook for scale.

overall plunge. The two sets of veins commonly merge with
one another and show conflicting crosscutting relation­
ships (Robert, 1990). Mineral fibers and the matching of
markers a n d /o r irregularities in walls across veins in both
vein sets indicate that their opening vector was steeply
plunging to the east, perpendicular to the plunge of the
stockwork (Fig. 23C). T he localized conjugate sets of
oblique-extension veins may have form ed in incipient
necks of megaboudins resulting from subvertical elonga­
tion o f die host sill (Fig. 23D; Robert, 1990).
The stockwork zones o f the San Antonio deposit at Rice
Lake (Fig. 8) are somewhat m ore complex (Fig. 22B, C;
 L O U V IC O U R T G O L D F IE L D

STOCKWORK ZONE

FELDSPAR PORPHYRY
m

h - .- .i GABBRO

J J VOLCANIC ROCK

WEST B ZONE LONGITUDINAL SECTION EAST

% II
4 -

B
B ZONE CROSS-SECTION
NORTH SOUTH
OBUQUE
EXTENSION
VEIN

dZ-

• sill

D
Fig. 29. Geology o f the Louvicouri Goldfield deposit, Val d'Or. Adapted from Robert (1990). A. Plan view o f level 225
(from Sauvfi, 1985). B. Longitudinal section through the Zone B orcbody (from Snuve, 1985), C. Line drawing o f a cross
section view o f the Zone B orcbody, showing the development o f conjugate seta o f oblique-extension veins. The open­
ing vectors o f the two vein sets are indicated, as well as the incremental axes o f elongation and shortening. D. Schematic
section showing how the conjugate sets o f oblique-extension veins can accommodate subvertical extension and incipi­
ent boudinage o f the host com petent gabbro sill.

Poulscn et al., 1986a; Lau, 1988). They form steeply dip­
ping, shallowly plunging elongate tabular bodies arranged
in an en echelon fashion within a moderately dipping gab-
broic sill (Fig. 8B, C). They attain a thickness of 10 m, are up
to 150 m high, and extend up to 500 m along their axis of
plunge, which corresponds to the line of intersection of die
tabular stockwork bodies wilh the host sill. The core o f the
stockwork bodies consist of an inner central quartz vein par­
allel to the stockwork, commonly surrounded by a central
breccia zone which is composed of angular, altered wall-rock
fragments cem ented by vein quartz, albite, and ankerite
(Fig. 22C). The fringes of the stockworks consist of two sets
 Fig. 24. Features o f vein intersections. A. Subhorizontal cxtcnsional vein culling a steeply dipping fault-fiU vein, but
itself truncated by a slip surface close to the right margin o f the fault-fill vein. Sigma deposit, Val d'Or. cross section view.
Field o f view - 2 m. B. Subhorizontal cxtcnsional vein merging with a steeply dipping fault-fill vein. T h e exiensional w in
cuts the intense foliation in the shear zone hosting the fault-fill vein (left), but is itself only slightly buckled, recording a
small amount o f shortening o f the com petent vein. Sigma deposit, Val d ‘Or, cross section view. Hammer for scale. C. Sig­
moidal cxtcnsional veins associated with the down-clip termination o f a fault-fill vein (dipping moderately to the right).
Note how the cxtensional veins are developed only in the hanging wall o f the down-dip termination, and die localized
development o f foliation in the footwall. Lucicn Belivcau deposit, Val d'Or, cross section view. Hammer for scale. D. lin e
drawing o f the area shown in Figure 24C.

of intersecting steep and flat extensional veins (Fig. 22B).
The flat extensional veins commonly display sigmoidal
shapes and make an angle of 45° to die stockwork zone as a
whole. The origin o f the stockworks at the San Antonio
mine is enigmatic but the fact that they occupy planes mak­
ing a dihedral angle of 45° to the host sill suggests they may
have formed as extensional meshes by shear along the sill
margins. This possibility is corroborated by the existence of
foliated sericite-carbonate schists at those margins (Fig. 8C).
Hydrothermal breccias and breccia veins are thought to
result from repeated individual increments o f fragmenta­
tion, driven by a combination of structural and hydraulic
processes, and infilling by hydrothermal material (Sibson,
1986;Jebrak, 1997). Jigsaw puzzle and fault breccias are two
common types o f hydrothermal breccias or breccia veins
associated with, or as a com ponent of, fault-fill veins (Table
2). The reader is referred to Jebrak (1997) for a more
extensive discussion of breccias in vein-type deposits. Jigsaw
puzzle or mosaic breccias, also referred to as implosion
breccias (Sibson, 1986), consist o f angular wall-rock frag­
ments, showing no evidence of rotation or large scale trans­
lation, set in a m atrix o f hydrotherm al m inerals such as
quartz and tourm aline (Fig. 22E, F). T he proportion of
hydrothermal matrix may be relatively small (Fig. 22E) or
may exceed that of wall-rock fragments, in which case the
jigsaw puzzle nature o f the breccias may be partly obscured
(Fig. 22F). These breccias are interpreted to indicate dila­
tion at specific structural sites related to discrete slip events
along faults and to result from hydraulic fracturing related
to sudden fluid pressure drops following slip (Sibson,
1986). They have been docum ented to occur at low-angle
dilational bends along shear zones in some gold deposits
(e.g., Nguyen et al., 1998). The fringes of such breccia veins
record their incipient development and show fractures of
diverse orientations. They resemble the “crackle breccias"
better known from porphyry-type deposits (Jebrak, 1997).
The 21 Zone deposit in the Star Lake district (Fig. 9) pro­
vides an example of a hybrid breccia and fault-fill vein ore-
body. It is hosted by a subvertical mylonitic shear zone (Fig.
9B, C) and plunges southwestward, at a moderate angle to
the shear zone's northeast-plunging elongation lineation Different vein sets display systematic crosscutting relationships:
(Poulsen, 1986; see also Robert et al., 1994). The ore zone Systematic crosscutting o f one vein set by another simply
is up to 20 m wide, 100 m high, and extends at least 250 m indicates that the veins are of different ages and that a sys­
down plunge. It consists o f a mosaic quartz breccia body tematic vein chronology can t>e established between them.
containing slab-shaped blocks o f mylonitic host rocks, This can provide im portant information about the struc­
defined by extensional veins mostly parallel to but also at a tural evolution of a deposit. It is common, but n o t always
high angle to the foliation (Fig. 22D), and locally accompar the case, for'systematically crosscutting vein sets to display
nied by discrete through-going laminated quartz veins. Dis­ differences in vein o r alteration mineral assemblages. It is
cordant extensional vein segments clearly cut the mylonitic also expected that, for example along a given shear zone,
foliation (Fig. 22D), indicating that the breccia vein formed veins o f different ages will show different degrees o f over­
very late in the developm ent o f the shear zone. T he slab­ printing deformation (e.g., Robert and Brown, 1986).
like character o f the mosaic breccia may be attributed to This situation is illustrated by the San Antonio deposit
the fact that the preexisting mylonitic foliation creates an which, in addition to the stockwork zones described above,
inherent anisotropy during late dilation o f the rock. also comprises a set o f fault-fill veins in reverse-sinistral
In contrast to the above examples, fault breccias consist shear zones striking at a high angle to the host sill (Fig. 8C).
o f angular clasts o f single o r varied composition, showing Both the stockworks and fault-fill veins consist o f quartz-
evidence o f rotation, translation, and attrition, set in a ankerite-albitc-pyrite, fringed by sericue-carbonate alter­
matrix o f hydrothermal minerals (Fig. 18G). They are anal­ ation selvages. However, fault-fill veins systematically cross­
ogous to attrition breccias described by Sibson (1986), c u t and offset the stockwork zones where they intersect
except for the presence o f a hydrothermal matrix, indicat­ (Lau, 1988), indicating that the two types have developed
ing that hydrothermal precipitation accompanied slip and at different times in the structural evolution o f the deposit.
fault brecciation along the host structure.
Inasmuch as the opening o f a single planar extensional Different vein sets display conflicting crosscutting relationships:
vein can lead to a unidirectional increm ent o f dilation, Conflicting crosscutting relationships are relatively com­
stockwork and breccia veins, particularly those of irregular m on am ong different vein sets within deposits o r within
shape, appear to accommodate simultaneous multi-direc­ stockwork zones (e.g., Robert and Brown, 1986; Ridley and
tional extension (Fig. 22B, C, E; see also Ridley and Men- Mengler, 2000). They can take the form of an extensional
gler, 2000). vein cutting across a fault-fill vein at one location and the
opposite at another location. They can also take the form
Associated ore shoots: Owing to the diversity of their settings o f an extensional vein cutting across some quartz laminae
and geometries, several types o f ore shoots are defined by in a fault-fill vein, but itself being truncated by a slip surface
stockwork zones. In some cases, as at the Louvicourt Gold­ within the fault-fill vein (Fig. 24A). These types o f conflict­
field deposit, the long axes of the stockwork bodies corre­ ing crosscutting relationships am ong two types o f veins
spond to the line of intersection between their main con­ indicate that they are broadly contemporaneous, and fur­
stituent veins (Fig. 23), and may correspond to die axes of ther indicate their cyclic, sequential development. Because
boudins of the host com petent layer (Robert, 1990). In conflicting relationships are not necessarily exposed at the
cases where overall tabular stockwork bodies are confined same location within a deposit, it is important not to base
to a specific lithologic unit, as at the San Antonio deposit, interpretations of age relationships among veins on a single
the long axes of the stockwork bodies correspond to the o r small num ber of non-representative observations. Fur­
line of intersection of the stockworks with their enclosing thermore! crosscutting relationships among veins represent
host unit. In contrast, the stockworks at the M ount Char­ the final products of their incremental development, and
lotte deposit have pipe-like shapes and subvertical plunges, not necessarily their crosscutting relationships at different
defined by fault-bounded blocks of their com petent host stages in their incremental development (see below). Simi­
unit (Ridley and Mengler, 2000). larities of vein or alteration assemblages between crosscut­
ting vein sets suggest that they have formed from the same
Relationships among vein sets fluids and are potentially contemporaneous.
In most greenstone gold deposits, multiple types and sets
of veins are present. They are commonly spatially associated Veins of one set merge with veins of another set: In a number of
with one another, but are not necessarily temporally related. cases, veins of two distinct sets merge with one another (e.g.,
There are only a few specific cases of closely genetically Moritz and Crocket, 1990). In some deposits of die Val d ’Or
related vein sets, such as arrays of extensional veins near the district, for example, extensional veins merge with specific
term inations of fault-fill veins (see below). In structural lamina within a fault-fill vein (Fig. 24B; Robert, 1990). The
analysis, it is critical to distinguish veins formed during the extensional vein may merge with a lamina on the edge of
same strain increment from those formed during successive the fault-fill vein, or with a lamina well within the center of
increments. The temporal relations between different vein die vein, cutting across outer laminae (Fig. 18E). This type
sets are best established by careful examination of relation­ of merging relationship suggests contemporaneous devel­
ships at vein intersections (Fig. 24). A number of recurring opm ent of the two vein types, especially if they have the
situations in gold deposits are described below. same vein and alteration mineral assemblages.
 Kinematically related fracture/vein sets: T here are special
cases where the generation o f two types o f fractures, and of
veins by analogy, are intimately genetically related. Because
of stress concentration due to die attenuation of displace­
m ent at the tips of propagating shear fractures, extensional
fractures (also termed wing cracks) will develop at their lat­
eral and frontal term inations (Pollard and Segall, 1987;
Scholz, 1989; see also Cox et al., 2001). As shown in Figure
25 for the case of a reverse fault, extensional fractures will
develop in the footwall o f the up-dip term inadon of the
fault, and in the hanging wall o f its down-dip termination
(Fig. 24C, D). Arrays o f planar to sigmoidal extensional
veins can also develop at lateral fault term inadons. The
same geometric configuradon can be expected with fault-
fill and extensional veins, and examples are n o t uncom­
mon in gold deposits. The extensional veins shown in Fig­
ure 20B represent an array developed at the lateral
term inadon of a reverse fault. The concentrauons o f sub-
horizontal extensional veins at the up-dip term inadon of
reverse shear zones at the Lamaque Main mine (Fig. 7C)
are also interpreted as wing cracks at a fault terminadon.
The dynamics of vein developm ent at the dps o f faults
has implications for interpreting resulting crosscutting rela­
tionships between intersecting vein sets. Figure 24C and D
illustrates an example o f extensional veins distributed at
the down-dip term ination o f a reverse fault-fill vein, at a
fixed time in the developm ent o f the host fault. The sig­
moidal shape of the extensional veins is compatible with,
and indicative of, reverse movement along the host fault.
Down-dip propagation of this fault will result in the trunca­
tion and offset of existing extensional veins by the fault or
any fault-fill vein it contains. It can also be accompanied by
the formation o f new extensional veins farther down dip.
i For example, vein orebodies within shear zones o r within
0 metric relationships between different types o f veins and
FlC. 25. Diagram showing the expected arrangements o f extensional
fractures around frontal and lateral dps o f reverse faults o r shear fractures.
Adapted from Scholz (1989).
The end-result o f this dynamic process is that, at any loca­
tion along a fault o r fault-fill vein other than its termina­
tion, extensional veins will be cut by the fault or the fault-fill
vein, although in fact they develop synchronously. These
relationships are illustrated in Figure 24C: the largest
extensional vein is clearly truncated by the fault-fill vein,
whereas other smaller extensional veins are not, and occur
beyond the down-dip termination of the fault-fill vein. The
truncated extensional vein probably formed at an earlier
stage at which the fault-fill vein had not yet propagated to
its current position. Similarly, if the fault-fill vein were to
propagate farther downward, it would truncate and offset
all extensional veins present at the point o f observation.
This can easily result in incorrect interpretation of age rela­
tionships among the two vein sets. However, identical vein
and alteration m ineral assemblages of different vein sets
and a continuous alteration halo around both vein types (as
seen in Fig. 24C, D) may be used as a first indication that
the veins are broadly contemporaneous.
A n a ly sis o f v e in netw orks
Many gold deposits, especially larger ones, are composed
of vein networks, which combine several vein sets and types.
This might include multiple sets of fault-fill veins and their
host shear zones commonly as conjugate pairs, extensional
veins as en echelon arrays o r planar veins outside shear
zones, and stockwork veins (Robert, 1990; Nguyen e t al.,
1998). Contem poraneous vein sets can be used to deter­
mine the axes o f the deposit-scale incremental strain (Figs.
13,26). In the cases of vein sets of multiple ages, incremen­
tal strain axes at different stages in the structural evolution
o f the deposit can also be constrained.
Determining some o r all of the incremental bulk strain
axes o f vein networks provides a framework for predicting
the possible plunges o f orebodies within a given structure.
com petent layers commonly plunge parallel to the axis of
intermediate incremental strain, dY, or rarely perpendicu­
lar to it. As discussed below, the increm ental bulk strain
axes of vein networks can also be used to test the structural
timing o f the deposit by comparison with the finite strain in
the host rocks, as recorded by folds, boudins, shear zones,
and regional foliation and lincation.
Strain axes of gold deposit networks can be determined
in a way similar to the reconstruction o f the three principal
stress axes from faults and shear zones, and from exten-
sional veins and vein arrays (Beach, 1975; Angelier, 1979;
Rickard and Rixon, 1983; Ramsay and H uber 1987). Geo­
vein arrays and incremental strain axes are shown in Figure
19 for the case of plane suain (i.e., no change along the Y
axis). The intermediate incremental strain axis, dY, lies a t
90® to the slip direction in the plane o f the fault-fill vein o r
shear zone and parallels the long axes o f planar and sig­
moidal extensional veins in en echelon arrays (Fig. 19). In
the case of conjugate sets o f fault-fill veins o r oblique-exten­
sion veins, dY is parallel to the line o f intersection between
the two sets, and the axes o f incremental shortening, dZ,
 dX ping shallowly to the west and extending outside the shear
zones into less strained rocks (Robert and Brown, 1986;
Robert, 1990). All these vein sets are interpreted to be con­
temporaneous on the basis o f conflicting crosscutting and
merging relationships (e.g., Robert and Brown, 1986). The
key geometric and suuctural elements o f the vein network
are represented in Figure 26, along with the interpreted
axes o f incremental strain recorded by the development of
the vein network. Following the principles outlined above,
the incremental elongation direction, dX, is detennined to
be subvertical (actually plunging steeply east), the shorten­
ing direction, dZ, to be horizontal and north-south, and the
intermediate direction, dY (alongwhich there is no change
in length), to plunge shallowly to the west. As discussed
below, there is a good correspondence between the incre­
mental axes o f the vein network with those o f the regional
D.» increment o f deformation (Robert, 1990).
In the stockwork zones o f the Louvicourt Goldfield
deposit, hosted in a subvertical, east-west gabbro sill (Fig.
Fig. 26. Relationships between conjugate shear zones. Fault-fill and 23), the line o f intersection between conjugate sets of
extensions! veins, and incremental bulk, straiu'axcs, based on the geomet*
ric and structural features o f the vein network at th e Sigma-Lamaque
oblique-extension veins defines an intermediate incremen­
deposit (Fig. 7). Adapted from Robert (1990). tal strain axis, dY, plunging shallowly to the west (-25°).
The corresponding axis of incremental elongation plunges
steeply to the east (~65°), whereas the axis of incremental
and incremental elongation. dX, lie in the bisecting angles shortening is subhorizontal (north-south; Fig. 23C). The
of the shortening and extending sectors, respectively (Fig. developm ent o f the vein stockworks at Louvicourt Gold­
26). The axis of incremental elongation, dX, is perpendic­ field can be interpreted to accommodate localized hori­
ular to planar extensional veins external to shear zones, zontal shortening and subvertical elongation in necks of
which, therefore, should contain dY and the axis o f incre­ incipient boudins within the subvertically extending sill
mental shortening, dZ. As a first approxim ation, dX can (Fig. 23C; Robert, 1990).
also be regarded as being perpendicular to extensional In a num ber of deposits, the formation o f contem pora­
veins in en echelon arrays (Ramsay, 1982; Ramsay and neous vein sets cannot be accounted for by considering
Huber, 1987). Examples o f determination of deposit-scale strain in only two dimensions, and requires consideration of
incremental strain axes (or stress axes in some cases) can strain in three dimensions (Dube et al., 1989; Ridley and
be found in Dube et al. (1989), Robert (1990), Robert et al. Mcngler, 2000). This is particularly the case for deposits con­
(1994), and Nguyen et al. (1998). fined to single competent layers: because of strain refraction
The Sigma-Lamaque deposit in the Val d ’Or district (Fig. in three dimensions, the axes of bulk increm ental strain
2) seizes to illusu*ate how deposit-scale incremental strain within the layer may depart from those external to the layer,
axes can be detennined from a vein network and compared especially if the layer is oblique to one o f the external axes
with the main increments o f deform ation in the district. o f strain (Treagus, 1983). The Norbeau deposit in the
The deposit consists of an extensive network of quartz-tour- Abitibi (D ubectal., 1989) and the M ount Charlotte deposit
maline veins, extending over 3 km- on surface and to a in Western Australia (Ridley and Mengler, 2000) provide
depth of 1.8 km (Fig. 7). It is developed in andesitic vol­ examples of such departures from the two-dimensional case.
canic rocks intruded by coeval subvolcanic porphyritic dior- The methods o f analyses of such cases are beyond the scope
ite bodies, cut by a swarm of feldspar porphyry dikes and o f this paper, but they have been discussed in Dube et al.
younger diorite-tonalitc stocks. Volcanic contacts in the (1989) and Robert etal. (1994).
area are subvertical and strike east-west, parallel to axial
Vein Deformation and Superimposed Strain
planes of tight to isoclinal F2 folds and to variably devel­
oped regional So foliation, which contains a down-dip elon­ This section reviews the effects of superimposed defor­
gation lineation (Robert, 1990). Quartz-tourmaline veins mation on veins, with a brief consideration of the effects of
and the regional S., foliation overprint all these rock types. deform ation on oth er styles o f gold m ineralization.
The vein network is associated with conjugate reverse- Because of their form ation in active structures, veins in
o b liq u e) shear zones, intersecting along a line plunging faults and shear zones nearly universally show some evi­
shallowly to the west and containing elongation lineations dence o f superimposed strain (Boullier and Robert, 1992;
raking steeply east. T he network consists mainly o f lami­ Mason and Helmstaedt, 1992; McCuaigand Kerrich, 1998).
nated fault-fill veins within die conjugate shear zones, con­ In structural analysis o f veins, it is critical to determ ine
taining slickenlines raking steeply east (parallel to shear w hether deform ation features of veins result from their
zone lineation), and o f subhorizontal extensional veins dip­ increm ental developm ent in active structures (in which
case they may provide im portant information on the con­
ditions of vein formation; McCuaig and Kerrich, 1998), or
from superimposed deformation. The distinction between
die two cases is commonly challenging, and is at the root of
many divergences o f interpretations o f the structural tim­
ing of greenstone gold deposits.
There is little literature specifically discussing the effects
of superimposed strain on vein-type gold deposits in green­
stone belts other than a few papers discussing deformation
features related to vein development in active shear zones
(McCuaig and Kerrich, 1998). The material presented here
is an application of the basic principles of structural geol­
ogy to the analysis of deform ed auriferous veins, drawing
for example on die work of Talbot (1970,1982), and Ram­
say and H uber (1983, 1987) on veins and dikes. It also
draws heavily on well-established principles o f structural
analysis o f deformed rocks in general (Ramsay, 1967,1982;
Talbot, 1970,1982; Ramsay and Huber, 1983,1987) as well
as on the experience of the authors. Geologic structures
and relationships observable at the scale o f the hand-speci­
mens, the outcrop, or the stope are emphasized.
Theoretical considerations
Veins and dikes are well known markers o f superimposed
strain (Talbot, 1970,1982). Their behavior during overprint­
ing deformation depends mainly on two factors: their com­
petence with respect to their hosts (Berger, 1971; Ramsay,
1982; Talbot, 1982), and their orientauons with respect to
external axes o f shortening and extension (Talbot, 1970,
1982). Where there is no competence contrast between veins
or dikes and their hosts, they will deform homogeneously
with tine enclosing rocks. Where more competent than their
host rocks, they will form buckles or boudins depending on
their orientation with respect to strain axes (Figs. 14 and
27A; see also fig. 2.14 in Ramsay and Huber, 1983). Where
' less competent, strong internal foliation and outward point-
ing cusps will develop in veins and dikes (Talbot and Souk-
outis, 1992), with the cusps being parallel with or perpendic­
ular to the foliadon. Veins o r dikes close to the direcdon of
zero finite elongadon in the rock (i.e., near the angle divid­
ing those buckled from those boudinaged; Fig. 14) may show
different geometries. Com petent layers may appear to be
: undeformed or may display boudinaged buckles or buckled
boudins (Ramsay, 1967). Inasmuch as the direcdon o f zero
; finite elongadon for most strains is close to that of maximum
v ,shear strain (Ramsay, 1967), it is likely that incom petent
veins or dikes in this orientation would be most susceptible
to lateral slip parallel to their walls. Characterized by oblique
internal foliadon, such incom petent veins and dikes can
become incipient shear zones (Berger, 1971).
Quartz veins tend to be more competent than their hosts
over a range o f metamorphic conditions (Talbot, 1970). In
a greenstone belt environm ent, quartz veins can be
, expected to be m ore com petent than greenschist-grade
•intrusive and volcanic rocks o f intermediate to ultramafic
composition, as well as fine-grained clastic sedim entary
rocks. As a result, quartz veins are generally expected to
deform into buckles and boudins (Fig. 27A). However, it is
possible that some host rocks, such as feldspar-rich por­
phyry dikes (except where sericidzed) which are commonly
associated with gold deposits, may be as competent o r more
com petent than the veins. Because o f their generally
restricted dimensions, either as isolated veins o r as part of
vein arrays, veins in equally com petent rocks are expected
to deform homogeneously with their host.
O f added relevance to gold deposits, however, is the fact
that selvages of hydrothermally altered wall rock commonly
intervene on both sides between veins and unaltered hosts
(Fig. 28A, B; McCuaig and Kerrich, 1998). Such alteration
may therefore be m odeled qualitatively as a multilayer
(host rock-altered rock-vein-altered rock-host rock) lead­
ing to a different style of deform ation (Fig. 27B; Ramsay
and Huber, 1987). The assumption here is that alteration
renders the rock directly adjacent to the vein incompetent
relative to both the unaltered host rock and the vein. This
scenario is reasonable in the case of a com petent host
lithology with alteration dom inated by m inerals such as
sericite, calcite, and chlorite (Fig. 28A). Dolomite, however,
is a comm on alteration m ineral associated with gold
deposits in greenstone belts and can lead to a relatively
com petent alteration selvage if present (with quartz) in
substantial quantities.
The main additional effects of incompetent alteration can
be expected to be the development of prominent mullions
on the in n er arcs of buckled veins, and o f bidirectional
cusps pointing outward from the alteration into both unal­
tered wall rock and the necks o f boudinaged veins (Fig. 27B;
Talbot and Soukoutis, 1992). Depending on the orientation
of the veins with respect to the increm ental shortening
direction, the presence of an incompetent alteration selvage
might also lead to significant refraction of external foliation
within the selvage, in a way similar to cleavage refraction in
graded greywacke-mudstone beds. In cases where veins are
suitably oriented with respect to the incremental shortening
direction, shear zones can develop in the alteration selvages
parallel to the vein, irrespective o f the original vein type. In
this case, the resulting geometric relationships between the
foliated alteration selvages and the vein might be very diffi­
cult to distinguish from those associated with a fault-fill vein
formed in an active shear zone.
Because shear zones in greenstone belts are commonly
sites o f hydrotherm al alteration (McCuaig and Kerrich,
1998), the foliated rocks within them may also behave
incompetently during subsequent deformation increments.
It is therefore reasonable to expect their reactivation, with
contem poraneous folding and boudinage o f veins con­
tained within the zone o f reactivation. W here preexisting
shear zones are oriented at a low angle to external axes of
shortening, the shear zones and the veins they contain
might also be expected to undergo wholesale folding (Fig.
27B). In this case, one m ight expect primary shear zone
fabrics (foliation, lineation, and shear bands) to be com­
pletely overprinted by the later generation of regional
cleavage and m inor folds.
Finally, it must be stressed that the hypothetical case illus­
trated here (Fig. 27) relates to an external bulk co-axial
140 ROBERT AND POULSEN

Extension
vein

h
i

Fig. 27. A. Isometric block diagram illustrating the expected dcformational behavior o f quartz veins more competent
than their hosts overprinted by significant coaxial strain (adapted from Talbot. 1082). B. Isometric block diagram illus­
trating the expected behavior o f a shear zone and extensions*! vein m esh, overprinted by a significant coaxial finite
strain. Note the folded shear zones and die development o f mntlions and cusps in the alteration haloes, modeled here
as less com petent than the veins and unaltered host rocks. See text for discussion.

strain, wherein finite strain axes X, Y, and Z have the same
directions as the incremental strains dX, dY, dZ. This is by no
means the only situation one m ight encounter in nature
where non-coaxial strains (i.e., where the finite strain axes lie
at some angle to the incremental strain axes) may predomi­
nate, particularly in shear zones (see Hanmer and Passchier,
1991). The same basic principles apply however, and similar
geomeuic relationships among veins, alteration selvages, and
foliation will be developed. An added complication is that
there will be a greater likelihood of encountering veins that
were first folded and then boudinaged, as they rotated from
the shortening field to the elongation field during progres­
sive shear zone development (Ramsay, 1980a).
In contrast to quartz veins, sulfide-rich ore, cither as con­
cordant lenses or discordant veins, is typically less compe­
tent than its enclosing lithologies. Consequently, sulfide
veins and layers will deform passively with respect to their
host rocks. They will fold with their host rocks and flow into
boudin neck areas and fold hinges, producing a variety of
cusps and piercemcnt structures (e.g., Maiden et al., 1986).
Deformation features of xleins and their aiialysis sional vein is compatible with the directions of elongation
and shortening within the host shear zone, suggesting it
Distinguishing features related to vein deformation
formed during progressive shear zone deformation. The
from those related to vein growth is a critical task in ana­
lyzing the structural significance of gold deposits. The fact that the quartz-tourmaline vein cuts the foliation and
foliation-parallel veinlets further indicates the vein formed
most common manifestations of vein deformation are
their external forms (folds and boudins), vein margin stri- relatively late during shear zone development and has only
ations, and internal deformational fabrics such as stylo- experienced a m inor amount of subsequent shortening.
lites. In the case o f an asymmetrically folded vein in a shear
zone (Fig. 28D), the fact that the shear zone foliation is
axial planar to the vein fold (and not folded with the vein;
Folding: It is common to encounter veins taking the see below) indicates that the vein has experienced a signif­
form of folds, particularly under greenschist to lower am­ icant amount of shear zone deformation. The fact that the
phibolite facies conditions. Within shear zones or highly fold axis is perpendicular to the elongation lineation within
strained rocks, two general cases o f folded veins need to the shear zone does not provide additional information on
be considered: one where the foliation of the host rock is the timing of vein formation relative to shear zone devel­
axial planar to the folds, and the other where both the opm ent. The same geom etric relationships would be
vein and foliation of the host rocks have been folded to­ expected by folding of a vein formed at a low angle to the
gether. shear plane early during shear zone development, and a
vein of similar orientation but entirely predating shear
Foliation axial planar to veinfolds: Implicit in this case is the zone development. In this example, the laminated nature
greater competency of the vein with respect to its foliated of the vein is typical of fault-fill veins, which may be used to
host and the possibility that the vein predates foliation suggest (but not prove) that the vein was related to shear
entirely. Folding ranges from incipient buckling (Fig. 28A, zone development rather than entirely predating it.
B) to tight folding (Fig. 28C, D) of the vein. Tight folding is Some gold orebodies consist of stockworks of variably ori­
commonly accompanied by boudinage of the vein along the ented veinlets. Where overprinted by deformation, veinlets
limbs (see below). In profile (section perpendicular to the at a high angle to foliation will be buckled (Fig. 28A, E, F),
fold axes), vein folds can be symmetric (Fig. 28C) or asym­ whereas those lying at a low angle to foliation will be boud-
metric (Fig. 28D). The distinction depends in part on where inaged (Fig. 28A). In both cases, the veinlets have clearly
the minor fold is located with respect to larger ones (limbs been overprinted by some o f the shortening across the foli­
or hinges); however, where all folds along a vein are of the ation plane, but it is n o t possible to determ ine if they have
same asymmetry, die enveloping surfaces of the folds reflect formed early during, o r entirely before, shear zone defor­
the overall orientation of the vein in the deform ed state. mation. In the case of the veinlets in Figure 28E and F, evi­
The axes o f folded veins further constrain the three-dimen­ dence indicating that the veinlets predate the development
sional orientation of the deformed vein as a whole, because o f the foliation comes from the existence o f the same,
they coincide with the line of intersection of the vein in its undeformed, stockwork veinlets extending outside the foli­
deformed state with the foliation plane. Finally, it is gener­ ated zone (see Robert, 1997).
ally possible to determine whether the folded vein was orig­
inally an extensional o r a fault-fill vein, based on its m or­ Folded veins and foliation: T here are a num ber o f cases
phological and textural characteristics as described above. where subparalle! foliation and veins are folded together
In shear zones, the internal features o f veins, the geome­ (Fig. 28G, H), suggesting that both features predate a sig­
try and orientation o f the vein folds, and the overall enve- nificant increm ent o f deform ation. In shear zones, two
lope of the veins can be used to assess their compatibility or main situations can lead to this wholesale folding o f a shear
' incompatibility with the strain axes or slip direction within zone: a subsequent increm ent o f regional deform ation,
the host structure. Compatibility o f original vein orienta- and progressive internal folding (Platt, 1993) either during
tions with strain axes might suggest that the vein formed continued movement o r during later reactivation.
during, b u t was outlasted by, shear zone deform ation. Map-scale folding o f both shear zone and veins during a
Incompatibility could be used to argue that the vein pre­ subsequent increment o f deformation is best documented
dated die development o f the host shear zone. The follow­ through systematic mapping. This mechanism can also be
ing examples serve to illustrate different situations. recognized at outcrop scale by the fact that a foliated zone
In the sample shown in Figure 28B, the planar walls of enveloping the veins is itself folded. The foliation is folded
the quartz-tourmaline vein and its internal fibers indicate together with the zone but, in detail, may locally be oblique
that this is an extensional vein. The overall angle between to both the margins o f the veins and o f the zone.
the vein and the foliation, as well as the retained perpen­ Asymmetric folds o f both veins and subparallcl foliation
dicularity of the foliation to vein walls, suggests that the vein are particularly common in intense shear zones (Fig. 28G).
formed at a high angle to the foliation. The trace o f the Such folds imply that, prior to folding, the veins and folia­
vein along the foliation plane and the fold axis are also per­ tion were parallel o r at low angles to each other, which can
pendicular to the elongation lineation in the foliation be the case for fault-fill veins, extensional veins formed by
plane Thus, the inferred original orientation of this exten­ dilation of foliation planes (e.g., Kerrich, 1989), or highly
 Fig. 28. Features o f folded veins. A. Small buckled quartz vein (shallow-dipping on photo) at a high angle to the foli­
ation in the host rock, with well-developed chloride alteration selvages (dark) forming small-scale muliions. Note the
incipient boudinage o f the vein at a low angle to the foliation on the left side o f photo. Copper Rand deposit, Chi-
bougnmau, cross section view. Steel plate is -1 5 cm across. B. Small buckled quam-tourmalinc vein with buff-colored
alteration selvages in a sample cut perpendicular to the foliadon but parallel to the elongation lincation o f die host shear
zone. The folded quartz-tourmaline vein cuts smaller foliation-parallel quartz-carbonate vcinlcts (3 and 4) and has an
overall orientation at a moderate to high angle to the foliation. Along the lower right limb o f the vein, the foliation itself
is deflected and retains an orthogonal orientation relative to the vein walls. Within the vein, there are tourmaline fibers
(arrows), now oblique to the vein, linking a pair o f foliation-parallel vcinlcts (1 and 2). Orenada Zone 4 deposit, Val
d'Or. C. Symmetrically folded quartz-carbonate vein in a shear zone. These tightly folded veins represent an earlier gen­
eration than the more abundant quartz-tourmaline veins. Sigma deposit, Val d ’Or, cross section view. Hammer for scale.
D. cross section view o f a asymmetrically folded vein in reverse shear in a view approximately perpendicular to the folia­
don and parallel to the elongation lincation o f the host reverse shear zone. The internal laminated character o f the vein
indicates that it is a fault-fill vein. The vein fold is strongly asymmetric and its vergence is compatible with reverse move­
m ent along the host shear zone. The envelope o f the vein is at a low angle to the foliadon, which is clearly axial planar
to the fold o f the vein; the fold axis is at a high angle to the clongadon lincation within the host shear zone. Con deposit,
Yellowknife, cross section view. Hammer head for scale. Courtesy o f the Geological Survey o f Canada, photo no. 1996-
015K. E. Folded siockwork vcinlcts (white) and associated aibitic alteration selvages (light brown) in moderately foliated
basalt. Holt-McDermott deposit, Harker-Holloway, cross section view. Field o f view -1 .5 m. F. Detail o f a folded vcinlct
and alteration halo from the deformed stockwork shown in Figure 28E. Note that the foliation is axial planar to the vein-
let folds. Holt-McDermou deposit, Harker-Holloway. G. Asymmetrically folded quartz vein (1) and mylonitic foliadon of
the host granitic intrusion (2). Jasper deposit, Star Lake district, map view. Pen for scale. H. Asymmetrically folded gray
quartz vcinlcts and foliation, defining a small intrnfolial fold domain within the Cadillac fault zone. Note the small vein
boudins formed prior to die folds (arrows). Orenada property, Val d'Or, map view. Compass for scale.

rotated extensional veins initially formed at high angles to ation. Symmetric boudins indicate that, prior to its extension
foliation. In all cases, the joint folding of veins and foliation (or boudinage), the vein was oriented subparallel to the foli­
might result either from continued shear zone movement ation plane and to the incremental elongation direction, dX.
(Platt, 1983) or from shear zone reactivation in a different Asymmetric boudins indicate that the vein was initially
sense or direction of movement. The distinction between oblique to the shear zone foliation and rotated toward the
the two situations may be difficult to establish in practice foliation plane and dX. The asymmetry o f the boudins can
but locally can be achieved by exam ining the geometric generally be used to infer the sense o f rotation o f the vein
relationships of the folds to the strain axes in the shear during its deform ation (H anm er and Passchier, 1991).
zone, as defined by its penetrative foliation and lincation. Boudins o f veins typically are elongated objects, with their
In the examples illustrated in Figure 28G and H, the axes long axes perpendicular to the elongation lineation, o r to
of the asymmetric folds of both the vein and foliation are the dX axis, in the related foliation plane.
subvertical, subparallel to the elongation lineation in the Boudinage of veins can also develop along the limbs o f
host shear zone. The close to tight character of the folds (as tight to isoclinal folds (as described above) in shear zones
opposed to isoclinal) suggests that shear strains were not and highly strained rocks. Fault-fill veins, because they gen­
sufficiently high to lead to significant rotation o f the fold erally form a t a low angle or parallel to their host shear
axes towards parallelism with the m ovem ent direction zone, will generally be boudinaged as a result of continuing
; along the shear zone. Based on the near parallelism o f fold movem ent and shortening across the shear zone, o r as a
axes with elongation lineation in the host shear zone, the result of reactivation. The orientation o f the long axes o f
;folds are best interpreted as resulting from reactivation of the boudins helps to discriminate between the two cases. If
•.the host shear zone by transcurrent movements. The asym­ produced during continued shear zone deformation, the
metric folds in the example shown in Figure 28H from the long axes o f the boudins will lie at a high angle to the elon­
Cadillac Tectonic Zone in the Val d ’O r district, have been gation lineation in the host foliation and typically parallel
interpreted by Robert (1990) as resulting from dextral tran­ to the axes of related folds (e.g., Robert and Brown, 1986).
scun ent reactivation o f the shear zone. If produced during later shear zone reactivation, the
boudins will have a different orientation relative to the
Boudinage and related features: Inasmuch as quartz veins elongation lineation in the related foliation. In the exam­
tend to be more competent than their surroundings, it is not ple shown in Figure 29C, the axes o f the boudins are sub­
(.uncommon to observe in them the effects o f vein-parallel vertical, subparallel to the elongation lineation in the folia­
extension. These effects may take the form of macroscopic tion plane, suggesting that boudinage took place during a
pinch and swell structures with cusps developed in boudin subsequent deformation increment.
/ necks (Fig. 29A) or of the development of new extensional Boudinage clearly indicates that a vein has been over­
veins perpendicular to an existing vein in order to accom- printed by strain. However, even in the case o f boudinage
i modate its extension (Fig. 29B). Extreme cases o f boudinage related to continued shear zone deformation, it is not pos­
, in shear zones also lead to development o f isolated quartz sible to determ ine w hether the vein predated the host
iipods along the foliation (Fig. 29C). Boudins can be symmet­ shear zone or if it formed during its development. Choco­
ric (Fig. 29A) or asymmetric (Fig. 29C, D) relative to the foli­ late-tablet boudinage o f a vein indicates its extension in two
ROBERT AND POUINEN
 Fig. 29. Features of boudinaged veins. A. Pinch and swell structure in boudinaged ankcritc veins (light gray)- Note
the development of cusps of less com petent foliated basalt (dark gray) in the necks of the vein boudins. Note also the
presence of small extensions!! quartz vcinlcis (white) at a high angle to the ankcritc veins. Dome deposit, Timmins, map
view. Steel plate is-15 cm across. Courtesy o f Geological Survey of Canada, photo no. 1999-0150. B. Boudinaged quartz-
tourmaline vein (outlined), folded together with its host foliation. Note the development of white extensional veinlets
perpendicular to the main vein in the necks of incipient boudins. The fact that the veinlets in boudin necks retain their
high angle to the main vein across the fold indicates that the vein was first boudinaged and then folded. Orenada Zone
4 deposit, Val d’Or district. Vein is 15 cm thick. C. Intense boudinage of quartz veins parallel to foliation, leading to sep­
aration of (lie boudins into individual isolated quartz pods in the shear zone. Note the asymmetry of the boudins, indi­
cating rotation of the vein during its boudinage. Orenada Property, Cadillac fault zone, Val d’Or. map view. Hammer for
scale. D. Development of asymmetric boudins in a quartz vein lying at a low angle to the foliation. The boudins arc -5 cm
thick. James Bay. E. Piercemcnt structures (arrows) of sulfides (light gray) in more competent andalusitc schist (dark
gray) developed along the walks of a massive sulfide band. La Rondc deposit, Bousquet district, map view. Steel plates arc
-15 cm across. F. Boudins of andatusilc schist (dark gray) within layers of massive sulfides (light gray), showing flow of
the sulfides in the boudin necks. La Rondc deposit, Bousquet, map view. Steel plates arc >15 cm across. G. Mylonilic
quartz in fault-fill vein (highlighted). Jasper deposit. Star Lake district, map view. Pen for scale. H. Disseminated sulfide
and suifidc-iich vcinlct ore in scricitc schist. Note the elongation of the pyritc grains parallel to the foliation (arrow).
Doyon deposit, Bousquet.

orthogonal directions and reflects bulk flattening o f the
shear zone. Again, it provides no information on the timing
of vein formation, other than to indicate the vein has been
overprinted by at least some of the shear zone deformation.
Because of their commonly incompetent character, sul-
fide-rich ores, either in the form of massive sulfide lenses or
veins, will display different deform ation features than
quartz veins. Shortening across sulfide-rich layers initially
produces piercement structures of the sulfides into incipi­
ent boudin necks in the adjacent competent material (Fig.
29E; Maiden et al., 1986). In more advanced states o f boud­
inage of the com petent layers, sulfide material flows into
the boudin neck areas (Fig. 29F). As in the case of boudi­
naged veins, sulfide cusps and piercement structures indi­
cate that the sulfides have experienced at least some of the
shortening across the host shear zone, but they provide no
information on the relative timing of sulfide introduction
and shear zone development.
Striated vein margins: Striations (slickenlines) on the walls
of fault-fill veins or of individual laminae within them are
relatively common (Fig. 19F), and more than one set may
be present in the same vein. Striations record the direction
of slip event(s) during vein development or during subse­
quent reactivation. Striations may define steps recording
the sense o f movement along the slip plane, and some may
be accompanied by fibrous mineral growth. As discussed in
the previous section, only in rare cases can striations be
unequivocally related to vein development: those in which
mineral fibers along the slip surface consist o f a specific
mineral (e.g., tourmaline) diagnostic o f the main stage fill­
ing of the vein. Striations overprinting a given mineral must
be clearly distinguished from those defined by fibers of the
same mineral.
' Striations subparallel to the movement direction along
tile host shear zone probably relate to the main period of
movement along the shear zone as deduced from other fea­
tures, especially if they contain steps compatible with the
sense of shear (Fig. 19F). In the absence o f diagnostic min­
eral fibers, it is not possible to determ ine if the striations
formed as a result of vein development or as a result o f their
overprinting during subsequent shear zone movement. In
the presence of multiple sets o f striations, those parallel to
the m ovem ent direction along the shear zone are most
likely to be related to its development. Striations produced
during reactivation o f a shear zone o r a vein need not be
parallel to the earlier movement direction along it.
Internal deformational features in veins: A variety o f o th er
deformation features, some o f which are only risible a t the
microscopic scale, are also comm only developed within
veins. These have been recently reviewed by McCuaig and
Kerrich (1998) and are only briefly considered here. Again,
such internal deformation features may be an integral part
of the progressive developm ent o f fault-fill veins, o r may
record subsequent overprinting deform ation. In m ost
cases, it is very difficult to discriminate between the two pos­
sibilities on the basis of these Internal structures alone.
Polygonization and recrystallization o f quartz, observed
at the microscopic scale, is very common in veins, suggest­
ing overprinting deformation a n d /o r metamorphism. This
is especially clear in the case o f extensional veins, in which
deform ation features do no t accompany vein growth.
Deformation o f vein quartz is also an integral part o f the
development of fault-fill veins in active shear zones (Bout-
lier and Robert, 1992; McCuaig and Kerrich, 1998). In this
case, however, polygonization and recrystallization are not
uniformly developed within the vein, varying between lam­
inae depending on their relative time o f growth during the
history of vein development (Boullier and Robert, 1992).
Mylonitic foliation o f quartz is present in a few cases and
represents extreme and pervasive recrystallization o f veins
in shear zones. It is defined in outcrops or hand specimens
by zones o f strong quartz banding and significant grain-size
reduction (Fig. 29G). It is also accompanied by destruction
of the prim ary vein textures. Such developm ent o f
mylonitic foliation is not expected to be part of vein forma­
tion in the case o f fault-fill veins and is, therefore, likely to
have developed during overprinting deformation or shear
zone reactivation.
 Stylolites marking trains of insoluble residues also pro­ suain increments in the districts (e.g., Table 1). The second
vide common evidence of internal deform ation o f veins. approach is based on a determination o f the timing of veins
They are most commonly developed in fault-fill veins and relative to specific fabrics and structures in penetratively
may form either during vein developm ent (Boullier and strained host rocks.
Robert, 1992) or during subsequent deformation.
Comparison of vein networks and regional strain axes
Pressure shadows around pyrite grains in altered wall rocks
adjacent to veins and in disseminated gold ores are also com­ In a num ber of districts, the axes of strain related to spe­
mon. They record deformation overprinting mineralization cific deformation increments can be determined from their
but provide no additional constraints on their relative tim­ corresponding penetrative structural fabrics (foliation and
ings. Pressure shadows may have a prolate shape that defines elongation lineations) and folds. The incremental strain axes
a lineation with die same significance as the elongation lin- of vein networks can then be compared with those of differ­
eation in the host foliation. In other cases, disseminated sul­ ent deformation increments. As a first approximation, com­
fide grains are parallel to the foliation (Fig. 29H), also patibility of network axes with those of a specific deformation
recording deformation overprinting the mineralization. increment suggests a temporal link between the two.
For example, in the eastern part of the Val d ’O r district,
Ore shoots resultingfrom vein deformation: The plunge of where not overprinted by Ds structures (Fig. 5), D2 is charac­
long axes of deformed veins and orebodies will reflect both terized an east-west subvertical S2 foliation, containing a sub­
their inidal morphology and its modification by deforma­ vertical L, elongation lineation (Robert, 1990). These struc­
tion in their host structures. Two main cases need to be tural fabrics record D2 strain in which the maximum finite
considered: deposits overprinted by shear zones o r by sig­ shortening direction, Z, is north-south and horizontal, and
nificant bulk strain, and those overprinted by folds. the maximum elongation direction, X, is subvertieal. In con­
Deposits and orebodies overprinted by shear zones will trast, D3 structures are marked by northeast to east-west sub­
tend to plunge parallel to the elongation lineation in the vertical crenulauon cleavages, steeply-plungi ng asymmetric
foliated host rocks. This is expected to be the case for pre­ folds, and subhorizontal striations on D2 shear zones and
shear zone deposits, as illustrated by the pre-deformation veins. They record a late increment of dextral transcurrent
sulfidic gold deposits of the Bousquet district, which plunge deform ation (Robert, 1990), in which Z is horizontal and
steeply to the west, parallel to the elongation lineation in trends broadly southeast, whereas X is subhorizontal and
their intense host shear zone (Marquis et al., 1990). If the trends broadly northeast. As discussed above, the develop­
deposits or orebodies have a well-defined primary plunge, m ent o f the quartz-tourmaline vein network at Sigma-
it will be variably modified by bulk rotation towards paral­ Lamaque clearly records incremental north-south horizontal
lelism with the elongation lineation or the movement direc­ shortening, dZ, and subvertical elongation, dX. (Fig. 26).
tion of the host shear zone, depending on the am ount of The cigar-shaped stockwork zones o f the Louvicourt Gold­
overprinting bulk su'ain. field deposit also define dX, dY, and dZ incremental strain
Deposits and orebodies overprinted by folds will tend to axes plunging steeply east, shallowly west, and horizontally
plunge parallel to the fold axis. Such oreshoots can be north, respectively (Fig. 23C). The axes o f incremental strain
defined by stope-scale concentrations of gold and sulfides recorded by quartz-tourmaline vein networks at both
remobilized in fold hinges, or by the plunge o f the entire deposits are compatible with those o f D2, suggesting that
orebody along the fold axis. The Orenada Zone 4 deposit they formed during this regional increment o f deformation.
at Val d ’Or provides such an example (Robert, 1990). The In contrast, the formation of quartz-tourmaline veins in both
O renada Zone 4 deposit is a small vein-type gold deposit deposits is inconsistent with the D3 transcurrent increment
located directly within the Cadillac fault zone (Fig. 2). It is of deformation, in which the elongation direction, X, is sub­
characterized by a cluster o f thin quartz-tourmaline veins horizontal. A D9 timing is also indicated by the overprinting
parallel to the intense S2 foliation within die fault zone, and o f quartz-tourmaline veins by F3 folds at the Orenada Zone 4
overprinted by mesoscopic F3 folds. The vein cluster occu­ deposit (see above), located within the first-order Larder
pies the core of a map-scale Fft fold and plunges moderately Lake Cadillac fault zone (Fig. 2). The local horizontal striar
to the west, parallel to the plunge of the fold. tions observed on fault-fill vein surfaces at Sigma-Lamaque
can also be explained by a weak D3 overprint (Robert, 1990).
Structural Timing of Veins The weak D3 overprinting of quartz-tourmaline veins north
Auriferous veins in greenstone belts occur in shear of the Larder Lake-Cadillac fault zone reflects the fact that
zones, folds, extended com petent layers, and in zones of this increment of deformation was largely localized along the
broader regional strain. Determining the timing o f vein for­ first-order structure in the district.
mation relative to the history of their host structures or host Timing constraints can also be placed even in the case of
districts is critical in determining key structural controls on sim pler vein networks with less well-defined incremental
mineralization and in selecting appropriate gold deposit strain axes. In the Star Lake district, quartz vein deposits
models. Two complementary approaches can be followed occur in a set of east-northeast- to northeast-trending, sub­
to constrain the structural timing of veins. The first is based vertical shear zones, overprinting a composite granitoid
on a comparison of the axes o f incremental strain recorded intrusion and parallel to the regional D2 structural trend
by the veins o r vein networks with those of well-established (Figs. 4 and 9A). All host shear zones are defined by a
 strong mylonitic foliation and contain a m oderately to
steeply northeast-plunging elongation lineadon, recording
oblique-slip south-side-up movements kinematically com­
patible with the regional D2 increm ent o f deform ation
(Poulsen et al., 1986b; Thomas and Heaman, 1994). The
deposits consist of laminated fault-fill veins (Jolu, Rod Main
zone; Fig. 18C), breccia vein bodies (21 Zone; Fig. 22D),
and rare, moderately dipping extensional veins external to
the shear zones (Jolu, Rod South zone). All deposits have
similar vein and wall-rock alteration mineral assemblages,
suggesting that they are of the same age. At the 21 Zone
deposit, discordant vein segments in the breccia body
clearly truncate the mylonitic foliation of the host shear
zone (Fig. 22D), indicadng the vein formed after foliauon
development. In o th er deposits such at Jasper, fault-fill
veins and the enclosing mylonitic foliation are overprinted
by asymmetric folds plunging subparallel to the shear
zone’s elongadon lineation (Fig. 28G). Given their open to
tight character, these folds are incompadble with the move­
ment direcdon in the host shear zone and are best ascribed
to the effects of the Ds transcurrent deformadon in the dis­
trict (Lewryetal., 1990). Orebodies within the shear zone
and the dip of external extensional veins all lie at a high
angle to the slip direction in the host shear zone. This
geometry is consistent with their occurrence at branches
and deflecdons in the shear zones at a high angle to the slip
direcdon (Fig. 19). Only in the Jasper deposit is this larger
pattern modified to include smaller ore shoots plunging
parallel to lineadon and die axes o f m inor folds. These rela­
tionships suggest that the veins in the Star Lake district
have form ed in existing shear zones, during (the late
stages?) of D2, but prior to D3.
Barren veins overprindng mineralized rocks can also be
used to place dme constraints on mineralization if the veins
can confidently be ascribed to aspecific deformation incre­
ment. In the Harker-Holloway district o f the Abitibi (Fig.
1), disseminated-replacement gold m ineralization a t the
Holt-McDermott and Holloway deposits is overprinted by a
set o f barren subhorizontal extensional veinlets (Fig. 30A;
Robert, 1997). These veinlets are perpendicular to the
local, weakly developed, subvertical elongation lineadon L*
and are interpreted as having formed during D2. T he fact
that these veins cut disseminated-replacement orebodies
suggests that gold mineralizarion a t these deposits predate
: some or all o f the D2 deformadon (Robert, 1997).
In a num ber o f cases, there may not be a unique solution
for the incremental strain axes o f a vein network, leaving
ambiguides in correlaung vein network axes with those o f
regional deformadon increments. This is illustrated by the
Mount Charlotte deposit, where two disdnet sets o f strain
(stress) axes can explain the development o f stockwork ore-
bodies (Ridley and Menglcr, 1998). Finally, this approach
can be used to highlight possible uncertainties in the tim­
ing o f mineralization. At the Revenge deposit, for example,
the vein network has been interpreted by Nguyen e t al.
(1998) to have formed in a congressional stress field (ver­
tical <J3), recording an increm ent of subvertical elongation
dX . However, the vein network is interpreted as haring
formed in a system of D3 sinistral wrench faults, for which
the elongation direction, by definition, is implied to be sub­
horizontal. Although this apparent incompatibility can be
explained by transpressional deform ation, this approach
serves to identify questions needing further attention.
Vein timing criteria in penetratively strained rocks
A second approach to establishing the structural timing
of veins relies on specific structural and geological relation­
ships at the mesoscopic, hand-specimen, and microscopic
scales. T hree general cases o f structural timing have typi­
cally been argued by geologists for veins in penetratively
strained rocks: (1) veins are synchronous with development
o f penetrative fabrics; (2) veins predate developm ent o f
penetrative fabrics and are unrelated to observed strain in
the host structure; o r (3) veins completely postdate fabric
development in the host rocks. These three situations are
discussed below, with an attem pt at formulating some crite­
ria for distinguishing among these different cases. The pro­
posed criteria can be regarded as a series o f questions that
should guide the observation and docum entation of spe­
cific features o r relationships in the field. Few o f the pro­
posed criteria are entirely diagnostic, and any interpreta­
tion of the timing of mineralization relative to deformation
must therefore be based on as many criteria as possible.
Veinsfanned synchronously with penetrativefabrics in their host
structure: Many auriferous quartz veins have been inter­
preted as having form ed synchronously with th eir host
shear zones (McCuaig and Kerrich, 1998). In some cases,
veins have been interpreted to develop late in the history of
their host structure (Robert and Brown, 1986), such that
they have retained their primary geometry and their inter­
nal textures are relatively well preserved. In oth er cases,
deform ation can outlast vein development, and the veins
will show some evidence o f deform ation overprinting
related to continued m ovement along tire host structure.
A num ber o f field criteria can be used to establish whether
o r not veins in shear zones are related directly to shear zone
displacement, even though they may be overprinted by con­
tinued shear zone deformation. Some o f the expected geo­
metric and structural relationships for veins formed synchro­
nously with their host structures are listed below:
1. W here both fault-fill veins and fringing extensional
veins are present, the overall vein configuration and geom­
etry will be com patible with the kinem atics o f th e host
shear zone (fig. 19), provided that the two sets o f veins are
contem poraneous. For example, the line o f intersection
between fault-fill and extensional veins is expected to be
perpendicular to th e slip direction along the host shear
zone, and parallel to the intersection between conjugate
sets o f shear zones if present (Fig. 19). This is case a t the
Sigma-Lamaque deposit, as illustrated in Figure 26,
where north- and south-dipping fault-fill veins and sub­
horizontal extensional veins all intersect along a line sub­
parallel to the slip direction along the related set o f con­
jugate shear zones.
 F ig . 3 0 . A . S u b h o r i z o n t a l b a r r e n e x t e n s i o n a l q u a r t z w i n s c r o s s c u t t i n g o r e - g r a d e a l b i t i z e d a n d p y r i t i z e d b a s a l t i n t h e
H o l l o w a y d c j j o s i t , H a r k e r - H o l l o w a y d i s t r i c t , c r o s s s e c t i o n v ie w , h a m m e r f o r s c a l e . C o u r t e s y o f t h e G e o l o g i c a l S u r v e y o f
C a n a d a , p h o t o n o . 1 9 9 9 -Q 1 5 C C . B . P h o t o m i c r o g r a p h o f t o u r m a l i n i z c d s h e a r z o n e i n m a f i c v o l c a n i c r o c k s a d j a c e n t t o a
q u a r t z - t o u r m a l i n e fa u lt- fill v e in o f t h e S i g m a d e p o s i t , V n l d 'O r . S h e a r z o n e f o l i a t i o n is m a r k e d b y t h e tr a i l s o f f i n e r u t i l e
g r a i n s ( b la c k ; s e e a r r o w ) ; i t is f o l d e d a n d c le a r ly o v e r g r o w n b y t o u r m a l i n e p o r p h y r o b l i u t s ( g r a y ) . F i e l d o f v ie w is 1 .3 5 m m
w id e . R e p r i n t e d w ith p e r m i s s i o n o f Economic Geology, R o b e r t , 1 9 9 6 , f i g . l . C . B a r r e n b u t f o l i a t e d d i o r i t c d ik e s ( a r r o w s ) c u t ­
t i n g a c r o s s q u a r t z - c a r b o n a t e - p y r i t c v e in s w i t h i n a n i n t e n s e s h e a r z o n e i n u l t m m a f i c r o c k s . F ir s t C a n a d i a n d e p o s i t , V al
d 'O r . O b l i q u e view , s t e e l p l a t e ~ 1 5 c m a c r o s s . D . I n t e n s e f o l i a t i o n ( r i g h t h a l f o f p h o t o g r a p h ) o v e r p r i n t i n g v e in s o n t h e
f r i n g e s o f a s t o c k w o r k o r e b o d y a t t h e S a n A n t o n i o d e p o s i t , R ic e L a k e . C r o s s s e c t i o n v ie w , h a m m e r f o r s c a le .

2. Planar extensional veins, if present, will cut the pen­ ing determined for the host shear zone (Fig. 19), as in the
etrative foliation and will be nearly perpendicular to the examples shown in Figures 20B and 25C.
foliation and elongation iineation in the shear zone or 5. Striations on slip surfaces within fault-fill veins, and
strained rocks. F urtherm ore, the line o f intersection especially those defined by hydrothermal minerals present
between the veins and the foliation will lie nearly per­ in the veins (Fig. 18F), will be subparallel to elongation lin-
pendicular to the elongation Iineation. M ineral fibers eations in the host foliated rocks (Fig. 19).
within the veins will also be subparallel to the elongation 6. Foliated wall-rock slivers within fault-fill veins are gen­
Iineation. erally intensely altered. In many cases, the foliated nature
3. If extensional veins arc buckled due to continued of the wall-rock slivers is lost due to progressive replace­
shear zone deformation, their enveloping surfaces remain m ent by hydrothermal minerals with increasing intensity of
at a high angle to the shear zone’s elongation, unless shear alteration (Robert and Brown, 1986). This observation indi­
strain is very' high. In this case, the penetrative foliation is cates that vein-related hydrothermal alteration postdates
expected to be deflected around rotated limbs o f buckle the bulk o f foliation development. At the Sigma-Lamaque
folds, as opposed to axial planar to the folds. The line of deposit, trails o f ilm enite and rutile grains defining the
intersection between the vein and the foliation plane, also foliation in wall-rock slivers within fault-fill veins are over­
corresponding to the axes o f the buckle folds, will lie at a grown by tourm aline porphyroblasts (Fig. 30B), clearly
high angle to die elongation Iineation, as is the case for the indicating that veins and associated alteration formed in a
folded veinlet shown in Figure 28B. zone o f already foliated wall rock (i.e., in a preexisting
4. The internal geometry in en echelon extensional vein shear zone; see Robert, 1996). O f similar significance are
arrays, if present, will be compatible with the sense o f shear­ the discordant segments o f veins truncating the myloniiic

inr
 foliation at the 21 Zone deposit in the Star Lake district the Hemlo (Robert and Poulsen, 1997), Bousquet (Cos-
(Fig.22D). selin et al, 1994), Val d ’O r (Couture e t al., 1994; Robert,
7. Contrasting degrees of strain will commonly be 1994) , and Chibougamau (Magnan and Blais, 1995) dis­
observed between adjacent quartz laminae in fault-fill veins. tricts in Canada.
Laminae with nearly strain-free quartz (showing only undu- 2. In the case o f veinlet and veinlel stockwork ores, the
lose extinction) will coexist with strained laminae in which presence offish-hooks, isolated fold hinges, and trails o f
quartz is polygonized to completely recrystallized (Boullier isolated boudins o f mineralized veinlets, indicates intense
and Robert, 1992; see also fig. 5 in Robert et al., 1995). The transposition (Hobbs e t al., 1976) and provide a good indi­
contrasting degree o f strain between different laminae in cation that the ores have been intensely deform ed. Such
shear veins reflects different degrees o f deformation over­ fish-hooks and isolated hinges are best observed looking at
printing due to their successive formation at different stages a plane perpendicular to the stretching lineation, as
of shear zone activity, the older laminae being more strained opposed to looking at a plane parallel to it.
than younger ones. In veins overprinted by subsequent 3. The gradation of a stockwork zone from a well-pre­
deformation, for example in the case o f shear zone reactiva­ served state outside a shear zone into a strongly deformed
tion, most if not all quartz laminae would be expected to stockwork within the shear zone indicates that the stock-
show similar, moderate to high degrees o f strain. work predates much if not all of the strain within the shear
zone. An example from the San Antonio stockwork ore-
Overall, veins form ed during progressive penetrative bodies is illustrated in Figure 30D, where intact stockwork
deformation will display some form of conflicting relation­ veinlets (above the ham m er) are increasingly deform ed
ship to deformation features of their host rocks. For exam­ toward the center of the shear zone on the right.
ple, some veins in a given set, or some part of a vein, may 4. The presence of a pervasive, strong foliation within a
truncate penetrative fabrics in the host rock, and die oppo­ vein, as is the case for the quartz vein from the Jasper
site relationship is noted for other veins of the set or for deposit at Star Lake shown in Figure 29G, is indicative of a
other parts of the same vein. This is the case for the 21 significant amount of strain overprinting the vein.
Zone deposit, where some discordant vein segments trun­ 5. Similarly, the joint folding of shear zone foliation and
cate the mylonidc foliauon in the core of the orebody, but foliation-parallel veins reflects the existence o f a significant
where the same veins are slighdy buckled and boudinaged overprinting strain. In the case of the Jasper deposit at Star
on the margins of the orebody, reflecting some weak over- Lake, the plunge of asymmetric open folds (Fig. 28G) is sub-
prinung strain. This type of conflicting relationship is also parallel to the elongation lineation in the host shear zone.
illustrated by the sample shown in Figure 28B, where the The plunge of the fold is incompatible with that of the folds
vein clearly truncates the foliation but is buckled by fur­ resulting from continued movement along the host struc­
ther shortening across it, while retaining geometric com­ ture, and the strain overprinting the vein is best interpreted
patibility with the contained elongation lineation, as dis­ as related to shear zone reactivation (during D3).
cussed above. 6. Folded extensional veins that predate penetrative foli­
ation are likely to display walls that are crenulated by the
; Veinsformed before penetrative deformation in their host struc­ foliation (Fig. 28A). In such cases, the enveloping surface
tures: A num ber o f vein gold deposits have been interpreted o f the veins can be at any angle to the elongation lineation
, as being overprinted by structural fabrics in their host (as opposed to a t a high angle for a vein synchronous with
/rocks, and to predate all o r a significant part o f the pene- fabric development). In the rotated limbs o f a folded vein,
-trative strain. Examples include the Campbell deposit in limited deflection o f the foliation is expected (Fig. 28C;
the Red Lake district (Penczak and Mason, 1997), the compare with Kg. 28B).
Hollinger-Mclntyre deposit in the Timmins district (Mason 7. T he geometric arrangem ent and kinematics o f over­
and Melnik, 1986), and sulfide-rich Cu-Au vein deposits in printed en echelon vein arrays (e.g., sigmoidal extensional
the Chibougamau district (Magnan and Blais, 1995; Pilote veins) or multiple vein sets (combined fault-fill and exten­
etal., 1995). Relationships that can be used to support such sional veins) will be incompatible with the kinematics o f the
interpretations include the following: host shear zone. This would be the case, for example, o f an
array of steeply-dipping, recognizably sigmoidal extensional
1. The presence o f postore dikes, cutting across veins orveins recording strike-slip shearing, overprinted by a pene­
other styles o f mineralization, which are themselves folded, trative foliation with down-dip elongation lineation.
transposed, intensely foliated, o r boudinaged in penetra­ 8. Several gold deposits occur in zones o f intense chlo­
tively strained rocks (shear zone o r regional foliation), pro­ rite o r sericite schists. Some of these schist zones are rela­
vides a clear time m arker in the history o f the host struc­ tively thick compared to their strike lengths and are diffi­
ture. Figure 30C shows a mafic dike crosscutting a vein at a cult to trace for any significant distance. Because shear
small angle within a shear zone; the dike itself is intensely zones along which displacement has taken place tend to be
foliated, which indicates that the vein form ed prior to at long and skinny, one can suspect that short “stubby” zones
least the last increm ent of penetradve deformation within o f schist around veins overprint and mimic the shapes of
the host shear zone. Additional examples o f deform ed preexisting phyllosilicate alteration zones (Pilote e t al.,
dikes cutting gold mineralizadon have been described in 1995) .
 Finally, it should be noted that shear zones and synchro­
nous veins within them may also be reactivated o r folded
together during subsequent generations o f deform ation
belonging to the same deformational event. Examples of
folded auriferous veins in the Larder Lake-Cadillac fault
zone have been described at the Orenada Zone 4 deposit,
Val d ’O r (Robert, 1990), and examples o f folded shear
zones and veins been described in the Rice Lake belt by
Brommecker et al. (1989). Shear zone-hosted veins that are
folded in subsequent deformation events are difficult to dis­
tinguish from those that have been folded during shear
zone developm ent
Veins postdating their host structure: There are at least two
lines o f evidence that have been used to argue for the for­
mation of auriferous veins well after regional deformation
and metamorphism and, thus, postdating the structures to
which they appear to be related.
The first argum ent is geological and commonly revolves
around the apparent late paragenesis of gold in some veins
(Mawdsley, 1938; W hite, 1943; McCuaig and Kerrich,
1998). These arguments closely parallel ideas set forward by
McKinstry and Ohle (1949), who regarded veins in part as
metasomatic replacem ents o f preexisting structural fea­
tures rather than exclusively dilational fillings of fractures.
The type of observation that supports tins view is illustrated
in Figure 21C, where an entire shear zone and its included
foliadon hits been tourmalinizcd to produce a tabular com­
petent rock mass overprinted further by an array of exten-
sional quartz veins. Taken on its own merits this observation
could lead one to conclude that the shear zone played a
passive role, serving only as a fluid conduit well after the
tectonites were formed within it. This type of argument has
also been applied at a microscopic scale to create a distinc­
tion between the timing of vein formation as opposed to
the timing of introduction of gold into the vein. Cases have
been made for late introduction of gold into previously bai^
ren veins, for example in the Star Lake district, typically
where quartz has recrystallized to finer grain size or subse­
quently been fractured (e.g., Hrdy and Kyser, 1995).
The second argum ent for late developm ent of gold
deposits in greenstone belts is geochronological. As
reviewed by Kerrich and Cassidy (1994) and McCuaig and
Kerrich (1998), the ages of auriferous veins in the Abitibi
Greenstone belt and the La Ronge Domain have been esti­
mated by a variety of methods (U-Pb, Pb-Pb, Ar-Ar, Sm-Nd)
applied to a variety of minerals. In nearly all cases, except
for U-Pb zircon data (Kerrich and King, 1993), the results
suggest vein formation postdating deformation, metamor-
phism, and plutonism by as m uch as 50 to 100 Ma after
regional deform ation and m etamorphism . These young
ages have been interpreted as representing either the age
of primary gold mineralization, selective resetting of iso­
topic systems during late hydrotherm al overprinting, or
remobilization (McCuaig and Kerrich, 1998).
The above arguments notwithstanding, there is evidence
of at least some deformation overprinting auriferous veins at
most deposits implying pre- or syn-kinematic timing, making
it difficult to rationalize a late, postkinematic timing for min­
eralization. Although one might dismiss the late paragenesis
o f gold in veins as evidence of rem obilization, and the
apparent young radiometric ages as evidence o f selective iso­
topic resetting, the observations are real and should never­
theless be considered in the structural analysis of veins in
gold deposits. O ne must acknowledge, however, the possi­
bility that some deposits may have formed very late in the
structural history of theregions in which they occur, partic­
ularly where evidence o f deformation o f veins i$ minimal.
Discussion and Concluding Remarks
T he determ ination o f the structural tim ing of veins in
greenstone gold deposits, either in relation to their host
structures or to the structural evolution of districts, is a chal­
lenging task. It is nevertheless essential to attem pt this in
order to highlight structural controls o f mineralization, to
select appropriate gold deposit models, and to formulate
valid predictions of the geometry and plunges o f deposits
and orebodies.
It is apparent from die above descriptions that nearly all
greenstone gold deposits show evidence for some level o f
pre-, syn-» and postm ineralization strain. T he task is to
assign the timing o f the main stage o f ore deposition cor­
rectly in relation to the structural evolution of districts and
o f host structures. Pre-deformation deposits will record the
effects o f all strain increments in the district, and their orig­
inal shapes and morphologies will be substantially modi­
fied. This is the case o f the Bousquct no. 2 gold-rich vol-
canic-hostcd massive sulfide deposit, which plunges parallel
to the elongation lineation in the highly strained host rocks
(Marquis et al., 1990). In deposits formed synchronously
with their host structures, there will be evidence that veins
and associated alteration overprint the host shear zone foli­
ation. For example, this can take the form of vein segments
truncating foliadon planes, as at die 21 Zone deposit (Fig.
22D), or of alteration minerals overgrowing the foliation in
wall-rock slivers within fault-fill veins, as at Sigma-Lamaque
(Fig. 30B). However, there will also be varying degrees of
deform ation overprinung the veins. This can result from
continued deformation along the host structure, as can be
argued for folds and boudins geometrically consistent with
the slip direction along the shear zone (e.g., the sample
shown in Fig. 28B). Overprinung deformation can also be
related to the effects of a subsequent deform ation incre­
ment, producing structural features inconsistent with the
movement direction along the host shear zone (e.g., the
asymmetric folds of vein and foliadon at the Jasper deposit;
Fig. 28G). Deposits can also postdate the development of
dieir host structure, such as where a vein is produced dur­
ing reactivadon of a preexisting shear zone during a later
increm ent of deform ation. In such cases, the geometric
and kinematic features of the veins will not be compadble
with those of the host shear zone.
Two main factors will influence the degree of preserva-
uon of gold deposits. The first is the timing of ore forma­
tion in the evolution of the district, with the youngcr-
formed deposits expected to be the least deform ed and
 best preserved. The second factor is the location o f the
deposits within the district, because of the domainal distri­
bution of the effects of some deformation increments, as
shown in Figure 5. As a result, for two deposits formed at
the same time, one may be well preserved if it is located in
an area where the effects o f a subsequent strain increment
are not well developed, whereas the other can show signifi­
cant overprinting. This is the case for the D«> quartz-tour­
maline veins o f the Val d ’O r district (Fig. 2): those in the
Sigma-Lamaque deposit, located in an area where the
effects of D3 are weak (Fig. 5), are relatively well preserved.
In contrast, those o f the O renada Zone 4 deposit, within
the Larder Lake-Cadillac fault zone, are overprinted by F<j
folds because of the reactivation of this fault during Ds. The
heterogeneous distribution of the effects of late (Ds) strain
increments can also explain why some veins in the Star
Lake district are well preserved, such as those in the 21
Zone deposit (Fig. 22D), whereas other veins, like those at
Jasper, are overprinted by a strong internal foliation (Fig.
29G) and asymmetric folds (Fig. 28G). An important impli­
cation is that it may be erroneous to conclude that a deposit
has formed very late in die evolution o f its host district on
the simple basis of the absence of overprinting deforma-
don: the deposit may have formed in an area where subse­
quent deformation increments had only very minor effects.
Knowledge of the structural timing of mineralization also
/ provides a basis for assessing interpretations of the age of
mineralizadon reached from other, or from perhaps incom­
plete, lines of evidence. For example, the very young ages
obtained for the quartz-tourmaline veins at Val d ’Or (see
. above), are inconsistent with the fact that the veins have
formed during D2, and prior to the last im portant incre­
ment of deformadon, Ds. In the Star Lake district, the feet
that gold occupies late brittle fractures in foliated and
folded vein quartz at Jasper has been used to argue for
introduction of gold into the veins after they were formed
: and deform ed (Hrdy and Kyser, 1995). Consideration o f
the 21 Zone deposit, where there is lim ited overprinting
deform ation and where there is an intim ate association
between the am ount o f coarse pyrite in the vein and the
gold grades (Poulsen et al., 1986b), rather suggests that
pyrite and gold were introduced together, during the for­
mation o f the veins.
Finally, with constraints on the liming and structural his­
tory o f a given deposit, it is possible to make some valid pre­
dictions concerning the possible plunges o f deposits and
orebodies, o r high-grade shoots within them , depending
pn the specific structural and lithologic context. In deposits
, synchronous with their host structures, orebodies in dila-
;•tional jogs o r at shear zone splays and intersections will
plunge at a high angle to the elongation lineation in the
host structure (Fig. 31A). In a few cases, however, orebodies
'have also been docum ented to plunge parallel to the slip
direction (see Poulsen and Robert, 1989). Orebodies in
bends caused by refraction o f the shear zone across a litho­
logic contact will instead plunge parallel to the line of inter­
section between the shear zone and the layer, and so will
orebodies formed by replacement o f chemically favorable
PLUNGE OF OREBODIES IN RELATION
TO THEIR STRUCTURAL SETTING
LONGITUDINAL SECTION STRUCTURAL SITE
Veins in dilationa! jogs
Veins at shear zone and vein
intersections
Some slockwork zones
- Veins in folds, in axial plane or
"AC" positions
- Folded orebodies
-intensely sheared orebodies
Veins in dilational jogs at
intersection with favorable layer
Orebodies at intersection with
favorable layers (e.g., BIF)
A Elongation Line of
J
Fold axis
T or slq»direction intersection
Fig. 31. Diagram showing the expected geometric relationships
between the plunges o f orebodies and the slip direction or elongation lin­
eation in shear zones in diverse structural settings (see text for discussion).
units (e.g., banded iron form ations) intersected by the
shear zone (Fig. 31B). Vein deposits formed in folds, con­
sisting o f veins parallel to the axial plane o r o f veins in AC
jo in t positions, tend to be concentrated within specific
lithologic units and will plunge parallel to the fold axes
(Fig. 31C). Orebodies overprinted by intense shear zones
will likely plunge parallel to the elongation lineation (Fig.
31C), whereas folded deposits and orebodies are expected
to plunge parallel to the fold axes.
In summary, careful structural analysis o f veins should
provide an accurate picture of the permissible dming of vein
formation relative to other deformational events, and, most
importantly, may highlight significant ore controls and pro­
vide some useful guidelines for local mineral exploration. In
addition, structural analysis provides an assessment o f the
degree to which a deposit has been modified by overprint­
ing deformation. This information can in turn be used in
determining possible and permissible plunges o f orebodies
o r o f high-grade ore shoots within them, even when only
limited structural information is available, information o f
particular value in the early stages of drilling an occurrence.
The timing of mineralization in deformed terranes is also an
essential paiam eter in the classification o f deposits o r occur­
rences and in the selection o f appropriate genetic o r explo­
ration models. This, in turn, relies heavily on the correct
interpretation of the structural significance o f veins. Char­
acteristics o f veins interpreted to be orogenic by some
authors (Groves et al., 1998) are similar to those for veins
argued to be intrusion-related by others (Sillitoc and
Thompson, 1998). Even though these designations are by
no means mutually exclusive, they cariy with them an over­
all genetic connotation, which commonly also guides explo­
ration. T he geologist working in this field must therefore be
willing to continually evaluate the reliability of genetic inter­
pretations o f veins based on further interpretations o f their
structural characteristics and their structural settings.
Acknowledgments
The review presented here relies heavily on more than
15 years o f structural studies in gold deposits and districts
by the authors. This work would never have been possible
without the collaboration and participation of numerous
colleagues from the Geological Survey o f Canada, provin­
cial geological surveys, and the private sector. We extend
our deepest appreciation to all o f these people, and partic­
ularly to B. Dube for constructive discussions of many of the
ideas presented here. Comments of the reviewers and J.P.
Richards’ editorial skills led to significant improvements of
the original manuscript. Finally, we thank the editors of this
volume for understanding the challenges o f writing such a
paper and of meeting deadlines while working in an indus­
try environment.
REFERENCES
Abraham, A.P.G., and Spooner, E.T.C., 1995, Late Archcan regional defor­
mation and structural controls on gold-quartz vein mineralization in the
northwestern Slave Province, N.W.T., Canada: Canadian Journal o f
Earth Sciences, v. 32, p. 1132-115-1.
Andrews, A.J., H ugon, II., Durochcr, M., Corfu, lr„ and Lavigne, M J..
1986, The anatomy o f a gold-bearing greenstone belt: Red Lake, north­
western Ontario, Canada, in Macdonald, A.J., cd„ Proceedings o f Gold
*86, An international symposium o n the geology o f gold deposits:
Toronto, Konsult International, p. 3-22.
Angelier, L979, Determination o f the moan principal directions o f stress
for a given fault population: Tcctonophysics, v. 56, p. T17-T26.
Beach, A., 1975, The geometry o f en-cchelon vein arrays: Tcctonophysics,
v. 28, p. 215-263.
Berger, A.R., 1971, Dynamic analysis using dikes with oblique internal foli­
ations: Geological Society o f America, v. 82, p. 781-786.
Blocker, W., and Parrish, R.R., 1996, Stratigraphy and U-Pb zircon
geochronology o f Kidd Creek: Implications for the formation o f giant
voleanogenic massive sulphide deposits and the tectonic history o f the
Abitibi greenston e belt: Canadian Journal o f Earth Sciences, v. 33,
p. 1213-1231.
Bouliier, A.-M., and Robert, F. 1992, Palcoscistnic events recorded in
Archean gold-quartz vein networks, Val d ’Or, Abitibi, Quebec, Canada:
Journal o f Structural Geology, v. M .p . 161-179.
Brommecker, R., Poulscn, K.H., and H odgson, C.J., 1989, Preliminary
report on the structural setting o f gold at the Gunnar mine in the Bcrcs-
ford Lake area, U chi subprovince, southeastern Manitoba: Current
Research, Part C, Geoogical Survey o f Canada, Paper 89-1C, p. 325-332.
Burrows, D.R., Spooner, E.T.C., W ood, P.C., and Jem iclita, R.A., 1993,
Structural conrol on formation o f the Hollinger-Mclntyre Au quartz
vein system! in the H ollinger shear zone, Tim m ins, southern Abitibi
greenstone belt, Ontario: Economic Geology, v. 88, p. 1643-1663.
Bursnall, J.T., cd., 1989, Mineralization and shear zones: Geological Asso­
ciation o f Canada. Short Course Notes, v. 6 .3 0 9 p.
Canadian Institute o f Mining and Metallurgy, 1948. Structural geolog)' o f
Canadian ore deposits:Jubilee Volume, 948 p.
Card, K.D.. 1992. A review o f the Superior Province o f the Canadian
Shield, a product o f Archcan accretion: Prccambrian Research, v. 48,
p. 99-156.
Card, K.D., Poulscn, K.H.. and Robert, F., 1989, T he Archean Superior
Province o f the Canadian Shield and its lod e gold deposits, in Keys.
R.R.. Ramsay, W.R.H., and Groves, D.I., cds.. T h e G eology o f Gold
Dc|>osi(s: The Perspective in 1988: Economic Geology Monograph 6,
p. 19-36.
Cassidy, K.F., Groves. D.I.. and M cNaughton, N J„ 1998, Late-Archcan
granitoid-hosted lode gold deposits, Yilgam Craton. Western Australia:
Deposit characteristics, crustal architecture, and implications for ore
generis: O re Geology Reviews, v. 13, p. 65-102.
Clark, M.E., Archibald, N.J., and Hodgson, C.J., 1986. The structural and
mctamorphic selling o f the Victor)'gold m ine, Kaiubaida, Western Aus­
tralia, in Macdonald, A.J.. cd.. Proceedings o f Gold *86. An international
symposium on the geology o f gold deposits: Toronto, Konsult Interna­
tional, p. 243-254.
Clout. Clcghorn, J.H., and Eaton, P.C., 1990, Geology o f the Kalgo-
orlie goldfield, in Hughes, F.E., cd.. Geology o f the mineral deposits o f
Australia and Papua New Guinea: Melbourne, Australasian Institute o f
Mining and Metallurgy, p. 411-431.
Colvinc, A.C., 1989, An empirical m odel for the formation o f Archean
gold deposits: Products o f final cratonization o f the Superior Province,
Canada, in Keys, R.R., Ramsay, W.R.H., and Groves, D.I., cds., T he Gcol-
ogy o f Gold Deposits: T he Perspective in 1988: Econom ic G eology
Monograph 6, p. 37-53.
Cooke. H.C., 1946, Canadian lode gold areas (summary account): Cana­
dian Department o f M ines and Resources, Econom ic G eology Series,
v. 15,86 p.
Corfu, F„ 1993, *I1ic evolution o f the Southern Abitibi greenstone belt in
light o f precise U-Pb geochronology: Econom ic Geology, v. 88,
p . 1323-1340.
Corfu, F., Jackson, S.L., and Sutcliffe, R.H., 1991, U-Pb ages and tectonic
significance o f late Archean alkalic magmatism and nonm arine sedi­
mentation: Timiskaming Group, southern Abitibi belt, Ontario: Cana­
dian Journal o f Earth Sciences, v. 28, p, 489-503.
Couture, J.F., Pilotc, P., Machado, N., and Desrochers, J.P., 1994, Timing o f
gold mineralization in the Val d'Or district, southern Abitibi belt: Evi­
d en ce for two distinct m ineralizing events: Econom ic Geology, v. 89,
p. 1542-1551.
Cox, S.F., 1991, Geomctiy and internal structures o f mesothennal vein sys­
tems: Implications for hydrodynamics and ore genesis during deforma­
tion, in Hronsky.J.MA., cd., Structural Geolog)’ in M ining and Explo­
ration: Perth, University o f Western Australia, University Extension,
Publication 25, p. 47-53.
------ 1995, Faulting processes at high fluid pressures: An example o f fault-
valve behavior from the Watdc Gully Fault, Victoria, Australia: Journal
o f Geophysical Research, v, 100, p. 12,841-12,859.
Cox, S.F., and Etheridge, M A , 1983. Crack-seal fiber growth mechanisms
and their significance in the developm ent o f oriented layer silicate
microsmictures: Tcctonophysics, v. 92, p. 147-170.
Cox. S.F., Etheridge, M.E., and Wall, VJ., 1987, T he role o f fluids in syn-
tectonic mass transport, and localization o f mctamorphic vein-type ore
deposits: Ore Geology Reviews, v. 2, p. 65-86.
Cox, S.F., Wall, VJ., Etheridge, M.E., and Potter, T.F., 1991, Dcformational
and m ctam orphic processes in the formation o f mcsothcrmnl vein-
hosted gold deposits—examples from the Lachlan fold belt in Central
Victoria, Australia: Ore Geology Reviews, p, 391-423.
Cox, S.F., Sun, S.-S., Etherdige, M.D., Wall, V.J., and Potter, T.F., 1995,
Structural and geochemical controls on the development o f eurbidite-
hotsted quartz vein deposits. Wattle Gully m ine, central Victoria, Aus­
tralia: Economic Geology, v. 90, p. 1722-1746.
Cox, S.F., Knackstedt, M A., and Braun, J., 2001, Principles o f structural
control on permeability and fluid flow in hydrothermal systems: Reviews
in Economic Geology, v. 14, p. 1-24.
Daigncault, R., and Archambault, G., 1990, Lcs grands couloirs dc deforma­
tion dc la Sous-Provincc de 1'Abitibi, in Rive, M„ Verpaetsi, P., Gagnon, Y.,
Lulin,J.M.. Riverin, G., and Simard, A., cds., The Northwestern Quebec
Polymetallic Belt: A Summary o f 60 Years o f Mining ami Exploration:
Canadian Institute of Mining and Metallurgy, Special Volume 13. p. -13— 6-1.
Dube, B.. Poulscn, K.H.. and Cuba.J., 1989. The effects o f layer anisotropy
on auriferous shear zones: The Norbeau mine, Quebec: Economic
Geology, v. 8-1. p. 871-878.
Diirncv, D.W., and Ramsay,J.G., 1973. Incremental strains measured by
symectonic crystal growths, in D cjon g. K_A., and Scholten, R.. eds.,
Gravity anti tectonics: New York, Wiley, p. 67-96.
Eiscnlohr, B.N., Groves, D.I., and Partington. G.A., 1989, Crustal-scale
shear zones and their significance to Archaean gold mineralisation in
Western Australia: Mineralium Deposits, v. 2-1. p. 1-8.
Etheridge, M.A., 1983, Differential stress magnitudes during regional
deformation and metamorphism—upper bound imposed by tensile
fracturing: Geology, v. 11, p. 231-235.
Etheridge, M.A.. Wall. VJ„ Cox, S.F., and Vernon, R.H., 1984. High fluid
pressures during regional metamorphism and deformation: Im|>lication
for mass transport and deformation mechanisms: Journal o f Geophysi­
cal Research, v. 89. p. 4344-4358.
Flinn, D„ 1962, On folding during three-dimensional progressive defor­
mation: QuarterlyJournal o f the Geological Society of London, v. 118.
p. 385-433.
Foster, R.P., 1989, Archean gold mineralization in Zimbabwe: Implications
for metallogenesis and exploration, in Key?. R.R.. Ramsay, W.R.H., and
Groves, D.I., eds.. The Geology o f Gold Deposits: The Perspective in
1988: Economic Geology Monograph 6. p. 5-1-70.
Foxford, K A , Nicholson, R., and Polya, DA., 1991, Textural evolution of
WCu-Sn-bearing hydrothermal veins at Minas da Panasqucira. Portugal:
Mineralogical Magazine, v. 55, p. 435—145.
Foxford, KA.. Nicholson, R., Polya, DA., and Hchhletlnvaiic, R.B.P.. 2000,
Exiensional failure and hydraulic valving at Minas da Panasqucira. Por-
uignl: Evidence from vein spatial distributions, displacements, and
geometries: Journal o f Structural Geology, v. 22, p. 1065-1086.
Gossclin, G., Savoie, A , and Dnigncault, R., 1994, Misc en place dcs vcincs
aurif&res ft la mine Doyon, Preissac, Abitibi Sud, in Scminairc d ’infor-
mation sur la recherche geologique, Programme ct resumes 1994: Min-
istere des Ressources Naiurellcs, Quebec, DV 94-09, p. 37.
Groves, D.I., and Foster, R.P., 1991, Archean lode gold deposits, in Fos­
ter, R.P., ed„ Gold M etallogeny and Exploration: London, Blackic,
p , 63-103.
Groves, D.I., Ridley, J.R., Bloem, E.M.J., Gcbre-Mariam, M„ Hagemann,
S.G., Hronsky,J.M.A., Knight, J.T., McNaughton, N.J., Ojala.J., Vielrc-
V idler, R.M., McCuaig, T.C., and Holyland, P.W., 1995, Lode gold
deposits o f the Yilgarn Block: Products o f laic-Archean crustal-scale
overpressured hydrothemtal systems, in Coward, M.P., and Ries, A.C„
cds,, Early Prccambrian Processes: Geological Society o f London Special
Publication 95, p. 155-172.
Groves, D.I., Goldfarb, R.J., Gcbre-Mariam, M„ H agem ann, S.G., and
Robert, F., 1998, Orogenic gold deposits: A proposed classification in
the context o f their crustal distribution and relationships to other gold
deposit types: Ore Geology Review, v. 13, p. 7-27.
Hagemann, S.G., and Cassidy, K.F., 2000, Archean orogenic gold deposits:
Reviews in Economic Geology, v. 13, p. 9-68.
Hagemann, S.G., Groves, D.L, Ridley, J.G., and Vearncombe, J.R., 1992,
The Archean lode gold deposits at Wiluna, Western Australia: high-level
brittle-style mineralization in a strike-slip regime: Economic Geology,
v. 87, p. 1022-1053,
Hanmer, S., and Passchier, C., 1991, Shear-sense indicators: A review: Geo­
logical Survey o f Canada, Paper 90-17, 72 p.
Heather, K.B., 1998, New insights on the stratigraphy and structural geol­
ogy o f the southwestern Abitibi greenstone belt: Implications for the tec­
tonic evolution and setting o f mineral deposits in the Superior Province,
in Bottrill, T.J., and Shcahan, P„ compilers, The first age o f giant ore for­
mation: Stratigraphy, tectonics and mineralization in the Late Archean
and Early Proterozoic: Technical session at the Annual Convention and
Trade Show o f the Prospectors and Developers Association o f Canada,
Toronto, 8 March 1998, p. 63-101.
Hobbs, B.E., Means, W.D., and Williams, P.F., 1976, An outline o f struc­
tural geology: New York, Wiley, 571 p,
H odgson, C.J., 1989, T he structure o f shear-related, vein-type gold
deposits: A review: Ore Geology Reviews, v. 4, p. 231-273.
■' 1993, Mesothermal lode gold deposits, »« Kirkham, R.V., Sinclair,
W.D.. Thorpe. RJ., and Duke, J.M, cds., Mineral deposit modeling: Geo­
logical Association o f Canada, Special Paper -10, p. 635-678.
Hodgson, C.J.. and Hamilton, J.V., 1989, Cold mineralization in the
Abitibi Greenstone belt: End-stage result o f Archean collisional tecton­
ics? in Keays. R.R., Ramsay, W.R.H., and Groves, D.I., eds.. The Geology
o f Gold Deposits: The Perspective in 1988: Economic Geology Mono­
graph 6, p. 86—100.
Hodgson, C.J., and Troop, D.G.. 1988, A new computer-aided methodol­
ogy for area selection in gold exploration: A case study from the Abitibi
greenstone belt, Ontario: Economic Geology, v. 83, p. 952-977.
Hrdy, F., and Kyser. T.K.. 1995, Origin, timing, and fluid characteristics o f
an auriferous event: The Proterozoic Jasper lode gold deposit,
Saskatchewan, Canada: Economic Geology, v. 90. p. 1918-1933.
Hubert, C., 1990, Geologic framework, evolution and structural setting o f
gold and base metal deposits o f die Abitibi greenstone belt, Canada, m
Ho, S.E., Robert, F., and Groves, D.I., compilers, Gold and base metal
mineralization in the Abitibi Subprovince, Canada, with emphasis on
the Quebec segment: Perth, University o f Western Australia, Geology
Department and University Extension. Publication No. 24, p. 53-62.
Hutchinson, R.W., 1993, A multi-stage, multi-process genetic hypothesis
for greenstone-hosted gold lodes: Ore Geology Reviews, v. 8. p. 349-382.
Jebrak, M., 1992. Lcs textures imra-fllonicnnes, marqueurs des conditions
hydrauliques et tcctontqucs: Chroniquc de la Recherche Minterc, no.
506, p. 25-35.
------1997, Hydrothermal breccias in vein-type ore deposits: A review o f
mechanisms, morphology, and size distribution: Ore Geology Reviews,
v. 12. p. 111-134.
Kcrrich, R., 1989, Gcodynamic setting and hydraulic regimes: Shear zone
hosted mesothermal gold deposits, in Bursnall.J.T.. cd.. Mineralisation
and Shear Zones: Geological Association o f Canada, Short Course
Notes, v. 6, p. 89-128.
Kcrrich, R., and Allison, L, 1978, Vein geometry and hydrostatics during
Yellowknife mineralization: Canadian journal o f Earth Sciences, v. 15,
p. 1653-1660.
Kerrich, R., and Cassidy. K.F., 1994, Temporal relationships o f lode gold
mineralization to accretion, magmatism, metamorphism and defor­
m ation-—Archean to present: A review: O re G eology Reviews, v. 9,
p. 263-310.
Kerrich, R., and King. R., 1993, Hydrothermal zircon and baddeleyite in
Val d'Or mesothermal gold deposits: Characteristics, compositions, and
fluid inclusion properties: Canadian journal o f Earth Sciences, v. 30,
p. 2334-2351.
Kuhns, R.J., Sawkins, FJ„ and Ito, E„ 1994, Magmatism, metamorphism,
and deformation at Hcmlo, Ontario, and the timing o f Au-Mo mineral­
ization in the Golden Giant mine: Economic Geology, v. 89, p. 720-756.
Kochn, D., and Passchier, C.S., 2000, Shear sense indicators in striped bed-
ding-veins: Journal o f Structural Geology, v. 22, p. 1141-1151.
Krapcz, B., Brown, S„ and Hand, J., 1997, Stratigraphic signatures of depo-
sitionnl basins in Archaean voleanosedimemary successions o f the East­
ern Goldfields Province, in Cassidy, K.F., Whitaker, A.J., and Liu, S.F.,
eds., Kalgooiiie’97: An international conference on crustal evolution,
metallogeny, and exploration o f the Yilgarn Craton—An update: Aus­
tralian Geological Survey Organization, Extended Abstracts, Record
1977/41, p . 33-38.
Labaumc, P„ Bcrty, C,., and Laurent, Pit., 1991, Syn-diagenetic evolution o f
shear structures in superficial nappes: An example from the Northern
Appcnincs (NW Italy): Journal o f Structural Geology, v. 13, p. 385-398.
Lau, 1988, Structural geology o f the vein system in the San Anto­
nio gold mine, Bissctt, Manitoba, Canada: Unpublished M .Sc Thesis,
Winnipeg, University o f Manitoba, 154 p.
Lewry, J.F., Thom as, D.J., M acdonald, R., and Chiarcnzelli, J., 1990,
Structural relations in accreted icrrancs o f the Trans-Hudson Orogcn,
Saskatchewan: Telescoping o f a collisional regime? in Lewry, J.F., and
Stauffer, M.R., eds., The Early Proterozoic Trans-Hudson Orogcn in
North America: Geological Association o f Canada, Special Paper 37,
p . 75-94.
Magnan, M.. and Blais, A., 1995, The Copper Rand Mine (Au-Cu-Ag), Day
4, in Pilote, R, eel., Prccambrian '95, Metallogcnic evolution and gcol-
 ogy o f the Chibougamau area—from porphyry Cu-Au-Mo to mesother-
rnal lode gold deposits: Geological Survey o f Canada, O pen File 31-13,
p. 87-9-1.
Maiden, K.J., Chimimba, L.R., and Smalley, TJ., 1986, Cuspate ore-wall
rock interfaces, picrccmciu structures, and the localization o f som e sul­
fide ores in deform ed sulfide deposits: Econom ic Geology, v. 81,
p. 1464-1472.
Marquis, P., Hubert, C.f Brown, A.C., and Rigg, D.M., 1990, An evaluation
o f genetic models for gold deposits o f the Bousquet district, Quebec,
based on their tmncralogic, geochemical, and structural characteristics,
i71 Rive, M., Verpaelst, P., Gagnon, Y., Lulin.J.M ., Riverin, G., and
Sitnard, A., eds., T he northwestern Quebec polymetallic belt: A sum­
mary o f 60 years o f mining exploration: Canadian Institute o f Mining
and Metallurgy, Special Volume 43, p. 383-399.
Mason, R„ and Helmstaedt, H.H., 1992, Structural controls during forma­
tion and deformation o f Archcan loclc gold deposits in the Canadian
Shield, in Bartholomew, M.J., Hyndman, D.W., Mogk, D.W., and Mason,
R„ eds., Basem ent tectonics 8: Characterization and com parison o f
ancient and M esozoic continental margins: Dordrecht. T he Nether­
lands, Kluwcr Academic Publishers, p. 625-631.
Mason, R., and Melnik, N„ 1986, The anatomy o f an Arcliean gold system—
The Mclmyrc-Hollinger complex at Timmins, Ontario, Canada. in Mac­
donald, A,)., cd., Proceedings o f Gold '86, An international symposium
on the geology o f gold deposits: Toronto. Konsult International, p. 40-55.
Mawdslcy, J.B., 1938, Late gold and som e o f its implications: Economic
Geology, v. 33, p. 194-210.
McCuaig, C.T., and Kerrich, R., 1998, P-T-t<lcformauon-fluid characteris­
tics o f lode gold deposits: Evidence from alteration systematic*: Ore
Geology Reviews, v. 12, p. 381-433.
McKinstry, H.E., 19-18, M ining Geology: New York, Prentice-Hall, 680 p.
McKinstry, H.E., and O hle, E.L., 1949, Ribbon structure in gold quartz
veins: Economic Geology, v. 44, p. 87-109.
McRitchic, W.D., 1971, Geology' o f die Wallace Lake - Siderock Lake area:
A reappraisal, in McRitchic, W.D., anc! Weber, W., eds, Geology and geo­
physics o f the Rice Lake region, southeastern Manitoba (Project Pio­
neer): Manitoba Mines and Natural Resources, Mines Branch Publica­
tion 71-1, p, 107-125.
Mi!esi,J.-P., Lcdru, P., Feybcsse,J.-L., Dommanget, A., and Marcoux, E.,
1992, Early Proterozoic ore deposits and tectonics o f die Birimian oro-
gcnic belt, West Africa: Precamlman Research, v. 58, p. 305-344.
Moritz, R.R, and Crocket, J.H., 1990, Mechanics o f formation o f the gold-
bearing quartz-fuchsitc vein at the Dome mine, Timmins area, Ontario:
Canadian Journal o f Earth Sciences, v. 27, p. 1609-1620.
Nelson, D.R., 1997a, SHRIMP U-Pb zircon gcochronologicai constraints
on the evolution o f the Eastern Goldfields granite-greenstone terranes,
in Cassidy, K.F., Whitaker, A.J., and Liu, S.F., eds., Kalgoorlic’97: An
international conference on crustal evolution, ineinllogcny, and explo­
ration o f the Yilgarn Craton—An update: Australian Geological Survey
Organization, Extended Abstracts. Record 1977/41, p. 11-14.
—— 1997b, Evolution o f the Archaean granite-greenstone terranes o f the
Eastern Goldfields, Western Australia: SHRIMP U-Pb zircon constraints:
Precambrian Research, v. 83, p. 57-81.
Newhouse, W.H., 1942, O re deposits as related to structural features:
Princeton University Press, 280 p.
Nguyen, P.T., Cox, S.F., Harris, L.B., and Powell, C.M., 1998, Fault-valve
behavior in optimally oriented shear zones: An example at die Revenge
gold mine, Kambalda, Western Australia: Journal ofStnictural Geology,
v. 20, p. 1625-16-10.
Nicholson, R., and Pollard, D.D., 1985, Dilation and linkage o f echelon
cracks: Journal o f Structural Geology, v. 7, p. 583-590.
Pan, Y., and Fleet, M.E., 1992, Calc-silicate alteration at H cm lo gold
deposit, Ontario: Mineral assemblages, P-T-X constraints, and signifi­
cance: Economic Geology, v. 87, p. 1104-1120.
Peacock, D.C.P., and Sanderson, D J., 1992, Effects o f layering and
anisotropy o f fault geometry: Journal o f the Geological Society o f Lon­
don. v. 149, p. 793-802.
Pcnczak, R.S., and Mason, R., 1997, Metamorphosed Arcliean epithermal
Au-As-Sb-Zn-(Hg) vein mineralization at the Campbell m ine, north­
western Ontario: Economic Geology, v. 92, p. 696-719.
Petit, J.-R, Wibbcrlcy, Cj VJ., and Ruiz. ]., 1999, “Crack-seal," slip: Anew fault-
valve mechanism? Journal o f Structural Geology, v. 21, p. 1199-1207.
Pilote, P„ Robert, F., Sinclair, W.D., Kirkham, R.V., and Daigneault, R.,
1995, Porphyry-type mineralisation in the Dore Lake complex: Clark
Lake and Merrill Island areas, Day 3, in Pilote, P., cd., Precambrian ’95,
M ctallogcnic evolution and geology o f the Chibougamau area—from
porphyry Cu-Au-Mo to mcsothermal lode gold deposits: Geological Sur­
vey o f Canada, Open File Report 3143, p. 65-86.
Platt, J.P., 1983, Progressive refolding in ductile shear zones: Journal o f
Structural Geology, v. G, p. 619-622.
Pollard, D.P., and Scgall, P., 1987, Theoretical displacements and stresses
near fractures in rocks: With applications to faults, joints, veins, dikes,
and solution surfaces, in Atkinson, B.K., cd ., Fracture M echanics o f
Rocks: London, Academic Press, p. 277-349.
Poulsen, K.H., 1986, Auriferous sliear zones with examples from the West­
ern Shield, in Clark L .L , cd ., G old in th e Western Shield: Canadian
Institute o f Mining and Metallurgy, Special Volume 38, p. 86-103.
Poulsen, K.H., and Robert. F.. 1989, Shear zones and gold: Practical exam­
ples from the southern Canadian Shield, in Bursnall, J.T., cd„ Mineral­
ization and shear zones: G eological Association o f Canada, Short
Course Notes 6, p. 239-266.
Poulsen, K.H., Ames, D.E., and Galley, A.G.. 1986a, Gold mineralization in
the Star Lake Pluton. La Rongc belt, Saskatchewan: A preliminary
report, in Current Research, Part A: Geological Survey o f Canada, Paper
86-IA, p. 205-212.
Poulsen, K.H., Ames, D.E., Lati, S„ and Brisbin, W.C., 1986b, Preliminary
report on the structural setting o f gold in the Rice Lake area, Uchi Sub-
province, southeastern Manitoba, in Current Research, Part B: Geolog­
ical Survey o f Canada, Paper 86-1B, p. 213-221.
Poulsen, K.H., Weber, \V„ Brommecker, R., and Scneshen, D.N., 1996,
Lithostratigraphic assembly and structural history o f gold mineraliza­
tion in the eastern Rice Lake greenston e belt, Manitoba: Field Trip
Guidebook A4, Geological Association o f Canada / Mirtcralogical Asso­
ciation o f Canada Annual Meeting, W innipeg May 27-29,1996, 106 p.
Poulsen, K.H., Robert, F„ and Dube, B., 2000, Geological classification o f
Canadian gold deposits: Geological Survey o f Canada, Bulletin 540,106 p.
Ramsay, J.G., 1967, Folding and fracturing o f rocks: New York, McGraw-
Hill, 568 p.
— 1980a, Shear zone geometry: A review:Journal o f Structural Geology,
v. 2, p. 83-100.
-------1980b, T h e crack-seal m echanism o f rock deform ation: Nature,
v. 284, p. 135-139.
------ 1982, Rock ductility and its influence o n the development o f tectonic
structures in m ountain belts, in Hsu, K.J., cd., M ountain Building
Processes: London, Academic Press, p. 111-127.
Ramsay,J.G., and Huber, M.I., 1983, T he Techniques ofM od cm Structural
Geology, Volume I: Strain Analysts: London, .Academic Press, 307 p.
------ 1987, The techniques o f M odem Structural Geology, Volume 2: Folds
and Fractures: London, Academic Press, 700 p.
Ramsay, J.G., and Wood, D.S., 1973, The geometric effects o f volume diangc
during dcfonnational processes: Tcctonophysics, v. 16, p. 263-277.
Rickard, M.J., and Rixon, L.K., 1983, Stress configurations in conjugate
quartz-vein arrays: journal o f Structural Geology, v. 5, p. 573-578.
Ridley, J., and Mcngtcr, F., 2000, Lithological and structural controls on
the form and setting o f vein stockwot k orcbodies at the Mount Char­
lotte gold deposit, Kalgoorlic: Economic Geology, v. 95, p. 85-98.
Riverin, G., Bernard, D„ and Body. B„ 1990, The Donaida gold deposit,
Rouyn-Noranda, Quebec, in Rive, M., Verpaelst, P., Gagnon, Y., Lulin,
J.M., Riverin, G., and Simard, A., eds.. T he Northwestern Quebec poly­
metallic belt: A summaiy o f 60 years o f mining and exploration: Cana­
dian Institute o f Mining and Metallurgy’, Special Volume 43, p. 199-209.
Rol>ert, F., 1990, Structural setting and control o f gold-quartz veins the Val
d ‘Or area, southeastern Abitibi Subprovince, in Mo, S.E., Robert, F„ and
Groves, D.I., eds., Gold and base metal mineralization in the Abitibi
Subprovince, Canada, with emphasis on the Quebec segment: Perth,
University o f Western Australia, G eology Key Centre and University
Extension, Publication 24, p. 16-1-209.
------ 1994, Vein fields in gold districts: The exam ple o f Val d ’Or, south­
eastern Abitibi subprovince, Quebec: Geological Survey o f Canada, Cur­
rent Research 1994-C, p. 305-302,
— 1996, A pre-2686 Ma intrusion-related gold deposit at die Kicna mine,
Val d'Or, Q uebec, southern Abitibi subprovince—A discussion: Eco­
nomic Geology, v. 91, p. 803-806.
------1997, A preliminary geological mode! for syenite-associated dissemi­
nated gold deposits in the Abitibi belt, Ontario and Quebec: Geological
Survey o f Canada, Current Research 1997-C, p. 201-210.
Robert, F„ and Brown. /VC., 1986, Archcan gold-bearing quartz veins at die
Sigma mine, Abitibi greenstone bell, Quebec. Part I. Geologic relations
and formation of the vein system: Economic Geology, v. 81, p. 578-592.
Robert. F.. and Poulsen, K.H., 1997, World-class Archaean gold deposits in
Canada: An overview: Australian Journal o f Earth Sciences, v. 44.
p. 329-351.
Robert. F., Poulsen, K.H., and Dube, B., 1994, Structural analysis oflodc
gold deposits in deformed tcrrancs: Geological Survey o f Canada, Open
File Report 2850,140 p.
Robert, F., Boullier, A.M., and Firdaous, K., 1995, Gold-quam veins in
mctamorphic tcrrancs and their bearing on the role o f fluids in fault­
ing: journal o f Geophysical Research, v. 100, B7, p. 12,861-12,879.
Same, P., 1985, Geologic <lc la mine Louvicourt Goldfield, region dcVal D’Or.
Ministcre de r£nergie ct des Rcssources de Quebec, MB 85-43,28 p.
Scholz, C.H., 1989. Mechanics o f faulting: Annual Review o f Earth and
Planetary Sciences, v. 7, p. 309-334.
Sibson, R.H., 1986, Brecciation processes in fault zones Inferences from
earthquake rupturing: Pure and Applied Geophysics, v. 124, p. 159-175.
------1990a, Faulting and fluid flow, in Nesbitt, B.E., cd., Short course on
fluids in tectonically active portions o f the crust: Mineralogical Associa­
tion o f Canada Short Course Handbook, v. 18. p. 93-129.
------1990b, Conditions for fault-valve behavior, in Knipc. R.J.. and Rutter,
E.H , cd s, Deformation mechanisms, rheology, and tectonics: Geologi­
cal Society, Special Paper 54, p. 15-28.
----- 2001, Scismogenic framework for hydrothermal transport and ore
deposition: Reviews in Economic Geology, v. 14, p. 25-50.
Sibson, R.H., Robert, F., and Poulsen, K.H.. 1988, High-angle reverse
faults, fluid-pressure cycling, and mesothermal gold-quart2 deposits:
Geology, v. 16, p. 551-555.
Silliioe, R.H., and Thompson, 1998, Intrusion-related vein gold
deposits: Types, icctono-magmatic settings and difficulties o f distinction
from orogcnic gold deposits: Resource Geology, v. 48. p. 237-250.
■Stauffer, M.R., 1990, The Missi Formation: An Aphcbian motassc deposit
In the Reindeer Lake Zone o f the Trans-Hudson Orogen, Canada, in
Lewry.J.F., and Stauffer, M.R., cds„ The Early Proterozoic Trans-Hudson
Orogen in North America: Geological Association o f Canada, Special
Paper 37, p. 121-141.
Swager. C.P., 1997, Tectono-straiigraphy of Late Archean greenstone ter-
ranes in the southern Eastern Goldfields, Western Australia: Precam-
brian Research, v. 83, p. 11—^42.
Swager, C.P., and Griffin TJ., 1990, An early thrust duplex in the Kalgoor-
lic-Kambalda greenstone belt. Eastern Goldfields Province, Western
Australia: Precambrian Research, v. 48, p. 63-73,
Talbot, C.J., 1970, The minimum strain ellipsoid using deformed quartz
veins: Tectonophysics, v. 9, p. 47-76.
------1982, Obliquely foliated dikes as deformed incompetent single layers:
Geological Society of America Bulletin, v. 93, p. 450-460.
Talbot, C.J., and Soukoutis, D., 1992, The importance o f incompetence:
Geology, v. 20, p. 951-953.
Thomas, D.J., and Hcaman, L.M., 1994, Geological setting o f the Jolu
Gold Mine, Saskatchewan: U-Pb age constraints on plutonism, defor­
mation, mineralization, and metamorphism: Economic Geology, v. 89,
p. 1017-1029.
Tourignv, G., Hubert, C., Brown, A.C., and Crepcau, R., 1989, Structural
controls o f gold mineralization at the Bousquet mine, Ai^itlbi, Quebec:
Canadian journal o f Earth Sciences, v. 26, p. 157-175.
Tourigny, G., Doucet, D., and Bourget, A., 1993, Geology o f the Bousquet
#2 mine: An example o f a deformed, gold-bearing, polymetallic sul­
phide deposit: Economic Geology, v. 88, p. 1578-1597.
Treagus,S.II., 1983. A theory o f finite strain variation through contrasting
layers, and its bearing on cleavage refraction:Journal o f Structural Geol­
ogy, v. 5. p. 351-368.
Vcamcombe. J.R., 1993, Quartz vein moiphology and implications for for­
mation depth and classification o f Archean gold-vein deposits: Ore
Geology Reviews, v. 8, p. 407-424.
— 1998. Shear zones, fault networks, and Archean gold: Geology, v. 26,
p .855-858.
Vcamcombe, J.R , Barley, M.E., Eiscnlohr, B.N., Groves, D.I., Hotistoun,
S.M.. Skwamccki, M.S., Grigson, M.W., and Partington, C A . 1989, Struc­
tural controls on mesothermal gold mineralization: Examples from the
Archean terranes o f Southern Africa and Western Australia, in Keays,
R.R., Ramsay, W.R.H., and Groves, D.I., eds.. The Geology o f Gold
Deposits: The Perspective in 1988: Economic Geology Monograph 6, p.
124-134.
White, W.H., 1943, The mechanisms and environment o f gold deposition
in veins: Economic Geology, v. 38. p. 512—532.
Wyman, D.A.. and Kernel]. R.. 1989, Archean shoshonilic lamprophyres
associated with $ti|>crior Province gold deposits: Distribution, tectonic
setting, noble metal abundance, and significance for gold mincraUzar
tion, in Keys. R.R., Ramsay, W.R.H.. and Groves, D.I., ed s. The Geology
o f Gold De|>osits: The Perspective in 1988: Economic Geology Mono­
graph 6, p. 651-667.
Society of Economic Geologists
Reviews v. 14, 200i, p. 15 7 - 18 1

Chapter 6

Magmatic and Structural Controls on the Developm ent o f

Porphyry Cu ± Mo ± Au Deposits
R. M . TOSDALt
Mineral Deposit Research Unit, University o f British Columbia, Vancouver, BC V6T17A, Canada

andJ . p. R ic h a r d s
Department o f Earth an d Atmospheric Sciences, University o f Alberta, Edmonton, AB T6G 2E3, Canada

Abstract

Porphyry Cu ± Mo ± Au deposits require the coincidence and positive interaction of a series o f individ­
ually commonplace geological processes. They, and all their genetically associated deposits, are a natural
consequence of convergent margin magniatism, and reflect the dynamic interplay between magmatic, hy­
drothermal, and tectonic processes. Magmas generated during subduclion rise into the upper crust, com­
monly along zones of lithospheric weakness, where they pond in tabular magma chambers at depths of 6
km or deeper. The chambers grow laterally by chamber floor depression (cantilever mechanism) and
some roof lifting (piston mechanism). Apophyses rise from the parental magma chamber and intrude to
within 1 to 3 km of die surface, where they may undergo volatile exsolution and crystallization as por-
phyritic stocks. Emplacement o f porphyry stocks is facilitated by structural anisotropy in the roof rocks. As­
cending hydrothermal fluids exsolved from the porphyry stocks and the underlying parental magma
chamber are focused into the cupola, taking advantage of vertical structural and rheological anisotropies
introduced either before o r during porphyry emplacement.
From a structural standpoint, three recurrent processes enhance permeability in die form of fracture or
breccia networks through which hydrothermal fluids pass and precipitate minerals. Fracture-producing
events are related to intrusion o f pre-, syn-, and post-mineral porphyry stocks or dikes to near-surface depths
(1-3 km), phase separation and volume expansion o f a hydrothermal fluid through a variety of mecha­
nisms, and tectonically induced failure. Concentric and radial fracture patterns reflect magmatic processes
whereas more linear arrays of veins reflect tectonic influences. The resulting different vein arrays are com­
monly vertically and temporally distributed in the porphyry system; concentric and radial arrays are more
common above o r in the upper parts o f the stocks, whereas linear arrays dominate at depth, forming as the
system cools and the pluton solidifies. Orthogonal and conjugate arrays o f veins characterize all scales and
all parts o f porphyry systems. Veins from a particular paragenetic stage do not have unique orientations, but
rather occur with all orientations typical o f that system. T he common conjugate to orthogonal intcr-vcin re­
lationships in porphyry Cu deposits requires repetitive exchange of principal stress orientations, events that
are facilitated by conditions of low differential horizontal stress. Such stress conditions indicate that many
porphyry Cu deposits form in specific environments where the magmatic arc is under a near-neutral stress
state. These conditions occur cither in areas removed from active deformation, or during periods of stress
relaxation and low strain in the magmatic arc. Achievement o f these conditions in time and space is likely
to be infrequent and transitory during the life of a convergent margin, which may explain the spatial and
temporal clustering o f deposits in large porphyry districts.

Introduction tion o f these deposits from a physical and geochemical

Porphyry C u ± Mo ± Au deposits, hereafter referred to as
porphyry Cu deposits, are associated with shallow-level plu-
tonic complexes emplaced in magmatic arcs at convergent
plate margins (Figs. 1, 2A; Sillitoe, 1972,1997; Sutherland
Brown, 1976; Tidey, 1982; Sawkins, 1990). Since their first
recognition as im portant sources of metals (Ransome,
1904, 1919; Lindgren, 1905, 1933, 19S7; Emmons, 1927),
much work has been devoted to understanding the forma-
'■ ^C orresponding a uthor: e-maii, rtosdal@ cos.ubc.ca
'fy
standpoint (Burnham , 1967,1979; Gustafson and H unt,
1975; Cathles, 1977; Burnham and Ohmoto, 1980; Beane
and Titley, 1981; Candela, 1989; Cline and Bodnar, ,1991;
Hedenquist and Richards, 1998). In addition, the question
o f why there is considerable variation in the size o f these
types o f deposits has been pondered (Clark, 1993). Other
studies have focused on structural controls on vein and
fracture distributions within porphyry Cu deposits, and
their spatial and paragenetic relationships to porphyry
stocks (R ehrigand Heidrick, 1972, 1976; Gustafson and
Hum, 1975; Tidey and Heidrick, 1978; Heidrick and Titley,
 Fig. 1. Distribution o f Mesozoic and Ccnozoic porphyry Cu ± Mo ± Au deposits along convergent plate margins in the
circum-Pacific region. PNG = Papua New Guinea. Adapted from Thom pson (1995).
1982; Wilkinson et al., 1982; Lindsay et al., 1995). These
studies dem onstrate that porphyry Cu deposits are influ­
enced not only by magmatic and hydrothermal processes,
but also by regional tectonics at the time of formation. In
fact, the prevailing tectonic setting and structural frame­
work may strongly influence not only die size and form of a
porphyry system (Sillitoe, 1994,1997; Skewes and Stern,
1995) but also its location (Rehrig and Heidrick, 1972,
1976; Titley and Heidrick, 1978; Heidrick and Titley, 1982;
Richards etal., 2001).
This chapter reviews the interplay between magmatism
and dynamic structural environments in the development
of porphyry Cu deposits. Critical to this goal is the general
model of porphyry Cu deposits, which highlights the need
for consu'uctive interaction between magmatic, magmatic-
hydrothermal, and tectonic events in the formation of por­
phyry Cu deposits. It is equally im portant to consider
em placem ent mechanisms of shallow-level plutonic com­
plexes in the upper crust because porphyry Cu deposits are
restricted to these environments. Throughout this chapter,
we draw liberally on well-described examples of porphyry
Cu deposits to illustrate various points. Most examples are
from the southwestern United States and from the Andes of
Peru and Chile. This geographic focus is not intended to
indicate that other porphyry Cu provinces are insignificant,
but merely reflects past history of geologic investigations,
availability o f literature, and our personal experiences.
General Porphyry Model
Physicochemical model
Important characteristics of porphyry Cu deposits summa­
rized herein are based on the primary physiochemical stud­
ies of Undgren (1937), Hemley and Jones (1964), Sheppard
etal. (1969,1971), Lowell and Guilbert (1970), Rose (1970),
Roedder (1971), Sillitoe (1973), Gustafson and H unt
(1975), Hollister (1975), Sheppard and Gustafson (1976),
Sutherland Brown (1976), Eastoe (1978), and Reynolds and
Beane (1985), and reviews by Gustafson (1978), Beane and
Titley (1981), Tilley and Beane (1981), Titley (1982), Hunt
(1991), Kirkham and Sinclair (1996), and Hedenqtiist and
Richards (1998).
Porphyry-type systems are so-named because of their spa­
tial and genetic association with porphyritic intrusions (Fig.
2A). Commonly, such intrusions are characterized by phe-
nocryst assemblages consisting of one or more of tlte min­
erals quartz, K feldspar, plagioclase, hornblende, and biotite
(rarely pyroxenes and olivine) in a fine-grained matrix.
Rock compositions vary from granitic to dioritic o r rarely
gabbroic, and are o f calc-alkaline to alkaline parentage
(Brown, 1982). Parental magmas are believed to be moder­
ately water-rich as indicated by the presence of hydrous phe-
nocryst phases such as amphibole and biotite, whose stabil­
ity requires at least 3 wt percent HsO in the melt (Holland,
1972; Burnham, 1979). Water concentrations are probably
 Generalized porphyry-related base metal systems

F ig . 2. Anatomy o f a porphyry Cu ± Mo ± Au deposit and other spatially related

deposits. A. General model o f a zoned magmatic hydrothermal system showing poten­
tial linkages between a porphyry Cu deposit centered on a composite porphyry intru­
sion at depth (adapted from Siilitoe, 1993). The porphyry Cu deposit is Ranked out­
ward and upward by a characteristic suite o f deposit types. N ot all deposits in the
general model are present in every zoned magmatic-hydrothermal system. Peripheral
deposits dominate some systems such as Copper Canyon in the Battle Mountain Min­
ing District (Theodore, 2000), whereas in other systems such as those in southern
Peru and Chile, the porphyry Cu deposit dominates and few peripheral deposits arc
known (Gustafson and Hunt, 1975; Clark c t ah, 1990). B. Simplified distribution o f
hydrothermal alteration and sulfide minerals in an idealized porphyry Cu deposit
(adapted from Lowell and Guilbert, 1970). Abbreviations: AA = advanced argillic alter­
ation; adul « adtilaria; alb = albite; anh = anhydrite; bi = biotitc; carb = carbonate min­
erals; chi * chlorite; cp = chalcopyrite (and bornitc); epi = cpidotc; gal * galena; kaol
= kaolinite; K-fcld » K feldspar; mag = magnetite; mb * molybdenite; py = pyrite; qtz *
quartz; scr ■ sericitc; si = sphalerite; sulf. = sulfidation.

not higher than 5 wt percent, however, because of the com­
mon presence of plagioclase phenocrysts (Gill, 1981). Fluid
exsolution from the melt upon reaching saturation leads to
a sudden rise o f the magma solidus tem perature, which
quenches the matrix and gives rise to the porphyritic tex­
ture. This magmatic-derived hydrothermal fluid is, in turn,
implicated in the transport and deposition o f metals to form
porphyry mineral deposits.
The hydrous nature o f the melt derives from magmagene-
sis in an arc environment where water is carried down into
the upper mantle by subduction of altered oceanic crust
(Fig. 3). Slab dehydration during descent to -100 km depth
metasomatizes the overlying mantle wedge, thereby generat­
ing conditions suitable for extraction of hydrous, basaltic
magmas by decompression melting (Hamilton, 1981; Arcu-
lus, 1994). As these magmas ascend toward the surface, a
combination of processes, including depressurization and
crystallization of anhydrous mineral phases, leads to a
decrease in volatile solubility and an increase in the volatile
content of die residual silicate melt, eventually leading to sat­
uration (Burnham, 1967,1979; Whitney, 1975). At this point,
a separate voladlc phase is exsolved, which becomes concen­
trated at die top of the magma chamber (Figs. 4 and 5A). In
calc-alkaline to alkaline arc magmas, this water-rich volatile
phase will also contain other water-soluble volatile compo­
nents such as Cl and S species. D epending on the depth
(pressure) of emplacement, initial voladle exsolution yields
either (1) a high-pressure, hom ogeneous, supercritical,
aqueous fluid with salinity in the range 2 to 10 wt percent
NaCl equiv, o r (2) at shallower depths (above approximately
5 km) two immiscible fluids, consisting of a low-salinity vapor
phase and a high-salinity brine (Burnham, 1979; Cline and
Bodnar, 1991; Shinohara, 1994; Gustafson and Quiroga,
1995; Hcdenquist and Richards, 1998; Hedenquist et al.,
1998). These fluid phases can transport base and precious
metals as chloride and possibly bisulfide complexes (Candela
and Holland, 1984,1986; Williams et al., 1995; Gammons
and Williams-Jones, 1997; Loucks and Mavrogenes, 1999).
 Three Dimensional Geometry of Magmatic Arcs
fixed
Upper Crust (UC): Strike-slip partitioning I Rising magma in
and distributed shearing T dikes / shear zones
Lower Crust (LC): Distributed shearing
Upper Crust (UM): Localized shearing
FMS: Forearc Mantle Sliver
Fie;. 3. Geometry o f a magmatic arc formed at an obliquely convergent
plate margin. Hydrous magmas arc generated in the mantle wedge and
intrude the lithosphere along broad deformation zones undergoing varying
amounts o f strike-slip motion depending upon the degree of obliquity along
the convergent margin. Modified from dc Saint Blanquat etal. (1998).
Exsolution of a volatile phase from magma necessarily
involves a large positive change in volume (Phillips, 1973;
Burnham, 1979; Burnham and Ohmoto, 1980). Depending
on the stress regime and the structure and permeability of
die confining host rocks, this volume change may be accom­
modated by pervasive brecciation of the volatile-saturated
carapace of the intrusion, or direct escape toward the surface
along fault zones or breccia pipes (Fig. 4). Both mechanisms
involve creation of fracture permeability and porosity into
which ore minerals may subsequently be precipitated,
thereby leading to formation of porphyry deposits.
Hydrothermal fluid exsolved at magmatic temperatures is
initially in equilibrium with the coexisting magma and its
mineral phases. High-temperature alteration assemblages
(Fig. 2B), therefore, consist of minerals such as quartz, K
feldspar, biotite, anhydrite, and magnetite (the “potassic”
alteration assemblage; Lowell and Guilbert, 1970; Gustafson
and Hunt, 1975). Base metal sulfide minerals, typically com­
binations o f chalcopyrite, bornite, and molybdenite with
pyritc (Fig. 2B), are deposited with these alteration minerals
as disseminations and in thin, commonly wavy or ductilely
deformed, wispy vcinlets that lack parallel vein walls. These
early veins were term ed “A veins’’ by Gustafson and H unt
(1975), based on their study o f the El Salvador porphyry
deposit in Chile. These veins are usually cut by intramineral
porphyry stocks where such stocks are present. Marginally
later “B veins” at El Salvador are quartz-rich with more abun­
dant m olybdenite and chalcopyrite, and straighter vein
walls; these veins cut all but post-mineral stocks. The
straighter vein walls and mineral textures within these veins
indicate formation at lower temperatures than the A veins.
lilgh-sulfulailon
alteration
Fig . 4. Fluid circulation patients in a porphyry copper deposit form­
ing in the cupola o f a stock at depths in the range o f 1 to S km. Exsolu­
tion and trapping o f magmatic fluids in a zone o f lithostatic fluid pres­
sure (P r) beneath an isotherm o f -3 7 0 * to 400®C, which marks the
brittle-ductile transition. W ithin this zone, brine and “steam" interact
with the country rocks form ing the potassic alteration stage. Above the
transition zone, circulating fluids u n der hydrostatic pressure alter the
country rocks to propyiitic mineral assemblages. In an unperturbed and
cooling porphyry Cu deposit, the brittle-ductile transition gradually col­
lapses on to th e early potassic alteration zone, leading to phyiiic over­
printing and the influx o f m eteoric fluids in to the system. If th e brit­
tle-d u c tile transition is suddenly perturbed or breached because o f
tectonism or later magmatism, the fluid ltd ruptures, leading to sudden
transition from lithostatic to hydrostatic fluid pressures (see left side o f
stock o n diagram ). T he fluid-saturated carapace un dergoes sudden
phase separation lead in g to “steam" plus brine plus crystals, with the
potential expulsion o f metal-bearing fluid upward into the epithermal
environm ent. W ithin the porphyry Cu environ m en t, th ese explosive
pressure transitions trigger the form ation o f hydrotherm al breccias.
Adapted from Fournier, 1999.
As the high-tem perature fluid migrates away from the
parent magma, it rises owing to buoyancy and begins to
cool. Two processes may happen, as follows:
1. If the fluid was exsolved originally its a homogeneous
supercritical phase, then upon cooling it will likely intersect
its solvus and separate into liquid and vapor phases. This
phase separation event involves a further increase in vol­
ume leading to additional fracture permeability and brec­
ciation, most commonly in die fluid-soaked carapace above
or at the top of die porphyry stock. Phase separation may
also occur during sudden pressure drops such as massive
hydrofracturing, sudden removal o f lithostatic load, o r rup­
turing o f the brittle-ductile transition zone separating the
deeper lithostatic from the overlying hydrostatic environ­
m ent (Fig. 4; Fournier, 1999).
2. Sulfur originally dissolved in the high-tem perature
fluid as SOs will react with water and dispropordonate to
form H3S and sulfuric acid (Burnham, 1979).
 N S

Domes
Domes
«. Palcosurface_

Sedimentary and volcanic rocks (Tri. and Jur.)

It Volcanics of Artesia Lake (Jur.) Granite porphyry dikes
'// a Quartzite (Jur.) 1 9 Luhr Hill porphyritic granite
. *. Siltstonc (Jur.) Border granite of Bear intrusion
1/• 1 Limestone (Tri.) Quartz monzonito of Bear intrusion
yVnv Sedimentary and volcanic rocks (Tri.) Quartz monzodiorite of McLeod Hill intrusion
Volcanic rocks (Tri.) V4 Shamrock batholith

Fig. 5. Relationship between porphyry Cu deposits, porphyry stocks, and dikes to underlying parental magma cham­
bers, which commonly lie at depths o f 6 km or greater. A. Evolutionary model for the exsolution o f high-density saline
aqueous fluid from tabular piutons and the upward escape o f this fluid into the roof zone where a porphyry Cu deposit
forms (Diltes, 1987). The saline aqueous fluid trapped along the roof o f a cooling batholith is channeled upward along :
dikes or stocks that emanate from the inwardly crystallizing magma. B. The Kelvin porphyry Cu prospect in Arizona is
associated with a small porphyry stock that emanates from die apical part o f the Tea Cup Granodioritc (modified from .
Cornwall and Kricgcr, 1975; Crcasey, 198-1). C. The Aim-Mason porphyry Cu deposit in the Yerington batholith is asso­
ciated with dikes emanating from successively deeper parts o f the Luhr Hill porphyritic granite (Dillcs, 1987). Jur. =
Jurassic; Tri. * Triassic.
 These processes generate a new set of veins and breccias
characterized by gangne and alteration minerals stable at
lower temperature and pH. In particular, an increase in activ­
ity of sulfate in the fluid owing to SOa disproportionation will
result in voluminous precipitation of anhydrite and further
release of more acid in the form of HC1 (Burnham, 1979). In
feldspar-rich host rocks, die characteristic alteration assem­
blage from these acidic fluids is fine-grained muscovite
(scricite) with quartz (“phyllic” alteration assemblage, Fig. 2B;
Lowell and Guilbert, 1970). In addition, sulfide minerals are
precipitated in response to the increase in H.»S activity. Asso­
ciated D veins (Gustafson and Hunt, 1975) are coarse-grained
and sharply defined with strong alteration haloes, and consist
of sulfide minerals (abundant pyrite, with lesser Cu Fe-sulfide
minerals, sulfosalt minerals, sphalerite, and galena), anhy­
drite, and minor quartz. Typically, die highest grades of hypo-
gene Cu ore are present dose to the boundary between die
potassic and phyllic alteration zones, again reflecting the
increased activity of H2S in the fluid as a result of SO., dispro­
portionation. The phyllic alteration zone is commonly
observed as a shell surrounding the potassic core of the por­
phyry system (Lowell and Guilbert, 1970), and may have a
diameter of several hundred meters to a few kilometers.
Separation of a vapor phase may lead to the formation of
a low-density vapor plume rich in acidic volatiles such as
S 0 2, HC1, HF, and COs (Fig. 4). Cooling and condensation
o f these vapors a t shallow depths causes intense acid leach­
ing (“advanced argillic” alteration; Fig. 4), characterized by
total destruction o f primary minerals to leave an insoluble
residue of quartz, clays (kaolinitc o r pyrophyllite), alu­
minum hydroxides (diaspore), and alunite. Such alteration
zones, being very porous and permeable, may serve as hosts
to high-sulfidation styles of epitherm al mineralization
(H edenquistetal., 1993; Hedenquist, 1995).
Emplacement of a hot body o f magma into the shallow
crust typically involves interaction with local ground waters
(Fig. 4). The degree of that interaction depends on crustal
permeability, but in many cases hydrothermal convection
cells driven by heat from the intrusion can be shown to
have operated over radii of many kilometers (Taylor, 1974;
Titley et al., 1986). The fluids involved are typically dilute
ground waters, and the alteration arising from their flow is
characterized by hydration, producing minerals such as
chlorite and epidote, commonly with carbonate minerals
(“propylitic” alteration; Fig. 2B).
Many variations upon the general porphyry model have
been recognized throughout the world, particularly in
term s of the sequence o f sulfide minerals deposited in
veins. For example, at Mineral Hill, Arizona, molybdenite is
the oldest sulfide mineral recognized and much o f the cop­
per sulfide mineralization is parageneticatly younger, gen­
erally associated with D-type veins (Wilkinson et al., 1982).
Nevertheless, the general sequence of alteration and sul­
fide mineral deposition described above characterizes the
majority o f porphyry Cu deposits, regardless of their size.
From a structural standpoint, three processes contribute
to fracturing during porphyry Cu form ation, and each
process enhances permeability in the form of fracture or
breccia networks through which hydrothermal fluids may
flow and precipitate minerals (Figs. 5-7): (1) intrusion of
pre-, syn-, and post-mineral porphyry stocks o r dikes, which
may be intruded to near-surface depths (1-3 km; Koide
and Bhattachariji, 1975); (2) exsolution of hydrothermal
fluid from a cooling parental magma chamber, with
upward focus into the solidified or partially solidified stocks
and warm wall and roof rocks (Fig. 5A); and (3) phase sep­
aration o f the hydrothermal fluid. Each event involves vol­
um e expansion and fracturing wherever the fluid pressure
exceeds the ambient minimum effective stress (cr3' = o 3- P,,
where Pf is the fluid pressure plus the tensile strength o f the
rock; Jaeger and Cook, 1979; Sibson, 2000). Multiple o r
recurrent pulses o f hydrothermal activity, which are com­
mon in porphyry systems, lead to repetition of fracturing
and superim position of veins and alteration, leading to
increased complexity.
Vein-dominated versus breccia-dominated porphyry Cu deposits
Porphyry Cu deposits developed within noncalcarcous
country rocks have a considerable range o f morphology
and show variations between vein-dominated and brcccia-
dom inated systems with considerable overlap between the
two. Vein-dominated porphyry Cu deposits represent a rel­
atively simple environment: veining events are consistent
with emplacement of a porphyry stock, exsolution of a mag­
matic-derived hydrothermal fluid with local influx of exter­
nal fluids, and cooling o f an evolving magmatic-driven
hydrothermal system. These deposits are the most common
variety of porphyry Cu deposits, with numerous well-docu­
mented examples (see volumes edited by Titley and Hicks,
1966; Sutherland Brown, 1976; Titley, 1982; Pierce and
Bolm, 1995; Schroeter, 1995).
The other end-m em ber is breccia-dominated (Sillitoe,
1985). Breccias in the porphyry Cu environments have a
range o f morphology and genesis, being form ed from
igneous, magmatic-hydrothermal, and phreatomagm atic
processes. The most common types of breccia are pre- to
syn-mineral bodies that are irregular to pipe-like in shape,
and are characterized by angular clasts, sharp or grada­
tional contacts against enclosing rocks, and an intimate
association with porphyry stocks o r dikes. The breccia
matrix consists o f igneous rocks o r hydrothermal silicate
and sulfide m inerals (Richard and Courtright, 1958;
Camus, 1975; Zweng and Clark, 1995). Post-mineral brec­
cias commonly but not always have a milled rock-flour
matrix and rounded clasts; they are referred to as pebble
breccias. These breccias essentially excise pans of a prior
hydrotherm al system (Richard and Courtright, 1958),
although locally, such as at Toquepala, the pebble breccias
are intram ineral and contain chalcopyrite in the matrix.
Zweng and Clark (1995) proposed that breccia-dominated
porphyry Cu deposits form at shallow levels based upon
fluid inclusion data from Toquepala. They suggested that
vein-dominated deposits would m ore commonly form at
slightly greater depths. Implied in their model is a primary
influence o f the lithostatic load o f the overlying column of
rock on breccia formation (see discussion below).
levels in th e crust. A. Com puter simulation o f growing tabular batholith fed from below (Australian Geodynamics
Research Group, 1997). Subhorizontal fracture and fault systems propagate outward from the expanding tips o f the tab­
ular pluton. Rising above the growing batholith arc stocks extending to shallow levels in the crust. T he surface o f the
Earth is displaced upward by the shallow stock, whose em placem ent develops concentric fracture and fault systems as
well as intense, steep fracture networks in the overlying volcanic carapace. B. A tabular pluton expands laterally and
grows dominantly by a cantilever mechanism. Lifting o f the pluton roof along steep dipping norma) faults accom m o­
dates die vertical expansion o f the stock. Upward expansion o f the pluton may be assisted by stoping o f roof rocks and
potentially by ballooning. R oof uplift is likely accompanied by extension, which leads to weak zones into which apophy­
ses or stocks may intrude. These apophyses focus the ascent o f exsolved magmatic-derived hydrothermal fluids, and sub­
sequent ore formation (see Fig. 5 for exam ples). PCD = porphyry Cu ± Mo ± Au deposit.
 A

Fig . 7. Simplified predicted stress trajectories above a rising, shallow-level stock where the stress state approaches uni­
axial extension above a cylindrical stock. In isotropic rocks, formation o f fractures, veins, and dikes would be guided by
the stress trajectories. In anisouopic rocks, older structural fabrics, such as faults, bedding, and folds, will strongly influ­
en ce the stress distribution, and favorably oriented planes o f weakness will be preferentially opened. In all eases, the max­
imum principal stress ( o {) is oriented vertically. Dikes and veins develop normal to the minimum principal stress (O,)
and in the plane containing o , and die intermediate principal stresses (o 2)- Other planes o f weakness can form conju*
gate systems o f shear fractures, distributed symmetrically distributed about a ,. A. Stress trajectories generated above a
domical slock. B, C. Map and cross-sectional views o f orientation o f stress for generation o f cone sheets or concentric
fractures. D. Map view o f stress trajectories for generation o f radial dikes and fractures. E. Cross-sectional view o f the ori­
entation o f shear fractures generated by a stress system above a domical stock. Modified from Park (1983).
 Convergent Plate Margin Magmatism
Porphyry Cu deposits are the e n d result o f a complex
sequence of events initiated by magmagenesis a t conver­
gent plate margins, and it is, therefore, useful to review the
process o f magma ascent and em placem ent in the lithos­
phere. Convergent margin magmatism is linked to subduc-
tion of an oceanic plate beneath an overriding continental
or oceanic plate, upon which the arc is constructed (Fig. 3).
In this environm ent, plate m otion vectors are commonly
oblique at varying angles to the plate margin, and stress
transmitted to the overriding plate, and the resulting defor­
mation, reflect the degree of obliquity. Strain is partitioned
into arc-normal (convergent o r divergent) and arc-parallel
(strike-slip) components. A wide range of structural styles
can, thus, be expected in the upper plate depending on the
degree o f convergence obliquity, convergence rate, and the
angle o f subduction (Fitch, 1972; Dewey, 1980; Jarrard,
1986). O f additional importance is whether trench retreat
or advance occurs at the plate margin (Ham ilton, 1988;
Royden, 1993). In the former situation (trench retreat), the
overall tectonic setting will be extensional, o r transten*
sional where a significant degree o f net convergence obliq­
uity is present. In the latter case (trench advance), contrac­
tion or transpression dominates. Regardless of convergence
geometry, some relationship between strike-slip tectonics
?and magmatism, including porphyry magmatism, is present
in convergent m argin arcs, although the relative impor­
tance of strike-slip motion will vary (e.g., Glazner, 1991; de
Saint Blanquat et al., 1998). The ubiquitous association of
plutonic and volcanic complexes and major fault zones in
arcs implies that rising magma concentrates in zones o f
deformation. A corollary to this observation is either that
deformation is crucial to the em placem ent o f magma, or
that magma assists the deform ation processes which, in
turn, helps its ascent to higher crustal levels (McCaffrey,
1992; de Saint Blanquat etal., 1998). Paterson and Schmidt
, (1999; see also Schmidt and Paterson, 2000) have argued
mat many plutonic complexes are not necessarily emplaced
, along faults but come to rest in the less deform ed rocks
between m ajor faults. However, their analysis applies to
contractional orogenic belts characterized by thrust faults,
rather than transpressional o r transtensional strike-slip
zones in volcanic arcs (Richards, 2000a).
Exactly how magma moves through the crust is poorly
understood. Nevertheless, there is general agreement that
'<the following factors are important (Hogan and Guilbert,
/ 1995; Hutton, 1997): (1) magmas, being fluid, will always
move down a hydrodynamic gradient toward regions o f low
pressure; (2) in the absence of significant magmatic over­
pressure, the density contrast between a granitic melt and
the surrounding rocks will drive the m elt upwards to an
“equilibrium ” level o f neutral buoyancy in the crust; (3)
increasing vapor pressure during late stages o f crystalliza-
lltion generally increases magma pressures; (4) the lithosta-
tic load of the overlying crust may act as a pressure seal hin­
dering the rise o f magma; and (5) temperature-dependent
viscous resistance to magma flow inhibits ascent Magmatic
overpressure may drive magmas beyond their level o f neu­
tral buoyancy, comm only resulting in violent eruptions.
Overpressured magmas may intrude by forcing open dikes
o r pipes, o r by penetrating old faults o r fractures (Bussell et
al., 1976; Delaney e t al., 1986; Ida, 1999). T he prevailing
crustal stress field controls the orientations o f dikes, favor­
ing intrusion along older faults oriented at high angles to
the least principal stress (<J3). Alternatively, tectonic stress
and resultant strain may episodically create reduced pres­
sure o r tensional zones in the crust, such as pull-apart zones
at fault intersections o r jogs, which may focus magma
ascent (Huuon, 1988; Tikoff and Teyssier, 1992; Grocott e t
al., 1994; Richards et al., 2001). Thus, the varied styles and
mechanisms o f magma em placem ent in the crust involve
an important dynamic interplay between tectonic and mag­
matic processes. T he final three-dimensional form o f the
stock is controlled by strain rate, magma supply, country
rock versus m agma rheologic differences, and tectonic
environm ent (Castro and Fernandez, 1998; Corriveau e t
al., 1998; Hogan et al., 1998; Paterson and Miller, 1998a, b;
Yoshinobu et al., 1998; Wilson and Grocott, 1999).
Growth of a magma chamber
Magmatism at the Earth's surface o r in near-surface envi­
ronm ents is a manifestation o f the relative effects o f tec­
tonic stress and magmatic overpressure. Whereas tectonic
or vertical stress acting on magma decreases with decreas­
ing depth, the effect of magmatic overpressure should
rem ain essentially unchanged because o f hydraulic con­
nectivity to depth. Hence, final emplacement of magma at
shallow crustal levels will, in large p a rt but n o t entirely,
reflect magma dynamics o r magma overpressure (Takada,
1994; H utton, 1997; d e Saint Blanquat e t al., 1998; Ida,
1999). C oherent m agm a cham bers recharge and grow
through the aggregation o f small pulses o f magma, perhaps
through the intrusion o f m ultiple dikes o r lobate bodies
(Petford et al., 1993,1994). Buoyant diapiric rise o f magma
o r lateral translation by pushing aside o f wall rocks helps
create room for an expanding pluton (Tikoff et al., 1999).
de Saint Blanquat e t al. (1998) proposed that magma
ascent into the crust stops where the magma becomes capa­
ble o f deform ing o r displacing its enclosing wall rocks,
thereby creating space for a growing or ballooning magma
chamber. Evidence for varying amounts o f wall-rock strain
around plutons at all depths of emplacement lends support
to this model.
Granitoids emplaced at high levels in the crust tend to
have tabular shapes with flat tops and bottoms and steep
sides, and lateral dim ensions that are m uch larger than
their vertical dimensions (Fig. 5B, C; Hamilton and Myers,
1967; Dilles, 1987; McCaffrey and Petford, 1997). Such
magma chambers grow above narrow vertical feeder zones
o r dikes and spread laterally as sill-like intrusions that
inflate by floor displacem ent o r by roof-lifting (Fig. 6;
Johnson and Pollard, 1973; Pollard and Johnson, 1973;
Corry, 1988). Floor depression leading to lopolithic com­
plexes seems to be the most comm on form o f granitoid
emplacement except at high-crustal levels (<3 km depth),
 where roof lifting is im portant and laccoliths may form
(Corry, 1988). Mechanisms of lateral growth of a magma
chamber are poorly understood, but they involve relative
downward transport of the country rocks as the magma
expands laterally. Mechanisms include cauldron subsi­
dence (Myers, 1975), stoping (Paterson et at., 1996), fold­
ing (Hamilton and Myers, 1967), and downward transport
on fringing brittle or ductile shear zones along the mar­
gins of the pluton (Paterson et al., 1996; Cruden, 1998).
Lateral growth of a tabular body is aided by faults and frac­
ture networks that extend laterally from the edges of the
plutons (Fig. 6A; Australian Geodynamics Research
Group, 1997). Such subhorizontal fracture systems could
be ductile shear zones or brittle fault zones depending on
the pressure-tem perature conditions and, hence, depth.
Emplacement o f additional dikes, sills, or veins along the
leading fracture tip would facilitate the lateral growth of
the pluton.
Cruden (1998) evaluated various mechanisms for the
thickening of tabular granitoid bodies in the upper crust.
He proposed that tabular plutons result from piston or can­
tilever mechanisms, o r a combination of the two (Fig. 6B).
Both mechanisms result in plutons that have flat roofs.
Sleep faults developed in the roof o r floor o f the growing
magma cham ber would also assist in vertical cham ber
growth (Fig. 6B). If chamber growth is accompanied by an
upward expansion somewhat akin to ballooning, or by a
piston mechanism on inward dipping faults, extension of
the roof rocks would be expected. Lateral growth in such
circumstances would result in the outward development of
a sequence of steep faults that presumably nucleated at the
tip of the expanding chamber. Alternatively, the steep faults
may nucleate along older planes of weakness such as intru­
sive contacts or older faults.
Plutons emplaced by a piston mechanism will have flat
floors and steep sides that correspond to britde o r ductile
faults or to feeder dikes. The bell-jar stocks in the Peruvian
Coastal batholith are excellent examples of this mechanism
(Myers, 1975). Em placem ent by a cantilever mechanism
results in displacement o f the floor of the chamber down­
a- ward, along steeply o r shallowly dipping shear zones, lead­
s-
ing to an inward sloping floor to the magma chamber.
Combinations of the two mechanisms are likely, and should
be expected for cham bers emplaced in the brittle upper
crust a t depths near or below ~3 km. Such depths are
believed to be typical for the parental magma cham bers
that source the shallower-level porphyry-related stocks and
dikes (Fig. 5B, C). Steep pluton margins are commonly
controlled by regional fracture sets, active faults, o r older
fault systems (Bussell, 1976; Dehlsetal., 1998). Growth and
replenishment of the chamber is via feeder dikes (Petford,
1996), which also act to recharge the magma system with
fresh magma. Each influx o f new magma could be o f dif­
ferent composition, and may carry additional volatiles
(Folch and Marti, 1998). This process leads to evolutionary
P :i changes in chemistry of the magma body (Matthews et al.,
l degree of structural anisotropy in the immediate area o f the
* 1995; Keith et al., 1998) that perhaps contribute to por­
I phyry Cu formation (Clark, 1993).
IU
11
I:5: ^j :
Volcanoes and Shallow-Level Intrusive Complexes:
The Environment o f Porphyry Cu Deposits
In the volcanic and nearaurface subvolcanic environ­
ment (<3 km depth) where porphyry Cu deposits form, the
free boundary provided by the Earth’s surface is a critical
constraint on magma emplacement, growth, and eruption
(Nakamura, 1977; Takada, 1994). The outward and upward
pressure exerted by" a rising magma cham ber lifts and
extends die roof and adjacent walls of the stock, leading to
characteristic patterns of brittle deformation in the overly­
ing carapace, inflation of wall rocks, and local emanation of
radial dikes or cone sheets from the central stock (Figs. 6
and 7; Koicle and Bhattachariji, 1975; Park, 1983). Fractur­
ing is most intense across the roof of the stock. Most of die
roof uplift is accom plished along faults that root in the
edges o f the stock, and outwardly progressing roof fault sys­
tems are expected above a series o f composite subvolcanic
stocks. The faults and fracture networks provide channel-
ways for intrusion o f later stocks and dikes, as well as
hydrothermal fluids (Fig. 5A).
Within this near-surface environment, several factors gov­
ern growth and final form o f a magma cham ber and the
overlying volcanic edifice if present (Nakamura, 1977),
and, by extension, the characteristics o f associated mag­
matic-hydrothermal mineral deposits. Volcanoes and sub­
volcanic stocks and dikes are commonly emplaced along
fault zones. These fault zones may have formed contempo­
raneously with magmatism, or may have been older struc­
tures rcactived during magmatism (Bussell, 1976; Shaw,
1980). W hether o r not these fault zones were active during
magmatism depends upon the regional tectonic stress field,
strain developed as a consequence of this stress, and mag­
matic pressure. The tectonic stress regime controls which
fault orientations m ight be dilatant, as well as whether or
not original formation of the structures was contem pora­
neous with magmatism. Favorably oriented faults provide
dilational channels perm itting em placem ent o f magma
and fluids (Sibson, 1985, 2000). Takada (1994) proposed
that the final form of volcanoes and their shallow-level
intrusive complexes reflects the degree o f magma supply
coupled with the differential horizontal stress o r strain rate
(Fig. 8). Areas of small differential horizontal stress and
high magma supply rate are conducive to the formation of
large plutonic complexes and overlying polygenedc strato-
volcanoes. In contrast, regions with a similar magma supply
b u t large differential horizontal stress (and correspond­
ingly higher strain rate) are characterized by num erous
largely monogenetic volcanoes and small, more dispersed,
subvolcanic plutonic complexes.
Subvolcanic plutons range from circular to elliptical in
map view. T he orientation o f the stock’s ellipticity and any
associated dikes will be at high angles to the orientation of
the finite extension direction and minimum principal stress,
whether that is a regional tectonic-induced o r local pluton-
induccd stress. T he exact angle depends gready on the
complex. W here the plutonic complex is em placed in
 Effect of differential horizontal stress and strain rate
on volcano-plutonic complex development
differential
stress

Fig. 8. Effect o f differential horizontal stress and magma supply on distribution o f volcanoes and subvolcanic intru­
sions; modified from Takada (1094). A large, near-surface, degassing magma chamber is required for the formation o f
a porphyry Cu deposit. Formation o f such magma chambers in the upper crust is seemingly favored by large to probably
intermediate magma input rates, coupled with small to intermediate differential horizontal stress and strain rates.

isotropic or little-faulted rocks, the long axis o f the complex
may approach orthogonality to o 3 depending upon the
magnitude and style o f deformation. However, where the
complex is emplaced into highly faulted crust, its long axis
could be at an oblique angle to o3, because favorably ori­
ented faults and o th er structural fabrics will be exploited
preferentially. T he ellipticity o f the stock also gives some
indication o f the differential horizontal stress and strain rate
during magmatism (Nakamura, 1977; Takada, 1994). For
example, where radial dikes are well developed, a uniform
horizontal stress field during magmatism is indicated (Fig.
9). This situation seems to dominate the near-vent areas of
many volcanoes (Nakamura, 1977), and implies that associ­
a te d subvolcanic stocks will be circular to slightly elliptical in
■;map view. In contrast, where significant differential hori­
zontal stress o r high strain rate is established, radial dike
, swarms characterize only the core o f the magmatic system
’ where magmatic pressure exceeds the tectonic stress (Fig.
9). Away from the volcanic core, dikes, peripheral intru­
sions, satellite volcanoes, and axial grabens develop along
fault systems that are aligned at high angles to the minimum
principal stress. An im portant part o f this transition from
the volcanic core to peripheral environments is the effect of
temperature. Cool rocks away from the magmatic center will
fracture under the influence o f a regional stress field,
whereas near the warm core of the magmatic system, inter­
action between regional and magmatic pressures will con­
trol the final form o f the volcano, subvolcanic stocks, and, at
still greater depths, the parental plutonic body.
O f additional im portance to the final form o f shallow
plutonic complexes is the interplay between the cooling,
solidified edges o f the chamber and the dynamic, partially
molten, and potentially recharged core. The importance o f
this interplay extends to the country rocks where the
mechanical contrast between a cooling pluton and its host
rocks governs the spatial development o f faults, fractures,
and shear zones, which may in turn control the locations o f
later intrusions and hydrothermal systems (Gow and Ord,
1999). Where a strong rheological contrast is present, such
as where plutons intrude sedim entary rocks, faults may
form along margins of the central stock, thereby permitting
access o f late magma and fluid to the country rocks. W here
there is a weaker contrast, however, such as where plutons
intrude volcanic rocks, deformation is most likely concen­
trated within the stock, thereby enhancing episodic mag­
matism and hydrotherm al circulation. In addition, the
num erical models o f Gow and O rd (1999) indicate that
conjugate zones o f volumetric dilation around the pluton
refract at the pluton margins, and do not reach its center.
The greatest dilatancy, therefore, occurs at the intersection
of these conjugate zones and above the pluton. Gow and
Ord (1999) further suggested that if horizontal compres­
sive stress is sufficient, fluid or magma will be drawn up out
of the stock and into the dilatant zone in the carapace.
External fluids may also be drawn into this zone leading to
fluid mixing, a prediction consistent with evidence for
influx o f external fluids during potassic alteradon in some
deposits (e.g., Dilles et al., 1995; Bouse et al., 1999).
 Cross-sectional views of conduits of subvoicanic plutonic and dike complexes

Uniform or isotivpic regional stress field Differential horizontal or anisotropic stress field
Fic. 9. Radial dike distribution, polygcnctic volcano morphology, and underlying map view o f plutonic complexes in
different idealized stress regimes; modified from Nakamura (1977, and references therein). A. Uniformly distributed
radial dikes, subcircular volcanic edifice, and underlying subvoicanic stock(s) form in an isotropic stress field where the
horizontal stresses at e subcqual (c.g., a 2 3). B. Asymmetric dike distribution and cutviplanar dikes developed under
an anisotropic stress field where a differential horizontal stress is extant (e.g., Oi> > 0 3). The volcano and underlying
stocks would have strongly elliptical oudincs in map view with the long axis at high angles to c s.

Where older structural anisotropy is present, faults may
be reactivated and may guide magma from the underlying
source chamber into the shallow crust o r volcanic environ­
m e n t Reactivation o f old structures is a common them e in
many porphyry Cu-rich regions such as Arizona (Heidrick
and Titley, 1982) and the Central Andes (Salfity, 1985;
Richards, 2000b; Richards et al., 2001). O n a regional scale,
reactivation o f old fault and fracture systems is enhanced if
they are oriented favorably with respect to the am bient
stress field. Furtherm ore, intersections o f such structures
may provide focal points for m agma em placem ent and
potential porphyry Cu generation if the conditions are
appropriate (Fig. 10; Heidrick and Tidcy, 1982; Richards,
2000b; Richards e t al., 2001). O n a local scale, however,
structures may be reactivated during magmausm and
hydrothermal circulation, even where they are not ideally
oriented with respect to the regional stress field, implying
that magmatic o r hydrotherm al fluid pressures locally
exceed tectonic stresses.
Examples of misoriented structural fabrics that have been
reactivated during intrusion and porphyry Cu formation
occur in Arizona (Heidrick and Tidey, 1982), theMaricunga
belt of Chile (e.g., compare convergence geometry oudined
by Pardo-Casas and Molnar, 1987, with vein orientations
described by Vila etal., 1991), and at El Salvador, Chile. The
alignm ent o f Eocene subvoicanic stocks along northeast-
striking Paleocene faults around the El Salvador porphyry
provides a particularly well-documented example of this
phenomenon (Fig. 11; Gustafson and Hunt, 1975; Cornejo
et al., 1997). During the Eocene, this region experienced
transpressive deformation, with the maximum compressive
stress oriented northwest-southeast, at high angles to north-
east-striking Paleocene faults that formed as a result o f ear­
lier volcanism and tectonism (Tomlinson e t al., 1993;
Cornejo et al., 1997). Under the Eocene stress field, north­
east-striking faults would not have been under tension and
should not, therefore, have localized magma intrusion. Nev­
ertheless, Eocene stocks are aligned parallel to Paleocene
faults and fracture systems, beginning with the Cerro Pelado
and Old Camp complexes to the northeast, and progressing
southwestward to Cerro Indio Muerto (Fig. 11A). A weak
porphyry Cu system developed in the older and slightly
m ore silicic C erro Pelado and O ld Cam p complexes
(Gustafson and Hunt, 1975), but the major porphyry Cu sys­
tem formed in association with granodioritic stocks under­
lying Cerro Indio Muerto. The latter stocks are elliptical in
plan view with long axes striking northwest (Fig. I IB;
Gustafson and Hunt, 1975), at high angles to the northeast­
erly alignm ent o f stocks at the district scale. Late dacitlc
(latite o f Gustafson and Hunt, 1975) dikes also strike north­
west, thereby implying northeasterly extension, consistent
with em placem ent within the inferred Eocene regional
stress field (Cornejo e t al., 1997). Thus, although misori­
ented Paleocene faults broadly localized plutonism on a dis­
trict scale, the form o f these intrusions reflects the prevail­
ing Eocene stress field. Either the horizontal differential
su-ess was not great, or magmatic pressures were sufficiently
high such that the tectonic stresses acting to close the north-
 FlC». 10. Geological sketch map o f the Central Andes showing the locations o f major Cu and An deposits (from Silli*
toe, 1992) and their spatial relationships to the West Fissure Zone and its intersections with cross*orogen lineaments
(from Salfity, 1985). Modified from Richards (2000b).
 A Generalized geology o f Co. Indio Mueno. Chile

2
km______ |

Granodioritc porphyry
(Eocene)
Quanz rhyolite
porphyry (Eocene)
Rhyolite (Palcocene)

Rhyolite & flows

(Psueoccne)
Tuff (Palcocene)
Llama Fm.
(Cretaceous)
Sierra Fraga Fm
(Jurassic)

Simplified geology, 2400 m level.

El Salvador mine
B
Eocene
magmatic-hydrothermal
complex
Pebble dike
Dacite
-N- 'n A porphyry
jififl Hydrothermal
breccia
Ji
□ L porphyry
E3 K porphyry
n X porphyry
— Quartz porphyry

r n Llama Fm.
*— J (Cretaceous)

Fig. 11. Geology o f Ccrro Indio Muerto and El Salvador porphyry copper deposit, northern Chile. A. Ccrro Indio
Muerto consists o f a Palcocene silicic dom e com plex and tuffs that unconfonnably overlie deformed Mesozoic volcanic,
volcaniclastic, and sedimentary rocks. The com plex sits on the edge o f the Palcocene El Salvador trap-door caldera.
Caldera margin faults and svibparalld faults strike northeast. Palcocene rhyolite dikes, such as the linear on e immedi­
ately east o f Ccrro Riolita, iniimdccl along northeast-trending faults and fracture systems. The inferred orientation o f the
Eocene stress field would have favored dilation along northwest-trending structures, but the strong northeasterly align­
ment o f the Eocene intrusions suggests that their emplacement was nevertheless controlled by the older Palcocene struc­
tural fabric. Quartz rhyolite or quartz porphyry intrusions are about 43 Ma old, whereas the granodioritc porphyries are
•I I to 42 Ma. Modified from Gustafson and Hunt (1975) and Cornejo c ta l. (1997). B. Simplified geological map o f the
2400-m level o f the El Salvador m ine showing the distribution o f porphyry intrusions, listed in order o f decreasing age
from l»ottom to top in the explanation. N ote at this scale the northwesterly long axis o f the piutonic com plex, compared
with the regional northeast-alignment o f plutons; the majority o f the com plex consists o f the laie-mincral L porphyry.
The oldest stocks, the X and K porphyries, seem to be largely semi-circular intrusions in the lower right-hand portion o f
the map area. Post mineral dacitc dikes and genetically related pebble dikes have prominent northwesterly strikes. The
orientation o f the dikes and long axis o f the stocks is consistent with the Eocene regional stress field shown in A. Modi­
fies! from unpublished m apping by A naconda geologists and provided by G uillerm o Mucllar, Com pania del
Cobre-Chilc (CODELCO) at FJ Salvador.
east-trending structures were overcome by the force o f intru­
sion. Because no evidence is present for any explosive vol­
canic activity associated with the £ocene complexes, it is
more likely that the differential horizontal stress was low.
Structural Sequence in Porphyry Systems
Early veins in porphyry systems (A-veins o f Gustafson and
H unt, 1975) form a t near-magmatic tem peratures and
under lithostatic pressures within a ductile environment,
which is surrounded and overlain by a brittle region domi­
nated by hydrostatic pressures (Fig. 4; Gustafson and Hunt,
1975; M untean, 1998; Fournier, 1999). D epending upon
rheologic contrast, the transition between the brittle and
ductile zones may lie within tem porally associated por­
phyries o r in the country rocks. As the system evolves and
cools, later veins at the same depth form under hydrostatic
pressures and u nder brittle conditions. Within the ductile
region, fractures may form in response to sudden changes
in internal fluid pressure (e.g., owing to chamber recharge
by fresh, volatile-rich m agm a), rapid strain rates that
exceed the capacity o f the rock to deform plastically, or rup­
turing o f the pressure seal leading to a sudden transition
from lithostatic to hydrostatic pressure conditions. In the
latter case, the volatile phase is likely to undergo phase sep­
aration and a large positive volume change, which may lead
to extensive hydrofracturing (crackle brecciation). As tem­
perature decreases, veins become m ore regular in orienta­
tion because they form by brittle failure under the influ­
ence of external tectonic stresses (Figs. 7 and 12). Fracture
formation and vein filling continues episodically so long as
;■sufficient hydrothermal fluid is channeled into the evolving
carapace from the tabular pluton at depth.
Variations in vein and mineralized fracture orientation
Comprehensive vein and mineralized-fracture analyses in
' porphyry Cu deposits are few. Those that have been under­
taken typically reveal a variation in orientation and style
; with inferred depth o f form ation of the porphyry Cu
deposit that is consistent with a magmatic and structural
framework of shallow subvolcanic stocks (Figs. 7 and 12;
Heidrick and Titley, 1982). At shallow depths, concentric or
' quasi-concentric and radial fracture patterns characterize
some deposits (Langerfeld, 1964; Baumer and Fraser, 1975;
Corn, 1975; Dunn, 1982; Heidrick and Titley, 1982). For
example, at the San Juan deposit, Arizona, concentric and
radial fractures and veins are centered around one lobe of
/ a stock, but dikes that form an orthogonal pattern are off-
: set from the center of the fracture and vein array (Fig. 12A;
Heidrick and Titley, 1982). In the shallow levels of El Sal­
vador, two overlapping radial arrays of post-mineral pebble
dikes, veins, hydrothermal breccias, and fractures reflect
em placem ent o f two apophyses or stocks at depth (Fig.
12B; Langerfeld, 1964). It is not clear at what point during
the life of the El Salvador porphyry system these fractures
formed, although they are now filled by late-D veins and
! post-mineral pebble dikes (Langerfeld, 1964; Gustafson
and Hunt, 1975). Beneath the concentric pattern in the
shallower levels of the system, Gustafson and H unt (1975)
docum ented a crude radial and concentric o r elliptical
arrangem ent o f late-D veins (Fig. 12C). T he patterns are
broadly centered on two L-porphyry stocks, which are late
intra-mineral intrusions that largely postdate sulfide min­
eral deposition. T he northw estern L-porphyry stock
directly underlies the center o f one o f the radial vein and
dike arrays, whereas a composite stock o f granodiorite por­
phyries cored by another L porphyry underlies the radial
vein and dike array to the southeast T he association with
the L porphyry implies a genetic linkage, a n d th at the
radial fractures formed late in the deposit’s evolution. How­
ever, because the southeast array lies above a composite
stock o f X, K, and L porphyry, formation of that radial array
could conceivably have been initiated earlier. A northwest­
erly elongation to the overall pattern suggests that late-D
veins at depth reflect not only magmatic stress associated
with em placem ent o f the composite porphyry stocks, b u t
also the northeast-southwest-oriented minimum principal
stress associated with regional transpressional deformation.
At deep levels in porphyry systems it has been proposed
that a singular trend dominates the veins and fractures, as
exemplified by the simple pattern at Sierrita, Arizona (Fig.
12D; Heidrick and Titley, 1982; Titley e t al., 1986; Titley,
1993). Here, veins and fractures are parallel to quartz mon-
zoniie porphyry dikes emplaced into the host granitic rocks
as part of (he magmatic evolution of the system. The densest
concentration o f fractures is centered on the apophyses of
quartz monzonite porphyry, which also localized the Sierrita
and Esperanza porphyry Cu deposits. Support for this mode!
is found at El Salvador, where late pebble dikes exploit older
fracture networks that become progressively more linear and
aligned to a northwesterly strike at depth (compare Figs. 11B
and 12B, C). Strong linear arrays o f veins are common in
many deposits (Britten and Marr, 1995; Pierce and Bolm,
1995; Schroeter, 1995), which, according to the m odel o f
Heidrick and Titley (1982), should therefore represent the
deep levels of porphyry Cu systems.
As with many models of porphyry systems, there are excep­
tions to these rules. Linear vein arrays at the shallow Marte
porphyry Au deposit are clearly inconsistent with the model
of Heidrick and Titley (1982), because that deposit can be
reasonably assumed to have formed at depths <1 km (Vila et
al., 1991). Exceptions are also found in several other deposits,
where vein and fracture networks vary with depth, parage-
netic stage, and location in die deposit (e.g., Wilkinson et al.,
1982; Clark, 1990; Lindsay et al., 1995). At Chuquicamata, for
example, interpretation of die vein geometry is complicated
by the apparent superimposition of two discrete porphyry sys­
tems separated by as much as 3 m.y. (Reynolds et ai., 1998).
Nonetheless, copper-bearing veins associated with potassic
alteration generally trend northeast and indicate formation
under a regional stress field, in this case dextral strike-slip
along die Domeyko fault system. Earlier quartz-molybdenite
veins, in contrast, form a crude elliptical fracture pattern
around the porphyry complex. Quartz-sericite-sulfide and
enargite veins are superimposed on die potassic alteration at
Chuquicamata, and their emplacement was controlled by old
fractures and veins as well as die regional tectonic stress field
D Veins Rock units cud Q u a r tz g r a in p o rp h y ry
— M a jo r le n d l . v . ) Ig n e o u s b re c c ia ED K p o rp h y ry
S u b o rd in a te tre n d l ■■■■■! A p o rp h y ry EZ! X p o rp h y r y
— M in o r tren d
ir ' 1 L p o rp h y ry □ Q u a r t z e y e p o rp h y ry
in - H S e c tio n " A " fe ld s p a r p o rp h y ry A n d e s ite

Fig. 12. Mineralized fracture and vein patterns developed in porphyry copper deposits. A. Concentric and radial miner­
alized fractures and veins developed at high levels in die San Juan mine area, Salford Mining District, Arizona. Note super­
position o f radial and concentric veins and mineralized fractures. In contrast, dikes do not conform to this patient but form
orthogonal sets. Modified from Heidrick and Titley (1982). B. Surface distribution ofdacitc dikes, pebble breccias, and
hydrothermal breccias on Cciro Indio Muerto as mapped by Langerfcld (1964). Pebble breccias are related to the post-min­
eral dacite dikes and D wins, and are paragenetically late. The pebble dikes decrease in abundance with depth and gener­
ally lie in the upper pans of the deposit above a zone that separates scriciic-dominatcd alteration from an overlying mixed
scricitc-kaolin alteration assemblage. With depth and where present, pebble dikes become more linear with dominantly
northwesterly strikes (see C and Fig. 1 IB). 'Hie hydrothermal breccias are cemented by hydrothermal minerals and, thus,
represent intermineral bodies. Superposed on the dikes arc interpreted concentric stress trajectories (dashed lines) that sug­
gest the presence of two domical stocks at depth. When compared with the geological map o f the subsurface (C), the center
o f the concentric tings correspond generally to the center o f die composite stock o f X. K, and L porphyries to the southeast,
and the main mass o f L porphyry to the northwest C. Distribution o f country rock units. Eocene granodioritc porphyries,
and cross-cutting D reins on the 2600- and 2660-m levels o f the El Salvador porphyry deposit. Note the roughly radial and
elliptical distribution o f veins, Modified from Gustafson and Hunt (1975). D. Generalized fracture distribution and orienta­
tion around the Sienitaand Esperanza porphyry deposits, Arizona. Modified from Tilley e l al. (1986) and U tley (1993).
 (Undsay et a!., 1995); evidently, the influence of magmatic Taking the simplest case as illustrated in Figure 12, an
pressure over tectonic stress was uansitory. The enargite veins orthogonal set o f veins characterizes many porphyry systems
may represent shallower levels of a younger porphyry Cu (see also Stanley et al., 1995) regardless o f depth o f forma­
deposit emplaced during uplift and unroofing o f the older tion. In some deposits such as El Salvador (Fig. 12C) and
system (Sillitoe, 1994). It is interesting that at Chuquicamata, locally in Sierrita (Titley et al., 1986), other veins bisect the
the dom inant faults and mineralized fractures within the orthogonal veins. During brittle failure of rocks, extensional
composite porphyry Cu deposit form a crude conjugate frac­ fractures are die easiest to form, and are oriented normal to
ture network, a pattern predicted by numerical modeling of the minimum effective principal stress, o f (Jaeger and
porphyry stocks (see above), despite the local structural con­ Cook, 1979; Sibson, 2000, 2001). Such fractures should
trol exerted by the Falla Oeste (Domeyko fault system). Con­ dominate fluid-saturated environments such as porphyry Cu
jugate vein patterns also characterize many other deposits, deposits, and they will be filled by hydrothermal minerals
such as Panguna (Papua New Guinea; Clark, 1990) and High­ whenever open. However, formauon o f extensional shear or
land Valley (British Columbia; Osatenko and Jones, 1976; compressional shear veins requires a higher differential
Casselman ctal., 1995). stress (Sibson, 2000, 2001). Such fractures and veins lie at
high angles to a f and should bisect the intersection of
Implications for stress during mineralization
orthogonal fractures; they will also be less common than
Despite the inherent beauty of Heidrick and Titley’s extensional fractures unless a planar fabric is already pre­
(1982) vertical zonation model, the variety of vein orienta­ sent within the rock, or, if the numerical models of Gow and
tions noted above suggests that it requires some modifica­ Ord (1999) are correct, little o r no rheologic contrast exists
tion. Concentric and radial fractures are expected to be con- between the stocks and host rocks. Hence, from simple rock
cenuated above and exterior to the intruding stock to which mechanical consideration, form auon o f veins in muldple
they arc related; where multiple intrusions are present, frac­ orientations during the life o f the porphyry Cu deposit
tures should form overlapping arrays (Fig. 12B, C). In con­ requires specific stress conditions. In view of the ease of
trast, after solidification, the stock and its host rocks will forming extensional fractures, the simplest interpretadon of
undergo brittle fracture associated with tectonic and the vein pattern in a porphyry Cu deposit, particularly of
hydrothermal activity. The orientations of resulting fractures orthogonal sets, is that G f changed orientadon many times
will reflect a new set of stress conditions controlled either by during the life of the system. In essence, the existence of
magmatic o r tectonic forces. Magmatic stress results from orthogonal veins requires that o f and the other horizontal
renewed intrusion, whereas tectonic stress is externally stress, either the maximum effective principal stress ( a /) or
imposed. Fractures developed under a tectonic stress field the intermediate effective principal stress (o2') depending
will reflect that field, and should be more ordered and linear. upon the stress field, swapped episodically during the life o f
They will also develop later in the system’s evolution, and will the system. This situation can be visualized by comparing
be concentrated within the stocks themselves. The elongate the inferred stress trajectories above a stock in anisotropic
arrays of type-D veins within Eocene granodtorite stocks at El rocks, such as is shown in Figure 7C and E.
Salvador (Fig. 12C) and the concentric and radial arrays in Porphyry Cu deposits form over a short period o f geo­
the overlying rocks can be attributed to supeiposition of frac­ logical time based upon thermal models o f cooling stocks
ture systems derived from early magmatic- and later tectonic- (Cathles, 1977) and geochronotogical data (e.g., Chesley
dominated stress fields (Fig. 12B, C). and Ruiz, 1997; Cornejo e t al., 1997; Marsh e t al., 1997).
Multiple vein sets are normal in porphyry Cu deposits. The Because o f this relatively short duration o f hydrothermal
veins are usually cogenetic, and open episodically and repeat­ activity, it seems unlikely that a significant and episodic
edly at different paragenetic stages during formation o f the rotation o f the regional stress field would occur d uring
deposits. In addition, there is usually either a consistent vein deposit form ation. T herefore, in o rd er to explain the
orientation persisting throughout the deposit laterally and repeated rotation of the effective minimum stress direction
vertically, o r the veins have a consistent angular relationship called for above, a low-differential stress field, modified by
between them, or both. Where documented, veins are com­ fluctuating fluid pressures, may be the norm during por­
monly orthogonal or conjugate, and are rarely completely phyry Cu formation.
randomly oriented with respect to one another. A common
angular pattern is not what would be expected in the case Origin o f Breccia Pipes in Porphyry Cu Deposits
where die veins reflect simple fracturing above a stock under As noted previously, hydrothermal breccia pipes are of
conditions approaching uniaxial extension, o r where they varying im portance in porphyry Cu deposits. They range
result from volume expansion caused by exsolution of from insignificant in vein-dominated systems, to forming
hydrothermal fluids. It is the consistent deposit-scale angular the major portion o f the ore body in breccia-dominated sys­
relationship among veins which suggests that intrusion or tems (Sillitoe, 1985). Most breccia pipes are pie- to synmin-
hydrothermal-related models for vein formation are simplis­ eral, but some pebble breccias are late and are usually bar­
tic. Recognizing a common angular relationship is also criti­ ren (Richard and Courtright, 1958; Sillitoe, 1985). All
cal to understanding the state o f stress that dominates during hydrothermal breccias in the porphyry Cu environment;
porphyry Cu formation, which in turn has important implica­ regardless of their genesis, reflect sudden expulsion of fluid
tions for the tectonic setting of these deposits. at pressures that exceed the lithostatic load, thereby caus­
ing hydraulic fracturing (Burnham, 1985; Sillitoe, 1985).
This fluid release is generally attributed to the process o f
second boiling in the porphyry magma, and rupturing of a
pressure seal a t the brittle-ductile transition (Fig. 4;
Phillips, 1973; Burnham, 1985; Fournier, 1999). T he resul­
tant catastrophic fragm entation o f the host rocks propa­
gates upwards from a narrow source toward areas of
reduced effective m ean stress. Most porphyry Cu-related
breccias are rooted in porphyry intrusions o r dikes but
apparently do n o t vent to the paleosurface (Sillitoe, 1985),
resulting in lensoid, ovoid, or irregular pipe-like geometry.
Hydrotherm al m inerals fill the resulting voids leading
locally to bonanza concentrations o f ore.
Zw engand Clark (1995) proposed that hydrotherm al
breccia pipes are more common in porphyry Cu deposits
form ed a t shallow levels because o f the decreased litho-
siatic load imposed by the mass of country rocks above the
stock. However, no t all shallow porphyry Cu deposits are
dom inated by breccias as shown by vein-dominated Cu-Au
and Au-Cu systems emplaced within 1 km o f the surface in
the Maricunga belt, Chile. Hence, simple magmatic and
fluid exsolution processes may not have been sufficient in
all deposits to perm it breccia formation. Because breccias
are present to varying degrees in all porphyry Cu deposits,
the obvious question is what controls their formation and
abundance? Porphyry Cu deposits represent dynamic sys­
tems where magmatism and fluid circulation are affected
by tectonic stresses, so there is no shortage o f potential
local causes that could trigger breccia form ation.
Amongst these local triggers are fluid-pressure build-up
(Burnham, 1985), stock or dike intrusion that disrupts the
evolving porphyry Cu system (Langerfeid, 1964), sector
collapse of the overlying volcanic edifice (Sillitoe, 1994),
and local seismic activity perhaps associated with intrusion
and volcanism. Skewes and Stern (1994) also proposed
that tectonic-driven changes in magmatism triggered
breccia form ation in the late-M iocene porphyry Cu
deposits o f central Chile. Still another potential trigger
for breccia form ation is rem ote (>500 km) earthquake
activity, which has been shown to generate enhanced seis­
mic activity in active geotherm al and magmatic systems
(Hill et al., 1993). Evidently, propagation o f seismic
energy into a magmatic or hydrotherm al system is suffi­
cient to induce gas separation leading to volume expan­
sion (Linde et al., 1994; Sturtevant et al. 1996). The mag­
matic pressure increase could give rise to volcanic
eruptions (Sahaglan and Proussevitch, 1992), whereas in
the hydrothermal environment of a porphyry Cu deposit,
increased fluid pressures could be sufficient to trigger
rupturing of the pressure seal above the porphyry system,
leading to generation of breccia pipes.
Tectonic Setting of Porphyry Cu Deposits
A simple model of porphyry Cu deposit genesis involves
formation from hydrotherm al fluids exsolved from calc-
alkaline to alkaline magmas in a subvolcanic arc environ­
ment. However, many shallow-level porphyritic plutonic
complexes are devoid of or have only weak porphyry-style
mineralization. The barren Paleocene ring complexes o f
the San Jeronim o superunit in the Peruvian Coastal
batholith are examples from an arc otherwise rich in por­
phyry Cu deposits o f broadly sim ilar age (Bussell, 1985;
Pitcher, 1985; Zweng and Clark, 1995). O n a m ore global
scale, porphyry Cu deposits are sparse in the western Pacific
(Uyeda,and Nishiwaki, 1980), .w ith notable exceptions in
Papua New Guinea and the Philippines (MacDonald and
Arnold, 1994; Sillitoe, 1997; Hedenquist et al., 1998). In
addition, porphyry Cu form ation is characteristically
episodic and localized, being concentrated during limited
time ranges within narrow, margin-parallel belts (Sillitoe,
1972, 1988; Titley and Beane, 1981; Clark et al., 1982;
Dam on e t al., 1983). These considerations suggest that
additional factors must be involved in determining whether
a porphyry Cu deposit will form or not, including details of
magma chemistry (e.g., water content, oxidation state) and
tectonic setting (Uyedaand Nishiwaki, 1980).
A review o f structural sellings o f porphyry Cu deposits
indicates that there are no unique environments into which
these deposits are emplaced. They arc found within strike-
slip fault zones with only limited displacement contempo­
raneous with mineralization (Maksaev and Zentilli, 1988; |
Clark e t al., 1990; Sapie and Cloos, 1994; Lindsay e t al., ||
1995), within regions affected by slightly older to concur- j
rent transpressional o r contractile strain and uplift (Tilley :1
and Heidrick, 1978; Heidrickand Titley, 1982; Olson, 1989; ;|
Tomlinson, 1994; Sillitoe, 1997), in areas undergoing liin- f
ited extensional deformation (Presnail, 1997), and in areas (1
lacking major active fault systems (Sillitoe, 1997). Many but I
not all deposits are localized along older fault systems that . J
provided crustal permeability (Heidrick and Titley, 1982;
Richards et al., 2001). Significant porphyry Cu deposits are,
however, seemingly absent from regions undeigoing large-
scale extensional strain, a negative association that has long
been recognized, ancl which seems to preclude those arcs
or times within arcs where such tectonics dominate. A com­
mon theme throughout these convergent arcs is the limited *
deformation that is demonstrably contem poraneous with
porphyry Cu development.
Times of porphyry Cu generation have been causally cor­
related with times of shallow-dipping subduction (Sillitoe, [
1972; Nielsen, 1976; Uyeda and Nishiwaki, 1980; Titley ancl
Beane, 1981), increased convergence rates (Clark et al.,
1990), subduction of aseismic ridges (Skewes and Stern, '
1995), subduction zone reversals (Solomon, 1990), or fol- '
lowing uplift (Sillitoe, 1997). Many but not all porphyry Cu :f/
deposits in Arizona (McCandless and Ruiz, 1993), southern :
Peru (Clark et al., 1990), Chile (Maksaev and Zentilli, 1988;
Skewes and Stern, 1995; Richards et al., 2001), and Papua
New Guinea (Titley and Heidrick, 1978; Sillitoe, 1997)
formed near the end of major periods of magmatism, con­
tractile or transpressional strain, and uplift. Sillitoe (1997).
suggested that rapid uplift in an arc is an essential precuiv
sor to the formation of giant porphyry Cu deposits, which ;
appear near to or just after the end of that event.
The above observations, combined with die near-circular '
plan of many porphyries and the common deposit-scale ori­
 entations o f associated veins and dikes, suggest that por­
phyry Cu deposits are developed during periods of low-dif­
ferential or ncar-isotropic horizontal stress in the arc. The
achievement o f such stress states is likely driven by funda­
mental changes in subduction geometry and convergence
rates. For exam ple, changes in the coupling of stress
between the subducting and overriding plate may lead to
shifts in the location o f zones of upper-plate deformation
with respect to die magmatic arc, either inboard toward the
back-arc or outboard toward the convergent margin. This
shift in the locus of deformation relative to magmatism may
permit the arc to approach a neutral stress state. The giant
Miocene and Pliocene porphyry Cu deposits in central
Chile (Los Pelambres, Rio Blanco-Los Bronces, El
Teniente) seem to fit this scenario because diey formed in
the hanging wall of a fold-and-thrust belt that lies to the
east in the A rgentinean foreland (Jordan et at., 1993;
Ramos et al., 1996). In an alternative scenario, relocation of
the volcanic arc may be accompanied by relaxation o f com-
pressional o r transpressional stress in an old, established,
plutonic-volcanic arc, resulting in late-stage development of
porphyry magmatism. The late-Eocene-early-Oligocene
porphyry Cu deposits of northern Chile are proposed to
have formed during such a period of stress relaxation fol­
lowing a prolonged episode o f Eocene transpression and
effusive volcanic activity; relaxation coincided with flatten­
ing o f the subduction angle and subsequent inland shift of
magmatism (Maksaev and Zentilli, 1988; Clark, 1993;
Richards et al., 2001).
These characteristics may be speculatively tied to models
of magmagenesis and crustal em placem ent as reviewed
above (Fig. 13). First, the simple textbook form o f a
destructive margin—that of orthogonal compression, deep-
crustal o r mantle melting, and widespread effusive magma­
tism—is not conducive to porphyry Cu form ation (Fig.
13A). N either is a state of arc rifting (extension), where
crustal m elting is short-circuited and mantle-derived
basaltic melts ascend directly to the surface (Fig. 13B) con­
ducive to porphyry Cu formation. Instead, the various tec­
tonic environments identified during porphyry Cu genera­
tion around the world suggest that ideal conditions form
during transitions between these end-member states, usu­
ally caused by changes in subduction geom etry o r rate.
Geochemical studies show that, even in island-arc settings,
porphyry Cu-related magmas are not direct products of
asthenospheric m elting in the m antle wedge above the
Benioff zone, but result from significant amounts of frac­
tionation and interaction of these magmas with the overly­
ing lithosphere. Isotopic and geochemical evidence for
crustal interaction in Central Andean magmas, for exam­
ple, is overwhelming (Tilton et al., 1981; Barreiro, 1984;
Harmon et al., 1984; Pankhurst et al., 1988; Walker et al.,
1991). A period o f “stewing” near the base o f the lithos­
phere is implied, involving extensive MASH (melting,
assimilation, storage, and homogenization; H ildreth and
Moorbath, 1988) processes in voluminous but probably dif­
fuse, batholithic magma chambers. Conditions appropriate
to this activity are achieved during “classic” convergence
periods in the arc, compressive stresses in the upper plate
acting to close m ajor orogen-parallel structures, and
inhibiting ascent of.all but the most overpressured magmas
(Fig. ISA).'.Those magmas that do eru p t do so violently,
producing widespread .effusive volcanic sequences but
largely unaccompanied by shallow-level plutonism.
If compressive stress in the lithosphere is locally relaxed,
however, magma dynamics subordinate tectonic stress, and
melts may ascend through the crust more passively driven
by buoyancy forces, rather than violently by magmatic over­
pressure (Fig. 13C, D). In consequence, such magmas tend
to rise to levels of neutral buoyancy in the shallow crust,
where they may undergo further fractionation and crustal
interaction, and potentially generate porphyry-style
deposits as these distillate magmas cool and saturate with
volatiles. Volcanism, where it occurs, is m ore localized,
forming discrete composite edifices directly above shallow-
source magma chambers.
An analog may be found in the use o f a pressure cooker.
Effective stewing is achieved at elevated temperatures and
pressures by sealing the pot with a lid and a safety;valve
(periodic violent eruptions occur only when high-pressure
fluids overcome the valve weight). Sudden removal.qf the
lid (equivalent to extensional tectonics, and not advised at
home) will result in catastrophic eruption o f the contents
o f the pot, equivalent to basaltic rift magmatism. To pre­
serve the stewed contents ju st right, the pot needs to be
depressurized (stress relaxation) before the lid is removed.
In this model, localization of porphyry magmatism in the
crust can be understood in terms of structural permeability.
During periods o f arc-normal compression, major arc-paral­
lel structures will be closed, and magma penetration will be
restricted. Upon relaxation of stress, however, these struc­
tures will proride planes o f elevated permeability through
the crust. More specifically, faultjogs or structural intersec­
tions may generate pull-apart volumes when oblique com­
pressive stress relaxes (Fig. 13C, D). Such volumes represent
discrete vertical columns o f maximum permeability through
the crust, which may serve as optim um loci for magma
ascent and potential porphyry Cu generation. Relatively low
degrees of differential horizontal stress are implied by this
model, because fault lock-up under high strain will progres­
sively destroy permeability (Sibson, 2001).
Concluding Remarks
Porphyry Cu deposits require the precise coincidence
and positive interaction of a series of individually common­
place geological processes. The role of magmas and their
exsolved hydrothermal fluids have long been recognized
and widely studied, and are now reasonably well under­
stood. In contrast, the detailed influence o f structure and
tectonic setting on deposit localization and evolution in the
porphyry environment is under-appreciated, despite several
careful studies. In part, this situation stems from the
immense scale of porphyry systems (commonly exceeding
diameters oflO km) and the inherent difficulties that this
poses to sutictural analysis. It also derives from the research
focus on mineralogic and geochemical processes during the
 Fig . 13. Cartoon depiction o f d ie relationship between convergent margin tectonics, upper plate structures, and mag-
matism. A. In orthogonally convergent settings, major arc-parallel fault zones are compressed. Magmas generated in die
mantle wedge pond near the base o f die lithosphere, and interact with it dirough processes o f melting, assimilation, stor­
age, anti homogenization (MASH processes; Hildredi and Moorbath. 1988). Ascent to the surface is restricted to mag­
mas under high overpressure, which erupt violcndy without significant residence in upper crustal magma chambers. B.
Arcs under tension provide multiple high-permeability paths for magma ascent, to the exten t that asthenosphere-
dcrived magmas may rise directly to the surface. T he distillation and crustal interaction processes that appear to b e nec­
essary for the development o f porphyry Cu-prospcctive magmas arc thereby short-circuited, and porphyry Cu deposits
arc not found in rifling arcs. C and D. Upon relaxation o f stress, localized zones o f extension may be generated during
transprcssional or. optimally, uanstensional strain along arc-related structures. These zones provide high-permeability
conduits for the passive ascent o f magmas to shallow crustal levels, driven primarily by buoyancy rather than magmatic
overpressure. Hypabyssal plutonism with associated localized volcanism is favored, and is an environment conducive to
porphyry Cu formation. PCD « porphyry Cu deposits.
period o f major porphyry Cu exploration and development
in the 1960s to 1980s (Tidey and Hicks, 1966; Sutherland
Brown, 1976; Tidey, 1982; Pierce and Bolm, 1995; Schroeter,
1995). Nevertheless, structure and tectonics clearly play a
major role in porphyry Cu emplacement and genesis, and
may be critical in determining where and when spatially and
temporally localized porphyry Cu provinces form within
convergent plate margins.
Porphyry Cu deposits and all their genetically associated
deposits are a natural consequence o f convergent margin
magmadsm. Nevertheless, the preceding review suggests
that significant porphyry Cu deposits form in specific struc­
tural environments under a near-neutral stress state. These
structural condidons appear to occur during periods of Hi­
de deformation o r in areas removed from deformation, or
during stress relaxation in the magmatic arc, conditions
likely to be infrequent and transitory during the life of a
convergent margin. This hypothesis suggests that detailed
research into arc history and dynamics may be fruitful in
predicting the loci not only of porphyry Cu provinces, but
also of individual porphyry centers within those provinces.
Acknowledgments
Writers of review articles benefit from the great minds,
careful work, and published literature that precedes them.
We are no different. We extend our appreciation to all of
those people who have lent their expertise to the study of
porphyry deposits; many of them are referenced herein, but
we apologize to any who have not been, either due to con­
straints of space or our oversights. RMT would like further
to acknowledge A.H. Clark, who first introduced him to
Andean porphyry copper deposits, and to geologists of Sevi-
cio Nacional de Geologia y Mineria (Chile), U.S. Geological
Survey, and Compania del Cobre-Chile for research oppor­
tunities in porphyry deposits in North and South America.
:JPR acknowledges RTZ Mining and Exploration Ltd., and
•Minera Outokumpu Chile S.A. for fieldwork support, and
research grants from the Natural Environm ent Research
Council (UK) and the Natural Sciences and Engineering
Research Council (Canada). The m anuscript benefited
from discussions with P. Erdmer and J. Waldron, and reviews
by T. Theodore and CJ. Hodgson; their insights and efforts
are greatly appreciated. This paper represents MDRU con­
tribution 141.
REFERENCES
Arculus, R.J., 1994, Aspects o f magma genesis in arcs: Lithos, v. 33,
p . 189-208.
.Australian Geodynamics Research Group, 1997, Annual Report 96/97:
••• Mechanical m odeling. Objective 4: www.agcrc.csiro.au/publications/
’■ rcports/AnnunlReport96-97/objcc tive4.html
Barreiro, B.A., 1984, Lead isotopes and Andean mngmagcncsis, in Har­
mon, R.S., and Barreiro, B.A., eds., Andean magmatism chemical and
isotopic constraints: Naiuwich, Cheshire, Shiva, p. 21-30.
Baumcr, A., and Fraser, R.B., 1973, Panguna porphyry copper deposit,
: Bougainville, in Knight, C.L„ ed„ Economic geology o f Australia and
; Papua New Guinea, 1: Metals: Australasian Institute o f Mining and Met­
allurgy Monograph 5, p. 855-866.
, Beane, R.E., and Tilley, S.R., 1981, Porphyry copper deposits. Part II.
; Hydrothermal alteration and mineralization: Economic Geology 75th
Anniversary Volume, p. 235-263.
Bouse, R.M., Ruiz,J„ Tidey, S.R., Tosdal, R.M., and W ooden, J.L.. 1999,
Lead isotope com positions o f Late Cretaceous and Early Tertiary
igneous rocks and sulfide minerals in Arizona: Im plications for the
sources o f plutons and metals in porphyry copper deposits: Economic
Geology, v. 94, p. 211-244.
Britten, R.M., and Marr, J.M ., 1995, T he Eaglchcad porphyry copper
prospect, northern British Columbia, in Schroeter, T.G., ed., 1995, Por­
phyry deposits o f the northwestern Cordillera o f North America: Cana­
dian Institute o f M ining, Metallurgy, and Petroleum, Special Volume
46. p. 467-472.
Brown, G.C., 1982, Calc-alkaline intrusive rocks: Their diversity, evolution,
and relation to volcanic arcs, in Thorpe, R.S., ed.. Andesites: New York,
Wiley, p. 437-461.
Burnham, C.W., 1967, Hydrothermal fluids at the magmatic stage, in
Barnes, H.L., ed„ Geochemistry o f Hydrothermal Ore Deposits: New
York. Holt, Rinehart and Winston, p. 34-76.
------ 1979, Magmas and hydrothermal fluids, in Barnes, H.L., ed., Geo­
chemistry o f Hydrothermal Ore Deposits, 2nd ed.: New York, Wiley,
p. 71-136.
------ 1985, Energy release in subvolcanic environments: Implications for
breccia formation: Economic Geology, v. 80, p. 1515-1522.
Burnham, C.W., and O hm oto. H., 1980, Latc-stagc processes in felsic mag­
matism: Mining Geology Special Issue, no. 8. p. 1-11.
Bussell, M A , 1976, Fracture control o f high-level piutonic contacts in the
Coastal batholith o f Peru: Proceedings o f the Geologists Association,
v. 87, p. 237-246.
------ 1985, T he centred com plex o f the Rio Huaura: A study o f magma
m ixing and differentiation In high-level magma chambers, in Pitcher,
W.S., Atherton, M.P., Cobbing, E.J., and Bcckinsalc, R.D., eds., Magma­
tism at a Plate Edge; T he Peruvian Andes: Glasgow, Blackie, p. 128-155.
Bussell, M A , Pitcher, W.S., and Wilson, PA.. 1976, Ring complexes o f die
Peruvian Coastal batholith: A long-standing subvolcanic regime: Cana­
dian Journal o f Earth Sciences, v. 13. p. 1020-1030.
Camus, F., 1975, G eology o f the El T cniem e orebody with emphasis on
wall-rock alteration: Economic Geology, v. 70, p. 1341-1372.
Candela, PA., 1989, Magmatic ore-forming fluids: Thermodynamic and
mass transfer calculations o f metal concentrations: Reviews in Eco­
nomic Geology, v. 4, p. 225-233.
Candela, P.A., and Holland. H.D., 198-1. T he partitioning o f copper and
molylKlcmim between silicate melts and aqueous fluids: Gcochimica e t
Cosmochimica Acta. v. 48, p. 373-380.
------ 1986, A mass transfer m odel for copper and molybdenum in mag­
matic hydrothermal systems: T he origin o f porphyry-type ore deposits:
Economic Geology, v. 81, p. 1-19.
Cassclman, M.J., McMillan. W J.,and Newman, K.M., 1995, Highland Vat-
Icy porphyry copper deposits near Kamloops. British Columbia: A
review and update with emphasis o n the Valley deposit, in Schroeter.
T.G., ed., 1995, Porphyry deposits o f the northwestern Cordillera o f
North America: Canadian Institute o f Mining. Metallurgy, and Petro­
leum, Special Volume 46, p. 161-191
Castro, A., and Fernandez, C„ 1998, Granite intrusion by externally
induced growth and deformation o f the magma reservoir. The example
o f the Plascnzuela ptuion Spain: Journal o f Structural Geology, v. 20,
p. 1219-1228.
Cathics, L.M., II, 1977, An analysis o f the cooling o f imrusives by ground-
water convection that includes boiling: Econom ic Geology, v. 72,
p. 804-826.
Chcslcy, J.T., and Ruiz, J., 1997, Preliminary Rc-Os elating on molybdenite
mineralization from the Bingham Canyon porphyry copper deposit,
Utah, in John , D.A., anti Ballantync, G.H., eds., G eology and ore
deposits o f the Oquirrh and Wasatch Mountains, Utah: Society o f Eco­
nomic Geologists Guidebook Scries, v. 29, p. 165-170.
Clark, A.I-L, 1993, Are outsize porphyry copper deposits cither anatomi­
cally or environmentally distinctive?, in Whiting, B.H., Mason, R., and
Hodgson, C.J., eds., Giant ore deposits: Society o f Economic Geologists,
Special Publication 2, p. 213-283.
Clark, A.H., Farrar, E., Kontak, D.J., Langridge, R.J., Arenas F., M.J.,
France, LJ„ McBride, S.L., Woodman, P.L., Wasiencys, H A ., Sande-
man, H.A., and Archibald, D.A., 1990, Geologic and geochronologic
constraints on the mctallogcnic evolution o f the Andes o f southeastern
Peru: Economic Geology, v. 85, p. 1520-1583.
 Clark, G.H., 1990, Panguna copper-gold deposit, in Hughes, F,E„ eel., Geol­
ogy o f die Mineral Deposits o f Australia and Papua New Guinea: Mel­
bourne, Australasian Institute o f Mining and Metalling)-, p. 1807-1816.
Clark, K.F., Foster, C.T., and Damon, P.E., 1982, Cenozoic mineral deposits
and subduction-rclated magmatic arcs in Mexico: Geological Society o f
America Bulletin, v. 93, p. 533-544.
Cline, J.S., and Bodnar, RJ.. 1991, Can econom ic porphyry copper miner­
alization be generated by a typical calc-a!ka!inc melt?: journal o f Geo­
physical Research, v. 96. p. 8113-8126.
Corn, R.M., 1975, Alteration-mineralization zoning. Red Mountain, Ari­
zona: Economic Geology, v. 70, p. 1437-1447.
Cornejo, P., Tosdai, R.M., Mpodozis, C., Tomlinson, A.J., Rivera, O., and
Fanning, C.M., 1997, El Salvador, Chile, porphyry copper deposit revis­
ited: Geologic and gcochronologic framework: International Geology
Review, v. 39. p. 22-54.
Cornwall, H.R., and Kricgcr. M.H., 1974, Geologic map o f the Grayback
quadrangle. Pinal County, Arizona: U.S. Geological Survey G eologic
Quadrangle Map GQ-1206, scale 1:24,000.
Corrivcau, L.. Rivard, B., and Van Brccmen, O., 1998, Rheological con­
trols on Grenviliian iiumsive suites: Implications for tectonic analysis:
Journal o f Structural Geology, v. 20, p. 1191-1204.
Corry, C.E., 1988, Laccoliths: M echanics o f em placem ent and growth:
Geological Society o f America, Special Publication 220.110 p.
Crcascy, S.C., 198-1, T lic Schultzc granite, the Tea Cup granodiorite, and
the Granite Basin porphyry: A geochemical comparison o f mineralized
and unmincralized stocks in southern Arizona: U.S. Geological Survey
Professional Paper 1303,41 p.
Crudcn, A.R., 1998, O n the em placem ent o f tabular granites: Journal o f
die Geological Society o f London, v. 155, p. 853-862.
Damon, P.E., Shafiqullah, M„ and d a rk . K.F., 1933, Geochronology o f the
porphyry copper deposits and related mineralization o f Mexico: Cana-
dtanjoum al o f Earth Sciences, v. 20, p, 1052-1071.
de Saint Blanquat, M., TikofF, B., Teyssier, C., and Vigncrcsse, J.L., 1998,
Transprcssional kinematics and magmatic arcs, in lioid sw on h , R.E.,
Strachan, R A , and Dewey, J.F., cds., Continental Transprcssional and
Tmmtensional Tectonics: Geological Society o f London Special Publi­
cations 135, p. 327-340.
Dehls, J.F., Cnulcn, A.R., and Vigncrcsse, J.L., 1998, Fracture control o f
late Archcan plm on em placem ent in the northern Slave Province,
Canada: Journal o f Structural Geology, v. 20, p. 11-15-1154.
Delaney, P.T., Pollard, D.D., Ziony.J.l., and McKee, E.H., 1986, Field rela­
tions between dikes and joints: Emplacement processes and paleostress
analysis: Journal o f Geophysical Research, v. 91, p. 4929-4938.
Dewey, J.F., 1980, Episodicity, sequ en ce, and style at convergent plate
boundaries, in Strangway, D.W., cd., The continental crust ami its min­
eral deposits: G eological Survey o f Canada, Special Publication 20,
p. 553-574.
Dilles, J.H., 1987, Petrology o f the Yeringion batholith, Nevada: Evidence
for evolution o f porphyry copper ore fluids: Economic Geology, v. 82,
p . 1750-1789.
Dilles,J.H., Fanner, G.L., and Field, C.W., 1995, Sodium-calcium alteration
by non-magmatic saline fluids in porphyry copper deposits: Results
from Ycringcon, Nevada, m Thompson, J.F.H., cd., Magmas, Fluids, and
Ore Deposits: Mineralogical Association o f Canada Short Course series,
v. 23, p. 309-338.
Dunn, P.G., 1982, Geology o f the Copper Flat porphyry copper deposit, in
Tilley. S.R., cd., Advances in geology o f the porphyry copper deposits,
southwestern North America: Tucson, University o f Arizona Press,
p . 313-325.
Eastoc. C.J., 1978. A fluid inclusion study o f the Panguna porphyry copper
deposit. Bougainville. Papua New Guinea: Econom ic Geology, v. 73,
p. 721-718.
Emmons, W.H.. 1927, Relations o f disseminated cop|>cr ores in porphyry
to igneous intrusions: American Institute o f Mining and Metallurgical
Engineers Transactions, v. 75, p. 797-809.
Fitch.J.T., 1972, Plate convergence, transcurrcnt faults, and internal defor­
mation adjacent to Southeast Asia and the Western Pacific: Journal o f
Geophysical Research, v. 101. p. 425-^78.
Fold), A., and Marti, J., 1998, T he generation o f overpressure in fclsic
magma chambers by replenishment: Earth and Planetary Science Let­
ters, v. 163, p. 301-314.
Fournier, R.O., 1999, Hydrothermal processes related to m ovement o f
fluid from plastic into brittle rock in the magmatic-epithermal environ­
ment: Economic Geology, v. 94, p. 119JM212.
Gammons, C.H., and Williams-Jones, A.E., 1997, Chemical mobility o f
gold in the porphyry-epithermal environm ent: Econom ic Geology,
v. 92, p. 45-59.
Gill. J.B., 1981, O togen ic andesites and plate tectonics: New York,
Springcr-Vcriag, 390 p.
Glazner, A.F., 1991, Plutonism, oblique subduction, and continental
growth: An example from the Mesozoic o f California: Geology, v. 19,
p. 784-786,
Gow, P., and Ord, A., 1999, Numerical m odelin g o f deform ation and
fluid-flow in shallow plutonic comprcssional environments: Halls Gap,
Victoria, Specialist Group in Tectonic and Structural Geology Confer­
ence, 7 p.
G io c o tt.j., Brown, M., Dallmeyer, R.D., Taylor, G.K., andTreloar, P.J.,
1994, Mechanisms o f continental growth in cxtcnsional arcs: An exam­
ple from the Andean plate-boundary zone: Geology, v. 22, p. 391-39-1.
Gustafson, L.B.. 1978, Som e major factors o f porphyry copper genesis:
Economic Geology, v. 73, p. 600-607.
Gustafson, L.B., and Hunt,J.P., 1975, The porphyry copper deposit at El
Salvador, Chile: Economic Geology, v. 70, p. 857-912.
Gustafson, L.B., and Quiroga, G., 1995, Patterns o f mineralization and
alteration below the porphyry copper orebody at Ei Salvador, Chile:
Economic Geology, v. 90, p. 2-16.
Hamilton, W., 1981, Crustal evolution by arc magmaiism: Royal Society o f
London Philosophical Transactions, Earth Sciences, Series A. v. 301.
p. 279-291.
Hamilton. W.B., 1988, Plate tectonics and island arcs: Geological Society
o f America Bulletin,v. 100,p. 1503-1527.
H am ilton, W.B., and Myers, W.B., 1967, T he nature o f batholiths: U.S.
Geological Survey Professional Paper 554-C, p. C1-C30.
Harmon, R.S., Barreiro, B A .. M oorbath, S„ H ocfs, J„ Francis, P.W.,
T horpe, R.S., D erucllc. B., M cHugh, J., and Viglino, J A , 1984,
Regional O-, Sr-, and Pb-isotopc relationships in Late Cenozoic calc-
alkaline lavas o f the Andean Cordillera: Journal o f tlic Geological Soci­
ety o f London, v. 141, p. 803-822.
H cdcnquist, J.W., 1995, T he ascent o f magmatic fluid: discharge versus
m ineralization, in T hom pson, J.F.H.. cd ., Magmas, fluids, and ore
deposits: M ineralogical Association o f Canada Short Course, v. 23.
p. 263-289.
Hcdcnquist. J.W., and Richards, J.P., 1998, The influence o f geochemical
techniques on the development o f genetic models for porphyry copper
deposits: Reviews in Economic Geology, v. 10, p. 235-256.
H cdcnquist, J.W., Simmons, S.F., Giggcnbach, W.F., and Eldridgc, C.S.,
1993, White Island, New Zealand, volcanic-hydrothermal system repre­
sents the geochemical environment o f high-sulfidation Cu and Au ore
deposition: Geology, v, 21, p. 731-734.
Hcdcnquist, J.W„ Arribas, A., Jr., and Reynolds, J.R., 1998, Evolution o f an
intrusion-centered hydrothermal system: Far Souihcast-Lepanto por­
phyry and epithermal Cu-Au deposits, Philippines: Economic Geology,
v. 93, p. 373-104.
Hcidrick, T.L„ and Titlcy, S.R., 1982, Fracture and dike patterns in
Laramidc plutons and their structural and tectonic implications: Amer­
ican southwest, in Titley, S.R., cd.. Advances in geology o f the porphyry
copper deposits: Tucson, University o f Arizona Press, p. 73-91.
Hemtcy, J.J., and Jones, W.R., 1964. Chemical aspects o f hydrothermal
alteration with emphasis on hydrogen metasomatism: Economic Geol­
ogy, v. 59, p. 538-569.
Hildreth, W., and Moorbath, S.. 1988, Crustal contributions to arc mag-
matism in the A id e s o f ccm ial Chile: Contributions to Mineralogy and
Petrology', v. 98, p. 455-489.
Hill, D.P., Rcascnbcrg. P A , mid 29 others, 1993, Seismicity remotely trig­
gered by the magnitude 7.3 Landers, California, earthquake: Science,
v. 260. p. 1617-1623.
H ogan. J.P., and G ilbert. M.C., 1995, T h e A-typc Mount Scott granite
sheet: Im portance o f crustal m agma traps: Journal o f Geophysical
Research, v. 100, p. 15,779-15,792.
Hogan, J.P., Price, J.D., and Gilbert, M.C., 1998, Magma traps and driving
pressure: C onsequences for phtlon shape and em placem ent in an
cxtensional regime: Journal o f Structural Geology, v. 20, p. 1155-1168.
Holland. H.D.. 1972. Granites, solutions, and base metal deposits: Eco­
nomic Geology, v. 67, p. 281-301.
Hollister, V.F., 1975, An appraisal o f the nature and source o f porphyry
copper deposits: Minerals Science and Engineering, v. 7, p. 225-233.
H unt.J.P., 1991, Porphyry copper deposits: Econom ic G eology Mono­
graph 8, p. 192-206.
Hutton, D.H.W., 1988, Granite em placem ent mechanisms and tectonic
controls: Inferences from deformation studies: Royal Society o f Edin­
burgh Transactions, Earth Sciences, v. 79, p. 245-255.
— 1997, Syntectonic granites and the principle o f effective stress: A gen­
eral solution to the space problem?, in Bouchcz, J.-L., Stephens. W.E.,
Hutton, D.H.W., eds., Granite: From melt segregation to emplacement
fabrics: Dordrecht, Kluwer Academic Publishers, p. 189-197.
Ida, Y., 1999, Effects o f the crustal stress on the growth o f dikes: Conditions
o f intrusion and extrusion o f magma: Journal o f Geophysical Research,
v. 104, p. 17,897-17,909.
Jaeger, J.C., and Cook, N.G.W., 1979, Fundamentals o f rock mechanics,
3rd ed.: London, Chapman and Hall, 593 p.
Jarrard, R.D., 1986, Terrane motion by strike-slip faulting o f forcarc slivers:
Geology, v. 14, p. 780-783.
Johnson. AM ., and Pollard, D.D., 1973, Mechanics o f growth o f som e lac-
colithic intrusions in the Henry Mountains, Utah, I. Held observations,
Gilbert's model, physical properties, and flow o f the magma: Tcctono-
physics, v. 18, p. 261-309.
Jordan, T.E., Allmendinger, R.W., Damanti, J.F., and Drake, R.E., 1993,
Chronology o f m otion in a com plete thrust belt: T he Precordillera,
30-31°S, Andes Mountains: Journal o f Geology, v. 101, p. 135—156.
Keith, J.D., Christiansen, E.H., M aughan, D.T., and Waite, K A ., 1998,
The role o f mafic alkaline magmas in fclsic porphyry-Cu and Mo sys­
tems, in Lentz, D.R., ed., Mineralized intrusion-related skarn systems:
M ineralogies! Association o f Canada Short Course Series, v. 26,
p . 211-243.
Kirkham, R.V., and Sinclair, W.D., 1996, Porphyry copper, gold, molybde­
num, tungsten, tin, silver, in Eckstrand, O.R., Sinclair, W.D., and
Thorpe, R.I., eds., Geology o f Canadian mineral deposits: Geological
Survey o f Canada, Geology of Canada, no. 8, p. 421-446.
Koidc, LI., and Bhattachariji, S„ 1975, Formation o f fractures around mag*
matic intrusions and their role in ore localization: Economic Geology,
v. 70, p. 781-799.
Langcrfeki, H., 1964, T he pebble dikes at El Salvador: Unpublished
report to Anaconda, 64 p,
Linde, A.T., Sacks, I.S.,Johnston, M.J.S., Hilt, D.P., and Bilham, R.G., 1994,
Increased pressure from rising bubbles as a mechanism for remotely
triggered seismicity: Nature, v. .371, p. 4QJM10.
Lindgrcn, W„ 1905, The copper deposits o f the Clifton-Morcnci district,
Arizona: U.S. Geological Survey Professional Paper 43,375 p.
— 1933, Mineral deposits, 4th ed.: New York, McGraw-Hill, 930 p.
— 1937, Succession o f minerals and temperatures o f formation in ore
deposits o f magmatic affiliation: American Institute o f Mining and Met­
allurgical Engineers Transactions, v. 126, p. 356-376.
Lindsay, D.D., Zentilli, M., and Rojas dc la Rivera, J„ 1995, Evolution o f an
active ductile to brittle shear system controlling mineralization at the
Chuquicamata porphyry copper deposit, northern Chile: International
Geology Review, v, 37, p. 945-958.
Loucks, R.R., and Mavrogcncs, J A ., 1999, Gold solubility in supercritical
hydrothermal brines measured in synthetic fluid inclusions: Science,
v. 284, p. 2159-2163,
Lowell, J.D., and Guilbert, J.M., 1970, Literal and vertical altcration-min-
cralization zoning in porphyry copper ore deposits: Economic Geolog);
v, 65, p. 373-408.
MacDonald, G.D., and Arnold, L.C., 1994, Geological and geochemical
zoning o f the Grnsbcrg igneous complex, Irian Jaya, Indonesia: Journal
o f Geochemical Exploration, v. 50, p. 143-178.
Maksaev, V„ and Zentilli, M„ 1988, Marco m ctalogcnico regional dc los
rnegadepositos dc lipo porfido cuprifcro del nortc grande de Chile:
Santiago, V Congrcsso Gcologico Chilcno, Tomo 1, p. B181—212.
Marsh, T.M., Einaudi, M.T., and McWilliams, M., 1997, 10A r /s<JAr
geochronology o f Cu-Au and Au-Ag mineralization in the Potrcrillos
district, Chile: Economic Geology, v. 92, p. 784-806.
Matthews, S.J., Sparks, R.SJ,, and Gardeweg, M.C., 1995, The relationships
between magma m ixing and volatile behaviour at Lascar volcano
(23“22'S, 67°44'W), northern Chile: Significance for the formation o f
copper sulphide and magnetite-apatite orcbodics, in Clark, A.H., ed.,
Giant ore deposits II, Controls o n the Scale o f O rogenic Magmatic-.
Hydrothermal Mineralization: Kingston, Ontario, QminEx Associates
and Q ueen’s University, p. 146-181.
McCaffrey, K.J.W., 1992, Igneous em placem ent in a iranspressive shear
zone: O x Mountains igneous complex: Journal o f the Geological Soci­
ety, London, v. 149, p. 221-235.
McCaffrey, K.J.W., and Petford, N., 1997, Are granidc intrusions scale vari­
ant?: Journal o f the Geological Society, London, v. 154, p. 1-4.
McCandlcss, T.E., and Ruiz, J., 1993, Rhenium-osmium evidence for
regional mineralization in southwestern North America: Science, v. 261,
p .1282-1286.
Muntcan, J.L., 1998, Magmatic-hydrothermal gold deposits o f the Mari-
cunga belt, northern Chile: Unpublished Ph.D. thesis, Stanford, Stan­
ford University, 400 p.
Myers, J.S., 1975, Cauldron subsidence and fluidization: Mechanism o f
intrusion o f the coastal batholith o f Peru into its own volcanic ejecta:
Geological Society o f America Bulletin, v. 86, p. 1209-1220.
Nakamura, K., 1977, Volcanoes as possible indicators o f tectonic stress ori­
entation— Principal and proposal: Journal o f Volcanology and Geot­
hermal Research, v. 2, p. 1-16.
Nielson, R.L., 1976, Recent developments in the study o f porphyry copper
geology—A review, in Sutherland Brown, A , ed.. Porphyry Deposits o f
the Canadian Cordillera: Canadian Institute o f Mining and Metallurgy
Special Volume, 15, p. 487-500.
Olson, S.F., 1989, The stratigraphic and structural setting o f the Potrcrillos
porphyry copper district. Northern Chile: Rcvista G cologico de Chile;
v. 16, p. 3-29.
O satenko, M.J., and Jones, M.B., 1976, Valley Copper, in Sutherland
Brown, A , ed.. Porphyry Deposits o f the Canadian Cordillera: Canadian
Institute o f Mining and Metallurgy Special Volume, 15, p. 130-143.
Pankhurst, R.J., Hole, M.J., and Brook, M., 1988, Isotope evidence for the
origin o f A ndean granites: Royal Society o f Edinburgh Transactions,
Earth Sciences, v. 79, p. 123-133.
Pardo-Casas, F., and Molnar, P., 1987. Relative motion o f the Nazca (Faral-
lon) and South Ancrican plates since Late Cretaceous time: Tectonics,
v, 6, p. 233-248.
Park, R.B., 1983, Foundations o f Structural Geology: Glasgow, Blackic, 135 p.
Paterson, S.R., and Fowler, T.K., 1993, Extensional pltiton-einplaccmciu
models: Do they work for large plutonic complexes?: Geology, v. 21,
p . 781-784.
Paterson, S.R., and Miller, R.B., 1998a, Magma emplacement during arc-
perpendicular shortening: An exam ple from the Cascades crystalline
core, Washington: Tectonics, v. 17, p, 571-586.
1998b, Mid-crustal magmatic sheets in the Cascade Mountains, Wash­
ington: Implications for magma ascent: Journal o f Structural Geology,
v. 20. p. 1345-1364.
Paterson, S.R., and Schmidt, K.L., 1999, Is there a close spatial relationship
between faults and plutons?: Journal o f Structural Geology, v. 21,
p. 1131-1142.
Paterson, S.R., Fowler, T.K., and Miller, R.B., 1996, Pluton emplacement in
arcs: A crustal-scale exchange process: Geological Society o f America
Special Paper, 315, p. 115-123.
Petford, N.. 1996, Dykes or diapirs?: Geological Society o f America Special
Paper, 315, p. 105-114.
Petford, N., Kerr, R.C., and Lister, J.R., 1993, Dike transport o f granitoid
magmas: Geology, v. 21, p. 845-848.
Petford, N., Lister,J.R., and Kerr, R.C., 1994, The ascent o f fclsic magmas
in dykes: Lilhos, v. 32, p. 161-168.
Phillips, W.J., 1973, Mechanical effects o f retrograde boiling and its prob­
able importance in the formation o f som e porphyry ore deposits: Trans­
actions o f the Institute o f Mining and Metallurgy, v. 82, p. 90-98.
Pierce, F.W., and Bolm, J.G., eds., 1995, Porphyry copper deposits o f the
American Cordillera: Tucson, Arizona Geological Society Digest, v. 20,
656 p.
Pitcher, W.S., 1985, A multiple and composite batholith, in Pitcher, W.S.,
Atherton, M.P., Cobbing, EJ., and Bcckinsalc, R.D., eds., Magmatism at
a Plate Edge; The Peruvian Andes: Glasgow, Blackic, p. 93-101.
Pollard, D.D., and Johnson, AM ., 1973, Mechanics o f growth o f some lac-
colithic intrusions in the Henry Mountains, Utah, II. Bending and fail­
 ure o f overburden layers and sill formation: Tectonophysics, v. 18,
p. 311-33-1.
Presnall, R.C., 1997, Structural controls on the plutonism and metallogeny
in the Wasatch and Oquirrh Mountains, Utah: Society o f Econom ic
Geologists Guidebook Series, v. 29, p. 1-10.
Ramos, VA., Cegarra, M., and Cristallini, E., 1996, Cenozoic tectonics o f
the High Andes o f west-central Argentina (30-36°S latitude): Tectono-
physics, v. 259. p. 183-200.
Ransome, F.L.. 190*1. Geology and ore deposits o f the Bisbee Quadrangle,
Arizona: U.S. Geological Survey Professional Paper 21,168 p.
------ 1919, The copper deposits o f Ray and Miami, Arizona: U.S. Geologi­
cal Survey Professional Paper 115,192 p.
Rehrig. WA., and Hcidrick, T.L., 1972, Regional fracturing in Laramidc
stocks o f Arizona and its relationship to porphyry copper mineraliza­
tion: Economic Geology, v. 67, p. 198-213.
------ 1976, Regional tectonic stress during the Laramidc and late Tertiaiy
intrusive periods, Basin and Range province, Arizona: Tucson, Arizona
Geological Society Digest, no. 10, p. 205-228,
Reynolds, P., Ravenhursl, C., Zctuilli, M„ and Lindsay, D., 1998, High-pre­
cision •,0A r /39Ar dating o f two consecutive hydrothermal events in the
Chuquicamata porphyry copper system, Chile: Chemical Geology,
v. 148, p. 45-60.
Reynolds, T.J., and Beane, R.E., 1985, Evolution o f hydrothermal fluid
characteristics at the Santa Rita, New Mexico, porphyry copper deposit:
Economic Geology, v. 80, p. 1528-1347.
Richard, K., and Courtright,J.H., 1938, Geology ofToquepala, Peru: Min­
ing Engineering, v. 10, p. 262-266.
Richards, J.P., 2000a, Is there a close spatial relationship between faults
and phiions? A discussion: journal o f Structural Geology, in press.
------ 2000b, Lineaments revisited: Society o f Economic Geologists Newslet­
ter, no. 42, p. 1,14-20.
Richards, J.P., Boyce, A.J., and Pringle, M.S., 2001, Geological evolution o f
the Escondida area, northern Chile: A model for spatial and temporal
localization o f porphyry Cu mineralization: Econom ic Geology, in
press,
Roedder, E., 1971, Fluid inclusion studies on the porphyry-type ore
deposits at Bingham, Utah, Butte, Montana, and Climax, Colorado:
Economic Geology, v. 66, p. 98-120.
Rose, AAV., 1970, Zona! relations o f wall rock alteration and sulfide distrib­
ution at porphyry copper deposits: Economic Geology', v. 65, p. 920-936.
Roydcn, L.H., 1993, The tectonic expression slab pull at continental con­
vergent boundaries: Tectonics, v. 12, p. 303-325.
Sahaglan, D.L., and Prousscvitch, AA ., 1992, Bubbles in volcanic systems:
Nature, v. 359, p. 485.
Salfity, J A ., 1985, L incam iem os transversalcs a! rumbo and ino en el
noroeste Atgcntino: Antofagasta, August 1985, IV Congrcso Gcoiogico
Chilcno, Part 2, 119-137.
Sapie, B., and Cloos, M., 1994, Strike-slip fault control on emplacement of
the Grasbcrg intrusive com plex, G um ing Bijih (Ertsberg) Cu-Au dis­
trict, Irianjayn, Indonesia (West New Guinea) [abs.): Geological Society
o f America Abstracts wit!) Programs, v. 26, p. A-186.
Sawkins, F.J., 1990, Meta! deposits in relation to plate tectonics, 2nd ed.:
Berlin, Springcr-Verlag, 461 p.
Schmidt, K.L., and Paterson, S.R., 2000, Analysis fails to find coupling
between deform ation and magmniism: EOS, v. 81, no.18, p. 197.
202-203.
Schroeter, T.G., ed., 1995, Porphyry deposits o f the northwestern
Cordillera o f North America: Canadian Institute o f Mining. Metallurgy,
and Petroleum. Special Volume 46,888 p.
Shaw, H.R., 1980, The fracture mechanisms o f magma transport from the
mantle to the surface, in Hargraves, R.B., ed.. Physics o f Magmatic
Processes: Princeton University Press, p. 201-26*1.
Sheppard, S.M.F., and Gustafson, L.B., 1976, Oxygen and hydrogen iso­
topes in the porphyry copper deposit at El Salvador, Chile: Economic
Geology, v. 71, p. 1549-1559.
Sheppard, S.M.F., Nielsen, R.L., and Taylor, H.P., Jr„ 1969, Oxygen and
hydrogen isotope ratios o f clay minerals from porphyry copper
deposits: Economic Geology, v. 64, p, 755-777.
Sheppard, S.M.F., Nielsen, R.L., and Thylor, H.P.,Jr., 1971, Hydrogen and
oxygen isotope ratios in minerals from porphyry copper deposits: Eco­
nomic Geology, v. 66, p. 515-542.
Shinohara, H., 1994. Exsolution o f immiscible vapor and liquid phases
from a crystallizing silicate melt: Implications for chlorine and metal
transport. Geochimica ct Cosmochimica Acta, v. 58, p. 5215-5221.
Sibson, R.H., 1985, A note on fault reactivation:Journal o f Structural Geol­
ogy, v. 7, p. 751-754.
—— 2000, A brittle failure m ode plot defining conditions for high-flux
flow: Economic Geology, v. 95, p. 41—18.
—— 2001, Seism ogenic framework for hydrothermal transport and ore
deposition: Reviews in Economic Geology, v. 14. p. 25-50.
Sillitoc, R;H., 1972, Relation o f metal provinces in western Americas to
subduction o f oceanic.lithosphcrc: Geological Society o f America Bul­
letin. v. 83. p. 813-818.*
—— 1973, The tops and bottoms of porphyry copper deposits: Economic
Geology, v. 68, p. 799-815.
—— 1985, Ore-reiatcd breccias in volcanoplutonic arcs: Economic Geol-
ogy, v. 80, p. 1467-1514.
— 1988, Epochs o f intrusion-related cop p er mineralization in the
Andes: journal o f South American Earth Sciences, v. 1, p. 89-108.
— — 1992, Gold and copper metallogeny o f the central Andes—Past, present,
and future exploration objectives. Economic Geology, v. 87, p. 2205-2216.
-i. ■■1993, Gold-rich porphyry copper deposits: G eological m odel and
exploration implications, in Kirkhain, R.V., Sinclair, W.D., Thorpe, R.I.,
and Duke.J.M., eds., Mineral Deposit Modeling: Geological Association
o f Canada Special Paper 40, p. 465—178.
—— 1994, Erosion and collapse o f volcanoes: Causes o f telescoping in
intrusion-centered ore dc|>osits: Geology, v. 22, p. 945-948.
— 1997, Characteristics and controls o f the largest porphyry copper-gold
and epithermal gold deposits in the ciicum-Pacific region: Australian
Journal o f Earth Sciences, v. 44. p. 373-388.
Skcwcs, MA , and Stern, C.R., 1994, Tectonic trigger for die formation o f
late Miocene Cu-rich breccia pipes in the Andes o f central Chile: Geol­
ogy, v. 22, p. 531-554.
— 1995, Genesis o f the giant Laic Miocene to Pliocene copper deposits
o f central Chile in the context o f Andean magmatic and tectonic evolu­
tion: International Geology Reviews, v. 37, p. 893-903.
Solom on, M.. 1990, Sulrduciiou, arc reversal, and the origin o f porphyry
copjvergold deposits in bland arcs: Geology, v. 18. p. 630-633.
Stanley. CR., Holbck, P.M., 1-Iuyck, H.L.O., Lang, J.R., Prcto, VA.G.. Blower,
S. J., and Botia»o,J.C, 1995, Geology o f the Copper Mountain alkalic cop­
per-gold porphyry deposits, Princeton, British Columbia, in Schroeter,
T. G., ed.. Porphyry Deposits o f the Northwestern Cordillera o f North
America: Canadian Institute o f Mining, Metallurgy; and Petroleum Special
Volume 46, p. 537-564.
Sturtcvant, B., Kanamori, H.. and Brodsky, E.E., 1996, Seismic triggering
by rectified diffusion in geotherm al systems: Journal o f Geophysical
Research, v. 101, no. B11, p. 25,269-25,282.
Sutherland Brown, A., ed ., 1976, Porphyry Deposits o f the Canadian
Cordillera: Canadian Institute o f Mining and Metallurgy Special Vol­
ume 15,510 p.
Takada, A , 1994, The influence o f regional stress and magmatic input on
styles o f monogcnetic anti polygcnetic volcantsm: Journal o f Geophysi­
cal Research, v. 99, p. 13,563-13.573.
Taylor, H.P.,Jr., 1974, The application o f oxygen and hydrogen isotope
studies to problems o f hydrothermal alteration and ore deposition: Eco­
nomic Geology, v. 69, p. 843-883.
Theodore, T.G., 2000, Geology o f pluton-rclatcd gold mineralization at Bat­
tle Mountain, Nevada: Tucson, Arizona, University o f Arizona and U.S.
Geologic.nl Survey Center for Mineral Resources Monograph 2, in press,
T hom pson, J.F.H., 1995, Exploration and research related to porphyry
deposits, in Schroeter, T.G., ed., Porphyry Deposits o f the Northwestern
Cordillera o f North America: Canadian Institute o f Mining, Metallurgy,
and Petroleum Special Volume 46, p. 857-870.
Tikoff, B., and Teyssier, C., 1992, Crustal scale en echelon ‘P-shear’ ten-
sional bridges: A possible solution to the batholithic room problem:
Geology, v. 20. p, 927-930.
Tikoff, B., do Saint Blanquat, M., and Teyssier, C., 1999, Translation and
the resolution o f the pluion space problem: journal o f Structural Geol­
ogy', v. 21, p. 1109-1117.
Tilton, G.R., Poliak, R.J., Clark, A.H., and Robertson, R.C.R., 1981, Iso­
topic com position o f Pb in central Andean ore deposits: Geological
Society o f America Memoir 154, p. 791-816.
Tilley, S.R., eel., 1982, Advances in G eology o f the Porphyry Copper
Deposits, Southwestern North America: Tucson, University o f Arizona
Press, 560 p.
Titlcy, S.R., 1993, Characteristics o f porphyry copper occurrence in the
American Southwest, in Kirkham, R,V„ Sinclair, W.D., Thorpe, R.I., and
Duke, J.M., cds„ Mineral Deposit Modeling: Geological Association of
Canada Special Paper 40, p. 433—164.
Titlcy, S.R., and Beane, R.K., 1981, Porphyry copper deposits. Pari 1, Geo­
logic sellings, petrology, and leciogcnesis: Econom ic G eology 75th
Anniversary Volume, p. 214-235.
Tilley, S.R., and Hcidrtck, T.L., 1978, Intrusion and fracture styles o f some
mineralized porphyry systems o f the southwestern Pacific and their rela­
tionship to plate interactions: Economic Geology, v. 73, p. 891-903.
Titlcy, S.R., and Hicks, C.L., cds., 1966, Geology o f the Porphyry Copper
Deposits, Southwestern North America: Tucson, University o f Arizona
Press, 287 p.
Titlcy, S.R., Thompson, R.C„ Haynes, F.M., Manskc, S„ Robison, L.C., and
White, J.L., 1986, Evolution o f fractures and alteration in the Sicrrita-
Esperanza hydrothermal system, Pima County. Arizona: Econom ic
Geology, v. 81, p. 343-370.
Tom linson, A.J., 199-1, R claciones entre el porfido cuprifero y la falla
iuversa de ia mina de Potrcrillos: Un caso de intrusion sime-ctonica:
Concepcion, 7 Congrcsso Geologico Chilcno, Adas v, 2, p. 1629-1633.
Tomlinson, A.J., Mpodozis, C., Cornejo, P„ and Ramirez, C.F., 1993, Struc­
tural geolog)' o f the Sierra Castillo-Agua Amarga fault system, Pre­
cordillera o f Chile, El Saivador-Potrerillos: Second Symposium Interna­
tional d e G fodynam iquc Andinc, Oxford: Paris, Edittoncs de
rORSTROM. Collection Colioques et Scminaires, p. 259-262.
Uycda, S., and Nisitiwnki, C., 1980, Stress field, metallogcnesis, and mode
o f subduction, in Strangway, D.W., cd., T he continental cn isi and its
mineral deposits: Geological Society o f Canada Special Publication 20.
p. 323-340.
Vila, T., Sillitoc, R.H., BctzhoId.J., and Vitcri, E., 1991, The porphyry gold
deposit at Mane, northern Chile: Economic Geolog)’, v. 86, p, 1271-1286.
W alker,JA, Moulds, T.N., Zenlilti, M., and Feigcnson, M.D., 1991, Spatial
and temporal variations in volcanics o f the Andean Central Volcanic
Zone (26 to 28°S), in Harmon, R.S., and Rapeln, C.W., eds„ Andean
Magmalism and its Tectonic Setting: Boulder, Colorado, Geological
Society o f America Special Paper 265, p. 139-155.
Whitney, J,A., 1975, Vapor generation in a quartz m onzonite magma, a
synthetic m odel with application to porphyry ore deposits: Economic
Geology, v. 70, p. 3-16-358.
Wilkinson, W.H., Jr., Vega, L.A., and Titlcy, S.R., 1982, The geology and
ore deposits at Mineral Park, Mohave County, Arizona, in Titlcy, S.R.,
cd., Advances in Geology o f the Porphyry Copper Deposits, Southwest­
ern North America: Tucson, University o f Arizona Press, p. 523-5-11.
Williams, TJ„ Candela, P A , and Piccoli, P.M., 1995, The partitioning o f
copper between silicate melts and two-phase aqueous fluids: An experi­
mental investigation at 1 kbar, 800°C, and 0.5 kbar, 85Q°C: Contribu­
tions to Mineralogy and Petrology, v. 121, p. 388-399.
Wilson, J„ and G rocott.J., 1999, T he em placem ent o f the granitic Las
Tazas com plex, northern Chile: The relationship between local and
regional strain: Journal o f Structural Geology, v. 21, p. 1513-1523.
Yoshinobu, A,S„ Okaya, D A , and Paterson, S.R., 1998, Modeling the ther­
mal evolution o f fault-controlled magma emplacement model: Implica­
tions for the solidification o f granitoid plutons: Journal o f Structural
Geology, v. 20, p. 1205-1218.
Zwcng, P.L., and Clark, A.H., 1995, Hypogcnc evolution o f the Toqucpala
porphyry coppcr-molylidenum deposit, Moqucgua. southeastern Peru,
in Pierce, F.W., and BoIm.J.G., cds., Porphyry Copper Deposits o f the
American Cordillera: Tucson, Arizona, Geological Society Digest, v. 20.
p. 566-612.