OrogenicGold Asjan2005 WorkshopUFRGS Brazil PDF

Exploration for Orogenic Gold Deposits, Porto Allegre, Brazil, January 2005
Contents
Contents....................................................................................................................................................................... 1
1. GOLD DEPOSITS IN METAMORPHIC BELTS: OVERVIEW OF CURRENT

UNDERSTANDING, OUTSTANDING PROBLEMS, FUTURE RESEARCH AND EXPLORATION
SIGNIFICANCE. .........................................................................................................................................1
Abstract ....................................................................................................................................................................... 1
1.1. Introduction......................................................................................................................................................... 2
1.2. Potential Diversity of Gold Deposit Types in Metamorphic Belts ........................................................ 4
1.3. Orogenic Gold Deposits: The Most Common Type?.................................................................................... 5

1.3.1 Definition and distinction from other gold deposit styles ............................................................................ 6
1.3.2 Current knowledge......................................................................................................................................... 8
1.3.3 Outstanding problems .................................................................................................................................. 12
1.4. Intrusion-Related Gold Deposits: A Coherent Group?.............................................................................. 14

1.4.1 Distinction from orogenic gold deposits..................................................................................................... 16
1.4.2 Current knowledge....................................................................................................................................... 17
1.5. Gold Deposits with Atypical Metal Associations: Where Do They Fit? ............................................ 20
1.5.1 Definition and contrasts with orogenic gold deposits ........................................................................ 21
1.5.2 Probable modified porphyry/epithermal systems....................................................................................... 21
1.5.3 Probable modified VHMS or submarine epithermal systems ................................................................... 23
1.6. Giant Gold Deposits in Metamorphic Belts?.......................................................................................... 24

1.6.1 Affiliation of giant gold deposits in metamorphic belts............................................................................. 25
1.6.2 Definition of special features....................................................................................................................... 25
1.7. Summary of Outstanding Problems for Classification and Genetic Models .................................... 27
1.7.1 Classification ................................................................................................................................................ 28
1.7.2 Orogenic gold deposits................................................................................................................................. 28
1.7.3 Intrusion-related gold deposits..................................................................................................................... 28
1.7.4 Gold deposits with atypical metal associations........................................................................................... 28
1.7.5 Temporal distribution of gold deposits in terms of crustal evolution........................................................ 29
1.8.1 Integrated province-scale and deposit-scale research................................................................................. 29
1.8.2 Research on fluid source.............................................................................................................................. 29
1.8.3 Research on fluid flow conduits .................................................................................................................. 30
1.8.4 Research on background gold concentrations ............................................................................................ 31
1.8.5 Transport and deposition of gold................................................................................................................. 31
1.8.6 Research on temporal distribution of gold deposits ................................................................................... 31
1.8.7 Development of four-dimensional models ................................................................................................. 32
1.9.1 Critical parameters for giant gold deposits ................................................................................................. 32
1.9.2 Better temporal and tectonic models ........................................................................................................... 32
1.9.3 Improved genetic models for gold deposits in metamorphic belts ............................................................ 32
1.9.4 Improved understanding of deposit overprinting ....................................................................................... 33
1.10. Acknowledgements......................................................................................................................................... 34
Page 1
1.11. References........................................................................................................................................................ 34
1.12. Appendix.......................................................................................................................................................... 46
2. OROGENIC GOLD DEPOSITS: A PROPOSED CLASSIFICATION IN THE CONTEXT OF

THEIR CRUSTAL DISTRIBUTION AND RELATIONSHIP TO OTHER GOLD DEPOSIT TYPES....1
Abstract ....................................................................................................................................................................... 1
2.1. Introduction......................................................................................................................................................... 1
2.2. Definition of So-Called Mesothermal Gold Deposits .............................................................................. 2

2.2.1. Geological characteristics ............................................................................................................................. 2
2.2.1.1. Geology of host terranes........................................................................................................................ 2
2.2.1.2. Deposit mineralogy................................................................................................................................ 2
2.2.1.3. Hydrothermal alteration......................................................................................................................... 3
2.2.1.4. Ore fluids................................................................................................................................................ 3
2.2.1.5. Structure ................................................................................................................................................. 3
2.2.2. Tectonic setting and timing of `mesothermal' vein emplacement .............................................................. 4
2.2.3. Crustal environment of `mesothermal' gold deposition ............................................................................ 12
2.2.4. Comparisons with other lode-gold deposit types ...................................................................................... 15
2.3. Problem of Nomeclature.................................................................................................................................. 18
2.4. Proposed Classification.................................................................................................................................... 20
2.5. Acknowledgements........................................................................................................................................... 21
2.6. References .......................................................................................................................................................... 21
3. ROTUND VS SKINNY OROGENS : WELL NOURISHED OR ANOREXIC GOLD? ............29
ASTRACT ................................................................................................................................................................ 29
3.1. Introduction....................................................................................................................................................... 29
3.2. Orogenic Gold Deposits In Space And Time................................................................................................ 30
3.3. Gold Deposits And Evolution Of Early Earth ............................................................................................. 33
3.4. Gold Distributions Under A More Modern-Style Plate Tectonics...................................................... 34
3.5. Conclusions: Control On Gold Distribution By Pattern Of Crustal Growth................................... 36
3.6. References .......................................................................................................................................................... 36

Konstantinov, M., Dankovtsev, R., Simkin, G. and Cherkasov, S., 1999, Deep structure ................ 37
4. TECTONIC SETTING OF SYNOROGENIC GOLD DEPOSITS OF THE PACIFIC RIM...........39
Abstract ..................................................................................................................................................................... 39
4.1. Introduction....................................................................................................................................................... 40
4.2 Tertiary synorogenic gold deposits of the Pacific Rim................................................................................. 47

4.2.1. Gulf of Alaska accretionary prism ............................................................................................................. 48
4.2.2. Near-arc and within-arc deposits of southern Alaska ............................................................................... 49
4.2.2.1. Juneau gold belt ................................................................................................................................... 50
4.2.2.2. Valdez Creek district ........................................................................................................................... 50
Page 2
4.2.2.3. Willow Creek district........................................................................................................................... 52
4.3. Mesozoic synorogenic gold deposits of the Pacific Rim .............................................................................. 54

4.3.1. Mid-Cretaceous deposits of the northern North American Cordillera..................................................... 54
4.3.1.1. Interior Alaska...................................................................................................................................... 54
4.3.1.2. Southwestern British Columbia and the south part of southeastern Alaska..................................... 56
4.3.2. Middle Jurassic to Early Cretaceous deposits of western North America ............................................... 57
4.3.2.1. Canadian Cordillera ............................................................................................................................. 57
4.3.2.2. Conterminous United States................................................................................................................ 58
4.3.3. Eastern Eurasia ............................................................................................................................................ 61
4.3.4. Otago, South Island, New Zealand............................................................................................................. 62
4.4. Palaeozoic synorogenic gold deposits of the Pacific Rim............................................................................ 62

4.4.1. Lachlan fold belt.......................................................................................................................................... 63
4.4.2. Buller terrane, Westland, New Zealand..................................................................................................... 65
4.4.3. New England Fold Belt............................................................................................................................... 65
4.5. Discussion........................................................................................................................................................... 66
4.6. Acknowledgements........................................................................................................................................... 68
4.7 References ........................................................................................................................................................... 68
5. ARCHEAN OROGENIC LODE GOLD DEPOSITS ..................................................................75
Abstract ..................................................................................................................................................................... 75
6. HYDROTHERMAL TRANSPORT AND DEPOSITIONAL PROCESSES IN ARCHEAN

LODE-GOLD SYSTEMS: A REVIEW ....................................................................................................77
Abstract ..................................................................................................................................................................... 77
6.1. Introduction....................................................................................................................................................... 77
6.2. Physicochemical conditions of the ore fluids ................................................................................................ 78
6.3. Gold transport in hydrothermal solutions.................................................................................................... 81

6.3.1. Gold transport at amphibolite fades conditions: Thermodynamic data.................................................... 82
6.4. Gold solubility and speciation in lode-gold ore fluids ................................................................................. 84
6.5. Ore precipitation mechanisms........................................................................................................................ 86

6.5.1. P-T changes ................................................................................................................................................. 87
6.5.2. Phase separation .......................................................................................................................................... 87
6.5.3. Fluid-rock interaction.................................................................................................................................. 89
6.5.4. Surface chemistry-driven processes ........................................................................................................... 89
6.6. Conclusions........................................................................................................................................................ 90
6.7 Acknowledgements............................................................................................................................................ 91
6.8 References ........................................................................................................................................................... 91
7. ALTERATION AND PRIMARY GEOCHEMICAL DISPERSION ASSOCIATED WITH THE

BULLETIN LODE-GOLD DEPOSIT, WILUNA, WESTERN AUSTRALIA .........................................97
Abstract ..................................................................................................................................................................... 97
Page 3
7.1. Introduction....................................................................................................................................................... 98
7.2. Geological setting .............................................................................................................................................. 99
7.3. Mapping, sampling and analytical procedures.......................................................................................... 102

7.3.1. Alteration mapping.................................................................................................................................... 102
7.3.2. Sampling .................................................................................................................................................... 102
7.3.3. Analytical procedures................................................................................................................................ 102
7.4. Host rocks ........................................................................................................................................................ 103

7.4.1. Petrography................................................................................................................................................ 103
7.4.2. Geochemistry............................................................................................................................................. 104
7.5. Hydrothermal alteration ............................................................................................................................... 107

7.5.1. Alteration zoning patterns......................................................................................................................... 108
7.5.2. Dolomite-present zoning patterns............................................................................................................. 108
7.5.2.1. Chlorite-calcite zone .......................................................................................................................... 109
7.5.2.2. Calcite-dolomite zone........................................................................................................................ 110
7.5.2.3. Sericite zone ....................................................................................................................................... 110
7.5.2.4. Dolomite-sericite zone....................................................................................................................... 111
7.5.3. Dolomite-absent zoning patterns.............................................................................................................. 111
7.6. Chemical changes related to alteration ....................................................................................................... 111

7.6.1. Mass transfer at Bulletin ........................................................................................................................... 112
7.7. Discussion......................................................................................................................................................... 115
7.8. Geochemical dispersion patterns.................................................................................................................. 119

7.8.1. Background thresholds for Au, Ag, As, Sb, Te and W........................................................................... 119
7.8.2. Primary dispersion defined by Au, Ag, As, Sb, Te and W..................................................................... 122
7.8.3. Major- and related trace-element dispersion patterns.............................................................................. 123
7.8.3.1. Carbonation........................................................................................................................................ 124
7.8.3.2. Sericitisation....................................................................................................................................... 124
7.8.3.3. Anomalies defined by Rb, S and Y .................................................................................................. 124
7.8.3.4. An extensive lateral geochemical vector .......................................................................................... 124
7.9. Conclusions...................................................................................................................................................... 125
7.10 Acknowledgements........................................................................................................................................ 127
7.11 References....................................................................................................................................................... 127
8. LATE-KINEMATIC TIMING OF OROGENIC GOLD DEPOSITS AND SIGNIFICANCE FOR

COMPUTER-BASED EXPLORATION TECHNIQUES WITH EMPHASIS ON THE YILGARN
BLOCK, WESTERN AUSTRALIA. ......................................................................................................131
Abstract ................................................................................................................................................................... 131
8.1. Introduction.................................................................................................................................................... 132
8.2. Geological characteristics of orogenic lode-gold deposits................................................................... 135

8.2.1 Geology of host terranes.................................................................................................................... 135
8.2.2 Deposit mineralogy and alteration .................................................................................................... 135
8.2.3 Structure.............................................................................................................................................. 136
8.2.4 Timing of gold mineralization........................................................................................................... 136
8.2.5 Yilgarn gold deposits as orogenic gold deposits ...................................................................................... 137
8.2 Timing of relationships in orogenic gold deposits....................................................................................... 137

8.3.1 Relative timing from structural evidence.......................................................................................... 137
Page 4
8.3.2 Geochronological evidence ............................................................................................................... 138

8.3.3 Geochronology : Yilgarn Block........................................................................................................ 139
8.3.4 Geochronology: Other Terranes Worldwide.................................................................................... 142
8.3.5 Implications of late timing................................................................................................................. 143
8.4. Structural geometry of orogenic lode-gold deposits................................................................................. 143
8.5. Stress mapping and its application to orogenic gold deposits.................................................................. 145
8.5.1. Introduction................................................................................................................................................ 146
8.5.2 Stress mapping principles ......................................................................................................................... 146
8.5.3 Computer program .................................................................................................................................... 147
8.5.4 Input for stress mapping .................................................................................................................... 148
8.5.5 Example of two-dimensional stress mapping: Kalgoorlie Terrane................................................. 148
8.5.5.1 Strategy................................................................................................................................................ 148
8.5.5.2. Application......................................................................................................................................... 149
8.5.5.3. Results ................................................................................................................................................ 150
8.5.6 Discussion........................................................................................................................................... 151
8.6. Computer-based prospectivity mapping............................................................................................... 152

8.6.1. Introduction............................................................................................................................................... 152
8.6.2 Principles of prospectivity mapping.......................................................................................................... 153
8.6.2.1 Step One: Identification of Spatial Relationships ............................................................................. 153
8.6.2.2 Step Two: Quantification of Identified Spatial Relationships.......................................................... 154
8.6.2.3 Step Three: Integration of Identified Spatial Relationships............................................................. 154
8.6.3 Prospectivity Analysis of the Kalgoorlie Terrane ............................................................................ 155
8.6.3.1 Identification of Spatial Relationships............................................................................................... 155
8.6.3.2 Quantification and Integration of Identified Spatial Relationships.................................................. 158
8.7. Comparison of Stress and Prospectivity Maps .......................................................................................... 160
8.8. Potential of Shape Analysis and Artificial Intelligence in Prospectivity Mapping............................. 161
8.9. Conclusions................................................................................................................................................ 162
8.10 Acknowledgements........................................................................................................................................ 164
8.11 References....................................................................................................................................................... 164
9. COMPUTER AIDED STRUCTURAL TARGETING IN MINERAL EXPLORATION: TWO

AND THREE-DIMENSIONAL STRESS MAPPING............................................................................173
Abstract ................................................................................................................................................................... 173
9.1. Introduction..................................................................................................................................................... 173
9.2 Methods And Theoretical Considerations................................................................................................... 174

9.2.1 Fluid focussing ........................................................................................................................................... 174
9.2.2 Stress-fluid pressure relationship at fracture initiation ............................................................................. 175
9.3 Stress Mapping................................................................................................................................................. 176

9.3.1 Computer programs.................................................................................................................................... 177
9.3.2 Input for stress mapping............................................................................................................................. 178
9.3.3 Example of two-dimensional stress mapping: Lennard Shelf ................................................................. 178
9.3.4 Example of three-dimensional stress mapping: the Granny Smith mine ............................................... 180
9.3.4.1 GEOLOGICAL MODEL .................................................................................................................. 181
9.3.4.2 MATERIAL PROPERTIES.............................................................................................................. 181
9.3.4.3 INITIAL STRESS STATE................................................................................................................ 182
9.2.4.4 RESULTS ........................................................................................................................................... 182
Page 5
9.4 Discussion.......................................................................................................................................................... 183
9.5 Acknowledgments............................................................................................................................................ 185
9.6 References ......................................................................................................................................................... 185
10. TOWARDS A HOLISTIC EXPLORATION STRATEGY: USING GEOGRAPHIC

INFORMATION SYSTEMS AS A TOOL TO ENHANCE EXPLORATION ......................................189
Abstract ................................................................................................................................................................... 189
10.1 Introduction.................................................................................................................................................... 189
10.2 GIS-Based Exploration Methodologies...................................................................................................... 191

10.2.1 Empirical approach .................................................................................................................................. 191
10.2.1.1 PRELIMINARY INVESTIGATION............................................................................................. 191
10.2.1.2 DATABASE CONSTRUCTION ................................................................................................... 192
10.2.1.3 EXTRACTION OF A SUBSET OF DEPOSITS .......................................................................... 192
10.2.1.4 IDENTIFICATION.......................................................................................................................... 193
10.2.1.5 QUANTIFICATION........................................................................................................................ 195
10.2.1.6 INTEGRATION............................................................................................................................... 195
10.2.1.7 TESTING OF PROSPECTIVITY MAP........................................................................................ 198
10.2.2. Conceptual approach............................................................................................................................... 198
10.2.2.1. DEVELOPMENT OF KNOWLEDGE BASE............................................................................. 199
10.2.2.2. DEVELOPMENT OF A GIS DATABASE ................................................................................. 200
10.2.2.3. ROUTINES FOR THE ASSESSMENT OF MINERALISATION POTENTIAL ................... 200
10.3 Discussion........................................................................................................................................................ 201

10.3.1. Comparison of methodologies................................................................................................................ 201
10.3.2. Assessment methodologies and other facets of exploration ................................................................. 201
10.3.3. Geological constraints............................................................................................................................. 202
10.3.4. GIS constraints ........................................................................................................................................ 202
10.3.5. Data accuracy .......................................................................................................................................... 202
10.3.6. The third dimension-depth...................................................................................................................... 203
10.3.7. The fourth dimension-time ..................................................................................................................... 203
10.4. Summary And Conclusions ........................................................................................................................ 203
10.5. Acknowledgements....................................................................................................................................... 203
10.6. References...................................................................................................................................................... 204
11. DEVELOPING THE TOOLS FOR GEOLOGICAL SHAPE ANALYSIS, WITH REGIONAL
TO LOCAL SCALE EXAMPLES FROM THE KALGOORLIE TERRANE OF WESTERN
AUSTRALIA. ..........................................................................................................................................209
Abstract ................................................................................................................................................................... 209
11.1. Introduction................................................................................................................................................... 210
11.2. Existing Geological-Shape Analysis Techniques ..................................................................................... 211
11.3. Pattern Recognition Processing Techniques ............................................................................................ 212
11.4 Shape Analysis In GIS .................................................................................................................................. 213
11.5. A Hybrid Gis-Pattern Recognition Approach ......................................................................................... 213

11.5.1 Pre-processing stage................................................................................................................................. 214
11.5.2. Classification stage.................................................................................................................................. 218
Page 6
11.6. Examples Of Map Data Analysis For Exploration ................................................................................. 219

11.6.1. Global-scale shape comparison .............................................................................................................. 219
11.6.2. Identification of similar geometries in the Kalgoorlie Terrane............................................................. 220
11.7. Conclusions.................................................................................................................................................... 221
11.8. Acknowledgements....................................................................................................................................... 222
11.9 References....................................................................................................................................................... 222
Page 7
1. Gold Deposits in Metamorphic Belts:

Overview of Current Understanding,
Outstanding Problems, Future Research and
Exploration Significance.
D.I. Groves a,*, R.J. Goldfarb b, F. Robert c, C.J.R Hart a,d

a
Centre for Teaching and Research in Strategic Mineral Deposits, Department of Geology
and Geophysics, University of Western Australia, Nedlands, WA 6907, Australia
b
U.S. Geological Survey, Box 25046, Mail Stop 973, Denver Federal Center, Denver, CO
80225, USA
c
Barrick Gold of Australia Ltd., 10th Floor, 2 Mill St., Perth W.A., Australia 6000
d
Yukon Geology Program, Box 2703 (K-10)Whitehorse, Yukon, Canada Y1A 2C6
Abstract
Metamorphic belts are complex regions where accretion or collision have added to, or thickened,
continental crust. Gold-rich deposits can be formed at all stages of orogen evolution, so that
evolving metamorphic belts contain diverse gold-deposit types that may be juxtaposed or overprint
each other. This partly explains the high level of controversy on the origin of some deposit types,
particularly those formed or overprinted/remobilized during the major compressional orogeny that
shaped the final geometry of the hosting metamorphic belts. These include gold-dominated orogenic
and intrusion-related deposits, but also particularly controversial gold deposits with atypical metal
associations.
Orogenic lode-gold deposits of Middle Archean to Tertiary age are arguably the predominant gold-
deposit type in metamorphic belts, and include several giant (>250t Au) and numerous world-class
(>100t Au) examples. Their defining characteristics and spatial and temporal distributions are now
relatively well documented, such that other gold-deposit types can be compared and contrasted
against them. They form as an integral part of the evolution of subduction-related accretionary or
collisional terranes in which the host rock sequences were formed in arcs, back-arcs or accretionary
prisms. Current unknowns for orogenic gold deposits include: (1) the precise tectonic setting and
age of mineralization in many provinces, particularly in Paleozoic and older metamorphic belts, (2)
the source of ore fluids and metals, (3) the precise architecture of the hydrothermal systems,
particularly the relationship between first- and lower-order structures, and (4) the specific
depositional mechanisms for gold, particularly for high-grade deposits.
Gold-dominant intrusion-related deposits are a less-coherent group of deposits, which are mainly
Phanerozoic in age, and include a few world-class, but no unequivocal giant, examples. They have
many similarities to orogenic deposits in terms of metal associations, wallrock alteration
assemblages, ore fluids, and, to a lesser extent, structural controls, and hence some deposits,
particularly those with close spatial relationships to granitoid intrusions, have been placed in both
orogenic and intrusion-related categories by different authors. Those that are clearly intrusion-
related ddeposits appear to be best distinguished by their near-craton setting, in locations more distal
from subduction zones than most orogenic gold deposits, and in provinces that also commonly
Page 1
contain Sn and/or W deposits; relatively low gold-grades (≤1-2 g/t Au); and district-scale zoning to
Ag-Pb-Zn deposits in distal zones. Outstanding problems for intrusion-related deposits include: (1)
lack of a clear definition of this apparently diverse group of deposits, (2) lack of a definitive link for
ore fluids and metals between mineralization and magmatism, (3) the diverse nature of both
petrogenetic association and redox state of the granitoids invoked as the source of mineralization,
and (4) mechanisms for exsolution of the CO2-rich ore fluids from the relatively shallow-level
granitoids implicated as ore-fluid sources.
Gold deposits with atypical metal associations are a particularly diverse and controversial group,
are most abundant in Late Archean terranes, and include several world-class to giant examples.
Most are probably modified Cu-Mo-Au porphyry, volcanic rock-hosted Zn-Pb-Ag-Au massive sulfide,
or Zn-Pb-Ag-Au or Ba-Au-Mo-Hg submarine epithermal systems, overprinted or remobilized during
the events in which orogenic gold deposits formed, but there is lack of consensus on genesis.
Outstanding problems for these deposits include: (1) lack of a clear grouping of distinctive deposits,
(2) lack of published well-integrated studies of their characteristics, (3) generally a poorly-defined
timing of mineralization events, and (4) lack of assessment of metal mass-balances in each stage of
the complex mineralization and overprinting events.
Both orogenic gold deposits and gold deposits with atypical metal associations contain a few giant
and numerous world-class examples, whereas the intrusion-related group contains very few world-
class examples, and no giants, unless Muruntau is included in this group. Preliminary analysis
suggests that the parameters of individual world-class to giant gold deposits of any type show
considerable variation, and that it is impossible to define critical factors that control their size and
grade at the deposit scale. However, there appears more promise at the terrane to province scale
where there are greater indications of common factors such as anomalous subduction-related
tectonic settings, reactivated crustal-scale deformation zones that focus porphyry-lamprophyre dike
swarms in linear volcano-sedimentary belts, complex regional-scale geometry of mixed
lithostratigraphic packages, and evidence for multiple mineralization or remobilization events.
There are a number of outstanding problems for all types of gold deposits in metamorphic belts.
These include: (1) definitive classifications, (2) unequivocal recognition of fluid and metal sources,
(3) understanding of fluid migration and focusing at all scales, 4) resolution of the precise role of
granitoid magmatism, (5) precise gold-depositional mechanisms, particularly those producing high
gold grades, and (6) understanding of the release of CO2-rich fluids from subducting slabs and
subcreted oceanic crust and granitoid magmas at different crustal levels. Research studies need to be
better coordinated and more integrated, such that detailed fluid-inclusion, trace element and isotopic
studies of both gold deposits and potential source rocks, using cutting-edge technology, are
embedded in a firm geological framework at terrane- to deposit-scales. Ultimately, four-dimensional
models need to be developed, involving high-quality three-dimensional geological data combined
with integrated chemical and fluid-flow modeling, to understand the total history of the hydrothermal
systems involved. Such research, particularly that which can predict superior targets visible in data
sets available to exploration companies before discovery, has obvious spin-offs for global- to deposit-
scale targeting of deposits with superior size and grade in the covered terrains that will be the
exploration focus of the 21st Century.
1.1. Introduction
Metamorphic belts are exceptionally complex regions of the Earths crust where accretionary or
collisional orogenies have added new continental crust and/or thickened existing continental crust.
These tectonic processes are of lithospheric scale, and, as such, involve thermal and stress anomalies
that progressively: (1) generate magmatic arcs and fore-arcs with associated thick sedimentary prisms
and back-arcs with associated extensional basins; then (2) deform and metamorphose these, normally
with continued extensive granitoid-plutonism; and, finally, (3) uplift and erode these with the
Page 2
generation of new sedimentary basins. Gold-bearing deposits can be generated and/or modified in
each evolutionary stage of orogen development. Therefore, it is not surprising that metamorphic belts
within an orogen can contain diverse auriferous deposits that may be juxtaposed with, or even
overprint, each other, leading to considerable controversy concerning their origin. The deposits may
include shallowly-formed ore systems preserved in less-eroded parts of an orogen and more deeply-
formed systems exposed within unroofed metamorphic belts.
Gold deposit types that formed during early volcanism and sedimentation in convergent margin
settings, or were accreted to cratons in oceanic arcs or obducted oceanic crust in such settings, include
epithermal Ag-Au, porphyry Cu-Au and Au-rich volcanic-hosted massive sulfide (VHMS) deposits
(Fig. l). Unless controversial, because of subsequent remobilization during
metamorphism/deformation or overprinting by contrasting deposit styles, these well-defined gold
deposit types are not considered further in this paper. The gold-rich deposits discussed below are
formed or overprinted/remobilized broadly synchronous with the deformation, metamorphism and
granitoid magmatism, normally in the fore-arc region of convergent margins, during the major
orogeny that produced the major compressional rock fabrics and shaped the final geometry of the
metamorphic terranes within the orogenic belts in which they reside.
Gold deposits that formed in this metamorphic environment are diverse in terms of their age (Middle
Archean to Tertiary), geometry (single veins to vein arrays to stratabound replacement to
disseminated deposits), structural controls (reverse to strike-slip, and, less commonly, normal faults),
host-rocks (mafic-ultramafic rock sequences, sedimentary rocks, or granitoid plutons), metamorphic
grade of host rocks (subgreenschist to granulite facies, although mainly greenschist facies),
temperature and pressure of formation (≈200 - 650°C, 0.5-5 kb), and consequent wallrock alteration
assemblages (carbonate to diopside, muscovite to biotite/phlogopite) and metal associations (Au with
variable Ag, As, B, Bi, Cu, Pb, Sb, Te, W and Zn).
As a result of this diversity, there have been various classification schemes developed for gold
deposits in metamorphic belts, as reviewed, for example, by Groves et al. (1998), Hagemann and
Cassidy (2000), and Bierlein and Crowe (2000). While it is generally accepted that there is diversity
of gold deposit styles in metamorphic belts (e.g., Robert and Poulsen, 1997; Robert et al., 1997; Witt,
1997; Sillitoe and Thompson, 1998; Poulsen et al., 2000), global reviews (e.g., Hodgson, 1993;
Kerrich and Cassidy, 1994; McCuaig and Kerrich, 1998; Goldfarb et al., 2001) indicate that the
dominant style is represented by orogenic gold deposits as defined by Groves et al. (1998). These are
deposits that formed during the late stages of orogenesis, within the main phase of crustal shortening
in compressional or transpressional regimes, in which the penetrative deformational and metamorphic
fabrics were generated and/or reactivated. These criteria are important because this deposit type can
be used as a basis for comparison with deposits assigned to other types by various authors, or for
deposits where possible overprinting of more than one style of mineralization, or alteration, has led to
controversy concerning their origin (e.g., Hemlo in Pan and Fleet, 1992, Michibayashi, 1995, and
Robert and Poulsen, 1997; Boddington in Roth, 1992, Allibone et al., 1998, and McCuaig et al.,
2001; and Bousquet in Valliant and Hutchinson, 1982, Marquis et al., 1990, Tourigny et al., 1993,
and Poulsen and Hannington, 1996). In turn, this classification can be important in attempting to
define the parameters that control whether these are world-class to giant deposits, and whether there
are specific end-member types that are more likely to provide world-class to giant deposits, as these
are the current focus of major exploration companies. For example, the overprinting or remobilizing
process may be particularly important with increasing deposit size, as several of the larger deposits
and districts show multiple episodes of gold mineralization (e.g. Kalgoorlie and Boddington, Western
Australia; Timmins and Hemlo, Canada).
The major objectives of this paper are to define the areas of uncertainty in the current understanding
of gold deposits in metamorphic rocks within orogenic belts, and to define those future research areas
that can potentially resolve these uncertainties and lead to better genetic and exploration models. In
order to achieve this, the major characteristics of the important deposit styles in metamorphic belts are
defined, with orogenic gold deposits used as a reference against which other deposit types are
Page 3
compared. From these considerations, equivocal parameters and/or genetic aspects and future
research needs are defined. An Appendix provides the location and classification of the deposits
discussed in the text, together with a recent reference for each.
Figure 1. Tectonic settings of gold-rich epigenetic mineral deposits. Vertical scale is

exaggerated to allow schematic depths of formation of various deposit styles to be shown. Adapted
from Groves et al. (1998).
1.2. Potential Diversity of Gold Deposit Types in Metamorphic Belts
Groves et al. (1998) reviewed the problems of nomenclature for gold deposits in metamorphic belts.
They emphasized the overall similarities in tectonic setting, structural controls, geochemistry of
alteration, ore-element association and fluid and isotopic composition of the most abundant Au-
dominant (Au>Ag, low Cu-Pb-Zn) deposits in metavolcanic rock (greenstone) and turbidite/slate
(accretionary prism) terranes. Groves et al. (1998) used orogenic gold deposit, following Bohlke
(1982), as a unifying term to describe this group of deposits, widely interpreted to form late in an
orogenic cycle from mid- to lower-crustal metamorphic fluids, although deeply-sourced magmatic
fluids are also possible. The term orogenic gold deposit is not universally accepted, but because no
better all-embracing term has been proposed, it is retained here.
Other authors have suggested that some gold deposits in metamorphic belts with broadly similar
characteristics to orogenic deposits, but which show a close spatial and temporal association with
granitoid intrusions, should be termed intrusion-related deposits (e.g., Sillitoe, 1991; Sillitoe and
Thompson, 1998; Lang et al., 2000; Thompson and Newberry, 2000). Some syenite-associated
deposits in the Abitibi Belt (e.g., Robert, 2001) may be a definable subgroup within the intrusion-
related group. Other workers (Mueller, 1991; Mueller et al., 1996; Mueller and McNaughton, 2000)
have suggested that certain calc-silicate-bearing gold deposits in amphibolite-facies settings of well-
documented orogenic gold provinces, particularly in the Yilgarn Craton of Western Australia, are
gold skarns. In summary, these intrusion-related gold deposits are suggested by some workers to be
representative of a distinct group of deposits because they all formed as a proximal part of a
magmatic-hydrothermal system, rather than from a fluid of more regional extent, favored by many
authors to be of a metamorphic origin, which circulated throughout much of an orogen.
Orogenic gold deposits are typically developed in terranes that have experienced moderate to high T
– low to moderate P metamorphism (Powell et al., 1991), with consequent generation of large
Page 4
volumes of granitic melts. Hence, a distal spatial relationship with certain intrusions is also expected
for these deposits, and there is a possibility of a genetic connection.
In addition, there are gold-bearing deposits in metamorphic belts with element associations unlike
typical orogenic deposits (e.g. Cu±Pb±Zn±Au±Ag±Ba; Cu±Au±Mo; W±Sn±Au), which include
some cited as intrusion-related deposits, gold-rich VHMS or gold-overprinted VHMS, and deformed
Cu±Au±Mo porphyry deposits or their overprinted equivalents. A recurrent theme in many of the
deposits in this group of anomalous and controversial deposits is the possibility of overprinting of one
ore system on pre-existing alteration or on a different type of ore system, and local remobilization of
pre-existing mineralization during deformation and metamorphism. Possible examples include an
orogenic gold overprint on VHMS deposits (Mt Gibson, Western Australia) or on seafloor-related
alteration (Campbell-Red Lake, Canada); an orogenic or intrusion-related gold overprint on
porphyry-style systems (Boddington, Western Australia; Hollinger-McIntyre at Timmins, Canada);
and metamorphic remobilization of porphyry or submarine epithermal style systems (Hemlo,
Canada).
Such diversity is not unexpected given the complex and lengthy (∼100-200 m.y.) evolution of
composite metamorphic belts along active continental margins. The orogenic gold deposits are likely
to be preserved in these belts unless the orogens are eroded down to their high-grade metamorphic
roots (Goldfarb et al., 2001). Porphyry Cu-Au ± Mo and epithermal Ag-Au deposits, generated at
relatively shallow levels in the volcano-plutonic arcs and back-arcs, are typically less likely to be
preserved in metamorphosed sequences. These types of deposits form prior to the major phase(s) of
orogenesis, involving compressional to transpressional deformation, regional metamorphism and
post-volcanic granitoid magmatism, during which the orogenic gold deposits form. During the main
phases of orogenesis, the porphyry, epithermal and VHMS deposits commonly will be deformed and
weakly metamorphosed, but would then typically be eroded during uplift of the orogens following
crustal thickening, leading to their rarity in preserved orogens older than Mesozoic (Kerrich et al.,
2000), although VHMS deposits that were formed in other settings may be preserved. However, the
rare occurrence of undoubted, little deformed Precambrian porphyry deposits (e.g., Clark Lake
porphyry Cu-Mo-Au and related Au-Cu veins at Chibougamau: Pilote et al., 1995), and even rarer
possible epithermal deposits (Groves and Barley, 1994), show that preservation, and hence
overprinting by younger hydrothermal systems, is possible given an appropriate P-T-t path.
The potential diversity of deposit types, and the controversial nature of some of them, means that the
research strategies required to resolve problems related to each group and each genetic controversy
are not necessarily the same in all cases.
1.3. Orogenic Gold Deposits: The Most Common Type?
Groves et al. (1998) defined the unifying term orogenic gold deposit to include those deposits widely
referred to as mesothermal (Nesbitt et al., 1986) in the past 20 years, but also classified in terms of
their ore associations (e.g. gold-only), their host sequences (e.g. greenstone-hosted, slate-belt style,
turbidite-hosted), their form (e.g. lode, quartz-carbonate vein or disseminated deposits) or even their
specific location (e.g. Mother Lode-style deposits). Following Gebre-Mariam et al. (1995), Groves et
al. (1998) further suggested the use of the terms epizonal, mesozonal and hypozonal to describe
specific depth segments of the vertically extensive orogenic-gold systems, with epizonal deposits <6
km, mesozonal deposits from 6 to 12 km, and hypozonal deposits >12 km depth (Fig. 2A)
Page 5
Figure 2. Schematic representation of crustal environments of orogenic gold deposits,

gold deposits with anomalous metal associations and intrusion-related gold deposits in terms
of depth of formation and structural setting. The figure is by necessity stylized. Adapted
partly from Groves et al. (1998) and Lang et al. (2000).
Giant deposits (>250t or >8 Moz Au) or districts, in the sense of Laznicka (1999), which have been
universally ascribed to orogenic gold systems include, from oldest (Archean) to youngest (Mesozoic):
Hollinger-McIntyre, Timmins, Canada; Golden Mile, Kalgoorlie, Western Australia; Kolar, India;
Ashanti, Ghana; Homestake, South Dakota, USA; Bendigo and Ballarat, Victoria, Australia;
Berezovsk, Russia; Kumtor, Kyrgyzstan; and Mother Lode, California, USA. Muruntau, Uzbekistan;
Vasil’kovsk, Kazakhstan; and Morro Velho, Brazil, are also giant deposits that probably belong to
this group, but there are conflicting data in the literature on classification of these deposits. The
poorly-described giant Russian deposits at Olympiada and Sukhoi Log have been assigned to this
group in some descriptions (Safonov, 1997; Goldfarb et al., 2001; Yakubchuk et al., 2001), but field
observations by one of us (F.R.) suggest that such a classification is far from certain. In addition,
there are numerous world-class deposits (>100t or 3 Moz Au) in more than 20 of the 75 metallogenic
provinces that contain orogenic gold deposits worldwide (Goldfarb et al., 2001).
1.3.1 Definition and distinction from other gold deposit styles
The features common to orogenic gold deposits have been extensively reviewed by Kerrich
and Cassidy (1994), Groves et al. (1998), McCuaig and Kerrich (1998), and Goldfarb et al.
(2001), and several papers in Hagemann and Brown (2000), and are only briefly
summarized below and in Table 1.
Page 6
These deposits form along convergent margins during terrane accretion, translation, or
collision, which were related to plate subduction and/or lithospheric delamination. They
formed typically in the latter part of the deformational-metamorphic-magmatic history of the
evolving orogen (Groves et al., 2000). Country rocks are most commonly regionally
metamorphosed into belts with extensive greenschist through lower-amphibolite facies
rocks. Importantly, ores developed synkinematically, with at least one stage of the main
penetrative deformation of the country rocks, and they inevitably have a strong structural
control involving faults or shear zones, folds and/or zones of competency contrast (Hodgson,
1989). They show vertical dimensions of as much as 1-2 km with only subtle metal zoning,
and distinctive, strong lateral zonation of wallrock alteration, which normally involves
addition of K, As, Sb, LILE, CO2 and S, with variable additions of Na or Ca in specific
cases, particularly in deposits sited in amphibolite-facies rocks (Ridley et al., 2000). Due to
considerable variation in the temperature of the hydrothermal systems, proximal wallrock-
alteration assemblages typically vary from sericite-carbonate-pyrite at high crustal levels
through biotite-carbonate-pyrite to biotite-amphibole-pyrrhotite and biotite/phlogopite-
diopside-pyrrhotite at deeper crustal levels (Ridley et al., 2000). Quartz ± carbonate veins
are ubiquitous and are commonly gold-bearing, although in many systems it is the
sulfidized, high Fe/Fe+Mg+Ca wallrocks adjacent to the veins that contain most ore (e.g.,
Bohlke, 1988).
A distinctive metal enrichment association (Au-Ag±As±B±Bi±Sb±Te±W) is characteristic

of the deposits, with Ag, Sb (e.g., Wiluna, Western Australia; Hillgrove, NSW, Australia;
Sarylakh, Russia) and As (e.g., Salsigne, France) present in high enough concentrations in
some places to be mined as by-products. The ores typically have background values to only
slight enrichments in Cu, Mo, Pb, Sn and Zn, and have high gold fineness (generally >900)
and high bulk Au:Ag ratios (generally ≥5:1). They were deposited from low-salinity, near-
neutral, H2O-CO2±CH4±N2 fluids, which transported gold as a reduced sulfur complex. The
CO2 concentrations are everywhere >5 mol%, with variable H2O:CO2:CH4 ratios commonly
caused by phase separation during extreme pressure fluctuations (Sibson et al., 1988).
Typical δ18O values for hydrothermal fluids are about 5-8 per mil in Archean greenstone
belts and about 2 per mil heavier in Phanerozoic gold lodes. Sulfur and C isotope ratios vary
because the fluids had a variable redox state, between the H2S:SO4 and CO2:CH4 buffers, at
the depositional sites (Mikucki, 1998), and there were inherent differences in the isotopic
composition of these species in fluid source regions.
A few gold deposits in well-defined orogenic gold provinces are characterized, however, by
atypical metal associations and ratios, alteration assemblages or fluid salinities (final two
columns of Table 1). These are differentiated from orogenic gold deposits in recent
overview papers (Robert et al., 1997; Groves et al., 1998; Kerrich et al., 2000; Goldfarb et
al., 2001), although these may result from preferential overprinting by orogenic gold
systems. There are also, however, other deposits with the geologic characteristics of
orogenic gold systems and which have been grouped with the orogenic deposit types by
Groves et al. (1998) and Goldfarb et al. (2001), but have been alternatively classified as
intrusion-related (Sillitoe 1991, Matthai et al., 1995, McCoy et al., 1997, Sillitoe and
Thompson, 1998, Brisbin, 2000, Lang et al., 2000, Thompson and Newberry, 2000; Lang
and Baker, 2001) or syenite-associated (Robert, 2001) deposit types (Fig. 2C: Table 1).
These latter two classifications are partly on the basis of the close spatial and temporal
association of gold deposits with granitoid intrusions. These deposits, with many features of
both orogenic and intrusion-related gold deposits, are particularly problematic and diverse,
Page 7
perhaps because of inconsistent model descriptions between the various studies. Hence, they
are singled out for discussion later in this paper.
Critical Characteristics Orogenic Gold Deposits Intrusion-Related Gold Deposits Gold Deposits with Anomalous Metal
Associations
Age Range Middle Archean to Tertiary: Peaks in Late Mainly Phanerozoic: Some Proterozoic: Mostly Late Archean: some Phanerozoic.
Archean, Paleoproterozoic, Phanerozoic. Rare Late Archean.
Teconic Setting Deformed continental margin mainly of Pericratonic terranes of the miogeocline Back-arc to arc early?: Accretionary to
allochthonous terranes. margin. collisional terranes late.
Structural Setting Commonly structural highs during later Compressional to extensional Late evolution similar to that of orogenic
stages of compression and transtension. transition in fold and thrust belts. deposits.
Host Rocks Variable: Mainly mafic volcanic or intrusive Major examples in grantioid intrusions: Variable: Commonly felsic intrusive or
rocks or greywacke-slate sequences. some in sedimentary rocks. volcanic rocks.
Metamorphic Grade of Mainly greenschist facies but sub- Mainly sub-greenschist to greenschist Greenschist to amphibolite facies.
Host Rocks greenschist to lower granulite facies. facies.
Association with Intrusions Commonly flesic to lamprophyre dikes Strong association with granitoid Strong association with granitoid stocks
or continental margin batholiths. stocks. Lamprophyre dikes. and/or felsic to lamprophyre dikes.
Mineralization Style Variable: Large veins, vein arrays, saddle Commonly sheeted veins, lesser High variable: Disseminated to vein
reefs, replacement of Fe-rich rocks. breccias, veins and disseminations. styles.
Timing of Mineralization Late-tectonic: Post (greenschist) to syn Very-late tectonic: Post regional Syn-volcanic and pre-metamorphic?: Late
(amphibolite) metamorphic peak. metamorphic peak. evolution syn-to post-metamorphic.
Structural Complexity Complexity common, particularly in Mainly simple vein arrays in relatively Complexity normal, causing extreme
of Ore Bodies brittle-ductile regimes. brittle regimes. controversy for this deposit style.
Evidence of Overprinting Strong overprinting in larger deposits: Minor evidence of overprinting by late Strong evidence of overprinting in most
Multiple veining events. structures. deposits.
Metal Association Au-Ag±As±B±Bi±Sb±Te±W. Au-Ag±As±B±Bi±Sb±Sn±Te±W Au-Ag±Ba±Cu±Hg±Mo±Pb±Zn
(Pb-Zn distal).
Metal Zoning Cryptic lateral and vertical zoning. Strong district-scale zoning: Variable, but strong in some deposits.
Au-W/Sn-Ag/Pb/Zn.
Proximal Alteration Varies with metamorphic grade: Normally Mica-K feldspar-carbonate-chlorite- Extremely variable due to different deposit
mica-carbonate-Fe sulfide. Fe sulfide. styles and metamorphic overprint (?)
P-T Conditions 0.5-4.5kb, 220-600ºC 0.5-1.5kb, 200-400ºC for Au-rich Variable: Now largely reflect metamorphic
Normally 1.5±0.5kb, 350±50ºC systems. conditions of host rocks.
Ore Fluids Low-salinity H2O-CO2±CH4±N2. Variable salinity H2O-CO2, very minor Variable: High-salinity H2O to low/
CH4±N2. moderate salinity H2O-CO2.
Proposed Heat Sources Varied: Asthenosphere upwelling to mid- High-level granitoids in gold Early igneous heat source?: Later deep
crustal granitoids. district. crustal/lithosphere heat source.
Proposed Metal Sources Subducted/subcreted crust and/or High-level granitoids and/or Variable: Magmatic, metamorphic or
supracrustal rocks and/or deep granitoids. supracrustal rocks. deep crustal sources proposed.
1.3.2 Current knowledge
With increasingly sophisticated geochronology, using robust isotopic systems, the temporal
distribution of orogenic gold deposits is now broadly known (Fig. 3). There is a
heterogeneous temporal distribution of formation ages marked by two Precambrian peaks
(2800-2550 Ma and 2100–1800 Ma), a singular lack of deposits for 1200 m.y. (1800-600
Ma), and a continuous genesis since 600 Ma. Combined with the general agreement that
orogenic gold deposits formed in areas of continental growth throughout geological time,
Goldfarb et al. (2001) described how the temporal distribution was related to evolution of
plate-tectonic processes as the Earth progressively cooled. They interpreted the two earlier
Page 8
Precambrian formation episodes to major, plume-influenced mantle overturn (e.g., Davies,

1995) in the hotter early Earth (ca. >1800 Ma), with major orogenic gold-forming events
correlated to major periods of crustal growth (Condie, 1998). These gold deposits have been
preserved in roughly equidimensional large masses of buoyant continental crust (e.g.
cratons). Evolution to a less episodic, more continuous, modern-style plate-tectonic regime,
beginning near the start of the Mesoproterozoic, is interpreted to have led to the more
common accretion of volcanosedimentary arcs and oceanic terranes as linear orogenic belts
surrounding the margins of the more-buoyant Archean to Paleoproterozoic cratons. Uplift
and erosion of these linear belts, and any contained orogenic gold deposits, to expose their
high-grade metamorphic cores can explain the noticeable lack of preserved deposits at 1800
– 600 Ma, their preferred exposure in 600 – 50 Ma orogens, the importance of placers
associated with Phanerozoic lodes older than ca. 100 Ma, and the general absence of
orogenic deposits in the still shallow levels of orogenic belts that are younger than ca. 50 Ma
(Fig. 3).
The district-scale controls on orogenic gold deposits are relatively well defined. Crustal-
scale deformation zones, particularly those hosting swarms of felsic porphyry intrusions,
serpentinized ophiolite fragments, and/or lamprophyre dikes, play an important role in
localizing deposits in many of the better endowed gold provinces (e.g., Abitibi belt, Canada;
Norseman-Wiluna belt, Australia; Ashanti belt, Ghana; Juneau and Mother Lode belts,
USA). On this scale, bends or jogs in these regional deep-crustal structural zones, their
interaction with lower-order shear zones, major competency contrasts in lithostratigraphic
sequences, anticlinal or uplifted zones, and irregularities along granitoid contacts may all
play a role in creating low minimum or mean stress zones into which ore fluid could be
focussed at the district or camp scale (Groves et al., 2000). In general, there is no consistent
spatial association between the orogenic gold deposits and specific granitoid composition
either within or between provinces, although abundant granitoid intrusions are a feature of
most orogenic gold provinces (e.g., Kerrich and Cassidy, 1994; Bierlein et al., 2001b;
Goldfarb et al., 2001) because they are a logical consequence of the collisional –
accretionary processes at convergent margins.
Page 9
Figure 3. Distribution of gold production from orogenic gold deposits with geological
time. Provinces in which intrusion-related gold deposits are considered by some authors to
be a major contributor to production are shown. Adapted from Goldfarb et al. (2001).
At the deposit scale, all orogenic-gold ore bodies show strong structural control. Although
the nature of that control varies within and between provinces, faults with a reverse
component of shear are more commonly mineralized than those with a normal or dominant
strike-slip shear component (Sibson et al., 1988). Host rocks are extremely variable,
although there is an overall trend from volcanic rock- or intrusion-hosted deposits in
Archean provinces to sedimentary rock-hosted deposits in Paleoproterozoic to Tertiary
provinces, and for the larger deposits to be hosted in a few of the physically and/or
chemically more-favorable units within the lithostratigraphy in any province.
Page 10
As noted above, orogenic gold deposits may be hosted in rocks that have been
metamorphosed from sub-greenschist to granulite facies. However, most giant and world-
class deposits that are ascribed to orogenic gold systems, with the notable exceptions of
Kolar and perhaps Muruntau, are hosted in greenschist facies rocks.
The relative timing of orogenic gold-deposit formation worldwide has recently been
summarized by Groves et al. (2000), with emphasis on the Yilgarn Craton of Western
Australia. They concluded that orogenic gold deposits most commonly form during
progressive D2 to D4 deformation events in a D1 to D4 deformational sequence, normally 20
to 100 m.y. after the deposition of volcano-sedimentary host rocks, although there may be a
greater hiatus in some provinces. There are very few Precambrian or early Paleozoic cases
where synchronicity between gold deposition and emplacement of adjacent granitoid
intrusions has been demonstrated unequivocally; adjacent or hosting intrusions may both
pre-date gold mineralization, as in most examples from the Yilgarn Block (Groves et al.,
2000) or post-date them by as much as 80 m.y. in the central Victorian goldfields of
Australia (Bierlein et al., 2001b), and actually hornfels pre-existing gold deposits in the
Stawell and Maldon gold camps (Foster et al., 1998; Bierlein et al., 2001a). Importantly,
Wilde et al. (2001) demonstrated that gold mineralization at Muruntau post-dates subjacent
granitoids by about 30 m.y., making its classification as an intrusion-related deposit dubious.
In some younger terranes, such synchronicity is better documented. For example, Eocene
orogenic gold deposits of the Juneau gold belt in southeastern Alaska formed simultaneously
with shallower emplacement of some plutons of the massive Coast batholith, which crops
out about 10 km landward of the ores (Miller et al., 1994). In the Chugach accretionary
prism of southern Alaska, Tertiary gold deposits and crustal-melt granitoids are coeval in a
belt extending for more than 2,000 km along the Gulf of Alaska margin (Haeussler et al.,
1995).
Depositional mechanisms for gold in orogenic deposits are summarized by Mikucki (1998).
For replacement deposits or deposits dominated by gold disseminated in altered wallrock,
desulfidation of reduced aqueous sulfur complexes of gold by reaction with rocks with high
Fe/Fe+Mg ratios is the most viable precipitation mechanism (Bohlke, 1988; Phillips and
Groves, 1983), although pH changes may be important in ultramafic rocks (Kishida and
Kerrich, 1987). For free-gold deposition in quartz-carbonate veins, large pressure
fluctuations during hydrofracturing, accompanied by phase separation, are the most likely
mechanisms for destabilizing aqueous sulfur complexes of gold, although there are
competing physicochemical changes, some of which increase, whereas others decrease, gold
solubility (Mikucki, 1998). Mixing of externally derived fluid (e.g., meteoric water) with
the main ore fluid has been suggested for some epizonal deposits (Nesbitt et al., 1986; Craw
and Koons, 1989; Hagemann et al., 1994), but, on the basis of existing H and O isotope data,
cannot be a widespread depositional mechanism. Fluid reduction and destabilization of gold
complexes by backmixing of fluid, which has reacted with local wallrocks, is also a
possibility (Cox et al., 1995), particularly in sedimentary rock sequences containing
carbonaceous matter, where fluid-rock reaction could contribute CH4, or other
hydrocarbons, and/or N2. This may explain why, in a single gold province, fluid inclusions
from gold-bearing veins hosted in metasedimentary country rocks are typically enriched in
these reduced volatiles, whereas CO2 comprises almost the entire non-aqueous fluid phase
where similar veins cut intrusions emplaced into the same metasedimentary sequence (e.g.,
Goldfarb et al., 1989).
Page 11
The fluid chemistry and source of ore fluids for orogenic gold deposits is reviewed
extensively by Ridley and Diamond (2000). They point out that careful scrutiny of fluid
inclusion data from deposits of this style leads to the unequivocal conclusion that the
deposits were formed from low-salinity, mixed aqueous-carbonic fluids, which are different
from those depositing all other major groups of gold deposits (e.g., epithermal, porphyry Cu-
Au, VHMS), although individual deposits may provide exceptions. The fact that broadly
synmetamorphic deposits occur in amphibolite-facies domains (Ridley et al., 2000) indicates
that the ore fluid must be derived from at least the depth represented by peak metamorphism,
and hence a deep source for the fluid is generally accepted. However, despite the extensive
stable and radiogenic isotope database available on both ore-related minerals and fluid
inclusions, there is no consensus on the source of this fluid. Ridley and Diamond (2000)
demonstrated that this is not surprising given the long fluid pathways and the siting and
availability in different wallrocks with variable elemental (e.g. K/Rb) or isotopic ratios.
Ridley and Diamond (2000) pointed out that certain elements which dominate the ore fluids,
such as N, Br, Cl, C and H, have isotope charactistics that may be useful in constraining the
fluid source. However, they also emphasized that H isotopes appear prone to resetting, the
deeper crustal chemistry of N, Br and Cl is poorly known, and earlier-formed reservoirs of C
in the form of graphite or carbonate alteration along fluid conduits may alter isotopic ratios.
Available fluid inclusion, geochemical and isotopic data cannot unequivocally distinguish
between a metamorphic and deep-magmatic source for the ore fluids in orogenic gold
systems. A deep magmatic source is feasible given the modeling by Cline and Bodnar
(1991), which demonstrates that salinity of magmatic fluids is strongly dependent on
pressure, with the existence of mixed aqueous – carbonic magmatic fluids possible above
about 3 kb. However, the mechanism of release of such fluids, given the high degree of
magma crystallization prior to water saturation at these pressures (e.g., Burnham, 1979), is
unresolved. In addition, unless the considered source is specifically a broad zone of deep-
crustal melting, the recognized intrusion bodies are too limited in extent in many gold
provinces to be the source of the required voluminous fluid flow and high gold tonnage.
1.3.3 Outstanding problems
For orogenic gold deposits, a major outstanding problem relates to the timing of
mineralization, as this also affects other components of genetic models including the
potential source(s) of ore fluid and the architecture of crustal-scale flow systems at the time
of gold mineralization. The unequivocal absolute age of deposits remains poorly established
in many cases, due, at least in part, to the common absence of readily datable and robust ore-
related minerals with sufficiently elevated closure temperatures (e.g., Kerrich and Cassidy,
1994). The structural timing (age of mineralization relative to host structures) of some
deposits, especially those hosted in rocks of amphibolite and higher metamorphic grades,
remains equivocal, owing to the difficulty in distinguishing those formed during shear-zone
movements from those overprinted by shear zones (e.g., Robert and Poulsen, 2001). The
question of the age of mineralization is further complicated by the existence of several
generations of gold deposits with orogenic characteristics in a number of districts, including
Val d’Or, Canada (Couture et al., 1994), Kalgoorlie, Australia (Clout et al., 1990), and
Muruntau, Uzbekistan (Kempe et al., 2001; Yakubchuk et al., 2002).
No single model for the fluid and metal sources explains all observations from the orogenic
gold deposits. Metamorphic Au-transporting fluids derived from deeper levels of the ore-
hosting volcanosedimentary or sedimentary rock-dominant terranes are favored by some
Page 12
authors (e.g., Powell et al., 1991; Stüwe, 1998), whereas others favor an even deeper
metamorphic source such as subducted oceanic crust or the residue of its previous melting
(e.g. Fyfe and Kerrich, 1985). The alternative of a deep (distal) magmatic source also
remains open (Ridley and Diamond, 2000). The lack of any known granitoid intrusions in
the Otago orogenic-gold province of South Island, New Zealand (Henley et al., 1976; Craw
et al., 1995), provides an argument for a metamorphic source in deeply underthrusted rocks,
if a single source is invoked for these deposits. The occurrence of several gold
mineralization episodes in the same region (e.g., Val d’Or, Kalgoorlie) also suggests that
simply metamorphism of hosting belts or basins is an unlikely source for the fluids in all
episodes. However, the exact mechanisms, if any, for fluid release from subducted or
subcreted oceanic crust are unknown, and this remains a major barrier to acceptance of these
as an ore fluid source. Similarly, the source of Au remains controversial, with some workers
suggesting a required crustal pre-concentration (Bierlein et al., 1998) and others discounting
the need for such in lieu of an effective Au extraction regime from common crustal
lithologies (e.g., Fyfe and Kerrich, 1984). In the latter case, pyrite in marine sedimentary
rocks and in greenstones is a likely leachable source mineral with high background
concentrations of Au. Similarly, radiogenic isotope studies at Muruntau suggest that
elements in scheelite and, by association, the gold itself, were derived from the Paleozoic
metasedimentary host-rock sequences (Kempe et al., 2001).
A somewhat related problem concerns the distinction between hypozonal orogenic gold
deposits, in the sense of Groves et al. (1998), and magmatic-related skarns, in the sense of
Mueller (1991), in the amphibolite facies terranes of Western Australia. Although described
as intrusion-related gold deposits developed in a continental margin setting (Mueller and
McNaughton, 2000), these controversial gold deposits are different to the intrusion-related
deposits that are discussed below. Skarns in the sense of Mueller (1991) represent deposits
formed at deep crustal levels (>10 km), whereas most intrusion-related deposits as defined
by Lang et al. (2000) are interpreted to form at relatively shallow crustal levels (<7 km).
A major unknown in the architecture of the fluid plumbing systems is the role that the
crustal-scale deformation zones play in fluid advection to middle and upper crustal levels.
Gold deposits are commonly, although not exclusively, in second- or third-order structures
that were probably connected to the first-order structures at the time of gold mineralization.
However, the precise fluid-flow paths in the systems are not well understood, nor is how the
fluid evolved as it passed from one part of the system to another, although some progress is
being made (e.g., Cox, 1999; Lonergan et al., 1999; Neumayr et al., 2000, Neumayr and
Hagemann, in press). The inclination of permeable faults and shear zones relative to the
regional hydraulic head gradient appears to be an extremely critical control on the focusing
of gold-bearing fluids (Cox et al., 2001)
The dominance of one specific depositional mechanism for gold (e.g., desulfidation, pH
change, phase separation, or back mixing) over all other mechanisms, particularly in high-
grade vein deposits and at different crustal levels, is also not always clear. The role of As-,
Sb- and Te-bearing ligands in gold transport requires more investigation (Wood and
Samson, 1998), particularly because initial research on Bi-Au systems shows liquid bismuth
to be a viable complexing agent in relatively S-poor systems and during Au-undersaturated
conditions (Douglas et al., 2000). Whereas it has long been suggested that carbonaceous
material within metamorphic belts may be important for destabilizing gold complexes,
Bierlein et al. (2001b) suggested that such material may not be important in depositing gold
Page 13
locally, as indicated by the poor correlation between high gold grades and carbonaceous
rocks at the deposit scale in the Victorian goldfields.
On a larger scale, the precise tectonic settings in which orogenic gold provinces formed, particularly
those containing giant deposits, requires better resolution. Evidence suggests that orogenic gold
deposits may form at times of change in plate motion (e.g., Goldfarb et al., 1991), in environments
where there were anomalous plate configurations, for example subduction reversals (e.g., Wyman et
al., 1999), and/or where there were anomalous thermal conditions related to upwelling asthenosphere
(e.g., Kerrich et al., 2000; Goldfarb et al., 2001).
1.4. Intrusion-Related Gold Deposits: A Coherent Group?
Hypothesized connections between hydrothermal ore deposits and granitoid intrusions has always
been widely argued. There is now almost universal acceptance that porphyry Cu-Mo or Cu-Au
deposits (e.g., Sillitoe, 1997), associated high-sulfidation Ag-Au epithermal deposits (e.g.,
Hedenquist and Lowenstern, 1994), and Au-bearing skarns (e.g., Meinert, 1993) are genetically
related to adjacent porphyry intrusions. Other gold-bearing hydrothermal deposits are more
controversial. The orogenic gold deposits are classic examples in which there is a broad spatial and, in
places, temporal connection to regional granitoid magmatism (Groves et al., 1998; Goldfarb et al.,
2001), but rarely a specific and proximal relationship to a given intrusion or intrusive suite.
During the past decade, there has been renewed emphasis on diversity in deposit styles within
provinces containing orogenic gold deposits (e.g. Robert and Poulsen, 1997), with emphasis on
intrusion-related gold deposits and their potential misclassification as orogenic gold deposits (e.g.,
Sillitoe and Thompson, 1998). Sillitoe (1991) described intrusion-related gold deposits as being
mainly restricted to accreted terranes in Phanerozoic convergent plate margins, spatially associated
with porphyry Mo or Cu-Mo mineralization, related to magnetite-series I-type intrusions,
characterized by an As-Bi-Te geochemical signature, and having formed from magmatic and/or
meteoric fluids. Sillitoe (1991) grouped these deposits into five distinct classes: (1) stockworks and
disseminated ores in porphyritic (e.g., Lepanto, OK Tedi, Boddington) and non-porphyritic (e.g.,
Zortman-Landusky, Salave, Gilt Edge, Kori Kollo) intrusions; (2) skarns (e.g., Fortitude, McCoy,
Nickel Plate, Red Dome) and replacement ores (e.g., Barney’s Canyon, Ketza River, Yanicocha) in
carbonate rocks; (3) stockworks, disseminated ores, and replacement bodies in country rocks to
intrusions (e.g., Porgera, Muruntau, Mt. Morgan, Quesnel River); (4) breccia pipes in country rocks
(e.g., Montana Tunnels/Golden Sunlight, Kidston, Chadbourne); and (5) mesothermal and low-
sulfidation epithermal veins in intrusions and country rocks (e.g., Charters Towers, Jiaodong
Peninsula, Majara): see Sillitoe (1991) for references to deposits. The classes obviously reflect many
different types of gold deposits that are suggested to show a relatively local, spatial zonation within
and surrounding a causative pluton. With some exceptions (e.g., Muruntau, Charters Towers,
Jiaodong), there is little debate that most of these gold deposits are genetically associated with a well-
defined igneous body and are thus well-classified as intrusion-related deposits.
Sillitoe and Thompson (1998) noted it is Sillitoe’s (1991) class 5 of intrusion-related gold vein
deposits that may have many characteristics identical to orogenic gold deposits. Of five geochemical
associations that they identify within this class of vein type deposits, only the deposits with the Au-
Te-Pb-Zn-Cu (e.g., Charters Towers, Linglong, Dongping) and Au-As-Bi-Sb (e.g., Ryan Lode)
associations have features resembling, and potentially confused with, orogenic gold deposits.
Subsequent overview descriptions of the intrusion-related gold deposit group (e.g., Thompson et al.,
1999; Lang et al., 2000; Thompson and Newberry, 2000; Lang and Baker, 2001) concentrate on the
Au-As-Bi-Sb group of Sillitoe and Thompson (1998), although deposits that would never be classed
as orogenic gold deposits based on the definition of Groves et al. (1998) (e.g., Salave, Spain; Kidston,
Queensland, Australia; and those of the Bolivian polymetallic belt) are also discussed in the
overviews.
Page 14
Figure 4. Photographs showing difference

in mineralization style between: A. the Fort Knox
intrusion-related deposit, Alaska, comprising
largely steeply-dipping sheeted veins in granitoid
with limited visible wallrock alteration, and B.
the Granny Smith orogenic gold deposit, Western
Australia, comprising gently-dipping conjugate-
vein sets with intense bleaching of the host
granitoid.
Figure 5. Close up of veins showing

contrasts in alteration between: A. a sheeted vein
in granitoid, with limited alteration, from the Fort
Knox intrusion-related gold deposit, Alaska and
B. a sheeted vein in dolerite, with associated
intensive alteration, from the Mount Charlotte
deposit, Kalgoorlie, Western Australia
Most descriptions of intrusion-related gold deposits as viewed here are not representative of a single
well-defined group, but rather of a number of different deposit styles with different tectonic settings,
metal associations, and ore fluids placed under a single umbrella term. For example, Wall (1999)
suggested that the environment above large plutons (roof zone) is prospective for gold mineralization
due to the presence of thermal, fluid and geochemical gradients and fluxes, as well as pre-existing
structures related to pluton emplacement. However, this concept, which has been argued to
accommodate a diverse group of Archean, Proterozoic, and Phanerozoic gold deposits (including Fort
Knox, Sukhoi Log, Pogo, Calie (Tanami), Muruntau, Telfer, and Kumtor), is genetically
unconstrained. Deposits can form from circulating meteoric or connate fluids, fluids produced by
thermal devolatilization within the aureole, or from those exsolved from a magma. As such, the
pluton serves as the heat engine, but fluids, metals, ligands, etc. may be derived from the pluton, the
aureole, or more distal environments,
Those deposits that occur in broadly similar tectonometamorphic settings to orogenic gold deposits
and have features that make distinction between the two deposit types equivocal, are discussed below.
Page 15
Into this category could also be added the potentially intrusion-related gold deposits of the Pine
Creek district, Northern Territory, Australia, including Cosmo Howley (e.g., Matthai et al., 1995), the
syenite-associated gold deposits of the Abitibi belt, Canada (e.g., Robert, 2001), and possibly the
Telfer gold deposits of Western Australia (e.g., Goellnicht et al., 1989; Rowins et al., 1997).
However, recent 40Ar/39Ar geochronology in the Tanami region of Northern Territory, indicates an
approximately 100 m.y. difference between magmatism and mineralization (Wygralak and Mernagh,
2001), such that lode gold deposits in this part of northern Australia no longer can be considered to
best fit an intrusion-related model.
1.4.1 Distinction from orogenic gold deposits
In perhaps the clearest refinement of their defining characteristics, Lang et al. (2000), utilizing the
studies of Sillitoe (1991), Newberry et al. (1995), McCoy et al., (1997), and Thompson et al. (1999),
among others, summarized the major characteristics of intrusion-related gold deposits as an
association of gold mineralization with: (1) metaluminous, subalkalic intrusions of intermediate to
felsic composition that span the boundary between ilmenite- and magnetite-series; (2) CO2-bearing
hydrothermal fluids, (3) a metal assemblage that variably includes Au with anomalous Bi, W, As,
Mo, Te and/or Sb, and typically has non-economic base-metal concentrations, (4) comparatively
restricted zones of hydrothermal alteration within granitoids, (5) a continental tectonic setting well
inboard of inferred or recognized convergent plate boundaries, and (6) a location in magmatic
provinces best or formerly known for W and/or Sn deposits. Lang et al. (2000) indicated that the
most characteristic style of deposit is an intrusion- or hornfels-hosted, sheeted array of low-sulfide
quartz veins with narrow alteration envelopes (Figs. 4A and 5A), but other styles occur and are
commonly zoned surrounding the intrusions (Hart et al., 2000). Examples might include the base-
metal bearing Keno Hill silver deposits and similar veins that are typically distal to many of the gold
prospects in the Yukon part of the Tintina gold province. The deposits of the Pine Creek, Tanami and
Telfer districts of northern Australia broadly fit these criteria, although none are actually hosted in the
associated granitoids. The syenite-associated group of deposits of Robert (2001) clearly does not fit
the above category, but might represent a distinct, yet related, style of intrusion-related deposit in
Archean greenstone belts.
In terms of the first four criteria above, there is clearly the potential for misclassifications between
intrusion-related deposits and orogenic gold deposits that are sited in small granitic bodies due to their
competency contrast with surrounding less-rigid supracrustal rocks (e.g., Ojala et al., 1993; Groves et
al., 2000). Perhaps the most useful distinctions are the last two criteria, which relate to the setting of
the intrusion-related deposits distal to subduction zones, commonly in carbonate rock-bearing shelf
environments rather than in turbidite-dominated accreted terranes seaward of old cratonic margins.
For example, the intrusion-related gold deposits of the Tintina gold province in Yukon occur in the
Selwyn Basin, part of the Neoproterozoic-early Paleozoic rifted margin of western North America, an
area well-recognized for its large W-bearing skarn deposits. Many gold deposits and small tungsten
skarns surround the same plutons that were emplaced in mid-Cretaceous times into the miogeoclinal
strata (Hart et al., 2000, 2002). However, these two criteria are not easily established for many gold
belts or districts, as the required knowledge of the overall tectonic setting of an area, particularly for
Paleozoic and older environments, is typically complex and controversial.
An additional criterion for discriminating between orogenic and intrusion-related deposits is the
timing of mineralization relative to penetrative deformation of the host rocks, even where both
deposits are structurally-controlled. In the case of orogenic gold deposits, as indicated above,
mineralization is synchronous with, or postdates, the development of penetrative (ductile) structures
such as shear zones and folds, and regional penetrative fabrics (typically during D2 and/or D3 of a
D1-D4 district evolution). In the case of many intrusion-related deposits, certainly including those in
Alaska and Yukon (e.g., Hart et al., 2002), the mineralization is younger than the penetrative, gneissic
fabrics of the host rocks as indicated by the fact that associated and mineralized intrusions cut
penetratively deformed host rocks. Yet in other cases, such as the syenite-associated gold deposits in
the Abitibi greenstone belt, the deposits are overprinted by penetrative fabrics (Robert, 2001). This
Page 16
relatively early timing is clearly different from that of the syntectonic orogenic gold deposits in the
area. Similar to the syenite-associated deposits of the Abitibi belt, the magmatism and Early to
Middle Triassic mineralization at Timbarra is interpreted by Mustard (2001) to be overprinted by later
deformation of the New England orogen.
1.4.2 Current knowledge
The defining characteristics of the intrusion-related gold deposit model are still being established
because these systems have only been suggested as a distinct mineral-deposit type within the last five
years. If only those deposits that are Au-dominant and base-metal poor are considered, some
generalizations can be made. The deposits placed in this group by Sillitoe and Thompson (1998),
Thompson et al. (1999), Lang et al. (2000), and Thompson and Newberry (2000) are exclusively
Phanerozoic. They are typically defined as situated near craton margins in magmatic provinces that
are distal to active convergent plate margins, explaining their overlap with Sn-W provinces that tend
to lie inboard of porphyry and epithermal provinces in arc and back-arc settings (Mitchell and
Garson, 1981; Sawkins, 1984).
Whereas the intrusion-related gold deposits are certainly not recognized in accretionary prisms, some
deposits assigned to this class are associated with subduction-related magmas emplaced after
accretion within allochthonous terranes located only a few hundred kilometers inland from an active
Phanerozoic continental margin. Deposits such as Timbarra in the New England fold belt of eastern
Australia (Mustard, 2001) and those of the Tian Shan in central Asia (e.g., Muruntau : Sillitoe, 1991;
Jilau : Cole et al., 2000) are such examples, according to some workers. These gold-hosting tectonic
settings are clearly part of the deformed continental margin and not part of older cratons; gold ores
are formed within rocks added during the same orogeny and not within rocks of the pre-accretionary
continent. Similarly, two deposits most commonly defined as intrusion-related within the Bohemian
Massif (e.g. Mokrsko and Petrackhova hora: Zacharias et al., 2001) were also emplaced in an actively
deforming collisional environment during the 360-320 Ma Variscan orogeny (e.g., Cliff and
Moravek, 1995). All the above tectonic settings for the ores also characterize many undoubted
orogenic gold deposits and, therefore, there is a clear spatial overlap that further hinders easy
distinction between the deposit groups.
The tectonic settings of the older Pine Creek and Telfer gold provinces are equivocal, but there is
overlap with Sn- or W-bearing deposits that suggests a setting inboard of the magmatic arc. The
syenite-associated, and potentially other intrusion-related deposits in Archean metamorphic belts,
would have formed in quite a different setting, for example, during syndeformational sedimentation
in a primitive back-arc setting, with no significant Sn or W mineralization.
The regional-scale structural controls on the proposed ore-related granitoids and associated deposits
are not well understood. In the Tintina gold province (Alaska, USA and Yukon, Canada), some of
the deposits in the Yukon lie within 10 km of the crustal-scale Tintina Fault, but at least part of the
movement on this structure is post-gold mineralization (Flanigan et al., 2000). Where the offset part
of the magmatic belt continues into eastern Alaska, plutons and associated mineralization (e.g. Fort
Knox) could be controlled by high-angle, NE-trending second-order faults between the Tintina and
Denali regional faults (Newberry, 2000). In the Telfer province, the deposits align along what is
interpreted to be a major basement structure (Rowins et al., 1997), and several of the larger deposits
in the Pine Creek province lie along the crustal-scale Pine Creek Shear Zone (e.g. Partington and
McNaughton, 1997). Although Robert (2001) documented an early timing for the syenite-associated
deposits in the Abitibi belt, they are nevertheless closely associated with the crustal-scale deformation
zones (breaks) that also controlled Timiskaming sedimentation and the gross distribution of later
orogenic gold deposits.
Page 17
The intrusions that are associated with these deposits, if considered on a global scale, show extreme
variation with most being I-type, although some with more evolved phases have some S-type
characteristics (e.g., Donlin Creek, Alaska : Goldfarb unpub. data). McCoy et al. (1997) and
Rombach and Newberry (2001) emphasized the reduced nature of intrusion-related gold deposits in
Alaska, noting that, in more oxidized magmatic systems, magnetite tends to remove Au that would
otherwise concentrate in the volatile phase from the melt. Lang et al. (2000) recorded that
granodiorites to granites of metaluminous subalkalic intrusions, with oxidation states intermediate
between magnetite- and ilmenite-series granitoids, are important in the Tintina gold province, but
note that intrusions at Timbarra in New South Wales, are more highly oxidized. The oxidized
syenites from the Abitibi belt also contrast markedly with the host intrusions in the Tintina province,
to a large degree reflecting a very different tectonic setting. In fact, Robert (2001) noted that
magnetite is common within the gold ores. Similarly many of the gold deposits along the northern
margin of the North China craton, also interpreted by some authors as intrusion-related (e.g.,
Dongping : Sillitoe and Thompson, 1998; Niuxinshan : Yao et al., 1999), are associated with highly-
oxidized mineral assemblages. Granitoids in the Telfer region include both ilmenite and magnetite
series, with the latter considered to be more closely related to gold mineralization (e.g., Goellnicht et
al., 1989, 1991). In the Pine Creek region, granitoids belong to a peraluminous magnetite to ilmenite
series (Klominsky et al., 1996). In both regions, the granitoids have high levels of heat-producing
elements. The broad range in inferred melt fO2 indicates that, if indeed all these deposits are taken as
parts of a coherent gold deposit class, then oxidation state is not a critical factor in determining
whether Au was present or absent within fluids exsolving from granitic melts.
At the deposit scale, there is variable structural control on intrusion-related gold mineralization (Fig.
2C). Within host intrusions, mineralization commonly occurs as sheeted veins (Figs. 4A and 5A),
typically extensional rather than shear veins. However, disseminated (e.g., Brewery Creek, Yukon:
Hart et al., 2000; Timbarra, New South Wales : Mustard, 2001) and/or stockwork styles of
mineralization (e.g., Donlin Creek, Alaska : Ebert et al., 2000; Shotgun : Rombach and Newberry,
2001) may be present at some deposits. Wallrock alteration surrounding the sheeted vein systems of
intrusion-related gold deposits is limited in intensity and distribution relative to that associated with
most orogenic gold deposits that are intrusion-hosted (compare Figs. 4A and 5A with 4B and 5B).
Rombach and Newberry (2001) suggested that porphyry-style stockworks systems are characteristic
of the intrusion-related gold systems emplaced at shallow levels (<0.5 kb). Country rocks, as well as
dikes and sills related to the larger granitoid intrusions, may also contain both gently-to steeply-
dipping sheeted veins or vein sets, as well as disseminated gold. In Proterozoic examples (Telfer and
Pine Creek), saddle reefs and/or other bedding-parallel veins formed in sedimentary host rocks. In
carbonate host-rocks, skarns may be developed. Apart from the typical high-level skarns, many of
these mineralization styles are similar to those ascribed to orogenic gold deposits. The Pogo deposit
in eastern Alaska, defined as an intrusion-related gold deposit by Smith et al. (1999), in particular,
contains quartz lodes with characteristics typical of a classic, shear zone-hosted orogenic gold
deposit.
In most cases, mineralization is approximately coeval with host or associated intrusions. At Clear
Creek, Yukon, in the Tintina gold province, Ar-Ar ages of hydrothermal mica of 91 and 90 Ma are
essentially identical to U-Pb ages of 92-91 Ma for the host stocks (Marsh et al., 2003). Similarly, Re-
Os ages from hydrothermal molybdenite at the Fort Knox deposit in the Alaskan part of the belt
overlap ages of granitoid crystallization (Hart et al., 2001; Selby et al., 2002). Overlapping ages also
characterize ores and host rocks at the Petrackhova hora deposit in the Bohemian Massif, where Re-
Os dating of molybdenite from a sheeted gold-bearing vein yields an age of 342±1.5 Ma and the Rb-
Sr whole-rock age on the host granodiorite is 348±23 Ma (Zacharias et al., 2001).
Fluid inclusion data (McCoy et al, 1997; Cole et al., 2000; Baker and Lang, 2001; Zacharias et al.,
2001) typically indicate low-salinity H2O-CO2±CH4±N2 fluids formed the intrusion-related gold
deposits, which is similar to those responsible for most orogenic gold deposits (e.g., Ridley and
Diamond, 2000). However, at least one fluid population of aqueous brines is reported from the
intrusion-related gold deposits in the Yukon (e.g., Baker and Lang, 2001, Marsh et al., 2003), and at
Page 18
the Tennant Creek (Zaw et al., 1994), Shotgun (Rombach and Newberry, 2001) and Petrackhova hora
(Zacharias et al., 2001) deposits. This may partly account for the high Cu concentrations in the latter
three examples. Pressure and temperature estimates for intrusion-related gold deposits are similar to
those of epizonal to hypozonal orogenic deposits, with ranges stated to be less than 0.5 kb and 200oC
to greater than 3 kb and 600oC, as reviewed by Lang and Baker (2001).
A major problem with the recognition of the intrusion-related gold deposits, as defined by Sillitoe and
Thompson (1998), is the diversity of tectonic settings, metal associations, wallrock alteration and
fluid chemistry, among other parameters. Even within the gold-dominant,
Au±Bi±W±As±Mo±Te±Sb association of Lang et al. (2000), there is a considerable diversity of
deposit characteristics that are typically ascribed to variations in depth, host rocks, or distance from a
causative pluton (Hart et al., 2002). If the intrusion-related deposits of the Pine Creek and Telfer
districts are included, as well as the syenite-associated deposits of Robert (2001), the range of
possible tectonic settings and associated intrusion types, alone, is very large.
Models for intrusion-related deposits typically stress their distinction from orogenic gold deposits
based on a distribution of gold ores surrounding a causative pluton. However, in many cases,
unequivocal supporting data to confirm the genetic link between mineralization and the host or
spatially-associated intrusion may not exist. Classifying a deposit as intrusion-related based upon its
geochemical signature may not be satisfactory because all the elements in association with Au in such
deposits may also be anomalous in orogenic gold deposits, and thus none are specific to magmatic
systems (Goldfarb et al., 2000). For example, the elements W, Bi, and Te are commonly regarded as
critical pathfinders for the intrusion-related gold systems (e.g., Lang et al., 2000). Yet, within the
class of orogenic gold deposits: (1) W in the Otago goldfields (South Island, New Zealand) has been
mined from many of the lodes (Paterson, 1977), where there are no granitoids (Henley et al., 1976);
(2) Bi and Te are highly anomalous in the Ouro Preto ores of the Quadrilatero Ferrifero, Brazil, also
without any spatially associated granitoids (Chauvet et al., 2001); (3) a Bi-Te-W signature with as
much as 470 ppm Bi, is recognized at the Beaver Dam turbidite-hosted gold deposit in the Meguma
terrane (Nova Scotia, Canada; Smith and Kontak, 1988); (4) all three elements are highly anomalous
at the Independence deposit (Willow Creek district, Alaska, USA) within veins that cut a batholith
emplaced 10 m.y. prior to hydrothermal activity (e.g. as much as 150 ppm Bi and 62 ppm Te;
Madden-McGuire et al., 1989); and (5) Bi-bearing telluride minerals commonly occur in the giant
Golden Mile deposit at Kalgoorlie (Western Australia). It is most likely that the common association
of these elements, as well as Sb, As, and Au, simply results from their high degree of mobility within
moderate temperature, low salinity, CO2- and H2S-bearing crustal fluids. This non-specific
association may also reflect some degree of crustal inheritance, which has led to repeated anomalous
concentrations of specific elements in diverse types and ages of ore deposits in certain regions (e.g.,
Titley, 2001).
Similarly, the geochronological evidence presented for the synchronicity of intrusions and
mineralization is not unique to intrusion-related gold deposits. As stated above, orogenic gold
deposits and granitoids are spatially and temporally associated in the majority of the orogenic gold
provinces. Well-constrained geochronology from the southern Alaskan accretionary prism, for
example, shows that spatially associated Eocene flysch-melt granitoids and orogenic gold deposits
formed at the same time along the 2,000 km-strike length of the belt (Haeussler et al., 1995).
Therefore, the presence of gold and widespread granitoids of the same age in the same belt cannot be
used as a criterion to identify intrusion-related gold systems. In fact, workers have even suggested
such an association to indicate a granitoid-gold connection between the Otago goldfields and the
major, roughly coeval Fiordland batholith that is exposed more than 200 km to the west (deRonde et
al., 2000). However, it remains speculative as to whether this magmatic arc continues at depth to the
east and thus could underlie the gold lodes. Deposits in the Pataz province in Peru were classified as
intrusion-related in recent models (e.g., Sillitoe and Thompson, 1998), but new geochronology has
Page 19
shown that the ores post-date the host batholith by about 15 m.y. and make such a genetic link
unlikely (Moritz, 2000; Haeberlin, 2002).
One outstanding problem deserving particular study is the structural timing of intrusion-related
deposits, that is their timing relative to the development of penetrative structures of the host rocks and
the timing of mineralization relative to the tectonic evolution of the gold district. This aspect of
intrusion-related deposits has yet to be studied adequately. As pointed out above, unequivocal
orogenic deposits can be demonstrated to have formed in structures produced or reactivated during
regional compressional or transpressional deformation. In contrast, it seems that many intrusion-
related deposits have formed tens of millions of years after the penetrative fabrics of their host rocks
and after the main compressional/transpressional stage in the evolution of the district (e.g., Yukon
part of the Tintina province; Poulsen et al., 1997). However, other deposits that are considered to be
intrusion-related (e.g., Tower Hill of the Leonora area, Western Australia; Witt, 2001) are interpreted
to pre-date much of the regional deformation and to have formed much earlier than orogenic gold
deposits hosted in younger sequences within the same metallogenic province.
Paleodepth estimates for formation of the intrusion-related gold systems (see figure 6 in Lang et al.,
2000) indicate pressures below those at which granitic magmas are likely to exsolve CO2-rich fluids
(i.e. >3 kb, Eggler and Kadik, 1979; Cline and Bodnar, 1991). Lang et al. (2000) compared their data
with those of Nabelek and Ternes (1996) from the Haney Peak Granite, but the fluids described by
the latter exsolved at pressures of about 3.5 kb, much greater than those expected at <7 km depths. In
almost all cases, the intrusion-related gold ores cut the granitoids, suggesting that CO2–bearing ore
fluids escaped from deeper, still uncrystallized parts of an evolving magmatic system, if indeed they
are magmatic in origin.
A major problem in all provinces containing intrusion-related gold deposits, therefore, is a

mechanism for the exsolution of carbonic fluids from the magmas at the depths depicted in the
deposit models. Most models from the better-studied intrusion-related gold systems show that ore
deposition occurs late in the history of the local magmatic-hydrothermal system and at epizonal
crustal levels. Yet, at such shallow crustal levels, volatile saturation typically should take place early
in the crystallization history of the causative melts (Candela, 1997) and is likely to be characterized
by H2O>>CO2 during emplacement of the more fractionated intrusions (Lowenstern, 2001). Baker
(2002), using relationships shown in Lowenstern (see figure 5, 2001), argues, however, that
significant amounts of CO2 in many intrusion-related gold systems may have exsolved throughout the
ascent of a magma from its source region.
An additional concern is that some proposed intrusion-related gold system fluids evolve from low
salinity to high salinity (e.g., Baker and Lang, 2001), whereas the opposite is recorded for other
intrusion-related deposits (e.g., Zacharias et al., 2001). The controls on these changing fluid
compositions, and what they mean for related gold tonnages, are poorly understood. A problem
specific to the syenite-associated deposits is why they are restricted to provinces in which there are
undoubted orogenic gold deposits, some of which are hosted by syenites (e.g. Duuring et al., 2000).
If they are magmatic-hydrothermal deposits, then at least anomalous gold concentrations should be
characteristic of syenites in other settings, for example, intracratonic environments, where orogenic
gold deposits are absent. Finally, it is important to determine the potential of intrusion-related gold
deposits as high-grade exploration targets. Where hosted by granitoids, or immediately adjacent
rocks, they are typically bulk-tonnage ores with grades below about 1 g/t Au (cf. other magmatic
deposits such as porphyry-style systems with similar grades), with Pogo (eastern Alaska), if indeed it
is correctly classified as intrusion-related, being a possible exception. However, it is noteworthy that
the Pogo deposit is located further seaward than the other intrusion-related gold lodes of the Tintina
province and distal to any recognized Sn-W province.
1.5. Gold Deposits with Atypical Metal Associations: Where

Page 20
Do They Fit?
1.5.1 Definition and contrasts with orogenic gold deposits
A number of enigmatic gold-bearing deposits occur in provinces dominated by orogenic gold

deposits. Broadly, these deposits fall into two groups, those enriched in Cu±Mo (e.g.,
McIntyre/Timmins, Canada; Boddington, Australia) and those enriched in Cu-Zn±Pb±Ag and/or
abundant pyrite (e.g., Bousquet, Canada; Mt Gibson, Australia; several deposits in Tanzania and
Kenya; Carolina slate belt, USA; a few deposits of the Mt. Read VMS province, Australia), as well as
those elements more normally associated with orogenic gold deposits (e.g. As, B, Bi, Sb, Te, W).
The Hemlo deposit, Canada, with its anomalous enrichment in Ba, Mo and Hg, among other
elements, does not fall neatly into either group. Collectively, these deposits are the source of much
controversy, with three main model types proposed for their genesis: (1) deformed and
metamorphosed Au-rich porphyry to epithermal or VHMS deposits, (2) porphyry or VHMS deposits
in which Au has been mobilized during deformation and metamorphism, or (3) porphyry or VHMS
deposits which have been overprinted by orogenic gold systems later in the history of the orogen.
The superimposition of two contrasting ore-deposit styles may seem fortuitous, but is to be expected
given that structures controlling the location of early mineralization may be reactivated during
subsequent events, and that wallrock alteration (± massive sulfide minerals) changes the physical
properties of rocks, generating new competency contrasts that can be exploited during later
hydrothermal activity (e.g., Groves et al., 2000). An excellent example is provided by the
coincidence of orogenic gold mineralization and massive Fe-Ni-Cu sulfide ores at the Hunt mine,
Kambalda deposit, Western Australia (Phillips and Groves, 1984), and another is the overprint of
VHMS-style mineralization by orogenic gold at the Mt. Gibson deposit, Western Australia (Yeats et
al., 1996).
It is difficult to generalize about such deposits because there is so much variation in detail between
individual examples. For this reason, a few of the better documented systems are very briefly
outlined below.
1.5.2 Probable modified porphyry/epithermal systems
The Hollinger-McIntyre deposit at Timmins is arguably the first of this type widely recognized as a
potential modified porphyry deposit in studies spanning the last 25 years. This deposit was the
largest producer in Canada (>1,000t Au), with the bulk of the ore derived from quartz-carbonate veins
and a minor proportion from earlier, porphyry-style, disseminated and stockwork Cu-Ag-Au-Mo
mineralization, mainly in the Pearl Lake porphyry (e.g., Smith and Kessler, 1985; Burrows et al.,
1993). The latter averaged 0.67% Cu, 0.59 g/t Au and 2.93 g/t Ag, with approximately 0.05% Mo
(Burrows and Spooner, 1986), making it completely unlike orogenic gold deposits with respect to Cu
content and Au/Ag ratio (≈ 0.2). In addition to sulfide minerals, anhydrite and hematite are present
(Burrows and Spooner, 1986). An early timing is suggested by cross cutting dikes that are about 15
m.y. younger than the host porphyry (e.g., Marmont and Corfu, 1989). Overprinting quartz-
carbonate vein systems are localized by the Hollinger shear zone, which cuts the Pearl Lake
porphyry, with the attitude of gold ore bodies partly controlled by anisotropies created by the rigid
porphyry body (Burrows et al., 1993). Their mineralogy, metal associations and wallrock alteration
are typical of orogenic gold deposits. The wealth of fluid inclusion and stable isotope data on the
Hollinger-McIntyre deposit does little to resolve the precise genetic relationships between the quartz-
carbonate veins and porphyry styles, but a reasonable interpretation is that a magmatic-hydrothermal
Cu-Ag-Au-Mo system was overprinted by an orogenic gold system, in part localized by the rigid
Pearl Lake porphyry. The other large deposit of the Timmins district, the Dome deposit, shows a
similar overprinting on an early Cu-Au±Mo system (Gray and Hutchinson, 2001).
Page 21
The Boddington gold deposit is the second largest gold resource (>800t Au) in the Yilgarn Craton of
Western Australia (e.g., Symons et al., 1990). Although mainly exploited to date for its supergene-
enriched gold in lateritic regolith, it contains a large hypogene resource, albeit low grade (<1.1 g/t
Au). The Cu-Au-Ag-Mo-Bi-W association, stockwork-style veins and variably saline, low-CO2 fluid
inclusions (Roth, 1992), for at least part of the mineralization, clearly distinguish the deposit from
Yilgarn orogenic gold deposits. Combined with the grade-tonnage characteristics, the occurrence of
mineralization in and near dioritic intrusions, the dominant amphibole-biotite alteration and
amphibole-rich veins, and the maximum mineralization temperatures being higher than peak
temperatures of the regional upper-greenschist facies metamorphism, are strong evidence in favor of
Boddington being a porphyry-style deposit, specifically of the diorite class, as interpreted by Roth
(1992). The detailed structural studies of Allibone et al. (1998), however, cast doubt on the syn-
diorite formation of the mineralization, and instead place most of the mineralization late in a Dl to D4
deformation sequence, in which there were successive generations of veins with different
mineralogical compositions. In this model, the timing of mineralization is broadly similar to that of
the Yilgarn orogenic gold deposits, but the fluid compositions and temperatures still imply a local
magmatic source. The lack of a recognized suitable magmatic source at the time of mineralization, as
proposed by Allibone et al. (1998), is a problem, although their data imply that Boddington is
another, albeit somewhat different, type of intrusion-related gold deposit. McCuaig et al. (2001) have
recently presented a well-constrained, two-stage ore genesis model for the Boddington deposit. They
define intrusion-related, gold-forming events at ca. 2700 and 2612 Ma, with the latter post-tectonic
magmatism responsible for the majority of the ore. It is stressed that, if this interpretation is correct,
Boddington would represent the first example of a major Archean Au-Cu intrusion-related deposit
that is associated with post-tectonic magmatism. The question remains whether the main
mineralization at Boddington post-dates the intrusion and it is simply coincidence that the deposit has
most of the characteristics of a diorite porphyry Cu-Au-Mo deposit (e.g., Hollister, 1978), whether an
original porphyry system has been reactivated, remobilized and/or overprinted during subsequent
orogenic events, or whether Boddington is an exotic example of the broadly-defined intrusion-related
group of gold deposits, as suggested by McCuaig et al (2001).
The Hemlo deposit, in the Wawa subprovince of Canada, containing approximately 600t Au at a
grade of >7 g/t Au (Harris, 1989) in three main deposits (Williams, Golden Giant and David Bell), is
an even more contentious deposit, mainly because of the complex magmatic, structural and
metamorphic history of the district (e.g. Muir, 2002). Similar to more poorly understood gold
deposits (e.g., Big Bell, Western Australia, Mueller et al., 1996), Hemlo is located in a high-grade
metamorphic domain, in this case of mid-amphibolite facies (e.g. Corfu and Muir, 1989b). The metal
association of Au-Mo-Sb-As-Hg-V-Tl-Ba, the high Hg and Ag contents of some native gold (e.g.,
Harris, 1989), and the unusual mineralogy, together with the moderate to high salinities of ore fluids
(Pan and Fleet, 1992) and consistently negative δ34S of ore minerals (Cameron and Hattori, 1985),
clearly distinguish this deposit from most orogenic gold deposits. However, the high gold grades,
high Au:Ag ratio and metal association also make correlations with normal porphyry, epithermal or
VHMS systems equivocal. Potassic and calc-silicate alteration are common, but there is a complex
array of alteration styles of both prograde and retrograde timing (Pan and Fleet, 1992).
Mineralization is located in the Moose Lake porphyry and an adjacent fragmental unit within a host
sequence of metasedimentary rocks, all of which are cut by later calc-alkaline granitoid bodies (Corfu
and Muir, 1989a). Four stages of deformation are recognized in the sequence (Michibayashi, 1995;
Lin, 2001).
Genetic models for Hemlo have ranged from those involving pre-metamorphic seafloor or epithermal
systems (Cameron and Hattori, 1985), to pre- to syn-metamorphic hydrothermal events (Burk et al.,
1986; Kuhns et al., 1986) that perhaps are magmatic in origin (Muir, 2002), to post-peak
metamorphic events related to emplacement of granitoid intrusions (Pan and Fleet, 1992) or to
mylonite development (Hugon, 1986; Corfu and Muir, 1989b). Recent structural studies
(Michibayashi, 1995; Powell and Pattison, 1997; Lin, 2001) support a pre-D2 and pre-amphibolite-
facies metamorphism timing for the main mineralization stage, or a pre- to early-D2 timing, but a
Page 22
robust genetic model has still not been established. Perhaps an analogy in terms of metal association,
if the Hemlo ores are pre-metamorphic, is the submarine epithermal-style mineralization at Conical
Seamount, near Lihir Island, Papua New Guinea (Herzig et al., 1999). It is still possible that part of
the complexity is due to overprinting of mineralization styles. For example, Michibayashi (1995)
argued for second-stage remobilization of Au (plus barite and stibnite) in late shear zones, and Pan
and Fleet (1992) suggested that some Au was deposited during late calc-silicate alteration.
1.5.3 Probable modified VHMS or submarine epithermal systems
The Bousquet group of gold deposits, with a total resource of about 280t Au and a grade of about
5.35 g/t Au (Marquis, 1990), situated immediately north of the Larder Lake-Cadillac Fault (a major
break) in the Abitibi belt, Canada, is suggested here as an example of a modified seafloor gold
system. As described by Marquis et al. (1990), Tourigny et al. (1989, 1993) and Robert and Poulsen
(1997), the deposits lie in a 500-m-wide shear zone within felsic volcaniclastic rocks, and mafic
volcanic and volcaniclastic rocks, which are locally intruded by a differentiated gabbro to
trondhjemite complex and metamorphosed to greenschist facies. The overall morphology of ore
bodies is strongly controlled by fabrics within the shear zone. Ore bodies comprise a variety of
styles, including highly deformed massive pyrite (with lesser galena, sphalerite or chalcopyrite)
lenses, various generations of foliation-oblique to foliation-parallel quartz-carbonate-sulfide veins
with >25% sulfide minerals, and disseminated (5–20% pyrite) mineralization in schistose host rocks.
From the nature of the mineralization, particularly the high sulfide content, high Au-Cu correlation,
and pre-shearing and pre-metamorphic timing for formation of the sulfide minerals, there has been
general agreement that the pyrite-rich ore bodies formed during synvolcanic sub-seafloor or
exhalative processes (e.g. Marquis et al., 1990; Tourigny et al., 1993) with similarities to high-
sulfidation epithermal systems (e.g. Poulsen and Hannington, 1996). The major area of controversy
concerns whether most of the Au was synvolcanic and remobilized by subsequent events (e.g.,
Tourigny et al., 1993) or was introduced during the overprinting orogenic gold event that affected the
surrounding district (e.g., Marquis et al., 1990). One of these processes must have operated given the
fact that some quartz-sulfide-gold veins cut all fabrics and that some gold is in textural equilibrium
with retrograde assemblages (Poulsen and Hannington, 1996).
The Paleoproterozoic Boliden Cu-Au-As deposit, within the 1.9 Ga Skellefte volcanic succession of
northern Sweden (Allen et al., 1996; Hannington et al., 1999), is another somewhat enigmatic
deposit. The Boliden deposit consisted of two, large, subvertical pyrite lenses enveloping several
smaller arsenopyrite-rich lodes surrounded by a narrow aluminous alteration envelope rich in
andalusite and quartz. The metal association of the ore was unusual, comprising Cu-Au-As (average
As grade 6.9%) with significant Ag, Bi and Se. Most Au (95%) was contained within fine-grained
massive pyrite and arsenopyrite, although bonanza-type gold mineralization was developed locally in
arsenopyrite lenses (>200 g/t Au ) and large crosscutting quartz - tourmaline veins (>600 g/t Au).
Such deposits could represent orogenic gold overprints, but Bergman Weihed et al. (1996) showed
that all ore bodies have been affected by all phases of deformation and by regional metamorphism,
and suggested that coarse-grained gold was derived from refractory gold in arsenopyrite during D1
brecciation. Currently, the Boliden deposit is considered to represent coincident submarine
epithermal-style and subsurface replacement-style mineralization developed in a shallow-water
environment (Allen et al., 1996).
The latest Proterozoic to earliest Paleozoic gold-rich VHMS deposits of the Carolina slate belt,
southeastern USA are similarly enigmatic. Deposits such as Ridgeway, Haile and Brewer are
associated with metamorphosed felsic volcanic units, which had previously been altered on the
seafloor to propylitic, argillic and advanced argillic assemblages surrounding pyrite–rich zones with
minor base metals, enargite and arsenopyrite. The gold ores in these massive sulfide bodies could
represent seafloor hot springs or epithermal mineralization superimposed on VHMS deposits
(Worthington et al., 1980; Crowe, 1995; Bierlein and Crowe, 2000). As with many VHMS deposits,
those of the Carolina slate belt occur in accreted terranes also characterized by post-accretionary
Page 23
orogenic lode-gold deposits. This has led to suggestions that the larger, low-grade gold resources
within the felsic volcanic rock sequences are also later syn-tectonic orogenic gold deposits related to
ductile shear zones (e.g. Tomkinson, 1988; Hayward, 1992). Dating of gold-stage molybdenite in
these deposits by Re-Os (Stein et al., 1997), however, indicates that ore formation was pre-
accretionary and approximately 200 m.y. prior to deformation and regional metamorphism.
A number of the VHMS deposits of the Cambrian Mount Read Volcanic Belt, western Tasmanian,
are also anomalous in Au. At the Henty deposit, gold ores in late brittle fractures have been related to
remobilization from seafloor concentrations during Devonian deformation, more than 100 m.y.
subsequent to the original VHMS deformation and prior to emplacement of nearby post-tectonic
granitoids (Halley and Roberts, 1997). In contrast, at the Rosebery deposit, Zaw et al. (1999)
suggested that fluids exsolved from the post-tectonic Devonian granitoids remobilized Au into
auriferous replacement zones. Huston (2001) favors an Ordovician age for the Cu-Au deposits
throughout the Mt. Lyell district and indicates these may not be modified VHMS deposits at all, but
rather porphyry Cu deposits and associated high-sulfidation epithermal ores.
Even some of the youngest orogens show some evidence of orogenic gold formation that can be
related in places to remobilization of the VHMS deposits during metamorphism and deformation. In
the Cordilleran orogen, the very high levels of H2S in gold-bearing quartz veins of the Sumdum Chief
gold deposit at the southern end of the Eocene Juneau gold belt, and in an area of extensive seafloor
sulfide mineralization, suggest that some of the volatiles in the epigenetic ores were derived from
syngenetic sources (Goldfarb et al., 1988). At the northern end of the same gold belt, Newberry et al.
(1997, p. 142) used Pb isotope data to show that components from VHMS deposits were remobilized
into veins during the Eocene gold-forming event.
Less-equivocal field, textural and geochronological evidence of overprinting of a weak VHMS

system by an orogenic gold system is recorded at the small Mt Gibson deposit in Western Australia
(Yeats et al., 1996). The realization that VHMS deposits can be overprinted by later orogenic gold
systems should initiate re-examination of anomalously Au-rich VHMS systems in orogenic gold
provinces, such as the Horne deposit in the Abitibi belt.
Despite considerable research effort, many of these deposits with atypical metal associations remain
enigmatic, with little consensus on their affinities and genesis. In many cases, there is a lack of the
well-integrated field-based structural, alteration, geochemical, isotope and fluid inclusion studies
needed to define the total history of the deposits and provide potential mass balances of metal
introduction (or removal) or remobilization at each stage of deposit evolution. In almost all cases,
there is a lack of knowledge regarding the precise timing constraints on the stages of ore development
needed to resolve models, mainly because of the apparent lack of ore-related minerals suitable for
dating by robust isotopic techniques. In general, each deposit within this loosely defined group has
been studied in isolation, such that there is little recognition of a potential grouping of deposits with
specific metal associations, alteration, fluid characteristics and overprinting relationships. For other
deposit styles, the ability to recognize genetically-related groups, and to integrate relationships across
those groups, has led to more robust deposit and genetic models, but this is lacking in this group of
deposits because of the complex and controversial nature of each of the deposits discussed above.
1.6. Giant Gold Deposits in Metamorphic Belts?
The defining parameters and origins of giant ore deposits of all deposit classes are currently
the most important topic for large exploration companies. Studies of giant mineral deposits,
Page 24
in general, reveal that there is no simple explanation for their giant size relative to adjacent
deposits in similar settings (e.g., Hodgson et al., 1993). It is more likely that giant deposits
form via the conjunction of a number of favorable physical and chemical processes, rather
than as a result of special processes (Phillips et al., 1996; Sillitoe, 2000). For example, a
study of the distribution of large versus small gold deposits in the Norseman-Wiluna Belt of
Western Australia, using an empirical approach within a GIS, showed that there are no
significant differences in individual parameters. However, only 15% of the deposits,
containing more than 80% of known gold resources, lie in zones of greatest overlap of the
critical defining parameters (Groves et al., 2000).
A brief discussion of those gold deposit types that contain significant giant deposits, or even
clusters of world-class deposits, is given below, together with a brief overview of those
special features that are potential defining parameters of giant deposits. The precise
definition of those parameters, however, remains one of the outstanding problems not only
of gold deposits in metamorphic rocks in many of the worlds major orogens, but also of
mineral deposits of all types globally.
1.6.1 Affiliation of giant gold deposits in metamorphic belts
If 2,500 t Au (≈ 75 Moz) is taken as the lower resource limit of supergiant gold deposits (e.g.
Laznicka, 1999), then it is evident that only the unique Witwatersrand deposits, possibly
Muruntau, and perhaps the undeveloped and poorly understood Sukhoi Log deposit, qualify
as supergiants, although the Kalgoorlie and Timmins goldfields are close to this limit as
districts. A large number of the giant deposits (some as clusters of world-class deposits—
e.g. the Mother Lode belt) in metamorphic belts are orogenic gold deposits ≥250 t Au (Table
4 in Appendix). There are also numerous world-class deposits (>100t Au or 3 Moz) within
gold provinces in metamorphic belts world-wide. There are also some giant deposits in the
category of deposits with atypical metal associations.
There are no individual, giant intrusion-related gold deposits (unless the classification of
Sillitoe, 2000, is used for Muruntau, Boddington, and Hemlo, or the inferred resource at
Donlin Creek, Alaska proves to be economic) and the provinces that contain deposits
ascribed to this class are generally less well-endowed than others (Fig. 3). In fact, there are
few world-class deposits of this type, despite the enormous tonnage of some of them,
because the ore grade is generally low. Exceptions include Fort Knox and Pogo in the
Alaskan side of the Tombstone gold province, if Pogo is indeed an intrusion-related deposit,
and perhaps Cosmo-Howley and Telfer in northern Australia. Although debatable, the 27
Moz gold resource of the Jiaodong Peninsula in eastern China is considered by some
workers to be an example of a giant intrusion-related gold province (Poulsen et al., 1990;
Sillitoe and Thompson, 1998). Only three individual deposits in this cluster of hydrothermal
systems, however, fit the world-class category (e.g., Zhou et al, 2002). Within the syenite-
associated group of Abitibi belt deposits, only Malartic is world-class; all others are
relatively small.
1.6.2 Definition of special features
As shown in Table 2, there are no easily discernable single factors that uniquely define giant
gold deposits at the deposit scale within broad belts of metamorphic rocks, as pointed out by
Page 25
Sillitoe (2000) for gold deposits of any type. For example, among orogenic gold deposits,
parameters such as deposit age, host rock lithology, structural control, type of spatially
associated intrusive rocks, and redox state of the ore fluid are variable. It is perhaps more
beneficial to examine the far-field tectonic controls on those gold provinces that contain
giant orogenic and other gold deposits; that is, examine the problem at a larger scale (Table
3).
Table 2. Comparison of some giant orogenic gold deposits in terms of deposit-scale

parameters. Listed in order of decreasing age.Length of bar indicates certainty of
interpretation.
(m = mafic; f = felsic; s = sedimentary; green. = greenschist; amph. = amphibolite; prox. =

proximal; dist. = distal; ox. = oxidised; n = normal; red. = reduced)
It is clear from modern arc-related mineral deposit types, such as porphyry Cu-Au and
epithermal Ag-Au deposits, that the geometry of subduction systems in convergent margin
settings is important in controlling the location of giant metallogenic provinces (e.g., Sillitoe,
1997; Kerrich et al., 2000; Kay and Mpodozis, 2001). There are indications that there may
be similar tectonic controls on orogenic gold deposits, with Goldfarb et al. (1991) first
demonstrating a relationship to changing plate motions in the generation of such deposits in
the relatively young accretionary collisional events of southern Alaska. Subsequently, other
workers have interpreted structural data to suggest similar tectonic events that controlled
formation of some of the oldest orogenic gold deposits (e.g., de Ronde and de Wit, 1994).
Wyman et al. (1999) demonstrated that the giant Abitibi gold province probably formed in
an environment where subduction reversal occurred, and Kerrich et al. (2000) and Goldfarb
et al. (2001) summarized a number of other scenarios where subducted spreading ridges,
subduction roll back, and other crustal processes leading to asthenospheric upwelling or
crustal thickening, and consequent thermal anomalies, played a key role in the generation of
giant orogenic gold provinces in metamorphic belts. Similarly, the presence of komatiites in
Archean and Paleoproterozoic belts that contain giant gold provinces signals the interaction
of plumes with subduction-related environments, again potentially producing anomalous
plate geometries (Dalziel et al., 2000) and thermal anomalies. A challenge is to develop
criteria to recognize some of these specific processes in the volcanic and intrusive rock
record in terranes where there is only indirect evidence of subduction.
The presence of crustal-scale deformation zones in linear volcano-sedimentary belts appears

a factor common to most gold provinces containing giant gold deposits of all ages. Such
crustal-scale structures also localize porphyry and lamprophyre dike swarms, and commonly
juxtapose volcanic and sedimentary sequences. Perhaps most importantly is that without
such structures, it is questionable whether enough fluid can be focused to form a giant gold
deposit. Unlike many sedimentary basins and carbonate platforms, where inherent
permeability of the rock sequences can favor migration of major fluid volumes, highly-
Page 26
deformed, metamorphosed greenstone belts and marine sedimentary sequences will have a
low primary permeability and major fluid flow must be along secondary structures (e.g.,
Jamtveit and Yardley, 1997).
Table 3. Comparison of some giant orogenic gold deposits in terms of province-scale

parameters. Listed in order of decreasing age. Length of bar indicates certainty of
interpretation.
(delamin. = delamination; green. = greenschist; amph. = amphibolite)
The giant orogenic gold deposits consistently have a complex regional-scale geometry,
commonly due to the reactivation or reorientation of complex thrust, commonly duplex,
structures (e.g., Timmins, Kalgoorlie, Ashanti), which may juxtapose competent and
incompetent segments of lithostratigraphy (Groves et al., 2000). In other cases, the deposits
are situated close to the closures of domal structures (e.g., Bendigo) in sequences situated in
the hangingwall of sole thrusts in thin-skinned tectonic belts (e.g., Gray, 1997). These
controlling geometries are large, complex and commonly well-sealed by relatively
incompetent rocks, making them zones of heterogeneous stress that focussed fluid flux.
Reactivation of suitably oriented structures late in the deformational history of an orogen,
commonly during regional uplift and retrograde metamorphism (Groves et al., 1988),
appears to be a common factor.
In all cases, complex lithostratigraphy provides paired strength and chemical contrasts that
allow very selective failure and consequent fluid flux into chemically reactive rocks (Phillips
et al., 1997; Groves et al., 2000). These rocks commonly have high Fe/Fe+Mg (± Ca) ratios
(e.g., tholeiitic dolerites or basalts atKalgoorlie, Timmins, Kolar, and some Mother Lode
deposits; BIF at Homestake) and/or anomalous carbon contents (e.g., Ashanti), although
these may be volumetrically minor (e.g., indicator slates in Victorian goldfields: Bierlein et
al., 2001c).
Finally, many giant deposits are the sites of multiple generations of gold mineralization. In
this case, the most obvious deposits are those in the category of deposits with atypical metal
associations, such as Hemlo and Boddington, where, in all probability, earlier mineralization
styles have been overprinted and/or remobilized during later events. The Hollinger-McIntyre
system at Timmins also fits into this category. At Kalgoorlie, the Golden Mile lode systems
are overprinted by the Oroya and Flat Lode systems which, in turn, are overprinted by the
Mt Charlotte style mineralization (Mueller et al., 1988; Clout et al., 1990; Bateman et al.
2001). At Bendigo, early, commonly Au-poor, quartz saddle-reefs are re-mineralized by
high-grade gold (Cox et al., 1991). A complex series of mineralization styles at Muruntau
are also suggestive of multiple, chemically distinct fluid-flow events (Graupner et al., 2001).
1.7. Summary of Outstanding Problems for Classification

Page 27
and Genetic Models
1.7.1 Classification
Many authors of recent papers appear to accept the classification of most lode-gold deposits in
metamorphic belts as orogenic gold deposits (Bouchot and Moritz, 2000; Hagemann and Brown,
2000), although specific examples are controversial (Sillitoe and Thompson, 1998). However, the
definition of intrusion-related gold deposits varies between papers and, because this model has been
proposed only recently, comprehensive lists of provinces or deposits ascribed to this type are lacking.
Most of these deposits are Phanerozoic, but Proterozoic (e.g., Pine Creek, Telfer) and Archean
examples (e.g., syenite-associated class of Robert, 2001) may be included in this type. It is probable
that there are several sub-types within this class of intrusion-related deposits, and this requires better
resolution and classification. The deposits with atypical metal associations are to some extent unique.
These different deposit types are well classified in the scheme proposed by Poulsen et al. (2000).
1.7.2 Orogenic gold deposits
The major outstanding problems in the understanding of orogenic gold deposits are: (1) the precise
timing of many deposits with respect to magmatic, metamorphic and deformational events in the host
terranes, (2) the exact nature and release mechanisms of the deeply-sourced ore fluids that dominate
these systems, (3) the configuration of the regional, crustal-scale hydrothermal conduits and fluid
flow within them, (4) the precise transport and depositional mechanisms, particularly in S-poor
systems, where As, Bi, Sb, W and/or Te may be abundant, and (5) the precise controls on the
generation of world-class to giant examples. A specific problem is the distinction between orogenic
gold deposits and intrusion-related deposits with similar ore-element associations and ore fluid types
in the same metamorphic belts. Another is the distinction between skarns and hypozonal orogenic
gold deposits in high metamorphic-grade settings.
1.7.3 Intrusion-related gold deposits
A major problem in the classification of intrusion-related gold deposits is the wide variety of deposits
included in this category by different authors. Outstanding problems in the understanding of this
deposit type include: (1) unequivocal evidence that connects individual deposits to their proposed
source intrusions, (2) the large range of granitic compositions and redox states that are ascribed to
source intrusions, (3) the mechanism for exsolution of aqueous-carbonic fluids from shallow-level
(<7 km deep) granitoids late in the differentiation history of the source magmas, and (4) the structural
timing and the structural and tectonic controls on deposit style.
1.7.4 Gold deposits with atypical metal associations
It is difficult to treat these deposits as a group because of their highly diverse characteristics and the
complications imparted by the overprinting deformation and metamorphism. For these atypical
deposits, major outstanding problems include: (1) the possible existence of multiple mineralizing
events and their variable importance, (2) timing of alteration and mineralization relative to their host
penetrative structures and to metamorphism, (3) distinction between local Au remobilization and
introduction of Au during an overprinting hydrothermal event, and (4) the geometry of the deposits
and host rocks in the case of severely deformed deposits (e.g. Lin, 2001). Resolution of the nature
and origin of these complex deposits requires integration of all geometric, structural, alteration,
Page 28
mineralogical and geochemical features of the deposits and their host rocks. This will provide the
basis for recognizing factors leading to the formation of giant examples of this anomalous group.
.
1.7.5 Temporal distribution of gold deposits in terms of crustal
evolution
Goldfarb et al. (2001) have drawn together all published, robust age data for orogenic gold deposits to
define their global temporal distribution. There are, however, outstanding uncertainties in the age of
some large deposits and gold provinces that, particularly in Brazil, Kolar, the Baikal region and the
orogens surrounding the southern Siberian craton, northern Africa, and China, need resolution.
Goldfarb et al. (2001) have also made some suggestions on the links between the formation of
orogenic gold provinces, the evolution of plate-tectonic processes and the periods of major
continental-crust formation. However, these models need to be further tested and refined by
examination of the temporal distribution of other ore deposit styles strongly influenced by tectonic
processes. The overall rarity of economic porphyry Cu-Au-Mo deposits and associated epithermal
systems beyond the Mesozoic, but their appearance, albeit modified, in Archean terranes (e.g.
possibly Hollinger-McIntyre, Boddington, Hemlo) also needs an explanation.

and Genetic Models
1.8.1 Integrated province-scale and deposit-scale research
Modern ore deposits research is based on sophisticated techniques using expensive research
equipment and tackles small, specific research problems within an ore deposit or province. What is
required to resolve the problems related to gold deposits in metamorphic belts is an integrated
approach, combining thorough field-based studies with careful structural and petrographic studies,
before more sophisticated geochemical and isotopic analysis is applied. Such studies are available for
some deposits or districts (e.g., Sigma-Lamaque: Robert and Brown, 1986a, b; northern Juneau gold
belt: Miller et al., 1995; central Victorian goldfields : Bierlein et al., 2001b), but there are few
integrated studies of the giant gold deposits carried out by a single research team working in concert,
in part contributing to current lack of knowledge of their defining characteristics.
For many deposits or provinces, the lack of precise dating of the mineralization, using robust isotopic
systems within the context of the regional events, limits the development of viable genetic models.
Such dating should be a priority, particularly for the giant and world-class deposits. Particularly
useful methodologies include SHRIMP U-Pb dating of zircon, monazite, titanite, allanite, rutile,
apatite, and xenotime to bracket both intrusive phases and ore-related mineral assemblages; Re-Os
analysis of molybdenite, other sulfide minerals, and gold itself; Sm-Nd measurements of ore-related
tungstates, fluorides, and uranium-rich oxides; and Ar-Ar studies on hydrothermal micas and
amphiboles, particularly if their robustness in specific provinces can be confirmed using other
methodologies (Richards and Noble, 1998).
1.8.2 Research on fluid source
There is clearly a need to better constrain the source of ore fluids in gold systems in metamorphic
belts. This can be achieved in three ways : (1) using the geochemistry and stable isotopic
compositions of relevant fluid inclusions in well-constrained ore-related minerals, (2) using
Page 29
radiogenic and stable isotope compositions of ore-related minerals as tracers, and (3) applying better
models of the mechanisms and P-T conditions for release of fluid of appropriate compositions from
proposed source rocks or magmas. These approaches should be integrated.
In order to characterize ore fluids, it is now possible to carry out sophisticated analysis of fluid
inclusions, for example by multi-element laser-ablation ICP-MS to measure Au, Ag, As, Bi, Sb and
other trace element concentrations, gas and ion chromatographic analyses to constrain fluid
compositions and halogen element ratios, and Os isotope and noble gas analyses. However,
techniques should be applied only when the inclusion and the host minerals are unequivocally related
to the gold mineralization phase. It must also be recognized that fluid pathways may be extensive,
leading to the possibility of fluid reaction with a variety of wallrocks, and that fluid immiscibility is
common, making any analyzed fluid only representative of the fluid evolution at a particular moment
in a particular location in the hydrothermal system.
Although a wide range of isotopic tracers can be measured in both fluid inclusions and hydrothermal
minerals (e.g., Pb, Sr, Nd, O, H, C, N, S, B, Br, Cl), most of these will not be diagnostic of a fluid
and/or metal source because of fluid-rock interactions along extensive fluid pathways and/or post-
entrapment modification of the fluid inclusions. Jia and Kerrich (1999) introduced N isotope
signatures as superior fluid-source tracers, and Ridley and Diamond (2000) emphasized that only N,
Br and Cl isotopes may be truly representative of source, but that isotopic composition of potential
source rocks are poorly characterized. Further investigations of the isotopic compositions of these
elements in potential source rocks are needed to allow these systems to be used effectively.
Where granitoid magmas are inferred to be the source of ore fluids, melt inclusion (e.g., Thomas et
al., 2000), as well as fluid inclusion, studies on the proposed source intrusions are necessary to
determine whether or not they are compatible with ore fluids inferred on the basis of fluid inclusion
studies of the ore-associated minerals. High-precision, ultra-low background analyses of the
granitoids for ore elements (e.g., Au, Ag, As, Bi, Sb, Te, W) would also help determine if these
granitoids were anomalous in terms of metal contents and ratios and help document their behavior
during fractionation and volatile phase separation.
In many cases, models are unconstrained by experimental studies or theoretical calculations of fluid
release from source rocks. For example, the release of aqueous-carbonic fluids from granitic magmas
at different crustal levels needs to be quantitatively modelled. Similarly, scenarios where degassing
of subducted oceanic crust (e.g., gently-dipping subduction zones), or flow of the residue of
previously partially-melted subducted crust, can contribute volatiles into subsequently formed melts,
need to be established.
1.8.3 Research on fluid flow conduits
If deeply-sourced ore fluids are accepted as necessary for the genesis of orogenic gold deposits, then
the conduits for fluid flux must be understood. As many giant gold provinces are located near
crustal-scale deformation zones in metamorphic belts, there is a need for thorough and integrated
studies of the structural evolution and associated magmatic activity, if any, along the structures,
combined with integrated fluid-inclusion and isotope studies. The nature of the connectivity and fluid
flow between the crustal-scale and lower-order structures that host the gold deposits must be better
understood in terms of their respective structural evolution (e.g., Neumayr et al., 2000, Neumayr and
Hagemann, in press).
Although many orogenic gold deposits are situated in faults or shear zones, many others are hosted in
adjacent rock bodies as stockworks, vein arrays or disseminations. Following from the studies of
Sibson (1990), the precise mechanisms for fluid flux in this type of ore deposit need to be more
clearly understood. Similarly, whether the presence of ultra-effective cap-rocks or seals to the system
can cause local convective flow (e.g., Etheridge et al., 1983) is still uncertain, as most ore systems
Page 30
show evidence of changes in mineralogy, and therefore probably fluid composition, with time. Such
features are difficult to explain in a one-pass advecting fluid system.
1.8.4 Research on background gold concentrations
Although several authors (e.g., Kerrich, 1986; Phillips et al., 1987) have shown that even giant gold
deposits can form from leaching and metal extraction of realistic volumes of any crustal rock type,
other authors have suggested that specific rocks (e.g., iron formations, exhalative sedimentary rocks,
komatiites, subducted oceanic crust, some granitoids) may be particularly enriched sources (Keays
and Scott, 1976; Bierlein et al., 1998). However, there is a dearth of high-precision, ultra low-level
background analyses of Au and related elements in potential source rocks. Such background studies
need to be carefully planned to avoid measuring gold in dispersion haloes related to the ore deposits.
This is particularly important for proposed sources for intrusion-related deposits; that is, can we tell if
a pluton hosting such deposits is itself inherently gold-rich away from the ore bodies?
1.8.5 Transport and deposition of gold
Our understanding of gold solubility in thiosulfide and chloride complexes is now adequate for
modelling of transport and deposition of gold complexes of this type particularly below 300oC and
1kb (e.g., Seward, 1991). However, in some deposits, Au may show an equally strong or even
stronger correlation to As, Te, Sb or Bi. Initial research on the Bi-Au association (Douglas et al.,
2000) explains the strong Bi-Au correlation in some deposits, potentially those formed in
environments with low fluid:rock ratios. There is an equal need to investigate Au-As, Au-Sb and Au-
Te complexes as alternative transport mechanisms for Au in systems rich in these elements (e.g.
Wood and Samson, 1998).
Orogenic and intrusion-related gold deposits commonly show evidence for H2O-CO2 phase
separation, but the precise chemical changes that lead to precipitation of gold during this process are
poorly understood, with competing processes complicating a thorough understanding of many
hydrothermal systems (e.g. Mikucki, 1998). Establishing the influence of volatile immiscibility on
the deposition of Au and associated elements is a priority to better understand orogenic gold systems,
particularly in vein style ores. In the many cases where H2O-CO2 phase immiscibility does not
accompany gold precipitation, it is critical to examine in more detail how other mechanisms,
especially fluctuations in pressure, without H2O-CO2 phase separation, may lead to destabilization of
gold-transporting complexes. In fact, the gold solubility experiments of Loucks and Mavrogenes
(1999) show that at 400oC, a 2 kb pressure drop during hydrofracturing of metamorphic rocks would
cause desulfidation of a fluid and an associated 90% decrease in gold solubility. The role of fluid
mixing or back-mixing also needs to be carefully assessed.
The processes by which Au can be remobilized during metamorphism of pre-existing deposits and
overprinting hydrothermal events must also be determined. The scales at which such processes
operate require more detailed documentation.
1.8.6 Research on temporal distribution of gold deposits
More robust dating of problematic gold deposits and provinces in Brazil, Russia, China and some
parts of Africa is required to improve understanding of the distribution of gold deposits through time
defined by Goldfarb et al. (2001). At the same time, Precambrian tectonic processes and
environments, and supercontinent geometries, must be better studied, in order to interpret the
temporal distribution of gold deposits in terms of the evolution of the Earth.
Page 31
1.8.7 Development of four-dimensional models
What is ultimately required is to develop four-dimensional space-time models for gold deposits in
metamorphic belts. This can be aided by the rapid improvements in physical modeling packages (e.g.,
ELLIPSIS, Moresi et al., 2001) that can eventually be used to produce realistic three-dimensional
models of the complex fluid flow systems. When these are combined with chemical modelling
packages into integrated models, real progress will be made in understanding of the total systems, as
has been done for sedimentary basins
Seismic traverses have proved exceptionally useful in delineating crustal structure and assisting better
understanding of the crustal-scale architecture of gold mineralizing systems, both in the Abitibi belt
(e.g., Wyman et al. 1999; Hynes and Ludden, 2000) and Western Australia (Drummond and Goleby,
1993), as have maps of mantle thermal anomalies in younger belts (e.g., de Boorder et al., 1998).
Such data are vital if breakthroughs are to be made in this area of research.

and Genetic Models
1.9.1 Critical parameters for giant gold deposits
The desire for giant gold deposits of superior size and /or grade dominates exploration by major
mining and exploration companies at the beginning of the 21st Century. However, to date, there is no
clear identification of the key parameters or conjunction of parameters that dictate their formation. It
appears at the province to deposit scale that a conjunction of several factors, rather than one or two
dominant factors, control the siting of giant orogenic gold deposits (Phillips et al., 1996). A
promising approach is to define those factors that control those provinces hosting numerous world-
class to giant deposits, and then apply these and more local parameters at the goldfield or deposit
scale. A database of the geophysical signatures of giant gold deposits would also be most useful
where GIS-based prospectivity analysis can facilitate the definition of factors which, in conjunction,
control giant deposits. Better definition of these factors has obvious benefits in terms of the choice of
terranes to explore and assignment of priorities to exploration target areas within them
1.9.2 Better temporal and tectonic models
Better models for the temporal distribution of gold deposits, particularly giant examples, would allow
better definition of these terranes of most interest to explorationists. For exploration, defining the
distribution of gold ores in metamorphic belts on the reconstructed Rodinian supercontinent, defining
gold-favorable tracts on such a reconstruction, and then rotating these tracts on to present-day
coordinates may present a novel approach to assessing Precambrian gold-resource potential. When
combined with improved tectonic models, and better Precambrian supercontinent reconstructions,
they allow refined predictive capacity for ground selection on a global scale. The capacity to overlay
global geological, geophysical and geochemical databases on superior supercontinent reconstructions
can be a powerful regional exploration tool.
1.9.3 Improved genetic models for gold deposits in metamorphic

belts
Page 32
Clearly, improved genetic models provide better understanding of the controls on deposit styles,
which can, in turn, improve exploration models. In the case of orogenic gold deposits, knowledge of
the fluid flow systems is far more critical to concept-driven exploration than knowledge of the fluid
source, at least at the goldfield scale. Better definition of the fluid conduits would allow more
quantitative, coupled modelling of fluid flow and Au deposition, bringing a predictive capability to
concept-driven exploration. Furthermore, because orogenic gold deposits are typically formed late in
the structural evolution of host terranes (Groves et al., 2000), better definition of flow conduits
improves the capacity of computer-based techniques, such as stress mapping (e.g., Holyland and
Ojala, 1997), and GIS-based prospectivity mapping using fuzzy logic (e.g., D’Ercole et al., 2000;
Knox-Robinson, 2000) and artificial neural-network (e.g., Brown et al., 2000) techniques, because
the parameters selected to be tested depend on the quality of the models as well as the quality of input
data.
In the case of intrusion-related gold deposits, improved geologic and, in turn, genetic models will
highlight key controls for the location of the deposits that will assist area selection in exploration. In
addition, recognition of the specific type(s) of granitoid intrusions responsible for the formation of
these systems becomes critical in area selection in provinces containing these deposits, if a magmatic
origin is valid.
1.9.4 Improved understanding of deposit overprinting
As overprinted deposits may combine the metal budgets of two separate mineralization episodes of
different style, they may be superior exploration targets (e.g. Hollinger-McIntyre; Hemlo;
Boddington). Even where two or more orogenic gold episodes overprint each other, because of
reactivation of controlling structures and or change in physical-chemical parameters relating to the
earlier mineralization or alteration episode, superior ore bodies may be developed (e.g. Golden Mile,
Kalgoorlie). Remobilization alone, without any input of new metal during that latter hydrothermal
event, may also increase grade and or improve the metallurgical qualities of the ore. Thus,
overprinting and or remobilization are processes that need to be better recognized and understood in
order to improve target generation in gold exploration.
Page 33
1.10. Acknowledgements
The authors thank the convenors of the JANUS I and II symposia, Dick Grauch, Lew
Gustafson, Murray Hitzmann and Dick Hutchinson, for the invitation to originally prepare
and present this paper. DIG and RJG thank staff and students of the Centre for Global
Metallogeny at UWA for extensive discussions that have improved our understanding of
gold deposits in orogenic belts. We also acknowledge fruitful discussions on orogenic gold
deposits with Neil Phillips. FR thanks Barrick Gold Corporation for permission to be
involved in this paper. The paper has been significantly improve by the incisive and helpful
reviews by Frank Bierlein, Tony Christie, Warren Day, Terry Klein, and the Editor, Mark
Hannington.
1.11. References
Allen, R.L., Weihed, P., and Svensson, S.A., 1996. Setting of Zn-Cu-Au-Ag massive sulfide deposits
in the evolution facies architecture of a 1.9Ga marine volcanic arc, Skellefte district, Sweden :
ECONOMIC GEOLOGY , v.91, p.1022-1053.
Allibone, A.H., McCuaig, T.C., Harris, D., Etheridge, M., Munroe, S., Byrne, D., Amanor, J. and
Gyapmg, W., 2002. Structural controls on gold mineralization at the Ashanti deposit, Obuasi,
Ghana : SOCIETY OF ECONOMIC GEOLOGISTS SPECIAL PUBLICATION, 9, p.65-94.
Allibone, A.H., Windh, J., Etheridge, M.A., Burton, D., Anderson, G., Edwards, P.W., Miller, A.,
Graves, C., Fanning, C.M., and Wysoczanski, R., 1998, Timing relationships and structural
controls on the location of Au-Cu mineralization at the Boddington Gold Mine, Western
Australia: ECONOMIC GEOLOGY, v. 93, p. 245-270.
Baker, T., 2002, Emplacement depth and carbon dioxide-rich fluid inclusions in intrusion-related
gold deposits: ECONOMIC GEOLOGY, v.97, in press.
Baker, T., and Lang, J.R., 2001, Fluid inclusion characteristics of intrusion-related gold
mineralization, Tombstone-Tungsten magmatic belt, Yukon Territory, Canada: Mineralium
Deposita, v. 36, p. 563-582.
Bateman, R.J., Hagemann, S.G., McCuaig, T.C. and Swager, C.P., 2001. Protracted gold
mineralization throughout Archaean orogenesis in the Kalgoorlie Camp, Yilgarn Craton, Western
Australia : structural, mineralogical and geochemical evolution: Geological Survey of Western
Australia Record 2001/17, p.63-98.
Bergman Weihed, J., Bergström, U., Billström, K., and Weihed, P., 1996. Geology, tectonic setting,
and origin of the Paleoproterozoic Boliden Au-Cu-As deposit , Skellefte district, northern
Sweden: ECONOMIC GEOLOGY, v.91, p.1073-1097.
Bierlein, F.P., Arne, D.C., Broome, J.M.N., and Ramsay, W.R.H., 1998, Metatholeiites and interflow
sediments from the Cambrian Heathcote greenstone belt, Australia: Source for gold mineralization
in Victoria?: ECONOMIC GEOLOGY, v. 93, p.84-101.
Bierlein, F.P., Arne, D.C., Foster, D.A., and Reynolds, P., 2001a, A geochronological framework for
orogenic gold mineralisation in central Victoria, Australia: Mineralium Deposita, v. 36, p. 741-
767.
Bierlein, F.B., Arne, D.C., Keay, M. and McNaughton, N.J., 2001b. Timing relationships between
felsic magmatism and mineralization in the central Victorian gold province, Southeast Australia:
Australian Journal of Earth Sciences, v.48, p. 883-899.
Bierlein, F.P., Cartwright, I., and McKnight, S., 2001, The role of carbonaceous “indicator” slates in
the genesis of lode gold mineralization in the western Lachlan orogen, Victoria, southeastern
Australia: ECONOMIC GEOLOGY, v. 96, p. 431-451.
Page 34
Bierlein, F.P. and Crowe, D.E., 2000. Phanerozoic orogenic lode gold deposits: REVIEWS IN
ECONOMIC GEOLOGY, v.13, p.103-139.
Bohlke, J.K., 1982, Orogenic (metamorphic-hosted) gold-quartz veins: United States Geological
Survey Open File Report, no. 795, p. 70-76.
Bohlke, K.K., 1988, Carbonate-sulfide equilibria and “stratbound” disseminated epigenetic gold
mineralization—a proposal based on examples from Allegheny, California, U.S.A.: Applied
Geochemistry, v. 3, p. 499-516.
Bohlke, J.K., and Kistler, R.W., 1986, Rb-Sr, K-Ar, and stable isotope evidence for ages and sources
of fluid components of gold-bearing quartz veins in the northern Sierra Nevada foothills
metamorphic belt, California: ECONOMIC GEOLOGY, v.81, p.296-322.
Bouchot, V., and Moritz, R., eds, 2000, A Geode-Geofrance 3D Workshop on Orogenic Gold
Deposits in Europe with Emphasis on the Variscides: Documents du BRGM 297, 118 pp.
Boyle, G.O., 1990, The Hillgrove antimony-gold deposit: Australia Institute of Mining and
Metallurgy, Monograph, no. 14(2), p1425-1427.
Brisbin, D.I., 2000, World class intrusion-related Archean vein gold deposits of the Porcupine gold
camp, Timmins, Ontario, in Cluer, J.K., Price, J.G., Struhsacker, E.M., Hardyman R.F., and
Morrise, C.L., eds, Geology and Ore Deposits 2000: The Great Basin and Beyond, The
Geological Society of Nevada, Reno, p. 19-35.
Brown, W.M., Gedeon, T.D., Groves, D.I., and Barnes, R.G., 2000, Artificial neural networks: a new
method for mineral prospectivity mapping: Australian Journal of Earth Sciences, v. 47, p. 757-
770.
Burk, R., Hodgson, C.J., and Quartermain, R.A., 1986, The geological setting of the Teck-Corona
Au-Mo-Ba deposit, Hemlo, Ontario, Canada, in Macdonald, A.J., ed., Proceedings of Gold ’86, an
International Symposium on the Geology of Gold Deposits: Toronto, p. 311-326.
Burnham, C.W., 1979, Magmas and hydrothermal fluids, in Barnes, H.L., ed., Geochemistry of
Hydrothermal Ore Deposits, 2nd Edition: New York, Wiley-Interscience, p.71-136.
Burrows, D.R., and Spooner, E.T.C., 1986, The McIntyre Cu-Au deposit, Timmins, Ontario, Canada,
in Macdonald, A.J., ed., Proceedings of Gold ’86, an International Symposium on the Geology of
Gold Deposits: Toronto, p. 23-39.
Burrows, D.R., Spooner, E.T.C., Wood, P.C., and Jemielita, R.A., 1993, Structural controls on
formation of the Hollinger-McIntyre Au quartz vein system in the Hollinger shear zone, Timmins,
southern Abitibi greenstone belt, Ontario: ECONOMIC GEOLOGY, v. 88, p. 1643-1663.
Caddey, S.W., Bachman, R.L, Campbell, T.J., Reid, R.R., and Otto, R.P., 1991, The Homestake gold
mine, an Early Proterozoic iron-formation-hosted gold deposit, Lawrence County South Dakota:
United States Geological Survey Bulletin, no. 1857-J, 67pp.
Cameron, E.M., and Hattori, K., 1985, The Hemlo gold deposit, Ontario: a geochemical and isotopic
study: Geochimica et Cosmochimica Acta., v. 49, p. 2041-2050.
Candela, P.A., 1997, A review of shallow, ore-related granites: Textures, volatiles, and ore metals:
Journal of Petrology, v. 38, p. 1619-1633.
Chauvet, A., Piatone, P. Barbanson, L., Nehlig, P., and Pedroletti, I., 2001, Gold deposit formation
during collapse tectonics: Structural, mineralogical, geochronological, and fluid inclusion
constraints in the Ouro Preto gold mines, Quadrilatero Ferrifero, Brazil: ECONOMIC
GEOLOGY, v. 96, p. 25-48.
Cliff, D.C., and Moravek, P., 1995, The Mokrsko gold deposit, Central Bohemia, Czech Republic, in
Pasava, J., Kribek, B., and Zak, K., eds., Mineral Deposits: From Their Origin to Their
Environmental Impacts: Balkema, Rotterdam, p. 105-108.
Cline, J.S., and Bodnar, R.J., 1991, Can economic porphyry copper mineralization be generated by a
typical calc-alkaline melt? Journal of Geophysical Research, v. 96, p. 8113-8126.
Clout, J.M.F., Cleghorn, J.H., and Eaton, P.C., 1990, Geology of the Kalgoorlie gold field, in
Hughes, F.E., ed., Geology of the Mineral Deposits of Australia and Papua New Guinea:
Australasian Institute of Mining and Metallurgy, Monograph 14, p. 411-431.
Cole, A., Wilkinson, J.J., Halls, C., and Serenko, T.J., 2000, Geologic characteristics, tectonic setting
and preliminary interpretations of the Jilau gold-quartz vein deposit, Tajikistan: Mineralium
Deposita, v. 35, p. 600-618.
Page 35
Condie, K.C., 1998, Episodic continental growth and supercontinents: A mantle avalanche
connection? Earth and Planetary Science Letters, v. 163, p. 97-108.
Corfu, F., and Muir, T.L., 1989a, The Hemlo-Heron Bay greenstone belt and Hemlo Au-Mo deposit,
Superior Province, Ontario, Canada: 1. Sequence of igneous activity determined by zircon U-Pb
geochronology: Chemical Geology, v. 79, p. 183-200.
Corfu, F., and Muir, T.L., 1989b, The Hemlo-Heron Bay greenstone belt and Hemlo Au-Mo deposit,
Superior Province, Ontario, Canada: 2. Timing of metamorphism, alteration and Au
mineralization from titanite, rutile, and monazite U-Pb geochronology: Chemical Geology, v. 79,
p. 201-223.
Couture, J.F., Pilote, P., Machado, N., Desrochers, J.P., 1994, Timing of gold mineralization in the
Val-d’Or District, southern Abitibi Belt; evidence for two distinct mineralizing events:
ECONOMIC GEOLOGY, v. 89, p. 1542-1551.
Cox, S.F., 1999, Deformationed continents in the dynamics of fluid flow in mesothermal gold
systems, in McCaffrey, K.J.W., Lonergan, L. and Wilkinson, J.J., eds, Fractures, Fluid Flow and
Mineralization: Geological Society Special Publication, v. 155, p. 123-140.
Cox, S.F., Knackstedt, M.A., and Braun, J., 2001, Principles of structural control on permeability and
fluid flow in hydrothermal systems: REVIEWS IN ECONOMIC GEOLOGY, v. 14, p. 1-14
Cox, S.F., Sun, S.S., Etheridge, M.A., Wall, V.J., and Potter, T.F., 1995, Structural and geochemical
controls on the development of turbidite-hosted gold quartz vein deposits, Wattle Gully Mine,
central Victoria, Australia: ECONOMIC GEOLOGY, v. 90, p. 1722-1746.
Cox, S.F., Wall, V.J., Etheridge, M.A., and Potter, T.F., 1991, Deformational and metamorphic
processes in the formation of mesothermal vein-hosted gold deposits - examples from the Lachlan
fold belt in central Victoria, Australia: Ore Geology Reviews, v. 6, p. 391-423.
Craw, D., Hall, A.J., Fallick, A.E., and Boyce, A.J., 1995, Sulphur isotopes in a metamorphic gold
deposit, Macraes Mine, Otago Schist, New Zealand: New Zealand Journal of Geology and
Geophysics, v. 38, p.131-136.
Craw, D., and Koons, P.O., 1989. Tectonically induced hydrothermal activity and gold mineralization
adjacent to major fault zones. ECONOMIC GEOLOGY MONOGRAPH v. 6, p. 463-470.
Craw, D., and Leckie, D.A., 1996, Tectonic controls on dispersal of gold into a foreland basin: an
example from the Western Canadian Basin: Journal of Sedimentary Research, v. 66, p. 559-566.
Crowe, D.E., 1995, An introduction to the Carolina slate belt: Society of Economic Geologists
Guidebook Series, v. 24, p. 1-20.
Dalziel, I.W.D., Lawyer, L.A., and Murphy, J.B., 2000, Plumes, orogenesis, and supercontinental
fragmentation: Earth and Planetary Science Letters, v. 178, p. 1-11.
Davies, G.F., 1995, Penetration of plates and plumes through the mantle transition zone: Earth and
Planetary Science Letters, v. 133, p. 507-516.
de Boorder, H., Sparkman, W., While, S.H., and Wortel, M.J.R., 1998, Late Cenozoic mineralization,
orogenic collapse and slab detachment in the European Alpine belt: Earth and Planetary Science
Letters, v. 164, p. 569-575.
de Ronde, C.E.J., and de Wit, M.J., 1994. The tectonothermal evolution of the Barberton Greenstone
Belt, South Africa: 490 million years of crustal evolution: Tectonics, v. 13, p. 983-1005.
de Ronde, C.E.J., Faure, K., Bray, C.J., and Whitford, D.J., 2000. Round Hill shear zone-hosted gold
deposit, Macraes Flat, Otago, New Zealand: Evidence of a magmatic ore fluid: ECONOMIC
GEOLOGY, v. 95, p. 1025-1048.
D’Ercole, C., Groves, D.I., and Knox-Robinson, C.M., 2000, Using fuzzy logic in a Geographic
Information System environment to enhance conceptually based prospectivity analysis of
Mississippi Valley-type mineralization: Australian Journal of Earth Sciences, v. 47, p. 913-928.
Douglas, N., Mavrogenes, J., Hack, A., and England, R., 2000, The liquid bismuth collector model:
an alternative gold depositional mechanism [abs]: Geological Society of Australia, Abstract no.
59, p. 135.
Drummond, B.J., and Goleby, B.R., 1993, Seismic reflection images of the major ore-controlling
structures in the Eastern Goldfields Province, Western Australia: Exploration Geophysics, v. 24,
p. 473-478.
Page 36
Dubé, B., Williamson K., and Malo, M., 2002. Geology of the Goldcorp Inc, High-Grade zone, Red
Lake mine, Ontario: an update : Geological Survey of Canada, Current Research, 2002-C26,
13pp.
Duuring, P., Hagemann, S.G., and Groves, D.I., 2000, Structural setting, hydrothermal alteration, and
gold mineralization at the Archean syenite-hosted Jupiter deposit, Yilgarn Craton, Western
Australia: Mineralium Deposita, v. 35, p. 402-421.
Ebert, S., Dodd, S., Miller, L., and Petsel, S., 2000, The Donlin Creek Au-As-Sb-Hg deposit,
southwestern Alaska, in Cluer, J.K., Price, J.G., Struhsacker, E.M., Hardyman, R.F., and Morris,
C.L., eds. Geology and Ore Deposits 2000: The Great Basin and Beyond: Geological Society of
Nevada Symposium Proceedings, p. 1069-1081.
Eggler, D.H., and Kadik, A.A., 1979, The system NaAlSi3O8 – H2O – CO2 to 20 kbar pressure;
Compositional and thermodynamic relations of liquids and vapors coexisting with albite:
American Mineralogist, v. 64, p. 1036-1048.
Etheridge, M.A., Wall, V.J., and Vernon, R.H., 1983, The role of the fluid phase during regional
metamorphism and deformation: Journal of Metamorphic Geology, v. 1, p. 205-226.
Flanigan, B., Freeman, C., McCoy, D., Newberry, R., and Hart, C., 2000, Paleo-reconstruction of the
Tintina gold belt – implications for mineral exploration: in The Tintina Gold Belt: Concepts,
Exploration and Discoveries: British Columbia and Yukon Chamber of Mines, Cordilleran
Roundup, January 2000, p. 35-48.
Foster, D.A., Gray, D.R. Kwak, T.A.P., and Bucher, M., 1998. Chronological and orogenic
framework of turbidite hosted gold deposits, in the Western Lachlan Fold Belt, Victoria: 40Ar/39Ar
results : Ore Geology Reviews, v.13, p. 229-250.
Fyfe, W.S., and Kerrich, R., 1984. Gold: Natural concentration processes, in Foster, R.P., ed.,
Gold’82: The Geology, Geochemistry and Genesis of Gold Deposits: A.A. Balkema, Rotterdam,
pp. 99-127.
Fyfe, W.S., and Kerrich, R., 1985, Fluids and thrusting: Chemical Geology, v. 49, p. 353-362.
Gebre-Mariam, M., Hagemann, S.G., and Groves, D.I., 1995, A classification scheme for epigenetic
Archean lode-gold deposits: Mineralium Deposita, v. 30, p. 408-410.
Genkin, A.D., Wagner, F.E., Krylova, T.L., and Tsepin, A.I., 2002, Gold-bearing arsenopyrite and its
formation condition at the Olympiada and Veduga gold deposits (Yenisei Range, Siberia):
Geology of Ore Deposits, v. 44, p. 52-68.
Goellnicht, N.M., Groves, D.I., and McNaughton, N.J., 1991, Late Proterozoic fractionated granitoids
of the mineralized Telfer area, Paterson Province, Western Australia: Precambrian Research, v.
51, p. 375-391
Goellnicht, N.M., Groves, D.I., McNaughton, N.J., and Dimo, G., 1989, An epigenetic origin for the
Telfer gold deposit: ECONOMIC GEOLOGY MONOGRAPH, v. 6, p. 151-167.
Goldfarb, R.J., Groves, D.I., and Gardoll, S., 2001, Orogenic gold and geologic time: a global
synthesis: Ore Geology Reviews, v. 18, p. 1-75.
Goldfarb, R.J., Hart, C.J.R., Miller, M.L., Miller, L.D., Farmer, G.L., and Groves, D.I., 2000, The
Tintina gold belt – a global perspective, in The Tintina Gold Belt: Concepts, Exploration and
Discoveries: British Columbia and Yukon Chamber of Mines, Cordilleran Roundup, January
2000, p. 5-34.
Goldfarb, R.J., Hofstra, A.H., Landis, G.P., and Leach, D.L., 1988, H2S-rich vein forming fluids at
the Sumdum Chief gold mine, southeastern Alaska, in Galloway, J.P., and Hamilton, T.D., eds,
Geologic studies in Alaska by the USGS during 1987: United States Geological Survey Circular,
p. 160-163.
Goldfarb, R.J., Leach, D.L., Rose, S.C., and Landis, G.P., 1989, Fluid inclusion geochemistry of
gold-bearing quartz veins of the Juneau gold belt, southeastern Alaska: Implications for ore
genesis: ECONOMIC GEOLOGY MONOGRAPH, v. 6, p. 363-375.
Goldfarb, R.J., Miller, L.D., Leach, D.L., and Snee, L.W., 1997, Gold deposits in metamorphic rocks
in Alaska: ECONOMIC GEOLOGY MONOGRAPH, v. 9, p. 151-190.
Goldfarb, R.J., Snee, L.W., Miller, L.D., and Newberry, R.J., 1991, Rapid dewatering of the crust
deduced from ages of mesothermal gold deposits: Nature, v. 354, p. 296-298.
Graupner, T., Kempe, U., Spooner, E.T.C., Bray, C.J., Kremenetsky, A.A., and Irmer, G., 2001,
Microthermometric, Laser Raman spectroscopic, and volatile-ion chromatographic analysis of
Page 37
hydrothermal fluids in the Paleozoic Muruntau Au-bearing quartz vein ore field, Uzbekistan:
Gray, D.R., 1997, Tectonics of the southeastern Australia Lachlan fold belt - Structural and thermal
aspects, in Burg J.P., and Ford M., eds, Orogeny Through Time: Geological Society of London,
Special Publication 121, p. 149-177.
Gray, M.D. and Hutchinson, R.W., 2001, New evidence for multiple periods of gold emplacement in
the Porcupine mining district, Timmins area, Ontario, Canada: ECONOMIC GEOLOGY, v. 96,
p. 453-475.
Groves, D.I., and Barley, M.E., 1994, Archean mineralization, in Condie, K.C., ed., Archean Crustal
Evolution: Elsevier, Amsterdam, p. 461-503.
Groves, D.I., Goldfarb, R.J., Gebre-Mariam, M., Hagemann, S.G., and Robert, F., 1998, Orogenic
gold deposits: A proposed classification in the context of their crustal distribution and relationship
to other gold deposit types: Ore Geology Reviews, v. 13, p. 7-27.
Groves, D.I., Goldfarb, R.J., Knox-Robinson, C.M., Ojala, J., Gardoll, S., Yun, G., and Holyland, P.,
2000. Late-kinematic timing of orogenic gold deposits and significance for computer-based
exploration techniques with emphasis on the Yilgarn block, Western Australia: Ore Geology
Reviews, v. 17, p. 1-38.
Groves, D.I., Phillips, G.N., Ho, S.E., Houstoun, S.M., and Standing, C.A., 1988, Craton-scale
distribution of greenstone gold deposits: predictive capacity of the metamorphic model:
ECONOMIC GEOLOGY, v.83, p. 2045-2058.
Haeberlin, Y., 2002, Geological and structural setting, age, and geochemistry of the orogenic gold
deposits at the Pataz province, Eastern Andean Cordillera, Peru: Universite de Geneve, Terre &
Environment, v. 36, 182 p.
Haeussler, P.J., Bradley, D., Goldfarb, R.J., Snee, L.W., and Taylor, C.D., 1995. Link between ridge
subduction and gold mineralization in southern Alaska: Geology, v. 23, p. 995-998.
Hagemann, S.G., and Brown, P.E., eds, 2000, Gold in 2000: REVIEWS IN ECONOMIC
GEOLOGY, v. 13, 559 pp.
Hagemann S.G., and Cassidy, K.F., 2000, Archean orogenic lode gold deposits: REVIEWS IN
Hagemann, S.G., Gebre-Mariam, M., and Groves, D.I., 1994, Surface-water influx in shallow-level
Archean lode-gold deposits in Western Australia: Geology, v. 22, p. 1067-1070.
Halley, S.W., and Roberts, R.H., 1997, Henty: A shallow-water gold-rich volcanogenic massive
sulfide deposit in western Tasmania: ECONOMIC GEOLOGY, v. 92, p. 438-447.
Hannington, M.D., Poulsen, K.H., Thompson, J.F.H., and Sillitoe, R.H., 1999. Volcanogenic gold in
the massive sulfide environment. REVIEWS IN ECONOMIC GEOLOGY, v.8, p.325-356.
Harris, D.C., 1989, The mineralogy and geochemistry of the Hemlo gold deposits: Geology, v. 15, p.
1107-1111.
Hart, C.J.R., Baker, T., and Burke, M., 2000, New exploration concepts for country-rock-hosted,
intrusion-related gold systems: Tintina gold belt in Yukon: in The Tintina Gold Belt: Concepts,
Exploration and Discoveries, British Columbia and Yukon Chamber of Mines, Cordilleran
Roundup, January 2000, p. 145-172.
Hart, C.J.R., McCoy, D.T., Goldfarb, R.J., Smith, M., Roberts, P., Hulstein, R., Bakke, A.A., and
Bundtzen, T.K., 2002, Geology, exploration and discovery in the Tintina gold province, Alaska
and Yukon: ECONOMIC GEOLOGY SPECIAL PUBLICATION, v. 9, in press.
Hart, C.J.R., Selby, D., and Creaser, R.A., 2001, Timing relationships between plutonism and gold
mineralization in the Tintina gold belt (Yukon and Alaska) using Re-Os molybdenite dating, in
2001 - A Hydrothermal Odyssey, New Developments in Metalliferous Hydrothermal Systems
Research, Program with Abstracts: James Cook University Townsville, Queensland, p. 71-72.
Hayward, N., 1992, Controls on syntectonic replacement mineralization in parasitic antiforms, Haile
gold mine, Carolina slate belt, U.S.A.: ECONOMIC GEOLOGY, v. 87, p. 91-112.
Hedenquist, J.W., and Lowenstern, J.B., 1994, The role of magmas in the formation of hydrothermal
ore deposits: Nature, v. 370, p. 519-527.
Henley, R.W., Norris, R.J., and Patterson, C.J. 1976. Multistage ore genesis in the New Zealand
geosyncline : a history of post-metamorphic lode emplacement: Mineralium Deposita, v.11, p.
180-196.
Page 38
Herzig, P.M., Petersen, S. and Hannington, M.D., 1999, Epithermal-type gold mineralization at
Conical Seamount: a shallow submarine volcano sout of Lihir Island, Papua New Guinea. In:
Stanley, C. et al. eds, Mineral Depoists: Processes to Processing, Balkema, Rotterdam, p. 527-530.
Hodgson, C.J. 1989, The structure of shear-related, vein-type gold deposits: a review: Ore Geology
Reviews, v. 4, p. 231-273.
Hodgson, C.J. 1993. Mesothermal lode-gold deposits, in Kirkham, R.V., Sinclair, W.D., Thorpe R.I.,
and Duke J.M., eds, Mineral Deposit Modeling: Geological Association of Canada Special Paper
40, pp. 635-678.
Hodgson, C.J., Love, D.A., and Hamilton, J.V., 1993, Giant mesothermal gold deposits, Descriptive
characteristics, genetic model and exploration area selection criteria, in Whiting, B.H., Hodgson,
C.J., and Mason R., eds., Giant Ore Deposits: ECONOMIC GEOLOGY SPECIAL
PUBLICATION, v. 2, p. 157-211.
Hollister, V.F., 1978, Geology of the porphyry copper deposits of the western hemisphere: Society of
Mining Engineers AIME, New York, 219 p.
Holyland, P.W., and Ojala, V.J., 1997, Computer aided structural targeting in mineral exploration:
two and three-dimensional stress mapping: Australian Journal of Earth Sciences, v. 44, p. 421-
432.
Hugon, H., 1986, The Hemlo deposits, Ontario, Canada: a central portion of a large scale, wide zone
of heterogeneous ductile shear, in Macdonald, A.J., ed., Proceedings of Gold ’86, an International
Symposium on the Geology of Gold Deposits: Toronto, p. 379-387.
Huston, D.L., 2001, Geology, alteration assemblages and geochemical dispersion about the Western
Tharsis deposit--Implications for the genesis of the Mt. Lyell Cu-Au district, western Tasmania:
2001: A Hydrothermal Odyssey, EGRU Contribution 59, p. 94.
Hynes, A., and Ludden, J.N., eds, 2000, The Lithoprobe Abitibi-Grenville transect: Canadian Journal
of Earth Sciences, v. 37, p. 115-516.
Jamtveit, B., and Yardley, B.W.D., eds, 1997, Fluid Flow and Transport in Rocks: Chapman & Hall,
London, 319 pp.
Jenchuraeva, R.J., Nikonorov, V.K., and Litvinov, P., 2001, The Kumtor gold deposit, in Seltmann,
R., and Jenchuraeva, R., eds., Paleozoic Geodynamics and Gold Deposits in the Kyrgyz Tien
Shan: IGCP-373 Field Conference Excursion Guidebook, p. 139-152.
Jia, Y. and Kerrich, R., 1999, Nitrogen isotope systematics of mesothermal lode gold deposits:
Metamorphic, granitic, meteoric water, or mantle origin: Geology, v. 27, p. 1051-1054.
Kay, S.M., and Mpodozis, C., 2001, Central Andean ore deposits linked to evolving shallow
subduction systems and thickening crust: Geological Society of America Today, March 2001, p.
4-9.
Keays, R.R., and Scott, R.B., 1976, Precious metals in ocean ridge basalts: Implications for basalts as
source rocks for gold mineralization: ECONOMIC GEOLOGY, v. 71, p. 705-720.
Kempe, U., Belyatsky, B.V., Krymsky, R.S., Kremenetsky, A.A., and Ivanov, P.A., 2001, Sm-Nd
and Sr isotope systematics of scheelite from the giant Au(-W) deposit Muruntau (Uzbekistan)—
implications for the age and sources of Au mineralization: Mineralium Deposita, v. 36, p. 379-
392.
Kerrich, R., 1986, Archean lode gold deposits of Canada; Part II, Characteristics of the hydrothermal
systems, and model of origin: Information Circular, University of the Witwatersrand, Economic
Geology Research Unit, v. 183, pp. 34.
Kerrich, R., and Cassidy, K.F., 1994, Temporal relationships of lode-gold mineralization to accretion,
magmatism, metamorphism and deformation - Archean to present: a review: Ore Geology
Reviews, v. 9, p. 263-310.
Kerrich, R., Goldfarb R.J., Groves D.I., and Garwin, S., 2000, The geodynamics of world class gold
deposits: characteristics, space-time distribution, and origins: REVIEWS IN ECONOMIC
GEOLOGY, v. 13, p. 501-551.
Kishida, A., and Kerrich, R., 1987, Hydrothermal alteration zoning and gold concentration at the
Kerr-Addison Archean lode gold deposit, Kirkland Lake, Ontario: ECONOMIC GEOLOGY, v.
82, p. 649-660.
Page 39
Kisters, A.F.M., Meyer,F.M., Seravkin, I.B., Znamensky, S.E., Kosarev, A.M., and Ertl, R.G.W.,
1999, The geologic setting of lode-gold deposits in central southern Urals: a review: Geologische
Rundschau, v. 87, p.603-616.
Klominsky, J., Partington, G.A., McNaughton, N.J., Ho, S.E., and Groves, D.I., 1996, Radiothermal
granites of the Cullen Batholith and associated mineralization (Australia): Czech Geological
Survey, Special Papers no.5, 44p.
Knox-Robinson, C.M., 2000, Vectorial fuzzy logic: a novel technique for enhanced mineral
prospectivity mapping, with references to the orogenic gold mineralisation potential of the
Kalgoorlie Terrane, Western Australia: Australian Journal of Earth Sciences, v. 7, p. 929-442.
Kuhns, R.J., Kennedy, P., Cooper, P., Brown, P., Mackie, B., Kusins, R., and Friesen R., 1986,
Geology and mineralization associated with Golden Giant deposit, Hemlo, Ontario, Canada: in
Macdonald, A.J. ed., Proceedings of Gold ’86, an International Symposium on the Geology of
Gold Deposits: Toronto, p. 327-756.
Lang, J.R., and Baker, T., 2001, Intrusion-related gold systems—The present level of understanding:
Mineralium Deposita, v. 36, p. 477-489.
Lang, J.R., Baker, T., Hart, C.J.R., and Mortensen, J.K., 2000, An exploration model for intrusion-
related gold systems; Society of Economic Geologists Newsletter no. 40, p.1-15.
Laverov, N.P., Lishnevskii, E.N., Distler, V.V., and Chernov, A.A., 2000, Model of the ore-
magmatic system of the Sukhoi Log gold-platinum deposit, eastern Siberia Russia: Doklady Earth
Sciences, v. 375A, p. 1362-1365.
Laznicka, P., 1999. Quantitative relationships among giant deposits of metals: ECONOMIC
GEOLOGY, v. 94, p. 455-473.
Lin, S., 2001. Stratigraphic and structural setting of the Hemlo gold deposit, Ontario, Canada:
Lobato, L.M., Ribeiro-Rodrigues, L.C., and Vireira, F.W.R., 2001, Brazil’s premier gold province.
Part II : geology and genesis of gold deposits in the Archean Riodas Velhas greenstone belt,
Quadrilatero Ferrifero, Mineralium Deposita, v.36, p. 249-277.
Lonergan, L., Wilkinson, J.J., and McCaffrey, K.J.W., 1999, Fractures, fluid flow and
mineralization—an introduction, in McCaffrey, K.J.W., Lonergan, L. and Wilkinson, J.J., eds,
Fractures, Fluid Flow and Mineralization: Geological Society Special Publication, v. 155, p. 1-6.
Loucks. R.R., and Mavrogenes, J.A., 199, Gold solubility in supercritical hydrothermal brines
measured in synthetic fluid inclusions: Science, v. 284, p. 2159-2163.
Lowenstern, J.B., 2001, Carbon dioxide in magmas and implications for hydrothermal systems:
Madden-McGuire, D.J., Silberman, M.L., and Church, S.E., 1989, Geologic relationships, K-Ar ages,
and isotopic data from the Willow Creek gold mining district, southern Alaska: ECONOMIC
GEOLOGY MONOGRAPH, v. 6, p. 242-251.
Marignac, C., and Cuney, M., 1999, Ore deposits of the French Massif Central--insight into the
metallogenesis of the Variscan Collision belt : Mineralium Deposita, v.34, p. 472-504.
Marmont, S., and Corfu, F., 1989, Timing of gold introduction in the late Archean tectonic
framework of the Canadian Shield; evidence from U-Pb zircon geochronology of the Abitibi
Subprovince: ECONOMIC GEOLOGY MONOGRAPH, v. 6, p. 101-111.
Marquis, P., 1990, Metallogenie des gisements archeen d’Au-Ag-Cu de la mine Donald J. LaRonde
(Dumagami), Cadillac, Abitibi, Quebec: Unpub. Ph.D. thesis, University of Montreal, 139p.
Marquis, P., Hubert, C., Brown, A.C., and Rigg, D.M., 1990, Overprinting of early, redistributed Fe
and Pb-Zn mineralization by late-stage Au-Ag-Cu deposition at the Dumagami mine, Bousquet
district, Abitibi, Quebec: Canadian Journal of Earth Sciences, v. 27, p. 1651-1671.
Marsh, E., Goldfarb, R., Hart, C., and Johnson, C., 2003, Geochemistry of the auriferous sheeted
quartz veins of the Clear Creek intrusion-related gold deposit, Tintina gold belt, Yukon, Canada:
Canadian Journal of Earth Sciences, submitted and being revised.
Matthai, S.K., Henley, R.W., Bacigalupo-Rose, S., Binns, R.A., Andrew, A.S., Carr, G.R., French,
D.H., McAndrew, J., and Kavanagh, M.E., 1995, Intrusion-related, high-temperature gold quartz
veining in the Cosmopolitan Howley metasedimentary rock-hosted gold deposit, Northern
Territory, Australia: ECONOMIC GEOLOGY, v. 90, p. 1012-1045.
Page 40
McCoy, D., Newberry, R.J., Layer, P.W., DiMarchi, J.J., Bakke, A.A., Masterman, J.S., and
Minehane, D.L., 1997, Plutonic-related gold deposits of interior Alaska: ECONOMIC
GEOLOGY MONOGRAPH, v. 9, p. 191-241.
McCuaig, T.C., Behn, M., Stein, H., Hagemann, S.G., McNaughton, N.J., Cassidy, K.F., Champion,
D., and Wyborn, L., 2001, The Boddington gold mine: A new style of Archaean Au-Cu deposit,
in Cassidy, K.F., Dunphy, J.M., and VanKranendonk, M.J., eds., 4th International Archaean
Symposium, Extended Abstracts: AGSO--Geoscience Australia, Record 2001/37, p. 453-455.
McCuaig, T.C., and Kerrich, R., 1998, P-T-t-deformation-fluid characteristics of lode-gold deposits:
Evidence from alteration systematics: Ore Geology Reviews, v. 12, p. 381-453.
Meinert, L.D., 1993, Igneous petrogenesis and skarn deposits, in Kirkham, R.V., Sinclair, W.D.,
Thorpe, R.J., and Duke, J.M., eds., Mineral Deposit Modeling: Geological Association of Canada,
Special Paper no. 40, p. 569-583.
Michibayashi, K., 1995, Two phase syntectonic gold mineralization and barite remobilization within
the main ore body of the Golden Giant mine, Hemlo, Ontario, Canada: Ore Geology Reviews, v.
10, p. 31-50.
Mikucki, E.J., 1998, Hydrothermal transport and depositional processes in Archean lode-gold
systems: A review: Ore Geology Reviews, v. 13, p. 307-321.
Miller, L.D., Goldfarb, R.J., Gehrels, G.E., and Snee, L.W., 1994. Genetic links among fluid cycling,
vein formation, regional deformation, and plutonism in the Juneau gold belt, southeastern Alaska:
Geology, v. 22, p. 203-206.
Miller, L.D., Goldfarb, R.J., Snee, L.W., Gent, C.A., and Kirkham, R.A., 1995, Structural geology,
age, and mechanisms of gold vein formation at the Kensington and Jualin deposits, Berners Bay
district, southeast Alaska: ECONOMIC GEOLOGY, v. 90, p. 343-368.
Mitchell, A.H.G., and Garson, M.S., 1981, Mineral Deposits and Global Tectonic Settings: Academic
Press, 405p.
Moresi, L., Mühlhaus, H., and Dufour, F., 2001, Particle-in-cell solutions for creeping viscous flow
with internal surfaces : in Mühlhaus, H.B., Dynskin, A., and Pasternak, E., eds, Bifurcation and
Localization in Soils and Rocks, Balkema, Rolterdam, p. 345-354.
Moritz, R., 2000, What have we learnt about orogenic lode gold deposits over the past 20 years? in A
Geode-GeoFrance 3D Workshop on Orogenic Gold Deposits in Europe with Emphasis on the
Variscides, Extended Abstracts, November 7-8, 2000, Orleans (BRGM), France, p. 17-23.
Mueller, A.G., 1991, The Savage lode magnesian skarn in the Marvel Loch gold-silver mine,
Southern Cross greenstone belt, Western Australia: Part 1. Structural setting, petrography and
geochemistry: Canadian Journal of Earth Sciences, v. 28, p. 659-685.
Mueller, A.G., Campbell, I.H., Schiotte, L., Sevingny, J.H., and Layer, P.W., 1996, Constraints on
the age of granitoid emplacement, metamorphism, gold mineralization, and subsequent cooling of
the Archean greenstone terrane at Big Bell, Western Australia: ECONOMIC GEOLOGY, v. 91,
p. 896-915.
Mueller, A.G., and McNaughton, N.J., 2000, U-Pb ages constraining batholith emplacement, contact
metamorphism, and the formation of gold and W-Mo skarns in the Southern Cross area, Yilgarn
craton, Western Australia: ECONOMIC GEOLOGY, v. 95, p. 1231-1258.
Mueller A.G., Harris, L.B., and Lungan, A., 1988, Structural control of greenstone-hosted gold
mineralization by transcurrent shearing: a new interpretation of the Kalgoorlie mining district,
Western Australia: Ore Geology Reviews, v. 3, p. 359-387.
Muir, T.L., 2002, The Hemlo gold deposit, Ontario, Canada—principal deposit characteristics and
constraints on mineralization: Ore Geology Reviews, in press.
Mustard, R., 2001, Granite-hosted gold mineralization at Timbarra, northern New South Wales,
Australia: Mineralium Deposita, v. 36, p. 542-562.
Nabelek, P.I., and Ternes, K., 1996, Fluid inclusion evidence for the evolution and solubility of fluids
in the Harney Peak leucogranite magma, Black Hills, South Dakota, in Brown, P.E., and
Hagemann, S.G., eds, PACROFI VI, Program and Abstracts: Biennial Pan-American Conference
on Research on Fluid Inclusions, v. 6, p. 95-96
Nesbitt, B.E., Murrowchick, J.B., and Muehlenbachs, K., 1986, Dual origin of lode-gold deposits in
the Canadian Cordillera: Geology, v. 14, p. 506-509.
Page 41
Neumayr, P., and Hagemann, S.G., 2002, Hydrothermal fluid evolution within the Cadillac Tectonic
Zone, Abitibi greenstone belt, Canada : relationship to auriferous fluids in adjacent second- and
third-order shear zones : ECONOMIC GEOLOGY, v.97, p. 1203-1225.
Neumayr, P., Hagemann, S.G., and Couture, J.F., 2000, Structural setting, textures and timing of
hydrothermal vein systems in the Val-d’Or camp, Abitibi, Canada: Implications for the evolution
of transcrustal, second- and third-order fault zones and gold mineralization: Canadian Journal of
Earth Sciences, v. 37, p. 95-115.
Newberry, R.J., 2000, Mineral deposits and associated Mesozoic and Tertiary igneous rocks within
the interior Alaska and adjacent Yukon portions of the ‘Tintina gold belt’, in The Tintina Gold
Belt: Concepts, Exploration and Discoveries, British Columbia and Yukon Chamber of Mines,
Cordilleran Roundup, January 2000, p. 59-88.
Newberry, R.J., Crafford, T.C., Newkirk, S.R., Young, L.E., Nelson, S.W., and Duke, N.A., 1997,
Volcanogenic massive sulfide deposits of Alaska: ECONOMIC GEOLOGY MONOGRAPH, v.
9, p. 120-150.
Newberry, R.J., McCoy, D.I., and Brew, D.A., 1995, Plutonic-hosted gold ores in Alaska: Igneous
versus metamorphic origin: Resource Geology, Special Issue no 18, p. 57-100.
Ojala, J.V., Ridley, J.R., Groves, D.I., and Hall, G.C., 1993, The Granny Smith gold deposit: the role
of heterogeneous stress distribution at an irregular granitoid contact in a greenschist facies terrane:
Pan, Y., and Fleet, M.E., 1992, Calc-silicate alteration in the Hemlo gold deposit, Ontario: mineral
assemblages, P-T-X constraints and significance: ECONOMIC GEOLOGY, v. 87, p. 1104-1120.
Partington, G.A., and McNaughton, N.J., 1997, Controls on mineralization in the Howley District,
Northern Territory: a link between granite intrusion and gold mineralization: Chronique de la
Recherche Miniere, v. 529, p. 25-44
Paterson, C.J., 1977, A geochemical investigation of the origin of scheelite mineralization,
Glenorchy, New Zealand: University of Otago, Dunedin, South Island, New Zealand.
Phillips, G.N., and Groves, D.I., 1983, The nature of Archean gold-bearing fluids as deduced from
gold deposits of Western Australia: Geological Society of Australia, Journal v. 30, p. 25-39.
Phillips, G.N., and Groves D.I., 1984, Fluid access and fluid-wallrock interaction in the genesis of the
Archean gold-quartz vein deposit at Hunt Mine, Western Australia, in Foster, R.P., ed., Gold’82:
The Geology, Geochemistry and Genesis of Gold Deposits, A.A. Balkema Publishers, Rotterdam,
p. 389-416.
Phillips, G.N., Groves, D.I., and Brown, I.J., 1987, Source requirements for the Golden Mile,
Kalgoorlie: significance to the metamorphic replacement model for Archean gold deposits:
Canadian Journal of Earth Sciences, v. 24, p. 1643-1651.
Phillips, G.N., Groves, D.I., and Kerrich, R., 1996, Factors in the formation of the giant Kalgoorlie
gold deposit: Ore Geology Reviews, v.10, p.295-317.
Pilote, P., Robert, F., Sinclair, W.D., Kirkham, R.V., and Daigneault, R., 1995. Porphyry-type
mineralisation in the Doré Lake complex: Clark Lake and Merrill Island areas, Day 3 in Pilote, P.,
ed., Precambrian ’95 Metallogenic Evolution and Geology of the Cibougamau area – from
Porphyry Cu-Mo-Au to Mesothermal Gold Deposits: Geological Survey of Canada, Open File
Report 3143. P. 65-86.
Poulsen, K.H., and Hannington, M.D., 1996, Volcanic-associated massive sulfide gold, in Eckstrand,
R.O., Sinclair, W.D., and Thorpe, R.I., eds, Geology of Canadian Mineral Deposit Types:
Geological Survey of Canada, no. 8, p. 183-196.
Poulsen, K.H., Mortensen, J.K., and Murphy, D.C., 1997, Styles of intrusion-related mineralization in
the Dawson-Mayo area, Yukon Territory, in Geological Survey of Canada Current research Paper
1997-A, p. 1-10.
Poulsen, K.H., Robert, F., and Dube, B., 2000, Geological classification of Canadian gold deposits:
Geological Survey of Canada, Bulletin 540, 106p.
Poulsen, K.H., Taylor, B.E., Robert, F., and Mortensen, J.K.. 1990, Observations on gold deposits in
North China Platform, in Current Research, Part A, Geological Survey of Canada, Paper 90-1A, p.
33-44.
Page 42
Powell, R., Will, T.M., and Phillips, G.N., 1991, Metamorphism in Archean greenstone belts:
Calculated fluid compositions and implications for gold mineralization: Journal of Metamorphic
Geology, v. 9, p. 141-150.
Powell, W.G., and Pattison, D.R.M., 1997, An exsolution origin for low-temperature sulfides at the
Hemlo gold deposit, Ontario, Canada: ECONOMIC GEOLOGY, v. 92, p. 569-577.
Qiu, Y., Groves, D.I., McNaughton, N.J., Wang, L-g, and Zhou,T., 2002, Nature, age and tectonic
setting of granitoid-hosted orogenic gold deposits at the Jiaodong Peninsular, eastern North China
cranton, China : Mineralium Deposita, v.37, p.283-305.
Radhakrishna, B.P., and Curtis, L.C., 1999. Gold in India. Geological Society of India, Bangalore,
India, 307 p.
Richards, J.P., and Noble, S.R., 1998, Application of radiogenic isotope systems to the timing and
origin of hydrothermal processes: REVIEWS IN ECONOMIC GEOLOGY, v. 10, p. 195-233.
Ridley, J.R., and Diamond, L.W., 2000, Fluid chemistry of orogenic lode-gold deposits and
implications for genetic models: REVIEWS IN ECONOMIC GEOLOGY, v. 13, p. 141-162.
Ridley, J.R., Groves, D.I., and Knight, J.T., 2000, Gold deposits in amphibolite and granulite facies
terranes of the Archean Yilgarn Craton, Western Australia: Evidence and implications for
synmetamorphic mineralization: REVIEWS IN ECONOMIC GEOLOGY, v. 11, p. 265-290.
Robert, F., 2001, Disseminated syenite-associated gold deposits in the Abitibi greenstone belt,
Canada: Mineralium Deposita, v. 36, p. 503-516.
Robert, F., and Brown, A.C., 1986a, Archean gold-bearing quartz veins at the Sigma Mine, Abitibi
greenstone belt, Quebec; Part I, Geologic relations and formation of the vein system:
---, 1986b, Archean gold-bearing quartz veins at the Sigma Mine, Abitibi greenstone belt, Quebec;
Part II, Vein paragenesis and hydrothermal alteration: ECONOMIC GEOLOGY, v. 81, p. 593-
616.
Robert, F., and Poulsen, K.H., 1997, World-class Archean gold deposits in Canada: an overview:
Australian Journal of Earth Sciences, v. 44, p. 329-351.
---, 2001, Vein formation and deformation in greenstone gold deposits: REVIEWS IN ECONOMIC
GEOLOGY, v. 14, p. 111-155.
Robert, F., Poulsen, K.H., and Dube, B., 1997, Gold deposits and their geological classification, in
Gubins, A.G., ed., Proceedings of Exploration 97: Fourth Decennial International Conference on
Mineral Exploration, p. 209-220.
Rombach, C.S., and Newberry, R.J., 2001, Shotgun deposit: Granite porphyry-hosted gold-arsenic
mineralization in southwestern Alaska, USA: Mineralium Deposita, v. 36, p. 607-621.
Roth, E., 1992, The nature and genesis of Archean porphyry-style Cu-Au-Mo mineralization at the
Boddington Gold Mine, Western Australia: Unpublished Ph.D. Thesis, University of Western
Australia, Perth, 126p.
Rowins, S.M., Groves, D.I., McNaughton, N.J., Palmer, M.R., and Eldridge, C.S., 1997. A
reinterpretation of the role of granitoids in the genesis of Neoproterozoic gold mineralization in
the Telfer dome, Western Australia: ECONOMIC GEOLOGY, v. 92, p. 133-160.
Safonov, Y.G., 1997, Hydrothermal gold deposits—distribution, geological/genetic types, and
productivity of ore-forming systems: Geology of Ore Deposits, v. 39, p. 20-32.
Sawkins, F.J., 1984, Metal Deposits in Relation to Plate Tectonics: Springer-Verlag, Berlin, 325pp.
Selby, D., Creaser, R.A., Hart, C.J.R., Rombach, C.S., Thompson, J.F.H., Smith, M.T., Bakke, A.A.,
and Goldfarb, R.J., 2002, Absolute timing of sulfide/gold mineralization: A comparison of Re-Os
molybdenite and Ar-Ar mica methods from the Tintina gold belt, Alaska: Geology, in press.
Seward, T.M., 1991. The hydrothermal geochemistry of gold : in Foster, R.P., ed., Gold Metallogeny
and Exploration. Chapman and Hall, London, p.37-62
Sibson, R.H., 1990, Faulting and fluid flow, in Nesbitt, B.E., ed., Fluids in Tectonically Active
Regimes of the Continental Crust: Mineralogical Association of Canada, Short Course 18, p. 93-
132.
Sibson, R.H., Robert, F., and Poulsen, K.H., 1988, High angle reverse faults, fluid-pressure cycling,
and mesothermal gold-quartz deposits: Geology, v. 16, p. 551-555.
Sillitoe, R.H., 1991, Intrusion-related gold deposits, in Foster, R.P., ed., Gold Metallogeny and
Exploration: Blackie and Son, Ltd., Glasgow, p. 165-209.
Page 43
Sillitoe, R.H., 1997, Characteristics and controls of the largest porphyry copper-gold and epithermal
gold deposits in the circum-Pacific region: Australian Journal of Earth Sciences, v. 44, p. 373-388.
Sillitoe, R.H., 2000, Enigmatic origins of giant gold deposits, in Cluer, J.K., Price, J.G., Struhsacker,
E.M., Hardyman, R.F., and Morrise C.L., eds, Geology and Ore Deposits 2000: The Great Basin
and Beyond, The Geological Society of Nevada, Reno, p. 1-18.
Sillitoe, R.H., and Thompson, J.F.H., 1998, Intrusion-related vein gold deposits: types, tectono-
magmatic settings and difficulties of distinction from orogenic gold deposits: Resource Geology,
v. 48, p. 237-250.
Smith, M.T., Thompson, J.F.H., Bressler, J., Layer, P., Mortensen, J.K., Abe, I., and Takaoka, H.,
1999, Geology of the Liese zone, Pogo property, east-central Alaska: Society of Economic
Geologists Newsletter, no. 38, p. 1, 12-21.
Smith, P.K., and Kontak, D.J., 1988, Bismuth-tellurium-silver-tungsten association at the Beaver
Dam gold deposit, eastern Nova Scotia: Nova Scotia Department of Minerals and Energy, Report
88-3, 11 p.
Smith, T.J., and Kesler, S.E., 1985, Relation of fluid inclusion geochemistry to wallrock alteration
and lithogeochemical zonation at the Hollinger-McIntyre gold deposit, Timmins, Ontario,
Canada: CIM Bulletin v. 78, no. 876, p. 35-46.
Spiridonov, E.M., 1996, Granitic rocks and gold mineralization of North Kazakhstan : in Shatov, V.,
ed., Granite-Related Ore Deposits of Central Kazakhstan and Adjacent Areas : Glagol Publishing
House, St. Petersburg, p.197-217.
Stein, H.J., Markey, R.J., Morgan, J.W., Hannah, J.L., Zak, K., and Sundblad, K., 1997, Re-Os dating
of shear-hosted Au deposits using molybdenite, in Papunen, H., ed., Mineral Deposits: Research
and Exploration—Where do They Meet?, A.A. Balkema, Rotterdam, p. 313-317.
Stüwe, K., 1998, Tectonic constraints on the timing relationships of metamorphism, fluid production
and gold-bearing quartz vein emplacement: Ore Geology Reviews, v. 13, p. 219-228.
Symons, P.M., Anderson, G., Beard, T.J., Hamilton, L.M., Reynolds, G.D., Robinson, J.M., Staley,
R.W. and Thompson, C.M., 1990, Boddington Gold Deposit, in Hughes, F.E., ed., Geology of the
Mineral Deposits of Australia and Papua New Guinea: Melbourne, Australasian Institute of
Mining and Metallogeny, Monograph 14, p. 165-169.
Thomas, R., Webster, J.D., and Heinrich, W., 2000, Melt inclusions in pegmatite quartz: complete
miscibility between silicate melts and hydrous fluids at low pressure: Contributions to Mineralogy
and Petrology, v. 139, p. 394-401.
Thompson, J.H.F., and Newberry, R.J., 2000, Gold deposits related to reduced granitic intrusions:
REVIEWS IN ECONOMIC GEOLOGY, v. 13, p. 377-400.
Thompson, J.H.F., Sillitoe, R.H., Baker, T., Lang, J.R., and Mortensen, J.K., 1999, Intrusion-related
gold deposits associated with tungsten-tin provinces: Mineralium Deposita, v. 34, p. 323-334.
Titley, S.R., 2001, Crustal affinities of metallogenesis in the American southwest: ECONOMIC
GEOLOGY, v. 96. p. 1323-1342.
Tomkinson, M.J., 1988, Gold mineralization in phyllonites at the Haile mine, South Carolina:
Tourigny, G., Brown, A.C., Hubert, C., and Crepeau, R., 1989, Synvolcanic and syntectonic gold
mineralization at the Bousquet mine, Abitibi greenstone belt, Quebec: ECONOMIC GEOLOGY,
v. 84, p. 1875-1890.
Tourigny, G., Douget, D., and Bourget, A., 1993, Geology of the Bousquet 2 mine: An example of a
deformed, gold-bearing polymetallic sulfide deposit: ECONOMIC GEOLOGY, v. 88, p. 1578-
1597.
Valliant, R.I., and Hutchinson, R.W., 1982, Stratigraphic distribution and genesis of gold deposits,
Bousquet region, Northwestern Quebec: in Hodder, R.W. and Petruk, W., eds, Geology of
Canadian Gold Deposits: Canadian Institute of Mining and Metallurgy, Special Volume 24, p. 27-
40.
Volkov, A.V., Sidorov, A.A., Goncharov, V.I., and Sidorov, V.A., 2002, Disseminated gold-sulfide
deposits in the Russian NE: Geology of Ore Deposits, v. 44, p. 179-197.
Wall, V.J., 1999, Pluton-related (thermal aureole) gold. Yukon Geoscience Forum, Short Course
Notes, Whitehorse, Yukon, November, 1999.
Page 44
Wilde, A.R., Layer , P., Mernagh, T., and Foster, J., 2001. The giant Muruntau gold deposit :
geologic, geochronologic, and fluid constaints on genesis : ECONOMIC GEOLOGY, v.96,
p.633-644.
Witt, W.K., 1997, Some atypical styles of gold mineralization and alteration in the Yilgarn Block, in
Cassidy, K.F., Whitaker, A.J., and Liu, S.F., eds., An International Conference on Crustal
Evolution, Metallogeny and Exploration of the Yilgarn Craton – An Update. Extended Abstracts:
Australian Geological Survey Organisation, Record 1997/41, p. 151-156.
Witt, W.K., 2001, Tower Hill gold deposit, Western Australia: an atypical, multiply deformed
Archaean gold-quartz vein deposit: Australian Journal of Earth Sciences, v. 48, p. 81-99.
Wood, S.A., and Samson, I.M., 1998, Solubility of ore minerals and complexation of ore metals in
hydrothermal solutions: REVIEWS IN ECONOMIC GEOLOGY, v. 10, p. 33-80.
Worthington, J.E., Kiff, I.T., Jones, E.M., and Chapman, P.E., 1980, Applications of the hot springs
or fumarolic model in prospecting for lode gold deposits: Mining Engineering, v. 32, p. 73-79.
Wygralak, A.S., and Mernagh, T.P., 2001, Gold mineralisation of the Tanami region: Northern
Territory Geological Survey, Record 2001-011.
Wyman, D.A., Kerrich, R., and Groves, D.I., 1999, Lode gold deposits and Archean mantle plume-
island arc interaction, Abitibi subprovince, Canada: Journal of Geology, v. 107, p. 715-725.
Yakubchuk, A., Cole, A., Seltmann, R., and Shatov, V., 2002, Tectonic setting, characteristics, and
regional exploration criteria for gold mineralization in the Altaid orogenic collage: The Tien Shan
province as a key example: SOCIETY OF ECONOMIC GEOLOGISTS SPECIAL
PUBLICATION, 9, p. 177-201.
Yakubchuk, A., Seltmann, R., Shatov, V., and Cole, A., 2001, The Altaids: Tectonic evolution and
metallogeny: Society of Economic Geologists Newsletter, No. 46, p. 1, 7-14.
Yao, Y., Morteani, G., and Trumbull, R.B., 1999, Fluid inclusion microthermometry and the P-T
evolution of gold-bearing hydrothermal fluids in the Niuxinshan gold deposit, eastern Hebei
province, NE China: Mineralium Deposita, v. 34, p. 348-365.
Yeats, C.J., McNaughton, N.J., and Groves, D.I., 1996, SHRIMP U-Pb geochronological constraints
on Archean volcanic-hosted massive sulfide and lode gold mineralization at Mount Gibson,
Yilgarn Craton, Western Australia: ECONOMIC GEOLOGY, v. 91, p. 1354-1371.
Zacharias, J., Pertold, Z., Pudilova, M., Zak, K., Pertoldova, J., Stein, H., and Markey, R., 2001,
Geology and genesis of Variscan porphyry-style gold mineralization, Petrackova hora deposit,
Bohemian Massif, Czech Republic: Mineralium Deposita, v. 36, p. 517-541.
Zaw, K., Huston, D.L., Large, R.R., Mernagh, T., and Hoffman, C.F., 1994. Microthermometry and
geochemistry of fluid inclusions from the Tennant Creek gold-copper deposits: Implications for
ore deposition and exploration: Mineralium Deposita, v. 29, p. 288-300.
Zaw, K., Huston, D.L., and Large, R.R., 1999, A chemical model for the Devonian remobilization
process in the Cambrian volcanic-hosted massive sulfide Rosebery deposit, western Tasmania:
Zhou, T., Goldfarb, R.J., and Phillips, G.N., 2002, Tectonics and distribution of gold deposits in
China—An overview: Mineralium Deposita, v. 37, p. 249-282.
Page 45
1.12. Appendix
The paper contains a large number of deposit descriptions. For ease of reference, there are listed
alphabetically in Table I and their locations are shown in Figure I. Only orogenic and gold-dominant
intrusion-related gold deposits and gold-rich deposits with atypical metal associations are listed.
References to other deposit types are given in the text.
Figure I. Location of orogenic and intrusion-related gold deposits and deposits with atypical metal
associations discussed in the text and listed in Table 4.
Page 46
Table 1: List, in alphabetical order are the orogenic and intrusion-related gold deposits and deposits with atypical metal associations that are
discussed in the text. The gold province is taken from Goldfarb et al. (2001) for consistency. Varying classifications are both given where
uncertainty exists. For deposits with atypical metal associations the probable major deposit style is given in brackets. Classification into
Super Giant (>2500 t Au), Giant (>250 t Au), World Class (>100 t Au) and Sub-World Class (<100 t Au) categories is based on best current
knowledge of production and resources for primary resources: alluvial production is excluded. A key recent reference is given, which where
possible, contains comprehensive reference lists: this is not possible for some deposits where there are limited publications in English. The
locations of the deposits are shown in Figure 6. For other deposit types discussed in the text, references are given that provide critical
geological and location data for individual deposits.
No. Deposit Gold Province Classification Size Key Reference
1 Ashanti (Obuasi) Birmian belts Orogenic Giant Allibone et al. (2002)
2 Ballarat Lachlan fold belt Orogenic World Class Bierlein et al. (2001a)
3 Bendigo Lachlan fold belt Orogenic Giant Bierlein et al. (2001a)
4 Berezovsk East-central Urals Orogenic Giant Kisters et al. (1999)
5 Boddington West Yilgarn Atypical (Porphyry) World Class McCuaig et al. (2001)
6 Boliden Svecofennian Province Atypical (VHMS) World Class Hannington et al. (1999)
7 Bousquet Southern Superior Province Atypical (VHMS) World Class Tourigny et al. (1993)
8 Brewery Creek Tombstone province Intrusion-related Sub-World Class Hart et al. (2000)
9 Calie Northern Territory Inliers Orogenic World Class Wygralak and Mernagh (2001)
10 Campbell-Red Lake Uchi Subprovince Orogenic Giant Dubé et al. (2002)
11 Charters Towers Thomson fold belt Orogenic or Intrusion-related World Class Sillitoe and Thompson (1998)
12 Clear Creek Tombstone province Intrusion-related Sub-World Class Marsh et al. (2003)
13 Cosmo Howley Northern Territory Inliers Orogenic or Intrusion-related World Class Matthai et al. (1995)
14 Dongping Yan-Liao/Changbaishan Orogenic or Intrusion-related Sub-World Class Sillitoe and Thompson (1998)
15 Donlin Creek Tombstone province Intrusion-related Giant? Ebert et al. (2000)
16 Fort Knox Tombstone province Intrusion-related World Class Selby et al. (2002)
17 Golden Mile Eastern Goldfields Province Orogenic Giant Bateman et al. (2001)
18 Hemlo Southern Superior Province Atypical (Epithermal) Giant Lin (2001)
19 Henty Mt Read belt Atypical (VHMS) Sub-World Class Halley and Roberts (1997)
20 Hillgrove New England fold belt Orogenic Sub-World Class Boyle (1990)
21 Hollinger-McIntyre Southern Superior Province Orogenic (+Porphyry) Giant Burrows et al. (1993)
22 Homestake Dakota segment Orogenic Giant Caddey et al. (1993)
23 Horne Southern Superior Province Atypical (VHMS) Giant Hannington et al. (1999)
24 Independence Southern Alaska Orogenic Sub-World Class Madden-McGuire et al. (1989)
25 Jilau Southern Tian Shan Intrusion-related World Class Cole et al. (2000)
26 Alaska-Juneau Juneau gold belt Orogenic World Class Goldfarb et al. (1997)
27 Kolar E. Dharwar block Orogenic Giant Radhakrishna and Curtis(1999)
28 Kumtor Southern Tian Shan Orogenic Giant Jenchuraeva et al. (2001)
29 Linglong Jiaodong Peninsular Orogenic or Intrusion-related Sub-World Class Qiu et al. (2002)
30 Malartic Southern Superior Province Intrusion-related or Orogenic World Class Robert (2001)
31 Maldon Lachlan fold belt Orogenic Sub-World Class Foster et al. (1998)
32 Mokrsko Bohemian Massif Orogenic or Intrusion-related World Class Cliff and Moravek (1995)
33 Morro Velho Rio das Velhas Orogenic Giant Lobato et al. (2001)
34 Mother Lode belt Sierra Foothills Orogenic Giant Bohlke and Kistler (1986)
35 Mt Charlotte Eastern Goldfields Province Orogenic World Class Bateman et al. (2001)
36 Mt Gibson West Yilgarn Orogenic (+VHMS) Sub-World Class Yeats et al. (1996)
37 Muruntau Southern Tian Shan Orogenic or Intrusion-related Super Giant Kempe et al. (2001)
38 Olympiada Yenisei fold belt Orogenic or Intrusion-related Giant Genkin et al. (2002)
39 Ouro Preto Quadrilatero Ferrifero Orogenic Sub-World Class Chauvet et al. (2001)
40 Oroya Eastern Goldfields Province Orogenic World Class Bateman et al. (2001)
41 Petrockhova Hora Bohemian Massif Orogenic or Intrusion-related World Class Zacharias et al. (2001)
42 Pogo Tombstone province Orogenic or Intrusion-related World Class Smith et al. (1999)
43 Ryan Lode Tombstone province Intrusion-related Sub-World Class Sillitoe and Thompson (1998)
44 Salsigne Massif Central Orogenic World Class Marignac and Cuney (1999)
45 Sarylakh Yara-Kolyma Orogenic World Class Volkov et al. (2002)
46 Shotgun Tombstone belt Intrusion-related Sub-World Class Romback and Newberry (2001)
47 Sigma-Lamaque Southern Superior Province Orogenic World Class Robert and Brown (1986a,b)
48 Stawell Lachlan fold belt Orogenic World Class Foster et al. (1998)
49 Sukhai Log Baikal fold belt Orogenic or Intrusion-related Giant Laverov et al. (2000)
50 Sumdum Chief Juneau gold belt Orogenic Sub-World Class Goldfarb et al. (1997)
51 Telfer Paterson Province Intrusion-related or Orogenic World Class Rowins et al. (1997)
52 Tennant Creek Tennant Creek Province Intrusion-related or Orogenic World Class Zaw et al. (1994)
53 Timbarra New England fold belt Intrusion-related Sub-World Class Mustard (2001)
54 Tower Hill Eastern Goldfields Province Intrusion-related or Orogenic Sub-World Class Witt (2001)
55 True North Tombstone province Intrusion-related or Orogenic Sub-World Class Hart et al. (2000)
56 Vasil'kovsk Kipchak arc Orogenic or Intrusion-related Giant Spiridonov (1996)
57 Wiluna Eastern Goldfields Province Orogenic World Class Hagemann and Cassidy (2000)
Page 47
2. Orogenic gold deposits: A proposed

classification in the context of their crustal
distribution and relationship to other gold
deposit types
D.I. Groves a,*, R.J. Goldfarb b, M. Gebre-Mariam a,c, S.G. Hagemann a, F. Robert d
a
b
U.S. Geological Survey, Box 25046, Mail Stop 973, Denver Federal Center, Denver, CO
80225, USA
c
Wiluna Gold Mines Limited, 10 Ord St., West Perth, WA 6005, Australia
d
Geological Survey of Canada, 601 Booth St., Ottawa, Ont., Canada KIA OE
Abstract
The so-called `mesothermal' gold deposits are associated with regionally metamorphosed terranes of
all ages. Ores were formed during compressional to transpressional deformation processes at
convergent plate margins in accretionary and collisional orogens. In both types of orogen, hydrated
marine sedimentary and volcanic rocks have been added to continental margins during tens to some
100 million years of collision. Subduction-related thermal events, episodically raising geothermal
gradients within the hydrated accretionary sequences, initiate and drive long-distance hydrothermal
fluid migration. The resulting gold-bearing quartz veins are emplaced over a unique depth range for
hydrothermal ore deposits, with gold deposition from 15-20 km to the near surface environment.
On the basis of this broad depth range of formation, the term `mesothermal' is not applicable to this
deposit type as a whole. Instead, the unique temporal and spatial association of this deposit type with
orogeny means that the vein systems are best termed orogenic gold deposits. Most ores are
post-orogenic with respect to tectonism of their immediate host rocks, but are simultaneously
syn-orogenic with respect to ongoing deep-crustal, subduction-related thermal processes and the
prefix orogenic satisfies both these conditions. On the basis of their depth of formation, the orogenic
deposits are best subdivided into epizonal (< 6 km), mesozonal (6-12 km) and hypozonal (> 12 km)
classes.
2.1. Introduction
This thematic issue of Ore Geology Reviews includes a wide variety of papers on a single type of
quartz-carbonate lode-gold deposit. The deposit type in this issue alone is referred to as synorogenic,
turbidite-hosted, mesothermal and Archaean lodegold. This reflects the proliferation of such terms
throughout the economic geology literature during the last ten years and a subsequent increase in
confusion for the readers. For example, is a synorogenic Mother-lode type gold deposit different from
an Archaean gold-only type or from a mesothermal greenstone-gold type? Many researchers working
on such deposits would recognize these as essentially a variety of subtypes of a single deposit type,
Page 1
i.e. epigenetic, structurally-hosted lode-gold vein systems in metamorphic terranes (Kerrich, 1993).
However, the consistent usage of a single and widelyaccepted classification term for this deposit type
as a whole is clearly warranted. `Mesothermal' is such a term that has been widely adopted during the
last ten years, but is a term that, as originally defined by Lindgren (1933) for deposits formed at about
1.2-3.6 km, is more applicable to sedimentary rock-hosted 'Carlin-type' deposits and the gold
porphyry/skarn environment (Poulsen, 1996).
A principal aim of this introductory paper is to present and justify a unifying classification for these
lode-gold deposits. An attempt is made to place these so-called `mesothermal' deposits into a broader
class that emphasizes their tectonic setting and time of formation relative to other gold deposit types.
A second aim is to review briefly their more significant defining features in the light of current
inconsistent terminology and the recognition that this deposit group may form over a wider range of
crustal depths and temperatures than commonly recognized (Groves, 1993; Hagemann and Ridley,
1993; GebreMariam et al., 1995). The term orogenic is introduced and justified as a term to replace
`mesothermal' and other descriptors for this deposit type. It is also suggested that the terms epizonal,
mesozonal and hypozonal be used to reflect crustal depth of gold deposition within the orogenic
group of deposits.
2.2. Definition of So-Called Mesothermal Gold Deposits
The so-called `mesothermal' gold deposits (Table 1) are a distinctive type of gold deposit which is
typified by many consistent features in space and time. These have been summarized in a variety of
comprehensive ore-deposit model descriptions that include Bohlke (1982), Colvine et al. (1984),
Berger (1986), Groves and Foster (1991), Nesbitt (1991), Hodgson (1993) and Robert (1996).
Kerrich (1993) summarizes many of the steps that led to these evolving modern-day models. A
unifying tectonic theme has recently been evaluated by workers such as Wyman and Kerrich (1988),
Barley et al. (1989), Hodgson and Hamilton (1989), Kerrich and Wyman (1990), Kerrich and
Cassidy (1994), and Goldfarb et al. (1998 - this issue).
2.2.1. Geological characteristics
2.2.1.1. Geology of host terranes
Perhaps the single most consistent characteristic of the deposits is their consistent association with
deformed metamorphic terranes of all ages. Observations from throughout the world's preserved
Archaean greenstone belts and most recently active Phanerozoic metamorphic belts indicate a strong
association of gold and greenschist facies rocks. However, some significant deposits occur in higher
metamorphic grade Archaean terranes (e.g. McCuaig et al., 1993) or in lower metamorphic grade
domains within the metamorphic belts of a variety of geological ages. In the Archaean of Western
Australia, a number of synmetamorphic deposits extend into granulite facies rocks (Groves et al.,
1992). Premetamorphic protoliths for the auriferous Archaean greenstone belts are predominantly
volcano-plutonic terranes of oceanic back-arc basalt and felsic to mafic arc rocks. Clastic marine
sedimentary rockdominant terranes that were metamorphosed to graywacke, argillite, schist and
phyllite host most younger ores, and are important in some Archaean terranes (e.g. Slave Province,
Canada).
2.2.1.2. Deposit mineralogy
These deposits are typified by quartz-dominant vein systems with <= 3-5% sulfide minerals (mainly
Fe-sulfides) and <= 5-15% carbonate minerals. Albite, white mica or fuchsite, chlorite, scheelite and
tourmaline are also common gangue phases in veins in greenschist-facies host rocks. Vein systems
Page 2
may be continuous along a vertical extent of 1-2 km with little change in mineralogy or gold grade;
mineral zoning does occur, however, in some deposits. Gold:silver ratios range from 10 (normal) to 1
(less common), with ore in places being in the veins and elsewhere in sulfidized wallrocks. Gold
grades are relatively high, historically having been in the 5-30 g/t range; modern-day bulk mining
methodology has led to exploration of lower grade targets. Sulfide mineralogy commonly reflects the
lithogeochemistry of the host. Arsenopyrite is the most common sulfide mineral in metasedimentary
country rocks, whereas pyrite or pyrrhotite is more typical in metamorphosed igneous rocks. In fact,
the Salsigne gold deposit in Cambrian sedimentary rocks of the French Massif Central is the world's
largest producer of arsenic (Guen et al., 1992). Gold-bearing veins exhibit variable enrichments in As,
B, Bi, Hg, Sb, Te and W; Cu, Pb and Zn concentrations are generally only slightly elevated above
regional backgrounds.
2.2.1.3. Hydrothermal alteration
Deposits exhibit strong lateral zonation of alteration phases from proximal to distal assemblages on
scales of metres. Mineralogical assemblages within the alteration zones and the width of these zones
generally vary with wallrock type and crustal level. Most commonly, carbonates include ankerite,
dolomite or calcite; sulfides include pyrite, pyrrhotite or arsenopyrite; alkali metasomatism involves
sericitization or, less commonly, formation of fuchsite, biotite or K-feldspar and albitization and
mafic minerals are highly chloritized. Amphibole or diopside occur at progressively deeper crustal
levels and carbonate minerals are less abundant. Sulfidization is extreme in BIF and Fe-rich mafic
host rocks. Wallrock alteration in greenschist facies rocks involves the addition of significant
amounts of CO2, S, K, H2O, Si02 ± Na and LILE.
2.2.1.4. Ore fluids
Ores were deposited from low-salinity, near-neutral, H20_CO2 ± CH, fluids which transported gold as
a reduced sulphur complex. Fluids associated with this gold deposit type are notable by their
consistently elevated CO2 concentrations of >= 5 mol%. Typical 8180 values for hydrothermal fluids
are about 5-8 per ml in the Archaean greenstone belts and about 2 per ml higher in the Phanerozoic
gold lodes.
2.2.1.5. Structure
There is strong structural control of mineralization at a variety of scales. Deposits are normally sited
in second or third order structures, most commonly near large-scale (often transcrustal) compressional
structures. Although the controlling structures are commonly ductile to brittle in nature, they are
highly variable in type, ranging from: (a) brittle faults to ductile shear zones with low-angle to high-
angle reverse motion to strike-slip or oblique-slip motion; (b) fracture arrays, stockwork networks or
breccia zones in competent rocks; (c) foliated zones (pressure solution cleavage) or (d) fold hinges in
ductile turbidite sequences. Mineralized structures have small syn- and post-mineralization
displacements, but the gold deposits commonly have extensive down-plunge continuity (hundreds of
metres to kilometres). Extreme pressure fluctuations leading to cyclic fault-valve behavior (Sibson et
al., 1988) result in flat-lying extensional veins and and mutually cross-cutting steep fault veins that
characterize many deposits (e.g. Robert and Brown, 1986).
Page 3
2.2.2. Tectonic setting and timing of `mesothermal' vein

emplacement
The so-called `mesothermal' gold deposits (Table 1) occupy a consistent spatial/temporal position
(Fig. 1), having formed during deformational processes at convergent plate margins (orogeny)
irrespective of whether they are hosted in Archaean or Proterozic greenstone belts or Proterozoic and
Phanerozoic sedimentary rock sequences (e.g. Barley and Groves, 1992; Kerrich and Cassidy, 1994).
The placing of these deposits in a plate tectonic setting was a logical outgrowth of the acceptance of
plate tectonic theory in the early 1970's. Guild (1971) initially discussed the "orogen-associated
endogenic mineral deposits of Mesozoic and Tertiary age on the sites of Cordilleran-type
(continent/ocean) collisions". Sawkins (1972) noted, soon after, how both these Circum-Pacific gold
ores and spatially associated felsic magmas were probable products of subduction-related tectonism.
Just as significant was Sawkins (1972) observation that Archaean gold lodes in the Superior Province,
Canada, may have some relationship to the southward younging of igneous ages, interpreted as being
reflective of a seawardmigrating trench. It would be, however, another sixteen years (cf. Wyman and
Kerrich, 1988) before workers would follow-up on this important concept and begin to widely look at
Archaean gold as a product of continental-margin deformational events.
The concept of a general spatial association between the gold deposits and subduction-related thermal
processes in accretionary orogens (oceanic-continental plate interactions) became commonplace in
the mid-1980's. Fyfe and Kerrich (1985) presented a model at that time to explain the massive fluid
volumes required for the numerous gold-bearing vein swarms adjacent to crustal-scale thrust zones of
continental margins. They hypothesized that underplated hydrated rocks contained the required water
and such water was released during thermal reequilibration as subduction ceased. Subsequent models
for the Mesozoic and Cenozoic gold fields of westernmost North America relied heavily on
correlating gold vein emplacement with subduction-driven processes (Bohlke and Kistler, 1986;
Goldfarb et al., 1988). Landefeld (1988), expanding on the ideas in Fyfe and Kerrich (1985), detailed
how the seaward stepping of subduction accompanying terrane accretion could have been crucial for
the formation of the Sierra foothills gold districts (including the Mother lode belt). With an
abundance of new geochronological data from western North America, recent models of gold genesis
in accretionary orogens have been able to look closely at specific processes (e.g. changing plate
motions, changing collisional velocities, ridge subduction, etc.) occurring during accretion/subduction
that tend to be most closely associated with veining (e.g. Goldfarb et al., 1991b; Elder and Cashman,
1992; Haeussler et al., 1995). Theoretically, as a subduction zone steps seaward, a series of gold
systems and plutonic bodies should develop and young towards the trench-part of a so-called Turkic-
type (Sengor and Okurogullari, 1991) orogen. This type of scenario crudely characterizes Alaska,
USA, a part of the North American margin almost entirely composed of accreted oceanic rock
sequences (Plafker and Berg, 1994).
Page 4
Table 1: Timing of orogenic gold vein formation and significant tectonic relationships from some gold provinces in metamorphic rocks (partly modified from Kerrich and
Cassidy, 1994; Goldfarb et al., 1998). Host terranes are mainly Archaean greenstone belts and younger oceanic sedimentary rock-dominant assemblages. Provinces are
ordered, from top to bottom of the table, in increasing age of formation
Province Age of Age of Spatially Metamorphic Other important events Geochron. Refs.
veining host associated events
(Ma) terranes magmatism (Ma)
(Ma) (Ma)
Mt. Rosa, upper <_ 33 Palaeozoic 310, 42-25 415, 90-60 hypothesized slab delamination at Curti (1987), Blanckenburg
nappes, W. Alps, (most (blueschist); 45 Ma and Davies (1995)
Italy abundant 44-40
at 33-29)
Chugach 57-49 L. Cretaceous 66-50 66-50 veining during subduction of Haeussler et al. (1995)
accretionary spreading ridge beneath growing
prism, S. Alaska prism
Juneau gold belt, 57-53 Permian- mid-Cret, mid-Cret, 70-60 emplacement of sill during Goldfarb et al. (1991b),
S. Alaska mid-Cretaceous 70-60 (sill), Barrovian metamorphism; change Miller et al. (1994)
60-48 from orthogonal to oblique
(batholith) convergence during veining
Willow Creek 66 Late Paleozoic 74-66 Jurassic veining during onset of oroclinal Madden-McGuire et al.
district, south- bending of Alaska; syn-veining (1989)
central Alaska accretion and subduction tens
of km seaward
Bridge River, SW 91-86 Late 270, 91-43 Jurassic veining during seaward collision Leitch et al. (1991)
British Columbia Paleozoic- of Wrangellia terrane and early
early stages of Coast batholith
Mesozoic formation
Page 5
(Ma) (Ma)
Fairbanks, east- 92-87, 77 Early 95-90 Early-Middle 120-110 Ma regional extension; McCoy et al. (1997)
central Alaska Paleozoic Jurassic syn-veining accretion and
subduction tens of km seaward;
veining continues into
unmetamorphosed rocks of craton
in Yukon
Nome, NW 109 Early 108-82 170-130 veining during regional Ford and Snee (1996)
Alaska Paleozoic (blueschist), extension and slab rollback; veins
108-82 40-50 km from high-T magmatic/
(Barrovian) metamorphic front
Russian Far East 135-100 Late 144-80 Late Jurassic- veining during increased Nokleberg et al. (1996)
Paleozoic- Early Cretaceous convergence rates between Goldfarb et al. (1998)
middle Eurasian and Izanagi plates
Mesozoic
Shangdong Early Archaean 190-170, Archaean veining during late stage of Trumbull et al. (1996),
Peninsula (E. Cretaceous 132-121 Yanshanian magmatism; Wang et al. (1996), Nie
China), NE hypothesized mantle plume during (1997)
China and Korea onset of post-collisional extension
Page 6
(Ma) (Ma)
Sierra foothills 144-108 Middle 177-135 Jurassic-Early 150-140 Ma seaward stepping of Bohlke and Kistler (1986),
and Klamath (127-108 = Paleozoic- (north), Cretaceous trench; 120 Ma onset of rapid, Landefeld (1988), Elder
Mts., California Mother Jurassic 150-80 orthogonal convergence and and Cashman (1992)
lode belt) (south) Sierra Nevada batholith emplacement
Otago, South Jurassic-Early Permian-Late none Early Jurassic- veining likely throughout last
McKeag and Craw (1989)
Island, New Cretaceous Triassic Early Cretaceous period of collisional deformation
Zealand along Gondwanan margin
SW Yukon and 180- >_ 134 Early 190-160 Late Triassic- younger dates on mineralization Rushton et al. (1993),
Interior British Paleozoic- Early Jurassic could be cooling ages; syn- Ash et al. (1996)
Columbia Triassic veining accretion and subduction
tens of km seaward
New England Permian-Early Carboniferous- 306-280, Permian-Triassic veining related to final period of
Ashley et al. (1994),
fold belt, E. Triassic Permian 255-245, Early accretion and subduction along Scheiber (1996)
Australia Triassic eastern Australia
Muruntau, Late Cambrian- 310, 271-261 Late deposits near suture of Hercynian Berger et al. (1994),
Uzbekistan and Carboniferous- Ordovician Carboniferous- continent-continent collision
adjacent central Early Permian Early Permian Drew et al. (1996)
Asia deposits
Page 7
(Ma) (Ma)
Variscan-related, 340-310 Late 360-320 350-340 Late Devonian(?)-Permian Bouchot et al. (1989),
Europe (Bohemia Proterozoic- subduction; Laurorussia-Africa Cathelineau et al. (1990),
Massif); early collision by 380-350 Ma Moravek (1995),
300±20 Paleozoic Stein et al. (1996)
(Massif Central)
Southern 343-294 Paleozoic Late Carboniferous veins emplaced at higher P-T and Stowell et al. (1996)
Appalachians, Ordovician to (main event); deeper crustal levels than other
USA Carboniferous lower grade Phanerozoic orogenic gold
episodes in Late deposits in North America
Ordovician and
Devonian
Meguma, Nova 380-362 Cambrian- 380-370, 316 415-377 host rocks obducted to Kontak et al. (1990),
Scotia Ordovician continental margin between Late Keppie and Dallmeyer
Silurian and Early Permain (1995)
Victoria, SE 460(?), Ordovician 415-390, 460-430 (Stawell- subduction event?; thin-skinned Arne et al. (1996), Foster
Australia 415-360 Early 370-360 Ballarat-Bendigo), tectonics; conflicting data on age et al. (1996), Phillips and
Devonian 410-400 of gold mineralization Hughes (1996)
(Melbourne)
Queensland, NE 408±30, Late Silurian- Middle Devonian subduction event?; thin-skinned

Australia Carboniferous Devonian Ordovician- tectonics Solomon and Groves
Middle (1994)
Devonian,
Carboniferous
Page 8
(Ma) (Ma)
Trans-Hudson 1807-1720 Early 1890-1834 1870-1770 perhaps a series of unrelated Ansdell and Kyser (1992),
orogen, central Proterozoic thermal and ore-forming events; Thomas and Heaman
Canada regional transpression continued (1994), Fayek and Kyser
until 1690 Ma (1995), Conners (1996)
Birimian belt of about 2100 2185-2150 2185-2150, veining in basinal rocks during Hirdes et al. (1996)
Ghana-eastern (volcanics); 2116-2088 oblique thrusting (Ebumean
Cote d'Ivorie- adjacent deformation) of these over
Burkina Faso basins are volcanic sequences
slightly younger
Dharwar craton, about 2400(?) 2700-2530 2550 mineralization during collision

Krogstad et al. (1989),
S. India and suturing of numerous terranes Balakrishnan et al. (1990)

to form the Kolar schist belt,
which is the site of the most
important ores; age of
mineralization poorly-constrained
Yilgarn craton, 2640-2620, 2750-2685 2690-2660, 2690-2660, youngest date on veining could Kent and McDougall
W. Australia 2602, 2565(?) 2650-2630 2650-2630 be cooling age; metamorphism
(1995), Kent et al. (1996),
poorly-constrained Kent and Hagemann (1996)
Slave craton, about 2670-2660 Middle and 2663, 2640- about 2690 100-m.y.-long subduction regime
Abraham and Spooner
Page 9
(Ma) (Ma)
NWT, Canada Late Archaean 2585 initiated by 2712 (1995), MacLachlan and
Helmstaedt (1995)
Zimbabwe craton, 2670, 2659, Early and Late 2700-2600, 2690(?) poorly dated crustal evolution Foster and Piper (1993),
Zimbabwe 2410(?) Archaean 2460 (Great Darbyshire et al. (1996),
Dyke), 2428 Vinyu et al. (1996)
Superior 2720-2670, Middle and 2720-2673, 2690-2643 young period for mineralization Kerrich (1994), Kerrich
Province, Canada 2633-2404(?) Late Archaean 2645-2611 might reflect thermal resetting of
and Cassidy (1994),
true ages Jackson and Cruden
(1995), Powell et al. (1995)
Kaapvaal craton, 3200-3064 3600-3200 (in 3437, 3106, > 3200, some at in Barberton, mineralization at deRonde et al. (1991),
South Africa (Barberton Barberton 3000-2700, 2850 least 100 m.y. after thrusting and Foster and Piper (1993)
belt); > 2700 belt) 2600-2500 regional metamorphism of hosts;
with perhaps some of the mineralization may
some at 2850 correlate with that of the Pilbara
(Murchison belt) block, western Australia
Page 10
Collisional orogens (continent-continent collision), including the Variscan, Appalachian and Alpine,
also are host environments for gold deposits. In fact, collisional (or internal) and accretionary (or
peripheral) orogens may represent end-members of a continuous process. Any continent-continent
collision will be preceded by closure of an ocean basin, and hence is nothing more than a final stage
of a peripheral orogen. The gold systems that are associated with the Phanerozoic internal orogens are
actually all spatially associated with marine rocks that have been caught up within the suture. In
addition, within peripheral orogens, accretion of microcontinents such as Wrangellia along western
North America (Plafker and Berg, 1994) or Avalonia along Laurentia (Keppie, 1993) may be viewed
as a type of small-scale continent-continent collision. A key point in all examples is that hydrated
marine sedimentary and volcanic rocks were added to continental margins and, at some time during
this growth, the accreted rocks experienced relatively high geothermal gradients.
Fig. 1. Tectonic settings of gold-rich epigenetic mineral deposits. Epithermal veins and gold-rich
porphyry and skarn deposits, form in the shallow (<5 km) parts of both island and continental arcs in
compressional through extensional regimes. The epithermal veins, as well as the sedimentary rock-
hosted type Carlin ores, also are emplaced in shallow regions of back-arc crustal thinning and
extension. In contrast, the so-called `mesothermal' gold ores (termed orogenic gold on this diagram)
are emplaced during compressional to transpressional regimes and throughout much of the upper
crust, in deformed accretionary belts adjacent to continental magmatic arcs. Note that both the
lateral and vertical scale of the arcs and accreted terranes have been exaggerated to allow the gold
deposits to be shown in terms of both spatial position and relative depth of formation.
Oligocene veins in the western European Alps (Curti, 1987) are the youngest recognized, economic
examples of this deposit type. They also serve to point out that more than simple plate subduction is
required for vein formation. The closure of an ocean basin between Europe and Adria (perhaps a part
of northern Africa) occurred during an 80-m.y.-long period of Early Cretaceous-early Tertiary
oceanic crust subduction without any preserved evidence of gold veining or magmatism; blueschist
metamorphic facies in the Alps now record the low thermal gradients. By the early Eocene, complete
closure of the ocean had led to continent-continent collision and a partial subduction of the European
continental margin between 55 and 45 Ma (Blanckenburg and Davies, 1995). It was not until almost
100 m.y. subsequent to the onset of convergence, perhaps due to slab delamination resulting in the
cessation of subduction at 45-40 Ma (Blanckenburg and Davies, 1995), that magmatism and high
temperature metamorphism impacted the obducted upper nappes of the western Alps near the
collisional suture. Much of the Alpine gold veining occurred during the early Oligocene peak of
magmatism (Curti, 1987).
The understanding of gold-forming processes and timing in older Phanerozoic orogens may be
complicated by the hundreds of millions of years of additional geological time, but certainly such
Palaeozoic continental margins were favorable environments for veining. Geochronological study of
Page 11
the gold deposits in the Meguma terrane of Nova Scotia, Canada, indicates veining between 380 and
362 Ma (Kontak et al., 1990), during the late part of Acadian deformation of the Appalachian orogen.
The Meguma was the final terrane accreted to the Atlantic margin during the poorly-understood late
Palaeozoic Laurentia-Gondwanaland collision. Keppie and Dallmeyer (1995), noting that magmatism
and high-temperature metamorphism were restricted to a narrow time range of about 380-370 Ma,
rather than the prolonged 100 m.y. of Meguma collision, suggest a distinct episode of lower
lithospheric delamination for the thermal perturbation. This brief thermal event, occurring at the same
time as gold veining, is also likely to be important to the ore-forming process. Whereas little is certain
about the subduction-related tectonics of the northern Appalachians, mesothermal-type gold ores such
as the Hammer Down in northwestern Newfoundland (Gaboury et al., 1996) indicate that a broad belt
of gold systems accompanied continental growth.
Palaeozoic gold veins of the Tasman orogenic system in eastern Australia make it clear that the ore-
forming process need not require a `Cordilleran style' of terrane accretion. Unlike the collage of small
terranes that formed the accreted margin of western North America, eastern Australia is mainly
composed of a single lithotectonic assemblage (the Lachlan `terrane') that represents a 2,000-km-wide
Palaeozoic turbidite fan sequence developed adjacent to the Gondwanan craton (Coney, 1992). Such
an environment lacks deep-crustal terrane-bounding faults located between accreted material and the
active margin, which, where present in the North American Cordillera, expose a variety of crustal
levels and often serve as the focus of hydrothermal fluid flow. Compression-related deformation is
solely intraplate rather than concentrated along sutures between terranes. The fact that such a large
percentage of gold has been concentrated in the Bendigo-Ballarat area of Victoria (Phillips and
Hughes, 1996; Ramsay, 1998 - this issue) indicates some significant and still poorly-understood, local
control on vein emplacement in the orogenic system. Nonetheless, similar to the North American
Cordillera, the Tasman orogenic system is characterized by significant growth of the eastern
Australian margin (addition of the Lachlan `terrane') and a subduction zone east of the Lachlan
assemblage throughout much of the Palaeozoic (Solomon and Groves, 1994).
The abundance of geological similarities between the gold ores of the Phanerozoic orogens and those
in Archaean greenstone belts began to be interpreted by the late 1980's as evidence of a similar
tectonic setting for ore formation. Wyman and Kerrich (1988) hypothesized that gold mineralization
in the Superior Province of Canada was "related to convergent plate margin-style tectonics". At
roughly the same time, Barley et al. (1989) independently reached the same conclusion to explain the
development of gold lodes in Western Australia. Subduction of oceanic rocks into the zone of partial
melting appeared to be significant in the development of these gold ores within orogens of all ages
(Hodgson and Hamilton, 1989). Major fault zones spatially associated with auriferous belts in the
Archaean terranes were now recognized by several researchers as ancient terrane boundaries. Kerrich
and Wyman (1990) pointed out that, as observed in present-day convergent margins, Archaean ore-
forming fluids were products of deeper crustal thermotectonic events which occurred subsequent to
magmatism and metamorphism in ore-hosting supracrustal rocks. Detailed geochronological studies
now recognize such lower- to mid-crustal, late deformational regimes in Archaean terranes (Jackson
and Cruden, 1995; Kent et al., 1996). Gold deposits in any given Archaean province may all be a part
of the same supercontinent cycle (cf. Barley and Groves, 1992), but can show a wide variation in age
(Table 1), reflecting a variety of thermal events during many tens of millions of years of accretion and
subduction.
2.2.3. Crustal environment of `mesothermal' gold deposition
The majority of deposits of this ore style are sited in ductile to brittle structures, have proximal
alteration assemblages of Fe sulfide-carbonate-sericite ± albite (in rocks of appropriate composition to
stabilise the assemblage) and were deposited at 300 ± 50°C and 1-3 kbar, as indicated by fluid
inclusion and other geothermobarometric studies (Groves and Foster, 1991; Nesbitt, 1991). They are
consistently syn- to post-peak metamorphic and were emplaced at temperatures generally within
100°C of peak metamorphic temperatures experienced by the surrounding host rocks. However,
Page 12
recent studies in mainly Archaean greenstone belts have extended the ranges of temperature and
pressure, and hence extended the inferred crustal range of formation of the deposits into higher- and
lower-grade metamorphic rocks (e.g. the crustal continuum model of Groves, 1993). The evidence for
formation of these gold deposits over P-T ranges of about 180-700°C and < 1-5 kbar (Groves, 1993;
Hagemann and Brown, 1996; Ridley et al., 1996) implies vertically extensive hydrothermal systems
that contrast sharply with other continental-margin gold systems that are apparently restricted to the
upper 5 km or so of crust (Fig. 2).
Studies in the early 1990's, summarized in McCuaig et al. (1993), identified higher P-T examples of
these gold ores in amphibolite facies terranes of Western Australia, the Superior and Slave Provinces
in Canada, India and Brazil. Most such mineralization occurred between 450-600°C and 3-5 kbar. A
few examples in granulite terranes formed at even higher P-T regimes (Barnicoat et al., 1991;
Lapointe and Chown, 1993). The gold ores were still precipitated from the same low salinity, COz-
and 80-rich fluids, but, because of the higher temperatures and different mineral stabilities, there is a
scarcity of carbonate phases and an abundance of calc-silicate minerals characterizing alteration
haloes (Mikucki and Ridley, 1993). Such assemblages are similar to those typifying skarn systems
(Mueller and Groves, 1991).
It is somewhat problematic as to why a similar continuum of gold deposits has not been widely
recognized in higher metamorphic-grade portions of Phanerozoic orogenic belts. Was there
something inherently different between the tectonics of Archaean and Phanerozoic continental
growth? Or do such gold deposits occur in high-grade terrains of the Phanerozoic and they have just
been classified differently? Perhaps a re-evaluation of the classification of some of the gold-bearing
`skarns' or contactmetamorphosed deposits in younger orogenic belts might help to solve this
problem. Ore fluid salinity might be a key discriminator in the case of the skarns, with relatively high
ore-fluid salinities being associated with typical gold skarn deposits that are more directly linked to
intrusive sources (Meinert, 1993). The late Palaeozoic Muruntau deposit in Uzbekistan is apparently
one example of a postArchaean, higher metamorphic grade `mesothermaltype' deposit. The
abundance of thin quartz layering, fluid inclusion data suggesting trapping temperatures in excess of
400°C (Berger et al., 1994) and a skarn-like, calc-silicate assemblage (Marakushev and Khokhlov,
1992) from deeper parts of the ore system all suggest that the deposit may represent a deeper part of
the crustal continuum.
Page 13
Fig. 2. Schematic representation of crustal environments of hydrothermal gold deposits in terms of

depth of formation and structural setting within a convergent plate margin. This figure is by necessity
stylised to show the deposit styles within a depth framework. There is no implication that all deposit
types or depths of formation will be represented in a single ore system. Adapted from Groves (1993),
Gebre-Mariam et al. (1995) and Poulsen (1996).
Ore formation at temperatures of 200-250°C and at crustal depths of only a few kilometers is not
uncharacteristic of these ores where hydrothermal fluids have migrated to shallower crustal levels.
However, a few anomalies from shallow gold systems in the Yilgarn block of Western Australia are
notable. Comb, cockade, crustiform and colloform textures at the Racetrack deposit, deposited from
C02-poor fluids in lower greenschist facies rocks at depths <= 2.5 km, are more like those developed
in classic epithermal vein deposits (Gebre-Mariam et al., 1993). Similar textures at the Wiluna gold
deposits in subgreenschist facies rocks, as well as s18Oquarcz measurements as light as 6-7 per ml,
provide some of the strongest evidence of meteoric water involvement in some of the `mesothermal'
hydrothermal systems (Hagemann et al., 1992, 1994).
Gold solubility relationships at temperatures below 200-250°C best explain the observation that the
continuum of this type of gold deposit does not continue into the uppermost few kilometres of the
crust. The moderately-reducing and only moderately sulphur-rich conditions likely to characterize
`mesothermal' gold ore-fluids at low temperature (Mikucki, 1998 - this issue), would favor low gold
solubilities at these low temperatures (e.g. Shenberger and Barnes, 1989). However, hydrothermal
fluids that have been depositing `mesothermal' gold along crustal-scale fault zones at depth, must still
advect along these faults to the surface. Such is probably the case in the westernmost part of North
America where C02-rich, isotopically heavy fluids migrated to near-surface environments of very low
P-T in the Cordilleran orogen. Cinnabar ± stibnitebearing epithermal, silica-carbonate veins, which
were deposited within the upper few kilometres of crust, define such flow (Nesbitt and
Muehlenbachs, 1989). Examples include the Hg-Sb deposits of the Kuskokwim basin in SW Alaska,
the Pinchi belt of British Columbia and the coast ranges of northern California. In fact, it has been
recognized now for thirty years that many of the thermal springs within the accreted margin of
western North America have a unique chemical character (White, 1967) and could be the surface
expression of deeper `mesothermal' gold deposits. 81809„artz values for Hg-rich veins emplaced in
the near surface are as heavy as + 30 per ml because of greater quartz-water fractionation, as
Page 14
temperatures of ore fluids cooled to as low 150°C. Such heavy oxygen values are very distinct from
818Oq„artz values of other types of vein systems deposited in classical epithermal environments,
such as those of the Nevada Basin and Range (Goldfarb et al., 1990). The identification of this type of
Hg-Sb epithermal system in a continental margin terrane with limited erosion may be a valuable
guide to the down-dip existence of a so-called `mesothermal' gold occurrence.
2.2.4. Comparisons with other lode-gold deposit types
Most deposit types that contain ore-grade gold (Table 2), whether with gold as the principal metal or
together with copper, are sited along convergent plate margins (Sawkins, 1990). There are notable
exceptions, such as gold-rich volcanogenic massive sulfide deposits developed along spreading ocean
ridges (e.g. Bousquet) and other deposit styles associated with possible anorogenic hot spots (e.g.
Olympic Dam). However, as a rule, many of the Phanerozoic gold-bearing epithermal vein, Carlin-
type sedimentary rock-hosted and porphyry/skarn deposits developed within the same active
continental margins as the so-called `mesothermal' deposits (Fig. 1). Notable distinctions, however,
can be made that relate to local changes in tectonism within a developing orogen and to crustal depth
range (a reflection of regional geothermal gradient) of the auriferous hydrothermal systems.
As shown schematically in Fig. 1, a significant proportion of epithermal and porphyry deposits are
distinct in that they form above subduction zones distal to continental margins or within continental
margins, but during post-collisional extension. Many other gold-rich epithermal and porphyry
systems develop in oceanic regimes within the top few kilometres of crust of volcano-plutonic island
arcs located above intermediate- to steeply-dipping subduction zones (e.g. Sawkins, 1990; Sillitoe,
1991), with a vertical transition from porphyry-style to classic epithermal vein-style mineralization
(e.g. White and Hedenquist, 1995). Other epithermal lodes, including some of the world-class
deposits (Muller and Groves, 1997), are associated with alkalic, mantle-related rocks that reflect
extensional episodes in a convergent orogen in either a near-arc region (e.g. Porgera: Richards et al.,
1990) or far inland of the accretionary wedge (e.g. Cripple Creek: Kelley et al., 1996). Certainly,
many of the well-studied Tertiary epithermal ores associated with volcanic rocks throughout Nevada
are products of post-orogenic Basin and Range extension. Geochronological evidence is beginning to
favour a similar temporal setting for Carlin-type mineralization (Hofstra, 1995; Emsboo et al., 1996).
The gold-bearing epithermal vein and porphyry systems that are, however, associated with
collisional, subduction-related tectonics (Sillitoe, 1993) are typically located in different crustal
regimes in the orogen than the so-called `mesothermal' gold systems. Whether in an island arc,
compressional orogen, or a zone of back-arc rifting, the porphyryskarn-epithermal vein continuum
normally is telescoped into the upper 2-5 km of crust (Figs. 1 and 2; Poulsen, 1996). Magmatism
(generally I-type) and high temperatures impose a very steep geothermal gradient on the upper crust,
often locally far in excess of 100°C/km. An abundance of subvolcanic to volcanic rocks necessitates
that much of the gold ore is hosted in lithologies of roughly equivalent age. The shallow level of the
hydrothermal activity restricts much of the lode-gold emplacement to rocks that are
unmetamorphosed to only slightly regionally metamorphosed.
Page 15
Table 2: Characteristics of epigenetic gold deposits. Summarized from Foster (1991), Robert et al. (1991), Kirkham et al. (1993), Hedenquist and Lowenstern (1994),
Richards (1995) and Poulsen (1996)
Deposit Examples Tectonic setting Temp. of Depth of Ore fluid Au:Ag Alteration types Other key features
type formation emplacement composition
(°C) (km)
Orogenic Kalgoorlie (Australia), continental margin; 200-700 2-20 3-10 eq. wt% 1-10 carbonation, hosted in deformed metamorphic
Val d'Or (Canada), compressional to NaCl, z 5 sericitization, terranes; <- 3-5% sulfide
Ashanti (Ghana), transpressional mol% COz; sulfidation; skam- minerals; individual deposits of
Mother lode (USA) regime; veins typically in traces of CH4 like assemblages in Z 1-2 km vertical extent; spatial
metamorphic rocks on and Nz higher temperature association with transcrustal fault
seaward side of deposits zones and granitic magmafsm
continental are
Epithermal high sulf. = Goldfield oceanic arc, continental 100-300 surface- < 1-20 eq. wt% 0.02-1 adularia-sericite- veins and replacements are similar
(low and high (USA), Summitville arc, or back arc 2 km NaCl; quartz (low sulf.) age as ore-hosting or nearby
sulfidation) (USA), Julcani (Peru), extension of continental early acidic versus quartz-alunite- volcanic rocks; ore zones
Lepanto (Philippines); crust; extensional condensate (high kaolinite (high generally 100-500 m in vertical
low sulf. = Comstock environments normal, sulf.) sulf.) extent; disseminated ore common
Lode (USA), Fresnillo but commonly in in high sulf. systems
(Mexico), Golden compressional regimes
Cross (New Zealand)
Epithermal Cripple Creek (USA);- post-subduction, back generally surface- <_ 10 eq. wt% very carbonation, K- Te-rich deposits associated with
(alkalic- Porgera (PNG); arc extension; extension 5 200 2 km NaCl high CO2; variable metasomatism, alkalic igneous rocks; ores
related) Emperor, Fiji can be adjacent to traces of CH Q and propylitic commonly in breccia pipes and as
magmatic arc or Nz assemblages manto-type replacements
hundreds of km
landward
Sedimentary- Carlin (USA), Jerritt back-arc extension and 200-300 2-3 <- 7 eq. wt% 0.1-10 intense very fine-grained gold in intensely
rock hosted Canyon (USA), thinning of continental NaCl; silicification; some silicified rock; dissolution of
Guizhou (PR China) crust kaolinization surrounding carbonate
Page 16
Deposit Examples Tectonic setting Temp. of Depth of Ore fluid Au:Ag Alteration types Other key features
type formation emplacement composition
(°C) (km)
Orogenic Gold-rich Bingham (USA), oceanic or continental 300-700 2-5 some fluids > 35 0.001-0.1 central biotite-KF disseminated sulfides and veinlets
porphyry Grasberg (Indonesia), arc; subduction-related eq. wt% zone surrounded within and adjacent to porphyritic,
Lepanto-Far Southeast but often associated NaCl; can mix by quartz-chlorite; silitic-to intermediate composition
(Philippines), with extensional with low salinity common sericite- intrusions; low oxidation state of
Kingking (Philippines) environments surface waters; pyrite overprinting; magmas may favor gold
often immiscible distal propylitic enrichments; generally 1-type
vapor alteration magmas; gold introduced with Cu
sulphides
Gold-rich Hedley (Canada), oceanic or continental 300-600 1-5 10 to > 35 eq. < 1-10 garnet-pyroxene- most occur as calcic exoskarus;
skaru Fortitude (USA), arc; subduction-related wt% NaCl epidote-chlorite- typically associated with mafic,
Crown Jewel (USA) but often associated calcite low-silica, very reduced plutons
with extensional
environments
Submarine Home (Canada), back-arc rift basins _< 350 on or near 3.5-6.5 eq. 0.0001- quartz-talc-chlorite laminated, banded, or massive
exhalative Bousquet (Canada), (Kuroko-type) or mid- seafloor wt% NaCl; 0.1 is most common fine-grained sulphides; commonly
Greens Creek (USA), ocean seafloor much higher with an outer zone both exhalative and
Boliden (Sweden) spreading (Cyprus- and salinities where of illite t smectite; synsedimentary replacement
Besshi-type) fluid interaction anhydrite or barite textures; gold relatively more
with brines cap in places important in back-arc regions
Page 17
In contrast, the so-called `mesothermal' ore deposit type is deposited over a very broad range of the
upper crust (Groves, 1993; Poulsen, 1996). Rather than bringing a concentrated heat source to the
near surface, the fluids, granitic magmas and heat are carried to higher crustal levels along major fault
zones that may have been suture zones between accreted terranes. Crustal geotherms of perhaps >=
30°C/km are elevated, but not to the levels of the more telescoped group of ore deposit types. Where
hydrothermal fluids reach the near-surface environment, their relatively low temperature hinders
significant gold transport; however, bisulphide complexes still may carry significant Sb and Hg into
the upper few kilometres of crust (Fig. 2). Where such fluids migrate into the realm of the typical
porphyryskarn-epithermal continuum, complex overlapping of deposit styles may develop. Such a
situation may characterize southwestern Alaska, where epithermal Hg-Sb ores that suggest so-called
`mesothermal' gold deposits at depth (Gray et al., 1997) are spatially associated with volcano-
plutonic-related gold deposits (Bundtzen and Miller, 1997), or northern California where the
McLaughlin gold deposit sits among a series of Hg-rich hot springs (Sherlock and Logan, 1995).
2.3. Problem of Nomeclature
Prior to 1980, the so called `mesothermal' group of Archaean through Tertiary deposits was not
widely recognized as a single special type of gold ore. Most classifications scattered the deposits
among the mesothermal and hypothermal regimes of Lindgren (1933). Others, such as Bateman
(1950), divided these deposits into groups within a very broad `cavity filling' type of epigenetic ore
deposit. Hence, many Archaean lodes were classified as fissure filling type deposits, Otago was a
shear zone deposit type, Bendigo was a saddle reef deposit type, Treadwell, Alaska was a stockwork
type deposit, etc. The relatively low price of gold correlated with a limited research interest in gold
genesis studies. In fact, in the 75th Anniversary Volume of Economic Geology (1981), there is
notably no chapter that is devoted to this economically important ore deposit type. Economic
geologists had begun to notice the basic association of the Phanerozoic deposits with subduction
zones and convergent margins during the growth of plate tectonic theories. However, books on
tectonics and ore deposits barely mentioned these gold systems (e.g. Mitchell and Garson, 1981).
As the price of gold increased dramatically in the late 1970's, so did interest in the understanding of
these gold deposits. `Mesothermal' lode-gold deposits began to receive extensive study by ore
geologists, and were subsequently described by a variety of terms during the last fifteen years as
workers began recognizing them as a single mineral deposit type. The abundance of terms that define
these ores reflects both the great expansion of knowledge about these systems accumulated during the
1980's (e.g. Robert et al., 1991) and the efforts by various groups to establish ore deposit model
volumes that classify deposits by type (e.g. Cox and Singer, 1986). One consequence of so many
terms for the same deposits is the resulting confusion for those not extremely familiar with the gold
literature. Certainly, a single deposit type title would be helpful for all workers.
The paper by Nesbitt et al. (1986) on lode-gold deposits of the Canadian Cordillera seemed to initiate
popularity of the phrase mesothermal. They define a group of Canadian `mesothermal' gold deposits
that formed between 200-350°C within a series of accreted terranes. Prior to this paper, the broad
class of `mesothermal' gold deposits did not exist. Major gold volumes such as `Gold '82' (Foster,
1984), `Turbidite-hosted Gold Deposits' (Keppie et al., 1986) and `Gold '86' (Macdonald, 1986)
lacked any mention of such a deposit type. However, since the Nesbitt et al. (1986) paper, the
`mesothermal' terminology has become well-entrenched in the literature. This may be a response, in
part, to the need to easily contrast this group of gold deposits with the generally more shallowly
deposited types of gold ores that had already been classified as èpithermal' for many years previous.
Because of this widespread acceptance of the mesothermal label, subsequent comprehensive
descriptions of these gold deposits have tended to group them under such a `mesothermal heading'
(Groves et al., 1989; Kerrich, 1991; Hodgson, 1993).
Whereas `mesothermal' has become the most common term used in referring to this type of deposit
during the last ten years, Poulsen (1996) has recently shown how it is very inconsistent with the
Page 18
meaning originally proposed by Lindgren (1907, 1933). Lindgren's description of such a deposit type
is for that which formed at depths of about 1,2003,600 m and at temperatures of 200-300°C. Because
of the restrictive temperature range, high-temperature alteration phases, including tourmaline, biotite,
hornblende, pyroxene and garnet, were stated as being absent in and surrounding mesothermal type
ores. Gold districts such as those of the California foothills belt, the Meguma domain of Nova Scotia,
central Victoria, and Charters Tower in Queensland were classified by Lindgren (1933) as
mesothermal.
Many other gold districts, however, that are routinely classified as `mesothermal' today were actually
termed `hypothermal' by Lindgren (1933). These deposits were described as having formed at
300500°C, thus exhibiting higher temperature alteration assemblages, and at depths below 3,600 m.
Most of the world's Archaean gold deposits were clearly stated as being hypothermal deposits. In
addition, some Phanerozoic lodes, including those of the Bohemian Massif and Juneau, Alaska, were
included in the class. The groupings into the mesothermal and hypothermal temperature ranges by
Lindgren are remarkably accurate in light of many modern fluid inclusion studies. But the key point
is that many of the deposits that are now termed `mesothermal' did not fit in the mesothermal category
in the early 20th century and still do not fit in the category today.
If one such Lindgren-type term was used to define the broad observed range for P-T conditions of
these deposits, it probably is `xenothermal'. The term, coined by Buddington (1935), covers the P-T
conditions from lepothermal (a vague P-T regime between epithermal and mesothermal) to
hypothermal. As such, it would include the broad range of ore forming pressures and temperatures
that is welldocumented in the Yilgarn block of Western Australia, as summarised by Groves (1993).
However, other factors, such as structural control, wall rock type and fluid chemistry play a major
role in the localization of a gold deposit and definition of a gold deposit type solely on P-T
environment is not advisable (Bateman, 1950).
The contrasting tectonic setting between the sites of most èpithermal' gold deposits and the sites of
all so-called `mesothermal' deposits presents another basic problem with usage of the Lindgren
model. As envisioned by Lindgren (1907, 1933), the epithermal, mesothermal and hypothermal terms
were intended to define a continuum among deposits. However, as implied in Fig. 2, the term
èpithermal' is now entrenched in the literature as a specific mineral-deposit type that most commonly
describes high-level veining and alteration broadly associated with volcanism or subvolcanic
magmatism (e.g. Berger and Bethke, 1985). As discussed above, such epithermal gold deposits may
form in oceanic arcs long before continental margin orogenesis or, as in the Basin and Range of the
USA, during post-orogenic extension, as shown schematically in Fig. 1. Hence, there are typically
neither consistent spatial nor temporal relations between the two gold deposit types.
Many other terms relating to host rocks, vein mineralogy or ore-fluid chemistry are equally
unacceptable in the overall description of these deposits. Commonly used terms, such as `greenstone
gold', `slate belt gold', or `turbidite-hosted gold', disguise the fact that the deposits have many
similarities despite their different hosting sequences (the theme of this special Ore Geology Reviews
issue) and should be used, if at all, to describe subgroups of the major deposit type, and not the
deposit type itself. The use of 'Archaean' or `Mother lode-type' gold deposits is also unacceptable,
clearly reflecting a specific temporal or spatial preference, respectively. `Metamorphic gold' implies
an understanding of the ore-forming process which is, however, still strongly under debate. The fact
that these deposits contain only a few percent sulfide minerals, in most cases, has led to classifications
referring to them as `low sulfide' (Berger, 1986), and the fact that gold is enriched by orders of
magnitudes over base metals and Au:Ag ratios are generally > 1 has led to their classification as `gold
only' (Hodgson and MacGeehan, 1982; Phillips and Powell, 1993) deposits. However, many other
types of gold deposits, including the sedimentary rock-hosted ores at Carlin and elsewhere in Nevada,
show the same low sulfide
content. Similarly, `lode-gold' (McCuaig and Kerrich, 1994) may be interpreted to contain a variety
of gold deposit types.
Page 19
A critical feature of all these deposits seems to be their common tectonic setting, as described in detail
above. These deposits were classified as `pre-orogenic' by Bache (1980, 1987), who recognized their
association with the world's orogenic belts. However, at the same time, the classification assumed a
syngenetic exhalative origin for the auriferous lodes, an assumption clearly in conflict with modern
geochronological data. Goldfarb et al. (1991a, 1998 this issue) have often preferred the term
`synorogenic', given the clear overlap of gold-forming events in the North American Cordillera with a
broad, 120-m.y.-long period of continental margin growth. The term `post-orogenic' has been used by
other workers (Gebre-Mariam et al., 1993; Groves, 1996) who emphasize that deformation and
metamorphism of ore host rocks commonly predate hydrothermal vein emplacement (Groves et al.,
1984; Colvine, 1989; Hodgson and Hamilton, 1989).
2.4. Proposed Classification
These gold deposits, throughout the world's collisional orogenic belts, can actually be viewed as both
syn- and post-orogenic in origin. Whereas host rocks for ore may already be undergoing uplift and
cooling (thus `post-orogenic'), the ore-forming fluids may be generated or set in motion by
simultaneous thermal processes at depth (thus `syn-orogenic') as described by Stuwe et al. (1993). For
example, Kent et al. (1996) show that the main episode of gold mineralization in the Yilgarn craton
postdates thermal events in the ore-hosting upper crust, but temporally correlates with melting and
magmatism of lower-middle Archaean crust. Because of this, it is suggested that the gold ores simply
be classified as òrogenic' lode types, as was originally suggested by Bohlke (1982).
A remaining problem is whether to classify many ìntrusion-related gold deposits' within this group
of orogenic gold deposits. Sillitoe (1991) places deposits such as Muruntau and Charters Tower in
such an intrusion-related deposit type. McCoy et al. (1997) distinguish `plutonic-related mesothermal
gold deposits' of interior Alaska, such as Fort Knox, as those where ore fluids are derived from
evolving magmas. The Proterozoic gold lodes of northern Australia and the Mesozoic deposits of the
north China craton and Korea are also commonly suggested to be genetically associated with igneous
processes. Are such deposits, with ore fluid chemistries essentially identical to those of typical
orogenic gold deposits, a different deposit type? Sillitoe (1991) indicated that the intrusion-related
gold deposits also form in Phanerozoic convergent plate margins above zones of active subduction,
although regional extension is stressed as an important characteristic and thus indicates some
difference from the orogenic class defined here. Sillitoe (1991) does stress that the apparent overlap
between orogenic and intrusion-related gold systems requires further attention. We would certainly
agree.
A convenient terminology that both retains the prefixes èpi', `meso', and `hypo' used by Lindgren
(1907, 1933), and subdivides the orogenic gold deposit type, is introduced by Hagemann and Ridley
(1993) and then further modified by Gebre-Mariam et al. (1995). Its continued usage is
recommended. In such a scenario, epizonal deposits form within 6 km of the surface at temperatures
of 150-300°C, mesozonal deposits form at depths of 6-12 km and at temperatures of 300-475°C and
hypozonal deposits form below 12 km and at temperatures exceeding 475°C. It is critical to note that
this terminology has been defined solely as a subdivision for orogenic gold deposits based on many
modern geothermobarometric studies. Because of this, the depth zones for these orogenic subclasses
do not correspond to those in Lindgren's epithermal, mesothermal, and hypothermal regimes.
Page 20
The authors acknowledge the input of past and present staff and students at the Key Centre at UWA,
particularly Mark Barley, Kevin Cassidy and John Ridley. The research was funded largely by
mining companies and supported by Key Centre Corporate Members, DEETYA, AMIRA,
MERIWA and UWA. The paper was inspired as a result of a course given by F. Robert in Perth in
February, 1996, and conferences on mesothermal gold deposits in Ballarat and Perth in July, 1996.
The encouragement of Ross Ramsay is greatly appreciated. This manuscript was much improved
through the exceptionally insightful comments of Kevin Cassidy, Rob Kerrich, Howard Poulsen, Ed
Mikucki and one anonymous journal reviewer.
2.6. References
Abraham, A.P.G., Spooner, E.T.C., 1995. Late Archean regional deformation and structural controls
on gold-quartz vein mineralization in the northwestern Slave province, N.W.T., Canada. Can. J.
Earth Sci. 32, 1132-1154.
Ansdell, K.M., Kyser, T.K., 1992. Mesothermal gold mineralization in a Proterozoic greenstone belt,
Western Flin Flon domain, Saskatchewan, Canada. Econ. Geol. 87, 1496-1524.
Ame, D.C., Bierlein, F.P., McNaughton, N.J., Wilson, C.J.L., Morard, V.J., Ramsay, W.R.H., 1996.
Timing of felsic magmatism in Victoria and its relationship to gold mineralization. In: Hughes,
M.J., Ho, S.E., Hughes, C.E. (Eds.), Recent Developments in Victorian Geology and
Mineralisation. Aust. Inst. Geosci. Bull. 20, 43-48.
Ash, C.H., Reynolds, P.H., Macdonald, R.W.J., 1996. Mesothermal gold-quartz vein deposits in
British Columbia oceanic terranes. New mineral deposit models of the Cordillera. B.C. Geol.
Surv., Cordilleran Roundup Short Course, pp. 01-032.
Ashley, P.M., Cook, N.D.J., Hill, R.L., Kent, A.J.R., 1994. Shoshonitic lamprophyre dykes and their
relation to mesothermal Au-Sb veins at Hillgrove, New South Wales, Australia. Lithos 32, 249-
272.
Bache, J.J., 1980. Essai de typologie quantitative des gisements mondiaux d'or. BRGM Note
SGN/GMX/GIT, No. 622. 9 PP.
Bache, J.J., 1987. World Gold Deposits: A Quantitative Classification. North Oxford Academic,
London, 179 pp.
Balakrishnan, S., Hanson, G.N., Rajamani, V., 1990. Ph and Nd isotope constraints on the origin of
high Mg and tholeftic amphibolites, Kolar Schist Belt, South India. Contrib. Mineral. Petrol. 107,
279-292.
Barley, M.E., Groves, D.I., 1992. Supercontinent cycles and the distribution of metal deposits
through time. Geology 20, 291294.
Barley, M.E., Eisenlohr, B.N., Groves, D.I., Perring, C.S., Vearncombe, J.R., 1989. Late Archean
convergent margin tectonics and gold mineralization: A new look at the Norseman-Wiluna belt,
Western Australia. Geology 17, 826-829.
Barnicot, A.C., Fare, R.J., Groves, D.I., McNaughton, N.J., 1991. Synmetamorphic lode-gold
deposits in high-grade Archean settings. Geology 19, 921-924.
Bateman, A.M., 1950. Economic Mineral Deposits, 2nd ed. Wiley, New York, 916 pp.
Berger, B.R., 1986. Descriptive model of low-sulphide Au-quartz veins. In: Cox, D.P., Singer, D.A.
(Eds.). Mineral Deposit Models. U.S. Geol. Surv. Bull. 1693, 183-186.
Berger, B.R., Bethke, P.M. (Eds.), 1985. Geology and geochemistry of epithermal deposits. Rev.
Econ. Geol. 2, 298 pp.
Berger, B.R., Drew, L.D., Goldfarb, R.J., Snee, L.W., 1994. An epoch of gold riches: The Late
Paleozoic in Uzbekistan, central Asia. Soc. Econ. Geol. Newslett. 16 (1), 7-11.
Von Blanckenburg, F., Davies, J.H., 1995. Slab breakoff: A model for syncollisional magmatism and
tectonics in the Alps. Tectonics 14, 120-131.
Page 21
Bohlke, J.K., 1982. Orogenic (metamorphic-hosted) gold-quartz veins. U.S. Geol. Surv., Open-file
Rep. 795, 70-76.
Bohlke, J.K., Kistler, R.W., 1986. Rb-Sr, K-Ar and stable isotope evidence for ages and sources of
fluid components of gold-bearing quartz veins in the northern Sierra Nevada foothills
metamorphic belt, California. Econ. Geol. 81, 296322.
Bouchot, V., Gros, Y., Bonnemaison, M., 1989. Structural controls on the auriferous shear zones of
the Saint Yrieix district, Massif central, France. Evidence from the Le Bourneix and Laurieras
gold deposits. Econ. Geol. 84, 1315-1327.
Buddington, A.F., 1935. High temperature mineral associations at shallow to moderate depths. Econ.
Geol. 30, 205-222.
Bundtzen, T.K., Miller, M.L., 1997. Precious metals associated with Late Cretaceous-Early Tertiary
igneous rocks of southwestern Alaska. In: Goldfarb, R.J., Miller, L.D. (Eds.), Mineral Deposits of
Alaska. . Econ. Geol. Monogr. 9, 242-286.
Cathelineau, M., Boiron, M.C., Hollinger, P., Poty, B., 1990. Metallogenesis of the French part of the
Variscan orogen. Part II: Time-space relationships between U, Au and Sn-W ore deposition and
geodynamic events. Mineralogical and U-Pb data. Tectonophysics 177, 59-79.
Colvine, A.C., 1989. An empirical model for the formation of Archean gold deposits. Products of
final cratonization of the Superior Province, Canada. In: Keays, R.R., Ramsay, W.R.H., Groves,
D.I. (Eds.), The Geology of Gold Deposits: The Perspective in 1988. Econ. Geol. Monogr., 6, 37-
53.
Colvine, A.C., Andrews, A.J., Cherry, M.E., Durocher, M.E., Fyon, J.A., Lavigne, M.J., Macdonald,
A.J., Marmont, S., Poulsen, K.H., Springer, J.S., Troop, D.G., 1984. An integrated model for the
origin of Archean lode-gold deposits. Ont. Geol. Surv. Open-File Rep. 5524, 98 pp.
Coney, P.J., 1992. The Lachlan belt of eastern Australia and Circum-Pacific tectonic evolution.
Tectonophysics 214, 1-25.
Conners, K.A., 1996. Unraveling the boundary between turbidites of the Kisseynew belt and volcano-
plutonic rocks of the Flin Flon belt, Trans-Hudson Orogen, Canada. Can. J. Earth Sci. 33, 811-
829.
Cox, D.P., Singer, D.A. (Eds.), 1986. Mineral Deposit Models. U.S. Geol. Surv. Bull. 1693, 379 pp.
Curti, E., 1987. Lead and oxygen isotope evidence for the origin of the Monte Rosa gold lode
deposits (Western Alps, Italy): A comparison with Archean lode deposits. Econ. Geol. 82, 2115-
2140.
Darbyshire, D.P.F., Pitfield, P.E.J., Campbell, S.D.G., 1996. Late Archean and Early Proterozoic
gold-tungsten mineralization in the Zimbabwe Archean craton: Rb-Sr and Sm-Nd isotope
constraints. Geology 24, 19-22.
deRonde, C.E.J., Hall, C.M., York, D., Spooner, E.T.C., 1991. Laser step heating 40Ar/39Ar age
spectra from early Archean Barberton greenstone belt sediments: A technique for detecting cryptic
tectono-thermal events. Geochim. Cosmochim. Acta 55, 1933-1951.
Drew, L.J., Berger, B.R., Kurbanov, N.K., 1996. Geology and structural evolution of the Muruntau
gold deposit, Kyzylkum desert, Uzbekistan. Ore Geol. Rev. 11, 175-196. Economic Geology,
1981. 75th Anniversary Volume. Econ. Geol. Publ. Co., Littleton, CO, 964 pp.
Elder, D., Cashman, S.M., 1992. Tectonic control and fluid evolution in the Quartz Hill, California,
lode-gold deposits. Econ. Geol. 87, 1795-1812.
Emsboo, P., Hofstra, A.H., Park, D., Zimmerman, J.M., Snee, L., 1996. A mid-Tertiary age constraint
on alteration and mineralization in igneous dikes on the Goldstrike property, Carlin Trend,
Nevada. Geol. Soc. Am. Abstr. Prog. 28 (7), 476.
Fayek, M., Kyser, T.K., 1995. Characteristics of auriferous and barren fluids associated with the
Proterozoic Contact Lake lode-gold deposit, Saskatchewan, Canada. Econ. Geol. 90, 385-406.
Ford, R.C., Snee, L.W., 1996. 4°Ar/39Ar thermochronology of white mica from the Nome district,
Alaska: The first ages of lode sources to placer gold deposits in the Seward Peninsula. Econ. Geol.
91, 213-220.
Foster, D.A., Kwak, T.A.P., Gray, D.R., 1996. Timing of gold mineralisation and relationship to
metamorphism, thrusting, and plutonism in Victoria. In: Hughes, M.J., Ho, S.E., Hughes, C.E.
(Eds.), Recent Developments in Victorian Geology and Mineralisation. Aust. Inst. Geosci. Bull.
20, 49-53.
Page 22
Foster, R.P. (Ed.), 1984. Gold '82: The Geology, Geochemistry and Genesis of Gold Deposits. A.A.
Balkema, Rotterdam, 753 PP.
Foster, R.P. (Ed.), 1991. Gold Metallogeny and Exploration. Blackie and Son, Glasgow, 432 pp.
Foster, R.P., Piper, D.P., 1993. Archaean lode-gold deposits in Africa: Crustal setting, metallogenesis
and cratonization. Ore Geol. Rev. 8, 303-347.
Fyfe, W.S., Kerrich, R., 1985. Fluids and thrusting. Chem. Geol. 49, 353-362.
Gaboury, D., Dube, B., Lafleche, M.R., Lauziere, K., 1996. Geology of the Hammer Down
mesothermal gold deposit, Newfoundland Appalachians, Canada. Can. J. Earth Sci. 33, 335-350.
Gebre-Mariam, M., Groves, D.I., McNaughton, N.J., Mikucki, E.J., Vearncombe, J.R., 1993.
Archaean Au-Ag mineralisation at Racetrack, near Kalgoorlie, Western Australia: A high crustal-
level expression of the Archaean composite lode-gold system. Miner. Deposita 28, 375-387.
Gebre-Mariam, M., Hagemann, S.G., Groves, D.I., 1995. A classification scheme for epigenetic
Archaean lode-gold deposits. Miner. Deposita 30, 408-410.
Goldfarb, R.J., Leach, D.L., Pickthorn, W.J., Paterson, C.J., 1988. Origin of lode-gold deposits of the
Juneau gold belt, southeastern Alaska. Geology 16, 440-443.
Goldfarb, R.J., Gray, J.E., Pickthorn, W.J., Gent, C.A., Cieutat, B.A., 1990. Stable isotope
systematics of epithermal mercury-antimony mineralization, southwestern Alaska. In: Goldfarb,
R.J., Nash, J.T., Stoesser, J.W. (Eds.), Geochemical Studies in Alaska by the U.S. Geological
Survey, 1989. U.S. Geol. Surv. Bull., 1950, pp. El-E9.
Goldfarb, R.J., Leach, D.L., Pickthorn, W.J., 1991a. Source of synorogenic fluids of the northern
Cordillera: Evidence from the Juneau Gold Belt, Alaska. In: Robert, F., Sheahan, P.A., Green,
S.B. (Eds.), Greenstone Gold and Crustal Evolution. Geol. Assoc. Can., Mineral Deposits Div.
Publ., pp. 160-161.
Goldfarb, R.J., Snee, L.W., Miller, L.D., Newberry, R.J., 1991b. Rapid dewatering of the crust
deduced from ages of mesothermal gold deposits. Nature 354, 296-298.
Goldfarb, R.J., Phillips, G.N., Nokleberg, W.J., 1998. Tectonic setting of synorogenic gold deposits
of the Pacific Rim. Ore Geol. Rev. 13, 185-218 (this issue).
Gray, J.D., Gent, C.A., Snee, L.W., Wilson, F.H., 1997. Epithermal mercury-antimony and gold-
bearing vein lodes of southwestern Alaska. In: Goldfarb, R.J., Miller, L.D. (Eds.), Mineral
Deposits of Alaska. Econ. Geol. Monogr. 9, 287-305.
Groves, D.I., 1993. The crustal continuum model for late-Archaean lode-gold deposits of the Yilgarn
Block, Western Australia. Miner. Deposita 28, 366-374.
Groves, D.I., 1996. Geological concepts in the exploration for large to giant late-orogenic
(mesothermal) gold deposits: The Archaean greenstone experience. In: Mesothermal Gold
Deposits: A Global Overview. Geol. Dept. (Key Centre) Univ. Ext., Univ. West. Aust. Publ. 27,
114-117.
Groves, D.I., Foster, R.P., 1991. Archean lode-gold deposits. In: Foster, R.P. (Ed.), Gold Metallogeny
and Exploration. Blackie and Son, Glasgow, pp. 63-103.
Groves, D.I., Phillips, G.N., Ho, S.E., Henderson, C.A., Clark, M.E., Woad, G.M., 1984. Controls on
distribution of Archaean hydrothermal gold deposits in Western Australia. In: Foster, R.P. (Ed.),
Gold '82: The Geology, Geochemistry and Genesis of Gold Deposits. A.A. Balkema, Rotterdam,
pp. 689-712.
Groves, D.I., Barley, M.E., Ho, S., 1989. Nature, genesis and tectonic setting of mesothermal gold
mineralization in the Yilgarn Block, Western Australia. In: Keays, R.R., Ramsay, W.R.H.,
Groves, D.I. (Eds.), The Geology of Gold Deposits: The Perspective in 1988. Econ Geol.
Monogr., 6, 71-85.
Groves, D.I., Barley, M.E., Barnicoat, A.C., Cassidy, K.F., Fare, R.J., Hagemann, S.G., Ho, S.E.,
Hronsky, J.M.A., Mikucki, E.J., Mueller, A.G., McNaughton, N.J., Perring, C.S., Ridley, J.R.,
Vearncombe, J.R., 1992. Sub-greenschist to granulitehosted Archaean lode-gold deposits of the
Yilgarn Craton: A depositional continuum from deep sourced hydrothermal fluids in crustal-scale
plumbing systems. Geol. Dept. (Key Centre) Univ. Ext., Univ. West. Aust. Publ. 22, 325-338.
Le Guen, M., Lescuyer, J.L., Marcoux, E., 1992. Lead-isotope evidence for a Hercynian origin of the
Salsigne gold deposit (Southern Massif Central France). Miner. Deposita 27, 129136.
Guild, P.W., 1971. Metallogeny: A key to exploration. Mining Eng. 23 (1), 69-72.
Page 23
Haeussler, P.J., Bradley, D., Goldfarb, R., Snee, L., Taylor, C., 1995. Link between ridge subduction
and gold mineralization in southern Alaska. Geology 23, 995-998.
Hagemann, S.G., Brown, P.E., 1996. Geobarometry in Archean lode-gold deposits. Eur. J. Mineral.
8, 937-960.
Hagemann, S.G., Ridley, JR., 1993. Hydrothermal fluids in epiand katazonal crustal levels in the
Archaean-Implications for P-T-X-t evolution of lode-gold mineralisation. In: Williams, P.R.,
Haldane, J.A. (Eds.), An international conference on crustal evolution, metallogeny and
exploration of the Eastern Goldfields. Extended Abstr. Aust. Geol. Surv. Org., Record 54, pp.
123-130.
Hagemann, S.G., Groves, D.I., Ridley, J.R., Vearncome, JR., 1992. The Archaean lode-gold deposits
at Wiluna, Western Australia. High level brittle-style mineralisation in a strike-slip regime. Econ.
Geol. 87, 1022-1053.
Hagemann, S.G., Gebre-Mariam, M., Groves, D.L., 1994. Surface-water influx in shallow-level
Archean lode-gold deposits in Western Australia. Geology 22, 1067-1070.
Hedenquist, JW., Lowenstern, J.B., 1994. The role of magmas in the formation of hydrothermal ore
deposits. Nature 370, 519527.
Hirdes, W., Davis, D.W., Ludtke, G., Konan, G., 1996. Two generations of Birimian
(Paleoproterozoic) volcanic belts in northeastern Cote d'Ivorie (West Africa): Consequences for
the `Birimian controversy'. Precambrian Res. 80, 173-191.
Hodgson, C.J., 1993. Mesothermal lode-gold deposits. In: Kirkham, R.V., Sinclair, W.D., Thorpe,
R.I., Duke, J.M. (Eds.), Mineral Deposit Modeling. Geol. Assoc. Can., Spec. Pap., 40, 635-678.
Hodgson, C.J., Hamilton, J.V., 1989. Gold mineralization in the Abitibi greenstone belt: End stage
result of Archaean collisional tectonics? In: Keays, R.R., Ramsay, W.R.H., Groves, D.I. (Eds.),
The Geology of Gold Deposits: The Perspective in 1988. Econ. Geol. Monogr., 6, 86-100.
Hodgson, C.J., MacGeehan, P.J., 1982. A review of the geological characteristics of `Gold Only'
deposits in the Superior Province of the Canadian Shield. In: Hodder, R.W., Petruks, W. (Eds.),
Geology of Canadian Gold Deposits. Can. Inst. Min. Metall., Spec. Vol. 24, 211-229.
Hofstra, A.H., 1995. Timing and duration of Carlin-type gold mineralization in Nevada and Utah:
Relation to back-arc extension and magmatism. Geol. Soc. Am. Abstr. Progr. 27 (6), 329.
Jackson, S.L., Cruden, A.R., 1995. Formation of the Abitibi greenstone belt by arc-trench migration.
Geology 23, 471-474.
Kelley, K.D., Romberger, S.B., Beaty, D.W., Snee, L.W., Stein, H.J., Thompson, R.B., 1996. Genetic
model for the Cripple Creek district: Constraints from 4°Ar/39Ar geochronology, major and trace
element geochemistry and stable and radiogenic isotope data. In: Thompson, T.B. (Ed.),
Diamonds to Gold. I. State Line Kimberlite district, Colorado. II. Cresson mine, Cripple Creek
district, Colorado. Soc. Econ. Geol., Guidebook Ser. 26, 65-83.
Kent, A.J.R., Hagemann, S.G., 1996. Constraints on the timing of lode-gold mineralisation in the
Wiluna greenstone belt, Yilgarn Craton, Western Australia. Aus. J. Earth Sci. 43, 573-588.
Kent, A.J.R., McDougall, L, 1995. Constraints on the timing of gold mineralization in the Kalgoorlie
goldfield, Western Australia, from 4°Ar/39Ar and U-Pb dating: Evidence for multiple
mineralization episodes. Econ. Geol. 90, 845-859.
Kent, A.J.R., Cassidy, K.F., Fanning, C.M., 1996. Archean gold mineralization synchronous with the
final stages of cratonization, Yilgarn Craton, Western Australia. Geology 24, 879-882.
Keppie, J.D., 1993. Synthesis of Palaeozoic deformation events and terrane accretion in the Canadian
Appalachians. Geol. Rundsch. 82, 381-431.
Keppie, J.D., Dallmeyer, R.D., 1995. Late Paleozoic collision, delamination, short-lived magmatism
and rapid denudation in the Meguma Terrane (Nova Scotia, Canada): Constraints from 40Ar/39Ar
isotopic data. Can. J. Earth Sci. 32, 644-659.
Keppie, J.D., Boyle, R.W., Haynes, S.J. (Eds.), 1986. TurbiditeHosted Gold Deposits. Geol. Assoc.
Can., Spec. Pap. 32, 186 pp.
Kerrich, R., 1991. Mesothermal gold deposits - A critique of genetic hypotheses. In: Robert, F.,
Sheahan, P.A., Green, S.B. (Eds.), Greenstone Gold and Crustal Evolution. Geol. Assoc. Can.,
Mineral Deposits Div. Publ., pp. 13-31.
Kerrich, R., 1993. Perspectives on genetic models for lode-gold deposits. Miner. Deposita 28, 362-
365.
Page 24
Kerrich, R., 1994. Dating Archaean auriferous quartz vein deposits in the Abitibi greenstone belt,
Canada: 4°Ar/39Ar evidence for a 70-100 m.y.-time gap between plutonism-metamorphism and
mineralisation. A discussion. Econ. Geol. 89, 679-686.
Kerrich, R., Cassidy, K.F., 1994. Temporal relationships of lodegold mineralization to accretion,
magmatism, metamorphism and deformation, Archean to present: A review. Ore Geol. Rev. 9,
263-310.
Kerrich, R., Wyman, D.A., 1990. The geodynamic setting of mesothermal gold deposits: An
association with accretionary tectonic regimes. Geology 18, 882-885.
Kirkham, R.V., Sinclair, W.D., Thorpe, R.I., Duke, J.M. (Eds.), 1993. Mineral Deposit Modeling.
Geol. Assoc. Can., Spec. Pap. 40, 798 pp.
Kontak, D.J., Smith, P.F., Reynolds, P., Taylor, K., 1990. Geological and 4°Ar/39Ar
geochronological constraints on the timing of quartz vein formation in Meguma Group lode-gold
deposits, Nova Scotia. Ad. Geol. 26, 201-227.
Krogstad, E.J., Balakrishnan, S., Mukhopadhyay, D.K., Rajamani, V., Hanson, G.N., 1989. Plate
tectonics 2.5 billion years age: Evidence at Kolar, south India. Science 243, 1337-1340.
Landefeld, L.A., 1988. The geology of the Mother Lode-gold belt, Sierra Nevada Foothills
metamorphic belt, California. In: Bicentennial Gold 88, Extended Abstracts, Oral Programme.
Geol. Soc. Aust. Abstr. 22, pp. 167-172.
Lapointe, B., Chown, E.H., 1993. Gold-bearing iron-formation in a granulite terrane of the Canadian
Shield: A possible deeplevel expression of an Archean gold-mineralizing system. Miner. Deposita
28, 191-197.
Leitch, C.H.B., Van der Hayden, P., Godwin, C.I., Armstrong, R.L., Harakal, J.E., 1991.
Geochronometry of the Bridge River camp, southwestern British Columbia. Can. J. Earth Sci. 28,
195-208.
Lindgren, W., 1907. The relation of ore deposition to physical conditions. Econ. Geol. 2, 105-127.
Lindgren, W., 1933. Mineral Deposits, 4th ed. McGraw Hill, New York and London, 930 pp.
Macdonald, A.J. (Ed.), 1986. Proceedings of Gold '86. An International Symposium on the Geology
of Gold Deposits. Toronto, 517 pp.
MacLachlan, K., Helmstaedt, H., 1995. Geology and geochemistry of an Archean mafic dike
complex in the Chan Formation: Basis for a revised plate-tectonic model of the Yellowknife
greenstone belt. Can. J. Earth Sci. 32, 614-630.
Madden-McGuire, D.J., Silberman, M.L., Church, S.E., 1989. Geologic relationships, K-Ar ages and
isotopic data from the Willow Creek gold mining district, southern Alaska. In: Keays, R.R.,
Ramsay, W.R.H., Groves, D.I. (Eds.), The Geology of Gold Deposits: The Perspective in 1988.
Econ. Geol. Monogr. 6, 242-251.
Marakushev, A.A., Khokhlov, V.A., 1992. A petrological model for the genesis of the Muruntau gold
deposit. Int. Geol. Rev. 34 (1), 59-76.
McCoy, D., Newberry, R.J., Layer, P., DiMarchi, J.J., Bakke, A., Masterman, J.S., Minehand, D.L.,
1997. Plutonic-related gold deposits of interior Alaska. In: Goldfarb, R.J., Miller, L.D. (Eds.),
Mineral Deposits of Alaska. Econ. Geol. Monogr. 9, 191-241.
McCuaig, T.C., Kerrich, R., 1994. P-T-t-deformation-fluid characteristics of lode-gold deposits:
Evidence from alteration systematics. In: Lentz, D.R. (Ed.), Alteration and Alteration Processes
Associated with Ore-forming Systems. Geol. Assoc. Can., Short Course Notes, 11, 339-379.
McCuaig, T.C., Kerrich, R., Groves, D.I., Archer, N., 1993. The nature and dimensions of regional
and local gold-related hydrothermal alteration in tholeftic metabasalts in the Norseman goldfields:
The missing link in a crustal continuum of gold deposits?. Miner. Deposita 28, 420-435.
McKeag, S.A., Craw, D., 1989. Contrasting fluids in gold-bearing quartz vein systems formed
progressively in a rising metamorphic belt: Otago Schist, New Zealand. Econ. Geol. 84, 22-33.
Meinert, L.D., 1993. Igneous petrogenesis and skarn deposits. In: Kirkham, R.V., Sinclair, W.D.,
Thorpe, R.I., Duke, J.M. (Eds.), Mineral Deposit Modeling. Geol. Assoc. Can., Spec. Pap. 40,
569-583.
Mikucki, E.J., 1998. Hydrothermal transport and depositional processes in Archaean lode-gold
systems: A review. Ore Geol. Rev. 13, 307-321.
Page 25
Mikucki, E.J., Ridley, J.R., 1993. The hydrothermal fluid of Archaean lode-gold deposits: Constraints
on its composition inferred from ore and wallrock alteration assemblages over a spectrum of
metamorphic grades. Miner. Deposita 28, 469481.
Miller, L.D., Goldfarb, R.J., Gehrels, G.E., Snee, L.W., 1994. Genetic links among fluid cycling, vein
formation, regional deformation, and plutonism in the Juneau gold belt, southeastern Alaska.
Geology 22, 203-206.
Mitchell, A.H.G., Garson, M.S., 1981. Mineral Deposits and Global Tectonic Settings. Academic
Press, London, 405 pp.
Moravek, P. (Ed.), 1995. Gold deposits of the central and SW part of the Bohemian Massif. In:
Excursion Guide, Third Biennial SGA Meeting. Mineral deposits: From their origin to their
environmental impacts. Prague, August 28-31, 104 pp.
Mueller, A.G., Groves, D.I., 1991. The classification of Western Australian greenstone-hosted gold
deposits according to wallrock-alteration assemblages. Ore Geol. Rev. 6, 291-332.
Muller, A.G., Groves, D.I., 1997. Potassic Igneous Rocks and Associated Gold-Copper
Mineralization, 2nd ed. SpringerVerlag, Berlin, 238 pp.
Nesbitt, B.E., 1991. Phanerozoic gold deposits in tectonically active continental margins. In: Foster,
R.P. (Ed.), Gold Metallogeny and Exploration. Blackie and Son, Glasgow, pp. 104132.
Nesbitt, B.E., Muehlenbachs, K., 1989. Geology, geochemistry and genesis of mesothermal gold
deposits of the Canadian Cordillera: Evidence for ore formation from evolved meteoric water. In:
Keays, R.R., Ramsay, W.R.H., Groves, D.I. (Eds.), The Geology of Gold Deposits: The
Perspective in 1988. Econ. Geol. Monogr. 6, 553-563.
Nesbitt, B.E., Murowchick, J.B., Muehlenbachs, K., 1986. Dual origin of lode-gold deposits in the
Canadian Cordillera. Geology 14, 506-509.
Nie, F., 1997. An overview of gold resources in China. Int. Geol. Rev. In press.
Nokleberg, W.J., Bundtzen, T.K., Dawson, K.M., Eremin, R.A., Goryachev, N.A., Koch, R.D.
Ratkin, V.V., Rozenblum, I.S., Shpikerman, V.I., Frolov, Y.F., Gorodinsky, M.E., Khanchuck,
A.I., Kovbas, L.I., Melnikov, V.D., Nekrasov, I.Ya., Ognyanov, N.V., Petrachenko, E.D.,
Petrachenko, R.I., Pozdeev, A.I., Ross, K.V., Sidorov, A.A., Wood, D.H., Grybeck, D., 1996.
Significant metalliferous lode deposits and placer districts for the Russian Far East, Alaska and the
Canadian Cordillera. U.S. Geol. Surv., Open-File Rep. 513-A, 385 pp.
Peters, S.G., Golding, S.D., 1989. Geologic, fluid inclusion, and stable isotope studies of granitoid-
hosted gold-bearing quartz veins, Charters Towers, northeastern Australia. Econ. Geol. Monogr.
6, 260-273.
Phillips, G.N., Hughes, M.J., 1996. The geology and gold deposits of the Victorian gold province.
Ore Geol. Rev. 11, 255-302.
Phillips, G.N., Powell, R., 1993. Link between gold provinces. Econ. Geol. 88, 1084-1098.
Plafker, G., Berg, H.C., 1994. Overview of the geology and tectonic evolution of Alaska. In: Plafker,
G., Berg, H.C. (Eds.), The Geology of Alaska. The Geology of North America, vol. G-1. Geol.
Soc. Am., pp. 989-1021.
Poulsen, K.H., 1996. Lode-gold. In: Eckstrand, O.R., Sinclair, W.D., Thorpe, R.I. (Eds.), Geology of
Canadian Mineral Deposit Types. The Geology of North America, vol. P-1. Geol. Soc. Am., pp.
323-328.
Powell, W.G., Carmichael, D.M., Hodgson, C.J., 1995. Conditions and timing of metamorphism in
the southern Abitibi greenstone belt, Quebec. Can. J. Earth Sci. 32, 787-805.
Ramsay, W.R.H., 1998. A review of Victorian mesothermal gold, regional setting, styles, and genetic
constraints. In: Keays, R.R., Ramsay, W.R.H., Groves, D.I. (Eds.), The Geology of Gold
Deposits: The perspective in 1988.
Richards, J.P., 1995. Alkalic-type epithermal gold deposits: A review. Turbidite - hosted gold
deposits of Central Victoria, Australia: their regional setting, mineralisation styles and some
genetic constraints. Ore Geol. Rev. 13, 131-151 (this issue) In: Thompson, J.F.H. (Ed.), Magmas,
Fluids and Ore Deposits. Min. Assoc. Can., Short Course Ser., 23, 367-400.
Richards, J.P., Chappel, BW., McCulloch, M.T., 1990. Intraplate-type magmatism in a continent-
island-arc collision zone-Porgera intrusive complex, Papua New Guinea. Geology 18, 958-961.
Ridley, J., Mikucki, E.J., Groves, D.I., 1996. Archean lode-gold deposits: Fluid flow and chemical
evolution in vertically extensive hydrothermal systems. Ore Geol. Rev. 10, 279-293.
Page 26
Robert, F., 1996. Quartz-carbonate vein gold. In: Eckstrand, O.R., Sinclair, W.D., Thorpe, R.I. (Eds.),
Geology of Canadian Mineral Deposit Types. The Geology of North America, vol. P-1. Geol.
Soc. Am., pp. 350-366.
Robert, F., Brown, A.C., 1986. Archean gold-bearing quartz veins at the Sigma mine, Abitibi
greenstone belt, Quebec. Part 1. Geologic relations and formation of the vein systems. Econ. Geol.
81, 578-592.
Robert, F., Sheahan, P.A., Green, S.B. (Eds.), 1991. Greenstone Gold and Crustal Evolution. Geol.
Assoc. Can., Mineral Deposits Div. Publ., 252 pp.
Rushton, R.W., Nesbitt, B.E., Muehlenbachs, K., Mortensen, J.K., 1993. A fluid inclusion and stable
isotope study of Au quartz veins in the Klondike district, Yukon Territory, Canada. A section
through a mesothermal vein system. Econ. Geol. 88, 647-678.
Sawkins, F.J., 1972. Sulfide ore deposits in relation to plate tectonics. J. Geol. 80, 377-397.
Sawkins, F.J., 1990. Metal Deposits in Relation to Plate Tectonics, 2nd ed. Springer-Verlag, Berlin,
451 pp.
Scheiber, E., 1996. Geology of New South Wales: Synthesis. Geol. Surv. N.S.W. Mem. Geol. 13 (1),
295 pp.
Sengor, A.M.C., Okurogullari, A.H., 1991. The role of accretionary wedges in the growth of
continents - Asiatic examples from Argand to plate tectonics. Eclogae Geol. Helv. 84 (3), 535-
597.
Shenberger, D.M., Barnes, H.L., 1989. Solubility of gold in aqueous sulfide solutions from 150 to
350°C. Geochim. Cosmochim. Acta 53, 269-278.
Sherlock, R.L., Logan, M.A.V., 1995. Silica-carbonate alteration of serpentinite: Implications for the
association of mercury and gold mineralization in northern California. Explor. Mining Geol. 4 (4),
395-409.
Sibson, R.H., Robert, F., Poulsen, K.H., 1988. High-angle reverse faults, fluid pressure cycling, and
mesothermal gold-quartz deposits. Geology 16, 551-555.
Sillitoe, R.H., 1991. Intrusion-related gold deposits. In: Foster, R.P. (Ed.), Gold Metallogeny and
Exploration. Blackie and Son, Glasgow, pp. 165-209.
Sillitoe, R.H., 1993. Gold-rich porphyry copper deposits: Geological model and exploration
implications. In: Kirkham, R.V.,
Sinclair, W.D., Thorpe, R.I., Duke, J.M. (Eds.), Mineral Deposit Modeling. Geol. Assoc. Can., Spec.
Pap. 40, 465-478.
Solomon, M., Groves, D.I., 1994. The Geology and Origin of Australia's Mineral Deposits. Oxford
Monogr. Geol. Geophys. 24, 951 pp.
Stein, H.J., Markey, R.J., Morgan, JW., Zak, K., Zachariad, J., Sundblad, K., 1996. Re-Os dating of
Au deposits in shear zones using accessory molybdenite: Bohemian Massif, Carolina slate belt,
and Fennoscandian Shield examples. Geol. Soc. Am. Abstr. Progr. 28 (7), 474.
Stowell, H.H., Lesher, C.M., Green, N.L., Sha, P., Guthrie, G.M., Sinha, A.K., 1996. Metamorphism
and gold mineralization in the Blue Ridge, southernmost Appalachians. Econ. Geol. 91, 1115-
1144.
Stuwe, K., Will, T.M., Zhou, S., 1993. On the timing relationship between fluid production and
metamorphism in metamorphic piles: Some implications for the origin of post-metamorphic gold
mineralization. Earth Planet. Sci. Lett. 114, 417-430.
Thomas, D.J., Heaman, L.M., 1994. Geologic setting of the Jolu gold mine, Saskatchewan: U-Pb age
constraints on plutonism, deformation, mineralization and metamorphism. Econ. Geol. 89, 1017-
1030.
Trumbull, R.B., Hua, L., Lehrberger, G., Satir, M., Wimbauer, T., Morteani, G., 1996. Granitoid -
hosted gold deposits in the Anjiayingzi District of Inner Mongolia, People's Republic of China.
Econ. Geol. 91, 875-895.
Vinyu, M.L., Frei, R., Jelsma, H.A., 1996. Timing between granitoid emplacement and associated
gold mineralization: Examples from the ca. 2.7 Ga Harare-Shamva greenstone belt, northern
Zimbabwe. Can. J. Earth Sci. 33, 981-992.
Wang, L.G., Luo, Z.K., McNaughton, N.J., Groves, D.I., Huang, J.Z., Miao, L.C., Guan, K., Liu,
Y.K., 1996. SHRIMP U-Pb in zircon studies of plutonic rocks from the Jiaodong Peninsular,
Shangdong Province, China: Constraints on crustal and tectonic evolution and gold metallogeny.
Page 27
In: Mesothermal Gold Deposits: A Global Overview. Geol. Dept. (Key Centre) Univ. Ext., Univ.
West. Aust. Publ. 27, 34-38.
White, D.E., 1967. Mercury and base-metal deposits with associated thermal and mineral waters. In:
Barnes, H.L. (Ed.), Geochemistry of Hydrothermal Ore Deposits. Holt, Rinehart and Winston,
New York, pp. 575-631.
White, N.C., Hedenquist, JW., 1995. Epithermal gold deposits: Styles, characteristics and
exploration. Soc. Econ. Geol. Newslett. 23 (1), 9-13.
Wyman, D., Kerrich, R., 1988. Alkaline magmatism, major structures, and gold deposits:
Implications for greenstone belt gold metallogeny. Econ. Geol. 83, 451-458.
Page 28
3. Rotund vs skinny orogens : well nourished or

anorexic gold?
R. J. Goldfarb a, D. I.Groves b and S. J.Gardoll b

a
U.S. Geological Survey, Box 25046, Denver Federal Centre, Denver, Co. 80225, USA
b
Centre for Global Metallogeny, Department of Geology and Geophysics, University of
Western Australia, Nedlands, WA 6907, Australia
ASTRACT
Globally-distributed orogenic gold vein deposits, because they require a particular conjunction of
processes to form and be preserved, can be sensitive indicators of evolving tectonic processes
throughout geologic time. A clearly heterogeneous temporal distribution of formation ages for these
mineral deposits is marked by two major Precambrian peaks (2800-2555 Ma and 2100–1800 Ma), a
singular lack of deposits for 1,200 m.y. (1800–600 Ma), and a continuous genesis since (<600 Ma).
The older parts to this distribution relate to major episodes of continental growth, perhaps controlled
by plume-influenced mantle overturn events, in the hotter early Earth (ca. ≥1800 Ma). This allowed
preservation of gold deposits in roughly equidimensional large masses of buoyant continental crust
(cratons). Evolution to a less episodic, more-continuous, modern-style plate-tectonic regime led to
the accretion of volcanosedimentary complexes as progressively-younger linear orogenic belts
surrounding the margins of the more-buoyant Archean-Paleoproterozoic cratons. The susceptibility
of these linear belts to uplift and erosion can elegantly explain the overall lack of orogenic gold
deposits at 1800-600 Ma, their exposure in 600-50 Ma orogens, the increasing importance of placer
deposits back through the Phanerozoic since ca. 100 Ma, and their absence in orogenic belts younger
than ca. 50 Ma.
3.1. Introduction
In the past two decades, there has been extreme controversy concerning the initiation and
evolution of styles of plate-tectonic processes as the Earth cooled (de Wit, 1998; Hamilton,
1998). Mineral deposits, because they commonly reflect the redox conditions of their
depositional environment (e.g. placer uraninite and pyrite in Witwatersrand, banded iron
formations, etc.), have been extensively used as evidence in the long-standing debate on the
evolution of the Earth’s biosphere and atmosphere (Barley and Groves, 1992), but mineral
deposits can be equally sensitive indicators of the tectonic environment during their
formation (Meyer, 1988). For example, specific mineral deposits (e.g. many Fe, Pb-Zn or
Cu deposits) can be related to the opening of internal oceans during the breakup of
continents, whereas other (e.g. Au-Ag, Cu-Au) record the closure of external oceans (Barley
et al., 1998). The temporal distribution of a type of mineral deposit types can, therefore,
contribute substantially to our understanding of the supercontinent cycle and their spatial
distribution can potentially provide insights into evolutionary changes in the nature of the
tectonic processes that contributed to their formation and controlled their preservation.
Page 29
3.2. Orogenic Gold Deposits In Space And Time
A problem with understanding the spatial-temporal distribution of many mineral-deposit

types is that, although they are sensitive to a particular plate tectonic setting, they form in
environments where they are highly susceptible to erosion. For example, shallow-level
porphyry-Cu-Au to subaerial epithermal Au-Ag deposits are the products of rapidly uplifting
magmatic arcs (Sillitoe, 1989), but their preservation in the geological record beyond the
Mesozoic is poor. Orogenic gold deposits are an important exception because they form
mainly at mid-crustal levels during the latter stages of the evolution of mountain belts
(Groves et al., 1998; Goldfarb et al., 1998; Groves et al., 2000), and hence have better
preservational potential, as shown by their global abundance within deformational belts as
old as Archean.
Figure 1. Distribution of major Mesozoic-early Tertiary orogenic gold provinces showing

strong control by linear accretionary or collisional orogenic belts. Most of these belts surround the
tectonically-active circum-Pacific margin. The lack of orogenic gold deposits along westernmost
South America reflects this being a consuming margin characterized by subduction-erosion and
exposure of shallow crustal levels in an uplifting arc, which are more favorable for young epithermal
gold deposits. The Alpine and Himalayan areas contain only rare, small orogenic gold systems.
Important deposits may have formed at mid-crustal levels during collisions within the last 50 m.y. of
Earth history, but most of the vein systems have yet to be unroofed.
These orogenic gold deposits were deposited from deeply-sourced metamorphic (±

magmatic) fluids generated in near-arc, volcano-sedimentary rock sequences, and they are
associated with plate subduction below, and (or) terrane collision on to, older blocks of
continental crust (Goldfarb et al., 1988, 1991; Landefeld, 1988; Barley et al., 1989; Kerrich
and Wyman, 1990). The deposits represent focused fluid flow that is an inherent product of
crustal heating during orogenesis (Thompson, 1997). Under typical mid-crustal temperature
Page 30
regimes, given sufficient availability of sulfide species, gold will be mobilized in the
upwardly migrating fluid systems.
The more recent orogenic gold deposits, therefore, coincide spatially with the margins of
external oceans where accretion of juvenile crust has taken place, as illustrated by the
distribution of most major Mesozoic-Tertiary orogenic gold provinces around the Pacific
margin (Fig. 1). They are commonly related to environments where large thermal anomalies
were caused by thrust-related thickening of relatively radiogenic crustal material in an
orogen core (Jamieson, 1998), or upwelling of asthenosphere due to processes that include
subduction of oceanic ridges (Haeussler et al., 1995), subduction roll-back (Landefeld, 1988)
or lithospheric delamination (Qiu and Groves, 1999). Older orogenic gold deposits will
reflect similar crustal fluid flux regimes. The formation and preservation of the gold
deposits are thus sensitive indicators of continental growth back through geologic time.
The abundance of recent geochronology for orogenic gold deposits, and the availability of
reliable gold resource data now for gold-producing areas in eastern and central Asia, allow
for a much improved estimate of the distribution of orogenic gold
(production+resource+associated placers) through time (Fig. 2). There are two distinct
major periods of early Precambrian orogenic-gold formation—one at 2800-2550 Ma, which
includes the giant Kalgoorlie, Australia, and Timmins, Canada, goldfields, and the second at
2100-1800 Ma, including the giant Ashanti, Ghana, and Homestake, USA, goldfields. There
is then a distinct lack of deposits until ca. 600 Ma where there is limited production from
well-documented modern mines, and then clusters of major gold provinces from ca. 450 to
ca. 50 Ma, including the giant Mother Lode, USA, goldfields, and the appearance of giant
gold placers (Henley and Adams, 1979) related to erosion of these deposits.
Page 31
Figure 2. Distribution of
orogenic gold provinces through time
indicating concentrations of ores at ca.
2800-2550 Ma, 2100-1800 Ma, and
600-50 Ma. Resource values combine
past production, known resources, and
associated placers from all available
data. The two older episodes correlate
with major periods of juvenile crustal
growth during episodes of whole-
mantle convection. Little gold is
preserved from between 1800-600 Ma,
as exposed parts of orogens from this
period reflect crustal levels typically
below gold-favorable depths (about 3-
20 km). Gold provinces spread
throughout the Phanerozoic were
mainly products of orogenesis around
the circum-Pacific rim, along the
Gondwanan continental margin, and
along the north and western margins of
the Paleo-Tethyan ocean.
Page 32
3.3. Gold Deposits And Evolution Of Early Earth
This distribution of gold ores cannot be explained simply by the lack of external oceans and
their convergent margins between 1800 and 600 Ma, because considerable new continental
crust was formed and preserved during some of this period, particularly culminating with the
formation of Rodinia at ca. 1300-1000 Ma (Dalziel, 1991; Hoffman, 1991; Moores, 1991).
The most logical conclusion, therefore, is that 1800–600 Ma represents a period for which
there is poor preservational potential for orogenic gold deposits, rather than simply a period
of non-deposition of gold ores. Thus, any evolutionary tectonic model for Earth must in
itself explain: i) the episodic nature of the extremely productive preserved orogenic gold
provinces of the early Precambrian, ii) the lack of preservation of deposits that would have
formed between 1800 and 600 Ma, iii) the more continuous nature of the provinces between
450 and 50 Ma, iv) the increasing number of placer goldfields in the Phanerozoic, and v) the
lack of major orogenic gold deposits in belts younger than 50 Ma.
Orogenesis and continental growth have occurred over a relatively broad time period
covering much of Earth history, including during 1.8-1.5 Ga events in the southwestern
USA (Hoffman, 1989) and during extensive 1.3-1.0 Ga events forming Rodinia (Hoffman,
1991). Differences, however, in both the styles of such growth and plate tectonic processes
themselves, affected the potential for preservation of the deposits at various stages in Earth
evolution. Based on the fact that the major periods of Precambrian orogenic gold formation
coincide with the two major episodes of crustal growth (“super-events”) at the end of both
the Archean and Paleoproterozoic, in which more than 75 percent of the juvenile crust on
Earth was developed (Stein and Hoffman, 1994; Condie, 1995, 1998), the solution to the
enigma of the uneven temporal distribution of orogenic gold must lie with the process that
caused such rapid crustal growth. This is further accentuated by the fact that the orogenic
gold events broadly coincide with Condie’s (1998) Archean crustal-growth “subevents”,
within the uncertainties of timing constraints, at 3.0, 2.7, and 2.6-2.5 Ga. The model
presented below is based on this premise.
The periods of early Precambrian rapid granite-greenstone crustal growth are now
commonly thought to be due to overturn of a layered mantle to one with a transient, whole-
mantle convection (Davies, 1992, 1995; Stein and Hoffman 1994). Resulting mantle plumes
would have generated vast amounts of new crust due to decompression melting at the base
of the lithosphere in these two “super-events”. Alternative ideas to the plume theory include
the presence of a modern style plate regime, but one in which a thickened early Precambrian
continental mantle roots reflect the underplating of relatively buoyant slabs produced on a
hotter Earth (e.g., Kusky and Polat, 1999; Smith and Lewis, 1999). Whereas such an
alternative isn’t as satisfactory for explaining the rapid and episodic Precambrian crustal
growth pattern, it too would lead to establishment of massive granite-greenstone terranes.
No matter what the process of generation of juvenile crust, the resulting amalgamation of the
volcano-sedimentary (greenstone) terranes produced relatively equidimensional, buoyant
continental cratons, in which there was exceptional preservation of the volcano-sedimentary
successions and the orogenic gold deposits created within them, particularly in the centers of
cratons, which were protected from later orogenesis. It could be argued that the cratons have
assumed their current shape largely by subsequent modification, but each well-described
craton has a distinctive history and, in particular, specific metallogenic associations, which
Page 33
are matched nowhere else on Earth. For example, the primitive ca. 2700 Ma Abitibi
volcanic rocks of the Superior Province, Canada, and their contained giant orogenic gold
and volcanogenic massive copper-zinc sulphide deposits are nowhere replicated, nor are the
giant orogenic gold and komatiite-hosted nickel provinces of the ca. 2700 Ma crustally-
contaminated volcanic successions of the Norseman-Wiluna Belt in Western Australia.
3.4. Gold Distributions Under A More Modern-Style Plate

Tectonics
In contrast to the relatively equidimensional Archean and Paleoproterozoic continental masses,

younger juvenile crust appears as long, relatively narrow, microcontinents or accretionary collages,
such as those of modern western North America or those in the various reconstructions of Rodinia at
ca. 1000 Ma (Fig. 3). This is interpreted to represent a shift from strongly episodic, perhaps plume-
influenced, plate tectonics in the hotter earlier Earth to a less-episodic style of plate tectonics as the
Earth cooled (e.g. Davies, 1995). Although formation of orogenic gold deposits was undoubtedly an
inherent part of the crustal fluid flow process in the medium-temperature parts to many post-1800 Ma
orogens, the deposits no longer were generally inherently “protected” in the interiors of the cratons.
They could now be more easily exposed to uplift and orogenic activity on the typically reworked
margins of the cratons. Progressive erosion down to the roots of these orogens, below auriferous
mid-crustal levels, is suggested to explain the general lack of orogenic gold deposits dated between
1800 and 600 Ma, and their increasing preservation post-450 Ma.
Figure 3. Rodinia at ca. 1000 Ma showing the location of main gold provinces at that time,
relatively-equidimensional Archean-Paleoproterozoic cratonic blocks, and linear Mesoproterozoic
orogenic belts. The change in shape of crustal units reflects the shift from episodic “super-events” to
a more present-day style of plate tectonics leading to continental growth. Whereas the gold deposits
in the cratons are still preserved today, reworking of the Rodinian collisional belts during the last 1
Ga has exposed high-grade basement rocks that are below zones favorable for formation of gold
Page 34
lodes. Shallower, gold-favorable levels of these types of linear belts are not well-preserved until
Phanerozoic times. Rodinia reconstruction from Unrug (1996).
Significant post-Paleoproterozoic orogenic gold ores are first noted from the latest Proterozoic. The
oldest extensive mid-crustal geosynclinal remnants surrounding cratonic cores include a number of
terranes that evolved in the Late Proterozoic Panthalassic Ocean and became important gold hosts
during collisions along ocean margins. The most significant of these include several deformed, gold-
rich Neoproterozoic terranes along the southern margin of the Siberian craton, such as the Yenisei
fold belt containing the large Olympiada and Sovetsk gold deposits. In is still uncertain as to whether
much of thermal activity here occurred ca. 800 Ma (Safanov, 1997) or ca. 600 Ma (Konstantinov et
al., 1999). Simultaneously, additional gold-forming events were recorded during the early evolution
of Gondwana, which were a part of the Pan-African tectonism defining Arabian-Nubian accretion in
northeastern Africa (Agar, 1992) and the Hoggar area collision in northwestern Africa (Ferkous and
Monie, 1995).
Relatively continuous gold-forming events are evident from the Ordovician throughout the remainder
of the Paleozoic along both globally extensive, subduction-related margins of a growing Pangea (e.g.,
the Paleozoic orogens of figure 1). To the south, these would have extended across Gondwana from
what were the historically productive goldfields of the Tasman orogen in eastern Australia (e.g.,
Lachlan, Thomson, Hodgkinson fold belts) and those of Westland New Zealand, to the Eastern
Cordillera of South America. Along the northern margin of the Paleo-Tethys Ocean, collisional
events along an active margin many thousands of kilometers in length ultimately led to formation of a
series of Caledonian and Variscan gold provinces. These included those of the southeastern U.S.,
Nova Scotia (Canada), United Kingdom, southern European massifs, Ural Mountains, central Asia,
Baikal region, and northern China. The breakup of Pangea and formation of the Mesozoic Pacific
basin led to a shift in location of the active margins and related Phanerozoic continental growth.
Associated with such are the great Pacific Rim goldfields of the western United States and eastern
Asia (fig. 1; Goldfarb et al., 1998). Collectively, the recognition of productive orogenic gold deposits
in the elongated collisional belts of the last 600-450 m.y. suggests this to be a broad-scale threshold
value for preservation of such deposits under the post-Paleoproterozoic plate tectonic pattern. That is,
for the most part, orogenic belts of Mesoproterozoic and Neoproterozoic age are likely to have been
too reworked and deeply eroded for the preservation of orogenic gold deposits. Many younger belts,
however, have yet to reach such a stage.
There is thus a roughly 500- to 600-m.y.-long “moving window” of time that is favorable for post-
Paleoproterozoic orogenic gold resources, which may occur in both lodes and/or downstream placers.
Certainly, the association of significant gold placers with the Phanerozoic orogenic goldfields (e.g.,
California Mother Lode, Kolyma and Lena regions of eastern Russia, Victorian goldfields of
southeastern Australia) emphasizes the preferential potential for erosion of primary gold deposits in
such linear mobile belts. In contrast to Archean and Paleoproterozoic gold deposits often located in
the centers of broad Precambrian cratons, ongoing tectonism along narrow Phanerozoic margins has
led to (e.g., Tasman orogen), or continues to lead to (western North America), remobilization of some
lode resources into economic placer accumulations. The lack of any economically significant
orogenic gold deposits in rocks younger than ca. 50 Ma (fig. 2) suggests that this is the minimum
period required to uplift and expose a significant goldfield with these types of deposits in such belts.
It is very probable that in a few tens of millions of years small Cenozoic orogenic gold vein
occurrences now being exposed in the European Alps (Pettke et al., 1999) and New Zealand Alps
(Koons and Craw, 1991) will appear as more extensive mineralized systems when hosting orogens
are further unroofed. At the same time in the future, presently exposed important gold systems of
Paleozoic or latest Proterozoic age may be lost from the geologic record, causing the approximately
1.2-b.y.-long gap between Paleoproterozoic orogenic gold deposits and the younger group of deposits
(fig. 2) to continue to increase.
Page 35
3.5. Conclusions: Control On Gold Distribution By Pattern Of

Crustal Growth
In summary, the temporal distribution of orogenic gold deposits is totally consistent with the notion
of evolving plate tectonic processes during Earth evolution. Whereas universally associated with
subduction-related processes, and presumably forming throughout Earth history, the deposits show a
specific temporal pattern due to selective preservation. Early Precambrian orogenic gold provinces
owe their preservation to probable strongly plume-influenced plate-tectonic processes, which caused
massive episodic lithosphere melting and incorporation of the gold-hosting orogens into broadly
equidimensional cratons, the centres of which were largely immune from later orogenesis.
Subsequently, the shift to a less plume-dominated, less-episodic style of plate tectonics in Earth
history is recorded by the presence of both gold-poor Mesoproterozoic-Neoproterozoic orogens and
the great Phanerozoic global orogenic gold belts of the circum-Pacific, Gondwanan margin, and
Paleo-Tethys. These more linear orogens, formed by terrane accretion to the older cratons. They
were more susceptible to one or more periods of later deformation, uplift and erosion, leading to little
preservation of any 1800-600 Ma mid-crustal gold deposits. A 550-m.y.-long time window (600-50
Ma) defines the concentration of preserved Phanerozoic lode and placer deposits, still geologically
young enough such that deeper crustal levels below the lodes do not yet dominate these orogens.
Giant gold deposits formed in both the early Precambrian and the Phanerozoic, further suggesting that
evolution in tectonic process and style was a greater influence on preservation than it was on
formation of the orogenic gold deposits.
3.6. References
Agar, R.A., 1992, The tectono-metallogenic evolution of the Arabian Shield:

Precambrian Research, v. 58, p. 169-194.
Barley, M.E., Eisenlohr, B.N., Groves, D.I., Perring, C.S., and Vearncombe, J.R., 1989, Late Archean
convergent margin tectonics and gold mineralization--A new look at the Norseman-Wiluna
belt, western Australia: Geology, v. 17, p. 826-829.
Barley, M.E., 1992, Groves, D.I. Super-continent cycles and the distribution of metal deposits
through time: Geology, v. 20, p. 291-302.
Barley, M.E., Krapez, B., Groves, D.I., and Kerrich, R., 1998, The Late Archaean bonanza:
Metallogenic and environmental consequences of the interaction between mantle plumes,
lithospheric tectonics and global cyclicity: Precambrian Research, v. 91, p. 65-90.
Condie, K.C., 1995, Episodic ages of greenstones: A key to mantle dynamics: Geophysical Research
Letters, v. 22, p. 2215-2218.
—, 1998, Episodic continental growth and supercontinents: A mantle avalanche connection?: Earth
and Planetary Science Letters, v. 163, p. 97-108.
Dalziel, I.W.D., 1991, Pacific margins of Laurentia and East Antarctica-Australia as a conjugate rift
pair: Evidence and implications for an Eocambrian supercontinent: Geology, v. 19, p. 598-
601.
Davies, G.F., 1992, On the emergence of plate tectonics: Geology, v. 20, p. 963-966.
—, 1995, Penetration of plates and plumes through the mantle transition zone: Earth and Planetary
Sciences Letter, v. 133, p. 507-516.
de Wit, M.J., 1998, On Archean granites, greenstones, cratons and tectonics: does the evidence
demand a verdict?: Precambrian Research., v. 91, p. 181-226.
Ferkous, K. and Monie, P., 1997, Petrostructural data and 40Ar/39Ar laser probe dating of the Pan-
African shearing and related gold-mineralization in the East In Ouzzal district (Western
Page 36
Hoggar, Algeria), in Papunen, H., ed., Research and Exploration--Where Do They Meet?:
Rotterdam, Balkema, p. 185-189.
Goldfarb, R.J., Leach, D.L., Pickthorn, W.J., and Paterson, C.J., 1988, Origin of lode-gold deposits of
the Juneau gold belt, southeastern Alaska: Geology, v. 16, p. 440-443.
Goldfarb, R.J., Snee, L.W., Miller, L.D., and Newberry, R.J., 1991, Rapid dewatering of the crust
deduced from ages of mesothermal gold deposits: Nature, v. 354, p. 296-298.
Goldfarb, R.J., Phillips, G.N., and Nokleberg, W.J., 1998, Tectonic setting of synorogenic gold
deposits of the Pacific Rim: Ore Geology Review, v. 13, p. 185-218.
Groves, D.I., Goldfarb, R.J., Gebre-Mariam, H., Hagemann, S.G., and Robert, F., 1998, Orogenic
gold deposits-A proposed classification in the context of their crustal distribution and
relationship to other gold deposit type: Ore Geology Reviews, v. 13, p. 7-27.
Groves, D.I., Goldfarb, R.J., Knox-Robinson, C.M., Ojala, J., Gardoll, S., Yun, G., and Holyland, P.,
2000, Late-kinematic timing of orogenic gold deposits and significance for computer-based
exploration techniques with emphasis on the Yilgarn block, Western Australia: Ore Geology
Reviews, v. 17, p. 1-38.
Haeussler, P.J., Bradley, D., Goldfarb, R.J., Snee, L.W., and Taylor, C.D., 1995, Link between ridge
subduction and gold mineralization in southern Alaska: Geology, v. 23, p. 995-998.
Hamilton, W.B., 1998, Archean magmatism and deformation were not products of plate tectonics:
Precambrian Research, v. 91, p. 143-179.
Henley, R.W., and Adams, J., 1979, On the evolution of giant gold placers: Transactions of the
Institution of Mining and Metallurgy, v. 88, p. B41-B50.
Hoffman, P.F., 1989, Precambrian geology and tectonic history of North America, in Bally, A.W.,
and Palmer, A.R., eds., The Geology of North America; An Overview: Boulder, Geological
Society of America, p. 447-512.
—, 1991, Did the breakout of Laurentia turn Gondwanaland inside-out?: Science, v. 252, p. 1409-
144.
Jamieson, R.A., Beaumont, C., Fullsack, P., and Lee, B., 1998, Barrovian regional metamorphism:
where’s the heat?, in Treloar, P.J., and O'Brien, P.J., eds., What Drives Metamorphism and
Metamorphic Reactions?, Volume 138, Geology Society of London, p. 23-51.
Kerrich, R., and Wyman, D.A., 1990, The geodynamic setting of mesothermal gold deposits-An
association with accretionary tectonic regimes: Geology, v. 18, p. 882-885.
Konstantinov, M., Dankovtsev, R., Simkin, G. and Cherkasov, S., 1999, Deep structure
of the north Enisei gold district (Russia) and setting of ore deposits: Geology of Ore
Deposits, v. 41, no. 5, p. 425-436.
Koons, P.O. and Craw, D., 1991, Gold mineralization as a consequence of continental collision:
An example from the Southern Alps of New Zealand. Earth and Planetary Science
Letters, 103: 1-9.
Kusky, T.M., and Polat, A., Growth of granite-greenstone terranes at convergent margins, and
stabilization of Archean cratons: Tectonophysics, v. 305, p. 43-73.
Landefeld, L.A., 1988, The geology of the Mother Lode gold belt, Sierra Nevada Foothills
metamorphic belt, California., Bicentennial Gold 88, Extended Abstracts: Melbourne,
Geological Society of Australia, p. 167-172.
Meyer, C., 1988, Ore deposits as guides to the geologic history of the Earth: Annual Review of Earth
and Planetary Science, v. 16, p. 147-172.
Moores, E.M., 1991, Southwest U.S.-East Antarctic (SWEAT) connection: A hypothesis: Geology,
v. 10, p. 425-428.
Pettke, T., Diamond, L.W. and Villa, I.M., 1999, Mesothermal gold veins and
metamorphic devolatilization in the northwestern Alps: The temporal link. Geology, v.
27, p. 641-644.
Qiu, Y., and Groves, D.I., 1999, Late Archean collision and delamination in the
southwest Yilgarn craton: The driving force for Archean orogenic lode gold mineralization:
Economic Geology, v. 94, p. 115-122.
Safonov, Y.G., 1997, Hydrothermal gold deposits--Distribution, geological/genetic types,
and productivity of ore-forming systems: Geology of Ore Deposits, v. 39, no. 1, p. 20-32.
Sillitoe, R.H., 1989, Gold deposits in western Pacific island arcs: The magmatic connection:
Page 37
Economic Geology, p. 274-291.

Smith, A.D., and Lewis, C., 1999, The planet beyond the plume hypothesis: Earth-Science
Reviews, v. 48, p. 135-182.
Stein, M., and Hoffmann, A.W., 1994, Mantle plumes and episodic crustal growth: Nature, v.
372, p. 63-68.
Thompson, A.B., 1997, Flow and focusing of metamorphic fluids, in Jamtveit, B., and Yardley,
B.W.D., eds., Fluid Flow and Transport in Rocks: London, Chapman and Hall, p. 297-314.
Unrug, R., 1996, Geodynamic map of Gondwana supercontinent assembly: Bureau de Recherches
Geologiques et Minieres, Orleans, France.
Page 38
4. Tectonic setting of synorogenic gold deposits

of the Pacific Rim
R.J. Goldfarb a, G.N. Phillips b, W.J. Nokleberg c

a
U. S. Geological Survey, Box 25046, Denver Federal Center, Denver, CO 80225. USA
b
Great Central Mines Limited, 1 Coppin St., East Malvern 3145. Vic., Australia U.S.
c
Geological Survey, 345 Middlefield Road, Menlo Park, CA 94025, USA
Abstract
More than 420 million oz of gold were concentrated in circum-Pacific synorogenic quartz lodes
mainly during two periods of continental growth, one along the Gondwanan margin in the
Palaeozoic and the other in the northern Pacific basin between 170 and 50 Ma. These ores have
many features in common and can be grouped into a single type of lode gold deposit widespread
throughout elastic sedimentary-rock dominant terranes. The auriferous veins contain only a few
percent sulphide minerals, have goldailver ratios typically greater than 1:1, show a distinct
association with medium grade metamorphic rocks, and may be associated with large-scale fault
zones. Ore fluids are consistently of low salinity and are CO2-rich.
In the early and middle Palaeozoic in the southern Pacific basin, a single immense turbidite sequence
was added to the eastern margin of Gondwanaland. Deformation of these rocks in southeastern
Australia was accompanied by deposition of at least 80 million oz of gold in the Victorian sector of
the Lachlan fold belt mainly during the Middle and Late Devonian. Lesser Devonian gold
accumulations characterized the more northerly parts of the Gondwanan margin within the
Hodgkinson-Broken River and Thomson fold belts. Additional lodes were emplaced in this flyschoid
sequence in Devonian or earlier Palaeozoic times in what is now the Buller terrane, Westland, New
Zealand. Minor post-Devonian growth of Gondwanaland included terrane collision and formation of
gold-bearing veins in the Permian in Australia's New England fold belt and in the Jurassic-Early
Cretaceous in New Zealand's Otago schists.
Collision and accretion of dozens of terranes for a 100-m.y.-long period against the western margin
of North America and eastern margin of Eurasia led to widespread, latest Jurassic to Eocene gold
veining in the northern Pacific basin. In the former location, Late Jurassic and Early Cretaceous
veins and related placer deposits along the western margin of the Sierra Nevada batholith have
yielded more than 100 million oz of gold. Additional significant ore-forming events during the
development of North America's Cordilleran orogen included those in the Klamath Mountains
region, California in the Late Jurassic and Early Cretaceous: the Klondike district, Yukon by the
Early Cretaceous; the Nome and Fairbanks districts, Alaska, and the Bridge River district, British
Columbia in the middle Cretaceous; and the Juneau gold belt, Alaska in the Eocene. Gold-bearing
veins deposited during the Late Jurassic and Early Cretaceous terrane collision that formed the
present-day Russian Far East have been the source for more than 130 million oz of placer gold.
The abundance of gold-bearing quartz-carbonate veins throughout the Gondwanan, North American
and Eurasian continental margins suggests the migration and concentration of large fluid volumes
during continental growth. Such volumes could be released during orogenic heating of hydrous
silicate mineral phases within accreted marine strata. The common temporal association between
gold veining and magmatism around the Pacific Rim reflects these thermal episodes.
Melting of the lower thick:ned crust during arc formation, slab rollback and extensional tectonism,
and subduction of a slab window beneath the seaward part of the forearc region can all provide the
Page 39
required heat for initiation of the ore-forming processes. 9 1993 Elsevier Science B.V. All rights
reserved.
4.1. Introduction
A key feature of many metasedimentary rockhosted gold deposits is their spatial and temporal
association with collisional orogens and, because of this, we refer to such deposits in this paper as
synorogenic gold deposits. The widespread distribution of these deposits throughout the accreted
terranes of western North America suggests a direct association between continental growth and ore
genesis (Barley et al.. 1989: herrich and Wyman, 1990). Gold ores have formed both in interior
orogens developed between lame land masses during continent-continent collision G.e,
AppalachianCaledonian. Hercynian. Uralian) and in peripheral orogens built during subduction of
oceanic crust along continental margins (i.e., Cordilleran, Tasman). Most significant is the fact that
collisional tectonism can eventually trigger crustal heating events and such heat is critical for
devolatilization reactions and ore fluid formation (Powell et al., 1991; Phillips, 1991; Goldfarb et al.,
1993). Simultaneously with or soon after the heating, transform movements along major crustal fault
zones may also be critical to the ore-forming process, leading to relaxation of regional compressive
forces and enhanced crustal-scale permeabilities.
Unlike Archaean gold vein deposits, synorogenic gold lodes of Phanerozoic age are predominantly
hosted by oceanic sedimentary rocks. In many Phanerozoic belts, however, subordinate mafic
volcanic rocks are interbedded with slates and graywackes. This reflects contemporaneous
sedimentation and oceanic arc or ridge volcanism prior to collision. The magmatic roots to such arc
volcanism are rarely observed in suture zones, appearing as minor mafic and/or ultramafic plutonic
bodies. In most auriferous slate belts, felsic to intermediate intrusive bodies occur within a few tens of
kilometres of the gold lodes. Modern dating methods have shown significant temporal overlap,
suggesting the intrusions and the gold to be products of a single, large-scale crustal thermal event.
These syn- to postcollision plutons, as well as the gold veins, can occur in the continental margin
fore-arc region, further inboard in the magmatic arc, or, less commonly, a few tens of kilometres
landward of the arc.
Fig. 1 . Major synorogenic gold districts along the Pacific plate margin.
Synorogenic gold deposits often referred to as mesothermal deposits despite a wide range in P-T of
ore formation, are widely-distributed in accreted terranes that were deformed along the circum-Pacific
margin (Fig. 1). More than 420 million oz of gold have been recovered from quartz vein and related
Page 40
placer deposits in this region, with greatest production equally divided between the North American
Cordillera, the Russian Far East and northeastern China, and the Tasman fold belt of eastern Australia
(Fig. 2). These regions contain most of the world's recognized Phanerozoic gold resources associated
with synorogenic lodes (the one main exception being the upper Palaeozoic deposits along the
southern side of the Angara craton in central Asia). The ores are ultimately the product of two
extensive diachronous periods of deformation along what were or had recently been growing
continental margins. These main periods of Pacific rim orogenesis include the Palaeozoic tectonism
along the southeastern Gondwanan margin and the Middle Jurassic to Early Tertiary plate collisions
in the northern Pacific basin. The detailed study of how gold formation relates to regional tectonism
for some of the youngest lodes located in Alaska (Goldfarb et al., 1997) is a key to helping decipher
the tectonic controls on genesis of older Phanerozoic ores.
Fig. 2. Relative production from synorogenic lodes and related placer deposits of the Pacific rim.
In the following summary, we describe the most accepted tectonic settings, the metamorphic and
igneous histories and the published dates of gold veining for the more important circum-Pacific
synorogenic deposits. Much of the detail is focused on the relatively well-studied Middle Jurassic to
Early Eocene ore districts formed in the accreted terranes of western North America (Table 1). A
briefer summary of the Early to middle Cretaceous eastern Eurasian gold districts (Table 2) reflects
the limited available information from that region of the Pacific margin. Similar geologic features are
described that characterize the mainly Palaeozoic lode districts of eastern Australia and South Island,
New Zealand (Table 2). Comparisons and contrasts between the regions are evaluated in a final
discussion to better establish significant relationships between gold veining and orogenesis.
Page 41
Table 1 Main tectonic/geochronological features of synorogenic gold deposits of the Cordilleran orogen, western Pacific Rim
Area a Major districts Major mines Lade Placer Host rocks b Age of Age of Associated fault zones Coeval tecmnism References c
production production veining (Ma) spatially
(reserves) (m. oz Au) associated
(m. oz Au) magmatism (Ma)
Southern coast, Kodiak, Nuka Bay, Chichagof, Hirst- 0.9 0.1 L K turbidities of the 57-49 611-50 Chichagof district = ridge subduction; 7,10,21
AK Moose Pass, Hope- Chichagof, Chugach terrane; Border Ranges and s-s at Chichagof
Sunrise, Cirdwood, Grande, Cliff E Tert plutons Queen Charlotte; other
Pt Wells, Pt districts = small brittle
Valdez, Chichagof to ductile faults and joints
Juneau gold belt, Berner's Bay, Alaska-Juneau, 6.8 ( > 5.0) minor J-K flysch of the Gravina 56-53 100-90, 70-60 subsidiary faults to the change from
SE AK Eagle River, Treadwell, belt; Perm metabasalt; Tr (mnalite sill Sumdum and Fanshaw orthogonal to
Juneau, Kensington phyllite and metagabbro belt), 611-50 systems oblique subduction;
Snettisham, of Taku terrane; mid-K (northern uplift of orogen
W indham Bay monzodior Coast core along eastern
buthulith) side of gold belt
Talkeetna Mts., Willow Creek Independence 0.6 minor L Pz schist of the 66, 57('?) 79-66 Castle Mountain(?) subduetion of 13
south-central Peninsular terrane; L K Chugach tertane;
AK tonalite possible s-s during
oroclinal bending
of AK
Maclaren glacier Valdez Creek Valdez Creek minor 0.5 J-K flysch of the 63-57 78. 70, 66-5.1 Valdez Creek shear zone; oroclinal bending 4,9,20
metamorphic bah, placer deposit Kahiltna terrane; E Tert Denali of AK; possible
south-central AK plutons
dextral s-s; uplift
of flysch belt
SW B.C. Bridge River, Pioneer, 4.3 (0.6) minor L Pz-E Mz basinal rocks, 91-86 270, 91-43 Yelakom orthogonal 12
Coyuihalla Bralorne, arc rocks, and ophiolites collision of
Carolin of Bridge River and Wrangellia, but
Cadwallader terranes possibly coeval
with onset of s-s
East-central AK Fairbanks, Circle, Fort Knox, Ryan 0.3 (5.5) > 11 E Pz quartzite and schist 92-85, IVS-85, 811-65, Stockwork veins and subduction of
Fonymile-Eagle, Lode, True North of the Yukon-Tanana 77 55-SU shears; regional Gravina tlysch belt
Kantishna terrane; mid K diorite association with NE- during collision of
trending s-s faultzones Wrangellia
between Denali and
Tinting faults
Area a Major districts Major mines Lade Placer Host rocks b Age of Age of Associated fault zones Coeval tecmnism References c
Page 43
production production veining (Ma) spatially

(reserves) (m. oz Au) associated
(m. oz Au) magmatism (Ma)
Sewurd Peninsula. Council-Solomon, Nome placer minor 6 E Pz phyllite of the 109 108-82 high-angle normal slab rollback and 6
Fairhaven- Kougarok deposits Seward terrane faults and extensional regional extension
joint sets
Sierra Foothills, Mother Lode, East Jamestown, 35(15) 65 mid-Pz-J grgillile, 144-141, 1511-80 Melones, Bear Mountains, seaward-stepping of 3,11
central CA Gold Belt, West Hammonton, graywacke, chen, and 127- 1118 and other steeply-dipping subduetion zone at
(fold Belt, Columbia. oceanic volcanic rusks of thrust laults 150-140 Ma: onset
Alleghany, Grass Jackson-Plymouth, the northern Sierra, of rapid, orthogonal
Valley Folsom Merded River, Sullivan subduclion at 120
('rrxk, mot I•muhilh Ma
tcrralles
Klamath Mts, French-Gulch, 3.5 3.5 complex group of Pz-1 >_ 147, 177-135 Soap Creek-Siskiyou rapid shifts in 5
northern CA Deadwood oceanic arcs and -< 136 convergence
accretionary prisms velocities
Omenica Klondike, Cassiar, Klondike placer >I > 12 E Pz pericratonic and z 140-13V I8U-190 perhaps regional-scale postcollisional 1,16,18,19
Gegnticline, Cariboo fields miogeoclinal strata of thrust faults in Klondike thermal
Yukon and B.C. Kootenay and Cassiar reequilibration
terranes; L Pz ophiolite
of Slide Mm. terrane;
Perm metafelsic
volcanics of Yukon
Tanana tertane
Northern and Atlin, Stewart Atlin placer fields minor 1 Perm-Tr ophiolite of the 170-165 172-162 outboard collision 2
central B.C. Lake Cache Creek terrane of the Yukon
Tanana terrane
'Areas: AK = Alaska, CA = California, B.C. = British Columbia.

bHost rocks: E = early, L = late, Pz = Palaeozoic, Mz = Mesozoic, K = Cretaceous, Tert = Tertiary, 1 = Jurassic, Tr = Triassic, Perm = Permian.
`References: (I ) Andrew et al. (1983); (2) Ash et al. (1996); (3) Bohlke and Kistler (1986); (4) Davidson et al. (1992); (5) Elder and Cashman (1992); (6) Ford and Snee
(1996); (7) Goldfarb et al. (1986): (8) Goldfarb et al. (1991); (9) Goldfarb et al. (1997); 10. Haeussler et al. (1995); (I I). Landefeld (1988); (12) Leitch et al. (1991); (13)
Madden-McGuire et al. (1989); (14) McCoy et al. (1997); (IS) Miller et al. (1994); (16) Mortensen (1990); (17) Mortensen et al. (1996); (I8) Rushton etal. (1993); (19)
Skelchley et al. (1986); (20) Smith (1981); (21) Taylor et al. (1994).
Page 44
Table 2: Main tectonic/ geochronological features of synorogenic gold deposits of the eastern Pacific Rim
Area Major districts lode Placer Host rexks c Age of Age of Coeval teclonisim c References c
production production veining spatially
(reserves) (m. oz. Au) (Ma) c associated
(m. oz. Au) magmatism
(Ma) c
Yana-Kolyma belt Omchak 8" > 130a L Pz-mid-Mz 135-100 144-134, increased collisional 5,9,12-14
Central part, turbidities of the 120-80 rates between
Russian northeast Kolyma-Omolon Eurasia and Izanagi
terrane plates in E K; slab
rollback and
extension in mid-K
Verkhoyansk belt, Carb.-J. passive mid-K mid-K increased collisional 12,13

western part and margin sedimentary rates between
Allakh-Yun, rocks Eurasia and Izanagi
southern part, plates
Russian northeast
Chukotka belt, numerous opening of the 12,13

northern part, continental margin Canada basin and
Russian northeast and oceanic terranes terrane collision in
mid-K; slab
rollback and
extension in mid-K
Selemdzha-Kerbi, accretionary prism I. K sufxluction 12,13

northwestern part, made up of and collision
Russian southeast Tukuringra-Dzhagdi
and Galam terranes
NE China Shangdong, Jian- > 30b Archaean gneisses 130-120` L Pz-L Tr, subduction-related 6,11,17,20
changgouliang, and schists, L Pz-L 165--150, basement uplift?;
Jinchangyu Tr granitoids 130-120 onset of extension?
Otago, South Macres Flat 1(1) >8 Perm-L Tr J-E K none collisional 10
Island, New turbidities of deformation
Zealand Torlesse and Caples
terranes
New England fold Gympie, Hillgrove > 4(I) Carb-Perm Perm-E Tr 3t16-280, subduction and 2,7,18
belt, eastern turbidities and 2.55-245, accretion; uplift'?
Page 45
Area Major districts lode Placer Host rexks c Age of Age of Coeval teclonisim c References c
production production veining spatially
(reserves) (m. oz. Au) (Ma) c associated
(m. oz. Au) magmatism
(Ma) c
Australia granites of t:'fr

Wandilla
Tablelands and
Gympie terranes
Lachlan fold belt, Bendigo, Ballarat, 32 48 Ord-E Dev flysch L Sil-E 410-390, intraplate thin- 1,16,19
SE Australia Castlemaine, Dev('I), 3711 360 skinned
Stawcll, Mill End M Dev-E deformation;
Carb perhaps subduction
related
Hodgkinson- Hodgkinson 9 Camb-Dev flysch, Sil-Dev, M Ord-M intraplate thin- 3,15,19

Broken River and Charters Tower, M Ord-M Dev L Carb Dev, skinned
Thomson fold Ravenswood granitoids L Carb-E deformation;
belts, NE Perm perhaps subduction
Australia related
Westland, South Reefton 2.5 (1) >5 Ord sandstone, Pz 370-310 uncertain 4,8
Island, New siltstone, and shale
Zealand
'Estimate for entire Russian Far East.

"Estimate for production and reserves.
`Host rock and ages: E = early, M = middle; L = late; Pz = Palaeozoic, Mz = Mesozoic, Camb = Cambrian, Ord = Ordovician, Dev = Devonian, Sil = Silurian, Perm =
Permian, Tr = Triassic, J = Jurassic, K = Cretaceous, Carb = Carboniferous.
'Some of the northeast China deposits are definitely older and probably Hercynian.
`References: (1) Arne et al. (1996); (2) Ashley et al. (1994); (3) Coney (1992); (4) Cooper and Tulloch (1992); (5) Eremin et al. (1994); (6) Geological Survey of Canada
(1991); (7) Gilligan and Barnes (1990); (8) Goldfarb et al. (1995); (9) Goryachev (1994); (10) McKeag and Craw (1989); (11) Nie and Wu (1995); (12) Nokleberg et al.,
1994b; (13) Nokleberg et al. (1996); (14) Parfenov et al. (1996); (15) Peters and Golding (1989); (16) Phillips and Hughes (1996); (17) Poulsen and Mortensen (1993); (18)
Scheiber (1996); (19) Solomon and Groves (1994); (20) Wang et al. (1996).
Page 46
4.2 Tertiary synorogenic gold deposits of the Pacific Rim
The youngest economically significant gold ores hosted in regionally metamorphosed terranes of the
Pacific Rim are those of the southern Alaska forearc (Fig. 3). Gold-bearing veins, now exposed in
districts throughout the Chugach and Kenai Mountains (Fig. 4b) and Alexander Archipelago (Fig.
4a), were emplaced within turbidities of a 2000-km-long accre. tionary prism between 57 and 49 Ma
(Goldfarb et al., 1986; Taylor et al., 1994; Haeussler et al., 1995), More productive districts, such as
near Juneau, are located a few hundred kilometres landward of the prism. They occur adjacent to or
are hosted by parts of the continental margin magmatic arc, and they formed between 66 and 53 Ma
(Goldfarb et al., 1993).
Fig. 3. Distribution of synorogenic gold districts within the Alaskan Cordillera. Absolute dates for
veining in some districts are listed, as determined by a number of workers using K-Ar and (or)
°°Ar/39Ar methods on hydrothermal micas. Generalized composite terranes after Plafker and Berg
(1994).
Page 47
Fig. 4. Gold districts within the Southern margin composite terrane and adjacent parts of the
Wrangellia terrane. (a) Deposits of the Chichagof district in southeastern Alaska. fib) Gold districts
in the Kenai and Chugach Mountains of south-central Alaska.
4.2.1. Gulf of Alaska accretionary prism
Subduction-related tectonics during the earliest Tertiary along the southern Alaska continental
margin, and along the northeasternmost part of the Pacific basin, led to the emplacement of
widespread gold veins in a near-trench environment. Alaska itself is composed of more than fifty
lithotectonic terranes of predominantly oceanic character (Jones et al., 1987; Monger and Berg,
1987). By the Late Cretaceous, most of these terranes were amalgamated and had already been
accreted to the northwestern corner of the North American continent. Subsequently, during the
Palaeocene and earliest Eocene, tectonism in the southern half of Alaska was driven by dextral-
oblique to orthogonal convergence of both the Kula and Farallon plates with the North America plate.
The convergence included emplacement of the outermost part of the Chugach terrane against and
below the southern Alaska margin (Plafker and Berg, 1994). This accretionary complex,
Page 48
predominantly Late Cretaceous to Palaeocene deep sea fan deposits, was highly deformed during
early Tertiary collision and underthrusting.
The Chugach terrane is bounded on its landward side by the steeply-dipping Border Ranges fault
system (Figs. 3 and 4). This crustal-scale thrust fault system separates the Chugach terrane from older
rocks of the inboard Wrangellia composite terrane. A similar fault system, the early Eocene contact
fault system (Fig. 4b), separates the Chugach terrane from slightly younger turbidities of the Prince
William terrane that are exposed along the continental margin of south-central Alaska. However, in
southeastern Alaska, the seaward boundary to the Chugach terrane has been dominated by dextral
translation along the Queen Charlotte-Fairweather fault system (Fig. 3) perhaps since early Eocene
(Plafker and Berg, 1994). As described below, this difference in structural setting along the length of
the accretionary complex likely played a major role in determination of the relative size of
synorogenic gold vein systems.
An early Tertiary high-T, low-P metamorphic event within the Chugach terrane (Hudson and Plafker,
198?) generated magmas and fluids, with the latter being directly responsible for the turbiditehosted
gold-quartz lode systems in the ChugachKenai Mountains of south-central Alaska (Goldfarb et al.,
1986; Fig. 4b) and Alexander Archipelago of southeastern Alaska (Taylor et al., 1994; Fig. 4a). The
event coincides with the final collisional deformation of the allochthonous terrane, but predates tlk
onset of regional uplift by about 5 m.y. (Plafker et al., 1989). The little documented postaccretionaiy,
strike-slip movement on the major terrane-bounding thrust-faulted suture zones in south-central
Alaska limited crustal permeabilities along these major structures. Synorogenic fluid flow was,
therefore, diffuse over a broad area of fracture networks between the suture zones, rather than
focussed into and migrating up the large faults. The resulting small, gold-bearing occurrences are
widespread, but generally subeconomic in south-central Alaska. Where developed, lodes and placers
of the Kodiak. Nuka Bay, Hope-Sunrise/Moose Pass, Girdwood. Port Wells and Port Valdez districts
(Figs. 3 and 4b) have a combined production of only 250.000 oz of Au. In contrast, in the
southeasternmost part of the Chugach terrane, orogen-parallel transcurrent motion along terrane-
bounding faults likely accompanied the Eocene thermal event. Such tectonism would have increased
permeabilities along the major faults and in subsidiary fault zones between the faults. Resulting large
fluid discharges could account for the 800,000 oz of gold recovered from the two main vein systems
of the Chichagof and Hirst-Chichagof deposits (Fig. 4a).
There is a general west to east younging of both 60- to 50-Ma flysch-melt plutons and 57- to 49-Ma
synorogenic gold veins along the entire 2000-km length of the Chugach complex. This delineates
migration of an active slab window (Bradley et al., 1993, 1994; Haeussler et al., 1995) or young crust
adjacent to an inactive window (Rick Saltus, oral commun., 1996). Where igneous rocks and veins
both occur in a district, veins always cut the plutons and suggest hydrothermal mineralization was
perhaps a few million years younger than magmatism. Subduction of a slab window in oceanic crust
migrating downward beneath the base of the prism, likely a part of the spreading Kula-Farallon
spreading ridge, is hypothesized to have contributed to anomalous heating of the prism that composed
the outer forearc (Haeussler et al., 1995). It is possible, however, that the gravity spreading along such
a topographic rise may have stopped once the ridge neared the trench. Nevertheless, simple heat
conduction from the underthrusting of hot, young oceanic lithosphere has the potential for
significantly increasing geothermal gradients at the base of the overlying accretionary prism (Molnar
and England, 1995)
4.2.2. Near-arc and within-arc deposits of southern Alaska
Simultaneous with early Tertiary gold formation along the continental margin of Alaska were the
final stages of magmatic arc development and major gold-veining episodes a few hundred kilometres
inboard from the margin (Goldfarb et al., 1993, 1997). The arc is dominated by Late Cretaceous
through early Eocene subduction-related batholiths that now predominate within the higher elevations
of the Talkeetna Mountains. Alaska Range and Coast Mountains. Productive gold-bearing veins
Page 49
formed along the seaward margin of the arc between about 66 Ma and 53 Ma (Fig. 3, No. 33. 34, 2, 8.
15, 32 and 35). Most were formed in sedimentary rock-dominant terranes within 5-10 km of the arc
rocks; some formed within felsic to intermediate calc-alkaline batholiths of the arc itself, about 5-10
m.y. after crystallization of igneous host rocks.
4.2.2.1. Juneau gold belt
The 200-km-long Juneau gold belt in southeastern Alaska (Berner's Bay, Eagle River, Juneau,
Snettisham and Windham Bay districts of Fig. 3) has yielded about seven million oz of gold and still
contains substantial resources. The deposits of the belt occur within a few kilometres of the Fanshaw
and Sumdum thrust faults, steeply dipping structures that separate rocks of the Yukon-Tanana terrane,
the composited Taku and Wrangellia terranes, and the Gravina overlap assemblage immediately west
of the Coast batholith (Fig. 5). Rocks in these terranes were regionally metamorphosed to lower
greenschist and subgreenschist fades during mid-Cretaceous and older deformational episodes above
the subducting Farallon plate (Himmelberg et al.. 1991). Thrusting along the two closely spaced
faults ceased at 83-71 Ma, but then stepped to the Coast shear zone generally located a few kilometres
eastward (Gehrels, 1998). Ductile deformation along this latter steep structure, which cuts both
sutures near the north end of the Juneau gold belt, occurred between 71 and 59 Ma. The deformation
included rapid regional uplift of the orogen core to the east of the Coast shear zone (Gehrels, 1998),
an area now exposing the immense Coast batholith.
Deformation along the Coast shear zone was accompanied by the deep (4-7.5 kb; i.e., McClelland et
al., 1991 and Gehrels et al., 1992) emplacement of a 700-km-long belt of 5-10-km-thick tonalite sills
at about 70-60 Ma. Sill emplacement was immediately followed by voluminous 60 to 50 Ma
magmatism to the east that formed the shallower (3-4 kb; i.e., McClelland et al., 1991) north part to
the Coast batholith. Most significantly, latest Cretaceous to Palaeocene sill emplacement was
accompanied by an 8-km-wide Barrovian metamorphism extending westerly to the two thrust faults
(Fig. 5; Himmelberg et al., 1991). The fluid volumes generated during this high-T episode are
hypothesized to have migrated into and along these faults, and deposited the gold in veins at 7-15 km
depth within greenschist fades units of the Barrovian sequence (Goldfarb et al., 1997). This 56-53 Ma
fluid flow event occurred simultaneously with a major change in regional stress fields (Goldfarb et
al., 1991). The model requires a period of at least five million years between initial pore fluid
generation and subsequent vein precipitation.
4.2.2.2. Valdez Creek district
A tectonic scenario similar to that observed in the Juneau gold belt also characterizes the gold-rich
Maclaren glacier metamorphic belt. This metamorphic belt is located along the southern side of the
Denali fault system in south-central Alaska. Auriferous lodes are widespread throughout the
greenschist fades units of the belt and have been eroded to yield the 500,000 oz of gold recovered
from placers in the Valdez Creek district (Fig. 3, No. 33; Fig. 6). The belt formed within the Jurassic
to mid-Cretaceous strata of the Kahiltna flysch basin. The basin was closed by middle Late
Cretaceous time during convergence between the Yukon-Tanana terrane to the north and the
subducting Wrangellia terrane to the south, with the latter being carried northward on the recently-
formed Kula plate (Plafker and Berg, 1994). Melt-enhanced deformation (Hollister and Crawford,
1986) of the pelitic rocks between 78 and 68 Ma, perhaps due to the continued subduction from the
south, was recorded within a major deep-crustal ductile shear zone, the 4-5-km-thick Valdez Creek
shear zone (Davidson et al., 1992). During this period, at depths of 25-40 km, a 1-km-thick tonalite
sill was emplaced at 78 Ma within the shear zone and the 70 Ma Susitna batholith crystallized along
the zone's eastern edge (Davidson et al., 1992). A steep Barrovian sequence was overprinted on the
flysch at 78 Ma, extending from the sill seaward (to the south) for 10 km (Smith, 1981). Deformed
metamorphic and igneous rocks began uplifting by 68 Ma and cooled below about 300°C by 62 Ma
(Davidson et al., 1992).
Page 50
Fig. 5. Gold deposits of the Juneau gold belt, southeastern Alaska. Listed dates include 40Ar/'9Ar
measurement on hydrothermal micas (from Goldfarb et al., 1997) and Li-Pb ages of adjacent
batholith (from Gehrels et al.. 1992 and Gehrels, 1990.
Additional magmatism and the synorogenic gold veining occurred during the uplift of the Kahiltna
flysch. A series of dioritic bodies were emplaced between 66 and 54 Ma throughout the greenschist
part of the metamorphic belt. Simultaneously, 63-57 Ma gold-bearing quartz veins (Fig. 6) were
deposited within the cooling diorites and flysch of the Valdez Creek shear zone (Goldfarb et al.,
1993, 1996). Veining was likely aided by the rapid reduction in lithostatic load during the
Palaeocene, which would have induced hydrofracturing and thus increased local permeabilities within
the deformed belt (for example, see Norris and Henley. 1976). The importance to the ore-forming
process of transpressive motions along major crustal structures at this same time is uncertain. But,
gold veining occurred during the onset, and near the axis, of the oroclinal bending of Alaska due to
North America-Eurasia plate collision (Plafker and Berg, 1994). The Denali fault system,
immediately landward of the Susitna batholith, underwent major dextral strike-slip motion during the
bending. Similarly, the Broxson Gulch thrust fault that comprises the suture between the Kahiltna and
Wrangellia terranes a couple of kilometres seaward of the Valdez Creek gold district, also was the
likely focus of Palaeocene strike-slip movement (Nokleberg et al., 1994a).
Page 51
4.2.2.3. Willow Creek district
More than 600,000 oz of gold were recovered from veins in the Willow Creek district of southcentral
Alaska (Fig. 3, No. 34), mainly from fissure veins at the Independence mine. Unlike the Juneau gold
belt and the Valdez Creek district, the ores of the Willow Creek district were almost entirely hosted
within the magmatic arc rocks. Jurassic pelitic rocks of the Peninsular terrane, amalgamated prior to
accretion within the Wrangellia composite terrane, were part of the southern Alaska continental
margin and regionally metamorphosed to greenschist facies by middle Late Cretaceous. Kula plate
subduction along a seaward trench led to accretion of the Chugach accretionary prism throughout the
latter half of the Late Cretaceous. Dioritic to tonalitic plutons, emplaced from 79 to 72 Ma in response
to the continued subduction, formed the Talkeetna Mountains batholith within the Peninsular terrane.
Tonalitic host rocks were rapidly cooled below 300°C by about 70 Ma (Csejtey et al., 1978).
Fig. 6. Distribution of gold-bearing quartz veins of the Valdez Creek district south-central Alaska.
Geology and metamorphic facies generalized from Smith (1981) and Dacidson et al. (1993).
Page 52
Fig. 7. Distribution of synorogenic gold districts of the Canadian Cordillera. Absolute dates for
deposits are listed, as determined by Leitch et al. (1991), Ash et al. (1996). Andrew et al. (1983),
Rushton et al. (1993). Sketchley et al. (1986) and Goldfarb et al. (1995) using K-Ar and (or)
4OAr/;9Ar methods on hydrothermal micas. Terranes generalized from Wheeler et al. (1991).
Gold veining occurred within the southern part of the batholith at 66 Ma (Madden-McGuire et al.,
1989; Steve Harlan, unpubl. data). As in the Valdez Creek district, ore formation post-dates thrusting
within the accreted terrane. Underthrusting of the Chugach terrane beneath the Peninsular terrane did
continue during ore formation (Burleigh, 1987), but gold veining clearly occurred relatively late
during the subduction-related thermal event. Because of this, we speculate that veining might have
been aided by the initiation of oroclinal bending in the Palaeocene. The Border Ranges thrust fault
system (Fig. 3) represents the suture between the Peninsular and Chugach terranes. Pavlis et al.
(1988) indicate that its probable northernmost strand, the Castle :fountain fault located within 7 km of
the gold veins, became inactive in the Late Cretaceous. Subsequently, unlike elsewhere along the
Border Ranges thrust fault system in south-central Alaska, this northern strand may have been
reactivated by strike-slip motion during the oroclinal bending (Pavlis et al., 1988).
Page 53
4.3. Mesozoic synorogenic gold deposits of the Pacific Rim
The Mesozoic in the northern Pacific basin was dominated by Kula-Farallon plate convergence with
North America on the east and Izanagi and Farallon plate collision with Eurasia on the west. In
addition to terrane accretion, many thousands of kilometres of lithosphere were subducted below the
continental margins. Synorogenic gold veining, predominantly in the Early and middle Cretaceous,
was an inherent part of the growth of both continental masses. Simultaneously in the southern Pacific
basin, a final stage of accretion and subduction along the Gondwanan margin temporally correlates
with veining in the Haast schists on New Zealand's South Island.
4.3.1. Mid-Cretaceous deposits of the northern North American

Cordillera
Two regions of the northern Cordillera are characterized by productive mid-Cretaceous gold
concentrations. A broad belt across interior Alaska, including the rich placer fields of the Nome (Fig.
3, No. 7 and 10) and Fairbanks (Fig. 3, No. 9) regions, contains Alaska's oldest gold vein deposits. A
smaller area in the south part of southeastern Alaska and southwestern British Columbia includes the
Bridge River district (Fig. 7), the most productive synorogenic lode system within the Canadian
Cordillera.
4.3.1.1. Interior Alaska
In interior Alaska, mid-Cretaceous gold-veining and felsic magmatism were widespread near
Fairbanks in the Yukon-Tanana terrane and near Nome in the Seward terrane. Although having
undergone a great deal of Mesozoic translation, these continental margin/carbonate platform
sedimentary rock-dominant terranes were already part of the North American cratonic margin for at
least 100 m.y. prior to ore deposition. The Council, Solomon, Fairhaven, and Kougarok districts near
Nome on northwestern Alaska's Seward Peninsula (Fig. 3) have combined to yield about six million
oz of gold from placer accumulations. Gold-bearing veins are very widespread, but are thin and
discontinuous and thus lodes have typically been uneconomical. Historically, Fairbanks and adjacent
smaller gold districts of east-central Alaska defined a major placer-producing region that yielded
more than 11 million oz of gold from the undulating terrain between the Denali and Tintina fault
systems. This part of the Yukon-Tanana terrane, however, has recently been recognized to contain
large-tonnage, bulk-mineable stockwork gold systems (e.g., the Fort Knox deposit) hosted in or near
early Late Cretaceous felsic to intermediate intrusions.
The tectonic regime during the mid-Cretaceous hydrothermal events is problematic. Pavlis (1989)
and Miller and Hudson (1991) indicate a broad extensional event across the northernmost Cordillera
due to lithospheric thinning and increased heat flow. The hypothesized control is slab retreat or
rollback of the subduction zone in the mid-Cretaceous northern Pacific basin. This large magnitude
crustal extention event may have continued into the Russian Far East forming the Okhotsk-Chukotka
magmatic belt.
In northern Alaska, contractional deformation due to Farallon-North America plate convergence

resulted in the 170-130 Ma Brookian orogenic event that occurred 20-60 million years before gold
veining. Deformation included oceanic crust being thrust above the Palaeozoic and older continental
margin pelitic rocks of what became the Seward Peninsula and Brooks Range. During the obduction
of the oceanic rocks, the lower continental plate experienced a high-P/low-T blueschist facies
metamorphism with a subsequent greenschist facies overprint (Till and Dumoulin, 1994).
Postkinematic Barrovian metamorphism and associated mid-Cretaceous magmatism were restricted
to a 250-km-long zone throughout the central and southeastern Seward Peninsula between 108 and 8?
Page 54
Ma (Armstrong et al., 1986; Amato et al., 1994). The initiation of the thermal episode correlates in
time with the Seward Peninsula gold veining located 40-50 km to the south and west (Ford and Snee,
1996). We believe that it is more than coincidence that magmatism and Barrovian metamorphism in
the Nome area were synchronous with veining. But the 40-50-km spacing between the veins and the
thermal front remains difficult to explain.
Trench rollback due to mid-Cretaceous opening of the Canada basin, with the North America and
Eurasia plates advancing on the Farallon plate, is hypothesized as the cause of large-magnitude
crustal extension and heating (Rubin et al., 1995). Most of the recognized gold-bearing veins on the
Seward Peninsula are located within associated small normal faults (Ford and Snee. 1996). The lack
of large, crustal-scale structures within the thermally upgraded part of the Seward Peninsula hindered
significant focusing of ore fluids and prevented formation of high tonnage ore lodes.
Fig. 8. Fairbanks and other gold districts within metamorphosed sedimentary rocks of the Yukon-
Tanana terrane, east-central Alaska and westernmost Yukon. Metamorphic fades after Dusel-Bacon
et al. (1993).
The tectonic setting for the mid-Cretaceous gold ores of the Fairbanks and other districts within the
Yukon-Tanana terrane of east-central Alaska is less clear; most recent geochronology suggests that
the Fairbanks ores formed subsequent to the regional extension. Palaeozoic and older sedimentary
and igneous rocks of the terrane were added to the continental margin and regionally metamorphosed
to both blucschist and amphibolite facies between the Late Triassic and Middle Jurassic. Gold veining
and related shallow-level ( _< 5 km), calc-alkaline magmatism did not begin for another 100 m.y.
(McCoy et al.. 1997). It is uncertain whether the veins and intrusions were products of the widespread
extension within the Yukon-Tanana terrane (Miller and Hudson. 1991), or slightly younger and
renewed subduction below an already thinned-crust (Stanley et al., 1990). In support of the latter,
Pavlis et al. (1993) use argon cooling ages to argue that extension in the terrane ceased by 110 Ma,
about 20 m.y. prior to ore deposition. Gold veining was roughly coeval with final docking of the
Gravina flysch basin and Wrangellia terrane along the Denali fault system with north wardly-directed
Page 55
subduction continuing below the Yukon-Tanana terrane. Where hosted by the mid-Cretaceous
intrusions, gold veins were generally emplaced within 2 m.y. of crystallization (McCoy et al., 1997).
The 600-km-long by 100-km-wide belt of midCretaceous plutons and gold ores, although offset 450
km along the Tinting fault system, continues east of the Yukon-Tanana tetrane and into the North
America craton of Yukon (Mortensen et al., 1996). In Yukon, bulk-mineable gold deposits include
the two million oz Dublin Gulch deposit. The deposits of the Canadian Cordillera indicate that
processes of gold ore formation that impacted the allochthonous rocks of western North America
may, in some cases, have continued inboard of the most landward suture zones.
A more favourable structural setting in the Yukon-Tanana tetrane relative to the Seward terrane may
account for the large, economical lode targets in the former. Northeast-trending strike-slip faults cut
the Yukon-Tanana terrane and link the extensive Tinting and Denali fault systems, and much of the
gold and magmatism is linked along these (Fig. 8; LeLacheur, 1991). More than 400-450 km of
dextral displacement on each of the two fault systems began in the Late Cretaceous (Nokleberg et al.,
1985; Plafker and Berg, 1994) and the NE-trending faults across the enclosed tetrane could be
dilational products of such motion. No matter what their origin, the presence of these large fault zones
crossing the terrane can not be ignored in attempting to understand the distribution of some of the
newly discovered lode deposits in the Yukon-Tanana terrane.
4.3.1.2. Southwestern British Columbia and the south part of southeastern Alaska
Whereas extension predominated in northwestern Alaska, and both extension and contraction were
important in interior Alaska, contraction was clearly dominant during mid-Cretaceous gold formation
along the continental margin to the southeast. The Middle Jurassic docking of the Yukon-Tanana
terrane to North America to form much of what is now the eastern mainland of Alaska was
accompanied by the docking of the juxtaposed Stikine terrane to form much of western British
Columbia (Plafker and Berg, 1994). At about the same time, two small and amalgamated terranes
composed of late Palaeozoic-early Mesozoic basinal and arc rocks and Permian ophiolites, the Bridge
River and Cadwallader terranes, were accreted to the seaward side of the Stikine terrane (Rusmore,
1987). Deformation within these terranes continued into the early Late Cretaceous because of the
outboard accretion of the Wrangellia terrane. The suture between the terranes to the east and the
Wrangellia terrane in the south part of southeastern Alaska and along the southern half of British
Columbia is marked by the oldest synkinematic plutons of the Coast batholith. During the final stages
of orthogonal convergence, immediately prior to the 85 Ma shift to a more northerly oblique
convergence with the birth of the Kula plate (Engebretson et al., 1985), gold veining occurred on both
sides of the south part of the batholith.
More than four million oz of gold were recovered from the lodes of the Bridge River district that are
hosted in rocks of the Bridge River and Cadwallader terranes about 10 km east of the south part of the
Coast batholith (Fig. 7). The veins formed between 91 and 86 Ma along a series of strands of the
terrane-bounding Yalakam thrust fault system (Leitch, 1990; Leitch et al., 1991). Although the suture
between the small host terrines and the Stikine terrane developed tens of millions of years earlier,
Wrangellia terrane collision from the west led to continued east-verging backthrusting and
deformation along these structural zones until about 85 Ma (Rusmore and Woodsworth, 1991).
Subsequently, major dextral slip continued into the Tertiary (Umhoefer and Schiarizza. 1996). Gold
veining appears to have occurred near the end of contractional deformation, but the close temporal
association with the onset of strike-slip motion could have been important. Less productive (1 X0,000
oz of gold production and reserves) gold deposits of the Coquihalla district (Fig. 7) located on the
opposite side of the terrane-bounding fault system and 115 km to the south may mark an offset part of
the same hydrothermal event (Leitch et al., 1991). The 90 Ma ages for gold veining in the Ketchikan
district (FiQ. 3, No. 17; 25,000 oz of Au produced) on the seaward side of the southern part of the
Coast batholith (Goldfarb, unpublished data) indicate ore-forming processes occurred on both sides of
the 100-km-wide batholith during docking of the Wrangellia terrane.
Page 56
4.3.2. Middle Jurassic to Early Cretaceous deposits of western North

America
As the Atlantic Ocean opened at about 180 Ma, tetrane collision and translation was initiated along
the entire length of the Pacific margin of North America. This oldest portion of significant continental
growth during Cordilleran orogenesis was accompanied by a number of major gold vein
emplacement episodes over a 70-million-year-long period. Resulting gold deposits are located
inboard of the subsequently emplaced batholiths of the Canadian Cordillera (Fig. 7) and outboard of
the syn- to postore Sierra Nevada batholith in the conterminous United States (Fig. 9). The latter gold
systems, located in central California, have been the most productive goldfields of the Cordilleran
orogen.
4.3.2.1. Canadian Cordillera
The oldest synorogenic lodes documented in the North American Cordillera occur along the western
margin of the Cache Creek terrane in northernmost and central British Columbia (Fig. 7). In the
former of these two areas, about one million oz of gold were recovered from placer accumulations in
the Atlin district. The Cache Creek terrane consists of Permian and Triassic oceanic crust that was
enclosed between the Stikine and Quesnel terranes in the Early Jurassic. These three terranes, as well
as the YukonTanana terrane to the north, were subsequently accreted to North America by about 180
Ma (Nelson and Mihalynuk, 1993). This final closure of the Anvil Ocean included westward
obduction of the Cache Creek terrane over the Stikine terrane.
Gold-bearing veins in the Atlin and Stewart Lake (central British Columbia) areas were emplaced in
the Cache Creek rocks within 10-15 m.y. of accretion and obduction (Ash et al., 1996). Also between
172 and 162 Ma, felsic plutons derived from underthrust Stikine terrane rocks intruded the Cache
Creek terrane-hosted gold districts (Ash et al., 1996; Mihalynuk et al., 1992). In the Atlin area, the
relatively undeformed nature of the igneous bodies (Mihalynuk et al., 1992) indicates that veining
clearly postdated local deformation. It is unclear how the localization of the lodes relates to the
poorly-exposed thrust contact between the terranes. Veining and plutonism, though, might be
ultimately related to outboard subduction and accretion of the Nisling assemblage (a part of the
pericratonic Yukon-Tanana terrane of Fig. 7) to the Stikine tetrane in the late Middle Jurassic (Currie
and Parrish, 1993).
A second and more extensive gold veining episode occurred sometime between the Middle Jurassic
and Early Cretaceous landward of or on the landward side of the Intermontane Superterrane (the
amalgamated Quesnel, Stikine, Cache Creek and YukonTanana terranes). Veins in the Klondike,
Cassiar and Cariboo districts (Fig. 7) are hosted by pericratonic and displaced lower Palaeozoic
miogeoclinal strata (Kootenay and Cassiar terranes), obducted upper Palaeozoic oceanic crust of the
Anvil Ocean (Slide Mountain terrane), and Permian schist with a metafelsic volcanic rock protolith
(Klondyke Schist assemblage of the Yukon-Tanana terrane). The veins of the Klondike district in
Yukon-Tanana rocks in Yukon, correlated by some with Kootenay temane rocks (Wheeler et al.,
1991), were eroded to form the more than ten million oz of gold recovered from placer
accumulations.
These gold deposits in the Kootenay, Cassiar and Yukon-Tanana terranes are scattered along the
length of the Omenica Geanticline, the uplifted region be-
tween the accreted terranes and older North American craton. Deformation and metamorphism of
rocks in the geanticline occurred during the Late Triassic and Early Jurassic collision of the
Intermontane Superterrane; the deformed rocks were later uplifted in the Early Cretaceous
(Mortensen, 1990). The veins in the geanticline all yield K-Ar dates on micas of about 140-130 Ma
(Andrew et al., 1983; Rushton et al., 1993; Sketchley et al., 1986), but regional magmatism is well
constrained to the Early Jurassic (Johnston et al., 1996).
Page 57
These ore systems in the Omenica Geanticline are notable in that they are the only major Cordilleran
gold vein systems that lack an obvious temporal association with magmatism. The lack of earliest
Cretaceous plutonism could be an indication of little thermal activity at this time. The 140-130 Ma
dates on the veins could be the true age of mineralization, which was caused by rising isotherms over
an extensive period of postthickening thermal reequilibration (Rushton et al., 1993). If this scenario is
correct, veining would postdate deformation and magmatism by perhaps 40-50 m.y. It can not be
ruled out, however, that the mica dates simply could record cooling from earlier vein emplacement
near the end of deformation. Biotites from schists in this part of the Canadian Cordillera have
typically yielded latest Jurassic to earliest Cretaceous K-Ar and Rb-Sr cooling ages, rather than
metamorphic ages (Johnston et al., 1996).
Fig. 9. Late Jurassic to Early Cretaceous synorogenic gold deposits of the Sierra foothills and
Klamath Mountains along the western margin of the conterminous United States. Farther inboard.
Late Cretaceous gold-bearing veins in some of the allochthonous terranes adjacent to the Idaho
batholith may also be associated with Cordilleran orogenesis. Terranes generalized from Silberling
et al. (1992).
4.3.2.2. Conterminous United States
As was much of the northern Cordillera, the southern part of the Cordillera comprisipg the western
edge of the conterminous United States was also a passive margin throughout the early Palaeozoic. In
the late Palaeozoic, magmatic arcs developed on oceanic crust, probably fringing a rifted North
American margin (Burchfiel et al., 1992). By the Early Triassic, long-term convergence and
westward growth had begun along the Pacific margin of the conterminous United States with
collision and accretion of the extensive northern Sierran-Klamath arc (Sonomia). These arc rocks
have been termed the northern Sierra terrane (or Shoo Fly Complex) and their eastern margin is now
marked by the Sierra Nevada batholith. During a 50-m.y.-long period in the Early and Middle
Jurassic, the Merced River (or Calveras Complex), Sullivan Creek and Foothills terranes were
Page 58
emplaced as narrow strips along the western edge of the northern Sierra terrane in what are now the
western foothills of the Sierra Nevada range in central California (Fig. 9; Paterson and Sharp, 1991).
In the Late Jurassic, subduction stepped outward with accretion of the Franciscan melange that is now
exposed in the California Coast Ranges and a magmatic arc began to develop landward on the
previously accreted terranes. The presence of Late Jurassic sediments in the Great Valley forearc
region indicates that uplift of the terranes was also beginning at this time. Ductile deformation and
metamorphism of the accreted oceanic arcs occurred between 155 and 123 Ma (Tobisch et al., 1989).
Plutons were emplaced in these accreted terranes between 151 and 115 Ma and also several tens of
kilometres to the east as the Sierra Nevada batholith between 150 and 80 Ma; the bulk of the batholith
was emplaced after 120 Ma. Magmatism is generally considered to be a consequence of the initiation
of the Late Jurassic, east-dipping subduction system outboard of the accreted terranes and rift basins
(Burchfiel et al., 1992).
The tectonic setting of gold deposits in what is now central California is remarkably similar to that of
the Juneau gold belt in southeastern Alaska; that is, auriferous lodes are located within a series of
extremely narrow accreted terranes and typically 1020 km seaward of a major Andean-type batholith
(Fig. 9). Steeply-dipping thrust faults between the terrane slivers in California, such as the mid-
Jurassic Melones fault zone, are sites of extensive gold veining. But unlike the Alaskan example,
timing of gold genesis throughout the Sierran foothills is poorly constrained. A few K-Ar and Rb-Sr
dates from scattered deposits suggest that lodes in the Grass Valley district may have formed during
early magmatism at 144 Ma (10 m.y. subsequent to the onset of regional uplift), and then from 125 to
110 Ma along terrane-bounding faults in some of the districts of the Mother Lode belt (Bohlke and
Kistler, 1986). The lodes in the Jurassic accreted terranes have combined to yield about 35 million oz
of gold. The age of gold veining in the northern Sierra terrane, the backstop to which other terranes
were accreted in the Jurassic, is unknown. Although lode production in this innermost part of the
forearc was relatively minor, most of the approximately 65 million oz of placer production was
derived from here.
The same Mesozoic terrane accretion event was continuous northward in rocks now exposed as the
Klamath Mountains of northern California and southern Oregon. Two composite terranes of
predominantly Palaeozoic island arc successions, the Eastern Klamath and the Central Metamorphic
terranes, had been added to the North American margin by Early Triassic time (Burchfiel et al.,
1992). A complex group of Palaeozoic to mid-Mesozoic terranes consisting of far-traveled oceanic
arcs and more locally derived accretionary prisms were accreted to these along the Siskiyou thrust
fault system during the Middle and Late Jurassic. Outboard accretion of the Franciscan complex
accompanied continued subduction in the Early Cretaceous. Magmatism was widespread on both
sides of the Jurassic suture with the older composite terranes between 177 and 135 Ma (Hacker et al.,
1995). Regional metamorphism and deformation along the suture also continued until about 135 Ma
and was likely continuous with Cordilleran orogenesis to the south in the Sierran foothills (Harper et
al., 1994). However, unlike that of the Sierra Nevada range and its western foothills, voluminous
magmatism in the Klamath Mountains did not continue through the remainder of the Cretaceous.
About seven million oz of gold were recovered from the districts in the Klamath Mountains (Fig. 9),
half from lodes and halt from associated placers. Some gold veining is at least as old as 147 Ma,
whereas other veining is probably younger than 136 Ma (Elder and Cashman, 1992). Most of the
veining, irregardless of age`, is concentrated in the backstop to the Middle Jurassic accretion.
Danielson et al. (1990) reported a K-Ar age of 396 Ma for gold veining in Devonian greenstone of
the eastern Klamath terrane. This date is roughly coeval with precollisional earliest deformation and
metamorphism of the terrane somewhere to the west of the Palaeozoic continental margin, and much
older than that of any of the other western North America gold ores. However, the K-Ar age is based
on analysis of hydrothetmal paragonite and it is unclear if the analyzed potassium was actually
derived from the K-poor mica rather than inherited from older wall rocks.
Page 59
Fig. 10. Major gold belts of northeastern Asia and tectonic map of the region generalized from
Sengor and Natal'in (1996) and Nokleberg et al. (1994b, 1996).
The far-field tectonic controls on latest Jurassic to Early Cretaceous gold formation in the
southwestern part of the Cordillera are problematic. Landefeld (1988) suggested that the 100 km
seaward stepping of the subduction zone between about 150 and 130 Ma was an important ore-
forming trigger in the Mother Lode belt. Such a process was shown to induce a rise in geotherms up
through the accreted terranes to the east. Elder and Cashman (1992) suggest 142-133 Ma changes in
Farallon and North America plate velocities and in the related stress fields, were the ultimate control
on earliest veining in the Klamath Mountains. They hypothesize an extremely low convergence rate
during this time as significant for an increased geothermal gradient that drove hydrothermal
circulation within upper plate rocks. Latest veins in the Klamath Mountains, however, were deposited
during a period of accelerated subduction, renewed regional contraction, and thrust faulting at about
130 Ma (Elder and Cashman, 1992). In contrast to the Juneau gold belt, where major changes in plate
convergence directions in the northern Pacific basin appeared important for ore genesis, these data
from the southern Cordillera suggest solely changes in subduction velocities and convergence angles
are adequate far-field controls on ore genesis.
Page 60
4.3.3. Eastern Eurasia
The major synorogenic gold deposits of the Russian Far East occur in five metallogenic belts (Fig.
10) that include Yana-Kolyma. Chukotka, Verkhoyansk, Allakh-Yun and Selemdzha-Kerbi
(Nokleberg et al., 1996). It remains difficult to get an estimate of tonnage for many of the
mesothermal-type deposits in these gold belts, but Goryachev (1994) has estimated a combined past
production of about 8 million oz lode gold and > 130 million oz placer gold for these belts in the
Russian Far East. Vein emplacement in all of the cited belts is mainly associated with Late Jurassic
and Early Cretaceous collisional, deformational, and metamorphic events along the northeastern
Eurasia margin (Nokleberg et al., 1994b). The majority of the known lodes are hosted in greenschist
facies turbidities comprising numerous accreted terranes. We speculate that the order-ofmagnitude
increase in convergence rates between the Eurasian and Izanagi plates at about 135 Ma (Engebretson
et al., 1985) was a driving force for the increased collision.
The specifics and timing of ore genesis in the Russian Far East are still poorly understood. Upper
Palaeozoic through middle Mesozoic turbidities of the Kular-Nera terrane of the Kolyma-Omolon
composite `superterrane' (Nokleberg et al., 1994b, 1996) host the most economically significant lode
ores in the Yana-Kolyma belt, mainly recognized in the deposits of the Omchak district. The gold-
bearing veins of the belt formed between 135-100 Ma (Eremin et al., 1994), during regional
deformation associated with final accretion of the superterrane to the Eurasian craton margin.
Widespread granite plutonism at 144-134 Ma accompanied the collision (Parfenov et al., 1996). Ore
formation and magmatism continue landward into the Carboniferous to Jurassic passive margin
sedimentary rocks of the Verkhoyansk and Allakh-Yun belts. Rocks of these belts were deformed
along the 2000-km-long margin of the eastern Siberian platform (Angara craton) during the collision
(Zonenshain et al., 1990).
At the same time, opening of the Canada basin to the northeast formed the Arctic Ocean and led to
collision of the Palaeozoic and early Mesozoic continental margin sedimentary rocks amalgamated
into the Chukotka superterrane and of a number of smaller oceanic terranes with the Kolyma-Omolon
superterrane (Nokleberg et al., 1994b, 1996). Gold-bearing veins of the Chukotka belt formed in
some of these terranes during the middle Cretaceous (Goryachev, 1996) as the terranes were being
deformed to generate the Anyuy fold belt (Harbert et al., 1990). Major crustal extension emplaced the
120-80 Ma Okhotsk -Chukotka magmatic arc within parts of the Anyuy fold belt and Kolyma-
Omolon superterrane during the final stages of accretion and deformation. Therefore, it is possible
that the same tectonic event that ultimately controlled Albian gold veining near Nome and elsewhere
on the Seward Peninsula of northwestern Alaska was also significant for ore formation in the Russian
northeast.
Continuing to the south in the Russian Far East, gold veins of the Selemdzha-Kerbi metallogenic belt
are hosted in the accretionary prism that includes rocks of the Tukuringra-Dzhagdi and Galam
terranes (Nokleberg et al., 1994b, 1996). Veining in this region is located a few tens of kilometres
seaward of the Precambrian Khingan-Bureya massif along the western side of the Sikhote-Alin-
Sakhalin fold belt. Most of the 140-110 Ma lodes produced only a few tens of thousands of ounces of
gold, but related placers have yielded more than 30 million oz of gold during the last seventy years
lRatkin, 1995). The fold belt is mainly comprised of terranes of Lower Cretaceous tlysch deposited in
back arc basins. It was deformed during Cretaceous subduction and collision of the Sikhote-Alin
terrane (Zonenshain et al., 1990).
Gold occurrences of the same age continue southward into the Archaean rocks of the eastern part of
the north China craton (Fig. 10), emplaced during what has been generally known in China as the
Middle Jurassic through mid-Cretaceous Yanshanian orogen. These lodes, notably concentrated on
the Shangdong Peninsula ( > 20 million oz of gold; Yang et al., 1996) and in provinces north and
northeast of Beijing, are China's most productive gold producers. They are associated with Early
Page 61
Cretaceous plutons generally stated to be subduction-related (Tian, 1992). However, Wang et al.
(1996) claim that by about 125 Ma, mantle plume activity initiated a 100-m.y.-long period of
extension along the eastern Asian continental margin and the 1?6-120 Ma deposits of the Shangdong
Peninsula overlap the onset of this tectonism. The Shangdong deposits are also spatially associated
with eastern strands of the perhaps 5000-km-long Tan Lu wrench fault system that extends from the
Yangtze River north to the Russian Far East (Xu, 1993). In the latter region, it may be associated with
some of the gold deposits of the Selemdzha-Kerbi metallogenic belt. Xu (1993) noted more than 700
km of strike slip motion along the fault system since Late Jurassic.
Similar ages for vein emplacement of 130-121 Ma have been determined in the Chifeng area of
southeastern Inner Mongolia and western Liaoning province (Geological Survey of Canada, 1991;
Trumbull et al., 1996). The spatial relation with alkalic igneous phases (Nie and Wu. 1995), the lack
of extensive alteration, consistent brittle nature of the host structures, Ag-rich nature of many of the
ores, and the hosting of ores in inboard basement uplifts along the northern border of the craton are
atypical of synorogenic lodes. These features have led some workers to suggest that these systems are
more like the precious metal-bearing veins of the Colorado mineral belt rather than Cordilleran-type
mesothermal veins (Poulsen et al., 1990). But some deposits, including the Jinchangyu gold deposit
in eastern Hebei (1.5 million oz of gold production and reserves), appear more like typical circum-
Pacific synorogenic gold lodes (Poulsen and Mortensen, 1993).
It is uncertain as to how far to the west do these Cretaceous gold deposits continue across north
China. Sulphide-poor gold vein deposits associated with alkalic igneous rocks extend 600 km west of
Beijing to the Batou and Bayan Obo areas of western Inner Mongolia. Whereas some of these
deposits and magmatic bodies to the west could also be products of Yanshanian orogenesis, many
formed during the late Palaeozoic Hercynian orogeny (Nie and Wu, 1995), as the Altaid Ocean was
consumed between the colliding North China and Angara cratons.
4.3.4. Otago, South Island, New Zealand
Turbidities of the Permian to Late Triassic Torlesse and Caples terranes along the eastern edge of
New Zealand were the final tectonostratigraphic units added to the Gondwanan margin in the
southwestern Pacific. Deformation and metamorphism of these terranes occurred during Early
Jurassic to Early Cretaceous collision (Bishop et al., 1985). Subsequently uplifted and exposed
auriferous greenschist fades rocks of the amalgamation, often termed the Haast schists, are mainly
exposed in the southeastern part of South Island. Final accretion of the terranes to the continental
margin took place in the mid-Cretaceous (Bradshaw, 1994).
Despite the production of more than eight million oz of placer gold, lode production from the Haast
schists has historically been negligible. However, during the last few years, mining of more than two
million oz of proven reserves has begun in the Macres Flat belt of deposits. McKeag and Craw (1989)
hypothesize that gold vein emplacement in the presently exposed Haast schists of the terranes was
continuous throughout much of the period of Jurassic and Early Cretaceous collisional deformation.
Absolute dates are lacking for the major deposits in the schists of the South Island, but Adams and
Graham (1993) have reported a range of Early Cretaceous dates for small auriferous veins in the
Torlesse terrane near Wellington on the North Island. Like North America's Klondike district, the
Mesozoic gold deposits of New Zealand also lack a spatial association with coeval igneous rocks.
Late Triassic to Early Jurassic arc magmatism is only recognized a few hundred kilometres west of
the major gold lodes.
4.4. Palaeozoic synorogenic gold deposits of the Pacific Rim
The Tasman Orogenic Belt formed along the eastern margin of Gondwanaland over a period of about
200 m.y. beginning in the Cambrian. Quartz-rich turbidite sequences were amalgamated and
deformed during much of this orogeny through episodic periods of contraction and strike-slip
Page 62
movement. This eastward growth of the Indo-Australian plate margin included early through middle
Palaeozoics development of the Lachlan, Thomson, and HodgkinsonBroken River fold belts and
middle to late Palaeozoic formation of the New England fold belt in what is now eastern Australia
(Fig. 11). The early to middle Palaeozoic deformation also impacted rocks now exposed in northern
Victoria Land, Antarctica and Westland, South Island, New Zealand. Significant goldfields
developed during orogenesis within the Lachlan fold belt, with less productive provinces also
recognized in the Hodgkinson-Broken River fold belt and Westland.
4.4.1. Lachlan fold belt
The Lachlan fold belt is part of the more extensive Tasman orogenic belt that comprises Australia on
the eastern side of the suture with its Precambrian craton. The Lachlan fold belt includes a succession
of Cambrian island-arc volcanic rocks (mostly basalts) overlain by a thick and widespread Ordovician
to Early Devonian flysch sequence representing a huge turbidite fan shed from the Ross-Delamerian
orogen to the west. Plutonism in the belt was most abundant between 420 and 360 Ma, with distinct
pulses near Stawell at 410-390 Ma and in central Victoria at 370-360 Ma (Ramsay and VandenBerg,
1986). The major period of Tabberabberan deformation (generally assumed to be 386-378 Ma) was
accompanied by mainly low-grade regional metamorphism, then followed rapidly by emplacement of
dyke swarms, I-type and S-type plutonism, and acid volcanism in central Victoria (Collins and
Vernon, 1992). Carboniferous molasse-style sedimentary rocks, including red beds, are only mildly
deformed. Unusual features of the Lachlan fold belt are its 800-km E-W width, widespread
greenschist fades rocks with a paucity of higher-grade units, lack of Precambrian basement, similar
structural levels exposed across the entire fold belt, and large volume of acid igneous rocks (Coney,
1992).
Fig. 11. Gold provinces and major gold districts of eastern Australia generalized from Solomon and
Groves (1994).
Production from the Victorian sector of the Lachlan fold belt (Fig. 11) has been outstanding on a
world scale: ca. 80 million oz of Au, mostly mined prior to 1910. Forty percent has come from quartz
vein deposits, and the rest from placer and palaeoplacer deposits that are traceable directly to known
lodes, or at least to the more highly mineralized gold districts. There are twelve districts that have
Page 63
yielded at least one million oz of gold, with Bendigo (23 million oz), Ballarat (13 million oz) and
Castlemaine (5.5 million oz) being the largest of these (Phillips and Hughes, 1995, 1996). These three
districts, most of the other one million ounce districts, and many of the smaller ones are within what
has been informally termed the Ballarat zone of central Victoria. Other geological zones to the east
and west of the Ballarat zone have considerably less gold, particularly in the far east and west.
Essentially all primary gold has come from rocks that predate the Carboniferous, and virtually all
rocks older than this age contain auriferous veins somewhere in the province. Major primary deposits
are in black slates and mafic dykes, and together these Candor Fe- rich hosts contained veins that
accounted for at least 75% of production (Phillips and Hughes, 1995, 1996). Alteration surrounding
deposits includes broad carbonate zones, muscovite or biotite envelopes, and narrow sulphide
intervals close to gold. All deposits are structurally controlled, but the type of local hosting structure
varies between goldfields and even within single fields. The major goldfields are characterized by a
pronounced repetition of hosting structures. Regionally, deposits are spatially associated with major
west-dipping reverse faults (Cox et al., 1991).
A diachronous period of gold mineralization in Victoria can be constrained for deposits that occur in
Devonian host rocks to being mainly Middle to Late Devonian in age, in part contemporaneous with
the major Tabberabberan thermal and deformational event. For deposits in older host rocks, a similar
age of mineralization is permissible but rarely demonstrated. Along the eastern edge of the Lachlan
fold belt, in the Hill End goldfield of New South Wales, argon geochronology indicates some
significant gold veining continued into the early Carboniferous (Lu et al., 1996). Arne et al. (1996)
have recently used U-Pb data from Devonian igneous rocks and crosscutting relationships between
these rocks and gold systems to suggest locally older ore deposition in the Lachlan fold belt. They
argue that some of the gold occurrences in the Stawell area are likely Late Silurian to Early Devonian
in age.
Two distinct periods of synorogenic gold veining appear to characterize the north part of the Tasman
orogenic belt - an economically significant Silurian-Devonian event and a less important
Carboniferous event. Flysch units of Silurian to Devonian age in the Hodgkinson-Broken River fold
belt (Fig. 1) are host to Late Carboniferous gold-quartz veins of the Hodgkinson district (Peters et al.,
1990). These are very similar in many ways to the Victorian gold deposits, but have had only minor
production. Although the deposits are hosted in flysch, they are surrounded by many of the
posttectonic Carboniferous batholiths of the north Queensland magmatic province.
The gold vein deposits of the Charters TowerRavenswood area (located a few tens of kilometres
south of the Hodgkinson-Broken River fold belt) have yielded 7.3 million oz of gold mostly from the
Charters Tower district. They might mark a southern extension of the synorogenic gold deposits in
the north part of the Tasman Orogenic Belt (Solomon and Groves, 1994). The veins at Charters
Tower cut Middle Ordovician to Middle Devonian plutons within the deformed flysch, which is often
termed the Thomson fold belt. Peters and Golding (1989) report a 40$ ± 30 Ma date for Charters
Tower ore formation, roughly coeval with early gold formation in the Stawell area of the Lachlan fold
belt. But, some of the gold in the area is definitely younger and of the same age as veins in the
Hodgkinson district. In the Ravenswood district, 313-296 Ma veins cut an early Palaeozoic tonalite
that intrudes an Ordovician sedimentary and volcanic rock sequence (Solomon and Groves. 1994).
Whereas the younger orogens of the northern Pacific basin were dominated by collision of multiple
allochthonous terranes and largely crustal thickening, the Lachlan fold belt and its northern
extensions, the Thomson and Hodgkinson-Broken River fold belts, were characterized by poorly
understood intraplate thin-skinned tectonics (Coney, 1992). Shortening and folding were major
components of deformation in the Tasman orogenic belt, but major uplift of deep crustal rocks did not
occur. Many workers now argue for a subduction-related event beginning in the Middle Devonian
during PacificAustralian plate collision (Powell and Li, 1994; Collins, 1996). But if such an event did
occur, crustal response within the Lachlan fold belt was certainly quite different than within the slices
of terranes along the North American Cordillera and northeastern Asia margin.
Page 64
4.4.2. Buller terrane, Westland, New Zealand
The outermost part of the Tasman orogenic belt throughout the early Palaeozoic was, in part, what are
now the rocks of the Buller terrane that are exposed in the Westland area (Fig. 1) along the western
margin of South Island of New Zealand. These rocks were deformed and metamorphosed along the
southeastern Gondwanan margin during the Silurian and Early Devonian and then were widely
intruded by S-type granites during the Late Devonian through Carboniferous (Cooper and Tulloch,
1992). It is uncertain whether the widespread gold veins within the Buller terrane, including the
productive lodes of the Reefton district (about 2.5 million oz of gold), were emplaced during the
magmatism or during earlier periods of deformation (Goldfarb et al., 1995). The Late Cretaceous
breakup of southern Gondwanaland further separated these Palaeozoic lodes from the other
Palaeozoic gold systems of southeastern Australia.
4.4.3. New England Fold Belt
The New England fold belt (Fig. 11) is the youngest part of the Tasman orogenic belt. It developed
during latest Devonian to Early Triassic accretion and subduction of oceanic arc and marine
sedimentary rock-dominant terranes along the eastern side of the Lachlan and Thomson fold belts
(Gilligan and Barnes, 1990; Scheiber. 1996). Within this part of the orogen, the Gympie deposits of
southeastern Queensland have yielded 3.4 million ounces of gold from lodes in hosted in the Gympie
tetrane (Kitch and Murphy, 1990), an accretionary prism of Permian rocks added to eastern Australia
by the Middle Triassic (Scheiber, 1996) along the seaward margin of the fold belt. The Gympie
terrane was accreted to the Wandilla-Tablelands terrane, a turbidite-rich sequence that had originally
collided with eastern Australia in the Late Carboniferous. The WandillaTablelands terrane
subsequently underwent Early Permian strike-slip movement prior to final collision in the Middle
Permian (Scheiber, 1996). The terrane hosts the Hillgrove gold district, in northeastern New South
Wales, comprised of a number of deposits with a combined historical production of 600,000 oz of
gold (Suppel, 1975).
Veining in the New England fold belt is Permian to Triassic in age and thus appears to be somehow
related to the final period of subduction and accretion along eastern Australia. Late Carboniferous
plutons were emplaced in the Wandilla-Tablelands terrane during its initial collision with the
continental margin and were deformed during the Middle Permian final collision. Immediately
thereafter, at about 250 Ma, gold veining in the Hillgrove district occurred during regional uplift and
perhaps along dilational zones between two major fault systems (Ashley et al., 1994). Age constraints
are poor for veins in the outboard Gympie terrane, but most likely veins are coeval with Early
Triassic dykes found in the region Witch and Murphy, 1990).
Fig. 12. A variety of tectonic scenarios (after Goldfarb et al., 1993 and Haeussler et al., 1995) are
associated with mid-crustal heating that eventually caused gold vein emplacement in the relatively
young Alaskan part of the Pacific rim. (A) A few hundred kilometres landward of the Pacific margin,
crustal thickening and melt-enhanced deformation are associated with Eocene ore formation during
changes in regional stress fields. Adjacent to the margin, subduction of an oceanic ridge led to ore
formation in the accretionary prism. (B) Mid- to Late Cretaceous extensional heating in interior
Alaska led to 110 Ma gold vein emplacement near Nome in the Arctic Alaska composite terrane.
Veining at 90 Ma near Fairbanks in the Yukon-Tanana terrane is related either to the final stages of
the extension or subsequent, renewed convergence.
Page 65
4.5. Discussion
Goldfields of the circum-Pacific rim (Fig. 1) have many features in common, and together define the
world's historically most productive region for the `gold-only' slate belt type of deposit as defined by
Phillips and Powell (1993). Auriferous quartz veins are consistently distributed in pelitic and
psammatic rock sequences that have been metamorphosed to low to medium grades. The more
productive lodes are spatially associated with deep crustal, steep thrusts that, in many cases, are
terrane boundaries. Adjacent to the first-order structures, relatively competent, older igneous rocks
and relatively carbonaceous sedimentary rocks typically provide secondorder structural or chemical
traps, respectively, for ore. Deposits are similar in that goldailver ratios of most productive systems
range from 1:1 to 10:1, deposits show great vertical continuity (1-2 km) despite limited surface
exposure, and low-salinity ore fluids are consistently CO,-rich and characterized by isotopically
heavy hydrogen and oxygen compositions. Plutonism is often coeval with gold vein emplacement.
But, with the exception of some intrusive bodies in east-central and south-central Alaska, the
temporally associated magmatic rocks rarely host ore-bearing veins. For example, about 20 percent of
the Lachlan fold belt is comprised of Palaeozoic granites, but none of the twenty largest Victorian
lode producers are hosted by the granites (Hughes et al., 1996).
In general, subduction and associated mountain building processes along the periphery of a continent
include extensive mid-crustal fluid flow events leading to the generation of synorogenic gold veins.
In the northern Pacific basin, such events characterized the period between 170 and 50 Ma along the
growing margins of North America and Eurasia. In the southwestern Pacific, veining was
concentrated along the active Gondwanan margin in the Devonian with lesser activity associated with
Permian and Jurassic-Early Cretaceous collisions. Significant synorogenic gold lodes are lacking
along the Andean margin of South America. This is, however, an eroding and not a growing margin,
and therefore lacks abundant young allochthonous terranes in the very narrow zone between the
trench and Andean arc. Perhaps this fact indicates that a key to ore formation is a growing margin;
i.e., one that contains a potential huge crustal fluid volume added into the margin within hydrous
silicate mineral phases of the accreted marine strata that undergo subsequent dehydration when
heated from below.
Observations from the various districts around the Pacific Rim make it clear that there is no single
type of tectonic process that provides the heat to drive the necessary volumes of fluid during
continental growth. The onset of orogenic processes above a downgoing slab are crucial for gold
genesis along a subductiondominated margin because very low heat flow typifies subduction zone
environments (cf. Hyndman and Lewis, 1995). Most commonly along western North America, it is
some type of Andean-style orogenesis above the subducting oceanic slab that causes the required
steep rise in geothermal gradients (Fig. 12). The partial melting of asthenospheric mantle above the
sinking slab initiates a series of thermal events that includes devolatilization of upper plate rocks
previously added to the continental margin and the resultant hydrothermal fluid circulation. From the
relatively detailed work in the Alaskan Cordillera, it is apparent that other processes such as
subduction of a spreading ridge or an extensional period during slab rollback also can provide heat
needed to initiate fluid flow and the resulting gold veining (Fig. 12).
Relations from the North American Cordillera indicate that ore formation can occur on the seaward
side of the subduction-related arc (Juneau gold belt, Mother Lode), directly within the magmatic arc
rocks (Willow Creek), and on the landward side of the arc (Bridge River). Where there is a lack of
temporal association with large-scale melting and magmatism, gold deposits are uncommon. But, the
Klondike and other 140 Ma districts in western Canada and the Otago region of New Zealand seem
to be exceptions. Although simply an increased thermal gradient following collisonal and
overthrusting events could have induced metamorphic dehydration in these latter areas, a deep-seated
heat source characterized by an abundance of unexposed intrusions is also certainly possible. In the
Canadian case, the 140 Ma dates may also simply be cooling ages and ore formation might have
occurred simultaneously with Early Jurassic magmatism.
Page 66
Data from the circum-Pacific goldfields indicate that subduction-related magmatism may predate, be
synchronous with, or postdate vein formation. Interpretation of detailed geochronology from the
relatively young Alaskan ores indicates the same range of possibilities for regional uplift. Certainly
when simultaneous, uplift can enhance fluid migration by reducing lithostatic loads. But data show
gold vein emplacement can predate uplift by 5 m.y. (Chugach-Kenai Mountains districts), occur
during a broad period of regional uplift (Juneau gold belt), or postdate uplift by 5 m.y. (Willow Creek
district). The spatial associations between greenschist facies and gold vein emplacement in the narrow
inverted Barrovian sequences of the Juneau gold belt and Valdez Creek district suggest ore genesis-
related to metamorphic processes. The required periods of about 5 m.y. and 15 m.y., respectively,
between metamorphic devolatilization and fluid focusing suggests a complex hydrologic history.
Such a history is strongly influenced by the evolving crustal thermal structure that would vary from
district to district, a structure whose understanding requires detailed numerical modeling specific for
each district.
Determination of absolute ages for the Mesozoic and Tertiary gold systems allow us to recognize
timing relationships between tectonic processes and ore genesis during orogeny. Along the
accretionary prism of southern Alaska, Tertiary veining coincident with the subduction of a ridge
segment was essentially coeval with accretion and deformation of the host terrane. A more typical
scenario may exist a few hundred kilometres inboard from accretionary prisms throughout the
Cordilleran orogen. Deposits formed here in host rocks tens of millions of years after their accretion
and the related initiation of deformation; i.e., at least 30-80 m.y. later in the Juneau gold belt, much of
the Canadian Cordillera, the Sierra foothills, and the Klamath Mountains. Dating of mid-Cretaceous
deposits in interior Alaska indicates that, in certain cases, gold veining can postdate accretion and
associated deformation by more than 100 m.y. Host rocks for ores in east-central Alaska and
northwestern Alaska had already cooled to relatively low temperatures and been uplifted to relatively
shallow depths by the time of mid-Cretaceous crustal heating. Also, thrusting along sutures or zones
of melt-enhanced deformation within the ore-host terranes ceased prior to veining.
In most cases, however, slab subduction and underthrusting continued beneath the earlier deformed
terranes during veining. Ores are typically emplaced in or near orogen cores after back-stepping of
the subduction zones along the growing continental margin (cf. Landefeld, 1988). Orogenic processes
taking place at depth above a sinking slab and 100-200 km landward of the subduction zone, will still
have a major impact on shallower rock sequences that already have undergone much of their
deformation. Additionally, reconfiguration of plate subduction patterns could be important in the
development of some of the larger gold districts in the overlying upper plate. For example, the onset
of orogen-parallel tectonism during a more oblique Eocene convergence along southeastern Alaska or
dramatic shifts in midMesozoic plate convergence rates along the western conterminous United
States may have been critical for the formation of the Juneau and California gold fields, respectively.
Gold veining is an inherent part of continental growth, whether by intraplate or interplate tectonics.
As pointed out by Coney (1992), such contrasting, tectonic styles characterize both the gold-rich
Tasman and Cordilleran circum-Pacific orogens. The Tasman system consists of a single composite
terrane in which deformation and magmatism were controlled by a subduction zone to the east. In
contrast, western North America and eastern Eurasia were built through the collision and deformation
of dozens of diverse terranes leading to widespread thrusting and the exposure of different crustal
levels. Hence, it is the thermal regime and not a single tectonic style of the growing margins that
played the major role in localizing the circum-Pacific gold ores.
Page 67
We thank David Groves and Ross Ramsey for suggesting the topic of this summary and for the
opportunity to present much of the material at the meeting on "Mesothermal Gold: A Worldwide
Connection", held in Perth during July, 1996. Discussions with Craig Hart, Larry Snee, Leslie
Landefeld, Rick Saltus, Feng-Jun Nie, Tony Christie, Lance D. Miller, Dwight Bradley, Peter P.
Haeussler, Roger Powell, Martin Hughes, Julian Vearncombe and Jonathan Law improved our
understanding of the Pac-Rim tectonics, metallogenesis, and geochronology in various geographic
areas. Vladimir I. Shpikerman and Nikolai A. Goryachev (Northeast Science Center, Russian
Academy of Sciences, Magadan) and Vladimir V. Ratkin (Far East Geological Institute, Russian
Academy of Sciences, Vladivostok) provided critical information on the Russian Far East deposits.
The manuscript has been greatly improved through the very constructive reviews by Rick Saltus,
Byron Bergen Andy Wilde and an anonymous journal reviewer.
4.7 References
Adams. C.J., Graham. LJ.. 1993. K-Ar and Rb-Sr age studies of the metamorphism and quartz vein
Au mineralisation on Terawhiti Hill, near Wellington. New Zealand. Chem. Geol. (Isotope
Geosci. Sect.) 103. 235-249.
Amato, J.M., Wright. J.E., Gans. P.B., Miller, E.L., 1994. Magmatically induced metamorphism and
deformation in the Kigluaik gneiss dome, Seward Peninsula, Alaska. Tectonics 13, 515-527.
Andrew, A., Godwin, C.L. Sinclair. A.J., 1983. Age and genesis of Cariboo gold mineralization
determined by isotopic methods. B.C. Ministry of Energy, Mines and Petroleum Resources, Pap.
1983-1, pp. 305-313.
Armstrong, R.L.. Harakal, J.E., Forbes, R.B., Evans, B.W., Thurston, S.P., 1986. Rb-Sr and K-Ar
study of metamorphic rocks of the Seward Peninsula and southern Brooks Range, Alaska. Geol.
Soc. Am. Mem. 164, 185-203.
Atne, D.C., Bierlein, F.P., McNaughton, N., Wilson, C.J.L., Morand, V.J., Ramsay, W.R.H.. 1996.
Timing of gold mineralisation in western and central Victoria. New constraints from SHRIMP II
analysis of zircon grains from felsic intrusive rocks. In: Ramsay, W'.R.H. (Ed.), Sedimentary-
hosted Mesothermal Gold Deposits. A Global Overview. Minerals Industry Research Institute,
Univ. Ballarat, Abstr. vol., pp. 20-24.
Ash, C.H., Reynolds, P.H.. Macdonald, R.W.J., 1996. Mesothermal gold-quartz vein deposits in
British Columbia oceanic tetranes. New mineral deposit models of the Cordillera, B.C. Geol.
Surv., 1996 Cordilleran Roundup Short Course, pp. O 1-032.
Ashley, P.M., Cook, N.D.J.. Hill. R.L.. Kent. A.J.R., 1994. Shoshonitic lamprophyre dykes and their
relation to mesothermal Au-Sb veins at Hill.-rove. New South Wales, Australia. Lithos 32, 249-
272.
Barley. M.E., Eisenlohr, B.N.. Groves, D.L, Perring, C.S., Vearncombe, J.R.. 1989. Late Archean
convergent margin tectonics and gold mineralization. A new look at the Norseman-Wiluna belt,
western Australia. Geology 17, 826-829.
Bishop, D.G., Bradshaw, J.D.. Landis, C.A., 1985. In: Howell, D.G. (Ed.), Tectonostratigraphic
tetranes of the circum-Pacific region. Circum-Pacific Council for Energy and Mineral Resources,
Earth Sciences Series, No. 1, Houston, pp. 515-521.
Bohlke, J.K., Kistler, R.W., 1986. Rb-Sr, K-Ar, and stable isotope evidence for ages and sources of
fluid components of gold-bearing quartz veins in the northern Sierra Nevada foothills
metamorphic belt, California. Econ. Geol. 81. 296322.
Bradley. D.C., Haeussler, P.J., Kusky, T.M., 1993. Timing of early Tertiary ridge subduction in
southern Alaska. U.S. Geol. Surv.. Bull. 2068, 163-177.
Page 68
Bradley. D.. Haeussler, P.. Kusky, R., Goldfarb, R., 1994. Near trench magmatism, deformation and
gold mineralization during Paleogene ridge subduction in Alaska. SUBCON Conf., Catalina
Island, CA, Abstr. vol., pp. 219-222.
Bradshaw, J.D.. 1994. Brook Street and Murihuku terranes of New Zealand in the context of a mobile
South Pacific Gondwana margin. J. South Am. Earth Sci. 7 (3-4), 325-332.
Burchfiel. B.C., Cowan. D.S.. Davis, G.A.. 1992. Tectonic overview of the Cordilleran orogen in the
western United States. In: Burchfiel, B.C.. Lipman. P.W.. Zoback, M.L. (Eds.), The Cordilleran
Orogen. Conterminous U.S. Geol. Soc. Am., The Geology of North America, G-3: 407-479.
Burleigh, R.E.~1987. A stable isotope, fluid inclusion and petrographic study of gold-quartz veins in
the Willow Creek mining district, Alaska. Unpubl. M.Sc. thesis, Univ. Alaska, 246 pp.
Collins, W.J.. 1996. Tectonic setting of gold deposits in eastern Australia. In: Mesothermal Gold
Deposits. A Global Overview. Univ. West. Aust., Publ. 27, 18-21.
Collins, W.J.. Vernon, R.H., 1992. Palaeozoic arc growth, deformation and migration across the
Lachlan fold belt, southeastern Australia. Tectonophysics 214, 381-400.
Coney, P.J., 1992. The Lachlan belt of eastern Australia and circum-Pacific tectonic evolution.
Cooper, R.A., Tulloch, A.J., 1992. Early Palaeozoic terranes in New Zealand and their relationship to
the Lachlan fold belt. Tectonophysics 214, 129-144.
Cox, S.F., Etheridge, M.A., Cas, R.A.F., Clifford, B.A.. 1991. Deformational style of the Castlemaine
area, Bendigo-Ballarat zone. Implications for evolution of crustal structure in central Victoria.
Aust. J. Earth Sci. 38, IS1-170.
Csejtey, B., Jr. and 8 others, 1978. Reconnaissance geologic map and geochronology, Talkeetna
Mountains quadrangle, northern part of Anchorage quadrangle, and southwest comer of Healy
quadrangle. Alaska. U.S. Geol. Surv.. Open-file Rep. 78-558A, 60 pp., scale 1:250.000.
Currie, L., Parrish, R.R., 1993. Jurassic accretion of Nisling terrane along the western margin of
Stikinia, Coast Mountains, northwestern British Columbia. Geology 21, 235-238.
Danielson, J., Silberman, M.L., Shafiqullah. M., 1990. Age of mineralization of gold-quartz veins at
the Reid mine, eastern Klamath terrane. Shasta County, northern California. Geol. Soc. Am.
Abstr. Prog. 22 (3). 17.
Davidson. C., Hollister, S.M., Schmid. S.M.. 1992. Role of melt in the formation of a deep-crustal
compressive shear zone. Tectonics 11. 348-359.
Dusel-Bacon. C., Csejtey. B.. Jr., Foster. H.L., Doyle. E.O., Nokleberg, W.J., Plafker. G., 1993.
Distribution, facies, ages, and proposed tectonic associations of regionally metamorphosed rocks
in east- and south-central Alaska.' U.S. Geol. Surv., Prof. Pap. 1497-C, 73 pp.
Elder, D., Cashman, S.M. 1992. 1992. Tectonic control and fluid evolution in the Quartz Hill.
California, lode gold deposits. Econ. Geol. 87, 1795-1812.
Engebretson, D.C., Cox, A.. Gordon, R.G., 1985. Relative motions between oceanic and continental
plates in the Pacific basin. Geol. Soc. Am.. Spec. Pap. 206. 64 pp.
Eremin, R.A., Voroshin, S.V., Sidorov, V.A.. Shakhtyrov, V.G., Pristavko. V.A., 1994. Geology and
genesis of the Natalka gold deposit, Northeast Russia. Int. Geol. Rev. 36, 1113-1138.
Ford, R.C., Snee. L.W., 1996. 4°Ar/39Ar thetmtochronology of white mica from the Nome district.
Alaska. The first ages of lode sources to placer gold deposits in the Seward Peninsula. Econ. Geol.
91, 213-220.
Gehrels, G.E., 1998. Geology and U-Pb geochronology of the western dank of the Coast Mountains
between Juneau and Skagway, southeastern Alaska. In: McClelland, W.C. (Ed.). Tectonics of the
Coast mountains, southeastern Alaska and coastal British Columbia. Geol. Soc. Am. Spec. Pap.,
in press.
Gehrels, G.E., McClelland, W.C., Samson. S.D., Patchett, P.J.. Orchard, M.J., 1992. Geology of the
western flank of the Coast Mountains between Cape Fanshaw and Taku Inlet, southeastern
Alaska. Tectonics 11, 567-585.
Geological Survey of Canada. 1991. Radiogenic age and isotopic studies. Geol. Surv. Can., Pap. 91-
2: 252.
Gilligan, L.B., Barnes, R.G., 1990. New England fold belt, New South Wales. Regional geology and
mineralisation. In: Hughes, F.E. (Ed.), Geology of the Mineral Deposits of Australia and Papua
New Guinea. Australas. Inst. Min. Metall., Melbourne, pp. 1417-1423.
Page 69
Goldfarb, R.J., Leach, D.L., Miller. M.L.. Pickthorn, W.J., 1986. Geology, metamorphic setting, and
genetic constraints of epigenetic lode-gold mineralization within the Cretaceous Valdez Group,
south-central Alaska. Geol. Assoc. Can. Spec. Pap. 32, 87-105.
Goldfarb, R.J., Snee, L.W., Miller, L.D.. Newberry. R.J., 1991. Rapid dewatering of the crust
Goldfarb, R.J., Snee. L.W., Pickthorn, W'.J., 1993. Orogenesis, high-T thermal events, and gold vein
formation within metamorphic rocks of the Alaskan Cordillera. Mineral. Mag. 57, 375-394.
Goldfarb, R.J., Christie, T.. Skinner, D., Haeussler. P., Bradley, D., 1995. Gold deposits of Westland.
New Zealand and southern Alaska. Products of the same tectonic processes? In: Mauk, J. (Ed.),
PACRIM '95. Symp. vol., pp. 239-244.
Goldfarb, R.J., Miller, L.D.. Leach, D.L.. Snee. L.W., 1997. Gold deposits in metamorphic rocks in
Alaska. In: Goldfarb, R.J., Miller. L.D. (Eds.), Mineral Deposits of Alaska. Econ. Geol. Monogr.
9, 151-190.
Goryachev. N.A., 1994. Mesothermal gold lode deposits of the Russian Far East. Alaska Miners
Assoc., 1994 Annual Convention. Anchorage. Abstr. and Pap.
Goryachev, N., 1996. Geologic setting and evolution of Au-quartz deposits in the geological history
of metallogenic province of Mesozoids, northeastern Asia. 30th IGC Abstr., vol. 1, Beijing, pp.
391.
Hacker, B.R., Donato, M.M., Barnes, C.G.. McWilliams, M.O., Ernst, W.G., 1995. Timescales of
orogeny. Jurassic construction of the Klamath Mountains. Tectonics 1-1. 677-703.
Haeussler, P.J., Bradley, D., Goldfarb, R., Snee, L., Taylor. C., 1995. Link between ridge subduction
and gold mineralization in southern Alaska. Geology 23, 995-998.
Harbert, W., Frei, L., larrard. R., Halgedahl. S.. Engebretson. D.. 1990. Paleomagnetic and plate-
tectonic constraints on the evolution of the Alaskan-eastern Siberian Arctic. In: Grantz, A.,
Johnson. L., Sweeny. J.F. (Eds.). The Arctic Ocean Region. Geol. Soc. Am.. The Geology of -
North America, v. L. pp. 567-592.
Harper, G.D., Saleeby, J.B.. Heizler. ML. 199-1. Formation and emplacement of the Josephine
ophiolite and the Nevadan orogeny in the Klamath Mountains, California-Oregon. U/Pb zircon
and 40Ar/'9Ar geochronology. J. Geophys. Res. 99. 4293-4321.
Himmelberg, G.R., Brew, D.A.. Ford. A.B.. 1991. Development of inverted metamorphic isograds in
the western metamorphic belt, Juneau, Alaska. J. Metamor. Geol. 9. 165-180.
Hollister, L.S., Crawford. M.L.. 1986. Melt-enhanced deformation. A major tectonic process.
Geology 1-1. 558-56L.
Hudson, T., Platker, G., 1982. Paleogene metamorphism of an accretionary flysch terrane, eastern
Gulf of Alaska. Geol. Soc. Am. Bull. 93, 1281-1290.
Hughes, M.J., Phillips, G.N., Gregory, L.. 1996. Victorian gold. II. Large gold deposits and granites.
Geol. Soc. Aust. Abstr. 41,204.
Hyndman, R.D.. Lewis, T.J., 1995. Review. The thermal regime along the southern Canadian
Cordillera Lithoprobe corridor. Can. J. Earth Sci. 32, 1611-1617.
Johnston, S.T., Mortensen, J.K.. Erdmer. P., 1996. Igneous and metaigneous age constraints for the
Aishihik metamorphic suite, southwest Yukon. Can. J. Earth Sci. 33, 1543-1555.
Jones, D.L., Silberling, N.J.. Coney. P.J.. Plafker, G., 1987. Lithotectonic terrane map of Alaska (west
of the 141st Meridian). U.S. Geol. Surv., Misc. Field Studies Map, MF-1874-A, scale
1:2,500,000.
Kerrich, R., Wyman, D.A.. 1990. The geodynamic setting of mesothermal gold deposits. An
association with accretionary tectonic regimes. Geology 18, 882-885.
Kitch, R.B.. Murphy, R.W., 1990. Gympie gold field. In: Hughes, F.E. (Ed.), Geology of the Mineral
Deposits of Australia and Papua New Guinea. Australas. Inst. Min. Vfetall., Melbourne, pp. 1515-
1518.
Landefeld, L.A., 1988. The geology of the :Mother Lode gold belt, Sierra Nevada Foothills
metamorphic belt. California. Bicentennial Gold 88. Geol. Soc. Australas. Abstr. 22, 167-172.
Leitch, C.H.B., 1990. Bralorne: A mesothermal, shield-type vein gold deposit of Cretaceous age in
southwestern British Columbia. Can. Inst. Min. Metall. Bull. 83 (941), 53-80.
Page 70
Leitch, C.H.B., Van der Hayden, P., Godwin. C.L. Armstrong. R.L.. Harakal, J.E.. 1991.
Geochronometry of the Bridge River camp, southwestern British Columbia. Can. J. Earth Sci. 28,
195-208.
LeLacheur, E.A., 1991. Brittle-fault hosted gold mineralization in the Fairbanks district, Alaska.
Unpubl. M.Sc. thesis, Univ. Alaska, 167 pp.
Lu, J., Seccombe, P.K., Foster. D., Andrew, A.S., 1996. Timing of mineralization and source of fluids
in a slate-belt auriferous vein system, Hill End goldfield, NSW, Australia. Evidence from
40Ar/39Ar dating and O- and H-isotopes. Lithos 38, 147-165.
Madden-McGuire, D.J.. Silberman. M.L.. Church, S.E., 1989. Geologic relationships, K-Ar ages. and
isotopic data from the Willow Creek gold mining district, southern Alaska. Econ. Geol. Mono2r.
6, 242-251.
McClelland, W.C., Anovitz, L.VL. Gehrels, G.E.. 1991. Thetmobarometric constraints on the
structural evolution of the Coast mountains batholith. central southeastern Alaska. Can. J. Earth
Sci. 28, 921-928.
McCoy. D.. Newberry. R.J.. Layer. P., DiMarchi, J.J., Bakke, A.. Masterman. J.S., Minehand. D.L..
1997. PI utonic-related gold deposits of interior Alaska. In: Goldfarb, R.J., Miller, L.D. (Eds.).
Mineral Deposits of Alaska. Econ. Geol. Monogr. 9, 191-241.
McKeag, S.A., Craw, D., 1989. Contrasting fluids in gold-bearing quartz vein systems formed
progressively in a rising metamorphic belt: Otago Schist, New Zealand. Econ. Geol. 84, 22-33.
Mihalynuk. M.G., Smith. M.T.. Gabites, J.E., Runkle, D., Lefebure, D.. 1992. Age of emplacement
and basement character of the Cache Creek terrane as constrained by new isotopic and
geochemical data. Can. J. Earth Sci. 29, 24632477.
Miller, E.L., Hudson. T.L., 1991. Mid-Cretaceous extensional fragmentation of a Jurassic-Early
Cretaceous compressional orogen, Alaska. Tectonics 10. 781-796.
Miller, L.D., Goldfarb, R.J.. Gehrels, G.E., Snee, L.W., 1994. Genetic links among fluid cycling, vein
formation, regional deformation, and plutonism in the Juneau gold belt, southeastern Alaska.
Geology 22, 203-306.
Molnar, P., England. P.. 1995. Temperatures in zones of steadystate underthrusting of young oceanic
lithosphere. Earth Planet. Sci. Lett. 131, 57-70.
Monger, J.W.H., Berg, H.C.. 1987. Lithotectonic terrane map of western Canada and southeastern
Alaska. U. S. Geol. Surv., Misc. Field Studies Map, MF- 1874-B, scale 1:2,500,000.
Mortensen, J.K.. 1990. Geology and U-Pb geochronology of the Klondike district, west-central
Yukon Territory. Can. J. Earth Sci. 27, 903-914.
Mortensen, J.K.. Murphy, D.C.. Poulsen, K.H., Bremner, T., 1996. Intrusion-related gold and base
metal mineralization associated with the early Cretaceous Tombstone plutonic suite, Yukon and
east-central Alaska. New mineral deposit models of the Cordillera. B.C. Geol. Surv., 1996
Cordilleran Roundup Short Course, pp. L1-L13.
Nelson, J.L., Mihalynuk, M.G., 1993. Cache Creek ocean. Closure or enclosure?. Geology 21, 173-
176.
Nie, F.-J., Wu, C.-Y.. 1995. Gold deposits related to alkaline igneous rocks in the north China craton,
People's Republic of China. In: Pasava, J., Kribek, B., Zak, K. (Eds.), Mineral Deposits: From
their Origin to their Environmental Impacts. Proc. of the 3rd Biennial SGA Meeting, Prague. A.A.
Balkema, Rotterdam, pp. 173-176.
Nokleberg, W.J., Bundtzen, T.K., Dawson, K.M., Eremin, R.A.. Goryachev, N.A., Koch, R.D.,
Ratkin, V.V., Rozenblum, LS., Shpikerman, V.I., Frolov, Y.F., Gorodinsky, M.E., Khanchuck.
A.L, Kovbas, L.L, Melnikov, V.D., Nekrasov, LYa., Ognyanov, N.V., Petrachenko, E.D.,
Petrachenko, R.I., Pozdeev, A.L, Ross, K.V., Sidorov, A.A., Wood, D.H., Grybeck, D., 1996.
Significant metalliferous lode deposits and placer districts for the Russian Far East, Alaska, and
the Canadian Cordillera. U.S. Geol. Surv., Open-File Rep. 96513-A,385 pp.
Nokleberg, W.J., Jones, D.L., Silberling, N.J., 1985. Origin and tectonic evolution of the Maclaren
and Wrangellia terranes, eastern Alaska Range. Alaska. Geol. Sue. Am. Bull. 96. 1251-1270.
Nokleberg, W.J., Parfenov, L.M.. Monger. J.W.H., Baranov, B.V.. Byalobzhesky, S.G., Bundtzen,
T.K., Feeney, T.D., Fujita, K., Gordey, S.P., Grantz, A., Khanchuk, A.L. Natal'in, B.A., Natapov,
L.M., Norton, LO., Patton, W.W., Jr., Plalker, G., Scholl, D.W., Sokolov, S.D., Sosunov, G.M.,
Stone, D.B., Tabor, R.W., Tsukanov, N.V., Vallier, T.L., Wakita, Koji, 1994b. Circum-North
Page 71
Pacific tectono-stratigraphic terrane map. U.S. Geol. Surv., Open-File Rep. 94-714, 221 pp., 2
sheets, scale 1:5,000,000; 2 sheets, scale 1:10,000,000.
Nokleberg, W.J.. Plafker, G., Wilson. F.H., 1994a. Geology of south-central Alaska. In: Plalker, G.,
Berg, H.C. (Eds.), The Geology of Alaska. Geol. Soc. Am., The Geology of North America, G-1,
pp. 311-366.
Norris, R.J., Henley, R.W., 1976. Dewatering of a metamorphic pile. Geology 8, 333-336.
Parfenov, L.M., Fujita, K., Layer, P.W., Stone, D.B., 1996. Main tectonic units, events, and tectonic
evolution of Mesozoic orogenic belts of northeast Russia. 30th IGC Abstr., vol. 1. Beijing, p. 217.
Paterson, S.R., Sharp, W.D., 1991. Comparison of structural and metamorphic histories of terranes in
the western metamorphic belt, Sierra Nevada, California. In: Sloan D.. Wagner, D.L. (Eds.),
Geologic Excursions in Northern California: San Francisco to the Sierra Nevada. California Div.
Mines and Geol., Spec. Publ. 109, pp. 113-130.
Pavlis, T.L., 1989. Middle Cretaceous orogenesis in the northern Cordillera. A Mediterranean analog
of collision-related extensional tectonics. Geology 17, 947-950.
Pavlis, T.L., Monteverde, D.H., Bowman, J.R., Rubenstone, J.L., Reason, M.D., 1988. Early
Cretaceous near-trench plutonism in southern Alaska. A tonal ite-trondhjemite intrusive complex
injected during ductile thrusting along the Border Ranges fault system. Tectonics 7, 1179-1199.
Pavlis, T.L., Sisson, V.B., Foster, H.L.. Nokleberg, W.J., Plafker, G.. 1993. Mid-Cretaceous
extensional tectonics of the Yukon-Tanana terrane, trans-Alaska crustal transect (TACT), east-
central Alaska. Tectonics 12. 103-122.
Peters, S.G.. Golding, S.D., 1989. Geologic, fluid inclusion, and stable isotope studies of ganitoid-
hosted gold-bearing quartz veins, Charters Towers, northeastern Australia . Econ. Geol. Monogr.
6, 260-273.
Peters, S.G., Golding, S.D., Dowling, K., 1990. Melange- and sediment-hosted gold-bearing quartz
veins. Hodgkinson gold field, Queensland, Australia . Econ. Geol. 85. 312-327.
Phillips, G.N., 1991. Gold deposits of Victoria: A major province within a Palaeozoic sedimentary
succession. World Gold '91, Cairns, pp. 237-245.
Phillips, G.N., Hughes. M.J., 1995. Victorian gold: A sleeping giant. Soc. Econ. Geol. Newslett. 21.
1-13.
Phillips, G.N., Hughes. M.J.. 1996. The geology and gold deposits of the Victorian gold province.
Ore Geol. Rev. 11, 255-302.
Phillips. G.N., Powell. R. 1993. Link between; gold provinces. Econ. Geol. 88, 1084-1098.
Plafker, G., Berg, H.C., 1994. Overview of the geology and tectonic evolution of Alaska. In: Plafker
G. Berg, H.C. (Eds.). The Geology of Alaska. Geol. Soc. Am.. The Geology of North America.
G-I. pp. 989-1021.
Platker. G.. Nokleberg. W'.J.. Lull. J.S., 1989. Bedrock geology and tectonic evolution of the
Wrangellia. Peninsular, and Chugach terranes along the Trans Alaska Crustal Transect in the
northern Chugach Mountains and southern Copper River basin, Alaska . J. Geophys. Res. 94.
4255-4_'95.
Poulsen, K.H.. Mortensen. J.K.. 1993. Observations on the gold deposits of Eastern Hebei Province,
China. In: Current Research. Part D. Geol. Surv. Can.. Pap. 93-1D, pp. 183-190.
Poulsen, K.H., Taylor. B.E., Robert. F., Mortensen, J.K., 1990. Observations on gold deposits in
North China Platform. In: Current Research, Part A. Geol. Surv. Can.. Pap. 90-1A, pp. 33-44.
Powell, C.McA., Li, Z.X.. 1994. Reconstruction of the Panthalassan margin of Gondwanaland. In:
Veever.. J.J. (Ed.), Permian-Triassic Pangean Basins and Foldbelts along the Panthalassan Margin
of Gondwanaland. Geol. Soc. Am., Mem. 184. pp. 5-10.
Powell, R.. Will, T.M.. Phillips, G.N., 1991. Metamorphism in Archaean greenstone belts. Calculated
fluid compositions and implications for gold mineralization. J. Vtetamor. Geol. 9, 141-1so.
Ramsay, W.R.H., VandenBerg, A.H.M.. 1986. Metallogeny and tectonic development of the Taxman
fold belt system in Victoria. Ore Geol. Rev. I. 213-257.
Ratkin. V., 1995. Pre- and postaccretionary metallogeny of the southern Russian Far East. Res. Geol.
Special Issue 18. 127133.
Rubin, C.M., Miller, E.L.. Toro. J.. 1995. Deformation of the northern circum-Pacific margin.
Variations in tectonic style and plate-tectonic implications. Geology ? :. 897-900.
Page 72
Rushton, R.W., Nesbitt. B.E.. Muehlenbachs, K.. Mortensen, J.K., 1993. A fluid inclusion and stable
isotope study of Au quartz veins in the Klondike district. Yukon Territory, Canada. A section
through a mesothetmtal vein system. Econ. Geol. 88. 647-678.
Rusmore. M.E., 1987. Geology of the Cadwallader Group and the Intermontane- Insular superterrane
boundary, southwestern British Columbia. Can. J. Earth Sci. 24, 2279-2291.
Rusmore, M.E., Woodswotth, G.J., 1991, Coast Platonic Complex. A mid- Cretaceous contractional
orogen. Geology 19, 941-944.
Scheiber, E., 1996. Geology of New South Wales. Synthesis. Geol. Surv. of New South Wales, Mem.
Geol. 13(1), 295 pp.
Sengor, A.M.C., Natal'in, B.A., 1996. Paleotectonics of Asia. Fragments of a synthesis. In: Yin, A.,
Harrison, M. (Eds.), The Tectonic Evolution of Asia. Cambridge University Press. Cambridge, pp.
486-640.
Silberling, N.J., Jones. D.L., Monger. J.W.H., Coney. P.J., 199'. Lithotectonic terrane map of the
North American Cordillera. U.S. Geol. Surv.. Misc. Invest. Map Series. I-2176, scale 1:5,000,000.
Sketchley, D.A., Sinclair, A.J.. Godwin, C.L, 1986. Early Cretaceous gold-silver mineralization in the
Sylvester allochthon. near Cassiar, north central British Columbia. Can. 1. Earth Sci. 23, 1455-
1458.
Smith, T.E., 1981. Geology of the Clearwater Mountains, southcentral Alaska. Alaska Div. Geol.
Geophys. Surv.. Rep. 60. 7= PP.
Solomon, M.. Groves, D.L, 1994. The geology and origin of Australia's mineral deposits. Oxford
Monogr. Geol. Geophys. 24, 951 pp.
Stanley, W.D., Labson, V.F., Nokleberg, W.J.. Csejtey, B. Jr.. Fisher, M.A., 1990. The Denali fault
system and Alaska Range of Alaska. Evidence for suturing and thin-skinned tectonics from
magnetotellurics. Geol. Soc. Am. Bull. 102. 160-173.
Suppel, D.W., 1975. Nambucca block. In: Markham, N.L., Basden, H. (Eds.), The Mineral Deposits
of New South Wales. Geol. Surv. New South Wales, Sydney, pp. 428-447.
Taylor, C.D., Goldfarb, R.J., Snee, L.W., Gent, C.A., Karl, S.M., Haeussler, P.P., 1994. New age data
for gold deposits and granites, Chichagof mining district. SE Alaska. Evidence for a common
origin. Geol. Soc. Am. Abstr. Prog. 26. A140.
Tian, Z.Y., 1992. The Mesozoic-Cenozoic East China rift system. Tectonophysics 208 (1-3), 341-
363.
Till. A.B., Dumoulin, J.A., 1994. Geology of Seward Peninsula and Saint Lawrence Island. In:
Plafker, G., Berg, H.C. (EdsJ. The Geology of Alaska. Geol. Soc. Am., The Geology of North
America, G-1, pp. 141-152.
Tobisch, O.T., Paterson, S.R., Saleeby, J.B., Geary, E.E., 1989. Nature and timing of deformation in
the Foothills terrane. central Sierra Nevada, California. Its bearing on orogenesis. Geol. Soc. Am.
Bull. 101, 401-413.
Trumbull, R.B., Hua. L., Lehrberger, G., Satin M., Wimbauer, T., Motteani, G., 1996. Econ. Geol.
91, 875-895.
Umhoefer, P.J., Schiarizza, P., 1996. Latest Cretaceous to early Tertiary dextral strike-slip faulting on
the southeastern Yalakom fault system, southeastern Coast Belt, British Columbia. Geol. Soc.
Am. Bull. 108. 768-785.
Wang, L.G., Luo, Z.K., McNaughton, N.J., Groves. D.L. Huang, J.Z., Miao, L.C., Guan, K., Liu,
Y.K., 1996. SHRIMP U-Pb in zircon studies of plutonic rocks from the Jiaodong Peninsular,
Shangdong Province, China. Constraints on crustal and tectonic evolution and gold metallogeny.
In: Mesothermal Gold Deposits. A Global Overview. Univ. West. Aust., Publ. 27, 34-38.
Wheeler, J.O., Brookfield, A.J., Gabrielse, H., Monger, J.W.H., Tipper, H.W., Woodsworth, G.J.,
1991. Terrane map of the Canadian Cordillera. Geol. Sun. Can., Map 1713A, scale 1:2,000,000.
Xu, J. (Ed.), 1993. The Tanchung-Lujiang wrench fault system. John Wiley and Sons, Chichester,
279 pp.
Yang, S., Su, S., Du, C., Zhou, M., 1996. Discussion on metallogenic patterns of large and super-
large gold mineral deposits in Jiaodong Peninsula, Shandong, China. 30th IGC Abstr., vol. 2,
Beijing, p. 784.
Zonenshain, L.P., Kuzmin, M.L, Natapov, L.M., 1990. Geology of the USSR: A plate-tectonic
synthesis. Am. Geophys. Union, Geodyn. Ser. 21, 242 pp.
Page 73
Page 74
5. Archean Orogenic Lode Gold Deposits
S.G. Hagemann a, K. F. Cassidy b

a
b
Australian Geological Survey Organisation, Canberra, ACT 2601, Australia
Abstract
Archean orogenic lode gold deposits are the result of large, complex mineralizing systems that h.c
developed within many Archean terrains. Mineralizing systems are defined to include all geologic
factosr that control the generation and preservation of mineral deposits and emphasize the processes
responsible for deposit formation at a variety of scales. Deposits belonging to Archean orogenic lode
gold mineralizing systems comprise epigenetic mineralization that formed as a result of focused fluid
flow late during active deformation and metamorphism of volcancrplutonic terranes. They can occur
in any lithology and formed at a range of paleocrustal levels through site-specific and local physical
and chemical processes. All Archean orogenic lode gold deposits formed through broadly similar
geologic processes, with the unique character of individual deposits resulting mainly from variations
at the depositional site The key feature of Archean orogenic lode gold systems is a broadly uniform
low-moderate salinity, mix, aqueous-carbonic fluid that is capable of carrying Au but has limited
capacity to transport base metals.
Models for development of orogenic lode gold mineralizing systems are generally poorly constrain
although geologic and geochemical characteristics are consistent with terrane- or larger-scale
processes Archean terranes containing orogenic lode gold systems include accretionary and
collisional settings. Mineralization is generally late in the tectonic evolution of the host terranes; is
typically syn- to postpeak metamorphism, becoming increasingly postpeak at higher paleocrustal
levels; and is indicative of clockwise metamorphic P-T paths and implicating processes involving
"deeper later"-type metamorphism. There are few, robust absolute ages on mineralization although,
in most terranes, available ages indicate mineralization follows major volcanic, sedimentarv, and
plutonic episodes. In many terranes, mineralization is coincident with mid-crustal felsic magmatism.
Young absolute ages recorded in some deposits probably reflect resetting and/or new mineral growth
during post-gold mineralization hydrothermal activity and/or slow cooling of host terranes. The
source (s) of fluids and metals in orogenic lode gold systems is poorly understood; however, mineral
equilibria and isotope tracers implicate sources deeper than presently exposed greenstones. Isotope
tracers and mineral equilibria are also consistent with derivation from and/or equilibration of the ore
fluid with felsic rocks during transport of hydrothermal fluids to depositional sites. Stable and
radiogenic isotope tracers alone do not distinguish between fluid derivation through metamorphic
devolatilization and magmatic fluid evolution. Some deposits formed -at high paleocrustal levels,
however, and record the influx of surface waters.
Transport of large volumes of broadly uniform hvdrothermal fluids over relatively long distances is
complicated and requires channelized fluid flow with minor modification of major molecular
components en route from source to depositional sites. Selected major deformation zones that are
truly "crustal scale," as demonstrated by deep seismic profiling, provide ideal fluid pathways for
deeply sourced hydrothermal fluids. The largest gold provinces show spatial proximity of world-class
lode gold deposits to "crustalscale" deformation zones (e.g., Boulder-Lefroy, Destor-Porcupine).
Linking of active faults is important for fluid focusing and effective transport of hvdrothermal fluids.
Page 75
The hvdrothermal fluids transport gold along; the pathways as one or more neutral and reduced
sulfide species. Chemical modeling demonstrates that the inferred hvdrothermal fluids can effectively
transport geld over long distances and over a significant crustal-depth range. Provided the fluids
remain effectively channelized during transport to higher crustal levels, camp- and deposit-scale
structural focusing and associated local gold precipitation mechanisms are required to ensure
development of economic gold mineralization at the trap site.
Archean orogenic lode gold mineralizing systems have distinctive depositional site characteristics at
both the camp and. deposit scale. Camp-scale features include geochemical signatures related to
regional major deformation zones, district-scale granitoid-greenstone contacts), a ~id the presence
or absence cat overlying rock successions that can act as barriers (e.g., seals, aquicludes) to fluid
movement. Deposit-scale variables include the local host rocks, structural traps (fault intersections,
contacts between contrastinlithologies) typically in zones of low mean stress, and the P-T conditions
in the host sequence. Resulting alteration assemblages primarily reflect interaction of host lithologies
with the hydrothermal fluid at a par ticular pressure and temperature. Alteration assemblages
generally show enrichment in K, COZ and S ir, deposits irrespective of paleocrustal level. A metal
association of Au, Ag ± As, B, Bi, Sb, Te, and W is di•played by most deposits. Fluid-incltLsion-
derived data at individual deposits show a range in composition. although salinity is generally low to
moderate, with mixed aqueous-carbonic compositions. Variations i:. CO, CH,, N.,, salinity, and
redox-state may reflect district- to local-scale processes and/or specific ho,; rocks rather than
differences in fluid source. Deposition at the trap site is likely to reflect catastrophic c!fects in
response to physical changes (e.g., large pressure fluctuations, seismic events), with resultat:t
chemical changes due to local fluid wall-rock interaction, phase separation, and or fluid mixing
.
The extreme diversity of Archean orogenic lode gold deposits reflects the complex interplay of
physical and chemical processes at a trap (depositional) site localized at various crustal levels
ranging from sttlgreenschist to upper-amphibolite facies metamorphic environments, with gold
precipitation occurring over a correspondingly wide range of pressures and temperatures. The
variability in deposit characteristics largely reflects the P-T conditions, variability in the host rock,
and local changes in ore fluid composition.
Page 76
6. Hydrothermal transport and depositional

processes in Archean lode-gold systems: A
review
E. J. Mikucki
Abstract
Compelling evidence now exists for the formation of Archean lode-gold deposits over a substantial
range in metamorphic conditions and crustal depth. The commonality of ore-fluid isotopic and
geochemical characteristics, the extreme fluid-rock ratios within actual ore zones and the observed
depth continuity within individual orebodies all suggest that these deposits formed within vertically
extensive, crustal-scale hydrothermal systems. Calculations based on new thermodynamic data and
estimated ore-fluid conditions suggest that gold would have been transported as Au(HS)2 over most
of this range. Gold transport in the ore fluids that formed some diopside- and diopside-K-feldspar
class deposits (T > 550°C) would have been as AuCl2, whereas the Au(HS)° complex may have been
significant in some of the lower temperature ( < 270°C), less alkaline ore fluids. The high solubility of
gold in lode-gold ore fluids at amphibolite fades conditions (10's to 1000's ppb) suggests that fluids
may have been undersaturated in gold within their source regions, promoting the ore fluid's capacity
to retain gold in solution over the large transport distances involved. Ore depositional mechanisms
were also likely to vary between crustal levels. Fluid-rock interactions and wallrock sulphidation in
particular, appear to be the most important type of precipitation mechanism. However, with
decreasing depth of formation, the scope of other depositional mechanisms, notably phase separation
and fluid mixing, becomes greater.
6.1. Introduction
Recent unifying models for the origin of Archean lode-gold systems (Colvine, 1989; Groves et al.,
1992) and the extended depth intervals over which mineralisation can be developed (e.g. 3200 m in
the Kolar Gold Field, India) necessarily imply extensive transport distances for the ore fluids and the
interaction of these fluids with a wide variety of host rocks as they ascend from deep-seated source
regions to high levels within the crust. The extent of the hydrothermal systems involved raises critical
questions about the nature of the hydrothermal transport and deposition of gold at different crustal
levels, or even whether or not such deeply-sourced solutions can remain potent ore-forming fluids
over the entire fluid-path length proposed in these models (e.g. Ridley et al., 1996). Until recently,
however, understanding of gold transport and depositional mechanisms in Archean lode-gold systems
has been limited largely to deposits formed at sub-amphibolite facies metamorphic conditions.
Speculations as to the nature of ore-forming processes in higher metamorphic grade settings have
been hampered by our limited understanding of the physicochemistry of the ore fluids responsible for
such deposits and by a dearth of reliable thermodynamic data for gold-transporting aqueous
complexes at markedly elevated temperatures and pressures. Recently, advances on both fronts
(Zotov et al., 1991; Mikucki and Ridley, 1993; Bloem et al., 1994; Benning and Seward, 1996;
Neumayr et al., 1996) warrant a re-examination of gold transport and depositional mechanisms over
the entire range of crustal settings at which these deposits formed. With this in mind, the following
contribution provides a brief summary of gold transport and depositional mechanisms in Archean
Page 77
lode-gold systems and also presents some new interpretations based on recent studies of gold
solubility in the low P-T supercritical field (Zotov et al., 1991; Benning and Seward, 1996).
6.2. Physicochemical conditions of the ore fluids
Knowledge of the physicochemical conditions for hydrothermal systems responsible for Archean
lodegold deposits, especially of those parameters which most affect gold solubility (T, foe, mss, mct-
and pH), is a fundamental prerequisite to understanding the geochemistry of gold in these systems. In
turn, estimating ore-fluid conditions from fluid inclusion studies (e.g. Smith et al., 1984; Robert and
Kelly, 1987; Guha et al., 1991) and from phase equilibria constraints (e.g. Neall and Phillips, 1987;
Mikucki and Groves, 1990a; Mikucki and Ridley, 1993) nec essarily requires that the fluid inclusions
and mineral phases used are representative of the ore-fluid conditions at the time of mineralisation.
Although relatively straight forward for sub-amphibolite-facies deposits (e.g. Phillips, 1986), the
timing of gold-related hydrothermal alteration with respect to peak metamorphism and deformation
remains somewhat controversial, with the bulk of the evidence suggesting that mineralisation can
occur either at pre-, syn- or post-peak metamorphic conditions, depending upon local circumstances.
Nevertheless, careful studies have indicated that the majority of Archean lode-gold deposits (with a
few notable exceptions; e.g. Phillips and de Nooy, 1988) formed synkinematically and either syn- to
post-peak metamorphism (Clark et al., 1989; Ridley and Barnicoat, 1990; Barnicoat et al., 1991;
Mueller and Groves, 1991; Mueller, 1992; Knight et al., 1993; McCualg et al., 1993; Neumayr et al.,
1993b; Bloem et al., 1994).
Table 1: Classes of lode-style gold deposits based on common ore and proximal alteration mineral
assemblages in mafic-hosted deposits from the Yilgarn Block, Western Australia
Deposit class Proximal alteration Proximal ore minerals Temperature range (°C) ganque minerals
Sericite-ankerite ankerite, dolomite, quartz, pyrite, arsenopyrite ~250-350 sericite, albite
Biotite-ankerite ankerite, dolomite, quartz, pyrite, pyrrhotite, arsenopyrite ~320-420 biotite, albite
Amphibole biotite, Ca-amphibole, plagioclase, pyrrhotite, arsenopyrite, (pyrite)~470-540 quartz, calcite
Diopside biotite, Ca-amphibole, pyrrhotite, arsenopyrite, ~520-580 plagioclase, quartz,

pyroxene, (loellingite) (garnet), (K-feld)
Diopside-K-feld quartz, K-feldspar, pyroxene pyrrhotite, arsenopyrite, loellingite ~570-700

garnet, biotite, (wollastonite)
Brackets indicate mineral not always present. Modified from Mueller and Groves (1991) and Ridley
et al. (1995).
Page 78
Fig. 1. 3 0xidation states for Archean lode-gold deposits relative to that of the C02-CH4 redox buffer
as a function of temperature. A log foe measures the displacement in log fo2 from the CO2-CH4 buffer.
Heavy black bars represent estimated ore-fluid conditions for the deposits considered based on
combined fluid inclusion,223 mineral assemblage and mineral chemistry data for proximal alteration
zones. Arrows represent deposits for which only maximum foe constraints are available. Important
redox buffers at P = 2 kbar (solid curves) are shown for reference. Data from: (a) Racetrack (Gebre-
Mariam et al., 1993); (b) North Kalgurlie (Sang, 1991); (c) Lady Bountiful (stages I and II) (Cassidy
and Bennett, 1993); (d) Wiluna (stage I) (Hagemann et al., in prep.); (e) Mt. Charlotte (Clark, 1980;
Mikucki and Heinrich, 1993); (f) Hunt (Neap and Phillips, 1987); (g) Norseman (Mueller, 1990;
McCuaig et al., 1993); (h) Main Hill (Neumayr et al., 1993b); (i) Zakanaker (Neumayr et al., 1993b);
(j) Nevoria (Mueller, 1990); (k) Westonia (Cassidy, 1992); (1) Marvel Loch (Mueller, 1991); (m)
Griffen's Find (Fare, 1989). Thermodynamic data from Clark (1960a,b), Barton and Skinner (1979),
Sharp et al. (1985) and Johnson et al. (1992). Abbreviations: As, arsenic metal; Aspy, arsenopyrite;
Fay, fayalite; Hem, hematite; Lo, loellingite; Mt, magnetite; Po, pyrrhotite; Py, pyrite; Qtz, quartz;
(S, As)Liq, S- and As-rich melt.
If these interpretations are correct, the observed alteration and vein assemblages largely reflect
equilibrium with the ore-forming fluids at the conditions at which mineralisation took place. Mikucki
and Ridley (1993) used this premise and the existing data for gold deposits from Western Australia to
constrain ore-fluid conditions for deposits formed at different metamorphic grades and crustal levels.
Their work supported earlier interpretations (Mueller and Groves, 1991; Witt, 1991) that the observed
mineralogical changes in deposits at different metamorphic grade (e.g. Table 1) primarily related to
changes in the temperatures of mineralisation and that ore-fluid compositions were broadly similar in
most deposits, regardless of metamorphic grade.
Page 79
Fig. 2. Calculated ranges of mH2s for several Archean lode-gold deposits from Western Australia.
Activity coefficients for aqueous HZS (YHZs) are assumed to be 1 for the purposes of these
calculations. See Fig. 1 for data sources.
Fluid inclusion studies and constraints based on proximal alteration assemblages are consistent with
neutral to weakly alkaline ore fluids of low salinity (typically <= 3 wt% NaCI equiv.) and high
C02(±CH4) (0.050.25XCo2) being responsible for mineralisation at conditions ranging from sub-
greenschist - to at least upper amphibolite-facies metamorphism (from 250 to => 600°C). Silicate-
sulphide and sulphideoxide assemblages within the ores are consistent with oxidation states for the
ore fluids remaining relatively constant with respect to major aqueous redox buffers ( fO2 - 0 to 3 log
units above the C02-CH4 buffer; Fig. 1) over this temperature range, whereas total sulphur contents
(m{s) probably varied from ~ 10 to 10-3.5 m with decreasing temperature (Fig. 2). Ore-fluid pH was
buffered to near-neutral or slightly alkaline values of between 5.2 and 6.8 by albite-sericite ±
paragonite mineral assemblages at low temperatures and by plagioclase-amphibole ± diopside
assemblages at amphibolite fades. In general, dissociation of aqueous species such as HCl° and H2C03
will decrease ore-fluid pH as the fluid cools (Heinrich et al., 1989). The fact that calculated pH values
for lode-gold ore fluids remain essentially constant may therefore result from hydrolysis reactions
between fluid and wallrock providing a sink for the excess H+ produced by acid dissociation
(Mikucki and Ridley, 1993). Fluid buffering, by itself, is likely to result in substantial pH changes
over the temperature range at which Archean lodegold systems are formed. For example, buffering by
carbonate equilibria would result in appreciably higher pH values for amphibolite facies conditions
due to the increase in pK for the H2C03 dissociation reaction with increasing temperature (Fig. 3).
Page 80
Fig. 3. Estimated pH ranges for Archean lode-gold deposits based on observed proximal alteration
and vein mineral assemblages. Predominance boundary for H+-OH- reflects neutral pH at elevated
temperatures and pressures. Data from Mikucki and Ridley (1993). Thermodynamic data from
SUPCRT92 (Johnson et al., 1992).
The above chemical considerations, plus the extreme fluid/rock ratios estimated for lode-gold systems
based on Si02 mass-balance calculations (Wall, 1987), can best be interpreted as resulting from well-
channelled, structurally focused fluid-flow under fluid-dominated conditions. Such conditions not
only permitted ore-fluid transport over the large distances required by crustal continuum models, but
the high fluid-rock ratios involved ensured that the ore fluids maintained essentially constant
composition, at least in terms of their major ore-fluid components (e.g. CO2). Concentrations of the
less abundant orefluid components such as S, Ca, H+, etc. were, however, likely to have been affected
by either fluid-rock exchange reactions and/or precipitation of ore and gangue minerals with
decreasing temperature and pressure.
6.3. Gold transport in hydrothermal solutions
Aqueous gold bisulphide and gold chloride complexes have long been considered the most likely
species for gold transport in hydrothermal solutions because of the ready availability of reduced S and
Cl ligands in most ore fluids, and because of the high thermodynamic stability of these complexes
(Seward, 1991). Modern consensus is that at least three gold bisulphide complexes, Au(HS)°,
HAu(HS)2° and Au(HS)2, are potentially important in transporting gold under hydrothermal
conditions (e.g. Seward, 1973; Shenberger and Barnes, 1989; Hayashi and Ohmoto, 1991). The most
recent study, that of Benning and Seward (1996), critically assesses the available bisulphide data and
provides an additional, internally consistent database from which to evaluate gold speciation in
sulphide-bearing solutions. Their results confirm earlier work indicating that Au(HS)2 is the dominant
gold bisulphide complex in neutral to weakly alkaline, HZS-rich solutions (Seward, 1973) and
suggest that gold is transported as Au(HS)° at low pH and low reduced-sulphur content. Although
previously proposed (e.g. Seward, 1984; Hayashi and Ohmoto, 1991), the presence of HAu(HS)Z in
acid, sulphur-poor solutions is inconsistent with the new, better constrained experimental data, and is
therefore discredited. Most studies, however, point to appreciable gold solubility as reducedsulphur
Page 81
complexes under moderately reducing conditions (at or below the SO4-H2S buffer), neutral to slightly
alkaline pH, and moderate to high total sulphur contents (m{s).
In contrast, the solubility of gold as gold-chloride complexes is generally considered to be negligible,

except under conditions of low pH, high foe, high mCl, and elevated temperatures (Shenberger and
Barnes, 1989; Hayashi and Ohmoto, 1991) and is therefore unimportant under typical sub-
amphibolite facies ore-fluid conditions. Until recently, such calculations relied heavily on
extrapolation of the low temperature data for AuCI2 to higher temperatures. Published experimental
data for high P-T conditions were either inadequate to characterise the stoichiometry and stability of
the gold-chloride complexes (e.g. Anderson and Burnham, 1967; Glyuk and Khlebnikova, 1982) or
resulted in major discrepancies between measured solubilities and those based on extrapolation of the
low-temperature data (e.g. Henley, 1973; Wood et al., 1987). The latter were variably attributed to
poor experimental design or to the presence of additional gold species at high P-T (e.g. AuCl°, Wood
et al., 1987, and AuCl(OH)-, Peshchevitsky et al., 1970; Gadet and Pouradier, 1972). More recent
studies (Zotov and Baranova, 1989; Zotov et al., 1991; Gammons et al., 1994) have, however,
confirmed the predominance of AuCl-2 as the major gold-transporting chloride complex under
hydrothermal conditions. Thermodynamic data provided by these latter studies are also consistent
with that of the low temperature studies of Nikolaeva et al. (1972) and the thermodynamic
extrapolations of Helgeson (1969), corroborating these earlier studies and extending them to
supercritical conditions (T <= 500°C and P <= 1.5 kbar).
6.3.1. Gold transport at amphibolite fades conditions:

Thermodynamic data
As outlined above, experimental gold solubility data now exists over the range of P-T conditions
applicable to the formation of sub-amphibolite, greenschist- or sub-greenschist-fades lode-gold
deposits (T <= 350°C, P <= 2 kbar). Calculation of gold solubility and speciation in the ore fluids
responsible for amphibolite-fades or even higher temperature deposits, some of which may have
formed at temperatures approaching 700°C and pressures of 5 kbar, generally requires extrapolation
of the currently available thermodynamic data. Zotov et al. (1991) and Matthaii et al. (1995) have
used the available experimental data to derive standard state thermodynamic parameters for AuC1-2
and the Au(HS)° and Au(HS)-2 complexes, respectively, using the modified HKF equation of state
models of Tanger and Helgeson (1988). Although still preliminary in nature, these parameters can be
used to provide firstorder estimates of gold solubility at supercritical conditions beyond those of the
experimental data (cf. Matthaii et al., 1995).
In an alternative approach, the current study used the empirically fit polynomial expressions of
Benning and Seward (1996) for the log Keg values of the reactions:
Au+ + H2S° + HS-= Au(HS)- 2 + H+ (1)

Au+ + HZS° = Au(HS)° + H+ (2)
and log Keq data for the reaction:
Au(s) + H+= Au+ + ½ H2(g) (3)
derived from the SUPCRT92 database (Johnson et al., 1992) to determine equilibrium constants for
the gold solubility reactions at elevated temperatures and pressures. Although extrapolation of
empirically fit equations beyond the limits of the experimental data used to derive them may in some
cases lead to significant errors, both Eqs. (1) and (2) are isocoulombic in form. Writing the reactions
in this manner minimises the errors associated with extrapolation of the log Keq values (Lindsay,
1980; Gu et al., 1994). Furthermore, no significant differences were found between the equilibrium
constants calculated above and those determined using conventional, theoretically-based methods
(e.g. Helgeson, 1969) and the yGr°, ySr° and yCpr° data (refit to a Meyer-Kelley format) from
Benning and Seward (1996). Comparison of the log Keq values calculated using the
Page 82
Fig. 4. Equilibrium constants for gold solubility-controlling reactions at P = 2 kbar. Data from Zotov
et al. (1991), Johnson et al. (1992) and Benning and Seward (1996). Reaction 1: Au(s) +H2S =
Au(HS)°+ (1/2)H2(g); reaction 2: An(s) +2H2S = Au(HS)-2 + H+ +(1/2)H2(g); reaction 3: Au(s) + 2Cl- +
H+ = AuC12 +(1/2)H2(g).
Table 2:
Reaction log Keg (2 kbar) (T, °C) Sources
150 200 300 400 500 550
Au+ + H2S° = Au(HS)° + H+ 12.32 9.97 5.58 1.08 -3.84 -6.52 2
Au+ + H2S° + HS-
= Au(HS)2 + H+ 17.37 14.64 10.33 6.97 4.20 2.97 2
+
Au + H + (1/4)O2(g)
=Au+ +(1/2)H2O -5.15 -4.35 -3.13 -2.24 -1.56 -1.28 1
- +
Au + 2C1 + H + (1/4)O2(g)
= AuCl-2 +(1/2)H2O 1.72 2.07 2.95 3.96 5.07 5.67 1, 3
Fe304 + 3H2S°
= 3FeS + (I /2)O2(g) + 3H20 -10.55 -9.97 -9.37 -9.21 -9.36 -9.52 1
Fe304 + 6H2S° + O2(g)
= 3FeS2 + 6H2O 81.27 68.67 49.45 35.30 24.30 19.63 1
FeS2 + H20
=FeS + H2S° + (1/2)O2(g) -30.6 -26.2 -19.6 -14.8 -11.2 -9.72 1
H2S° = H+ + HS- 6.08 6.20 6.70 7.39 8.23 8.70 1
CO2(g) + 2H2O
= CH4(g) + 2O2(g) -96.6 -85.2 -68.5 -56.9 -48.4 -44.9 1
Sources: (1) SUPCRT'92 database (Johnson et al., 1992); (2) Benning and Seward (1996); (3) Zotov
et al. (1991).
data of Benning and Seward (1996) with those calculated using the standard state parameters of
Matthaii et al. (1995) show no appreciable difference for Eq. (1) over the temperature range of
150600°C. Log Keq values for Eq. (2), however, diverge significantly at high temperatures. The cause
of this divergence is presently unknown. Because the fits of Matthaii et al. (1995) incorporate data
from widely different sources, the expressions of Benning and Seward (1996) are used in preference
in the subsequent calculations. Equilibrium constants for reactions involving gold transport as AuC1-2
were calculated using the data of Zotov et al. (1991) and the software package SUPCRT92 (Johnson
et al., 1992). Equilibrium constants for gold solubility-controlling reactions at 2 kbar calculated in this
Page 83
study are summarised in Fig. 4 and in Table 2. Except where otherwise noted, activity coefficients
have been ignored in the following calculations ( = 1). Given the probable level of uncertainty in
the extrapolated thermodynamic data, detailed calculations of activity coefficients are probably
unwarranted.
6.4. Gold solubility and speciation in lode-gold ore fluids
Several studies have addressed the likely concentrations and mode of transport of gold in
subamphibolite-fades lode-gold ore-fluids using the then available thermodynamic data and ore-fluid
conditions based on fluid inclusion and mineral stability constraints (e.g. Phillips and Groves, 1983;
Mikucki and Groves, 1990b). In most cases, gold was determined to be transported in solution as
Au(HS)-2 in concentrations ranging from about 1-100 ppb. Speculations as to the speciation and gold
content of the ore-fluids responsible for forming amphibolite-facies deposits were understandably less
certain. Most considered that the role of Au(HS)-2 would diminish with increasing temperature due to
several factors, including: (1) the shift in the predominance boundary between Au(HS)° and Au(HS)-2
to higher pH with temperature; (2) the general tendency of increasing temperature to favour neutral,
fully associated aqueous complexes as a consequence of changes in the dielectric constant of the
solvent; and (3) the rapid increase in the stability of the AuC1-2 complex with temperature (Seward,
1989; Mikucki and Groves, 1990b; Seward, 1991; Benning and Seward, 1996).
The newly derived equilibrium constants were used to evaluate the changes in gold speciation and
concentrations within Archean lode-gold ore fluids over a range of conditions previously
unattainable. The results of these calculations are summarised in Figs. 5 and 6. The equilibrium
constants used for Figs. 4-6 are given in Table 2. Fig. 5 compares the predominance fields for the
three most important gold complexes, Au(HS)-2, Au(HS)° and AuCl-2 , with the field for Archean ore
fluids. Despite previous speculations, the present calculations indicate that Au(HS)-2 remains the
predominant gold bisulphide species at moderate to high mH2s and high temperature and will
dominate gold transport in most lode-gold fluids. As shown, the expected expansion with increasing
temperature of the Au(HS)° predominance field at the expense of Au(HS)-2 is, in fact, limited to
temperatures below about 300°C.
Page 84
Fig. 5. Predominance-field boundaries for the major gold-transporting ligands as a function of

temperature and H2S. .Data from Zotov et al. (1991), Johnson et al. (1992) and Benning and Seward
(1996). Field for Archean gold-ore fluids generalised from Mikucki and Ridley (1993). Chloride
activities ( Cl-) at elevated temperatures were estimated from dissociation constants for the NaCI°
ion pair, the activity coefficient expressions of Helgeson et al. (1981) and solute species data from
Shock and Helgeson (1988).
This fact, plus the dramatic increase in the H2S- content of Archean gold ore fluids with increasing
temperature, ensure that gold transport as Au(HS)° will be significant only in low-temperature,
relatively low-pH ore fluids. The calculations also suggest that at high temperatures AuC1-2 will
eventually supersede Au(HS)-2 as the dominant gold-transporting complex. Thus for diopside and
diopside-K-feldspar class deposits formed at temperatures in excess of 550°C, AuC1-2 is likely to
dominate gold transport even in the low-salinity ore fluids that typify Archean gold systems.
Solubility contours for gold transported as Au(HS) -2 are presented in Fig. 6 for typical Archean lode-
gold conditions (see above). These results confirm earlier estimates and suggest that gold solubilities
could have exceeded 1-100 ppb Au in low-temperature and 10-10,000 ppb Au in high-temperature
ore fluids, even without accounting for AuC1-2 contributions at high temperature. These extreme
solubilities at amphibolite-fades conditions are supported by Loucks and Hibberson (1996), who
measured gold solubilities approaching 1000 ppm in pyrite-pyrrhotite-magnetite and muscovite-
orthoclase-quartz-buffered solutions at 625°C and 3.7 kbar. The high solubility of gold at high
temperatures has several important consequences. First, the calculated solubilities far exceed
minimum concentrations required to form an effective gold ore fluid (~5 ppb Au; Heinrich et al.,
1989), and therefore negate the need to call on èxotic' gold complexes (e.g. gold thioarsenite
complexes) to transport gold. Far more importantly, the calculations support earlier speculations that
most Archean ore fluids were probably undersaturated with gold in their source regions (cf. Phillips
and Powell, 1993). Provided fluids remained effectively channelled and largely fluidbuffered on their
rise to higher crustal levels (Mikucki and Ridley, 1993), modest undersaturation of the fluids in gold
would help prevent deeply sourced fluids from being rapidly depleted, thus ensuring that the fluids
Page 85
remained effective as ore fluids over the entire crustal extent at which Archean gold deposits were
formed. Furthermore, greater degrees of initial gold-undersaturation in the ore fluids responsible for
amphibolite-fades deposits probably accounts for the relatively low gold production from
amphibolitefacies terrains as compared to those that host greenschist-facies deposits (cf. Ridley et al.,
1996): greater degrees of undersaturation would lead to less efficient ore-depositional processes
Fig. 6. Gold solubility contours (dashed lines) as a function of temperature and HzS content.
Calculations assume typical lodegold ore fluid with COZ /CH4 = 10 pH = 5.5 and P = 2 kbar (see
text). Magnetite (Mt), pyrrhotite (Po) and pyrite (Py) stability fields are bounded by solid lines. Field
for Archean gold-ore fluids generalised from Mikucki and Ridley (1993). Thermodynamic data from
Johnson et al. (1992) and Benning and Seward (1996). Calculations assume Au(HS)-2 = 1.
6.5. Ore precipitation mechanisms
Gold is precipitated from hydrothermal ore fluids in response to changes in the physicochemical
conditions of the fluid at the site of ore deposition. These specific changes in ore-fluid chemistry,
pressure or temperature can result from a number of geological processes, including: (1) adiabatic and
conductive cooling of the ore fluids; (2) interaction between ore fluids and their surrounding host
rocks; (3) phase separation in response to pressure decreases during the rise or throttling of the ore
fluid; or (4) by the mixing of two or more different fluids. Most of these mechanisms have been
invoked as an important cause of gold precipitation within Archean lode-gold deposits at one time or
another. In addition, surface chemistry-driven processes, such as the chemisorption of gold onto
sulphide, sulpharsenide and/or arsenide mineral surfaces (Starling et al., 1989) and the precipitation of
gold from colloidal suspensions (Herrington and Wilkinson, 1993), have also been proposed.
Sections 5.1, 5.2, 5.3 and 5.4 briefly review these mechanisms for gold precipitation in light of the
wide range of crustal conditions over which Archean lode-gold deposits form.
Page 86
6.5.1. P-T changes
Changes in gold solubility as a function of pressure and temperature are important, not only because
cooling has been proposed as a possible depositional mechanism for lode-gold mineralisation (e.g.
Phillips and Powell, 1993), but also because of the ramifications such changes will have on the ability
of deeply sourced ore fluids to transport gold from upper amphibolite to sub-greenschist conditions.
Pressure changes, without resultant phase separation, will have little effect on gold solubility (Seward,
1973; Benning and Seward, 1996).
As a first approximation, whether or not decreasing temperature will precipitate gold in the lode-gold
environment will depend on the speciation of gold in solution and, therefore, on the P-T conditions at
which the deposit formed (Fig. 5). As indicated in Fig. 4, only gold soluble as the chloride complex
will precipitate in response to lowering temperature at constant fluid composition. In fact, changes in
log Keq for the reaction:
Au + 2H2S° = Au(HS)-2 + H+ + 2 H2(g) (4)
will actually promote greater gold solubility at lower temperatures (cf. Fig. 4). Thus cooling of the ore
fluid in isolation from changes in ore-fluid chemistry is unlikely to be an important depositional
mechanism for gold in all but the highest temperature diopside- and diopside-K-feldspar-class
deposits.
Decreasing temperature can, however, affect other equilibria and thereby indirectly bring about gold
deposition, even in ore fluids in which gold is transported by Au(HS) -2. As discussed above, the only
parameter of consequence to gold solubility that is likely to change appreciably with decreasing
temperature is mH2S, due to sulphide precipitation from the ore fluids as they rise through the crust.
Here, however, the effects of H2S loss during cooling are broadly counterbalanced by changes in the
gold solubility constants at T => 350°C. This fact can be seen by comparing the trends of gold
solubility contours with that of Archean gold-ore fluids in Fig. 6. Cooling as a depositional
mechanism from gold bisulphide-bearing ore fluids is therefore only promoted in deposits formed
below about 350°C. At these temperatures gold solubility contours trend more steeply than the loss of
H2S in gold ore fluids with decreasing temperature. Ore fluids that reach saturation in both pyrite and
pyrrhotite will also precipitate gold as a consequence of temperature-induced sulphide precipitation.
In general, however, there is a paucity of pyrite in the majority of deposits formed at amphibolite or
higher metamorphic grades (Mueller and Groves, 1991; Mikucki and Ridley, 1993). Thus cooling of
the ore fluid is deemed to be a significant depositional mechanism only at the extremities of the
temperature range at which Archean lode-gold deposits form (i.e. at uppermost amphibolite to
granulite conditions and at subgreenschist grades). In the deeper settings, the thermodynamic
arguments presented above are further supported by the unlikelihood of the most commonly called
upon means of cooling, fluid mixing or phase separation, within metamorphic environments (Wall,
1987). Heat loss in most lode-gold ore-fluids is therefore limited to the relatively inefficient
mechanisms of conduction and adiabatic cooling. Temperature decreases are therefore more likely to
result in low grade, subeconomic mineralisation spread over large vertical distances, rather than
concentrated orebodies.
6.5.2. Phase separation
Although phase separation has clearly been recognised as an important gold precipitation mechanism
within geothermal systems and epithermal deposits, existing evidence indicates that its role in
mesothermal systems is somewhat more variable, most likely due to the higher ambient P-T
conditions and CO2 contents involved. Regardless of its effects on gold solubility, phase separation
will only take place if ambient conditions fall below those of the solvus for the fluid compositions) in
question. This can take place either through advection of the fluid to higher crustal levels (Spooner et
Page 87
al., 1985), a catastrophic drop in Pfluid due to fault-valve mechanisms (Cox et al., 1995; Robert et al.,
1995) or by expansion of the solvus due to interaction of the fluid with graphitic host rocks or mixing
with CH4-rich fluids (Naden and Shepherd, 1989). Because high-temperature lode-gold deposits
formed at well above critical temperatures for even CH4-rich fluids (Fig. 7), phase separation will
become increasingly more unlikely in amphibolite fades deposits.
Phase separation has, however, been proposed as an important mechanism for forming high-grade
oreshoots and for depositing free gold within laminated quartz reefs, vein arrays and vein breccias
formed at sub-amphibolite-fades conditions (Robert and Kelly, 1987; Walsh et al., 1988; Guha et al.,
1991). These and other studies document close spatial and temporal associations between fluid
inclusion evidence for fluid immiscibility and the precipitation of coarsegrained gold and, therefore,
provide unambiguous evidence that phase separation was, in some cases, an effective gold
depositional mechanism in Archean lode-gold deposits. Gold deposition as a consequence of phase
separation was not, however, universal. Examples exist where gold is hosted primarily in wallrock
alteration haloes despite ample evidence for the occurrence of phase separation in adjacent veins (e.g.
the Victory mine, Clark et al., 1989 and the Wiluna deposits, Hagemann et al., 1994).
Fig. 7. Solvus curves for fluids in the COZ ( f CH4)-H20-NaCl system. Data from Bowers and
Helgeson (1983) and Naden and Shepherd (1989). Temperature ranges for deposit classes of
Archean lode-gold systems taken from Ridley et al. (1995). Diopside- and diopside-Kfeldspar classes
(not shown) occur at temperatures and pressures too high for phase separation of typical ore fluids to
occur.
These examples highlight the fact that the chemical and physical consequences of phase separation
often have opposing effects on gold solubility. Whether or not gold deposition results from the
separation of a C02-rich vapour from the ore fluid will depend on initial fluid compositions and both
the magnitude and relative rates at which pH, m{s, foe and temperature vary during the event
(Drummond and Ohmoto, 1985). Of these factors, only the change in m{s during phase separation
will generally favour gold precipitation. Heat loss, increase in fO2 and pH during initial phase
separation all act to inhibit gold deposition. It is not surprising, therefore, that fluid immiscibility in
Page 88
Archean lode-gold fluids did not always result in the deposition of free gold.
6.5.3. Fluid-rock interaction
A substantial proportion of the gold hosted by sub-amphibolite- and amphibolite-fades gold deposits
is present as microscopic to submicroscopic particles hosted in intensely altered wallrock (Groves et
al., 1985; Ridley et al., 1995). This association has led many authors to suggest that fluid-wallrock
reactions were the principal gold depositional mechanism in Archean lode-gold deposits (e.g. Groves
and Phillips, 1987). Perhaps the most widely applicable of these mechanisms over the entire P-T
spectrum of Archean lode-gold deposits involves the destabilisation of the Au(HS) -2 complex by the
reaction between S-rich ore fluids and Fe-bearing host rocks. As originally envisioned (Phillips and
Groves, 1983), wallrock sulphidation models effectively account for the close association between
wallrock Fe-sulphides and gold and for the overall importance of mafic host rocks (high Fe and/or
Fe/Mg) within deposits of the Yilgarn Craton. Later studies (Neall and Phillips, 1987; Mikucki and
Heinrich, 1993) demonstrated gradients in H2S during wallrock alteration that could easily result in
10- to 100-fold decreases in gold solubility, and therefore strongly support sulphidation reactions as a
primary means of gold precipitation. Gold deposition is thought to occur through coupled reactions of
the type:
‘Fe0’rock + 2H2S° = FeS2(py) + H2O + H2(g) (5)
and
Au(HS)-2 + H+ + 1/2H2(g) = Au + 2H2S° (6)
Increases in whole-rock Fe3+/Fe2+ within proximal sulphide alteration haloes at the Mt. Charlotte and
the Golden Kilometre mines (Mikucki et al., 1995) suggest concomitant reduction of the ore fluids,
indicating that both reduction and H2S loss from the infiltrating ore fluid during wallrock sulphidation
are important in precipitating gold from solution.
The deposition of gold as a result of other types of fluid-rock interaction is also possible. For
example, the intense COz and Ca metasomatism that is characteristic of many sub-amphibolite fades
deposits hosted in ultramafic rocks can lead to ore-fluid acidification through reactions such as:
`Mg0'rock + Ca 2+ + 2H2C03 = CaMg(C03)2(carb) + 2H+ + H2O (7)
(e.g. Kishida and Kerrich, 1987). The acid produced by this process could enhance gold precipitation
in host rocks whose pH buffering capacity is low. Although gold-related carbonatization also occurs
in other rock types, carbonate formation in these rocks is less likely to involve acid production
because Ca is generally not added to the rocks during alteration and carbonatization is far less intense
than it is in ultramafic host rocks (Kerrich and Fyfe, 1981). Higher A1 contents in mafic, intermediate
and felsic host rocks also stabilise mineral assemblages that buffer pore-fluid pH. Significant
decreases in orefluid pH during wallrock alteration are therefore unlikely to be an important gold
depositing mechanism in other than ultramafic-hosted lode-gold deposits formed at sub-amphibolite-
fades conditions. The general absence of intense carbonate alteration haloes associated with
amphibolite-fades deposits (Ridley and Barnicoat, 1990) precludes carbonatization reactions as a
means of depositing gold in these systems.
6.5.4. Surface chemistry-driven processes
The intimate association between gold grains and sulphide minerals within many mesothermal
lodegold veins and detailed morphological studies of gold-sulphide mineral intergrowths from these
deposits, have led a number of workers to stress the importance of sulphide mineral surfaces as
Page 89
essential substrates for gold deposition (Jean and Bancroft, 1985; Hyland and Bancroft, 1989;
Starling et al., 1989; Knipe et al., 1992; Michel and Giuliani, 1996). Gold deposition in these cases is
thought to involve a two step process initiated by the adsorption of gold complexes onto receptive
sulphide mineral surfaces:
AuL4 + Sf => L3AuSf + L (8)
followed by reduction of the gold by surface electrons:
L3AuSf + 3e- => AuSf + 3L- (9)
Here, Sf represents the surface site and L the goldcomplexing ligand (Hyland and Bancroft, 1989).
Similar surface chemistry-driven processes may also be responsible for the high contents of ìnvisible'
gold present in some pyrite, arsenopyrite and loellingite (Cathelineau et al., 1989; Neumayr et al.,
1993a). If so, then adsorption-reduction precipitation mechanisms may occur in both low- and high-
temperature deposits. At present, the relative importance of these mechanisms compared to deposition
resulting from major physicochemical changes in the ore fluid remains unclear. Michel and Giuliani
(1996), for example, noted that diffusion and adsorption of negatively charged Au(HS) -2 complexes
on similarly charged sulpharsenide surfaces (S-rich arsenopyrite from the Fasenda Brasileiro and
Maria Lazara mines, Brazil) would be impossible and suggested that destabilisation of Au(HS) -2,
presumably during phase separation in the deposits, must have occurred prior to adsorption on
arsenopyrite grains. Adsorption-reduction mechanisms are therefore likely to account for the spatial
association between gold and sulphide minerals, but need not be the primary driving force for gold
deposition, at least in cases where visible gold is present.
6.6. Conclusions
The relatively limited range of effective ore-depositional mechanisms for Archean lode-gold systems
reflects the regional metamorphic settings in which these deposits formed (Wall, 1987) and, although
data are limited for high-temperature deposits, there does appear to be some systematic variations in
the way in which deposits at different crustal levels were formed (Fig. 8). Nevertheless if all deposits
are considered together, fluid-rock interactions and wallrock sulphidation in particular, appear to be
the most important type of precipitation mechanism at all crustal levels. With decreasing depth of
formation, the scope for other ore-forming processes to occur becomes greater. For example, in the
dominantly brittle regimes characteristic of the higher level Archean lode-gold deposits (e.g. Wiluna
and Racetrack), hydrothermal fluids from a number of sources, including surficial (e.g. meteoric,
marine or basinal fluids) and deep-seated metamorphic or magmatic origins, may have been present
(Hagemann et al., 1994). Although the exact role that surficial waters play in gold deposition remains
unclear, they have been identified on the basis of isotopic evidence in several deposits, thus opening
up the possibility of fluid mixing as a gold precipitation mechanism. Just as importantly, in the brittle
and brittle-ductile regimes, potentially large, local gradients in Pfluid may occur due to fault rupture
processes. In lower amphibolite- to sub-greenschistfacies deposits, rapid pressure fluctuations may
have been sufficient to cause phase separation and to have resulted in the precipitation of free gold in
quartz veins. In contrast, incursion of fluids from shallower levels is unlikely at deeper crustal levels
where Pfluid ~ Plithostatic and where Pfluid gradients are normally a few hundred bars or less (Etheridge et
al., 1983). Fluid mixing and phase separation are therefore unlikely processes for the formation of
amphibolite fades deposits.
Page 90
Fig. 8. Relative importance of ore depositional mechanisms for different Archean lode-gold deposit
classes. Line thickness represents relative importance of that mechanism. Dashed lines indicate
potentially important precipitation mechanisms whereas question marks indicate possible or
uncertain significance. Deposit classes of Archean lode-gold systems taken from Ridley et al. (1995).
6.7 Acknowledgements
This paper is based on the research of staff and students at the Key Centre for Strategic Mineral
Deposits, the University of Western Australia and has benefited from discussions with John Ridley
and David Groves, among others. Careful reviews by David Cooke and Jennifer Mikucki are
gratefully acknowledged.
6.8 References
Anderson, G.M., Burnham, C.W., 1967. Reactions of quartz and corundum with aqueous chloride
and hydroxide solutions at high temperatures and pressures. Am. J. Sci. 265, 12-27.
Barnicoat, A.C., Fare, R.J., Groves, D.L, McNaughton, N.J., 1991. Syn-metamorphic lode-gold
Barton, P.B., Skinner, B.J., 1979. Sulfide mineral stabilities. In: Barnes, H.L. (Ed.), Geochemistry of
Hydrothermal Ore Deposits, 2nd ed. Wiley-Interscience, New York, pp. 278-403.
Berating, L.G., Seward, T.M., 1996. Hydrosulphide complexing of gold(I) in hydrothermal solutions
from 150 to 500°C and 500 to 1500 bars. Geochim. Cosmochim. Acta 60, 1849-1871.
Bloem, E.J.M., Dalstra, H.J., Groves, D.L, Ridley, J.R., 1994. Metamorphic and structural setting of
Archaean amphibolitehosted gold deposits near Southern Cross, Southern Cross Province, Yilgarn
Block, Western Australia. Ore Geol. Rev. 9, 183-208.
Bowers, T.S., Helgeson, H.C., 1983. Calculations of the thermodynamic and geochemical
consequences of non-ideal mixing in the system H20_CO2-NaCI on phase relations in geologic
systems: Equation of state for non-ideal mixing in the system H20_CO2-NaCI fluids at high
pressures and temperatures. Geochim. Cosmochim. Acta 47, 1247-1275.
Cassidy, K.F., 1992. Archaean granitoid-hosted gold deposits in greenschist to amphibolite facies
terrains: A high P-T to low P-T depositional continuum equivalent to greenstone-hosted deposits.
Ph.D. Thesis, unpublished. Univ. Western Australia.
Page 91
Cassidy, K.F., Bennett, J.M., 1993. Gold mineralisation at the Lady Bountiful Mine, Western
Australia: An example of a granitoid-hosted Archaean lode gold deposit. Miner. Deposita 28, 388-
408.
Cathelineau, M., Boiron, M.-C., Holliger, P., Marion, P., Denis, M., 1989. Gold in arsenopyrites:
Crystal chemistry, location and state, physical and chemical conditions of deposition. In: Keays,
R.R., Ramsay, W.R.H., Groves, D.I. (Eds.), The Geology of Gold Deposits: The Perspective in
1988. Econ. Geol. Monogr. 6, 328-341.
Clark, L.A., 1960a. The Fe-As-S system: Phase relations and applications. Econ. Geol. 55, 1345-
1381.
Clark, L.A., 1960b. The Fe-As-S system: Phase relations and applications. Econ. Geol. 55, 1631-
1652.
Clark, M.E., 1980. Localization of gold, Mt. Charlotte, Kalgoor-lie, Western Australia. B.Sc.
(Honours) Thesis. The University of Western Australia, Nedlands, 128 pp. Unpublished.
Clark, M.E., Carmichael, D.M., Hodgson, C.J., Fu, M., 1989. The structure and metamorphic setting
of the Victory gold mine, Kambalda, Western Australia. In: Keays, R.R., Ramsay, W.R.H.,
Groves, D.I. (Eds.), The Geology of Gold Deposits: The Perspective in 1988. Econ. Geol.
Monogr. 6, 445-459.
Colvine, A.C., 1989. An empirical model for the formation of Archean gold deposits: Products of
final cratonization to the
Superior Province, Canada. In: Keays, R.R., Ramsay, W.R.H., Groves, D.I. (Eds.), The Geology of
Gold Deposits: The Perspective in 1988. Econ. Geol. Monogr. 6, 37-53.
Cox, S.F., Sun, S.S., Etheridge, M.A., Wall, V.J., Potter, T.F., 1995. Structural and geochemical
Central Victoria, Australia. Econ. Geol. 90, 1722-1746.
Drummond, S.E., Ohmoto, H., 1985. Chemical evolution and mineral deposition in boiling
hydrothermal systems. Econ. Geol. 80, 126-147.
Etheridge, M.A., Wall, V.J., Vernon, R.H., 1983. The role of the fluid phase during regional
metamorphism and deformation. J. Metamor. Geol. 1, 205-226.
Fare, R.J., 1989. Nature, timing and genesis of a granulate-fades gold deposit at Griffin's Find,
Western Gneiss Terrain, Western Australia. B.Sc. (Honours) Thesis, The University of Western
Australia, Nedlands, 63 pp. Unpublished.
Gadet, M.C., Pouradier, J., 1972. Hydrolyse des complexes de for(1). C. R. Acad. Sci. Ser. C: 275,
1061-1064.
Gammons, C.H., Williams-Jones, A.E., Yu, Y., 1994. New data on the stability of gold(I) chloride
complexes at 300°C. Min. Mag. 58A, 309-310.
Gebre-Mariam, M., Groves, D.L, McNaughton, N.J., Mikucki, E.J., Vearncombe, JR., 1993.
Archaean Au-Ag mineralisation at Racetrack, near Kalgoorlie, Western Australia: A high crustal-
level expression of the Archaean composite lode-gold system. Miner. Deposita 28, 375-387.
Glyuk, D.S., Khlebnikova, A.A., 1982. Gold solubility in water, HCI, HF and sodium and potassium
chloride, fluoride, carbonate and bicarbonate solutions at a pressure of 1000 kg/cm3. Dokl. Akad.
Nauk SSSR 254, 190-194.
Groves, D.L, Phillips, G.N., 1987. The genesis and tectonic controls on Archaean gold deposits of the
Western Australian Shield: A metamorphic-replacement model. Ore Geol. Rev. 2, 287-322.
Groves, D.L, Barley, M.E., Barnicoat, A.C., Cassidy, K.F., Fare, R.J., Hagemann, S.G., Ho, S.E.,
Hronsky, J.M.A., Mikucki, E.J., Mueller, A.G., McNaughton N.J., Pernng, C.S., Ridley, J.R.,
Vearncombe, J.R., 1992. Sub-greenschist- to granulitehosted Archaean lode-gold deposits of the
Yilgarn Craton: A depositional continuum from deep-sourced hydrothermal fluids in crustal-scale
plumbing systems. In: Glover, J.E., Ho, S.E. (Eds.), The Archaean: Terrains, Processes and
Metallogeny. Proceedings of the Third International Archaean Symposium, Perth, 1990. Geol.
Dep. Univ. Ext., Univ. West. Aust. Publ. 22, 325-338.
Groves, D.L, Phillips, G.N., Ho, S.E., Houstoun, S.M., 1985. The nature, genesis and regional
controls of gold mineralisation in Archaean greenstone belts of the Western Australian shield: A
brief review. Trans. Geol. Soc. S. Afr. 88, 135-148.
Page 92
Gu, Y., Gammons, C.H., Bloom, M.S., 1994. A one-term extrapolation method for estimating
equilibrium constants of aqueous reactions at elevated temperatures. Geochim. Cosmochim. Acta
58, 3545-3560.
Guha, J., Lu, H.-Z., Dube, B., Robert, F., Gagnon, M., 1991. Fluid characteristics of vein and altered
wallrock in Archean mesothermal gold deposits. Econ. Geol. 86, 667-684.
Hagemann, S.G., Brown, P.E., Groves, D.L, Ridley, J.R., Valley, J., 1994. The Wiluna lode-gold
deposits, Western Australia: Surface water influx in a shallow level Archean lode-gold system.
Geol. Soc. Aust. Abstr. 37, 160.
Hayashi, K., Ohmoto, H., 1991. Solubility of gold in NaCI and 112S-bearing aqueous solutions at
250-350°C. Geochim. Cosmochim. Acta 55, 2111-2126.
Heinrich, C.A., Henley, R.W., Seward, T.M., 1989. Hydrothermal Systems. Australian Mineral
Foundation, Adelaide, 74 pp.
Helgeson, H.C., 1969. Thermodynamics of hydrothermal systems at elevated temperatures and
pressures. Am. J. Sci. 267, 729-804.
Helgeson, H.C., Kirkham, D.H., Flowers, G.C., 1981. Theoretical prediction of the thermodynamic
behavior of aqueous electrolytes at high pressures and temperatures. IV. Calculation of activity
and osmotic coefficients and apparent molal and standard and relative partial molal properties to
600°C and 5 kbar. Am. J. Sci. 281, 1249-1493.
Henley, R.W., 1973. Solubility of gold in hydrothermal chloride solutions. Chem. Geol. 11, 73-87.
Herrington, R.J., Wilkinson, J.J., 1993. Colloidal gold and silica in mesothermal vein systems.
Geology 21, 539-542.
Hyland, M.M., Bancroft, G.M., 1989. An XPS study of gold deposition at low temperatures on
sulphide minerals: Reducing agents. Geochim. Cosmochim. Acta 53, 367-372.
Jean, G.E., Bancroft, G.M., 1985. An XPS and SEM study of gold deposition at low temperatures on
sulphide mineral surfaces: Concentration of gold by adsorption/reduction. Geochim. Cosmochim.
Acta 49, 979-987.
Johnson, J.W., Oelkers, E.H., Helgeson, H.C., 1992. SUPCRT92: A software package for calculating
the standard molal thermodynamic properties of minerals, gases, aqueous species, and reactions
from 1 to 5000 bar and 0 to 1000°C. Comput. Geosci. 18, 899-947.
Kernch, R., Fyfe, W.S., 1981. The gold-carbonate association: Source of COz fixation reactions in
Archaean lode deposits. Chem. Geol. 33, 265-294.
Kishida, A., Kerrich, R., 1987. Hydrothermal alteration zoning and gold concentration at the Kerr-
Addison Archean lode gold deposit, Kirkland Lake, Ontario . Econ. Geol. 82, 649690.
Knight, J.T., Groves, D.L, Ridley, J.R., 1993. The Coolgardie Goldfield, Western Australia: District-
scale controls on an Archaean gold camp in an amphibolite fades terrane. Mineral. Deposita 28,
436-456.
Knipe, S.W., Foster, R.P., Stanley, C.J., 1992. Role of sulphide surfaces in sorption of precious
metals from hydrothermal fluids. Trans. Inst. Min. Metall. B101, 83-88.
Lindsay, W.T. Jr., 1980. Estimation of concentration quotients for ionic equilibria in high temperature
water: The model substance approach. Proc. Int. Water Conf. 41, 284-294.
Loucks, R.R., Hibberson, W., 1996. Experimental solubility of gold in sulphidic brine to 625°C and 4
kbar measured in synthetic fluid inclusions by ICPMS-ULTEMA. Geol. Soc. Aust. Abstr. 41,
258.
Matthaii, S.K., Henley, R.W., Heinrich, C.A., 1995. Gold precipitation by fluid mixing in bedding-
parallel fractures near carbonaceous slates at the Cosmopolitan Howley gold deposit, Northern
Australia. Econ. Geol. 90, 2123-2142.
McCuaig, T.C., Kernch, R., Groves, D.L, Archer, N., 1993. The nature and dimensions of regional
and local gold-related hydrothermal alteration in tholeiitic metabasalts in the Norseman goldfields:
The missing link in a crustal continuum of gold deposits?. Miner. Deposita 28, 420-435.
Michel, D., Giuliani, G., 1996. Habit and composition of gold grains in quartz veins from greenstone
belts: Implications for mechanisms of precipitation of gold. Can. Mineral. 34, 513528.
Mikucki, E.J., Groves, D.L, 1990a. 2.2.1 Constraints on genesis of primary gold deposits:
Mineralogical constraints. In: Ho, S.E., Groves, D.L, Bennett, J.M. (Eds.), Gold Deposits of the
Archaean Yilgarn Block, Western Australia: Nature, Genesis and Exploration Guides. Geol. Dep.
Univ. Ext., Univ. West. Aust. Publ., 20, 212-220.
Page 93
Mikucki, E.J., Groves, D.L, 1990b. 3.1 Genesis of primary gold deposits: Gold transport and
depositional models. In: Ho, S.E., Groves, D.L, Bennett, J.M. (Eds.), Gold Deposits of the
Archaean Yilgarn Block, Western Australia: Nature, Genesis and Exploration Guides. Geol. Dep.
Univ. Ext., Univ. West. Aust. Publ., 20, 278-284.
Mikucki, E.J., Heinrich, C.A., 1993. Vein- and mine-scale wallrock alteration and gold mineralisation
in the Archeaen Mount Charlotte deposit, Kalgoorlie, Western Australia. An international
conference on crustal evolution, metallogeny and exploration of the Eastern Goldfields. Ext.
Abstr., AGSO Record, 54, pp. 135-140.
Mikucki, E.J., Ridley, JR., 1993. The hydrothermal fluid of Archaean lode-gold deposits at different
metamorphic grades: Compositional constraints from ore and wallrock alteration assemblages.
Miner. Deposita 28, 469-481.
Mikucki, E.J., Gebre-Mariam, M., Heinrich, C.A., 1995. Material-balance constraints on precipitation
mechanisms in sulphide-rich lode-gold alteration haloes. Precambrian '95, Progr. and Abstr., p.
160.
Mueller, A.G., 1990. The nature and genesis of high- and medium-temperature Archaean gold
deposits in the Yilgarn Block, including a specific study of scheelite-bearing gold skarn deposits.
Ph.D. Thesis, The University of Western Australia, Nedlands, 144 pp. Unpublished.
Mueller, A.G., 1991. The Savage Lode magnesian skarn in the Marvel Loch gold-silver mine,
Southern Cross greenstone belt, Western Australia. Part 1: Structural setting, petrography, and
geochemistry. Can. J. Earth Sci. 28, 659-685.
Mueller, A.G., 1992. Petrogenesis of amphibole-biotite-calciteplagioclase alteration and laminated
quartz veins in four Archean shear zones of the Norseman District, Western Australia. Can. J.
Earth Sci. 29, 388-417.
Mueller, A.G., Groves, D.L, 1991. The classification of Western Australian greenstone-hosted gold
deposits according to wallrock alteration mineral assemblages. Ore Geol. Rev. 6, 291331.
Naden, J., Shepherd, T.J., 1989. Role of methane and carbon dioxide in gold deposition. Nature 342,
793-795.
Neap, F.B., Phillips, G.N., 1987. Fluid-wall rock interaction in an Archean hydrothermal gold
deposit: A thermodynamic model for the Hunt mine, Kambalda. Econ. Geol. 82, 1679-1694.
Neumayr, P., Groves, D.L, Ridley, J.R., 1993a. Syn-amphibolite fades Archaean lode gold
mineralisation in the Mt. York District, Pilbara Block, Western Australia. Miner. Deposita 28,
457-468.
Neumayr, P., Cabri, L.J., Groves, D.L, Mikucki, E.J., Jackman, J.A., 1993b. The mineralogical
distribution of gold and relative timing of gold mineralization in two Archean settings of high
metamorphic grade in Australia. Can. Mineral. 31, 711725.
Neumayr, P., Ridley, JR., Groves, D.L, 1996. Physico-chemical conditions of fluid-wallrock
interaction at amphibolite-fades conditions in two Archean hydrothennal gold deposits in the Mt.
York district, Pilbara Craton, Western Australia. Can. J. Earth Sci. 32, 993-1016.
Nikolaeva, N.M., Yerenburg, A.M., Antinina, V.A., 1972. Temperature dependence of the standard
potential of halide complexes of gold. Izv. Sib. Otd. Akad. Nauk SSSR Ser. Khim. 4, 126-129.
Peshchevitsky, B.L, Yerenburg, A.M., Belevantsev, V.I., Kazakov, V.P., 1970. Stability of gold
complexes in aqueous solutions. Izv. Akad. Nauk SSSR Ser. Geol. 4, 126-129.
Phillips, G.N., 1986. Geology and alteration in the Golden Mile, Kalgoorlie. Econ. Geol. 81, 779-
808.
Phillips, G.N., de Nooy, D., 1988. High-grade metamorphic processes which influence Archaean
gold deposits, with particular reference to Big Bell, Australia. J. Metamor. Geol. 6, 95-114.
Phillips, G.N., Groves, D.L, 1983. The nature of Archaean gold bearing fluids as deduced from gold
deposits from Western Australia. J. Geol. Sue. Aust. 30, 25-40.
Ridley, J.R., Barnicoat, A.C., 1990. Nature and setting of primary gold deposits: Wallrock alteration
in amphibolite-facies gold deposits. In: Ho, S.E., Groves, D.L, Bennett, J.M. (Eds.), Gold
Deposits of the Archxan Yilgarn Block, Western Australia: Nature, Genesis and Exploration
Guides. Geol. Dep. Univ. Ext., The Univ. West. Aust. Publ., 20, 79-86.
Ridley, J., Groves, D.L, Hagemann, S.G., 1995. Exploration and Deposit Models for Gold Deposits
in Amphibolite/Granulite Fades Terrains. MERIWA Rep. 142, 126 pp.
Page 94
Ridley, J., Mikucki, E.J., Groves, D.L, 1996. Archean lode-gold deposits: Fluid flow and chemical
evolution in vertically extensive hydrothermal systems. Ore Geol. Rev. 10, 279-293.
Robert, F., Kelly, W.C., 1987. Ore-forming fluids in Archean gold-bearing quartz veins at the Sigma
Mine, Abitibi greenstone belt, Quebec, Canada. Econ. Geol. 82, 1464-1482.
Robert, F., Boullier, A.M., Firdaous, K., 1995. Gold-quartz veins in metamorphic terranes and their
bearing on the role of fluids in faulting. J. Geophys. Res. 100 (B7), 12861-12879.
Sang, J.H., 1991. The evolution of a giant hydrothermal system: A study of main/flat lode at North
Kalgurli, Kalgoorlie, Western Australia. M.Sc. thesis, The University of Western Australia,
Nedlands, 92 pp. Unpublished.
Seward, T.M., 1973. Thio complexes of gold and the transport of gold in hydrothermal ore solutions.
Geochim. Cosmochim. Acta 37, 379-399.
Seward, T.M., 1984. The transport and deposition of gold in hydrothermal systems. In: Foster, R.P.
(Ed.), Gold '82: The Geology, Geochemistry and Genesis of Gold Deposits. A.A. Balkema,
Rotterdam, pp. 165-182.
Seward, T.M., 1989. The hydrothermal chemistry of gold and its implications for ore formation:
Boiling and conductive cooling as examples. In: Keays, R.R., Ramsay, W.R.H., Groves, D.I.
(Eds.), The Geology of Gold Deposits: The Perspective in 1988. Econ. Geol. Monogr. 6, 398-404.
Seward, T.M., 1991. The hydrothermal geochemistry of gold. In: Foster, R.P. (Ed.), Gold
Metallogeny and Exploration. Chapman and Hall, London, pp. 37-62.
Sharp, Z.D., Essene, E.J., Kelly, W.C., 1985. A re-examination of the arsenopyrite geothermometer:
Pressure considerations and applications to natural assemblages. Can. Mineral. 23, 517534.
Shenberger, D.M., Barnes, H.L., 1989. Solubility of gold in aqueous sulfide solutions from 150 to
350°C. Geochim. Cosmochim. Acta 53, 269-278.
Shock, E.L., Helgeson, H.C., 1988. Calculation of the thermodynamic and transport properties of
aqueous species at high pressures and temperatures: Correlation algorithms for ionic species and
equation of state predictions to 5 kbar and 1000°C. Geochim. Cosmochim. Acta 52, 2009-2036.
Smith, T.J., Cloke, P.L., Kesler, S.E., 1984. Geochemistry of fluid inclusions from the McIntyre-
Hollinger gold deposit, Timmins, Ontario, Canada. Econ. Geol. 79, 1265-1285.
Starling, A., Gilligan, J.M., Carter, A.H.C., Foster, R.P., Saunders, R.A., 1989. High-temperature
hydrothermal precipitation of precious metals on the surface of pyrite. Nature 340, 298-300.
Tanger, J., Helgeson, H.C., 1988. Calculation of the thermodynamic and transport properties of
aqueous species at high pressures and temperatures: Revised equations of state for the standard
molal properties of ions and electrolytes. Am. J. Sci. 288, 19-98.
Wall, V.J., 1987. The geochemistry of gold in regional metamorphism. In: The Second Eastern
Goldfields. Geological Field Conference, Abstracts and Excursion Guide, pp. 8-23.
Walsh, J.F., Kesler, S.E., Duff, D., Cloke, P.L., 1988. Fluid inclusion geochemistry of high-grade
vein-hosted gold ore at the Pamour Mine, Porcupine Camp, Ontario. Econ. Geol. 81, 681-703.
Witt, W.K., 1991. Regional metamorphic controls on alteration associated with gold mineralization in
the Eastern Goldfields Province Western Australia: Implications for the timing and origin of
Archaean lode-gold deposits. Geol. 19, 982-985.
Wood, S.A., Crerar, D.A., Borcsik, M.P., 1987. Solubility of the assemblage pyrite-pyrrhotite-
magnetite-sphalerite-galenagold-stibnite-bismuthinite-ar gentite-molybdenite in HZONaCI-CO2
solutions from 200 to 350°C. Econ. Geol. 82, 1864-1887.
Zotov, A., Baranova, N., 1989. Thermodynamic feature of the aurochloride solute complex, AuCl2,
at temperatures from 350 to 500°C and under pressure from 500 to 1500 bar. Sci. Geol.
Zotov, A., Baranova, N., Dar'yina, T., Bannykh, L., 1991. The solubility of gold in aqueous chloride
fluids at 350-500°C and 500-1500 atm. Thermodynamic parameters of AuCIZ~s°1) up to 750°C
and 5000 atm. Geochem. Int. 28, 63-71.
Page 95
7. Alteration and primary geochemical

dispersion associated with the Bulletin lode-
gold deposit, Wiluna, Western Australia
Pasi Eilu a and Edward J. Mikucki b
a Department of Geology, University of Turku, FIN-20014 Turku, Finland

b Centre for Strategic Mineral Deposits, Department of Geology and Geophysics,
University of Western Australia, Nedlands, WA6907, Australia
Abstract
The Bulletin lode-gold deposit is within the northernmost part of the Norseman-Wiluna greenstone
belt in the Archaean Yilgarn Block, Western Australia. It is located within a brittle-ductile shear zone
and hosted by tholeiitic metavolcanic rocks. Syn-metamorphic wallrock alteration envelops the gold
mineralisation and is pervasive throughout the entire shear zone and extends up to 150 m into the
undeformed wallrocks. Alteration is characterised by the sequence of distal chlorite-calcite,
intermediate calcite-dolomite, outer proximal sericite and inner proximal dolomite-sericite zones.
The thickness of the alteration envelope, and the occurrence of dolomite in the alteration sequence,
can be used as a rough guide to the width, extent and grade of gold mineralisation, because a
positive correlation exists between these variables.
Mass transfer evaluations indicate that chemical changes related to the wallrock alteration are
similar in all host rocks: in general, Ag, As, Au, Ba, CO2, K, Rb, S, Sb, Te and W are enriched, Na
and Y are depleted, and Al, Cr, Cu, Fe, Mg, Mn, Nb, Ni, P, Se, V, Zn and Zr are immobile, while Ca,
Si and Sr show only minor or negligible relative changes. The degree of mobility of each component
increases with proximity to gold mineralisation.
The largest potential exploration targets are possibly defined by regional As (>6 ppm) and Sb (>0.6
ppm) anomalies. These anomalies, if real, extend laterally for >150 m from the mineralised shear
zone into areas of apparently unaltered rocks. Anomalies defined by Te (>10 ppb), W (>0.6 ppm),
carbonation indices, local enrichment of Sb (>2.0 ppm) and As (>28 ppm), and potassic alteration
indices also form significant exploration targets extending beyond the HJB shear zone and the Au
anomaly (>6 ppb) and, locally, into apparently unaltered rock. Gold, itself, has a restricted
dispersion, with an anomaly extending for 1-35 m from ore, and being restricted to within the shear
zone itself.
Amongst individual geochemical parameters, only As and Sb define significant, consistent and
smooth trends (vectors) when laterally approaching the ore. However, the respective dimensions of
individual geochemical anomalies can be used as an extensive, though stepwise, vector towards ore.
Page 97
7.1. Introduction
Wallrock alteration and related primary geochemical and isotopic dispersion haloes envelop most
hydrothermal ore deposits, including Archaean lode-gold deposits. A number of studies have
demonstrated that these alteration patterns can represent significant targets at both regional and mine-
scale (e.g. Rose and Burt, 1979; Perrault et al., 1984; Smith and Kesler, 1985; Kishida and Kerrich,
1987; Barley et al., 1990; Davies et al., 1990). Even so, the potential of primary dispersion in
targeting lode-gold mineralisation has not been fully appreciated. This reluctance is partly due to the
complex geochemical patterns that arise from the large variety of host rocks, mineralisation and
alteration styles and the extended range of metamorphic grades that characterise deposits of this type
(Groves, 1993; Eilu et al., 1995). In addition, many earlier efforts relied heavily on a single
parameter, or failed to fully account for the effects of closure and pre-metasomatic geochemical
variations within host rock sequences.
The Bulletin lode-gold deposit at Wiluna is situated within the northernmost part of the Norseman-
Wiluna greenstone belt in the Archaean Yilgarn Block, Western Australia, about 550 km north of
Kalgoorlie (Fig. 1). It is hosted by mafic volcanic rocks which were metamorphosed and altered
under sub- to lower-greenschist facies P-T conditions (Hagemann, 1992; Hagemann et al., 1994;
Chanter et al., 1997).
Granitoid and gneiss 200 km

Greenstones Yilgarn
Fault/shear zone
Bulletin
Gold deposit
Kalgoorlie
Perth
Wiluna
township N
Squib
Moonlight
Bulletin
Happy Jack
North Lode
West Lode
East Lode
3 km
Felsic and intermediate volcanic Ultramafic rocks
and sedimentary rocks
Mafic volcanic rocks Granitoid
Fault or shear zone Gold mine
Page 98
Fig. 1. Location of the Bulletin deposit at the Yilgarn Block and geological map of the Wiluna region
(Hagemann, 1992) showing the sites of main gold deposits.
The Bulletin deposit offers a good opportunity to investigate the extent and style of alteration and
primary geochemical dispersion related to mesothermal gold mineralisation, because its geological
and structural setting is relatively simple and well understood (Hagemann et al., 1992; Chanter et al.,
1997) and because a large amount of sample material from ore and its wallrocks is available. The
primary impetus for the present study is, however, the fact that relatively few studies have
concentrated on the use of primary geochemical dispersion in the exploration for mesothermal gold
deposits. In those studies that are available, the majority have been restricted to relatively few of the
potentially mobile elements.
To approach the problem at Bulletin, rock types and alteration were mapped, sampled and analysed
for several sections across the deposit and its wall rocks. Staining for carbonates, XRD-analyses and
thin section investigations were used to determine mineral assemblages and textures in unaltered
rocks and all alteration zones. This was followed by whole-rock geochemical analyses and
evaluation of mass transfer during alteration. The relationships between chemical changes and
changes in mineral assemblages were then assessed. Finally, the values of the geochemical
parameters potentially related to gold mineralisation were plotted on alteration zone maps to asses the
extent of primary geochemical dispersion, the relationship between the dispersion and alteration, and
the potential of these parameters to form vectors towards ore.
7.2. Geological setting
The Bulletin lode-gold deposit is hosted by the subvertical to steeply SE dipping, SW-NE striking
Happy Jack-Bulletin (HJB) shear zone, which cross-cuts a tholeiitic volcanic sequence (Fig 1). This
volcanic sequence has a thickness in excess of 2250 m, with individual flow units up to 100 m thick.
According to Hagemann (1992) and Hagemann et al. (1994), the rocks were metamorphosed under
sub-greenschist facies P-T conditions between 250-350°C and ≤2 kbar. The sample material for the
present study is from deeper parts of the area (>200 m below the present surface) than material
available for the previous studies and exhibits typical lower-greenschist facies mineral assemblages.
Because all host rocks are metamorphosed, the prefix "meta" is implied but omitted from the names
of the rock types in this paper.
Massive and pillowed varieties of lavas, both essentially unstrained, are common in the area studied.
The medium-grained massive basalts are commonly termed dolerites at Bulletin (Fig. 2). Bedding
within the volcanic pile strikes NW-SE and dips approximately 60° to the SW. Inter-flow sedimen-
tary rocks, and felsic and intermediate dykes occur throughout the area as 0.1-5 m and 0.2-1 m thick
units, respectively, but form less than 1% of the total rock volume. The sedimentary rocks locally
host ore but there are no porphyries within the shear zone at Bulletin, although felsic units contain
significant amounts of ore in the East and West Lode deposits (Hagemann et al., 1992), which are
only 3 km from the Bulletin deposit (Fig. 1).
Dextral faulting, with at least 750 m of horizontal displacement (S. Chanter, pers. comm. 1995), has
taken place along the HJB shear zone. On a regional scale, the shear zone is almost perpendicular to
bedding, but the latest investigations in underground parts of the mine suggest that the structure is not
as simple as that shown in Figure 2 but includes several minor faults with unknown magnitudes of
displacement (possibly up to a few tens of metres) that subparallel the major shear zone within the
hangingwall (SE side of the shear zone) sequence (S. Chanter, pers. comm. 1996). The occurrence of
0.5-5 m wide, partially sericitised, vein-rich, foliated shear zones in the hangingwall indicates the
sites of the minor faults (Fig. 3). There is no information on whether similar faults exist in the
footwall (NW side), because diamond drilling and underground development are unavailable.
Page 99
At Bulletin, the HJB shear zone is subvertical to steeply SE dipping and 20-60 m wide (Figs 2 and
3). Where it is branched, it encloses weakly altered and relatively unstrained blocks of wallrock ≤10
m wide. A striking change in strain state, from massive wallrock to foliated shear zone, occurs at the
margins of the shear zone. The distribution of primary lithologies within the shear zone is complex:
variably strained blocks of pillowed and massive lava, medium-grained doleritic rock and interflow
sedimentary rocks are juxtaposed, and the intensity of shearing has obliterated the primary textures of
the mafic host rocks.
The gold mineralisation is confined to the central part of the HJB shear zone, consisting of one major
mineralised zone and, locally, two subsidiary mineralised zones that parallel the main ore zone (Fig.
2). The main ore zone extends for 300-500 m along strike and at least 600 m down dip within the
HJB shear zone (Figs 2 and 3), varying in thickness from 1 to 25 m and plunging at 60-70° to the
SW. The minor ore zones are generally 1-5 m wide and are less continuous along strike.
A
WD140
D
WD135
B WD138
WD126
WD108 WD130
WD117
100 m
Mafic lava Dolerite Inter-flow sedimentary rock
Shear zone Ore Strike of the Happy Jack-

Bulletin fault
Drill hole sampled and logged Pit outline
in this study
Fig. 2. Simplified geological map of the Bulletin area, showing outlines of the open pit, after a map
prepared by Wiluna Mines Ltd exploration staff. The location of the cross sections shown in Fig. 12
is marked by the line AB and CD. Also the collars and drilling directions of holes logged are
indicated.
Page 100
NW SE
130
WD
Happy Jack-
117
Bulletin
WD
shear zone
6
12
WD
?
Ore b
?
cklo
Hangingwall
100 m
Footwall
Fe tholeiite
Mg tholeiite-1
Mg tholeiite-2
Minor shear zone
Diamond drill h ole: no. WD117,

logged and sampled during this
study
117
WD
Fig. 3. Mafic rock suites, shear zones and ore blocks in the NW-SE cross section AB shown in Figure
2.
Page 101
7.3. Mapping, sampling and analytical procedures
7.3.1. Alteration mapping
Mineral assemblages were logged in seven diamond drill cores in order to define the alteration zones
and their extents, both within the shear zone and in adjacent wallrocks. Textures and mineral
assemblages were documented in both host rocks and veins. Thin sections were examined for 67
samples, while dilute HCl and carbonate staining techniques (Hutchinson, 1974) were used on all
samples. In selected cases, qualitative XRD analysis was used to confirm these studies. In this paper,
"dolomite" refers to Fe-bearing dolomite. Due to scale limitations, alteration zones presented in
figures in this paper represent the predominant alteration type present. Most of the alteration zones
depicted contain minor intercalations of adjacent alteration zones.
7.3.2. Sampling
A total of 190 samples along profiles were collected from diamond drill core on both sides of the
shear zone. Due to the very minor extent and significance of the other rock types in the area, only
mafic rocks were sampled. Sampling was concentrated within the shear zone and the hangingwall
because alteration extends much further into the hangingwall (up to 150 m) than into the footwall
(≤20 m).
Sampling was focussed on five sections across the shear zone and its wallrocks (Figs 2 and 3) and
was designed: (1) to obtain unweathered samples representing both lateral and vertical profiles across
the shear zone and its wallrocks; (2) to obtain samples representing unaltered and altered types of
lavas and dolerites; (3) to cover the extensive alteration halo in the hangingwall; (4) to provide
information approximately 200 m along the strike of the HJB shear zone and adjacent wallrocks in an
inclined longitudinal section that cross-cuts the lode at approximately 350 m below the present
surface; and (5) to provide information from areas below the economic gold mineralisation.
Drill core logging was used as a guide to sampling. At least one sample was taken from each
alteration zone that extended over more than 5 m in drill core, but narrower zones (down to 5 cm
wide) were also sampled in order to obtain comprehensive sample sets from alteration sequences. To
avoid mixing of different alteration zones in a single sample, and to provide a larger areal coverage
with a limited number of samples, sampling was not undertaken at regular intervals, nor with regular
sample sizes. Most of the samples selected represent typical examples of the alteration zones, and,
commonly, only a single sample was taken from each zone. However, if a zone was wide (>10-20 m)
or suspected to include distinct subzones, several samples at 0.5-15 m intervals were taken. In the
cases where zones of apparently homogeneous alteration or of unaltered rock were more than 10-20
m wide, multiple samples were also taken at roughly equal distances between the sample sites. In
addition, some sampling concentrated on, and close to, boundaries between alteration zones. The size
of the sample taken during this study varied from 5 cm to 1 m of diamond-drill core - half or quarter
core, depending on how much core was available.
7.3.3. Analytical procedures
Whole-rock geochemical analyses were made on 92 samples selected on the basis of the alteration
mapping and thin section investigations. All possible precautions were taken in order to analyse only
samples of uniform alteration. However, this was not possible for a minority of samples from
proximal alteration zones. The latter samples (5% of the total number of samples analysed) may have
Page 102
contained up to 20 vol.% of other alteration assemblages intermixed. Nevertheless, analyses of these

mixtures were considered separately in modelling of metasomatism.
Major element oxides, Cr and Zr were analysed by XRF, loss of ignition (LOI) by gravimetry, As,
Ba, Cu, Nb, Ni, Pb, Rb, Sb, Sr, Y, Zn and W by ICPMS and V by ICPOES at the Analabs Pty. Ltd.
laboratories in Perth, Australia. Gold and the potential pathfinder elements, Ag, Bi, Se and Te, were
analysed at the Geological Survey of Finland Rovaniemi laboratories by graphite-furnace atomic
absorption spectrometry (GAAS; Niskavaara and Kontas, 1990).
7.4. Host rocks
Assessment of primary host rock mineral and chemical composition is essential for establishing
alteration zoning sequences and for modelling metasomatism related to mineralisation. The general
trends in mineralogical and chemical composition of unaltered rocks at Bulletin are therefore
assessed first, in this section, prior to evaluation of mineralogical and chemical changes related to
gold mineralisation.
7.4.1. Petrography
Unaltered mafic host rocks are massive, undeformed and dark green in colour. Textures are primary
magmatic in origin, intergranular and, if the rock is fine-grained, generally microporphyritic and
amygdaloidal, although, in some places, seriate and variolitic as well. The most common phenocryst
mineral is plagioclase, although completely chloritised olivine and clinopyroxene phenocrysts occur
sporadically. Groundmass olivine and clinopyroxene have been partially to completely replaced by
chlorite and actinolite and calcic plagioclase by albite and epidote. The dominant Ti mineral is
titanite. In the pillow lava units, the pillow margins are epidote-dominated, and the inter-pillow
matrices consist of epidote, calcite, quartz, chlorite and traces of pyrite.
Epidote- and chlorite-filled veins, the contents of the pillow margins and inter-pillow matrices, and
the partial to complete, pseudomorphous replacement of olivine and clinopyroxene by chlorite and
actinolite suggest that "unaltered" rocks are in fact weakly altered, probably during an early
synvolcanic, submarine, spilitic stage in the evolution of the volcanic sequence (c.f., Barley et al.,
1990; McCuaig and Kerrich, 1994).
Page 103
7.4.2. Geochemistry
Mg
Fe-tholeiite
tho
lei
ite
Basaltic
komatiite
Calc-alkaline field
Al2O3 MgO
Fe tholeiite Mg tholeiite-1 Mg tholeiite-2
Unaltered rock Altered rock Least altered Mg tholeiite-2
Fig. 4. Jensen cation plot (Jensen, 1976) for mafic host rocks from Bulletin. Only the lower left
corner of the Jensen plot is shown, as all analysed samples, both altered and unaltered, are located
on that area. The least altered samples are marked separately for Mg tholeiite-2 because no
unaltered samples were obtained from that suite. The hatched areas cover the domains of each
basalt suite at Bulletin.
All analyses of the mafic host rocks at Bulletin are displayed on a Jensen cation plot (Jensen, 1976) in
Figure 4, where they form three separate groups. As the mass balance evaluations below show, all of
the components of the Jensen plot have remained immobile at Bulletin. The sample suites lie within
three fields interpreted as separate suites and referred to as Fe tholeiite, Mg tholeiite-1 and Mg
tholeiite-2, respectively. The median chemical compositions of the dominant suites, Fe tholeiite and
Mg tholeiite-1, are given in Tables 1 and 2. Geochemical data from the Mg tholeiite-2 are not
presented, as the suite is not a major host to ore nor a major rock type in the area. The three groups
represent separate basalt suites with distinct primary chemical characteristics and which could not
have been produced by fractionation of a single parent magma or by alteration of originally
homogeneous rocks. There are, however, no apparent differences between the basalt suites in terms
of structure, textures or mineral assemblages.
In Figure 5, the three basalt suites are presented on two plots of potentially immobile elements, where
they again occupy separate trends. As a whole, the data show a clear separation into the three groups,
each having characteristic Al/Ti and Al/Zr ratios. Some igneous fractionation within the Mg
tholeiites, as evidenced by the relatively wide range of Mg, Ti and Zr values in unaltered to least
altered samples, is apparent (Figs 4 and 5). The primary variation within Mg tholeiite-2 is larger than
for the other suites, as suggested by the least altered samples in Figure 5. The distribution of the
various basalt suites at Bulletin is shown in Figure 3. The suites cover relatively coherent domains,
and are only locally chaotically intermixed due to deformation within the HJB shear zone.
Page 104
18
16
14
12
10
8
6
4
2
0
0 0.2 0.4 0.6 0.8 1.0 1.2
TiO2
18
16
14
12
10
8
6
4
2
0
0 20 40 60 80
Zr
Rock and alteration types:
Fe tholeiite, unaltered Mg tholeiite-1, unaltered
Fe tholeiite, altered Mg tholeiite-1, altered
Mg tholeiite-2, least altered Mg tholeiite-2, intermediate

and proximal alteration
Fig. 5. Al2O3 vs. TiO2 and Zr for mafic host rocks from Bulletin. Diagonal lines indicate the suite-
characteristic conserved element ratios, i.e., trends on which samples within the suites ideally plot
when elements considered are conserved. Error bars define the total error range.
Page 105
Table 1. Median compositions of the Fe tholeiite at Bulletin, classified according to alteration zone
and style. Concentrations of major oxides, LOI, CO2 and S are in wt.% and of trace elements in
ppm, except for Au and Te, which are in ppb. Trace elements after Ag are presented in the order of
potentially decreasing mobility during alteration (Rose et al., 1979). N is the number of samples
included in a sample group.
Alteration Unaltered Chlorite-calcite Calcite- Sericite Dolomite-

zone rock dolomite sericite
Subzone Chlorite- Calcite- Calcite Calcite- Calcite- Dolomite-
actinolite epidote chlorite- dolomite- chlorite-
sericite chlorite- sericite
sericite
SiO2 49.4 48.0 46.9 46.0 44.3 46.4 45.0 42.0 41.2
TiO2 1.06 1.06 1.04 0.96 1.01 0.92 0.93 0.94 0.75
Al2O3 13.65 13.85 13.50 12.60 13.00 11.95 12.40 13.70 11.00
Fe2O3tot 13.88 14.08 12.86 11.91 12.75 12.00 12.09 12.65 10.97
MnO 0.22 0.21 0.21 0.19 0.19 0.18 0.19 0.15 0.17
MgO 6.46 6.45 5.30 5.45 5.15 5.34 5.71 6.43 4.65
CaO 9.52 8.66 9.90 10.12 9.64 9.20 8.63 9.59 11.67
Na2O 2.11 1.97 2.15 2.11 1.77 0.45 0.57 0.47 0.95
K2O 0.13 0.11 0.09 0.18 0.33 0.69 0.66 0.59 1.00
P2O5 0.09 0.09 0.08 0.08 0.08 0.08 0.07 0.08 0.04
LOI 3.20 5.31 9.01 9.84 11.59 9.29 12.06 12.05 13.48
sum 99.7 99.8 101.0 99.4 99.7 96.5 98.3 98.7 95.9
S 0.109 0.062 0.076 0.150 0.145 0.365 0.310 0.390 3.570
CO2 0.73 2.09 6.23 7.26 8.18 7.17 9.57 8.36 13.79
Au 0.7 n.a. 5.8 3.5 11 269 166.5 21.9 15000
Te 13 n.a. 12 10 16 33 56 17 930
Ag 0.05 n.a. 0.06 0.04 0.03 0.10 0.16 0.06 0.22
Se 0.41 n.a. 0.39 0.38 0.23 0.21 0.35 0.25 0.25
Sb 3.3 3.3 5.2 3.1 7.8 7.7 6.0 14.6 25.4
As 56 65 97 76 148 341 305 209 19200
Bi 0.01 n.a. 0.01 0.01 0.01 0.01 0.01 0.01 0.02
W 0.2 0.2 0.2 1.1 1.4 6.5 5.6 8.0 12.0
Rb 6.0 4.8 5.3 8.1 13.4 26.1 21.6 21.2 40.6
Ba 43 26 31 26 32 41 49 44 104
Sr 138 128 117 79 81 75 77 82 105
Pb 5 1 2 2 2 2.5 2 2 4
Cu 158 155 152 146 139 145 132 139 108
Zn 117 120 114 110 105 99 100 131 63
Ni 112 126 116 106 108 103 105 144 100
Nb 3.5 3.3 3.1 3.0 3.1 2.9 2.8 3.8 2.4
Y 27.8 28.1 24.2 21.1 11.5 13.1 10.9 11.2 10.6
V 263 291 273 255 257 251 252 236 231
Cr 147 151 137 126 133 125 130 145 119
Zr 70 67 64 62 64 58 60 62 47
N 2 2 5 7 3 8 5 1 5
n.a.: not analysed
Page 106
7.5. Hydrothermal alteration
Wallrock alteration is pervasive throughout the entire width of the HJB shear zone and extends into
the surrounding wallrocks for 25-150 m from the shear zone boundary on the hangingwall and 10-20
m on the footwall side. There seems to be a positive correlation between the maximum lateral extent
of alteration and the width of the HJB shear and its contained ore zones.
Table 2. Median compositions of the Mg tholeiite-1 at Bulletin, classified according to alteration

zone and style. Concentrations of major oxides, LOI, CO2 and S are in wt.% and of trace elements in
ppm, except for Au and Te, which are in ppb. Trace elements after Ag are presented in the order of
potentially decreasing mobility during alteration (Rose et al., 1979). N is the number of samples
included in a sample group.
Alteration Unaltered Chlorite-calcite Calcite- Sericite Dolomite-

zone rock dolomite sericite
Subzone Chlorite- Calcite- Calcite Calcite- Calcite- Dolomite-
actinolite epidote chlorite- dolomite- chlorite-
sericite chlorite- sericite
sericite
SiO2 48.2 49.0 45.0 45.0 45.7 46.3 42.0 40.9 42.4
TiO2 0.84 0.81 0.77 0.77 0.81 0.77 0.74 0.82 0.68
Al2O3 14.35 14.70 13.20 13.20 13.80 13.15 13.00 14.70 11.45
Fe2O3tot 11.52 11.01 10.46 10.52 10.33 10.36 10.11 11.67 9.58
MnO 0.18 0.17 0.17 0.16 0.16 0.14 0.16 0.15 0.18
MgO 8.12 8.15 7.84 7.43 6.06 7.21 6.40 6.33 5.56
CaO 9.89 10.56 10.17 9.53 9.67 8.92 10.52 10.50 12.21
Na2O 2.09 1.98 1.82 1.66 2.64 0.89 1.03 0.58 0.92
K2O 0.10 0.07 0.10 0.12 0.14 0.62 0.60 0.83 0.77
P2O5 0.07 0.06 0.06 0.06 0.06 0.06 0.06 0.07 0.06
LOI 4.18 3.60 9.84 11.16 10.55 10.60 14.18 13.10 15.44
sum 99.5 100.1 99.4 99.6 99.9 99.0 98.8 99.7 99.2
S 0.100 0.087 0.073 0.068 0.047 0.340 0.140 0.140 0.470
CO2 1.01 0.81 5.98 6.49 7.13 7.04 11.11 8.76 13.35
Au 3 2.4 1.3 3.2 2.95 273 3.5 122 1900
Te 12 11 12 12 8 82 13 19 74
Ag 0.04 0.06 0.04 0.03 0.02 0.14 0.04 0.08 0.15
Se 0.38 0.38 0.27 0.27 0.25 0.42 0.34 0.63 0.25
Sb 1.1 1.8 2.2 1.8 5.3 6.5 5.5 12.3 15.0
As 18 17 18 32 29 183 48 289 319
Bi 0.01 <0.01 <0.01 <0.01 <0.01 0.01 <0.01 0.01 0.02
W 0.2 0.2 0.2 0.8 0.5 4.4 0.6 5.8 4.5
Rb 2.5 2.1 3.9 5.2 6.1 22.6 21.5 28.4 28.5
Ba 29 27 27 28 30 61 41 62 57
Sr 111 130 137 66 86 91 72 104 112
Pb 2 2 2 1 1 2 2 2 2
Cu 149 156 147 124 133 121 125 102 101
Zn 101 96 85 90 108 102 104 122 92
Ni 189 145 145 134 78 124 134 194 110
Nb 2.6 2.6 2.4 2.4 3.8 2.3 2.6 2.3 2.2
Y 21.4 21.3 19.4 15.3 10.3 11.3 10.4 10.6 9.8
V 231 235 209 214 220 213 203 233 191
Cr 338 377 358 344 162 289 266 179 190
Zr 53 50 46 47 50 46 49 55 42
N 6 9 3 11 2 4 5 1 2
n.a.: not analysed
Page 107
Alteration mineral assemblages are distinctly zoned, although the coalescence of overlapping
alteration haloes surrounding both the main shear zone and minor hangingwall faults has resulted in
complex zoning patterns (as shown later, in Fig. 12). In addition, mineral assemblages in individual
alteration zones suggest that at least two separate zoning sequences occur within the Bulletin mine
environment. Because such zoning sequences themselves reflect the progressive interaction between
an infiltrating fluid and its wallrocks, any differences in the alteration zoning patterns can be
attributed to only any one of four possibilities: 1) multiple episodes of alteration along the same
localising structure, 2) changes in bulk rock composition of the host rock, 3) differences in the initial
ore fluid composition, or 4) variation in the P-T conditions of the environment (Meyer and Hemley,
1967; Rose and Burt, 1979). For Bulletin, the origin of contrasting zoning patterns is still unclear,
and remains the subject of ongoing research (S. Chanter, unpubl. data). There are no indications of
post-alteration metamorphism at Bulletin, but rather the textures indicate syn-metamorphic and syn-
peak deformation alteration and gold mineralisation, as the minerals formed during alteration define
all foliations present in the area and envelop both veins that are deformed and veins that truncate
foliated texture.
7.5.1. Alteration zoning patterns
Mineral assemblages within the alteration envelope associated with the HJB shear zone, hangingwall
faults and the Bulletin orebody can be divided into a number of alteration zones depending upon the
mineral assemblages present. The distributions of the major alteration zones is presented in
geochemical maps in Figure 12.
Two main alteration zone sequences have been identified, one being dolomite-bearing and the other
dolomite-free. The progressive mineralogical changes associated with each of the sequences are
summarised in Figures 6 and 7, respectively, and are discussed in further detail below. Both are
characterised by identical assemblages in the distal alteration zones, but differ significantly in their
intermediate and proximal zone assemblages. The dolomite-bearing alteration sequence dominates
and characterises virtually all of the high-grade ore zones in the Bulletin orebody. The dolomite-free
sequence, on the other hand, is more prevalent in lower grade, more peripheral portions of the gold
mineralisation and in unmineralised hangingwall faults. It is also clear that the distribution of the two
alteration zoning sequences is not related to that of the three basalt suites. Individual alteration zones
can be present at all scales, and range in width from 2 cm or less to 75 m. Regardless of scale, the
mineralogical changes that can be traced laterally from a controlling structure produce consistent and
repeatable patterns, suggesting that the alteration zones form part of a single zoned alteration halo
rather than separate and superimposed alteration events.
7.5.2. Dolomite-present zoning patterns
The dolomite-bearing alteration sequence can be divided into four major alteration zones and a
number of subzones (Fig. 6). Gold is, for the most part, associated with the dolomite-sericite zone,
but minor gold mineralisation may also occur within the sericite zone. Sericite alteration zones are
subdivided into outer- and inner-sericite subzones on the basis of mineral assemblages. The outer
sericite zone is dominated by chlorite-sericite-carbonate (both Fe-calcite and dolomite), whereas the
inner zone contains either sericite-Fe-calcite-dolomite or sericite-chlorite-dolomite mineral
assemblages. The different mineral assemblages of the inner subzone are mutually exclusive, and
most probably reflect whether chlorite or calcite is depleted during advancing alteration. This, in turn,
depends upon the modal proportions of chlorite and calcite in the rock and ultimately relates to subtle
differences in original bulk rock chemistry. The main alteration zones are described in detail below.
Page 108
7.5.2.1. Chlorite-calcite zone
Chlorite-calcite zone mineral assemblages dominate in unsheared and weakly sheared host rock in
the foot- and hangingwall where the primary magmatic and metamorphic textures and fabrics are
commonly preserved. The width of this zone ranges from 0.5 to 75 m.
The appearance of dark green chlorite-actinolite spots, 2-20 mm in diameter, is the first, most distal
indication of hydrothermal alteration at Bulletin. These spots form 20-30% of the total rock volume,
and appear gradually over an interval of 5-25 m. Because the spots are only developed in medium-
grained rocks and are the only indication of incipient alteration visible in hand specimen, it is not
always possible to accurately define the outermost extent of alteration in fine-grained rocks at
Bulletin. Microscopic observations indicate a corresponding gradual replacement of relict olivine and
clinopyroxene of the unaltered rocks by chlorite and actinolite.
The boundary between the chlorite-actinolite subzone, described above, and the adjacent calcite-
epidote subzone is marked by a dramatic increase in disseminated calcite. Actinolite is completely
replaced by chlorite in the calcite-epidote subzone. Spotting in medium-grained rocks, however,
persists as a distinct feature until the onset of moderate to intense foliation within calcite subzone or
calcite-dolomite zone.
INTER-
DISTAL PROXIMAL
MED.
Zone
Chlorite-calcite Sericite
Unaltered Calcite- Dolomite-
Subzone dolomite Calcite- Calcite- Dolomite- sericite
Chlorite- Calcite-
dolomite- dolomite- chlorite
actinolite epidote Calcite chlorite- sericite sericite
Mineral sericite
SILICATES
Quartz
Albite
Olivine
Clinopyroxene
Actinolite
Epidote
Titanite
Chlorite
Sericite
CARBONATES
Calcite, Fe-free
Calcite, Fe-bearing
Dolomite
OXIDES
Magnetite1
Rutile
SULPHIDES
Pyrite
Arsenopyrite
1 Very minor, not present in most cases
Fig. 6. Schematic summary of the dominant, dolomite-bearing, paragenetic alteration sequence in

mafic rocks at Bulletin. Proximity to mineralisation increases to the right, except for calcite-
dolomite-sericite and dolomite-chlorite-sericite subzones, which are both developed next to the most
proximal zones, but are mutually exclusive. Zone widths shown here bear no resemblance to actual
widths in the field.
Page 109
Calcite subzones represent the most intense alteration of the chlorite-calcite zone rocks, and are
recognised by the complete disappearance of epidote and titanite (replaced by chlorite, calcite, quartz
and rutile) and the replacement of Fe-free calcite by Fe-bearing calcite.
7.5.2.2. Calcite-dolomite zone
With increasing deformation, primary magmatic and metamorphic features and chlorite spotting
gradually disappear. Calcite-dolomite alteration zones up to 5 m wide are developed within
hangingwall faults and at the margins of the HJB shear zone. Within the zone, calcite gradually
decreases in abundance while dolomite increases towards the more proximal parts of the alteration
sequence.
7.5.2.3. Sericite zone
Sericite-bearing alteration zones form the proximal part of alteration sequences at Bulletin, and occur
chiefly within the HJB shear zone itself or, less commonly, as narrow (<3 m wide) domains within
the hangingwall shear zones. The transition between calcite-dolomite and sericite zones is clearly
marked by the emergence of yellow sericite. Where ore-grade gold mineralisation is widest, sericite
zones form relatively thin, 0.1-3 m wide, envelopes around the dolomite-sericite alteration of the ore
zone. Where gold mineralisation is of lower grade or less extensive, sericite zones may form
continuous domains up to 25 m wide and represent the dominant alteration type within the HJB shear
zone.
DISTAL PROXIMAL
Zone
Unaltered Chlorite-calcite Calcite-chlorite-
sericite
Subzone Chlorite-
Calcite-epidote Calcite
Mineral actinolite
SILICATES
Quartz
Albite
Olivine
Clinopyroxene
Actinolite
Epidote
Titanite
Chlorite
Sericite
CARBONATES
Calcite, Fe-free
Calcite, Fe-bearing
OXIDES
Magnetite1
Rutile
SULPHIDES
Pyrite2
Arsenopyrite
1 Very minor, not present in most cases
Fig. 7. Schematic summary of the dolomite-absent paragenetic alteration sequence in mafic rocks at
Bulletin. Proximity to mineralisation increases to the right. Zone widths shown here bear no
resemblance to actual widths in the field. Note that while the distal assemblages are identical to that
Page 110
of the dolomite-bearing sequence, the intermediate calcite-dolomite zone is not present and that
dolomite is absent in the proximal zone, too.
The most distinguishing feature of the sericite zone alteration is the presence of yellow sericite
lamellae and disseminated sericite in an otherwise dark green rock characterised by relatively
abundant (15-30 vol.%) chlorite. The modal abundance of sericite is between 5 and 20 vol.%.
Sulphide minerals comprise between 1 and 6 vol.% (mostly in the low portion of this range). Pyrite is
dominant but arsenopyrite may be important locally. With increasing proximity to ore, fluid-rock
interaction resulted in the formation of sericite-dolomite assemblages. Sericite-chlorite-dolomite is
the most common inner assemblage, forming 1-20 m wide zones. Calcite-dolomite-sericite subzones
are, by contrast, narrow (≤5 cm in most cases) and relatively scarce.
7.5.2.4. Dolomite-sericite zone
Dolomite-sericite zone represents the most intense alteration at Bulletin. It is dominant in those parts
of the HJB shear zone where ore is most extensive, comprising domains up to 20 m wide. In barren
or low grade areas, the zone is absent or of limited extent. The mineral assemblage is characterised
by high sulphide contents and complete destruction of chlorite and calcite (Fig. 6). Fine-grained,
idioblastic arsenopyrite and pyrite comprise 5-8 vol.%, imparting a darkish grey colour to the rocks.
High sulphide concentrations generally correlate with high gold grades (up to 50-100 ppm). Sericite
comprises between 20-40 vol.%, and quartz-carbonate veins make up between 15 to 50 vol.% of the
rock.
7.5.3. Dolomite-absent zoning patterns
Dolomite-absent alteration sequences are subordinate to dolomite-bearing haloes and typify the
peripheral, low grade portions of Bulletin lode and some of the unmineralised hangingwall shear
zones. Chlorite-calcite zone assemblages are identical to those in dolomite-present haloes. However,
the intermediate and proximal zones of the dolomite-present sequence are replaced by a proximal
alteration assemblage characterised by chlorite-calcite-sericite (Figs 6 and 7). Rocks with the latter
mineral assemblage are characterised by anastomosing sericite lamellae and disseminated sericite in
an otherwise dark green rock containing relatively abundant (15-30 vol.%) chlorite. Both
arsenopyrite and pyrite may be present, but are generally less abundant than in the dolomite-sericite
alteration zones.
7.6. Chemical changes related to alteration
Geochemical features of the altered host rocks, mass balance between alteration zones and unaltered
rocks, and the relationship between metasomatic changes and changes in mineral assemblages at
Bulletin are described and discussed in this section, followed by a further discussion of the dispersion
of mobilised components.
Median chemical compositions of the dominant suites, Fe tholeiite and Mg tholeiite-1, including both
unaltered and altered types, are presented in Tables 1 and 2. Chemical mass balance for Bulletin was
evaluated by using simple variation diagrams, the methods of Gresens (1967) and MacLean and
Barrett (1993), and the Pearce Element Ratio method (Stanley and Madeisky, 1993).
In this paper, elements not affected by metasomatic changes are described by two terms, ‘conserved’
and ‘immobile’. Conserved elements are those which were neither affected by igneous fractionation
nor by metasomatic processes (Stanley and Madeisky, 1993). Immobile elements include all
Page 111
elements which were neither added or lost from the rock during alteration, whether or not they were
affected by igneous fractionation (Gresens, 1967).
The Mg tholeiite-2 suite is not included in the detailed mass transfer evaluations, because it
represents only a small percentage of the host rocks, and no unaltered or weakly altered samples were
obtained from that suite. In addition, the variation in primary chemical composition is large which,
combined with small number of samples obtained from the suite, makes the potential error in mass
transfer evaluations large. On the other hand, the alteration zones and changes in mineral
assemblages detected in Mg tholeiite-2 are the same as in the other basalt suites. It is therefore most
probable that the mass transfer related to alteration in Mg tholeiite-2 was similar to that in the other
two suites.
7.6.1. Mass transfer at Bulletin
Variation diagrams were used to check which components are conserved, which are affected by
fractionation, and to qualitatively evaluate the mobilities of elements during alteration. Selected
variation diagrams, in addition to Figure 5, are presented in Figure 8. These diagrams, and similar
plots not presented here, show that Nb, Ti, V and Zr are conserved in all suites and that, in the Fe
tholeiite, Al, Fe and Cr are also conserved. In contrast, at least Au, Ag, As, CO2, K, Rb, Sb, S, Te
and W were enriched and Na and Y depleted during alteration (see, as examples of mobilised
elements, variation diagrams of K2O and Y vs. TiO2 in Fig. 8).
For tholeiitic basalts, where the fractionation of olivine, clinopyroxene and plagioclase accounts for
most of the mass transfer prior to any alteration, the Pearce Element Ratio (PER) method (Stanley
and Madeisky, 1993) can be used to check the mobilities of the major elements Al, Ca, Fe, Mg, Na
and Si during alteration while taking the effects of fractionation into account. Figure 9 presents the
olivine-clinopyroxene-plagioclase fractionation diagram for the Bulletin Fe tholeiites. Both Fe
tholeiite and Mg tholeiite-1 show similar trends in the PER fractionation diagrams, although, in
detail, the trends are not simple and most samples fall within error of the crystal fractionation control
lines. Samples from distal and intermediate alteration zones scatter on both sides of the modelled
igneous fractionation trend, possibly indicating both slight mass gain or slight mass loss of the
components in question during alteration. On the other hand, samples from the outer proximal,
sericite alteration zones show the most substantial net material transfer, but the most proximal
samples tend to plot close to the fractionation line (Fig. 9).
Page 112
16
Fe tholeiite 14
Mg tholeiite-1
14 12
12
10
10
8
8
6
6
4
4
2 2
0 0
2.5 1.2
1.0
2.0
0.8
1.5
0.6
1.0
0.4
depletion
0.5 0.2
enrichment
0 0
30 25
25
20
20
15
15
10
10
5 5
0 0
80 70
60
60 50
40
40
30
20
20
10
0 0
0 0.2 0.4 0.6 0.8 1 1.2 0 0.2 0.4 0.6 0.8 1
TiO2 TiO2
Unaltered rock Proximal alteration, sericite zone

Distal alteration Proximal alteration, dolomite-sericite zone
Intermediate alteration
Fig. 8. Selected variation diagrams for Fe tholeiite and Mg tholeiite-1. Solid diagonal lines define
the trends where samples ideally plot if the considered elements are conserved. Dashed line in the
Fe-Ti variation diagram defines the fractionation trend based on unaltered samples. Error bars
define the total error range.
Page 113
70
Fe tholeiite
60
50
40
30
20 Unaltered rock
Distal alteration
10 Intermediate alteration
l
-P Proximal alteration, sericite zone
px
-C
Proximal alteration, dolomite-sericite zone

Ol
0
0 20 40 60 80 100 120 140
Si/Ti
Fig. 9. Olivine-clinopyroxene-plagioclase fractionation PER model diagram for Fe tholeiite. The
heavy diagonal line represents the trend on which fractionation of olivine, clinopyroxene and
plagioclase would put any sample from the suite. The light diagonal lines define the range of error in
the fractionation line.
The distance of a sample from the PER model fractionation line is directly proportional to the net
material transfer during alteration of all elements on which the PER diagram is based. This distance is
termed the absolute residual value (Stanley and Madeisky, 1993). The effect of a single element in a
PER diagram of the type shown in Figure 9 can be investigated by plotting a ratio between the
element under consideration and the conserved element, e.g. Si/Ti, vs. the absolute residual value.
This was done for Al, Fe, Mg, Ca, Na, and Si for both Fe tholeiite and Mg tholeiite-1. These plots
show that the complicated trend defined by altered rocks (Fig. 9) is best explained by the combined
net effect of two mobile components, Na and Si (Fig. 10). Local Na enrichment is related to distal
and intermediate alteration, and significant Na depletion to the outer proximal sericite alteration
zones, while Na is enriched relative to the sericite zones in the inner proximal dolomite-sericite
zones. Silica shows significant mobility and enrichment, locally in both proximal zones. These
transfers, as well as insignificant mobility of Ca and immobility of Al, Fe and Mg, are confirmed by
the mass balance calculations below.
Mass balance for each sample and element analysed, except Bi and Pb, was calculated by the method
of Gresens (1967) in Fe tholeiite and by the method of MacLean and Barrett (1993) for Mg tholeiite-
1. Bismuth and lead were excluded, because a significant number of samples throughout the
alteration sequence (approx. 40% for Bi, 30% for Pb) have concentrations below the respective
detection limits (0.01 ppm for Bi, 1 ppm for Pb). The more simple method of Gresens was used for
Fe tholeiite, because the degree of fractionation within the suite is negligible. For Mg tholeiite-1, the
use of the MacLean-Barrett method was necessary due to the presence of significant igneous
fractionation within the suite.
To illustrate the mass transfer trends during alteration, the median enrichment factors for each
component within each alteration zone are presented in Figure 11. When inspecting these histograms,
the effect of analytical error on the enrichment factors should be taken into account. It is reasonable
to consider the value of the enrichment factor from approximately 0.9 to 1.1 for the major
components, 0.8-1.2 for CO2 and S, and 0.7-1.3 for the trace elements, as variations due to the total
analytical error, rather than an indication of gain or loss due to alteration. The results of mass balance
Page 114
calculations for individual samples are not reproduced here, but are considered in the discussion
below.
gai
130 10
n
120
8
110
100
6
90
in
ga
Na/Ti
Si/Ti
80 4
Si
70
2
60
50 0
-20 0 20 40 60 80 -20 0 20 40 60 80
Absolute Residual PER Absolute Residual PER
Unaltered rock
Distal alteration
Proximal alteration, sericite zone
Proximal alteration, dolomite-sericite zone
Fig. 10. The mobilities of Si and Na shown by the Si/Ti and Na/Ti PER ratios vs. the absolute
residual value diagrams for Fe tholeiite. The area where apparently immobile cases of the element
considered (Si or Na) would plot is enveloped by a dashed rectangle. Error bars present the total
error.
7.7. Discussion
Mass transfer at Bulletin is considered here by grouping the elements according to their geochemical
behaviour: (1) conserved elements, (2) immobile elements, (3) silica, (4) alkali and alkali-earth
elements, (5) volatiles, (6) Au and potential pathfinder elements, and (7) other trace elements (Cu, Zn
and Y).
1) Variation diagrams, and PER and other mass balance calculations (Figs 5, 8-11) show that,
in Fe tholeiite, Ti, Fe, Nb, V, Cr and Zr are conserved elements, while Ti, Nb, V and Zr are
conserved in Mg tholeiite-1.
2) In Fe tholeiite, Al, Mn, Mg and Ni are immobile, although possibly not conserved elements,
while in Mg tholeiite-1, Al, Fe, Mn, P, Ni and Cr are immobile, but clearly not conserved.
For most of the elements, immobility is shown by both PER evaluation and mass balance
calculations. Although mass balance calculations suggest minor Mg depletion in
intermediate and proximal alteration zones in Mg tholeiite-1, PER values clearly indicate
Mg immobility for that suite as well. Phosphorus is generally immobile, but depleted in
areas of most intense alteration in Fe tholeiite, probably reflecting the replacement of apatite
by sericite and carbonate. There are indications of erratic Ni and Cr mobility in Mg
tholeiite-1 (Fig. 11B). However, much of the mass transfer in Ni and Cr is within the limits
of analytical error, while the remainder could be explained by fractionation that could not
be modelled with mass balance calculations based on Al and Ti alone.
3) The relative transfer in Si during alteration in the host rocks is generally minor or
negligible. However, a few individual samples from proximal alteration zones show
substantial gains in silica (Fig. 10). This enrichment is, however, related to intense quartz
vein formation in the HJB shear zone, not silicification of the host rock. It was not possible
Page 115
to remove all vein material from samples prior to geochemical analyses, as most of the
veins are <1 cm wide and, in the proximal areas, their relative volume of the rock can be
large.
A
Fe tholeiite
1000
Distal alte ration
Intermedia te alteration
Proximal alteration,
100 sericite zone
dolomite-sericite zone
10
0.1
Y
V
W
Rb
Se
Ni
Sr
Cr
Ba
Zn
Sb
Cu
Ag
As
Te
Zr
P2O 5
Au
Nb
CaO
SiO2
TiO 2
CO2
MnO
MgO
K 2O
Na 2O
Al2O3
Fe2O3
B
Mg tholeiite-1
100
Distal alte ration
Intermedia te alteration
sericite zone
10 dolomite-sericite zone
0.1
Y
V
Ni
S
Sb
Zr
Se
Nb
Cr
As
Ba
Cu
Te
Au
Ag
Sr
Zn
Rb
SiO 2
P2O 5
CO 2
K 2O
CaO
MnO
MgO
Fe2O 3
Na2O
Fig. 11. Median gains and losses (enrichment factors) in Fe tholeiite (A), calculated by the method of
Gresens (1967), based on the conserved elements Ti and Zr, and Mg tholeiite-1 (B), calculated by the
method of MacLean and Barret (1993), based on the fractionation line defined by Al and Ti which
are, hence, excluded in the figure. The median unaltered composition has enrichment factor values
equal to 1; values below 1 indicate depletion and values above 1 enrichment. The potential mobility
of trace elements, except for Au, Ag and Te, decreases from left to right (Rose et al., 1979).
4) Although not obvious from Figure 11, individual samples show local enrichment for Ca and
Na in areas of distal alteration. This may be the result of minor spilitic alteration during the
early evolution of the volcanic sequence at Wiluna. This is supported by the significant Ca
and Na enrichment in some samples from pillow lava units which would represent the most
permeable rock units during seafloor hydrothermal circulation, as compared to the more
massive units. The mobility of Ca is not uniform in the proximal zones either, most samples
showing either minor enrichment or no change at all; enrichment is probably related to
carbonation.
Proximal alteration zones are generally depleted in Na relative to unaltered host rock, with
the greatest Na loss in the outer proximal, sericite zone. The Na depletion is most probably
due to sericitisation of albite, as it coincides with prominent K and Rb enrichment. The
trend in Na mobility reverts to enrichment in the inner proximal zone compared to the outer
proximal zone, although the rocks remain Na-depleted compared to unaltered areas. This
Page 116
apparent lessening of the amount of Na transfer is attributed to the presence of narrow

albite-bearing quartz veins, rather than albitisation within the wallrocks.
Both K and Rb are enriched throughout the alteration sequence. The enrichment becomes
gradually more prominent through the distal and intermediate zones, and is discernible in
every sample at the boundary between intermediate and proximal alteration, where sericite
becomes visible to the naked eye. The gains in K and Rb are related to sericitisation of
albite and chlorite.
Barium is either immobile or slightly depleted in distal and intermediate zones, but is
typically enriched in proximal zones. The reasons for depletion are unclear but enrichment
in proximal zones is easily explained by the intense sericitisation.
Strontium is, on average, slightly depleted in intermediate and proximal zones, except in the
most proximal zones where it shows slight relative enrichment, attaining levels similar to
those in unaltered rocks. Depletion is possibly related to replacement of calcite and albite by
dolomite and sericite, respectively, while the proximal enrichment may be related to
formation of vein-albite.
5) CO2 is enriched in all alteration zones, reflecting the carbonation of silicates (Fig. 11).
Sulphur shows limited erratic mobility in the distal and intermediate alteration zones. Part
or all of this mobility may be related to the early spilitic alteration, which is consistent with
the presence of pyrite in inter-pillow matrices and in some fractures in areas of distal
alteration and unaltered rock. In contrast to distal and intermediate alteration zones,
prominent S enrichment covers the proximal zones, reflecting the formation of gold-related
pyrite and arsenopyrite.
6) Median enrichment factors suggest enrichment for Au roughly as extensively as that for the
pathfinder elements As, Sb, Te and W. In detail, however, there are significant differences
between Au and these pathfinder elements, and between all of the pathfinder elements.
Arsenic and antimony are probably enriched in the entire study area, as is shown by their
anomalous concentrations compared to typical unaltered basalts (see Tables 1-3 and
discussion of background thresholds below). Therefore, the immobility of As and Sb
implied by Figure 11 for the distal alteration zones is apparent rather than real.
Mass balance calculations give erratic trends for Te and Ag in distal and intermediate zones,
but consistent enrichment in proximal zones. For Ag, this probably indicates an absence of
significant mobility outside proximal alteration, where the range of Ag concentration falls
within that of the normal unmineralised background values (Tables 1-3). For Te, however,
individual samples in distal alteration zones and unaltered rocks have abundances that are
predominantly above, and only locally within, the range of unmineralised igneous rocks,
indicating widespread enrichment and, probably, no depletion at all.
Page 117
Table 3. Background threshold value estimates based on cumulative frequency plots for Au and
selected pathfinder elements at Bulletin, with published average concentrations of these elements in
unmineralised mafic rocks. Data from Lahtinen and Lestinen (1996) are from Svecofennian
(Palaeoproterozoic) southern Finland; data from Eilu et al. (1996) are from the Bronzewing deposit
and its wallrocks, Archaean Yandal greenstone belt, Western Australia. All values are in ppm.
Element Threshold Average mafic rock Average Average Range for mafic Range for meta-
at Bulletin 1 oceanic basalt metavolcanic tholeiites (Eilu
( Tilling et al., 1973; Rose
tholeiite (Nurmi et al., rocks (Lahtinen et al. 1996)
et al., 1979; Perrault et al.,
(Ringwood, 1991) and Lestinen
1984;
1979) 1996)
2
Taylor, 1968;
3
Wedepohl, 1969)
Au 0.006 1 0.002 0.0002-0.003 0.0006-0.004
0.003-0.004
Ag 0.08 2 0.03 0.02-0.05 0.062-0.071 0.04-0.16
0.10
As 28 2 1.0 0.5-2.5 0.6-5.8 0.50-1.10
2.0
Sb 4 2 0.029 0.1-0.4 0.084-0.126 0.10-0.40
0.2
Te 0.023 3 0.002-0.008 <0.002-0.005
0.002
W 0.6 3 0.08 0.5 0.2-1.9
1.0
Tungsten exhibits relatively consistent enrichment within distal and intermediate a

alteration zones and substantial enrichment within proximal alteration, reflecting the
formation of scheelite in the altered rocks.
Selenium, although a potential pathfinder element, does not show any consistent
metasomatic trends. The Se-Ti variation diagrams (not presented) and mass balance
calculations, suggest that Se is relatively immobile, with variations in its concentration
probably being related to crystal fractionation and dilution by gains in mobile major
elements.
7) Copper and zinc show some very local enrichments and depletions that appear to be
independent of the gold-related alteration. They probably are immobile during the gold-
related alteration, the erratic changes reflecting primary variation and local enrichments and
depletions which have taken place during early spilitic alteration of the volcanic sequence.
As the gold mineralisation-related alteration has overprinted most of the weak spilitisation,
this cannot be confirmed in altered areas, but it is in agreement with the occurrence of
chalcopyrite and sphalerite in fractures, amygdales and inter-pillow matrices in unaltered
basalt at Bulletin.
Variation diagrams and mass balance calculations (Figs 8 and 11) show that Y, which is
quite commonly considered to be an immobile element during hydrothermal alteration (e.g.
Lesher et al., 1986; MacLean and Kranidiotis, 1987; Huston, 1993), is indeed immobile in
the outer areas of distal alteration, but is locally depleted in inner distal areas and
consistently depleted in intermediate and proximal alteration zones. As the most prominent
change in Y mobility takes place at the onset of intermediate alteration, the depletion seems
to be chiefly related to the formation of dolomite. To mobilise Y, dolomite has replaced
either Y-bearing chlorite or some accessory minerals, such as apatite and monazite, or both.
Page 118
7.8. Geochemical dispersion patterns
7.8.1. Background thresholds for Au, Ag, As, Sb, Te and W
Log-normal cumulative frequency plots were constructed for Au and pathfinder elements to provide
a first approximation for threshold values between local background and anomalous values
associated with primary dispersion related to ore (Table 3). The justification of the use of cumulative
frequency plots is discussed by Sinclair (1976, 1991), Rose et al. (1979) and Perrault et al. (1984),
among others. Bismuth and selenium, potentially useful pathfinder elements in lode-gold systems,
are excluded here because Bi analyses either plot within the normal background range for unaltered
mafic rocks or below detection limit (≤0.2 ppm - e.g., Taylor, 1968 and Nurmi et al. 1991, and <0.01
ppm, respectively), and Se behaved as an immobile element.
For Ag, Au and W, the threshold values obtained from the cumulative frequency plots agree fairly
well with data representing average unmineralised mafic rocks from the literature, while, for As, Sb
and Te, the thresholds defined are exceptionally high (Table 3).
None of the analysed samples has an As concentration and only one sample has Sb concentration low
enough to be considered within normal background range: the lowest concentrations detected are 9
ppm for As and 0.4 ppm for Sb. This suggests that 28 ppm for As and 4.0 ppm for Sb represent only
local thresholds and that larger, regional-scale anomalies probably exist. For Sb, dispersion map and
the Sb profiles (Fig. 12D) suggest that about 2.0 ppm would be more suitable than 4.0 ppm as the
local threshold between background and anomalous values. Starting from roughly 2 ppm level, there
clearly is an increasing Sb trend towards ore across oscillating variations in the alteration zoning
patterns. For As, the concentration trends or dispersion maps (Fig. 12C) do not suggest any obvious
threshold value below 28 ppm. Consequently, 2.0 ppm for Sb and 28 ppm for As have been used as
the thresholds defining local anomalies.
Arsenic and antimony concentrations define oscillating but, on average, increasing trends towards
ore across all the cross sections studied (Figs 12C and 12D). Data from the literature (Table 3)
suggests that the threshold values of 2-6 ppm for As and 0.2-0.5 ppm for Sb are probably realistic in
defining the total extent of anomalies for As and Sb in mafic rocks. From the As and Sb
concentration data from and trends at Bulletin, and from data from the literature, it is, therefore,
suggested that the regional As and Sb anomalies at Wiluna are defined by the maximum threshold
values of 6 and 0.5 ppm, respectively. This is also supported by regional data from the Wiluna
greenstones: a large number of unweathered, unaltered mafic metavolcanic rocks have As
concentrations below 2 ppm and Sb concentrations below 0.5 ppm (Steffen Hagemann and Eduardo
Videla, personal communications 20/4/1998). Unfortunately, our sample material does not extend
more than up to 150 m from the HJB shear zone. This is because fresh rock in the region is covered
by 40-150 m of weathered material and colluvium and is only available where drilled and in mines.
Hence, it is not, presently, possible to estimate the extent of the regional As and Sb anomalies more
accurately.
Page 119
ORE
A Section CD Section AB HJB Shear Zone
1
1 Ag, ppm
Ag, ppm
0.1
0.1
0.01 0.01
0.001 0.001
Alteration
ORE
0 50 100 150 200 250 300 350 0 50 100 150 200 250 300 350
Down-hole depth (m) Down-hole depth (m)
Section AB HJB Shear Zone
B Section CD Au, ppb
10000 10000
Au, ppb
1000 1000
100 100
10 10
1 1
0.1 0.1
ORE
0 50 100 150 200 250 300 350 0 50 100 150 200 250 300 350
C Section CD Section AB HJB Shear Zone
10000 As, ppm
As, ppm
10000
1000
1000
100
100
10
10
1 1
Alteration ORE
0 50 100 150 200 250 300 350 0 50 100 150 200 250 300 350
100
Sb,
D ppm
Section CD
Sb, ppm
10 10
1 1
0.1 0.1
Alteration ORE
0 50 100 150 200 250 300 350 0 50 100 150 200 250 300 350
Unaltered rock Proximal sericite zone
Distal alteration Proximal dolomite-sericite zone
Fig. 12. Two sections (AB and CD of Fig. 2) across ore, HJB shear zone and its wallrocks with
alteration zonation and Au and trace element concentrations and alteration index values. The legend
for the alteration zones for all sections is given in figure A. Note that, for the trace elements, the
concentration scales are logarithmic and that other elements are given in ppm but Au and Te in ppb.
The hatched domains in the Au and trace element diagrams indicate the concentration ranges of the
trace elements in unmineralised mafic rocks (Table 3). The light dashed lines indicate the local
background thresholds as given in Tables 4 and 5. The first sample in the CD section has a different
primary composition from the rest of the data; this is reflected by its high Y/Ti ratio.
Page 120
10000
Section AB HJB Shear Zone E Section CD
Te, ppb 10000
Te, ppb ORE
1000 1000
100 100
10 10
1 1
Alteration
ORE
0 50 100 150 200 250 300 350 0 50 100 150 200 250 300 350
Section AB HJB Shear Zone F Section CD
W, ppm W, ppm
10
10
1 1
0.1 0.1
Alteration ORE
0 50 100 150 200 250 300 350 0 50 100 150 200 250 300 350
2.0
Section AB HJB Shear Zone G Section CD
1.8
CO2/Ca 1.0
1.6 CO2/Ca
1.4 0.8
1.2
1.0 0.6
0.8 0.4
0.6
0.4 0.2
0.2
0.0 0.0
Alteration
ORE
0 50 100 150 200 250 300 350 0 50 100 150 200 250 300 350
0.7
3K/Al
H Section CD
0.6 0.30
3K/Al
0.5 0.25
0.4 0.20
0.3 0.15
0.2 0.10
0.1 0.05
0.0 0.00
Alteration
ORE
0 50 100 150 200 250 300 350 0 50 100 150 200 250 300 350
I
Section CD
40 Section AB HJB Shear Zone
10000*Y/TiO2 28
35 26
ORE 24
30 22
25 20
18
20 16
15 14
12 10000*Y/TiO2
10 10
Alteration
ORE
0 50 100 150 200 250 300 350 0 50 100 150 200 250 300 350
Down-hol e depth (m) Down-hole depth (m)
Fig. 12. Continued.
Page 121
Table 4. Background threshold concentrations applied at Bulletin during this study and the lateral
extent of Au and pathfinder element anomalies based on these thresholds.
Element Background Lateral extent of anomaly away from

threshold used at Ore HJB shear zone
Bulletin
Au 6 ppb 5-20 m 0(-20) m
Ag 0.08 ppm 0(-10) m 0m
As, local 28 ppm 15-50 m 10-40 m
As, regional(?) 6 ppm >150 m >150 m
Sb, local 2 ppm 5-70 m 0-60 m
Sb, regional(?) 0.6 ppm >150 m >150 m
Te 10 ppb 10-115 m (->150 m?) 10-100 m (->150 m?)
W 0.6 ppm 5-25 m 0-20 m
A significant number of samples (23%) have Te concentrations within the “normal” background
range of <10 ppb (Nurmi et al., 1991; Lahtinen and Lestinen 1996), but there appears to be no
obvious separation of these samples from those with 10-23 ppb Te in the cumulative frequency
diagram or indications of extensive trends in the sections studied (Fig. 12E). In any case, the 23 ppb
threshold given by the cumulative frequency diagram is so high that it most probably represents a
local threshold inside a larger Te-enrichment halo enveloping the gold deposit. Therefore, 10 ppb has
been used as the threshold between anomalous and background values.
7.8.2. Primary dispersion defined by Au, Ag, As, Sb, Te and W
Primary dispersion at Bulletin, defined by Au and selected pathfinder elements, is superimposed on

the alteration zones in Figure 12 and summarised in Table 4. All elements considered here define
larger anomalies than ore. The main anomalies are tabular in shape and follow the strike and dip of
the HJB shear zone extending along the shear zone beyond the area studied. Arsenic, Sb, Te and W
define larger anomalies than Au. Only Ag exhibits less dispersion than Au and is, consequently, of
minor importance in exploration (Figs 12A and 12 B).
Arsenic and antimony apparently define the largest anomalies (>150 m from ore) extending beyond
the limits of drilling in fresh rock. Even the more limited dispersion haloes defined by the local As
and Sb threshold values extend up to 60 m from ore and the HJB shear zone, into areas of distal
alteration and apparently unaltered rock. These are widest around the thickest part of ore; smaller
anomalies 5-50 m wide are present in the hangingwall (Figs 12C and 12D). The As and Sb
concentrations define somewhat oscillating but, on average, fairly consistently increasing trends
towards the HJB shear zone. Closer to ore, both As and Sb concentrations show more uniform
increases as mineralisation is approached. In cross section, these trends may provide vectors to ore
over distances of up to 35 m from the HJB shear zone and rough trends towards ore for up to 100 m
or more (Figs 12C and 12D).
The main Te anomaly extends up to 100 m away from the HJB shear zone, while smaller anomalies,
10-50 m wide, cover most of the area beyond the main anomaly (Fig. 12E). Erratic concentrations
appear to dominate within the Te anomalies, except within the HJB shear zone where values within
30 m from ore show a gradual increase towards ore. Only locally do these lateral gradients extend
into the wallrocks of the shear zone.
Anomalous W values extend laterally up to 20 m beyond the HJB shear zone (Fig. 12F). Narrow
anomalies up to 15 m wide are essentially restricted to the boundaries of the hangingwall shear
zones. The concentration gradients shown by W are similar to those of Te.
Page 122
The width of the major Au anomaly at Bulletin, with concentrations from 8 ppb to >50 ppm, varies
from 10 to 70 m. However, dispersion outside the HJB shear zone is very limited. Only in one
section (CD, Fig. 12B), does the anomaly clearly extend beyond the borders of the HJB shear zone,
≥20 m into the footwall. The hangingwall shear zones generally show minor gold enrichment, with
Au concentrations ranging up to 22 ppb. Across strike, Au concentrations locally define gradients
over areas which extend for 5-35 m from the ore.
7.8.3. Major- and related trace-element dispersion patterns
A number of empirical alteration indices have been established to equate alteration intensities to
mineralogy, and to provide directional vectors to areas of likely lode-gold mineralisation (e.g. Dubé
et al., 1987; Kishida and Kerrich, 1987; Davies et al., 1990). In most cases, they reflect carbonate,
potassic or sodic alteration, but an index based on any mobile component or combination of them
may be useful. Their major advantages over pathfinder element dispersion is that they reflect
metasomatism of the major ore-fluid components, and therefore may be less sensitive than trace
components to perturbations in fluid composition caused by interaction with variable lithologies
along fluid flow paths or variations reflecting processes related to the fluid source. The abundances
of the major elements and those trace elements (e.g. Rb, Ba, and Sr) that are present in significant
concentrations within the host rocks prior to alteration are, however, partly a function of closure
effects and the consequence of pre-metasomatic material transfer processes (e.g. crystal
fractionation), which potentially complicates interpretation of the indices.
For Bulletin, geochemical dispersion has been investigated using potassic, sodic and carbonate
alteration indices, enrichment factors and Pearce Element Ratios. The components considered
include CO2, K, Na, S, Rb, Ba and Y, and the absolute residual calculated with the PER evaluations.
Of these, K, CO2, Rb and Y define significant dispersion or depletion haloes (Table 5). Indices based
on the other elements show material transfer only within the HJB shear zone and, locally, within the
hangingwall shear zones, and do not exhibit systematic gradients; accordingly, they are unlikely to
aid in target delineation.
Table 5. The extent of the alteration index anomalies at Bulletin. Enrichment factors are derived
from mass balance calculations. All ratios presented are on molar basis, except 10000xY/TiO2
which is based on wt.%. The "Absolute residual" is obtained using the Pearce Element Ratio method
(Stanley and Madeisky, 1993).
Index Lateral extent of anomaly away from Background

Ore HJB shear zone range
CO2/Ca 10-80 m 3-60 m <0.3
CO2/(Ca+Mg+Fe-0.5(S+As)) 10-80 m 3-60 m <0.15
CO2/Ti 10-80 m 3-60 m <5
3K/Al 1-30 m 0-15 m <0.04
K/Ti 1-30 m 0-15 m <0.5
Na/Al 0-15 m 0-5 m 0.2-0.3
Enrichment factor for Na 0-15 m 0-5 m 0.85-1.15
Absolute residual 0-15 m 0m 0 ±5
Enrichment factor for S 1-20 m 0-5 m <1.2
Enrichment factor for Rb 3-30 m 0-15 m <1.3
Enrichment factor for Ba 0-30 m 0-5 m <1.3
Enrichment factor for Y 2-40 m 0-15 m >0.7
10000xY/TiO2 5-40 m 0-15 m >24
Page 123
7.8.3.1. Carbonation
At Bulletin, carbonation indices, the molar ratios CO2/Ca, CO2/(Ca+Mg+Fe-0.5(S+As)) and CO2/Ti
(Kishida and Kerrich, 1987; Davies et al., 1990; Stanley and Madeisky, 1993; Eilu et al., 1995), the
define the widest anomalies of all the alteration indices (Table 5, Fig. 12G). The main anomaly
extends for 3-60 m beyond the HJB shear zone, with its greatest extent occurring around the thickest
part of the orebody. In addition, narrow anomalies cover most of the studied footwall and
hangingwall. The extent of the carbonation index anomalies is due to the wide extent of carbonate
alteration at Bulletin; essentially, the extent and location of the carbonated areas and the carbonation
index anomalies is the same. Consequently, the use of the carbonation indices does not extend the
exploration target any further than the alteration zone mapping. The potential added advantage of
using the indices is their ability to define gradients towards gold mineralisation. However, such
gradients were not detected.
7.8.3.2. Sericitisation
The sericitisation indices, 3K/Al (Kishida and Kerrich, 1987) and K/Ti, both define essentially the
same anomalies. Both show a major anomaly enveloping the ore, although it extends for no more
than 15 m beyond the HJB shear zone (Table 5, Fig. 12H). The indices also define narrow anomalies,
which are dominantly restricted to within the bounds of the hangingwall shear zones. Increases in the
index value towards ore are common. In cross section, these gradients extend for up to 30 m from the
ore and 15 m from the HJB shear zone.
7.8.3.3. Anomalies defined by Rb, S and Y
As indicated by the mass transfer evaluations, Rb and S are enriched and Y depleted within altered
rocks, commonly rivalling or even exceeding the relative metasomatic changes shown by K, CO2
and the pathfinder elements (Fig. 11). Both Y/TiO2 and the enrichment factor for Y show the same
distribution (Table 5). Yttrium depletion defines prominent anomalies within the shear zones (Fig.
12I), but these anomalies do not extend into unstrained areas, nor do the yttrium indices define any
significant gradients towards ore. Rubidium defines both significant anomalies and vectors to ore,
imitating the anomalies defined by the potassic alteration indices. Sulphur anomalies largely coincide
with those of K and Rb. The most significant differences between S and K and Rb dispersion haloes
are the almost total absence of S enrichment in the footwall and in a 10-20 m wide segment of the
HJB shear zone next to its footwall contact, and the less regular nature of the S enrichment gradient.
7.8.3.4. An extensive lateral geochemical vector
The lateral extents of the geochemical anomalies at Bulletin can be presented as a sequence in order
of increasing dispersion distance: Ag <Ba <Au, Na, Y <W, S, K, Rb <As (local) <CO2, Sb (local)
<Te <Sb (regional), As (regional) - the latter two supposing that the As and Sb anomalies extend
beyond the area studied. This stepwise sequence can be used as a lateral vector to ore at Bulletin and
in analogous areas, and it can, possibly, also define an extensive along-strike vector.
Page 124
7.9. Conclusions
Gold mineralisation at Bulletin is hosted by three distinct suites of Mg- and Fe-tholeiitic metabasalts.
Concentric sequences of alteration zones, with diagnostic colour and mineral assemblage envelop the
mineralisation and, where strained, indicate the sites of shear zones in the area. Wallrock alteration is
pervasive throughout the entire width and extent of the Happy Jack-Bulletin (HJB) shear zone, which
is the structure hosting the ore, and extends up to 150 m beyond the shear zone.
The alteration sequence can be divided into two types, one containing dolomite in the intermediate
and proximal zones and another which is completely dolomite-free. These comprise four or two
major alteration zones, respectively, which can be further divided into as many as 8 subzones. The
former type is dominant, is related to most of the economic ore, and is present in all high-grade gold
ore. The latter is common in areas of low-grade ore and in unmineralised, narrow shear zones in the
hangingwall. The occurrence of the two alteration sequence types is possibly related to variation in
the chemical composition within the tholeiite suites and changes in the hydrothermal system; both
sequences are detected in each of the three basalt suites.
The width of the alteration envelope and the presence of dolomite in the alteration sequence may be
used as a broad guide to the width of the shear zone and size and grade of the gold mineralisation, as
there is a positive correlation between these variables. Irrespective of the extent of the alteration halo,
or the presence or absence of dolomite, progressive changes in alteration mineralogy always form a
consistent sequence towards zones of highest fluid flux and, thus, potential gold mineralisation.
However, this sequence is similar for both the HJB and the hangingwall shear zones. Hence, in any
given section across the studied area, there is a complex series of changes in alteration zones that
relate to the frequency of the hangingwall shear zones, even though there is a broad overall trend
towards the HJB shear zone.
A number of pathfinder elements and geochemical alteration indices, independent of the primary
rock composition, can be used to define anomalies that extend beyond the HJB shear zone and
beyond the area of gold dispersion itself. The anomalies are tabular in shape, follow the strike and dip
of the mineralisation, and tend to be widest around the thickest part of the ore. Only the complete
shape of the possible regional As and Sb anomalies remains unclear, as they extend beyond the area
available for sampling.
The largest potential exploration targets at Bulletin are defined by the possible regional As and Sb
anomalies (Table 6). Anomalies defined by Te and local enrichment of Sb and As also form
significant exploration targets. Other potential pathfinder elements do not show significant dispersion
extending beyond the HJB or minor shear zones, although the main W anomaly is clearly larger than
the Au anomaly. Gold itself has a restricted dispersion, with anomalies extending for 1-35 m from
ore, and ≤5 m from the HJB shear zone.
Page 125
Table 6. Background thresholds or ranges for, and the extent of anomalies defined by, Au, significant
pathfinders and alteration indices at Bulletin.
Background Typical extent of anomaly away from

threshold
or range Ore HJB shear zone
Au 6 ppb 20 m 0m
As, local 28 ppm 25-40 m 20-30 m
As, regional(?) 6 ppm >150 m >150 m
Sb, local 2 ppm 10-70 m 0-50 m
Sb, regional(?) 0.6 ppm >150 m >150 m
Te 10 ppb 50-100 m 50-100 m
W 0.6 ppm 5-15 m 0-5 m
CO2/Ca <0.3 15-60 m 10-50 m
CO2/Ti <5.0 15-60 m 10-50 m
CO2/(Ca+Mg+Fe- <0.15 15-60 m 10-50 m
0.5(S+As))
3K/Al <0.04 20-30 m 5-10 m
K/Ti <0.5 20-30 m 5-10 m
Enrichment factor for Rb <1.3 20-30 m 5-10 m
Of all the alteration indices based on major and other closure-affected elements, only carbonation
indices outline exploration targets significantly larger than the HJB shear zone itself (Table 6). The
carbonation index anomalies do not, however, extend the exploration target beyond alteration
detectable during drill core logging. Indices related to potassic alteration (K, Rb) define anomalies
that extend only up to 10-15 m from the HJB shear zone, and are more restricted than those defined
by As, Sb or Te, but larger than those defined by W or visible sericitisation. Indices calculated by the
Pearce Element Ratio method, or describing the mobility of Na, S or Y, define anomalies that do not
essentially extend beyond the HJB shear zone.
Only As, Sb, K and Rb consistently define gradients (vectors) towards ore and extend beyond the
HJB shear zone. Overall, the best vectors are defined by As, followed by Sb, Rb and K. Vectors
defined by As and Sb are the most consistent. The sequence of dispersion, Ag <Ba <Au, Na, Y <W,
S, K, Rb <As (local) <CO2, Sb (local) <Te <Sb (regional), As (regional), can potentially be used as a
stepwise vector to ore at Bulletin and similar areas, possibly as both a lateral and along-strike vector.
The results of this study indicate that:
1) The pre-alteration geochemical nature of the host rocks must be defined first, before the
mobilities of elements during alteration can be defined reliably, and that reliable
evaluation of chemical changes can only be carried out by applying mass balance
calculation methods which take into account the pre-alteration composition variation.
2) The reliability of the background threshold determinations for pathfinder elements can
only be guaranteed by using a combination of cumulative frequency plots, with careful
examination of geochemical dispersion maps and by comparing pathfinder element
concentrations with data for unmineralised rocks from the literature.
3) By mapping alteration and analysing certain potential pathfinder and other mobile
elements, by techniques that yield low detection limits, exploration targets significantly
larger than the Au anomaly itself can be defined, even if Au is analysed with a ppb-level
detection limit. Some pathfinder elements can potentially be used to define target areas
which are much larger than the mappable alteration halo.
Page 126
4) The lateral extent of the largest geochemical anomalies may be >150 m from a lode-gold
deposit and extend far into areas of apparently unaltered rock, while the elements
mobilised during gold-related alteration tend to define extensive gradients (vectors)
towards ore within the plane of the structure hosting the mineralisation.
Mr Shaun Chanter and Dr Greg Jones, Wiluna Mines Ltd, provided general geological information
for the Bulletin deposit, access to the underground mine, drill core and core cutting services. Their
support and help, as well as critical comments by Dr Neal McNaughton and Dr John Ridley and,
especially, by Prof. David Groves (Centre for Strategic Mineral Deposits, Department of Geology
and Geophysics, University of Western Australia (UWA)), Mr Esko Kontas, Dr Pekka Nurmi and Dr
Peter Sorjonen-Ward (Geological Survey of Finland), and Prof. Heikki Papunen (University of
Turku) are acknowledged. Mr Stephen Gardoll, Ms Sandra Dominguez and Mr Eduardo Videla
(UWA) helped to produce the digitised geochemical maps. Eduardo Videla and Dr Steffen
Hagemann are also thanked for providing insights on the regional geochemical data. Two Journal of
Geochemical Exploration reviewers, Dr. Arthur W. Rose and one anonymous, are thanked for their
critical comments which were useful in improving the paper. This research was funded by a grant
from the Minerals and Energy Research Institute of Western Australia (project M243), the Academy
of Finland and the Geological Survey of Finland. Logistic support was provided by the Centre for
Strategic Mineral Deposits, Department of Geology and Geophysics, UWA, and the Department of
Geology, University of Turku.
7.11 References
Barley, M.E., Cassidy, K.F., Golding, S.D., Groves, D.I. and McNaughton, N.J., 1990. Assessment
of mineralized zones: Alteration haloes. In: S.E. Ho, D.I. Groves & J.M. Bennett (Editors), Gold
deposits of the Archaean Yilgarn Block, Western Australia: nature, genesis and exploration
guides. Geology Department (Key Centre) & University Extension, The University of Western
Australia, Publ. 20: 317-327.
Chanter, S.C., Eilu, P., Erickson, M.E., Jones, G.F.P. and Mikucki, E.J. (1997). Bulletin lode-gold
deposit, Wiluna. In: Berkman, D.A. and Mackenzie, D.H. (Eds) Geology of the Mineral deposits
of Australia and Papua New Guinea. Australasian Institute of Mining and Metallurgy, Monograph
22.
Davies, J.F., Whitehead, R.E.S., Huang, J. and Nawaratne, S., 1990. A comparison of progressive
hydrothermal carbonate alteration in Archean metabasalts and metaperidotites. Mineralium
Deposita 25: 65-72.
Dubé, B., Guha, J. and Rocheleau, M., 1987. Alteration patterns related to gold mineralization and
their relation to CO2/H2O ratios. Mineralogy and Petrology 37: 267-291.
Eilu, P., Groves, D.I., Mikucki, E.J., McNaughton, N.J. and Ridley, J.R., 1995. Alteration indices and
pathfinder elements in wallrock alteration zones around Archaean lode-gold deposits. In: J.
Pasava, B. Kribek and K. Zak (Editors) Mineral deposits: From their origin to their environmental
impacts. Proceedings of the Third Biennial SGA Meeting, Prague, 28-31 August 1995. Balkema,
Amsterdam, pp. 113-116.
Eilu, P., Mikucki, E.J., Dugdale, A.L., Rak, D. and Wright, J., 1996. Primary geochemical dispersion
around the Central Zone of the Bronzewing lode-gold deposit, Yandal greenstone belt, Western
Australia. Exploration ‘96 Congress, 24-26 July 1996, Gold Coast, Australia. Abstracts.
Gresens, R.L., 1967. Composition - volume relationships of metasomatism. Chem. Geol. 2: 47-65.
Groves, D.I., 1993. The crustal continuum model for late-Archaean lode-gold deposits of the Yilgarn
Block, Western Australia. Mineralium Deposita 28: 366-374.
Page 127
Hagemann, S.G., 1992. The Wiluna lode-gold deposits, Western Australia: A case study of a high
crustal level Archaean lode-gold system. Unpublished PhD thesis, Department of Geology, The
University of Western Australia.
Hagemann, S.G., Groves, D.I., Ridley, J.R. and Vearncombe, J.R., 1992. The Archean lode-gold
deposits at Wiluna, Western Australia: high-level brittle-style mineralization in a strike-slip
regime. Econ. Geol. 87: 1022-1053.
Hagemann, S.G., Brown, P.E., Groves, D.I., Ridley, J.R. and Valley, J., 1994. The Wiluna lode-gold
deposits, Western Australia: surface water influx in a shallow level Archean lode-gold system.
Geological Society of Australia Abstracts 37: 160-161.
Huston, D.L., 1993. The effect of alteration and metamorphism on wall rocks to the Balcooma and
Dry River South volcanic-hosted massive sulfide deposits, Queensland, Australia. J. Geochem.
Explor. 48: 277-307.
Hutchinson, C.S., 1974. Laboratory Handbook of Petrographic Techniques. Wiley, New York. 527
pp.
Jensen, L.S., 1976. A new cation plot for classifying subalkalic volcanic rocks. Ontario Div. Mines,
Misc. Paper no. 66. 22 pp.
Kishida, A. and Kerrich, R., 1987. Hydrothermal alteration zoning and gold concentration at the
Kerr-Addison Archean lode gold deposit, Kirkland Lake, Ontario. Econ. Geol. 82: 649-690.
Lahtinen, R. and Lestinen, P., 1997. Background variation of ore-related elements and regional-scale
mineralization indications in Palaeoproterozoic bedrock in the Tampere-Hämeenlinna area,
southern Finland. Geol. Surv. Finland Bull. 390. 38 p.
Lesher, C.M., Gibson, H.L. and Campbell, I.H., 1986. Composition-volume changes during
alteration of andesite at Buttercup Hill, Noranda District, Quebec. Geochim. Cosmochim. Acta
50: 2693-2705.
MacLean, W.H. and Barrett, T.J., 1993. Lithogeochemical techniques using immobile elements. J.
Geochem. Explor. 48: 109-133.
MacLean, W.H. and Kranidiotis, P., 1987. Immobile elements as monitors of mass transfer in
hydrothermal alteration: Phelps Dodge massive sulphide deposit, Matagami, Quebec. Econ. Geol.
82: 951-962.
McCuaig, T.C. and Kerrich, R., 1994. P-T-t-deformation-fluid characteristics of gold deposits:
evidence from alteration systems. In: Lentz, D.R. (ed.) Alteration and Alteration Processes
Associated with Ore-forming Systems. Geol. Assoc. Canada, Short Course Notes, 11: 339-379.
Meyer, C. and Hemley, J.J., 1967. Wall rock alteration. In: H.L. Barnes (Editor) Geochemistry of
Hydrothermal Ore Deposits. Holt, Rinehart and Winston. New York, pp. 166-235.
Niskavaara, H. and Kontas, E., 1990. Reductive coprecipitation as a separation method for the
determination of gold, palladium, platinum, rhodium, silver, selenium and tellurium in geological
samples by graphite furnace atomic absorption spectrometry. Analytica Chemica Acta 231: 273-
282.
Nurmi, P.A., Lestinen, P. and Niskavaara, H., 1991. Geochemical characteristics of mesothermal
gold deposits in the Fennoscandian Shield, and a comparison with selected Canadian and
Australian deposits. Geological Survey of Finland, Bulletin 351. 101 pp.
Perrault, G., Trudel, P. and Bedard, P., 1984. Auriferous halos associated with the gold deposits at
Lamaque mine, Quebec. Econ. Geol. 79: 227-238.
Ringwood, A.E., 1979. Origin of the Earth and Moon. Springer-Verlag, New York. 295 pp.
Rose, A.W. and Burt, D.M., 1979. Hydrothermal alteration. In: H.L. Barnes (Editor) Geochemistry of
Hydrothermal Ore Deposits. 2nd ed., pp. 173-235. Wiley. New York.
Rose, A.W., Hawkes, H.E. and Webb, J.S., 1979. Geochemistry in Mineral Exploration. 2nd edition.
Academic Press, London.
Sinclair, A.J., 1976. Probability graphs in mineral exploration. Assoc. Exploration Geochemists Spec.
Vol. 4. 95 pp.
Sinclair, A.J., 1991. A fundamental approach to threshold estimation in exploration geochemistry:
probability plots revisited. J. Geochem. Explor. 41, 1-22.
Smith, T.J. and Kesler, S.E., 1985. Relation of fluid inclusion geochemistry to wallrock alteration and
lithogeochemical zonation at the Hollinger-McIntyre gold deposit, Timmins, Ontario, Canada.
CIM Bull. 78: 35-46.
Page 128
Stanley, C.R. and Madeisky, H.E., 1993. Pearce element ratio analysis: applications in
lithogeochemical exploration. MDRU Short Course Notes: SC-13. The University of British
Columbia, Vancouver.
Taylor, S.R., 1968. Geochemistry of andesites. In: L.H. Ahrens (Editor) Origin and Distribution of
the Elements. Pergamon Press, Oxford, pp. 559-583.
Tilling, R.I., Gottfried, D. and Rowe, J.J., 1973. Gold abundance in igneous rocks: bearing on gold
mineralization. Econ. Geol. 68: 168-186.
Wedepohl, K.H., 1969. Handbook of Geochemistry, Volume 1. Springer-Verlag, Berlin.
Page 129
8. Late-Kinematic Timing of Orogenic Gold

Deposits and Significance for Computer-Based
Exploration Techniques with Emphasis on the
Yilgarn Block, Western Australia.
David I. Groves a, Richard J. Goldfarb a,b, Carl M. Knox-Robinson a, Juhani Ojala c, Stephen
Gardoll a, Grace Y. Yun a and Peter Holyland c.
a
Centre for Strategic Mineral Deposits, Department of Geology and Geophysics, University of
Western Australia, Nedlands, W.A., Australia 6907.
b
United States Geological Survey, Denver Federal Center, Denver, Colorado, 80225 USA
c
Terra Sancta, Unit 6, 4 Brodie Hall Drive, Bentley, W.A., Australia 6104.
Abstract
Orogenic gold deposits are a widespread coherent group of epigenetic ore deposits that are sited in
accretionary or collisional orogens. They formed over a large crustal-depth range from deep-seated
low-salinity H2O-CO2±CH4±N2 ore fluids and with Au transported as thio-complexes. Regional
structures provide the main control on deposit distribution. In many terranes, first-order faults or
shear zones appear to have controlled regional fluid flow, with greatest ore-fluid fluxes in, and
adjacent to, lower-order faults, shear zones and/or large folds. Highly competent and/or chemically
reactive rocks are the most common hosts to the larger deposits. Focusing of supralithostatic ore
fluids into dilatant zones appears to occur late during the evolutionary history of the host terranes,
normally within D3 or D4 in a D1-D4 deformation sequence. Reactivation of suitably-oriented pre-
existing structures during a change in far-field stress orientation is a factor common to many
deposits, and repeated reactivation may account for multiple mineralization episodes in some larger
deposits. Absolute robust ages of mineralization support their late-kinematic timing, and, in general,
suggest that deposits formed diachronously towards the end of the 100 to 200 m.y. long evolutionary
history of hosting orogens. For example, in the Yilgarn Block, a region specifically emphasised in this
study, orogenic gold deposits formed in the time interval between 40 and 90 m.y., with most about 60
to 70 m.y., after the youngest widespread basic-ultrabasic volcanism and towards the end of felsic
magmatism.
The late timing of orogenic gold deposits is pivotal to geologically-based exploration methodologies.
This is because the present structural geometries of: i) the deposits, ii) the hosting goldfields and iii)
the enclosing terranes are all essentially similar to those during gold mineralization, at least in their
relative position to each other. Thus, interpretation of geological maps and cross sections and three-
dimensional models can be used to accurately simulate the physical conditions that existed at the time
of ore deposition. It is particularly significant that the deposits are commonly related to repetitive
and predictable geometries, such as structural heterogeneities within or adjacent to first-order
structures, around rigid granitoid bodies, or in specific “locked-up” fold-thrust structures.
Importantly, the two giant greenstone-hosted goldfields, Kalgoorlie and Timmins, show a remarkably
similar geometry at the regional scale.
Computer-based stress mapping and GIS-based prospectivity mapping are two computer-based
quantitative methodologies that can utilize and take advantage of the late timing aspect of this deposit
type to provide important geological aids in exploration, both in broad regions and more localized
goldfields. Both require an accurate and consistent solid geology map, stress mapping requires
Page 131
knowledge of the far-field stresses during mineralization, and the empirical prospectivity mapping
requires data from a significant number of known deposits in the terrane.
The Kalgoorlie Terrane, in the Yilgarn Block, meets these criteria, and illustrates the potential of
these methodologies in the exploration for orogenic gold deposits. Low minimum stress anomalies,
interpreted to represent dilational zones during gold-related deformation, coincide well with the
positions of known goldfields rather than individual gold deposits in the terrane, and there are
additional as-yet unexplained anomalies. The prospectivity analysis confirms that predictable and
repetitive factors controlling the siting of deposits are: i) proximity to, and orientation and curvature
of, granitoid-greenstone contacts, ii) proximity to segments of crustal faults which strike in a
preferred direction, iii) proximity to specific lithological contacts which have similar preferred strike,
iv) proximity to anticlinal structures and v) the presence of preferred reactive host rocks (e.g.,
dolerite). The prospectivity map defines a series of anomalous areas, which broadly conform to those
of the stress map (>78% correspondence). The most prospective category on this map covers less
than 0.3% of the greenstone belts and yet hosts 16% of the known deposits, which have produced
>80% of known gold. Thus, it discriminates in favour of the larger economically more-attractive
deposits in the terrane.
The successful application of stress mapping and prospectivity mapping to geology-based

exploration for orogenic gold deposits indicates that more quantitative analysis of geological map
data is a profitable line of research. The computer-based nature of these methodologies is ideal for
the production of an ultimate, integrated, deposit target map, which can be compared to other, more
conventional, targeting parameters such as geophysical and geochemical anomalies. Such an
integrated strategy appears the way forward in the increasingly difficult task of cost-effective global
exploration for orogenic gold deposits in poorly exposed terranes.
8.1. Introduction
In a recent review article, Groves et al. (1998) suggest that lode-gold deposits worldwide, which have
been variously termed mesothermal, greenstone-hosted, slate-belt hosted, turbidite-hosted, Mother
Lode-type or gold-only deposits, are a coherent group of deposits with a common origin. They favour
the term “orogenic gold deposit” for this deposit class, because all of the deposits ascribed to this
class were formed during compressional to transpressional deformation processes at convergent plate
margins in accretionary or collisional orogens. The deposits occupy a unique depth range for
hydrothermal ore deposits, with gold deposition from 15-20 km depth to within 2-3 km of the surface,
where stibnite and cinnabar may become dominant within the same hydrothermal cells (Fig. 1).
Page 132
Figure 1. Schematic diagram showing the setting and nature of orogenic lode-gold deposits. A.
Plate tectonic environments of formation of orogenic lode-gold deposits and other largely syn-
volcanic or syn-intrusive gold-rich deposit styles.
Orogenic gold deposits may occur in the same orogens as gold-rich porphyry-style and epithermal
vein-deposits. However, the porphyry and epithermal deposits occur over a very narrow depth range
above, and landward of, a continental margin arc, whereas the orogenic gold lodes extend over a
continuum of depths in the growing forearc region (Fig.1). Auriferous volcanic-hosted massive
sulfide (VHMS) deposits may have a gross spatial association with orogenic gold deposits, but
typically formed tens of millions of years earlier, and in the oceanic rocks prior to their collision with
the growing continental margin.
The chemistry of vein-forming fluids is remarkably similar between orogenic gold ores of all ages.
Detailed fluid inclusion studies have shown ore fluids consistently to contain from about 5 mole
percent CO2 to equal amounts of H2O and CO2 (e.g., Roedder, 1984; Smith et al., 1984; Goldfarb et
al., 1986; Robert and Kelly, 1987; Goldfarb et al., 1989; Ho. et al., 1990; Diamond, 1990; de Ronde
et al., 1992). The relatively high CO2 content of these ore fluids, responsible for the common
widespread carbonate-rich alteration zones, contrasts with those that deposited other gold deposit
types. In addition, ore fluids from orogenic gold deposits, especially where hosted in terranes
dominated by metasedimentary rock units, may contain nitrogen and methane at the mole percent
level. The greater the non-aqueous volatile content of the original ore fluid, the more likely the
occurrence of at least episodic immiscibility between aqueous and non-aqueous fluids during
decreases in temperature into the two-phase fluid field. The transport of gold in these H2O-
CO2±CH4-N2 fluids is facilitied by reduced sulfur complexes, as is supported by measurements of
minor amounts of H2S in some of the studied fluid inclusions (Goldfarb et al., 1989, Yardley et al.,
Page 133
1993). Pressure-temperature conditions of veining, as determined from the above listed and many
other fluid inclusion investigations of orogenic gold systems, are concentrated between 1 and 3 kb
and 250o and 350oC. However, conditions for some near-surface gold deposits may be as low as 0.5
kb and 150oC, and as high as 4-5 kb and 700oC for some of the deeper deposits formed within
evolving orogens (Groves, 1993; Groves et al., 1998).
The relatively non-aqueous volatile- and H2S-rich fluids are also characterized by low total-solvent
volumes. The above-listed fluid inclusion studies from both Phanerozoic and Precambrian
hydrothermal systems consistently report salinities of less than about 6 wt. percent equiv. NaCl. The
low salinities are responsible for the well-recognized lack of base metal enrichments within orogenic
lode-gold deposits (Kerrich and Fryer, 1981). A few gold provinces containing veins with many
features suggestive of an orogenic gold classification, such as Sabie-Pilgrim’s Rest along the eastern
margin of the Transvaal Basin in South Africa (Boer et al., 1993; Tyler and Tyler, 1996) and Tennant
Creek (Khin et al, 1994) and Telfer (Rowins et al., 1997), contain high-salinity fluids that may have
deposited gold ore along with abundant copper. The spatial association between the gold ores and
basinal sedimentary sequences appears important in these circumstances. Yardley (1997) points out
that high-salinity pore waters, such as developed in basinal settings, will persist in the crust until
anatexis and, therefore, may be incorporated into many deep-crustal fluid-flow cells.
A consistently heavy and narrow range characterizes oxygen isotope measurements of ore-hosting
quartz of Precambrian (Kerrich, 1987; Golding and Wilson, 1987), Proterozoic (Oberthur et al.,
1996), Paleozoic (Kontak and Kerrich, 1997), and Mesozoic-Cenozoic age (Bohlke and Kistler,
1986; Nesbitt et al., 1989, Goldfarb et al., 1997). Most measured values for δ18O quartz of
Precambrian age range between about 11 and 14 per mil, and younger veins are typically 2-3 per mil
heavier, with the difference reflecting variations in host rock compositions and (or) vein formation
temperatures. For both age groups, ore fluids are calculated to have δ18O compositions between 6
and 11 per mil. Values of fluid δD from these studies, when calculated from measurements made on
H-bearing hydrothermal silicates, are between about -70 and -30 per mil. Collectively, these data
implicate a deep-crustal fluid source for orogenic lode-gold, although whether this is a metamorphic
(Kerrich and Fryer, 1979; Phillips and Groves, 1983; Goldfarb et al., 1988) magmatic (Wood et al.,
1986; Spooner, 1993), or mantle fluid (Perring et al., 1987; Cameron, 1988) is still poorly understood
after more than a decade of debate on the scenarios. Carbon and sulfur isotope data are highly
variable between orogenic gold districts of all ages, indicating local crustal contributions of at least
some of the ore fluid carbon and sulfur, as well as variations in the oxidation state of ore fluids at the
final depositional site in the case of sulfur (Kerrich, 1989).
As a result of deposition from low-salinity, H2O-CO2±CH4±N2 fluids, orogenic lode-gold deposits

characteristically have high Au:Ag ratios and low base-metal and tin contents (see more detailed
discussion below). Lode-gold deposits which have abundant silver (Au/Ag ratios less than 1),
significant base-metal contents or tin-rich mineral phases should not be classified as orogenic deposits
unless there are convincing reasons to do so based on detailed research. Some of the so-called
intrusion-related gold deposits (e.g., Mantos de Punitaqui, Chile; Snip, Canada; Pacoy-Pataz, Peru;
Palpa-Ocona, Peru; Kidston, Australia; Kari Kolla, Bolivia; of Sillitoe and Thompson, 1998 and
Thompson et al., 1999) have metal characteristics that clearly distinguish them from orogenic gold
deposits as defined by Groves et al. (1998), whereas others have metal ratios (and other features)
appropriate to their classification as orogenic deposits (e.g., Charters Towers, Australia; Ryan Lode,
USA; Mokrsko, Czech Republic; Vasil’kovskoye, Kazakhstan). Most of the lode-gold deposits with
relatively high base-metal, silver and/or tin contents were emplaced at shallow depths only. They are
gold systems that are better linked to the continental arc setting of Figure 1, as many of the referenced
ore deposits are spatially associated with continental-margin Andean batholiths.
The purpose of the first part of this paper is to briefly review the consistent features of the orogenic
gold deposits (see Groves et al., 1998 for more exhaustive review) and to document that the timing of
gold deposition is typically late within the tectonic evolution of its host terranes. Demonstration of
Page 134
late timing is vital to the utilization of geology in exploration because, if demonstrated, it means that
the current geometries of the deposits and their enclosing host rocks are essentially the same in map
and cross-sectional view as they were at the time of gold mineralization. As the primary controls on
the siting of the gold deposits are the geometry (including orientation) of the controlling structures
and the nature and geometry of host sequences, methodologies that utilize such geometries are thus a
potentially powerful tool in mineral exploration for orogenic lode-gold deposits. The second part of
the paper reviews the role of stress mapping and GIS-based prospectivity analysis, two such
methodologies, in exploration for these deposits. To provide a consistent background, examples of
the application of the two methodologies are given for the Yilgarn Block of Western Australia, but
they equally can be applied to other terranes hosting this deposit type.
8.2. Geological characteristics of orogenic lode-gold

deposits
8.2.1 Geology of host terranes
Perhaps the single most consistent characteristic of the deposits is their consistent association with
deformed metamorphic terranes of all ages. Observations from preserved Archaean greenstone belts
and most-recently active, Phanerozoic sedimentary rock-dominant metamorphic belts throughout the
world indicate a strong association of gold and greenschist facies rocks. However, some significant
deposits occur in higher metamorphic-grade Archaean terranes (e.g., McCuaig et al., 1993) or in sub-
greenschist grade domains within the metamorphic belts of a variety of geological ages. In the
Archaean rocks of Western Australia, some synmetamorphic deposits extend into upper amphibolite
to granulite facies domains (Groves 1993). In a few areas, where parts of older cratons are reworked
during younger orogens, gold ores may be sited also in high-grade metamorphic rocks (e.g., northern,
eastern and southern margins of the North China Craton). However, in these cases, gold ores post-
date the metamorphism of immediate host-rocks by more than 1 b.y. Pre-metamorphic protoliths for
the auriferous Archaean greenstone belts are predominantly volcano-plutonic terranes of oceanic
back-arc basalt and felsic to mafic arc rocks. Terranes dominated by clastic marine sedimentary rocks,
generally metamorphosed under greenschist facies conditions, host mostly younger ores. They also
are, however, important in some Archaean terranes (e.g., Slave Province, Canada) and Proterozoic
(e.g., Ashanti belt, West Africa) terranes. Banded iron formation or ferruginous chert is an additional
favourable ore host in many Precambrian gold provinces (e.g., Morro Velho, Quadrilatero Ferrifero,
Brazil; Homestake, Northern Black Hills, USA).
8.2.2 Deposit mineralogy and alteration
The orogenic gold deposits in greenschist facies terranes are typified by quartz-dominant vein
systems with ≤3-5 percent sulfide minerals (mainly Fe- and As-bearing sulfides) and ≤5-15 percent
carbonate minerals. Albite, white mica or fuchsite, chlorite, scheelite and tourmaline are also common
gangue phases in veins in greenschist-facies host rocks. Vein systems may be continuous over a
vertical extent of 1-2 km with little change in mineralogy or gold grade; subtle mineral zoning does
occur, however, in some deposits (e.g., Mt Charlotte, Golden Crown, Hill 50 in Western Australia
and Alaska-Juneau, SE Alaska, USA). Gold:silver ratios range from 10 (normal) to 1 (less common),
with ore in places being in the veins and elsewhere in sulfidized wallrocks. Gold grades are relatively
high, historically having been in the 5-30 g/t range; modern-day bulk-mining methods have led to
exploitation of lower grade targets. Sulfide mineralogy commonly reflects the lithogeochemistry of
the host. Arsenopyrite is the most common sulfide mineral in metasedimentary country rocks,
whereas pyrite or pyrrhotite is more typical in metamorphosed igneous rocks. Gold-bearing veins
exhibit variable enrichment in As, B, Bi, Hg, Sb, Te and W; Cu, Pb and Zn concentrations are
Page 135
generally only slightly elevated above regional backgrounds, although Pb may be anomalous in some
deposits.
Deposits exhibit strong lateral zonation of alteration phases from proximal to distal assemblages on
the scale of metres. Mineralogical assemblages within the alteration zones, and the width of these
zones, generally vary with wallrock type and crustal level. Most commonly, carbonates include
ankerite, dolomite or calcite; sulfides include pyrite, pyrrhotite or arsenopyrite; alkali metasomatism
involves sericitization or, less commonly, formation of fuchsite, biotite or K-feldspar, and
albitization; and mafic minerals are highly chloritized. Amphibole or diopside occurs at progressively
deeper crustal levels, and carbonate minerals are less abundant. Sulfidation is extreme in banded iron-
formation (BIF) and iron-rich mafic host rocks. Wallrock alteration in greenschist facies rocks
involves the addition of significant amounts of CO2, S, K, H2O, SiO2 ± Na and LILE.
8.2.3 Structure
The strong structural control of mineralization is very apparent at a variety of scales. Many would
argue that structure is the single most-important control on gold mineralization. Deposits are normally
sited in second- or third- order structures, typically near large-scale, commonly transcrustal, first-
order compressional or transpressional structures. Although the controlling structures are commonly
ductile to brittle, they are highly variable in type, including: (a) brittle faults to ductile shear zones
with both low-angle to high-angle reverse motion and strike-slip or oblique-slip motion; (b) fracture
arrays, stockwork networks or breccia zones in competent rocks; (c) foliated zones; and (d) fold
hinges and overturned limbs in ductile turbidite sequences. Mineralized structures have small syn-
and post-mineralization displacements (e.g., Ojala et al., 1993), but the gold deposits commonly have
extensive down-plunge continuity (hundreds of metres to as much as 1-2.5 km). Extreme pressure
fluctuations, leading to cyclic fault-valve behavior (Sibson et al., 1988), result in flat-lying
extensional veins and mutually cross-cutting steep fault veins that characterize many deposits (e.g.,
Robert and Brown, 1986; Miller et al., 1994).
8.2.4 Timing of gold mineralization
The late timing of formation of orogenic gold deposits is taken, in a relative sense, to suggest that
most of the gold veining post-dates regional metamorphism, plutonism and early phases of orogenic
deformation in the immediate host rocks. Simultaneously, however, veining is coeval with
metamorphism and magmatism at deeper levels within the orogen (e.g., Groves et al., 1995). Ore
fluids may have formed at these deeper levels, or even below the deepest supracrustal rocks, and
migrated to retrograding and uplifting parts of the orogen during continuing episodes of deformation
(e.g., Sibson, 1985; Stuwe et al., 1993; Phillips and Powell, 1993; Kerrich and Cassidy, 1994). The
deposits are, therefore, commonly late in the overall deformational cycle, although, in some
provinces, they may be overprinted by deformational fabrics (e.g., Morasse et al., 1995).
In an absolute time sense, the late timing can represent a range of scenarios from veining just a few
million years after regional metamorphism and deformation of host rocks (e.g., Chugach accretionary
prism, Gulf of Alaska) to many tens of millions of years after initial collisional and orogenic events
affecting the host sequences (e.g., Juneau gold belt, SE Alaska). Many of the worlds great
Phanerozoic orogenic belts evolved over periods of 100 to 200 m.y., with orogenic gold lodes having
been emplaced throughout diachronous post-collisional deformation and uplift episodes in the
accreted crust. Timing relationships may be more complex in older collisional belts. Robust
geochronological data commonly place formation of Precambrian gold deposits some 20 to 70 m.y.
after youngest nearby volcanism (e.g., Kerrich and Cassidy, 1994; Yeats and McNaughton, 1997),
although deposits sited in some older greenstone belts may be formed in the deformation event
related to accretion of younger greenstone belts in the same terrane some hundreds of millions of
years after formation of the host rocks (e.g., Yilgarn Block, Australia; Groves et al., 1995). In
Page 136
exceptional cases (e.g., eastern, northern and southern margins of North China craton; Variscan of
southern Europe), Phanerozoic orogenic gold deposits may be emplaced, within reworked
Precambrian basement, hundreds of millions to billions of years subsequent to initial deformation of
the host rock.
This late timing is discussed in more detail below as it is pivotal to the discussion of exploration
methodologies based on geological models presented in the second half of the paper, and is critical to
their implementation.
8.2.5 Yilgarn gold deposits as orogenic gold deposits
The gold deposits of the Yilgarn Block of Western Australia are the best-documented examples
where stress mapping and GIS-based prospectivity mapping have been applied to the exploration for
orogenic gold deposits. Because of this, they are chosen for review in this paper. Based on the
parameters outlined above, the majority of gold deposits in the Yilgarn Block, including the
somewhat controversial porphyry-hosted (e.g., Kanowna Belle) and syenite-hosted (e.g., Jupiter)
deposits, fit into the class of orogenic gold deposits. There is no evidence for different ore fluids,
significantly different alteration assemblages (where allowance is made for variable host-rock
composition and P-T conditions), or different metal ratios to indicate a different fluid source type or
genetic model for any of the gold deposits in the eastern or central part of the craton.
The only large deposit that is clearly anomalous is Boddington in the western part of the Yilgarn
Block, where the Au-Cu-Mo association, the high-temperature biotite-amphibole alteration
assemblages in a relatively low-grade greenschist-facies metamorphic environment, and the high-
temperature, high-salinity fluid inclusions are unlike those in any other well-studied Yilgarn gold
deposit (Roth, 1992). The timing of mineralization is controversial (e.g., Allibone et al., 1998), but
the style of mineralization appears best to fit a diorite porphyry-style model.
8.2 Timing of relationships in orogenic gold deposits
8.3.1 Relative timing from structural evidence
As mentioned above, most structural studies of orogenic lode-gold deposits indicate ore formation
during D3 or D4 events in terranes for which D1-D3 or D1-D4 deformation events are recognized (e.g.,
Colvine, 1989; Vearncombe et al., 1989; Cox et al., 1991).
The Eocene Juneau gold belt in southeastern Alaska, USA provides an excellent example of such a
relative timing scenario. Gehrels et al. (1992) noted four sets of regional fabrics, developed over
about 50 m.y., within metasedimentary units of some of the gold-host terranes. These included in
order: (1) an initial foliation; (2) fabric related to ductile thrusting along the major, terrane-bounding
first- order shear zones; (3) polygenetic, mainly dip-slip structures; and (4) later shear fabric related to
sill emplacement during regional uplift. Gold veins surround first-order fault zones, but were
emplaced in third-order subsidary shear veins and tension fractures during the final of the four
deformational events (Goldfarb et al., 1991; Miller et al., 1994). The relatively late timing of gold
veins is also evident from more detailed mine-scale structural studies at the deposits. For example,
Miller et al. (1992) show that veins at the Alaska-Juneau deposit cut schistosity (S1), isoclinal folding
(S2) and crenulation cleavage (S3), although boudinaged veins and offset ore zones indicate minor
post-gold deformation.
Such a relatively late timing for orogenic gold deposits characterizes many orogens. Also within the
Cordilleran orogen and along the accretionary prism of southern Alaska, gold lodes are reported to
Page 137
have been emplaced during a fourth stage of five deformational events that have affected the turbidite
host rocks of the Chugach terrane (Nokleberg et al., 1989). Farther south within the orogen along
western North America, metasedimentary rocks of the Foothills terrane in central California were
progressively deformed over a period of about 50 m.y. (Tobisch et al., 1989). Dating of gold veins of
the Mother Lode belt (Bohlke and Kistler, 1986) indicates that they were emplaced during the last
part of the deformation period. In the Tasman orogen, the Victorian goldfields of eastern Australia
occur in classic saddle reef structures that were formed during a D2 event, whereas the gold was
introduced into these structures during D4 (Forde, 1991; Forde and Bell, 1994). Similarly, gold in D2
reef structures in the Meguma terrane, of the Acadian orogen, Nova Scotia, was mainly deposited
during D3 deformation (Ryan and Smith, 1998). In older sequences, such as the Paleoproterozoic
Birimian orogenic belt of West Africa, gold ores were similarly emplaced during D4 events within an
extensive history of deformational episodes (Milesi et al., 1992).
It is critical to realize that various workers use D1, D2,... nomenclature to describe a variety of
features that are not always equivalent from one area to another. Although many studies note that
gold is D3 or D4 in timing, and thus is emplaced late during deformation in an orogen, some studies
indicate a D1 or D2 event as being temporally associated with formation of ores in some relatively-
small orogenic-gold provinces. For example, Murphy and Roberts (1997) indicate early D1 Variscan
gold-forming events in the Iberian Massif of Spain, Stowell et al. (1996) favour a D2 timing in the
Blue Ridge of the Appalachian orogen, southeastern U.S., and LeAnderson et al. (1995) describe
some D2 veining in the Pan-African goldfields of Saudi Arabia. It is important to realize that the
features described by these authors are relatively broad-scale features, such as a D1 regional
metamorphic event. This may still be a relatively late feature in the deformation history of the rocks
(confirmed for Blue Ridge by T.H. Bell, pers. comm., 1999), but either is much more well-developed
or completely overprints early pre-accretionary to collisional structures that certainly must have been
present prior to metamorphism. Such complexities of relative timing between gold provinces
highlight the need for absolute dating of ores and deformation events to confidently establish the
temporal setting of gold ores in the evolving orogen.
Generally in brittle-ductile regimes, orogenic-gold vein formation is commonly related to reactivation

of earlier-formed faults or shear zones in a contrasting far-field stress regime (e.g., Goldfarb et al.,
1991; de Ronde and de Wit, 1994) or to reactivation of tightly-folded, “locked-up” folds (e.g., Cox et
al., 1995). In some cases, there is evidence for overprinting foliation on pre-existing gold deposits in
brittle-ductile regimes (e.g., Morasse et al., 1995), but, in such cases, there is little evidence to suggest
that there has been drastic changes to the structural geometry of the deposits after formation. In more
ductile regimes, gold-bearing veins, particularly where hosted by shear zones, are commonly
deformed, but they still formed late in the overall deformational history of the host terrane (e.g.,
Knight et al., 1993; Bloem et al., 1994). Thus, most field evidence suggests that orogenic lode-gold
deposits formed under conditions of low effective stress and limited displacement during reactivation
of earlier-formed structures. The nature and orientation of quartz veins and their selective
development as vein arrays in more competent units in the host successions suggest the existence of
supalithostatic pressures during mineralization (e.g., Robert and Brown, 1986; Sibson et al., 1988;
Miller et al., 1994; Groves et al., 1995).
8.3.2 Geochronological evidence
In many terranes, the relative structural evidence has been confirmed by absolute dating of gold
mineralization based on robust geochronology (e.g., U-Pb studies) and/or consistent data from other
techniques (e.g., Ar-Ar plateaus, Pb-Pb or Sm-Nd techniques): see, for example, review by Kerrich
and Cassidy (1994). It is not the object of this paper to thoroughly review the extensive and, in
places, controversial literature on this topic. However, as the examples of repetitive structural
geometry, stress mapping and prospectivity mapping presented below are mainly from the Yilgarn
Block of Western Australia, the geochronological evidence from there is summarized in some detail.
Subsequently, less detailed data, which confirm the same recurring temporal patterns between gold-
Page 138
forming hydrothermal events and other processes in developing orogens are briefly discussed from
other terranes.
8.3.3 Geochronology : Yilgarn Block
The ages of gold mineralization of Yilgarn orogenic lode-gold deposits (Fig. 2) have been constrained
by three different methods. A maximum age has been determined by dating felsic intrusive rocks that
are cut by gold-bearing veins and/or are altered, a minimum age is defined by the age of crosscutting
felsic dykes, and the age of mineralization has been dated directly by using appropriate, gold-related
minerals (e.g., hydrothermal zircon, titanite, rutile, mica and garnet) in relevant alteration zones. A
summary of these ages is given in Figure 3.
Most of the larger gold deposits are hosted by ca 2.7 Ga greenstone belts in the Norseman-Wiluna
Belt, although some deposits are also hosted by greenstone belts as old as 3.0 Ga in the Murchison
and Southern Cross Provinces (Fig. 2). The best constraints on the maximum age of gold
mineralization are provided by pre-mineralization granitoid plutons and felsic porphyry dykes and
sills, including diorite at Granny Smith, with a U-Pb in zircon SHRIMP age of 2665±5 Ma (Campbell
et al., 1993) and tonalite and leucogranite at Lawlers with a SHRIMP age of 2666± 3 and 2666± 7
Ma, respectively (Fletcher et al., in Yeats and McNaughton, 1997). Various porphyry bodies which
are cut by gold-bearing veins at the Mount Charlotte, Mount Percy, Racetrack and Porphyry deposits
have U-Pb in zircon SHRIMP ages between 2670 and 2660 Ma (Yeats et al., in press). Thus, gold
mineralization in these deposits must be younger than about 2660 Ma.
Page 139
Figure 2. Simplified geological map of the Yilgarn Block, Western Australia, showing the location of
gold deposits discussed in the text. Adapted from Groves (1993) and Myers (1993).
Page 140
Figure 3. Diagram illustrating geochronological constraints on orogenic lode-gold deposits of the

Yilgarn Block, Western Australia. Ages are shown as ±2σ. Shaded area is 2640-2625 Ma,
interpreted to be the period of a widespread gold mineralizing event in the Yilgarn Block.
Greenstone belt ages are not shown, but the youngest mineralized mafic and ultramafic rocks are ca
2700 Ma in age. Adapted from Yeats and McNaughton (1997).
Minimum ages for gold mineralization are provided by dykes and sills which post-date gold deposits,
or are not gold mineralized, but show minor alteration of similar type to that of the adjacent gold
deposit, or are slightly offset by late movement on gold-bearing structures. Such intrusions are
extremely rare in the Norseman-Wiluna Belt. However, in the more western Yilgarn provinces and
the southern most tip of the Norseman-Wiluna Belt, U-Pb in zircon SHRIMP dates late- to post-
mineralization pegmatite dykes at 2623±7 Ma at Mt Gibson (Yeats et al., 1996), 2637±7 Ma at
Westonia (Kent et al., 1996) and 2637±3 Ma at Griffins Find (Qui et al., 1997). Similarly, Sm-Nd
isochrons date late- to post-mineralization pegmatite intrusions at Westonia at 2640±11 Ma, at
Nevoria at 2628±10 Ma and Scotia at 2620±36 Ma (Kent et al., 1996). At Corinthia, Bloem et al.
(1995) record a Pb-Pb isochron age of 2620±6 Ma for a post-mineralization pegmatite dyke. Thus,
the minimum age for mineralization of some Yilgarn deposits lies between 2620 and 2640 Ma.
Exceptions to this are provided by U-Pb in zircon SHRIMP ages of porphyry dykes which are
interpreted to cut gold mineralization at Jundee and Mt McClure in the northeastern Eastern
Goldfields Province to the east of the Keith-Kilkenny Fault, in an area interpreted to be a distinct
terrane within the province (e.g., Myers, 1993). These dykes have ages of 2656±7 Ma and 2663±4
Ma, respectively (Yeats et al., in press), which are virtually indistinguishable from the ages of
porphyry bodies that are cut by gold mineralization in the Norseman-Wiluna Belt to the west of the
Fault. This suggests that at least some of the gold deposits east of the Keith-Kilkenny Fault are older
than the majority of gold deposits to the west of the Fault and older than other deposits within the
same eastern terrane (e.g., Porphyry, Granny Smith; Fig. 2).
Direct dating of gold deposits by robust techniques, such as U-Pb or Pb-Pb of alteration minerals, and
by less-robust Ar-Ar techniques for muscovite, mainly produce ages in the range ca. 2625 to 2640
Ma, consistent with the maximum and minimum age constraints provided by most pre- and post-
mineralization intrusive rocks. Examples include a U-Pb in zircon SHRIMP age of 2627±13 Ma for
Mt Gibson (Yeats et al., 1996), a Pb-Pb age of rutile from Victory at Kambalda of 2627±14 Ma
Page 141
(Clark et al., 1988), Pb-Pb ages of titanite from Griffins Find and Reedys of 2636±3 Ma (Barnicoat et
al., 1991) and 2639±4 Ma(Wang et al., 1993), respectively, and Ar-Ar muscovite ages of 2629±9 Ma
at the Golden Mile (Kent and McDougall, 1995), 2622±12 Ma at Matilda (Kent and Hagemann,
1996), and 2639±16 Ma at Big Bell (Mueller et al., 1996). Significantly younger ages, which are still
geologically reasonable, include an Ar-Ar muscovite age of 2602±8 Ma for Mt Charlotte (Kent and
McDougall, 1995) and U-Pb in titanite SHRIMP ages of minor, second-stage mineralization at Big
Bell at 2614±2 Ma (Muller et al., 1996) and Lawlers at 2590±9 Ma (Fletcher et al., in Yeats and
McNaughton, 1997). An Ar-Ar muscovite age of 2590±9 Ma for East Lode, Wiluna (Kent and
Hagemann, 1996) appears unreasonable in that it post-dates the interpreted age of cratonization at ca
2600 Ma. whereas the hosting faults are D3 or D4 structures.
It thus appears that gold mineralization took place between about 2660 and 2610 Ma, with the most
voluminous and widespread episode between 2640 Ma and 2630 Ma. This is some 40 to 90 m.y. after
the latest major episode of basic-ultrabasic magmatism, and late during the major episode of craton-
wide felsic magmatism. The geochronological data are entirely consistent with the relative structural
data, both confirming the late timing of gold mineralization. To the east of the Keith-Kilkenny Fault,
some gold formed at the old end of the 50 m.y. long period of hydrothermal activity, suggesting that
this terrane (Kurnalpi Terrane of Myers, 1993) could have been subjected to an earlier gold
mineralization event, although structural data from the apparently older deposits suggest that they are
still late within local deformational sequences.
8.3.4 Geochronology: Other Terranes Worldwide
The growing body of absolute geochronological data from orogenic gold deposits worldwide is
consistent with relationships in the Yilgarn Block, (e.g., see summaries in Kerrich and Cassidy, 1994;
Groves et al., 1998; McCuaig and Kerrich, 1998). Most orogens tend to evolve over about 100- to
200-m.y.-long periods, and it is during the later deformational events that gold veining occurs,
generally forming a diachronous pattern across the entire evolving orogen within the uplifting and
cooling parts of the host terranes.
In the Precambrian, outside of Western Australia, timing of events are particularly well-established
for the Superior Province in central Canada (Kerrich and Cassidy, 1994), where ages of Late Archean
rocks range from >2.9 Ga to ca. 2.65 Ga, among a variety of amalgamated terranes. The last
accretionary events included the ca. 2.7 to ≤2.67 Ga progressive collision of the Quetico, Abitibi-
Wawa, Pontiac, and Minnesota River Valley terranes on to the southern margin of the previously
formed part of the Superior Province (Card, 1990). Also, within the gold-rich Abitibi-Wawa terrane
itself, there is a younging southward of the greenstone sequences from 2730-2720 Ma in the north to
as young as 2680 Ma in the south (Jackson and Cruden, 1995). Deformation, regional
metamorphism, and voluminous calc-alkaline magmatism were widespread in the terrane at 2700-
2670 Ma, with much of the deformation along the major Porcupine-Destor and Larder Lake-Cadillac
crustal shear zones at ca. 2680-2670 Ma. Because absolute dates on gold ores cluster near the end of
the 30-m.y.-long deformation range (Kerrich and Cassidy, 1994) and gold shows a spatial association
with these faults that formed at about this same time (Jackson and Cruden, 1995), it is apparent that
the major orogenic gold ores of the Superior Province are also late kinematic.
Not all gold deposits in the Superior Province may have formed by ca. 2670 Ma. There are a series
of more controversial younger dates on some of the gold deposits within the Abitibi-Wawa terrane
that give ages of mineralization of 2630 Ma and younger (Kerrich and Cassidy, 1994). Final ductile
deformation and formation of subvertical faults, prior to cratonization, also occurred within the
southern Superior Province to as late as 2660-2640 Ma (Jackson and Craden, 1995). This is certainly
an expected deformational event, as final collision and subduction of the Pontiac and Minnesota
River Valley terranes seemingly occurred throughout this period. Therefore, other episodes of fluid
flow and gold veining after the initial ≥2670 Ma event are not inconsistent with the late tectonism.
Page 142
As far back as the Middle Archean, gold veining is tied to late orogenic events. In the Barberton
greenstone belt of South Africa, collisional processes are recognized between about 3230 and 3080
Ma, but gold veining did not occur until about 3084 Ma, when a major change to strike-slip tectonics
occurred (deRonde and deWit, 1994). The Proterozoic Birimian belt in western Africa evolved
between 2190 and 2080 Ma, with most of the deformation concentrated during the last 40 m.y.
(Oberthur et al., 1998). The most extensive period of metamorphism and magmatism in this period
was at about 2105 Ma, with the Ashanti belt gold veins being emplaced at ≥2098-2085 Ma near the
end of Eburnian orogenesis.
Late timing of vein emplacement also characterizes many of the Phanerozoic orogens. In the
Variscan of southern Europe, orogeny, and subsequent extension and uplift extended from 440 to 280
Ma (Rey et al., 1997). Orogenic lodes within areas of basement uplifts, such as the Bohemian
Massif, Massif Central, and Iberian Massif, were deposited coevally with the later part of 360-290
Ma magmatism (Bouchot et al., 1989, Le Guen et al., 1992, Murphy and Roberts, 1997; Stein et al.,
1998). In the southern part of the Cordilleran orogen of North America, initial folding of sedimentary
rocks and associated ductile deformation occurred at about 160-150 Ma in the Foothills terrane
(Tobish et al., 1989), with gold veining some 25-50 m.y. after initial deformation (Bohlke and
Kistler, 1986). Initial terrane collision and resulting deformation of rocks that host the ores of the
Juneau gold belt began at about 100-90 Ma (Gehrels et al., 1992), with ore deposition some 35-45
m.y. later during changes in plate motions at about 55 Ma (Goldfarb et al., 1991).
8.3.5 Implications of late timing
The late timing of most Archaean, Proterozoic, Paleozoic, Mesozoic and Tertiary orogenic lode-gold
deposits worldwide carries with it implications that the present structural geometry within the
deposits, and the structural geometry within the host terranes, approximate those at the time of gold-
deposit formation. Post-ore strike-slip faults may offset coeval gold deposits on opposite sides of
major structures, but the within-terrane structural geometry is little affected. Transtensional and
uplift-related normal faulting is normally the only significant feature superimposed on the late-
forming orogenic gold deposits. Thus, geological maps, and computer-based methodologies based
on their analysis, can be used as a predictive tool in gold exploration where the structural controls on
mineralization are understood and repetitive structural geometries are produced.
8.4. Structural geometry of orogenic lode-gold deposits
Structure is widely considered to be the single most-dominant control on localisation of lode-gold

deposits (e.g., Sibson et al., 1988; Hodgson, 1989; Cox et al., 1991). Hence, the structural geometry
of mineralized environments should be predictable given the late timing of mineralization. This is
well known for strike-slip faults and shear zones in brittle-ductile transpressional regimes (e.g.,
Hodgson, 1989; Sibson, 1990), where deposits occur in predictable dilational sites, such as in
dilational jogs and at overstepping faults. For example, this is well illustrated for the Wiluna gold
deposits in the Yilgarn Block (Hagemann et al., 1992). Younger gold deposits, such as those in the
Fairbanks area of central Alaska and the Chicagoff district of SE Alaska, may also be sited along
cross faults between first-order strike-slip faults (Le Lacheur, 1991; Goldfarb et al., 1997). In
compressional brittle-ductile regimes, the lode-gold deposits are commonly sited in dilational zones
where the shear zones or faults are more gently dipping than elsewhere. For example, this occurs at
the Granny Smith (Ojala et al., 1993) and Sons of Gwalia (Smith and Gardoll, 1997) deposits in the
Yilgarn Block. Flower structures (e.g., at the Ashanti deposits, Ghana) are also predictable sites of
gold mineralization in compressional regimes, as are suitably oriented structures in zones of
heterogeneous stress fields around rigid granitoid plutons (e.g., Granny Smith, Yilgarn; Ojala et al.,
1993: Kensington, SE Alaska; Miller et al., 1995) or complex batholith margins (e.g., Coolgardie,
Yilgarn; Knight et al., 1993: Willow Creek, south-central Alaska; Goldfarb et al., 1997), or in the
pressure shadows at the terminations of elongate batholiths: see summary diagrams in Groves et al.
Page 143
(1995). Fluctuations in local stress fields can lead to a complex pattern of coeval shear and
extensional vein networks (Miller et al., 1994; Nguyen et al., 1998).
Apart from such deposits preferentially hosted in structural heterogeneities near major crustal faults or
shear zones, other orogenic lode-gold deposits may be confined to specific lithologies (i.e. they are
essentially stratabound) within complex lithostratigraphic sequences. These can also show repetitive
and predictable structural geometries. Markedly linear greenstone belts are bounded by crustal-scale
shear zones, forming elongate crustal blocks in dominantly granitoid-gneiss terrains. Greenstone belts
with a large percentage of relatively competent greenstone lithologies oriented at a high angle to the
maximum principal far-field stress that was active during progressive deformation, are the most
common sites of large gold deposits (e.g., Groves and Barley, 1994). This is because these more
competent rock units tend to fail selectively in this orientation with respect to the far-field stress
(Ridley, 1993). Within the linear greenstone belts, structurally isolated units of the most competent
volcanic rocks especially tend to form zones of structurally-enhanced permeability at all scales. The
geometry of these structurally-isolated units varies, but there are repetitive geometries shown
schematically on geological maps. In the Yilgarn Block, for example, one repetitive geometry is
shown in Figure 4, where broadly NW-trending competent units are isolated by broadly NE-trending
late faults or shear zones. These isolated blocks are selectively fractured as the confined ore fluids
reach supalithostatic pressures, leading to the formation of large vein arrays and associated alteration.
The mineralization is typically brittle in style, although local stress reorientation can lead to shear
failure (Nguyen et al., 1998). In brittle-ductile regimes, thick units of such competent rocks may be
structurally isolated by complex interactions between early compressional (e.g., thrust) structures and
later transpressional structures to form curvilinear blocks which close on one or both ends. Such
structural geometries are the sites of highly anomalous fluid flux and can produce giant gold deposits.
Example of such geometries are shown in Figure 5, which illustrates the overall similar structural
geometry of the two giant orogenic lode-gold goldfields, Kalgoorlie and Timmins, in Late-Archaean
greenstone belts. The geometry of the geological setting of the Pampalo deposit from Finland (Nurmi
et al., 1993) is shown to illustrate how similar structural geometry may also characterise a smaller
deposit.
Figure 4. Schematic diagram showing examples of structurally controlled lode-gold deposits from
the Yilgarn Block, shown in the same orientation at the same scale. The NE- trending faults have
divided the sequence into a number of isolated or semi-isolated blocks of competent units which are
selectively mineralized in a brittle-ductile to brittle regime. Adapted from Knight et al. (1996),
Groves et al. (1997) and Newton et al. (1997).
In Proterozoic and Phanerozoic sedimentary rock-dominant terranes, gold deposits are, in places,
located in fold hinges in belts that are oriented at a high angle to the far-field stress, because the
Page 144
weaker rocks, such as graded greywacke units, selectively fail in such fold zones where the layering
is oriented sub-parallel to the far-field stress (Ridley, 1993). Examples are especially well-
documented for the Pine Creek region (Northern Territory, Australia), the Meguma terrane (Nova
Scotia, Canada) and the Bendigo and Ballarat districts (Victoria, Australia). The deposits are
commonly sited in anticlinal hinges and particularly at domal culminations (e.g., Thomas, 1953;
Kontak et al., 1990; Partington and McNaughton, 1997). Other deposits may be located in, and
above, reactivated thrust ramps at the more gently-dipping segments of steep reverse faults that are
active late in, or postdate, the major folding episode (e.g., Cox et al., 1991). Commonly, the gold
deposits in greywacke sequences are sited in “locked-up” structures, such as steep reverse faults or
tight overturned anticlines, which show extreme fault-valve behavior, and are reactivated and fail
under high fluid pressure relative to minimum principal stress (e.g., Sibson et al., 1988; Cox et al.,
1991). Anticline-hosted lode-gold deposits thus tend to show repetitive structural geometries in
similar rock sequences, with common parameters for deposits in greywacke sequences including low
dihedral angles, overturned back limbs, front-limb parallel thrusts, which break through on to the
back limb, and doubly plunging closures (e.g., Nesbitt, 1991).
Figure 5. Comparison of the structural geometries of the two giant greenstone-hosted orogenic lode-
gold deposits at Kalgoorlie, Western Australia, and Timmins, Canada, and a smaller gold deposit,
Pampalo from Finland. Note the similar geometry of competent units surrounded by less competent
units on a variety of scales caused by the subsequent folding and faulting of early-formed thrust
faults. Adapted from Phillips et al. (1996), Groves et al. (1997) and Nurmi et al. (1993).
In summary, it is evident that orogenic lode-gold deposits, hosted both in Precambrian greenstone
belts and Precambrian to Phanerozoic accretionary sedimentary belts, commonly show predictable
and repetitive structural geometries. Most importantly, their late timing relative to the evolution of
the host terranes means that this geometry is depicted on geological maps and/or cross sections
derived from geological mapping, which may be combined with interpretation of remotely sensed
images and airborne geophysical data sets. Exploration guides include structural heterogeneities near
major fault zones, competency contrasts within the broad lithostratigraphic sequences themselves,
and major fold hinge zones. This is the geological basis for the computer-based exploration
technologies, stress mapping and prospectivity mapping, which are described and discussed below
and are advantageous for recognition of such favourable geometric patterns.
8.5. Stress mapping and its application to orogenic gold

Page 145
deposits
8.5.1. Introduction
Stress mapping is a numerical modelling technique developed to target structurally-controlled

hydrothermal mineralization located in dilatant (or low stress) sites. It is based on rock mechanics
principles and stress-strain relationships (Holyland 1990 a, b; Holyland and Ojala 1997). The
technique is used to transform strain data, in the form of a solid geology map, to stress data. In order
to use geological maps directly as an input, it must be shown that the present rock geometries are
those that existed at the time of mineralization. As discussed in detail above, this is certainly the case
for the orogenic gold deposits globally, and for those of the Yilgarn Block, specifically.
The strong structural control of the orogenic gold deposits, particularly the association with fault and
shear zones, indicates that they are characterized by channelized fluid flow and high fluid/rock ratios.
Under medium- to high-grade metamorphic conditions, fluid pressures are buffered close to, or
commonly exceed, lithostatic pressures and fluid flow is generally upward (Etheridge et al., 1984).
Focusing of the upward fluid flow into a discrete channelway, as required to form an ore deposit, is
due to lateral fluid-pressure gradients. These lateral gradients may be induced structurally by either
variations in fracture permeability of active fault zones, or by variations in mean rock stress.
Mineralized extensional veins indicative of supralithostatic fluid pressures are common in orogenic
gold systems (Sibson et al., 1988), and these are compatible with fluid focussing into zones of low
mean rock stress. Relative fluid pressures are higher in these zones than local lithostatic rock
pressure, but absolute fluid pressures are lower than in the surrounding rocks (Ridley, 1993). Sites of
low mean stress can, therefore, be simultaneously sites of fluid focussing and of low effective mean-
stress. At most crustal levels characterized by orogenic-gold vein formation, rocks are too plastic to
allow larger-scale free convection, and flow must be driven by fluid pressure gradients.
At higher crustal levels, ambient fluid pressures are closer to hydrostatic pressures, but fluid pressures
along major conduits are still likely to be in excess of hydrostatic pressure and probably episodically
exceed lithostatic pressures. Consequently, fluid flow is strongly controlled by permeability which,
in turn, is influenced by rock heterogeneities. Although high permeability pathways can be
lithologically determined, structurally-controlled ore deposits are more common in crystalline rocks
and, in general, they have formed in reactivated fault systems (e.g., Phillips, 1972; Sibson et al., 1975;
Sibson et al., 1988): this is certainly the case for many orogenic gold deposits. A failure of a pre-
existing weakness, or intact rock, can be initiated by an increase of differential stress, by an increase
in the pore fluid pressure, or by a combination of both. In most cases, stress changes lead to decrease
of mean stress or, more commonly, decrease of minimum principal stress.
A model in which fluid focussing in the crust is due to variations in mean stress, or in which failure of
pre-existing weaknesses and, therefore, enhanced fracture permeability, is due to lowered mean or
minimum stress, best explains the wide variety of structural settings of many vein-type mineral
deposits, including orogenic gold deposits. Thus, a technique that measures variations in rock stress,
specifically where the geometry of the ore-hosting structure is complex, has the potential to generate
viable exploration targets. Complex geometry (e.g. variations in orientations of structures) and large
rheological contrasts are important because different structures and rock types accommodate strain
via different modes and different rates, resulting in heterogeneous stress distribution and hence,
potentially compressional and dilational zones in the same domain.
8.5.2 Stress mapping principles
Page 146
Stress mapping is a form of computer-based structural analysis which examines the variation in local
strain and stress within an inhomogeneous terrane resulting from the imposition of a regional stress
field (e.g. Holyland and Ojala, 1997). When stressed, an inhomogeneous material develops zones of
strain partitioning whose components vary with the geometry and rheological properties of
discontinuities and blocks, with variable deformation behavior between discontinuities. The
modelling considers only elastic and elasto-plastic stresses and strains, and, therefore, deformation
dilatency due to viscous strains is not considered.
Stress mapping as a basis for prediction of hydrothermal fluid flow is based on the following
reasonable assumptions:
1) Low minimum principal (σ3) stress indicates proximity to failure and therefore, the
likelihood of deformation-enhanced increased permeability. This is more important in
modelling of high crustal-level deformation because absolute fluid pressure is higher in the
fluid conduits (commonly faults) than in the wall rocks, and the fracture permeability/rock
permeability ratio is higher than at deeper levels.
2) At depths of more than a few kilometers, fluid pressure is consistently close to lithostatic
pressure and the control on fluid pressure is mean stress.
3) Variations in mean stress will cause variations in fluid pressure.
4) Fluid flow is both upwards and towards zones of low mean stress.
Combined with the knowledge that structurally-controlled orogenic gold mineralization is commonly
late in the tectonic history of a terrain and that typical Archaean orogenic gold deposits show
evidence of fluid overpressuring (e.g., Sibson et al., 1988; Groves et al., 1995), this enables stress
analysis of two-dimensional map patterns of rock units and faults to predict those zones of low mean
(σm= (σ1+σ2+σ3) /3) or minimum principal (σ3) stress, provided that the orientation of the far-field
stress can be reasonably assumed.
The most severe limiting assumption of two-dimensional stress modelling is that the plane of a map
does not accurately reflect the stress pattern in an area with complex three-dimensional geometry
(Holyland et al., 1993). In many terranes, such as those of the Yilgarn Block, this may not be a
critical restriction, since the vast majority of the recorded gold-hosting structures are near vertical
(Hronsky et al., 1990; Libby et al., 1990). The accuracy of the three-dimensional geological
interpretation becomes especially important in deposit-scale modelling where the extent of interest of
the vertical dimension is similar to horizontal dimensions (Holyland and Ojala, 1997) and/or where
deposits are controlled by changes in dip of the hosting structures. However, the example presented
below is restricted to two dimensions.
8.5.3 Computer program
Details of the computer program are described elsewhere (Holyland, 1990d) and only a brief
summary is presented here. Two-dimensional modelling is by a distinct element code (UDEC
program) and the method has three distinguishing features which make it well suited for
discontinuum modelling (Holyland, 1990a, b). These are:
1) An assemblage of blocks, which interact through corner and edge contacts, is simulated.
2) Discontinuities are regarded as boundaries between blocks; discontinuity (fault) behavior is

prescribed for interactions between these.
Page 147
3) An explicit timestepping (dynamic) algorithm is utilized that does not limit displacements or
rotations, and assumes general non-constitutive behavior for both the matrix and
discontinuities.
Modelling can be done at any scale and, in addition to minimum, maximum and mean stresses, total
displacement and fault displacements can be computed. The total displacement is the amount of
movement experienced by the rock blocks, and their movement directions.
8.5.4 Input for stress mapping
Stress mapping requires an accurate geological map or, if a three-dimensional model is desired, then
information on the subsurface is also critical. It also requires estimates of the magnitudes and
orientations of the far-field horizontal stresses and rock and fault deformation properties. Geological
map data are converted to a solid-geology computerized base, which provides continuous lithological
and structural information (e.g., Fig. 6). The modelled area is treated as a mosaic of polygonal blocks
(rock units) and joins (discontinuities: contacts, faults and shear zones). When external stress is
applied to this system, the blocks are juggled and deform internally until equilibrium is attained. This
identifies areas of lower stress (i.e. dilation zones), which represent favoured areas for ore-fluid flux
and hence mineralization, and therefore targets in exploration.
8.5.5 Example of two-dimensional stress mapping: Kalgoorlie

Terrane
8.5.5.1 Strategy
The geological map used to illustrate the value of stress mapping in exploration is the 1:250 000 solid
geology map of the Kalgoorlie Terrane (Swager and Griffin, 1990). Figure 6 is a simplified scale-
reduction version of this map. This map was chosen for a variety of generic reasons. First, the
terrane is a mature one in terms of geological mapping and mineral exploration, such that there is a
large geological database. Second, the area is uniformly covered by airborne geophysics, which has
aided interpretation of the solid geology map in widespread areas of deep regolith. Third, the map
was produced by a small team of experienced geologists, and hence is a relatively consistent product.
From the viewpoint of demonstration of the potential, or otherwise, of the technique, it is also
important that there are numerous gold deposits and goldfields that have produced more than one
million ounces of gold in the terrane (Fig. 6). Finally, detailed research by a number of workers in the
terrane has confirmed the late timing of the gold deposits (see summary in Witt, 1993). At the
1:250,000 scale of the solid geology map, it is not expected to define the location of individual gold
deposits, normally <1 km² in area, although the giant Golden Mile deposit (approx. 8 km²) is an
exception. Rather, the output from the technique at this scale will define the location of major
goldfields, potentially containing several individual gold deposits.
Page 148
8.5.5.2. Application
For the stress mapping, discontinuities on the 1:250,000 scale map of the Kalgoorlie Terrane were
classified according to their nature, length and continuity as first- or second-order D3-D4 faults, D1
thrusts or lithological contacts. In this order, the discontinuties define a relative stiffness scale from
lowest (first-order fault) to highest (lithological continuity) stiffness. First-order structures include
regional-scale faults that generally have strike lengths of more than 10 km (e.g., Ida Fault, Boulder-
Lefroy Fault : Fig. 6); whereas second-order structures are of shorter strike length.
Figure 6. A digitized solid-geology map of the Kalgoorlie Terrane. The geology is simplified to
facilitate modelling. The position of major goldfields, or the major deposits within them, is shown for
reference. For the geomechanical units, felsic intrusive rocks are treated as the most competent rock
type, followed by undifferentiated mafic rocks, undifferentiated sedimentary rocks (includes felsic
volcanic rocks) and, finally, undifferentiated ultramafic rocks. First-order structures are modelled as
having the lowest stiffness and lithological contacts the highest stiffness. The maximum principal
stress, σ1 is simulated at E-W, based on the history of the Yilgarn Craton and the place of gold
mineralization within it. Abbreviations for mines in the terrane : BA, Broad Arrow; Bd, Bardoc; Bl,
Page 149
Bayleys; Ch, Chalice; CR, Cave Rocks; Du, Durkin; Gh, Golden Hope; GM, Golden Mile; GS,
Gimlet South; Hg, Higginsville; Hu, Hunt; Jn, Junction; Ju, Jubilee; Kb, Kanowna Belle; Kd,
Kundana; Ki, Kintore; Ku, Kunanalling; LB, Lady Bountiful; Mtc, Mt Charlotte; MtM, Mt Morgan;
MtP, Mt Pleasant; MtPe, Mt Percy; NC, New Celebration; Pad, Paddington; RW, Prince of Wales;
Ot, Otter; Re, Revenge; Si, Siberia; Stl, St Ives; TMH, Three Mile Hill, Ti, Tindals; VD, Victory-
Defiance.
Rock types on the geological map of the Kalgoorlie Terrane are grouped by their interpreted
geomechanical properties into undifferentiated mafic rocks, undifferentiated ultramafic rocks,
undifferentiated sedimentary rocks (including sedimentary rocks and felsic volcanic rocks) and felsic
intrusive rocks. The felsic intrusive rocks are treated as the most competent rock type, followed by the
mafic rocks, sedimentary rocks and finally ultramafic rocks. The rock parameters used reflect more
than ten years experience of stress mapping in greenstone belts. The maximum principal stress (σ1)
direction used in the simulation is E-W. This is based on the gross greenstone belt geometry and
published structural studies (e.g., Swager 1989), and more detailed structural studies of the orogenic
gold deposits themselves (e.g., Knight et al., 1993). The technique is relatively insensitive to small
differences (<30°) in orientation of σ1. Studies of orogenic gold deposits in the Yilgarn Block and
elsewhere indicate low displacements on controlling structures during mineralization, as discussed
above. Therefore, the stress modelling was carried out with relatively low imposed compressive
stresses, 100 Mpa for σ1, and 70 Mpa for σ3, such that overall strain was less than 0.25 percent after
the models were cycled until a steady state of deformation was reached, corresponding to a constant
stress distribution where no new stress anomalies developed. The northern and southern haloes of the
map, with some overlap to avoid edge effects, were stressed separately due to size and shape
constraints. This has led to small differences in the magnitude of the stress anomalies in each half of
the map, although, in areas of overlap, the stress anomalies are in identical locations.
8.5.5.3. Results
Sites of low minimum principal stress (σ3) simulated in the model for the Kalgoorlie Terrane are
shown in Figure 7. Low stress anomalies constitute less than ten percent of the map area, and are
typically related to changes in strike of first-order faults, largely from the dominant NNW trend to a
more restricted NW trend, and to intersections of two or more first-order faults or of first-order and
second-order faults. Despite the relatively large scale of the modelling, anomalously-low minimum
stress zones define all of the major goldfields in the area, including the Kalgoorlie (Golden Mile, Mt
Charlotte, Mt Percy), Coolgardie (Bayleys, Tindals, Three Mile Hill), Kanowna Belle, Mt Pleasant
(Lady Bountiful, Mt Pleasant), New Celebration (New Celebration, Jubilee, Golden Hope),
Paddington and Kambalda-St Ives (Victory-Defiance, St Ives, Junction) and Higginsville goldfields.
Most of these are adjacent to first-order faults in the modelled map (Fig. 7). Smaller deposits not sited
within low stress anomalies (e.g., Siberia, Prince of Wales) do not have significant faults or fault
intersections nearby (Fig. 7), and it is probable that less significant controlling structures were simply
not shown on the 1:250,000 scale map which was modelled.
Page 150
Figure 7. Equivalent map to Figure 6, showing contoured values of minimum principal stress (σ3).
Based on stressing the digital map of the Kalgoorlie Terrane with σ, in an E-W orientation. The first-
order faults are shown to assist correlation between Figures 6 and 7. Anomalously low values of σ3
have a strong correlation with the locations of known goldfields in the Terrane, but there are also
anomalies which potentially define as-yet undiscovered goldfields or gold deposits.
8.5.6 Discussion
The two-dimensional, regional-scale stress modelling of the Kalgoorlie Terrane shows the ability of
stress mapping to define zones of low minimum principal stress (dilation zones), which correspond
very well with locations of known goldfields. It also defines several new target areas of high
prospectivity. The modelling of palaeo-stress patterns in the Kalgoorlie Terrane is in agreement with
the earlier-described orogenic gold model in which the fluid flow and location of mineralization are
mainly controlled by major structures and rheological heterogeneities. The fact that this simulation
successfully located the major goldfields is attributed to the fact that the gold lodes formed late in the
tectonic evolution of the Kalgoorlie Terrane. As a result, the goldfields remain in essentially the same
spatial setting as that when they were first formed, in close proximity to complexities in the first-order
structures of the terrane.
Page 151
Numerical modelling of rock deformation is not an exact science and there will always be controlling
variables which are not fully understood. That is, rock mechanics models are "data-limited", and
there are seldom enough data to simulate the rockmass behaviour unambiguously (Starfield and
Cundall, 1988). However, two-dimensional modelling of the Kalgoorlie Terrane indicates that results
which are useful in exploration for orogenic gold deposits can be obtained with a good geological
database. Critical to the successful modelling are reasonably accurate input parameters for rock
properties developed from extensive experience in the terranes, and the far-field stress field. In the
example from the Kalgoorlie Terrane, stress mapping demonstrates which faults and which segments
of the faults are most likely to be dilational at a regional scale, and hence the predicted location of
major goldfields. Such knowledge also assists with future data collection and interpretation, because,
for example, it defines the direction and magnitude of fault strikes which are likely to be critical in
controlling the siting of goldfields and gold deposits. Thus, stress mapping can focus attention on
those geological parameters that are critical to goldfield location and thereby aid in discriminant data
collection of value for conceptual exploration.
It is evident that stress mapping is a viable tool for exploration of broad crustal tracts to define zones
of greatest favourability for orogenic gold deposits, provided that there are adequate geological maps
and other geological data available. The technique also has been applied successfully on a regional
scale to Archaen rocks in the Mt Pleasant area north of Kalgoorlie (Holyland et al., 1993), in the
goldfields of the Murchison greenstone belt of South Africa (Vearncombe and Holyland 1995), in a
segment of the Midlands greenstone belt of Zimbabwe (as reported in Campbell and Pitfield, 1994),
and to the Rouyn-Noranda district, Abitibi, Canada (Jébrak et al., 1995). In all of these cases, the
majority of known gold deposits are coincident with broad low minimum-stress anomalies and, most
significantly, additional anomalies were defined, identifying the most permissive areas for further
exploration.
As discussed above, the regional stress mapping is capable of defining anomalies at the scale of
goldfields. Once these potentially mineralized areas have been defined, the next stage is to carry out
a more detailed analysis at a smaller scale (e.g. 1:25,000 to 1:50,000 scale) in order to define more
specific gold targets within the defined potential goldfields. At this scale, it is important to further
subdivide the rock units in terms of their strength parameters. For example, dolerite, the host to many
of the large gold deposits in the Kalgoorlie Terrane, is simply one rock type grouped within the
category undifferentiated mafic rocks in the stress model at 1:250,000 scale shown in Figures 6 and 7.
This can be done in two dimensions, as for the example of the Kalgoorlie Terrane above, or in three
dimensions if there is sufficient drilling data in the area (Holyland and Ojala, 1997). An example of
such a stress mapping study is that of the area surrounding the Granny Smith gold deposit in the
Eastern Goldfields Province of the Yilgarn Block (Fig. 2). In this case, low minimum stress
anomalies and anomalous local σ1 orientations correlate with changes in orientation (both strike and
dip) of a reverse-sheared contact between diorite and metasedimentary rocks, which, in turn, correlate
with ore zones in one or both host rocks dependant on the contact configuration (Holyland and Ojala,
1997: figs 5-8). In this case, progressive three dimensional stress mapping greatly assisted siting of
drill-holes during ongoing exploration.
8.6. Computer-based prospectivity mapping
8.6.1. Introduction
Since the late 1970s, a computer-based technology has been steadily evolving that has the potential to
revolutionize the mineral exploration industry. Geographic Information Systems (GIS) are computer-
based databases for spatial and associated non-spatial data that provide the necessary tools for the
efficient storage and analysis of map and image data. For over a decade, the use of GIS as an
exploration tool has been examined by a number of research groups throughout the world (Bonham-
Page 152
Carter et al., 1989, Knox-Robinson, 1994, Wyborn et al., 1994) and for a number of different deposit
types, which importantly include orogenic gold deposits (Robinson, 1989, Knox-Robinson, 1994).
Although GIS is a very powerful tool in exploration, there has been a reluctance, until recently, to
accept the technology. Reasons for this include the dearth of digital map data, the expense of the GIS
software and associated hardware, and the user-unfriendliness of GIS that necessitated the services of
a specialized operator. Fortunately, the past few years have witnessed the emergence of so-called
‘desktop’ GIS that are inexpensive, PC-based, and very user-friendly, although less capable than
conventional systems. In addition, a number of tools are now available that allow sophisticated
exploration-relevant analyses to be applied to stored map data and imagery.
Specifically, GIS has been used, in conjunction with typical exploration data sets, to test concepts
related to the formation of orogenic gold deposits and to identify and quantify spatial associations
between known deposits and other geological features (Knox-Robinson, 1994; Knox-Robinson et al.,
1996; Knox-Robinson and Groves, 1997; Knox-Robinson and Wyborn, 1997; Yun et al., 1998). The
ultimate aim of the research is to develop the methodology to construct ‘prospectivity’ maps that have
a predictive capability in assessing the likelihood that a given area could host an orogenic gold
deposit. Attention has focused on regional-scale prospectivity mapping, in which the goal is to
identify prospective areas at a goldfield scale using primarily conventional geological map data. An
application is presented below for the orogenic gold deposits of the well-documented Kalgoorlie
Terrane in order to link this methodology to the stress mapping technology discussed above.
8.6.2 Principles of prospectivity mapping
There are two general approaches to prospectivity mapping: conceptual and empirical (Knox-
Robinson & Wyborn, 1997). Using a conceptual approach (e.g., Wyborn et al., 1997), genetic
concepts are represented as mappable criteria and these are ultimately integrated into a single
prospectivity map. An empirical approach, on the other hand, relies on the existence of a significant
number of known deposits which are characterized by a series of defining factors. Such factors are
spatially quantified and integrated into a single prospectivity map using the same techniques as for the
conceptual approach.
The Yilgarn orogenic gold deposits form a coherent group and share a common genesis, but vary in
their physical parameters due to differences in structural style, host rock, and degree of alteration, and
therefore, it is not appropriate to employ a purely conceptual approach. Fortunately, exploration in the
Yilgarn Block has resulted in the discovery of more than 2000 gold deposits. Furthermore, the current
structures and geometries on geological maps are similar to those that existed at the time of
mineralization, as discussed above. Consequently, an appropriate predominantly-empirical
methodology has been developed that exploits knowledge of known gold deposits and high-quality
map data. The methodology is flexible so that relevant concepts can be tested and, if important,
included in the analysis and the subsequent construction of a prospectivity map (Knox-Robinson,
1994). The methodology comprises three main analytical steps of equal importance: identification of
spatial relationships on maps, quantification of those identified relationships, and integration of
multiple relationships into a single prospectivity map.
8.6.2.1 Step One: Identification of Spatial Relationships
The first step involves the identification of factors that may constrain the location of known gold
deposits. Some features, such as crustal-scale faults and shear zones, can control the linearity of
orogenic belts, and typically serve as ore-fluid pathways. Since fluid pressure variations are
important, zones of maximum uplift or extension (decreased lithostatic pressure) may be more
prospective than others. Along these crustal-scale structures and in uplifted belts, a number of smaller
structures can represent zones of enhanced permeability and fluid flow. These include dilational jogs
along shear zones, dilation at fold hinges, and portions of deformed lithological contacts that separate
Page 153
rock types of strongly contrasting rheologies. Consequently, particular strike-ranges of faults, shear
zones and lithological contacts may be more important than others, and hinge zones may also be
highly prospective. In addition, gold deposits may simply be associated with an increased abundance
of permeable faults, shears and lithological contacts. Specific rock types that are most competent
and/or have a very reactive geochemistry (e.g., high Fe/Fe+Mg+Ca ratios) should also be favoured
sites for orogenic gold deposits.
A number of techniques have been developed to test, amongst others, the above relationships (Knox-
Robinson, 1994). At the scale of investigation (normally between 1:100,000 and 1:500,000), gold
deposits are adequately represented in a GIS as point entities, and other geological structures can be
represented as point, line, or polygon entities. With such a representation, there are three broad spatial
relationships that can exist between orogenic gold deposits and other geological features. A proximity
relationship is one in which deposits are hosted preferentially closer to a feature than expected. An
association relationship is one in which deposits are preferentially hosted in a particular polygon type
(e.g., a particular rock type). An abundance relationship is one in which deposits are more likely in
areas where there is a high spatial density of a particular feature (e.g., faults or shear zones). By using
the data-query functionality afforded by most GIS, a subset of geological features can be selected to
determine, for example, if the strike of faults or shear zones is related to prospectivity. Apart from
controlling the location of orogenic gold deposits, particular geological features may, in addition,
have a bearing on the size of deposit. Consequently, techniques have also been developed to examine
size relative to proximity, association, and abundance relationships. All of the GIS-based software
routines employ non-parametric statistical tests, such as the Kolmogorov-Smirnov and Chi-square
tests, to determine the significance of the spatial relationships (Knox-Robinson, 1994; Knox-
Robinson et al., 1996).
8.6.2.2 Step Two: Quantification of Identified Spatial Relationships
Once critical factors are identified, a geological map or image needs to be constructed that depicts
how the spatial relationships between these factors vary. Such an approach uses a metric that denotes
the favourability for discovery of new deposits based on the characteristics of the known gold
deposits. Suitable metrics include the number of known gold deposits of a given minimum production
per square kilometer or, preferably, the tonnage of gold production plus reserves per square kilometer.
Even if a statistically significant spatial relationship has not been identified, the contribution of a
particular geological feature can be quantified and incorporated into a prospectivity map. Stress maps,
discussed above, could also be quantified for inclusion into the final prospectivity map.
Vector-based GIS display map features as point, line or polygon entities, and are best suited for the
management of discrete data. With such a GIS, a map that depicts a quantified spatial relationship
comprises a number of polygon features separated by distinct boundaries. The other kinds of GIS
(raster- and quad-tree-based) store map data as pixilated images, or some encoded equivalent thereof,
similar to that produced by a computer monitor. Each cell in the image has a value that denotes the
state of a particular feature at that location. With this kind of GIS, a spatial relationship can be
quantified as a continuous surface. It is noteworthy that, as geological maps comprise sharp
boundaries, a vector GIS is the most suitable for the identification of spatial relationships. However, a
raster format is preferred for the quantitative representation of an identified relationship. Fortunately,
most GIS software, including desktop packages, are hybrid, and can accept and process, with
different degrees of flexibility, spatial data in both map and image form.
8.6.2.3 Step Three: Integration of Identified Spatial Relationships
The third step is the integration of multiple quantified relationships into a single prospectivity map.
The majority of papers in the GIS research field (e.g., Reddy, 1987, Bonham-Carter et al., 1989,) tend
to concentrate on this end-product step. Consequently, there are several different integration
Page 154
techniques available, ranging from very simple Boolean techniques to more complex index overlay,
Bayesian, algebraic, and fuzzy logic methods. Knox-Robinson & Wyborn (1997) discuss these in
some detail and this is not repeated here. However, the main features of these methods are
summarized in Table 1.
Table 1. Summary of the main techniques used to combine multiple spatial datasets into a single
prospectivity map. These techniques are discussed in more detail in Knox-Robinson and Groves (1997).
Boolean Integration
Description Input maps can only have two states (prospective and unprospective). Boolean
logic rules are used to combine datasets. The resultant map has only two states.
Pros Easy to implement in most GIS.
Cons Boolean "AND" is too conservative in the definition of prospective areas, whereas
Boolean "OR" is too liberal.
Index Overlay
Description Input maps comprise two or more levels of prospectivity: the higher the
prospectivity, the larger the number. Inputs are combined by spatial overlay and
summation of index values.
Pros Easy to implement in most GIS. Better than Boolean method in that much internal
structure is retained in the resultant prospectivity map.
Cons Prospectivity is represented by an ordinal value: A region with an index value of
10 is more prospective than a region with an index value of 5, but not twice as
much. Input maps that are split into a large number of prospectivity levels will
contribute more to the resultant map.
Algebraic Method
Description Essentially an extension of the index overlay method, except that prospectivity
values are represented by ratio-scale numbers. That is, a region with an output
value of 10 is twice as prospective as a region with an output value of 5.
Pros Deposit size can be incorporated into the analysis.
Cons Difficult to integrate spatial relationships quantified as continuous surfaces.
Weights of Evidence
Description Uses Bayes' Rule of conditional probability to update a prospectivity map based
upon new information.
Pros Provides a probabilistic result.
Cons Difficult to incorporate deposit size into the analysis. Difficult to implement in a
vector GIS.
Fuzzy Logic
Description Uses the rules of fuzzy logic, a superset of Boolean logic, to combine datasets.
Unlike Boolean logic, in which a variable can have only one of two possible states
(`0' or '1'), fuzzy logic variables can have an infinite number of values between `0'
and Ì', inclusive.
Pros Can be used to integrate spatial relationships that have been quantified as
continuous surfaces
Cons Dependent on the fuzzy logic rule used to combine datasets.
8.6.3 Prospectivity Analysis of the Kalgoorlie Terrane
To provide continuity with the section on stress mapping above, a prospectivity analysis was
conducted for the Kalgoorlie Terrane of the Yilgarn Block, based predominantly on a 1:250,000-scale
solid geology map produced by Swager and Griffin (1990). Importantly, the geological data and
degree of detail of presentation are consistent throughout the entire region: i.e., mineralized areas will
not simply be defined on the basis that they contain superior data – a problem inherent in
prospectivity mapping.
8.6.3.1 Identification of Spatial Relationships
A database of more than 230 known orogenic gold deposits was constructed for the area. Of these, a
subset of 150 deposits was extracted that included the largest five gold deposits in the region. This
subset, which is not used in any of the analyses, is used to test the applicability of the resultant
Page 155
prospectivity map. With the aid of custom-built GIS routines (Knox-Robinson, 1994, Knox-Robinson
et al., 1996), a number of important spatial relationships are identified in the region.
Selected relationships are illustrated in Figure 8. First, analysis shows that deposits within 3 km of a
granitoid-greenstone contact are controlled by a different set of factors than more distal deposits. The
locations of granitoid-proximal gold deposits are spatially related to the strike of the granitoid-
greenstone contact, with the strike range of 116°-146° being the most prospective. This is similar to
the preferred strike of lithological contacts (see Fig. 8b) and fault segments. In addition, the curvature
of the granitoid-greenstone contact is important, with sections of higher curvature being more
prospective than others. This may be due to zones of dilatancy being produced in pressure shadows
around rigid granitoid bodies during post-emplacement deformation.
Distal to the granitoid-greenstone contacts, different factors control the location of gold deposits.
Crustal-scale faults and shears, below referred to simply as faults, provide the strongest control on the
location of granitoid-distal gold deposits. Not only are gold deposits more likely to occur close to
faults, but fault-proximal deposits tend to be larger (Fig. 8a). The strike of faults is also important,
with significantly more deposits than expected located close to sections of faults striking within the
range of 119°-150°. Lithological contacts are identified as the second main control, although a size-
proximity relationship cannot be identified. Sections of lithological contacts that strike within the
range of 125°-148° are more prospective than others (Fig. 8b). There is also a positive abundance
relationship between the location of gold deposits and lithological contacts, with gold deposits more
likely to be sited where there is a high spatial density of lithological contacts. Furthermore,
lithological contacts that separate rocks of strongly contrasting rheologies are more important than
others (Fig. 8c) About 70 percent of deposits are in anticlinal zones or broad horst-like structures, and
75 percent of gold occurs within 1.5 km of an anticlinal hinge (Fig. 8d). Finally, for both granitoid-
distal and granitoid-proximal deposits, mafic rocks, particularly differentiated mafic intrusions, tend
to be the most prospective host rock (Fig. 8e).
Page 156
Figure 8. Selected spatial relationships identified between orogenic gold deposits and particular
geological features in the Kalgoorlie Terrane.
a) Graph displaying the size-proximity relationship that exists between gold deposits and
regional-scale faults. Large deposits are those with >1.5t production and small deposits those with
<1.5t production.
b) Results of strike-proximity analysis conducted between gold deposits and lithological
contacts. Sections of such contacts within a strike range of between 120° and 150° are more
prospective than others.
c) Normalised gold production according to the lithological contact nearest to the gold
deposit, with numbers indicating rheological contrast (0 = no contrast, 4 = maximum contrast).
Page 157
Figure 8. Continued
d) Proportion of discovered gold located within any given distance of an anticlinal hinge.
e) Importance of rock-types based on gold production (left graph) and gold production
normalised by area of each rock-type (right graph).
8.6.3.2 Quantification and Integration of Identified Spatial Relationships
For each of the identified spatial relationships, one or more images show how the distribution of
known gold deposits relates to the features of interest. In the case of faults, one raster image quantifies
the identified proximity relationship between known gold deposits and sections of faults that strike
within the range 119°-150°. Prospectivity is depicted in terms of known gold per square kilometer.
Similarly, the second image quantifies the relationship between known gold deposits and the
remaining faults. The contribution of faults to overall prospectivity is then determined by the
combination of these two images using a simple maximum operator. That is, each output pixel is
assigned the larger of the two corresponding input pixel values. The other identified spatial
relationships are quantified in a similar way. Finally, the quantified factors are scaled to fit the range
between 0 and 1, and are integrated using a new technique, termed vectorial fuzzy logic (Knox-
Robinson, in press).
Page 158
Fuzzy logic (Zadeh, 1973) represents a superset of conventional Boolean logic, and is perfectly suited
to the task of combining multiple spatial datasets, especially those that are represented as continuous
surfaces. A number of fuzzy logic combinatorial operators have been used to combine exploration
datasets into a single prospectivity map (Bonham-Carter, 1995; D’Ercole, 1998), but, owing to
limitations in these simple functions, a new operator, the gamma function, is normally used in
prospectivity mapping (Bonham-Carter, 1995).
Although the gamma function is superior to conventional Boolean and simple fuzzy logic techniques,
it has a number of limitations (Knox-Robinson, in press). It is unable to address null data values and,
more importantly, it is not possible to ascertain from conventional techniques whether areas of
intermediate prospectivity are the result of combining several similar input values, or whether all of
the input values are low except for one or two extremely high values. These and other limitations
have acted as the catalyst for the development of the new vertical fuzzy-logic technique in which a
measure of confidence can be incorporated into the analysis (Knox-Robinson, in press).
The resultant prospectivity map of the Kalgoorlie Terrane is presented in Figure 9, with an artificial
illumination to enhance the visualization of prospectivity. Both height and color are related to the
calculated prospectivity: purples and blues and topographic lows are areas of low prospectivity,
whereas reds and whites and topographic highs represent highly prospective regions. Gold deposits
with more than five tonnes of total contained gold are depicted on this map as red circles. A section of
the prospectivity map, represented by the orange box on Figure 9, is presented in Figure 10 as a
perspective view.
Figure 9. Prospectivity map of Kalgoorlie Terrane generated using a new vectorial fuzzy logic
technique (Knox-Robinson, in press). Purples and blues represent low prospectivity, greens and
yellows represent moderate prospectivity, and reds and whites represent high prospectivity. An
Page 159
artificial sun-angle has been applied to this image, such that topography is proportional to
prospectivity also: i.e. high topography corresponds to high prospectivity.
Figure 10. A section of the prospectivity map for the Kalgoorlie area, presented as a perspective
view. The area represented by this image is illustrated in Figure 9.
8.7. Comparison of Stress and Prospectivity Maps
Prospectivity maps generated using stress mapping have a conceptual premise: the foundation of this
method is that areas of low minimum stress are those most likely to host goldfields. Consequently, to
construct a prospectivity map using stress mapping, only a solid-geological map is required and a
database on the siting and size of gold deposits in the terrane is not needed. In contrast, a
prospectivity map generated using the fuzzy logic method has a predominantly empirical basis.
Although a number of concepts related to gold mineralisation are tested, this method involves the
analysis of a subset of known gold deposits, and the relationship of these deposits to particular
geological features. Without a significant number of known gold deposits, a prospectivity map of this
type cannot be constructed. Comparison of prospectivity maps for the Kalgoorlie Terrane
demonstrates that the stress mapping and fuzzy logic methods are complementary. The majority of
low minimum stress centres (>70%) are coincident with areas considered highly prospective on the
fuzzy logic prospectivity map. In the northern part of the Terrane, where the greenstone belt is very
narrow, there are a few areas where centres of low minimum stress correspond to the position of
small goldfields at Bardoc and Siberia but are represented by zones of only moderate prospectivity on
the fuzzy logic map. However, in other areas, the fuzzy logic map provides a better definition of the
location of known gold fields. The best example is that of the Coolgardie goldfields. In this region,
areas of low minimum stress are centred on undifferentiated mafic and ultramafic rocks, bounded by
Page 160
first-order faults and surrounded by undifferentiated sedimentary rock. The fuzzy logic method
correctly highlights the area further west as the most prospective, based on the proximity to the
granitoid and the curvature of the greenstone-granitoid contact.
It can be demonstrated that prospectivity maps, constructed using the above methodology, have
definite predictive capability with the aid of the subset of known deposits that were not used in any of
the analyses, and that the results closely parallel those of stress mapping. For the prospectivity map
constructed for the Kalgoorlie Terrane, which implemented an algebraic method of integration
(Knox-Robinson, 1994), the most prospective category occupies less than 0.3 percent of the
greenstone belt, yet hosts more than 16 percent of known gold deposits. This represents a 55-fold
increase for the likelihood of the discovery of an orogenic gold deposit. Most significantly, this area
defined by the GIS approach accounts for more than 80 percent of known gold production for the
region (Knox-Robinson and Groves, 1997), demonstrating that the prospectivity map also
discriminates in favour of the most-desired larger gold deposits in the region. Furthermore, only
deposits that lie within the northern half of the terrane (north of 6,570,000 mN) were used in the
analysis, and subsequent construction of the prospectivity map. Known prospective areas in the south,
such as the Kambalda-St Ives goldfield, are highlighted as being prospective, thus indicating that the
high prospectivity areas identified were correctly defined to best identify favourable tracts for gold
deposits in the Kalgoorlie Terrane. However, although similar factors control the location of gold
deposits in all of the constituent greenstone belts of the Yilgarn Block, specifics of these factors, such
as important strike-ranges and important lithological contacts, differ due to factors such as differences
in far-field stress directions and different lithostratigraphic sequences. Consequently, it is not possible
to export identical factors from one terrane and apply them, directly, to another, although the same
principles will apply.
As shown above, GIS are useful tools for the quantitative analysis and integration of diverse
exploration data sets for late-hydrothermal deposits, such as the orogenic gold deposits. However,
although GIS have a good topological foundation and are designed to effectively store and
manipulate the spatial relationships between map features, they are not well suited for the analysis of
the shapes of objects. As noted above, structural geometry (essentially shape) appears to be a
significant factor in the focussing of ore-fluids and the formation of deposits. Current research is
addressing this limitation inherent in current GIS, and has progressed towards the development of
new techniques and software tools for the quantitative analysis of shape and geometry with respect to
mineralization potential, particularly for orogenic gold deposits.
8.8. Potential of Shape Analysis and Artificial Intelligence in

Prospectivity Mapping
From the above, all evidence suggests that structural geometry, and the shape of mappable geological
units that result from it, are the most important parameters which can be used to define the location of
orogenic gold deposits. Although the shapes and geometries of map features are stored within the
GIS, there are very few tools available within present-day systems that allow for their quantitative
analysis in relation to the location of gold deposits. This deficiency in GIS is being addressed by the
development of a number of software tools that allow aspects of shape and geometry to be quantified
(Elliott, 1997; Gardoll et al., in press). The output from these tools can be used to highlight areas that
share similar shape and geometry properties, and can be incorporated into a prospectivity analysis and
subsequent prospectivity map.
Another technique gaining favourablilty is the application of Artificial Neural Networks (ANN’s). An
ANN represents a type of adaptive computing system that can learn from the data and is particularly
suited to the tasks of pattern recognition and classification. As such, ANN’s are used routinely for
optical character recognition and speech recognition, and, by their very nature, they are also suited to
the task of prospectivity mapping. Unlike the conventional three-step approach to prospectivity
Page 161
mapping, in which the importance of each spatial feature is examined individually, a neural network
analyzes all data simultaneously. Essentially, using a neural network requires the construction of a
‘training set’ that is representative of both mineralized and unmineralized areas (output values).
Geological information is collected for each of these locations, such as distance to nearest fault and
rock-type (input values). The ANN iteratively attempts to build a network that maps the input values
to the associated output value for each record in the training set. This is known as the ‘training’ of the
network. Once trained, the validity of the network is tested. If adequate, information for every
location is presented to the trained neural network, and a prospectivity map is constructed from the
output values.
There is a wealth of different ANN structures, philosophies and algorithms available, and, for a
particular kind of network, there are a large number of parameters that can be changed. One of the
biggest challenges in the use of ANN’s is to determine the best configuration for prospectivity
mapping. This is at an early stage of investigation, yet is showing signs of promise (Brown et al., in
press).
8.9. Conclusions
The following conclusions can be drawn from this integrated study of the nature, controls and timing
of orogenic gold deposits, and the application of these parameters to computer-based exploration
methodologies for the deposit class.
1. Orogenic gold deposits represent a coherent group of widely-distributed deposits in

accretionary or collisional orogens of all ages. They formed over an extended depth
range (<5 km to >15 km), although most commonly in greenschist-facies host rocks,
from a relatively uniform, deeply sourced, low-salinity H2O-CO2±CH4±N2 fluid with
low a H2S and, as a result, have high Au:Ag ratios and low base-metal contents and a
distinctive geochemical signature of enhanced Si, K±Na, LILE, CO2, H2O, Au, As, B,
Sb, Te and W. They differ from other gold-rich deposit groups in the same terranes,
which formed at uniformly shallow depths (normally 5 km to the surface) from highly-
saline low-CO2 fluids (e.g., VHMS and porphyry-style deposits) or acidic low-CO2
fluids (epithermal deposits) and which, characteristically, are either base-metal
(±Sn±Mo) and/or silver enriched.
2. The orogenic gold deposits are epigenetic. Structure is the first-order control on deposit
distribution, as high ore-fluid flux into permeable structures or fractured rock bodies is
an essential ore control. In many terranes, first-order crustal-scale faults and shear zones
control the regional distribution of the deposits, which, however, are normally sited in,
or adjacent to, second- and third-order structures where fluid focussing was most intense.
Other important controls on the siting of the ores are fold hinges, competency contrasts
between adjacent rock units, with vein arrays preferentially developed in the most
competent units, and chemically reactive host rocks which commonly coincide with the
most competent units (e.g., dolerite or BIF with high Fe/Fe+Ma+Ca ratios). Focusing of
supralithostatic ore fluids into dilatent zones in suitable structures and host rocks appears
to be the dominant control.
3. Structural studies in the terraces which host orogenic gold deposits normally indicate
their relative timing as towards the end of the deformational cycle (D3 or D4 in a D1-D4
cycle), after accretion or collision, and normally during uplift and exhumation of the
orogen. Reactivation of pre-existing, suitably-orientated fault and shear zones during a
change in far-field stress is a common scenario.
4. Robust geochronological studies indicate that most orogens evolved over about 100 to
200 m.y. periods, and that orogenic gold deposits formed diachronously over the
Page 162
terranes in the latter half of this evolutionary period, thus confirming the relative
structural timing. In Archaean greenstone terranes, such as those of the Yilgarn Block,
the orogenic gold deposits formed some 20-80 m.y. after the youngest major volcanic
event, but in younger, largely sedimentary, terranes, the time lag may be longer. In rare
instances, the orogenic gold deposits may be related to deformation events that post-date
cratonization of the host terrane by more than 1 b.y.
5. The late timing of the orogenic gold deposits has important implications for exploration
methodologies which are geology-based. It means that the structural geometries of the
gold deposits, hosting goldfields and enclosing terranes, as shown in geological maps
and sections, and three-dimensional models, where available, are essentially those which
were developed at the time of gold mineralization. This is pivotal to the use of geology-
based qualitative methodologies and more-quantitative computer methodologies for
exploration for this deposit class.
6. As expected from their dominant structural control and late timing, there are repetitive
predictable patterns to the siting of orogenic gold deposits. Deposits or groups of
deposits are commonly sited near structural heterogeneities within, or adjacent to, first-
or lower-order faults and shear zones, adjacent to contacts across which there are large
competency contrasts, in the pressure shadows of rigid granitoid plutons or batholiths, or
in fold-hinges, commonly those “locked-up” by pre-mineralization deformation. The
two largest greenstone-hosted goldfields, Kalgoorlie and Timmins, show a remarkably
similar structural geometry, attesting to the significance of this parameter in geology-
based exploration.
7. Computer-based stress mapping is one methodology which can utilize the dominant
structural control and late timing of orogenic gold deposits to quantify the expected
siting of dilational zones during late deformation of a gold-mineralized terrane. The
geological requirements are an accurate solid-geology map of uniform quality and
knowledge of the far-field stress at the time of gold mineralization. All rock units, faults
and contacts are assigned strength or stiffness parameters, and the stress modelling is
carried out in two-dimensions using a distinct element code to determine displacements
and movement vectors, as well as the distribution of minimum, maximum and mean
stresses. The stress map of the Kalgoorlie Terrace in the Yilgarn Block of Western
Australia, used as an example of the potential of stress mapping, clearly demonstrates
that low minimum-stress anomalies coincide with known goldfields at the 1:250,000
scale of the modelling. This supports a model in which the location of gold deposits is
strongly controlled by the distribution and orientation of first-order structures. There are
also additional anomalies that may represent as-yet undiscovered deposits. The next step
would be to perform two-dimensional stress mapping at a smaller scale on more-detailed
solid-geology maps in the areas (goldfields?) defined by the regional low minimum-
stress anomalies. Three-dimensional stress mapping could be performed if sufficient
drill data are available.
8. GIS-based prospectivity mapping is another computer-based methodology which relies

on current structural geometry, as shown on geological maps, representing that at the
time of mineralization. Thus, it is ideally suited to the search for orogenic gold deposits.
Using an empirical approach, it is possible to semi-quantitatively define the spatial
relationship between parameters in the map data and the gold deposits, quantify them,
and then integrate them into a prospectivity map. Again using the Kalgoorlie Terrace
1:250,000 scale map as an ideal example, the prospectivity analysis confirms expected
relationships, with the siting of orogenic gold deposits controlled by such factors as
proximity to, and orientation and curvature of, granitoid-greenstone contacts, proximity
to segments of the crustal-scale faults which strike in a preferred north-westerly
direction, proximity to certain lithological contacts which strike in a similar direction,
Page 163
proximity to anticlinal hinges or horst-like structures, and the presence of chemically-

reactive host rocks such as differentiated mafic (dolerite) sills. The prospectivity map,
produced from the interpretation of these parameters, closely conforms to the stress map
in defining the major goldfields, although there are differences in detail. In this map, the
most prospective category occupies less than 0.3% of the greenstone belts, yet hosts over
16% of known gold deposits with over 80% of known gold production from the region.
This demonstrates the potential of the methodology on a regional scale, particularly in
the targeting of the larger, most -desired, orogenic gold deposits in the analyzed terrace.
9. The success of both the qualitative structural geometry approach and the semi-
quantitative stress-mapping and GIS-based prospectivity mapping in regional targeting
for orogenic gold deposits, at least for the terranes so far researched, indicates that
quantitative analysis of geological map data is a potentially profitable line of research.
Stress mapping has an advantage in previously unexplored terranes as it relies wholly on
prediction of dilatent sites and requires no database of existing gold deposits such as is
required by the empirically-based prospectivity mapping. Artificial neural networks and
quantitative shape analysis are two additional methodologies currently being
investigated.
10. The ultimate deposit target map should be an integration of as many geologically-based
prospectivity guides as possible, such that coincident stress anomalies, high prospectivity
categories and, in the future, critical shape parameters can be compared to other, more
conventional, targeting parameters such as geophysical and geochemical anomalies. In
combination, these methodologies should be capable of pinpointing potentially
mineralized zones containing orogenic gold deposits in poorly exposed terraces with
complex regolith or thick superficial cover, the ever increasing locations of modern
global exploration.
We gratefully acknowledge colleagues in the Centre for Strategic Mineral Deposits at the University
of Western Australia, who have contributed to our collective knowledge of orogenic gold deposits
and the Yilgarn Block and to development of prospectivity mapping methodology. Thanks are due to
Terra Sancta for use of stress mapping software packages and permission to publish. We are grateful
for funding for gold research from many supportive Australian mining and exploration companies.
The paper has been improved by the reviews of Tim Bell and Peter Laznicka.
8.11 References
Alibone, A.H., Windh, J., Etheridge, M.A., Burton, D., Anderson, G., Edwards, P.W., Miller, A.,
Graves, C., Fanning, C.M., Wysoczanski, R., 1998. Timing relationships and structural controls
on the location of Au-Cu mineralization at the Boddington Gold Mine, Western Australia. Econ.
Geol. 93, 245-270.
Barnicoat, A.C., Fare, R.J., Groves, D.I., McNaughton, N., 1991. Synmetamorphic lode-gold
Bloem, E.J.M., Dalstra, H., Groves, D.I., Ridley, J.R., 1994. Metamorphic and structural setting of
amphibolite-hosted gold deposits between Southern Cross and Bullfinch, Southern Cross
Province, Yilgarn Block, Western Australia. Ore Geol. Rev. 9, 183-208.
Bloem, E.J.M., McNaughton, N. J., Groves, D.I., Ridley, J.R., 1995. An indirect lead isotope age
determination of gold mineralization at the Corinthia mine, Yilgarn Block, Western Australia.
Austr. J. Earth Sci. 42, 447-451.
Page 164
Boer, R.H., Meyer, F.M., Robb, L.J., Graney, J.R., Vennemann, T.W., 1993. The nature of gold
mineralization in the Sabie-Pilgrim’s Rest goldfield, eastern Transvaal, South Africa. Univ. of
Witwatersrand, Econ. Geol. Res. Unit, Inform. Circ. 262, 33 p.
Bohlke, J.K., Kistler, R.W., 1986. Rb-Sr, K-Ar, and stable isotope evidence for the ages and sources
of fluid components of gold-bearing quartz veins in the northern Sierra Nevada foothills
metamorphic belt, California. Econ. Geol. 81, 296-322.
Bonham-Carter, G.F., 1995. Geographic information systems for geoscientists: modelling with GIS.
Computer Methods in the Geosciences 13. Pergamon Press, New York.
Bonham-Carter, G.F., Agterberg, F.P., Wright, D.F., 1989. Weights of evidence modelling: a new
approach to mapping mineral deposits. In: Agterberg, F.P., Bonham-Carter, G.F., (Eds.) Statistical
Applications in Earth Sciences. Geol. Surv. Canada Paper 89-9, 171-183.
Bouchot, V., Gros, Y., Bonnemaison, M., 1989. Structural controls on the auriferous shear zones of
the Saint Yrieux district, Massif Central, France. Evidence from the Le Bourneix and Laurieras
gold deposits. Econ. Geol. 84, 1315-1327.
Brown, W.M., Gedeon, T.D., Groves, D.I., Barnes, R.G., in press. Artificial neural networks: a new
method for mineral prospectivity mapping. Austr. J. Earth Sci.
Cameron, E.M., 1988. Archean gold: relation to granulite formation and redox zoning in the crust.
Geology 16, 109-112
Campbell, I.H., Bitmead, R.I., Hill, R.I., Schiotte, L., Thom, A.M., 1993. Implications of zircon dates
for the age of granite rocks in the Eastern Goldfields Province. AGSO Record 1993/54, 47-48.
Campbell, S.D.G., Pitfield, P.E.J., 1994. Structural controls on gold mineralization in the Zimbabwe
craton-exploration guidelines. Zimbabwe Geol. Surv. Bull. 101, 270p.
Card, K.D., 1990. A review of the Superior Province of the Canadian Shield, a product of Archean
accretion. Precamb. Res. 48, 99-156.
Clark, M.E., Carmichael, D.M., Hodgson, C.J., Fu, M., 1989. Wall-rock alteration, Victory gold
mine, Kambalda, Western Australia: processes and P-T-XCO2 conditions of metasomatism.
Econ. Geol. Monogr. 6, 445-459.
Colvine, A.C., 1989. An empirical model for the formation of Archean gold deposits: products of
final cratonization of the Superior Province, Canada. Econ. Geol. Monogr. 6, 37-53.
Cox, S.F., Sun, S.S., Etheridge, M.A., Wall, V.J., Potter, T.F., 1995. Structural and geochemical
central Victoria, Australia. Econ. Geol. 90, 1722-1746.
Cox, S.F., Wall, V.J., Etheridge, M.A., Potter, T.F., 1991. Deformational and metamorphic processes
in the formation of mesothermal vein-hosted gold deposits - examples from the Lachlan fold belt
in central Victoria, Australia. Ore Geol. Rev. 6, 391-423.
D’Ercole, C. 1998. An integrated model for lead-zinc mineralisation on the southeastern Lennard
shelf based on geographic information systems (GIS). Univ. Western Australia (unpublished),
195pp.
Diamond, L., 1990. Fluid inclusion evidence for P-V-T-X evolution of hydrothermal solutions in
Late-Alpine gold-quartz veins at Brusson, Val-D’Ayas, Northwest Italian Alps. Am. J. Sci. 290,
912-958.
de Ronde, C.E.J., de Wit, M.J., 1994. Tectonic history of the Barberton greenstone belt, South
Africa—490 million years of Archean crustal evolution. Tectonics 13, 983-1005.
de Ronde, C.E.J., Spooner, E.T.C., de Wit, M.J., Bray, C.J., 1992. Shear zone-related, Au quartz vein
deposits in the Barberton greenstone belt, South Africa: field and petrographic characteristics,
fluid properties, and light stable isotope geochemistry. Econ Geol. 87, 366-402.
Elliott, N., 1997. Quantification of geological features stored within a geographic information system
(GIS). B.Sc. Hons., Univ. Western Australia (Unpublished), 50pp.
Forde, A., 1991. The late orogenic timing of gold mineralization in some slate belt gold deposits,
Victoria, Australia. Mineralium Deposita 26, 257-266.
Forde, A., Bell, T.H., 1994. Late structural control of mesothermal vein-hosted gold deposits in
central Victoria, Australia: mineralization mechanisms and exploration potential. Ore Geol. Rev.
9, 33-59.
Page 165
Gardoll, S.J., Yun, G.Y., Groves, D.I., Knox-Robinson, C.M., Elliott, N., in press. Developing the
tools for geological shape analysis: a regional to local scale analysis of the Kalgoorlie Terrane.
Austr. J. Earth Sci.
Gehrels, G.E., McClelland, W.C., Samson, S.D., Patchett, P.J., Orchard, M.J., 1992. Geology of the
western flank of the Coast Mountains between Cape Fanshaw and Taku Inlet, southeastern
Alaska. Tectonics 11, 567-585.
Goldfarb, R.J., Leach, D.L., Miller, M.L., Pickthorn, W.J., 1986. Geology, metamorphic setting and
genetic constraints of epigenetic lode-gold mineralization within the Cretaceous Valdez Group,
south-central Alaska. In: Keppie, J.D., Boyle, R.W., and Hanes, S.J. (Eds.), Turbidite-hosted
Gold Deposits. Geol. Assoc. Can. Spec. Pap. 32, 87-105.
Goldfarb, R.J., Leach, D.L., Pickthorn, W.J., Paterson, C.J., 1988, Origin of lode-gold deposits of the
Juneau gold deposit, southeast Alaska. Geology 16, 440-443.
Goldfarb, R.J., Leach, D.L., Roses, S.C., Landis, G.P., 1989. Fluid inclusion geochemistry of gold-
bearing quartz veins of the Juneau gold belt, southeastern Alaska: implications for ore genesis. In:
Keays, R.R., Ramsay, W.R.H., Groves, D.I. (Eds.), The Geology of Gold Deposits: The
Perspective in 1988. Econ. Geol. Monogr. 6, 363-375.
Goldfarb, R.J., Miller, L.D., Leach, D.L., Snee, L.W., 1997. Gold deposits in metamorphic rocks of
Alaska. In: Goldfarb, R.J., Miller, L.D. (Eds.), Mineral Deposits of Alaska. Econ. Geol. Monogr.
9, 151-190.
Goldfarb, R.J., Snee, L.W., Miller, L.D., Newberry, R.J., 1991. Rapid dewatering of the crust
Golding, S.D., Wilson, A.F., 1987. Oxygen and hydrogen isotope relationships in Archaean gold
deposits of the Eastern Goldfields Province, Western Australia: constraints on the source of
Archaean gold-bearing fluids. In, Ho, S.E., Groves, D.I. (Eds). Recent Advances in Understanding
Precambrian Gold Deposits. Geol. Dep. Univ. Extension, Univ. West. Austr. Publ. 11, 203-214.
Groves, D.I., 1993. The crustal continuum model for late Archaean lode-gold deposits of the Yilgarn
Block, Western Australia. Mineralium Deposita 28, 366-374.
Groves, D.I., Barley, M.E., 1994. Archean mineralization. In: Condie, K.C. (ed.) Archaean Crustal
Evolution. Elsevier, Amsterdam, 461-503.
Groves, D.I., Goldfarb, R.J., Gebre-Mariam, M., Hagemann, S.G., Robert, F., 1998. Orogenic gold
deposits: A proposed classification in the context of their crustal distribution and relationship to
other gold deposit types. Ore Geol. Rev. 13, 7-27.
Groves, D.I., Ojala, V.J., Holyland, P.W., 1997. Use of geometric parameters of greenstone belt in
conceptual exploration for orogenic lode-gold deposits. AGSO Record 1997/41, 103-108.
Groves, D.I., Ridley, J.R., Bloem, E.J.M., Gebre-Mariam, M., Hronsky J.M.A., Knight, J.T.,
McCuaig, T.C., McNaughton, N.J., Ojala, J., 1995. Lode-gold deposits of the Yilgarn Block:
products of late-Archaean crustal-scale over-pressured hydrothermal systems. In: Coward, R.P.,
Ries, A.C. (eds.) Early Precambrian Processes. Geol. Soc. London, Spec. Publ. 95, 155-172.
Ho, S.E., Bennett, J.M., Cassidy, K.F., Hronsky, J.M.A., Mikucki, E.J., Sang, J.H., 1990. Fluid
inclusion studies. In: Ho, S.E., Groves, D.I., Bennett, J.M. (Eds.), Gold deposits of the Archaean
Yilgarn Block, Western Australia: Nature, Genesis and Exploration Guides. Geol. Dep. Univ.
Extension, Univ. West. Austr. Publ. 20, 198-211.
Hodgson, C.J. 1989. The structure of shear-related, vein-type gold deposits: a review. Ore Geol.
Rev. 4, 231-273.
Holyland, P.W., 1987. Dynamic hydrothermal modelling of the Renison Tin Mine, Tasmania.,
Pacific Rim 87 Congress, 1987, 189-193.
Holyland, P.W., 1990a. The nature of the lithosphere: cracks and blocks?, Terra Sancta Newsletter:
Perth, Terra Sancta, unpaginated.
Holyland, P.W., 1990b. Simulation of the dynamics of Archaean deformation in the Yilgarn Block,
Western Australia., Abstract Volume. Third International Archaean Symposium, Perth,
Geoconferences W.A. Inc., Perth. Geoconferences W.A. Inc., Perth, 347-349.
Holyland, P.W., 1990c. Stress mapping in the Mt. Isa region: Mount Isa Inlier Geology Conference,
Melbourne, 76-77.
Holyland, P.W., 1990d. Targeting of epithermal ore deposits using stress mapping techniques.:
Pacific Rim 90 Congress, Melbourne, 337-341.
Page 166
Holyland, P.W., 1996. Controls on fracture permeability, and self-sealing in the Renison Tin Mine,
Tasmania., in Herbert, H.K., Ho, S.E. (Eds). Stable Isotopes and Fluid Processes in
Mineralization., 23. Geology Department (Key Centre) Univ. Extension, Univ. of West. Austr.
Publ. 23, 118-121.
Holyland, P.W., Ojala, V.J., 1997. Computer aided structural targeting in mineral exploration: two
and three-dimensional stress mapping. Austr. Earth Sci. 44, 421-432.
Holyland, P., Ridley, J. R., Vearncombe, J.R., 1993, Stress mapping technology (STMTM):
Geofluids '93: Contributions to an International Conference on Fluid Evolution, Migration and
Interaction in Rocks, Torquay, England, 272-275.
Jackson, S.L., Cruden, A.R., 1995. Formation of the Abitibi greenstone belt by arc-trench migration.
Geology 23, 471-474.
Jébrak, M., Holyland, P.W., Carrier, A., Angelier, J., 1995. Modelisation des gîtes d'or de district de
Rouyn-Noranda (Abitibi): Des mesures structurales aux cartes paleobarometriques previsionelles.:
Cami '95: 3rd Canadian Conference on Computer Applications in the Mineral Industry, Montreal,
189-193.
Kent, A.J.R., Cassidy, K.F., Fanning, C.M.F., 1996. Gold mineralization synchronous with the final
stages of cratonization, Yilgarn Craton, Western Australia: Evidence from Sm-Nd and U-Pb ages
of crosscutting (post-gold) dykes. Geology 24, 879-882.
Kent, A.J.R., Hagemann, S.G., 1996. Constraints on the timing of lode-gold mineralization in the
Wiluna greenstone belt, Yilgarn Craton, Western Australia. Austr. J. Earth Sci. 43, 573-588.
Kent, A.J.R., McDougall, I., 1995. 40Ar/39Ar and U-Pb constraints on the timing of gold
mineralization in the Kalgoorlie Goldfield, Western Australia. Econ. Geol. 90, 845-859.
Kerrich, R., 1987. The stable isotope geochemistry of Au-Ag vein deposits in metamorphic rocks.
In: Kyser, T.K. (Ed.), Min. Assoc. Can. Short Course 13, 287-336.
Kerrich, R., 1989. Geochemical evidence on the sources of fluids and solutes for shear zone hosted
mesothermal Au deposits. In: Bursnall, J.T. (Ed.), Mineralization and Shear Zones. Geol. Assoc.
Can. Short Course Notes 6, 129-197.
Kerrich, R., Cassidy, K.F., 1994. Temporal relationships of lode-gold mineralization to accretion,
magmatism, metamorphism and deformation - Archaean to present: a review. Ore Geol. Rev. 9,
263-310.
Kerrich, R., Fryer, B.J., 1979. Archean precious-metal hydrothermal systems, Dome Mine, Abitibi
greenstone belt. II. REE and oxygen isotope relations. Can. J. Earth Sci. 16, 440-458.
Kerrich, R., Fryer, B.J., 1981. The separation of rare elements from abundant base metals in Archean
lode-gold deposits: implications of low water/rock source regions. Econ. Geol. 76, 160-166.
Khin, Z., Huston, D.L., Large, R.R., Mernagh, T., Hoffman, C.F., 1994. Microthermometry and
geochemistry of fluid inclusions from the Tennant-Creek gold-copper deposits : Implications for
ore deposition and exploration. Mineralium Deposita 29, 288-300.
Knight, J.T., Groves, D.I., Ridley, J.R., 1993. District-scale structural and metamorphic controls on
Archaean lode-gold mineralization in the amphibolite-facies Coolgardie Goldfield, Western
Australia. Mineralium Deposita 28, 436-456.
Knight, J.T., Ridley, J.R., Groves, D.I., McCall, C., 1996. Syn-peak metamorphic gold
mineralization in the amphibolite-facies, gabbro-hosted Three Mile Hill deposits, Coolgardie
Goldfield: a high temperature analogue of mesothermal gabbro-hosted gold deposits. Trans.
Instn. Min. Metall. (London) 105, B175-199.
Knox-Robinson, C.M., 1994. Archaean lode-gold mineralisation potential of portions of the Yilgarn
block, Western Australia: development and implementation of methodology for the creation of
regional-scale prospectivity maps using conventional geological map data and a geographic
information system (GIS). PhD Thesis (unpublished). 178pp.
Knox-Robinson, C.M., in press. Vectorial fuzzy logic: a new and novel technique for enhanced
mineral prospectivity mapping, with references to the orogenic gold mineralisation potential of the
Kalgoorlie Terrane, Western Australia. Austr. J. Earth Sci.
Knox-Robinson, C.M., Groves, D.I., 1997. Gold prospectivity mapping using a geographic
information system (GIS) with examples from the Yilgarn block of Western Australia. Chronique
de la Recherche Miniere 529, 127-138.
Page 167
Knox-Robinson, C.M., Groves, D.I., Robinson, D.C., Wheatley, M.R., 1996. Improved resource
evaluation using geographic information systems (GIS): a pilot study. AMIRA Project P383 Final
Report,196pp
Knox-Robinson, C.M., Wyborn, L.A.I., 1997. Towards a holistic exploration strategy: using
geographic information systems as a tool to enhance exploration. Austr. J. Earth Sci. 44, 453-463.
Kontak, D.J., Kerrich, R., 1997. An isotopic (C, O, Sr) study of vein gold deposits in the Meguma
terrane, Nova Scotia: implication for source reservoirs. Econ. Geol. 92, 161-180.
Kontak, D.J., Smith, P.F., Kerrich, R., Williams, P.F., 1990. Integrated model for Meguma Group
lode-gold deposits, Nova Scotia, Canada. Geology 18, 238-242.
Le Guen, M., Lescuyer, J.L., Marcoux, E., 1992. Lead-isotope evidence for a Hercynian origin of the
Salsigne gold deposit (Southern Massif Central France). Mineralium Deposita 27, 129-136.
LeAnderson, P.J., Yoldash, M., Johnson, P., Offield, T.W., 1995. Structure, vein paragenesis, and
alteration in the Al Wajh gold district, Saudi Arabia. Econ. Geol. 90, 2262-2273.
LeLacheur, E.A., 1991. Brittle-fault hosted gold mineralization in the Fairbanks district, Alaska.
M.S. thesis, Univ. Alaska Fairbanks (unpublished), 167 pp.
McCuaig, T.C., Kerrich, R., 1998. P-T-t-deformation-fluid characteristics of lode-gold deposits:
evidence from alteration systematics. Ore Geol. Rev. 12, 381-453.
McCuaig, T.C., Kerrich, R., Groves, D.I., Archer, N., 1993. The nature and dimensions of regional
and local gold-related hydrothermal alteration in theoleiitic metabasalts in the Norseman
goldfields: the missing link in a crustal continuum of gold deposits? Mineralium Deposita 28,
420-435.
Milesi, J.P., Ledru, P., Feybesse, J.L., Dommanget, A., Marcoux, E., 1992. Early Proterozoic ore
deposits and tectonics of the Birimian orogenic belt, West Africa. Precamb. Res. 58, 305-344.
Miller, L.D., Barton, C.C., Fredericksen, R.S., Dressler, J.R., 1992. Structural evolution of the
Alaska-Juneau gold deposit, southeastern Alaska. Can. J. Earth Sci. 29, 865-878.
Miller, L.D., Goldfarb, R.J., Gehrels, G.E., Snee, L.W., 1994. Genetic links among fluid cycling,
vein formation, regional deformation, and plutonism in the Juneau gold belt, southeastern Alaska.
Geology 22, 203-206.
Miller, L.D., Goldfarb, R.J., Snee, L.W., Gent, C.A., Kirkham, R.A., 1995. Structural geology, age,
and mechanisms of gold vein formation at the Kensington and Jualin deposits, Berners Bay
district, southeast Alaska. Econ. Geol. 90, 343-368.
Morasse, S., Wasteneys, H.A., Cormier, M., Helmstaedt, H., Mason, R., 1995. A pre-2686 Ma
intrusion-related gold deposit at the Kiena mine, Val d’Or, Quebec, Southern Abitibi
Subprovince. Econ. Geol. 90, 1310-1321.
Mueller, A.G., Campbell, I.H., Schiotte, L., Sevingny, J.H., Layer, P.W., 1996. Constraints on the
age of granitoid emplacement, metamorphism, gold mineralization, and subsequent cooling of the
Archaean greenstone terrane at Big Bell, Western Australia. Econ. Geol. 91, 896-915.
Murphy, P.J., Roberts, S., 1997. Evolution of a metamorphic fluid and its role in lode- gold
mineralization in the Central Iberian Zone. Mineralium Deposita 32, 459-474.
Myers, J.S., 1993. Precambrian history of the West Australian Craton and adjacent orogens. Ann.
Rev. Earth, Planet. Sci. 21, 453-485.
Nesbitt, B., 1991. Phanerozoic gold deposits in tectonically active continental margins. In: Foster,
R.P. (Ed.) Gold Metallogeny and Exploration. Blackie, Glasgow, 104-132.
Nesbitt, B.E., Muehlenbachs, K., Murowchick, J.B., 1989. Genetic implications of stable isotope
characteristics of mesothermal Au deposits and related Sb and Hg deposits in the Canadian
Cordillera. Econ. Geol. 84, 1489-1506.
Newton, P.G.N., Ridley, J.R., Groves, D.I., Khosrowshahi, S., Smith, B., 1997. Integration of
directional variography and structural geology: an example of the Santa-Craze BIF-hosted Au
deposit, near Kalgoorlie, Western Australia. Chronique de la Recherche Minière 529, 105-126.
Nguyen, P.T., Cox, S.F., Harris, L.B., Powell, C.M., 1998. Fault-valve behaviour in optimally
oriented shear zones. An example at the Revenge gold mine, Kambalda, Western Australia. J.
Struc. Geol. 20, 1625-1640.
Nokleberg, W.J., Plafker, G., Lull, J.S., Wallace, W.K., Winkler, G.R., 1989. Structural analysis of
the southern Peninsular, southern Wrangellia, and northern Chugach terranes along the Trans-
Alaska Crustal Transect, northern Chugach Mountains, Alaska. J. Geophys. Res. 94, 4297-4320.
Page 168
Nurmi, P.A., Sorjonen-Ward, P., Damsten, M., 1993. Geological setting, characteristics and
exploration history of mesothermal gold occurrences in the late Archaean Hattu schist belt,
Ilomantsi, eastern Finland. Geol. Surv. Finland, Spec. Pap. 17, 193-231.
Oberthur, T.U., Mumm, A.S., Vetter, U., Simon, K., Amanor, J.A., 1996. Gold mineralization in the
Ashanti belt of Ghana: genetic constraints of the stable isotope geochemistry. Econ. Geol. 91,
289-301.
Oberthur, T., Vetter, U., David, D.W., Amanor, J.A., 1998. Age constraints on gold mineralization
and Paleoproterozoic crustal evolution in the Ashanti belt of southern Ghana. Precamb. Res. 89,
129-143.
Ojala, J.V., Ridley, J.R., Groves, D.I., Hall, G.C., 1993. The Granny Smith gold deposit: the role of
heterogeneous stress distribution at an irregular granitoid contact in a greenschist facies terrane.
Mineralium Deposita 28, 409-419.
Perring, C.S., Groves, D.I., Ho, S.E., 1987. Constraints on the source of auriferous fluids for Archean
gold deposits. In, Ho, S.E., Groves, D.I. (Eds). Recent Advances in Understanding Precambrian
Gold Deposits. Geol. Dep. Univ. Extension, Univ. Western Australia Publ. 11 203-214.
Partington, G.A., McNaughton, N.J., 1997. Controls on mineralization in the Howley District,
Northern Territory: a link between granite intrusion and gold mineralization. Chronique de la
Recherche Miniere 529, 25-44.
Phillips, G.N., Groves, D.I., 1983. The nature of Archaean gold-bearing fluids as deduced from gold
deposits of Western Australia. Geol. Soc. Aust. J. 30, 25-39.
Phillips, G.N., Groves, D.I., Kerrich, R., 1996. Factors in the formation of the giant Kalgoorlie gold
deposit. Ore Geol. Rev. 10, 295-317.
Qiu, Y., McNaughton, N.J., Groves, D.I., Dahlstra, H.J., 1997. SHRIMP U-Pb in zircon and lead-
isotope constraints on the timing and source of an Archaean granite-hosted lode-gold deposit at
Griffin’s Find, Yilgarn Craton, Western Australia. Chronique de la Recherche Miniere 529, 91-
104.
Reddy, R.K., 1987. A generalised model for evaluating area-potential in a mineral exploration
program. A case study of silver exploration in parts of Idaho and Montana. PhD thesis, Univ.
Georgia (unpublished), 119pp.
Ridley, J.R., 1993. The relations between mean rock stress and fluid flow in the crust: with reference
to vein- and lode-style deposits. Ore Geol. Rev. 8, 23-37.
Robert, F., Brown, A.C., 1986. Archaean gold-quartz veins at the Sigma mine, Abitibi greenstone
belt, Quebec. Part I. Geological relations and formation of the vein systems. Part II. Vein
paragenesis and hydrothermal alteration. Econ. Geol. 81, 578-616.
Robert, F., Kelly, W.C., 1987. Ore-forming fluids in Archean gold-bearing quartz veins at the Sigma
mine, Abitibi greenstone belt, Quebec, Canada. Econ. Geol. 82, 1464-1482.
Robinson, D.C., 1990. Gold resource potential of the Southern Cross greenstone belt, Western
Australia: mineral resource assessment using deposit models and a geographic information
system. B.Sc. Hons., Univ Western Australia (Unpublished), 84pp.
Roedder, E., 1984. Fluid inclusion evidence bearing on the environments of gold deposition. In:
Foster, R.P. (Ed.), Gold ‘82. Geol. Soc. Zimbabwe, Spec. Publ. No. 1, 129-163.
Roth, E., 1992. The nature and genesis of Archaean porphyry-style Cu-Au-Mo mineralization at the
Boddington Gold Mine, Western Australia. PhD Thesis, University of Western Australia, Perth
(unpublished) 126p.
Rowins, S.M., Groves, D.I., McNaughton, N.J., Palmer, M.R., Eldridge, C.S., 1997. A
reinterpretation of the role of granitoids in the genesis of Neoproterozoic gold mineralization in
the Telfer Dome, Western Australia. Econ. Geol. 92, 133-160.
Ryan, R.J., Smith, P.K., 1998. A review of the mesothermal gold deposits of the Meguma Group,
Nova Scotia, Canada. Ore. Geol. Rev. 13, 153-183
Sibson, R.H., 1985. A note on fault reactivation. J. Struct. Geol. 7, 751-754.
Sibson, R.H., 1990. Faulting and fluid flow. In: B.E. Nesbitt (Ed.), Fluids in Tectonically Active
Regimes of the Continental Crust, Mineral. Assoc. Can. Short Course 18, 93-132.
Sibson, R.H., Robert, F., Poulsen, K.H., 1988. High angle reverse faults, fluid-pressure cycling, and
mesothermal gold-quartz deposits. Geology 16, 551-555.
Page 169
Sillitoe, R.H., Thompson, J.F.H., 1998. Intrusion-related vein gold deposits: Types, tectono-
magmatic settings and difficulties of distinction from orogenic gold deposits. Resource Geol. 48,
237-250.
Smith, A.B., Gardoll, S.J., 1997. Structural analysis in mineral exploration using a Geographic
Information Systems-adapted stereographic-projection plotting program. Austr. J. Earth Sci. 44,
445-452.
Smith, T.J., Cloke, P.L., Kesler, S.E., 1984. Geochemistry of fluid inclusions from the McIntyre-
Hollinger gold deposit, Timmins, Ontario, Canada. Econ. Geol. 79, 1265-1285.
Spooner, E.T.C., 1993. Magmatic sulphide/volatile interaction as a mechanism for producing
chalcophile element enriched, Archean Au-quartz, epithermal Au-Ag and Au skarn hydrothermal
ore fluids. Ore Geol. Rev. 7, 359-380.
Stein, H.J., Morgan, J.W., Markey, R.J., Hannah, J.L., 1998. An introduction to Re-Os. What’s in it
for the mineral industry? SEG Newsl. 32, 1, 8-15.
Stowell, H.H., Lesher, C.M., Green, N.L., Peng, S., Guthrie, G.M., Krishna, S.A., 1996.
Metamorphism and gold mineralization in the Blue Ridge, southernmost Appalachians. Econ.
Geol. 91, 1115-1144.
Stuwe, K., Will, T.M., Zhou, S., 1993. On the timing relationship between fluid production and
metamorphism in metamorphic piles: some implications for the origin of post-metamorphic gold
mineralization. Earth Planet. Sci. Lett. 114, 417-430.
Swager, C., 1989. Structure of Kalgoorlie greenstones – Regional deformation history and
implications for the structural setting of the Golden Mile gold deposits. Geol. Surv. West. Austr.
Rep. 25, 59-84p.
Swager, C., Griffin, T. 1990. Geology of the Archaean Kalgoorlie Terrane. 1:250 000 scale
geological map. Geological Survey of Western Australia.
Thomas, D.E., 1953. Mineralization and its relationship to the geologic structure of Victoria. In:
Edwards, A.B. (Ed.), Geology of Australian Ore Deposits. 5th Empire Min. Metall. Congr. 971-
985.
Thompson, J.F.H., Sillitoe, R.H., Baker, T., Lang, J.R., Mortensen, J.K., 1999. Intrusion-related gold
deposits associated with tungsten-tin provinces. Mineralium Deposita 34, 323-334.
Tobisch, O.T., Paterson, S.R., Saleeby, J.B., Geary, E.E., 1989. Nature and timing of deformation in
the Foothills terrane, central Sierra Nevada, California. Its bearing on orogenesis. Geol. Soc. Am.
Bull. 101, 401-413.
Tyler, R., Tyler, N., 1996. Stratigraphic and structural controls on gold mineralization in the Pilgrim’s
Rest goldfield, eastern Transvaal, South Africa. Precamb. Res. 79, 141-169.
Vearncombe, J.R., Barley, M.E., Eisenlohr, B.N., Groves, D.I., Houston, S.M., Skwarnecki, M.S.,
Grigson, M.W., Partington, G.A., 1989. Structural controls on mesothermal gold mineralization:
examples from the Archean terranes of Southern Africa and Western Australia. Econ. Geol.
Monogr. 6, 124-134.
Vearncombe, J.R., Holyland, P.W., 1995. Rheology, stress-mapping and hydrothermal
mineralisation-the example of structurally-controlled gold-antimony deposits in South Africa.
South African Geology 98, 415-429.
Wang, L.G., McNaughton, N.J., Groves, D.I., 1993. An overview of the relationship between
granitoid intrusions and gold mineralization in the Archaean Murchison Province, Western
Australia. Mineralium Deposita 28, 482-494.
Wilkins, C., 1993. A post-deformational, post-peak metamorphic timing for mineralization at the
Archean Big Bell gold deposit, Western Australia. Ore Geol. Rev. 7, 439-484.
Witt, W.K., 1993. Gold mineralization in the Menzies-Kambalda region, Eastern Goldfields,
Western Australia. Geol. Surv. West. Austr. Rept. 39, 165p.
Wood, P.C., Burrows, D.R., Thomas, A.V., Spooner, E.T.C., 1986. The Hollinger-McIntyre Au-
quartz vein system, Timmins, Ontario, Canada: geologic characteristics, fluid properties, and light
stable isotope geochemistry. In: Macdonald, A.J. (Ed.), Proceedings of Gold ‘86: An
International Symposium on the Geology of Gold, Konsult International Inc., Ontario, 56-80.
Wyborn, L.A.I., Gallagher, R., Jagodzinski, E.A., 1994. A conceptual approach to metallogenic
modelling using GIS: examples from the Pine Creek inlier. In: Proceedings of a Symposium on
Australian Research in Ore Genesis. Australian Mineral Foundation, 15.1-15.5.
Page 170
Yardley, B.W.D., 1997. The evolution of fluids through the metamorphic cycle. In: Jamtveit, B.,
Yardley, B.W.D. (Eds), Fluid Flow and Transport in Rocks, Chapman & Hill, London, 99-121.
Yardley, B.W.D., Banks, D.A., Bottrell, S.H., Diamond, L.W., 1993. Post-metamorphic gold-quartz
veins from N.W. Italy: The composition and origin of the ore fluid. Min. Mag. 57, 407-422.
Yeats, C.J., McNaughton, N.J., 1997. Significance of SHRIMP II U-Pb geochronology on lode-gold
deposits of the Yilgarn Craton. AGSO Record 1997/41, 125-130.
Yeats, C.J., McNaughton, N.J., Groves, D.I., 1996. SHRIMP U-Pb geochronological constraints on
Archaean volcanic-hosted massive sulphide and lode-gold mineralization at Mount Gibson,
Yilgarn Craton, Western Australia. Econ. Geol., 41, 1354-1371.
Yeats, C.J., McNaughton, N.J., Ruettger, D., Bateman, R., Groves, D.I., Harris, J.L., Kohler, E., in
press. Evidence for diachronous Archaean lode-gold mineralization in the Yilgarn Craton,
Western Australia: A SHRIMP U-Pb study of intrusive rocks. Econ. Geol. in press.
Yun, G.Y., Groves, D.I., Knox-Robinson, C.M., Gardoll, S.J., 1998. GIS-based methodology for
prospectivity analysis of orogenic lode gold deposits: a preliminary study of the Kalgoorlie terrane
as an example. In: Zhou, Q., Li, Z., Lin, H., Shi W., (Eds.),Proceedings of Geofinformatics ’98
Conference: Spatial Information Technology Towards 2000 and Beyond., Beijing, 288-298.
Zadeh, L., 1973. Outline of a new approach to the analysis of complex systems and decision
processes. IEEE Transactions on Systems, Man and Cybernetics, 3, 28-44.
Page 171
9. Computer Aided Structural Targeting in

Mineral Exploration: Two and Three-
Dimensional Stress Mapping.
P.W. HOLYLAND a and V.J. OJALA b
a Terra Sancta Research, 48 Peoples Avenue, Gooseberry Hill WA 6076,

Australia
b Key Centre for Strategic Mineral Deposits, Department of Geology and
Geophysics, the University of Western Australia, Nedland WA 6907, Australia
Abstract
This paper examines a geomechanical approach to predict the location of structurally-controlled

hydrothermal Mississippi-type Pb-Zn mineralisation (MVT) and Archaean gold mineralisation using
the Stress Mapping Technique. Two-dimensional modelling was using the finite difference/distinct
element code in the UDEC program and three-dimensional using the indirect boundary element
method in the MAP3D program. Two-dimensional, regional scale, stress-modelling of the Lennard
Shelf, Western Australia shows that most of the known MVT deposits and prospects are located
within or near the modelled minimum principal-stress anomalies using the gross geology of the
Lennard Shelf and an ENE-WSW extension direction.
Three-dimensional stress modelling was used to model stress fields around the eastern margin of the
Granny Smith Granodiorite under E-W compression at the Granny Smith mine, Laverton, Western
Australia. At a deposit scale, the patterns of the simulated minimum principal-stress correlate well
with the known areas of gold mineralisation near the contact between the Granny Smith Granodiorite
and sedimentary rocks.
9.1. Introduction
Many types of hydrothermal ore deposits are formed in structurally controlled sites in permeable
fracture-systems. Typically, mineralisation occurs in discrete segments of individual structures and,
within the mineral deposits, some parts of the host structures are better mineralised than others.
Epigenetic mineral deposits, are related, in general, to structurally focussed fluid flow during active
deformation (eg. Hodgson, 1989). Hronsky et al. (1990) reviewed the common structural controls on
localisation of mineralisation and oreshoots:
i) the intersection of host structure with a particular lithological unit (eg. BIF hosted gold
deposits),
ii) intersection of two synmineralisation structures,
iii) dilational jogs, divergent bends in faults or en echelon fault segmentation,
iv) fold hinge zones,
Page 173
v) flexures in the host structure, with axes oblique to, and commonly at a high angle to, the
movement direction, and
vi) zones that plunge subparallel to the stretching lineation, but in which the specific
controls on their location are unclear.
If only one control applies to mineralisation, it may be possible to predict the location of the next
mineralised zone in a particular area or structure. However, commonly several controls interact, and
the complex three-dimensional geometry of the controlling structures makes interpretation very
difficult. In addition to geometry, competency contrasts and rheological properties of the host
lithologies may have a significant influence. Therefore, even a purely kinematic approach to
structural interpretation of a geological map is commonly based more on interpreter’s intuition than
true geometrical relationships. Several computer codes have been developed for geomechanical
purposes to model structural behaviour of complex rock masses and this paper examines the
applications of two such codes in computer simulations to achieve more consistent structural
interpretations and for locating structurally controlled Mississippi Valley-type Pb-Zn mineralisation
(MVT) and Archaean mesothermal lode-gold deposits. However, the modelling is applicable to
geometrically similar deposits of other mineral commodities. This paper demonstrates that a
geomechanical approach using two and three-dimensional computer simulations (eg. Holyland
1990a, Oliver et al., 1990) is a practical way in which to attempt to predict the locations of the
dilational sites which result from deformation of a complex rock and fracture geometry. Regional
two-dimensional stress modelling is applied to a part of the Lennard Shelf in Canning Basin, Western
Australia which hosts significant structurally controlled MVT mineralisation (eg. Eisenlohr et al.,
1994, Vearncombe et al., 1995). A deposit scale three-dimensional modelling example is the late
Archaean gold mineralisation at the Granny Smith mine in the Eastern Goldfields Province, Western
Australia.
9.2 Methods And Theoretical Considerations
9.2.1 Fluid focussing
Hydrothermal ore deposits are characterised by channelised fluid flow and high fluid/rock ratios.
Under low to medium-grade metamorphic conditions, fluid pressures are buffered close to lithostatic
pressures and fluid flow is generally upward directed (Etheridge et al., 1984). Focussing of the
upward fluid flow into a discrete channelway, as required to form an ore deposit, is due to lateral
variations in hydraulic head. These lateral gradients may be induced structurally by either variations
in fracture permeability of active fault zones, or by variations in mean rock stress. Mineralised
extension veins indicative of supralithostatic fluid pressures are commonly present in mesothermal
gold systems, and these are compatible with fluid focussing into zones of low mean rock stress, where
relative fluid pressures are higher than the local rock pressure but absolute fluid pressures are lower
than in surrounding rock (Ridley, 1993). Sites of low mean stress can, therefore, be simultaneously
sites of fluid focussing and of low effective mean-stress.
At higher crustal levels ambient fluid pressures are closer to hydrostatic pressures but fluid pressures
in fluid channels might be lithostatic and both relative and absolute fluid pressures higher than in the
surrounding rock. Consequently, fluid flow is strongly controlled by permeability. Although high
permeability pathways can be lithologically determined, structurally controlled ore deposits are more
common in crystalline rocks and, in general, they have formed in reactivated fault systems (eg.
Phillips, 1972; Sibson et al., 1975; Sibson et al., 1988). A failure of a pre-existing weakness, or intact
rock, can be initiated by an increase of differential stress, by an increase in the pore fluid pressure, or
by combination of both. In most cases, stress changes lead to decrease of mean stress or, more
commonly, decrease of minimum principal stress.
A model in which fluid focussing in the crust is due to variations in mean stress, or in which failure of
pre-existing weaknesses and, therefore, enhanced fracture permeability, is due to lowered mean or
Page 174
minimum stress, allows for the wide variety of structural settings of mineral deposits. Thus, a
technique that measures variations in rock stress has the potential to generate viable exploration
targets.
9.2.2 Stress-fluid pressure relationship at fracture initiation
Three different modes of brittle failure for homogeneous, isotropic, intact rock are shown in relation
to loading stress fields on the Mohr diagram in Figure 1 under condition of high relative fluid
pressure (Sibson, 1989). The type of failure depends on the magnitude of the differential stress
relative to the tensile strength of the rock. Depending on the stress state and the shape of the failure
envelope, failure may be by:
i) extensional hydrofracture perpendicular to σ3,
ii) shear failure along conjugate zones about 30° to σ1, or
iii) extensional shear failure as conjugate zones at, or less than, 30° to σ1 (Sibson, 1989).
Figure 2 shows the effect of a plane of weakness for instance on failure using the Mohr-Coulomb
failure criterion. The failure envelopes are constructed assuming any orientation of the plane of
weakness. A stress regime in which the Mohr circle overlaps the failure envelope for planes of
weakness, but does not touch the envelope for intact rock, will fail along that plane of weakness.
Analysis of the likely mode of failure requires consideration of the orientation of the plane relative to
the imposed stress. As for the case of intact rock, failure may be in one of three modes; extension
hydrofracturing, shear failure, or extensional shear failure.
Generally, there will be a range of orientations of planes for which failure will occur (Fig. 2) before
fracturing or shearing of adjacent intact rock. Pure extensional hydrofracturing in a plane
perpendicular to σ3 will occur at low differential stress states in which the Mohr circle touches the
failure envelope at the normal stress axis (Fig. 1). In this case, the fluid pressure exceeds σ3 plus the
tensile strength of the fracture. With increasing fluid pressure or differential stress, the range of
orientations of planes which can fail increases. As long as the intersection between the Mohr circle
and the failure envelope is in the tensional field, failure will involve hydrofracturing and shearing.
During deformation, failure along the plane, or through intact rock, can be induced by:
i) increasing the fluid pressure, and hence reducing the effective stresses and moving the
Mohr circle to the left (Fig. 2a),
ii) increasing σ1, and hence increasing mean and differential stresses (Fig. 2b),
iii) increasing the differential stress at constant mean stress (σ = σ1+ σ3/2) (Fig. 2c), and or
iv) decreasing 3, and hence decreasing mean and increasing differential stresses (Fig. 2d).
The diagrams in Figure 2 demonstrate that a change in stress state from stable to unstable results in a
decrease of mean stress in cases a and d, an increase in case b, and a constant mean stress in case c.
However, the minimum principal stress (σ3) is constant only in case b and decreases in all other
cases. Therefore, it is assumed in the analysis that variations in σ3 give a better indication of
proximity to failure than the mean stress.
Page 175
Figure 1. Modes of failure in relation to differential stress in homogeneous, isotropic rock. Mohr
diagrams illustrate the general failure envelope for intact rock and the stress conditions for the three
modes of failure. Modified from (Sibson 1989). a Extensional fracturing (hydraulic fracturing) b
Extensional shear (combination of hydraulic fracturing and shear) c Shear fracturing.
9.3 Stress Mapping
Stress mapping examines the variation in strain and stress through an inhomogeneous terrain on
imposition of a regional stress field. When stressed, an inhomogeneous material develops an
inhomogeneous stress field whose components vary with rheological properties and geometry. The
modelling is considering only elastic and elasto-plastic stresses and strains, and, for instance,
deformation dilatancy due to viscous strains is not considered.
Stress mapping as a basis for prediction of hydrothermal fluid flow is based on the following
reasonable assumptions:
1. Low minimum principal (σ3) stress indicates proximity to failure and therefore possible
deformation enhanced permeability (this is more important in modelling of high crustal level
deformation).
2. At depths more than a few kilometres fluid pressure is buffered to be close to lithostatic
pressure and the control on fluid pressure is mean stress.
3. Variations in mean stress will be followed by variations in fluid pressure.
4. Fluid flow is both upwards and towards zones of low mean stress.
Combined with the knowledge that structurally controlled mineralisation is commonly late in the
tectonic history of a terrain and that typical Archaean lode-gold deposits and MVT deposits show
evidence of fluid overpressuring (eg. Archaean: Groves et al. 1995; MVT: Eisenlohr et al. 1994), this
enables stress analysis of two-dimensional map patterns of rock units and faults to predict those zones
Page 176
of low mean ((σ m=( σ1+ σ2+ σ3)/3) or the minimum principal (σ3) stress, provided that the
orientation of the far field stress can be determined or a reasonable orientation assumed.
The most severe limiting assumption of two-dimensional stress modelling is that the plane of a map
does not accurately reflect the stress pattern in an area with complex three-dimensional geometry
(Holyland et al., 1993). In many terrains, such as the Yilgarn Block, this may not be a critical
restriction, since the vast majority of the observed structures are upright and the majority of gold
deposits are hosted in steep structures, although significant exceptions do occur (Hronsky et al., 1990;
Libby et al., 1990). The accuracy of the three-dimensional geological interpretation becomes
especially important in deposit scale modelling where the extent of interest of the vertical dimension
is similar to horizontal dimensions.
Figure 2. Schematic Mohr diagrams illustrating possible

changes in the stress state which can lead to failure of a
plane of weakness. a Increasing fluid pressure, for the stress
state 2, tensile failure will occur along fractures
perpendicular to σ3, and shear failure along a wide range
of orientations of pre-existing planes of weakness. b
Increasing σ 1leads to increasing differential stress and
shear failure will occur when Mohr circle touches the
failure envelope. c Increasing differential stress with
constant mean stress leads to shear failure. d Decreasing σ
3 leads to increasing differential stress and shear failure.
9.3.1 Computer programs
Two-dimensional modelling is by a distinct element code (UDEC program) and the method has three
distinguishing features which make it well suited for discontinuum modelling (Holyland, 1990a;
Holyland, 1990b). These are:
Page 177
1. The method simulates an assemblage of blocks which interact through corner and edge
contacts.
2. Discontinuities are regarded as boundary interactions between these blocks; discontinuity

(fault) behaviour is prescribed for these interactions.
3. The method utilises an explicit timestepping (dynamic) algorithm which does not limit
displacements or rotations, and general non-constitutive behaviour for both the matrix and
discontinuities.
Modelling can be done at any scale and in addition to minimum, maximum and mean stresses, total
displacement and fault displacements are computed. The total displacement is the amount of
movement experienced by the rock blocks, and their movement directions.
MAP3D® version 1.29 (Mining Analysis Program in 3-Dimensions, Copyright Mine Modelling
Limited, 1993) is used in the three-dimensional stress analysis to simulate rockmass response under
the imposed external stress. The rockmass can include multiple zones with different moduli, and fault
slip as well as crack opening can be simulated.
The formulation of the code is based on the Indirect Boundary Element Method of Banarjee and
Butterfield (1981), and incorporates simultaneous use of both fictitious force and displacement
discontinuity elements. The elastic rockmass may contain multiple non-homogeneous regions, and
can be intersected by multiple fault planes and joint sets. Results of the simulations are presented on
user specified grids which may slice through the three-dimensional model at any desired location and
orientation.
9.3.2 Input for stress mapping
Stress mapping requires the following input data:
1. An interpreted, accurate geological map or three-dimensional geological model. Geological

map data are converted to a solid geology interpretation which provides continuous
lithological and structural information (eg. Fig. 3a). The study area is treated as a mosaic of
polygonal blocks (rock units) and joins (faults and shear zones). When external stress is
applied to this system the blocks are juggled and internally deform until equilibrium is
attained. This juggling and internal deformation results in areas of lower stress (ie. dilation
zones) which represent favoured areas for mineralisation and therefore targets in exploration.
2. The production of a stress map requires estimates of the magnitudes and orientations of the
far-field horizontal stresses.
3. Rock deformation properties including strength and moduli, and friction angle and stiffness
for fault deformation are required for modelling.
9.3.3 Example of two-dimensional stress mapping: Lennard Shelf
A 1:250 000 solid geology interpretation of an area including MVT deposits at Cadjebut, Blendevale
and Twelve mile Bore produced by Dörling (1995) as a part of his PhD study was used in the two-
dimensional simulation (Fig. 3a is a simplified scale reduction of this map). Structures were divided
according to their size as first or second order. First order structures include regional faults such as the
Pinnacles Fault and have strike lengths over 10 km; second order structures are generally under 10
km in strike length. First order structures were modelled as having lowest stiffness. Rock types
include conglomerates, shale, platform limestones, and basement granite. The basement granites were
Page 178
treated as the most competent rock type, followed by the platform limestones, shale and finally
unconsolidated conglomerates (for commercial reasons the exact parameters are not published here).
The maximum principal stress (σ1) direction used in the simulation is at 110°-290°, equivalent to an
extension at 020°-200° based on the regional geology of the Fitzroy Trough and Lennard Shelf.
Sites of low minimum principal stress (σ3) simulated in the model for the southeast Lennard Shelf are
shown in Figure 3b. Low stress anomalies show a close correlation with known sites of MVT
mineralisation and a significant anomaly occurs near Blendevale. The Cadjebut deposit has a small,
less intense low stress anomaly, and at Twelve Mile Bore, there is again a prominent anomaly. In
addition to Blendevale, Cadjebut and Twelve Mile Bore, other sites of known mineralisation which
correspond with low stress anomalies include the Brooking Spring, Brooking Springs Station, Fossil
Downs, Virgin Hill, and Gap Creek prospects.
Figure 3 a. A digitised solid geology

interpretation of the Lennard Shelf between
Fitzroy Crossing and Cadjebut (after
Dörling 1995). The geology is simplified to
four mechanical units, the basement granites
are treated as the most competent rock type,
followed by the platform limestones, shale
and finally unconsolidated conglomerates.
First order structures are modelled as
having lowest stiffness. The maximum
principal stress σ1 was simulated at 110°
based on the regional geological history of
the Fitzroy Trough and Lennard Shelf. b. A
plot showing contoured values of minimum
principal stress (σ3), only the contours
below far-field minimum principal stress are
shown. Anomalously low values of σ3 show
a close correlation with known sites of MVT
mineralisation. A significant anomaly occurs
near Blendevale, with the deposit positioned
approximately in the northwest segment of
the low stress anomaly. Other sites of MVT
mineralisation which correspond with low
σ3 anomalies include the Brooking Spring
prospect, Brooking Springs Station prospect,
Cadjebut, Fossil Downs prospect, Virgin
Hills prospect, Twelve Mile Bore deposit,
and the Gap Creek prospect.
Page 179
9.3.4 Example of three-dimensional stress mapping: the Granny

Smith mine
At Granny Smith, gold mineralisation is located along a N-S striking, moderately east dipping,
reverse fault, which partly follows a granitoid-sedimentary rock contact (Fig. 4). In different sections
of the fault, mineralisation may be developed in the sedimentary rocks, in the granitoid, and/or along
the contact between them. The orientations of conjugate sets of mineralised fractures in the granitoid
are variable, indicating that the local stress field was heterogeneous and the orientation of the
maximum principal stress (σ1) varied by up to 90 degrees from the far field ENE-WSW orientation.
It is interpreted that these variations were controlled by the geometry of the irregular granitoid contact
(Ojala, 1995; Ojala et al., 1993a,b).
After mining and extensive drilling at the Granny Smith mine, it has been possible to construct a
detailed three-dimensional geological model of the mineralisation and the eastern contact of the
Granny Smith (Fig. 5). As mentioned above, the structural observations indicate that the local stress
field was heterogenous, therefore, this deposit should be a good test for three-dimensional stress
analysis on the scale of a single mineral deposit. The main aims of the computer simulation were to
test correlation between:
1) computed minimum principal stress lows and dilations and the areas of mineralisation at the
granitoid margin, and
2) computed variations in local stress fields under ENE - WSW compression and observed
variations of orientations of mineralised structures near the granitoid margin.
Figure 4. Geology of environs of the Granny

Smith mine showing the distribution of main
rock types and major structures.
Page 180
9.3.4.1 GEOLOGICAL MODEL
For the geological model, a simple linear triangulation was used to interpolate the granitoid contact
between the drill hole intersections and mapped contact locations to create the eastern contact of the
Granny Smith Granodiorite. To minimise boundary effects in the area of interest, the granitoid block
was extended from the present erosion surface 420 RL (metres above sea level) to 600 RL and from
the deepest drill hole intersection at 54 RL to -800 RL. The dimensions of the granitoid model used
in the simulations are about 1.4 km high, 2.1 km long and 1.4 km wide, and it contains 938 surfaces
defined by 237 control points (ie. diamond drill hole intersections and mapped contacts).
Figure 5. Contours of gold grade

thickness (sum of gold assays above 0.5
ppm along the hole, contact ± 50 m)
and inferred maximum principal stress
direction with respect to the structure
contour map of the east dipping contact
between the granitoid and sedimentary
rocks at Granny Smith. The fracture
vein orientations discussed in this study
have been measured from oriented drill
cores and from the south and west
walls of the Granny pit. Diamond drill
cores have been oriented using
downhole spear, and holes have been
surveyed using Eastman single-shot
camera to record dip and azimuth of
the hole. Oriented drill cores intersect
most of the eastern contact of the
granitoid to a depth of about 150
metres.
9.3.4.2 MATERIAL PROPERTIES
The rock properties used for the simulation are shown in Table 1. Physical properties of rocks were
estimated using the compilations of Clark (1966) and Engelder (1993). All materials were modelled
as Coulomb materials. Young's modulus of 50x109 Pa was used for the Granny Smith Granodiorite,
and is typical of granitic rocks. Assuming Young's modulus ratio of two, as used by Strömgård
(1973) in his photo-elastic and finite-element models of a strong body within a weak matrix, this
gives a value of 25x109 Pa for sedimentary rocks. This value can be considered a reasonable
approximation as it is within the lower end of the range measured for shales and siltstones (Clark,
1966). These values define a relative rheological scale, such that the granitoid was stronger than the
surrounding sedimentary rocks. The rock units were assumed to be homogeneous and their material
Page 181
properties isotropic and constant. The lithological contact was modelled to be weaker than the
surrounding rocks.
Table 1 Summary of material properties for the stress simulation of a granitoid intrusion in a
sedimentary rock host at Granny Smith. Values are approximate from Clarke (1996).
9.3.4.3 INITIAL STRESS STATE
The far-field minimum principal stress (σ3) was defined as vertical and the maximum and
intermediate principal stresses (σ1 and σ2) to be horizontal and oriented 080° and 170°, respectively.
These are based on the orientations of regional foliations and the dominant mineralised fracture vein
orientation within the Granny Smith Granodiorite (Ojala, 1995). The values for σ1, σ2 and σ3 were
100 MPa (1 kb), 80 MPa (0.8 kb) and 50 MPa (0.5 kb), respectively. Assuming a hydrostatic pore
pressure, the 50 MPa confining pressure (ie. σ3) corresponds approximately to the effective stress at
a depth of about 4 km (Engelder, 1993). A far-field differential stress of 50 MPa, used to simulate
deformation during mineralisation, is within the lower end of the range of the differential stress
estimated for thrust faulting (Engelder, 1993). This was considered as a reasonable assumption
considering that the observed displacements along mineralised fractures in the granitoid are at most
on a decimetre scale (Ojala, 1995). If a much higher differential stress is used, the computer model
became very unstable, and simulation a lower differential stress yielded similar results but with
longer computing times.
9.2.4.4 RESULTS
Minimum stress (σ3) patterns and σ1 orientations produced after 10 load steps, which resulted in
about 0.5 m maximum displacement along the contact, are shown in Figures 6 and 7, respectively.
The most obvious result of the deformation simulations is, as expected for a thrust fault regime, that
σ3 values are low in the areas of shallower dipping granitoid contact. The σ3 values are also low
within the granitoid close to areas steeper dipping contact and a high σ3 anomaly is developed in the
sedimentary rock adjacent to the contact at these sites. Discrete low- σ3 anomalies also occur in
locations where the change in dip is large in steeply dipping areas. In addition, within the areas of a
shallowly dipping contact, variations in the magnitude of σ3 are correlated with small changes in the
shape of the surface. Within the shallowly dipping but smooth parts of the contact, the σ3 anomalies
are not as strong as in the areas in which the contact is concave or convex.
The simulated stress field is also heterogeneous with respect to σ1 stress orientations. These
heterogeneities for the deformation simulation at the 380RL are shown in Figure 7. Variations in the
calculated σ1 orientations are greatest in the areas within the Granny pit area where the contact is
shallowly dipping and irregular. These areas correlate with areas of low σ3 values. At the Windich
deposit, the contact shape is less complex and, although there are up to 90° variations in the modelled
σ1 orientation, variations are restricted to very close to the contact and cover smaller volumes than in
the Granny deposit. Importantly, the modelled σ1 orientations correlate very well with σ1
orientations inferred from fracture vein orientation measurements (Fig. 7).
Page 182
Figure 6 shows that low σ3 correlate with the gold mineralisation over the eastern contact of the
Granny Smith Granodiorite. The anomaly east of the Windich pit is at a deeper level than the pit and
it is solely a result of the small area of shallowly-dipping contact in the three-dimensional geological
model which is outside of the drilling data. There is an especially good correlation between the wide
low-stress areas in the granitoid and the wide low gold-grade halo. In cross-section, this broad low-
stress volume, which develops in the granitoid where the contact sharply steepens, correlates with a
wide mineralised zone of fracturing of lower average grade than deeper levels in the deposit (Fig 8).
Figure 6. Contours of σ3 (shaded) and gold

grade thickness illustrating the good
correlation of the simulated low stress areas
and gold mineralisation. Pit outlines show
the broad extent of economic mineralisation.
In summary, the deformation simulation suggests that:
i) the gold distribution gold mineralisation is structurally controlled, especially by the

shape of the granitoid surface,
ii) the maximum compressive far-field stress was oriented at about 080°-260°, and
iii) the variations in the orientations of mineralised fracture veins might be related to
heterogenous local stress fields which could have resulted from irregularities in the
shape of the granitoid surface.
9.4 Discussion
The two-dimensional, regional-scale stress modelling of the Lennard Shelf between Fitzroy Crossing
and Cadjebut shows the ability of Stress Mapping Technology to simulate zones of low minimum
Page 183
principal stress (dilation zones) which correspond with known MVT deposits, suggesting a structural
control on fluid focussing. The modelling of palaeo-stress patterns on the Lennard Shelf supports the
model in which the fluid flow and location of mineralisation is largely the result of gross regional
structure and rheological variation of different rock types during 020°-200° directed extension.
Figure 7. σ1 orientations at 380 RL

calculated using the MAP3D® program
and σ1 orientations inferred from
conjugate fracture vein measurements. Note
that the greatest variations in inferred and
calculated orientations occur in the same
areas.
At Granny Smith, the general patterns of the simulated three-dimensional minimum principal stress
correlate very well with gold mineralisation. Also the computed variations of the σ1 orientations
correlate with the observed variations of σ1 as determined by orientations of mineralised conjugate
fractures. This also suggests that the important input parameters (three-dimensional geological model,
rock parameters, stress orientations) are realistic. In some respects, this correlation is quite surprising
as the modelling was completed only to simulate the stress pattern before failure. This indicates that
the low stress zones, where failure is likely to occur first, remain as low stress zones even after failure
and during deformation. It also suggests that no through-going failures, which would have
significantly changed the geometries of structures or rock units, formed during the mineralisation.
The results further support the conceptual model that low rock stress focuses the ore-fluids
(Holyland, 1990c; Holyland, 1990d; Ridley, 1993). Two main explanations for this correlation can be
drawn:
i) variations in mean rock stress result in gradients in fluid pressure; or
ii) rising fluid pressure decreases effective stress and the tensile or extensional shear failure
occurs first in the areas of low σ3, which, following failure, are more permeable.
Numerical modelling of rockmass deformation is not an exact science and there will always be
parameters of the rockmass which are not understood. In other words, rock mechanics models are
Page 184
"data-limited", and there is seldom enough data to simulate the rockmass behaviour unambiguously
(Starfield and Cundall, 1988). However, two-dimensional modelling of the Lennard Shelf indicates
that good results, which are useful in exploration, can be obtained with a quite limited geological
database. Furthermore, the results of three-dimensional stress modelling using the Granny Smith data
show that it is possible to simulate a stress distribution that is very similar to the stress distribution
which is inferred from field observations of fracturing. Critical to the successful modelling of stresses
during epigenetic mineralisation and deformation are reasonably accurate input parameters (rock
properties, far-field stress field, and especially the three-dimensional geometry).
Figure 8. Section through the northern part of the Windich deposit showing the gold mineralisation
and low σ3. Note the widening of the gold mineralisation and the low σ3 area within the granitoid
where the dip of the contact steepens. The low σ3 anomalies deeper in the section are the result of a
curve on the modelled surface caused by interpolation beyond drill hole intersections.
9.5 Acknowledgments
Thanks are due to Simon Dörling for providing the Lennard Shelf solid geology map for the
modelling. Comments by Mark Barley, William Power and John Ridley improved the manuscript
considerably. We thank David Groves for the support at the Geology Key Centre and Terra Sancta
Research for the use of equipment and permission to publish the results.
Three-dimensional stress modelling formed part of VJO's PhD project which was supported by
University of Western Australia and Australian Overseas Postgraduate Research scholarships,
supplemented by a Outokumpu Säätiö stipend. The permission to publish this data from Placer
Pacific Limited and Delta Gold N.L. is gratefully acknowledged.
9.6 References
Banerjee, P. K. and Butterfield, R. 1981. Boundary Element Methods in Engineering Science.,

McGraw-Hill Book Company Ltd, London.
Clark, S. P. 1966. Handbook of Physical Constants., The Geological Society of America
Incorporated, New York, pp. 600.
Page 185
Dörling, S. L. 1995. Structural Evolution of Fitzroy Trough and Lennard Shelf During the Devonian -
Early Carboniferous Pillara Extension Phase: Implications for facies distribution and MVT
deposit formation. The University of Western Australia. Unpublished PhD. pp. 163.
Eisenlohr, B.N, Tompkins, L.A., Cathles, L.M., Barley, M.E., Groves, D.I. 1994. Mississippi Valley-
type deposits: Products of brine expulsion by eustatically induced hydrocarbon generation? An
example from northwestern Australia. Geology 22, 315-318
Engelder, T. 1993. Stress Regimes in the Lithosphere, Princeton University Press, Princeton, pp. 457.
Etheridge, M. A., Wall, V. and Cox, S. F. 1984. High fluid pressures during regional metamorphism
and deformation: Implications for mass transport and deformation mechanisms. Journal of
Geophysical Research 89, 4344-4358.
Groves, D.I., Ridley, J.R., Bloem, E.M.J., Gebre-Mariam, M., Hagemann, S.G., Hronsky, J.M.A.,
Knight, J.T., McNaughton, N.J., Ojala, J., Vielreicher, R.M., McCuaig, T.M., Holyland, P.W.
1995. Lode-gold deposits of the Yilgarn Block: products of Late-Archaean crustal-scale
overpressured hydrothermal systems. Journal of the Geological Society Special Pulication 95,
155-172.
Hodgson, C. J. 1989. The structure of shear-related, vein-type gold deposits: a review. Ore Geology
Reviews 4, 231-273.
Holyland, P. W. 1990a. The nature of the lithosphere: cracks and blocks?, unpaginated. Terra Sancta,
Perth.
Holyland, P. W. 1990b. Simulation of the dynamics of Archaean deformation in the Yilgarn Block,
Western Australia. In Glover, J. E. and Ho, S. E., eds. Third International Archaean Symposium,
Perth. Abstracts, pp. 347-349. Geoconferences (W.A.) Inc. Perth.
Holyland, P. W. 1990c. Stress mapping in the Mt. Isa region. Mount Isa Inlier Geology Conference,
Melbourne. Abstracts, pp. 76-77. Monash University.
Holyland, P. W. 1990d. Targeting of epithermal ore deposits using stress mapping techniques. Pacific
Rim 90 Congress, Melbourne, Proceedings, III, 337-341. Australasian Institute of Mining and
Metallurgy.
Holyland, P., Ridley, J. R. and Vearncombe, J. R. 1993. Stress mapping technology (STMTM), In
Parnell, J., Ruffel, A.H. and Moles, N.R. eds. Geofluids '93: Contributions to an International
Conference on Fluid Evolution, Migration and Interaction in Rocks. Torquay, England, Abstracts,
pp. 272-275.
Hronsky, J. M. A., Cassidy, K. F., Grigson, M. W., Groves, D. I., Hagemann, S. G., Mueller, A. G.,
Ridley, J. R., Skwarnecki, M. S. and Vearncombe, J. R. 1990. Deposit- and mine-scale structure.
In Ho, S.E, Groves, D.I., and Bennet, J.M. eds. Gold Deposits of the Archaean Yilgarn Block,
Western Australia: nature, genesis and exploration guides, 20, pp. 38-54. The University of
Western Australia, Nedlands.
Libby, J. W., Barley, M. E., Eisenlohr, B. N., Groves, D. I., Hronsky, J. M. A. and Vearncombe, J. R.
1990. Craton-scale deformation zones. In Ho, S.E, Groves, D.I., and Bennet, J.M. eds. Gold
Deposits of the Archaean Yilgarn Block, Western Australia: nature, genesis and exploration
guides, 20, pp. 30-37. The University of Western Australia, Nedlands.
MAP3D® Version 1.28 - User's Manual, Mine Modelling Limited, Copper Cliff, Ontario.
Ojala, V. J. 1995. Structural and Depositional Controls on Gold Mineralisation at the Granny Smith
Mine, Laverton, Western Australia. The University of Western Australia. Unpublished PhD
Thesis. pp. 184
Ojala, V. J., Ridley, J. R., Groves, D. I. and Hall, G. C. 1993a. The Granny Smith gold deposit: the
role of heterogeneous stress distribution at an irregular granitoid contact in a greenstone facies
terrane. Mineralium Deposita 28, 409-419.
Ojala, V. J., Ridley, J. R., Groves, D. I. and Hall, G. C. 1993b. Granny Smith: an example of a
granitoid-hosted Archaean lode-gold deposit. Geological Society of Australia, Abstracts 34, 55-
56.
Oliver, N. H. S., Valenta, R. K., Wall, V. J. 1990. The effect of heterogeneous stress and strain on
metamorphic fluid flow, Mary Kathleen, Australia, and a model for large-scale fluid circulation.
Journal of Metamorphic Geology 8, 311-331.
Page 186
Phillips, W. J. 1972. Hydraulic fracturing and mineralization. Journal of the Geological Society of
London 128, 337-359.
Ridley, J. R. 1993. The relationship between mean rock stress and fluid flow in the crust: With
reference to vein- and lode-style gold deposits. Ore Geology Reviews 8, 23-37.
Sibson, R. H. 1989. Structure and Mechanics of Fault Zones in Relation to Fault-Hosted
Mineralization., The Australian Mineral Foundation, Glenside, pp. Pages.
Sibson, R. H., Moore, R. M. and Rankin, A. H. 1975. Seismic pumping - a hydrothermal fluid
transport mechanism. Journal of the Geological Society of London 131, 653-659.
Sibson, R. H., Robert, F. and Poulsen, K. H. 1988. High angle reverse faults, fluid-pressure cycling,
and mesothermal gold-quartz deposits. Geology 16, 551-555.
Starfield, A. M. and Cundall, P. A. 1988. Towards a methodology for rock mechanics modelling.
International Journal of Rock Mechanics and Mining Sciences and Geomechanics, Abstracts 25,
99-106.
Strömgård, K. E. 1973. Stress distribution during deformation of boudinage and pressure shadows.
Vearncombe, J.R., Dentith, M.C., Dörling, S.L., Reed, A., Cooper, R., Hart, J., Muhling, P.,
Windrim, D., Woad, G. 1995. Regional- and prospect-scale controls on Mississippi Valley-type
Zn-Pb mineralization at Blendevale, Canning Basin, Western Australia. Economic Geology 90,
181-186.
Page 187
10. Towards a holistic exploration strategy: using

Geographic Information Systems as a tool to
enhance exploration
C. M. KNOX-ROBINSON a AND L. A. I. WYBORN b

a
Department of Geology and Geophysics, University of Western Australia, Nedlands,
WA 6907, Australia.
b
Australian Geological Survey Organisation, PO Box 378, Canberra, ACT 2601,
Australia.
Abstract
The exploration and mining process, from grass-roots exploration to mine-site development is a
multidisciplinary task and involves the collection, integration and analysis of datasets from
many different sources. Geographic Information Systems (GIS) have been used to coordinate
and manage the large amounts of spatial and related nonspatial data associated with modern
exploration programs. Once suitably captured in a GIS, these spatial data can be queried,
analysed, and by the application of various techniques, maps that depict mineralisation potential
or prospectivity, can be defined. Methodologies for the construction of prospectivity maps can
be split into two complementary types: empirical and conceptual. Empirical methodologies
analyse for spatial relationships between known deposits and surrounding features. Identified
spatial relationships are quantified and ultimately integrated into a single map which highlights
areas similar to those known to contain significant mineralisation. Conceptual methodologies,
which are suited to areas that contain few known deposits, use current knowledge about the
orebody formation to identify those areas which are most likely to contain significant
mineralisation. This paper focuses on the use of a GIS for the analysis of geological exploration
datasets and the construction of prospectivity maps using both empirical and conceptual
methodologies.
10.1 Introduction
The development and implementation of efficient exploration programs is becoming an

increasingly complex task. Not only are exploitable deposits becoming harder to identify, but
non-geological factors, including conservation, environmental and native-title issues must now
be incorporated into any exploration program as a matter of routine. There are two general
strategies employed in the search for new deposits. In mature mining areas, most ore bodies,
regardless of commodity type, have been located through the identification of deposit-related
surface anomalies. However, with the majority of such exposed deposits already discovered, the
focus of exploration in these areas has shifted towards the search for deposits completely
obscured by cover and which exhibit no appreciable surface expression. Alternatively, many
exploration companies have expanded their search areas to include the assessment of mineral
potential in previously under-explored areas. The two general exploration approaches involve
the assimilation and integration of large amounts of spatial and related non-spatial data.
Although the increased use of geophysical and remote sensing datasets adds significantly to the
volume of data to be processed, the necessary consideration of non-geological issues adds to the
Page 189
diversity of spatial information that needs to be collected and assessed during a typical
exploration program.
Geographic Information Systems (GIS) provide an ideal platform on which all spatial and
related non-spatial data can be stored, manipulated, analysed and integrated. Importantly, the
functionality of a GIS can be used in all facets of the mineral exploration and mining industry,
from grass-roots prospecting through to the mine development. Primarily, a GIS allows the
construction and maintenance of a database in which all spatial data, including maps, plans and
remotely sensed imagery can be stored, along with any related non-spatial data such as
lithological descriptions and structural information. Once in a GIS, these data can be
manipulated, and the contents of the database queried and rigorously analysed with a spatial
context. More importantly, a GIS allows the spatial interaction of two or more datasets to be
visualised and quantified: a task which is difficult to conduct manually, especially if the
component maps are of different scales and projections.
Although GIS have the potential to revolutionise the exploration process (Gallagher 1995) the
geological community has been relatively slow and hesitant to embrace this technology for
several reasons. Early GIS packages were very expensive and required the power of mainframe
computers, which were outside the reach of most exploration budgets. These GIS were also
userunfriendly to the extent that they required a dedicated operator. However, perhaps the most
important reason why exploration companies could not justify the cost of a GIS was the lack of
geological data in appropriate digital form, and the expense involved to convert conventional
hand-drafted maps into a suitable format. Even those companies that held data in digital
databases were unable to rapidly convert to GIS as much of the data were in a format that was
not easy to incorporate within a GIS. Most commonly, the data were held in spreadsheet formats
and there were too many composite free-text fields. More importantly the accuracy of the
locational
information held on each sample site was inadequate for accurate positioning of the site relative
to other datasets held within the GIS (Wyborn et al. 1994a).
Fortunately, most of the above reasons are no longer applicable. The price-to-performance ratio
of computers has dropped dramatically over the past decade and a reasonable platform can be
obtained for a few thousand dollars. In terms of software, although commercial systems remain
expensive, there are a number of powerful GIS packages in the public domain. The most notable
of these are the GRASS and Moss GIS programs, produced by the United States Army Corps of
Engineers and the United States Bureau of Land Management respectively, and which can be
obtained from the Internet.* Unfortunately, most GIS software still has a poor learning curve
and requires that users are moderately computer literate. However, user-friendly Graphical
User-Interfaces (GUI) are now commonplace. These both simplify the use of a GIS and reduce
the time required to become acquainted with the system. The availability of digital data has also
improved. Throughout Australia, and the rest of the world, most geological surveys are in the
process of converting their paper maps into computer-readable formats, and many now produce
all their new compilations in a digital format. With the availability of inexpensive Global
Positioning Systems (GPS), many exploration companies also produce digital map products
without the requirement of first producing a paper map.
As it is almost possible to create ìnstant' geological maps in the field with the aid of GPS, the
potential role of GIS is now finding acceptance within the geological community. Although a
GIS can be used in all stages of exploration, this paper focuses on how a GIS can be used in the
identification of prospective target areas and presents empirical and conceptual GIS-based
exploration methodologies that can be used to map mineralisation potential.
Page 190
10.2 GIS-Based Exploration Methodologies
Assessment methodologies fall into two broad groups: empirical and conceptual. Empirical
methodologies examine the spatial relationship of known deposits to the surrounding geology in
an attempt to identify the best combination of features that signify where deposits lie. Once
identified, these factors are combined and applied to the entire study region to highlight under-
explored areas similar to those known to contain deposits. Conceptual methodologies, on the
other hand, examine the essential processes and factors which led to the formation of a
particular deposit type. Such a methodology involves a good understanding of deposit
formation, and the `conversion' of these factors and processes into criteria mappable digitally
within a GIS. Whereas empirical methods can be applied only to those areas which contain a
significant number of discovered deposits, conceptual models can be applied also to
underexplored regions which contain few or no known deposits.
10.2.1 Empirical approach
An empirical approach to mineral-potential mapping, or prospectivity mapping, has been

developed in the Key Centre for Teaching and Research in Strategic Mineral Deposits at the
University of Western Australia (KnoxRobinson & Robinson 1993). The methodology was
developed to gain a better understanding of the regionalscale factors which control the location
of Archaean lode-gold deposits in the Yilgarn Craton of Western Australia, but can be applied
to other deposit styles. An empirical methodology was adopted for two main reasons. First,
there is a large number of known gold deposits within the study areas with which to identify
important spatial relationships. Second, Archaean lodegold deposits vary in their physical
appearance due to differences in structural style, host rock and degree of alteration, hence it
would be difficult to apply a conceptual-based approach. However, studies show that the
majority of deposits form a coherent group and are the result of a single, Yilgarn-wide
epigenetic mineralisation event (Groves et al. 1995). Consequently, an empirical approach based
on analogy should be able to identify new prospective areas.
The empirical methodology comprises seven main steps: (i) preliminary investigation of
deposit characteristics; (ii) collection of relevant spatial and non-spatial data and construction of
a GIS database; (iii) selection of a subset of deposits with which to identify spatial relationships;
(iv) identification of spatial relationships between known deposits and surrounding geological
features; (v) quantification of identified spatial relationships; (vi) integration of spatial
relationships into a unified prospectivity map; and (vii) test of prospectivity map using deposits
excluded in step (iii). These steps are discussed in greater detail below and are further illustrated
in Figure 1.
10.2.1.1 PRELIMINARY INVESTIGATION
A preliminary investigation of the nature of the deposits in the study region needs to be
conducted as the first step of the prospectivity analysis. As an empirical methodology is one of
analogy and aims to identify areas most likely to contain deposits of the same type as those used
in the construction of the prospectivity map, it is important that the deposits conform to a
coherent genetic group. If there are two or more distinct types of deposit in the region, each
spatially controlled by different factors, then using all of the deposits in the analysis will result
in a poor prospectivity map. This is because important factors which may be used to identify
likely sites of mineralisation for one kind of deposit may be `masked' by the lack of similar
control in the second kind of deposit. If two or more distinct groups of deposit are identified, all
deposits need to be appropriately categorised and a single prospectivity map constructed for
each group.
Page 191
10.2.1.2 DATABASE CONSTRUCTION
The empirical approach to mineral-potential mapping fully utilises the layer concept of most
GIS. That is, all features of a particular type are contained within a separate data layer, or
theme. For example, lithological boundaries are stored in one layer while lineaments are stored
in another. For maximum flexibility, attribute information is stored in a flat-file format, and
wherever possible, hierarchical codes are used to maximise the query capability of the database.
For example, metamorphic grade may be represented by a number: the greater the metamorphic
grade, the larger the number. By doing this, all rocks, say those with a metamorphic grade
below the greenschist-amphibolite transition, can be selected by a single query.
10.2.1.3 EXTRACTION OF A SUBSET OF DEPOSITS
Once it has been confirmed that deposits to be used in the analysis share a common genesis (and
hence are likely to be controlled by the same geological factors), and after the construction of a
suitable GIS database, the next step of the process is to extract a subset of deposits from the
database. These deposits are used to test the validity of the final prospectivity map. There are
two strategies with regard to the selection of a suitable subset of deposits to extract from the
database. First, a number of deposits could be selected at random for exclusion. Alternatively,
only the smaller deposits could be extracted from the database, based upon the assumption that
smaller deposits are less likely to be as spatially constrained to particular geological features (or
combination thereof) than larger deposits. If the resultant prospectivity map proves to be valid
using the smaller deposits as a test set, then attention should be turned to highly prospective
areas which contain small deposits, as ore-bearing fluids are known to have passed through
these areas.
Page 192
Figure 1 Generalised flow chart illustrating the main steps in the production of a prospectivity
map using an empirical methodology.
10.2.1.4 IDENTIFICATION
The first main step of an empirical prospectivity analysis involves the systematic and
quantitative identification of spatial relationships between known deposits and the other
geological themes stored within the GIS. At a regional scale, deposits are adequately
represented in a vector GIS as points, while other geological features can be represented as
points, lines or polygons. Consequently, there are three general spatial relationships that may
exist between known deposits and surrounding geological features: (i) proximity; (ii)
association; and (iii) abundance relationships. Methods to investigate for the existence of these
relationships have been developed, all of which implement non-parametric statistical tests.
Proximity relationship A proximity relationship exists when deposits lie preferentially closer to
a particular geological feature. For example, in the Yilgarn Craton of Western Australia,
Page 193
Archaean lode-gold deposits tend to lie close to contacts which separate rocks of strongly
contrasting rheology (Knox-Robinson 1994). One method developed to examine for a proximity
relationship involves measuring the distance of each deposit to the nearest feature of interest and
the production of a cumulative frequency diagram which shows the proportion of deposits
within any given distance of a feature. A set of randomly distributed points is then generated
over the study area and a similar cumulative frequency diagram constructed. As the randomly
distributed points are spatially unrelated to the features of interest, the associated cumulative
frequency diagram can be used as a basis to statistically determine whether gold deposits are
spatially related to the features. A two-sample Smirnov test (Rock 1988) is used to compare the
two distributions. This method should be used only when deposit clustering is not considered to
be a problem (i.e. when only the largest 20% of deposits are used in the analysis). A method
which implements a Monte Car16 simulation can be applied when deposit clustering is severe
(Knox-Robinson 1994).
A variant of the standard proximity relationship is a sizeproximity relationship, in which larger

deposits are spatially related to a particular geological feature, whereas smaller deposits are not
so constrained, or vice versa. Such a relationship has been identified within the Murchison
Province of the Yilgarn Craton in which crustal-scale faults constrain the location of large gold
deposits but have no significant influence over the siting of smaller deposits (Knox-Robinson &
Robinson 1993; Knox-Robinson 1994).
Once a proximity relationship has been identified between deposits and a particular linear
feature, the spatial relationship between the two can be further refined. Techniques have been
developed to determine whether sections of lineaments oriented within a particular strike range
are more prospective than others. A strike-proximity relationship has been identified between
gold deposits and crustal-scale faults in a portion of the Murchison Province. Along the
Meekatharra to Reedy shear zone, 16 out of 19 deposits lie in close proximity to sections of
crustal-scale faults striking within the range 013-023°. Faults within this strike range constitute
less than 25% of the total strike length of crustal-scale faults in the region.
Association relationship Whereas a proximity relationship can exist between deposits and
geological features which are represented as point or line features, an association relationship
may exist between deposits and polygon features such as rock types. Put simply, an association
relationship is one in which deposits tend to lie within a particular polygon type. The method
employed to test for the existence of an association relationship involves the use of the spatial
overlay functionality of the GIS. The points within the deposit database are spatially intersected
with the features in the polygon coverage under investigation. Once intersected, the polygon in
which each deposit lies is known and the number of deposits within each polygon type (e.g.
rock type) can be tabulated. The number of deposits expected to be associated with each
polygon type, assuming no spatial relationship exists, can be calculated based on the proportion
of area occupied by each polygon type and the total number of deposits in the study area. The
observed and expected values can then be compared using a xz test (Rock 1988). A size-
association relationship may also exist and can be quantitatively identified using a simple
variant of this method.
Abundance relationship The third type of relationship that may exist is one in which deposits
are related to the spatial abundance of a particular feature and is consequently termed an
abundance relationship. For example, rather than exhibiting just a proximity relationship,
deposits may lie preferentially in areas of high fault density. The identification of an abundance
relationship involves the production of a map which expresses the spatial density of the feature
of interest. The featuredensity map comprises a mesh of square or hexagonal cells, with each
cell assigned a value which relates to the spatial abundance of the feature of interest. An
association analysis is then conducted to determine whether deposits are spatially associated
with the feature abundance map.
Page 194
10.2.1.5 QUANTIFICATION
Once a spatial relationship has been identified between known deposits and a particular
geological feature, the relationship needs to be quantified into a map which expresses the extent
of the spatial association over the entire area of investigation. Depending on the method of
integration, discussed below, identified spatial relationships may be quantified in two ways. The
first involves the quantification of a spatial relationship as a polygon coverage comprising well-
defined boundaries and at least two types of polygon (i.e. high- and lowprospectivity polygons).
The second way involves the construction of a continuous surface which represents the
identified spatial relationship.
Discrete representation of spatial relationships Within a GIS there are two main ways to
quantify an identified spatial relationship as a discrete polygon coverage: buffering and
reclassification. Proximity relationships are quantified by buffering features of interest to a
suitable distance. Buffering involves the creation of a polygon coverage from either a point or
line precursor coverage. The resultant buffered coverage contains two types of polygon: those
within the buffering distance of the feature of interest, and those outside this distance. Graphical
methods are available which allow the selection of an appropriate distance at which to buffer
features to best represent a proximity relationship (Knox-Robinson 1994). Association and
abundance relationships are quantified simply by appropriately reclassifying the associated map
into fewer categories.
Continuous representation of spatial relationships Both proximity and abundance relationships

can be appropriately quantified as continuous surfaces within a raster GIS. In the case of a
proximity relationship, diagrams showing the proportion of deposits and proportion of area
within any given distance of a feature can be used to characterise prospectivity anywhere on the
map. For example, a surface expressing distance to features of interest can be produced and a
graph showing cumulative proportion of deposits divided by proportion of area can be used to
classify prospectivity for each cell.
10.2.1.6 INTEGRATION
Once two or more spatial relationships have been identified and quantified, they can be
integrated into a single prospectivity map. There are many ways in which several spatial
relationships can be integrated, four of which are discussed here: (i) simple; (ii) Bayesian; (iii)
algebraic; and (iv) fuzzy logic. The first three require spatial relationships to be represented as
discrete polygons; the fourth method can implement spatial relationships expressed as both
discrete maps and as continuous surfaces.
Simple integration If spatial relationships are quantified as `binary' maps comprising only two
states, and areas of low- and high-prospectivity are given values of `0' and `1', respectively, then
a simple prospectivity map can be quickly constructed. The quantified maps are spatially
ùnioned'. That is, the polygons from each input coverage are spatially merged and topology is
re-established. Each polygon of the ùnioned' coverage contains information about each input
coverage within its attribute table. A combined prospectivity can then be calculated simply by
summing the individual prospectivities for each polygon in the ùnioned' coverage. For
example, if there are five binary input maps, the resultant ùnioned' map will contain polygons
with prospectivity values ranging between `0' and `5', inclusive.
Bayesian integration The simple method is useful for appreciating the general prospectivity of
an area. However, some spatial relationships are better at targeting deposits than others and this
realisation is not accommodated in the above method. The Bayesian methodology developed by
Bonham-Carter et al. (1989) is a probabilistic one which implements the log-odds formulation
Page 195
of Bayes' Rule of conditional probability. Bayes' Rule is used to update probabilities as new
spatial relationships are integrated. The method is best applied in a raster GIS, although it can be
suitably adapted for use within a vector system. Full details of the mathematical derivation of
the Bayesian methodology are presented in Bonham-Carter (1995).
For each identified spatial relationship, R, quantified as a binary map, `weights of evidence' for
the relationship are calculated for the low- and high-prospectivity areas:
respectively, where P(B\D) is the probability that a cell which contains a known deposit is
tagged as being prospective, P(B\D) is the probability that a cell which does not contain a
known deposit is marked as a lowprospectivity cell, etc.
Using the log-odds formulation of Bayes' Rule, the spatial relationship R can be combined with
another (D):
ln [O(D\R)] = lm[O(D)] + W(R) ±
where W(R)± are W(R)- or W(R)+, as appropriate. The log-odds can then be converted to
actual probabilities.
Algebraic integration Although the Bayesian method considers the fact that some identified
spatial relationships are more significant than others, the method does have some disadvantages.
Firstly, a mathematical requirement of Bayes' Rule is that the input coverages are conditionally
independent with respect to the deposits. If this is not the case then the calculated prospectivities
will be biased and the resultant prospectivity map will either under- or overestimate the number
of deposits expected to occur within the study area (Bonham-Carter 1995). The second
limitation of the Bayesian method is that it is probabilistic and deposit size cannot be
incorporated in the calculation of prospectivity.
Another approach, which combines most of the benefits of the Bayesian method with the
relative ease-of-use of the simple method, is non-probabilistic and allows deposit size to be
incorporated. This algebraic method involves the calculation of prospectivities as weighted
means. The weights used in calculating the resultant prospectivities are determined by the way
in which the quantified relationships relate spatially (Knox-Robinson 1994). Suitable measures
of prospectivity include the number of known deposits per unit area or amount of the known
extractable gold per unit area.
Calculation of revised prospectivities using the algebraic method is illustrated in Figure 2. In

this figure, two hypothetical and very simplistic quantified spatial relationships are integrated
into a single map. As each component map comprises only two types of area (i.e. low and high
prospectivity) the resultant integrated map can contain a maximum of four different
prospectivities. The way in which the two input maps spatially interact can be summarised by a
quasi-Venn diagram and, based on the area of each prospectivity in the integrated map, revised
prospectivities can be calculated as the mean of the component weights (Knox-Robinson 1994).
Fuzzy logic integration With conventional Boolean logic, a property can have one of only two
states (`0' and `1'; òff' and òn'). In the case of a point on a map which depicts a quantified
spatial relationship, the point either lies entirely within a low prospective or high prospective
zone. Fuzzy logic breaks the requirement of `total membership', and instead of having only two
states, an infinite number of possible values exist ranging from `0' (complete disassociation)
through to `1' (complete association). Fuzzy equivalents of Boolean ÀND' and ÒR' operators
Page 196
have been defined (Zadeh 1973, 1983) as have a fuzzy algebraic product and fuzzy algebraic
sum (Table 1).
Figure 2 Cartoon illustrating the spatial combination of two identified and quantified spatial
relationships into a single map. Shown also is a quasi-Venn diagram summarising how the
regions of the component maps interrelate.
It is possible, using techniques similar to those mentioned above, to represent a proximity or

abundance relationship as a continuous surface within a raster GIS. Several spatial relationships
can be combined on a cell by cell basis using any one of the fuzzy operators listed in Table 1.
However, each operator has a limitation. When combining datasets using the fuzzy ÀND' or
fuzzy ÒR' operators, the value in the combined coverage is obtained from only one of the input
datasets. Therefore, it is possible for noise, which often has extreme high or low values, to
easily permeate through the process. Although the fuzzy algebraic sum and fuzzy algebraic
Page 197
product use all of the input datasets in their calculation, they are increasive and decreasive,
respectively. That is, in the case of the fuzzy algebraic product, as all values are less than or
equal to one, the output value is always less than or equal to the smallest input value. In the case
of the fuzzy algebraic sum, which is not a true algebraic sum, the output value is always larger
than or equal to the largest input value. As a compromise, a function has been defined, called
the gamma function, that combines the fuzzy algebraic sum with the fuzzy algebraic product.
The gamma function has the form:
If the value of gamma (y) is set equal to `0', the gamma function is equivalent to the fuzzy
algebraic product. If the value of gamma is set to `1', the gamma function is equivalent to the
fuzzy algebraic sum. Judicious choice of gamma will result in output values that are neither
increasive nor decreasive. For prospectivity mapping, a high value of gamma (y=0.95) is
usually preferred (Bonham-Carter 1995).
10.2.1.7 TESTING OF PROSPECTIVITY MAP
Once constructed, the prospectivity map can be tested using the deposits excluded from the
analysis at step three (Extraction of a subset of deposits). The test involves examination for an
association relationship between these deposits and the prospectivity categories; more deposits
should lie within areas deemed highly prospective.
The above methodology, using the algebraic method to combine spatial relationships,
has been used to construct prospectivity maps for the constituent greenstone belts of the Yilgarn
Craton. As this work was funded through an industry-sponsored AMIRA project, the results of
these analyses are confidential. However, the methodology was applied successfully to all areas
and resulted in prospectivity maps with a significant predictive capability. In the case of the
Kalgoorlie terrane, the highest prospectivity category occupies less than 0.3% of the greenstone
belt, yet this area contains in excess of 16% of known gold deposits, including the giant Golden
Mile. This represents an increased likelihood of finding a deposit in the order of over 55:1.
10.2.2. Conceptual approach
An empirical approach can be applied only when there are enough discovered deposits with
which to identify significant spatial relationships. In under-explored regions, which contain few
known deposits, a conceptual approach must be applied to assess mineralisation potential. The
Australian Geological Survey Organisation has developed a conceptual methodology which can
be used in any adequately mapped area, regardless of the number of discovered deposits
(Wyborn et al. 1994b, 1995a, b). The methodology is based on the awareness that an ore body
represents only a single component of a regionally extensive mineral system, and comprises
three
separate, but integrated steps: (i) development of a knowledge base of processes which lead to
deposit formation and their translation into criteria mappable within a GIS at a district or
regional scale; (ii) development of a high-quality GIS database of relevant information that
allow the mappable criteria of step (i) to be applied; and (iii) development and application of
routines for the assessment of mineral potential that are not statistically driven, and thus do not
rely upon the analysis of a minimum number of known deposits.
Page 198
10.2.2.1. DEVELOPMENT OF KNOWLEDGE BASE
The first step involves the development of a knowledge base of the factors and processes that
lead to the formation of ore deposits, and their subsequent translation into criteria that are
mappable at a district (tens of kilometres) or regional scale (hundreds of kilometres) within a
GIS. Most ore bodies have areal extents of less than 3 kmz, and hence do not offer a particularly
large target either for a district- or regional-scale exploration programme. However, most ore
deposits are the product of a coincidence of several common geological processes, many of
which are mappable at such large scales (Wyborn et al. 1994c, 1995a). Therefore, to adequately
analyse the mineral potential of an area it is crucial that ore deposits are considered to be only a
minor part of a complete regional-scale mineralisation system which operates on all scales from
regional to local. Based upon this realisation, a mineralisation system can be considered to
comprise six major components (Figure 3): (i) energy to drive system; (ii) source of ligands;
(iii) source of metals; (iv) transport pathway; (v) trap zone; and (vi) outflow zone.
The mineral systems approach developed by the Australian Geological Survey Organisation
attempts to identify all of the above major components for a particular deposit type and aims to
define individual criteria that are mappable within a GIS. It is emphasised that this approach is
fundamentally different from searching for specific ore deposits models (Cox & Singer 1986),
which only define the district- or local-scale features of the trap zone [component (v) of the
above list].
The methodology requires that for each major component, with the possible exception of the
outflow zone [component (vi)], potential ìngredients' are identified that may be conducive to
mineralisation. For example, a metamorphic or magmatic event may have provided the energy
to drive the system [component (i)], while a dilatant zone along a fault structure may provide a
suitable trap [component (v)]. Once identified, the essential ingredients need to be translated
into criteria that are mappable directly within a GIS. This is not always an easy task. Many
essential ingredients for ore deposit formation are expressed in terms of fluid characteristics
such as pH, Eh, salinity and fluid temperature. These ore-fluid characteristics have been derived
from detailed fluid-inclusion studies conducted at the mine scale, and an extensive database of
fluidinclusion data would be required to highlight areas on a regional scale through which
similar brines have passed.
Furthermore, even if such a database was available, it may not be reliable as the appropriate
fluid inclusions may not have been trapped or their existence may have been masked by other
generations of inclusions. To map these fluid characteristics, it is essential to translate them into
either the likely effects of particular fluid/rock interactions, or else to make some simple
assumptions on external geological parameters that would have been required to generate fluids
of specific compositions. For example, if the ore-deposit geologist has characterised the fluid in
a particular unconformity-style U-Au-PGE deposit as being highly oxidised (hematite-bearing),
of meteoric origin, and at temperatures of at least 140°C, several mappable criteria can now be
determined. First, the related fluid would have had to descend to a depth of at least 4.5 km,
based on the assumption of a normal geothermal gradient of 30°C/km, and second, the rocks
through which the fluid passed would have to be hematite-stable to prevent precipitation of U-
Au-PGE from the fluid. Consequently, at a regional scale, major thick, neutral to oxidised rock
packages should be targeted as important components in the mineral system associated with
unconformity-style uranium deposits. As appropriate trap lithologies could include
magnetitebearing units, interaction of hematite-bearing fluids with such units should produce
magnetic lows along major fault structures. The presence of these anomalous magnetic lows
would be stored within the magnetic interpretation layer of a properly constructed GIS database.
Page 199
Figure 3 The mineral systems concept of ore-body formation.
10.2.2.2. DEVELOPMENT OF A GIS DATABASE
Once the essential ingredients of a mineral system have been identified and translated into
mappable criteria, a GIS database of relevant data must be constructed such that information is
readily extractable. The construction of a GIS database has been discussed above. To further
speed up metallogenic analysis, a series of specific themes can be established from the original
map data. For example, to facilitate the measurement of distances from a particular feature, such
as a fault or granite boundary, a series of themes can be constructed which comprise concentric
buffers at varying distances from these specific features.
10.2.2.3. ROUTINES FOR THE ASSESSMENT OF MINERALISATION POTENTIAL
The final step involves the development and refinement of methodologies for the evaluation of
mineralisation potential. Three menu-driven modules have been developed for the ARC/INFO
GIS: specific modelling, interactive modelling, and generic modelling.
Specific modelling module The specific modelling module contains a suite of programs that
extract from a set of themes those regions which satisfy all geological parameters considered
essential for the formation of a specific mineral system and its associated deposit types. In
essence, this module is a simple `black box' expert system in which present knowledge of
Australian mineral systems has been previously coded and stored. The module allows the user
Page 200
to select a particular deposit type, such as vein-related tin, and the name of a GIS database
which contains the area to be assessed. The module then interrogates the GIS database and
highlights areas in which all of the deposit criteria are satisfied. Two main drawbacks are
identified with this module. First, the characteristics of a deposit must be defined, and
consequently, new deposit types will not be located using this module. Second, the GIS
database of the region has to be constructed in a rigid manner, with all themes and attributes
named identically to those within the expert system database.
Interactive modelling module The interactive modelling module is more flexible and allows the
geologist to develop a user-model by defining specific search parameters within a set of
specified themes. This methodology gives the geologist the ability to define unusual rock types
and other geological features that may make up an unknown deposit type. However, unlike the
specific modelling module, this module requires the user to have a sound knowledge of mineral
systems. Once an analysis has been completed, any known deposits in the region can be used to
test the results of an interactive modelling session.
Generic modelling module The generic modelling module is used to examine known areas of
mineralisation or important geophysical anomalies within the GIS. The module allows the user
to select an area of interest and proceeds to determine a specific geoscientific expression for the
selected area based on the contents of all themes in the GIS database. The module then identifies
all other regions that share similar characteristics, and ultimately results in a binary map that
segments the selected area into regions that are either similar or dissimilar to the area of interest.
10.3 Discussion
10.3.1. Comparison of methodologies
The two methodologies, empirical and conceptual, are not in competition with one another; they
are complementary. The empirical methodology requires a statistically viable number of
deposits to be present and cannot
be used in areas with only a few deposits/ prospects. In contrast, the conceptual methodology
can be used regardless of the number of ore deposits/prospects that are available. Once an
empirical determination has been made on controls on the location of mineral deposits in a
certain area, a conceptual approach can then be applied to vary some of the conditions that have
been predetermined to be of significance. Known deposits can then be incorporated to help
determine the relative importance of the individual criteria empirically determined to be
associated with the deposits.
10.3.2. Assessment methodologies and other facets of exploration
GIS have a great capacity to refine the exploration process, by being able to quantify (without
human bias) the associations around known areas of mineralisation. An additional use for a GIS
in the exploration industry is that it can be built to be a project management system and can
store exploration-relevant information such as lease boundaries, previous exploration results,
land tenure applications. Datasets from other project-related nongeological disciplines can be
readily incorporated, for example locations of rare faunal habitats and important cultural sites.
The ability to readily call up the location of cultural or heritage sites relevant to a planned
exploration program will help ensure that a drilling pad is not placed on a site of recognised
value.
Page 201
10.3.3. Geological constraints
Most GIS packages have the capacity to perform logical searches and analyse for aspects related
to distribution of mineral deposits. None have the capacity to ensure that the searches being
carried out will produce results that are geologically meaningful and this geological control
must come from the operator. Further, geoscientists can be so prejudiced towards their own
particular geological model that they may not see obvious associations. Empirical analysis of
exactly what occurs at a particular site will remove this bias.
Computer-driven analysis of data is usually not biased, although there is a danger that a
computer-driven geological analysis can lead to geologically meaningless interpretations. For
example, although in a GIS it is relatively easy to analyse all deposits within a certain radius of
a granite, unless the GIS has been carefully constructed, it is relatively difficult to select only
those deposits within a certain radius of a granite that are hosted by rock units that are older
than or equivalent in age to the granite. By not being able to eliminate relationships which are
geologically meaningless, an empirical association may be determined by the computer between
Proterozoic granites and deposits that are hosted by Cambrian sediments, for example.
The more geological constraints that are in-built into the original GIS database, the more
realistic the results of the analysis will be. Regardless of how the GIS is constructed, it is
essential in any analytical GIS that all results be verified independently by a geologist to ensure
they are geologically realistic.
10.3.4. GIS constraints
Currently only a few constraints in GIS development can be related to either the hardware
and/or the software capabilities. The constraints provided by the GIS are more related to the
quality and type of data incorporated, and the care that is taken in the initial stages of data
compilation-in particular, standards of digital data capture and data attribution (Hazell &
Wyborn 1995; Wyborn et al. 1994a, 1995b). Currently the quality of data that are available
limits the development of a GIS as a tool for assisting in mineral exploration programs simply
because many of the common geological attributes required for even the most simple geological
analysis are not attached to the digital data in such a way that the data can be readily
interrogated.
10.3.5. Data accuracy
A critical constraint that a GIS can offer is the scale at which the data are compiled. In reality,
GIS are scaledependent, in that only data captured in a certain range of scales will contain
adequate information for specific types of analysis. Because of the ease with which datasets can
be put together, it is essential that all data in all layers are in the same projection, and that data
in a specific layer have locational accuracy relative to similar data within other layers.
Individual data points or features must also have similar absolute accuracy. Thus if the GIS is
used to measure the distance of mineral deposits from faults, if the faults have been digitised
from a 1:10 000 scale map and the source of the mineral deposit locations is a 1:500 000 scale
map, then the distances measured are essentially meaningless.
Page 202
10.3.6. The third dimension-depth
Most GIS packages publicly available are two-dimensional. Three-dimensional packages such
as MICROMINE and VULCAN readily present geological drillhole information in the third
dimension, but in reality, these are a series of interlinked two-dimensional projections within the
vertical plane. For an advanced exploration program which will inevitably have this type of data
available, this approach will give insights into the third dimension.
In contrast, for the more greenfields-type exploration program, three-dimensional GIS is

difficult, if not impossible. Cross-sections shown on geological maps are heavily interpretative
requiring considerable intellectual input from the geologist who is in reality considering many
possible data interpretations and using geological knowledge to eliminate most of these
possibilities. Such interpretative capabilities are not possible by current GIS software. Currently
this type of analysis requires complex construction of interrelated geological datasets and
databases so as to perform even the most simple threedimensional analysis.
10.3.7. The fourth dimension-time
Properties of rocks, particularly those attributes that relate to metallogenesis such as porosity
and mineralogy, can be stored as attributes associated with each individual geological unit.
Where this has been done in GIS packages, the properties stored are usually those that pertain to
the individual unit as it currently exists. Hence, if a unit is strongly metamorphosed, it is
difficult to model the potential of that unit for synsedimentary mineral deposits.
There are no easy solutions to modelling continuous changes in rock properties with time.
However, rock properties and mineral compositions can be documented at set time intervals and
by using a series of attribute tables, the properties of a specific unit at T0 (deposition), T1
(diagenesis), T2 (deformation/metamorphism 1), T3 (deformation/metamorphism 2), etc. can be
linked within a relational database. This would then enable the potential of a highly
metamorphosed unit to be assessed for synsedimentary deposits, by selecting the properties of
that unit at T0.
10.4. Summary And Conclusions
A GIS provides endless possibilities for better and more effective targeting and focusing of
mineral exploration programs. GIS technology is currently vastly underutilised. This is in part
because of the limitations provided by digital data that are currently available. Increasingly
there is a changing emphasis towards providing `digital maps' and `databases' that are suitable
for analysis using the methodologies described above.
A considerable proportion of the development work described in the conceptual methodologies

section was carried out by Robyn Gallagher of GISolutions. LAIW would like to acknowledge
her colleagues in the Australian Geological Survey Organisation who have contributed to GIS
development over the years. Reviews by Ollie Raymond and an anonymous reviewer are
gratefully acknowledged. LAIW publishes with the permission of the Executive Director,
Australian Geological Survey Organisation.
Page 203
10.6. References
BONHAM-CARTER G. F. 1995. Geographic information systems for geoscientists: modelling

with GIS. Computer Methods in the Geosciences 13. Pergamon Press, New York.
BONHAM-CARTER G. F., AGTERBERG F. P. & WRIGHT D. F. 1989. Weights of evidence
modelling: a new approach to mapping mineral deposits. In: Agterberg F. P & Bonham-
Carter G. F. eds. Statistical Applications in the Earth Sciences, pp. 171-183. Geological
Survey of Canada Paper 89-9.
Cox D. P. & SINGER D. A. 1986. Mineral Deposit Models. United States Geological Survey
Bulletin 1693.
GALLAGHER R. 1995. Geographic Information Systems-an essential tool box for exploration.
Preview 19-27.
GROVES D. L, RIDLEY J. R., BLOEM E. M. J. ET AL. 1995. Lode gold deposits of the
Yilgarn Block: products of late Archaean crustal-scale overpressurised hydrothermal
systems. In: Coward
M. P. & Ries A. C. eds. Early Precambrian Processes, pp. 155-172. Geological Society of
London Special Publication 95.
HAZELL M. & WYBORN L. A. I. 1995. The rigorous steps required in data capture for
creating a useable GIS and the need for national data standards. In: Proceedings of the Third
National Conference and Trade Exhibition on the Management of Geoscience Information
and Data, Adelaide, 1995, pp. 22.1-22.2. Australian Mineral Foundation, Glenside, South
Australia.
KNOX-ROBINSON C. M. 1994. Archaean lode-gold mineralisation potential of portions of the
Yilgarn Block, Western Australia: development and implementation of methodology for the
creation of regional-scale prospectivity maps using conventional geological map data and a
geographic information system (GIS). PhD thesis, University of Western Australia, Perth
(unpubl.).
KNOX-ROBINSON C. M. & ROBINSON D. C. 1993. Application of geographic information
systems (GIS) to regional-scale gold prospectivity mapping. In: Williams P. R. & Haldane J.
A. eds. An international conference on crustal evolution, metallogeny and exploration of the
Eastern Goldfields, pp. 211-215. Australian Geological Survey Organisation Record
1993/54.
ROCK N. M. S. 1988. Numerical geology: a source guide, glossary, and selected bibliography
to geological issues of computers and statistics. Lecture Notes in Earth Sciences 18. Springer
Verlag, New York.
WYBORN L. A. L, GALLAGHER R., JAQUES A. L., JAGODZINSKI E. A., THOST D. &
AHMAD M. 1994a. Developing metallogenic Geographic Information Systems: examples
from Mount Isa, Kakadu, and Pine Creek. Australasian Institute of Mining and Metallurgy
Publication Series 5/94, 129-133.
WYBORN L. A. L, GALLAGHER R. BL JAGODZINSKI E. A. 1994b. A conceptual approach
to metallogenic modelling using GIS: examples from the Pine Creek Inlier. In: Proceedings
of a Symposium on Australian Research an Ore Genesis, December 12-14, pp. 15.1-15.5.
Australian Mineral Foundation, Genvale, South Australia.
WYBORN L. A. L, GALLAGHER R. & MERNAGH T. P. 1995a. Using GIS for mineral
potential evaluation in areas with few known mineral occurrences. Proceedings of the
Second National Forum on GIS in the Geosciences. Australian Geological Survey
Organisation Record 1995/46, 6-24.
WYBORN L. A. L, GALLAGHER R. & RAYMOND O. L. 1995b. Creating mineral province
GIS packages-issues, problems, solutions, and opportunities. Proceedings of the Second
National Forum on GIS in the Geosciences, Australian Geological Survey Organisation
Record 1995/46, 199-211.
Page 204
WYBORN L. A. L, HEINRICH C. A. BL JAQUES A. L. 1994c. Australian Proterozoic

Mineral Systems: essential ingredients and mappable criteria. Australasian Institute of
Mining and Metallurgy Publication Series 5/94, 109-115.
ZADEH L. 1973. Outline of a new approach to the analysis of complex systems and decision
processes. IEEE Transactions on Systems, Man and Cybernetics 3, 28-44.
ZADEH L. 1983. The role of fuzzy logic in the management of uncertainty in expert systems.
Fuzzy Sets and Systems 11, 199-227.
Page 205
Page 207
11. Developing the tools for geological shape

analysis, with regional to local scale
examples from the Kalgoorlie Terrane of
Western Australia.
S. J. GARDOLL,1 D. I. GROVES,1 C. M. KNOX-ROBINSON,1 G. Y. YUN1 AND N.

ELLIOTT2
1
Centre for Strategic Mineral Deposits, Department of Geology and Geophysics,
University of Western Australia, Perth, WA 6907, Australia.
2
Department of Computer Science, University of Western Australia, Perth, WA 6907,
Australia.
Abstract
Geological map data are often under-used in mineral exploration programs, which rely increasingly
on regolith geochemistry and geophysical and other remotely-sensed data to generate exploration
targets. However, solid geology maps, which are progressively being upgraded due to improved
interpretations of superior, remotely sensed images and airborne geophysical data, can be useful in
targeting specific types of mineral deposits, which formed late in the evolutionary history of the host
terrane. In such terranes, the present map geometry is essentially the same as that at the time of
deposit formation. This is the case for orogenic lode-gold deposits, which commonly show
predictable structural controls and/or structural geometry. Thus, the shape of a rock body, or
combinations of structures and rock bodies, may provide an important guide to the exploration
potential for orogenic lode-gold deposits.
However, until recently, there has been a dearth of techniques to quantify the various properties of
shape, and hence test the potential of the two-dimensional shape of geological bodies in map view as
an exploration tool. Integrating techniques from the field of pattern recognition with a modern
Geographical Information System (GIS) can provide the shape analysis tools required to investigate
the geometries of geological shapes. Two-dimensional shape analysis is now possible through the
calculation of several shape metrics including, but not restricted to, aspect ratio, blockiness,
elongation, compactness, complexity, roundness, spreadness and squareness. Methods are
developed for describing the geometries of rock units about mineral deposits, or any geological
features, at any scale, which for the first time makes it possible to compare shapes. These shape
analysis techniques are tested using orogenic lode-gold deposits, particularly those in the Kalgoorlie
Terrane of the highly auriferous late-Archaean Norseman-Wiluna Belt of Western Australia. On a
global scale, shape analysis indicates that those greenstone belts whose volcanic rock sequences
have high elongation and relative low roundness, complexity and aspect ratio (e.g. Kalgoorlie
Terrane) are likely to be the most richly endowed in gold. On a more local scale, characteristics of
the shape of geological features around the Golden Mile deposit are calculated and used to test the
likelihood of occurrence of gold deposits with similar geometry elsewhere in the Kalgoorlie Terrane.
The area with the most closely matching shape, on the basis of a 2km clipping-circle radius, chosen
on the basis of available proximity-analysis data, corresponds to the recently discovered Ghost Crab
deposit, illustrating the potential of the shape analysis methodology in mineral exploration.
Shape analysis is, at least in part, scale dependant, due to the inherent problem of being able to
define rock boundaries more precisely in units that have strong geophysical signatures than those
with weak signatures in poorly exposed terrains. Overcoming this problem is a challenge to the
application of this methodology.
Page 209
11.1. Introduction
Orogenic lode-gold deposits represent an ideal mineral deposit type on which to test shape analysis
methods. They are an important and widespread deposit style occurring in about 75 gold provinces,
spanning a wide geographic and temporal range within subduction-related volcano-sedimentary belts
worldwide (Groves et al. 1998). They are epigentic structurally-controlled deposits, which formed
within compressional to transpressional stress regimes associated with convergent-margin settings,
particularly during accretion, but also during continent-continent collision (Kerrich & Cassidy 1994).
Importantly, most studies show that they are late in the structural evolution of the host terrane (e.g.
Yilgarn, Western Australia, Yeats & McNaughton, 1996; Northern American Cordillera, Goldfarb et
al. 1991; Victoria, Australia: Cox et al. 1991). This means that the current geometry of the hosting
sequence and controlling sequences, as viewed in maps and cross-sections, approximates that at the
time of their formation. Thus, geological maps represent a primary mineral exploration tool if the
structural geometries of the orogenic lode-gold deposits are repeated. This was emphasised by
Groves (1996) and Groves et al. (1997), and is essentially the basis for the successful application of
stress mapping (Holyland et al. 1993; Holyland & Ojala 1997) and GIS-based prospectivity mapping
techniques (e.g. Knox-Robinson & Groves, 1997). An overview is provided by Groves et al. (in
press).
Groves (1996) and Groves et al. (1997) briefly illustrated the repetitive nature of the structural
controls on orogenic lode-gold deposits in the Yilgarn Block of Western Australia. Basically,
selective failure of competent, commonly highly chemically-reactive rocks occurs where they are
oriented at a high angle to the maximum principle stress (Ridley 1993; Figure 1a), and hence strong
orientation and elongation of the greenstone belts is an important regional factor. On a smaller scale,
within these linear belts, structurally isolated competent rock-bodies often selectively fail, forming
zones of structurally enhanced permeability, and hence high fluid flux, associated with gold
mineralisation (Figure 1b). Complex interactions of earlier and later structures may be particularly
important in producing geometries that allow greatly enhanced fluid flux through specific bodies,
potentially leading to the generation of giant gold deposits: note the similar overall geometry of the
giant Kalgoorlie and Timmins gold camps in Figure 1c. Similarly, large rigid granitoid bodies with
complex contacts may produce local heterogeneities in regional stress fields, leading to anomalously
high fluid-flux in suitably oriented structures and/or rock bodies (e.g. Ojala et al. 1993; Knight et al.
1993; Smith & Gardoll, 1997).
Thus, there is a need to develop methods to identify prospective geometries, which are sought in
exploration and expressed in two-dimensional geological maps. The approach used here is to
quantify parameters which define the shape of rock units in terms of metrics using maps stored in a
GIS database. It is the objective of this paper to show that this methodology, although in the
developmental stage, has considerable potential in mineral exploration. Structural geometry, or
shape, is an important parameter controlling the location of orogenic lode-gold deposits. Hence, this
type of deposit is a suitable type on which to test shape-analysis methodology. The Kalgoorlie
Terrane Map at 1:250,000 scale of the Geological Survey of Western Australia provides an excellent
medium to test the methodology as it: (a) represents a solid geology map with an excellent
underpinning geological and airborne geophysical database, (b) was produced in a consistent format,
by outstanding geologists (C. Swager, T. Griffin and W.K. Witt), and (c) represents a region where
GIS-based prospectivity analysis (Groves et al. in press; Knox-Robinson, this volume) has been
highly successful.
Page 210
Figure 1 Schematic reproduction of the geological controls and structures of orogenic lode-gold
deposits: (a) selective failure of units of contrasting competency with varying geometrical orientation
with respect to the maximum principal stress of the regional stress field. Adapted from Ridley (1993),
(b) structurally-controlled gold deposits from the Kalgoorlie-Coolgardie area, that have been divided
by faults into a number of isolated or semi-isolated blocks of competent units, which are selectively
gold mineralised, and (c) comparison of similar geometries of the two giant orogenic lode-gold
deposits, Kalgoorlie in the Yilgarn Block, and Timmins in the Abitibi Belt of Canada. Adapted from
Groves (1996) and Phillips et al. (1996).
11.2. Existing Geological-Shape Analysis Techniques

Many exploration companies employ strategies which are based on defining spatial factors (e.g.
faults, folds, lithological boundaries, in part defined by magnetic signatures) related to mineralisation
in known mineral fields, and then identifying similar features in under-explored areas of strategic
interest. The key to such exploration is a good understanding of the geological processes required to
form the deposits, and the ability to identify important host structures and settings responsible for
mineralisation in new areas. This type of exploration, based on the identification of key features and
the location of similar features, is a form of pattern recognition.
Several geological methods used in mineral exploration are based on shape analysis methods. One of
these is stress mapping (Holyland et al. 1993), which, by a process known as finite element analysis,
assigns rheological characteristics to digital geological map data and then simulates the deformation
of the geometries and calculates areas of low principle stress. Areas of low principle stress are
associated with possible deformation-enhanced permeability, which are, in turn, potential zones of
enhanced fluid flow and gold mineralisation. To a lesser extent, conventional GIS prospectivity
mapping, which relates the probability of occurrence of mineral deposits to measured map features
(Knox-Robinson 1994), by either empirical or conceptual relationships, can be considered to be a
simple form of geometric pattern recognition. Finally, autocorrelation is a self-comparison of either
time or spatial data to determine whether there are cyclicities or periodicities in the data. This method
has been applied to pattern recognition in the spatial distribution of gold mineralisation in the Mount
Pleasant area, Kalgoorlie Terrane (Vearncombe & Vearncombe 1999).
Pattern recognition can potentially be applied to exploration for deposits, such as orogenic lode-gold
deposits, which are structurally controlled in a limited range of rock bodies by a specific range of
structures, and developed late in the evolutionary history of the host terrane. The limiting assumption
Page 211
of two-dimensional geological-shape analysis is that the surface geology does not reflect the true
complexity of the three-dimensional geometry. Fortunately for the orogenic lode-gold deposits of the
Kalgoorlie Terrane, and deposits from other Archaean greenstone belts with which they are
compared, this may not be a critical restriction, since the vast majority of the observed structures are
upright, with most gold deposits sited within steep structures, although some significant exceptions
do occur (Vanderhor & Groves, 1998).
11.3. Pattern Recognition Processing Techniques

Pattern recognition involves the study, design and operation of systems that can recognize patterns in
digital data. Sub-disciplines of pattern recognition include: discrimination analysis, feature
extraction, cluster analysis, image analysis, character recognition, speech analysis and person
identification. Typical pattern recognition problems in non-geological shape analysis include
resolving shading, texture, shadows, blurring and resolution of various images. All computer pattern
recognition systems consist of two stages: a pre-processing stage to measure and quantify features
from the data, and a classification stage to identify and extract key elements. The typical components
of the pre-processing and classification stages of a pattern recognition system are outlined in Figure 2,
and are discussed below.
Figure 2 The procedural components of a

typical pattern recognition problem,
illustrated using a hypothetical example
involving thin-section mineral identification.
In the pre-processing stage; (a) image capture
and digitization, (b) digital image processing,
which involves filtering and enhancing the
data to emphasize specific characteristics of
interest. To recognize grain boundaries, the
image is smoothed and the edges enhanced,
and (c) analysis and description, which
quantifies image features. In this case the
grain boundaries are defined with edge
detection. In the classification stage; (a) data
modelling involves building a digital model of
the preprocessing results, and (b)
classification involves assigning the data to an
appropriate class.
The pre-processing stage is usually ad-hoc

and methods vary greatly. However, all
preprocessing stages of any pattern recognition problem comprise the following three stages: (a)
image capture and digitization, which involves making a digital image of the data, using either a
digital camera, video camera or a scanner, and converting the recorded information into a digital
Page 212
representation (typically, digital image data are represented using a two-dimensional array); (b)
digital image processing, which involves filtering and enhancing the data to emphasize specific
characteristics of interest (Figure 2), and (c) analysis and description of the image, which divides it
into constituent parts and assigns a numeric value to each part.
The classification stage is dependent on the output generated in the preprocessing stage. There are
two steps in the classification stage of a pattern recognition problem: (a) data modeling, where the
image data are normally organised as a list of objects with associated features, and (b) classification,
which involves determining to which class a pattern belongs. Often, multivariate statistical analyses,
such as multiple regression, cluster analysis and principal component analysis, are employed as the
classification technique. For example, Figure 2 illustrates how the digital image of a rock in thin
section is classed into its constitution mineralogy.
11.4 Shape Analysis In GIS
Over the past decade, the emergence of GIS has revolutionised the means by which geological map
data are collected, stored, manipulated and analyzed. However, there has been little change in the
functionality of GIS even though systems have become simpler to use. There are two common
methods for recording geographic information; vector and raster format. Vector based systems
record data using a Cartesian coordinate system, with, points, lines and polygons. Points are stored as
x,y coordinate pairs, lines are stored as list of points, and polygons are stored as one or more lines
depending on the GIS application used. Typically, the only shape attributes assigned to vector data
are the length of line features and the area and perimeter of polygon features. There are exceptions,
including TNTmips and the GIS package called MAP, which have some rudimentary shape
processing functionality.
In a raster based GIS, the spatial locations of geographic features are recorded by using rows and
columns of grid cells of a uniform size. Although most of the work involving pattern recognition is
performed with raster data, these pattern recognition methods are not well developed in raster GIS.
The problem is the lack of analysis and description tools for calculating shape attributes. However,
more advanced GIS contain multivariate statistical analysis techniques, which are required in the
classification stage of most pattern recognition problems.
11.5. A Hybrid Gis-Pattern Recognition Approach

Principal moments of inertia are amongst the most important global features of a shape, and are
commonly used in the pattern recognition processing of two-dimensional shapes (Gonzalez &
Woods, 1992). The advantage of moment-based calculations is that they are independent of scale,
and offer the best means of identifying similar shapes at different scales. Most calculations of
moments are done using raster data, as the majority of pattern recognition work involves digital
imaging, typically from video input.
Although some GIS support raster data-types, a vector format is most commonly used, as its
implementation maximizes the storage and analysis capabilities of the software. Essentially, a
geological map in a GIS environment can be considered to be a two-dimensional vector image, and
can be treated in a similar fashion to most binary images in pattern recognition processes. For each
polygon in a vector GIS, the first- and second- order moments can be calculated using the Leu
Triangle Method (Leu, 1991), which works on polygon data types. Using this and other methods
(Elliott, 1997), the area and perimeter (Parea and Pperim respectively; Figure 3a) of a polygon can be
calculated. Although area and perimeter information is usually stored as an attribute of a polygon in
the GIS database, it is necessary to recalculate these values when only part of a shape is extracted and
examined. The minimum and maximum magnitudes of inertia, Mininertia and Maxinertia,
respectively (Figure 3b), are calculated. The angle theta describes the orientation of the measured
Page 213
shape along the minimum magnitude of inertia axis, and is used to calculate the orientation of the
minimum inertia bounding-rectangle (MIBR; Figure 3c). The MIBR is the smallest rectangle that
circumscribes the measured polygon. Finally, the convex hull (Figure 3d), which is the minimum
enveloping surface of the measured polygon, can be determined, and its area (Carea) and perimeter
(Cperim) defined.
Figure 3 Illustration of some shape properties that can be calculated using moments. (a) Area and
perimeter of a polygon, Parea and Pperim, respectively. (b) The minimum magnitude of inertia,
Mininertia. (c) The minimum inertia bounding rectangle, MIBR. (d) The convex hull area and
perimeter, Carea and Cperim, respectively.
The shape attributes outlined above, and illustrated in Figure 3, are limited in the extent to which they
describe a shape. However, using these primitive shape attributes, eight shape metrics, which are
more descriptive of shape properties, are calculated. These eight shape metrics are; aspect ratio,
blockiness, elongation, compactness, complexity, roundness, spreadness and squareness. All metrics
are normalised between 0 and 1. A description of each metric and their associated formulas are given
in Table 1.
11.5.1 Pre-processing stage
Pre-processing is the name given to a family of procedures for smoothing, enhancing, filtering,
cleaning-up and generally preparing a digital image to improve the accuracy of algorithms used to
produce a final classification. Within a GIS, the pre-processing stage is different to that shown in
Figure 2, because a geologist performs the equivalent of the image processing step by collating the
map from as many data sources as possible (Figure 4). Thus, there is no need to filter and enhance
the data to emphasize specific characteristics of interest in the image processing step, as the
geological mapper has resolved these issues. Image capture and digitization occurs when the map
data is digitized and stored as vector data.
Page 214
Figure 4 The procedural components of the pre-processing stage of a pattern recognition problem in
a GIS environment, when applied to geological maps.
The final step of pre-processing, analysis and description is partly completed with the data already
subdivided into it constituent parts, the rock types, as defined by the geological mapper. Thus, there
is no need for thresholding and edge detection functions to help define rock boundaries. The only
decision that remains is which objects or parts of objects are to be measured. Determining which
features to extract, and quantify, is known in the field of pattern recognition as the feature extraction
process. To examine the geometry of rock units around a deposit, a clip operation is required to select
the shapes for analysis. Figure 5 illustrates the progressive steps in a feature extraction process and
the subsequent calculation of the elongation metric. Shapes that are extracted and measured are
termed the subject polygons. A circle is chosen as the clipping shape since it is rotationally invariant.
This means that similar shapes in different orientations will always return the same value for a given
metric. The circle used for clipping the shapes is called the clipping circle. The value for each metric
is the sum of all the clipped polygons, normalized by their polygon area (Parea), divided by the area
of the clipping circle, as shown in Figure 5 and Equation (1). The resulting metric is assigned to a
point at the centre of the clipping circle. Sampling the data at regular intervals generates a regular
distribution of points, which can be stored as a two-dimensional array containing the values of the
measured metric.
Page 215
Figure 5 Components of the analysis and description step in the pre-processing stage of a pattern
recognition problem. Shown diagrammatically is the feature extraction process for the rock units
about the Golden Mile deposit, and the feature description of the elongation metric on the selected
subject polygons. The resulting elongation-metric value for the rock units about the Golden Mile
deposit is assigned back to a point at the sample location.
xj = ∑ F ( P ) * ( P area / clipping circle area )

i=0
j i i (1)
where, xj is the measured value of metric j,

where j = { aspect ratio, blockiness, elongation, compactness,
complexity, roundness, spreadness, and squareness}
Fj is a function that maps a polygon to a metric j
n is the number of polygons inside the clipping circle
Pi is the measured polygon i
The limiting factor in this approach is that it is scale dependent, and the shapes that are detected are
highly dependent on the diameter and position of the clipping circle used. This effect is illustrated in
Figure 6, which displays the results of using at 2 and 8 km radius on the elongation metric. The 2 km
radius clipping circle clearly identifies long, sliver-like greenstone belts with high elongation values
(Figure 6). The granitoids and poorly subdivided sedimentary units are assigned very low elongation
values at this scale, because neither have significant internal boundaries. In contrast, the 8 km radius
clipping circle assigns high elongation to a combination of greenstone-dominated areas with multiple
stacked thin units, thinner belts of sedimentary rocks and some elongate granitoids bodies,
particularly those internal to the greenstone belts. Thus, different geological significance is ascribed
to areas defined by the same shape parameter at different scales.
Another example of how the shape properties of rock units vary with scale is shown for rock units
adjacent to the Golden Mile gold deposit at 2, 4, 8, and 16 km in Figure 7. These were calculated
using clipping circles at 2, 4, 8, and 16 km centred on the Golden Mile gold deposit, and clipping out
the intersecting rock units defined in the northern and southern map sheet of the 1:250,000 scale
Kalgoorlie Terrane map (Swager and Griffin, 1990). The GIS geological map was first simplified by
reducing the geology to its basic rock types, which eliminates fault contacts between the same rock
units. Fault boundaries within the same rock unit are regarded as less favourable for gold
mineralisation, as they represent contact boundaries without the large chemical and rheological
contrast, normally important for mineralisation (Groves et al. 1997). In Figure 7, it is evident that the
elongation of lithological units about the Golden Mile gold deposit decreases as the clipping circle
size increases from 2 to 16 km. This is interpreted to reflect the fact that as the scale of observation
Page 216
increases, the area of undifferentiated sedimentary rocks on the map increases, and this affects the
shape parameters because, as discussed above, internal boundaries within the sedimentary rocks
cannot be mapped, due to their low geophysical response and poor outcrop in the area. Similarly, the
roundness metric decreases with increasing diameter of the clipping circle, in this case probably
controlled by distance from the major anticline defined by well-differentiated volcanic units near the
Golden Mile deposit.
Figure 6 Diagram to illustrate that shapes which are detected using shape analysis are highly
dependent on the diameter and position of the clipping circle used. This effect is illustrated here with
elongation values over the Kalgoorlie Terrane calculated using clipping circles set at 2 and 8 km
radius. High elongation values are show in red.
Figure 7 Shape feature vectors for the Golden Mile deposit at

2, 4, 8 and 16 Km clipping-circle radius are represented
diagrammatically on spider diagrams. The length of the lines
on the spider diagrams is proportional to the value of the
defined metric value: i.e., the longer the line, the higher the
value.
Page 217
11.5.2. Classification stage
Classification is concerned with making decisions concerning the class membership of the patterns in
question. In this study, one area of interest is to compare the similarity of a given area to the
geometry of the giant Golden Mile gold deposit, Kalgoorlie. To do this, a decision rule must be
defined that is easy to compute and will minimize the probability of misclassification. Here, the data
collected from the pre-processing stage are collected and stored in a two-dimensional array for each
metric. The resolution of the two-dimensional arrays is determined by the sample interval between
clipping circles. For any location within the sampled area, an eight-dimensional vector
V=(x1,x2,…,x8), where x is a defined metric (aspect ratio, blockiness, elongation, compactness,
complexity, roundness, spreadness and squareness), can be defined (Figure 8a). V is referred to as a
feature vector, and the space in which V lies is called the feature space. A class representing a
particular type of shape can be visualised as occupying a region or sub-space of the feature space.
Figure 8 Simplified schematic diagram of the metrics in three-dimensional space: (a) The
relationship between feature vector and feature, (b) The relationship of the classification score to the
observed feature vector and the defined feature vector.
At any given location on the map, the feature vector for the shapes, at the scale defined in the pre-
processing stage, can be defined. This is known as the defined feature vector. To determine how
close any other pattern is to the defined feature vector requires calculation of the Euclidean distance
between the defined feature vector and the observed feature vector. Figure 8b illustrates the
relationship between the defined feature vector and the observed feature vector for a case involving
only three metrics, which can be represented in a three-dimensional feature space. A classification
score can be calculated based on the Euclidean distance between the defined and observed feature
vectors. The value assigned to the classification score is shown in Equation (2). A low classification
score implies a high degree of similarity between the measured shapes, as both feature vectors will be
in close proximity in the feature space.
Page 218
2
⎛ 8 ⎞
Classification score = ⎜⎜ ∑ (V j − X j ) ⎟⎟ (2)
⎝ j =0 ⎠
Where, Vj is the defined feature vector for the Golden Mile deposit.
Xj is the observed feature vector.
Where j = { aspect ratio, blockiness, elongation, compactness,
complexity, roundness, spreadness, and squareness}
11.6. Examples Of Map Data Analysis For Exploration

Shape analysis is in the developmental stage, and the precise applications to mineral exploration are
currently under review. However, two examples, both involving the Kalgoorlie Terrane map
discussed above (Figure 6), are briefly described below in order to demonstrate the potential of the
methodology. The first example is a global scale comparison of several gold-mineralised cratons,
terranes or provinces, and the second concerns the likelihood of an area with similar geometry to the
Golden Mile occurring elsewhere in the Kalgoorlie Terrane.
11.6.1. Global-scale shape comparison
Through the PhD study of G.Y. Yun (e.g. Yun et al., 1998), digital maps are available of the
Kalgoorlie Terrane (compiled at 1:250,000), Abitibi Belt (1:500,000), Yilgarn Craton (1:1,000,000),
Zimbabwe Craton (1:1,000,000) and Superior Province (1:1,000,000). These regions of the Earth are
all anomalously endowed with orogenic lode-gold deposits, but have decreasing prospectivity or
endowment from the Kalgoorlie Terrane (158 Kg Au/Km2), to Abitibi Belt (134 Kg Au/Km2), to
Yilgarn Craton (42.6 Kg Au/Km2), to Zimbabwe Craton (44 Kg Au/Km2) and Superior Province
(21.6 Kg Au/Km2). Visual inspection of the maps suggests that there may be a broad trend from
elongate, less domal terranes to less elongate, more domal terranes from the more gold endowed to
the less gold-endowed terranes. This can be tested using the shape analysis techniques on the digital
map data. In this case, the two-dimensional shapes of the volcanic-dominated segments of the
greenstone belts are used as they have the greatest internal resolution, and are hence interpreted to
best represent the true shape of the belts. No clipping circle is used, but rather the total shape of the
units is recorded. Due to difficulties of subdivision during regional mapping, the sedimentary
sequences are displayed as broad, largely undivided, units on the regional-scale maps, and hence are
less suitable for the analysis.
Figure 9 represents the variations in the selected mean shape-parameters of the volcanic-dominated
segments of greenstone belts for the five regions. Its is clear that it defines a broad relationship
between shape and gold endowment for the regions, with elongation decreasing, whereas convexity,
aspect ratio and roundness all increase, with declining gold endowment across the scales represented.
The shape parameters are, in fact, quantifying the qualitative observation that increasing elongation
and less domal form equates with greater gold mineralisation potential. This, in turn, potentially
agrees with theoretical considerations (e.g. Ridley, 1993: Figure 1a) that greenstone belts with a high
percentage of rocks oriented at a high angle to Sigma1 in the far-field stress will be better mineralised
than those with a more random orientation, due to the enhanced mechanical failure of the more
suitably-orientated competent units during deformation in this stress field. Thus, regional-scale shape
analysis has the potential to provide semi-quantitative predictions of the orogenic gold potential of
greenstone belts where there are adequate maps for the analysis.
Page 219
Figure 9 Mean total shape parameters for the volcanic-dominated segments of the greenstone belts
in the Kalgoorlie Terrane, Abitibi Belt, Yilgarn Craton, Zimbabwe Craton and Superior Province.
Note that elongation decreases and convexity, aspect ratio and roundness all increase with
decreasing gold endowment.
11.6.2. Identification of similar geometries in the Kalgoorlie Terrane.
As mentioned above, and illustrated in Figures 6 and 7, the limiting factor in identifying similar
geometries via shape analysis is that the pre-processing stage of the technique is scale dependent, and
thus the shape values that are returned are highly dependent on the diameter and position of the
clipping circle used. Thus, if the geometry about a deposit is to be determined, there is a need to
resolve the scale at which the host rock units are measured. Knox-Robinson (1994), from
prospectivity mapping work conducted on the same map of the Kalgoorlie Terrane, identified that
over 95% of the deposits fall within 2 km of any contact type, and concluded that lithological contacts
have a strong spatial control on mineralisation. For this reason, analysis of the shapes in the
Kalgoorlie Terrane area was conducted with a 2 Km radius of the clipping circle. The sample interval
was set at 1 km, and the regular distribution of points that were generated were stored as two-
dimensional arrays for each metric. It is considered that a 1 km sample interval, with a 2 km clipping
circle radius, ensures sufficient overlay between samples to produce a reliable image over the entire
map.
Using the defined feature vectors for the Golden Mile deposit (Figure 7), a comparison to any
observed feature vector on the Kalgoorlie Terrane map can be calculated, and the corresponding
classification score calculated. The defined feature vector, with a 2km clipping circle radius, for the
Golden Mile gold deposit (aspect ratio 0.21, blockiness 0.51, compactness 0.12, convexity 0.93,
elongation 0.76, roundness 0.72, spreadness 0.62, squareness 0.65) is compared with observed feature
vectors at 2, 4, 8 and 16 km clipping circle radii at 1 km intervals across the Kalgoorlie Terrane map.
Comparing the defined feature vector measured at 2 km, with the observed feature vectors at 2, 4, 8,
and 16 km, partially resolves the limiting effect of scale between these intervals. For each sample
location, only the best classification score, from the four scales recorded, is assigned to the output.
The results of this shape classification are presented in Figure 10, and are shown using a shape match
score, which is defined as the inverse percentage of the classification score mentioned above (e.g.
100% is a complete shape match). It is evident that, in the Kalgoorlie-Coolgardie region, although
Page 220
much of the greenstone belt has a shape match of about 80%, there are a small number of areas that
have a shape match greater then 95%. The highest shape match to the defined feature vector of the
Golden Mile gold deposit, in the Kalgoorlie Terrane map, is to the recently discovered 400,000oz
Ghost Crab gold deposit with a 96.5% shape match. Visually, the geometries of the rock units about
the Ghost Crab deposit resemble the geometries about the Golden Mile, except that the rock units are
inverted at Ghost Crab. This is clearly not an exhaustive study, but simply illustrates the potential
value of the methodology in exploration.
Figure 10 Results of the shape match of the defined feature vector for the Golden Mile deposit at 2,
4, 8, 16 km radii to every other observed feature vector on the Kalgoorlie Terrane map. See text for
detailed exploration of methodology.
11.7. Conclusions
The ability to describe and identify characteristic shapes or geometries, reflected in the surrounding
geology, which, in turn, reflect the structural controls on mineral deposits could assist mineral
exploration. This is not possible in existing GIS, but can be achieved by integrating the functionality
of a GIS with the processing techniques of computer pattern recognition. Within the GIS
environment, there are unique opportunities to develop useful pattern recognition tools for use in
geological shape analysis.
It is now possible to describe the geometries of rock units around an ore deposit, and make
comparisons with other areas. The shape tools developed here offer new exploration opportunities
for assessing the potential of terranes to contain ore deposits that have specific structural controls. At
the craton to terrane scale, for example, regions with decreasing gold endowment (Kalgoorlie
Terrane, Abitibi Belt, Yilgarn Craton, Zimbabwe Craton, Superior Province) are shown to have
decreasing elongation but increasing aspect ratio, roundness and complexity. On a goldfield to
greenstone belt scale, an assessment of the potential of an area to host an ore deposit within host units
of a specific two-dimensional geometry can be achieved by using a clipping circle as the feature
Page 221
extraction process. This method makes it possible to examine the geometry of rock units about a
deposit at any scale. Although moments are scale independent, the feature extraction method
employed here limits the scale of the measured shapes to the radius of the clipping circle. Critical to
this type of investigation is defining the scale, or range of scales, of the geological structures that host
the deposits, including orogenic lode-gold deposits. This is because shape parameters change with
scale, partly because of real changes in geometry of the rock units, but also partly because of inherent
problems in ascribing rock boundaries in different types of sequences: for example, volcanic-
dominated vs sediment-dominated portions of greenstone belts. Shape analysis must be controlled by
investigations that define the critical scale of the controlling features about orogenic lode-gold
deposits. This work has shown a commonality between the shapes of the geological units at the
Golden Mile deposit and the recently discovered Ghost Crab gold deposit (400,000oz). Ghost Crab is
shown to have 96.4% shape match with that of the defined Golden Mile shape, measured with a 2 km
clipping circle radius, and compared with observed feature vectors at 1km intervals across the
Kalgoorlie Terrane map measured at the 2,4,8 and 16 km scale range. The 2km clipping radius is
based on previously defined proximity relationships.
This research provides an initial, but significant, step in increasing the practical functionality of GIS
in mineral exploration by including the analysis of geological shapes. This work offers new tools to
enhance current prospectivity mapping techniques by providing the potential to identify the geometry
of specific geological structures that control late-formed ore deposits. Scale dependency is a
outstanding problem which needs to be resolved.
This paper has been developed from co-operative research at the Centre for Strategic Mineral
Deposits (CSMD). We are particularly grateful to, in alphabetical order, Michael Barrett-Lennard,
Warick Brown, Howard Carr, Cecilia D’Ercole, Peter Kovesi and Tom Ridsill-Smith for their
invaluable comments.
11.9 References
COX S. F., WALL V. J., ETHERIDGE M. A. & POTTER T. F. 1991. Deformational and metamorphic
processes in the formation of mesothermal vein-hosted gold deposits- examples from the Lachlan
Fold Belt in Central Victoria, Australia. Ore Geology Reviews 6, 391-324.
ELLIOTT N. 1997. Quantification of geological features stored within a Geographic Information
System (GIS). Unpublished BSc Honours thesis. The University of Western Australia, Perth, 92p.
GOLDFARB R. J., LEACH D. L. & PICKTHORN, W. J. 1991. Source of synorgenic fluids of the
Northern Cordillera: Evidence from the Juneau Gold Belt, Alaska. In: ROBERT F., SHEAHAN P. A.
& GREEN S. B. eds. Greenstone Gold and Crustal Evolution. Geological Association Canada,
Mineral Deposits Division Publication, 160-161.
GONZALEZ R. C. & WOODS R. E. 1992. Digital Image Processing. Addison-Wesley, New York,
235p.
GROVES D. I. 1993. The crustal continuum model for Late Archaean lode gold deposits of the
Yilgarn Block. Western Australia. Mineralium Deposita 18, 366-374.
GROVES D. I. 1996. Geological concepts in the exploration for large to giant late-orogenic
(mesothermal) gold deposits. In: Peru: Second International Gold Symposium. Peru, 374-379.
GROVES D. I., GOLDFARB R. J., GEBRE-MARIAM M., HAGEMANN S. G. & ROBERT F. 1998.
Orogenic gold deposits: a proposed classification in the context of their crustal distribution and
relationship to other gold deposit types. Ore Geology Reviews 13, 7-27.
GROVES D. I., GOLDFARB R. J., KNOX-ROBINSON C. M., OJALA V. J., GARDOLL S. J., YUN G. Y.
and HOLYLAND P. W. in press. Late-kinematic timing of orogenic gold deposits and significance
Page 222
for computer based exploration techniques with emphasis on the Yilgarn Block, Western Australia.
Ore Geology Reviews.
GROVES D. I., OJALA V. J. & HOLYLAND P. W. 1997. Use of geometric parameters of greenstone
belts in conceptual exploration for orogenic lode-gold deposits. Australia Geological Survey
Organisation 103-108.
HOLYLAND P. W. & OJALA V. J. 1997. Computer aided structural targeting in mineral exploration:
two and three-dimensional stress mapping. Australian Journal of Earth Sciences 44, 421-432.
HOLYLAND P. W., RIDLEY J. R. & VEARNCOMBE J. R. 1993. Stress mapping technology (SMT). In:
PARNELL J., RUFFEL A. H. & MOLES N. R. eds. Geofluids 93: Contributions to an International
Conference on Fluid Evolution, Migration and Interaction in Rocks. Torquay, England, 171-175.
KERRICH R. & CASSIDY K. F. 1994. Temporal relationships of lode-gold mineralisation to accretion,
magmatism, metamorphism and deformation- Archaean to present: a review. Ore Geology Reviews
9, 263-310.
KNIGHT J. T., RIDLEY J. R., GROVES D. I. & MCCALL C. 1996. Syn-peak metamorphic gold
mineralisation in the amphibolite-facies, gabbro-hosted Three Mill Hill deposits, Coolgardie
Goldfield: a high temperature analogue of mesothermal grabbo-hosted gold deposits. Transactions
of the Institution of Mining and Metallurgy (London) 105, B175-B199.
KNOX-ROBINSON C. M. 1994. Archaean lode-gold mineralisation potential of portions of the
Yilgarn Block, Western Australia: development and implementation of the methodology for the
creation of regional-scale prospectivity maps using conventional geological map data and a
Geographic Information System (GIS). Unpublished PhD thesis, University of Western Australia,
Perth, 178pp.
KNOX-ROBINSON C. M. & GROVES D. I. 1997. Gold prospectivity mapping using a geographic
information system (GIS), with examples from the Yilgarn Block of Western Australia. Chronique
de la Recherche Minière 529, 127-138.
KNOX-ROBINSON C. M. & WYBORN L. A. I. 1997. Towards a holistic exploration strategy: using
Geographic Information Systems (GIS) as a tool to enhance exploration. Australian Journal of
Earth Sciences 44, 452-464.
LEU J. 1991. Computing a shapes moment from its boundary. Pattern Recognition 24, 949-957.
OJALA J. V., RIDLEY J. R., GROVES D. I. & HALL G. C. 1993. The Granny Smith gold deposit: the
role of heterogeneous stress distribution at irregular granitoid contact in a greenstone facies terrane.
Mineralium Deposita 28, 409-419.
PHILLIPS G. N., GROVES D. I. & KERRICH R. 1996. Factors in the formation of the giant Kalgoorlie
gold deposit. Ore Geology Reviews 10, 295-317.
RIDLEY J. R. 1993. The relations between mean rock stress and fluid flow in the crust: with reference
to vein lode-style deposits. Ore Geology Reviews 8, 23-27.
SMITH A. & GARDOLL S. J. 1997. Structural analysis in mineral exploration using a GIS-adapted
stereographic-projection plotting program. Australia Journal of Earth Sciences 44, 445-452.
SWAGER C. & GRIFFIN T. 1990. Geology of the Archaean Kalgoorlie Terrane, Northern and
Southern Sheets. Geological Survey of Western Australia, Perth.
VANDERHOR F. & GROVES D. I. 1998. Project M195, Systematic documentation of Archaean
gold deposits of the Yilgarn Block. Mineral and Energy Research Institute of Western
Australia, Perth, MERIWA Report 193, 319p.
VEARNCOMBE J. R., BARLEY M. E., EISENLOHR B. N., GROVES D. I., HOUSTOUN S. M.,
SKWARNECKI M. S., GRIGSON M. W. & PARTINGTON G.A. 1989. Structural controls on
mesothermal mineralization: examples from the Archaean terranes of Southern Africa and Western
Australia. Economic Geology Monograph 6, 124-134.
VEARNCOMBE J. R. & VEARNCOMBE S. 1999. The spatial distribution of mineralization: applications
of the Fry Analysis. Economic Geology 94, 475-486.
YEATS C. J. & MCNAUGHTON N. J. 1996. SHRIMP U-Pb geochronological constraints on Archean
volcanic-hosted massive sulfide and lode gold mineralization at Mount Gibson, Yilgarn Craton,
Western Australia. Economic Geology 91, 1354-1371.
YUN, Y. G., GROVES, D. I., KNOX-ROBINSON, C. M. & GARDOLL, S. J. 1998. GIS-based
methodology for prospectivity analysis of orogenic lode-gold deposits: a preliminary study of the
Kalgoorlie Terrane as an example. In: Zhou, Q., Li, Z., Lin, H., Shi, W., (Eds), Proceedings of
Page 223
Geoinformatics '98 Conference: Spatial Information Technologu Towards 2000 and Beyond.,
Beijing, 288-298.
Page 224

OrogenicGold Asjan2005 WorkshopUFRGS Brazil PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

OrogenicGold Asjan2005 WorkshopUFRGS Brazil PDF

Uploaded by

Copyright:

Available Formats

Exploration for Orogenic Gold Deposits, Porto Allegre, Brazil, January 2005

1. GOLD DEPOSITS IN METAMORPHIC BELTS: OVERVIEW OF CURRENT

1.2. Potential Diversity of Gold Deposit Types in Metamorphic Belts ........................................................ 4

1.3. Orogenic Gold Deposits: The Most Common Type?.................................................................................... 5

1.4. Intrusion-Related Gold Deposits: A Coherent Group?.............................................................................. 14

1.6. Giant Gold Deposits in Metamorphic Belts?.......................................................................................... 24

2. OROGENIC GOLD DEPOSITS: A PROPOSED CLASSIFICATION IN THE CONTEXT OF

2.2. Definition of So-Called Mesothermal Gold Deposits .............................................................................. 2

2.3. Problem of Nomeclature.................................................................................................................................. 18

2.4. Proposed Classification.................................................................................................................................... 20

2.6. References .......................................................................................................................................................... 21

3. ROTUND VS SKINNY OROGENS : WELL NOURISHED OR ANOREXIC GOLD? ............29

3.2. Orogenic Gold Deposits In Space And Time................................................................................................ 30

3.3. Gold Deposits And Evolution Of Early Earth ............................................................................................. 33

3.4. Gold Distributions Under A More Modern-Style Plate Tectonics...................................................... 34

3.5. Conclusions: Control On Gold Distribution By Pattern Of Crustal Growth................................... 36

3.6. References .......................................................................................................................................................... 36

4. TECTONIC SETTING OF SYNOROGENIC GOLD DEPOSITS OF THE PACIFIC RIM...........39

4.2 Tertiary synorogenic gold deposits of the Pacific Rim................................................................................. 47

4.2.2.3. Willow Creek district........................................................................................................................... 52

4.3. Mesozoic synorogenic gold deposits of the Pacific Rim .............................................................................. 54

4.4. Palaeozoic synorogenic gold deposits of the Pacific Rim............................................................................ 62

4.7 References ........................................................................................................................................................... 68

5. ARCHEAN OROGENIC LODE GOLD DEPOSITS ..................................................................75

6. HYDROTHERMAL TRANSPORT AND DEPOSITIONAL PROCESSES IN ARCHEAN

6.2. Physicochemical conditions of the ore fluids ................................................................................................ 78

6.3. Gold transport in hydrothermal solutions.................................................................................................... 81

6.4. Gold solubility and speciation in lode-gold ore fluids ................................................................................. 84

6.5. Ore precipitation mechanisms........................................................................................................................ 86

6.8 References ........................................................................................................................................................... 91

7. ALTERATION AND PRIMARY GEOCHEMICAL DISPERSION ASSOCIATED WITH THE

7.2. Geological setting .............................................................................................................................................. 99

7.3. Mapping, sampling and analytical procedures.......................................................................................... 102

7.4. Host rocks ........................................................................................................................................................ 103

7.5. Hydrothermal alteration ............................................................................................................................... 107

7.6. Chemical changes related to alteration ....................................................................................................... 111

7.7. Discussion......................................................................................................................................................... 115

7.8. Geochemical dispersion patterns.................................................................................................................. 119

7.9. Conclusions...................................................................................................................................................... 125

7.10 Acknowledgements........................................................................................................................................ 127

7.11 References....................................................................................................................................................... 127

8. LATE-KINEMATIC TIMING OF OROGENIC GOLD DEPOSITS AND SIGNIFICANCE FOR

Abstract ................................................................................................................................................................... 131

8.1. Introduction.................................................................................................................................................... 132

8.2. Geological characteristics of orogenic lode-gold deposits................................................................... 135

8.2 Timing of relationships in orogenic gold deposits....................................................................................... 137

8.3.2 Geochronological evidence ............................................................................................................... 138

8.4. Structural geometry of orogenic lode-gold deposits................................................................................. 143

8.6. Computer-based prospectivity mapping............................................................................................... 152

8.7. Comparison of Stress and Prospectivity Maps .......................................................................................... 160

8.9. Conclusions................................................................................................................................................ 162

8.10 Acknowledgements........................................................................................................................................ 164

8.11 References....................................................................................................................................................... 164

9. COMPUTER AIDED STRUCTURAL TARGETING IN MINERAL EXPLORATION: TWO

Abstract ................................................................................................................................................................... 173

9.1. Introduction..................................................................................................................................................... 173

9.2 Methods And Theoretical Considerations................................................................................................... 174

9.3 Stress Mapping................................................................................................................................................. 176

9.4 Discussion.......................................................................................................................................................... 183

9.5 Acknowledgments............................................................................................................................................ 185

9.6 References ......................................................................................................................................................... 185

10. TOWARDS A HOLISTIC EXPLORATION STRATEGY: USING GEOGRAPHIC

Abstract ................................................................................................................................................................... 189