Geology of the World’s Major Gold Deposits and Provinces [Special Publication ed.] 9781629493121, 9781629496429

Citation preview

Geology of the World’s Major Gold Deposits and Provinces

Richard H. Sillitoe, Richard J. Goldfarb, François Robert, and Stuart F. Simmons, Editors

Special Publication Number 23 Commemorating the 100th Anniversary of The Society of Economic Geologists, Inc.

Special Publication Number 23

Geology of the World’s Major Gold Deposits and Provinces

Richard H. Sillitoe, Richard J. Goldfarb, François Robert, and Stuart F. Simmons, Editors

Special Publication Number 23 Commemorating the 100th Anniversary of The Society of Economic Geologists, Inc.

Special Publications of the Society of Economic Geologists

Special Publication Number 23 Geology of the World’s Major Gold Deposits and Provinces Richard H. Sillitoe, Richard J. Goldfarb, François Robert, and Stuart F. Simmons, Editors

First Edition 2020 Printed by Allen Press, Inc. 810 East 10th St. Lawrence, Kansas 66044 www.allenpress.com

Additional copies of this publication can be obtained from Society of Economic Geologists, Inc. 7811 Shaffer Parkway Littleton, CO 80127 www.segweb.org/store

ISSN 1547-3112 (Print) 2639-1910 (Online) ISBN 978-1-629493-12-1 (Print) 978-1-629496-42-9 (Online)

© 2020 Society of Economic Geologists, Inc. On the cover: Front cover photo displaying high-grade Timiskaming conglomerate hosting folded carbonate-pyritequartz veins with visible gold, Dome mine (see paper by Dubé et al.); photo on the back cover is the Grasberg mine, taken in 2015, looking northeast and showing the ~2.5-km-diameter open pit and the surrounding area (see Leys et al.).

SPONSOR

The Society of Economic Geologists Publications Board thanks Barrick Gold Corporation for generous financial support of this volume.

SEG Publications Board Shaun L.L. Barker, Chair Pilar Lecumberri Sanchez Patrick Mercier-Langevin

Elizabeth R. Sharman Ross L. Sherlock

Brian G. Hoal, Executive Director, ex officio Lawrence D. Meinert, Editor, Economic Geology J. Bruce Gemmell, Editor, SEG Discovery

SOCIETY OF ECONOMIC GEOLOGISTS, INC.

Special Publication Number 23

Contents Papers and Appendices are on the flash drive attached to the inside back cover of this publication Sponsor

iii

Forward

ix

Preface

xi

Introduction

Richard H. Sillitoe

1

Stéphane De Souza, Stéphane Perrouty, Benoît Dubé, Patrick Mercier-Langevin, Robert L. Linnen, and Gema R. Olivo

29

   1 Gold Deposit Types: An Overview Precambrian Gold Deposits

   2 Metallogeny of the Neoarchean Malartic Gold Camp, Québec, Canada

   3 Gold Deposits of the World-Class TimminsPorcupine Camp, Abitibi Greenstone Belt, Canada    4 Hemlo Gold System, Superior Province, Canada    5 The Low-Grade, Neoproterozoic, Vein-Style, Carbonaceous Phyllite-Hosted Paracatu Gold Deposit, Minas Gerais, Brazil

Benoît Dubé, Patrick Mercier-Langevin, John Ayer, Jean-Luc Pilote, and Thomas Monecke

53

K. Howard Poulsen, Rodney Barber, and François Robert

81

Nicholas H.S. Oliver, Brian Thomson, 101 Flavio H. Freitas-Silva, and Rodney J. Holcombe

   6 The Super-Giant, High Grade, Paleoproterozoic Metasedimentary Rock- and Shear Vein-Hosted Obuasi (Ashanti) gold deposit, Ghana, West Africa

Nicholas H.S. Oliver, Andrew Allibone, 121 Michael J. Nugus, Carlos Vargas, Richard Jongens, Richard Peattie, and Vaughan A. Chamberlain

   7 Paleoproterozoic Gold Deposits of the Loulo District, Western Mali

Andrew Allibone, David Lawrence, John Scott, 141 Mark Fanning, James Lambert-Smith, Paul Stenhouse, Reinet Harbidge, Carlos Vargas, Rose Turnbull, and Joel Holliday

   8 The World-Class Gold Deposits in the Geita Greenstone Belt, Northwestern Tanzania

P.H.G.M. Dirks, I. V. Sanislav, M. R. van Ryt, 163 J-M Huizenga, T. G. Blenkinsop, S. L. Kolling, S. D. Kwelwa, and G. Mwazembe

   9 Orogenic Gold Deposits of the Kibali District, Neoarchean Moto Belt, Northeastern Democratic Republic of Congo

Andrew Allibone, Carlos Vargas, Etienne Mwandale, 185 Justus Kwibisa, Richard Jongens, Sarah Quick, Nathan Komarnisky, Mark Fanning, Philip Bird, Doug MacKenzie, Rose Turnbull, and Joel Holliday

   10 Olympiada Gold Deposit, Yenisei Ridge, Russia

A. M. Sazonov, K. V. Lobanov, E. A. Zvyagina, 203 S. I. Leontiev, S. A. Silyanov, N. A. Nekrasova, A. Y. Nekrasov, A. B. Borodushkin, V. A. Poperekov, V. V. Zhuravlev, S. S. Ilyin, Yu.A. Kalinin, A. A. Savichev, and A. S. Yakubchuk v

   11 The Telfer Gold-Copper Deposit, Paterson Province, Western Australia

Alan J. Wilson, Nick Lisowiec, Cameron Switzer, Anthony C. Harris, Robert A. Creaser, and C. Mark Fanning

227

   12 Geologic Setting and Gold Mineralization of the Kalgoorlie Gold Camp, Yilgarn Craton, Western Australia

Jordan A. McDivitt, Steffen G. Hagemann, Matthew S. Baggott, and Stuart Perazzo

251

   13 Boddington: An Enigmatic Giant Archean GoldCopper (Molybdenum-Silver) Deposit in the Southwest Yilgarn Craton, Western Australia

Stephen J. Turner, Graeme Reynolds, and Steffen G. Hagemann

275

Warwick S. Board, Duncan F. McLeish, Charles J. Greig, Octavia E. Bath, Joel E. Ashburner, Travis Murphy, and Richard M. Friedman

289

Paul J. Dobak, François Robert, Shaun L.L. Barker, Jeremy R. Vaughan, and Douglas Eck

313

   16 Giant Carlin-Type Gold Deposits of the Cortez District, Lander and Eureka Counties, Nevada

Mark A. Bradley, L. Page Anderson, Nathan Eck, and Kevin D. Creel

335

   17 Epithermal Gold Deposits Related to Alkaline Igneous Rocks in the Cripple Creek District, Colorado, United States

Karen D. Kelley, Eric P. Jensen, Jason S. Rampe, and Doug White

355

   18 Geology of Round Mountain, Nevada: A Giant Low-Sulfidation Epithermal Gold Deposit

David A. Rhys, Nadia St. Jean, Rodolfo Lagos, David Emmons, George A. Schroer, and Richard Friedman

375

Omar Dromundo, Sigfrido Robles, Thomas Bissig, Claudio Flores, Maria del Carmen Alfaro, and Lorenzo Cardona

399

Jeremy Vaughan, Carl E. Nelson, Guillermo Garrido, Jose Polanco, Valery Garcia, and Arturo Macassi

415

Stephen Leary, Richard H. Sillitoe, Jorge Lema, Fernando Téliz, and Diego Mena

431

Richard Pilco and Sean McCann

451

T. Baker, S. Mckinley, S. Juras, Y. Oztas, J. Hunt, L. Paolillo, S. Pontual, M. Chiaradia, A. Ulianov, and D. Selby

467

Reimar Seltmann, Richard J. Goldfarb, Bo Zu, Robert A. Creaser, Alla Dolgopolova, and Vitaly V. Shatov

497

G. L. Vursiy, I. A. Zibrov, S. G. Lobov, and A. S. Yakubchuk

523

Takayuki Seto, Yu Yamato, Ryota Sekine, and Eiji Izawa

545

Phanerozoic Gold Deposits

   14 The Brucejack Au-Ag Deposit, Northwest British Columbia, Canada: Multistage Porphyry to Epithermal Alteration, Mineralization, and Deposit Formation in an Island-Arc Setting    15 Goldstrike Gold System, North Carlin Trend, Nevada, USA

   19 The Peñasquito Gold-(Silver-Lead-Zinc) Deposit, Zacatecas, Mexico    20 The Pueblo Viejo Au-Ag-Cu-(Zn) Deposit, Dominican Republic    21 Geology of the Fruta del Norte Epithermal Gold-Silver Deposit, Ecuador    22 Gold Deposits of the Yanacocha District, Cajamarca, Peru   23 Alteration, Mineralization, and Age Relationships at the Kıs¸ladag˘ Porphyry Gold Deposit, Turkey    24 Muruntau, Uzbekistan: The World’s Largest Epigenetic Gold Deposit    25 The Sukhoi Log Gold Deposit, Russia    26 Geology of the Hishikari Gold Deposit, Kagoshima, Japan

vi

   27 Geology of the Porgera Gold Deposit, Papua New Guinea

Jonathan P. Hay, Mark M. Haydon, and François Robert

559

   28 Lihir Alkalic Epithermal Gold Deposit, Papua New Guinea

David R. Cooke, Stephanie Sykora, Erin Lawlis, Jacqueline L. Blackwell, Mathieu Ageneau, Nicholas H. Jansen, Anthony C. Harris, and David Selley

579

   29 Grasberg Copper-Gold-(Molybdenum) Deposit: Product of Two Overlapping Porphyry Systems

Clyde Leys, Adam Schwarz, Mark Cloos, Sugeng Widodo, J. Richard Kyle, and Julius Sirait

599

   30 Geologic Evolution of Late Ordovician to Early Silurian Alkalic Porphyry Au-Cu Deposits at Cadia, New South Wales, Australia

Anthony C. Harris, David R. Cooke, Ana Liza Garcia Cuison, Malissa Groome, Alan J. Wilson, Nathan Fox, John Holliday and Richard Tosdal

621

Hartwig E. Frimmel and Glen T. Nwaila

645

Benoît Dubé and Patrick Mercier-Langevin

669

Gerard I. Tripp, Richard M. Tosdal, Thomas Blenkinsop, Jamie R. Rogers, and Scott Halley

709

Nicolas Thébaud, Andrew Allibone, Quentin Masurel, Aurélien Eglinger, James Davis, Anne-Sylvie André-Mayer, John Miller, Morou François Ouedrago, and Mark Jessell

735

Kun-Feng Qiu, Richard J. Goldfarb, Jun Deng, Hao-Cheng Yu, Zong-Yang Gou, Zheng-Jiang Ding, Zhao-Kun Wang, and Da-Peng Li

753

Gold Provinces

   31 Geologic Evidence of Syngenetic Gold in the Witwatersrand Goldfields, South Africa    32 Gold Deposits of the Archean Abitibi Greenstone Belt, Canada    33 Neoarchean Eastern Goldfields of Western Australia

   34 The Paleoproterozoic (Rhyacian) Gold Deposits of West Africa

   35 Gold Deposits of the Jiaodong Peninsula, Eastern China

   36 Carlin-Type Gold Deposits in Nevada: Geologic Characteristics, Critical Processes, and Exploration    37 Giant Placers of the Upper Kolyma Gold Fields, Yana-Kolyma Province, Russian Northeast

John L. Muntean

775

N. A. Goryachev, A. S. Yakubchuk, I. S. Litvinenko, A. V. Lozhkin, Yu.V. Pruss, and V. N. Smirnov

797

Stuart F. Simmons, Benjamin M. Tutolo, Shaun L.L. Barker, Richard J. Goldfarb, and François Robert

823

Epilogue

   38 Hydrothermal Gold Deposition in Epithermal, Carlin, and Orogenic Deposits

vii

©2020 Society of Economic Geologists, Inc. SEG Special Publications, no. 23, pp. ix

Foreword Mark Bristow Ph.D. President and Chief Executive Barrick Gold Corporation

Discovery is the source of real value creation in the gold mining industry, which makes exploration our equivalent of a pharmaceutical company’s research and development department: the foundation of the business as well as the driver of its growth. The 100th anniversary of the SEG is an opportune occasion to highlight the critical role of geology in the complexity of modern mining, both as an essential pioneering endeavor and as the provider of the research, analysis, and modeling that will determine the viability of a new discovery and then point the way for its optimal development. At a time when greenfields discoveries are becoming scarcer, it plays an equally important part in extending known asset bases. For Barrick and for me as a geologist, it is a great pleasure to sponsor this landmark volume, thereby continuing our long association with exploration’s foremost professional

doi: 10.5382/SP.23.fwd; 1 p.

ix

organization. Barrick and our legacy company, Randgold, have a long history of exploration success—we contributed to seven of the 29 deposit descriptions featured in these pages—while I have always been committed to putting the geosciences at the core of our operations. This special volume gives geology its deserved due and provides a timely insight into the world’s major gold deposits and provinces. It will be a highly valuable, long-lasting reference for all geoscience practitioners of this and future generations. I would like to express my appreciation to the editors, Richard H. Sillitoe, Richard J. Goldfarb, François Robert, and Stuart F. Simmons, and to thank the authors for their contributions, along with Brian G. Hoal, Alice Bouley, and Mabel Peterson of the Society of Economic Geologists. In conclusion, I also want to pay tribute to the men and women who walk the hills in search of their next discovery.

©2020 Society of Economic Geologists, Inc. SEG Special Publications, no. 23, pp. xi–xii

Geology of the World’s Major Gold Deposits and Provinces: Preface For economic geologists, be they from academia or industry, gold is an important and fascinating commodity that demands a great deal of their collective research and exploration attention. The metal accounts for roughly half of the world’s nonferrous exploration activity, which equates to corporate expenditure on the order of US$5 billion annually. Although present-day research and exploration employ a myriad of specialized tools and techniques, the basic geologic features of gold deposits—lithologic and structural controls, alteration zoning, lithogeochemistry, and mineralogy—provide the foundation stones for both laboratory studies of and search for gold deposits, and will undoubtedly continue to do so. It is for this reason that successive SEG Publications Board chairs, Stuart Simmons and Rich Goldfarb, promoted the idea of a volume focused on basic geologic descriptions of the world’s major gold deposits and provinces as an ideal contribution to the Society’s 100th anniversary celebrations. Selection of the deposits and provinces for inclusion in the volume—SEG Special Publication 23—was no easy task, but total gold content and high average gold tenor were the most influential parameters. Although a number of abandoned deposits, including Homestake in the United States, Kolar in India, and Morro Velho in Brazil, are historically important, it was decided to focus on currently or soon-to-be producing mines and provinces. Some of these (Cripple Creek, Kalgoorlie, Obuasi, Timmins, Witwatersrand) have long and illustrious histories whereas others, including all the low-grade deposits, entered production in the last few decades. Indeed, Fruta del Norte came on stream since the paper was written and Sukhoi Log, although long known, is still at the feasibility stage. Orogenic gold deposits and provinces dominate the volume’s content, in keeping with their preeminence for global gold endowment. Nonetheless, the geology of the world’s largest gold concentration, the Witwatersrand goldfields of South Africa, also features prominently. It could be perceived that epithermal and porphyry deposits are somewhat underrepresented, given their worldwide abundance, the reason being the relatively small size and low grade, respectively, of many of them.

doi: 10.5382/SP.23.pre; 2 p.

The volume is introduced with a series of thumbnail sketches that attempt to give a general flavor of the principal currently recognized gold deposit types, several of them insufficiently important globally to merit representation, and concludes with a consideration of gold transport mechanisms relevant to the major deposit types. The meat of the volume, however, is in the 29 deposit descriptions and seven province overviews. Each description summarizes exploration history and regional and local geologic settings preparatory to synthesizing the salient lithologic, structural, alteration, and mineralization features of the deposit itself. The province papers address terrane-scale geologic parameters and their controls on the localization, styles, and timing of gold mineralization. As editors of the volume, we have enjoyed the disparate challenges presented over the past three years and would like to express our appreciation to the authors of the papers— compliant and recalcitrant alike—for their contributions, many prepared during serious competition with corporate duties. We are also indebted to the 63 specialists, listed below, who peer reviewed the manuscripts in timely fashion, to Alice Bouley, Managing Editor of SEG publications, for overseeing production of the volume, to Mabel Peterson, for careful copyediting, and to Laura Doll and Vivian Smallwood, for meticulous handling of layout. Last but not least, the Society offers its sincere thanks to Barrick Gold Corporation for generous sponsorship of the volume. The Editors Richard H. Sillitoe Independent consultant, London Richard J. Goldfarb China University of Geosciences, Beijing and Colorado School of Mines, Golden, Colorado François Robert Independent consultant, Montreal Stuart F. Simmons Hot Solutions Ltd., Auckland and University of Utah, Salt Lake City, Utah

xi

Reviewers of SEG Special Publication 23 Andrew Allibone Tim Baker Marc Bardoux Dave Boden Dave Braxton Patrick Browne Graham Carman Rob Chapman Odin Christensen Tony Christie Dave Craw Stéphane DeSouza Lluís Fontboté Louis Gauthier Lynnette Greyling David Groves Steffen Hagemann Scott Halley Jeff Hedenquist Ken Hickey David John

Imants Kavalieris Steve Kesler Doug Kreiner Bruno Lafrance James Lambert-Smith Ross Large Bernd Lehmann Shoufa Lin Robert Linnen Rael Lipson Tony Longo James MacDonald Quentin Masurel Peter Megaw Cam McCuaig Lawrie Minter Robert Moritz Jim Mortensen John Muntean Miguel Nassif Evgeniy Naumov

xii

Gema Olivo Pepe Perelló Stéphane Perrouty Franco Pirajno Howard Poulsen Stewart Redwood Mike Ressel Dave Rhys Robbie Rowe Norman Russell Ross Sherlock Mike Skead Sachihiro Taguchi John Thompson Tommy Thompson Steve Turner Peter Vikre Yasushi Watanabe Willy Williams-Jones Wally Witt Roberto Xavier

©2020 Society of Economic Geologists, Inc. SEG Special Publications, no. 23, pp. 1–28

Chapter 1 Gold Deposit Types: An Overview Richard H. Sillitoe† 27 West Hill Park, Highgate Village, London N6 6ND, England

Abstract Gold is either the only economically important metal or a major by-product in 11 well-characterized deposit types—paleoplacer, orogenic, porphyry, epithermal, Carlin, placer, reduced intrusion related, volcanogenic massive sulfide (VMS), skarn, carbonate replacement, and iron oxide-copper-gold (IOCG), arguably more than for those of any other metal; it also dominates a number of deposits of uncertain or unknown origin. Major gold concentrations formed worldwide from the Mesoarchean to the Pleistocene, from Earth’s surface to midcrustal paleodepths, alone or in association with silver, base metals, and/or uranium, and from hydrothermal fluids of predominantly metamorphic, magmatic, meteoric, seawater, or, uncommonly, basinal origins, as well as from mafic magma or ambient surface water. Most of the Neoproterozoic and Phanerozoic deposits unequivocally formed in accretionary orogens. As an introduction to this compilation of the world’s major gold deposits and provinces, this paper provides a thumbnail sketch of each gold deposit type, including geologic and economic characteristics and widely accepted genetic models, as well as briefly discusses aspects of their spatial and temporal associations and distributions.

Introduction Gold can be concentrated hydrothermally, magmatically, and mechanically and is either the principal commodity or major by-product in more ore deposit types worldwide than any other metal. This geochemical versatility reflects its well-known siderophile, chalcophile, and sulfophile character along with the fact that it can be transported under hydrothermal conditions in both liquid and vapor states and as aqueous sulfur and chloride complexes (Williams-Jones et al., 2009, and references therein). Notwithstanding its resistance to supergene oxidation and high density (19.3 g/ cm3)—leading to surficial mechanical concentration, gold mobilization can also be microbially mediated (Sanyal et al., 2019), a process that is especially effective during lateritic weathering (Freyssinet et al., 2005). Nanoparticulate (colloidal) gold transport is an additional possibility in both the hypogene and supergene environments (Hough et al., 2011). Furthermore, there also are a variety of effective mechanisms for hydrothermal gold deposition, including boiling, mixing, sulfidation, and sorption onto precipitating arsenian pyrite (Simmons et al., 2020, and references therein), as well as the possibility of further upgrading by fluid-mediated chalcophile element-rich melt remobilization during subsequent deformation and metamorphism (Hastie et al., 2020). Much of Earth’s gold endowment is considered by Frimmel (2008) to have been delivered from the mantle to the crust in Mesoarchean times, with ~40% of the known total having been preserved in the Witwatersrand Basin of South Africa. Since that time, gold deposits of different types have formed in diverse plate tectonic settings, but particularly during subduction and collision events in accretionary orogens. In order of decreasing endowment and overall economic importance, gold deposits may be assigned to several widely recognized categories: † Corresponding

author: e-mail, [email protected]

doi: 10.5382/SP.23.01; 28 p. 1

paleoplacer, orogenic, porphyry, epithermal, Carlin type, placer, reduced intrusion related, volcanogenic massive sulfide (VMS), skarn, carbonate replacement, iron oxide-copper-gold (IOCG), and other minor types. In addition, some deposits— including a number of giants—either have equivocal geologic features that hinder unique classification or are simply indecipherable because of the effects of intense postmineral deformation and metamorphism. As an introduction to the major gold deposits and provinces described in this volume, a brief overview of this spectrum of deposit types is presented, with emphasis on their main subtypes, mineralization styles, origins, exploration factors, and exploitability. Formational depths and relationships among deposit types (Fig. 1) are emphasized throughout. A selection of the largest and/or highest grade gold deposits of each type, including those covered in this volume that are reliably assignable, is tabulated, with measured and indicated (but not inferred) resources plus any past production used as the basis for the overall size and grade estimates. The worldwide distribution of the tabulated deposits is shown in Figure 2. Paleoplacer Gold Deposits Quartz pebble conglomerate-hosted paleoplacer deposits may be divided into two types: those older than ~2.8 Ga, principally in the Witwatersrand Basin, South Africa, that include highgrade gold concentrations in stratiform, kerogen-rich “carbon seams”—fossilized microbial mats (Mossman et al., 2008), which were mechanically reworked to generate the intimately associated but unusually fine-grained (submillimeter) placer gold; and the much smaller and younger deposits containing exclusively placer gold (e.g., Frimmel and Hennigh, 2015; Table 1). In some Witwatersrand conglomerate horizons (reefs), the carbon seams can account for as much as 50% of the contained gold (Hallbauer and Joughin, 1973). The diametrically opposed placer and epigenetic hydrothermal origins for the earliest paleoplacer deposits

2

RICHARD H. SILLITOE

0

HS EPITHERMAL Au LS EPITHERMAL Au IS EPITHERMAL Au VMS Au

PLACER Au

km

Volcanic rocks

V

V V

V V

V

V V

V

V V

V

V V

V

V

V

V V

V V

V V

V

V V

V

V V

CRD Au SKARN Au CARLIN-TYPE Au REDUCED INTRUSIONRELATED Au Carbonate (± SKARN) rocks Limestone

1

V V

Volcanic rocks

Turbidite Reduced felsic equigranular intrusion Greenstone

Seawater

PORPHYRY Cu-Au/Au

IOCG

Porphyry intrusion POSSIBLE SKARN/CRD Au Inferred felsic intrusion

5

10

V V

Inferred felsic intrusion

Dioritic intrusion

OROGENIC Au Granitoid

Shear zone

Fig. 1. Schematic geologic settings and interrelationships of principal gold deposit types (except paleoplacer), inspired by Robert et al. (2007). The approximate depth scale is logarithmic. Note that most placer gold is derived by erosion of orogenic deposits. CRD = carbonate-replacement deposit, HS = high sulfidation, IS = intermediate sulfidation, IOCG = iron oxidecopper-gold, LS = low sulfidation, and VMS = volcanogenic massive sulfide.

can be partially rationalized in the context of a syngenetic model that accords rather better with geologic observations (Horscroft et al., 2011; Frimmel, 2014; Frimmel and Hennigh, 2015; Heinrich, 2015; Frimmel and Nwaila, 2020).

Much of the early gold is believed to have been leached from the Mesoarchean land surface by volcanically triggered acid rain, delivered in solution by braided river systems to lowenergy, riverine and nearshore environments, and efficiently

Fort Knox Dublin Gulch Nome Klondike Greens Creek

Pogo Pebble Donlin Creek

Brucejack Nickel Plate

Boliden Hemlo

Homestake

GREAT BASIN

Leadville Cripple Creek Peñasquito

Twin Creeks Getchell Phoenix Grass Valley Sierra Foothills Comstock

Morelos Goldstrike + Deep Star Bingham Canyon Gold Quarry Goldrush Round Mountain Goldfield

Los FilosBermejal Lagunas Norte Yanacocha (La Quinua) Pataz-Parcoy Tipuani Candelaria Pascua-Lama

ABITIBI BELT LaRonde Penna + Horne Bousquet 2 Kirkland Lake Canadian Timmins Malartic

Rosia ¸ Montana Kisladag ¸

Pardo La Colosa

Loulo

Beiya

Salobo Igarapé Bahia-Alemão

Hishikari

Kolar Ok Tedi Grasberg Lihir

Kibali Tarkwa Obuasi

Geita Fortescue Basin

Jacobina Morro Velho Paracatu

Linglong Shuiyindong

Far Southeast

Morila

Tongon

Upper Kolyma (Natalka)

Zarmitan

Olympias

Nambija Fruta del Norte

El Indio

Vasilkovskoye Muruntau

Salave

Côté Gold Pueblo Viejo

Olympiada Sukhoi Log

ˆ

Eskay Creek

Porgera

Callie Prominent Hill Olympic Dam Carrapateena Cadia East

Telfer Witwatersrand

Golpu Solwara 1

Kalgoorlie Boddington

Waihi Victoria (Bendigo)

Paleoplacer Orogenic Porphyry

Epithermal Carlin-type Placer

Reduced intrusion related Volcanogenic massive sulfide Skarn + carbonate replacement

Iron oxide-copper-gold Disputed: orogenic or intrusion related

Fig. 2. Worldwide distribution of gold deposits tabulated in this paper (and described elsewhere in this volume), categorized by deposit type.



Whymark and Frimmel (2018) Reduced Minor provided by H. Frimmel (writ. commun., 2018) Creek prospect 2Beatons

1Estimate

Lower greenschist Polymictic conglomerate 2450–2400

Fortescue Basin, Western Australia Pardo, Ontario, Canada

Jacobina, Brazil

3

precipitated on the microbial mats during prolonged sedimentary hiatuses (Horscroft et al., 2011; Frimmel, 2014; Frimmel and Hennigh, 2015; Heinrich, 2015). The microbial mats, although now represented by carbon seams only millimeters to centimeters thick (Fig. 3), can be highly auriferous over tens of square kilometers (Horscroft et al., 2011). Such major, early gold concentrations, now widely preserved only in the Witwatersrand Basin because of concealment beneath thick flood basalts, along with some of the oldest orogenic and perhaps other gold deposit types (e.g., VMS; Long et al., 2011), were reworked to form the younger paleoplacers: the earlier ones under anoxic conditions and characterized by detrital and synsedimentary pyrite and uraninite and those later than the Great Oxidation Event (~2.4 Ga) containing detrital iron oxides and no uraninite (e.g., Tarkwa; Table 1). The paleoplacer deposits were variably modified by local metamorphic and hydrothermal gold remobilization and redistribution, as exemplified by the Witwatersrand (e.g., Robb and Meyer, 1991). The relatively small number of known paleoplacer gold deposits and prospects worldwide is the main reason for a general lack of greenfield exploration attention, although recently there has been a notable resurgence of interest in 2.8 to 2.6 Ga marine paleoplacer prospects in the Fortescue Basin of the Pilbara craton, Western Australia (Hennigh, 2018; Table 1). Most paleoplacer gold, especially that in the Witwatersrand carbon seams and derivative pebble lags, is recovered by highly selective, labor-intensive underground mining, with grades in parts of the Witwatersrand Basin sufficiently high (20–40 g/t) to enable extraction to extreme depths of ~4 km, the absolute economic limit. However, bulk, open-pit mining is conducted at the Tarkwa paleoplacer (Wright, 1997; Table 1). Although the Witwatersrand paleoplacer system accounts for ~40% of Earth’s known gold endowment (Frimmel, 2008), mined production has diminished dramatically over the past decade as a result of declining ore grades, high production costs, labor unrest, and an increasingly restrictive regulatory environment, resulting in South Africa losing its long-standing status as the world’s leading producer; it is currently the seventh largest.

Absent

Hennigh (2018) Reduced Present Polymictic conglomerate 2780–2750

Absent

Teles et al. (2015) Reduced Present Absent

Greenschist-lower amphibolite Subgreenschist Quartz pebble conglomerate

~2105–2075

>2400

Oxidized Absent Upper greenschist

Absent

Au-bearing carbon seams Important Metamorphic facies Lower greenschist

Host rock Quartz pebble conglomerate, sandstone Quartz pebble conglomerate Age (Ma) ~2940–2714

Size, Moz Au; grade, g/t Au 3,086; ~101 41; 1.2 10.3; 2.6 0.5; 2.12 No resources Deposit, location Witwatersrand, South Africa Tarkwa, Ghana

Table 1. Noteworthy Paleoplacer Gold Deposits

Detrital uraninite Present

Redox state Reduced

Reference Frimmel and Nwaila (2020) Wright (1997)

GOLD DEPOSIT TYPES: AN OVERVIEW

Orogenic Gold Deposits Formation of lode gold deposits in metamorphic terranes— now widely referred to as orogenic deposits as they typically formed during late stages of regional-scale orogeny (Groves et al., 1998)—first peaked in the Neoarchean (2.75–2.55 Ga; Goldfarb et al., 2001, 2005), mainly in the Superior and Slave Provinces of Canada and Eastern Goldfields Province of Western Australia, as well as other cratons in Africa, India, and South America (Groves et al., 2005; Robert et al., 2005; Table 2; Fig. 2). The other major orogenic gold deposits formed during the Paleoproterozoic (2.1–1.75 Ga) and Neoproterozoic-Phanerozoic (Goldfarb et al., 2001, 2005; Table 2; Fig. 2). Two broad lithologic hosts for major orogenic gold deposits are widely recognized (Fig. 1): volcano-sedimentary greenstone sequences, including chemically receptive banded iron formations (BIFs), and turbidite-dominated accretionary terranes (slate belts). The first lithologic association accounts for much of the Neoarchean and Paleoproterozoic orogenic

4

RICHARD H. SILLITOE

Au

Fig. 3. High-grade gold (Au) in kerogen-rich carbon seam from Witwatersrand paleoplacer deposit (from Frimmel, 2014). The seam (black) rests on an erosion surface and is overlain by conglomerate.

gold, including that in the main provinces noted above as well as in the Birimian belts of West Africa (Goldfarb et al., 2017; Thébaud et al., 2020), whereas the turbidite-hosted deposits are more characteristic of the Paleozoic (e.g., Bendigo, Australia, and Meguma terrane, eastern Canada) and Mesozoic-Eocene (e.g., Natalka, Russian Far East; Table 2). Many deposits comprise relatively sulfide-poor quartz ± carbonate veins, commonly associated with white mica (sericite)/biotite-carbonate-pyrite alteration and an As-Sb-W ± Te geochemical signature. The veins contain high-grade (5–>12 g/t Au) ore shoots, with large downplunge extents suitable for selective underground mining, in places to depths of >2 km (3.2 km at Kolar; Table 2). However, a minority of deposits, or parts thereof, contain quartz-poor, disseminatedor replacement-style sulfide mineralization, which can be metallurgically refractory (e.g., Callie, Kibali, Natalka, and Obuasi; Table 2). Orogenic gold deposits are widely believed to have formed in accretionary or collisional orogens at paleodepths of ~5 to 15 km (Fig. 1) from low-salinity, gold- and arsenic-bearing, aqueous carbonic fluids generated by devolatilization reactions accompanying the downward transition from regional greenschist- to amphibolite-facies metamorphism (Goldfarb et al., 2005; Pitcairn et al., 2006, 2014; Phillips and Powell, 2010; Tomkins, 2013; Gaboury, 2019). Nonetheless, even more deeply sourced metamorphic fluids released from oceanic crust and overlying sediments in subduction zones or overlying fertilized mantle lithosphere are being increasingly considered (Goldfarb and Groves, 2015; Groves et al., 2020a). Indeed, irrespective of their ultimate origin, gold-rich fluids were recently documented in amphibolite-facies rocks at midcrustal (17–18 km) paleodepths (Prokofiev et al., 2020). The auriferous fluids are normally channeled upward through dilatant portions of orogen-parallel, deep-crustal fault systems in which lower order splays and associated favorable structures (e.g., anticlinal closures) are mineralized during repetitive, seismically assisted reverse displacement under brittle-ductile conditions (Sibson et al., 1988; Robert et al., 1995, 2005;

Goldfarb et al., 2005; Table 2). Late-orogenic changes from compression to transpression may have been instrumental in triggering the formation of some deposits (e.g., Goldfarb et al., 2008; Allibone et al., 2020b; Goryachev et al., 2020; Thébaud et al., 2020). Instead of being late orogenic in timing, a subset of metasedimentary rock-hosted deposits formed early in the orogenic cycle, as documented in the case of Paracatu and early stages of Obuasi (Oliver et al., 2020a, b; Table 2), or could even have preorogenic precursors (e.g., Sukhoi Log and Olympiada; Large et al., 2007; Sazonov et al., 2020; Vursiy et al., 2020; Table. 2). Multiple, vertically extensive veins mined to great depths account for the large gold endowments of several deposits in the Superior and Eastern Goldfields Provinces and elsewhere (e.g., Hollinger-McIntyre, Kolar, Morro Velho, and Obuasi; Table 2). Nonetheless, since the 1980s, open-pit mining of the shallower parts of deposits has become commonplace in the Eastern Goldfields Province as well as elsewhere (e.g., Birimian belts, Guiana shield, and Paracatu). These bulk, lowgrade (28 Moz of gold, is the most notable exception (Lodder et al., 2010; Table 3). Gold-rich porphyry deposits have the same overall geologic settings and characteristics as most porphyry deposits. They are concentrated in continental (Cordilleran) and island arcs and, although most formed during active subduction (Sillitoe, 1972), a minority postdate subduction and are synto postcollisional in timing (Sillitoe, 2000; Hou et al., 2009; Richards, 2009). The deposits are generally centered on multiphase, chemically oxidized, calc-alkaline porphyry stocks or dikes (Fig. 4), typically dioritic/monzodioritic to quartz dioritic in composition (Table 3); however, they can also be more felsic and, locally in British Columbia and elsewhere,

50.3; 4–4.6 50.0; 1.4

66; ~122

23; 4.11

15; ~1.5

17; 4.5

25.4; 16.5

15.2; 7.2 16.3; ~0.42

Olympiada, Siberia, Russia Natalka, Magadan, Russia

Obuasi (Ashanti), Ghana

Kibali, D.R.Congo

Geita, Tanzania

Loulo, Mali

Kolar, India

Morro Velho, Brazil Paracatu, Brazil

2

1

63; 2.1

Sukhoi Log, Siberia, Russia

Schistose quartz-carbonate rock (lapa seca) Carbonaceous phyllite

Schistose amphibolite, BIF

Metasedimentary rocks

Ironstone, diorite

Carbonaceous phyllite, slate, metasandstone, volcanic rocks Siliciclastic metasedimentary rocks, chert, BIF

Calcareous and carbonaceous schists, marble Carbonaceous shale

Carbonaceous slate + phyllite

Graywacke, slate

Fe-rich and Fe-poor metasiltstone

Tholeiitic basalt, minor carbonaceous mudstone

Granodiorite, serpentinite, mélange

Host rock Carbonate-facies BIF

Resource grade Estimate provided by N.H.S. Oliver (writ. commun., 2019)

~630

~2670

~2550

2090–2060

~2640

2630–2600

2100–2090

140–135

810–750(?)

~530(?)

~445

22; 8.0

2660–2640

33.6; 9.3 1805

~160

13; 6.2

14.2; 5.61

Age (Ma) 1730

Size, Moz Au; grade, g/t Au 39.8; 8.8

Callie, Northern Territory, Australia Bendigo, Victoria, Australia

Deposit, location Homestake, South Dakota, United States Grass Valley, California, United States HollingerMcIntyre, Ontario. Canada

Low-angle thrust zone

Steep isoclinal folds

Shear zone intersections

Shear zones, sandstone bed

Dike contacts, fault intersections, F3 fold hinges Fold axes, foliation, shear zones, BIF horizons Folds, late faults, intrusive contacts

Reverse faults, extension fractures + saddle reefs Recumbent anticlinal crest, parasitic folds Fold hinges + intersecting faults Strike-slip fault zone

Reverse-oblique fault, D3 shearing, anticline, porphyry intrusions Anticline cut by D2 high-strain zone

Structural control Moderately plunging isoclinal synclinal folds Minor thrust faults

Lower greenschist

Greenschist

Middle to upper amphibolite

Upper greenschist

Upper greenschist

Lower-middle greenschist

Greenschist

Greenschist to amphibolite Greenschist

Lower greenschist

Lower greenschist

Greenschist

Zeolite-lower greenschist

Metamorphic facies Upper greenschist-lower amphibolite Greenschist

White mica, quartz, siderite, ankerite

Carbonate, white mica

Diopside, hornblende, biotite, calcite

Albite, white mica, carbonate, tourmaline

Carbonate, quartz, K-feldspar, biotite

Ankerite/siderite ± quartz

Quartz, carbonate, white mica Quartz, albite, white mica, chlorite, carbonates Quartz, carbonate, white mica, chlorite

Siderite, white mica

White mica

Chlorite, white mica

Ankerite-calcitewhite mica±albite

White mica

Ore-related alteration Chlorite, siderite, white mica

Table 2. Noteworthy Orogenic Gold Deposits

Disseminated to semimassive sulfide, minor quartz veins Quartz-carbonate stockworks, breccias, disseminations Quartz-calcite veins, subordinate sulfidized BIF horizons Massive + disseminated sulfide lodes, quartz veins Boudinaged quartzcarbonate-sulfide veins + veinlets

Pyrite veinlets + disseminations

Pyrite-quartz veinlets + disseminations, minor quartz ± carbonate veins Strata-bound disseminated sulfides Quartz ± carbonate stockworks + veins, disseminations Quartz veins, disseminated sulfides

Sheeted quartz veins + strata-bound disseminations Quartz veins + saddle reefs

Quartz-carbonate vein network

Mineralization style Tabular to pipe-like bodies, including quartz veins + sulfidic replacements Quartz-carbonate veins

Oliver et al. (2020b)

Vial et al. (2007)

Hamilton and Hodgson (1986)

Allibone et al. (2020b)

Dirks et al. (2020)

Allibone et al. 2020a)

Oliver et et al. (2020a)

Eremin et al. (1994)

Sazonov et al. (2020)

Vursiy et al. (2020)

Phillips and Hughes (1996)

Petrella et al. (2020)

Dubé et al. (2020)

Taylor et al. (2015)

Reference Caddey et al. (1991)

GOLD DEPOSIT TYPES: AN OVERVIEW

5

2Breccia-dominated

porphyry deposit deposit

2740

10.0; 0.87

1Gold-only

8

28.3; 0.84

La Colosa1, Colombia Côté2, Ontario, Canada

14.5

17.4; 0.62

38

31.4; 0.39

1.4

19.8; 0.70

90

8.7

18.6; 0.70

71; 0.34

1.1

18.1; 0.50

~440

0.5

27; 0.49

38.7; 0.38

Age (Ma) 3.3

Size, Moz Au; grade, g/t Au 96; 0.57

Kişladağ1, Turkey

Deposit, location Grasberg, Papua, Indonesia Onto, Sumbawa, Indonesia Ok Tedi, Papua New Guinea Golpu, Papua New Guinea Far Southeast, Philippines Cadia East, NSW, Australia Pebble, Alaska, United States Bingham Canyon, Utah, United States Schist, gneiss, volcaniclastic rocks Greenschist-facies metasedimentary rocks Tonalite, diorite

Greenschist-facies metasedimentary rocks Basaltic volcanic and volcaniclastic rocks Volcaniclastic conglomerate, basaltic lava Granodiorite, diorite, siltstone Precursor monzonite, quartzite

Host rocks Limestone, siliciclastic sedimentary rocks Diatreme in andesitic volcanic rocks Siltstone, limestone

None exposed

Diorite

Monzonite

Quartz monzonite

Alkalic quartz monzonite Granodiorite

Quartz diorite

Diorite

Diorite/quartz diorite (?) Monzodiorite

Main mineralized porphyry Monzodiorite

Potassic, white mica, albite

Potassic, white mica-tourmaline Potassic, sodic-calcic

Potassic, sodic-potassic, pyrophyllite-white mica Potassic

Potassic, chlorite-white mica-albite Potassic, calcic-potassic

Potassic

Potassic

Main ore-related alteration Potassic, chlorite-white mica Advanced argillic

Cu, Mo

Cu

Mo

Cu, Mo, Ag

Cu, Mo, Ag

Cu, Mo

Cu

Cu

Cu

Cu

Other contained metals Cu, Mo

Table 3. Noteworthy Gold-Rich Porphyry Deposits, Including Breccias

Vein Au

Skarn Cu-Au, CRD + vein Zn-Pb-Ag-Au, Carlin-type Au

HS + IS disseminated Au HS Cu-Au bodies, IS Au-Zn-Pb veins Skarn Fe-Cu-Au

Skarn Cu-Au

Genetically related deposit type(s) Skarn Cu-Au

Katz et al. (2017)

Lodder et al. (2010)

Baker et al. (2020)

Porter et al. (2012)

Lang et al. (2013)

Harris et al. (2020)

Gaibor et al. (2013)

Rush and Seegers (1990) Rinne et al. (2018)

Burrows et al. (2020)

Reference Leys et al. (2020)

6 RICHARD H. SILLITOE



GOLD DEPOSIT TYPES: AN OVERVIEW

200 m Cu-Au MINERALIZATION

PORPHYRY

Subore grade

Late mineral

Low grade

Intermineral

High grade

Early

Fig. 4. Schematic sections to show typical distribution of early, intermineral, and late-mineral intrusions in a multiphase porphyry stock and two possible ore-shell configurations. Note that the high-grade ore is confined to the early porphyry and its immediate wall rocks, low-grade ore is hosted by intermineral porphyry and more distal wall rocks, and late-mineral porphyry is subore grade and commonly barren.

include their alkaline counterparts. The early phases of these intrusions and their immediate host rocks contain much of the copper and gold in intermediate to low sulfidation-state sulfide assemblages associated with potassic, sodic, and/or overprinted chlorite-white mica (sericite) alteration zones and their contained quartz veinlet stockworks (Sillitoe, 2000; Fig. 4; Table 3). Although most gold-rich porphyry deposits appear to contain limited volumes of well-mineralized, magmatic-hydrothermal breccia, a subset of deposits is breccia dominated (e.g., Cȏté; Katz et al., 2017; Table 3). In telescoped systems, upper parts of the porphyrytype alteration and mineralization can be overprinted and reconstituted by advanced argillic assemblages, containing high sulfidation-state copper-gold ore (Sillitoe, 2010; e.g.,

A

7

Onto; Burrows et al., 2020; Table 3), which can transition upward and outward to the even shallower high-sulfidation epithermal environment (e.g., Far Southeast-Lepanto; Hedenquist et al., 1998; Table 3; see below). In carbonate host rocks, porphyry deposits can be flanked by copper-gold skarns that typically contain higher grade ore, as exemplified by Big Gossan in the Ertsberg porphyry-skarn district, Indonesia (Meinert et al., 1997), in which, exceptionally, approaching half of the >200 Moz of gold is skarn hosted (applying a 0.1% Cu cutoff; Leys et al., 2012). Gold-rich porphyry deposits, like all such deposits, are the products of focused ascent of magmatic fluids controlled by the host stocks or dikes. The gold enrichment seems to be favored by relatively shallow (140 Moz Au; Goryachev et al., 2020), are Cenozoic in age and formed during uplift, weathering, and erosion of orogenic, particularly turbidite-hosted gold deposits and occurrences either in their hinterlands or nearby (Table 6). Other gold deposit types made relatively minor placer contributions, and where gold is submicroscopic or extremely fine grained, as in the case of Carlin-type and porphyry deposits, there is minimal placer development. Geochemical characterization of placer gold grains can be used to determine the source deposit type (e.g., Chapman and Mortensen, 2006). Most large gold placers have a fluvial origin (Table 6), occurring along channel floors and/or in paleochannels partially preserved as terraces (Fig.



Su et al. (2009) 12 thin, strata-bound bodies

Heitt et al. (2003) Structurally controlled breccias

provided by P.J. Dobak (writ. commun., 2020) 1Estimate

Decarbonatization, silicification, dolomitization None ~240–200 8.5; 5.1

Bioclastic limestone, calcareous siltstone

Quartz-kaolinite Rhyolite dikes 38.3 1.7; 34.0

Skarnified carbonate rocks

None 39.5 6.9; 10–12

Silty limestone, calcareous siltstone, basaltic sills, tuffs, + flows Limestone, calcareous mudstone, debris flows 41 20; ~2

11

8), but a few formed by reworking of glacial till in a nearshore marine environment (e.g., Nome; Kaufman and Hopkins, 1989; Table 6) or in fluvioglacial settings (e.g., La Quinua in the Yanacocha high-sulfidation epithermal district; Mallette et al., 2004; Pilco and McCann, 2020; Table 6). Placer gold concentrates mainly as detrital grains in lag or accumulation gravels, typically only a few meters thick, along the floors of fluvial channels. The gold was commonly repeatedly reworked throughout the Cenozoic (Henley and Adams, 1979; Hérail et al., 1989; Hughes et al., 2004; Garnett and Bassett, 2005; Craw, 2010; Fig. 8), or from even older placers (e.g., Black Hills, South Dakota, United States; Paterson et al., 1988), unless isolated as deep leads beneath postmineral volcanic and/or sedimentary material (e.g., Lindgren, 1911; Hughes et al., 2004; Fig. 8). Major deposits result from an optimized interplay between tectonic uplift, bedrock competence, depth of weathering, gold availability, storm intensity, and, hence, climate (Henley and Adams, 1979: Hughes et al., 2004; Roy et al., 2018). There is increasing evidence for biogeochemical mobility of gold during placer formation (Reith et al., 2012). Placer deposits are not a present-day exploration target for most major and junior companies because of difficulties in estimating compliant resources and environmental concerns; they are mainly exploited by either placer specialists or small-scale and artisanal miners, the latter commonly illegal and responsible for environmental defilement. The placers are typically exploited by means of open-cut mining using heavy earth-moving equipment, rather than by dredging or hydraulicing as was more common in the past, with gravity separation being the usual beneficiation method. However, placer gold in deep leads, terraces, and raised beaches has also been mined underground. In contrast, La Quinua (Table 6) was treated as an oxidized gold deposit, employing cyanide heap leaching.

Releasing bend on dextral-normal fault Anticlinal axis, normal faults

Cassinerio and Muntean (2011)

Breit et al. (2005)

Decarbonatization, silicification (jasperoid), dolomitization, illite, adularia Silicification, argillization, decarbonatization None

Overturned antiStrata-bound bodies in clines, steep faults anticlinal hinge, some beneath sills Moderate-angle Fault-controlled + strafault, fracture ta-bound bodies zones

Bradley et al. (2020) Strata-bound solution-collapse breccias + decalcified horizons Silicification, sulfidation, decarbonatization 35.7 11.3; 9.5

Limestone, calcareous mudstone, debris flows

Felsic dikes

Thrust faults + associated folds

Rota (1993) Fault-controlled and strata-bound bodies Decarbonatization, argillization, silicification Felsic dikes Calcareous mudstone, silty limestone ~38

Deposit, location Goldstrike (Carlin trend), Nevada, United States Gold Quarry (Carlin trend), Nevada, United States Goldrush (Battle Mountain-Eureka trend), Nevada, United States Twin Creeks (Getchell trend), Nevada, United States Getchell-Turquoise Ridge (Getchell trend), Nevada, United States Deep Star (Carlin trend), Nevada, United States Shuiyindong, Guizhou, China

>25; ~1.5

Ore-related alteration Decarbonatization, argillization, silicification Coeval intrusion Felsic dikes Host rock(s) Calcareous mudstone, debris flows Age (Ma) 39.1 Size, Moz Au; grade, g/t Au 58; ~61

Table 5. Noteworthy Carlin-Type Gold Deposits

Ore controls Normal faults, anticlines, dikes, preore sills Normal + reverse faults, fold hinges

Mineralization style Strata-bound + fault-controlled bodies

Reference Dobak et al. (2020)

GOLD DEPOSIT TYPES: AN OVERVIEW

Reduced Intrusion-Related Gold Deposits Reduced intrusion-related gold deposits are genetically related to moderately reduced felsic granitoids, mostly assignable to the ilmenite-series of Ishihara (1977), which can also display a tungsten skarn association (Fig. 1). They were distinguished as a discrete deposit type on the basis of gold deposits and prospects in the Late Cretaceous Tintina belt of Alaska and the Yukon, but examples occur worldwide (Thompson et al., 1999; Thompson and Newberry, 2000; Lang and Baker, 2001; Hart, 2007; Fig. 2). Tectonic settings are varied and range from postcollisional back-arc extension in the Tintina belt, including Fort Knox and Dublin Gulch (Hart, 2007; Table 7), as well as at Salave (Martínez-Terente et al., 2018) to intraplate in the case of Telfer (A.J. Wilson et al., 2020). Some deposits, including Fort Knox, Dublin Gulch, and Vasilkovskoye, comprise intrusion-hosted arrays of sheeted quartz veins and stockworks, whereas others include quartzveined siliciclastic metasedimentary wall rocks within the thermal aureoles (Table 7). In the case of Telfer, the parental granitic pluton could lie at a depth of several kilometers (A.J. Wilson et al., 2020). Aplites and pegmatites are typical accompaniments, but the porphyry stocks and their distinctive quartz ± magnetite veinlet stockworks (Fig. 5) that define

12

RICHARD H. SILLITOE

V

V

V

V

V

Fluvial terrace

Deep lead

V

V

V

V

Volcanic unit Siliciclastic rocks Au in basal conglomerate Bedrock

Orogenic Au veins

500 m

Streambed gravel

Gold-bearing material

Fig. 8. Schematic section showing progressive recycling of gold in Cenozoic placer deposits. Gold from eroded orogenic veins is concentrated in basal conglomerate of siliciclastic sedimentary sequence and, during subsequent valley incision, is further concentrated in basal streambed gravels. The older basal gravels are preserved as fluvial terraces, whereas the youngest basal gravels reside in the present streambed. Deep leads are formed by concealment of placer deposits beneath volcanic and sedimentary cover.

gold-rich porphyry deposits are notably absent. The reduced intrusion-related deposits are associated with potassic ± sodic and/or white mica (sericitic) ± carbonate alteration and typically have a low-sulfide, base metal-deficient, W-Bi-Te-As ± Mo ± Sb signature (Table 7), in keeping with the low redox states of the related intrusions. In contrast to the hypersaline brines involved in shallow porphyry deposit formation, most, but not all, reduced intrusion-related gold deposits formed from low-salinity, aqueous carbonic fluids, typically at depths of ≥~5 km (Baker, 2002; Fig. 1). The gold tenor of reduced intrusion-related deposits is dependent on vein density and is typically low (Table 7) and amenable to open-pit mining and heap leaching. However, beyond known concentrations of such deposits and prospects, like those in the Tintina belt and greater Telfer district, there is little exploration specifically targeting this deposit type. Thus, new discoveries tend to stem from follow-up of geologic and geochemical indications of gold mineralization. Gold-Rich Volcanogenic Massive Sulfide (VMS) Deposits Gold-rich VMS deposits—defined by Mercier-Langevin et al. (2011) as those containing >~3.5 g/t Au—are largely confined to rifted magmatic arcs and immature back-arc basins characterized by submarine bimodal volcanism of tholeiitic to calc-alkaline affinity (Hannington et al., 1999). Felsic flow domes and coeval autoclastic and volcaniclastic products, commonly occupying caldera settings (Sillitoe, 1980; de Ronde et al., 2019), host many deposits (Fig. 9), which commonly

coincide with volcanic or sedimentary facies changes or hiatuses. Synvolcanic intrusions and extensional faults underlie the VMS districts (Dubé et al., 2007; Fig. 9). Although the largest gold-rich VMS deposits are Precambrian in age, some are much younger (Dubé et al., 2007; Mercier-Langevin et al., 2011; Table 8), including small seafloor examples like the Solwara 1 deposit in the Manus Basin, Papua New Guinea, which has been touted as mineable at a present-day water depth of ~1,500 m (Yeats et al., 2014; Table 8). Gold-rich VMS deposits have broadly the same geologic characteristics as their gold-poor brethren and comprise syngenetic, massive to semimassive sulfide lenses or sheets underlain by large, subconcordant to discordant stockwork feeder zones marked by alteration pipes (Huston, 2000; Dubé et al., 2007). The massive sulfide bodies, commonly displaying stacked configurations, may form either by replacement immediately beneath and/or precipitation directly on the seafloor. The gold can be part of two VMS metal associations: Au-Cu and Au-Zn (Hannington et al., 1999; Huston, 2000; Dubé et al., 2007), which are commonly enriched in the epithermal suite (As, Sb, Hg), as particularly noteworthy in the bonanza-grade Eskay Creek deposit (Roth et al., 1999; Table 8). Although the giant Horne system, in common with many VMS deposits, has massive, near-ore chloritic alteration surrounded by an extensive white mica zone (Monecke et al., 2017), many goldrich examples are associated with advanced argillic alteration or its aluminosilicate-rich metamorphic equivalent (Table 8), in keeping with the concept of a magmatic fluid contribution

Table 6. Noteworthy Placer Gold Deposits Deposit, location Upper Kolyma, Siberia, Russia Lena, Siberia, Russia Victoria, Australia Tipuani, Bolivia Klondike, Yukon, Canada Sierra Foothills, California, United States Nome, Alaska, United States La Quinua, Peru 1Areal

Gold production1 (Moz) 87 40.5 44 ~30 >20 ~70 4.7 10.22

Age Mainly Quaternary

Placer type Mainly fluvial

Gold deposit type source Orogenic

Reference Goryachev et al. (2020)

Quaternary Pre-Eocene-Quaternary Miocene-Quaternary Pliocene-Quaternary Eocene, Quaternary Quaternary

Fluvial Fluvial Fluvial Fluvial Fluvial

Orogenic (Sukhoi Log type) Orogenic Orogenic Orogenic Orogenic

Vursiy et al. (2020) Hughes et al. (2004) Hérail et al. (1989) Lowey (2006) Lindgren (1911)

Beach

Orogenic

Quaternary

Glaciofluvial

High-sulfidation epithermal

Kaufman and Hopkins (1989) Mallette et al. (2004)

extents of these placer fields vary by orders of magnitude 0.75 g/t Au

2Averaging

Metasedimentary rocks

Granodiorite + metasedimentary rocks Granodiorite

466–465 Granodiorite + diorite

~2074

293

93

Age (Ma) Host rock 645–620 Neoproterozoic metasiltstone + sandstone 92 Granite

Metarhyolitic flows and sills

20; 0.75 10.5; 0.46 4.4; 0.66 ~6.4; 2.91 7; 3.2 12.6; ~1.7

Deposit, location Horne, Quebec, Canada

1 Current

Vasilkovskoe, Kazakhstan

Morila, Mali

Deposit, location Telfer, Western Australia Fort Knox, Alaska, United States Dublin Gulch, Yukon, Canada Salave, Spain

Table 7. Noteworthy Reduced Intrusion-Related Gold Deposits

Yeats et al. (2014)

Taube (1986)

Roth et al. (1999)

Mercier-Langevin et al. (2017) Mercier-Langevin et al. (2013) Taylor et al. (1999)

Dubé et al. (2014)

Reference Monecke et al. (2017)

Dolgopolova et al. (2015)

Martínez-Terente et al. (2018) McFarlane et al. (2011)

Maloof et al. (2001)

Bakke (1995)

Reference A.J. Wilson et al. (2020)

GOLD DEPOSIT TYPES: AN OVERVIEW

13

14

RICHARD H. SILLITOE

X

X

X

Seawater

X

Au-rich VMS deposit

X

Volcanic dome

2 km

X

X

X

Au-poor VMS deposit

Sill complex

Fig. 9. Schematic section showing contrasting formational settings of felsic volcanic domes associated with gold-rich and -poor VMS deposits. Note that the gold-rich deposit received greater magmatic fluid input as a consequence of relatively shallow intrusion and formed under shallower submarine conditions that gave rise to fluid boiling conducive to gold precipitation.

to ore genesis (Sawkins, 1986; Sillitoe et al., 1996; Hannington et al., 1999; Huston, 2000; Large et al., 2001; Dubé et al., 2007; Piercey, 2011). The enhanced gold tenor is generally ascribed to magmatic fluid input and/or boiling induced by formation under relatively shallow-water conditions (20 vol %) deposits, most of Neoarchean or Proterozoic age, but also including important Mesozoic examples in the Coastal Cordillera of Chile and Peru (Hitzman et al., 1992; Williams et al., 2005; Groves et al., 2010). Many of them do not contain major gold inventories because of either their relatively small sizes (10; 9.8

25; 15.3

2881

~1701; 3.5–4

2665–2660

70

39; 2.3

17.9; 0.972

Age (Ma) 104

Size, Moz Au; grade, g/t Au 7.9; 12.9

Veinlets, subsidiary quartz veins

Quartz veins + tabular disseminated zones Quartz ± carbonate veins

Sheeted veins, breccias, stockwork Quartz-carbonate disseminated-stockwork replacements Quartz veins

Quartz veins + stockworks

Mineralization style Stacked low-angle quartz veins Stockwork

Oxidized diorite porphyry, oxidized monzogranite

Oxidized granodiorite + monzodiorite

Oxidized granodiorite

Oxidized porphyritic quartz monzodiorite-granodiorite Oxidized syenite porphyry

Reduced rhyolite-rhyodacite dike complex Reduced granite + granodiorite dikes Reduced syenogranite

Potentially ore-related intrusion Reduced granite dikes

estimate, probably including inferred resources; see also Seltmann et al. (2020) grade of 10.7-Moz open-pit resource 3 See also Dubé and Mercier-Langevin (2020) 4 And the rest of the deposits in the Jiaodong gold province (see Qiu et al., 2020) 5 See also Tripp et al. (2020) 6 Estimate provided by S. Turner (writ. commun., 2019) 7 See also Turner et al. (2020)

2 Average

1 Imprecise

Canadian Malartic, Quebec, Canada Kirkland Lake, Ontario, Canada Linglong4, Shandong, China Kalgoorlie (Golden Mile), Western Australia Boddington, Western Australia

Deposit, location Pogo, Alaska, United States Donlin Creek, Alaska, United States Muruntau, Uzbekistan Zarmitan, Uzbekistan

Cu, Ag, Mo, Bi, Te

Cu, Pb, Zn, Te

Cu, Zn, Pb, As

Te, W, Ag ± Bi, Pb, Mo Te, Ag, Mo, Pb

W, Bi, As, Sb

As, W, U

As, Sb

Metal/ metalloid signature As

Table 11. Major Gold Deposits Interpreted as Either Orogenic or Intrusion Related

Allibone et al. (1998)7

Vielreicher et al. (2016)5

Watson and Kerrich (1983)3 Qiu et al. (2020)

Goldfarb et al. (2001, 2014) De Souza et al. (2020)3

Goldfarb et al. (2014)1

Goldfarb et al. (2004)

Orogenic interpretation Goldfarb et al. (2007)

Hagemann et al. (2007)7

Mueller et al. (2020)5

Fan et al. (2007)

Ispolatov et al. (2008)

Helt et al. (2014)

Abzalov (2007)

Wall et al. (2004)1

Szumigala et al. (1999)

Intrusion-related interpretation Rhys et al. (2003)

18 RICHARD H. SILLITOE



GOLD DEPOSIT TYPES: AN OVERVIEW

styles documented from the mineralized trends in northcentral Nevada, classification of geologically similar deposits elsewhere remains more controversial. The Carlin-type deposits in the Golden Triangle of southwestern China, for example, have been considered by some investigators to possess affinities with orogenic gold deposits (e.g., Cline et al., 2013; Goldfarb et al., 2019). In contrast to the situation in northern Nevada and southwestern China, many similar sedimentary rock-hosted gold deposits occupy distal sites in porphyry intrusioncentered districts. Such deposits, exemplified by Barneys Canyon and Melco in the Bingham district (Babcock et al., 1995), Sepon in Laos (Smith et al., 2005), and Bau in Sarawak, East Malaysia (Sillitoe and Bonham, 1990), have been distinguished from true Carlin-type deposits and designated by some investigators as distal-disseminated or Carlin-like (e.g., Hofstra and Cline, 2000), although the only substantive difference is their observed spatial association with porphyry intrusions and base metal deposits and their smaller sizes (Fig. 10). However, recent work in Nevada, specifically at North Bullion in the southern Carlin trend and McCoy-Cove in the Battle Mountain-Eureka trend, has confirmed the contemporaneity of Carlin-type mineralization and exposed stocks and dikes associated with porphyry, skarn, and carbonate-replacement mineralization (Henry et al., 2015; Thompson et al., 2015; Muntean et al., 2017), further emphasizing the proposed identity of Carlin-type and distaldisseminated/Carlin-like deposits (e.g., Sillitoe and Bonham, 1990; Ressel et al., 2000; Sillitoe, 2020). Classification of low- and intermediate-sulfidation deposits can also be problematic because some epithermal deposits contain both low- and intermediate-sulfidation veins (e.g., Castor et al., 2003; Camprubí and Albinson, 2007; Leary et al., 2020), and even individual veins in a few deposits can contain low- and intermediate-sulfidation paragenetic stages (e.g., Permuy Vidal et al., 2016). These observations suggest a close spatial and temporal association between the two fluid types responsible and a potentially transitional relationship. Deposit type associations, transitions, and superpositions Some of the gold deposit types characterized herein can occur in close association either with gold-poor examples of the same deposit types or with other gold deposit types. Alternatively, transitional formational environments or superposition of mineralization events can enable formation of hybrid deposits that may display features of two deposit types. Gold-rich porphyry and gold-rich VMS deposits can be parts of deposit clusters that contain geologically similar deposits deficient in gold. In the case of porphyry copper deposits, molybdenum- and gold-rich examples of the same age can occur nearby or even in direct contact with one another (e.g., Los Pelambres, Chile, and Grasberg, Indonesia; Perelló et al., 2012; Leys et al., 2020). The prolific Hokuroku, Japan, Skellefte, Sweden, and Noranda, Quebec, districts provide good examples of the association of a few gold-rich and numerous gold-poor VMS deposits formed at broadly the same time (Yamada, 1988; Allen et al., 1996; Monecke et al., 2017). The reasons for the proximity of gold-rich and -poor porphyry and VMS deposits requires further work but may be a consequence of emplacement depth of associated

19

intrusions combined, in the VMS case, with overlying water depth (Hannington et al., 1999; Murakami et al., 2010; Fig. 9). In porphyry systems emplaced into carbonate sequences, porphyry copper-gold deposits can transition laterally and vertically to copper-gold and/or gold skarns, which, in turn, can grade outward into carbonate-replacement deposits and, on the distal fringes of some systems, Carlin-type gold deposits (Sillitoe, 2010; Fig. 10; Table 3). The Bingham district is the type example of this proximal to distal sequence of gold-bearing deposit types (Babcock et al., 1995). Hybrid gold deposits, representing the transition between carbonatereplacement and Carlin-type environments, have also been proposed (Sillitoe, 2020). In the case of the Cove gold-silver deposit, Nevada, however, base metal-bearing gold-silver mineralization, including carbonate-replacement bodies, is overprinted by the Carlin-type gold mineralization (Thompson et al., 2015; Muntean et al., 2017). The shallower parts of porphyry systems are typically characterized by advanced argillic lithocaps, which can host high-sulfidation epithermal gold ± silver ± copper deposits and transition outward to intermediate-sulfidation gold ± silver ± zinc ± lead deposits (Sillitoe and Hedenquist, 2003; Sillitoe, 2010; Fig. 1). Nevertheless, existence of high-sulfidation epithermal gold deposits is not necessarily dependent on the underlying porphyry centers being auriferous (e.g., Famatina district, Argentina; Lozada-Calderón and McPhail, 1996). Similarly, at Peñasquito, the two gold-rich, intermediatesulfidation diatreme breccias overlie weakly developed, molybdenum-dominated porphyry mineralization (Dromundo et al., 2020). Exceptionally, as observed at the giant Lihir deposit, even low-sulfidation gold mineralization can overprint a low-grade porphyry copper-gold center, albeit following extreme perturbation and telescoping of the system by volcano sector collapse (Sillitoe, 1994; Cooke et al., 2020; Table 4). Telescoping, consequent upon synhydrothermal uplift and erosion, has also been invoked at Porgera for overprinting of alkaline intrusion-related carbonate-base metal mineralization by the Zone VII bonanza-grade epithermal veins and breccias (Richards, 1992; Hay et al., 2020; Table 4). Transitions between terrestrial and submarine environments, favored on partly emergent island-arc volcanoes (Naden et al., 2005), may give rise to submarine epithermal deposits (Alfieris et al., 2013) or shallow-water VMS deposits, in which stockwork feeder zones have textural and mineralogic similarities to epithermal deposits (e.g., Eskay Creek; Roth et al., 1999; Table 8). The epithermal deposits are likely to be of intermediate- rather than lowsulfidation type, in keeping with the enhanced fluid salinities consequent upon seawater involvement (Alfieris et al., 2013). Although overprinting of two or more gold-bearing paragenetic stages and upgrading by gold remobilization (Hastie et al., 2020) are commonplace in many gold deposit types, there is also the possibility that either two discrete but genetically similar gold-forming events or two different gold deposit types, formed at widely different times, are superimposed on one another. Indeed, Meffre et al. (2016) went as far as to suggest that such superpositions are a basic requirement for major gold deposit formation. For example, several major orogenic gold deposits, including Red Lake, Obuasi, and Bendigo (Table 2), have been suggested

20

RICHARD H. SILLITOE

as formed during two separate stages (Dubé et al., 2004; Fougerouse et al., 2017; C.J.L. Wilson et al., 2020) or, in the case of the Golden Mile at Kalgoorlie—irrespective of its orogenic and/or intrusion-related origin (Table 11)—in at least three stages (Bateman and Hagemann, 2004; Robert et al., 2005; McDivitt et al., 2020; Mueller et al., 2020; Tripp et al., 2020). Examples of two or more different ore-forming processes at a single locality may include: Boddington, where porphyry copper-gold and even orogenic gold may have been overprinted ~100 m.y. later by oxidized intrusion-related gold prior to eventual incorporation of its uppermost parts into a Cenozoic laterite profile mined nearby for bauxite (Symons et al., 1990; Ciobanu et al., 2013; Turner et al., 2020; Table 11); Hollinger-McIntyre, where a small porphyry coppergold-molybdenum deposit is overprinted by the preeminent orogenic gold veins (Mason and Melnik, 1986; Dubé et al., 2020; Table 2); and Aitik, where the porphyry coppergold deposit is believed to have been overprinted by minor IOCG mineralization (Wanhainen et al., 2012). It has even been proposed that early-stage gold in several Neoarchean deposits is intrusion related, whereas that introduced later is classically orogenic (Bateman and Hagemann, 2004; Mériaud and Jébrak, 2017; Thébaud et al., 2018; Tripp et al., 2020): a metallogenic progression that might be anticipated in Neoarchean greenstone belts (Robert et al., 2005), but not necessarily at the same sites unless fundamental structural conduits remained at least intermittently effective for tens of millions of years (e.g., Tripp et al., 2020). Provinciality In addition to the clustering of individual gold deposit types in districts or camps, as pointed out above, there is also a clear tendency for relatively restricted volumes of the upper crust to be especially rich in gold deposits, thereby defining metallogenic gold provinces. Such provinces are apparent both in the Archean cratons (e.g., Abitibi belt and Eastern Goldfields Province; Robert et al., 2005; Dubé et al., 2017; Dubé and Mercier-Langevin, 2020; Tripp et al., 2020) and mobile belts (e.g., Birimian of West Africa; Goldfarb et al., 2017; Thébaud et al., 2020) as well as in Phanerozoic accretionary orogens (Sillitoe, 2008, and references therein). The gold provinces may have formed in either single, relatively short-lived (e.g., Deseado massif, southern Argentina) or multiple, spatially confined metallogenic epochs (e.g., Abitibi belt, Eastern Goldfields Province, Birimian belts, and Great Basin, Nevada). In the latter case, gold province formation can span 90 to 120 m.y. (Sillitoe, 2008; Dubé and Mercier-Langevin, 2020; Thébaud et al., 2020; Tripp et al., 2020). Where the provinces are products of multiple gold-forming epochs, several deposit types are generally present although a single type may predominate. Hence, the preeminent orogenic quartz-carbonate vein deposits of the Abitibi belt are accompanied by gold-rich VMS, porphyry, and a variety of other intrusion-related vein and disseminated types (Dubé and Mercier-Langevin, 2020, and references therein), whereas in the Great Basin, low-, intermediate-, and high-sulfidation epithermal, porphyry, reduced and oxidized skarn, reduced intrusion-related, and orogenic gold deposit types are present besides the dominant Carlin-type deposits (Sillitoe, 2008, and references therein).

This clustering of different deposit types strongly suggests that discrete portions of the upper crust were (and in some cases may still be) predisposed to the formation of gold deposits irrespective of their different formational mechanisms. It has been proposed that gold provinces reflect subcontinental lithospheric mantle or lowermost crust that was either preenriched in gold or from which gold was more readily or efficiently extractable (Sillitoe, 2008; Hronsky et al., 2012; Wang et al., 2020; Fig. 12). The heterogeneous distribution of gold deposits could reflect a spatially similar concentration of gold in the lithospheric mantle, possibly due to random additions of highly siderophile metals during bombardment of Earth by large plenetesimals (solid astronomical objects) at ~4 Ga (Frimmel, 2014; Marchi et al., 2018) and/or metasomatic refertilization during one or more previous, potentially ancient subduction cycles (Richards, 2009; Hronsky et al., 2012; Griffin et al., 2013). These hypotheses gain some support from reports of gold particles in mantle-derived rocks in the Chinese Golden Triangle and Deseado massif gold provinces (Zhang et al., 2006; Tassara et al., 2017). There are also alternative, possibly complementary hypotheses to account for gold provinces, particularly those in (meta)sedimentary rock-dominated terranes. One, used to account for the Carlin-type gold dominance of the western Great Basin and possibly also applicable to the Central Cordillera of Colombia and elsewhere, posits that the buffering effect of an organic carbon- and pyrite-rich crustal profile results in relatively reduced intrusions favoring gold (rather than copper) deposit formation (Johnson et al., 2020). A second hypothesis considers that certain mid- to upper-crustal rock volumes contain early, diffuse, low-grade gold concentrations that can be upgraded during orogenesis, metamorphism, and/or magmatism to generate gold deposits, particularly turbidite-hosted orogenic and Carlin types

Crust

Su

bd

uc

ted

Lithospheric mantle Asthenosphere

oc

ea

nic

Au province

lith

os

ph

er

e

Convective mantle flow Partial melting Oxidized fluids/melts from slab

Au-rich mantle lithosphere + lowermost crust Fig. 12. Schematic section of an active accretionary orogen to illustrate how mantle lithosphere and lowermost crust rich in gold (or from which gold is particularly easy to extract) can give rise to upper-crustal gold provinces during single (several m.y.) or multiple (up to at least 120 m.y.) metallogenic epochs. Ascendant magmas and/or fluids (in red) scavenge and transport gold into the middle to upper crust to generate the variety of gold deposit types summarized herein. See text for additional explanation.



21

GOLD DEPOSIT TYPES: AN OVERVIEW

(Pitcairn et al., 2006; Large et al., 2007, 2011; Meffre et al., 2016; Vursiy et al., 2020). Secularity It has long been appreciated that over geologic time the various gold deposit types are heterogeneously distributed and their relative importance changed dramatically (e.g., Groves et al., 2005). Frimmel (2018) proposed that the first major gold concentration at Earth’s surface was a direct result of the unique reduced and acidic conditions that prevailed from ~2.9 to 2.8 Ga. The resultant auriferous algal mats and derivative paleoplacers, still preserved in the Witwatersrand, were hydrothermally and magmatically reworked to generate other major gold deposit types: first orogenic gold (2.75–2.55 and 2.1–1.75 Ga; Goldfarb et al., 2001) and subsequently, after the initiation of modern-day “cold” plate tectonics, at ~0.75 Ga, most of the porphyry, epithermal, Carlin, skarn, and reduced intrusion-related types. This hypothesis does not necessitate preexisting mantle heterogeneity of gold. Distinctive formational peaks of orogenic gold and gold-rich VMS deposits coincided with the assembly of supercontinents (Goldfarb et al., 2001; Huston et al., 2010). Cawood and Hawkesworth (2015) pointed out that this could be an apparent correlation caused by the superior preservation potential of rocks and contained mineralization in collisional tectonic settings. Once embedded in the assembled supercontinent, gold (and other) deposits are partially protected from the removal and recycling inherent in plate margin settings. This preservational bias might offer a viable explanation for the existence of so many major Paleozoic gold deposits, including orogenic, epithermal, gold-rich porphyry, reduced intrusionrelated, VMS, and genetically disputed types (e.g., Muruntau and Zarmitan; Table 11), in the Central Asian orogenic belt (e.g., Wan et al., 2017). The preservation potential of rocks and contained mineralization is low at convergent margins, particularly those undergoing contraction and consequent rapid rock uplift and exhumation (Sawkins, 1972; Cawood and Hawkesworth, 2015). Shallowly emplaced gold mineralization, especially epithermal and gold-rich porphyry deposits generated at paleodepths of 10 cm) rock. About 50% of the sequence is deformed into meter-scale folds. The garnet-two mica schist includes boudinaged and mylonitic rocks. There is common cleavage in combination with microfolds and microlensing. Mineralogically, the schist consists of biotite (15–35%), muscovite (5–40%), and quartz (20–40%). Purple almandine and green chlorite are present in a fine-grained groundmass. Garnet is regularly present, forming 3 to 5% of the rock volume. Quartz and sericitized plagioclase make up 5 to 30% of rock volume, with carbonate up to 5%, mainly due to carbonatization. The accessory minerals are tourmaline (10), corresponding to crustal values (Naumov et al., 2015) and consistent with its Pb isotope composition (Fig. 11a). Osmium in arsenopyrite-2 could be sourced from crustal granitoid melts or such osmium could be incorporated into fluids from the metamorphic or magmatic country rocks. Helium isotope studies: The helium isotope studies were completed in fluid inclusions in quartz, stibnite, arsenopyrite-1, and arsenopyrite-2 (Gibsher et al., 2019). The fluids from short prismatic arsenopyrite-2 inclusions in quartz reveal crustal (1,564 t Au, grading 4 to 4.6 g/t Au. This could be supplemented by the deep discoveries made in early 2018. The auriferous oxidized ore, developed to a depth of 400 m, contained ca. 200 t of nonrefractory gold, grading 11.1 g/t. In addition, the ores contain economic concentrations of silver and antimony (Olympiada East), as well as tungsten (in oxidized ore). Acknowledgments We are grateful to R.G. Sharipov, M.N. Fominykh, and A.L. Popov from Polyus Gold for access to mineral and rock collections as well as for detailed information on historic gold production. A.M. Likhman, A.V. Pridannikov, and A.N. Logachev helped with graphic data. We acknowledge the important contributions to the understanding and development of the deposit by Yu.M. Stragis, who directed the geologic department of Polyus Gold during early mining, and A.A. Plekhanov, S.I. Savushkina, and other geologists. We extend our appreciation to the achievements of Khazret Sovmen, founder and first managing director of Polyus. A.I. Averchenkov and V.P. Bordonosov, the geologists of the Severnaya Expedition, initiated exploration in the 1960s. A.Y. Kurilin, together with L.V. Li, G.P. Kruglov, and N.F. Gavrilov, discovered Olympiada, whereas V.A. Lopatin, V.I. Arefieva, V.A. Nevolin, and other geologists of Krasnoyarskgeologia identified the main geologic controls during the early years. The group of A.S. Borisenko from the Institute for Geology and Mineralogy, Siberian Branch, Russian Academy of Sciences in Novosibirsk, contributed to the understanding of mineralization and setting of the deposit.

225

REFERENCES Bernatonis, P.B., 1999, Oxidation zone at the Olympiada disseminated gold-sulfide deposit: Abstract of Candidate of Science Dissertation, Tomsk, Tomsk Polytechnic University, 19 p. (in Russian). Borisenko, A.S., Sazonov, A.M., Nevolko, P.A., Naumov, E.A., Tessalina, S., Kovalev, K.R., and Sukhorukov, V.P., 2014, Gold deposits of the Yenisei Ridge (Russia) and age of its formation [abs.]: Acta Geologica Sinica (English ed.), v. 88 (supp. 2), p. 686–687. Buryak, V.A., 1982, Metamorphogenic ore mineralization: Nedra, Moscow, 256 p. (in Russian). Cline, J.S., Hofstra, A.H., Muntean, J.L., Tosdal, R.M., and Hickey, K.A., 2005, Carlin-type gold deposits in Nevada: Critical geologic characteristics and viable models: Economic Geology 100th Anniversary Volume, p. 451–484. Dobretsov, N.L., 1977, Metamorphism of Mesozoic and Cenozoic fold belts, in Sobolev, V.S., Lepezin, G.G., and Dobretsov, N.L., eds., Metamorphic complexes of Asia: Novosibirsk, Nauka, p. 211–221 (in Russian). Ernst, R.E., Hamilton, M.A., Soderlund, U., Hanes, J.A., Gladkochub, D.P., Okrugin, A.V., Kolotilina, T., Mekhonoshin, A.S., Bleeker, W., LeCheminant, A.N., Buchan, K.L., Chamberlain, K.R., and Didenko, A.N., 2016, Long-lived connection between southern Siberia and northern Laurentia in the Proterozoic: Nature Geoscience, v. 9, p. 464–469. Genkin, A.D., 1998, Auriferous arsenopyrite from gold deposits: Inner structure of the grains, composition, mechanisms of growth and forms of gold: Geologia Rudnykh Mestorozhdeniy, v. 40, p. 551–557 (in Russian). Genkin, A.D., Wagner, F., Krylova, T.L., and Tsepin, A.I., 2002, Auriferous arsenopyrite and conditions of its formation at the Olympiada and Veduga gold deposits (Yenisei Ridge, Siberia): Geologia Rudnykh Mestorozhdeniy, v. 44, p. 59–76 (in Russian). Gibsher, N.A., Tomilenko, A.A., Sazonov, A.M., Bulbak, T.A. Ryabukha, M.A., Nekrasova, N.A., Khomenko, M.O., and Shaparenko, E.O., 2019, Olympiada gold deposit (Yenisei Ridge): Temperature, pressure, composition of ore-forming fluids, δ34S of sulfides, 3He/4He of fluids, Ar-Ar age and continuity of formation: Russian Geology and Geophysics, v. 60, no. 9, p. 1043–1059. Grigorov, V.T., 2003, The largest gold deposits of the Yenisei Ridge and Kuznetsk Alatau and their economic assessment from the point of view of stratiform mineralization: Moscow, Nauchnyi Mir, 168 p. (in Russian). Holdaway, M.J., 2000, Application of new experimental and garnet Margules data to the garnet-biotite geothermometer: American Mineralogist, v. 85, p. 881–892. Henry, D., Guidotti, C., and Thomson, J., 2005, The Ti-saturation surface for low-to-medium pressure metapelitic biotites: Implications for geothermometry and Ti-substitution mechanism: American Mineralogist, v. 90, p. 316–328. Konstantinov, M.M., Cherkasov, S.V., and Dankovtsov, R.F., 1999, Specific crustal features for large and superlarge endogenic gold deposits (Siberia and Far East regions): Geology Tectonics and Metallogeny, v. 7, p. 143–147 (in Russian). Korobeinikov, A.F., Ananiev, Y.S., Gusev, A.I., Voroshilov, V.G., Nomokonova, G.G., Pshenichkin, A.Y., and Timkin, T.V., 2013, Ore-metasomatic system and geochemical zonation of goldfields and deposits in the fold belts of Siberia: Tomsk, Tomsk Polytechnic University, 458 p. (in Russian). Kretschmar, U., and Scott, S.D., 1976, Phase relations involving arsenopyrite in the system Fe–As–S and their application: Canadian Mineralogist, v. 14, p. 364–386. Kryazhev, S.G., 2017, Genetic models and forecast criteria for gold deposits in the carbonaceous-clastic formations: Doctor of Science Dissertation, Moscow, Central Research Institute of Geological Prospecting for Base and Precious Metals (TSNIGRI), 288 p. (in Russian). Kun, L., Ruidong, Y., Wenyong, C., Rui, L., and Ping, T., 2014, Trace element and REE geochemistry of the Zhewang gold deposit, southeastern Guizhou Province, China: Chinese Journal of Geochemistry, v. 33, p. 109–118. Lambert, J.M., Simkovich, G., and Walker, P.L., Jr., 1998, The kinetics and mechanism of the pyrite-to-pyrrhotite transformation: Metallurgical and Materials Transformations, v. 29B, p. 385–396. Large, R.R., Maslennikov, V.V., Robert, F., Danyushevsky, L.V., and Chang, Z., 2007, Multistage sedimentary and metamorphic origin of pyrite and gold in the giant Sukhoi Log deposit, Lena gold province, Russia: Economic Geology, v. 102, p. 1233–1267. Li, L.V., 2003, The Olympiada disseminated gold-sulfide deposit: Krasnoyarsk, KNIIGiMS, 120 p. (in Russian). Likhanov, I.I., and Reverdatto, V.V., 2014, P-T-t constraints on the metamorphic evolution of the Transangarian Yenisei Ridge: Geodynamic

226

SAZONOV ET AL.

and petrological implications: Russian Geology and Geophysics, v. 55 (3), p. 385–416. Likhanov, I.I., Polyansky, O.P., Reverdatto, V.V., and Memmi, I., 2004, Evidence from Fe- and Al-rich metapelites for thrust loading in the Transangarian region of the Yenisey Ridge, eastern Siberia: Journal of Metamorphic Geology, v. 22, p. 743–762. Likhanov, I.I., Reverdatto, V.V., Sukhorukov, V.P., Kozlov, P.S., and Khiller, V.V., 2013, Three metamorphic events in the Precambrian P-T-t history of the Transangarian Yenisey Ridge recorded in garnet grains in metapelites: Petrology, v. 21, p. 561–578. Likhanov, I.I., Nozhkin, A.D., Reverdatto, V.V., and Kozlov, P.S., 2014, Grenvillian tectonic events and evolution of the Yenisei Ridge at the western margin of the Siberian craton: Geotektonics, v. 48 (5), p. 371–389. Metelkin, D.V., Vernikovskiy, V.A., and Kazanskiy, A.Y., 2007, Neoproterozoic evolution of Rodinia in the light of new paleomagnetic data at the western margin of the Siberian craton: Russian Geology and Geophysics, v. 48, p. 32–54. Mkrtychian, A.K., 2007, The source of gold and general characteristics in distribution of gold deposits in the Yenisei Ridge in Mkrtychian, A.K., ed., Geology and mineral resources of central Siberia and adjacent areas: Krasnoyarsk, Krasnoyarskgeols’emka Publishing, p. 52–56 (in Russian). Monecke, T., Kempe, U., Monecke, J., Sals, M., and Wolf, D., 2002, Tetrad effect in rare earth element distribution patterns: A method of quantification with application to rock and mineral samples from granite-related rare metal deposits: Geochimica et Cosmochimica Acta, v. 66, p. 1185–1196. Naumov, E.A., Borisenko, A.S., Nevolko, P.A., and Kovalev, K.R., 2015, Gold-sulfide (Au-As) deposits of the Yenisei Ridge (Russia): Age, sources of metals and nature of fluids: Mineral Resources in a Sustainable World: Biennial Society for Geology Applied to Mineral Deposits (SGA) Meeting, 13th, Nancy, 2015, Proceedings, v. 1, p. 165–168. Novozhilov, Y.I., and Gavrilov, A.M., 1999, Gold-sulfide deposits in carbonaceous-clastic formations: Moscow, TsNIGRI, 175 p. (in Russian). Novozhilov, Y.I., Gavrilov, A.M., Yablokova, S.V., and Arefieva, V.I., 2014, Unique economic Olympiada gold-sulfide deposit in the upper Proterozoic clastic formations: Rudy i Metally, no. 3, p. 51–64 (in Russian). Nozhkin, A.D., Turkina, O.M., Maslov, A.V., Dmitrieva, N.V., Kovach, V.P., and Ronkin, Y.L., 2008, Sm-Nd isotopic systematics of Precambrian metapelites from the Yenisei Ridge and variation of the ages in the source of denudation: Doklady Earth Science, v. 423A, no. 9, p. 1495–1500. Nozhkin, A.D., Borisenko, A.S., and Nevolko, P.A., 2011, Stages of the Late Proterozoic magmatism and ages of gold mineralization in the Yenisei Ridge: Russian Geology and Geophysics, v. 52, p. 158–181. Petrov, V.G., 1974, Gold metallogeny of the northern Yenisei Ridge: Novosibirsk, Nauka, 140 p. (in Russian). Polyakov, G.V., Tolstykh, N.D., Mekhonoshin, A.S., Izokh, A.E., Podlipskiy, M.Y., Orsoev, D.A., and Kolotilina, T.B., 2013, Ultramafic-mafic igneous complexes of the Precambrian East Siberian metallogenic province (southern framing of the Siberian craton): Age, composition, origin, and ore potential: Russian Geology and Geophysics, v. 54, p. 1319–1331. Savichev, A.A., Shevchenko, S.S., Rozinov, M.I., et al., 2006, Isotopic and petrogeochemical characteristics of the Olympiada gold-sulfide deposit and its satellites (Yenisei Range): Regional Geology and Metallogeny, v. 28, p. 122–143. Sazonov, A.M., 1998, Geochemistry of gold in metamorphic formations: Tomsk, Tomsk Polytechnical University Publishing, 166 p. (in Russian). Sazonov, A.M., Ananyev, A.A., Poleva, T.V., Khokhlov, A.N., Vlasov, V.S., Zvyagina, E.A., Fedorova, A.V., Tishin, P.A., and Leontyev, S.I., 2010, Goldore metallogeny of the Yenisey Ridge: Geological-structural provincialism, structural types of ore fields: Journal of Siberian Federal University, Engineering and Technologies, v. 4, p. 371–395 (in Russian). Sazonov, A.M., Nekrasova, N.A., Zvyagina, E.A., and Tishin, P.A., 2016, Geochronology of granites, host schists and ores of the Panimba deposit (Yenisei Ridge): Journal of Siberian Federal University, Engineering and Technologies, v. 9 (2), p. 174–188 (in Russian). Scott, S.D., 1983, Chemical behaviour of sphalerite and arsenopyrite in hydrothermal and metamorphic environments: Mineralogical Magazine, v. 47, p. 427–435. Serdyuk, S.S., Komorovskiy, Yu.E., Zverev, A.I., Oyaber`, V.K., Vlasov, V.S., Babushkin, V.E., Kirilenko, V.A., and Zemlyanskiy, S.A., 2010, Models of gold deposits in the Yenisei area of Siberia: Krasnoyarsk, Institute Gornogo Dela, Geologii i Geotekhnologiy, 582 p. (in Russian).

Shenfil’, V.Yu., 1991, The Late Cambrian of the Siberian platform: Novosibirsk, Nauka, 185 p. (in Russian). Sovmen, V.K., Stragis, Y.M., Plekhanov, A.A., Bibik, S.M., Krovyakova, L.P., Savushkina, S.I., Lokhmakov, V.A., Zvezdin, I.G., and Logachev, V.S., 2009, Geological structure of gold deposits and experience of geological work at Polyus Company in the Krasnoyarsk region: Krasnoyarsk, Verso, 208 p. (in Russian). Storozhenko, A.A., Vasiliev, N.F., and Diner, A.E., 2002, Explanatory note to the state geological map of the Russian Federation at a scale of 1:200 000, 2nd ed., Yenisei Series, Sheet O-46-III: Moscow, VSEGEI (in Russian). Toulmin, P., and Barton, P.B., 1964, A thermodynamic study of pyrite and pyrrhotite: Geochimica et Cosmochimica Acta, v. 288, p. 641–671. Vernikovskaya, A.E., Metelkin, D.V., et al., 2016, Neoproterozoic structure of the Yenisei Ridge and formation of the western margin of the Siberian craton constrained by geological, paleomagnetic and geochronological data: Geologia i Geofizika, v. 57, p. 63–90 (in Russian). Vernikovskaya, A.E., Vernikovsky, V.A., Salnikova, E.B., Kotov, A.B., Kovach, V.P., Travin, A.V., and Wingate, M.T.D., 2007, Leucocratic A-type magmatism in evolution of the continental crust at the western margin of the Siberian craton: Russian Geology and Geophysics, v. 48, p. 3–16. Vernikovsky, V.A., and Vernikovskaya, A.E., 2006, Tectonics and evolution of the granitoid magmatism in the Yenisei Ridge: Russian Geology and Geophysics, v. 47, p. 35–52. Vernikovsky, V.A., Vernikovskaya, A.E., Kotov, A.B., Salnikova, E.B., and Kovach, V.P., 2003, Neoproterozoic accretionary and collisional events on the western margin of the Siberian craton: New geological and geochronological evidence from the Yenisey Ridge: Tectonophysics, v. 375, p. 147–168. Vernikovsky, V.A., Metelkin, D.V., Vernikovskaya, A.E., Matushkin, N.Yu., Kazanskiy, A.Yu., Kadilnikov, P.I., Romanova, I.V., Wingate, M.T.D., Larionov, A.N., and Rodionov, N.V., 2016, Neoproterozoic tectonic structure of the Yenisei ridge and formation of the western margin of the Siberian craton based on new geological, paleomagnetic, and geochronological data: Russian Geology and Geophysics, v. 57 (1), p. 63–90. Volkov, A.V., Sidorov, A.A., Savva, N.E., Prokofiev, V.Y., Kolova, E.E., Savchuk, Y.S., Murashov, K.Y., Sidorova, N.V., Zemskova, M.I., Aristov, V.V., and Wolfson, A.A., 2016, Gold-quartz deposits of the Yana-Kolyma fold belt: Geochemical characteristics of ores and fluids, formation settings: Vestnik SVNTs DVO RAN, no. 3, p. 3–21 (in Russian). Volobuev, M.I., Stupnikova, N.I., and Zykov, S.I., 1973, Yenisei Ridge, in Polovinkin, Yu.I., ed., Geochronology of the USSR, v. 1, Precambrian: Leningrad, Nedra, p. 189–201 (in Russian). Vrublevskiy, V.V., Nikitin, R.N., Tishin, P.A., and Travin, A.V., 2017, Metamafic rocks of the Middle Transangara region, Yenisei Ridge: E-MORB relics of Neoproterozoic lithosphere: Litosfera, v. 17 (5), p. 67–84 (in Russian). Winkler, H.G.F., 1979, Petrogenesis of metamorphic rocks: New York, Springer, 339 p. Wu, C.-M., Zhang, J., and Ren, L., 2004, Empirical garnet-biotite-plagioclase-quartz (GBPQ) geobarometry in medium- to high-grade metapelites: Journal of Petrology, v. 45, p. 1907–1921. Yakubchuk, A., Stein, H., and Wilde, A., 2014, Results of pilot Re-Os dating of sulfides from the Sukhoi Log and Olympiada orogenic gold deposits, Russia: Ore Geology Reviews, v. 59, p. 21–28. Zabiyaka, A.I., Kurgankov, P.P., Gusarov, Yu.V., et al., 2004, Tectonics and metallogeny of the Lower Angara region: Krasnoyarsk, KNIIGiMS, 322 p. (in Russian). Znamenskiy, S.E., Michurin, S.V., Velivetskaya, T.A., and Znamenskaya, N.M., 2014, Structural setting and possible sources of mineralization in the Ganeevskoe gold deposit (South Urals): Litosfera, no. 6, p. 118–131 (in Russian). Zuev, V.K., Kachevskiy, L.K., Kachevskaya, G.I., Komarov, V.V., Minaeva, O.A., Markovich, L.A., Shatalina, T.N., and Potapenko, L.Y., 2009, Explanatory note to the state geological map of the Russian Federation at a scale of 1:1000000 (3rd generation): Angara-Yenisei Series, Sheet O-46 Krasnoyarsk: St.-Petersburg, VSEGEI, 500 p. Zvyagina, E.A., 1989, Metamorphism and gold metallogeny of the Upper Enashimo ore cluster: Candidate of Science Dissertation, Krasnoyarsk, Siberian Federal University, 275 p. (in Russian).

©2020 Society of Economic Geologists, Inc. SEG Special Publications, no. 23, pp. 227–249

Chapter 11 The Telfer Gold-Copper Deposit, Paterson Province, Western Australia Alan J. Wilson,1,† Nick Lisowiec,2 Cameron Switzer,2 Anthony C. Harris,2 Robert A. Creaser,3 and C. Mark Fanning4 1GeoAqua 2 Newcrest 3University

Consultants Limited, PO Box 316, The Valley, Anguilla, British West Indies

Mining Limited, Level 8, 600 St Kilda Road, Melbourne, Victoria 3004, Australia

of Alberta, Department of Earth and Atmospheric Sciences, 126ESB, Edmonton, Alberta T6G 2R3, Canada

4Research

School of Earth Sciences, Australian National University, 142 Mills Road, Acton, ACT 2601, Australia

Abstract The giant (>20 Moz) Telfer Au-Cu deposit is located in the Paterson Province of Western Australia and is hosted by complexly deformed marine Neoproterozoic metasedimentary siltstones and quartz arenites. The Telfer district also contains magnetite- and ilmenite-series granitoids dated between ca. 645 and 600 Ma and a world-class W skarn deposit associated with the reduced, ~604 Ma O’Callaghans granite. Based on monazite and xenotime U-Pb geochronology, Telfer is estimated to be older than O’Callaghans, forming between 645 and 620 Ma. Au-Cu mineralization at Telfer is hosted in multistage, bedding-parallel quartz-dolomite-pyrite-chalcopyrite reefs and related discordant veins and stockworks of similar composition that were emplaced into two NW-striking doubly plunging anticlines or domes. Mineralization is late orogenic in timing, with hot (≤460°C), saline (20 million ounces (Moz) Au and 0.7 million metric tons (Mt) Cu at an average grade of 0.75 g/t Au, 0.12% Cu has been defined to date, with historic production accounting for 14.5 Moz Au and 0.4 Mt Cu and current resources of 5.4 Moz Au and 0.3 Mt Cu (Newcrest Mining Limited, 2020). Economic Au mineralization is localized in complexly folded and faulted metasedimentary rocks of Neoproterozoic age, occupying an elongate surface footprint of 5 by 1 km. Mineralization in Main Dome (Fig. 2A, B), the largest deposit in the Telfer district, has been drill defined to approximately 1.5 km below the surface and remains open at depth. Forty-eight years of exploration activity and 43 years of mining, in combination with >2,090 km of drilling, have generated substantial data to permit refinement of the Telfer genetic model. Early syngenetic, sediment-hosted, exhalative models (Tyrwhitt, 1979, 1985; Turner, 1982) were superseded by an epigenetic model with possible input of mineralizing fluids from surrounding granitic intrusions (e.g., Goellnicht et al., 1989, 1991; Dimo, 1990; Rowins et al., 1997). Goellnicht et al. (1989) suggested a mix of magmatic and meteoric fluids † Corresponding

author: e-mail, [email protected]

were responsible for ore development and a granitic Au-Cu source, whereas Rowins et al. (1997) subsequently proposed the S and metals in Telfer were derived from convection and leaching of a sedimentary rock package by hydrothermal fluids circulating above a granitic heat source. More recently, Schindler et al. (2016) concluded, based on detailed fluid inclusion studies, that Telfer mineralization formed from purely magmatic fluids, with ore deposition occurring distal to the magmatic fluid source and with significant fluid buffering by local wall rocks. Previous studies focused primarily on the local mine area or individual prospects in the Telfer district, including Telfer (e.g., Goellnicht et al., 1991), the O’Callaghans W skarn (Schindler et al., 2012), and the 17 Mile Hill Au-Cu prospect (Rowins, 2000, Fig. 2A), rather than taking a holistic view of the entire system. This paper presents the evolution of the Telfer system in the context of the mineralized district, relating ore-forming events to the regional deformation and intrusive events that culminated in the formation of this giant Au-Cu deposit. Exploration and Mining History Regional mapping by the Bureau of Mineral Resources was completed in the Telfer area by 1959 (Wells, 1959); however, it was not until the late 1960s and early 1970s that prospectors recognized the Au and Cu potential of the region (Dimo, 1990). In 1971, Day Dawn Minerals NL undertook a regional sampling program in the district and outlined an area of Au

doi: 10.5382/SP.23.11; 23 p. Digital appendices are on the USB drive attached to the inside back cover and are also available online. 227

228

WILSON ET AL.

Fig. 1. Location of Telfer and regional geology of the Paterson Province, Western Australia. Modified after Whitaker et al. (2010).



TELFER Au-Cu DEPOSIT, PATERSON PROVINCE, WA

anomalism that subsequently led to the staking of claims over the identified gossans at Telfer by Newmont Australia Pty Ltd. in May 1972 (Dimo, 1990). Follow-up exploration activity by Newmont identified significant Au mineralization in two bedding-conformable layers that were termed the Middle Vale reef and the Eastern reefs (Fig. 2B). Exposure of gossanous material showed Middle Vale reef continuity in a large doubly plunging anticline called Main Dome, whereas the Eastern reefs were developed on the east flank of Main Dome. Similar gossans in an identical stratigraphic position were also identified in an adjacent anticline termed West Dome (Dimo, 1990; Fig. 2B). An intensive exploration and resource drilling program was undertaken by Newmont from 1972 to 1975 that defined an oxide, open-pit reserve of 3.8 Mt @ 9.6 g/t Au containing in excess of 1 Moz Au, hosted principally by the Middle Vale reef (Turner, 1982). In 1975, as a result of foreign ownership legislation, BHP purchased a 30% interest in the project. Mining commenced during 1975 at Main Dome, with Newmont as operator, and full production of 0.5 million tonnes per annum (Mtpa) was achieved in 1977. Initially, ore processing was by milling, cyanidation, and Merrill-Crowe gold recovery. During the 1980s, the potential was recognized for a large, low-grade oxide resource in the Main and West Domes. This supported a mill expansion in 1986 to increase crushing and grinding capacity, including conversion from Merrill-Crowe gold recovery to a carbon-in-leach (CIL) circuit. Further extensive metallurgical test work led to the establishment of a dump-leach operation that commenced in 1988, with an initial processing rate of 4 Mtpa and by the late 1990s, Telfer was treating 2.5 Mtpa of high-grade oxide ore through the mill and CIL circuit, and 15 Mtpa of low-grade oxide ore by dump leaching. In 1989, a sulfide flotation circuit was established to process Middle Vale reef chalcocite ore at a nominal 0.25 Mtpa. Corporate activity in 1990 led to the formation of Newcrest Mining Ltd. via a merger of Newmont Australia with BHP Gold Ltd. In May 1992, based on new geologic concepts hypothesizing an underlying granitoid source for mineralization, Newcrest board and management supported deep diamond drilling into central Main Dome. Drill holes were located on the availability of bench space in the active open pit. The first hole of the program (MRC09538; Fig. 2B) intersected the I30 deep reef system in the targeted structural position, returning intercepts of 14.3 m @ 7.8 g/t Au and 0.61% Cu from 902.2 and 11.55 m @ 19.3 g/t Au and 0.56% Cu from 955.65 m. Additionally, low-level Au (500 ppm) anomalism in the upper stratigraphic portion of the hole was followed up in 1993, leading to the discovery of the M10, M12, and M30 reefs. An exploration decline commenced from the existing Middle Vale reef underground workings and confirmed the presence of further reefs with economic Au grades, thus identifying the Telfer Deeps deposit. Resource-definition programs continued throughout the 1990s but in 2000 production ceased due to a combination of low gold prices and high cyanide-soluble copper material that could not be economically processed through the existing treatment facilities. In late 2002, a revised sulfide feasibility study was approved allowing for the development of a 20 Mtpa sulfide processing facility incorporating ore from both

229

open-pit (13–19 Mtpa) and sublevel-caving (4 Mtpa) operations. Telfer recommenced commercial production in 2005, with the mine producing ~450,000 oz Au in the 12-month period to June 2019. Regional Geologic Setting Telfer is located in the Paleo- to Neoproterozoic Paterson Province, a composite orogenic belt defined by geological and magnetic data that extends 2000 km in a northwest-southeast direction from the northern West Australian coast through to central Australia (Bagas et al., 2002; Bagas, 2004). In the Telfer region, the Paterson Province is flanked to the west and southwest by the Archean Pilbara craton and unconformably overlain to the northeast by sedimentary rocks of the Phanerozoic Canning Basin (Fig. 1). The oldest rocks of the Paterson Province are the Paleoto Mesoproterozoic Rudall Complex that is composed of the tectonically juxtaposed and lithologically distinct Talbot, Connaughton, and Tabletop terranes (Bagas, 2004; Fig. 1). Unconformably overlying the Rudall Complex are the broadly time correlative Neoproterozoic (ca. 850–830 Ma) Yeneena and Officer Basins, which are separated by the Vines-Southwest-McKay fault zone (Czarnota et al., 2009; Fig. 1). The Yeneena Basin has been subdivided into a stratigraphically lower Throssell Range Group and overlying Lamil Group, the latter host to the Telfer deposit. The Throssell Range Group has a minimum stratigraphic thickness of 7 km and comprises basal quartz-rich sandstones of the Coolbro Sandstone that are conformably overlain by turbiditic graywackes and carbonaceous shales of the Broadhurst Formation (Whitaker et al., 2010). The overlying Isdell Formation comprises dolomitic siltstones and local limestones; the nature of the contact with the underlying Broadhurst Formation remains unclear due to Cenozoic cover. The Lamil Group is divided, from bottom to top, into the Malu, Telfer, Puntapunta, Gardens, and Wilki Formations (Bagas, 2000; Fig. 2A), which are described more fully below. The Malu Formation is the main host to the Telfer AuCu deposit (Fig. 3) although mineralization also occurs in the Telfer and Puntapunta Formations. The Lamil Group sediments were laid down in depositional environments ranging from carbonate shelf through prograding turbidite fans to deep marine. The environment was interpreted by Williams (1990) to be an intracontinental basin, marked by slow, even subsidence, with major NW-trending faults such as the Camel-Tabletop fault (Fig. 1) having acted as important basin-controlling structures (Langsford, 2000). The Yeneena Basin has been subjected to two major deformation events, the Miles (650 Ma) and Paterson (1 km thick. The unit is conformable with the overlying Malu Formation, but stratigraphic relationships between the Isdell Formation and Throssell Range Group remain uncertain as the contact is either faulted or unexposed (Maidment et al., 2017). The Malu Formation is an ~1,500-m-thick sequence of massive to bedded, fine- to medium-grained sandstone, minor interbedded shale, and dolomite, and is the primary host to Telfer mineralization. Regionally, the Malu Formation also includes the ~580-m-thick Telfer Formation, but this is described as a separate unit here. Based on detailed work around the Telfer deposit, the Malu Formation has been divided into the following units, with the mineralized reefs noted below described in detail in the subsequent section on mineralization and alteration: Lower Malu Member (~500 m thick): Fine- to medium-grained quartz arenite with fine siltstone in the middle portion of the unit. Host to the Telfer B-reefs. The Lower Limey unit (on avg 10 m thick) is a lithologically distinct unit comprising a calcareous banded siltstone with interbedded calcareous and dolomitic grainstone and is situated at the transition between the Lower Malu and Middle Malu Members. Middle Malu Member (~290–320 m thick): Banded/laminated siltstone with lesser sandstone. Host to the Telfer A-reefs. Upper Malu Member (~620 m thick): Massive to thickly bedded quartz arenite. Host to the Telfer M-reefs. The top 100- to 120-m of the Malu Formation comprises an intercalated siltstone/sandstone sequence (referred to at the mine as Upper, Middle, and Lower Vale Siltstones, and Rim, Median, and Footwall Sandstones; Fig. 3). This includes the economically significant Middle Vale reef at the base of the Middle Vale Siltstone. The Telfer Formation lies immediately above the Upper Vale Siltstone of the Malu Formation and includes two main components: Outer Siltstone Member (~500 m thick): Fine-grained siltstone with minor interbedded quartz arenite and dolomitic grainstone.

Camp Sandstone Member (~80 m thick): Fine-grained quartz arenite. The upper parts of the Telfer Formation can also be correlated regionally with the basal parts of the Puntapunta Formation, indicating a transitional contact. The Puntapunta Formation is a sequence of dominantly clastic carbonate rocks interbedded with siltstone and shale, with minor chert, and is estimated to be around 2,000 m thick. The Puntapunta Formation is deeply leached resulting in a silcrete/calcrete duricust and as such is variably exposed, with large portions covered by Paleozoic sedimentary rocks. The Gardens Formation is a 700-m-thick sequence of thinly bedded, purple siltstone, shale, and sandstone. The unit is transitional between the Puntapunta and overlying Wilki Formation. The Wilki Formation is approximately 1,300 m thick and represents the upper part of the Lamil Group sequence observed in the Telfer district. The unit is characterized by fine- to medium-grained quartz sandstone, massive to bedded, minor shale and laminated sandstone. Deformation history The sedimentary rocks of the Yeneena Basin in the Telfer district, and broader Paterson region as a whole, record a complex deformation history associated with multiple episodes of horizontal and vertical crustal shortening, the principal of which is related to the Miles and Paterson Orogenies (Hewson, 1996; Bagas and Smithies, 1998; Bagas et al., 2002; Bagas, 2004). These events are summarized in Table 1. The first recognized event in the district (D1) produced recumbent monoclinal folds with SW-dipping axial planes and short northeast limbs and attenuated long limbs. Kinematic indicators imply top-to-the-south shear sense and infers the existence of an earlier upright folding event (DEF), which is not observed locally at Telfer, but has been recognized elsewhere in the Yeneena Basin (Hewson, 1996). This episode (DEF) was related to early northeast-southwest horizontal compression followed by vertical crustal shortening and SW-directed horizontal movement (D1). The large-scale, upright, open to tight NW-SE-trending folds (F2) and domes characteristic of the district (Fig. 2A) were formed during northeast-southwest horizontal shortening in D2 (D4 of Bagas, 2004). This event was associated with lower to subgreenschist-facies metamorphism (Bagas, 2000) and has been correlated to early stages of the Miles Orogeny (Hewson, 1996). This was followed by a period of NE-directed tectonic transport related to vertical crustal shortening (D3) that locally formed minor, mesoscale, asymmetric recumbent folds and/or refolded/tightened earlier D2 folds. Folds associated with D3 typically verge to the northeast, although vergence changes have been observed and are related to localized shear reversals or deformation partitioning. Horizontal shortening recommenced in D4 as sinistral transpression, resulting in cross folding of D2 folds and mesoscale upright open folding. NNW-SSE-trending axial planes indicate clockwise rotation of the principal horizontal stress orientation from northeast-southwest (D2 trend) to eastnortheast-west-southwest. The D4 event has been loosely correlated with the initial stages of the Paterson Orogeny and likely commencement of mineralization at Telfer and early



TELFER Au-Cu DEPOSIT, PATERSON PROVINCE, WA

233

Table 1. Deformation Events of the Telfer Region Based on Work by Hewson (1996), Bagas and Smithies (1998), Bagas et al. (2002), and Bagas (2004) Deformation event DEF

Regional characteristics Southwest and west directed thrust stacking and complex folding; northwest-southeast trending symmetrical upright folding

Principal horizontal stress orientation NE-SW compression

Correlation/references DE-F, Hewson (1996) D1-2, Bagas (2004)

D1

Recumbent monoclinal folds, fold axes striking from NW to NE

Vertical compression / NE-SW extension

D1, Hewson (1996) D3, Bagas (2004) D2, Hewson (1996) D4, Bagas (2004) D3, Hewson (1996)

Miles Orogeny

D2

Regional NW-SE striking folds

NE-SW compression (Telfer)

 

D3

Weak, asymmetric recumbent folding, subhorizontal axial planes; granite intrusion?

Vertical compression; NE-directed thrusting

Paterson Orogeny?

D4

 

D5

Early NE-SW compression, rotating clockwise to ENE-WSW ESE-WNW compression

 

D6

Mesoscale upright open folding striking NNW-SSE; sinistral movement along NW-SE-trending faults Monoclinal kink folds, interference folds; subvertical axial planes Dextral reactivation of D4 sinistral faults; NNWto NNE-striking brittle structures; dolerite diking

granite emplacement (Maxlow and Wilson, 2005; Batt et al., 2013). Continued clockwise rotation of principal compressive stress during the Paterson Orogeny generated D4-D5 monoclinal kink folds and crenulations consistent with minor ESEto WNW-directed compression and late, predominantly brittle deformation associated with extensional veins and normal faulting. The final major deformation event in the Telfer district, D6, was associated with N-S-directed compression, resulting in development of S- to SE-dipping, fault-hosted veins and conjugate horizontal extension veins (Batt et al., 2013), and localized reactivation of earlier structures. Granite emplacement continued up to this time, as evidenced by locally developed foliation in the O’Callaghans granite, located below premineralization cover rocks some 10 km southeast of Telfer (Maidment et al., 2010). This foliation has not been assigned to a deformation event as drill core was unoriented, but it is noteworthy that foliation is not commonly observed in either granites or sedimentary rocks of the Telfer district. Continued movement along right-lateral faults generated the largely postmineralization Graben and North faults that transect Telfer Main Dome and the Graben series faults at Telfer West Dome (see below). This event is also interpreted to be responsible for the emplacement of N-trending dolerite dikes in the region (Hewson, 1996; Langsford, 2000). The maximum age of the Miles Orogeny is constrained by the ca. 950 Ma age of detrital zircons in the Throssell Range and Lamil Groups (Bagas et al., 2002) and it is generally agreed that deformation and lower greenschist-facies metamorphism generated by the Miles Orogeny concluded prior to the emplacement of the Mount Crofton granite (ca. 650 Ma), which crosscuts the regional D2 folds north of Telfer (Bagas, 2004; Maidment et al., 2010; Whitaker et al., 2010; Fig. 2A). The Paterson Orogeny is even more loosely constrained at between 650 and 550 Ma based on a 40Ar/39Ar K-feldspar age obtained from the Rudall Complex near the Kintyre deposit (Durocher et al., 2003; Fig. 1) and the aforementioned foliation in the ca. 604 Ma O’Callaghans granite (Maidment et al., 2010).

NNW-SSE to N-S compression

D4, Hewson (1996) D5, Hewson (1996) D5, Bagas (2004) D5b, Hewson (1996) D6, Bagas (2004)

Granite intrusions Goellnicht et al. (1991) determined felsic intrusions of the Telfer district to be highly fractionated, metaluminous, I-type monzogranites to alkali-feldspar granites of calc-alkaline composition and of intraplate syn- to postorogenic setting. Two suites of granite have been defined, based on their oxidation state (Goellnicht et al., 1991). The Crofton suite (Fig. 2A) consists of highly fractionated, magnetite-series granites with 1 to 2 vol % biotite, whereas the Minyari suite granites are less-fractionated, ilmenite-series granites with 5 to 8 vol % biotite. Schindler et al., (2016) considered the Minyari granites to be part of the oxidized, magnetite-series granites, highlighting the likely composite nature of the two suites in the district. The O’Callaghans granite, considered part of the ilmenite-series Minyari suite, is a reduced, K-feldspar-biotite granite that has been intercepted only in drilling at the O’Callaghans W skarn deposit. Granites were emplaced along two broadly linear NE-striking trends to the northwest and southeast of the Telfer deposit (Fig. 2A), orientations that may represent deep, long-lived, basin-controlling or transfer structures (Langsford, 2000). These trends cut across the dominant regional NW-oriented D2 fold axes, although Minyari suite granites also occupy a third, NW-oriented belt between the two northeast trends, lying roughly subparallel to F2 (Fig. 2A). Additionally, regional gravity data (Fig. 2D) suggest that a large body of ilmenite-series granite underlies the broader O’Callaghans area, with a less-intense gravity low- and subtle low-intensity magnetic feature (Fig. 2C) that extends under the Telfer district north to the 17 Mile Hill prospect, raising the possibility that the entire Telfer district may be underlain by a composite ilmenite-series felsic batholith. Contrasting Pb isotope initial ratios imply that the two suites are not cogenetic (Goellnicht, 1991), and Schindler et al. (2016) noted that different granites show different degrees of mantle-derived magma, indicating the variable role of crustal melting and contamination during granite emplacement. Abundant mafic enclaves in the Minyari suite of granites (Langsford, 2000) also point to the involvement of an as yet poorly documented mafic melt during this magmatic event.

234

WILSON ET AL.

U-Pb geochronology (Fig. 4) clearly indicates the ilmenite-series O’Callaghans granite has a discrete emplacement age of ca. 604 Ma. The Minyari and some Wilki granites form a cluster of zircon ages around 630 Ma, suggesting this was a significant period of intrusion, largely also of ilmenite-series granites. Magnetite-series Wilki and Mount Crofton granites appear to be older, at ca. 645 Ma, although the high U content of the Mount Crofton granite has historically rendered this intrusion challenging to date (Nelson, 1995; Dunphy and McNaughton, 1998). Deposit Structure, Mineralization, and Alteration Structural architecture At surface, Main Dome comprises an open, doubly plunging, ENE-verging F2 anticline with a steeper northeast limb, a steeply SW-dipping axial plane, and shallow plunges to both the northwest and southeast (Figs. 5, 6). The F2 dome structure is locally overprinted by ENE-verging F4 folds with axial planes that strike overall northwest but are coincident with the F2 anticline in the central part of the dome (Fig. 7A, B).

This refolding event is best observed in Main Dome underground area, in particular, in the I30 Monocline Corridor located ~1,000 m below surface (discussed in more detail below; Figs. 6, 7B). These relatively late, crosscutting fold corridors are considered to be fault-propagated folds generated at the tip of a major ENE-verging D4 thrust at depth, and together with associated reefs, host the bulk of the economic mineralization at Main Dome. West Dome, located to the west of Main Dome, is structurally more complex, comprising two right-stepping, locally tight, ENE-verging anticlines with steep to overturned limbs (Pit 9 and Pit 10 anticlines; Figs. 5, 6), both of which have steep SW-dipping axial planes. The NW-striking axial planes broadly parallel the later Main Dome D4 fold event outlined above. Fold plunge varies from horizontal to 50° to the north-northwest (Pit 10 anticline) and horizontal to 45°to the south-southeast (Pit 9 anticline; Fig. 5). Mineralization styles Deep drilling and underground development since 2003 improved the understanding of the mineralization and paragen-

Fig. 4. Age constraints on the Telfer district intrusion, alteration, and mineralization. The Mount Crofton and Minyari granites form reasonably well-defined clusters of ages around 645 and 630 Ma, respectively. The O’Callaghans granite has a well-constrained age of ca. 604 Ma. Mineralization at Telfer has proven challenging to constrain, with multiple monazite and xenotime analyses returning a range of ages, the most reliable of which cluster between 645 and 620 Ma. Muscovite K-Ar ages from Telfer reef samples are considered thermally reset at ca. 600 Ma. Re-Os analysis of two molybdenite samples from the O’Callaghans W skarn returned ages in the 606 to 604 Ma range, consistent with zircon U-Pb ages from the causative O’Callaghans granite. Data sources: 1 Nelson, 1995; 2 Dunphy and McNaughton, 1998; 3 Nelson, 1999; 4 Maidment et al., 2010; 5 Nelson, 2001;6 Newcrest internal data. Unpublished data for monazite (Table A1) and molybdenite (Table A2) and analytical methodologies are available in the Electronic Appendix.



TELFER Au-Cu DEPOSIT, PATERSON PROVINCE, WA

235

Fig. 5. Geology of Main and West Domes at Telfer. Main Dome comprises a relatively simple NW-trending, doubly plunging anticline, whereas West Dome is structurally more complex, comprising two right-stepping, locally tight, ENE-verging anticlines with steep to overturned ENE-striking limbs. The location of the Middle Vale reef and E-reefs are highlighted, as well as the late-stage Leader Hill veins. Modified from Batt et al. (2013), original data source is Langsford (2000).

esis of Telfer (Maxlow and Wilson, 2005; Maxlow, 2012; Batt el al., 2013; Schindler et al., 2016). The two dominant styles of mineralization at Telfer, reefs and stockworks, are mainly composed of dolomite, quartz, and sulfides, principally pyrite and chalcopyrite. A summary of the main mineralizing events, described in detail below, is provided in Table 2. Reefs are strata-bound, bedding semiconcordant veins that show a complex textural and mineralogical zonation related to the host unit and the local structural geometry. They are best developed in laminated silty and carbonate-rich units immediately adjacent to thick quartzite units, and range in thickness from a few centimeters to >10 m. Occurring at the same stratigraphic position in Main and West Domes, reefs are texturally, mineralogically, and geochemically zoned depending on their proximity to the axial plane of fault-propagated folds. The highest grade (>10 g/t Au) and the thickest (to 10 m) reefs are typically associated with intense hydrothermal and structural brecciation of early-formed quartz veins in zones of localized steepening of bedding, typically in the hinge zones of thrust-propagated folds. Early-formed, layer-parallel, milky

white quartz veins are intensely brecciated and cemented by pyrite-chalcopyrite infill (Fig. 8A) and exhibit vertical zonation from a quartz-rich base through a zone of increasingly abundant sulfide cement to a sulfide-dominant top that has a greater amount of brecciation due to synmineralization shear strain (Fig. 8B). Local inclusion of strongly folded siltstone beds internal to the reefs with partially milled clasts of hydrothermal pyrite and quartz recemented by a second phase of sulfide mineralization (Fig. 8C) provides further evidence for synmineralization deformation. As structural complexity decreases laterally away from the D4 fault-propagated fold hinge, reef textures and composition change, displaying a greater component of preserved, well-bedded siltstones, with less sulfide alteration and hydrothermal quartz, and a concomitant decrease in gold grade. The downdip terminations of reefs are contorted, laminated, carbonate-altered, silty carbonate rocks. Deformation of bedding traces grades from highly contorted nearest the hinge regions to weakly deformed down the limbs. Decalcification and open space are also observed downdip. Open-space sul-

236

WILSON ET AL.

Fig. 6. Schematic oblique section through the Telfer deposit. Reef development at both Main and West Domes took place over a vertical distance of >1,400 m, from the E-reefs and Middle Vale reef that were exposed at the premining surface through to the B-reefs and Vertical Stockwork Corridor (VSC) in the core of Main Dome. Evident at Main Dome is the shallow, SW-dipping I30 Monocline Corridor that represents the locus of thrust-propagated folding emanating from the tip of an ENE-verging thrust fault farther downdip. Local back-thrusting associated with this synmineralization compressive deformation is evident in the I30 Thrust. Figure modified from Batt et al. (2013) whose section was based on structural detail and stratigraphic form lines derive d from Maxlow (2003). Geology should be regarded as schematic and for illustrative purposes only.

fide content decreases downdip, with sulfides frequently paralleling deformed bedding traces. Thickness variations are common, ranging from a few centimeters to 10 m, and individual reefs have distinct ore shoots and ore-shoot controls, depending on local lithological, rheological, and structural conditions. There are in excess of 20 individual reefs stacked over a vertical kilometer in Telfer Dome (Figs. 3, 6), many of which have historically provided high-grade ore (>10 g/t Au). The most economically significant was the Middle Vale reef, which produced an estimated 1.5 Moz of Au at an average grade of 30 g/t and had a horizontal footprint at Main Dome of 2.5 by 1.0 km (Fig. 2B). Veins and stockworks have an identical mineralogy to their adjacent reefs, and mineralization intensity relates to the degree of local structural complexity as well as the host-rock rheology. These occur as sheeted F4 axial-planar arrays (e.g., veins within the I30 Monocline), sheeted arrays discordant to both stratigraphy and the F4 (e.g., North Dipping veins), discrete vertical veins (e.g., Leader Hills veins) and strata-bound arrays, with stockwork veins best developed in massive quartzite units. Mutually crosscutting relationships between these vein

sets highlights a complex origin for mineralization although the deposit paragenesis has been established using the most frequently observed crosscutting relationships. Eastern reefs: The Eastern reefs occur in the lower 40 to 50 m of the Outer Siltstone Member (Fig. 6) and are defined by up to eight bands or layers of ferruginous siltstones (between 0.2 and 1 m thick) cut by variable amounts of associated quartz-sulfide veins. Extension veins between reefs are considered to have formed in response to bedding-plane shear and, locally, form pods of quartz-sulfide infill >10 m thick. Eastern reef mineralization is best developed in the hinge and western limb of West Dome anticlines and on the eastern limb of Main Dome, where the hinge zone has been eroded. Much of the hydrothermal mineralogy has been obscured by weathering and related oxidation. Middle Vale reef: The Middle Vale reef occurs near the top of the Upper Malu Formation, at the base of the Middle Vale Siltstone (Figs. 3, 6). The Middle Vale reef has been recognized over the entire extents of Main and West Domes, has an average thickness of 1 m and is up to 3 m thick on the eastern flank of Main Dome where Au grades are best developed

Bedding concordant veins

Bedding concordant veins, local carbonate replacement Bedding concordant veins, local carbonate replacement Discrete discordant veins and stockworks Sheeted discordant veins

I30 Quartz reef

Lower Limey unit (LLU)

Qtz-py-cp-dol (gray, white, pink) ab-Au Pink dol-qtz-py-cp-ab- Au Gray dol-qtz-py-cpy - Au

Py-qtz-dol-ab-Au

Py-qtz-gray dol-cp

Qtz-dol (gray, white, pink)-py-cpyAu-bn-cv-cc-gal-stb-sch-Bi tel Py-qtz-cp-gray dol - Au

Principal mineralogy1,2 Qtz-goe-Au; surface oxidation obscures primary mineralogy Py-qtz-cpy-Au; qtz-py-cpy; cc-cv-dg - mal Qtz-dol (gray and pink)-py-cpysch-Au; cc-bn Qtz-gray dol-py-cpy-Au

Ser Ser

Ser-py -ab

Ser - py

Ser

Ser

Ser - dol

Ser - py

Wall-rock alteration Ser

1-km-long, moderately SW-dipping (30°–50°) corridor of monoclinal folding on the northeast limb of Main Dome (Fig. 6). Monocline development was mostly restricted to the central portion of Main Dome and locally displaced the F2 axial plane to the east-northeast (Fig. 7B). The corridor extends from the refolded F4 hinge of the Lower Limey unit toward the surface, locally folding the stratigraphy of the Middle and Upper Malu Members into a subvertical orientation (Figs. 6, 9A). Au-Cu



TELFER Au-Cu DEPOSIT, PATERSON PROVINCE, WA

mineralization occurs in localized zones of stockwork veins and is best developed in sandstone to quartzite units as opposed to siltstones. Reef-style mineralization is poorly developed within the corridor and the monocline effectively acts as the lower limit to economic mineralization for the majority of the M-reefs. The I30 Thrust (Fig. 6) is a shallowly (10°–20°) NE-dipping zone of fracture-controlled pyrite-chalcopyrite-quartz-dolomite mineralization, which locally grades into crackle to mosaic breccia. The zone broadly emanates from the top of the Lower Limey unit fold hinge and extends into the northwest quadrant of Main Dome. The I30 Thrust is interpreted to be a back thrust that preferentially accommodated strain in the northwest quadrant of Main Dome. Leader Hill veins: Leader Hill veins (Fig. 5) are a series of subvertical, NE-striking structures with massive pyrite or pyrite-cemented breccia, and coarse quartz-carbonate infill. These features are observed in West Dome pits and host significant (>10 g/t) Au grades. NW-trending veins: The NW-trending veins (orientation relative to mine grid, which is 42.8° west of grid north) are ENE-trending veins that are typically subvertical to steeply N-dipping. They are most commonly recognized in the northwest quadrant of Main Dome, generally below the Lower Limey unit, but veins of similar orientation and mineralogy are recognized elsewhere in Main Dome.

241

The veins are up to 1 m thick and contain coarse quartz-pyrite-chalcopyrite with pink, gray, and/or white dolomite. Veins are commonly laminated, display evidence of multiple phases of reactivation, and have a mutual offset and truncation relationships with minor reefs below the Lower Limey unit. The NW-trending veins have a similar structural orientation to the Leader Hill veins at West Dome and may be genetically equivalent. N-dipping veins: The N-dipping veins (orientation is relative to mine grid) are a series of NE-trending veins that dip moderately (30°–50°) to the northwest. They are mineralogically similar to the NW-trending veins, being composed of coarsely crystalline pink and gray dolomite, quartz, pyrite, and chalcopyrite (Fig. 8D, E) and also generally occur below the Lower Limey unit on the western limb of Main Dome. The veins display a complex, commonly mutually crosscutting relationship with adjacent structures. Stockwork veins and breccias: Discrete corridors of stockwork veins and localized breccia extend to at least 500 m below the Lower Limey unit at Main Dome (1,500 m below surface) in an area referred to as the Vertical Stockwork Corridor (VSC) (Fig. 6). Mineralization is focused into several steeply SW-dipping fault corridors localized along the overturned limbs of two or more stacked fault-propagated folds. These corridors extend laterally for up to 1 km and their orientation mirrors the D4 fold axis. Gold grades may also be locally

Fig. 9. A. View looking north at the north end of Main Dome pit. A gently W-dipping corridor of fault-propagated anticlines is seen rising up the pit wall. The location of the Middle Vale Reef (MVR) is noted in the lower left of the photograph. Brittle quartz arenite beds have hematite staining due to sulfide oxidation, whereas interbedded siltstones weather white to pale gray as they generally lack hydrothermal sulfides. Discordant veins and stockworks preferentially developed in brittle quartz arenites. True width of the mineralized corridor is approximately 200 m. B. View looking northeast in Pit 9, West Dome. Gently dipping, partially oxidized E-reefs (E1 and E2) cut by discordant dolomite-quartz-pyrite veins. Note the increase in intensity of sulfide mineralization at the intersection of the E-reefs with the discordant veins, a structural setting that generated economically important ore shoots. Person for scale in lower right-hand corner.

242

WILSON ET AL.

enhanced at the intersections of D4 and D5 veins (NW-trending and N-dipping vein series). The breccia and stockwork vein mineralogy is variable, with the presence of pink and white dolomite-chalcopyrite-quartz-pyrite, suggesting an intimate association with all three sets of discordant vein mineralization and deformation. Of note, the VSC is distinctly more chalcopyrite rich than other stockwork and vein zones (Fig. 6). Stockwork veins also occur in the SW-dipping I30 Monocline corridor and are preferentially hosted in more brittle quartz arenites of either the Middle or Upper Malu Formation (Figs. 6, 9A). Examples include the Eastern Stockwork corridor, which is an elongate zone of dominantly shallowly SW-dipping veins, constrained to the quartz-rich sandstones between the M30 and M35 reefs within the I30 Monocline corridor (Fig. 6). The zone is only 30 to 40 m wide and 50 m high but extends for >1 km with a shallow northerly plunge. Stockwork veins and localized breccias are typically associated with and better developed adjacent to faults, particularly where a rheological contrast exists between host rocks or where high-pressure fluids have been focused into selected structures. Conjugate veins: A distinct set of S- to SE-dipping reverse fault veins and associated flat-lying tension veins also occur within parts of Main Dome (Batt et al., 2013). These veins are characterized by gray dolomite infill with lesser quartz, pyrite, and chalcopyrite. The veins both crosscut and terminate into earlier formed D4 reefs and discordant D5 veins implying late, likely D6 timing, associated with minor reactivation of earlier structures. Intrusions Direct evidence of a physical connection between Telfer mineralization and felsic intrusions is recognized only locally, with Sillitoe (2016) first recording the occurrence of thin leucogranite intrusions at depth (>1,000 m vertically) in the western limb of West Dome (drill hole WRC32801A; Fig. 5). This texturally complex felsic body, intersected over 10 m in drill core (WRC32801A, 1,053–1,063 m; Fig. 8F), comprises thin dikes of intimately and irregularly intergrown K-feldspar and quartz with lesser biotite and chalcopyrite (Fig. 8G, H). These thin intrusive units are underlain by a small massive quartz body cut by aplite and leucogranite stringers, with blebs of chalcopyrite (Fig. 8I). The best developed leucogranite interval, from 1,154.65 to 1,158.0 m, averages 0.35 g/t Au and 0.28% Cu, with individual samples of 0.8-m length grading 1% Cu. The presence of Au-Cu mineralization intimately associated with these leucogranite intrusions provides strong evidence in support of an intrusion-related origin for the Telfer deposit. Metamorphism, hydrothermal alteration, and metal zonation Prior to formation of Telfer, lower greenschist-facies metamorphism associated with the Miles Orogeny resulted in sutured quartz grains and recrystallization of feldspar to albite in the quartz arenite units. Minor metamorphic white mica grew in clay-bearing units and the matrix of carbonate-bearing rocks recrystallized, forming new calcite or dolomite. Minor sparsely disseminated pyrite and trace chalcopyrite were also deposited during this event.

The distribution of hydrothermal alteration assemblages at Telfer is strongly influenced by protolith composition. At the deposit scale, wall rocks proximal to reefs, discordant veins, and breccias are commonly replaced by a fine-grained white mica-albite-tourmaline-quartz-dolomite-chlorite-pyrite ± chalcopyrite ± rutile ± biotite assemblage. Fine-grained white mica and minor dolomite alteration is typically localized around the M-reefs and associated stockwork zones in the Upper Malu Member. Yellow-green, fine-grained mica alteration is also a characteristic feature adjacent to discordant quartz-carbonate-sulfide veins in the Middle and Lower Malu Units. Hydrothermal albite, characterized by a buff to palepink color related to fine hematite inclusions (Mason, 2010), and tourmaline are developed extensively in the deeper parts of the deposit and may have formed early in the deposit paragenesis (Mason, 2010). Distal to the deposit (20 m.y. older than this ca. 604 Ma W skarn and its causative granitic intrusion. A punctuated development of mineralization at Telfer is implied from the poor correlation and decoupling of Cu and Au at the reef scale, reflecting episodic mineralization pulses and variable structural reactivation in response to changes in regional stress fields. The linking of discrete mineralizing events at Telfer to changes in regional stress fields suggests that ore formation may have occurred episodically over many millions of years. This is in contrast to other ore deposit types genetically related to felsic intrusive complexes, such as porphyry Cu deposits, where geochronological data indicates the duration of magmatic-hydrothermal systems to be commonly 50 m, with a spacupdip along the Towns fault at the upper contact of the Wil- ing from 1 to >2 m in the orebodies (Clark, 1980; Clout et liamstown Dolerite with the Paringa Basalt (Fitzgerald and al., 1990; Ridley and Mengler, 2000; Mueller, 2015). The Nixon, 2016). The upper Hidden Secret orebody is hosted in veins predominantly occur as two contemporaneous sets: the upper, granophyric portion of the Williamstown Doler- (1) steeply NW- or SE-dipping, and (2) shallowly to moderite and the lower Paringa Basalt intercalated with shale; the ately N-dipping (Fig. 10E). Both vein sets are characterized lower Hidden Secret orebody is predominantly hosted within texturally by buck quartz (dominant), fibrous quartz, and

Fig. 9. Geologic plan map (A) and longitudinal section (B) of the Mt. Charlotte deposit (adapted from Ridley and Mengler, 2000, and Mueller, 2015), illustrating the distribution of stockwork ore in the Charlotte (COB), Reward (ROB), Maritana (MOB), and Northern (NOB) orebodies. The stockwork orebodies have a strong affinity for Unit 8 of the Golden Mile Dolerite, which is segmented into fault-bounded blocks by SW-dipping thrusts such as the Neptune and Flanagan faults as well as N-striking dextral faults such as the Charlotte, Reward, and Maritana faults. Geologic plan map (C) and cross section (D) of the Mt. Percy deposit (adapted from Sully, 2010), illustrating two distinct settings where overlapping gold-telluride lodes and stockwork-type veins define orebodies: (1) in feldspar-quartz porphyries within the Hannan’s Lake Serpentinite where small hornblende-albite porphyries and kersantites occur and (2) along the contact between Williamstown Dolerite and Devon Consols Basalt in association with the Kapai Slate.

266

MCDIVITT ET AL.



SETTING, Au MINERALIZATION, KALGOORLIE Au CAMP, YILGARN CRATON, WA

crack-seal laminations (Clark, 1980; Clout et al., 1990; Ridley and Mengler, 2000). The veins are surrounded by symmetrical alteration halos that extend 0.05 to 1.0 m from the vein and overprint the regional chlorite-calcite alteration in Unit 8 of the Golden Mile Dolerite, which is the primary host for stockwork-type mineralization at Mt. Charlotte, Drysdale, and Golden Pike (Clark, 1980; Clout et al., 1990; Bateman and Hagemann, 2004). There is vertical zoning at the Mt. Charlotte mine (Clark, 1980; Bateman and Hagemann, 2004; Mueller, 2015): In the upper 450 m, the vein alteration halos comprise an inner, mineralized (6–12 ppm Au), bleached selvage of ankerite-sericite-albite-pyrite-siderite-rutile, and an outer, barren halo of albite-chlorite-magnetite-pyrite. From 600 to 800 m in depth the abundance of albite increases from 20 to 30 vol % in the inner bleached alteration zone at the expense of muscovite, and pyrrhotite becomes dominant over pyrite in the outer albite-chlorite-magnetite-pyrite alteration zone. At depths of >800 m, pyrrhotite becomes dominant over pyrite in both the inner and outer alteration zones, and albite increases to 40 vol % of the inner bleached zone. Gold grades in the Mt. Charlotte-type stockwork veins generally vary from 0.02 to 2 ppm, with the highest grade intercepts having values from 3 to 19 ppm (Table 1). The veins are geochemically distinct from the Fimiston- and Hidden Secret-type lodes as they lack the typical Au-Hg correlation and Te/Mo enrichment (Fig. 10K-M; Table 1). Timing of mineralization Structural studies constrain the timing of mineralization relative to the development of faults, folds, cleavages, and the emplacement of intrusions (Mueller et al., 1988; Swager et al., 1989; Bateman and Hagemann, 2004; Gauthier et al., 2004a, b; Weinberg et al., 2005; Mueller, 2015, 2020a). Despite variations in deformation schemas and the interpreted structural timing of mineralization events (cf. Fig. 7A), shared conclusions include the following: (1) the Fimiston lodes are the structurally earliest mineralization event; (2) the Oroya shoot formed after the emplacement of the Fimiston lodes; and (3) the Mt. Charlotte stockwork-type mineralization is

267

the youngest mineralization event, developed within faultbounded blocks related to the D4 deformation event. These conclusions are supported by the fact that the regional, NNW-trending cleavage deforms and overprints the Fimiston lodes (Fig. 7C); in contrast, the Mt. Charlotte stockwork veins are typically undeformed (e.g., Fig. 10D). Additionally, in the Golden Pike, Drysdale, and Hidden Secret areas, the stockwork mineralization is observed to crosscut the lode mineralization (e.g., Fig. 10B). Although Bateman and Hagemann (2004) suggested that the Fimiston lodes formed during the development of early structures such as the Golden Mile fault and Kalgoorlie anticline, most workers consider the Fimiston lodes to have formed synchronously with later sinistral deformation (i.e., D2-D3, Fig. 7A) in Riedel shear-type structures that correspond to the different lode orientations (i.e., Main, Caunter, Cross; Fig. 10I; Mueller et al., 1988; Swager, 1989; Mueller, 2015, 2020a). The Oroya shoot is interpreted to be a later stage of Te-rich mineralization than the Fimiston lodes because its bounding structures—the Oroya hanging-wall and footwall shear zones—crosscut and offset earlier lode structures (Fig. 8C; Bateman and Hagemann, 2004; Mueller, 2020a). Southwest-dipping thrust faults similar to the Oroya shear zones, such as the Cadiz and Salamanca faults, crosscut Ag-rich lodes in the lower Hidden Secret orebody (Fitzgerald and Nixon, 2016), suggesting that Hidden Secret mineralization also predates Oroya mineralization. These structural relationships, combined with local and regional U-Pb zircon constraints (Fig. 7B), place the formation of the Fimiston-, Hidden Secret-, and Oroya-type lodes within the D2-D3 time period (ca. 2675–2655 Ma), with the formation of the Fimiston lodes and the Oroya shoot early and late, respectively, within this interval. The overprinting Mt. Charlotte-type stockwork mineralization must have formed subsequent to 2648 ± 6 Ma, when the Liberty Granodiorite was affected by the late, D4 dextral faulting event (Fig. 5B). Direct dating of ore- and alteration-stage monazite, xenotime, and sericite from Fimiston and Oroya mineralized zones yields age ranges that are younger than the structurally inferred minimum age for Te-rich mineralization of ca. 2655

Fig. 10. Gold mineralization types of the Kalgoorlie gold camp. A. Fimiston-type lode mineralization comprising a banded, carbonate-quartz vein with a distinctive colloform texture from the Golden Mile Super Pit; DDH OLGD001G at 1,330 m. B. Hidden Secret-type lode mineralization as a large banded vein at the end of the Hidden Secret 310 South Footwall. Gray and green fragments of Williamstown Dolerite are included in the vein, with the green fragments reflecting carbonate-fuchsite alteration. A late, stockwork-type vein (SWV) cuts across the lode mineralization. C. Oroya-type mineralization from the 160-m level of the Oroya shoot. The mineralization is banded in nature and of high sulfide content, with the deep-green coloration of wall-rock fragments reflecting alteration assemblages that include V-bearing muscovite. D. Mt. Charlotte stockwork-type veins in Unit 8 of the Golden Mile Dolerite. The veins are from the 1680 level of the underground Mt. Charlotte mine and display well-defined alteration halos expressed by the tan-colored wall rock. E-G. Contoured poles (equal area stereonet; 1% area contours; start = 2%; CI = 2%) to vein and bedding-plane measurements from Kalgoorlie Consolidated Gold Mines drill core database. E. Stockwork veins from the Mt. Charlotte underground mine; steeply NW- or SE-dipping, and shallowly to moderately N-dipping orientations are dominant. F. Hidden Secret-type lodes in the lower Hidden Secret orebody. A dominant NW-striking and steeply SW-dipping orientation is well defined. G. Kapai Slate bedding in the lower Hidden Secret orebody. The π axis to a best-fit great circle defines a shallowly SSE-plunging fold hinge, parasitic to the Kalgoorlie anticline. This hinge is a feature that controls the distribution of stockwork-type mineralization in the orebody. H. Schematic three-dimension diagram of the geologic setting and geometry of the Oroya shoot (adapted from Lungan, 1986). Localized deformation at the Oroya Shale (OS) contact between the Golden Mile Dolerite (GMD) and Paringa Basalt (PB) resulted in the nucleation of the bounding Oroya hanging- and footwall shear zones (OHW and OFW, respectively). I. Stereographic depiction of the different Fimiston lode orientations (after Mueller, 2020a). J-M. Bivariate scatterplots of Ag, Te, Hg, and Mo vs. Au. Data are from the KCGM drill-core database and include gold-mineralized and barren samples from the southern Fimiston, Hidden Secret, Golden Pike, and Mt. Charlotte areas. Arrays of data points near and parallel to the plot axes are artifacts of the analytical detection limits for the elements portrayed.

268

47200N

47400N

47600N

MCDIVITT ET AL.

100m Alteration Types Sericite

Shear Zone

Vanadian Sericite Ephesite Ankerite-Dolomite

Main Lode West Leg Lake View Mine

Fig. 11. Longitudinal section of the main lode, west leg from the Lake View mine adapted from Clout (1989). The section displays vertical zonation in wall-rock alteration related to Fimiston mineralization: (1) upper, low-grade ankerite-quartzephesite-pyrite alteration (green); (2) intermediate, high-grade ankerite-quartz-vanadian muscovite-pyrite alteration (blue); and (3) lower, ankerite-quartz-sericite-pyrite alteration of variable gold grade (orange and pink). The alteration zones are developed in Units 9 and 10 of the Golden Mile Dolerite. Northing markers correlate to the local KLV mine grid.

Ma, and conflict with the observed overprinting relationships (Fig. 5B). SHRIMP U-Pb monazite ages and 40Ar/39Ar sericite ages from the Fimiston-type lodes overlap in a time interval from 2655 to 2620 Ma, with SHRIMP U-Pb zircon (2642 ± 6 Ma) and monazite (2637 ± 20 Ma) ages from an altered, synmineralization lamprophyre dike in the Oroya hanging-wall shear zone (Kent and McDougall, 1995; McNaughton et al., 2005; Vielreicher et al., 2010). An even younger age for the Fimiston gold event of 2610 ± 4 Ma has been inferred from 40Ar/39Ar sericite geochronology (Heath, 2003). SHRIMP U-Pb xenotime ages from the Mt. Charlotte stockwork-type mineralization define an age range from 2670 to 2630 Ma (Rasmussen et al., 2009; Vielreicher et al., 2010; Mueller, 2020a). This age range largely overlaps with the monazite-defined age range for the Fimiston- and Oroya-type mineralization; however, the maximum age of the xenotime-defined range for the stockwork mineralization is 12 m.y. older. A second cluster of xenotime and sericite ages delineate a younger age range from ca. 2620 to 2585 Ma for the stockwork mineralization (Phillips and Miller, 2006; Vielreicher et al., 2010). These geochronologic results have been interpreted by Vielreicher et al. (2010, 2015) to indicate a broadly synchronous timing of formation for the Fimiston, Oroya, and Mt. Charlotte types of mineralization at ca. 2.64 Ga (Fig. 5B). This model of a single mineralization episode conflicts with the outcomes of the structural studies and field observations, which support multiple, distinct mineralization events over the course of ca. 20 to 50 m.y. (cf. Bateman and Hagemann, 2004; Robert et al., 2005; Weinberg et al., 2005; Bateman and Jones, 2015; Mueller, 2020a). One possibility to consider, but difficult to assess, is that the young ages from phosphate minerals and sericite actually reflect complete or partial resetting by subsequent thermal or hydrothermal events. Clearly, further work is required to better constrain the timing of mineralization in the camp.

Genetic Considerations Numerous genetic models have been proposed for the origin of gold mineralization in the camp. Early syngenetic models have been discarded in favor of epigenetic ones (cf. Clout, 1989), which remain debated, in particular concerning the derivation of ore fluids from magmatic (Clout, 1989; Bateman and Hagemann, 2004; Mueller and Muhling, 2013; Mueller 2007, 2015) or metamorphic processes (Golding and Wilson, 1983; Phillips, 1986; Phillips et al., 1987; Vielreicher et al., 2010, 2016). A summary of different interpretations for epigenetic mineralization is presented in Figure 12, grouped into the following categories: (1) a model that favors predominantly metamorphic fluids and a single hydrothermal episode for all ore types in the camp (Phillips, 1986; Phillips et al., 1987; Vielreicher et al., 2010, 2016 ); (2) a model that recognizes multiple mineralization events and suggests that both early and late mineralization formed by similar magmatic-hydrothermal processes associated with high Mg, monzodiorite-tonalite intrusions (Mueller, 2007, 2015, 2020a; Mueller and Muhling, 2013); and (3) a model that identifies multiple mineralization events and favors a magmatic-hydrothermal ± surficial fluid model for the early, Te-rich Fimiston and Oroya mineralization, but recognizes the late Mt. Charlotte mineralization as typical mesozonal orogenic mineralization (Clout, 1989; Bateman and Hagemann, 2004), which may be the product of metamorphic fluids. Textural, mineralogical, and geochemical characteristics of mineralization and inferred P-T conditions Principal differences among genetic models relate to different interpretations of the types of mineralization in terms of single or multiple hydrothermal events/genetic mechanisms. Additional characteristics and constraints are considered here



269

SETTING, Au MINERALIZATION, KALGOORLIE Au CAMP, YILGARN CRATON, WA Table 1. Textural, Geochemical, and P-T Estimate Summary of The Fimiston, Hidden Secret, Oroya, and Mt. Charlotte Mineralization Types

Mineralization type Textural features

Hidden Secret lodes Banded and breccia veins

[Au] (20–80th percentile):1

Fimiston lodes Coliform, crustiform, banded, and breccia textures n = 166, 0.2 to 12 ppm

n = 303, 0.1 to 6 ppm

-

n = 227, 0.02 to 2 ppm

[Au] (90–100th percentile):1

n = 166, 20 to 222 ppm

n = 303, 11 to 1,640 ppm

-

n = 227, 3 to 19 ppm

[Te] (20–80th percentile):1

n = 103, 0.1 to 10 ppm

n = 175, 0.2 to 32 ppm

-

n = 227, 0.05 to 0.4 ppm

[Te] (90–100th percentile):1

n = 103, 17 to 370 ppm

n = 175, 66 to 500 ppm

-

n = 227, 0.6 to 2 ppm

Trace metal correlations (Log10 Pearson ≥0.50):1

n = 103, Ag (0.70), Te (0.71), Hg (0.68), Mo (0.51)

n = 175, Ag (0.78), Te (0.88), Bi (0.53), Hg (0.79), Mo (0.63), Pb (0.54), Sb (0.59)

-

n = 227, Ag (0.64), Te (0.65)

Distinguishing elemental signature:1,2

Ag, Te, Hg enrichment and Au correlation; B, Mo, Sb enrichment

Major Ag, Te enrichment and Au correlation; Bi, Hg, Mo, Pb, Sb enrichment and Au correlation

O Isotopes – quartz/carbonate (δ18O):2

~10 to 15‰ (quartz) and ~9 to 14‰ (carbonate)

-

D Isotopes – sericite/ inclusion waters (δD):2

~–45 to –18‰ in sericite; high variability in fluid inclusion waters from ~–60 to –30‰

-

Fluid inclusions:2

Inferred parental H2O-CO2 fluid with X CO2 = 0.02 to 2 and wt % NaCl 1 to 8; 2-phase H2O inclusions of variable salinity

C Isotopes – carbonate (δ13C):2 S Isotopes – pyrite

(δ34S;

Pressure-temperature estimates:2

Δ33S):2

Mt. Charlotte Planar buck quartz-carbonate, lack of banding

Similar to Fimiston and Hidden Secret with major relative enrichments in S and V

Lacks the Te, Hg, Mo, and Sb enrichments characteristic of the Fimiston and Hidden Secret

(Limited data) ~10 to 13‰ (carbonate)

~10 to 13‰ (quartz) and ~9 to 13‰ (carbonate)

-

(Limited data) H2O-CO2 inclusions and 2-phase H2O inclusions

~–5 to –2‰

Oroya lodes Breccia, silica mottling, and banded veins

-

(Limited data) CO2-rich (±CH4) CO2-H2O inclusions and 2-phase H2O inclusions

Inferred parental H2O-CO2 fluid with XCO2 = 0.2 to 0.3 and wt % NaCl ~3; Carbonic inclusions can be CH4-rich; 2-phase H2O inclusions of variable salinity

-

(Limited data) ~–3‰

~–8 to –2‰

δ34S

-

δ34S

δ34S range 0 to 15‰; Majority Δ33S range 0 to 0.5‰

Pressure variation from ~3 to 1 Kb; Temperature range from ~400° to 200°C

-

Pressure variation from Pressure variation from ~3 to 1 Kb; ~2.5 to 0.5 Kb (limited Temperature range from ~450° data); Temperature range to 200°C from ~425° to 250°C

range –10 to +5‰; Majority Δ33S range 0 to 0.5‰

range –10 to +5‰; Wide range Δ33S range from ~–1 to +1‰

Notes: Dash indicates a lack of available data 1Data sourced from the Kalgoorlie Consolidated Gold Mines multielement database 2Compiled data and references are available in the Electronic Appendix; distinguishing elemental signature is based on Kalgoorlie Consolidated Gold Mines multielement data and the mass balance calculations; reported isotopic ranges encompass the majority of the data but exclude outliers

to further assess these different interpretations, as summarized in Table 1 and presented in the Electronic Appendix. There are clear textural differences between the Te-rich Fimiston, Hidden Secret, and Oroya mineralization types and the Te-poor Mt. Charlotte stockwork veins. The former often display colloform-crustiform textures and/or are banded or brecciated in nature, and the latter are typically characterized by planar buck quartz infill (e.g., Fig. 10A-D). Mineralogically, the Fimiston lodes are distinguished by telluride minerals, sulfosalts, tourmaline, and mineral species such as hematite, magnetite, and anhydrite, reflecting oxidized conditions (Clout, 1989; Clout et al., 1990; Mueller, 2007). In contrast, the Mt. Charlotte stockwork veins display a more reduced mineral assemblage, locally including pyrrhotite, and are distinguished by the presence of scheelite (Clark, 1980; Mueller, 2015). The bivariate elemental plots (Fig. 10J-M) illustrate that the Fimiston and Hidden Secret lode

mineralization types show distinct Te-Hg-Mo enrichment and Au-Hg/Au-Mo correlation in comparison to the Mt. Charlotte stockwork veins. Quartz and carbonate oxygen isotope data from Fimiston lodes and Mt. Charlotte veins show significant overlap (Table 1; Fig. A1B, C, E), but δ18O values from the Fimiston samples tend to be higher than those from Mt. Charlotte, which may indicate formation from an isotopically heavier fluid or, alternatively, at a lower temperature. Evidence for surficial fluid involvement is equivocal as δ18OQuartz-δ18OH2O and δDSericiteδDH2O fractionation is temperature dependent. On the basis of δ18OQuartz and δ18OCarbonate values, a surficial fluid component could be invoked in both Fimiston and Mt. Charlotte mineralization types over a 200° to 300°C temperature range. Using the δD values of sericite in the Fimiston lodes, the calculated δDH2O fluid values over a 300º to 400°C temperature range are permissible with surficial fluid infiltration, as a number

270

MCDIVITT ET AL.

ca. 2640 Ma Broadly Synchronous Mineralization

ca. 2680-2640 Ma Multiple Mineralization Events

1 Fimiston 2 Mt. Charlotte

Meteoric Fluids Bateman and Hagemann (2004)

Clout (1989)

Magmatic Fluids

1

1

2

Mueller (2007, 2015) Phillips (1986); Phillips et al. (1987)

Clout (1989)

Metamorphic Fluids

Mueller and Muhling (2013) Bateman and Hagemann (2004) 2

Vielreicher et al. (2010, 2016)

Genetic Models Fig. 12. Summary diagram of competing absolute timing and genetic models for gold mineralization in the Kalgoorlie gold camp. See text for further details.

of the fluid values fall within the δD range of –20 to 0‰. Fluid inclusion studies of Fimiston and Mt. Charlotte mineralization document similar fluid inclusion types and compositions (Table 1; Figs. A2–4, 6). In both cases, the parental ore-forming fluid is considered a low-salinity, H2O-CO2 fluid. Fluid inclusions in Fimiston lodes are generally devoid of CH4, whereas those in the lower level of the Mt. Charlotte mine have a carbonic component of variable XCH4 from 0 to 1. In both mineralization types, two-phase liquid-vapor H2O inclusions of variable salinity coexist with carbonic inclusions, interpreted in some instances to record phase immiscibility. Low-salinity varieties of H2O inclusions in Fimiston mineralization have been interpreted to record the influx of a surficial fluid (Clout, 1989). The carbon isotope signature differs between the Fimiston and Mt. Charlotte mineralization types (Fig. A1E). Fimiston samples tend to have higher δ13C values than the majority of Mt. Charlotte samples (Table 1). As δ13C fractionation is affected by intensive fluid properties such as T, fO2, and pH, these contrasting carbon isotope signatures may record differences in hydrothermal fluid chemistry between the Fimiston and Mt. Charlotte systems. Alternatively, the difference in δ13C carbonate values may be attributable to different carbon source reservoirs. There are significant ranges in the δ34S pyrite data from the Fimiston, Oroya, and Mt. Charlotte mineralization types (Table 1; Fig. A1F). However, the majority of Fimiston δ34S pyrite values and a large number of Oroya δ34S pyrite values are significantly 0‰. The majority of Δ33S values for Fimiston and Mt. Charlotte mineralization overlap from 0 to 0.5‰ (Fig. A1G). The Oroya samples are the most distinct in their Δ33S values and display a large range from ~–1.1 to +1.1‰. These values have been interpreted as predominantly juvenile S (Δ33S = 0 ± 0.2‰), mixed with lesser supracrustal S characterized by a mass independent fractionation signature (Δ33S≠0 ± 0.2‰) sourced from local country rocks (Godefroy-Rodríguez et al., 2020). The range

of δ34S pyrite values in the Fimiston samples is suggested to record a variety of processes (Phillips et al., 1986; Clout, 1989; Bateman and Hagemann, 2004; Godefroy-Rodríguez et al., 2020), including: (1) boiling of ore fluids, (2) variance in fluid oxidation state, and (3) mixing of S from isotopically distinct sources. The prevalence of negative δ34S values in Fimiston pyrite is typically attributed to oxidized fluid conditions. The nature of the oxidized fluid for the Fimiston event is debatable, and suggestions include: (1) an initially oxidized fluid of likely magmatic origin (Clout, 1989; Bateman and Hagemann, 2004), (2) mixing of a reduced fluid with a surficial fluid (Clout, 1989; Bateman and Hagemann, 2004), and (3) interaction of a reduced fluid with magnetite-bearing portions of the Golden Mile Dolerite (Phillips et al., 1986; Evans et al., 2006). Godefroy-Rodríguez et al. (2020) documented negative δ34S values in the earlier paragenetic stages of Fimiston mineralization and δ34S values in a late stage of the event that are generally near 0‰ and higher. This supports the presence of an oxidized fluid during the early stages of the Fimiston mineralization event. Deuterium and oxygen isotope data indicate that fluid oxidation related to surficial fluid infiltration is plausible. Because Mt. Charlotte mineralization is primarily hosted within magnetite-bearing portions of the Golden Mile Dolerite and does not display the same hematite-magnetite-sulfate mineralogy and prevalent negative δ34S values, it is difficult to invoke wall-rock interaction as a fluid oxidation mechanism. Pressure-temperature estimates for mineralization have been made using fluid inclusion-, mineral equilibria-, and stable isotope-based methods (Clark, 1980; Golding and Wilson, 1983; Ho, 1986; Clout, 1989; Bateman et al., 2001b; Heath, 2003; Mernagh et al., 2004; Mueller, 2015, 2020b). Estimates for the Fimiston, Oroya, and Mt. Charlotte mineralization types generally overlap in a P range from 300 to 100 MPa and T range from 400° to 200°C (Table 1; Fig. A5), suggesting that the various types of mineralization may have formed in broadly similar P-T environments. Evidence for phase immiscibility infers pressure cycling and the P range from 300 to



SETTING, Au MINERALIZATION, KALGOORLIE Au CAMP, YILGARN CRATON, WA

100 MPa may record transitions from lithostatic to hydrostatic pressures. One notable outlier in the P-T estimate dataset is that of Clout (1989), wherein it is suggested that Fimiston mineralization may have formed in a near-surface environment (i.e., ≤1-km depth), with fluid temperatures ranging from ~100° to 250°C. These conditions are not supported by the P-T inferences cited above. However, high-level textures, the indication of a surficial fluid component, and vertical zonation of the Fimiston lode alteration, together with the documented presence of low-temperature minerals such as realgar, are supportive of a shallow level of emplacement for Fimiston mineralization (Clout, 1989). Therefore, it remains difficult to reconcile the textural, mineralogical, and geochemical differences between the Fimiston and Mt. Charlotte mineralization types in the context of the foregoing P-T considerations. Metamorphism and metamorphic fluid models These models favor a single mineralizing fluid generated by the metamorphism of mafic-volcanic and/or sedimentary rocks, with the formation of the different mineralization types contemporaneously or penecontemporaneously at ca. 2.64 Ga. Devolatilization during the greenschist to amphibolite transition is invoked in the production of low-salinity, H2O-CO2 (~0.2 XCO2) ore fluids, with gold deposition in greenschist-facies metavolcanic rocks at the brittle-ductile transition, with a broadly syn- to postpeak metamorphic timing (cf. Goldfarb and Groves, 2015, and references therein). Peak greenschist-facies metamorphic conditions in the gold camp are defined by actinolite/tremolite-chlorite-albite-zoisite-epidote-white mica-carbonate assemblage in mafic rocks. Because the axial planar cleavage to the Kalgoorlie anticline is defined by an actinolite-chlorite-albite assemblage, greenschist-facies metamorphism is considered to have been coeval with the development of D1 structures (e.g., Phillips, 1986; Mueller, 2020a). The widespread chlorite-calcite alteration overprints the greenschist-facies mineralogy and is considered to be related to the Fimiston lodes (Phillips, 1986), which are inferred to occupy post-D1 structures (Swager, 1989; Mueller, 2020a); this suggests a postpeak metamorphic timing for the Fimiston gold event (Phillips, 1986; Mueller, 2020a). A similar timing relative to peak metamorphism is suggested for the Mt. Charlotte stockwork-type mineralization, as the alteration halos of these veins overprint the regional chlorite-calcite alteration (Clark, 1980; Clout et al., 1990; Bateman and Hagemann, 2004). Therefore, a metamorphic fluid model is permissive given the nature of metamorphism and the relative timing relationships between the mineralization events and peak metamorphism. In addition, fluid inclusion studies suggest that the main ore fluids associated with both the Fimiston and Mt. Charlotte mineralization types are comparable low-salinity, H2O-CO2 fluids, which may be considered as further supporting evidence for a metamorphic fluid model and a single mineralizing episode in the camp (Phillips, 1986; Phillips et al., 1987; Vielreicher et al., 2010, 2016). Magmatism and magmatic fluid models A critical difference among the proposed magmatic-hydrothermal models is the nature of magmatism invoked in the mineralizing process: Clout (1989) and Bateman and Hagemann (2004) considered the dacitic, subalkaline, leucocratic

271

porphyries to be subvolcanic feeders to the Black Flag Group that may have represented the volcanic edifice developed cotemporally with Fimiston-type mineralization. In contrast, Mueller and Muhling (2013) and Mueller (2015) considered the andesitic, alkaline, mesocratic porphyries to be critical in the mineralization process, largely due to the similarities these intrusions share with the high Mg, monzodiorite-tonalite suite in the Mt. Shea area to the south (i.e. Ca, Fe, Mg, P, V, Ni, Cr enrichments and similar types of associated mineralization). Whereas the leucocratic porphyries are consistently crosscut by Fimiston mineralization, the mesocratic porphyries both crosscut and are crosscut by Fimiston mineralization, suggesting temporal overlap with Te-rich mineralization (Gauthier et al., 2004b; Mueller, 2020a). Both models consider the Te-rich and oxidized nature of the ore as an indication of magmatic affinity (Clout, 1989; Bateman and Hagemann, 2004; Mueller and Muhling, 2013). The km-scale lateral Au-Ag zonation in Te-rich lode mineralization from the Golden Mile to the Hidden Secret area is interpreted by Mueller and Muhling (2013) to be indicative of a magmatic-hydrothermal origin. Whereas Clout (1989) and Bateman and Hagemann (2004) considered the possibility of a distinct genesis for the Te-poor, Mt. Charlotte stockwork mineralization type, Mueller (2015) proposed a genetic connection between the Mt. Charlotte stockwork mineralization and the mesocratic porphyries. This is premised on inferred temporal overlap between the mesocratic porphyries and Mt. Charlotte mineralization based on U-Pb geochronology (Fig. 5). However, the large errors associated with these ages prevent any firm conclusion (i.e., 2651 ± 9 Ma age for Mt. Charlotte mineralization and 2663 +11/–9 Ma for the mesocratic porphyries; Mueller, 2020a). Presently, no examples of mesocratic porphyries crosscutting stockwork veins have been documented, suggesting that the porphyries may have been emplaced prior to the Mt. Charlotte mineralization event. Concluding Remarks Gold mineralization in the Kalgoorlie gold camp can be broadly subdivided into structurally early, telluride-bearing lodes (Fimiston, Oroya, and Hidden Secret) and structurally late, telluride-poor quartz-carbonate vein sets (Mt. Charlotte). All mineralization types have a clear affinity for the competent, quartz-rich, granophyric Unit 8 of the Golden Mile Dolerite sill, which hosts the majority of mineralization in the Golden Mile Super Pit and the Mt. Charlotte deposit. To a lesser degree, basalts and ultramafic units—such as the Paringa Basalt in the eastern lode system, the Williamstown Dolerite at Hidden Secret, and the Hannan’s Lake Serpentinite at Mt. Percy—provide favorable hosts to mineralization, particularly where intruded by preore, quartz-feldspar-phyric, dacitic porphyries that also host mineralization. These porphyries occur in ore zones throughout the camp, including the Golden Mile Super Pit and the Mt. Charlotte deposit where they are accompanied by volumetrically minor hornblende-phyric andesites and lamprophyre dikes that overlap temporally with the telluride-bearing mineralization. It remains difficult to define a precise geochronological framework for mineralization. From a structural point of view, it is likely that the early gold-telluride Fimiston, Hidden Secret, and Oroya lodes formed during regional D2-D3

272

MCDIVITT ET AL.

sinistral transpression (ca. 2675–2655 Ma). Their maximum age is constrained by ca. 2675 Ma, preore quartz-feldspar porphyries and their minimum age is constrained by the crosscutting Mt. Charlotte stockwork-type veins, with the latter inferred to have formed during the transition to regional D4 dextral shearing, which affected the 2648 ± 6 Ma Liberty granodiorite pluton. Zircon U-Pb dating of the hornblende-phyric andesitic porphyries emplaced synchronously with the gold-telluride lodes has proved problematic, and conclusive crystallization ages remain elusive, in part due to the large errors associated with the reported ages, and also due to the interpretive nature of the calculated ages. Dating of the synore Oroya lamprophyre has proved to be similarly difficult, with the 2642 ± 6 Ma age obtained from U-Pb zircon SHRIMP geochronology first considered to represent the age of lamprophyric magmatism and Oroya mineralization (McNaughton et al., 2005), and now considered to be a minimum age for those events (Mueller, 2020a; Mueller and Muhling, 2020). The age of the Mt. Charlotte gold mineralization event is less ambiguous and it is indicated to be ca. 2650 Ma by U-Pb SHRIMP geochronology of hydrothermal xenotime within Mt. Charlotte ore (Mueller et al., 2020), which agrees with the timeframe established for regional D4 deformation. Gold mineralization in the Kalgoorlie gold camp may be considered orogenic in the sense that it formed contemporaneously with protracted magmatism, metamorphism, and deformation during a ca. 2685 to 2640 Ma period of orogeny. However, as noted by Robert et al. (2005) and further illustrated by Mueller et al. (2020), magmatism, metamorphism, and deformation overlap both spatially and temporally in this orogenic framework making it difficult to fingerprint the specific processes responsible for gold mineralization events. It is clear from structural, geochemical, mineralogical, and textural data that early, Au- and Te-rich, aqueous-carbonic hydrothermal fluids of an oxidized nature were responsible for the bulk of gold deposition in the Golden Mile and Hidden Secret deposits. Later aqueous-carbonic fluids of a Te-poor and reduced nature contributed to this earlier gold endowment by forming stockwork veins at Mt. Charlotte, Drysdale, and Golden Pike. The evidence invokes temporally and geochemically distinct hydrothermal systems that differed in their gold and telluride endowment, and suggests that multiple genetic mechanisms contributed to the formation of ore in the Kalgoorlie gold camp with both magmatic and metamorphic ore-forming processes likely to be represented. Although mineralization is spread along the 8- × 2-km, NNW-trending corridor that defines the camp, the bulk of the estimated 2,300 t Au endowment is hosted in a much smaller area represented by the Golden Mile Super Pit. The majority of gold production from the Golden Mile is sourced from Fimiston-type gold telluride lodes, implying that the later overprinting Mt. Charlotte mineralization event was not a critical contributor to this exceptional endowment. Hence, the highly concentrated gold endowment within the Golden Mile is likely a reflection of numerous optimized factors related to host-rock preparation and ore formation (cf. Groves et al., 2016) that affected the circulation and focusing of Au- and Te-rich hydrothermal fluids during D2-D3 regional sinistral transpression. Additionally, it is likely that the hydrothermal fluids responsible for gold endowment in the Kalgoorlie gold camp ultimately stem

from anomalously enriched source regions, such as the fertile uppermantle proposed by Hronsky et al. (2012). Acknowledgments The authors express sincere gratitude to François Robert, Howard Poulsen, and Scott Halley for their meticulous reviews of this manuscript. Chris Hesford and Gerard Tripp are thanked for their major contributions toward production of the updated camp-scale geology map. Numerous members of the Kalgoorlie Consolidated Gold Mines geology team contributed to the production of this manuscript and their efforts are greatly appreciated. REFERENCES Bateman, R., and Hagemann, S., 2004, Gold mineralization throughout about 45 Ma of Archaean orogenesis: Protracted flux of gold in the Golden Mile, Yilgarn craton, Western Australia: Mineralium Deposita, v. 39, p. 536–559. Bateman, R., and Jones, S. 2015, Discussion: The timing of gold mineralization across the eastern Yilgarn craton using U-Pb phosphate geochronology of hydrothermal minerals: Mineralium Deposita, v. 50, p. 885–888. Bateman, R., Costa, S., Swe, T., and Lambert, D., 2001a, Archean mafic magmatism in the Kalgoorlie area of the Yilgarn craton, Western Australia: A geochemical and Nd isotopic study of the petrogenetic and tectonic evolution of a greenstone belt: Precambrian Research, v. 108, p. 75–112. Bateman, R., Hagemann, S.G., McCuaig, T.C., and Swager, C.P., 2001b, 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. Claoué-Long, J.C., Compston, W., and Cowden, A., 1988, The age of Kambalda greenstone resolved by ion-microprobe: Implications for Archaean dating methods: Earth and Planetary Science Letters, v. 89, p. 239–259. Clark, M.E., 1980, Localization of gold, Mt. Charlotte, Kalgoorlie, Western Australia: B.Sc. (Honors) thesis, Perth, University of Western Australia, 128 p. Clout, J.M.F., 1989, Structural and isotopic studies of the Golden Mile gold-telluride deposit, Kalgoorlie, Western Australia: Ph.D. thesis, Melbourne, Monash University, 352 p. Clout, J.M.F., Cleghorn, J.H., and Eaton, P.C., 1990, Geology of the Kalgoorlie gold field: Australasian Institute of Mining and Metallurgy Monograph 14, v. 1, p. 411–431. Eaton, P.C., 1986, The regional geology of Kalgoorlie: Unpublished report to Kalgoorlie Mining Associates, 66 p. Evans, K.A., Phillips, G.N., and Powell, R., 2006, Rock-buffering of auriferous fluids in altered rocks associated with the Golden Mile-style mineralization, Kalgoorlie gold field, Western Australia: Economic Geology, v. 101, p. 805–817. Fitzgerald, M., and Nixon, D.G., 2016, The exploration and geology of the Hidden Secret Au-Ag lode orebody, Mount Charlotte, Kalgoorlie [ext. abs.]: Australian Institute of Geoscientists Bulletin 62, p. 16–20. Fletcher, I.R., Dunphy, J.M, Cassidy, K.F., and Champion, D.C., 2001, Compilation of SHRIMP U-Pb geochronological data, Yilgarn craton, Western Australia, 2000–2001: Geoscience Australia Record 2001/47, 111 p. Gauthier, L., Hagemann, S., Robert, F., and Pickens, G., 2004a, Structural architecture and relative timing of Fimiston gold mineralization at the Golden Mile deposit, Kalgoorlie: Geological Society of Western Australia Record 2004/16, p. 53–60. ——2004b, New constraints on the architecture and timing of the giant Golden Mile deposit, Kalgoorlie, Western Australia [ext. abs.]: Centre for Global Metallogeny, University of Western Australia Publication 33, p. 353–356. Godefroy-Rodríguez, M., Hagemann, S., LaFlamme, C., and Fiorentini, M., 2020, The multiple sulfur isotope architecture of the Golden Mile and Mount Charlotte deposits, Western Australia: Mineralium Deposita, v. 55, p. 797–822, https://doi.org/10.1007/s00126-018-0828-y Goldfarb, R.J., and Groves, D.I., 2015, Orogenic gold: Common or evolving fluid and metal sources through time: Lithos, v. 233, p. 2–26.



SETTING, Au MINERALIZATION, KALGOORLIE Au CAMP, YILGARN CRATON, WA

Golding, S.D., and Wilson, A.F., 1983, Geochemical and stable isotope studies of the No. 4 lode, Kalgoorlie, Western Australia: Economic Geology, v. 78, p. 438–450. Goscombe, B., Blewett, R.S., Czarnota, K., Groenewald, P.B., and Maas, R., 2009, Metamorphic evolution and integrated terrane analysis of the eastern Yilgarn craton: Rationale, methods, outcomes and interpretation: Geoscience Australia Record 2009/23, 270 p. Groves, D.I., Goldfarb, R.J., and Santosh, M., 2016, The Conjunction of factors that lead to the formation of giant gold provinces and deposits in nonarc settings: Geoscience Frontiers, v. 7, p. 303–314. Heath, C.J., 2003, Fluid flow at the giant Golden Mile deposit, Kalgoorlie, Western Australia: Ph.D. thesis, Canberra, Australian National University, 173 p. Ho, S.E., 1986, A fluid inclusion study of Archean gold deposits of the Yilgarn Block, Western Australia: Ph.D. thesis, Perth, University of Western Australia, 90 p. Hronsky, J.A., Groves, D.I., Loucks, R.R., and Begg, G.C., 2012, A unified model for gold mineralization in accretionary orogens and implications for regional-scale exploration targeting methods: Mineralium Deposita, v. 47, p. 339–358. Kent, A.J.R., and McDougall, I., 1995, 40Ar-39Ar and U-Pb age constraints on the timing of gold mineralization in the Kalgoorlie gold field, Western Australia: Economic Geology v. 90, p.845–859. Krapež, B., and Hand, J.L., 2008, Late Archaean deep-marine volcaniclastic sedimentation in an arc-related basin: The Kalgoorlie Sequence of the Eastern Goldfields superterrane, Yilgarn craton, Western Australia: Precambrian Research, v. 161, p. 89–113. Krapež, B., Brown, S.J.A., Hand, J, Barley, M.E., and Cas, R.A.F., 2000, Age constraints on recycled crustal and supracrustal sources of Archaean metasedimentary sequences, Eastern Goldfields Province, Western Australia: Evidence from SHRIMP zircon dating: Tectonophysics, v. 322, p. 89–133. Lungan, A., 1986, The Structural controls of the Oroya shoot: Implications for the structure of the Kalgoorlie region, Western Australia: B.Sc. (Honors) thesis, Perth, University of Western Australia. McNaughton, N.J., Mueller, A.G., and Groves, D.I., 2005, The age of the giant Golden Mile deposit, Kalgoorlie, Western Australia: Ion-microprobe zircon and monazite U-Pb geochronology of a synmineralization lamprophyre dike: Economic Geology v. 100, p. 1427–1440. Mernagh, T.P., Heinrich, C.A., and Mikucki, E.J., 2004, Temperature gradients recorded by fluid inclusions and hydrothermal alteration at the Mount Charlotte gold deposit, Kalgoorlie, Australia: Canadian Mineralogist, v. 42, p. 1383–1403. Morris, P.A., Pescud, L., Thomas, A., Gamble, J., Tovey, E., Marsh, N., and Everett, R., 1991, Geochemical analyses of Archaean mafic and ultramafic volcanics, Eastern Yilgarn craton, Western Australia: Geological Survey of Western Australia Record 1991/8, 78 p. Mueller, A.G., 2007, Copper-gold endoskarns and high-Mg monzodiorite-­ tonalite intrusions at Mt. Shea, Kalgoorlie, Australia: Implications for the origin of gold-pyrite-tennantite mineralization in the Golden Mile: Mineralium Deposita, v. 24, p. 737–769. ——2015, Structure, alteration, and geochemistry of the Charlotte quartz vein stockwork, Mt. Charlotte gold mine, Kalgoorlie, Australia: Time constraints, down-plunge zonation, and fluid source: Mineralium Deposita, v. 50, p. 221–244. ——2020a, Structural setting of Fimiston- and Oroya-style pyrite-telluride-gold lodes, Paringa South mine, Golden Mile, Kalgoorlie: 1. Shear zone systems, porphyry dikes and deposit-scale alteration zones: Mineralium Deposita, v. 55, p. 665–695, https://doi.org/10.1007/s00126-017-0747-3. ——2020b, Paragonite-chloritoid alteration in the Trafalgar fault and Fimiston-and Oroya-style gold lodes in the Paringa South mine, Golden Mile, Kalgoorlie: 2. Muscovite-pyrite and silica-chlorite-telluride ore deposited by two superimposed hydrothermal systems: Mineralium Deposita, https:// doi.org/10.1007/s00126-018-0813-5. Mueller, A.G., and Muhling, J.R., 2013, Silver-rich telluride mineralization at Mount Charlotte and Au-Ag zonation in the giant Golden Mile deposit, Kalgoorlie, Western Australia: Mineralium Deposita, v. 48, p. 295–311. ——2020, Early pyrite and late telluride mineralization in vanadium-rich gold ore from the Oroya shoot, Paringa South mine, Golden Mile, Kalgoorlie: 3. Ore mineralogy, Pb-Te (Au-Ag) melt inclusions, and stable isotope constraints on fluid sources: Mineralium Deposita, v. 55, p. 733–766, https://doi.org/10.1007/s00126-019-00876-6. Mueller, A.G., Harris, L.B., and Lungan, A., 1988, Structural control of greenstone-hosted gold mineralizaion by transcurrent shearing: A new

273

interpretation of the Kalgoorlie mining district, Western Australia: Ore Geology Reviews, v. 3, p. 359–387. Mueller, A.G., Hagemann, S.G., and McNaughton, N.J., 2020, Neoarchean orogenic, magmatic and hydothermal events in the Kalgoorlie-Kambalda area, Western Australia: Constraints on gold mineralization in the Boulder Lefroy-Golden Mile fault system: Mineralium Deposita, v. 55, p. 633–663, https://doi.org/10.1007/s00126-016-0665-9. Nelson, D.R., 1997, Evolution of the Archaean granite-greenstone terranes of the Eastern Goldfields, Western Australia: SHRIMP U-Pb zircon constraints: Precambrian Research, v. 83, p. 57–81. O’Connor-Parsons, T., and Stanley, C.R. 2007. Downhole lithogeochemical patterns relating to chemostratigraphy and igneous fractionation processes in the Golden Mile dolerite, Western Australia: Geochemistry: Exploration, Environment, Analysis, v. 7, p. 109–127. Phillips, G.N., 1986, Geology and alteration in the Golden Mile, Kalgoorlie: Economic Geology, v. 81, p. 779–808. Phillips, D., and Miller, J.M., 2006, 40Ar/39Ar dating of mica-bearing pyrites from thermally overprinted Archean gold deposits: Geology, v. 34, p. 397–400. 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 Science, v. 24, p. 1643–1651. Phillips, G.N., Groves, D.I., Neall, F.B., Donnelly, T.H., and Lambert, I., 1986, Anomalous sulfur isotope compositions in the Golden Mile, Kalgoorlie: Economic Geology, v. 81, p. 2008–2015. Rasmussen, B., Mueller, A.G., and Fletcher, I.R., 2009, Zirconalite and xenotime U-Pb age constraints on the emplacement of the Golden Mile Dolerite sill and gold mineralization at the Mt. Charlotte mine, Eastern Goldfields Province, Yilgarn craton, Western Australia: Contributions to Mineralogy and Petrology, v. 157, p. 559–572. Redman, B.A., and Keays, R.R., 1985, Archaean basic volcanism in the Eastern Goldfields Province, Yilgarn block, Western Australia: Precambrian Research, v. 30, p. 113–152. Ridley, J., and Mengler, F., 2000, Lithological and structural controls on the form and setting of vein stockwork orebodies at the Mount Charlotte gold deposit, Kalgoorlie: Economic Geology, v. 95, p. 85–98. Robert, F., Poulsen, K.H., Cassidy, K.F., and Hodgson, C.J., 2005, Gold metallogeny of the Superior and Yilgarn cratons: Economic Geology 100th Anniversary Volume, p. 1001–1033. Ross, A.A., Barley, M.E., Brown, S.J.A., McNaughton, N.J., Ridley, J.R., and Fletcher, I.R., 2004, Young porphyries, old zircons: New constraints on the timing of deformation and gold mineralization in the Eastern Goldfields from SHRIMP U-Pb zircon dating at the Kanowna Belle gold mine, Western Australia: Precambrian Research, v. 128, p. 105–142. Said, N., and Kerrich, R., 2009, Geochemistry of coexisting depleted and enriched Paringa Basalts, in the 2.7 Ga Kalgoorlie terrane, Yilgarn craton, Western Australia: Evidence for a heterogeneous mantle plume event: Precambrian Research, v. 174, p. 287–309. Sellman, L., 2016, Hidden Secret: The characteristic mineralogy and geochemistry of a unique Ag-rich Au-Ag-Te Golden Mile deposit, Kalgoorlie, Western Australia: B.Sc. (Honors) thesis, Perth, Curtin University, 144 p. Smail, T., 2016, Porphyry dikes at the Hidden Secret Au-Ag-Te deposit, Kalgoorlie, Western Australia: B.Sc. (Honors) thesis, Perth, Curtin University, 151 p. Sully, D., 2010, The Mt. Percy gold deposit and the role and significance of porphyry intrusions in the gold mineralization process: B.Sc. (Honors) thesis, Perth, University of Western Australia, 19 p. Sund, J.O., Schwabe, M.R., Hamlyn, D.A., and Bonsall, E.M., 1984, Gold mineralization at the north end of the Kalgoorlie field, Mount Percy-Kalgoorlie, Western Australia, in Gold-mining, metallurgy and geology: Australasian Institute of Mining and Metallurgy, Perth and Kalgoorlie Branches Regional Conference, p. 397–404. Swager, C., 1989, Structure of Kalgoorlie greenstone—regional deformation history and implications for the structural setting of the Golden Mile gold deposits: Geological Survey of Western Australia Report 25, p. 59–84. Travis, G.A., Woodall, R., and Bartram, G.D., 1971, The geology of the Kalgoorlie goldfield: Geological Society of Australia Special Publication 3, p. 175–190. Tripp, G.I., 2013, Stratigraphy and structure in the Neoarchaean of the Kalgoorlie district, Australia: Critical controls on greenstone-hosted gold deposits: Ph.D. thesis, Townsville, Queensland, James Cook University, 476 p.

274

MCDIVITT ET AL.

Vielreicher, N.M., Groves, D.I., Snee, L.W., Fletcher, I.R., and McNaughton, N.J., 2010, Broad synchroneity of three gold mineralization styles in the Kalgoorlie gold field: SHRIMP, U-Pb, and 40Ar/39Ar geochronological evidence: Economic Geology, v. 105, p. 187–227. Vielreicher, N.M., Groves, D.I., McNaughton, N., and Fletcher, I., 2015, The timing of gold mineralization across the eastern Yilgarn craton using U-Pb geochronology of hydrothermal phosphate minerals: Mineralium Deposita, v. 50, p. 391–428. Vielreicher, N.M., Groves, D.I., and McNaughton, N.J., 2016, The giant Kalgoorlie gold field revisited: Geoscience Frontiers, v. 7, p. 359–374. Weinberg, R.F., Van der Borgh, P., Bateman, R.J., and Groves, D.I., 2005, Kinematic history of the Boulder-Lefroy shear zone system and controls on associated gold mineralization, Yilgarn craton, Western Australia: Economic Geology, v. 100, p. 1407–1426.

Woodall, R.W., 1965, Structure of the Kalgoorlie goldfield: Commonweath Mining and Metallurgy Congress, 8th, Melbourne, Proceedings, p. 71–79. Woods, B.K., 1997, Petrogenesis and geochronology of felsic porphyry dikes in the Kalgoorlie terrane, Kalgoorlie, Western Australia: B.Sc. (Honors) thesis, Perth, Curtin University, 154 p. Yeats, C.J., McNaughton, N.J., Ruettger, D., Bateman, R., Groves, D.I., Harris, J.L., and Kohler, E., 1999, Evidence for diachronous Archean lode gold mineralization in the Yilgarn craton, Western Australia: A SHRIMP U-Pb study of intrusive rocks: Economic Geology, v. 94, p. 1259–1276.

©2020 Society of Economic Geologists, Inc. SEG Special Publications, no. 23, pp. 275–288

Chapter 13 Boddington: An Enigmatic Giant Archean Gold-Copper (Molybdenum-Silver) Deposit in the Southwest Yilgarn Craton, Western Australia Stephen J. Turner,1,† Graeme Reynolds,2 and Steffen G. Hagemann3 1Chief 2 Superintendent 3Centre

Geologist, Newmont Corporation, Australia, 3 Bellows Street, Welshpool 6106, Australia Geology, Newmont Boddington Gold, P.O. Box 48, Boddington, Western Australia 6930, Australia

for Exploration Targeting, School of Earth Sciences, University of Western Australia, M006, 35 Stirling Highway, Crawley, Western Australia 6009, Australia

Abstract Boddington is a giant, enigmatic, and atypical Archean Au-Cu deposit hosted in a small, remnant greenstone belt within granite-gneiss and migmatite of the Southwest terrane of the Yilgarn craton, Western Australia. Primary Au and Cu (and Mo) mineralization consists of a network of thin fractures and veins, controlled by shear zones, and dominantly hosted by early dioritic intrusions and their immediate wall rocks, which comprise felsic to intermediate- composition volcanic and volcaniclastic rocks. The pre-~2714 Ma host rocks are typically steeply dipping and strongly deformed, with early ductile and overprinting brittle-ductile fabrics, and have been metamorphosed at mid- to upper greenschist facies. Features consistent with porphyry-style mineralization, classic orogenic shear zones, and intrusion-related Au-Cu-Bi mineralization are all recognized, giving rise to a variety of genetic interpretations. It is clear that Boddington does not fit any classic Archean orogenic gold deposit model, having a general lack of quartz veins and iron carbonate alteration, a Cu (Mo and Bi) association, zoned geochemical anomalism, and evidence of high-temperature, saline ore-forming fluids. Detailed petrographic, geochemical, and melt inclusion studies suggest a late-stage ~2612 Ma, monzogranite intrusion as one of the principal sources of the mineralizing fluids. However, there is also local evidence for older, perhaps protore, porphyry-style Cu (±Au) in the dioritic intrusions and patchy, locally high-grade, orogenic-style gold mineralization associated with enclosing shear zones and brittle-style deformation, which was focused on the relatively competent dioritic intrusions. The relative contributions of metals from these components to the system may not be resolvable. It appears that the Boddington deposit has been a locus for multiple episodes of intrusion, alteration, and mineralization over an extended period of time, as has been demonstrated in a number of other large Canadian and Australian gold deposits, including the Golden Mile near Kalgoorlie.

Introduction The Boddington gold mine is located in the Darling Ranges, 130 km south-southeast of Perth in Western Australia at 32.74º S and 116.35° E, at elevations between 250 and 400 m above sea level (asl; Fig. 1). With an endowment of ~30 million ounces (Moz) Au the Boddington mine is a giant gold deposit on a world scale and includes a significant copper endowment (Table 1). Past production includes >6 Moz Au from a series of open pits within laterite and saprolite that capped the primary bedrock Au and Cu deposit, which is named Wandoo. The current primary resource and reserve include 545 t Au (17.5 Moz), 1.1 Mt Cu, plus Ag, which are recovered and significant Mo, which is not recovered. Boddington is an active open-pit mining operation owned 100% by Newmont Corporation. Exploration and Mining History The Saddleback greenstone belt was discovered by the Geological Survey of Western Australia in 1975 during regional mapping (Wilde, 1976). During subsequent geochemical sampling by the Survey in 1977 to 1978, rock, soil, and stream-­sediment samples were collected, which outlined a 3-km-long zone with anomalous Cu, Pb, Zn, As, Au, Mo, and Sn. Exploration †Corresponding

author: e-mail, [email protected]

sampling of surface lateritic hard cap and soils (gravel) by Reynolds Australia Mines discovered potentially economic gold grades in July 1980. An initial resource of 15 Mt @ 2.70 g/t Au was defined simply by reanalysis of existing bauxite vacuum drill hole samples. Laterite- and saprolite-hosted gold was mined at Boddington between 1987 and 2001 by Worsley Alumina Pty. Ltd. and by Hedges Gold Pty. Ltd. (a subsidiary of Alcoa) between 1988 and 1998. A small supergene copper resource was discovered in 1988. The high-grade Jarrah quartz veins were first identified in 1989 with initial mining of several oxide open pits, followed by about 400 m of underground development and mining between 1992 and 1996. The underlying Au (Cu-Mo) Wandoo mineralization was first recognized in the early 1980s, with initial resources defined in late 1993, but was not extensively drill tested until 1995. By 1999 the Wandoo resource was estimated at ~19.57 Moz of Au (+Cu). Oxide mining ceased in 2001. Primary resources continued to expand, with associated feasibility studies, and production from the primary hard rock zone commenced in 2009 by the Boddington Gold Mine Joint Venture (AngloGold and Newmont). Previous Work The Boddington primary mineralization forms a very large Au-Cu deposit with characteristics that have variably been

doi: 10.5382/SP.23.13; 14 p. 275

276

TURNER ET AL.

Fig. 1. Location map and simplified regional geologic setting for the Saddleback greenstone belt and Boddington gold mine in the Southwest terrane of the Yilgarn craton (modified from Wilde et al., 1996).

described as porphyry (Roth, 1992), structurally controlled (Allibone et al., 1998), and intrusion related (McCuaig et al., 2001). Aspects of the Boddington deposit and setting have been covered by a number of publications (Roth, 1992; Allibone et al., 1998; Kalleske, 2010; Crawford, 2011), as well as many internal studies (McCuaig and Behn, 1998; Davies et al., 2010; Augenstein et al., 2013; and others). There have been few recent (268,000 m at surface (740 holes), >182,000 m (1,929 holes) underground, and driven >30 km of underground development to date. Regional Geologic Setting The Brucejack deposit is situated on the western side of the Stikine terrane (Stikinia), an island-arc terrane that forms part of the Intermontane morphogeologic belt of the Canadian Cordillera (Fig. 1; Monger and Price, 2002). This part of Stikinia is bounded to the west by the Coast Plutonic Complex and east by rocks of the Bowser Lake Group (Fig.1). Stikinia is considered to be an allocthonous terrane that was initially developed in the northern parts of a composite island-arc-rifted microcontinent chain that was outboard of ancestral North America during the late Palaeozoic (Belasky et al., 2002; Colpron et al., 2006). A complex and protracted multistage history of tectonism, volcanism, and magmatism is recorded in Stikinia rocks (Nelson and Kyba, 2014). Arc construction during the Late Triassic and again in the Early Jurassic superimposed younger rocks on older arc architecture: Paleozoic arc-related volcanic, plutonic, and associated sedimentary rocks of the Stikine assemblage form the basement to marine arc-related volcanic, volcaniclastic, and arc-derived clastic rocks of the Triassic Stuhini and latest Triassic to Middle Jurassic Hazelton Groups, and their associated comagmatic plutonic rocks (Nelson and Kyba, 2014; Nelson et al., 2018). Angular unconformities to disconformities between and within stratigraphic units indicate episodic deformation and erosion (Greig, 1992; Henderson et al., 1992). Numerous Late Triassic to Early-Middle Jurassic Cu-Au (±Mo) porphyry, volcanogenic massive sulfide (VMS), and associated deposits were emplaced while Stikinia was in a state of compression or sinistral transpression (Nelson and Colpron, 2007). The abundance of these deposits in Stikinia highlights its importance as a rich metallotect comparable to modern-arc settings like the Philippines (Nelson and Colpron, 2007). Accretion of the more southerly parts of the arc chain to western Laurentia occurred during the Permo-Triassic (Nelson and Colpron, 2007). Oroclinal bending of the chain,

292

BOARD ET AL.

centered on the Yukon-Tanana terrane, from the Late Permian through the Middle Jurassic resulted in Stikinia rotating in an anticlockwise direction (Mihalynuk et al., 1994). A tectonic setting similar to the modern Philippine archipelago has been proposed for Stikinia in the Early Jurassic: two opposing island arcs separated by a marine trough with associated development of widespread and voluminous volcanogenic strata (Marsden and Thorkelsen, 1992). Closure of the intervening Cache Creek ocean through progressive oroclinal bending culminated in a Halmahera-style arc-arc collision between Stikinia and the already-accreted Quesnel terrane (Quesnellia) in the Middle Jurassic (~172 Ma; Mihalynuk et al., 2004), resulting in cessation of arc activity (Nelson and Colpron, 2007). Northeasterly sourced and southwesterly younging coarse clastic strata of the Middle Jurassic to Lower Cretaceous Bowser Lake Group unconformably overlie the arc-related sequence, and are considered to have been deposited in a subsiding basin that developed in response to the arc-arc collision (Evenchick et al., 2007). Postaccretion deformation in the mid-Cretaceous (~120– 90 Ma), associated with the accretion of outboard terranes of the Insular belt, variably affected rocks in Stikinia and the mid-Cretaceous Coast Plutonic Complex to the west (Evenchick, 1991; Rusmore and Woodsworth, 1991; Journeay and Friedman, 1993; Nelson and Kyba, 2014). District Geology The Sulphurets mineral district (Fig. 1) contains numerous important structurally controlled porphyry and epithermal deposits along a narrow northerly trend that extends 60 km from Treaty Creek south to near the town of Stewart (Nelson and Kyba, 2014). The deposits, hosted in Hazelton Group rocks and its subvolcanic intrusions, include Kerr-SulphuretsMitchell-Iron Cap, Brucejack, Premier, Big Missouri, Scottie Gold, and Red Mountain. The Kerr-Sulphurets-Mitchell-Iron Cap and Brucejack deposits occur on the eastern side of the McTagg anticlinorium and display a strong spatial association to the unconformity between the Stuhini and Hazelton Group rocks and north-south structures to the east of this contact (Fig. 1), suggesting that these features were important for deposit genesis (Nelson and Kyba, 2014). Stratigraphy The mid-Cretaceous McTagg anticlinorium dominates the district-scale geology in the Sulphurets mineral district (Fig 1). The core of the anticlinorium consists of Upper Triassic arc and marine sedimentary strata of the Stuhini Group, which are the oldest exposed rocks in the district (Fig. A1, Electronic Appendix). The Stuhini Group consists of dark-gray turbiditic siltstone interbedded with minor micritic limestone (Kirkham and Margolis, 1995), subaqueous mafic crystal-lithic lapilli tuff, lapilli tuff-breccia, ash tuff (Brown and Greig, 1990), and pyroxene-phyric breccia flows interlayered with massive andesite and coarse-bladed feldspar porphyry flows (Logan and Koyanagi, 1994). Closer to the Brucejack property, the Stuhini Group is characterized by dark turbiditic sandstone and interbedded mudstone. During the Late Triassic to Early Jurassic the Stuhini Group rocks were deformed, uplifted, and eroded prior to deposition of the unconformably overlying Hazelton Group (Brown

and Greig, 1990). The Jack Formation marks the base of the Hazelton Group (Fig. A1) and consists of cobble- to boulder-bearing, granitoid-clast conglomerate, with arkosic sandstone grading to thinly bedded siltstone and local interbedded intermediate volcaniclastic and pyroclastic rocks (Nelson and Kyba, 2014). Nelson et al. (2018) considered the upper parts of the Jack Formation as having been formed by ~197 to 196 Ma, putting its base at about the Upper Triassic-Lower Jurassic boundary. The Jack Formation locally marks the Stuhini-­Hazelton unconformity on the limbs of the McTagg anticlinorium (Figs. 1, A1). The Betty Creek Formation overlies the Jack Formation paraconformably to the east and is subdivided into the Unuk River andesite and Brucejack Lake units (Nelson et al., 2018; Fig. A1). A sharp contact between the Jack and Betty Creek Formations, which collectively form the lower Hazelton Group (Fig. A1), marks an abrupt change from siliciclastic sedimentation to primarily andesitic pyroclastic and epiclastic accumulation (Nelson et al., 2018). The ~198 to 187 Ma Unuk River andesite is characterized by well-bedded, intermediate-composition epiclastic, pyroclastic, and locally massive volcanic rocks in contrast with the overlying ~185 to 178 Ma Brucejack Lake felsic unit (Lewis et al., 2001), which consists of potassium feldspar-plagioclase-hornblende-phyric flows, breccia, and welded to nonwelded tuffs (Nelson et al., 2018). Lower Hazelton Group rocks occur on both limbs of the anticlinorium, but the unit is significantly thicker on the eastern side (Fig.1). The change in thickness is considered to be due to differences in primary sedimentary basin development on either side of the current McTagg anticlinorium, resulting in it being considered as a structural highland (the McTagg highland) during Hazelton Group deposition (Nelson and Kyba, 2014). Intermediate-composition porphyry intrusions of the Texas Creek suite, which include the KerrSulphurets-Mitchell-Iron Cap-associated Mitchell suite intrusions, are considered as the intrusive equivalents of the Unuk River andesite unit (Nelson et al., 2018). The Betty Creek Formation is overlain by bimodal volcanics and associated sedimentary facies of the ~179 to 173 Ma Iskut River Formation (upper Hazelton Group; Fig. A1; Nelson et al., 2018). The Bajocian to Callovian Quock Formation, partly a lateral facies equivalent of the Iskut River Formation, is the uppermost unit of the Hazelton Group (Fig. A1; Nelson et al., 2018). This unit consists of alternated bands of thinly bedded argillite and felsic tuff laminae and represents the waning stages of Hazelton arc volcanism (Gagnon et al., 2012). Rocks of the upper Hazelton Group occur to the east of the Brucejack project area (Fig. 1) and are considered important marker horizons for regional exploration programs. Synorogenic siliciclastic rocks of the Middle Jurassic to Lower Cretaceous Bowser Lake Group gradationally overlie the Hazelton Group rocks to the north and east of the McTagg anticlinorium (Evenchick et al., 2010). Structure and metamorphism Rocks in the Sulphurets mineral district have been affected by ductile and brittle deformation and regional subgreenschist facies metamorphism (Kirkham and Margolis, 1995). Stuhini Group rocks in the core of the N-plunging McTagg anticlinorium have been deformed into a series of upright folds and associated subvertical, N-striking elongation and



BRUCEJACK Au-Ag DEPOSIT, NW BRITISH COLUMBIA, CANADA

flattening foliations (Kirkham and Margolis, 1995). These structures are truncated by the unconformity between the Stuhini and Hazelton Groups, as well as by tabular Hazelton Group intrusions. This indicates that the Stuhini Group rocks were deformed, uplifted, and eroded prior to deposition of the Hazelton Group (Kirkham and Margolis, 1995). The flanks of the anticlinorium are formed by Hazelton Group rocks overlain by the Bowser Lake Group. Numerous N- and NW-trending, steeply dipping faults are developed in the thick sequence of Hazelton Group rocks on the eastern limb of the McTagg anticlinorium (Fig. 1). The largest and most significant of these is the N-trending Brucejack fault, which strikes for >11 km across the Sulphurets mineral district (Fig. 1). The fault is interpreted as a reactivated growth fault that was active during Early Jurassic volcanism and mineralization, as described below, and remained active as recently as the Eocene (Kirkham and Margolis, 1995). West-directed thrusts and W-vergent overturned folds affect rocks on the western limb of the anticlinorium, whereas rocks on the eastern limb display E- to SE-directed thrusts and E-vergent overturned folds (Kirkham and Margolis, 1995). These thrusts and overturned folds are kinematically linked to the mid-Cretaceous Skeena fold and thrust belt. Eastand SE-vergent thrusts of the Skeena fold and thrust belt are common in the Sulphurets mineral district and affect Stuhini, Hazelton, and Bowser Lake Group rocks (Evenchick, 1991; Kirkham and Margolis, 1995). The Sulphurets and Mitchell thrust faults postdate porphyry-style mineralization and locally juxtapose deeper and shallower parts of magmatic-hydrothermal systems (Kirkham and Margolis, 1995). A penetrative foliation of variable orientation is preferentially developed in altered Hazelton Group rocks in the Sulphurets mineral district, with fabric intensity proportional to phyllosilicate content (Kirkham and Margolis, 1995). Pyrite grains in porphyry-related phyllic alteration zones commonly display pressure shadows, indicating that the foliation postdated the pyrite (Margolis, 1993). Timing of penetrative fabric development is difficult to ascertain due to the absence of unambiguous crosscutting relationships and appropriate nonreset geochronologic data. Although Early Jurassic (~185 Ma) fabric development is mentioned by several investigators (e.g., Roach and Macdonald, 1992; Margolis, 1993; McPherson, 1994), most considered it to have been developed during deformation associated with the mid-Cretaceous Skeena fold and thrust belt (e.g., Alldrick, 1993; Davies et al., 1994; Kirkham and Margolis, 1995; Nelson and Kyba, 2014; Febbo et al., 2015, unpub. data; Febbo, 2016). Regional subgreenschist facies metamorphism (chlorite + prehnite + epidote + albite + calcite) in the Sulphurets mineral district is associated with development of the mid-Cretaceous Skeena fold and thrust belt (Alldrick, 1993). Maximum temperatures and pressures reached ~290°C and 4.5 kbars, respectively, corresponding to thermally reset K-Ar and 40Ar-39Ar ages for foliation-parallel white mica (sericite) in older porphyry-related phyllic alteration zones at ~110 Ma (Alldrick, 1993; Kirkham and Margolis, 1995). Porphyry-style mineralization Four large porphyry-style Cu-Au-Mo deposits are associated with intrusive rocks of the ~196 to 190 Ma Mitchell suite: the Kerr-Sulphurets-Mitchell-Iron Cap deposits, part of the

293

Texas Creek intrusive suite. The deposits are located on the eastern limb of the McTagg anticlinorium, proximal to the Stuhini-Hazelton Group unconformity (Fig.1; Kirkham and Margolis, 1995). Campbell and Dilles (2017) noted that the deposits are broadly contemporaneous and have similar mineralogy, alteration, and textures. Large areas of hydrothermal alteration affected rocks in and around the Mitchell suite intrusions, with overprinting alteration relationships indicating that the magmatic-hydrothermal systems underwent telescoping: (1) early potassic alteration closely associated with porphyry Cu-Au mineralization (~196–192 Ma in the Mitchell deposit; Febbo et al., 2015, unpub. data); (2) locally overprinted propylitic, albitic, and chlorite-sericite alteration; and (3) widespread, well-developed, late quartz-sericite-­ pyrite alteration and associated molybdenite mineralization that pervasively overprinted earlier alteration and extends into the surrounding Stuhini and Hazelton host rocks (~192–190 Ma; Febbo et al., 2015, unpub. data). The later stage phyllic alteration does not appear to have been grade destructive. Late, poorly mineralized massive pyrite veins and associated pervasive advanced argillic alteration are locally developed in the Mitchell deposit (~190 Ma; Febbo et al., 2015, unpub. data), and were overprinted by high-level, gold-rich veins. Deposit Features More than 40 mineral showings are recognized on the Brucejack property, with the Brucejack deposits (sensu lato) currently referring to 10 main zones between the Bridge zone in the south and the Hanging Glacier zone in the north (Fig. 2). Five of these zones have been explored in some detail (West, Valley of the Kings, Bridge, Gossan Hill, and Shore), with another, the Flow Dome zone, the target of current near-mine exploration. This study focuses on the West and Valley of the Kings zones (the Brucejack deposit sensu stricto) as they are the only ones sufficiently drilled to support a mineral reserve estimate and mine plan at the time of writing. Physical characteristics-Valley of the Kings and West zones With a surface extent of 1,000 × 600 m, the Valley of the Kings zone (Fig. 3) occupies the central portion of a rugged, NE-sloping alpine plateau located 1 km southwest of Brucejack Lake. Moderate to strongly gossanous, variably silicified, glacially polished, NW-trending ridges and crags on the tens to hundreds of meters scale punctuate the plateau, surrounded and partially obscured by glacial till. Glacial ice covers the southeastern quadrant of the zone. All mineralized vein generations crop out, as do rare hydrothermal breccia and intermediate to mafic dikes. Crosscutting relationships between veins, dikes, structures, alteration, and lithology observed underground (see below) are visible on the surface. Deep drilling has indicated that the alteration, mineralization, and veining in this zone extend to a depth of at least 1,100 m. Mineralization is open to the east, west, and at depth. The West zone lies approximately 750 m north of the Valley of the Kings and ~250 m southwest of Brucejack Lake. A ~250-m-wide zone of well-developed gossans, strongly silicified host rocks, and Au-Ag-rich quartz stockwork and vein swarms define a NW-trending cliff exposure that extends for >570 m between the mine camp and Brucejack fault (Fig. 3). The zone is generally less obscured by till cover than the

Sulphurets Glacier

1 2

3

8

U D

7 10

6

4 5

9

Brucejack Lake

Hanging Glacier

1km

6257000 N

6258000 N

Knipple Glacier

6259000 N

6260000 N

6261000 N

N

Fig. 2. Geologic map of the Brucejack property, showing locations of main mineralized zones (modified after Tombe et al., 2018). Zones are labeled numerically: 1 = Hanging Glacier, 2 = Golden Marmot, 3 = Bonanza/SG, 4 = Gossan Hill, 5 = Shore, 6 = West, 7 = Valley of the Kings, 8 = Waterloo, 9 = Flow Dome, 10 = Bridge. *U-Pb geochronologic data for P2 presented in Appendix 1 (sample SU-450), other geochronologic data = see text for details.

Heterolithic volcaniclastic conglomerate, and minor andesite flows

Stuhini Group

Sandstones, conglomerate, and minor shales

Jack Formation

Hornblende and/or feldspar phyric latite to trachyandesite flow. (’P1 - Office porphyry’). U-Pb zircon 194.1±0.9 Ma

Hornblende feldspar phyric latite to trachyandesite flow. (’P1-Bridge zone porphyry’). U-Pb zircon 189.2±0.5 Ma

Potassium feldspar, hornblende and plagioclase phyric latite to trachyandestie flows and fragmental rocks (’P2’). U-Pb zircon 186.5±0.7 Ma

Undifferentiated sedimentary rocks including heterolithic coarse pebble to boulder conglomerate and sandstone (’Cong’). Youngest detrital U-Pb zircon 187.5±2.6 Ma. Volcanic derived mudstone, siltstone, or fine-grained sandstone and pebble conglomerate with carbonate concretions (’Vsf’). Youngest detrital U-Pb zircon 195.1±2.8 Ma

Tuff, and reworked tuffaceous ash-lapilli-block intermediate volcanic fragmental (’Andx’) U-Pb zircon 184.3±1.6 Ma

Flow-foliated hornblende and/or feldspar phyric latite to trachyandesite, fragmental rocks and undifferentiated sediments. U-Pb zircon 184.3±1.7 Ma

Betty Creek Formation

Undivided volcanic (principally flows) and subordinate sedimentary rocks

Hazelton Group

Quartz-Illite-Pyrite-Calcite alteration

425000 E

Brucejack Brucejack Fault U D Fault

426000 E

Glacial Ice

429000 E 429000 E

Fault

427000 E 427000 E

428000 E 428000 E

294 BOARD ET AL.



295

BRUCEJACK Au-Ag DEPOSIT, NW BRITISH COLUMBIA, CANADA

N

Brucejack Fault

Brucejack Lake

West Zone A’

U D 6258500 N

Brucejack Camp

Waterloo Zone

6258000 N

A

Valley of the Kings

FDZ

U D

Glacial Ice

6257500 N

426500 E

426000 E

Fault

Bridge Zone

Surface Veining

250 m

Fig. 3. Detailed geologic map of the Brucejack deposit, showing the Valley of the Kings and West zones. Refer to legend in Figure 2 for lithology key. White dashed line marks the approximate location of the south-north section presented in Figure 12d. FDZ = Flow Dome zone.

Valley of the Kings zone and is completely free of glacial ice. Deep drilling by previous operators demonstrated continuity of the West zone to at least 650 m below surface; mineralization is open to the northwest, southeast, and at depth (Roach and Macdonald, 1992; Macdonald et al., 1996). Mine stratigraphy The key lithologic sequence at the Brucejack mine is characterized by a basal marine volcano-sedimentary

(Volcanosediment) package unconformably overlain by an immature polylithic volcanic conglomerate (Conglomerate) that grades upward through a sandy epiclastic unit (Transition) into a predominantly pyroclastic trachyandesite (latite) fragmental unit (Andesite) (Figs. 4, 5). The mine stratigraphic unit terms Volcanosediment, Conglomerate, Transition, and Andesite are used here to provide additional resolution to a relatively complex volcanic stratigraphy characterized by rapid lateral facies changes. The mine sequence shows a general

Elevation m a.s.l.

1000m

1100m

1200m

1300m

1400m

6258100N

6258000N

6257900N

6257800N

Volcano-sedimentary Facies (’Vsf’)

Epithermal veining and stockwork Foliaon Silicified conglomerate (’Silcap’)

100m

S

ion ect

426700E

N

at 2

95

deg

th

426600E

imu Az

one st z We

426500E

Fig. 4. Representative three-dimensional section through the Brucejack deposit, showing key geologic, structural, and mineralization relationships developed in the Valley of the Kings and West zones.

Monzonic Intrusive and late flows Office zone Porphyry (P1)

Monzonic Intrusive and late flows Bridge zone Porphyry (P1)

Pyroclasc trachyandesite fragmental (’Andx’) Transional sandy epiclasc (’Trans’) Polylithic conglomerate (’Cong’)

Thrust Fault

ult

a tF

P1 rus

h yT

in

Ra

Alteraon / Structure Normal Fault

Volcanosedimentary

Section along 205 deg Azimuth

6258200N

Intermediate to Mafic dike

Intrusive

Valley Of The Kings zone

Vsf

Trans Cong

Andx

6258300N

Mafic dike

P1

hru

KT VO

a

F st

ult

6258400N

Brucejack Fault

6258500N

D

U

296 BOARD ET AL.

426400E



BRUCEJACK Au-Ag DEPOSIT, NW BRITISH COLUMBIA, CANADA

297

Fig. 5. Representative surface exposure and drill core photos of the local mine stratigraphy, which hosts the Brucejack deposit. Stratigraphic layers are arranged from youngest at the top. (a). Andesite unit: pyroclastic trachyandesite (latite) fragmental unit. (b). Transition unit: sandy epiclastic transitional between the Andesite and Conglomerate units. (c). Conglomerate unit: immature polylithic volcanic conglomerate. (d). Volcano-sediment unit: basal marine volcano-sedimentary clastic unit, showing different facies variations from coarse clast-supported conglomerate to fine mudstone. (e). P1 porphyry: plagioclase ± potassium feldspar ± hornblende phyric porphyry rocks.

younging direction upward and to the east, and is interpreted to have been deposited in a series of small fault-bounded half grabens on the eastern side of the Brucejack fault (Fig. 4). The mine sequence is bounded to the south and northwest by massive and relatively fine grained plagioclase ± potassium feldspar ± hornblende-phyric rocks (P1 porphyry) of the Bridge zone and Office porphyries (Figs. 5, 6e). The Office P1 and Bridge zone P1 porphyry bodies display sharp contacts with the volcaniclastic rocks and have been variably interpreted as comagmatic subvolcanic/hypabyssal monzonitic intrusions or latite flows (Kirkham and Margolis, 1995; Pretium, 2013b). Coarser grained feldspar-hornblende-phyric porphyry rocks (P2 porphyry) are locally present within the

mine sequence, especially north and east of the West zone (Figs. 2, 3). The Volcanosediment unit consists of intercalated immature volcaniclastic conglomerate, gritstone, sandstone, siltstone, and argillite, showing considerable lateral and vertical variations (Fig 5d). The unit is 25 to 240 m thick in the Brucejack deposit. Conglomerate horizons in the Volcanosediment unit are largely clast supported, well sorted, and monomict to oligomict, with pebble- to boulder-sized clasts set in a medium-grained silt matrix. Clasts are most commonly cobble sized, altered, and deformed into oblate shapes parallel to the dominant foliation. Locally, the conglomerate units can be matrix supported, with clasts making up 1,000 m downdip. Dikes in the West zone trend northwest-southeast, have strike lengths of at least 500 m, and extend for >450 m downdip. Banded chilled margins are commonly developed (Figs. 6d, A2) and small (400-m-thick Upper Imnyakh Subformation (PR3im2) comprises pale-gray to -green limestone with rare shale, siltstone, and sandstone beds. Structural geology The recumbent Sukhoi Log anticline is the principal host for gold mineralization. It extends for approximately 50 km in a west to west-northwest direction (Fig. 2), with its axial plane dipping 20° to 25° north and the fold hinge plunging west at 2° to 9°, with minor 3° to 5° undulations (Fig. 6). The upper limb is consistent along the entire strike length, dipping 15° to 17° north (Fig. 6C, D) and flattening to 10° in the east. The lower limb dips 30° to 35° north in the western part and 45° to 50° in the east. The anticline is linked to the N-dipping Kadali-Sukhoi Log shear zone, which occurs some 700 to 800 m south of the deposit (Fig. 6B). The shear zone dips north at 75° to 85° near surface and, according to the geophysical surveys, it may flatten at depth. The Sukhoi Log anticline formed during north-northeastsouth-southwest compression. A blind thrust fault in the overturned, lower limb of the anticline displaces the stratigraphy for at least 100 to 200 m. Southern and northern thrust zones were identified at the contact between the Imnyakh and Khomolkho Formations (Fig. 6B; Martynenko et al., 2009) and are subparallel to the limbs of the Sukhoi Log anticline. The fracture zones along these thrusts host the Central and Northern quartz vein zones (Fig. 6C, D). Parasitic folds whose axial planes parallel that of the Sukhoi Log anticline vary in amplitude from several millimeters to

531

a few meters. Asymmetric drag folds dominate (Fig. A3A-C) on the limbs and symmetric folds, with significant pressure-­ solution enhancement developed in the core of the anticline (Fig. A3D-F). Within Units 4 and 5, there are synsedimentary slump folds (Fig. A3G, H). Minor folds are best developed in Unit 2 and in the core of the anticline in thin interbeds of Units 1, 2, and 3 in the Upper Khomolkho Member. Ductile deformation prevails over brittle. In previous studies, three types of cleavage (axial, interbedding, and bedding) were described. All types of cleavage are related to compressional folding and are axial planar to the fold. They represent variations in orientation related to hinge versus limb positions on the fold and to refraction across the different lithologies on the fold limbs, where the dip of cleavage is oblique to bedding. Microfolds, axial cleavage, and reverse faults are most intense in the most deformed axial part of the fold along its hinge zone, forming a higher strain zone that is the principal ore-controlling structure in the deposit (Fig. 6). Locally, quartz-pyrite mineralization is controlled by the closures of microfolds and cleavage planes. There is a direct correlation between intensity of deformation and gold mineralization. The plan view configuration of the Kadali-Sukhoi Log shear zone and Sukhoi Log anticline can be clearly seen on magnetic and electrical maps (Fig. 7; Galanov et al., 1974; Suprunenko et al., 1975). Due to contrasting lithologies, abundance of sulfides, carbonaceous matter, and fold structure, electrical (self-potential, gradient induced polarization (IP), pole-dipole IP) and magnetic surveys were the most effective for outlining the structural architecture of the deposit (Fig. 7A-C). The IP survey records well the presence of sulfides. In addition, the lithologic contrast between limestones of the Imnyakh Formation and highly carbonaceous, sulfide-rich shales of the Khomolkho Formation can be seen in contrasted electrical properties. In cross sections (Fig. 7D, E), the IP survey detected zones of veinlet-­ disseminated mineralization as areas of moderate resistivity (500–1,000 ohm-m) that correlate with areas of high chargeability (up to 50%), as confirmed by drilling. In highly carbonaceous rocks (in the east of the area), sulfides can be recognized only in combination with the gravity survey data if density variations are >0.15 g/cm3. The Sukhoi Log anticline is cut by a set of WNW-striking, N-dipping (20°–35°) reverse faults and by NW-striking, NE-dipping (35°–65°) oblique faults (Fig. 6B), which can be interpreted as either dextral strike-slip or rotational normal faults. Both fault sets can be identified in the magnetic data (Fig. 7A), and 2017 to 2019 drilling identified several tens of meters of displacement along these fault sets. Quartz veins in the core of the anticline within the Khomolkho Formation are parallel with these two fault sets, which broadly coincide with the regional northwest lineament described above. Morphology of gold mineralization The four mineralized zones in the Sukhoi Log deposit are Main, Central vein, Northwestern, and Zapadnoe (Fig. 8A). They occur along the axial plane of the Sukhoi Log anticline, which dips 20° to 25° north. Mineralization has been traced for >5 km along strike and 1.5 km downdip, from +1000 mRL to +0 mRL, and remains open at depth.

532

VURSIY ET AL.

Fig. 6. Geology of the Sukhoi Log gold deposit. A. Satellite image. B. Geologic map. C. Cross section along exploration line –45, as shown in the map. D. Cross section along exploration line +5. COG = cut-off grade. Courtesy of Polyus Gold.



SUKHOI LOG Au DEPOSIT, RUSSIA

Fig. 6. (Cont.)

533

534

VURSIY ET AL.

Fig. 7. A. Gold orebodies (>0.6 g/t Au) plotted in the total magnetic intensity map. Quartz-sulfide mineralization corresponds to a gradient zone between magnetic high and low. At the same time, magnetic highs correspond to pyrrhotite-rich rocks of the Imnyakh Formation, controlled by main WNW-trending thrust faults and crosscut by minor NW-trending faults. B. Apparent resistivity based on pole-dipole data maps position of the >0.6 g/t Au disseminated orebody at Sukhoi Log (compiled by Polyus Gold based on Suprunenko et al., 1975). C. Apparent resistivity based on gradient survey data reveals best correlation with highly resistive rocks of the Imnyakh Formation and low-resistivity rocks of the Khomolkho Formation (compiled by Polyus Gold based on Suprunenko et al., 1975). D. Pole-dipole chargeability, and E. resistivity sections along Exploration Line +5. See Figure 6B for position of the exploration line. All images are courtesy of Polyus Gold.



SUKHOI LOG Au DEPOSIT, RUSSIA

Fig. 7. (Cont.)

535

536

VURSIY ET AL.

Fig. 8. A. Three-dimensional model of the four mineralized domains at Sukhoi Log. B. and C. Grade distribution in cross sections along exploration lines –45 (B) and +5 (C) (looking west). Courtesy of Polyus Gold.



SUKHOI LOG Au DEPOSIT, RUSSIA

537

Fig. 8. (Cont.)

In the central part of the deposit, the Main ore zone forms a simple 100- to 150-m-thick slab, broadly coincident with the axial plane of the fold (Fig. 8B), but its morphology becomes more complex near the surface where the orientation of the cleavage is more variable. Toward its eastern and western ends, the gold grade of the Main ore zone decreases and the orebody splits into isolated low-grade lodes. However, the volume of pyrite-quartz mineralization does not decrease as much. The internal structure of the ore zone and its shape are lithologically controlled. This control is illustrated by the tongues of quartz-pyrite mineralization in Units 1 and 2 of the Upper Khomolkho Member, extending along the limbs of the fold (Fig. 6D), and by the crescent or fold shape of the highest grade mineralization (Fig. 8B). The ore zone is thickest (up to 150–170 m) and contains its highest concentration of quartz-pyrite mineralization (>5 %) in the Central part of the deposit where the axial plane of the Sukhoi Log anticline intersects Units 1, 2, and 3 of the Upper Khomolkho Member, these units being the most favorable hosts for mineralization. The Western (or Zapadnoe) and blind Northwestern zones form isolated lodes. The Western lode also occurs on the axial plane of the Sukhoi Log anticline, within Unit 5 of the Upper Khomolkho Member and the stratigraphically overlying Imnyakh Formation. The Northwestern lode also occurs within the axial plane of the Sukhoi Log anticline, some 700 to 900 m downdip from the Western lode. It is mostly located in the

highly carbonaceous rocks of Units 1 to 3 of the Upper Khomolkho Member and does not crop out (Fig. 8C). The extent and shape of the lodes is readily detectable in IP data (Fig. 7D, E). The pole-dipole survey revealed a marked contrast along the boundary between the highly resistive Imnyakh Formation and the low-resistivity Upper Khomolkho Member. The N-dipping mineralized zone corresponds to zones of high chargeability and low resistivity. Mineralogical characteristics Mineralized zones at Sukhoi Log include two texturally distinct styles: (1) veinlet and disseminated pyrite-quartz and pyrite mineralization, which hosts 99% of the gold resource within the sheared rocks along the axial plane of the Sukhoi Log anticline; (2) quartz veins which host 1% of the gold resource as native gold, focused along the boundary between the Khomolkho and Imnyakh Formations within the overturned limb of the Sukhoi Log anticline. The two styles of mineralized rocks are the products of a complex paragenesis that spans the diagenesis, metamorphism, and deformation of the host rocks (Buryak, 1982; Large et al., 2007; Yudovskaya et al., 2016). Neither style of mineralized rock is the product of a single event. Pyrite: Pyrite constitutes 98 to 99% of sulfides in the mineralized rocks, where it is accompanied by trace amounts of pyrrhotite, arsenopyrite, chalcopyrite, sphalerite, and galena.

538

VURSIY ET AL.

Although the sulfide mineralogy of the mineralized rocks at Sukhoi Log is monotonous, its morphology is diverse (Fig. 9). Pyrite is the principal host for gold at Sukhoi Log. Detailed studies of pyrite morphology and geochemistry (Buryak, 1982; Large et al., 2007, Yudovskaya et al., 2016) revealed six generations, ranging from early, fine-grained, synsedimentary framboidal to late postdeformational varieties (Fig. A4). The early pyrite commonly forms the cores of grains, overgrown by late crystals. Historically, geologists distinguished quartz-pyrite and pyrite-only varieties. However, it is possible to recognize similar generations of pyrite in both quartz-­pyrite veinlets and in disseminated ore (Large et al., 2007). Gold was first introduced with the synsedimentary pyrite. However, most

gold occurs in free form in the late diagenetic to early metamorphic pyrite (Large et al., 2007). More rarely, native gold occurs in quartz and quartz-pyrite veinlets. Disseminated pyrite mineralization (Fig. 9) typically occurs outside the Main ore zone, within the limbs of the anticline, mainly in Units 4 and 5 of the Upper Khomolkho Member. It forms very fine, but abundant disseminated, idiomorphic porphyroblasts (Fig. 9E, F). Less abundant pyrite of this type occurs within the gold zone, where it is often overprinted by later pyrite generations. Pyrite-quartz mineralization (Fig. 10) dominates the Main ore zone in the core of the Sukhoi Log anticline, reaching 1.5 to 3 vol %, locally 3 to 5 vol %. Large et al. (2007)

Fig. 9. Morphology of pyrite mineralization in the Upper Khomolkho Member (drill hole numbers and corresponding gold grade in brackets). A.-B. Disseminated pyrite in shales of Unit 2 (SL1078; A = 2.43 ppm; B = Au 1.41 ppm). C.-D. Pyrite metacrystals. C. Pyrite with quartz rim in shales of Unit 5 (SL1112). D. Pyrite in sandstone of Unit 1 (SL1112; Au = 0.47 ppm). E.-F. Densely disseminated banded pyrite. E. Disseminated pyrite dust in shales of Unit 5 (SL1005; Au = 0.56 ppm). F. Densely disseminated pyrite in shales of Unit 4 (SL1006; Au = 0.28 ppm). G.-F. Mineralized sandstone boudins in shale. G. In shales of Unit 4 (SL3032; Au 800 m. The strike length of individual veins does not exceed a few tens of meters and the average thickness is 1 to 1.2 m. Their resources were originally evaluated separately from the rest of the deposit (Bobrov et al., 1990). Quartz veins in the core of the Sukhoi Log anticline occur mostly at appreciable depths in the Middle Khomolkho Member (Fig. 6C, D) where the veins and vein zones are 3 to 5 m thick in contrast to only 20 to 30 cm thick at surface. The veins and vein zones have numerous apophyses, pinches, and swells, contain xenoliths of host rock, and have brecciated selvages. The downdip extents of such zones are >500 to 600 m. Native gold: Native gold is sporadically present in all types of quartz veins, more commonly in the layer-parallel to oblique veins of the Central vein zone along the contact between the Imnyakh and Khomolkho Formations and in the Middle Khomolkho Member in the core of the Sukhoi Log anticline. Native gold in quartz veins is present in their most fractured and vuggy parts, as well as in the parts with polymetallic sulfides. Gold particles commonly occur in the selvages of veins, in thin veinlets and lenses of quartz in quartz vein zones, along the contacts between different minerals, within sulfide aggregates, locally within the crystals, around xenoliths, and in cavities. The size of individual gold grains varies from 0.01 to 1 to 2 mm. Genesis The fundamental question concerning the origin of the Sukhoi Log gold deposit is whether the gold was mobilized from early auriferous synsedimentary pyrite (with as much as 350 ppm Au) during diagenesis, folding, and metamorphism (Buryak, 1982; Large et al., 2007), or if the Khomolkho black shales acted as a chemical trap for later, metamorphic- or intrusion-generated fluids associated with proximal granitoids (Laverov et al., 2007). A combination of the two processes may also have occurred given the complex paragenesis (Fig. A4). Although the sequence of multiple mineralization events seems to be well understood (Large et al., 2007), they must be carefully dated as significant time gaps were previously proposed or recognized between them (Rundqvist et al., 1992; Laverov et al., 2000).



SUKHOI LOG Au DEPOSIT, RUSSIA

The first key question concerns the age of the host rocks and, therefore, the age of early syngenetic sulfides. Recent studies of detrital zircon constrained the age of the host rocks to the Neoproterozoic and Late Cambrian (Palenova et al., 2019). The second key question concerns the timing of metamorphism and postmetamorphic events, as well as their correlation with the timing of mineralization. Rundqvist et al. (1992) and Laverov et al. (2000) reported a 516 ± 22 Ma age for the main phase of metamorphism, e.g., soon after deposition of the host rocks. This metamorphic event was confirmed with a U-Pb age of 516 ± 10 Ma for metamorphic monazite (Meffre et al., 2008), probably marking the obduction of the OlokitMuya ophiolites onto the Bodaibo passive margin (de Boisgrollier et al., 2009). This age invalidates the 421 Ma age of metamorphism proposed by Zorin et al. (2008) unless, of course, there were several metamorphic events. Field relationships at Sukhoi Log clearly indicate that the bulk of the gold mineralization predates formation of the orogenic quartz veins, which are related to the main shortening event. Laverov et al. (2007) reported 447 ± 5 Ma Rb-Sr ages for bulk ore samples, which they interpreted as the time of mineralization in relationship to a hidden intrusive body (Yudovskaya et al., 2011). Based on geophysical data, such a body, unexposed at surface in the deposit area, was interpreted to occur at a ~3-km depth (Lishnevsky and Distler, 2004). Rb-Sr dating of bulk ore-stage quartz yielded a middle Paleozoic (320 ± 16 Ma) age (Laverov et al., 2000), which partly overlaps with those of 354 to 320 Ma for the nearby Konkuder-Mamakan granites (Large et al., 2007), including a 325 Ma U-Pb age for the Konstantinovsky stock, 6 km southwest of Sukhoi Log (Fig. 2; Ivanov et al., 1995). Yudovskaya et al. (2011) excluded this complex as a major source of ore-forming fluids due to its association with the postmetamorphic Angara-Vitim batholith. Recent geochronological studies have focused on the correlation of isotopic ages for individual minerals. These studies also fail to provide a conclusive answer to the genetic question because of the multiple nature of events that could have reset the original element ratios. For example, Yudovskaya et al. (2011) reported SIMS U-Pb results for two generations of monazite. Five metamorphic monazite grains yielded a 637 ± 19 Ma age, interpreted to represent the metamorphism peak. One monazite domain yielded a 728 ± 26 Ma age, interpreted as inherited detrital monazite. One grain of hydrothermal monazite showed a 437 ± 15 Ma age. Ages for zircons with overgrowth rims scatter between 670 and 450 Ma, with two peaks at around 650 and 580 Ma, interpreted to reflect hydrothermal zircon growth that disturbed the U-Th-Pb isotope system during the early or main metamorphism (Yudovskaya et al., 2011). Two U-Pb discordia ages have lower intercepts with concordia at 466 ± 29 and 447 ± 32 Ma. The youngest zircon in the dataset has a 448 ± 19 Ma U-Pb age, very close to the youngest age for the hydrothermal monazite and above-mentioned Rb-Sr age for bulk ore samples (Laverov et al., 2007). The U-Pb age estimates for synore monazite cover the interval of 486 ± 18 to 439 ± 17 Ma (Yudovskaya et al., 2011). The discordant isotopic ages indicate that the U-Th-Pb isotope system of ancient detrital zircon was disturbed at 470 to 440 Ma.

541

This agrees with the isotopic age for the younger monazite and the RbSr whole-rock isochron age of black shales (447 ± 6 Ma). Yakubchuk et al. (2014) conducted a pilot Re-Os study of pyrite from the Zapadnoe orebody at Sukhoi Log, which revealed that the metamorphogenic sulfides formed between 508 and 470 Ma, again closer to the interpreted timing of metamorphism, provided that the isotopic system was not reset. The direct studies of sulfides from the Sukhoi Log orebodies by Chernyshev et al. (2009) revealed model Pb-Pb ages of 455 and 130 Ma, indicating a crustal source for Pb. The available isotopic data clearly show the complexity of the Sukhoi Log system, but in general confirm multiple events, with some sulfides (and gold) formed during diagenesis, some during folding and thrusting, and some significantly after regional metamorphism, as supported by the timing relationships documented in this paper. These broadly confirm the Buryak’s (1982) concept of multistage mineralization at Sukhoi Log, with the main events taking place in the early and middle Paleozoic. This does not exclude input of “external” gold to the system, especially during the quartz vein stage(s), although Buryak (1982) considered them as formed at the expense of disseminated gold mineralization during late phases of metamorphism. Buryak and Khmelevskaya (1997) believed that regionally developed black shales and a local site of seafloor exhalative activity were controlled by a syndepositional normal fault, which was later transformed into the Kadali-Sukhoi Log thrust zone. Large et al. (2007) estimated that 15 km3 of black shales grading 30 ppb Au would be necessary to form the giant gold resource at Sukhoi Log. Furthermore, the appearance of black shale strata at several levels in the Patom passive margin sequence (Fig. 4B), both in the Bodaibo synclinorium and adjacent structural units, is viewed as a favorable factor for the discovery of additional bedrock gold deposits. These strata are interpreted to represent anoxic events, some of which may correspond to Neoproterozoic glaciations. Further studies are necessary to estimate the proportion of “internal” and “external” gold in the deposit. The southwest-vergent fold-and-thrust deformation that controls the mineralization is traditionally linked to the collision of a magmatic arc and turbidite terranes with the Patom passive margin. This collision also involved fragments of Siberian continental crust, which were pushed up within the passive margin (Chuya-Nechera uplift) and also incorporated into external terranes (e.g., Muya terrane). The structural pattern (Fig. 4A) suggests that the Muya terrane, located south of the Olokit-Muya suture, acted as an indenter that was pushed northward toward the Patom passive margin. Although it was pushed together with other terranes, it was the most rigid unit in the collage and may have defined the position of both the WNW-trending imbricate thrusts in the entire Bodaibo synclinorium and oblique north-northwest structural trends (Fig. 2), crosscutting the folds, thrusts, and mineralization, and controlling areas with higher grade ore shoots at Sukhoi Log. Conclusions The Neoproterozoic Patom passive margin sequence, with black shale horizons at several stratigraphic levels—many known to host gold deposits—can be considered as an

542

VURSIY ET AL.

especially favorable factor for development of giant orogenic gold accumulations, in contrast to the typically smaller orogenic gold deposits present in accretionary wedges elsewhere in the world. Over the years, the understanding of the Sukhoi Log deposit evolved from its being a small quartz vein occurrence to the large disseminated gold resource in Russia. The deposit was originally perceived as a series of quartz veins related to a distal granitic intrusion. Subsequently, it was reinterpreted as metamorphogenic in origin, with a significant contribution of gold present in synsedimentary pyrite that was remobilized during the metamorphism to produce a huge volume of auriferous disseminated sulfides with only minor syntectonic quartz veins. Recent isotopic studies showed that there may have been more than one synmetamorphic gold-mineralizing event, assuming that the isotopic systems were not reset. One of them might have been contemporaneous with regional granitoid magmatism. Sukhoi Log can therefore be considered as a multistage gold deposit. A significant role is assigned to the black shales, which may have acted either as a synsedimentary and/or synmetamorphic geochemical trap for gold transported by transgressive fluids. Structurally, the mineralization is controlled by the recumbent Sukhoi Log anticline, conjugate with the Kadali-Sukhoi Log shear zone, thrust southward along the major boundary between two regional tectonic features: the Chuya-Nechera anticlinorium and Bodaibo synclinorium. Acknowledgments We thank Polyus Gold for permission to publish data from the ongoing prefeasibility study. We are grateful to Nikolai Goryachev and his team for statistical data on Bodaibo placer gold production and comments on genesis of the Sukhoi Log deposit. Alexander Borisov (Polyus Gold) and Anna Zelenskaya (VIMS) are acknowledged for their help during preparation of the manuscript. We thank Marina Yudovskaya for sharing her views and for help with the most recent published data. We are grateful to Andrew Allibone and François Robert for their detailed and constructive comments and corrections. REFERENCES AMC Consultants Pty Ltd [AMC], 2018, Sukhoi Log resource definition and data collection: Independent Reviewer Report, 17 December 2018 (http://www.polyus.com/en/media/press-releases/sukhoi-log-mineralresources-update/). Anonymous, 2016, The Sukhoi Log saga or the record of non-development of Sukhoi Log: Vestnik Zolotopromyshlennika, Prime, 84 p. (in Russian). Benevolskiy, B.I., 2002, Gold of Russia: Moscow, Geoinformmark, 464 p. (in Russian). Bobrov, B.A., Bobrova, T.A., Glazkov, V.P., 1990, Report on exploration results for hardrock gold in quartz vein zones at the western and southern flanks of the Sukhoi Log deposit in 1987-1990: Irkutsk, Russian Federal Geological Fund (in Russian). Budyak, A.E., Parshin, A., Spiridonov, A.M., Volkova, M.G., Tarasova, Y.I., Bryukhanova, N.N., Zarubina, O.V., Reutsky, V.N., Damdinov, B.B., and Abramova, V.A., 2017, Geochemical controls on the formation of unconformity-type Au-U deposits (Northern Transbaikalia): Geochemistry International, v. 55, p. 184–194, doi.org/10.1134/S0016702917010049. Bukharov, A.A., Khalilov, V.A., Strakhova, T.M., and Chernikov, V.V., 1992, Geology of the Baikal-Patom Highlands from new data on U-Pb dating of accessory zircon: Geologia i Geofizika, v. 33 (12), p. 29–39 (in Russian). Buryak, V.A., 1959, Preliminary results of studying the gold-sulphide mineralization in the Lena gold district (exemplified by Golets Vysochaishyi):

Transactions on geology and mineral deposits of eastern Siberia, Irkutsk, No. 5 (26) (in Russian). ——1982, Metamorphism and ore formation: Moscow, Nauka, 256 p. (in Russian). Buryak, V.A., and Khmelevskaya, N.M., 1997, Sukhoi Log: One of the world’s largest gold deposit: Genesis, localization of ore, and forecasting criteria: Vladivostok, Dal’nauka, 156 p. (in Russian). Chernyshev, I.V., Chugaev, A.V., Safonov, Yu.G., Saroyan, M.P., Yudovskaya, M.A., and Eremina, A.V., 2009, Lead isotopic composition from data of high-precision MC-ICP-MS and sources of matter in the large-scale Sukhoi Log noble metal deposit, Russia: Geology of Ore Deposits, v. 51, p. 496–504. de Boisgrollier, T., Petit, C., Fournier, M., Leturmy, P., J.-C. Ringenbach, J.-C., San’kov, V.A., Anisimova, S.A., and Kovalenko, S.N., 2009, Palaeozoic orogeneses around the Siberian craton: Structure and evolution of the Patom belt and foredeep: Tectonics, v. 28 (1), TC1005. Doi: 10.1029/2007TC00221 . Distler, V.V., Mitrofanov, G.L., Nemerov, V.K., Kovalenker, V.A., Mokhov, A.V., Semeikina, L.K., and Yudovskaya, M.A., 1996, Modes of occurrence of the platinum group elements and their origin in the Sukhoi Log gold deposit (Russia): Geology of Ore Deposits, v. 38, p. 413–428. Distler, V.V., Yudovskaya, M.A., Mitrofanov, G.L., Prokof’ev, V.V., and Lishnevsky, E.N., 2004, Geology, composition, and genesis of the Sukhoi Log noble metals deposit, Russia: Ore Geology Reviews, v. 24, p. 7–44. Ernst, R.E., Hamilton, M.A., Soderlund, U., Hanes, J.A., Gladkochub, D.P., Okrugin, A.V., Kolotilina, T., Mekhonoshin, A.S., Bleeker, W., LeCheminant, A.N., Buchan, K.L., Chamberlain, K.R., and Didenko, A.N., 2016, Long-lived connection between southern Siberia and northern Laurentia in the Proterozoic: Nature Geoscience, v. 9, p. 464–469. Galanov, S.I., Karavaev, Yu.A., Radchenko, K.M., 1974, Geophysical criteria for prognosis and exploration of quartz-sulphide gold in the Lena district at a scale of 1:50 000 to 1:10 000: Irkutsk, VostSibNIIGGiMS. Russian Federal Geological Fund (in Russian). Golubev, V.N., Makar’ev, L.B., and Bylinskaya, L.V., 2008, Deposition and remobilization of uranium in the North Baikal region: Evidence from the U-Pb isotopic systems of uranium ores: Geology of Ore Deposits, v. 50, p. 482–490. Ivanov, A.I., 2017, Gold of Baikal-Patom area (geology, mineralisation, perspectives): Moscow, TSNIGRI, 215 p. (in Russian). Ivanov, A.I., Lifshits, V.I., Perevalov, T.M., Strakhova, T.M., Yablonovsky, B.V., Il’yinskaya, M.I., Graizer, H.G., and Golovenok, V.K., 1995, Precambrian of the Patom Highlands: Moscow, Nedra, 262 p. (in Russian). Karpenko, I.A., Kulikov, D.A., Cheremisin, A.A., and Avdeev B.V., 2007, Report on reserve estimate at the Sukhoi Log deposit as of 01.06.2007: Moscow, TsNIGRI (in Russian). Kazakevich, Yu.P., and Storozhenko, A.A., 1957, 1:200,000 State Geological Map of the USSR, Sheet O-50-XIV: Moscow, TsNIGRI. Large, R.R., Maslennikov, V.V., Robert, F., Danyushevsky, L., and Chang, Z., 2007, Multistage sedimentary and metamorphic origin of pyrite and gold in the giant Sukhoi Log deposit, Lena gold province, Russia: Economic Geology, v. 102, p. 1233–1267. Larin, A.M., Sal’nikova, E.B., Yakovleva, S.Z., Kovach, V.P., Kotov, A.B., and Makar’ev, L.B., 2006, Early Proterozoic syn- and postcollision granites in the northern part of the Baikal fold area: Stratigraphy and Geological Correlation, v. 14, p. 463–474. Laverov, N.P., Prokof’ev, V.Yu., Distler, V.V., Bairova, E.D., Gozman, Yu.V., Golubev, V.N., Chugaev, A.V., and Yudovskaya, M.A., 2000, New data on conditions of ore deposition and composition of ore-forming fluids in the Sukhoi Log gold-platinum deposit: Doklady Earth Sciences, v. 371, p. 357–361. Laverov, N.P., Chernyshev, I.V., Chugaev, A.V., Bairova, E.D., Golzman, Yu.V., Distler, V.V., and Yudovskaya, M.A., 2007, Formation stages of the large-scale noble metal mineralization in the Sukhoi Log deposit, East Siberia: Results of isotope-geochronological study: Doklady Earth Sciences, v. 415, p. 810–814. Lishnevsky, E.N., and Distler, V.V., 2004, Deep structure of the earth’s crust in the district of the Sukhoi Log gold-platinum deposit (eastern Siberia, Russia) based on geological and geophysical data: Geology of Ore Deposits, v. 46, p. 76–90. Martynenko, V.G., Verkhozin, A.V., and Altynnikov, A.M., 2009, Zapadnoe gold deposit. Report on exploration results on the flanks and deep levels in 2007–2008 with estimate of resources as of 01.01.2009: Bodaibo, ZAO GRK Sukhoi Log. Russian Federal Geological Fund (in Russian).



SUKHOI LOG Au DEPOSIT, RUSSIA

Meffre, S., Large, R.R., Scott, R., Woodhead, J., Chang, Z., Gilbert, S.E., Danyushevsky, L.V., Maslennikov, V., and Hergt, J.M., 2008, Age and pyrite Pb isotopic composition of the giant Sukhoi Log sediment-hosted gold deposit, Russia: Geochimica et Cosmochimica Acta, v. 72, p. 2377–2391. Micon, 2010, Audit review of the Verninskoe gold deposit, Bodaibo region, Russian Federation, December 2010: Norwich, UK, Micon International Co. Limited. Migachev, I.F., Karpenko, I.A., and Ivanov, A.I., 2008, The Sukhoi Log gold deposit: Reappraisal and estimation of forecasting of ore field and district: Otechestvennaya Geologia, no. 2, p. 55–67 (in Russian). Mungalov, N.N., 2007, The Lena gold fields: A historical essay. Book II: Irkutsk, OOO Reprotsentr A1, 272 p. (in Russian). Neymark, L.A., Rytsk, E.Yu., Gorokhovsky, B.M., Amelin, Yu.V., Ovchinnikova, G,V., Smirnov, M.Yu., and Gracheva, T.V., 1993, Geochronology and isotope geochemistry of gold deposits in the Baikal fold region, in Bibikova, E.V., ed., Isotopic dating of endogenic rocks: Moscow, Nauka, p. 124–146 (in Russian). Obruchev, V.A., 1935, A problem of gold in pyritic shales of the Lena-Vitim region: Problems of Soviet Geology, v. 5, p. 60–69 (in Russian). Palenova, E.E., Yudovskaya, M.A., Frei, D., and Rodionov, N.V., 2019, Detrital zircon U-Pb ages of Paleo- to Neoproterozoic black shales of the Baikal-Patom Highlands in Siberia with implications to timing of metamorphism and gold mineralization: Journal of Asian Earth Sciences, v. 174, p. 37–58. Powerman, V., Shatsillo, A., Chumakov, N., Kapitonov, I., and Hourigan, J., 2015, Interaction between the Central Asian orogenic belt and the Siberian craton as recorded by detrital zircon suites from Transbaikalia: Precambrian Research, v. 267, p. 39–71. Rundqvist, D.V., Mints, M.V., Larin, A.M., Nenakhov, V.M., Rytsk, E.Yu., Turchenko, S.I., and Chernyshov, N.M., 1999, Metallogeny early Precambrian geodynamic settings: Moscow, Ministry of Natural Resources of the Russian Federation, 399 p. (in Russian). Rundqvist, I.K., Bobrov, V.A., Smirnova, T.N., Cmirnov, M.Y., Danilova, M.Y., and Ascheulov, A.A., 1992, Stages of formation of the Bodaibo ore district: Geology of Ore Deposits, v. 34, p. 3–15 (in Russian). Rytsk, E.Yu., Kovach, V.P., Yarmolyuk, V.V., Kovalenko, V.I., Bogomolov, E.S., and Kotov, A.B., 2011. Isotopic structure and evolution of the continental crust in the East Transbaikalian segment of the Central Asian fold belt: Geotectonics, v. 45, p. 349–377 (in Russian). Storozhenko, A.A., Nabrovenkov, O.S., Kazakevich, Yu.P., and Shofman, I.L., 1965, 1:200,000 State Geological Map of the USSR, Sheet O-50-VIII: Moscow, TsNIGRI.

543

Suprunenko, A.N., Ermolov, A.I., and Baranov, A.G., 1975, Report on results of geophysical works at the Sukhoi Log deposit in 1971–1974: Irkutsk, Russian Federal Geological Fund (in Russian). Tsygankov, A.A., Matukov, D.I., Berezhnaya, N.G., Larionov, A.N., Posokhov, V.E., Tsyrenov, B.Ts, Khromov, A.A., and Sergeev, S.A., 2007, Magma sources and stages of emplacement of the late Paleozoic granitoids in the West Transbaikal region: Geology and Geophysics, v. 48, p. 156–180. Wilde, A., Edwards, A., and Yakubchuk, A., 2003, Unconventional deposits of Pt and Pd: A review with implications for exploration: Society of Economic Geologists Newsletter 52, p. 1, 10–18. Wood, B.L., and Popov, N.P., 2006, The giant Sukhoi Log gold deposit: Geologia i Geofizika, v. 47, p. 315–341. Yakubchuk, A., 2000, Bodaibo mesothermal goldfields in Siberia and Carlintype gold mineralization in South China: A similarity of regional structural pattern, in Cluer, J.K., Price, E.M., Struhsacker, E.M., Hardyman, R.F., and Morris, C.L., eds., Geology and ore deposits 2000: The Great Basin and beyond: Reno, Geological Society of Nevada, v. 1, p. 539–547. Yakubchuk, A.S., Shatov, V.V., Kirwin, D., Edwards, A., Tomurtogoo, O., Badarch, G., and Buryak, V.A., 2005, Gold and base metal metallogeny of the Central Asian orogenic supercollage: Economic Geology 100th Anniversary Volume, p. 1069–1096. Yakubchuk, A., Stein, H., and Wilde, A., 2014, Results of pilot Re-Os dating of sulphides from the Sukhoi Log and Olympiada orogenic gold deposits, Russia: Ore Geology Reviews, v. 59, p. 21–28. Yudovskaya, M.A., Distler, V.V., Rodionov, N.V., Mokhov, A.V., Antonov, A.V., and Sergeev, S.A., 2011, Relationship between metamorphism and ore formation at the Sukhoi Log gold deposit hosted in black slates from the data of U-Th-Pb isotopic SHRIMP dating of accessory minerals: Geologiya Rudnykh Mestorozhdenii, v. 53, p. 32–64 (in Russian). Yudovskaya, M.A., Distler, V.V., Prokofiev, V.Yu., and Akinfiev. N.N., 2016, Gold mineralisation and orogenic metamorphism in the Lena province of Siberia as assessed from Chertovo Koryto and Sukhoi Log deposits: Geoscience Frontiers, v. 7, p. 453–481. Zorin, Yu.A., Mazukabzov, A.M., Gladkochub, D.P., Donskaya, T.V., Presnyakov, S.L., and Sergeev, S.A., 2008, Silurian age of major folding in Riphean deposits of the Baikal-Patom zone: Doklady Earth Sciences, v. 423, p. 1235–1239. Zorin, Y.A., Sklyarov, E.V., Belichenko, V.G., and Mazukabzov, A.M., 2009, Island arc-back-arc basin evolution: Implications for Late Riphean-Early Paleozoic geodynamic history of the Sayan-Baikal folded area: Russian Geology and Geophysics, v. 50, p. 149–161.

©2020 Society of Economic Geologists, Inc. SEG Special Publications, no. 23, pp. 545–558

Chapter 26 Geology of the Hishikari Gold Deposit, Kagoshima, Japan Takayuki Seto,1 Yu Yamato,2 Ryota Sekine,3 and Eiji Izawa4,† 1Sumitomo 2Sumiko

Metal Mining Co., Ltd., Hishikari Mine, 3844, Hishikarimaeme, Isa, Kagoshima, 895-2701 Japan

Resources Exploration and Development Co., Ltd., 8-21,3-Chome, Toranomon Minato, Tokyo, 105-0001 Japan 3Sumitomo

Metal Mining Co., Ltd., 11-3 Shimbashi 5-Chome, Minato, Tokyo, 105-8716 Japan 42-15-4

Sakurayamate, Shingu, Fukuoka 811-0113 Japan

Abstract The bonanza-grade, low-sulfidation epithermal Hishikari gold deposit is located in the Plio-Pleistocene volcanic area of southern Kyushu, Japan. The concealed veins were discovered in 1981 and the mine has since produced 5.462 million metric tons (Mt) of ore averaging 44.3 g/t Au (242 t Au) from 1985 to the end of 2018, at which time reserves were 7.98 Mt at 20.9 g/t Au. The Hishikari deposit consists of the Honko, Sanjin, and Yamada ore zones, which occur in a NE-trending area 2.8 km long and 1.0 km wide. The veins are hosted by basement sedimentary rocks of the Cretaceous Shimanto Supergroup and by overlying Hishikari Lower Andesites of Pleistocene age. Sinter occurs about 100 m above the Yamada ore zone. Temperature-controlled hydrothermal alteration zones occupy an area of >5 km long and 2 km wide. The Honko and Sanjin veins occur within a chlorite-illite alteration zone (paleotemperature >230°C), whereas the Yamada veins occur within an interstratified clay mineral zone (150°–230°C). The marginal alteration comprises quartz-smectite (100°–150°C) and cristobalite-smectite (150 m thick and crop out in the eastern part of the mine area. The unit is made up of lightgray, hornblende-bearing dacite lavas with well-­ developed flow banding and interbedded tuff breccias. The lavas contain phenocrysts of plagioclase, augite, hypersthene, and magnetite, which are set in a glassy groundmass containing cristobalite and tridymite. The Maeda Dacite is very similar, although it is defined as a separate stratigraphic unit, occurring northwest of the Hishikari vein zone (Fig. 2). The Hannyaji welded tuff is widely distributed in the mine area and is the product of a distal volcanic eruption. The tuff is 50 to 90 m thick and sandwiched between underlying Hishikari Lower Andesites and Shishimano Dacites and the overlying Ito pyroclastic flow. The deposit is gray, glassy, and dacitic, with plagioclase, hornblende, magnetite, quartz, cristobalite, and tridymite phenocrysts.

The Hishikari Upper Andesites occur east of the mine, on top of the Shishimano Dacites. The unit consists of dark-gray, fine- to medium-grained andesite lava flows and pyroclastic rocks, with plagioclase, augite, hypersthene, olivine, and magnetite phenocrysts. The youngest deposits are the late Pleistocene Kakuto Group lacustrine sediments to the east and the Ito pyroclastic flow deposit to the west (Fig. 2). The Kakuto Group was deposited within the Kakuto caldera (0.56–0.57 and 0.33– 0.34  Ma; Machida and Arai, 2003), located on the northern margin of the Kagoshima graben (Fig. 1). The widespread Ito deposit is up to 70 m thick and made up of pumiceous hornblende dacite, which erupted 29,000 years ago (Okuno, 2002) from the Aira caldera (Fig. 1). Age of volcanic rocks and epithermal mineralization Radiometric ages of volcanic rocks and vein minerals (Fig. 3) constrain the timing of volcanic activity and mineralization. Andesitic volcanism in the Hishikari area began at about 1.62 ± 0.09 Ma and continued to 0.78 ± 0.08 Ma. An older age of 2.34 ± 0.05 Ma was obtained for andesite from the base of the volcanic pile, and this sample may represent older volcanic activity to the west. Dacite domes were emplaced from 1.16 ± 0.21 to 0.66 ± 0.04 Ma, and this period of magmatism overlaps the period of epithermal gold mineralization, which occurred between 1.25 ± 0.12 and 0.606 ± 0.009 Ma. Based on the available age data, gold mineralization formed first in the Yamada zone, followed by the Honko and then Sanjin zones (Fig. 4). The paleosurface at the time of hydrothermal activity is preserved beneath the Hannyaji welded tuff, which was deposited between 0.73 ± 0.11 and 0.58 ± 0.20 Ma. The Hishikari Upper Andesites erupted between 0.58 ± 0.10 and 0.51 ± 0.06 Ma and were probably sourced from volcanic centers associated with the Kagoshima graben. Structural geology The basement Shimanto Supergroup is one of the principal host-rock units for vein mineralization. In the vicinity of Hishikari, the unconformable contact with overlying volcanic rocks forms a structural high, which appears to have developed mostly during emplacement of the Hishikari Lower Andesites (Fig. 2). A period of extension and normal faulting appears to be responsible for the elevational offset on the basement contact, which ranges from 120 m asl in the mine area to 800 m below sea level (bsl) outside the mine area (Fig. 2). By the time the Shishimano Dacites were emplaced, extensional fault activity was waning. Young volcanic stratigraphy conceals the older formed fault structures, and only two faults have been detected on the surface, the Shishimano and the Matsuyama West (Fig. 2). The Shishimano fault strikes N 80° E and dips 85° N and is located 2  km southeast of the Hishikari vein zone where it cuts the Shishimano Dacites. The Matsuyama West fault strikes N 30° W and dips 65° E and is located about 3 km east-­southeast of the Hishikari veins, near the eastern limit of the Hishikari volcanic area. This fault also cuts the Shishimano Dacites and is likely related to formation of the Kagoshima graben. Mineralization and Hydrothermal Alteration The Hishikari low-sulfidation epithermal deposit is represented by a system of steeply dipping, subparallel veins,



549

HISHIKARI GOLD DEPOSIT, KAGOSHIMA, JAPAN

Uono Shinkawa

C V

S M

Hirasawatu

Yuno

Se nd ai -r er iv

A

400

B

Hishikari Deposit

400

200

200

0mL

0mL

-200

-200

-400

-400

-600

-600

-800

-800

-1000

Early Pleistocene Holocene / Late Pleistocene Shishimano Dacites Alluvium / Kakuto Group Hishikari Middle Andesites Ito pyroclastic flow Maeda Dacite Middle Pleistocene Hishikari Lower Andesites Hishikari Upper Andesites Cretaceous Hannyaji welded tuff Shimanto Supergroup

Faults Shishimano fault (S) Matsuyama west fault (M) Resistivity discontinuity

-1000

Veins projected to the surface Cobble of quartz-alunite rock (C) Quartz vein outcrop (V)

Fig. 2. Geology of the Hishikari mine area. Surface geology simplified from mapping by New Energy and Industrial Technology Development Organization (1991), with vein locations projected to surface. Cross section was constructed using drill information and data from Schlumberger depth sounding (Metal Mining Agency of Japan, 1994). M = Matsuyama West fault, S = Shishimano fault.

550

SETO ET AL.

2.5 2.34 Ma

2

Yamada zone

1.62 Ma Maeda Dacites

Age (Ma)

1.5

1.16 Ma

1

Sanjin zone

Hannyaji welded tuff

1.25 Ma 1.11 Ma

1.10 Ma

1.05 Ma

0.79 Ma 0.84 Ma

0.5

Honko zone

0.73 Ma

0.78 Ma

Hishikari Middle Andesites

Hishikari Lower Andesites

Hishikari Upper Andesites

0.75 Ma

0.66 Ma

0.61 Ma

0.65 Ma

0.58 Ma

0.58 Ma

Shishimano Dacites

0.51 Ma

0 Fig. 3. Compilation of K-Ar dates of volcanic rocks and ore veins (open symbols) and 40Ar/39Ar ages of vein minerals (solid circles). The x-axis has no dimension, and the results are stretched horizontally simply for clarity. Ages of the Hannyaji welded tuff were obtained by fission-track dating. Error (2σ) of age data ranges from 0.02 to 0.17 Ma (avg 0.08 Ma; n = 92) for K-Ar dates, from 0.004 to 0.13 Ma (avg 0.02 Ma; n = 45) for 40Ar/39Ar ages, and from 0.11 to 0.20 Ma (avg 0.15 Ma; n = 4) for fission-track ages. Data sources: Izawa et al. (1990), New Energy and Industrial Technology Development Organization (1991), Sekine et al. (2002), Sanematsu (2005), Tohma et al. (2010), and Suga (2013).

2.8  km long and 1.0 km wide, that trend N 50° E (Figs. 4, 5). Three ore zones are distinguished, called Honko, Sanjin, and Yamada. The Honko zone is composed of Daisen, Zuisen, Ryosen, Hosen, and Kinsen veins, and the Sanjin ore zone is composed of the Keisen and Shosen veins. Most of the veins in these two ore zones transect the basement unconformity, and gold mineralization is hosted by both the Hishikari Lower Andesites and the Shimanto Supergroup metasedimentary rocks. The Yamada ore zone consists of Seisen, Eisen, Kosen,

W120B

W160B

W80B

Yusen, and Fukusen veins, and gold mineralization is entirely hosted by volcanic rocks of the Hishikari Lower Andesites. Individual veins generally strike N 50° E and dip between 70° and 90° to the northwest, with strike lengths of 200 to 500 m and widths of 0.2 to 4 m (Ibaraki and Suzuki, 1993). In the Yamada ore zone, several veins branch to N 30° E and form a cymoid loop, whereas other veins in the northern part strike N 70° E (Sekine et al., 1998; Yamato et al, 2002; Murata et al., 2008).

W40B

E80B

E40B

0B

±

50

50

Daisen Zuisen

Eisen

Seisen

1 0 0 100

0G

Kinsen

Hosen

50

Keisen

-20 0

0

0

5

Kosen

-5 0

N40G

0

Ryosen

50

-150

Honko zone

50

Yamada zone

Yusen Shosen -10

Fukusen

-450

-35

-25

0

-30

0

0

0

0 100 200

50

-400

Sanjin zone 400 m

0

-50

Fig. 4. Vein system hosting ores and structure contour surface (m asl) of the contact with the underlying Shimanto graywacke basement.

S40G



551

HISHIKARI GOLD DEPOSIT, KAGOSHIMA, JAPAN

C

W120B

W160B

400

W40B

W80B

E80B

D

400

0.7 - 0.6 Ma

200

200

Maa .84 M .62 - 00.84 1.62 1

0

0

Yamada zone 1.25 - 0.61Ma

-200 S40G

0G

E

200

200

0 -200

E40B

1.1 - 0.66 Ma 29 ka

F

0B

0

0

200m

-200

100 - 84 Ma

Silica sinter / siliceous mudstone Vein Bottom of Lacustrine layer Ito pyroclastic flow Hannyaji welded tuff Shishimano Dacites

Honko - Sanjin zones 1.11 - 0.75 Ma

-200

Hishikari Lower Andesites Lacustrine layer

(Mudstone / sandstone / tuff)

Tuff Andesite lava Andesitic pyroclastics

Shimanto Supergroup

Fig. 5. Geologic map and schematic long and cross sections of epithermal veins in the Hishikari deposit. The structure contour elevations of the lacustrine sedimentary sequence over the Yamada zone is shown, along with occurrences of silica sinter.

The vertical interval of ore-grade mineralization (>20 g/t Au) occurs between –60- and 120-m elev, and high ore grades generally occur in the highest part of the interval. In the Honko and Sanjin zones, stratigraphy and host rocks appear to have limited influence on the development of exceptionally high-grade mineralization (>100 g/t Au), which is spatially associated with the unconformity; that is, the high-grade mineralization occurs below the basement unconformity in the Hosen-1, Hosen-2, and Keisen-3, above it in the Daisen-1, and across it in the Zuisen-1 veins. The average gold grade of the Yamada veins is generally lower than that of the Honko and Sanjin zones (Table 1), although there are several highgrade veins in the Seisen group and the eastern part of the Yusen group. In the latter, the main production levels are between 10- and 110-m elev and the upper limit of highgrade ore correlates with a clay-altered tuff layer (Yamato et al., 2002). The clay is characterized by interstratified chlorite/ smectite and interstratified illite/smectite. Overall, the Ag/Au

mass ratio of Hishikari ores is low, ranging from 0.3 to 2.0, and exceptionally low values (0.15–0.25) occur in the northwestern part of the Yamada zone. Vein structures pinch out upward and along strike. For example, in Hosen 2, the uppermost part of the vein terminates abruptly in clay (chlorite)-altered pyroclastic rocks above the unconformity, or it splits into several veinlets. With depth, the structures hosting quartz veins continue downward and the gold grade generally decreases with few exceptions. Below –100-m elev, few drill hole data exist, except for deep drilling beneath the Fukusen vein in the Yamada ore zone, which intersected several quartz veins in the basement Shimanto Supergroup. Laterally, veins narrow near their tips and pinch out abruptly or split into several veinlets. Ore mineralogy Figure 6 shows typical examples of vein fillings and minerals, which consist mainly of quartz (60–92 wt %) and adularia

552

SETO ET AL.

a

b

c

d

e

Fig. 6. Hishikari vein fillings. a. Hosen-1 at –20 mL (E2B) in the Honko ore zone. b. Bladed quartz in the IA band of Hosen1. c. Sketch vein filling paragenesis of Hosen-1. d. Keisen-6 at 10 mL (E45B) in the Sanjin ore zone, showing a 3.8-m-wide vein zone, containing 15.8 g/t Au that is hosted by brecciated shale of the Shimanto Supergroup. e. Yusen-3 at 30 mL (W150B) in the Yamada ore zone showing a 2.6-m-wide, N-dipping quartz vein, with grades of 15.1 g/t Au and 16.8 g/t Ag.

(6–30 wt %) plus smectite (2–8 wt %). Abundant adularia is a distinctive feature, and the amounts of vein adularia is estimated to be 20 wt % in the case of theHonko veins, 15 wt % for the Sanjin veins, and 7 wt % for the Yamada veins, based on whole-rock chemical analyses (484 samples; Ibaraki and Suzuki, 1993) and X-ray diffraction analyses (1,424 samples). The principal metals, sulfides, and sulfosalts in high-grade ore consist of electrum, chalcopyrite, pyrite, naumannite-aguilarite, galena, and sphalerite, with minor amounts of pyrargyrite, hessite, marcasite, tetrahedrite, miargyrite, acanthite, stibnite, and Cd sulfide (Izawa et al., 1990). Polybasite and clausthalite were also reported in the Yusen 1 vein group (Sekine et al., 1998).

Electrum occurs as discrete grains in banded vein fillings made up of quartz-adularia-smectite, quartz-adularia, and quartz-smectite. Electrum also occurs with sulfide minerals such as chalcopyrite, galena, sphalerite, naumannite, pyrargyrite, or hessite. Less frequently, electrum occurs with truscottite, prehnite, stibnite, pyrite, or marcasite, and rarely occurs with columnar adularia or bladed quartz. In the Honko-Sanjin zones, gold grades correlate with chalcopyrite, galena, and sphalerite contents. Electrum grains range from 3 to 50 μm, and the typical size is 5 to 25 μm in the Honko veins, 10 to 50 μm in the Sanjin veins, and 3 to 15 μm in the Yamada veins (Ibaraki and Suzuki, 1993). Electrum contains 66 to 81 wt % Au (avg 70 wt %; Izawa et al., 1990). In addition, Ag-poor gold



HISHIKARI GOLD DEPOSIT, KAGOSHIMA, JAPAN

(84.5–92.1 wt % Au) was reported from the eastern part of the Yamada ore zone (Sekine et al., 1998; Yamato et al. 2002). The iron content of sphalerite ranges from 0.1 to 2.7 mol % FeS (Sekine et al., 1998; Honda, 2003), and fluid inclusions in quartz and adularia give homogenization temperatures that range from 175° to 230°C, with salinities of 0.2 to 2.7 wt % NaCl equiv (Etoh et al., 2002a; Tohma et al., 2010; Takahashi et al. 2017; Shimizu, 2018). The activity of S2 can be estimated from the mol % FeS in sphalerite and temperature values. The aS2 ranges in the stability field of chalcopyrite + pyrite + hematite that corresponds to the range of the “main line” sulfidation state of Barton (1970) and to the range of typical Japanese epithermal vein-type deposits (Shikazono, 1986, 1987). Quartz occurs principally as aggregates with adularia and smectite. Bladed quartz is a typical characteristic of the Honko-Sanjin ores, whereas in the Yamada veins it is observed only in the northeastern area (the Seisen vein group). Bladed, lamellar, or platy quartz is found in many epithermal gold veins and is regarded as a platy calcite replaced by quartz and adularia (Lindgren, 1933; Dong et al, 1995). Bladed quartz at Hishikari is formed from successive precipitation and dissolution, first involving deposition of platy calcite, followed by precipitation of quartz-adularia on calcite plates, and thirdly calcite dissolution leaving cavities in the quartz-adularia aggregates (Etoh et al., 2002b). Urashima (1956) also considered that lamellar quartz in the Koˉnomai veins was formed by the preferential dissolution of platy calcite in aggregates of calcite and quartz. Platy calcite forms under boiling conditions due to CO2 loss and rapid calcite saturation, as observed in New Zealand geothermal systems (Simmons and Christenson, 1994); therefore, bladed quartz is an indicator of fluid boiling (Dong et al., 1995; Simmons and Browne, 2000). Bladed quartz at Honko-Sanjin forms platy aggregates consisting mainly of fine-grained crystals of anhedral quartz (10–200 μm in size) and rhombic adularia (5–70 μm in size). Smectite occurs as minute ( 3 g/t Gold Feldspar Porphyry Augite Hornblende Diorite Hornblende Diorite Phyllic Altered Mudstone Black Mudstone 10,500mN Calcareous Mudstone Interpreted Contact/Fault

Legend

v

22,500mE 22,500mE

22,000mE

22,000mE

750mRL

R

T UL FA E AN M OA

250 m

1,250mRL

1,750mRL

2,250mRL

2,750mRL

A

O ZONE AHD

Zone 7

RAMBARI AHD

EAST ZONE

Mt Roamane

2,750mRL

B

750mRL

1,250mRL

1,750mRL

2,250mRL

Posion of level plan

Fig. 3. Geology of the Porgera deposit. A. Level plan at 2,370-m elevation interpreted mainly from drill holes, showing the distribution of intrusive bodies of the Porgera Igneous Complex, alteration, main faults, and mineralization. B. Vertical section looking west-southwest through the main part of the deposit.

A

Posion of cross-secon

Grid N

23,000mE 23,000mE

12,000mN 12,000mN

True N

23,500mE 23,500mE

11,000mN 11,000mN

11,500mN 11,500mN

564 HAY ET AL.



A

C

565

GEOLOGY OF THE PORGERA GOLD DEPOSIT, PAPUA NEW GUINEA

2 cm

2 cm

B

D

2 cm

2 cm

Fig. 4. Core samples of major rock types. A. Black mudstone of the Chim Formation. B. Augite-hornblende diorite. C. Hornblende diorite. D. Feldspar porphyry.

and 60 m vertical based on underground exposures (Standing, 1994). The fault zone ranges from 1 to 30 m in width and generally consists of intensely foliated fault gouge and cataclasite, developed mainly in strongly deformed black mudstone of the Chim Formation, with variable amounts of thick calcite veins, siliceous fault breccia, and mineralized stockwork breccia (Standing, 1994; Munroe, 1996). The fault zone is generally thicker within more competent intrusions, and narrowest where developed wholly within mudstone (Standing, 1994). Palaeomagnetic studies completed on the Porgera Intrusive Complex by Clark and Schmidt (1993) indicated that a 60° progressive tilting toward the south had occurred within the local geology, suggesting that the Roamane fault zone and sedimentary bedding may have originally dipped approximately 10° to 15° to the southeast. If correct, this may indicate that the Roamane fault zone may represent a reactivated early thrust decollement (Standing, 1994). An array of synmineral, low-displacement normal-(dextral) faults and subvertical extensional fractures hosts the bulk of epithermal gold mineralization, which is mostly developed in the footwall of the Roamane fault zone (Fig. 3). These normal faults typically dip 40° to 60° south and have displacements of 30,000 ppm SO42−). These fluids have high concentrations of dissolved salts and trace metals, and deuterium and oxygen isotope signatures consistent with derivation from mixtures of magmatic ± meteoric fluids, potentially with a minor seawater component (Williamson, 1983; Carman, 1994; Heinrich, 2006; Simmons and Brown, 2006).



LIHIR ALKALIC EPITHERMAL Au DEPOSIT, PAPUA NEW GUINEA

The ~250-m-thick, subhorizontal clay blanket that defines the steam-heated zone at Lihir is a product of recent geothermal activity. The clay blanket is predominantly composed of rocks that underwent argillic alteration (montmorillonite ± illite ± fine-grained carbonate) and occurs across the entire floor of the Luise amphitheater. It has overprinted the Lihir volcanic-­hydrothermal breccia complex. A steam-heated origin is supported by the limited distribution of limonite at the top of the clay blanket. Oxidation of sulfides to limonite was restricted to 50 m bsl at Lienetz (Carman, 1994). The argillic-altered rocks include domains of higher temperature illite alteration that occur above the current upflow zones of the geothermal system. Locally, advanced argillic alteration produced kaolinite ± opal ± alunite ± dickite ± halloysite ± native sulfur along major structures and at structural intersections. Notably, advanced argillic-altered rocks overlie deeper, highgrade (>5 g/t Au) zones in Minifie, Lienetz, Coastal, and Kapit. The intensity and thickness of the clay alteration was greatest onshore and diminished seaward due to topographic effects. This implies that meteoric water in the vadose zone played a significant role in clay alteration, probably due to the condensation of H2S(g) into oxygenated meteoric water (e.g., Sillitoe, 1993). Oxygen-deuterium analyses of downhole water samples from geothermal wells at Lihir confirm the involvement of a mixture of local meteoric water (d18O = −0.6‰; dD = −37‰) and a deep geothermal brine (d18O = 6.0‰; dD = −25‰) in modern geothermal activity (Carman, 1994). The clay blanket has overprinted the sulfide zone and mostly coincides with low gold grades, although elevated gold contents (~1.5–2.5 g/t Au) have been detected locally in alunite-­altered areas within the clay blanket (Clark et al., 2015). In the sulfide zone immediately underlying the clay blanket, gold-rich rinds around pyrite grains may have been produced by leaching of gold from the clay blanket and reprecipitation at the base of oxidation. This could explain the tabular high-grade gold zones beneath the clay blanket that were traditionally referred to as the “boiling zone” by early workers (e.g., Davies and Ballantyne, 1987; Moyle et al., 1990). Downhole analyses of the modern geothermal fluids responsible for the steam-heated zone revealed aqueous gold contents of ~16 ppb in sulfate-chloride water samples (Simmons and Brown, 2006). If the gold flux remained constant over time (currently around 24 kg/yr), then the gold resource at Lihir could have formed in as little as 55,000 yr, based on the current fluid flow regime (Simmons and Brown, 2006). These calculations give useful insights into the overall time frame required to form Lihir’s giant epithermal gold resource. However, given that much of the high-grade gold mineralization was associated with hydrothermal breccia formation, it is likely that mineralization was episodic rather than continuous over this time interval. Genetic Considerations Early recognition of the coexistence of porphyry- and epithermal-style alteration features at Lihir was made during the exploration and delineation of the Lihir gold deposit. This led to the development of models that related sector collapse of the Luise Volcano to the instantaneous superposition of shallow epithermal mineralization onto deeper seated

591

porphyry-style alteration features (Moyle et al., 1990; Sillitoe, 1994; Carman, 1994; Fig. 4). Later geologic and geochemical studies of individual mineralized zones refined this model, providing a greater understanding of the distribution of volcanic and hydrothermal breccias (Blackwell, 2010; Ageneau, 2012; Blackwell et al., 2014) and suggesting that sector collapse may have occurred during a hiatus in magmatic-­hydrothermal activity (Sykora, 2016; Sykora et al., 2018a, b). Recent studies delineated the major early-stage magmatic-­hydrothermal breccia body at Lienetz (Blackwell, 2010; Sykora, 2016; Sykora et al., 2018a) and the series of late-stage volcanic-hydrothermal breccias that crosscut several orebodies and postdated most of the epithermal mineralization (Lawlis et al., 2015). The following sections highlight key aspects of the major geologic events in the formation and evolution of the Lihir gold deposit, which are schematically illustrated in Figure 10. Porphyry stage Porphyry-style hydrothermal activity at Lihir occurred within the core of the Luise Volcano prior to sector collapse (Fig. 10A, B). Northern Lienetz appears to have been a major focus of porphyry-style activity, with widespread potassic alteration assemblages (anhydrite-biotite ± K-feldspar ± magnetite) and related veins and cemented breccias (Fig. 8A, B) that transition outward to a propylitic alteration assemblage (chlorite-calcite ± epidote; Sykora et al., 2018a; Fig. 8C). Hydrothermal biotite K-Ar ages range from 0.917 ± 0.100 to 0.336 ± 0.027 Ma (Davies and Ballantyne, 1987; Moyle et al., 1990; Rytuba et al., 1993; Table 1). No obvious intrusive center for porphyry-style hydrothermal activity has been identified, although syenite-­ cemented breccias at Lienetz (e.g., Fig. 8D) imply proximity to an intrusive complex. Minor biotite-altered pebble dikes at Minifie and Kapit NE imply the presence of additional intrusive complexes at both locations (Blackwell et al., 2014; Lawlis, 2020). Porphyry-style biotite alteration in Lienetz was partly localized by NE-striking faults (Sykora et al., 2018a). The porphyry-stage veins have both low- and high-angle dips and formed in a compressional setting under low differential stress, with the maximum principal stress (σ1) oscillating from subhorizontal to subvertical and fluid pressures temporarily elevated from mineral sealing (Sykora et al., 2018a). A major biotite-anhydrite-cemented magmatic-hydrothermal breccia complex formed late in the evolution of the porphyry stage at Lienetz (Figs. 8G, 10B). The breccia complex is centered on porphyritic syenite dikes (Blackwell, 2010; Sykora et al., 2018a; Figs. 6I, 10B) and contains clasts of anhydrite veins with biotite alteration halos (Fig. 8G, I). The syenite dikes have not been dated. Low-grade porphyry-stage mineralization at Lienetz averages ~0.5 g/t Au and 20 volume %

Ekmai Fm. Shale Member

D a l a m Volcanics Vol c a n i c s Dalam

Quartz Vein Stockwork >50 volume % Limestone Dissolution Breccia >10 volume %

Ekmai Fm. Limestone Member Other Features Ekmai Fm. Sandstone Member

c

Anhydrite Leaching Front

Massive Sulfide (Heavy Sulfide Zone) H y d r ot h e r m a l Hydrothermal Fault B r e cci a Breccia

Limit of Marbleization

d

GRS-4

GRS-4

Cu wt %

Au ppm

>5

>5

2.0 - 5.0

2.0 - 5.0

1.5 - 2.0

1.5 - 2.0

1.0 - 1.5

1.0 - 1.5

0.5 - 1.0

0.5 - 1.0

0.25 - 0.5

0.25 - 0.5

0.1 - 0.25

0.1 - 0.25

< 0.1

< 0.1

Fig. 4. Grasberg northeast-southwest cross sections (location in Fig. 3a), showing a. premining lithology, b. alteration, c. copper distribution, and d. gold distribution.

604

LEYS ET AL.

of the Grasberg deposit (Widodo et al., 1999). The alreadymentioned Big Gossan skarn was also discovered during this period as was the Wabu gold skarn (O’Connor et al., 1994) located 40 km north-northwest of Grasberg (Fig. 2).

adiabatic decompression melting of asthenospheric mantle that upwelled into the slab gap. Additional mafic magma that was potassium rich and had a high Au/Mo ratio was generated by partial melting of stretched subcontinental lithospheric mantle (Müller and Groves, 2000; Cloos and Housh, 2008). Regional Geologic Setting Most magma that rose from the mantle pooled near the Moho The Ertsberg-Grasberg mining district is located near Puncak beneath the collision-generated mountain range (Huppert Jaya, which at 4,884 m is the highest peak in New Guinea, an and Sparks, 1988) and this heated and partly melted the felsic island long recognized as the product of Cenozoic subduction lower continental crust to generate hundreds of cubic kiloculminating in an oceanic arc-continent collision (Dewey and meters of intermediate composition magma that intruded the Bird, 1970). The Grasberg porphyry and nearby Ertsberg-­ Central Ranges (O’Connor et al., 1994; Housh and ­McMahon, related skarn copper-gold deposits formed during the final 2000). stage of this collisional orogenesis (Fig. 2). Geologic studies of the Central Ranges of Papua ProvDistrict Geology ince have revealed timing relationships that, combined with mechanical considerations, lead to a refined understanding of Sedimentary rocks the tectono-magmatic effects of collisional orogenesis (Cloos The Ertsberg-Grasberg district is underlain by sedimenet al., 2005). Prior to 30 Ma, the oceanic end of the Australian tary rocks that are cut by a variety of phaneritic to porphyplate began to subduct beneath the nonaccretionary (Mari- ritic intrusions (Fig. 3; Leys et al., 2012). The northern and ana-type) inner Melanesian island arc (Quarles van Ufford predominantly mineralized central portion of the district and Cloos, 2005). Tectonic reconstructions indicate that the around Grasberg is dominated by the >1.7-km-thick, largely Australian and Pacific plates converged obliquely along a carbonate rocks of the Lower to Middle Cenozoic New west-southwest to east-northeast axis at about 12 cm/yr, with Guinea Limestone Group. From top to bottom this group underthrusting of around 5 cm/yr. Shelf limestone (New comprises the >1,200-m-thick Kais Formation, clean and in Guinea Limestone Group) strata that form the host rocks for places highly fossiliferous limestone; the ~30-m-thick Sirga Grasberg at surface accumulated on top of Australian conti- Formation quartz-carbonate sandstone; the ~150-m-thick nental basement until at least 15 Ma. Petroleum exploration Faumai Formation massive-bedded, clean limestone; and the wells across the southern half of New Guinea reveal that ~300-m-thick Waripi Formation, an anhydrite nodule-­bearing widespread siliciclastic sedimentation began around 12 Ma dolomitic limestone with thin quartz sandstone beds. due to erosion of emerging islands with a change to conglomThe New Guinea Limestone Group is underlain by silicieratic molasse at about 5 Ma (Quarles van Ufford and Cloos, clastic rocks of the Jurassic-Cretaceous Kembelangan Group, 2005). The southern flank of the Central Range is capped by the upper part of which comprises the three-member Ekmai bulldozed carbonate-rich, shallow-water strata deposited on Formation. From top down these members are the 3- to Australian passive margin strata. The northern flank is under- 5-m-thick Ekmai Shale (hornfelsed within the mineral dislain by metamorphosed, largely siliciclastic, deeper water trict), the ~100-m-thick Ekmai Limestone (a calcareous mudstrata that pass northward into the Papuan ophiolite belt, the stone), and the ~600-m-thick Ekmai Sandstone (Leys et al., upturned forearc basement of the Melanesian island arc that 2012). extends to the north coast of the island but is largely buried by Igneous rocks recent alluvial sediments. Near the top of the Central Ranges the carbonate strata The district contains intrusive rocks generated during two and underlying Mesozoic siliciclastics are deformed into a geochemically distinct magmatic events: an early (3.7–3.3 Ma) series of kilometer-scale, en echelon, open folds affected by and lower K suite of small dikes and sills, the largest of which thrust or reverse faulting. Mapping along the mining district is the Wanagon sill, and a later (3.5–2.7 Ma), high K suite of access road reveals an upright, N-dipping, ~12-km-thick sed- intrusions including Kay, Karume, North Grasberg, Grasberg, imentary section, of Cretaceous age and older, that overlies a and Ertsberg (Fig. 3; Leys et al., 2012). Isotopic studies (Sr, basement of greenschist facies slates and phyllites of probable Nd, Pb) by Housh and McMahon (2000) indicate that magPrecambrian age within the northern limb of the 50-km-wide, mas were primarily derived from the asthenosphere, with 300-km-long Mapenduma anticline (Nash et al., 1993). Apa- significant contributions derived from lithospheric mantle tite fission track thermochronology indicates the mid-slope (enriched in potassium by ancient episodes of metasomatism) region was unroofed at rates of 1 to 2 km/m.y. (Weiland and and modified by substantial (±50%) assimilation of lower Cloos, 1996), showing that southward displacement (totaling crust. ~10 km) and unroofing of the Mapenduma anticline began at McMahon (1999) showed that the Grasberg and Ertsberg ~8 Ma. intrusions are chemically similar but that textures and mafic Volumetrically minor, but widespread volcanism including minerals vary considerably. Most Grasberg rocks are porphyat Grasberg occurred along the spine of the western Central ritic, with plagioclase, biotite, and hornblende phenocrysts Ranges from 8 to 3 Ma after the subduction zone jammed around 3 to 5 mm in size that are typically ten times coarser when Australian continental crust became too buoyant to than the groundmass. In contrast, >95% of the Ertsberg continue subducting (Cloos et al., 2005). After jamming, intrusion is equigranular, with clinopyroxene and sphene as lithospheric delamination and breakoff of the subducted end dominant mafic minerals. Marble aureoles occur around all of the Australian plate occurred and magma formed by the the intrusions in the district, and at Grasberg the aureole



GRASBERG Cu-Au-(Mo) DEPOSIT: PRODUCT OF TWO OVERLAPPING PORPHYRY SYSTEMS

605

extends ~250  m from the contact (Fig. 4b) and up to 1 km where major faults provided fluid pathways (Leys et al., 2012).

deep porphyry system that predates the main stage of Grasberg porphyry copper-gold mineralization.

Folding and faulting Two phases of deformation occurred in the Ertsberg-­ Grasberg district, the first between 12 and 4 Ma caused by regional shortening resulting from the development and subsequent jamming of the subduction zone (Quarles van Ufford and Cloos, 2005). This deformation phase formed kilometer-scale folds characterized as open synclines and anticlines with locally tight chevron subsidiary anticlines, with steep to vertical axes trending 300°, and subsidiary reverse and strikeslip faults (Sapiie, 1998; Fig. 2). The most prominent of these folds is the Yellow Valley syncline into whose axis the Grasberg Igneous Complex was emplaced and whose northern limb rolls over to become part of the regional Mapenduma anticline. The strike-parallel and steeply NE-dipping thrust faults ramped into their high-angle reverse orientation at their leading edges and display stratigraphic offsets in cross section of up to 1,000 m. For historic reasons some strike parallel faults have different names in the west than in the east side of the district, and these names are combined in this report. The major orebodies of Kucing Liar and Ertsberg East skarn system are located along one of these faults, Idenberg 1-Ertsberg 2, which cuts the southern limb of the Yellow Valley syncline (Fig. 3). The second deformation phase was active from 4 to 2 Ma (Sapiie, 1998) and resulted in strike-slip movement mostly along preexisting high-angle reverse faults (Leys et al., 2012). Major strike-slip fault zones are subparallel to upturned bedding, a few tens of meters wide, and internally are highly brecciated. Movement on these faults was predominantly left-lateral, with offsets of tens to several hundred meters. Subvertical, NE-striking tear faults with minor displacements, such as New Zealand Pass and Carstenz Valley, cut these major faults, and one (the Grasberg fault) intersects the Grasberg Igneous Complex (Fig. 3). These northeast-southwest faults as well as others of north-south orientation are interpreted to be Riedel shears (Sapiie and Cloos, 2013). Strike-slip faulting and igneous activity in the district were contemporaneous, and local extension on left-stepping structures between these faults was a major control on intrusion and subsequent hydrothermal activity.

Grasberg Igneous Complex The Grasberg Cu-Au deposit is hosted by the Grasberg Igneous Complex that comprises volcanic rocks, subvolcanic breccias, and monzodioritic intrusions within an upward-flared pipe emplaced into the axis of the Yellow Valley syncline (Figs. 3, 4). Although the center of the pipe is a phreatomagmatic diatreme (MacDonald and Arnold, 1994), wall-rock carbonate dissolution may have contributed significantly to its enlargement (Lambert, 2008). The pipe is ~1,800 m in diameter at surface, ~1,100 m at 3,400 m, and 700 m at 2,400-m elevation (Fig. 4). Near surface, the contact flares to ~45° where the Grasberg Igneous Complex cuts the Kais Formation limestone strata. MacDonald and Arnold (1994) divided the Grasberg Igneous Complex into three main igneous phases that were recently extensively U-Pb zircon dated by Trautman (2013) and Wafforn (2017) [see Geochronology section]. The first phase, the Dalam, occupies the bulk of the Grasberg Igneous Complex and comprises multiple intrusions and subordinate volcanic units emplaced from 3.6 to 3.3 Ma. The second phase, the Main Grasberg intrusion, is a porphyritic, cylindrical to conical plug emplaced at 3.2 Ma. The third and final phase, the Kali dikes were emplaced between 3.2 and 3.0  Ma and form a NW-SE-trending, wedge-shaped body in the southeast quadrant of the Grasberg Igneous Complex. Dalam: Above 3,850-m elevation and cropping out on the premining surface, Dalam rocks consist of well-bedded tuffs and tuffaceous sediments (Fig. 5a), with breccias and flow domes (MacDonald and Arnold, 1994). Below these and extending to about 3,000-m elevation, Dalam rocks can be divided into two units: an inner monomict fragmental magmatic breccia and an outer polymict subvolcanic breccia (Fig. 5b; Paterson and Cloos, 2005a). Below 3,000-m elevation Dalam rocks are generally equigranular but locally porphyritic, fine- to medium-grained monzodiorites that Pollard and Taylor (2005) classified into five intrusion types. However, they could not map these out due to pervasive, texture-­destructive alteration and gradational contacts, and for the same reasons Freeport geologists classify these as undifferentiated Dalam Diorite (Fig. 4a). The Gajah Tidur intrusion (Fig. 4a) is found below 2,800-m elevation and is a distinctive coarsely porphyritic monzonite (Fig. 5d), with up to 50% of feldspar and biotite phenocrysts in a fine-grained K-feldspar and quartz groundmass (Trautman, 2013). The intrusion extends to at least 1,600-m elevation but is poorly defined at depth because of a lack of drilling. No direct observations of the intrusive contact between the Gajah Tidur intrusion and surrounding units have been made because the contact is obscured by intense quartz veining and texture-destructive phyllic alteration (Figs. 4b, 5e). However, based on zircon U-Pb dating (Trautman, 2013) and supported by the distribution of quartz veining and related alteration that is centered on it, the Gajah Tidur intrusion is younger than all Dalam phases. Main Grasberg intrusion: The Main Grasberg (Figs. 4a, 5h) is a vertical plug of fine- to medium-grained porphyritic monzodiorite, with distinctive acicular hornblende phenocrysts. It

Deposit Geology MacDonald and Arnold’s (1994) benchmark study described the upper and middle parts of the Grasberg deposit. This report focuses on the middle and deeper parts of the deposit and illustrates these using geology and alteration cross sections (Fig. 4), rock photographs (Fig. 5), and mineral and elemental distribution models (Fig. 6). Detailed models of mineralization and alteration as well as recent extensive U-Pb zircon dating of intrusions by Trautman (2013) and Wafforn (2017) have greatly improved our understanding of the deep geology of Grasberg. A key finding is that the Gajah Tidur intrusion, associated low-grade quartz stockwork and surrounding porphyry copper-molybdenum mineralization is older than the main Grasberg mineralization stage and not younger as reported by Leys et al. (2012), supporting MacDonald and Arnold’s (1994) identification of a

606

LEYS ET AL.

b

a

c sed

1 cm

1 cm

d

e

f

1 cm

g

cp

mo

py-cv

1 cm

5 cm

h

ap

5 cm

i py-cp

bn-cp

1 cm

1 cm

2 cm

Fig. 5. Hand samples and core showing selected Grasberg igneous rock textures, alteration, and mineralization. a. Dalamstage well-laminated, fine-grained tuffs from ~4,000-m elevation interpreted as reworked crater lake deposits, overlain by coarser lithic-rich lapilli tuff with load structures at its base. b. Moderately propylitic-altered (chlorite-carbonate-quartz) Dalam fragmental andesite from 3,800-m elevation: a polymict breccia with hornblende-rich, commonly rounded igneous clasts and rare angular sedimentary fragments (sed) in a finer matrix of similar composition. c. Strongly potassic-altered (K-feldspar-anhydrite-biotite) Dalam Diorite in core from 2,700-m elevation: a medium-grained, equigranular monzodiorite cut by quartz ± molybdenite veins from the outer part of the Gajah Tidur stockwork. d. Weakly potassic + phyllic-altered Gajah Tidur intrusion from southwest Grasberg at 2,000-m elevation: a crowded monzodiorite porphyry with K-feldspar and hornblende phenocrysts, the latter largely replaced by biotite, and groundmass plagioclase partly altered to sericite. e. Abundant gray quartz stockwork veins containing trace molybdenite, hosted by strongly phyllic-altered (sericite-quartz-pyrite) Dalam Diorite in the center of the Gajah Tidur stockwork at 2,640-m elevation. f. Quartz veins containing molybdenite cut by irregular pyrite-covellite (py-cv) veinlets in the outer edge of the Gajah Tidur stockwork accompanied by phyllic-altered (sericite-quartz-pyrite) Dalam Diorite from southwest Grasberg at 3,130-m elevation. g. Potassic-altered (quartz-biotite-magnetite) Dalam Diorite near its contact with the Main Grasberg intrusion at 3,800-m elevation, cut by early biotite and magnetite veinlets, later coarsely crystalline quartz-anhydrite-(magnetite-biotite) stockwork veins, and lastly by a 1-cm chalcopyrite (cp) vein containing a euhedral apatite (ap) crystal. Late microfractures throughout the rock contain disseminated chalcopyrite deposited at the same time as the chalcopyrite veins. h. Potassic-altered (K-feldspar-biotite-magnetite) Main Grasberg intrusion from central Grasberg at 3,800-m elevation: fine-grained, porphyritic monzodiorite with characteristic biotite-altered acicular hornblende. The intrusion is cut by early biotite and magnetite veinlets, later quartz-anhydrite-(magnetite-biotite) stockwork veins, and finally bornite-chalcopyrite (bn-cp) veins. i. Strongly potassic-altered (K-feldspar-biotite-magnetite-anhydrite) and stockwork-veined Dalam Andesite from southeast Grasberg at 3,900-m elevation, cut by an Early Kali dike that truncates most veins including some quartz-pyrite-chalcopyrite veinlets that cut earlier quartz-anhydrite-(magnetite-biotite) stockwork veins. The dike is a moderately K-feldspar-biotite-altered, fine-grained monzonite with no chilled margin that is cut by a few late quartz-pyrite-chalcopyrite (py-cp) veinlets, indicating it is a late intermineral intrusion.

lies in the center of the Grasberg Igneous Complex pipe and has a narrow cone shape, which reduces from about 800  × 500 m at 4,000-m elevation to 400 × 250 m at 3,000-m elevation, and is absent below 2,700-m elevation where only Kali dikes are present (Fig. 4a). A N-striking dike of weakly mineralized Main Grasberg at 2,800-m elevation cuts highly quartzveined and mineralized Dalam rocks (Fig. 6a), providing

further evidence for pre-Main Grasberg hydrothermal activity. Above 3,000 m, the Main Grasberg is strongly potassic altered and overprinted by an intense quartz-magnetite stockwork that is cut by abundant chalcopyrite-bornite veins, and this zone hosts most of the Grasberg high-grade Cu-Au ore. The Plagioclase dikes (Pollard et al., 2005) are rare but distinctive dikes, with >1-cm plagioclase phenocrysts in a fine-grained



GRASBERG Cu-Au-(Mo) DEPOSIT: PRODUCT OF TWO OVERLAPPING PORPHYRY SYSTEMS

matrix, which are spatially associated with and cut the Main Grasberg. Kali: The Kali dikes form a steeply SW-dipping, wedgeshaped composite monzodiorite body that cuts the Main Grasberg. The body has an apex near the center of the Grasberg Igneous Complex, strikes southeast, and dips approximately 80° to the southwest (Figs. 3, 4a). It is 300 m across where it cuts the edge of the Grasberg Igneous Complex, then widens to a maximum of 600 m toward the southeast where in drill holes it is truncated by the Ertsberg intrusion. There are two main Kali phase intrusions, the older of which is a volumetrically minor biotite monzodiorite dike called the Early Kali that is mineralized near its contact with Main Grasberg-and Dalam-stage rocks (Fig. 5i). The Early Kali dike is intruded by the Late Kali dike, a wedge-shaped, larger hornblende monzodiorite body that postdates all significant mineralization (MacDonald and Arnold, 1994). Marginal Breccia and Heavy Sulfide Zone: From surface down to ~3,700-m elevation, the northeast and southwest contacts of the Grasberg Igneous Complex with wall-rock limestones are 100- to 200-m-wide zones of breccia called the Marginal Breccia (Figs. 3, 4a; Sapiie, 1998). These hydrothermal breccias have sharp contacts with the Grasberg Igneous Complex and comprise an inner polymict, matrix-supported breccia with minor igneous and skarn fragments that grades sharply outward into a monomict, clast-supported breccia containing partly dissolved, marbleized limestone clasts that passes into marbleized carbonates. From 3,700- to at least 2,700-m elevation, igneous and sedimentary wall rocks adjacent to the Grasberg Igneous Complex country-rock contact are replaced by a 20- to 40-m wide zone of semi- to massive pyrite named the Heavy Sulfide Zone (Fig. 4) that contains locally abundant chalcopyrite, minor covellite, and trace enargite. The outer parts of the Heavy Sulfide Zone contain varied, but generally minor zinc mineralization (Fig. 6h) as sphalerite and lesser lead mineralization as galena, and rare meter-scale lenses of these sulfides are found in limestone up to a few hundred meters outboard of the Grasberg Igneous Complex. At lower elevation the Heavy Sulfide Zone contains pods of semimassive magnetite that are veined and partly replaced by pyrite, indicating that early skarn formation was succeeded by a sulfidation event. Below about 3,500-m elevation, the Kali dikes extend outside the Grasberg Igneous Complex and cut the Heavy Sulfide Zone and the Marginal Breccia. Banded Clay: Two distinctive bodies of halloysite clay with highly contorted laminations, named “Banded Clay” (Fig. 3), cropped out at the premining surface and interfingered with marginal breccias along the easr-northeast and west-­southwest edges of the Grasberg Igneous Complex. The latter body is by far the larger; it measures up to 600 m long by 200 m wide and 200 m high and rapidly pinches out at depth. The clay contains significant amounts of gold and minor copper and silver associated with small amounts of pyrite, believed to have been deposited by postvolcanic hot springs (Lambert, 2008). Hydrothermal breccias: Small zones of hydrothermal breccia are common within highly altered and stockwork-veined Grasberg Igneous Complex units. However, a much larger NW-SE-trending, tabular breccia body occurs from 3,000to 2,600-m elevation along the southern edge of the early

607

Kali dike intrusion and extends beyond its tip (Fig. 4a). The body measures 500 m long by up to 100 m wide and contains millimeter- to meter-sized fragments of Early Kali as well as quartz-veined and mineralized Dalam rocks in an anhydrite, sericite, and pyrite cement. The breccia occupies reactivated NW-SE-trending strike-slip faults that controlled the Kali emplacement (Sapiie and Cloos, 2013) and is interpreted to be of tectonic-hydrothermal origin. Pebble dikes: These linear, rounded-fragment breccias are common in many porphyry deposits (Sillitoe, 2010) and are locally abundant close to the edges of the Grasberg Igneous Complex, where some cut the Marginal Breccia and extend up to 200 m into the surrounding sedimentary rocks. Within the Grasberg Igneous Complex, dike compositions vary from rounded igneous clasts of Dalam andesite and diorite, up to tens of centimeters across, in a finer matrix of similar composition. Near the edge of the complex, dikes comprise mixed igneous and sedimentary clasts and beyond the complex margin dikes contain predominantly rounded sedimentary clasts of variable sizes. Farthest from the Grasberg Igneous Complex, within carbonate country rocks, the pebble dikes pass laterally and vertically into karst-like dissolution breccias and can only be confidently distinguished from these by the presence of rare igneous or very rare sulfide-bearing clasts (Fig. 4a). Structure Sapiie and Cloos (2013) grouped numerous small faults with limited movement that cut the Grasberg Igneous Complex into three NW-trending structural domains based on pit mapping in the upper part of the deposit, above 3,900-m elevation. The northeast domain is dominated by nearly northsouth, right-lateral strike-slip faults, the southwest domain by 070° left-lateral strike-slip faults, and the central domain by 300° left-lateral strike-slip faults as well as many small faults with trends typical of the other two domains. This reinforced their earlier conclusion that the Grasberg Igneous Complex was emplaced during local extension on left-­ stepping, ­northeast-southwest connecting structures between district-scale, west-northwest to east-southeast, left-lateral strike-slip faults (Sapiie and Cloos, 2004). The northeast-southwest Grasberg fault is the only structure that fully crosses the Grasberg Igneous Complex and becomes more prominent in the deeper part of the pit below 3,300-m elevation, where it is an up to 200-m-wide zone of minor faults, some filled with clay gouge, and small-scale breccias. This fault is cut by the west-northwest to east-southeast Kali dikes within the central structural domain and which contain very few faults except at their margins. Many of the veins associated with hydrothermal alteration and mineralization have trends similar to faults within the three structural domains and in places faults are infilled by veins. Grasberg Alteration and Mineralization Grasberg hydrothermal alteration and mineralization is the product of two partly overlapping porphyry systems, the older and deeper Gajah Tidur copper-molybdenum porphyry and the younger and shallower Main Grasberg copper-gold porphyry. Their mineralized and phyllic-altered zones overlap within Dalam rocks but their potassic-altered and stockwork-veined centers are distinct, allowing many of their

608

LEYS ET AL.

HSZ

Pebble dike

3400L

Tk1

Tgd fa Tgdd

cc

XRD samples within 100m of block diagram

Geology Tf

N

2800L

Kkes Tgpd

200m

DDH within 100m of block diagram

Tgm

Tigt

Kkel

bb

2 Tk Tk3

Tsk

a

Tw 2400L

MGI 1% Cu shell

d

Base of GT Cu shell

e

Au

Cu

g

f

i

h

Mo

j

Zn

Chalcopyrite

l

k

Bornite

Ag

Covellite

PyritePyrrhote

Fig. 6. Block diagrams through the Grasberg deposit viewed looking west, illustrating geology, sample spacing, metal grade, and mineralogy. a. Lithology, including scale and view direction. Major intrusion outlines are shown in each panel of the figure. Grasberg Igneous Complex contact with sedimentary country rocks in black, Gajah Tidur porphyry in white, Main Grasberg in dark red, and Kali in orange. b. Diamond drill holes within 100 m of the block diagram, whose assay data support Cu, Au, Ag, Mo, and Zn block models, and whose assay pulp magnetic susceptibility measurements support magnetite wt % block models. c. Approximately 50-m spaced drill hole samples within 100 m of the block diagram that were analyzed by XRD to support mineralogy block models. d. Copper (wt %), showing base of Gajah Tidur copper shell and approximate limit of Main Grasberg 1% Cu shell. e. Gold (ppm). f. Silver (ppm). g. Molybdenum (wt %). h. Zinc (wt %). i. Chalcopyrite (wt %). j. Bornite (wt %). k. Covellite (wt %). l. Pyrite + pyrrhotite (wt %).

m

o

n



m

o

n

K-feldspar

Quartz

p Quartz

q

p

s

r

Plagioclase

r

q

BiotePhlogopitet

Plagioclase

K-feldspar

BiotePhlogopite

s

609

GRASBERG Cu-Au-(Mo) DEPOSIT: PRODUCT OF TWO OVERLAPPING PORPHYRY SYSTEMS

AnhydriteGypsum

Magnete Magnete

u

t

Amphibole Amphibole

AnhydriteGypsum

u

MuscoviteMuscoviteIllite (ser) Illite (ser) (sericite) (sericite)

Kaolinite Kaolinite

Bt-P hBlt-P hl Anh nh--Gyp Anh nh--Gyp t% Au pp ppm pp Mo wt% Zn wt% Cp wt% Bn wt% Cv wt% P y-P o wt% Qtz wt% Kfs wt% P lg wt% % % Kaol wt% Mus-Ill wt% Cu wt% CAu uwpp ppm m Appm gmpp ppm mAg M omwt% Zn wt% Cp wt% Bn wt% Cv wt% P y-P o wt% Qtz wt% Kfs wt% P lg wt% wt% wt%Mg wtM g wt%wt% wAtm %ph wtA mph wt% Kaol wt% Mus-Ill wt% 5 5 50 0.1 5 5 5 5 50 75 75 75 20 20 15 20 20 40 5 5 50 0.1 5 5 5 5 50 75 75 75 20 20 15 20 20 40

2 1.5

2

2

2

25

25

1.5

1.5

10

1

1

5

1.5

10

0.05 0.04

0.05

0.04 0.03

2 1

2

1 0.5

1

0.51

0.5 5

0.5

0.5 0.25

0.252.5

2.5 0.03 0.02 0.5 0.2 1 0.02 0.01 0.2 0.1

0.25

0.25 0.1

0.1 1

0.5 0.01 0.005 0.1

0.1

00.1

0

0 0.5

0 0.005

0

-

0

4 3

4

3 2

4 3

4

3 2

4 3

4

3 2

20

50

50

50

10

40

40

40

5

30

30

30

20

10

50 40

40

2030

50 40

2 1 1 0.5

2 1 1 0.5

21

53

10.5

31

20 10

1020

20 30 10 20

0.50.25

0.50.25

0.5 0.25

1-

510

5 10

0.25 0

0.250

-0

05

05

0 0.25

30 20

50

10

7.5 5

10 7.5

10

7.5 5

10 7.5

10

7.5 5

10

10

10

30

10

10

30

7.5

7.5

25

2

52

515

20

2.5 1

21

210

15

7.5

5

7.5 5

25 20

2

5

2

5

1

2

1

2

1

5 10

-

1

-

1

-

1

-

1-

15

10

0 5

0

-

0

-

0

-

0

-0

-0

5

2.5

5

7.5

Unknown Unknown Unknown Unknown Unknown Unknown Unknown 0 0 0 Unknown0 Unknown0 Unknown 0Unknown Unknown 0 0 0 0 0 Unknown 0 Unknown0 Unknown 0 Unknown 0Unknown Unknown 0

Unknown Unknown Unknown Unknown Unknown Unknown Unknown Unknown Unknown Unknown Unknown Unknown Unknown Unknown

Unknown

0

Unknown Unknown Unknown

Fig. 6. (Cont.). m. Quartz (wt %). n. K-feldspar (wt %). o. Plagioclase (wt %). p. Biotite + phlogopite (wt %). q. Magnetite (wt %). r. anhydrite + gypsum (wt %). s. Amphibole (wt %). t. Kaolinite (wt %). u. Muscovite + illite (wt %).

610

LEYS ET AL.

features to be described separately. In addition to describing their alteration and mineralization, details of gold deportment and geochemistry as well as a summary of fluid inclusion results are provided. Hydrothermal alteration Previous alteration studies focused on the open pit, such as that by Paterson and Cloos (2005b) at elevations above 3,700  m where they mapped a ~1-km-diameter zone of quartz-biotite-magnetite–K-feldspar veining and alteration that surrounds a central 500-m-diameter intense quartz stockwork, all cut by much fresher Late Kali dikes. Surrounding the potassic alteration is a concentric zone of texture-destructive sericite-pyrite ± anhydrite alteration, up to 700 m wide, that overlaps the outer 100 to 200 m of the inner biotite + magnetite zone and extends to the edge of the Grasberg Igneous Complex, except where relict zones of epidote-chlorite ± albite alteration occur. Sapiie and Cloos (2013) complemented this work by defining ring-like domains of different vein types, particularly the outer limit of 1 and 5% quartz veining and the inner limit of pyrite ± quartz ± anhydrite veins with sericite selvages. This simple alteration pattern in which a tall inner cylindrical core of higher temperature potassic alteration and stockwork veining is surrounded by texturally destructive, lower temperature phyllic alteration extends from the top of the open pit to about 3,500-m elevation. Below this, however, the pattern is more complex due to the Gajah Tidur porphyry system, with a dome-shaped stockwork in the deep southwestern part of the Grasberg Igneous Complex that is centered on the upper part of the Gajah Tidur intrusion and is quite separate from the stockwork zone centered on the Main Grasberg (Fig. 4b). Quantitative XRD-based mineralogy: Since MacDonald and Arnold’s (1994) paper on geochemistry and mineralogy, ~700 km of core has been drilled at Grasberg (Fig. 6b), most of which was assayed in 3-m intervals for Cu, Au, Ag, Mo, As, Zn, and Pb. Recently, to significantly improve confidence in geometallurgical ore characterization and, in particular, to obtain accurate estimates of pyrite content in future mill feed, PT Freeport Indonesia analyzed assay pulps throughout the Grasberg block cave and surrounding deposits for multielement geochemistry and mineralogy by X-ray diffraction (XRD). These XRD results are more accurate than the calculated mineral content from geochemical analyses reported by MacDonald and Arnold (1994) and more representative than thin section-based estimates reported by Paterson and Cloos (2005a, b). More than 4,000 Grasberg diamond drill hole assay pulps spaced at approximately 50 m within the to-be-mined central and lower parts of the deposit, below ~3,500-­elevation (Fig. 6c), were analyzed for 65 elements by a commercial laboratory, and the data used to constrain XRD mineralogy determined at Freeport-McMoRan’s technical center in Arizona. XRD detection limits are 0.3% for sulfides and approximately 2% for other minerals, achieved by preparing samples by micronizing, Rietveld refinement techniques and cation exchange capacity methods for clay minerals, together with careful quality control by in-house experts. Potassic alteration and stockwork veining: Gajah Tidur and Main Grasberg intrusions as well as surrounding Dalam units were originally intensely K-feldspar and biotite altered, but

potassic alteration is now mainly preserved in and around the Main Grasberg due to only localized overprinting by lower temperature assemblages (Fig. 4b). In Dalam rocks, especially those strongly affected by both Gajah Tidur and Main Grasberg porphyry alteration, potassic alteration is often texture destructive and protolith identification is difficult. The Early Kali dikes are weakly potassic altered and the Late Kali dikes mainly propylitic altered, and both contain abundant relict plagioclase (Fig. 6o). K-feldspar alteration occurs as fine-grained replacement of porphyry intrusion groundmass minerals and alteration of plagioclase and igneous K-feldspar phenocrysts, thus Figure 6n represents the sum of igneous and alteration K-feldspar. K-feldspar is most abundant in an annular zone just outside the Main Grasberg, locally exceeding 40% in the surrounding Dalam rocks where it replaces almost all the plagioclase. Trace amounts of andalusite occur with K-feldspar in Main Grasberg rocks at the center of the Grasberg Igneous Complex, around 3,800-m elevation (Patterson and Cloos, 2005b). Below 2,400-m elevation the Gajah Tidur porphyry is pervasively altered to biotite and K-feldspar where it has escaped later phyllic alteration. Here, however, XRD results only average 10 to 20% K-feldspar because porphyry wall rocks are diluted by the intense quartz stockwork that locally comprises >75% of the rock volume (Fig. 6m). Biotite alteration (Fig. 6p) occurs throughout the groundmass of Dalam rocks as small shreddy clots that give them a spotted appearance, replacements of mafic phenocrysts, and as coarse aggregates along quartz vein margins. Within and surrounding the Main Grasberg quartz stockwork, biotite content is 5 to 10% and decreases outward as sericite becomes more dominant; however, the Gajah Tidur stockwork has a much lower biotite content of 1 to 2%. Magnetite (Fig. 6q) data are much more extensive than other minerals because its abundance is calculated from assay pulp magnetic susceptibility measurements calibrated from XRD results. Substantial magnetite occurs in a 20% quartz veins (Figs. 4b, 6m) has a cylindrical shape that widens from 200 m in diameter at surface to a maximum of 600 m at ~3,150-m elevation and then narrows before being wedged apart by the Late Kali dike at 2,750-m elevation. The stockwork coincides with the highest grade ore (Fig. 4c, d) and contains an average of 30% quartz veins and 5% magnetite (Fig. 5g, h), although locally the former can reach 75% (see below). Magnetite ± quartz occurs both in veins (Fig. 5h) and disseminations that predate chalcopyrite ± bornite deposition (Pollard et al., 2005). Multistage quartz and quartz-magnetite stockwork veins display variable textures including laminated, massive, fine to coarse grained, and with center-line vugs, and in places veins are sheeted and trend northwest parallel to the Kali dikes. Most of the copper sulfide mineralization that overprints the Main Grasberg stockwork is in crosscutting chalcopyrite and bornite veins (Fig. 5g, h), but sulfides commonly fill reopened quartz veins. Anhydrite (Fig. 6r) occurs as a fine-grained replacement of groundmass or feldspar phenocrysts, in vein fillings alone or with quartz and sulfides, and as very thin, sheeted veinlets. Strong anhydrite alteration is common within Dalam rocks



GRASBERG Cu-Au-(Mo) DEPOSIT: PRODUCT OF TWO OVERLAPPING PORPHYRY SYSTEMS

surrounding the Main Grasberg intrusion, exceeding 15% in the wide potassic alteration zone, but within the intrusion anhydrite averages 2 to 3%. Anhydrite also occurs in veins within phyllic-altered zones described below and is usually inversely proportional to sericite abundance. However, supergene processes have completely removed anhydrite within the Grasberg Igneous Complex from surface to a maximum depth of 800 m (Fig. 4b). Actinolite (Fig. 6s) is another mineral that occurs within potassic-altered Dalam rocks where it reaches 10%. It is also common in skarns, both in country rocks adjacent to the Grasberg Igneous Complex as well as xenoliths found deep within the Dalam Diorite and is particularly abundant in the deep skarn on the northeast side of the Grasberg Igneous Complex. The Gajah Tidur porphyry also has well-developed potassic alteration characterized by an intense but largely low-grade quartz stockwork (Figs. 4b, 5e, 6m) that overprints the contact of the Gajah Tidur and Dalam units from below 2,400- up to 2,800-m elevation and contains 30 to 90% quartz veins. A few historic deep drill holes intersected the margins of this stockwork, which MacDonald and Arnold (1994) correctly interpreted as an older and deeper Dalam-stage porphyry system. Although copper and gold are very low in the heart of the Gajah Tidur stockwork, some quartz veins contain molybdenite (Fig. 5f) and resemble the B veins at San Salvador (Gustafson and Hunt, 1975). Above the most intense stockwork, these B veins and crosscutting anhydrite-molybdenite veins are locally abundant enough to form a ~750- × 260-m zone at 3,150-m elevation grading 500 to 1,000 ppm Mo (Fig. 6g). The igneous host rocks to the Gajah Tidur stockwork contain no plagioclase (Fig. 6o) because it was replaced by K-feldspar, and in contrast to the Main Grasberg stockwork the Gajah Tidur stockwork contains no magnetite (Fig. 6q). Phyllic alteration: An annulus of phyllic (quartz-sericite-pyrite ± anhydrite) alteration surrounds the core of Main Grasberg stockwork and potassic alteration (Figs. 4b, 6l, u). The transition from potassic to phyllic alteration is marked by sparse pyrite ± quartz veins with thin sericite (muscovite and illite) selvages that overprint potassic-altered rocks, and these veins became increasingly abundant away from the Main Grasberg intrusion until the rock mass is pervasively replaced by texturally destructive phyllic alteration. Pyrite and sericite increase in concert toward the margins of the Grasberg Igneous Complex, culminating in the massive replacement pyrite zone of the Heavy Sulfide Zone. Above the Gajah Tidur quartz vein stockwork, phyllic alteration is extensively developed, with sericite reaching 40% near 3,400-m elevation. Within the Gajah Tidur stockwork, all wall rocks between the stockwork veins are pervasively overprinted by sericite and pyrite, and pyrite veins commonly cut and offset the stockwork quartz veins (Fig. 5f). We have been unable to identify features or crosscutting relationships that distinguish phyllic alteration related to Gajah Tidur versus that related to the younger Main Grasberg hydrothermal systems. However, the abundance, intensity, and depth extent of this alteration in the southwest part of the Grasberg Igneous Complex compared to other quadrants suggests it is predominantly related to Gajah Tidur rather than from the later Main Grasberg porphyry system, although some overprinting undoubtedly occurred, especially at higher elevations. Pervasive texture-destructive

611

quartz-sericite-pyrite ± kaolinite alteration occurs deep in the Gajah Tidur intrusion to 2,000-m elevation and extends up to the open pit where pit slope angles are shallowed to accommodate these weaker phyllic-argillic-altered rocks. The upper part of this zone is demonstrated by areas with >2% kaolinite around 3,400-m elevation (Fig. 6t). Near the edges of the Grasberg Igneous Complex, quartz veins are rare but quartz content is high due to silicification associated with the intense quartz-sericite-pyrite alteration. The fluids that caused phyllic alteration deposited massive pyrite ± minor chalcopyrite-bornite-covellite-­sphalerite-galena of the Heavy Sulfide Zone (Figs. 4a, b, 6l) where they encountered carbonate wall rocks around the upper Grasberg Igneous Complex margins or replaced magnetite skarn around the edges of the lower Grasberg Igneous Complex. Advanced argillic alteration: MacDonald and Arnold (1994) reported that kaolinite is common in the upper parts of the Grasberg Igneous Complex (Fig. 6t) where it is associated with sericite. This association continues below 3,400-m elevation although kaolinite is much less abundant. The related advanced argillic minerals pyrophyllite and alunite are detected in 2% kaolinite (Fig. 6t) occurs in a hydrothermal breccia that contains fragments of unmineralized Early Kali as well as quartz-veined and mineralized Dalam. The breccia is cut by native sulfur veins that are in places accompanied by anhydrite, covellite, and pyrite, and these mark the final stage of Grasberg hydrothermal activity because the Early Kali is cut nearby by the unbrecciated and unveined Late Kali dike. Propylitic alteration: Epidote-chlorite-albite alteration of plagioclase and mafic phenocrysts in Dalam units (Paterson and Cloos, 2005b) occurs only in a narrow band just inside the Grasberg Igneous Complex contact above 3,800-m elevation. Elsewhere within the Complex propylitic alteration is very limited except for chlorite-hematite alteration at the margins of the Early Kali where it breaches the Complex margin. Skarn and marble: The Kais and Faumai Formation carbonate wall rocks to the upper Grasberg Igneous Complex are commonly altered to calc-silicate skarn up to a few meters thick. However, within the underlying Waripi Formation and especially its lower, quartz sand-rich and dolomitic carbonates the skarn widens to >100 m (Fig. 4b) and contains both calc-­ silicate and magnetite skarn with anhydrite (Fig. 6q, r; Leys et al., 2012). Hydrous minerals, such as actinolite, phlogopite, and chlorite, are common retrograde phases in mineralized Grasberg Igneous Complex skarns and often replace garnet and diopside, whereas forsterite generally alters to serpentine and talc. The well-developed skarns at the margins of the Complex are clearly visible in mineral maps of amphibole (actinolite; Fig. 6s) and biotite/phlogopite (Fig. 6p). On the southwest side of the Complex these Waripi-hosted

612

LEYS ET AL.

skarns form the eastern edge of the giant Kucing Liar skarn, and the same formation is also the principal host to the giant Ertsberg-related skarns (Leys et al., 2012). Nevertheless, ­ while significant Waripi-hosted calc-silicate skarn with abundant phlogopite, anhydrite, and actinolite occurs deep on the northeast side of the Grasberg Igneous Complex (Figs. 4b, 6p, r, s), mineralization here is weak, presumably because this area was not accessed by Gajah Tidur- or Main Grasberg-­ related mineralizing fluids. Carbonate wall rocks around the edge of the Complex are marbleized to 250 m beyond the contact (Fig. 4b) but this alteration extends farther around major faults that are close to the Complex margin, particularly on its southwest side. Mineralization Copper, gold, and molybdenum metal distribution (Figs. 4c, d, 6d-g) clearly outlines the two overlapping porphyry systems: (1) the Gajah Tidur porphyry system has a broad, inverted cup-shaped, high-grade copper zone containing weak gold mineralization, whose upper part contains high-grade molybdenum, overlying a very weakly mineralized, domal quartz stockwork (Figs. 4b, 6m) centered on and overprinting the upper part of the Gajah Tidur intrusion; and (2) the Main Grasberg porphyry system comprises a vertically extensive, cylindrical column of high-grade copper and gold mineralization centered on the Main Grasberg and its quartz-magnetite stockwork. Silver is strongly correlated with copper in both systems (Fig. 6f) and is commonly associated with minor lead, zinc, and arsenic mineralization within both the Heavy Sulfide Zone and proximal Grasberg Igneous Complex skarns (Fig. 6h). The Early Kali dikes are well mineralized at their contacts with high-grade Main Grasberg and Dalam-hosted ore, but grades drop rapidly within Early Kali away from this contact and it is only weakly mineralized where it is cut by the barren Late Kali intrusion. From these relationships we interpret that the Early Kali is a late intermineral intrusion emplaced during the waning stages of Main Grasberg-stage porphyry mineralization and that mineralization ceased before the Late Kali was emplaced. Main Grasberg-related mineralization principally comprises chalcopyrite veins with associated digenite and rare chalcocite, but bornite increases relative to chalcopyrite with depth (Fig. 6g, h, i, j). Pyrite (Fig. 6l) is virtually absent within the potassic-altered central portion of the orebody and only becomes significant within phyllic alteration toward the edge of the Grasberg Igneous Complex as well as in the Heavy Sulfide Zone and surrounding skarns. Covellite (Fig. 6k) is associated with pyrite and native sulfur in minor, very late-stage (postEarly Kali) veins and structurally controlled breccias that cut high-grade ore within and around mineralized Early Kali. The potassic-altered core of the Gajah Tidur stockwork is overprinted by intense phyllic and subordinate advanced argillic alteration above 2,400-m elevation (Fig. 4b-d) and is weakly mineralized except toward and outboard from its margins where disseminated and quartz vein-hosted chalcopyrite becomes increasingly abundant. Gold grades are low throughout Gajah Tidur and no bornite is found even in the deepest drill holes, but chalcopyrite is overprinted by lesser pyrite and covellite (Fig. 6i, k, l) within phyllic-altered rocks. The highest molybdenum grades (Fig. 6g) occur above the apex

of the most intense Gajah Tidur quartz vein stockwork, with two stages of molybdenite mineralization: (1) early anhydrite-­ molybdenite ± quartz-chalcopyrite veins (Fig. 5f) that postdate most quartz-only stockwork veins and contribute most of the molybdenum in Gajah Tidur; and (2) minor molybdenite associated with pyrite-covellite mineralization and related phyllic and advanced argillic alteration. Gajah Tidur and Main Grasberg porphyry system characteristics are described and compared in Table 2. In summary, in the Grasberg Igneous Complex above 3,700-m elevation, the Main Grasberg-related, high copper-gold mineralization system dominates and is associated with a quartz-magnetite stockwork. In the southwestern Grasberg Igneous Complex below 3,400-m elevation, the older Gajah Tidur copper-­ molybdenum system occurs in the upper parts of and above a large, low-grade quartz stockwork that was overprinted by phyllic alteration. The general lack of gold, bornite, and magnetite and presence of significant molybdenite in Gajah Tidur mineralization compared to Main Grasberg indicates that their hydrothermal fluids may have had different chemistries. Similarly, the distinct geometries of their host intrusions, stockworks, and mineralization suggest that their magmas and associated hydrothermal fluids were emplaced under contrasting structural conditions. Although not the subject of this paper, it is worth mentioning the spatial and temporal relationship of the Gajah Tidur porphyry system with the adjacent giant Kucing Liar skarn (Fig. 4) that extends >1 km from the edge of the Gajah Tidur low-grade stockwork into the adjacent Waripi and other sedimentary rocks (Widodo et al., 1999). Like Gajah Tidur mineralization, Kucing Liar contains significant covellite and molybdenite, but unlike Gajah Tidur it contains significant gold mineralization especially around the Idenberg 1 fault at the southwest edge of the deposit (Fig. 4d). Copper in Kucing Liar (Fig. 4c) is concentrated in two zones, one adjacent to Gajah Tidur and the other around the Idenberg 1 fault. Based on these spatial relationships, we infer that some of the Kucing Liar skarn alteration and subsequent copper and molybdenum mineralization is related to the adjacent Gajah Tidur porphyry system and some alteration, some copper, and all the gold to a different event whose hydrothermal fluids ascended along the Idenberg 1 fault. Supergene effects MacDonald and Arnold (1994) reported that copper was leached by meteoric water to about 30 m below surface, and that minor supergene sulfides coat hypogene sulfides down to 100-m depth. Oxide copper minerals are absent from the thin leached capping, but gold grades appear to be unmodified. Supergene processes mainly affect anhydrite, which within the Grasberg Igneous Complex is completely leached out in a bowl-shaped zone that extends to 3,400-m elevation (Fig. 4b). Within the leached zone, rocks are vuggy and highly disaggregated because up to 30% of their volume was removed, as indicated by the up to 30% anhydrite content of Dalam rocks immediately below the leached zone (Fig. 6r). Pockets of leached rocks also occur below the main leaching front, where meteoric water penetrated along faults or lithologic contacts and dissolved anhydrite up to 1.3 km below surface; this sulfate-enriched ground water may have deposited the gypsum



GRASBERG Cu-Au-(Mo) DEPOSIT: PRODUCT OF TWO OVERLAPPING PORPHYRY SYSTEMS

613

Table 2. Summary of Gajah Tidur and Main Grasberg Porphyry System Characteristics, Including Estimated Contributions to Overall Grasberg Metal Endowment Porphyry system

Porphyry intrusion composition and texture

Ages of intrusion, alteration, and mineralization

Intrusion geometry and dimensions

Stockwork characteristics

Gajah Tidur Coarse-grained Intrusion: 3.4 ± 0.05 Ma (zircon U-Pb) Dike elongated NW-SE at NW-SE-trending elliptical zone   porphyritic quartz Alteration: 3.33 ± 0.12 (biotite Ar-Ar)   apex, probable SW dip,   of sheeted to orthogonal sets of   monzonite with qz-kf- Mineralization: 3.36 ± 0.01 and 3.41 ±   geometry undefined at   gray translucent quartz veins   bi phenocrysts   0.03 Ma (molybdenite Re-Os)   depth; 100-m-wide   (±molybdenite above 2,800-m   near-apex, widening at   elevation) centered on upper   depth but poorly defined;   part of Gajah Tidur porphyry;   top at 2,700 m, base at   zone of >50% quartz veins is   1,200 m high   550 × 375 m at 2,600-m Base undefined due to   elevation   no deep drilling Main Grasberg Fine- to medium- Intrusion: 3.22 ± 0.04 (zircon U-Pb) Conical-to-cylindrical plug Cylindrical zone of milky quartz   grained, porphyritic Alteration: 3.02 ± 0.03 and 3.07 ±   plunges 80° to SW; 800 ×   ± magnetite veins overprints   monzodiorite with   0.01 Ma (biotite Ar-Ar)   500 m at 4,000 m, 400 ×   and surrounds Main Grasberg   pl-bi-hb(acicular) Mineralization: 3.09 ± 0.01 Ma   250 m at 3,000 m; top at   intrusion at 3,900-m elevation,   phenocrysts   (molybdenite Re-Os)   4,200 m, base at 5% quartz veins is   >1,500 m high; base poorly   500 × 400 m and >1% quartz   defined, largely cut out by   veins ca. 1 × >1 km; quartz   Kali dikes   veined zones trend NW-SE as   do sheeted veins locally Porphyry system Alteration and mineralization Mineralization geometry and dimensions

Estimated metal contribution to Grasberg deposit1

Gajah Tidur 1. Early chalcopyrite and molybdenite in Dome-shaped zone of Cu mineralization   upper part of (and above) largely barren   surrounding barren quartz vein stockwork;   quartz vein stockwork,   Mo mineralization mainly in upper part of 2. Later covellite-pyrite ± enargite-   the Cu zone; 1% Cu grade shell maximum   (±molybdenite) associated with phyllic and   700 × 400 m at 3,100-m elevation   minor advanced argillic alteration   0.05% Mo grade shell maximum  750 × 260 m at 3150 m elevation

Cu: 30% Au: 10% Mo: 75%

Main Grasberg 1. Inner zone of cp-bn-(chalcocite-digenite- Cylindrical, vertically extensive (>1.5 km) of   covellite) overprints stockwork and potassic   Cu-Au mineralization centered on the   alteration,   Main Grasberg intrusion; 1% Cu grade shell 2. Outer zone of py-covellite ± enargite   maximum 800 × 600 m at 3,550-m elevation   associated with phyllic and minor advanced Concentric grade shells in level plan with   argillic alteration   radial extensions along interpreted structures

Cu: 70% Au: 90% Mo: 25%

Abbreviations: bi = biotite, hb = hornblende, kf = k-feldspar, pl = plagioclase, qz = quartz 1 Gajah Tidur and Main Grasberg intrusion porphyry contribution to overall Grasberg metal content is estimated from elevation and grade: below 3,600 m: Cu >0.25%, Au 0.5 g/t to Main Grasberg intrusion; above 3,600 m: Cu >0.25% and all Au and Mo is assigned to Main Grasberg porphyry system and none to Gajah Tidur

that commonly fills late extensional fractures and faults in the central part of the deposit. Geochronology Grasberg igneous biotite 40Ar/39Ar and zircon U-Pb ages, Grasberg and Kucing Liar hydrothermal mica and actinolite 40Ar/39Ar ages as well as molybdenite Re-Os ages from both deposits are listed in Table 3 and shown in Figure 7. Grasberg 40Ar/39Ar ages were reported by Pollard et al. (2005) and Pollard (unpub. report, 2009), Kucing Liar 40Ar/39Ar ages by New (2006), zircon U-Pb ages by Trautman (2013), Cocker (2016), and Wafforn (2017), and molybdenite Re-Os ages by Mathur et al. (2005) and Pollard (unpub. report, 2009). Fourteen samples of Gajah Tidur intrusive rocks have zircon U-Pb ages of 3.57 to 3.33 Ma, similar to the age range of 29 Dalam stage rocks. Five samples of undifferentiated dikes that cut Gajah Tidur and surrounding Dalam units also

have ages that lie within the age range of Dalam rocks. Sixteen Main Grasberg intrusion samples have ages of 3.31 to 3.07 Ma, with a composite age (n = 6) of the Trautman (2013) and Wafforn (2017) samples with 50% of them were halite bearing, with the main types being L + V + halite (39%), L + V (37%), and L  + V + multidaughter crystal (18%). The last multiphase inclusions are believed to represent the trapped supercritical fluid exsolved from magma at depth. However, between ~3,000- and 2,900-m, halite-bearing inclusions comprise >80% of the population, and at higher elevations vapor-rich inclusions are more abundant (Harrison, 1999). Fluid inclusions in quartz associated with potassic alteration deep within the Grasberg Igneous Complex have homogenization temperatures of 520° to >700°C (mean 630°C) and salinities of 56 to 84 wt % NaCl equiv (mean 68 wt%). Fluid inclusions associated with phyllic alteration in the Grasberg Igneous Complex have lower homogenization temperatures, averaging ~500°C and salinities of 37 to 69 wt % (avg 44 wt %) NaCl equiv (Baline, 2007). Quartz veins around the margins of the Grasberg Igneous Complex typically have halite-bearing fluid inclusions that homogenize at ~400° ± 100°C. Late-stage quartz veins that cut copper-gold ore contain liquid-vapor inclusions that homogenize at ~300° ± 50°C and contain ≤10 wt % NaCl equiv fluids. Late anhydrite that fills voids in Grasberg Igneous Complex sulfide veins contains abundant liquid + vapor inclusions that homogenize at ~250°  ± 20°C with 8 to 11 wt % NaCl equiv fluids. Fluid inclusions thus exhibit a consistent outward decrease in homogenization temperatures and salinities as hydrothermal fluids radiated from a source in the center of the Grasberg Igneous Complex. Inclusions in crosscutting veins also show significant temporal decreases in fluid temperatures and salinities as the hydrothermal system decayed and influx of meteoric fluid increased. SEM-CL patterns in Grasberg quartz veins record cooling of the hydrothermal fluids from submagmatic temperatures, through dissolution events, to repeated crosscuts by younger veins related to multiple tectonic, intrusive, and hydrothermal fluid events. Fractures in Grasberg Igneous Complex quartz veins above 3,000-m elevation contain fluid inclusions recording temperatures from 500° to >700°C that trapped exsolved Na-K-Ca brines and vapor-rich fluids with high concentrations of copper and other base metals. Supercritical fluids are locally trapped in fluid inclusions below 3,000-m elevation, and all copper sulfides postdate quartz stockwork veining and were thus deposited after fluid temperatures had dropped. Mernagh and Mavrogenes (2019) carried out the most recent study of Grasberg fluid inclusions from Main Grasberg stockwork quartz veins at 3,800-m elevation. They reported

617

that the veins exhibit crack-seal textures and contain many different types of inclusions, including K-feldspar-rich silicate melt, sulfide melt, virtually water-free salt melt, and coexisting hypersaline and vapor-rich types. The hypersaline fluid inclusions have high copper contents (up to 6.3 wt %) but display unrealistically high homogenization temperatures of up to 1,300°C that they attributed to entrapment of either immiscible silicate melt, hypersaline fluid and vapor or cumulates. They proposed that the melt and fluid inclusions record cycles of transitory, high-temperature (>700°C) hydrofracturing, melt and fluid release, and vein formation in a cooler (500°–600°C) background host-rock thermal regime. Genetic Model and Discussion The formation of the supergiant Grasberg copper-gold-(molybdenum) deposit has been the subject of considerable scientific debate, especially the origin, abundance, and grade of metals and particularly of gold (Mathur et al., 2005). Since the detailed geology of the Gajah Tidur copper-(molybdenum) porphyry was only recently documented, research to date has focused almost exclusively on the Main Grasberg porphyry and thus we discuss here only its origin and controls. It is unclear what changes occurred in structural, magmatic, and hydrothermal conditions so that these two very different porphyry systems were formed close to one another within 14,000 km of drill core from >5,000 drill holes. These workers characterized the porphyry Au-Cu and associated Cu-Au-Fe skarn deposits (Green, 1999; Close, 2000; Harper, 2000; Holliday et al., 2002; Wilson, 2003; Wilson et al., 2003, 2007a, b; Forster et al., 2004; Finn, 2006; Reynolds, 2007; Cuison 2010; Fox, 2012; Fox et al., 2015; Morrison, 2016; Harraden et al., 2019), their volcano-sedimentary host rocks (Holliday et al., 2002; Wilson, 2003; Kitto, 2005; Harris et al., †Corresponding

author: e-mail, [email protected]

2009b, 2014; Fox et al., 2015), associated high-K calc-alkalic to alkalic intrusive rocks (Blevin, 2002; Wilson, 2003; Wilson et al., 2007a; Fox, 2012), and the postmineralization history of sedimentation, burial, deformation, and uplift (Washburn, 2008). Geochronological and geochemical analyses of the intrusive rocks (Blevin, 2002; Wilson, 2003; Wilson et al., 2007a; Lowczak et al., 2018) together with paleontological data (Packham et al., 1999; Harris et al., 2014) have provided a detailed geochemical and temporal framework for understanding the timing and evolution of these porphyry deposits. This manuscript reviews the geology of the Cadia district, its porphyry and skarn deposits, and the regional- and district-scale geologic setting. We also discuss the postmineralization history that led to their disruption and preservation, and we propose an exploration model for the diverse alkalic porphyry Au-Cu deposits exposed in the district. Exploration and Mining History In 1851, prospecting at large gossanous outcrops of ironstones at Big Cadia (formerly Iron Duke) and Little Cadia (Fig. 2) led to the discovery of Au and Cu in the district. This was followed by the development of many small open-cut and underground mining and processing operations in the mid1800s through to the early 1900s. The small-scale mining of Au and Cu proved only intermittently sustainable during this period. Larger scale mining of magnetite for iron ore from

doi: 10.5382/SP.23.30; 23 p. 621

622 31º00'S

HARRIS ET AL. QLD

N

30ºS

Brisbane

NSW

Cadia district

SA 35ºS

VIC Melbourne

Nyngan

Gilgandra

40ºS

145ºE

155ºE

32º00'S

Legend Macquarie Arc Dubbo

Tomingley

VB

Cowra Trough

Calc-alkaline porphyry Au-Cu

Hill End Trough

MVB

Parkes

Alkalic porphyry Au-Cu

Peak Hill Yeoval

Mordialloc

33º00'S

VB

Boda

JN Northparkes district

Silurian basins

RG

Kingswood

Porphyry Au-Cu / Cu-Au (unspecified)

Copper Hill

Low sulfidation epithermal Au

Orange Bathurst

Cargo Cowal E39

34º00'S

Sydney ACT

West Wyalong

E43 Marsden

Cadia district

Junction Reefs

High sulfidation epithermal Au Fe-Cu-Au skarn Active mine

Bald Hill

Yiddah Estoril, Mandamah, Cullingerai The Dam Gidgingbung

147º00'E

148º00'E

0

50 miles

0

50

149º00'E

100 km

150º00'E

Fig. 1. Geology and ore deposits of the Late Ordovician Macquarie Arc of New South Wales (modified from Fox, 2012). The Ordovician Molong volcanic belt is one of the four segments of the dismembered oceanic arc. These volcanic belts formed together as part of one arc and were disrupted during rifting of the Hill End and Cowra troughs. Middle to Late Ordovician rocks of the Junee-Narromine volcanic belt preserve the center of the volcanic arc, where basaltic andesitic and andesitic stratovolcanoes were produced by suprasubduction zone magmatism (Simpson et al., 2007). The Molong volcanic belt comprises multiple packages of volcanic and volcaniclastic rocks, including the Forest Reefs Volcanics, host to the Cadia district porphyry Au-Cu deposits. The arc-related andesitic succession is dominated by thick sequences of volcano-sedimentary breccias with volumetrically minor lavas (Squire, 2001). The volcaniclastic rocks conformably overlie and interfinger with fine-grained turbidites deposited in a restricted sedimentary basin east-southeast of an arc-like volcanic succession. A thicker apron of deep marine turbidites and fine-grained volcaniclastic rocks filled a more extensive basin to the east of the Cadia district, cropping out today in the Rockley-Gulgong volcanic belt. Note that the Kiandra volcanic belt, the fourth component of the Macquarie Arc, is located to the south of the map area. Inset map in the top right corner shows the location of the Cadia district in eastern Australia. Abbreviations: ACT = Australian Capital Territory, JNVB = Junee-Narromine volcanic belt, MVB = Molong volcanic belt, NSW = New South Wales, QLD = Queensland, RGVB = Rockley-Gulgong volcanic belt, VIC = Victoria.

Big Cadia occurred during the first and second world wars (Wood and Holliday, 1995; Wood, 2012a, b). Renewed exploration at Cadia by Pacific Copper with various joint venture partners, including Homestake Ltd., in the 1970s and early 1980s focused mainly on Big and Little Cadia, but expanded to include core drilling at Cadia Quarry, Cadia Hill (then called Marrs Forest), and Cadia East (then called Sulphide Lode). Significant Au anomalies were detected at Cadia Hill

in soils and shallow diamond drilling, including 96 m @ 0.6 g/t Au (drill hole HC8) and 97 m @ 0.95 g/t Au (drill hole PC406; Fig. 2; Wood, 2012a), although the porphyry potential of the district remained effectively unrecognized. The potential for significant porphyry Au-Cu discoveries at Cadia was finally realized in 1991 to 1992 after Newcrest Mining Limited acquired the Cadia exploration licenses in 1990 (Wood and Holliday, 1995). Big Cadia and Little Cadia



ALKALIC PORPHYRY Au-Cu DEPOSITS, CADIA, NSW, AUSTRALIA

623

Table 1. Alkalic Porphyry and Skarn Deposits of the Cadia District, NSW Deposit

Mineralization style

Grade and tonnage data1

References

Ridgeway Intrusion-centered, volcanic hosted, 155 Mt @ 0.73 g/t Au, 0.38% Cu Holliday et al. (1999), Harper (2000),   quartz-carbonate-sulfide stockwork veins   Holliday et al. (2002), Wilson (2003),   Wilson et al. (2003, 2007a, b), Cuison (2010) Cadia Quarry Intrusion-hosted, sheeted quartz-carbonate- 83 Mt @ 0.35 g/t Au, 0.20% Cu Holliday et al. (2002), Wilson (2003),   sulfide veins, breccias and pegmatites   Wilson et al. (2004, 2007a, b) Cadia Hill Intrusion-hosted, sheeted quartz-carbonate- 427 Mt @ 0.43 g/t Au, 0.12% Cu Holliday et al. (2002), Wilson (2003),   sulfide veins   Wilson et al. (2007a, b) Cadia East Intrusion-centered, volcanic hosted, 3,000 Mt @ 0.38 g/t Au, 0.26% Cu Tedder et al. (2001), Holliday et al. (2002),   quartz-carbonate-sulfide sheeted veins   Wilson (2003), Wilson et al. (2007a, b),   Fox (2012), Fox et al. (2015) Big Cadia Massive magnetite-Cu-Au skarn in 42 Mt @ 0.38 g/t Au, 0.40% Cu Green (1999), Wilson (2003), Forster et al. (2004)   calcareous volcanic sandstone Little Cadia Magnetite-Cu-Au skarn in calcareous 8 Mt @ 0.30 g/t Au, 0.40% Cu Wilson (2003), Forster et al. (2004), Fox (2012)   volcanic sandstone 1Resource

data: Ridgeway, Cadia Quarry (Cadia Extended) and Big Cadia = Newcrest (2010), Cadia Hill = Newcrest (2009), Cadia East = Newcrest (2017), Little Cadia = Forster et al. (2004)

were interpreted by Newcrest geologists to be distal magnetite skarns associated with a larger (+4 km2) porphyry-related hydrothermal system (Wood and Holliday, 1995). Exploration by Newcrest defined an approximately 800- × 200-m soil anomaly of >0.4 ppm Au and >250 ppm Cu associated with outcrops of sheeted quartz veins and potassic-altered quartz monzonite at Cadia Hill (Wood, 2012a). From 1992 to 1996, exploration drilling led to the discovery of four porphyry Au-Cu deposits, starting with Cadia Hill and Cadia Quarry in 1992 and 1993. Step-out drilling along a structural corridor to the southeast led to the discovery of Cadia East (1994) and Cadia Far East (1996), now collectively known as the Cadia East deposit, the largest porphyry Au-Cu resource in Australia. The discovery of Cadia East was achieved by deep drilling southeast of Cadia Hill across the Gibb fault (Figs. 2, 3) and through tens to hundreds of meters of postmineralization Silurian sedimentary cover rocks (Wood and Holliday, 1995). A small (0.1-g/t Au intercepts in magnetic intrusive and volcanic rocks with >0.5% disseminated pyrite were also encountered in the first drill hole. Coherent alteration-metal zoning patterns, including elevated Zn concentrations, led to a deeper drilling program that intersected increasingly higher concentrations of Cu, including 118 m @ 0.10% Cu in drill hole NC371 from 396 m to the end of the hole (Fig. 2; Wood, 2012b). Subsequent deepening of drill hole NC371 to 858.4 m in early 1996 intersected a zone of chalcopyrite-bearing sheeted quartz veins grading 0.13 g/t Au and 0.40% Cu from 610 to 711 m, with several higher grade (>2 g/t Au) intersections outside the sheeted vein zone. The Ridgeway discovery hole (NC498) was the fourth of a set of 200-m step-out holes drilled around drill hole NC371 (Wood, 2012b). Drill hole NC498 was positioned based on prediction of vein directions and interpretation of the extent of porphyry-related magnetite alteration. The hole intersected two exceptionally high-grade stockwork zones, including 145 m @ 4.30 g/t Au and 1.20% Cu from 598 and 84 m @ 7.40 g/t Au and 1.27% Cu from 821 m (Wood, 2012b), triggering a shift in focus of drilling to the Ridgeway area, where continued exceptional drilling results resulted in fast tracking of approval for an underground mining operation. Large-scale mining commenced with open-pit operations at Cadia Hill in June 1998, followed by the development of a sublevel cave underground mine at Ridgeway in 2002. Cadia Hill operated until June 2012 and the pit is now used as a tailings reservoir. Ridgeway was converted to block caving in Ridgeway Deeps, then operated until March 2016, with minor production resuming for a short period during 2017. The Cadia East large-scale underground panel-cave mining operation commenced commercial production in January 2013. The orebody has an estimated mine life of >30 years. From the commencement of mining in 1998 to December 2018, a total of 11.8 Moz Au and 1.15 Mt Cu was produced from the Cadia Hill, Cadia Quarry (Cadia Extended), Ridgeway, and Cadia East mines (www.newcrest.com.au). The total mineral resources remaining for the Cadia district (including ore reserves) in December 2018 were 3,170 Mt @ 0.34 g/t Au

624

HARRIS ET AL.

Projected to surface from the 0 m absl level

Fig. 2. District geology map of the Cadia district, based on surface mapping by Newcrest Mining Ltd. between 1997 and 2002 and modified after Harris (pers. commun., 2007) and Washburn (2008). Geologic boundaries were further constrained using top-of-collar drill hole information (unpub. data, Newcrest Mining Ltd). Washburn (2008) mapped the Silurian cover rocks. Felsic intrusions and most of known ore deposits form a broad northwest trend that parallels the broader district trends evident in Figure 1 map. Systematic logging of the drill holes across the Cadia district by the senior author constrained the volcanic architecture. As has been shown elsewhere (Holliday et al., 2002; Wilson et al., 2003), the exposed stratigraphy becomes younger to the east. The Forest Reefs Volcanics vary laterally and vertically in terms of eruptive style and composition of volcanism. At Ridgeway, a narrow zone (300 m) of porphyry-style alteration and mineralization occurs in and around multiphase monzonitic stocks. By contrast, the Cadia East deposit occupies a mineralized zone 2 km long, 600 m wide, and >1,500 m in vertical extent. Here, hydrothermal alteration and mineralization is lithologically controlled at surface, becoming more structurally controlled (as stockwork and sheeted vein arrays) at depth. Cadia East 0.7 g/t Au shell projected to surface from 0 m asl level.

and 0.26% Cu for a total of 38 Moz Au and 8.3 Mt Cu (Newcrest, 2019). Regional Geologic Setting The Macquarie Arc is part of the Lachlan Orogen of eastern Australia and is exposed in four N-trending structural belts that crop out poorly over a strike length of ~450 km (Fig. 1). Paleotectonic reconstructions suggest that the belts composed an intraoceanic island-arc outboard of the eastern margin of Gondwana on the proto-Pacific plate (Glen et al., 2007a), with submarine to subaerial volcanic and intrusive rocks emplaced from the Early Ordovician to the early Silurian (Oversby, 1971; Solomon and Griffiths, 1972; Scheibner, 1973; Glen et

al., 1998; Percival and Glen, 2007; Crawford et al., 2007a, b). Repeated magmatic hiatuses and regional uplift and exhumation characterized the development of the Macquarie Arc, possibly related to the subduction of seamounts or a spreading ridge (Meffre et al., 2007). Remnants of what were once emergent volcanic islands composed of intermediate to mafic volcanic and derivative volcaniclastic rocks surrounded by fringing shallow-water carbonate rock are preserved in the Junee-Narromine and Molong volcanic belts (Fig. 1)—these are the oldest preserved segments of the Macquarie Arc (~489–474 Ma; Percival and Glen, 2007; Glen et al., 2007b). Phase 1 volcanic rocks include high-K calc-alkaline to shoshonitic basalts, basaltic andesites, and



ALKALIC PORPHYRY Au-Cu DEPOSITS, CADIA, NSW, AUSTRALIA

625

Fig. 3. Schematic NW-trending long section (looking NE) through the Cadia district. Long northwest-southeast section across the Cadia district. Section line A-A' shown in Figure 2. Skarn-bearing and bedded units in the Forest Reefs Volcanics (FRV) defined the uppermost parts of the Ordovician stratigraphy. Lower portions of the Forest Reefs Volcanics are dominated by polymict volcaniclastic breccias and conglomerates and are underlain by laminated volcaniclastic siltstones of the Weemalla Formation. The Ridgeway and Cadia intrusive complexes intruded the Ordovician stratigraphy and produced porphyry Cu-Au mineralization in the district (Wilson et al., 2007b). Middle Silurian marine sedimentary rocks of the Waugoola Group and Tertiary basalts unconformably overlie the Ordovician stratigraphy. Ages from Packham et al. (1999), Rickards et al. (2002), Holliday et al. (2002), and Wilson et al. (2007a). Figure modified after Harris (pers. commun., 2007) and Washburn (2008).

andesites that have chemical characteristics typical of intraoceanic arcs. These rocks were deposited on old oceanic crust, based on whole-rock geochemical and radiogenic isotope data and the lack of inherited zircons of cratonic derivation (Glen et al., 2007a; Crawford et al., 2007b; Kemp et al., 2020). The second phase of Macquarie Arc magmatism initiated in the Middle to Late Ordovician (Darriwilian to Gisbornian; ~468–455 Ma; Glen et al., 2007b, 2011) with the development of arc-wide, submarine to locally emergent, volcanic centers also flanked by shallow-marine carbonates (Percival and Glen, 2007; Crawford et al., 2007b). Reconstruction of the volcanic paleogeography indicates that widely spaced and large submarine volcanoes existed along a ~100-km-long segment of the arc (Simpson et al., 2005, 2007). Phase 2 rocks in the Junee-Narromine volcanic belt include the Goonumbla and Wombin volcanics, host to the Northparkes porphyry Cu-Au deposits (Fig. 1; Wells et al., 2020). Medium- to high-K basalts, andesites, rare dacites, and lesser tholeiitic mafic rocks with compositions typical of oceanic arc-type rocks formed eruptive centers (Crawford et al., 2007b; Simpson et al., 2007). In the Molong volcanic belt (Fig. 1), minor lavas interbedded with more laterally extensive turbiditic siltstone and sandstone sequences deposited in deeper marine environments suggest the preservation of adjoining back-arc basins. Crawford et al. (2007b) suggested that Middle Ordovician basaltic magmatism occurred at this time due to rifting of the Macquarie Arc. Phase 3 magmatism at ~455 to 450 Ma produced adakite-like, quartz-phyric dacite intrusions associated with calc-alkaline Cu-Au porphyry deposits and prospects at Copper Hill, Cargo, Kingswood, and several other localities (Fig. 1; Cooke et al., 2007; Crawford et al., 2007b, c; Glen et al., 2007b). The most economically significant magmatic cycle in the Macquarie Arc was phase 4 when latest Ordovician to early

Silurian (444–435 Ma) intrusions produced alkalic porphyry mineralization in the Cadia and Northparkes districts (Glen et al., 2007b). The Cadia porphyry Au-Cu deposits formed in the Molong volcanic belt (Holliday and Wood, 1993; Wilson et al., 2007a), and the Northparkes porphyry Cu-Au deposits formed at a similar time in the Junee-Narromine volcanic belt (Lickfold et al., 2007; Pacey et al., 2019, 2020; Wells et al., 2020; Fig. 1). The mineralized intrusive complexes at Cadia and Northparkes represent the youngest magmatic event in the Macquarie Arc and record the transition from mediumto high-K calc-alkaline and shoshonitic magma compositions. Strontium and Nd radiogenic isotope compositions imply some crustal contamination of phase 4 magmas emplaced during the Benambran orogeny relative to the more strongly mantle-derived phase 1 to 3 magmas (Crawford et al., 2007b). Crawford et al. (2007b) interpreted the mafic to ultramafic shoshonitic Nash Hill and Bushman Volcanics that crop out immediately north of Parkes to be part of phase 4 based on their unusual chemistry. These rocks are enriched in U-Th-Pb and La-Ce and have low initial eNd values, between +5.00 and –0.91, suggesting that they are lamprophyric (Crawford et al., 2007b). Phase 4 magmatism in the Macquarie Arc coincided with the early parts of the Benambran orogeny. Glen et al. (2007c) argued for two significant deformation events as part of the Benambran orogeny, ~443 Ma (late Bolidian to early Llandovery) and 430 Ma (late Llandovery). These compressional deformation events occurred during closure of the backarc basin, which caused cessation of magmatism as the Macquarie Arc collided with Gondwana (Glen et al., 2007c). Subsequent rifting of the arc was associated with Siluro-Devonian extension and transpressional deformation (Glen, 1992). The Middle Devonian Tabberabberan orogeny caused closure of

626

HARRIS ET AL.

the Siluro-Devonian basins and final amalgamation of the Macquarie Arc to Gondwana.

Sil T

District Geology Up to 2.5 km of Ordovician volcano-sedimentary stratigraphy has been intersected by deep drilling in the Cadia district (Figs. 2–4). The basal Weemalla Formation is between 1,300 and 1,500 m thick and consists primarily of laminated feldspathic siltstones and sandstones (Pogson and Watkins, 1998; Squire and Crawford, 2007; Fig. 5A, B), with ~500 m of intercalated pillow basalts (Mount Pleasant Basalt Member; Harris et al., 2014; Fig. 4). The Weemalla Formation is overlain by andesitic to basaltic andesitic volcanic, volcaniclastic, and locally calcareous rocks of the Forest Reefs Volcanics (Figs. 3–5C-H). The Forest Reefs Volcanics are the main host to the Cadia Au-Cu deposits and were erupted as a relatively low-­relief, multiple-vent submarine volcanic complex in the Late Ordovician (Harris et al., 2014). The continuity of sedimentation between the basal deep-marine turbidite deposits Basalt

Boulder conglomerate

Waugoola Group

Calcareous volcaniclastic sandstone / magnetite skarn

Forest Reefs Volcanics

Ordovician

Cadia Intrusive Complex (438 Ma)

Weemalla Formation v v v v

Mount Pleasant Basalt Member

Fig. 4. Simplified stratigraphic column for the Cadia district, showing the major Paleozoic and Tertiary rock units (modified after Washburn, 2008). The Late Ordovician calcareous volcaniclastic sandstone horizon of the upper Forest Reefs Volcanics contains Eastonian fossils (~452 Ma; Packham et al., 1999); it is the protolith for magnetite skarn at Big Cadia and Little Cadia and defines the lower marker horizon at Cadia East (Fox et al., 2015). The Cadia Intrusive Complex was emplaced around ~438 Ma (Early Silurian) based on U-Pb zircon dating (Wilson et al., 2007a). Macrofossils in siltstones of the Waugoola Group constrain Silurian sedimentation to the Late Wenlock (426–424 Ma; Rickards et al., 2001). The basal boulder conglomerate of the Waugoola Group contains clasts of Cadia Hill Monzonite and skarn, providing evidence of significant uplift and exhumation of the Cadia district between 438 and 426 Ma, prior to Silurian marine sedimentation (Green, 1999). The combined thickness of the Paleozoic section encountered by drilling in the Cadia district is approximately 2.5 km. Abbreviations: Sil = Silurian, T = Tertiary.

of the Weemalla Formation and the breccias and sandstones of the overlying Forest Reefs Volcanics (Figs. 3, 4), coupled with the predominance of sheet-like, laterally continuous debris-flow and other coarse-grained sedimentary deposits, implies submarine deposition in an active sedimentary basin, interpreted to be part of a larger basin system on the flank of an oceanic island arc (Harris et al., 2014). This basin extended over 100 km to the north and south of the Cadia district (e.g., Percival and Glen, 2007; Fig. 1). Stacked lava sequences, including hyaloclastite, massive lava flows, and their reworked equivalents, are up to 1 km thick, forming significant intra-basinal topography (Harris et al., 2009b). Proximity to volcanic vents is inferred by the presence of mafic to intermediate lava flows, cryptodomes, and subvolcanic intrusions (dikes and sills). Explosive volcanism characterized the upper parts of the Forest Reefs Volcanics, with tuffaceous rocks deposited from subaerial phreatomagmatic eruptions (Fig. 5G, H), providing permeable horizons for mineralization and alteration at Cadia East (Kitto, 2005; Harris et al., 2014; Fox et al., 2015). The upper parts of the Forest Reefs Volcanics also include local calcareous and fossiliferous sandstones and minor limestones (Harris et al., 2014; Fox et al., 2015; Figs. 4, 5E, F), consistent with a progressively shallowing basin through the Late Ordovician. Spatial relationships of observed volcanic and sedimentary facies, combined with stratigraphic thickening, show that volcanic detritus accumulated in sub-basins between volcanic landforms at Cadia (Fox et al., 2009, 2015). At Ridgeway, basin-fill facies grade laterally from thick (up to 200 m), coarse-grained volcanic breccias and conglomerates through mixed breccia and sandstone facies toward a thin (~50 m) sequence of finer grained sandstones and siltstones (Cuison, 2010). This rock package is wedge-shaped across 600 m, where the thickest point is immediately adjacent to a major NNE-trending fault. Up to 300 m of sedimentary detritus occurs in the center of elongate, NW-trending basins; laterally, the basin-fill stratigraphy thins to 500 m of altered Ordovician volcanics preserved above Ridgeway (Fig. 3).



ALKALIC PORPHYRY Au-Cu DEPOSITS, CADIA, NSW, AUSTRALIA

Fig. 6. Examples of intrusive rocks from the Cadia district. (A). Ridgeway intrusive complex: premineralization monzodiorite with euhedral plagioclase, biotite-altered hornblende, and rare euhedral K-feldspar phenocrysts in a fine-grained K-feldspar-hematite-altered groundmass. (B). P1 mafic monzonite porphyry, Ridgeway: K-feldspar, hornblende, and plagioclase phenocrysts in a fine-grained orthoclase-hematite-altered groundmass cut by quartz-magnetite veins. (C). P2 quartz monzonite porphyry, Ridgeway: plagioclase, K-feldspar, hornblende, and rare quartz phenocrysts in a K-feldspar-hematite-altered groundmass. (D). Intrusive contact between P1 and P2 porphyries, Ridgeway. Early quartz-magnetite-chalcopyrite veins in actinolite-orthoclase-magnetite-altered P1 porphyry truncated at P2 contact. P2 contains angular fragments of quartz-­ chalcopyrite-biotite veins (xenoliths). (E). P3 quartz monzonite, Ridgeway: interlocking crystals of plagioclase, hornblende, and K-feldspar, cut by thin veinlet with K-feldspar-hematite alteration halo, and cutting an early quartz- chalcopyrite vein with a K-feldspar alteration halo in the Forest Reefs Volcanics. (F). Monzodiorite, Cadia East: phenocrysts of plagioclase (weakly epidote-altered), hornblende (chlorite-altered) and K-feldspar in a sparse, fine K-feldspar-rich groundmass. (G).  Quartz monzonite porphyry, Cadia East: illite-altered plagioclase and hematite-dusted K-feldspar and chlorite-epidote-altered hornblende phenocrysts in a strongly hematite-orthoclase altered groundmass. (H) Variably red-rock-altered Cadia Hill quartz monzonite porphyry with K-feldspar-biotite-magnetite alteration and weak illite dusting of plagioclase, cut by milky quartz-chalcopyrite-pyrite vein with intense K-feldspar-hematite alteration halo. (I). Late-stage chlorite-calcite-quartz vein cutting hematite-dusted orthoclase-biotite-magnetite-chalcopyrite-altered Cadia Hill quartz monzonite porphyry. Abbreviations: act = actinolite, bi = biotite, cc = calcite, chl = chlorite, cp = chalcopyrite, epi = epidote, FRV = Forest Reefs Volcanics, hm = hematite, hb = hornblende, Kf = K-feldspar, mt = magnetite, pl = plagioclase, xeno = xenolith.

629

630

HARRIS ET AL.

A

B

Ca

di

Fo

ys

0.1 km

fa

ul

t

an

SSE

gu

llo

ng

fa

ul

t

Legend Waugoola Group (Silurian)

Ordovician rocks V

Fig. 7. Panoramic views (looking SSE) across the Cadia Hill open pit. The yellow box highlights the approximate location of panel B. (A). Silurian sedimentary rocks unconformably overlying the Cadia Hill Monzonite at Cadia Hill, with large-scale low-angle thrust faults (black lines) disrupting the western side of the Cadia Hill deposit. (B). Close-up of the high wall on the eastern side of the Cadia Hill open pit. Small-scale faults controlled Silurian Waugoola Group sedimentation and disruption of the underlying Cadia Hill Monzonite and related mineralization. Propagation of the low-angle thrust faults into the Silurian sedimentary rocks generated multiple low-angle bedding plane thrust detachments, including numerous thrust ramps and ramp anticlines (modified from Washburn, 2008).

Tertiary basalts of the Canobolas Volcanic Complex overlie the Paleozoic rocks unconformably in many parts of the Cadia district (Fig. 2; Wellman and McDougall, 1974). The Tertiary basalts cover Ridgeway where they are between 20 and 80 m thick (Figs. 2–3, 8). Deposit Geology In the Cadia district, porphyry Au-Cu mineralization is associated with composite quartz-saturated alkalic monzonite intrusions. Mineralization at Ridgeway is characterized by quartz-magnetite-sulfide vein stockworks whereas Cadia Hill, Cadia Quarry (Cadia Extended), and Cadia East are dominated by sheeted quartz-carbonate-sulfide veins (Holliday et al., 2002; Wilson et al., 2007a). Although the ore deposits vary in size, metal content, and host rocks (Table 1), their alteration assemblages are similar, characterized by subtle but mineralogically complicated Na-K-Ca metasomatic zones (Table 2). General features Hydrothermal alteration and sulfide minerals are centered on multiphase monzonitic stocks and dikes (Holliday et al., 2002; Wilson et al., 2003, 2007a; Figs. 8, 9). In general, the potassic

alteration assemblages are zoned from a core of quartz ± biotite-K-feldspar-magnetite, outward to domains dominated by K-feldspar-biotite ± magnetite (Wilson et al., 2003, 2007b; Table 2; Fig. 9). A calc-potassic alteration assemblage composed of actinolite + magnetite + biotite forms an important part of the Ridgeway alteration zone (Fig. 8) and is also present to a lesser extent at Cadia East (Fig. 9; Wilson, 2002; Wilson et al., 2003, 2007b). Distal alteration around each porphyry deposit is characterized by propylitic assemblages (Wilson, 2002; Table 2). Multiple stages of chlorite-epidote-hematite alteration, with submicron hematite dusting of feldspars, produced inner alteration domains that locally overprinted potassic alteration assemblages (Wilson et al., 2007b; Figs. 8–10). Weakly mineralized, chlorite-altered, matrix-rich breccias localized chalcopyrite-Au in Cadia Quarry and to a lesser extent at Cadia Hill. An outer halo of sodic (albite-K-feldspar-tourmaline; Fig. 10F) and propylitic alteration (epidote-chlorite-albitecalcite-­hematite) assemblages extends up to 1 km away from Cadia East (Fig. 9) but for lesser distances at Ridgeway (Fig. 8). Isolated chalcopyrite- and lesser Au-bearing magnetite-­ epidote (garnet-pyroxene) skarns and carbonate-replacement



631

ALKALIC PORPHYRY Au-Cu DEPOSITS, CADIA, NSW, AUSTRALIA

22,400mN

23,200mN B’

22,800mN

B

22,400mN

22,800mN

B’

5,800mRL 5,400mRL

5,400mRL

23,200mN

North fault

th Sou lt fa u

hin Delp fault

5,800mRL

B

5,000mRL

5,000mRL

Allana fault

4,600mRL

4,600mRL

Claudia fault

Pamela fault 200 m

A – Geology Ridgeway intrusive complex

Legend

B – Alteration Alteration assemblages

Late mineralization quartz monzonite (P3)

Upper sodic (albite – quartz – pyrite)

Syn-mineralization quartz monzonite porphyry (P2)

Outer propylitic (epidote – albite – calcite – chlorite)

Syn-mineralization mafic monzonite porphyry (P1) Pre-mineralization monzodiorite Feldspar-phyric dikes and sills Pyroxene-phyric dikes and sills

Volcano-sedimentary units Tertiary basalt

Forest Reefs Volcanics Massive to bedded volcaniclastic sandstone, siltstone Monomictic volcaniclastic lithic breccia Polymictic volcaniclastic lithic breccia

Weemalla Formation Transitional unit (sandstone, siltstone, breccia) Feldspathic sandstone and siltstone

Inner propylitic (chlorite – albite – hematite – magnetite – epidote) Late potassic (orthoclase – albite – hematite) Potassic (orthoclase – biotite – albite ± magnetite Calc-potassic (actinolite – biotite – orthoclase) Deep calc-potassic (magnetite – actinolite – albite – biotite) Calc-silicate (garnet – epidote ± calcite) Deep sodic (quartz – albite ± hematite ± pyrite) > 0.2 g/t Au > 1.5 g/t Au

Calcareous sandstone and siltstone Laminated and siliceous siltstone

Faults

Fig. 8. Cross section 11,050mE through Ridgeway (local mine grid) after Cuison (2010), with upper sodic, inner propylitic, and potassic alteration shells from Wilson (2003). Section B-B' location highlighted in Figure 2. (A). Geology, highlighting lithofacies in the Weemalla Formation and Forest Reefs Volcanics, and quartz monzonite porphyry and monzodiorite dikes of the Ridgeway intrusive complex. (B). Alteration, showing a central core of calc-potassic and potassic alteration with an outer halo of inner propylitic alteration. Sodic alteration occurs in both the deep root zone and upper, outer halo to the Ridgeway system.

Alteration

References

Abbreviations: ab = albite, act = actinolite, ap = apatite, bi = biotite, bn = bornite, cc = calcite, chl = chlorite, cp = chalcopyrite, epi = epidote, fl = fluorite, gl = galena, hm = hematite, ill = illite, mt = magnetite, ms = muscovite, or = orthoclase, plag = plagioclase, preh = prehnite, py = pyrite, qz = quartz, rut = rutile, ser = sericite, sp = sphalerite, zeo = zeolites

Ridgeway Early: qz-mt-bn-cp- Transitional: qz-cp- Peripheral: epi-chl- Deep: (act)-bi-or- Shallow: ab-chl- Peripheral: chl-ab- Harper (2000),  or-bt-(act) veins  or-(chl) veins  py-cc-qz-preh-zeo –  ab-qz-mt  qz-py-cc  hm-mt-epi;  Wilson (2003),   (cp)-(bn) Late: ser-py alteration   Wilson et al. (2003), Late: qz - py-ser-cc-   halos around   Wilson et al. (2007a),  (chl)-(cp)-(fl)  late-stage faults  Reynolds (2007),   Cuison (2010) Cadia Quarry Early: ab flooding with Transitional: sheeted Peripheral: epi-cc Early-stage: mt-ab Main-stage: bi-or Peripheral: chl-cc- Holliday et al. (2002,   bn-cp-epi stringers;   qz-sulfide ± cc veins   veinlets   (mt-destructive)   ab-hm-epi ± act   Wilson (2003),   mt stringers and Late: qz-cc-base   alteration   Wilson et al. (2004),   qz-mt ± cp veinlets   metal sulfide veins Late-stage: qz-ser-py-   Wilson et al. (2007a, b)   rut-chl halos around   veins and faults Cadia Hill Early: mt-bi-chl-cp- Transitional: qz-cc- Peripheral: preh- Early: ab, or-bi-ap, Transitional: or-epi- Peripheral: hm-chl-epi Wilson (2003),   or-qz, mt-qz-chl-cp-   cp-bn, bn-cp-chl-   cc-epi veins   also minor cp and   ab and minor qz, cc Late: ill-ms-py-qz-cc   Wilson et al. (2007a)   py-or, or-qz-bi-cp-   epi-ab and or-qz- Late: chl-bi-py, epi-   mt locally   or bi locally (primary   mt, qz-mt-or-cc-cp-   bi-cc-cp-py veins   cc-cp, ill-ms-cc-py-   mt destroyed)   bn veins   qz-sp-gl-cp veins Cadia East Early: mt-act-cp-py, Transitional: qz-cc- Peripheral: preh-cc- Early: ab, or-bi, also Transitional: or-ap-bi Peripheral: hm-chl-epi Wilson (2003),   qz-mt-cp-py-or, or-   cp-bn, and bi-sp-py-   epi-(fl) veins   minor cp and mt   (primary mt Late: ill-ms-py-qz-cc   Finn (2006),   qz-bn-cp-mt and   qz veins Late: ill-ms-cc-py-qz-   locally   destroyed)   Wilson et al. (2007a),   qz-mt-cp-cc-bn veins   sp-gl-cp veins   Fox (2012),   Fox et al. (2015),   Morrison (2016)

Deposit Mineralization

Table 2. Cadia Porphyry Au-Cu Deposits: Mineralization and Alteration Assemblages

632 HARRIS ET AL.



ALKALIC PORPHYRY Au-Cu DEPOSITS, CADIA, NSW, AUSTRALIA

Fig. 9. Cross section 15,820mE through Cadia East (local mine grid) after Wilson (2003). Section C-C' location highlighted in Figure 2. (A). Geology, highlighting lithofacies in the Forest Reefs Volcanics and quartz monzonite porphyry and diorite dikes of the Cadia East intrusive complex. (B). Alteration, showing a central core of Potassic I (biotite-orthoclase ± magnetite) alteration that transitions upward to Potassic III (biotite-tourmaline) alteration. Inner propylitic (chlorite-hematite-albite) alteration forms an asymmetric halo to the potassic alteration core, while an upper, mineralogically complex zone of Potassic IV (orthoclase), Sodic II (albite) and phyllic (illite-muscovite-pyrite) alteration overprints the upper extent of earlier potassic alteration.

633

634

HARRIS ET AL.

A

C

B

bn mt

Kf

cc

chl

qz

cc mt epi

epi

1 cm

bn

1 cm

D

1 cm

E

F

tm

bi

1 cm mt

qz

cp

Kf bn

pyy

cp p

1 cm

G

Kf Mo

1 cm

2 cm

H

I

J

epi mt pr

Kf cal

pyy fl

1 cm

1 cm

1 cm

Fig. 10. Examples of altered and mineralized rocks from the Cadia district. (A). Quartz-magnetite-bornite vein stockwork cut by epidote-calcite-quartz-chlorite vein in hematite-dusted Ridgeway quartz monzonite porphyry, Ridgeway. (B). Sheeted quartz-bornite-calcite veins in chlorite-epidote-hematite-altered Cadia Hill Monzonite porphyry. Thin (1 mm) K-feldspar halos surround the sheeted quartz veins. (C). K-feldspar-plagioclase-phyric quartz monzonite porphyry dike, deep levels of Cadia East, cut by quartz-bornite-magnetite-K-feldspar vein and thin quartz veinlets that define a stockwork. This sample lacks the hematite dusting associated with red-rock alteration, and instead has a pale pink wash of orthoclase alteration. (D). Quartz-bornite-magnetite veins in actinolite-orthoclase-altered Forest Reefs Volcanics, deep levels of Cadia East. (E). Pervasive biotite-chalcopyrite alteration in Forest Reefs Volcanics, upper levels of Cadia East. (F). Pervasive, texturally destructive albite-K-feldspar-tourmaline alteration of Forest Reefs Volcanics, upper levels of Cadia East. (G). Coarse-grained ­ K-feldspar-quartz-chalcopyrite-molybdenite-pyrite cemented pegmatite breccia with clasts of K-feldspar-and chlorite-illitealtered quartz monzonite porphyry, Cadia Quarry. (H). Disrupted quartz-magnetite-chalcopyrite veins cut by epidote-­ cemented breccia associated with hematite-dusting (inner propylitic) alteration (Ridgeway). (I). Distal outer propylitic alteration of Forest Reefs Volcanics, Ridgeway. The chlorite-altered volcanics are cut by a calcite-prehnite-pyrite-cemented breccia. Note the lack of hematite dusting; this is typical of the outer propylitic halos at Cadia. (J). Magnetite-K-feldspar-­ epidote-chlorite-fluorite retrograde skarn alteration of calcareous volcaniclastic unit in Forest Reefs Volcanics, Big Cadia. Abbreviations: bi = biotite, bn = bornite; cc,= calcite, chl = chlorite, cp = chalcopyrite, epi = epidote, fl = fluorite, Kf = K-feldspar, Mo = molybdenite, mt = magnetite, pr = prehnite, py = pyrite, qz = quartz, tm = tourmaline.



ALKALIC PORPHYRY Au-Cu DEPOSITS, CADIA, NSW, AUSTRALIA

mineralization formed in calcareous volcaniclastic horizons in the upper parts of the Forest Reefs Volcanics (Green, 1999; Forster et al., 2004; Fig. 10J). Propylitic and potassic assemblages occur together at Cadia Hill (e.g., Figs. 6H, 10B), with thin (mm-scale) orthoclase alteration halos surrounding quartz-sulfide veins in epidote-chlorite-hematite-altered quartz monzonite porphyry. Epidote occurs locally in some quartz-chalcopyrite ± bornite ± chalcocite (digenite) veins. Cadia East has well-developed inner and outer propylitic zones, and a well-developed propylitic zone is associated with skarn at nearby Little Cadia (Fig. 9B). The upper levels of Cadia East consist of disseminated, volcanic-hosted Cu-Au ore associated with early-formed biotite alteration (Figs. 9B, 10E) that has been overprinted by pervasive sericite-albite-orthoclase-pyrite-tourmaline alteration (Fig. 10F; Kitto, 2005). Irregular, pervasive domains of barren sodic alteration composed of albite-pyrite-quartz occur above and peripheral to several of the Cadia deposits (Wilson, 2002; Figs. 8, 9). Late-stage faults that locally contain carbonate-base metal veins with pervasive phyllic alteration halos are recognized in all of the Cadia porphyry systems (Wilson, 2003; Finn, 2006; Fox, 2012; Fig. 9B). Quartz-sulfide ± magnetite ± carbonate ± chlorite veins and minor breccias form the mineralized rocks throughout the district (Figs. 6B, D-I, 10A-D, G-I). The principal ore minerals are chalcopyrite, bornite, and native Au, which occur primarily in veins. Gold is intergrown with bornite (± digenite) and, to a lesser extent, chalcopyrite (Wilson et al., 2003). Ore zones in the Cadia porphyry deposits also locally contain rare Ag-Pd tellurides. Overall, the deposits are characterized by low Cu/Au ratios, in contrast to the more Cu-rich alkalic porphyry deposits at Northparkes (Cooke et al., 2007). Ridgeway Ridgeway is a small, high-grade, Au-rich porphyry Au-Cu deposit (155 Mt @ 0.73 g/t Au, 0.38% Cu; Newcrest, 2010; Table 1; Figs. 8, 11). The high-grade upper part of the deposit is hosted by K-rich volcano-sedimentary rocks of the Forest Reefs Volcanics (Fig. 8, 10I), whereas lower grade mineralization in Ridgeway Deeps is hosted by laminated siltstones and sandstones of the Weemalla Formation (Figs. 5A, B, 8). A quartz-bornite-chalcopyrite-magnetite vein stockwork (e.g., Fig. 10A) is centered in the Ridgeway intrusive complex (Wilson et al., 2003; Fig. 8), which consists of a premineralization monzodiorite (Fig. 6A), narrow, 1.7 km (Cuison, 2010; Figs. 3, 8, 11). There is a strong protolith control on alteration assemblages at Ridgeway (Fig. 8). In the roots of the deposit, laminated

635

and siliceous siltstones of the Weemalla Formation have undergone distal quartz-albite-hematite-pyrite-chlorite alteration and proximal deep-level calc-potassic alteration around the premineralization monzodiorite (Figs. 5A, B, 8). Calcareous horizons have locally undergone calc-silicate alteration to garnet and epidote (Fig. 9). At shallower levels, hydrothermal alteration assemblages in the Forest Reefs Volcanics are broadly zoned from an inner calc-potassic (actinolite-­biotiteorthoclase; Fig. 6D) and potassic (orthoclase-biotite-quartz; Fig. 6C) core centered in the Ridgeway intrusive complex (Fig. 9), outward and upward through propylitic (chlorite-­ epidote-albite-hematite-magnetite-pyrite ± calcite; Fig. 10H, I) and shallow-level upper sodic assemblages (albite-pyritequartz; Fig. 8; Table 2). Sulfide minerals are similarly zoned from the Au-rich core where bornite is more abundant than chalcopyrite, outward and upward through a Au-bearing, chalcopyrite-rich halo to an outer, unmineralized pyrite-rich domain that overlies the deposit (Wilson et al., 2003). Bornite is intergrown locally with hypogene chalcocite or chalcopyrite and has in some cases been partially replaced by covellite. Locally, free Au grains occur in quartz-magnetite veins (Wilson, 2003). Gold occurs as micron- to submillimeter-sized grains either within quartz or as round inclusions within bornite. The mineralized alteration assemblages have a limited spatial distribution, extending a few tens of meters from the intrusions (Figs. 8, 11), and are flanked by inner pyrite-­bearing and outer sulfide-deficient propylitic alteration assemblages that extend hundreds of meters away from the ore zone (Fig. 8). The known vertical extent of the mineralized system is >1,100 m (Cuison, 2010; Fig. 8). At Ridgeway, the highest Au and Cu grades are associated with calc-potassic (actinolite-biotite-magnetite) and potassic (orthoclase-biotite-magnetite) alteration assemblages (Fig. 8B). The Au and Cu grades decrease dramatically with distance from the orebody at the transition to the inner propylitic alteration domain assemblages (chlorite-albitehematite-­ magnetite-epidote with lesser calcite-prehnitepyrite-quartz-­leucoxene; Wilson et al., 2003; Fig. 10H). The inner propylitic assemblage transitions to the more distal, metal-poor outer propylitic alteration assemblage (epidotealbite-calcite-­ chlorite; Fig. 10I)—this transition coincides with the disappearance of hematite-dusted hydrothermal albite (Wilson et al., 2003). Dusting of feldspars by submicron hematite occurs in propylitic-altered rocks up to 400 m outboard of the 0.2 g/t Au and 0.2% Cu grade contours (Fig. 8B). Early recognition of the potential importance of this red-rock alteration as a near-miss indicator provided encouragement to persist with exploration drilling at Ridgeway (Holliday and Cooke, 2007). The hydrothermal footprint of Ridgeway is relatively subtle—weak to moderately developed propylitic alteration occurs within hundreds of meters of the 0.2 g/t Au grade shell (Figs. 8, 11; Wilson et al., 2003; Cuison, 2010). Potassic (K-feldspar ± biotite ± magnetite) alteration extends up to 300 m above the ore zone (Fig. 8B). Aggregates of hydrothermal magnetite can occur up to 350 m from the 0.2 g/t Au shell (Reynolds, 2007). A narrow (100 m wide) annulus of pyrite surrounds the chalcopyrite-bornite domain (Wilson et al., 2003). Beyond the potassic zone, a K-feldspar-destructive alteration zone occurs across the top and sides of the deposit.

636

HARRIS ET AL. 22

75

0m

Merkin Ft

d2

North Fault

d1

d7

d3

d4

A

d1

d2

N

d8

d6

Legend Monzonite Intrusions

d5

Late-mineralization quartz monzonite (P3)

d3 d8 22

Quartz monzonite porphyry (P2)

85

0m

Mafic monzonite porphyry (P1)

50 m

E

d7

Magnetite-rich USTs

11 2

d4

Quartz-rich USTs

N

Pre-mineralization intrusions

South Fault

0m

11 0

11 15

50

5330L

Monzodiorite

Delphin Fault

mE

10

95

0m

d5

E

E

d6

22

Pyroxene- and feldspar-phyric dikes

65

0m

N

Fault d2 d3

d1

d2

d4

B

d1 d7

d6 750m

N

d5

< 0.2 g/t Au

d8

d7

> 0.2 g/t Au

E

d3

Veins Domain boundary

22

11 15

0m

> 4.0 g/t Au

North Fault

85

N

mE

Purple Fault

11 2

d6

Rose diagram indicates strike directions and frequency of veins for each domain (10 degrees arc intervals)

0m

50

Delphin Fault

22

South Fault

d4

n = 811 number of samples

95

0m

E

Merkin Fault

10

22

mE

75

0m

11 0

22

North Fault

85

N

Merkin Ft

65

N

d1 d2 d7 d3 d6 d4 d5 d8

N

d3

0m

C

0

50 m

11 15 0m E

22

N

d8

0m

d1

d2

Delphin Ft

0m

65

d7

N

Purple Fault Purple-splay

d4 d6

95

0m E

11 25

North Fault

E

75

22

AMG North CMLG North

0m

22

Rimmers Fault

50

5255L

d5

0m

11 05 0

5100L

mE

10

d8

d5

Rimmers Fault Weem Fm FRV

Fig. 11. Plan views showing vein orientations on three underground mine levels (A: 5330L; B: 5255L; C: 5100L) from Ridgeway (modified after Cuison, 2010). Data from unpublished underground mapping by Newcrest geologists. Vein orientations summarized using rose diagrams for each structural domain defined in the inset boxes on the upper right side of each panel. Inset panels on lower right side of each panel summarize vein trends for each structural domain. Local mine grid coordinates. Abbreviations: d = domain, FRV = Forest Reefs Volcanics, Ft = fault, UST = unidirectional solidification textures, Weem Fm = Weemalla Formation.



ALKALIC PORPHYRY Au-Cu DEPOSITS, CADIA, NSW, AUSTRALIA

This zone is weakly enriched in Zn and Pb (>100 ppm Zn and >20 ppm Pb) and is associated with propylitic and locally pervasive albite-sericite alteration. Chlorite-, hematite-, and epidote-rich propylitic assemblages occur up to 500 m from the bornite-chalcopyrite-rich core. Mapping of vein distributions on three underground mine levels at Ridgeway demonstrated that vein orientations vary over a 230-m vertical interval (Fig. 11; Cuison, 2010). At the 5330 level, above the intrusive complex (Fig. 11A), sheeted veins striking northwesterly dominate the center of the ore zone and are surrounded by veins showing a range of orientations that can be broadly resolved into a radial pattern. Veins at this level are hosted largely in the volcanic and volcaniclastic country rocks, where they are distributed above the synmineralization P1 and P2 quartz monzonite porphyry intrusions. A similar distribution of veins characterizes the 5100 level (Fig. 11C), but there is also a strongly developed sheeted set of veins parallel to the synmineralization P1 and P2 monzonite porphyries, which likely reflects far field-driven fracturing in advance of the rising magmas. In contrast, on the 5255 level, where veins are localized around the tops of the P1 and P2 monzonite porphyry intrusions, east-northeast and north-northwest strikes predominate that are oblique to the dominant vein orientations at shallower and deeper levels and imply that local stresses have influenced stockwork development at this level (Fig. 11B). The 5255 level corresponds to a position near the tips of the synmineralization P1 and P2 monzonite intrusions that form a NW-elongated, dike-like composite body.

637

in the highest grade zones (>2.0 g/t Au) are characterized by bornite > chalcopyrite. There is a transition from deepseated K-feldspar-dominated early alteration (potassic I) to shallow, later stage biotite-tourmaline assemblage (potassic III; Wilson, 2003; Figs. 9B, 10E). Quartz-sulfide veins are widespread throughout Cadia East, where they typically have thin (2 km below present surface) include a weakly mineralized sodic-calcic assemblage (albite-actinolite-magnetite ± epidote ± chalcopyrite) that is cut by early, magnetite-actinolite veins and zones upward to the central domain of mineralized calc-potassic alteration (biotite-magnetite-K-feldspar-­ albite-actinolite ± chalcopyrite). Subtle relict textures show that calc-potassic alteration extended to shallow levels and may have been in part related to strata-bound disseminated mineralization associated with the potassic III alteration Cadia East domain, which is hosted by tuffaceous planar laminated volCadia East is the largest porphyry deposit in the Macqua- caniclastic siltstones, sandstones, and breccias that define an rie Arc in terms of contained metal. In December 2016, the upper marker horizon of the Forest Reefs Volcanics (Figs. Cadia East resource was reported to be 3,000 Mt @ 0.38 g/t 5G, H; 9; Fox et al., 2015). The disseminated strata-bound Au and 0.26% Cu for 36 Moz Au and 7.6 Mt Cu (Newcrest, Cu zone is approximately 400 m wide and is characterized 2017; Table 1). Cadia East has produced over 2.7 Moz Au by grades of 1.9 km in vertical extent (Fox et al., 2015). genic breccias, and conglomerates of the upper Forest Reefs Gold and Cu are hosted by the Forest Reefs Volcanics and are Volcanics. associated with a swarm of narrow alkaline quartz monzonite Late-stage potassic IV, sodic, and phyllic alteration assemand monzodiorite dikes (Wilson, 2003; Wilson et al., 2007b; blages consist of texturally destructive albite, orthoclase, Fox, 2012; Fox et al., 2015; Figs. 2, 3, 6F, G, 9). Over 70% of quartz, tourmaline, apatite, pyrite, and muscovite that overCadia East is covered by barren, postmineralization Silurian printed the disseminated strata-bound Cu mineralization and sandstones and siltstones that are locally up to 200 m thick, or older alteration assemblages (Wilson, 2003; Kitto, 2005; Finn, by Tertiary basalt (Figs. 2, 3, 9). 2006; Fox, 2012; Figs. 9, 10F). Although minor tourmaline The distribution of altered and mineralized rocks in the has been recognized throughout the vertical extent of Cadia Cadia East orebody reflects the paleohydrology of the hydro- East (Fox, 2012), pervasive, texturally destructive, bleached thermal system, which was strongly influenced by host-rock K-feldspar-albite-tourmaline alteration is largely confined to permeability and a protracted history of structural reacti- the upper parts of the stratigraphy and was localized in pervation (Fox et al., 2015; Fig. 9). At deep levels, high-grade meable tuffaceous volcanic breccia horizons below a subvolAu-Cu, typically with grades of >0.5 g/t Au, is hosted in vol- canic hypabyssal basaltic andesite sill. The assemblage extends caniclastic breccias and sandstones of the Forest Reefs Vol- regionally as a laterally extensive (several km2), broadly stracanics by sheeted quartz-sulfide veins associated with the ta-bound alteration zone. Deep-seated tourmaline-­albite-Kpotassic I alteration assemblage (biotite-K-feldspar-quartz-­ feldspar replacements and matrix-rich hydrothermal breccias anhydrite-carbonate; Wilson, 2003; Table 2; Figs. 9, 10C, D). with no associated mineralization also cut through Cadia East Shallow levels in the deposit are more Cu-rich, with sulfide (Fox, 2012)—these late-stage features may be feeders to the minerals disseminated in porous and permeable volcaniclastic shallow strata-bound feldspar-albite-tourmaline alteration units, and also hosted in sheeted quartz-sulfide veins. Veins domain.

638

HARRIS ET AL.

Although generally barren of Au and Cu, a restricted volume of albite ± sericite alteration occurs directly above the Cadia East orebody (Fig. 9). It is weakly anomalous with respect to Cu (>300 ppm) and Mo (~50 ppm), in marked contrast to its immediate surroundings of albite-K-feldspar alteration that are depleted in Cu (10–50 ppm). Although the origin of the albite ± sericite domain remains poorly understood, oxygen deuterium isotope analyses are consistent with magmatic fluids (Wilson, 2003; Fox, 2012) that dispersed laterally above the steeply dipping ore zones, with lateral flow controlled by permeable tuffaceous horizons in the Forest Reefs Volcanics. The presence of sericite implies a moderately acidic fluid and raises the possibility that albite ± sericite alteration was originally overlain, at shallower, now-eroded levels, by advanced argillic alteration typical of lithocaps (e.g., Sillitoe, 1995; Chang et al., 2011). Narrow zones of structurally controlled, hypogene kaolinite alteration could represent the roots of a more typical lithocap that was eroded from above Cadia East. Alternatively, fluids never evolved to conditions acidic enough to produce advanced argillic alteration, in which case the strata-bound K-feldspar-albite alteration zone above Cadia East might be an alkalic version of a lithocap (Holliday and Cooke, 2007). Maps of vein orientations along the 2-km northwest strike length of the Cadia East deposit show a curvilinear trend (Fig. 2), with strikes varying slightly and dips changing, but generally remaining close to vertical (Fox et al., 2015). As sheeted vein systems in porphyry Cu deposits typically reflect the regional far-field or tectonic stress state at the time of formation (Tosdal and Richards, 2001; Tosdal and Dilles, 2020), the nonuniform and curvilinear strike to the sheeted veins suggest postmineralization deformation of the orebody. Major inflections in the trend of the veins broadly correlate with known locations of the postmineralization inverted normal faults (Figs. 2, 3). Complications in the grade shells in the upper part of the western end of Cadia East have been interpreted to reflect shortening associated with the postmineralization Gibb-Cadiangullong fault system (Fox et al., 2015). Cadia Hill and Cadia Quarry The Cadia Hill (427 Mt @ 0.43 g/t Au, 0.12% Cu; Newcrest, 2009) and Cadia Quarry (83 Mt @ 0.35 g/t Au, 0.20% Cu; Newcrest, 2010; Table 1) porphyry Au-Cu deposits are localized in a complex domain where steep reverse faults juxtapose different levels of porphyry mineralization (Figs. 2, 3). Multiple, moderately dipping (30°–60°) reverse faults, including the Gibb-Cadiangullong, Cadia Extended, and Powerline faults, cut the Cadia Intrusive Complex, the large composite stock at the center of the Cadia district (Wilson, 2003; Figs. 2, 3). Both Cadia Hill and Cadia Quarry are located in the hanging wall of the Gibb-Cadiangullong fault system and are hosted primarily by the Cadia Intrusive Complex (Figs. 2, 3, 6H, I). Reverse motion along the Gibb fault thrust the deeper seated Cadia Intrusive Complex over volcanic rocks that host shallower level, intrusion-centered porphyry mineralization at Cadia East (Fig. 3). At Cadia Quarry, the Ordovician volcano-sedimentary sequence dips between 20° and 30° west-northwest. Multiple, narrow porphyry intrusions plunge eastward at approximately 65°, cutting well-bedded strata, and are preserved in

fault-bounded blocks (Fig. 3). Up to 20° of clockwise rotation is required to bring the pencil-like porphyries at Cadia Quarry to vertical orientations. Similar block rotations are likely to be required to return the mineralized thrust-dismembered segments of Cadia Hill back to their original orientations. Sheeted quartz veins host Cu and Au mineralization at Cadia Hill and Cadia Quarry (Table 2; Fig. 10B). Mineralization is restricted mostly to thrust-dismembered, medium-grained, equigranular monzodiorite to coarsely orthoclase porphyritic quartz monzonite (Figs. 6H, I, 10B) and, locally, fine-grained syenite of the Cadia Intrusive Complex. Minor mineralization occurs in narrow domains of adjacent Forest Reefs Volcanics. Alteration is dominated by propylitic assemblages, with thin potassic selvages around the sheeted quartz-sulfide-carbonate veins (Table 2; Figs. 6H, 10B). At Cadia Quarry, a small zone of high-grade Cu mineralization is associated with an unusual pegmatitic breccia (Fig. 10G), in which weakly mineralized and veined monzonite clasts are cemented by coarse-grained orthoclase, quartz, biotite, calcite, and sulfides (Wilson et al., 2004). Early-formed sodic (albite) and potassic (biotite-orthoclase) alteration assemblages are irregularly distributed. Faults that cut Cadia Hill juxtapose zones with high Cu grades (chalcopyrite and bornite rich) against pyrite-dominated rocks, the latter like those normally present along the margins of an intact porphyry Au-Cu deposit. This is particularly striking in the footwall of Cadia Hill, where 600 m of west-over-east movement placed ore-grade rocks above lower grade pyrite-rich zones (Wilson, 2002; Fig. 3). Big Cadia skarn The Big Cadia deposit crops out as the largest ironstone in eastern Australia (Figs. 2, 3, 5F). It is an elongate, faultbounded, WNW-trending, tabular (approximately 1 km long and 250 m wide), magnetite ± epidote Cu-Au skarn deposit that dips toward the southeast (Green, 1999; Forster et al., 2004; Fig. 10J). Chalcopyrite and minor Au are closely associated with bladed hematite, magnetite, pyrite, and epidote with lesser chlorite-quartz-calcite, which occur as replacements of calcareous sandstone in the Forest Reefs Volcanics. Garnet is rarely observed and has typically been replaced by retrograde calcite ± epidote (Green, 1999). Complex fault and fold relationships are visible in the historic open-cut workings, and a significant portion of the deposit was downdropped tens of meters along the PC40 fault (Fig. 2), so that only part of the deposit actually crops out. A total resource of 42 Mt @ 0.38 g/t Au and 0.40% Cu has been reported for Big Cadia (Newcrest, 2010). Little Cadia skarn The Little Cadia Au-Cu skarn prospect crops out in Copper Gully at Cadia East (Figs. 2, 3). Little Cadia is hosted by the same bedded, calcareous, volcaniclastic sandstones that host Big Cadia (Packham et al., 1999; Fig. 5E). The style of Cu-Au mineralization is similar to that at Big Cadia, with Au and chalcopyrite associated with epidote ± quartz in the interstices of bladed hematite-magnetite-pyrite aggregates that have replaced calcareous volcaniclastic sandstones (Forster et al., 2004). Replacement produced banded magnetite-­epidotecalcite retrograde skarn with an envelope of pervasive,



ALKALIC PORPHYRY Au-Cu DEPOSITS, CADIA, NSW, AUSTRALIA

hematite-bearing epidote-chlorite-calcite-pyrite alteration. Forster et al. (2004) reported a resource of 8 Mt @ 0.30 g/t Au and 0.40% Cu for Little Cadia. Timing of magmatism and mineralization Wilson (2003) and Wilson et al. (2007a) reported two ages for porphyry Au-Cu systems in the Cadia district. Ridgeway was considered to be older, with U-Pb ages of 456.9 ± 7.2 Ma obtained from the premineralization monzodiorite and 455.8  ± 4.4 Ma from the P2 quartz monzodiorite porphyry. Younger intrusion and mineralization ages were obtained for Cadia Hill, Cadia Quarry, and Cadia East. Wilson et al. (2007a) reported a zircon 206Pb/238U age of 435.9 ± 3.7 Ma, Re-Os ages of 442.9 ± 1.4 and 443.0 ± 1.5 Ma were obtained from two aliquots of a single molybdenite-bearing ore-related vein sample, and a 40Ar/39Ar age for hydrothermal sericite of 435.4 ± 1 Ma was obtained from Cadia Hill (C. Perkins, unpub. data, 1994). Kemp et al. (2020) reported similar Re-Os molybdenite ages from quartz-molybdenite-chalcopyrite veins for Cadia Hill (443.7 ± 2.2, 443.1 ± 2.2 Ma), and a U-Pb zircon age of 441.4 ± 2.9 Ma for the Cadia Hill Monzonite. For Cadia East, relatively young ages were obtained for the intrusions and molybdenite-bearing veins by Wilson et al. (2007a), although two U-Pb SHRIMP ages demonstrate that a mixed zircon population is present in the intrusions. A sample of quartz monzodiorite porphyry has a bimodal distribution of 206Pb/238U ages with an older, dominant 451.0 ± 1.4  Ma group and a subordinate 440.1 ± 4.2 Ma age population. We now interpret the older age to be inherited from the host volcanic rocks, although Wilson et al. (2007a) interpreted it to be the intrusive age. The younger zircon age population overlaps within uncertainty of a Re-Os molybdenite age of 441.8 ± 1.4 Ma for Cadia East (Wilson et al., 2007a). A quartz monzonite porphyry dike from Cadia East (NC556, 1,289 m; Squire, 2001) is dominated by zircons with a mean age of 437.1 ± 1.6 Ma, consistent with an early Silurian age for monzonite emplacement. Cuison (2010) documented additional U-Pb LA-ICP-MS analyses of zircons from Ridgeway monzonites and demonstrated a bimodal population of zircon ages, with 444 to 442 Ma ages predominant, and a subordinate older population (458–455 Ma). The older zircon grains are interpreted to be xenocrystic, most likely inherited from the Forest Reefs Volcanics and/or the Weemalla Formation. Re-Os dating of ore-related quartz-molybdenite veins from Ridgeway provided ages between 445.9 ± 2.1 and 442.8  ± 2.2 Ma (Cuison, 2010). The available data therefore suggest that all of the porphyry Au-Cu deposits in the Cadia district formed between 444 and 437 Ma, with Ridgeway forming at ~443 Ma (Cuison, 2010). Genetic Considerations Alkalic Au-Cu porphyry deposits in the Cadia district formed in postsubduction, transtensional environments after the initial accretion of the Macquarie Arc to Gondwana (Cooke et al., 2007; Glen et al., 2007b, c). Volcanism in the district ceased more than 10 m.y. before mineralization occurred. Highly oxidized and K-rich, ore-related magmas were derived from an enriched mantle source previously modified by subduction processes. Their emplacement was facilitated by deep-crustal, arc-normal transverse zones (Glen and Walshe, 1999; Glen et

639

al., 2007b), which provided permeable pathways for magma ascent to upper crustal levels (2- to 3-km depth) favorable for porphyry ore formation (Wilson et al., 2003). Rifting of the oceanic crust in a back-arc setting may have facilitated the supply of primitive mantle-derived magmas to evolving upper-crustal magma bodies, now preserved as mafic and ultramafic xenoliths and intrusions associated with mineralized porphyritic intrusions. These mafic magmas could have supplied additional metals (Cu, Au, and PGEs) and S to the Cadia mineralizing system, ingredients critical to ore formation. Keith et al. (1995) proposed a similar contribution by alkaline mafic dikes to mineralization at the giant Bingham Canyon porphyry Cu-Mo-Au deposit, Utah, and Lickfold et al. (2007) provided geochemical evidence for similar mafic magma recharge of synmineralization alkalic monzonite porphyries at Northparkes. Postsubduction emplacement of the Cadia porphyry deposits in the Macquarie Arc implies that decompressive partial melting sourced melts from subcontinental lithospheric mantle that had been preconditioned to generate oxidized, low-volume alkalic melts during arc magmatism (e.g., Richards, 2009). Only limited fluid inclusion studies have been conducted at Cadia—fluid inclusion preservation is poor due to postmineralization deformation. Wilson et al. (2003) identified halite-saturated fluid inclusions associated with calc-potassic and potassic alteration from Ridgeway that recorded homogenization temperatures between ~490° and ~415°C (Wilson et al., 2003; Table 3). Halite-homogenizing fluid inclusions imply some postentrapment modification, probably due to deformation during the Tabberabberan orogeny, preventing accurate pressure-depth estimates (Wilson et al., 2003). Based on geologic evidence, the depth of emplacement of the mineralized porphyries is interpreted to have been within 3 km of the paleosurface (Wilson et al., 2003). Oxygen-deuterium isotope analyses of white mica from Cadia East and Cadia Quarry are consistent with a predominance of magmatic water during late-stage alteration (Wilson, 2003; Finn, 2006; Fox, 2012). Strontium isotope analyses of the mineralizing intrusions yielded 87Sr/86Sri values of 0.70419 to 0.70452 for the monzonites, 0.70440 to 0.70467 for epidote, and 0.70403 to 0.70443 for tourmaline (Cooke et al., 2007; Fox, 2012; Table 3), providing evidence for primitive (mantle-derived) sources of strontium for the intrusions as well as for propylitic alteration assemblages. These data preclude any significant crustal contamination of the intrusions and discount seawater as a source of Sr during alteration (Cooke et al., 2007; Fox, 2012). Sulfur isotope compositions of pyrite, chalcopyrite, and bornite at Ridgeway, Cadia East, and Cadia Hill vary from –9.8 to +9.2‰, with most between –4 and –1‰, implying a predominance of magmatic S and oxidizing (sulfate-­dominant) fluid conditions (Wilson et al., 2007b; Table 3). Two anomalous analyses of pyrite from distal propylitic and sodic alteration yielded elevated δ34S values of 9.1 and 9.2‰ from Cadia Hill and Ridgeway, respectively, providing limited evidence for a seawater S source on the peripheries of the porphyry deposits (Wilson et al., 2007b). The S isotope data from Ridgeway, Cadia East, and Cadia Hill define systematic spatial variations from more negative values in the core of each deposit to less negative values above

640

HARRIS ET AL. Table 3. Cadia Porphyry Deposits: Fluid Inclusion and Isotopic Data

Deposit

T (°C)

Salinity

Sulfur isotopes

Strontium isotopes (87Sr/86Sr) References

Ridgeway Average 440° to 450° 53 to 63 equiv wt% Sulfides (n = 76): Ridgeway mafic Harper (2000),   NaCl + KCl   −6.3 to +9.2‰,   monzonite: 0.70452   Wilson (2003),   with most between Ridgeway monzonite: 0.70433   Wilson et al. (2003),   −4 and −1‰   Wilson et al. (2007a),   Cooke et al. (2007),   Cuison (2010) Cadia Quarry 415° to 490° 53 to 61 equiv wt % No data available Cadia Quarry monzonite: Holliday et al. (2002),   NaCl + KCl   0.70426   Wilson (2003),   Wilson et al. (2004),   Wilson et al. (2007a, b),   Cooke et al. (2007) Cadia Hill 415° to 490° 53 to 61 equiv wt % Sulfides (n = 28): Cadia Hill Monzonite: 0.70419 Wilson (2003),   NaCl + KCl   −9.8 to +9.1‰, Epidote: 0.70451   Wilson et al. (2007a),   with most between   Cooke et al. (2007)   −6 and −2‰ Cadia East 415° to 490° 53 to 61 equiv wt % Sulfides (n = 51): Cadia East monzonite: 0.70426 Wilson (2003),   NaCl + KCl   −5.4 to −1.0‰ Tourmaline: 0.70403-0.70443   Finn (2006), Epidote: 0.70440-0.70467   Wilson et al. (2007a),   Cooke et al. (2007),   Fox (2012)

and to the sides of the ore zone (Harper, 2000; Wilson, 2003; Wilson et al., 2007b). Spatial zoning patterns of increasing δ34S values outward from the deposit cores are the opposite to what would be expected if cooling alone was responsible for S isotope fractionation. Instead, the sulfate-dominant magmatic-hydrothermal fluids are interpreted to have undergone water-rock interaction as they migrated away from the intrusive complexes, causing Fe2+ that occurs in solid solution in igneous feldspars to be oxidized to Fe3+, producing the redrock alteration halos (hematite dusting) that characterize the inner propylitic halos of the Cadia porphyry deposits (Wilson et al., 2007b). Oxidation of Fe2+ in feldspars during water-rock interaction could have caused reduction of SO42–(aq) to H2S(aq), providing additional reduced sulfur for ore deposition and resulting in sulfide δ34S values increasing systematically with distance away from the fluid source at each porphyry deposit (Wilson et al., 2007). Exploration Model Figure 12 is a schematic model for alkalic porphyry Au-Cu deposits based on the geologic relationships observed at Cadia East (modified from Holliday and Cooke, 2007). It highlights the subtle variations in potassic, calc-potassic, sodic, and propylitic alteration that can be encountered in alkalic porphyry environments. Both Cadia East and Ridgeway have high-level strata-bound alteration zones characterized by albite-, K-feldspar-, sericite-, and tourmaline-bearing alteration assemblages that may be the alkalic equivalents of lithocaps. This upper strata-bound alteration zone at Cadia East is laterally extensive and its exploration significance was not appreciated prior to the discovery of this giant porphyry deposit. The inner propylitic alteration zones around each of the Cadia porphyry deposits are marked by pronounced reddening, caused by submicron-sized hematite dusting of secondary feldspars, indicative of wall-rock oxidation by the mineralizing fluids. This red-rock alteration halo provides a useful exploration vector to mineralization, as it is

a near-miss indicator during drilling, providing confidence that an oxidized magmatic-hydrothermal fluid source was nearby. Conclusions Alkalic porphyry deposits of the Cadia district were the products of multiple phases of intrusive activity that occurred during the early stages of the Benambran orogeny, at the culmination of magmatic activity in the Ordovician to early Silurian Macquarie Arc. Hydrothermal fluids sourced from multiphase monzonite intrusive complexes produced highgrade Au-Cu mineralization in sheeted quartz-sulfide-carbonate veins (Cadia East, Cadia Hill, Cadia Quarry) and stockworks (Ridgeway). Complex and subtle overprinting and zoning relationships characterize the potassic, calc-potassic, sodic, and propylitic assemblages that coincide with and surround each deposit. Mineralizing fluids were oxidized (sulfate-dominant) brine and vapor, with water-rock interaction producing distinctive red-rock alteration halos via hematite dusting of feldspars around each deposit. The deposits vary from pluton-hosted (Cadia Hill, Cadia Quarry) to volcanic-hosted, intrusion-centred (Ridgeway, Cadia East), with variations in the host rocks and permeability characteristics producing considerable diversity between the deposits. Phyllic alteration is restricted to late-stage faults and, locally, associated with carbonate-base metal veins. A distinctive and unusual K-feldspar-albite-tourmaline blanket covers the giant Cadia East deposit (Holliday and Cooke, 2007; Wilson et al., 2007b; Fox, 2012; Fig. 12). This upper, distal alteration assemblage warrants more consideration by explorers because it could represent the upper, lithocap environment of alkalic porphyry systems and is distinct from the lithocaps that typically overlie calc-alkaline porphyry deposits (e.g., Sillitoe, 1995; Chang et al., 2011). Similar strata-bound feldspar alteration domains in alkalic provinces should be explored thoroughly as they may conceal a significant alkalic porphyry resource.



ALKALIC PORPHYRY Au-Cu DEPOSITS, CADIA, NSW, AUSTRALIA

641

Alkalic lithocap

(albite – K-feldspar – tourmaline ± sericite ± quartz ± carbonate): chargeability high, magnetic low

Skarn (mt skarn may form in limestone or reactive volcanic and volcaniclastic rocks, magnetic high) Reddened propylitic halo

(hematite dusting of feldspars, negative S isotopes in sulfides, increasing magnetic susceptibility towards ore, Zn and Pb geochemical anomaly)

Legend – Alteration

Distal propylitic (chlorite sub-zone: chl - carb ± hm ± epi)

Skarn (py - hm - mt - chl - carb - gt)

Pyrite halo

Skarn propylitic (epi - py) Alkalic lithocap (ab - kf - tm ± ser ± qz ± carb)

Sodic (ab - qz - hm)

Calc-potassic or potassic core (magnetic high, Cu - Au ± Mo ore zone)

Intrusive complex

500 m

Outer propylitic (albite - actinolite subzone: ab - act - qz - carb - py) Inner propylitic (actinolite - hematiteepidote subzone: ab - chl - act - epi - hm) Outer calc-potassic

(Kf - chl - bi - ab - act - qtz - cp)

Inner calc-potassic

(bi - act - mt - kf - ab - qtz - bn)

Fig. 12. Schematic illustration of alteration zoning and overprinting relationships in a volcanic-hosted, intrusion-centered alkalic porphyry system, based on geologic relationships from the Cadia East porphyry Au-Cu deposit (Tedder et al., 2001; Wilson 2003; Holliday and Cooke, 2007). Cadia East contains what may be the alkalic equivalent of a lithocap with less acidic alteration assemblages (albite-K-feldspar-tourmaline ± sericite ± quartz ± carbonate). The propylitic subfacies around alkalic porphyry Au-Cu deposits are more complicated than around calc-alkaline porphyry deposits. Calcium-bearing alteration minerals (calcite, actinolite, epidote, garnet) occur in the core of alkalic porphyry deposits, in contrast to calc-alkaline porphyries. The dashed line shows the approximate location of the pyrite halo (defined by the outer limit of disseminated pyrite). The position of the pyrite halo varies from deposit to deposit, but because the propylitic halo typically extends beyond the pyrite halo, mapping can identify outer pyrite-deficient and inner pyrite-bearing propylitic assemblages. Abbreviations: ab = albite, act = actinolite, anh = anhydrite, Au = gold, bn = bornite, bt = biotite, cb = carbonate, chl = chlorite, cp = chalcopyrite, epi = epidote,gt = garnet, hm = hematite, Kf = K-feldspar, lm = laumontite, mt = magnetite, pr = prehnite, py = pyrite, qz = quartz, ser = sericite, tm = tourmaline.

Acknowledgments We thank all of our colleagues and students over the years who have contributed to our understanding of the Cadia porphyry deposits, including Ian Tedder, Dan Wood, Paul Dunham, Ben Harper, Tony Crawford, Ron Berry, Jocelyn McPhie, Ross Large, Joel Kitto, Dave Finn, Janina Micko, Dean Fredrickson, Dean Collett, Garry Davidson, Cari Deyell, Joanne Morrison, Cassady Harraden, Laura Jackson, William Reynolds, Andrew Beattie, and Mitchell Bland. Thanks to Newcrest Mining Ltd. for their sustained support of research in the Cadia district. We thank the Australian Research Council for their support of our research through the Industrial Transformation Research Program, and through the Centre of Excellence funding scheme. We thank Dick Sillitoe and Pepe Perelló for their insightful and constructive review comments that helped to significantly improve this manuscript. REFERENCES Blevin, P.L., 2002, The petrographic and compositional character of variably K-enriched magmatic suites associated with Ordovician porphyry Cu-Au

mineralisation in the Lachlan fold belt, Australia: Mineralium Deposita, v. 37, p. 87–99. Chang, Z., Hedenquist, J.W., White, N.C., Cooke, D.R., Roach, M., Deyell, C.L., Garcia, J., Gemmell, J.B., McKnight, S., and Cuison, A.L., 2011, Exploration tools for linked porphyry and epithermal deposits: Example from the Mankayan intrusion-centered Cu-Au district, Luzon, Philippines, Economic Geology, v.106, p.1365–1398 Close, D.I., 2000, A geophysical study of the Ridgeway gold-copper deposit, New South Wales: B.Sc. (Honours) thesis, Hobart, University of Tasmania, 132 p. Cooke, D.R., Wilson, A.J., House, M.J., Wolfe, R.C., Walshe, J.L., Lickfold, V., and Crawford, A.J., 2007, Alkalic porphyry Au–Cu deposits and associated mineral deposits of the Ordovician to early Silurian Macquarie Arc, New South Wales: Australian Journal of Earth Sciences, v. 54, p. 455–463. Crawford, A.J., Glen, R.A., Cooke, D.R., and Percival, I.G., 2007a, Geological evolution and metallogenesis of the Ordovician Macquarie Arc, Lachlan Orogen, New South Wales: Australian Journal of Earth Sciences, v. 54, p. 137–479. Crawford, A.J., Meffre, S., Squire, R.J., Barron, L.M., and Falloon, T.J., 2007b, Middle and Late Ordovician magmatic evolution of the Macquarie Arc, Lachlan Orogen, New South Wales: Australian Journal of Earth Sciences, v. 54, p. 181–214. Crawford, A.J., Cooke, D.R., and Fanning, C.M., 2007c, Geochemistry and age of magmatic rocks in the unexposed Narromine, Cowal and Fairholme

642

HARRIS ET AL.

Igneous Complexes in the Ordovician Macquarie Arc, New South Wales: Australian Journal of Earth Sciences, v. 54, p. 243–271. Cuison, A.L.G., 2010, Geology and genesis of the Ridgeway porphyry Au-Cu deposit, NSW: Ph.D. thesis, Hobart, University of Tasmania, 305 p. Finn, D.J., 2006, Late stage phyllic alteration in the Cadia East copper-gold porphyry deposit, NSW: Implications to mineralisation: B.Sc. (Honours) thesis, Hobart, University of Tasmania, 108 p. Forster, D.B., Seccombe, P.K., and Phillips, D., 2004, Controls on skarn mineralization and alteration at the Cadia deposits, New South Wales, Australia: Economic Geology, v. 99, p. 761–788. Fox, N., 2012, Controls on alteration and mineralization at the Cadia East alkalic porphyry Au-Cu deposit, NSW: Ph.D. thesis, Hobart, University of Tasmania, 222 p. Fox, N., Harris, A.C., Cooke, D.R., Collett, D., and Faure, K, 2009, Hydrothermal fluid sources in the development of shallow potassic-sodic alteration at the Cadia East alkalic porphyry Au-Cu deposit, NSW: Proceedings of the 10th Biennial Meeting of the Society for Geology Applied to Mineral Deposits, August 17–20, Townsville, Australia, p. 222–224. Fox, N., Cooke, D.R., Harris, A.C.H., Collett, D., and Eastwood, G., 2015, Porphyry Au-Cu mineralization controlled by reactivation of an arc-transverse volcanosedimentary subbasin: Geology, v. 43, p. 811–814. Glen, R.A. 1992, Thrust, extensional and strike-slip tectonics in an evolving Palaeozoic orogen—a structural synthesis of the Lachlan Orogen of southeastern Australia: Tectonophysics, v. 214, p. 341–380. ——2005, The Tasmanides of eastern Australia: Geological Society, London, Special Publication 246, p. 23–96. Glen, R.A., and Walshe, J.L., 1999, Cross-structures in the Lachlan Orogen: the Lachlan transverse zone example: Australian Journal of Earth Sciences, v. 46, p. 641–658. Glen, R.A., Walshe, J.L., Barron, L.M., and Watkins, J.J., 1998, Ordovician convergent-margin volcanism and tectonism in the Lachlan sector of east Gondwana: Geology, v. 26, p. 751–754. Glen, R.A., Crawford, A.J., Percival, I.G., and Barron, L.M., 2007a, Early Ordovician development of the Macquarie Arc, Lachlan Orogen, New South Wales: Australian Journal of Earth Sciences, v. 54, p. 167–179. Glen, R.A., Meffre, S., and Scott, R.J., 2007b, Benambran orogeny in the eastern Lachlan Orogen, Australia: Australian Journal of Earth Sciences, v. 54, p. 385–415. Glen, R.A., Crawford, A.J., and Cooke, D.R., 2007c, Tectonic setting of porphyry Cu-Au mineralisation in the Ordovician-early Silurian Macquarie Arc, eastern Lachlan Orogen, New South Wales: Australian Journal of Earth Sciences, v. 54, p. 465–479. Glen, R.A., Saeed, A., Quinn, C.D., and Griffin, W.L., 2011, U-Pb and Hf isotope data from zircons in the Macquarie Arc, Lachlan Orogen: Implications for arc evolution and Ordovician paleogeography along part of the east Gondwana margin: Gondwana Research, v. 19, p. 670–685. Gray, D.R., and Foster, D.A., 2004, Tectonic evolution of the Lachlan Orogen, southeast Australia: Historical review, data synthesis and modern perspectives: Australian Journal of Earth Sciences, v. 51, p. 773–817. Green, D.G., 1999, The geology and origin of the Big Cadia skarn, NSW, Australia: B.Sc. (Honours) thesis, Hobart, University of Tasmania, 154 p. Harper, B.L., 2000, Hydrothermal alteration at the Ridgeway porphyry gold-copper deposit, N.S.W.: B.Sc. (Honours) thesis, Hobart, University of Tasmania, 130 p. Harraden, C.L., Cracknell, M.J., Lett, J., Berry, R.F., Carey, R., and Harris, A.C., 2019, Automated core logging technology for geotechnical assessment: A study on core from the Cadia East porphyry deposit: Economic Geology, v. 114, p. 1495–1511. Harris, A.C., Cooke, D.R., Washburn, M., Tosdal, R.M., and Fox, N., 2009a, Reconstruction of the Macquarie Arc: A review of the tectonic setting of the porphyry ore deposits of central New South Wales: Final Report. Cadia Valley alkalic porphyry ore deposits: Deconstructing the system architecture: Orange, NSW, ARC Centre of Excellence in Ore Deposits and Newcrest Mining Limited, p. 6.1–6.13. Harris, A.C., Percival, I., Allen, C., Tedder, I., McMillan, C., Cooke, D.R., McPhie, J., Dunham, P., and Collett, D., 2009b, Volcanic stratigraphy and architecture of the Cadia Valley, New South Wales: Final Report. Cadia Valley alkalic porphyry ore deposits: Deconstructing the system architecture: Orange, NSW, ARC Centre of Excellence in Ore Deposits and Newcrest Mining Limited, p. 2.1–2.18. Harris, A.C., Percival, I.G, Cooke, D.R., Tosdall, R.M., Fox, N., Allen, C., Tedder, I., McMillan, C., Dunham, P., and Collett, D., 2014, Marine

volcanosedimentary basins hosting porphyry Au-Cu deposits, Cadia Valley, New South Wales, Australia: Economic Geology, v. 109, p. 1117–1135. Holliday, J.R., and Cooke, D.R., 2007, Advances in geological models and exploration methods for copper ± gold porphyry deposits, in Milkereit, B., ed., Proceedings of Exploration 07: Fifth Decennial International Conference on Mineral Exploration: Toronto, Prospectors and Developers Association of Canada, p. 791−809. Holliday, J., and Wood, D., 1993, Discovery and exploration of porphyry-­ associated gold/copper mineralisation at Cadia, in New South Wales—the state of exploration: Sydney, New South Wales Department of Mineral Resources, Proceedings, p. 19–22. Holliday, J., McMillan C., and Tedder, I., 1999, Discovery of the Cadia Ridgeway gold-copper deposit, in New generation gold mines—case histories of discovery: Perth, Australian Mineral Foundation, p. 101–107. Holliday, J.R., Wilson, A.J., Blevin, P.L., Tedder, I.J., Dunham, P.D., and Pfitzner, M., 2002, Porphyry gold-copper mineralisation in the Cadia district, eastern Lachlan Fold Belt, New South Wales, and its relationship to shoshonitic magmatism: Mineralium Deposita, v. 37, p. 100–116. Keith, J. D., Whitney, J. A., Canna, T. M., Hook, C. and Hattori, K., 1995, The role of magmatic sulphides and mafic alkaline magmatism in the formation of giant porphyry and vein systems: Examples from the Bingham and Tintic mining districts, Utah: in Clark A. H., ed., Giant ore deposits—II, QMinEx: Kingston, Ontario, Department of Geological Sciences, Queen’s University, p. 316-339. Kemp, A.I.S., Blevin, P.L., and Norman, M.D., 2020, A SIMS U-Pb (zircon) and Re-Os (molybdenite) isotope study of the early Paleozoic Macquarie Arc, southeastern Australia: Implications for the tectono-magmatic evolution of the paleo-Pacific Gondwana margin: Gondwana Research, v. 82, p. 73–96. Kitto, J.C., 2005, Lithostratigraphy, alteration, and geochemistry at the Cadia East Au-Cu porphyry deposit, NSW: B.Sc. (Honours) thesis, Hobart, University of Tasmania, 136 p. Lickfold, V., Cooke, D.R., Crawford, A.J., and Fanning, C.M., 2007, Shoshonitic magmatism and the formation of the Northparkes porphyry Cu-Au deposits, New South Wales: Australian Journal of Earth Sciences, v. 54, p. 417–444. Lowczak, J.N., Campbell, I.H., Cocker, H., Park, J.W., and Cooke D.R., 2018, Platinum-group element geochemistry of the Forest Reefs Volcanics, southeastern Australia: Implications for porphyry Au-Cu mineralisation: Geochimica et Cosmochimica Acta, v. 220, p. 385–406. Meffre, S., Scott, R.J., Glen, R.A., and Squire, R.J., 2007, Re-evaluation of contact relationships between Ordovician volcanic belts and the quartz-rich turbidites of the Lachlan Orogen: Geological evolution and metallogenesis of the Ordovician Macquarie Arc, Lachlan Orogen, New South Wales: Australian Journal of Earth Sciences, v. 54, p. 363–383. Morrison, J., 2016, The association of fluorine recovery with lithogeochemical domains at Cadia East: Master of Economic Geology thesis, Hobart, University of Tasmania, 72 p. Newcrest Mining Ltd., 2009, Newcrest annual report 2009: Melbourne, Australia, Newcrest Mining Ltd., 116 p. (http://www.annualreports.com/ HostedData/AnnualReportarchive/N/ASX_NCM_2009.pdf). ——2010, Newcrest annual report 2010: Melbourne, Australia, Newcrest Mining Ltd., 116 p. (http://www.annualreports.com/HostedData/AnnualReportarchive/N/ASX_NCM_2010.pdf). ——2017, Annual report: Melbourne, Australia, Newcrest Mining Ltd., 152 p. (https://www.newcrest.com/sites/default/files/2019-10/171009_Newcrest%20Annual%20Report_0.pdf). ——2019, Annual report: Melbourne, Australia, Newcrest Mining Ltd., 148  p. (https://www.newcrest.com/sites/default/files/2019-10/191004_New crest%20Annual%20Report_0.pdf). Oversby, B., 1971, Palaeozoic plate tectonics in the southern Tasman geosyncline: Nature (Physical Sciences), v. 234, p. 45–48. Pacey, A., Wilkinson, J.J., Owens, J., Priest, D., Cooke, D.R., and Millar, I.L., 2019, The anatomy of an alkalic porphyry Cu-Au system: Geology and alteration at Northparkes mines, New South Wales, Australia: Economic Geology, v. 114, p. 441–472. Pacey, A., Wilkinson, J.J., and Cooke, D.R., 2020a, Chlorite and epidote mineral chemistry in porphyry ore systems: A case study of the Northparkes district, New South Wales, Australia: Economic Geology, v. 115, p. 701–727. Pacey, A., Wilkinson, J.J., Boyce, A.J., and Millar, I.L., 2020b, Isotopic constraints on the origin of propylitic alteration in porphyry deposits: Insights from the Northparkes district, New South Wales, Australia: Economic Geology, v. 115, p. 729–748.



ALKALIC PORPHYRY Au-Cu DEPOSITS, CADIA, NSW, AUSTRALIA

Packham, G., Percival, I., and Bischoff, G., 1999, Age constraints on strata enclosing the Cadia and Junction Reefs ore deposits of central New South Wales, and tectonic implications: Quarterly Notes—Geological Survey of New South Wales, v. 110, p. 1–12. Packham, G.H., Keene, J.B., and Barron, L.M., 2003 Middle to early Late Ordovician hydrothermal veining in the Molong volcanic belt, northeastern Lachlan Fold Belt: Sedimentological evidence: Australian Journal of Earth Sciences, v. 50, p. 257–269. Percival, I.G., and Glen, R.A., 2007, Ordovician to earliest Silurian history of the Macquarie Arc, Lachlan Orogen, New South Wales: Geological evolution and metallogenesis of the Ordovician Macquarie Arc, Lachlan Orogen, New South Wales: Australian Journal of Earth Sciences, v. 54, p. 143–165. Pogson, D.J., and Watkins, J.J., eds., 1998, Bathurst 1:250 000 Geological Sheet SI/55-8 explanatory notes: Sydney, Geological Survey of New South Wales, 430 p. Reynolds, W.J., 2007, Characterisation of distal alteration assemblages at the Ridgeway gold-copper deposit, New South Wales: B.Sc. (Honours) thesis, Hobart, University of Tasmania, 86 p. Richards, J.P., 2009, Postsubduction porphyry Cu-Au deposits and epithermal Au deposits: Products of remelting of the subduction-modified lithosphere: Geology, v. 37, p. 247–250. Rickards, R.B., and Wright, A.J., 2004, early Silurian graptolites from Cadia, New South Wales: Proceedings of the Linnean Society of New South Wales, v. 125, p. 173–175. Rickards, R.B., Percival, I.G., Simpson, A.J., and Wright, A.J., 2001, Silurian biostratigraphy of the Cadia area, south of Orange, New South Wales: Proceedings of the Linnean Society of New South Wales, v. 123, p. 173–191. Scheibner, E., 1973, A plate tectonic model of the Palaeozoic tectonic history of New South Wales: Journal of the Geological Society of Australia, v. 20, p. 405–426. Sillitoe, R.H., 1995, Exploration of porphyry copper lithocaps, in Mauk, J.L., and St George, J.D., eds., Australasian Institute of Mining and Metallurgy Pacific Rim Congress, Auckland, 1995, Proceedings, p. 527–532. Simpson, C.J., Cas, R.A.F., and Arundell, M.C., 2005, Volcanic evolution of a long-lived Ordovician island-arc province in the Parkes region of the Lachlan fold belt, southeastern Australia: Australian Journal of Earth Sciences, v. 52, p. 863–886. Simpson, C.J., Scott, R.J., Crawford, A.J., and Meffre, S., 2007, Volcanology, geochemistry and structure of the Ordovician Cargo Volcanics in the Cargo-Walli region, central New South Wales: Australian Journal of Earth Sciences, v. 54, p. 315–352. Solomon, M., and Griffiths, J.R., 1972, Tectonic evolution of the Tasman orogenic zone, eastern Australia: Nature (Physical Sciences), v. 237, p. 3–6. Squire, R.J., 2001, Volcanological and tectono-magmatic evolution of the Cadia-Neville region, NSW: Ph.D. thesis, Hobart, University of Tasmania, 221 p. Squire, R.J., and Crawford, A.J., 2007, Magmatic characteristics and geochronology of Ordovician igneous rocks from the Cadia-Neville region, New South Wales: Implications for tectonic evolution: Australian Journal of Earth Sciences, v. 54, p. 293–314.

643

Tedder, I.J., Holliday, J., and Hayward, S., 2001, Discovery and evaluation drilling of the Cadia Far East gold-copper deposit, in NewGenGold 2001, Conference Proceedings: Adelaide, Australian Mineral Foundation, p. 171–184. Tosdal, R.M., and Dilles, J.H., 2020, Creation of permeability in the porphyry Cu environment: Reviews in Economic Geology, v. 21, p. 205–214. Tosdal, R.M., and Richards, J.P., 2001, Magmatic and structural controls on the development of porphyry Cu ± Mo ± Au deposits: Reviews in Economic Geology, v. 14, p. 157–181. Washburn, M., 2008, Architecture of the Silurian sedimentary cover sequence in the Cadia porphyry Au-Cu district, NSW, Australia: Implication for post-mineralization deformation: M.Sc. thesis, Vancouver, University of British Columbia, 165 p. Wellman, P., and McDougall, I., 1974, Cainozoic igneous activity in eastern Australia: Tectonophysics, v. 23, p. 49–65. Wells, T.J., Meffre, S., Cooke, D.R., Steadman, J.A., and Hoye, J.L., 2020, Porphyry fertility in the Northparkes district: Indicators from whole-rock geochemistry: Australian Journal of Earth Sciences, v. 67, p. 717–738. Wilson, A.J., 2002, Diverse styles of porphyry gold-copper mineralisation, Cadia district, New South Wales, Australia [abs.]: Giant Ore Deposits, Centre for Ore Deposit Research (CODES), University of Tasmania, Hobart, Abstract Volume, p. 29–31. ——2003, The geology, genesis and exploration context of the Cadia gold-copper porphyry deposits, New South Wales, Australia: Ph.D. thesis, Hobart, University of Tasmania, 335 p. Wilson, A.J., Cooke, D.R., and Harper, B.L., 2003, The Ridgeway gold-copper deposit: A high-grade alkalic porphyry deposit in the Lachlan fold belt, New South Wales, Australia: Economic Geology, v. 98, p. 1637–1666. Wilson, A.J., Cooke, D.R., and Richards, T., 2004, Veins, pegmatites and breccias: Examples from the alkalic Cadia Quarry gold-copper porphyry deposit, NSW, Australia: Hobart, Australia, Centre for Ore Deposit Research (CODES) Special Publication 5, p. 45–55. Wilson, A.J., Cooke, D.R., Stein, H.J., Fanning, C.M., Holliday, J.R., and Tedder, I.J., 2007a, U-Pb and Re-Os geochronologic evidence for two alkalic porphyry ore-forming events in the Cadia district, New South Wales, Australia: Economic Geology, v. 102, p. 3–26. Wilson, A.J., Cooke, D.R., Harper, B.J., and Deyell, C.L., 2007b, Sulfur isotopic zonation in the Cadia district, southeastern Australia: Exploration significance and implications for the genesis of alkalic porphyry gold-copper deposits: Mineralium Deposita, v. 42, p. 465–487. Wood, D., 2012a, Discovery of the Cadia deposits, NSW, Australia (Part 1): Society of Economic Geologists Newsletter 88, p. 1, 13-18. ——2012b, Discovery of the Cadia deposits, NSW, Australia (Part 2): Society of Economic Geologists Newsletter 89, p. 1, 17-22. Wood, D., and Holliday, J., 1995, Discovery of the Cadia gold/copper deposit, New South Wales, in New generation gold mines: Case histories of discovery: Adelaide, Australian Mineral Foundation, p 11.1–11.10.

©2020 Society of Economic Geologists, Inc. SEG Special Publications, no. 23, pp. 645–668

Chapter 31 Geologic Evidence of Syngenetic Gold in the Witwatersrand Goldfields, South Africa Hartwig E. Frimmel1,† and Glen T. Nwaila 1 Bavarian Georesources Centre (BGC), Department of Geodynamics and Geomaterials Research, Institute of Geography and Geology, University of Würzburg, Am Hubland, D-97074 Würzburg, Germany, and Department of Geological Sciences, University of Cape Town, Rondebosch 7700, South Africa 2 School

of Geosciences, University of the Witwatersrand, Private Bag 3, Wits, 2050, South Africa

Abstract The Mesoarchean Witwatersrand Basin in the central Kaapvaal craton, South Africa, has been the largest gold-producing province in history. Although mining has reached a very mature state, this ore province remains the biggest regional gold anomaly in the world. Most recent research on the Witwatersrand gold deposits has focused on postdepositional processes, often on a microscale, thereby constraining conditions of gold transport in the host conglomerates. Here we review past and current observations on the geologic setting of the orebodies and first-order controls on gold mineralization, all of which strengthen the argument for a primarily syngenetic model. The Witwatersrand deposits are regarded as remnants of a gold megaevent at 2.9 Ga when environmental conditions are suggested to have been suitable for intense gold flux off the Archean land surface and early photosynthesizing microbes could act as trap sites for riverine and possibly shallow-marine gold. Sedimentary reworking of gold-rich microbial mats led to rich placer deposits which, in turn, became sources of younger placers higher up in the stratigraphy. The same gold concentration mechanism most likely operated on all Mesoarchean land masses, not only on the Kaapvaal craton. The uniqueness of the Witwatersrand gold province is explained by exceptional preservation of these easily erodible, largely continental sediments beneath a thick cover of flood basalt and a later impact melt sheet in the middle of a buoyant craton, with little tectonic overprint over the past two billion years.

Introduction The Witwatersrand Basin, stretching for ~350 km in a northeasterly and ~200 km in a northwesterly direction near the center of the Kaapvaal craton (Fig. 1A), is by far the largest known gold province in the world and accounts for some 53,000 metric tons (t) Au, close to one-third of the gold that has been mined throughout history. For many decades the Witwatersrand gold fields (Fig. 1B) were the world’s leading gold producers. Although the mines there have reached a very mature state, with annual production having decreased continuously since 1970 (Fig. 2) and more and more mines facing closure, the Witwatersrand Basin still contains close to 30% of known global gold resources (Frimmel, 2014). More than a century of mining and exploration, largely underground, turned the Witwatersrand Supergroup rocks into probably the most intensely investigated Mesoarchean sediment archive in the world, notwithstanding very limited surface outcrop. A unique geologic history prevented total erosion or tectonic recycling of these ancient deposits. Not surprisingly, the Witwatersrand Basin fill has become the best available record of Mesoarchean environmental conditions and serves as a key reference for the reconstruction of secular changes in the composition of the Archean atmosphere, hydrosphere, and biosphere. There is hardly any imaginable genetic model that has not been proposed at some time or other to explain the origin of the Witwatersrand gold deposits. Proposed models range from synsedimentary placer to epigenetic, magmatic, and metamorphic. The genesis of Witwatersrand gold has been †Corresponding

author: e-mail, [email protected]

one of the most debated issues in economic geology (e.g., Muntean et al., 2005), and well over 1,000 publications on the topic have failed to provide a universally accepted consensus. Bearing in mind the mature state of the mining operations and the fact that many key sections are either inaccessible or mined out, it is opportune to summarize the critical geologic features that might help to clarify the metallogeny of this unparalleled gold province; this is therefore the aim of this contribution. The number of publications and theses devoted to the Witwatersrand gold province is far too large to be properly credited here and the interested reader is referred to the most recent review papers for further information and references (Robb and Meyer, 1995; Robb and Robb, 1998; Phillips and Law, 2000; Frimmel et al., 2005; Hayward et al., 2005; McCarthy, 2006; Frimmel, 2014; Tucker et al., 2016). Mining History Most historians regard George Walker and George Harrison as the discoverers of Witwatersrand gold because they had found gold on farm Langlaagte 224 IQ on 28 March 1886 in a position that later became known as the Main reef. The true discoverer was, however, Fred Strubens, who, with financial and logistic backing from his brother Harry, discovered the Confidence reef, a quartz vein, on farm Wilgespruit 192 IQ as early as 1884, which was followed a year later by their discovery of auriferous conglomerate on farm Groot Paardekraal 226 IQ (later recognized as the Main reef) and several other farms (Handley, 2004). First descriptions of the local stratigraphy and geographic distribution of gold occurrences were provided by Fred Strubens in April 1885. Soon thereafter, on 1 August 1885, the first mining concession over Witwatersrand

doi: 10.5382/SP.23.31; 24 p. 645

646

FRIMMEL AND NWAILA

Fig. 1. A. Surface and subsurface distribution of the main Archean stratigraphic units of the Kaapvaal craton and the Bushveld Igneous Complex, showing the extent of the Witwatersrand Basin, the fill of which comprises the West Rand and Central Rand groups. Blue line delineates trace of the Black Reef Quartzite Formation at the base of Neoarchean to early Paleoproterozoic Transvaal Basin (from Frimmel and Hennigh, 2015). Subsurface extent of West Rand Group and correlating Pongola Supergroup extrapolated from drilling and geophysical surveys (from Corner and Durrheim, 2018). Major greenstone belts: 1 = Barberton, 2 = Murchison, 3 = Giyani, 4 = Pietersburg, 5 = Kraaipan, 6 = Amalia; SIFS = Saddleback-Inyoka fault system. B. Simplified surface and subsurface geologic map of the Witwatersrand Basin, also showing the distribution of Archean granitoid domes, the location of the gold fields, and major faults (from Frimmel and Minter, 2002).



GEOLOGIC EVIDENCE OF SYNGENETIC Au, WITWATERSRAND GOLDFIELDS, SOUTH AFRICA

Fig. 2. A. South Africa’s gold production (almost exclusively from the Witwatersrand) in comparison to global gold production since the discovery of the Witwatersrand gold fields. B. Variation in gold grade and tonnage of ore milled over Witwatersrand mining history. C. Hubbert linearization applied to Witwatersrand gold production: P = annual production, Q = cumulative production, Q∞ = extrapolated total recoverable amount of gold.

rocks was granted to Harry Strubens. In contrast to other gold fields elsewhere in the world at that time, there was hardly any alluvial gold. Instead, the ore was hard weathered rock exposed at surface, which necessitated the erection of stamp batteries to pulverize the ore. Thus, significant production only began in 1887, leading to the founding of Johannesburg. The obstacle caused by the refractory nature of the pyritic conglomerate reefs was overcome by John S. MacArthur and Robert and William Forrest from Scotland, who invented the gold cyanidation process that was first employed on the Witwatersrand in 1890. It did not take long before the first diamond drilling (as early as 1889) established the downdip extension of rich gold-bearing conglomerate to depths of 2 km. It soon became clear that mining of the Witwatersrand

647

ores would require large amounts of capital, and the rush for prospective ground by mining companies commenced. A steady increase in production from the Central, East, and West Rand gold fields, with average grades of ~20 g/t, was briefly interrupted by the Anglo-Boer war (1899–1902), only to continue its rise thereafter, partly thanks to the arrival of cheap labor from China in 1904, until the period from 1924 to 1933. At that point, growth slowed due to global financial problems, marked by the 1929 Wall Street crash and difficulties that arose in local labor relationships (Fig. 2A). In 1933 the Union of South Africa abandoned the gold standard and introduced new mining legislation, while in 1934 the United States increased the price of gold to US$35/oz. These events triggered renewed interest in the Witwatersrand, with the registration of new well-financed companies and opening of new mines at a time when the rest of the world was experiencing its greatest economic depression. The Carletonville gold field was discovered and the first shafts were sunk in the Klerksdorp gold field in 1952. Critical to the Witwatersrand’s emergence as the world’s leading gold producer for most of the 20th century was the foresight of a young geophysicist, Rudolf Krahmann, who in 1930 came up with the idea of delineating hidden magnetite-bearing shales below the main gold-bearing conglomerates with a magnetometric field balance (Handley, 2004). This technique proved so successful that in the following years drill holes could be sited to intersect their targets beneath hundreds of meters of younger cover rocks. This pioneering geophysical method soon resulted in the development of the Carletonville gold field and, once exploration programs resumed after a pause during World War II, in the Welkom gold field. Shortly thereafter, the first aeromagnetic survey flown in South Africa resulted in the opening of the Evander gold field at the opposite end of the basin (Fig. 1B), where detailed magnetic and gravimetric surveys were instrumental in locating orebodies and their exploitation from 1955 to 2018. From 1955 to 1970 Witwatersrand gold output increased steadily, reaching its peak in 1970 with an annual production of 988,933 kg. Although termination of the Bretton Woods agreement by the United States in 1971 had a strong effect on the global gold market, as it led to a steeply rising gold price and to a series of gold sales by Central Banks and the International Monetary Fund in the late 1970s, the reasons for the decline of Witwatersrand gold production since 1970 are more complex. As evident from Figure 2B, the grade began to drop, because some of the rich orebodies had already been mined out and the higher gold price enabled reduction of cut-off grades. Correspondingly, the tonnage of milled ore increased until it reached a peak of 130 Mt in 1990. This period also coincided with the employment of more than half a million workers at the mines. Since then both the tonnage and gold grade have decreased continuously for many reasons, including geologic, sociopolitical, and economic. With the advent of a democratic “new” South Africa in 1994, South African major companies shifted their focus to other countries, which diverted much needed investment capital away from the Witwatersrand. Production costs rose as mines became deeper and labor costs increased. The number of operating mines dwindled from 60 in the heydays of the 1960s to 21 at present, operated by no more than four major (AngloGold

648

FRIMMEL AND NWAILA

Ashanti, Harmony, Sibanye Stillwater, Gold Fields) and two smaller companies (Pan African Resources, Village Main reef). According to the Minerals Council South Africa, about half of the present mines are economically marginal. Labor strife and falling productivity add to the difficulties and illegal mining (also underground!) is becoming an increasing problem. A new mining charter introduced by the government in 2017 did not help to inspire confidence in South Africa’s mining industry either, quite the contrary. Much of the challenge of the Witwatersrand mines revolves around the increasing depth of mining. Today some of the mines in the Witwatersrand are the deepest in the world with operations at 4 km below surface. Ambient temperatures underground are well in excess of 50°C and costly procedures to cool active stopes are required, routinely using chilled (and later ice-loaded) water since 1975. Applying a Hubbert-linearization approach to Witwatersrand gold production results in a very gloomy prospect for the future (Fig. 2C). The theoretically extrapolated total cumulative production (Q∞) is approximately 56,000 t of gold. Of these, 53,000 t have already been produced. The difference, although smaller than the official total reserve of ~6,000 t, underscores the extreme maturity of the Witwatersrand gold province and implies that only ~20 years of mining remain at current production levels. This depressing outlook stands in stark contrast to the huge resources that are inferred to remain in the Witwatersrand but too deep to be recovered economically with current mining methods. A dramatic change in mining methods as well as in politics, safety, and productivity would be required to breathe new long-term life into the Witwatersrand gold mines. Regional Geologic Setting The Witwatersrand Basin (Fig. 1) evolved after the proto-Kaapvaal craton had largely consolidated by amalgamation of the Witwatersrand and Swaziland blocks along the Saddleback-Inyoka fault system at ca. 3.23 Ga (Dziggel et al., 2006). Some gold was introduced along late shear zones in the greenstone belts at ~3.1 Ga (Dziggel et al., 2010). The change from an overall compressional to transtensional stress regime heralded the beginning of intracontinental basin formation that eventually led to the Witwatersrand Basin. Initially, an up to 2,250-m-thick, bimodal volcanic sequence with a thin basal siliciclastic unit (Dominion Group) was laid down between 3086 ± 3 Ma (the youngest age of pre-Dominion basement; Robb et al., 1992) and 3074 ± 6 Ma (the age of volcanism; Armstrong et al., 1991). The basal siliciclastic unit includes conglomerate beds (Dominion reef) with abundant detrital uraninite and pyrite but relatively low gold tenors (Rantzsch et al., 2011). Paleocurrent data consistently point to a source to the north or northeast (Frimmel and Minter, 2002), but the original extent of this basin, for which both a continental and intra-arc setting has been suggested (Frimmel et al., 2009), remains unconstrained. After a hiatus of some 90 m.y., deposition of predominantly siliciclastic sediments of the lower Witwatersrand Supergroup, the West Rand Group, commenced on the peneplaned surface of the proto-Kaapvaal craton on top of Mesoarchean granitoids and greenstone and, in places, Dominion Group rocks. Initial drowning of the post-Dominion cratonic land surface was followed by repeated cycles of largely eustatic

regression and transgression. Most of the West Rand Group comprises quartzite and shale, with only very minor conglomerate. Many of the quartzite beds have been interpreted as shallow-marine shelf sand accumulations (Eriksson et al., 1981; Winter and Brink, 1991). The shale beds are marine, generally rich in iron, with four of them representing true iron formation (Smith et al., 2013). The latter overlie diamictite beds, which represent the oldest known glacial deposits in the world. The West Rand Group contains a single volcanic unit, a basaltic andesite (Crown Formation) near the top of the group, the age of which, 2914 ± 8 Ma (Armstrong et al., 1991), provides the best available minimum constraint on the timing of West Rand Group sedimentation. Stratigraphically equivalent units occur in the up to 4,200-m-thick, volcanosedimentary but predominantly volcanic Nsuze Group and the lowermost parts of the up to 4,800-m-thick, predominantly siliciclastic Mozaan Group of the Pongola Supergroup in the southeastern part of the Kaapvaal craton (Fig. 1A). These were deposited between 2.99 and 2.95 Ga and derived from a source to the north (Wilson and Zeh, 2018). Exploration drilling and geophysical surveys revealed that the West Rand Basin (and its Pongolan equivalent) covered most of the southern Kaapvaal craton (Fig. 1A) and was much larger than indicated by known outcrop and subcrop (Corner and Durrheim, 2018). The tectonic nature of this basin has been a matter of debate. The proximal northwestern part is considered a foreland basin formed behind an Andean-type arc along the northern margin of the protocraton in the north (Zeh et al., 2013), whereas the southeastern part represents a continental margin. This interpretation is suggested by uniform paleocurrent directions from the north, northeast, and northwest in the West Rand Group (Frimmel and Minter, 2002) and its equivalents in the Pongola Supergroup (Wilson and Zeh, 2018), an upward decrease in complexity of detrital zircon age spectra (Kositcin and Krapez, 2004), and the spatial distribution of sedimentary facies, including stromatolitic and microbial units in the Pongola Group (Wilson et al., 2013). With the extrusion of the 2914 Ma Crown Formation lavas, the general deepening of the basin was reversed, and shoreline propagation and eventual fluvial and braid-plain deposition heralded a change to predominantly continental sedimentation represented by the overlying Central Rand Group. The contact between the West Rand and Central Rand groups can be traced across the entire basin and probably reflects a hiatus of some 10 m.y. The 2.90 to 2.79 Ga Central Rand Group is dominated by fluvial to fluviodeltaic sandstone and conglomerate from former alluvial braid plains, especially in the lower Central Rand Group (Els, 1998), and alluvial fans toward the top of the group (Kingsley, 1987). A single marine shale unit of basin-wide extent (Booysens Formation) led to subdivision of the group into the Johannesburg and Turffontein subgroups. Two diamictite units are present within the group but differ from those in the West Rand Group by being nonglaciogenic debris flow deposits. A further difference to the West Rand Group is the gold endowment of the conglomerate beds, which in certain stratigraphic and paleogeographic positions are exceptionally rich, with 93% of all Witwatersrand gold produced to date coming from the Central Rand Group, especially from proximal areas along the basin edge where a



GEOLOGIC EVIDENCE OF SYNGENETIC Au, WITWATERSRAND GOLDFIELDS, SOUTH AFRICA

number of paleoriver entry points have been mapped out. Most of the conglomerate beds are the result of uplift triggered by intermittent tectonism in the hinterland, followed by periods of tectonic stability and thus low-energy degradation of the conglomerate beds. With the exception of a minor basalt layer in the middle of the group (Bird Member), limited to the northern and eastern parts of the basin, no volcanic activity is recorded in the Central Rand Group. Deepening of the Central Rand Basin toward the center of today’s known domain occupied by Central Rand Group rocks (Fig. 1B) is indicated by (1) the position of the gold fields on the eastern, northern, and northwestern to southwestern margin; (2) paleocurrent directions from all around the currently known surface and subsurface extent of the Central Rand Group (Minter and Loen, 1991); and (3) the spatial distribution of sedimentary facies within a given stratigraphic level, specifically grain size and sorting trends downslope (Minter, 1978). The Central Rand Basin is generally interpreted as a retroarc foreland basin (e.g., Catuneanu, 2001) for the following reasons: (1) dominance of continental clastic sediments, (2) numerous low-angle unconformities within the group, (3) synsedimentary folding of the older Central Rand Group strata toward the western margin of the basin, (4) geochemistry of the sediments indicating derivation from a volcanic arc, (5) progressive unroofing of granites during sediment deposition (Nwaila et al., 2017), (6) large range in paleocurrent directions from all around the basin (Minter and Loen, 1991), and (7) an overall increase in the complexity of detrital zircon age spectra upsection (Kositcin and Krapez, 2004). The corresponding arc developed in the course of amalgamation of the Witwatersrand block with the Kimberley block to the west (Fig. 1) between 2.93 and 2.88 Ga (Schmitz et al., 2004). Shortly after Central Rand Group sedimentation, constrained by the youngest detrital zircon age of ~2.79 Ga, the Witwatersrand rocks were subjected to erosion and peneplanation before being covered by an up to 1,800-m-thick succession of erosion-resistant, partly komatiitic, flood basalt (Klipriviersberg Group) and an up to 5,800-m-thick bimodal suite of felsic and mafic volcanic rocks of the Platberg Group, together constituting the Ventersdorp Supergroup. Recently obtained U-Pb baddeleyite ages for feeder dikes of the Klipriviersberg Group lavas lie between 2789 ± 4 and 2787 ± 2 Ma (Gumsley et al., 2020), representing the currently best available constraint on the onset of this major volcanic episode that covered some 300,000 km2 of the craton. The cause of Ventersdorp volcanism remains a matter of debate but, considering the timing and tectonic position relative to coeval magmatic arc formation in the Pietersburg block (Laurent and Zeh, 2015), it may be speculated that it was related to back-arc rifting. During and subsequent to collisional tectonism along the northern margin of the craton in the course of the 2.69 to 2.60 Ga first Limpopo orogeny (Zeh et al., 2009), erosion led to craton-wide peneplanation before deposition of the Transvaal Supergroup (integrated maximum thickness ~15 km) in the Transvaal Basin on top of the Witwatersrand Basin fill (and equivalent structural basins elsewhere on the craton). This supergroup encompasses four sequences: (1) a laterally limited, volcanosedimentary sequence at the base, possibly

649

syntectonic with respect to ongoing continental collision along the north of the craton; (2) locally auriferous basal conglomerate, quartzite, and shale of the ~2642 ± 2 Ma Black Reef Formation reflecting alluvial, braid-delta, lacustrine to shallow-marine environments, (3) the first major sequence of platform carbonate in an epeiric sea on the craton, followed by iron formation ( ~2.58–2.43 Ga Chuniespoort Group), and (4) sandstone, mudstone, and minor basaltic-andesitic flows of the ~2.43 to 2.06 Ga Pretoria Group (Zeh et al., 2016). Sedimentation in the Transvaal Basin was terminated by bimodal volcanism of the 2061 ± 2 Ma Rooiberg Group and the emplacement of the world’s largest layered intrusive complex, the >350,000-km3 Bushveld Complex between 2056 and 2055 Ma (Zeh et al., 2015), immediately north of the Witwatersrand Basin (Fig. 1). The effect on the Witwatersrand Basin fill must have been profound, certainly in terms of heat flow (Frimmel, 1997) and also in terms of thermally induced fluid mobilization through the country rocks, including those of the Witwatersrand Supergroup (Rasmussen et al., 2007; Gleason et al., 2011). The Witwatersrand gold province hosts in its center the world’s largest and second oldest known impact structure, the 2023 ± 2 Ma (Kamo et al., 1996) Vredefort Dome (Fig. 3). Although the exceptional metal endowment of the Witwatersrand strata cannot be related to this impact because it predates this catastrophic event by 800 to 900 m.y., the impact had some major consequences for the auriferous sediments. The resulting impact melt sheet probably helped to preserve the underlying, much older sediments from subsequent erosion. The impact shattered the surrounding crustal rocks, especially those of the Witwatersrand Basin, whose areal extent corresponds closely to the diameter of the original astrobleme, estimated as some 300 km (Therriault et al., 1997). This enabled the widespread percolation of fluids and partial remobilization of ore components, including gold, which, in turn, led to textural and geochemical features that confused many of those who have attempted to establish the genesis of this gold province and thought that all gold therein is epigenetic. Since then, the Witwatersrand Basin fill has not been subjected to any major tectonothermal overprint. Geology of the Gold Fields Although the Witwatersrand Supergroup is often referred to as an exceptionally well-preserved Mesoarchaean sediment succession, it should be borne in mind that it experienced some degree of metamorphic overprint. Based on metamorphic phase relationships, lower greenschist-facies conditions of 350° ± 50°C at ~2.5 to 3 kbars are indicated for most parts (Wallmach and Meyer, 1990; Frimmel, 1994; Phillips and Law, 1994). Only locally, in the deeper parts of the basin fill now upturned around the Vredefort Dome, is the metamorphic grade higher, reaching amphibolite facies. The P/T gradient in the latter area is distinctly lower, possibly due to heating in the course of Bushveld magma emplacement (Gibson and Wallmach, 1995; Frimmel, 1997). In all other areas, the metamorphic grade is surprisingly uniform across the gold fields, although this should not mask the fact that metamorphism was caused by more than one event, diachronously across the basin: burial, syntectonic dynamic metamorphism during continental collision along the northern

650

FRIMMEL AND NWAILA

Fig. 3. A. Syndepositional structures during Central Rand Group sedimentation, showing entry points of major rivers (paleocurrent directions). B. Vredefort impact-related structures in the Transvaal and Witwatersrand Basins. Ventersdorp and Karoo Supergroup rocks are omitted to better reveal the underlying strata (modified after McCarthy, 2006).

craton margin, syn-Bushveld thermal overprint, and Vredefort-related impact (see Frimmel et al., 2005, and references therein). All of these events triggered varying degrees of fluid flow through the Witwatersrand rocks. The overall structure differs markedly from one gold field to another, thereby highlighting that there is very little structural

control on gold grade. This is illustrated below by a few examples from different parts of the Witwatersrand Basin. Welkom gold field In the Welkom (also referred to as Free State) gold field in the southwestern corner of the Witwatersrand Basin, most of the



GEOLOGIC EVIDENCE OF SYNGENETIC Au, WITWATERSRAND GOLDFIELDS, SOUTH AFRICA

651

Fig. 4. A. West-east section across the western margin of the Welkom gold field (after Minter et al., 1986). B. Detailed section of westernmost part as derived from exploration bore holes (numbers, stippled curves), showing syndepositional (w.r.t. Elsburg Formation) up- and overturning of strata along the basin margin as well as synorogenic thrust and syn-Ventersdorp normal faults; modified after Tweedie, K.A.M. (1986). Original formation names as used by Minter et al. (1986) are shown but note the unified stratigraphic nomenclature now in use as shown in Figure 5.

strata are upturned and, in places, overturned along the western basin margin (Fig. 4). Farther (north)east, the strata dip gently toward the basin center, and syn-Ventersdorp extensional faults dominate the overall structure (Fig. 4). A comprehensive description of the geology of this gold field was provided by Minter et al. (1986). Note that large gaps between the Welkom and other gold fields farther north made correlation at formation level all but impossible, leading Minter et al. (1986) to use gold field-specific formation names. Since then, further exploration drilling in the gap areas made it possible to refine stratigraphic correlation, which led to a proposed stratigraphic subdivision, based on a series of basin-wide disconformities, that attempts to standardize stratigraphic terminology across the entire Witwatersrand Basin (McCarthy, 2006). To avoid proliferation of stratigraphic names, the formation names originally proposed for the Welkom gold field have been changed (see stratigraphic columns in Fig. 5). The overall structure of the gold field is a N-S-trending synform

that is divided along its center by two major faults, the De Bron and Homestead faults (Fig. 4). All orebodies in the Welkom gold field are located in the Central Rand Group, and little is known about the underlying West Rand Group rocks. Within the Central Rand Group, a number of low-angle unconformity-bound sediment packages can be distinguished, most of which represent individual alluvial fans or reworked parts of underlying fan deposits. Each unconformity is overlain by an auriferous and uraniferous conglomerate, constituting the orebodies. From the northsouth section shown in Figure 6, relative to the position of the mine lease areas, it can be seen that the younger formations onlap onto older ones, reflecting an expansion of the area of deposition until the Eldorado Formation (upper Turffontein Subgroup) times, and an overall northward dip of the stratiform reefs. Detailed mapping of sedimentary facies and paleocurrent directions (Minter et al., 1986) made it possible to differentiate not only between different reefs at different

Fig. 5. Stratigraphy of the Witwatersrand Supergroup and attempted correlation between the various gold fields. Modified after McCarthy (2006), from Frimmel (2018).

652 FRIMMEL AND NWAILA



653

GEOLOGIC EVIDENCE OF SYNGENETIC Au, WITWATERSRAND GOLDFIELDS, SOUTH AFRICA

Paleoflow direction Pebble size Tectonic displacement Paleochannel

Gold grade high (15-60 g/t) medium (5-15 g/t) low (1-5 g/t)

Fig. 6. The Welkom gold field. A. Current mine lease areas, depositional limits of the coeval Basal and Steyn placers as well as paleocurrent directions (rose diagrams) for the various placer units (based on Minter et al., 1986). B. North-south cross section, illustrating progressive erosional truncation of lower stratigraphic units toward the south of the gold field. C. Subcrop maps of the Basal and Steyn placer paleosurfaces in the Welkom gold field, illustrating the principal drainage pattern (as derived from mean cross-bed azimuths), decrease in pebble size from proximal to distal positions, and the distribution of gold grade. Based on Minter et al. (1986) and W.E.L. Minter (unpub. data).

stratigraphic levels but also between different, penecontemporaneous paleoriver systems and fans that had entered the basin from the west (Fig. 6). The stratigraphically lowest orebody is the Beisa reef, a pebble layer above a paleosurface that is also locally marked by a thin layer of kerogen. Due to the general dip toward the north(east), this relatively low-grade but uraniferous reef is reached at mineable depths only in the south of the gold field (Fig. 6). The most productive orebodies have been the Basal and Steyn reefs at the base of the Krugersdorp Formation. They represent two shifting systems of braided rivers, the former having transported sediment to the east, the latter to the northeast (Fig. 6). They drained two different source areas, reflected by an oligomictic pebble assemblage in the Basal reef and a polymictic assemblage in the Steyn placer. Both contain large amounts of uranium, mainly in the form of semirounded detrital uraninite grains, the chemistry of which is distinctly different in the two reefs (e.g., Ta/Nb) and thus provides an excellent provenance indicator (Frimmel et al., 2014). Fluvial channels degraded the fan surface of the underlying formation to depths of as much as 4 m, which led to the concentration of placer minerals, such as widespread rounded pyrite, chromite, zircon, leucoxene (after ilmenite), uraninite, and

gold, on bottom-scour surfaces and also bottom-lag surfaces where multiple channel sequences were stacked during slow aggradation. In places, thin kerogen layers are preserved on the scour surfaces. The maximum pebble (including rounded pyrite) size decreases toward the more distal deposits from, on average, 40 mm in proximal to, on average, 20 mm in the quartzite-dominated distal facies (Minter et al., 1986). Along the same direction, the U/Au ratio increases systematically. The Saaiplaas reef is composed of sandy fluvial channel fills representing a network of semiperennial braided streams (Buck, 1983). Volumetrically subordinate conglomerate contains pebbles derived from erosion of the Steyn placer higher up the paleoslope. Gold and uraninite are limited to trough crossbedding and foresets and trough scour surfaces that are rich in rounded pyrite. Locally, Au- and U-rich kerogen seams occur in the lower parts of horizontally laminated quartzite. However, the Saaiplaas reef lacks the lateral continuity and sufficient thickness of the conglomerates to be of major economic significance. The Leader reef is part of a sequence that fines upward to shale of the Booysens Formation. It is a composite placer, similar to the Basal/Steyn reefs, but is only locally mineralized where it truncates older reefs, especially the Basal reef.

654

FRIMMEL AND NWAILA

It represents a braided channel system that spread out toward the east like a fan. Again, an increase in U/Au has been noted down the paleoslope (Minter et al., 1986). Further disconformities or low-angle unconformities higher up in the succession are all marked by conglomerate, which can be variably mineralized where they reworked parts of the older reefs. Of these, the Elsburg (Eldorado) reefs are of particular significance as they represent the most important reserve at Target mine, one of very few highly mechanized Witwatersrand mines with a projected life exceeding 25 years.

Klerksdorp gold field The Klerksdorp gold field on the northwestern rim of the Witwatersrand Basin contains gold and uranium deposits in both the West Rand and Central Rand groups as well as the younger Ventersdorp Supergroup (see below). Most of the ore is hosted by Central Rand Group rocks, which are almost completely concealed beneath younger cover of the Ventersdorp Supergroup and, in general, dip toward the southeast and the basin center (Fig. 7A). West of the town of Klerksdorp, rocks of the West Rand Group crop out and have been

Fig. 7. The Klerksdorp gold field. A. Current mine lease areas, isopachs for the Gold Estates Member of the Kimberley Formation, illustrating synsedimentary folding, and a northwest-southeast cross section, illustrating progressive erosional truncation of lower stratigraphic units toward the southeast of the gold field (based on Antrobus et al., 1986); also shown are measured paleocurrent directions (rose diagrams) in the Vaal reef, its hanging wall, the Zandpan Member, and its footwall, the Mapaiskraal Member (after Minter, 1972). B. Vaal reef isopach plan, showing relationship between reef thickness and gold grade in mined out subcrop of the Vaal reef. C. Gold isocon trend surface plan (based on one million channel samples), also showing mapped paleochannels. D. Diagrammatic west-east section through the Vaal reef (for position see line A-B in (B) and (C)), relating reef geology and footwall shape to gold grade; length of section is ~7 km (from Frimmel et al., 2005, based on Minter, 1976).



GEOLOGIC EVIDENCE OF SYNGENETIC Au, WITWATERSRAND GOLDFIELDS, SOUTH AFRICA

mined intermittently since 1887 for gold, and later for uranium. The Central Rand Group is present in subcrop in an intensely faulted sub-basin to the east of the Klerksdorp gold field. Syn-Central Rand Group thrusting from the northwest led to periodic uplift and erosion of lower units, evident in numerous low-angle unconformities and disconformities. Southeast-directed compression also caused synsedimentary folding, as shown by isopachs of the Kimberley Formation in Figure 7A. After deposition of the Witwatersrand strata, the gold field was affected mainly by syn-Ventersdorp extension, resulting in numerous SE-dipping listric faults and NW-dipping antithetic normal faults, and to a lesser extent by syn- to post-Transvaal faulting. The various erosion surfaces within the Central Rand Group are typically overlain by prograding braid-plain deposits. At the base of the group is the Ada May reef, a poorly mineralized unit that varies from a 1.5-m-thick conglomerate to a kerogen layer. By far the most important orebody has been the Vaal reef, an oligomictic pebbly quartz arenite bed on top of an angular unconformity at the base of the Krugersdorp Formation. Isopachs of the reef and measured paleocurrent directions indicate a meandering drainage pattern incised into a regional truncation surface down a SE-dipping slope (Fig. 7B). The reef varies from >1-m-thick channel fills to a thin winnowed pebble lag that is marked by abundant ventifacts, reflecting an eolian deflation surface. From differences in the pebble assemblage (predominantly vein quartz, subordinately chert), Minter (1976) distinguished two different sources and thus two paleoriver systems. Kerogen layers are a conspicuous feature of this reef and typically formed as columnar, 2- to 50-mm-thick seams on the already consolidated surface of the underlying unconformity, less commonly on sedimentary partings within the reef. In the latter case, they are draped over the tops of pebbles, suggesting derivation from in situ microbes. Although several auriferous conglomerates exist in the upper Central Rand Group, they are of limited economic interest and reflect local reworking of essentially the Vaal reef. Noteworthy, however, is the Crystalkop (or C) reef for its abundance of ventifacts in the form of dreikanters (Minter, 1999). The second-most important orebody in the gold field is the Ventersdorp Contact reef at the base of the Ventersdorp Supergroup (see below). Carletonville and West Rand gold fields Along the northern margin of the Witwatersrand Basin, a string of mines has been developed, comprising the Carletonville gold field (also referred to as the West Wits Line) to the south of Carletonville and the West Rand gold field stretching from Randfontein to Fochville. They are separated from one other by the West Rand fault (Fig. 8A). The western sector shows a general east-northeast strike whereas the eastern sector, separated by the Bank fault, is dominated by north-northeast strike. The two gold fields are overlain by rocks of the Transvaal Supergroup and intervening Ventersdorp Supergroup (see profile in Fig. 8A). Folding, evident from the traces of marker beds in subcrop, and much of the N-S-striking faulting were due to syn-Central Rand Group compression from the west. The major north-south faults, however, also contain an extensional component that is related to subsequent syn-Ventersdorp normal faulting. The

655

complex deformational history, unravelled from field observations by Myers et al. (1990), could be confirmed through three-dimensional seismic surveys by Manzi et al. (2013), who also recognized the development of an E-verging fold-thrust belt at the end of Central Rand sedimentation. Locally developed bedding-parallel mylonite zones are, however, younger and probably related to the Vredefort impact (Killick et al., 1988). Dextral wrench faulting was the last significant deformation and is post-Transvaal in age. Similarly, as in the gold fields described above, several low-angle unconformities can be distinguished, each of which is overlain by a conglomerate, in places auriferous and uraniferous, with progressive truncation of the underlying units toward the north (Fig. 8A). Conglomerate beds in the upper West Rand Group are of only limited economic significance. The richest orebodies have been the Carbon Leader and Main reefs on a disconformity at the base of the Main Formation in the lower Central Rand Group in the Carletonville and West Rand gold fields, respectively. Today, the Carbon Leader reef is largely mined out. It consisted of a 1- to 30-mm-thick, laterally continuous kerogen layer, overlain by a thin (typically 10 cm) quartz pebble conglomerate (mean pebble size 11.3  mm, with little variation) and intercalated quartzite, which emanated from two fluvial paleodrainage systems (Buck and Minter, 1985) with sediment transport to the south and southeast (Fig. 8A). The kerogen layer was particularly rich in gold, accounting for as much as 50% of the total gold tenor of the reef in places (Hallbauer and Joughin, 1973). Thicker, clast-supported conglomerate beds or lenses, representing lag gravels on former scour surfaces and gravel bars in braided rivers, contain high concentrations of gold and uranium. In places, kerogen layers drape top surfaces of pebbles and former sand bars. Several major SE-directed erosion channels, up to 2,200 m wide and 100 m deep, cut into the Carbon Leader and Main reefs down to shale of the Roodepoort Formation (West Rand Group). One such channel is at the western margin of the Carletonville gold field, whereas the other lies between the Driefontein mine and southern end of the Kloof mine (Fig. 8B-D). These erosion channels are largely filled with shale that corresponds to that of the so-called Green Bar, a chloritoid-rich shale above the Carbon Leader reef, but also with quartzite and conglomerates. Overall, the gold grade in these shale-filled channels is low compared to the adjacent conglomeratic units and kerogen seams (Fig. 8E). Genetically critical is the observation that the channel fill contains detrital clasts of mineralized Carbon Leader fragments (Buck and Minter, 1985), which clearly indicates that gold mineralization must be older than deposition of the upper parts of the Blyvooruitzicht Formation. The Green Bar is overlain by quartzite and, in turn, by the Middelvlei reef, a 7-m-thick set of multiple conglomerates with intercalated quartzite. In places, it contains economic grades. Similarly, locally mineralized conglomerate horizons exist higher up in the succession, such as the Livingstone reef at the base of the Luipardsvlei Formation. Discontinuous conglomerate lenses also mark the Kimberley Formation as well as the upward-coarsening Elsburg and Mondeor formations above the stratigraphic marker of the Booysens Formation shale, and these have been mined at several places. The Elsburg reefs carry the hopes for the future as they are

Fig. 8. The West Wits Line and West Rand gold fields. A. Current mine lease areas (former mine names in brackets), the traces of subcrop of selected Witwatersrand Supergroup stratigraphic markers and reefs and major faults; also shown is a north-south cross section, illustrating progressive erosional truncation of lower stratigraphic units toward the north of the gold field (based on Engelbrecht et al., 1986) and measured paleocurrent directions (rose diagrams) for the Carbon Leader, Middelvlei, and Livingstone reefs (Minter and Loen, 1991). B. Plan of sedimentary facies across the Carletonville gold field (modified from Tucker et al., 2016). C. Cross section through a paleoerosion channel cutting through the Carbon Leader deep into the bedrock at the Driefontein mine. D. Facies types, and (E) gold grade distribution in the Driefontein lease area.

656 FRIMMEL AND NWAILA



GEOLOGIC EVIDENCE OF SYNGENETIC Au, WITWATERSRAND GOLDFIELDS, SOUTH AFRICA

Fig. 9. A. Geologic map of the West, Central, and East Rand gold fields with post-Witwatersrand cover removed, also showing operating and defunct mine lease areas and major faults. B. Plan showing control of NW-SE-trending folds on position of channels and ore shoots in the Main Reef Leader (after Tucker et al., 2016).

657

658

FRIMMEL AND NWAILA

the prime target at the South Deep mine, one of few (highly mechanized) mines with >25 years life in the entire province. Central, East, and South Rand gold fields Located immediately south of Johannesburg (Fig. 9), the Central Rand gold field supported some 46 mines and produced >8,000 t of gold but is now largely mined out. Compared to the other gold fields, the Central Rand is structurally simple, with the Witwatersrand strata dipping moderately to the south. However, toward the Rietfontein fault (Fig. 9), a syndepositional (with respect to Central Rand Group), sinistral oblique-slip reverse fault (Myers et al., 1990), the West Rand Group strata steepen to the south, whereas the Central Rand Group units thin northward toward the fault. During Ventersdorp extension, this fault was reactivated as an oblique-slip normal fault. Moreover, post-Transvaal, N-vergent thrusting has been documented but without significant displacement of units (McCarthy et al., 1986). Similarly, the adjacent and contiguous East Rand gold field, where some 28 mines produced ~10,000 t of gold, is also largely mined out. It represents a separate sub-basin marked by NW-striking, syndepositional open folds (Fig. 10A). The principal gold-producing units were the Main reef, Main reef leader, and South reef (also referred to as the Nigel reef). They overlie an erosional unconformity that cuts into the footwall as deep as the Jeppestown Subgroup in the East Rand sub-basin. Analogously, in the Central Rand gold field, erosion channels beneath the Main reef leader cut through the Main reef into the Jeppestown Subgroup and are largely filled with shale, similar to that in the Carletonville and West Rand gold fields. The South reef was deposited on a degradational braid plain by SE-flowing rivers. Gold was concentrated in narrow pay shoots that represent the fluvial channels. No kerogen seams have been reported from the Central and West Rand gold fields. Farther south, separated by a basement dome, the Devon dome (Fig. 1B), is the South Rand gold field in which the South/Nigel reef constituted the main orebody. Today, the focus is on several conglomerate bodies (labelled UK#9A/B) in the Kimberley Formation (collectively referred to as the Kimberley reefs), currently being explored in the Burnstone Project. Extensive drilling and past underground sampling revealed a complex network of northwest-draining fluvial paleochannels with essentially two main tributaries. Economic gold grades (6–23 g/t) are confined to paleochannels, whereas interchannel deposits contain an order of magnitude less gold (Fig. 10B). Interestingly, only the northern paleoriver system is rich in gold, which reflects the significance of source area for the ore grade. Evander gold field To the northeast of the Central Rand Basin, separated by the Devon dome, is a further small basin whose fill constitutes the Evander gold field (Fig. 1B). The bulk of production stems from the period 1958 to 1996 and came from four mines (Winkelhaak, Bracken, Leslie Gold, and Kinross), totalling ~1,342 t of gold. Currently, efforts are underway by Taung Gold to revive the gold field. Little is known about the West Rand Group there, but from borehole intersections it appears as if the entire Jeppestown

Subgroup had been eroded prior to Johannesburg Subgroup deposition (Tweedie, E.B., 1986). The Central Rand Group, which reaches a maximum thickness of not more than 650 m here, is in parts overlain by rocks of the Ventersdorp and Transvaal supergroups, and all of them are covered by rocks of the Karoo Supergroup (Fig. 10C). The regional disconformity at the base of the Central Rand Group is overlain by distal quartzite with minimal conglomerate (and little economic potential), fining upward into marine shale of the Booysens Formation. Sediment transport was mainly from the northeast. Within the overlying Kimberley Formation, the depositional setting changed from shallow marine to fluvial, with a farther major disconformity, on top of which rests an up to 3-m-thick conglomerate, the Kimberley reef, which is the only orebody in the gold field. Sediment transport from the southwest at that level points to major uplift of the southern domain. The full extent of the gold field to the north remains unknown because of the general northerly dip of the strata to inaccessible depths (Fig. 10C). Whereas the style of mineralization is similar to that in the other gold fields, the Evander gold field differs in its metamorphic history, which is well displayed by mineralogic changes in the Booysens Formation shale. Toward the northeast of the Evander gold field, a series of metamorphic zones have been differentiated, ranging from lowest greenschist to hornblende-hornfels facies (Tweedie, E.B., 1986). Most likely the cause of this contact metamorphic overprint was the emplacement of the Bushveld Complex, whose southern lobe extends under younger Karoo Supergroup cover to the northeastern vicinity of the Evander gold field. Post-Witwatersrand auriferous conglomerates Conglomerates with economic gold grades are not limited to the Witwatersrand Supergroup but are also present at the base of the two younger megasequences, the Ventersdorp and Transvaal supergroups. They have contributed significantly to the overall gold production of Witwatersrand mines and their mode of occurrence and stratigraphic position provide useful information on the metallogeny of the entire province. The northwestern portion of the Central Rand Basin fill is unconformably overlain by thick (up to 1,830 m), predominantly komatiitic flood basalt of the Klipriviersberg Group (lower Ventersdorp Supergroup). At the base of this group are immature conglomerates with minor quartzite of the Venterspost Formation. Toward the east, the angle of unconformity decreases and the basal conglomerate thins out. Thus, in the Evander gold field, only a very thin, uneconomic conglomerate bed rests paraconformably on the Witwatersrand Supergroup strata. The elimination of the Central Rand Group from syn-Central Rand Group anticlinal paleohighs (e.g., in the eastern section of the Driefontein mine in the Carletonville gold field), prior to deposition of the Venterspost Formation, indicates a major hiatus. By the time of Venterspost deposition, the underlying sediments of the Central Rand Group had already been lithified (Hall, 1997). Thus, the Venterspost Formation represents a gravel sheet, related to a number of depositional events, on a variably incised old pediment (Fig. 11). Various sedimentary facies have been distinguished; they are controlled by the underlying paleotopography and represent former terraces and intervening broad valleys with fluvial



GEOLOGIC EVIDENCE OF SYNGENETIC Au, WITWATERSRAND GOLDFIELDS, SOUTH AFRICA

659

Fig. 10. A. Geologic map of subcropping footwall of the South reef in the East Rand gold field (after De Jager, 1986). B. Paleochannels in the UK9A and UK9B reefs (Kimberley Formation) in the Burnstone project, South Rand gold field, as mapped by conglomerate thickness and distribution of gold grade, illustrating maximum gold concentrations along narrow, high-energy channels in both the UK9A river system and in partly eroded remnants of the UK9B river system (from Stewart, 2016). C. North-south profile across the Evander Basin (modified after Tweedie, E.B., 1986).

channels as well as proximal and distal alluvial fans (Henning et al., 1994; Fig. 12A). In the waning stages of sedimentation, first komatiitic lavas erupted onto the still wet, soft sediments (e.g., emplacement of pillow lava onto unconsolidated

pebbles; see Tucker et al., 2016, their fig. 32) before more widespread volcanism over large parts of the craton put an end to sedimentation. All of these features clearly indicate that the Venterspost Formation is the basal unit of the Ventersdorp

660

FRIMMEL AND NWAILA

Fig. 11. Morphostratigraphic model for the Venterspost Formation across the Carletonville and West Rand gold fields, disregarding postdepositional block faulting (modified after Germs and Schweitzer, 1994).

Supergroup and not part of the Witwatersrand Supergroup (Hall, 1997). Along the northwestern and western margins of the Central Rand Basin, the basal conglomerate of the overlying Venterspost Formation can attain 4 m in thickness and represents a high-grade orebody, referred to as the Ventersdorp Contact reef. This orebody has contributed as much as 7.6% of the total Witwatersrand gold production. From extensive underground mapping and sampling it appears that economic gold grades are spatially linked to areas in which the Ventersdorp Contact reef directly overlies auriferous Witwatersrand reefs (Engelbrecht et al., 1986). A high-energy regime is implied by the very large pebble size, on average 2 to 5 cm and, depending on facies type, reaching a maximum of 45 cm. The pebble assemblage varies strongly with facies, with quartzite, quartz, and various shales dominating. Rounded pyrite and pyrrhotite pebbles in hydraulic equilibrium can make up as much as 6 vol %. The complexity of sedimentary facies (Fig. 12) is a transient feature and would have degraded into a uniform conglomerate layer (with likely diluted gold grades) if it had not been for the sealing by erosion-resistant volcanic rocks shortly after sedimentation. At the base of the Neoarchean to Paleoproterozoic Transvaal Supergroup is the 2642 ± 2 Ma Black reef formation, a typical basal siliciclastic unit developed over a largely peneplaned surface. At its base, it consists of a largely quartzitic unit with locally developed pyritic quartz pebble conglomerate and scattered pebbles, representing former fluvial channels, fining upward into siltstone and, in most places, carbonaceous shale. Locally, the conglomerates can be auriferous and have been mined as the Black reef at several places in the East

Rand and, to a lesser extent, in the West Rand and Klerksdorp gold fields. The past production of this reef, 40 t of gold, is, however, miniscule in comparison to the main Witwatersrand reefs and the Ventersdorp Contact reef. Although the areal extent of the Black reef formation over the Kaapvaal craton is wide (Fig. 1A), elevated gold contents have been reported only from areas in which the Black reef conglomerates were deposited in channels that had cut into sedimentary bedrock near exposed gold-rich Witwatersrand reefs, such as the Main, Bird, and Kimberley reefs or the Ventersdorp Contact reef. Controls on Mineralization The almost exclusively underground exploitation of this gigantic gold province would not have been so effective, or even possible, if the miners did not have reliable geologic leads, based on sedimentology, in their daily underground exploration. There is no doubt whatsoever that the principal control on the gold (and uranium) mineralization in the Witwatersrand gold fields is sedimentologic. This holds true on almost all scales, except for the microscopic. The first-order control is host lithology—all orebodies are within conglomerates of various sedimentary facies or pebbly arenite. Mineralization was truly strata-bound and stratiform with the ore being bound to specific conglomerate horizons that can be correlated across the entire basin over a distance of >300 km (Fig. 5). Arenitic metasedimentary rocks (quartzite) in the hanging- and footwalls of the conglomeratic orebodies typically contain two to three orders of magnitude less gold. Background gold concentrations in Witwatersrand shale are close to Clarke values, i.e. between 1 and 6 ppb (Nwaila and Frimmel, 2019). Similar values have been reported from Central



GEOLOGIC EVIDENCE OF SYNGENETIC Au, WITWATERSRAND GOLDFIELDS, SOUTH AFRICA

661

Fig. 12. Sedimentary facies and gold grade in the Ventersdorp contact reef: A = schematic section as mapped in the Carletonville gold field; B = plan of facies distribution in the Driefontein mine; C = channel width; D = gold grade distribution; numbers refer to mines as in Figure 8E.

Rand quartzite (Law et al., 1988). A huge amount of geospatial data (estimated >11M) have been collected by different mining companies over decades from all Witwatersrand gold operations, largely from conglomerates, some of which have already been digitized. The average sampling interval in Witwatersrand mines is 5 m. For many reefs >300,000 sample points are available but a block size with a search area, including all of these, invariably yields no interpretable pattern. However, if the search radius is reduced to 15 m, a positive correlation between gold grade and clast size becomes apparent, as exemplified by the Carbon Leader, Middelvlei, and Ventersdorp Contact reefs (Fig. 13). This clearly illustrates that even the average gold grade (irrespective of whether calculated as g/t or cm.g/t) between footwall/hanging wall and actual orebody (conglomerate) within such a radius differs by more than an order of magnitude. The somewhat elevated (relative to background) gold concentrations in the hangingand footwalls are explained by (1) sampling bias toward the immediate contacts with the conglomeratic orebodies during mining operations, (2) the close spatial proximity of more than one reef in some places, and (3) the preferential sampling of

quartzite bodies that are intercalated with, and belong to the same depositional events as, the conglomerates. On a basin-wide scale, proximity to the former basin margin is the overriding control on mineralization. In the producing gold fields along the northern, northwestern, and southwestern margins of the currently known extent of the Central Rand basin fill (Fig. 1), some 8 vol % of the total Central Rand Group consists of conglomerate. In contrast, in the upturned Witwatersrand strata around the Vredefort Dome, this figure is not more than 1 vol % and conglomerates are effectively absent along the southeastern side. This clearly illustrates that the northern, northwestern, and southwestern limits of the currently known extent of Central Rand Group rocks corresponds closely to the margin of the original depositional basin that deepened toward the southeast. The Evander gold field is an exception as it represents a separate sub-basin fill. All around the northern and western margins of the Central Rand Basin, the strata dip to the basin center. Although the original stratification was complicated by syn- and postdepositional folding, local thrusting, and normal faulting, the overall monoclinal geometry, with steep dips at the margin and flattening

662

FRIMMEL AND NWAILA

Fig. 13. Average gold grade vs. clast size in conglomerate ore, footwall, and hanging wall of the Carbon Leader reef (A), Middelvlei reef (B), and Ventersdorp Contact reef (C), based on >300,000 individual data points for each; to obtain interpretable and low-noise plots, raw data were spatially averaged over clast-size bins. Optimal autocorrelation between Au and clast size was achieved with a neighborhood search radius of 15 m as defined using Euclidian distance. Data Z-scores (Zi) were calculated to trim outliers and reduce nugget effect using Zi = (x – μ)/σ where μ = population mean and σ = population standard deviation. Samples with Zi >4 were excluded.

of the strata toward the basin center, must have already been a primary feature at the time of basin development. The position of the gold fields relative to the architecture of the Central Rand Basin is clear testimony that the mineralization was restricted to the most proximal domains of that basin. More distal regions, not to mention open marine environments, are effectively barren, irrespective of their tectonic and/or metamorphic history. On a more local, gold field-wide scale, the most important control on gold grade is sedimentary facies. In terms of gold endowment, two facies types stand out as particularly rich: first, thin pebble lags representing, at least in places, eolian deflation surfaces, many of which are covered by kerogen seams, and second, conglomeratic, high-energy fluvial channel deposits. This situation is beautifully illustrated by the largely mined-out Vaal reef in the Klerksdorp gold field (Fig. 7) and Carbon Leader reef in the Carletonville gold field (Fig. 8), where the highest grades were achieved in the kerogen-rich deflation surfaces, followed by the channel facies. Gold grade in the mined-out sections of the Carbon Leader reef averaged 25 g/t and reported grades in the currently mined parts and remaining reserves are between 17.5 and 22.5 g/t (AngloGold Ashanti, 2014). This kerogen seam-dominated reef constituted the richest orebody after the Main reef leader, a similarly thin sheet-like conglomerate that has been mined for >40 km along strike, with grades in excess of 40 g/t Au (Tucker et al., 2016). Particularly noteworthy is the confinement of gold ore to a sedimentary unit hardly thicker than 10 cm over such an extensive strike length, independent of whatever crosscutting structures may be present. In the channel facies, gold grade is directly related to reef package thickness, pebble size, proportion of conglomerate, and sorting, which equates to the sediment load and energy of a given stream. This is evident from all gold fields where detailed underground mapping and sampling over many decades made it possible to reconstruct the positions of former streams and channels. Without exception, in all areas the highest gold concentrations overlap spatially with these channels (see Figs. 7–9, 10B, 12). This mineralized-facies type exists next to the kerogen seam/thin conglomerate sheet facies in the oldest and richest reefs of the basal Central Rand Group and is characteristic of the younger reefs of the Turffontein Subgroup (e.g., the Elsburg reefs). In the latter position, the extent of sedimentary reworking of older reefs becomes a critical controlling parameter. This applies also to the post-Witwatersrand reefs at the base of the Ventersdorp and Transvaal Supergroups, where Witwatersrand reef exposure and energy level of reworking streams on the eroded angular unconformity are the critical grade-determining factors. A further decisive control on gold grade is the presence of erosional unconformities. All reefs rest on such paleosurfaces that represent sedimentary hiatuses, most of which are angular unconformities with some of the respective bedrock having been tilted prior to sedimentation, thus exposing parts of the immediately underlying stratigraphy, including older reefs, to erosion. In the view of the conflicting hypotheses for the genesis of the gold mineralization (see below), the paucity of orebodies that cut across stratigraphy cannot be emphasized enough. Although the gold fields are dissected by numerous



GEOLOGIC EVIDENCE OF SYNGENETIC Au, WITWATERSRAND GOLDFIELDS, SOUTH AFRICA

faults, veins, and dikes of variable age, these structures had no significant control on the primary mineralization. Locally, however, increase in grade can be discerned where a given reef is hydrothermally altered near a fault, and visible gold is most often observed in conglomerate-hosted quartz veins. An example is the Leader reef at the Tshepong Mine (Welkom gold field), which typically contained 100 to 200 cm.g/t Au but >1,000 cm.g/t where it was chloritized near the intersection with a low-angle thrust (A. Brown, pers. commun., 2015). However, the extent of such hydrothermally mineralized zones is too small to be of economic significance. Nevertheless, evidence of bedding-parallel fluid flow in conglomerate beds is common in the form of sericitization and chloritization (e.g., Gartz and Frimmel, 1999). Application of purely postdepositional hydrothermal models has so far failed to identify a single orebody. In contrast, daily mapping of sedimentary facies distribution, combined with delineating structures that offset auriferous conglomerate units, has been extremely successful in recovering some 53,000 t of gold over >130 years of mining and in identifying remaining reserves. On the scale of a stope face, the main control on grade remains lithology (i.e., conglomerate and, to a lesser extent, intercalated quartzite). Visible gold is rare and most commonly found in crosscutting quartz-sulfide veins, as thin platelets on the surface of columnar kerogen structures, rarely as heavy mineral concentrate in pebble lags on top of degradation surfaces or at the base of clast-supported conglomerate units, and on crossbedded foresets, bottomsets, and coset boundaries in the form of rounded, disk-shaped to toroidal micronuggets (Minter et al., 1993), recently revealed in situ by microXRD-CT methods (Hölzing et al., 2015; Frimmel, 2018). In places where quartz veins carry gold, the veins invariably cut reefs in which the gold grade in the conglomerate exceeds that in the vein. The gold content in the conglomerate does not increase with proximity to the vein but the gold content in the vein decreases away from the intersected reef. It is clear from these observations that gold in veins is the product of remobilization of originally conglomerate-hosted gold and not vice versa. There is also a temporal control on gold grade. Conglomerates that are older than 2.9 Ga, that is those in the West Rand Group and underlying Dominion Group, locally contain somewhat elevated gold contents, but their overall gold tenor is miniscule compared to the conglomerates that are younger than 2.9 Ga, even if they occur in similar tectonic positions. The time at around 2.9 Ga represents a gold megaevent with close to 80% of all Witwatersrand gold having been (or still being) located within 2.90 to 2.86 Ga conglomerates of the lower Central Rand Group and both gold endowment and ore grade decreasing exponentially from 2.90 Ga with decreasing age (Frimmel, 2018). The marked difference in gold content between the West Rand and Central Rand Group conglomerates goes hand in hand with two major mineralogic and lithologic differences: (1) the highest gold grades in the lower Central Rand Group are spatially associated with kerogen, which is conspicuously absent in the West Rand Group; and (2) the gold-rich conglomerates lack detrital feldspar, which is in stark contrast to the West Rand Group metasedimentary rocks, in which detrital feldspar is common (Law et al., 1990). This latter difference reflects a paleoclimatic control

663

(Frimmel, 2018). High chemical indices of alteration and weathering, especially in the footwall of paleoerosion surfaces (Frimmel and Minter, 2002) but also in marine shale of the Central Rand Group (Nwaila et al., 2017), attest to intense chemical weathering under a warm and aggressive (acidic) atmosphere at that time (Frimmel, 2005), whereas the abundance of detrital feldspar, the presence of the glaciogenic diamictite and overlying banded iron formation in the West Rand Group indicate cold conditions prior to 2.90 Ga (Smith et al., 2013). The postulated warmer climatic conditions during Central Rand Group times would have favored microbial growth and thus explain the temporal restriction of kerogen seams to that time. Metallogenic Concepts Since the discovery of the Witwatersrand gold fields their metallogeny has been a matter of intense debate, which has mainly revolved around the question of whether mineralization was syn- or epigenetic. Early on, following the recognition of a strong stratigraphic and sedimentologic control on the ore, it was hypothesized that the Witwatersrand deposits represent ancient placers (Mellor, 1916), but soon thereafter, a variety of epigenetic, hydrothermal models were proposed. Initially, a magmatic source of the postulated ore fluids was envisaged (Graton, 1930), later a metamorphic (Phillips and Myers, 1989), with parallels drawn to orogenic gold deposits (Phillips and Powell, 2011). Sometimes it was suggested that gold mineralization was connected with the 2.056 Ga Bushveld event (Phillips and Law, 1994), at others with the 2.785 Ga Ventersdorp Supergroup volcanism (Phillips et al., 1997). Much of this debate on syngenesis versus epigenesis hinges on the apparently conflicting interpretation of large- and smallscale observations. At the latest since the detailed petrographic studies by Ramdohr (1958), it has become evident that much of the gold occurs in late paragenetic textural positions and in association with a range of metamorphic and hydrothermal phases, notably secondary sulfides (pyrite, galena, chalcopyrite, gersdorfite), chlorite, pyrophyllite, or pyrobitumen (e.g., Feather and Koen, 1975; Frimmel, 2005; Fuchs et al., 2016a). In rare instances, gold inclusions within postdepositional fluid inclusions in quartz overgrowths around detrital quartz grains bear evidence of gold precipitation from a hydrothermal fluid (Frimmel et al., 1993). While agreement exists that much of the Witwatersrand gold is, in its current form, the product of hydrothermal precipitation, opinions on the source of this hydrothermal gold diverge. Was the source syngenetic gold that had existed already within the host conglomerates? Or was the gold introduced into its host rocks by postdepositional fluids from a distant external source? A wealth of observations speaks for the former. This includes the presence of clearly mechanically deformed micronuggets with minute secondary gold overgrowths (Minter et al., 1993; Hölzing et al., 2015). The coexistence of detrital particles with postdepositional secondary phases on a microscale is analogous to most other ore components, such as the coexistence of detrital, syn- and postdepositional pyrite (England et al., 2002; Guy et al., 2010; Large et al., 2013), detrital and secondary uranium minerals (Rantzsch et al., 2011; Depiné et al., 2013; Frimmel et al., 2014), as well as in situ kerogen layers and secondary pyrobitumen (Spangenberg and Frimmel, 2001; Mossman et al.,

664

FRIMMEL AND NWAILA

2008; Frimmel and Hennigh, 2015). All of the large-scale controls on ore summarized in this paper support syngenetic rather than epigenetic gold concentration. None of the hydrothermal models can explain the geometry of the orebodies and the principal controls on ore grade. If the gold (and other ore components, such as uranium) had been introduced into the host conglomerates by fluid flow along unconformity surfaces and into overlying conglomerates, an overall antiformal geometry of the orebodies, as for instance in orogenic gold systems, would be expected and not the synformal geometry present in the Witwatersrand gold fields, as already pointed out by Groves et al. (2003). Furthermore, many of the mineralized conglomerates are matrix supported, with an arenitic matrix similar in texture and composition to arenitic rocks in the hanging- and footwalls. The permeability in these conglomerates, especially tens to hundreds of millions of years after sediment deposition, should have been similar to that of the over- and underlying arenitic rocks. Consequently, perfectly channelled postdepositional fluid flow exclusively through conglomerate beds is highly unlikely. The features presented in support of hydrothermal models are local phenomena, not those that control gold distribution on a mine- to district-scale. The apparent conflict between large- and small-scale observations, such as those under a microscope or the local presence of gold in veins and dikes that cut a conglomeratic orebody, can be resolved by assuming short-range (micrometer- to meter-scale) mobilization of syngenetic gold by postdepositional fluids—a concept that has become known as the “modified paleoplacer model.” The biggest challenge to the (modified) paleoplacer model comes from an apparent lack of a viable source for some 90,000 t of gold (Phillips and Powell, 2011). The recognition of primary fluvial dispersal patterns of gold in the conglomerates in all gold fields (e.g., Figs. 6–10, 12) necessitates some source that delivered coherent gold particles. As correctly pointed out by Phillips and Powell (2011), erosion of greenstone-hosted, vein-type deposits cannot explain the Witwatersrand gold, simply because the required amount of such vein-type deposits in the hinterland would have been entirely unrealistic. Apart from this mass-balance problem, the composition of greenstone-hosted gold in the Kaapvaal craton (Barberton greenstone belt) is markedly different from that of Witwatersrand gold (Frimmel et al., 2005), and evidence of detrital quartz clasts with primary gold inclusions and of larger gold nuggets—typical features of younger placer gold derived from orogenic gold deposits—is missing. As an alternative, Hutchinson and Viljoen (1987) proposed erosion of hydrothermal exhalative deposits, similar to volcanic-hosted massive sulfide deposits, along the former basin margin as a possible source. No evidence exists, however, of such deposits, neither in the bimodal volcanic suite of the Dominion Group nor anywhere in the West Rand Group. Erosion of such hypothetical deposits should have left behind not only gold but also elevated base metal concentrations, which are not evident in the Witwatersrand sediments. A third possibility suggested in the past is derivation of the gold from hydrothermally altered pre- and synsedimentary granites in the hinterland (Robb et al., 1992). Although this hypothesis was doomed to failure for the same mass-balance problem mentioned above, it became untenable after it had been

recognized that the alteration was related to post-Witwatersrand basinal fluids (Klemd, 1999). This lack of suitable discrete deposits in the hinterland as likely sources of the Witwatersrand gold leaves leaching of background gold concentrations from a very large rock volume as the only alternative. Theoretically, such large-scale leaching could have been achieved by metamorphic fluids as suggested by Phillips and Powell (2011). This hypothesis, which would involve basin-wide fluid flow into anticlinal structures, shear zones and, locally, tension gaps, is, however, contradicted by the strong sedimentologic control on the gold distribution as outlined above. An alternative source was proposed by Large et al. (2013), who suggested reworking of gold-bearing pyrite from carbonaceous shales of the underlying West Rand Group. This hypothesis suffers from a conspicuous lack of highly metalliferous carbonaceous shales in the Witwatersrand Supergroup and a resulting mass-balance problem in spite of an elevated gold content in synsedimentary pyrite, constituting 6,100 metric tons (t) Au produced and a total endowment, including production, reserves, and resources (measured and indicated), of >9,375 t Au. The Abitibi belt records continuous mafic to felsic submarine volcanism and plutonism from ca. 2740 to 2660 Ma. A significant part of that gold is synvolcanic and/or synmagmatic and was formed during the volcanic construction of the belt between ca. 2740 and 2695 Ma. However, >60% of the gold is hosted in late, orogenic quartz-­carbonate veinstyle deposits that formed between ca. 2660 and 2640 ± 10 Ma, predominantly along the Larder Lake-Cadillac and Destor-Porcupine fault zones. This ore-forming period coincides with the D3 deformation, a broad northsouth main phase of regional shortening that followed a period of extension and associated crustal thinning, alkaline to subalkaline magmatism, and development of orogenic fluvial-alluvial sedimentary basins (ca. 9,375 t Au, as of December 31, 2018. This significant gold endowment resides in a number of key factors that cumulated into forming a diversity of styles of mineralization through time and space, more particularly in the southern part of the Abitibi greenstone belt. †Corresponding

author: e-mail, [email protected] *Natural Resources Canada contribution 20190283

Despite decades of exploration and research (Goldfarb et al., 2005; Robert et al., 2005; Dubé and Gosselin, 2007; Wyman et al., 2016), there is no consensus on the source of metals in greenstone-hosted gold deposits. Different models and ideas have been proposed to explain how and why such deposits form, and the diverse range of interpretations and hypotheses can in part be attributed to a poor preservation of primary features, which are commonly severely obscured by overprinting deformation and metamorphism. However, field observations and high-precision geochronology help constraining relative timing with regard to deformation events, an essential step in understanding how and where deposits are formed. This overview of gold deposits of the Abitibi greenstone belt is based on firsthand knowledge of most of the deposits discussed and on an exhaustive compilation of available

doi: 10.5382/SP.23.32; 40 p. Digital appendices are on the USB drive attached to the inside back cover and are also available online. 669

670

DUBÉ AND MERCIER-LANGEVIN

radiometric ages and whole-rock geochemical analyses. The emphasis is put on key field relationships gathered during the last 30 years. The main objective is to highlight and describe the parameters that have influenced the formation and distribution of gold deposits in space and time. Deposits are grouped by style rather than by genetic type or by specific deposit, which highlights variations in ore styles that result from the same processes but with contrasting characteristics, or those with similar characteristics but of different origins. Presenting such an overview in a concise contribution represents a challenge and consequently, numerous relevant references could not be cited in the main body of text and had to be moved to Appendices 1 and 2. Geology of the Abitibi Greenstone Belt The Abitibi greenstone belt covers an area ~515 km (eastwest) by 300 km (north-south) (Fig. 1; App. 1, Fig. A1A-B) and consists of Neoarchean supracrustal rocks and intervening mafic to felsic intrusive rocks. The supracrustal rocks were deposited over a period of ca. 125 m.y. from 2795 to 2670 Ma and the intrusions cover a slightly longer time span with the youngest batholiths dated at ca. 2640 Ma (Thurston et al., 2008; Monecke et al., 2017a, and references therein; Fig. 1). The belt is in tectonic contact to the north with the high-grade granites and gneisses of the Opatica subprovince. To the east, it is truncated by the Mesoproterozoic Grenville tectonic zone, and it is structurally bounded by the 0.5 mm