0361-0128/07/3709/1415-26 1415

IntroductionIN HIS REVIEW of iron oxide copper-gold (IOCG) deposits,Hitzman (2000) noted that a comprehensive genetic modelthat can help distinguish productive from barren or subeco-nomic systems is still lacking. Although IOCG deposits glob-ally share many features, such as abundant Fe oxides (mag-netite or hematite) and widespread sodic-calcic, potassic orhydrolytic alteration (Hitzman et al., 1992; Hitzman, 2000),the family of IOCG deposits is very diverse. Many IOCG dis-tricts also contain magnetite-apatite deposits of so-calledKiruna type, but no set of geologic features distinguishes

economic Cu-Au systems from these large accumulations ofFe oxides containing only anomalous amounts of these metals(Barton and Johnson, 2004). They speculated whether thiscontrast in ore metal abundance reflects fundamental differ-ences in the genesis of these hydrothermal systems or the lackof efficient traps or metal sources. The iron oxides may haveformed concurrently with Cu-Au mineralization (e.g., Hayneset al., 1995) or may have largely predated the introduction ofthe ore metals (e.g., Skirrow, 2000).

The eastern Gawler craton, South Australia (Fig. 1), is hostto the 3.81 billion metric ton (Gt) Olympic Dam Cu-U-Au-Ag-REE deposit (Reeve et al., 1990) and the recently discov-ered Prominent Hill Cu-Au deposit (Belperio, 2002). Nu-merous subeconomic Fe oxide-rich hydrothermal alteration

Fluid Evolution and Origins of Iron Oxide Cu-Au Prospects in the Olympic Dam District, Gawler Craton, South Australia

EVGENIY N. BASTRAKOV, ROGER G. SKIRROW, Geoscience Australia, GPO Box 378, Canberra, A.C.T., Australia 2601

AND GARRY J. DAVIDSONUniversity of Tasmania, Private Box 79, Hobart, Tasmania, Australia 7001

AbstractGeologic observations suggest two stages of hydrothermal activity at a number of presently subeconomic iron

oxide copper-gold systems in the Olympic Dam district, eastern Gawler craton. They contain high-, and mod-erate- to low-temperature Fe oxide-rich hydrothermal alteration. The mineral assemblages include magnetite-calc-silicate-alkali feldspar Fe-Cu sulfides and hematite-sericite-chlorite-carbonate Fe-Cu sulfides U,REE minerals. In all documented prospects, the minerals of the hematitic assemblages replace the minerals ofthe magnetite-rich assemblages.

The bulk of the subeconomic Cu-Au mineralization is associated with the hematitic alteration assemblages.Microanalysis by proton ion probe (PIXE) of hypersaline fluid inclusions in magnetite-rich assemblages, how-ever, demonstrates that significant amounts of copper (>500 ppm) were transported by the early-stage high-temperature (>400C) fluids responsible for the magnetite-rich alteration. These brine inclusions contain mul-tiple solid phases (liquid + vapor + multiple solids) including chalcopyrite in some cases. In comparison,inclusions of the hematitic stage are relatively simple liquid + vapor types, with homogenization temperaturesof 200 to 300C and containing 1 to 8 wt percent NaCl equiv. The Br/Cl ratios of the magnetite-forming flu-ids measured by PIXE lie beyond the range of typical magmatic and/or mantle values, allowing for the possi-bility that the fluids originated as brines from a sedimentary basin or the crystalline basement.

Sulfur isotope compositions of chalcopyrite and pyrite demonstrate that sulfur in both alteration assemblageswas derived either from cooling magmas and/or crystalline igneous rocks carried by relatively oxidized fluids(SO4

2 H2S, 34Ssulfides from 5 to +2) or from crustal sedimentary rocks (34Ssulfides from +5 to +10).Oxygen and hydrogen isotope compositions of waters calculated for minerals of the magnetite-rich assemblagehave 18O values of +7.7 to +12.8 per mil and D values of 15 to 21 per mil. The only available 18O andDfluid values for the hematitic assemblage are +4.7 and 9 per mil, respectively. The isotopic compositions ofboth fluids, coupled with the available literature data, can be explained in terms of fluid reequilibration withfelsic Gawler Range Volcanics or other felsic igneous rocks in the region and with metasedimentary rocks ofthe Wallaroo Group at low water-to-rock ratios prior to their arrival at the mineralization sites.

The lack of significant copper mineralization associated with magnetite-forming fluids that carried coppersuggests that there was no effective mechanism of saturation of copper minerals or the quantity of these fluidswas not sufficient to produce appreciable copper mineralization. Association of the copper-gold mineralizationwith the hematitic alteration in the subeconomic prospects can be explained by a two-stage model in whichpreexisting hydrothermal magnetite with minor associated copper-gold mineralization was flushed by late-stageoxidized brines that had extensively reacted with sedimentary or metamorphic rocks. The reduction of thesebrines, driven by conversion of magnetite to hematite, resulted in precipitation of copper and gold. The oxi-dized brines may have contributed additional copper and gold to the system in addition to upgrading preexist-ing subeconomic Cu-Au mineralization. When compared to published models for the Olympic Dam deposit,the new data for fluids in subeconomic Fe-oxide Cu-Au prospects of the Olympic Dam district indicate the di-versity of origins of iron oxide-copper-gold systems, even within the same geologic region.

Corresponding author: e-mail, evgeniy.bastrakov@ga.gov.au

2007 Society of Economic Geologists, Inc.Economic Geology, v. 102, pp. 14151440

systems are also known from the eastern Gawler craton,within the Olympic Cu-Au province (Skirrow et al., 2002).

For mineral explorers, an additional challenge is in dis-criminating between major mineralized IOCG systems andsubeconomic or minor systems concealed by deep cover,where minimal geologic and geochemical data are available.The lack of comprehensive information published on thenumerous IOCG alteration systems elsewhere in the OlympicCu-Au province adds to this challenge. The present contri-bution aims to provide a mineralogical, geochemical, andfluid evolution context for understanding the relationships

between subeconomic IOCG prospects and the major eco-nomic IOCG deposits in the Olympic Cu-Au province.

Previous Work and Scope of the Present StudyThe main focus of previously published descriptions and

research within the Olympic Cu-Au province has been theOlympic Dam deposit itself (e.g., Oreskes and Einaudi, 1990,1992; Reeve et al., 1990; Haynes et al., 1995; Johnson andCross, 1995; Johnson and McCulloch, 1995; Reynolds, 2000,2001) and Cu-Au vein deposits in the Moonta-Wallaroo dis-trict (Conor, 1995; Morales et al., 2002). Few studies have

1416 BASTRAKOV ET AL.

0361-0128/98/000/000-00 $6.00 1416

Coastline

MAR

GIN

OF

GAW

LER

CRAT

ON

6

Area of Figure2

2

3

Olympic Dam

Mt Woods Inlier

Moonta-Wallaroo district

Olympic Damdistrict

Polda Basin

FIG. 1. Location and simplified geology of the Gawler craton (Skirrow et al., 2007), showing the Olympic Cu-Au province(Skirrow et al., 2002) and major Cu-Au and Au deposits and prospects. The two boxes correspond to Figures 2 and 3. ThePolda basin (Neoproterozoic to Mesozoic) overlies the Archean-Mesozoic basement.

documented in detail the geology, hydrothermal alteration,and mineralization of the many IOCG prospects in theOlympic Dam district or in the Olympic Cu-Au province as awhole (Figs. 1, 2). Valuable information is available fromcompany reports (e.g., Paterson and Muir, 1986; Curtis et al.,1993), and summary descriptions of the Acropolis and Wirrda

Well prospects are given in Drexel et al. (1993). Gow et al.(1994) and Gow (1996) considered in detail the hydrothermalsystem at Emmie Bluff, including oxygen isotope fingerprint-ing of ore fluids. Knutson et al. (1992) studied alteration pat-terns of the Gawler Range Volcanics in the Mount Gunsonarea and reported results of petrological, fluid inclusion, and

FLUID EVOLUTION & ORIGINS OF IOCG PROSPECTS IN THE OLYMPIC DAM DISTRICT, GAWLER CRATON 1417

0361-0128/98/000/000-00 $6.00 1417

Weednanna Au

320

MenninnieDam Pb-Zn

Middl b k R F

Roopena

Magnetite - K-feldspar -calcsilicate

Hematite - sericite - chlorite -carbonate sulfides+

Magnetite-bearing alteration overprinted by hematitic alteration

IOCG alteration in drill holesHiltaba Suite granitoid & maficintrusions (1595-1575 Ma)

Mafic & anorthositic intrusions(~1760 Ma) Hiltaba Suite

Gawler Range Volcanics(mainly felsic,~1590 Ma)

Mafic Gawler Range VolcanicsMESO

PRO

TERO

ZOIC

Wallaroo Gp - metasedimentary& metavolcanic rocks; BIF? (~1740-1760 Ma)Donington Suite metagranitoids(~1850 Ma)

Undifferentiated Archean to Paleoproterozoic rocks

PALE

OPR

OTE

ROZO

IC

Fault

SAR9

Interpreted basement geology (pre-Pandurra Formation)

Drill hole number - sample location

Emmie Bluff IOCG prospect

FIG. 2. Interpreted pre-Neoproterozoic basement geology and IOCG hydrothermal alteration in the Olympic Dam dis-trict of the Olympic Cu-Au province. See Figure 1 for location.

300

310

WOOMERA

Wirrda WellAcropolis

DRD1

CSD1

TD2

SAE6

SAR9

OLYMPIC DAM

Emmie Bluff

Oak Dam

BLD1

NHD1

TitanBD1, TI2, TI6

PY3

HUD1

MRD1

SAE7

WRD1

CarrapateenaWOOMERA

township

30 00o

31 00o

136 30o

138 00o

50 km

stable isotope (S, C, and O) studies. Davidson and Paterson(1993) summarized a study of the Oak Dam deposit, reportedin detail in 2007 (Davidson et al., 2007). In the Mount Woodsinlier, Hampton (1997) described the Manxman and JoesDam IOCG prospects, whereas only summary informationhas been available on the Prominent Hill deposit (Belperio,2002). The Moonta-Wallaroo Cu-Au district has receivedmore attention (e.g., Conor, 1995; Morales et al., 2002). Basedon new geologic and petrological observations and on compi-lation of the available literature data, Skirrow et al. (2002)provided a summary of characteristics of hydrothermal IOCGalteration and mineralization in the Olympic Cu-Au province.

The current paper updates the descriptions presented bySkirrow et al. (2002) and focuses on the Olympic Dam dis-trict, one of the worlds most important IOCG districts. Wereport fluid inclusion and stable isotope data and consider theevolution and sources of fluids responsible for alteration andmineralization in a number of subeconomic IOCG systems inthe region and the implications of these results for models ofCu-Au mineralization.

Regional Geologic SettingThe geology of the Olympic Dam district (Fig. 2) is briefly

described below. For details, the reader is referred to Direenand Lyons (2007), Hand et al. (2007), and Skirrow et al. (2007).

Mesoproterozoic and older crystalline basement of theOlympic Dam district (Fig. 2) is covered by >200 m (com-monly >500 m) of Neoproterozoic, Cambrian, and youngersedimentary basin rocks, collectively known as the StuartShelf. The principal basement units are interpreted from po-tential field data and relatively sparse drilling (Direen andLyons, 2002, 2007). The oldest known lithologic units are vo-luminous, intensely deformed syntectonic granitoids of theDonington Suite, emplaced between 1850 3.5 and 1860 4Ma (Jagodzinski, 2005). Older Paleoproterozoic metasedi-mentary rocks and Archean crust have not been directly ob-served in the region but are inferred from geophysical mod-eling (Direen and Lyons, 2007). Large areas of basement areinferred to comprise the Wallaroo Group, a compositionallydiverse package of metasedimentary rocks deposited between~1740 and ~1760 Ma (Jagodzinski, 2005). This group com-prises thin-bedded feldspathic to biotite-rich metasiltstones,metacarbonate rocks, minor banded iron formation, and rarecarbonaceous metapelitic rocks. Underlying these rocks arered K-feldspar-bearing meta-arkosic rocks, with maximumdepositional ages of ~1850 to 1855 Ma (Jagodzinski, 2005).They may represent the lower part of the Wallaroo Group ora separate older unit. The only igneous rocks of possibly Wal-laroo Group age identified in the Olympic Dam district areleucogabbros in the Bills Lookout area (e.g., drill holesBLD1, BLD2, Fig. 2), which contain zircons with an age of1764 12 Ma (Johnson, 1993). The age of metamorphismand deformation of the Paleoproterozoic rocks is poorly con-strained in the Olympic Dam district but may correspond tothe Kimban orogeny, the effects of which were widespreadelsewhere in the eastern Gawler craton about ~1700 to 1730Ma (Daly et al., 1998).

The Paleoproterozoic basement in the Olympic Dam dis-trict was unconformably overlain by generally flat-lying andessentially unmetamorphosed units of the Gawler Range

Volcanics, a large-volume, felsic-dominated bimodal igneousprovince (Drexel et al., 1993). In the Olympic Dam districtthe only published U-Pb zircon determination for the GawlerRange Volcanics is 1591 10 Ma from the Wirrda Wellprospect (Fig. 2; Creaser and Cooper, 1993). The GawlerRange Volcanics was comagmatic with the Hiltaba Suite,which comprises granites, granodiorites, quartz monzonites,quartz syenite, and quartz monzodiorites (Drexel et al.,1993). The Hiltaba Suite is widespread throughout the cen-tral and eastern Gawler craton including the Olympic Damdistrict, where it hosts the Olympic Dam deposit. The mostprecise zircon U-Pb ages for Hiltaba Suite granitoids in thisdistrict cluster around 1588 to 1596 Ma (Mortimer et al.,1988; Creaser and Cooper, 1993; Johnson and Cross, 1995;Jagodzinski, 2005). Mafic intrusions of comparable age havebeen increasingly recognized in the Olympic Dam district(Johnson and Cross, 1995; Jagodzinski, 2005), as well as ineach of the other two main IOCG districts of the OlympicCu-Au province (Skirrow et al., 2007). The nature of defor-mation in the early Mesoproterozoic, during IOCG forma-tion, has been characterized by Direen and Lyons (2007), andHand et al. (2007). Deformation of the Hiltaba Suite granitesand Gawler Range Volcanics in the Olympic Dam district ap-pears to have been restricted to local brittle-style fracturing,brecciation, and faulting.

Alteration and MineralizationTaking into account the absence of outcrop in the studied

area, documentation and mapping of alteration in this studywere based on the examination of drill core (App. 1). The drillholes selected for detailed examination were initiallyscreened using the drill hole logs published by the Depart-ment of Primary Industries and Resources of South Australia(PIRSA; Curtis et al., 1993). Drill holes were examined in thePIRSA core library in Adelaide and sampled for petrographicwork and fluid inclusion analysis (Conventional micother-mometry and proton induced X-ray emission), and stable (S,O, H) and radiogenic (Sm-Nd, 39Ar-40Ar, Re-Os) isotope stud-ies (see Skirrow et al., 2007). This sample set was supple-mented with proprietary drill core from the Titan prospect(Fig. 2) provided by Tasman Resources.

The main challenge in understanding the spatial distribu-tion of the alteration types is correlating alteration assem-blages intersected in widely separated diamond drill holes dis-persed over a distance of 200 km (Fig. 2). Historically, most ofthe drilling targeted geophysical anomalies with coincidentmagnetic and gravity anomalies resulting from magnetite-richhydrothermal alteration. Recent discoveries of hematite-dom-inated IOCG systems characterized by gravity highs with littleor no magnetic response (e.g., Prominent Hill and Carrap-ateena) mean that many of the drill holes in the present studywere biased toward magnetite-bearing alteration.

Alteration types and paragenesis

The key hydrothermal alteration and ore mineral assem-blages recognized in the Olympic Dam district are magnetite-alkali feldspar-calc-silicate Fe-Cu sulfides and hematite-sericite-chlorite-carbonate Fe-Cu sulfides U and REEminerals (abbreviated to hematitic alteration). The mag-netite-biotite alteration assemblage observed in other parts of

1418 BASTRAKOV ET AL.

0361-0128/98/000/000-00 $6.00 1418

the Olympic Cu-Au province (Skirrow et al., 2002) has notbeen recognized in the Olympic Dam district. All three alter-ation assemblages occur as veins and replacements of hostrocks ranging from felsic to mafic in composition.

The magnetite-feldspar-calc-silicate assemblage is furthersubdivided into two types: magnetite-albite-calc-silicate (Fe-Na-Carich) alteration and a K-rich alteration in which K-feldspar replaced albite. This magnetite-K-feldspar-calc-sili-cate assemblage is by far the dominant magnetite-bearingassemblage in the IOCG systems studied. The calc-silicates

are generally actinolite or diopside; minerals present in bothmagnetite-bearing assemblages in minor to trace abundancesare quartz, pyrite, apatite, titanite, chalcopyrite, scapolite, andallanite.

The regional distribution and local temporal relationshipsbetween magnetite-bearing and hematitic alteration are sum-marized in Figures 2 and 3. In many examined drill holes (Figs.24, App. 1), both assemblages are present, with hematiticassemblages replacing preexisting magnetite-bearing assem-blages to different degrees. Nevertheless, it is possible to

FLUID EVOLUTION & ORIGINS OF IOCG PROSPECTS IN THE OLYMPIC DAM DISTRICT, GAWLER CRATON 1419

0361-0128/98/000/000-00 $6.00 1419

10 km

Hiltaba Suite (Balta Granite);mafic intrusions (~1590-1575 Ma)

Undifferentiated Paleoproterozoicmetasedimentary rocksUndifferentiated Archean toMesoproterozoic rocks

Fault - major

Engenina Adamellite (~1690 Ma)

Drill hole andsample location

PKN1

Fault - subsidiary

550000 E525000 E

6725

000

N67

5000

0N

Manxman Cu-Au prospect

Cairn Hill Cu-Au prospect

Magnetite - albite - calcsilicateand/or magnetite - biotite sulfides+

Hematite - sericite - chlorite -carbonate Cu-Fe sulfides+

Magnetite-bearing alterationoverprinted by hematitic alteration

Prominent HillCu-Au deposit

PKN1

DD86EN33

DD89EN61

Joes Dam Cu-Au prospect

DD85EN17

IOCG alteration in drill holes

FIG. 3. Interpreted pre-Neoproterozoic basement geology and IOCG hydrothermal alteration of part of the MountWoods inlier of the Olympic Cu-Au province. Modified and simplified from Betts et al. (2003). See Figure 1 for location.

sample drill holes and hydrothermal systems where the end-member alteration types are represented quite distinctlyforexample, at Murdie Murdie and Titan (magnetite-K-feldspar-calc-silicate alteration) and Emmie Bluff and Torrens(hematite-sericite-chlorite-carbonate alteration; Figs. 2, 4).Typical altered rocks, ore mineral assemblages, and some par-agenetic relationships are shown in Figure 5. For details ofthe key geologic and mineralogical features of the studiedprospects the reader is referred to table 1 of Skirrow et al.(2007).

Murdie Murdie

The hydrothermal system at Murdie Murdie is located ap-proximately 7 km east of the southern end of Andamooka Is-land in Lake Torrens (Fig. 2). Drill hole MRD1 targeted co-incident magnetic and gravity anomalies (Paterson and Muir,1986) resulting from the magnetite-bearing alteration (Fig.

2). This drill hole contains the least intense overprint by thehematitic assemblage and is used here to represent end-member magnetite-K-feldspar-calc-silicate alteration (Figs.4, 5A-F). The drill hole is characterized by abundant mag-netite, developed as replacement zones and veins within al-tered fine-grained quartzofeldspathic metasedimentary rock.Magnetite is accompanied by hematite-stained red feldspar(K-feldspar + albite), actinolite, apatite, titanite, pyrite, andtraces of chalcopyrite (Fig. 5A-E). The prominent feature ofthis drill hole is a zone of coarse-grained, subhedral mag-netite associated with pink centimeter-scale subhedral mi-crocline that replaced albite and coarse-grained pink apatite(Fig. 5A). Carbonate, quartz, pyrite, and chalcopyrite are in-terstitial to these minerals; some quartz and carbonate infillfractures in the coarse-grained minerals (Fig. 5C-E). Asnoted by Paterson and Muir (1986), the coarse-grained mag-netite-rich zone may represent either total replacement of

1420 BASTRAKOV ET AL.

0361-0128/98/000/000-00 $6.00 1420

A) MRD1 (Murdie-Murdie)Lithology Alteration Mineralization

Dep

th (m

) 800

810

820

830

840

850

860

870

880

890

900

910

K-Feldspar-Magnetite-Amphibole/Clinopyroxene-Apatite;Alteration distribution: network, pervasive

Magnetite-Amphibole/Clinopyroxene-K-Feldspar; Alteration distribution: network, pervasive

K-Feldspar-Amphibole/Diopside-Apatite

BXSL

STM

-pLx

(Py, Ccp)

B) BD1 (Titan)Lithology Alteration Mineralization

Dep

th (m

)

680690700

710720730

740750760

770780790

800810820

830840850

860870880

890900910

920930940

LxM

a/M

amLx

Lx

FPY

BXAR

KS

Hematite-Chlorite-Carbonate over Magnetite-Quartz-K-Feldspar-Amphibole-Apatite-Pyrite; Alterationdistribution: Network

Hematite-Chlorite-Carbonate over Magnetite-Quartz-K-Feldspar-Amphibole-Apatite-Pyrite; Alterationdistribution: Network, Pervasive

Py(-Ccp)

Py-Ccp

Py(-Ccp)

Ma/

Mam

ARKS

ARKS

ARKS

FPY

FIG. 4. Reference IOCG alteration types in the Olympic Dam district, based on the drill logs from the Murdie-Murdie(A), Titan (B), Emmie Bluff (C), and Torrens prospects (D). (A) and (B). Drill holes dominated by high-temperature mag-netite-K-feldspar-calc-silicate alteration. (C) and (D). Medium- to low-temperature hematite-sericite-chlorite-carbonate al-teration superimposed on magnetite-bearing alteration. Lithostratigraphic units: Lx = Wallaroo Group, M-p = Pandurra For-mation, Ma/Mam = interdigitated felsic and mafic Gawler Range Volcanics, Ppd = Donington Granitoid Suite. Lithologicunits: ARKS = arkose; BX = breccia; CNGL = conglomerate; FPY = felsic porphyry dike, GRT = granite, SLST = siltstone.Mineral abbreviations: Bn = bornite; Ccp = chalcopyrite; Cv = covellite; Py = pyrite. Brackets denote minerals in minorand/or trace abundances.

the host metasedimentary rock or a zone of strong veining.Dating of hydrothermal titanite in this alteration by theSHRIMP U-Pb method yielded an age of 1576 5 Ma, in-terpreted as the age of magnetite-bearing hydrothermal al-teration at Murdie Murdie (Skirrow et al., 2007).

Later alteration is represented by partial martitization ofmagnetite (Fig. 5E), chloritization of amphibole, and deposi-tion of calcite. The extent of this hematite-chlorite-carbonateassemblage within drill hole MRD1 is extremely limited.

There are no significant assays associated with the alter-ation within drill hole MRD1, but the magnetite-enrichedzone shows weak enrichment in copper (30120 ppm: Pater-son and Muir, 1986).

Titan

The Titan prospect is another example of an IOCG systemdominated by magnetite-K-feldspar-calc-silicate alteration,with minor hematitic alteration. This prospect is situated ~25km north-northeast of the Olympic Dam deposit (Fig. 2). Hy-drothermal alteration and sulfide mineralization are hostedmainly by meta-arkose. Alteration is also hosted by thin-bed-ded feldspathic to carbonaceous and biotite-rich metasilt-stone, similar to that in the upper Wallaroo Group farthersouth, and by unmetamorphosed felsic porphyry dikes similarto Gawler Range Volcanics. The metasiltstones are steeply

dipping with a strong shear foliation developed subparallel tocompositional layering. Fine-grained mafic dikes up to 30 mthick, but typically

or during magnetite-K-feldspar-calc-silicate alteration (Fig.5F), although the occurrence of rare pyrrhotite inclusions inpyrite points to at least locally reducing conditions during for-mation of magnetite-bearing assemblages.

The central alteration zone in drill hole TI2 is occupied bymultistage breccias, mafic dikes, and hematite-carbonate-chlorite alteration that cut and replace the magnetite-bearingassemblages. A narrower but similar zone is present in drill

1422 BASTRAKOV ET AL.

0361-0128/98/000/000-00 $6.00 1422

C

E

Mgt

Kfs

Host rock: Qtz+Bt+Fsp

Ap

Kfs

Amp,Chl+CalQtz

MagPy

Qtz Cal

Mag

Cal

Qtz

Ccp

Mag

Cal vein

Mar

Ccp

Mag

Hem

Qtz

Carb

F

D

BA

Hem

Carb

cm

cm

cm cm

FIG. 5. Common alteration assemblages associated with IOCG prospects of the Olympic Dam district, represented in drillcore samples and thin sections: magnetite-K-feldspar-calc-silicate assemblages (A-F), hematite-sericite-chlorite-carbonateassemblages (G-N). (A) and (B). Magnetite-K-feldspar-calc-silicate alteration, drill hole MRD1. (C) and (D). Magnetite-quartz-calcite-sulfide veins, drill hole MRD1. (E). Martite developing along margins and fractures of vein magnetite, drillhole MRD1. (F). Bladed magnetite developed after hematite, drill hole TI6. (G). Pyrite-chalcopyrite vein with specularhematite, drill hole TI6. (H). Vuggy hematitization zone, drill hole SAE6. (I). Early pyrite stringers cut by a quartz-chal-copyrite vein and hematite-chalcopyrite veinlets, drill hole SAE6. (J). Same sample as (I), showing detail of early fracturedpyrite with hematite-chalcopyrite fracture infill. (K) and (L). Pyrite-chalcopyrite veinlets cutting magnetite. (M). Quartz-hematite-chlorite-sulfidealtered rock, drill hole TD2. (N). Replacement of magnetite and pyrite by chalcopyrite-hematiteassemblage, drill hole TD2. Mineral abbreviations: Amp = amphibole; Ap = apatite; Bt = biotite; Cal = calcite; Carb = car-bonate; Ccp = chalcopyrite; Chl = chlorite; Hem = hematite; Kfs = K-feldspar; Mag = magnetite; Mar = martite; Py = pyrite;Qtz = quartz.

FLUID EVOLUTION & ORIGINS OF IOCG PROSPECTS IN THE OLYMPIC DAM DISTRICT, GAWLER CRATON 1423

0361-0128/98/000/000-00 $6.00 1423

Mag

Ccp

Py

Qtz

Hem

MagJ

Ccp

Py

HemMar

IHem-Ccp veinlet

Early Py

Qtz-Ccp vein

CcpHem

L

N

Hem-ChlPy-Ccp

cmM

cm

K

PyCcp

Ccp

Mag

Py-Ccp

cm

G H

Py

Hem/Mar

Hem

vugs

Qtz Chl-Qtzhost

vugs

Hem-Py-Ccp

Specular Hem vein

FIG. 5. (Cont.)

hole BD1. Angular clasts of previously magnetite-alteredmeta-arkose and largely unaltered mafic dike rock occur in amatrix of comminuted wall rock, mafic silicate and/or maficigneous rock, hematite-altered magnetite, carbonate, andchlorite. In contrast to the enclosing magnetite-K-feldspar-calc-silicatedominated alteration zones, which typically con-tain 700 to 900 ppm Cu and 1.4 to 1.7 wt percent sulfur, thiscentral zone of hematitic alteration contains only trace chal-copyrite and negligible pyrite (62 m at 200 ppm Cu, 700 ppmS: Tasman Resources, unpub. data). Microtextural evidenceindicates that pyrite and chalcopyrite were replaced byhematite and carbonate in this central alteration zone. Incomparison to other IOCG prospects in the Olympic Damdistrict, the low abundance of Cu associated with this centralhematitic alteration zone at Titan is unusual and suggests hy-drothermal leaching. The significance of this observation isconsidered below.

The intensity of martite replacement of magnetite de-creases away from the central hematitic alteration zone, al-though minor late-stage hematite, carbonate, muscovite, andchlorite (replacement of amphiboles) persist throughout themagnetite-rich zones. Minor chalcopyrite is commonly pre-sent in the hematitic alteration, suggesting a second stage ofcopper deposition during hematitic alteration.

Emmie Bluff

As at the Olympic Dam and Prominent Hill deposits, thehematite-sericite-chlorite-carbonate assemblage is also welldeveloped at the Emmie Bluff prospect (Gow et al., 1994).Two major zones of Fe oxide alteration about 3 km apart (Fig.6) are characterized by coincident positive gravity and mag-netic anomalies. The southern, discordant alteration zone con-tains veins and replacements within rhyodacite of the GawlerRange Volcanics; the veins contain skarnlike high-tempera-ture mineral assemblages with combinations of clinopyrox-ene, actinolite, magnetite (

quartz, pyrite, K-feldspar, and apatite. Although no calc-silicateminerals are preserved, clinopyroxene and garnet are foundas trapped phases within fluid inclusions in quartz. The rela-tionships suggest the former presence of magnetite-K-feldspar-calc-silicate alteration at Torrens. Hydrothermal carbonate islocally abundant as clasts in pseudobreccia with magnetite-rich matrix, suggesting that some carbonate may have formedin association with the magnetite-bearing alteration.

Hematitic alteration is well developed in several zones upto 30 m wide, overprinting the magnetite-bearing assem-blages and, in places, associated with brecciation. Magnetitein these zones is extensively replaced by martite and fine-grained specular hematite with chlorite and white mica. Late-stage disseminated carbonate is also associated with chloriteand martite and/or hematite. Whereas chalcopyrite is inti-mately intergrown with pyrite that formed part of the mag-netite-bearing assemblage, in detail some chalcopyrite ap-pears to have replaced pyrite (Fig. 5N). We suggest that thischalcopyrite most likely formed during the hematitic alter-ation stage.

Analytical MethodsThe nature of the fluids at several IOCG prospects in the

Olympic Dam district was investigated by petrographic, mi-crothermometric, and laser Raman microprobe methods atGeoscience Australia. Microthermometry was completed usinga Linkam MDS 600 stage (196 to +600C), and Raman mi-croanalysis was performed on a Dilor SuperLabram laserRaman microprobe.

A microanalytical study of selected fluid inclusions wascompleted at the CSIRO-GEMOC Nuclear Microprobe inSydney. The nondestructive method of proton induced X-rayemission (PIXE) analysis of fluid inclusions is described byRyan et al. (1993, 1995). Wherever possible, we selected shal-low (10 m in diameter, to opti-mize analytical precision and accuracy.

Sulfur isotope compositions of sulfide minerals were deter-mined at the University of Tasmania stable isotope laboratory.From the Olympic Dam district, 24 samples were analyzedusing a Conventional technique (Robinson and Kusabe, 1975)with an overall precision of 0.1 per mil. The samples forConventional analysis were prepared by drilling sulfide min-erals with a dental drill or by handpicking. To better under-stand the relationships between sulfur isotope compositionand paragenetic position of the sulfides, 20 additional mea-surements were determined in situ on polished thin sectionsby the Laser ablation method (Huston et al., 1995), with anoverall precision of 0.3 per mil. For comparative purposes,seven additional samples of pyrite, chalcopyrite, andpyrrhotite were analyzed (three by Conventional technique,four by Laser ablation) from the Mount Woods inlier IOCGprospects (Fig. 3). Isotopic compositions of samples are re-ported in Table 1 and shown in Figure 10 relative to the CDTstandard.

The oxygen and hydrogen isotope compositions of 20 sam-ples of silicate minerals were determined at the Institute ofGeological and Nuclear Sciences, New Zealand. Oxygen wasextracted from silicates for isotope analyses using a CO2 laserand BrF5, similar to the method described by Sharp (1990).Values are reported in the familiar 18O notation relative to

VSMOW (Table 2). Samples were normalized to the interna-tional quartz standard NBS-28, using a value of 9.6 per mil(). Values for four NBS-28 standards analyzed with thesamples had values that varied by less than 0.1 per mil(K. Faure, pers. commun., 2004).

Hydrogen isotope (D) values of the samples were deter-mined using the method of Vennemann and ONeil (1993).Samples were degassed at 200C overnight to remove ad-sorbed water. Water was extracted from the hydrosilicates byheating to temperatures of about 1,300C. All sample valuesare reported relative to VSMOW, and reproducibility of in-ternal standards was better than 3 per mil (Table 2). Samplevalues are normalized to the international biotite standard(NBS-30), assuming a value of 65 per mil.

Fluid Inclusions

Fluid inclusion types and microthermometry

The fluid inclusions from quartz and calcite in hydrother-mal alteration and mineralization can be divided into threemajor types, based on phase relationships at room tempera-ture and microthermometric behavior: (1) type A, vapor-rich,high-temperature inclusions; (2) type B, medium- to low-temperature, liquid-vapor inclusions; and (3) type C, high-temperature, halite-saturated hypersaline multiphase fluidinclusions (Fig. 7). In our study, type A and C inclusions weredocumented within magnetite-bearing assemblages at Titan(drill hole BD1), Torrens (drill hole TD2), and Emmie Bluff(drill hole SAE7). Some type A and type C inclusions survivein quartz grains in magnetite-bearing zones that were over-printed by hematitic assemblages (e.g., Torrens). Type B in-clusions are ubiquitous in association with hematitic assem-blages but also occur as trails of secondary inclusions inquartz of magnetite-bearing assemblages.

High-temperature brine inclusions have been previouslyreported only at Olympic Dam (Oreskes and Einaudi, 1992).Our work and that of Davidson et al. (2007) at Oak Damdemonstrates that the high-temperature hypersaline fluidswere widespread throughout the Olympic Dam IOCG dis-trict, in association with magnetite-K-feldspar-calc-silicate al-teration. An extensive dataset for the medium- to low-tem-perature type B liquid-vapor inclusions associated withhematitic alteration is available from the Mount Gunson area(Knutson et al., 1992), as well as for Olympic Dam (Oreskesand Einaudi, 1992) and Oak Dam (Davidson et al., 2007).

The gas phase in type A inclusions is dominated by watervapor with no detectable CO2 or CH4 (based on detectionlimits of about 0.1 mol % for CO2 and 0.03 mol % for CH4,using the laser Raman microprobe).

In addition to the ubiquitous halite crystals in type C hy-persaline multiphase fluid inclusions, a number of other salt,silicate, and opaque daughter phases are present. Some ofthese minerals were identified using the Raman microprobeand include ferropyrosmalite ((Fe,Mn)8Si6O15(OH,Cl)10),white mica (Fig. 7C, D), magnetite, hematite, and chalcopy-rite. In some cases, the multiphase brine inclusions containminerals captured accidentally during inclusion growth(e.g., drill hole TD2, Fig. 7G, H). This is deduced from therelative size of the captured minerals and from inconsistencyof their occurrence within the inclusion populations. Such

FLUID EVOLUTION & ORIGINS OF IOCG PROSPECTS IN THE OLYMPIC DAM DISTRICT, GAWLER CRATON 1425

0361-0128/98/000/000-00 $6.00 1425

minerals detected in quartz from the hematitic alteration as-semblage at Torrens included clinopyroxene and garnet (Figs.5M, 7G, H) and can be interpreted as relics of the early mag-netite-bearing assemblages that were overprinted by hematiticalteration. In microthermometric studies, complete homoge-nization was not attained in type C inclusions. They decrepi-tated at temperatures above ~400C, and nonsalt daughterminerals did not dissolve fully prior to decrepitation.

The results of microthermometric measurements for thetype B inclusions are shown in Figure 8. Our data indicate

salinities from 1 to 6 wt percent NaCl equiv and homoge-nization temperatures from 150 to 300C. Overall, includingthe available literature data (Knutson et al., 1992; Oreskesand Einaudi, 1992), there is a significant spread of the freez-ing point depression data for hematite-stage fluids, reflectingvariable fluid salinities (up to NaCl saturation). Most type Bfluid inclusions homogenized between 150 and 250C.

Hypersaline liquid-vapor-multiple solid (L-V-nS) fluid in-clusions are widely documented in IOCG systems of the Clon-curry district in association with regional albite-calc-silicate

1426 BASTRAKOV ET AL.

0361-0128/98/000/000-00 $6.00 1426

TABLE 1. Sulfur Isotope Values of Sulfides from the Olympic Dam District

34Ssulfide()

Prospect Drill hole Depth (m) Method Spot Py Ccp Po Sph

Mt Woods region

Manxman DD85 EN17 Laser ablation 6.5 6.3Manxman DD89 EN61 322.10 Conventional 2.0 2.3Manxman DD89 EN61 456.80 ConventionalManxman DD89 EN61 Laser ablation 1 3.6 5.4 6.4Manxman EN33 Laser ablation 1 0.4Peculiar Knob PK1 326.90 Conventional 4.3

Olympic Dam region

Titan BD1 685.50 Conventional 2.2 3.2BD1 685.50 Laser ablation 1 4.0 3.9BD1 799.40 Conventional 2.2 2.8BD1 799.40 Laser ablation 1 0.5 2.5BD1 801.80 Conventional 2.9BD1 868.00 Laser ablation 1 1.1 0.2BD1 868.00 Laser ablation 2 4.1BD1 897.40 Conventional 2.1BD1 917.00 Conventional 3.1 1.4

Bills Lookout BLD1 581.30 Conventional 2.8BLD1 730.70 Conventional 1.4CSD1 960.00 Laser ablation 1 0.5 0.1CSD1 960.00 Laser ablation 1 1.2CSD1 960.00 Laser ablation 2 1CSD1 960.00 Laser ablation 3 1.5DRD1 1,182.35 Conventional 2.8

Murdie Murdie MRD1 859.30 Conventional 1.2MRD1 Conventional 2.7

Mount Gunson area PY3 942.55 Conventional 2Emmie Bluff SAE6 959.70 Laser ablation 1 5.4 5.2

SAE6 959.70 Laser ablation 2 3.1SAE6 986.20 Laser ablation 1 10.3 12.1SAE6 986.20 Laser ablation 2 9.6 9.4SAE6 1,022.00 Laser ablation 1 6.3 8.3SAE6 1,022.00 Laser ablation 2 7.9SAE6 1,029.00 Conventional 7.6 8.5SAE6 1,041.00 Laser ablation 1 2.6 12.5SAR8 1,141.10 Conventional 3.1SAR8 1,149.40 Conventional 3.3SAR8 1,160.10 Conventional 2.8SAR8 1,160.10 Laser ablation 1 3.1 6.4SAR8 1,160.10 Laser ablation 2 6.4

Torrens Dam TD2 724.40 Conventional 0.3TD2 724.40 Laser ablation 1 0.1 0.5TD2 724.40 Laser ablation 2 3.4TD2 724.40 Laser ablation 3 3.8TD2 735.10 Conventional 0.1TD2 746.60 Conventional 0.6TD2 768.20 Conventional 0.4TD2 806.10 Conventional 0.2

Note: Mineral abbreviations: Ccp = chalcopyrite, Po = pyrrhotite, Py = pyrite, Sph = sphalerite

magnetite alteration (e.g., Williams et al., 2001). Although K-feldspar is the dominant alkali feldspar in magnetite-bearingalteration assemblages in the Olympic Dam district, ratherthan albite, there appear to be similarities in fluid character-istics in the two IOCG districts. Especially striking is the sim-ilarity in the set of daughter minerals that persistently occurthrough the fluid inclusion mineral assemblages: halite, cal-cite, ferropyrosmalite, and magnetite. Interestingly, the hy-persaline (L-V-nS) fluid inclusions in the Cloncurry districtare associated with high-density vapor-rich CO2 inclusions,whereas in our study vapor-rich type A inclusions lack CO2.

Microanalytical (PIXE) results

The main aim of the PIXE study was to directly measurethe copper contents of the fluids associated with IOCG sys-tems in the Olympic Cu-Au province. A second goal was tocharacterize fluids in terms of the Br-Cl systematics to tracepossible fluid sources. Selected results of PIXE analyses offluid inclusion compositions are shown in Table 3 and Figure9. Copper concentrations and Br/Cl ratios were successfullymeasured in the multiphase type C fluid inclusions associatedwith magnetite-bearing alteration (Table 3, Fig. 9A). Type Bfluid inclusions associated with hematitic alteration did notyield quantitative data for copper, although the hematite-stage fluids were found to contain very low concentrations ofcopper compared to the magnetite-stage fluids (i.e., belowthe detection limits for the particular inclusions measured).

The fact that some of the magnetite-related fluids were ex-tremely high in copper can be directly deduced from the ob-servation that many of the multiphase fluid inclusions containchalcopyrite crystals (confirmed by laser Raman microprobe).Thus, the extremely high measurements of the copper (24wt %) for some inclusions are not surprising. The absence of

chalcopyrite from other brine fluid inclusions from the sametrails suggests that the observed chalcopyrite may have beentrapped accidentally (cf. Fig. 7B). However, the measure-ments of copper concentrations of inclusions that lack visiblechalcopyrite are still high (300600 ppm copper) with poten-tially important implications for Cu-Au mineralization associ-ated with magnetite-bearing alteration. Only minor copperwas deposited in magnetite-rich assemblages, despite thebrines carrying more than 300 to 600 ppm Cu. This suggeststhat saturation of copper minerals did not occur (e.g., the flu-ids were too hot, or redox conditions, pH, or sulfur concen-trations were unfavorable), or the quantity of these fluids wasnot sufficient to produce appreciable copper mineralizationat the magnetite alteration stage.

The Br/Cl ratios for magnetite-related fluids in theOlympic Dam district (Fig. 9B) lie beyond the range of typi-cal magmatic and/or mantle values (Johnson et al., 2000;Kendrick et al., 2001), allowing for the possibility that the flu-ids originated as brines from a sedimentary basin or the crys-talline basement.

Stable Isotopes

Sulfur isotopes

Within the Olympic Dam district our new datatogetherwith those of Knutson et al. (1992) from the Mount Gunsonarea, Eldridge and Danti (1994) from Olympic Dam, andDavidson et al. (2007) from Oak Damindicate a signifi-cant range of sulfide 34S values from 14 to +12.5 per mil(Fig. 10). As a first approximation, these data imply twotypes of hydrothermal systems, the first characterized bynegative to slightly positive 34S values (14 to +4; GroupI, Fig. 10) and the second characterized by a component

FLUID EVOLUTION & ORIGINS OF IOCG PROSPECTS IN THE OLYMPIC DAM DISTRICT, GAWLER CRATON 1427

0361-0128/98/000/000-00 $6.00 1427

TABLE 2. Oxygen and Hydrogen Isotope Compositions of Alteration Minerals from the Olympic Dam District

Prospect Drill hole Depth (m) Rock Mineral 18O D T1 (C) T2 (C 18Owater Dwater

Emmie Bluff SAE7 1,123.9 Magnetite alteration zone Amp 7.7 42 500 9.5 21Emmie Bluff SAE7 1,168.0 Magnetite alteration zone Amp 6.5 40 500 8.3 19

Magnetite alteration zone Amp 9.5 500 11.3Magnetite alteration zone Quartz 10.0 500 7.7

Murdie Murdie MRD1 821.7 Magnetite alteration zone Amphibole 8.8 36 500 10.6 15834.3 Quartz-magnetite vein Quartz 12.4 617 500 10.1

Magnetite 4.4 500 12.2857.3 Quartz-magnetite vein Quartz 13.0 595 500 10.7

Magnetite 4.6 500 12.4

Titan BD1 722.7 Quartz-magnetite vein Quartz 10.9 773 500 8.6Magnetite 5.0 500 12.8

TI006 620.9 Brecciation zone with iron- Hematite 0.9 250 8.6oxide-pyrite-chalcopyrite matrix (hematitic alteration)

TI006 622.9 Quartz-magnetite vein Magnetite 4.7 500 12.5TI006 732.2 Quartz-hematite-sulfide vein Hematite 0.3 250 9.8

(hematitic alteration)

Torrens Dam TD2 751.0 Hematitic alteration zone Muscovite 7.7 44 250 4.7 9HUD1 481.2 Gawler Range Volcanics dacite Biotite 5.3 60 800 8.0 39

1 Calculated isotopic temperature based on the combination of 18O factors from Clayton and Kieffer (1991) for quartz and from Cole et al. (2004) formagnetite

2 Temperature assumed for calculation of isotopic composition of water (see text)

1428 BASTRAKOV ET AL.

0361-0128/98/000/000-00 $6.00 1428

A

C D

E F

Type A

Type C

Type B

G H

B

Fsp

CcpCal

NaCl

?Mica

?Mica

Fsp

Ccp

Ccp

Cal

Cal

CcpCpx

Udp

Grt

Udp

NaCl

10 m25 m

10 m

10 m

10 m25 m

10 m10 m

FIG. 7. Fluid inclusions of the IOCG prospects of the Olympic Dam district. (A). Three types of identified inclusions. (B)-(D). Type C inclusions (Titan prospect). (E). Type A inclusions filling a pseudosecondary fracture (Titan prospect). (F). TypeB inclusions from calcite cutting fractured magnetite (Murdie Murdie prospect). (G) and (H). Multiphase fluid inclusionsfrom Torrens prospect with relict minerals from magnetite-K-feldspar-calc-silicate alteration stage. Identified solid phases:Cal = calcite, Ccp = chalcopyrite, Cpx = clinopyroxene, Fsp = ferropyrosmalite, Grt = garnet, Udp = unidentified solidphases.

with positive isotopic values from about 6 to 13 per mil(Group II, Fig. 10).

The widest range of sulfide 34S values within a singleprospect was documented at Emmie Bluff (2.612.5; Fig.10). The lowest value (2.6) corresponds to a generation ofpyrite intergrown with magnetite and localized along compo-sitional layering within the metasedimentary host rocks. Thispyrite and magnetite predate the hematitic alteration assem-blage because they are cut by chalcopyrite-hematite veins,and the magnetite is martitized (Fig. 5I, J). They may haveformed as a part of the precursor magnetite-bearing assem-blage, although most chalcopyrite and possibly minor pyritegrew during later hematitic alteration. The chalcopyrite val-ues of 5.2 to 10.5 per mil are, overall, higher than the pyritevalues of 3.1 to 10.3 per mil. Thus, the chalcopyrite is not inisotopic equilibrium with pyrite (cf. Ohmoto, 1972). Takinginto account the textural evidence that the bulk of the chal-copyrite was deposited after pyrite, this indicates evolution ofthe bulk isotopic composition of sulfur in the fluids towardhigher values. The highest 34S value (12.5) was measuredin chalcopyrite from a specular hematite-chalcopyrite vein(Fig. 5J). This trend cannot be explained by simple redox re-actions, as one would expect higher 34S values associatedwith the most reduced assemblages (i.e., chalcopyrite associ-ated with magnetite) compared to the more oxidizedhematite-bearing assemblages. Rather, the increase in the34S values of chalcopyrite suggests that two differentprocesses and/or sulfur sources could have contributed to theobserved variation in the sulfur isotope composition atEmmie Bluff. These are considered further below.

Data for prospects in the Mount Woods inlier were ob-tained from Manxman and Cairn Hill prospects (Figs. 3, 10;drill holes DD89EN61, DD85EN17, and DD85EN33). Theobserved mineral assemblages are quite reduced, withpyrrhotite stable in drill holes DD89EN61 and DD89EN33,

and the analyzed minerals, including pyrrhotite, are charac-terized by negative 34S values (from 6.5 to 0.38). Asfractionation between aqueous H2S and pyrrhotite is insignif-icant, these values indicate that 34Sfluid values were negative.Such values cannot be attributed to changing redox condi-tions and point to negative isotopic values of the sulfursource. Magmatic pyrrhotite from the Peculiar Knob gabbro(drill hole PK1) has a positive 34S value of 4.3 per mil (Fig.10, Table 1), suggesting that some mafic magmas of HiltabaSuite age in the Mount Woods inlier (Jagodzinski, 2005)could have assimilated crustal sulfur. By comparison, pyrite inthe Bills Lookout gabbro in the Olympic Dam district hassomewhat lower 34S values of 1.3 and 2.8 per mil (Fig. 10,Table 1). In either case, sulfur sourced directly from suchmagmas or leached from the equivalent crystallized igneousrocks would also have a positive isotopic signature. If repre-sentative of the sulfur compositions in mafic magmas in thisdistrict, such sulfur would be an unlikely source of sulfur inthe Manxman IOCG prospect.

Oxygen and hydrogen isotopes

We have attempted to constrain the formation tempera-tures of the magnetite-bearing alteration using quartz-mag-netite geothermometry in samples from Murdie Murdie,where these minerals are intergrown (Table 2). The quartz-magnetite fractionation curve was based on the combinationof 18O factors from Clayton and Kieffer (1991) for quartzand from Cole et al. (2004) for magnetite, as recently recom-mended by Cole et al. (2004). The three analyzed mineralpairs gave temperature estimates between 595 and 773C. Inthe absence of the quantitative homogenization temperaturedata for fluid inclusions associated with magnetite formationin the studied prospects (see earlier), these temperatures can-not be independently verified. However, textural observationssuggest that quartz postdates magnetite; thus, it is likely that

FLUID EVOLUTION & ORIGINS OF IOCG PROSPECTS IN THE OLYMPIC DAM DISTRICT, GAWLER CRATON 1429

0361-0128/98/000/000-00 $6.00 1429

0

50

100

150

200

250

300

350

400

0 5 10 15 20 25

wt % NaCl equiv

Murdie-MurdieSAR8Mt Gunson area (Knutson et al., 1992)Emmie BluffOlympic Dam (Oreskes and Einaudi, 1992)

T h, C

FIG. 8. Temperatures of final homogenization of B-type fluid inclusions vs. apparent salinities expressed in wt percentNaCl equiv.

the calculated temperatures reflect isotopic disequilibriumand are unrealistically high. Postmagnetite deposition ofquartz must have been from a fluid with 18O values some-what lower than those of the fluid that precipitated mag-netite.

Oreskes and Einaudi (1992) previously reported isotopicequilibrium temperatures for quartz-magnetite from OlympicDam within the range ~420 to 540C. Thus, we used 500Cfor calculation of oxygen isotope composition of fluids inequilibrium with the analyzed magnetite-stage minerals(Table 2).

The limited data permit preliminary conclusions to bedrawn on the origins of the water component of the fluids re-lated to magnetite-bearing and hematitic alteration (Fig. 11).The composition of local magmatic water, based on a singleanalysis of fresh igneous biotite in the Gawler Range Vol-canics, fits well within the generic box of magmatic watercompositions associated with felsic magmas (Taylor and Shep-pard, 1986). In contrast, compositions of fluids associatedwith the magnetite-bearing assemblage from Murdie Murdieand Emmie Bluff are consistent with fluids of metamorphicorigin (Valley, 1986), which include fluids that have reacted

1430 BASTRAKOV ET AL.

0361-0128/98/000/000-00 $6.00 1430

TAB

LE

3. R

epre

sent

ativ

e R

esul

ts o

f PIX

E A

naly

ses

of E

lem

enta

l Abu

ndan

ces

for

Mul

tipha

se B

rine

Inc

lusi

ons

(Typ

e C

) fr

om Q

uart

z fr

om K

-fel

dspa

rC

alc-

Silic

ate-

Alk

ali F

elds

par-

Mag

netit

e (C

AM

)A

ltera

tion

Ass

embl

ages

1

Pros

pect

Dri

ll ho

leC

lK

Ca

TiM

nF

eN

iC

uZn

As

Br

Rb

SrB

aA

uPb

Tita

nB

D1

122,

429

69,2

4740

,015

1,50

06,

532

277,

179

with metamorphic rocks at elevated temperatures. The singledatum for the hematitic assemblage lies well outside the mag-matic water box (Fig. 11).

Discussion

Two-stage evolution of subeconomic IOCG systems

The observed paragenetic relationships in subeconomicIOCG systems in the Olympic Dam district indicate a re-markably consistent pattern of hydrothermal evolution, withearly magnetite-bearing alteration assemblages overprintedby hematitic assemblages. Our results support the two-stagemodel proposed by Gow (1996) for Emmie Bluff. In contrast,Haynes et al. (1995) argued that magnetite and hematite atOlympic Dam formed essentially coevally in different parts ofthe Olympic Dam Breccia Complex, during multiple over-printing hydrothermal cycles.

The general two-stage pattern of hematitic alteration over-printing magnetite-bearing assemblages is also observed inthe other parts of the Olympic Cu-Au province at Moonta-Wallaroo and in the Mount Woods inlier (Skirrow et al.,2002). However, in addition to the magnetite-alkali feldspar-calc-silicate and hematitic alteration, deposits and prospectsin these other two areas also involved magnetite-biotite alter-ation. The few cases where magnetite-biotite alteration hasbeen observed together with magnetite-alkali feldspar-calc-silicate and hematitic alteration assemblages suggest that it

FLUID EVOLUTION & ORIGINS OF IOCG PROSPECTS IN THE OLYMPIC DAM DISTRICT, GAWLER CRATON 1431

0361-0128/98/000/000-00 $6.00 1431

-14

-12

-10

-8

-6

-4

-2

0

2

4

6

8

10

12

14

Prospect or Drill-hole Number

34 S

sulf

Py Ccp

Po Bn

Cc Sph

Olympic Dam (Eldridge & Danti, 1994: mean values)

Group II

Group I

Oak Dam (Davidson et al., this volume)

Knutson et al (1992)

BLD1

Emmie Bluff

TitanMurdie-Murdie

PK1

CSD

PY3

ES21SAR

Manxman, Joes Dam,

and Cairn Hill

Torrens Dam

FIG. 10. Compilation of 34S data for sulfides from IOCG prospects of the Olympic Dam district. See Figures 1 to 3 forlocation of prospects. Data from Oak Dam (Davidson et al., 2007), regional alteration samples (Knutson et al., 1992, drillholes PY3 and EC21), and the median values for the Olympic Dam deposit (Eldridge and Danti, 1994) are included for com-parison. Note that the horizontal axis does not have geologic meaning, but the locations are arranged from north (left) tosouth (right).

-120

-100

-80

-60

-40

-20

0

20

-5 0 5 10 15 20

18Owater,

Dw

ate

r,

GRV BtAmph: Murdie-Murdie, Emmie BluffSer: Torrens Dam

SMOW

Meteoricwaters

Metamorphic H2O(300 - 600C)

H2O - felsic magmas

Cloncurry Cu-Au

FIG. 11. 18O vs. D systematics of IOCG-related fluids of the OlympicDam district. Reference fields for magmatic and metamorphic waters arefrom Taylor (1986) and Sheppard (1986), respectively. Isotopic compositionsof the fluids are calculated for the assumed temperatures of 800C (GRV Bt= magmatic biotite from the Gawler Range Volcanics), 500C (Amph = acti-nolite from magnetite-K-feldspar-calsilicate assemblages), and 250C (Ser =sericite from hematite-sericite-chlorite-carbonate assemblage). Range ofdata for the Cloncurry IOCG district of Queensland is based on Mark et al.(2004).

postdated the calc-silicatebearing alteration and predatedhematitic alteration (Conor, 1995; Hampton, 1997; Skirrow etal., 2002). The absence of magnetite-biotite alteration in theOlympic Dam district may reflect the relatively shallowcrustal levels exposed in comparison with deeper levels in theother two areas (Skirrow et al., 2002, 2007). Currently avail-able geochronology does not resolve the ages of magnetite-bearing and hematitic alteration in the Olympic Cu-Auprovince. The data are permissive of more than one cycle ofmagnetite to hematite alteration within the overall age brack-ets of 1595 to 1570 Ma (Skirrow et al., 2007).

Sources of fluids

The measured Br/Cl ratios of fluids related to magnetite-bearing alteration in the Olympic Dam district do not corre-spond to typical magmatic values, allowing for the possibilitythat the fluids originated as brines from a sedimentary basinor crystalline basement (Fig. 9A). The high salinities of themagnetite-related fluids could be explained either by theirderivation from metasedimentary rocks of the WallarooGroup, which contain scapolite or by liquid-gas phase separa-tion of an initially supercritical fluid of moderate salinity (cf.Heinrich et al., 2004). Indeed, coexistence of type A and C in-clusions in some samples (Fig. 7A) indicates that phase sepa-ration may have occurred.

Oxygen and hydrogen isotope compositions calculated forfluids related to magnetite-bearing and hematitic alterationare not consistent with a primary magmatic water source(Table 2, Fig. 11). Although the straightforward explanation isthat these fluids are of metamorphic origin, they may also beproducts of local 18O-enriched magmatic waters related to theGawler Range Volcanics and Hiltaba Suite granites, reequili-bration of primary magmatic waters with the host-rock se-quences (i.e., metasedimentary rocks in the case of mag-netite-related fluids, felsic igneous rocks in the case ofhematite-related fluids), or mixing of the primary magmaticwaters with fluids of nonmagmatic origin (i.e., metamorphicin the case of magnetite-related fluids, meteoric or seawaterin the case of hematite-related fluids). Oreskes and Einaudi(1992) reported high 18O values for quartz from two samplesof unaltered Hiltaba Suite granite from the vicinity of theOlympic Dam deposit (14.9 and 16) and suggested thatthese values characterize the granite, although whole-rock18O values are quite typical of granitic rocks worldwide (9and 9.1: Oreskes and Einaudi, 1992). It is not possible todeduce whether the high values are a result of subsolidusfluid-rock reequilibration. Taking into account that theHiltaba Suite granite in this region can be classified as A-type(Creaser, 1989), high 18O values are unlikely to be primary,as these are typical for S-type magmas (Taylor and Sheppard,1986).

Isotopic exchange with local felsic igneous rocks wouldhave resulted in a substantial decrease of 18O values and anincrease of D values of the fluids (i.e., a trend toward theTorrens point in Fig. 11). Reaction with metamorphic rocks inthe Olympic Dam district, which are likely to be enriched in18O, could account for the high 18O values calculated formagnetite-related fluids (Fig. 11).

Using the oxygen isotope composition of fresh igneous bi-otite from a Gawler Range Volcanics sample (5.3) and the

plagioclase-biotite fractionation factor at near-magmatic tem-peratures (e.g., ~800C), we can reconstruct the primarywhole-rock isotopic composition (~6.8) of a hypotheticalGawler Range Volcanics dacite. As another constraint, we canuse whole-rock values of ~9 per mil for the Hiltaba Suitegranite host at Olympic Dam, published by Oreskes and Ein-audi (1992). The compositional evolution of a fluid in equi-librium with each of these rocks at rock-dominated (R >> W,rock-buffered) conditions can be calculated simply as 18Ofluid= 18Orock rock-fluid(T) (Fig. 12; Campbell et al., 1984;Matveyeva et al., 1991).

The calculated fluid compositions based on literature val-ues for magnetite-bearing assemblages from Emmie Bluffand Acropolis, and magnetite at Olympic Dam, are similar(Oreskes and Einaudi, 1992; Gow et al., 1994; Gow, 1996).The 18O values of some of the magnetite-bearing assem-blages lie close to the values of the unbuffered pristine flu-ids of magmatic-like composition (Fig. 12). However, manycompositions are significantly heavier. This is consistent withequilibration of a fluid with a country rock that is isotopicallyheavier than a hypothetical Gawler Range Volcanics dacite(i.e., >7). In contrast, values of 18Ofluid calculated for thehematitic alteration stage show a considerable range. AtEmmie Bluff, Gow et al. (1994) and Gow (1996) used whole-rock and mineral isotope data to suggest a shift toward lowertemperatures and lower 18Ofluid values due to the influx ofevolved meteoric-hydrothermal waters. However, the data forminerals of the hematitic alteration stage (hematite andquartz) suggest isotopic temperatures that are too high(>390C: Gow, 1996) when compared to our fluid inclusiondata and temperatures of the hematite-related fluids else-where. We recalculated the mineral data of Gow (1996) usinga temperature of 250C (Fig. 12). As a result, the data splitinto two groups: high values associated with hematite and lowvalues associated with quartz. The heavy oxygen in fluids inequilibrium with hematite is very similar to that associatedwith magnetite-related alteration, both at Emmie Bluff andelsewhere in the Olympic Dam district. We believe that thiscan result from hematite inheriting magnetite 18O valuesfrom the magnetite stage.

The single data point for hematite from the Torrensprospect also may reflect reequilibration with local volcanicrocks. Although no Gawler Range Volcanics are knownfrom this area, Donington Suite granites are widespread.The relatively low 18Ofluid values are unlikely to have re-sulted from meteoric water influx, as the sulfur isotopecompositions of sulfides are inconsistent with this hypothe-sis (see below).

Oxygen isotope compositions of fluids at Olympic Damwere calculated by Oreskes and Einaudi (1992), using thetemperatures determined from magnetite-quartz andhematite-quartz pairs and assuming isotopic equilibrium be-tween iron oxide minerals and quartz. The lower end of theirisotopic temperature range (200360C) is consistent withtheir fluid inclusion data (mean homogenization temperature~200C, see Fig. 8), and they favored input of meteoric wa-ters to explain the relatively low 18Ofluid values associatedwith hematite. Figure 12 shows that the Olympic Dam dataare also consistent with a scenario of buffering of the fluids bythe Hiltaba Suite granite host rock.

1432 BASTRAKOV ET AL.

0361-0128/98/000/000-00 $6.00 1432

Morales et al. (2002) presented oxygen isotope analysesfrom quartz veins and chlorite from the Moonta Cu-Au de-posits in the southern Olympic Cu-Au province. The resultsare plotted in Figure 12 based on the mean fluid inclusion ho-mogenization temperatures reported by these authors. Thesedata are also consistent with the trend constructed for reequi-libration with the Gawler Range Volcanics and Hiltaba Suitegranites.

Sources of sulfur

The Olympic Dam district contains Fe-Cu sulfides associ-ated with magnetite and hematite with a wide range of 34Svalues, from 14 to +12.5 per mil (Fig. 10). The large nega-tive values in this range are associated with hematitic alter-ation in the Olympic Dam deposit (Eldridge and Danti, 1994)and in the Oak Dam prospect (Davidson et al., 2007). Most34S values from other IOCG systems in the Olympic Damdistrict, whether for sulfides associated with magnetite-bear-ing or hematitic assemblages, lie in the range of 4 to +4 permil. In the Mount Woods inlier, pyrrhotite, pyrite, and chal-copyrite in magnetite-bearing assemblages of the IOCGprospects have 34S values ranging from 6.5 to 0.4 per mil.Sulfur in this system was unlikely to have been sourced fromlocal mafic rocks (e.g., the Peculiar Knob gabbro, if itspyrrhotite value of +6.4 is representative). The availabledata suggest that a metasedimentary rock source of sulfur ismore likely. Except for the Titan prospect, IOCG alterationand mineralization assemblages in the Olympic Dam districtare more oxidized and lack pyrrhotite. Pyrite and chalcopyritethat formed as part of magnetite-bearing assemblages, for ex-ample at Titan, Murdie Murdie, and in drill holes CSD1 andDRD1, have 34S values of 4.1 to +1.5 per mil (Table 1).Given the likely formation temperatures of 400 to 500C and

moderately reduced conditions of magnetite-pyrite-chalcopy-rite deposition where we would expect relatively low [HSO4 +SO4

2]/ [H2S + HS] ratios, these sulfide 34S values are con-sistent with a 34Sfluid value of around 2 per mil (Fig. 13A).Although nonunique (see below), possible sources of this sul-fur include magmatic-hydrothermal fluids or sulfur leachedfrom igneous rocks (Ohmoto, 1986). Sulfides in magnetite-bearing assemblages at the Emmie Bluff prospect have quitedifferent sulfur isotope compositions. Paragenetically earlypyrite, some of which is associated with magnetite, has 34Svalues from 2.6 to as high as 10.3 per mil (Table 1). Chal-copyrite, which commonly replaced pyrite and formed pre-dominantly during the hematitic alteration and Cu-Au miner-alization event, has either similar or higher 34S values thanthe pyrite. A possible explanation of these characteristics ispresented in Figure 13B. Early pyrite associated with mag-netite may have formed from fluids similar to those associatedwith magnetite-bearing assemblages in other IOCGprospects (~2) but was modified during the hematitic al-teration event, when a different, isotopically heavy fluid wasinvolved. At the relatively oxidized conditions of hematitic al-teration and mineralization, the chalcopyrite 34S values of upto 12.5 per mil at Emmie Bluff are consistent with a 34Sfluidvalue ~20 per mil. This requires either an oxidized metasedi-mentary rock source or seawater or evaporitic water source ofsulfur.

Figure 13B shows that the 34S values of pyrite and chal-copyrite associated with magnetite could theoretically resultfrom deposition from a highly oxidized fluid with a 34SSvalue of around 20 per mil. However, we consider this veryunlikely because the magnetite-bearing assemblages consis-tently point to relatively reduced conditions. In contrast,Eldridge and Dante (1994) reported consistently negative

FLUID EVOLUTION & ORIGINS OF IOCG PROSPECTS IN THE OLYMPIC DAM DISTRICT, GAWLER CRATON 1433

0361-0128/98/000/000-00 $6.00 1433

-6

-4

-2

0

2

4

6

8

10

12

14

16

100 200 300 400 500 600 700 800 900

T(C)

18 O

wate

r,

GRV BtMgt (this study)Hem (Torrens Dam)Hem (Emmie Bluff, Gow, 1996)Mgt (Acropolis, Oreskes & Einaudi, 1992)Mgt & Hem (Olympic Dam, Oreskes & Einaudi, 1992)Moonta (Morales et al., 2002)

Re-equlibration trends

(R >> W)

Cooling of "pristine"magmatic water

(2)

(1)

(3)

Possible mixing

FIG. 12. Calculated 18O composition of the IOCG-related fluids as a function of temperature. The solid lines track hy-pothetical changes in isotopic composition of fluids in equilibrium with the Gawler Range Volcanics and Hiltaba Suite gran-ites at rock-buffered conditions. Numbers adjacent to the outlined fields identify possible scenarios controlling oxygen iso-tope composition of hydrothermal fluids. See text for discussion. Abbreviations: GRV Bt = magmatic biotite from the GawlerRange Volcanics, Mgt = magnetite from magnetite-K-feldspar-calc-silicate assemblages, Hem = hematite from hematite-sericite-chlorite-carbonate assemblages. Data for Acropolis and Olympic Dam (Oreskes and Einaudi, 1992) are for Fe oxideand silicate minerals of the magnetite and hematite alteration.

isotopic values for sulfides from Olympic Dam (mean valuesfrom 34Spyrite = 5.2 to 34Schalcocite = 10, standard de-viation of 2; Fig. 10) and a mean value for barite of 11 permil consistent with deposition from a fluid with a 34S valueof around 2 per mil, if fluid [HSO4 + SO4

2] contents approxi-mately equalled [H2S + HS] contents (Fig. 13C). This is rea-sonable for chalcopyrite-pyrite-hematite assemblages, partic-ularly where early magnetite was replaced by hematite sulfides(Oreskes and Einaudi, 1990; Reeve et al., 1990; Haynes et al.,1995). Alternatively, barite, chalcocite, and perhaps some bor-nite at Olympic Dam were precipitated from an SO4

2-domi-nant fluid with bulk 34S = 13 per mil (Fig. 13C).

Further evidence for a 34S-enriched sulfur source in theOlympic Dam district is provided by the data of Knutson et

al. (1992) for the Mount Gunson area. The studied drill holes,EC21 and PY3 (Fig. 2), intersect strongly altered GawlerRange Volcanics. Although most of the contained pyrite has34S values of 5.2 to +2.6 per mil, a few samples had dis-tinctly different 34S values from 7.1 to 11.7 per mil (Fig. 10).Knutson et al. (1992) interpreted these positive values interms of fluid interaction with an oxidized host rock and pre-cipitation of sulfate minerals that were later remobilized andlocally reduced. We believe that these data can be equallywell explained by a sedimentary rock surce of heavy sulfur.

In the Moonta-Wallaroo district, 34S data for sulfides inWallaroo Group metasedimentary and metavolcanic rocks (1.5to +4.6) and Cu-Au vein deposits (2.3 to +6.4) are simi-lar (Morales et al., 2002). If the values for metasedimentary

1434 BASTRAKOV ET AL.

0361-0128/98/000/000-00 $6.00 1434

400C

-30

-20

-10

0

10

20

30

40

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

X(SO42)tot

34 S

()

H2S

SO42

Isotopically closed system

H2Stot SO42tot

H2S in equilibrium with sulfides in magnetite-bearing alteration

34Stot = +2

A

250C

-30

-20

-10

0

10

20

30

40

50

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

X(SO42)tot

(a) 34Stot = +2

(b) 34Stot = +20 Sulfides

Sulfates

B

34 S

()

300C

-30

-20

-10

0

10

20

30

40

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

X(SO42)tot

34 S

()

(a) 34Stot = +2

(b) 34Stot = +13

H2S

SO42

SO42

H2S

Isotopically closed system

H2Stot SO42tot

C

FIG. 13. Relationship between the observed isotopic composition of sul-fur from minerals, sulfate to sulfide ratio in the fluid (SO4

2total/(SO4

2total +

H2Stotal), and an assumed isotopic composition of total sulfur in the fluids.The hatched ellipses represent 34SH2S values calculated from isotopic com-positions of pyrite, chalcopyrite, and bornite; the open ellipses represent sul-fate values based on the isotopic composition of barite. The observed valuesand sulfide-sulfate fractionation can be explained either from precipitation ofsulfides from a fluid with approximately equal quantities of sulfide and sul-fate species with 34Stotal = 2 or from a sulfate-dominated fluid with = 2or from a sulfate-dominated fluid with 34S = 13. (A). Conditions for mag-netite-feldspar-calc-silicate alteration. (B). Conditions for hematitic alter-ation and Cu-Au mineralization at Emmie Bluff. (C). Conditions pertinent tohematitic alteration and Cu-U-Au mineralization at Olympic Dam (based onsulfur isotope data of Eldridge and Danti, 1994).

rock-hosted sulfides are representative of the Wallaroo Groupelsewhere, it cannot account for the strongly positive 34S val-ues associated with hematitic alteration and Cu-Fe sulfidemineralization at Emmie Bluff, Mount Gunson area, orOlympic Dam. Other potential 34S-enriched sulfur sourcesare scapolite in the Wallaroo Group, Proterozoic seawater, orsulfate-rich evaporitic waters.

In summary, sulfur in fluids associated with magnetite-pyrite-chalcopyritebearing assemblages was sourced domi-nantly from magmas or leached from igneous rocks in theOlympic Dam district, whereas in the Mount Woods inliermetasedimentary rock sources (biogenic sulfur) were in-volved in some IOCG systems. Sulfur in hematitic alterationand mineralization assemblages had diverse origins and prob-ably included strongly 34S-enriched metasedimentary rocksources and/or the Proterozoic hydrosphere.

Thermodynamic modeling of Cu-Au transport and deposition

Several geologic and chemical models of ore formationhave been published for the Olympic Dam IOCG deposit(Oreskes and Einaudi, 1990, 1992; Reeve et al., 1990; Hayneset al., 1995), whereas models for subeconomic mineralizationin the Olympic Dam district have been presented only for theEmmie Bluff prospect (Gow et al., 1994; Gow, 1996). Thecommon theme of most models is the interplay between rel-atively reduced, magnetite-stable, high-temperature fluid andoxidized, hematite-stable, lower temperature fluid (Table 4).Key differences between models relate to the role of granitesversus other sources of metals and fluids, the timing of fluid-rock interaction, and which of the two fluids carried the oremetals. The viability of fluid mixing as an ore-forming processat Olympic Dam was demonstrated by Haynes et al. (1995).Although new thermodynamic data for Fe, Cu, and Au havebecome available since 1995, they do not dramatically changethe conclusions of that paper. A two-stage water-rock interac-tion hypothesis, advocated by Gow et al. (1994) and Gow(1996) for Emmie Bluff has not been modeled thermody-namically. Our geologic and mineralogical studies of othersubeconomic prospects in the district suggest that a two-stagemodel of water-rock interaction may be widely applicable tothe formation of IOCG systems. In the following, we usethermodynamic modeling to evaluate whether two-stage

water-rock interaction can account for the observed mineralassemblages and Cu-Au grades in subeconomic IOCG sys-tems of the Olympic Cu-Au province.

In this paper, we consider the interaction of a preexistingmagnetite-bearing assemblage, formed at high temperatureand from high-salinty fluids, with lower temperature and lesssaline oxidized fluids, using the HCh package for geochemi-cal modeling (Shvarov and Bastrakov, 1999; Shvarov et al.,1999). Thermodynamic properties of the model species weretaken from a Geoscience Australia version of the UNI-THERM database (Shvarov and Bastrakov, 1999; Shvarov etal., 1999) consistent with the web-enabled FreeGs database(Bastrakov et al., 2004). The modeling was completed at250C and saturation pressures of pure water. The salinity ofthe fluid prior to water-rock equilibration was set to 3 mNaCl; other fluid compositional parameters were constrainedby rock-buffered granite-rock equilibrium (at a water/rock wtratio of ~0.03) in a closed-box system open to oxygen (a slid-ing fO2 scale). We have considered a simplified chemical sys-tem of H-O-Cl-S-Na-K-Mg-Fe-Si-Al-Cu-Au, which would bea reasonable approximation for a felsic rock type, with thepreequilibration composition of granite specified as a mixtureof quartz, K-feldspar, albite, muscovite, phlogopite, annite,magnetite, and pyrite chalcopyrite. Starting compositionsfor the modeling are provided in Appendix 2.

Reduction of the oxidized fluid by interaction with mag-netite-bearing alteration: In this model (Table 4), a mag-netite-rich body acts as a chemical trap for copper and goldtransported in an oxidized, sulfate-dominated fluid. Precipi-tation of ore components results from the reduction of sul-fate due to magnetite oxidation (Fig. 14A). As a first approx-imation, the impact on copper solubility is modeled for asystem that is closed with respect to all components exceptfO2. Figure 14A shows the change in the ore assemblage andcopper solubility in response to the change of redox condi-tions. A dramatic drop in the solubility of copper (~2 ordersof magnitude) occurs at the chalcocite-bornite reactionboundary, followed by a minor drop toward the hematite-magnetite transition and then toward the pyrrhotite stabilityfield. Thus, simple reduction of an oxidized fluid similar tothat associated with hematitic alteration can be an effectivedriving force for precipitation of copper, if that fluid carriedcopper.

FLUID EVOLUTION & ORIGINS OF IOCG PROSPECTS IN THE OLYMPIC DAM DISTRICT, GAWLER CRATON 1435

0361-0128/98/000/000-00 $6.00 1435

TABLE 4. Possible Geochemical Models for IOCG Mineralization in the Olympic Dam District

Model group Mechanism Ore fluid Precipitating agent/substrate Reference

Fluid mixing Oxidized Cu-Au-Ubearing fluid Reduced fluid without Haynes et al. (1995)ore metals

Fluid mixing Reduced Cu-Aubearing fluid Oxidized U-bearing fluid Reeve et al. (1990)

Fluid-rock interaction Simple fluid reduction Oxidized Cu-Au-Ubearing fluid Reduced barren magnetite- Haynes et al. (1995)rich assemblage

Fluid-rock interaction Upgrading of preexisting Oxidized fluid without ore metals Reduced magnetite-rich This study; after Oreskes mineralization assemblage with subeconomic and Einaudi (1990,

Cu-Au mineralization 1992) and Gow (1996)

Note: Models of Haynes et al. (1995), Reeve et al. (1990), Oreskes and Einaudi (1990, 1992) were specifically developed for the Olympic Dam deposit,whereas that of Gow (1996) was for the Emmie Bluff prospect; these models do not necessarily apply to other IOCG mineralization in the district (see text)

Upgrading of preexisting Cu-Au mineralization: Dissolu-tion of preexisting Cu-Au mineralization and reprecipitationat higher grades (upgrading) also has been examined. The

oxidized H2O-NaCl fluid is chemically equilibrated with anoxidized Gawler Range Volcanics felsic rock lacking Cu andAu; the initial composition of the fluid was calculated as dis-cussed above, with the fO2 value set ~8 log units above thehematite-magnetite buffer (log fO2 = 27, in the chalcocitestability field, compared to 34.9 for the hematite-magnetitebuffer at 250C). This fluid, lacking Cu and Au, is isolatedfrom its source and progressively reacted with magnetiteusing the step-flow-through reactor numerical technique(Fig. 14B; cf. Heinrich et al., 1996). In this model, it is as-sumed that the magnetite-rich body acted as a source of Cuand Au (0.5 wt % Cu, 0.5 ppm Au), which were progressivelyupgraded by a dissolution-reprecipitation mechanism. Thismodel offers a plausible explanation for the apparent lack ofmeasurable amounts of copper in the fluids related tohematitic alteration examined in this study. The infiltration ofthe oxidized fluids into a magnetite domain oxidizes mag-netite and leaches ore minerals and sulfur; the fluid reductionalong the infiltration path results in the zoned reprecipitationof sulfides in the sequence: chalcocite, bornite + chalcopyrite,and chalcopyrite + pyrite (Fig. 14C). The model results ingrades of up to 3 wt percent Cu and 5 ppm Au, similar tothose observed in narrow intervals in the studied subeco-nomic IOCG systems. However, the total mass of ore metalsaccumulated in the model will be limited by the initialamount of copper and gold introduced during the early mag-netite-forming event.

The upgrading mechanism can account for the parageneticand mineralogical observations at most of the subeconomicIOCG systems in the Olympic Dam district, including Titan,Torrens, and Emmie Bluff. The simple reduction mechanismin which additional copper is introduced at the hematitic al-teration stage may have been important for the formation ofmajor IOCG deposits.

Summary and ConclusionsThis study of fluid evolution in subeconomic Fe oxide Cu-

Au systems of the Olympic Dam district indicates a diversityof possible origins for these deposits. The available geologicdata suggest two broad stages of hydrothermal activity. WeaklyCu-Au mineralized bodies of magnetite-bearing alterationwere present prior to formation of subeconomic as well assome major Cu-Au bodies in the Olympic Cu-Au province(Magnetite stage, Fig. 15). Our study has shown that the flu-ids associated with relatively reduced, high-temperature(~450500C) magnetite-bearing alteration were Cu-rich hy-persaline brines, yet they failed to produce extensive high-grade Cu mineralization. The Br/Cl ratios, and S, O, and Hisotope compositions of the magnetite-related fluids indicatemagmatic or leached igneous rock sources of sulfur but alsosuggest extensive exchange reactions with metasedimentaryrock sequences in the fluid source and/or pathway regions.

The new stable isotope, fluid inclusion, and geochemicaldata indicate that the hematite-forming fluids experiencedvarying amounts of chemical exchange with host rocks in theOlympic Dam district, resulting in diverse sulfur and oxygenisotope signatures of the oxidized fluids (Hematite stage, Fig.15). The oxidized fluids associated with Cu-U-Au mineraliza-tion and hematization at the Olympic Dam deposit appear tohave reequilibrated more extensively with igneous rocks (mafic

1436 BASTRAKOV ET AL.

0361-0128/98/000/000-00 $6.00 1436

HematiteMagnetite

Pyrite

Pyrrhotite

ChalcociteChalcopyrite Bornite

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

-34 -33 -32 -31 -30 -29 -28 -27 -26 -25 -24 -23 -22

log fO2

Opaq

ue m

iner

alog

y, m

ol%

1.E-05

1.E-04

1.E-03

1.E-02

1.E-01

1.E+00

Cu s

olub

ility

, mol

/kgCu,tot (aq)

Waves(Fluid

batches)Wave 1 1 12 i-1 i...

1 12 i-1 i... N ...

...

Wave N

Hematite Magnetite

PyriteChalcocite

Chalcopyrite

ChalcopyriteBornite

Au

Cu

80%

85%

90%

95%

100%

1 2 3 4 5 6 7 8 9 10 11 12 13 14Reactors (> direction of the fluid flow)

Opaq

ue m

iner

alog

y, m

ol%

0

1

2

3

Cu, %

; Au,

5 (x

10,00

0)

A

C

B

Magnetite-bearing

alteration

Hematiticalteration

Ti, n, Pi, n, fj ... k, i, n, [*] i, n

Steps (Reactors)

FIG. 14. Water-rock interaction models for Cu and Au precipitation inIOCG systems. (A). Copper precipitation by reduction of a copper-goldbearing oxidized (hematite-stable) fluid interacting with a magnetite-bearing assemblage. (B). Cartoon showing the calculation principle of pro-gressive fluid-rock interaction along a one-dimensional fluid-flow pathfollowing the step-flow-through reactor modeling technique. Each reactoris characterized by its own temperature (Ti), pressure (Pi), chemical poten-tials of perfectly mobile components (fi, if any), and bulk chemical composi-tion ([*]). Equilibrium compositions of consecutive reactors are calculatedthrough a number of passes (waves) of fluid batches through the system. (C).Copper and gold precipitation by progressive oxidation and upgrading of pre-existing copper-gold mineralization. The figure is a snapshot of the mineralzoning developing in response to the infiltration of the oxidized fluid. Notethe enrichment fronts of copper minerals and the predicted sequence of thecopper minerals. An intrinsic water-to-rock ratio is set to 10. The fluid infil-tration front is at reactor 9.

and felsic) than in the subeconomic IOCG systems, at temper-atures as low as ~150C (Oreskes and Einaudi, 1992; Hayneset al., 1995; Johnson and McCulloch, 1995). This comparisonsuggests that the process of relatively low temperature fluid-rock reaction was crucial for generation of Cu-U-Aurich oxi-dized ore fluids. Previous chemical modeling has shown thatreduction of Cu-Aurich oxidized fluids is a viable mechanism

of formation of major hematite-dominated IOCG deposits. Incontrast, our observations of the weakly mineralized IOCGsystems are consistent with a process of upgrading of preex-isting low-grade Cu-Au mineralization in magnetite-bearingalteration zones by Cu-Aupoor oxidized brines.

Sulfur isotope data for sulfides at Olympic Dam are consis-tent with late-stage influx of highly oxidized (surficial?) waters

FLUID EVOLUTION & ORIGINS OF IOCG PROSPECTS IN THE OLYMPIC DAM DISTRICT, GAWLER CRATON 1437

0361-0128/98/000/000-00 $6.00 1437

Murdie,Manxman

Magnetite stage Hematite stage

weak hematitic alteration

Titan

leaching/upgrading

upgradingTorrens

EmmieBluff

upgrading

OlympicDam (Hayneset al., 1995)

mixing? and/orboiling?

to ~ -6

34S S ~2

34S S ~2

34S S ~2

34S S ~2 34S S ~20

34S S ~2

CuAu

CuAu

CuAu

CuAu

CuAu

CuAu

CuAu

CuUAu

KEYMagnetite-bearingalterationMagnetite + weakhematitic alterationIntense hematiticalterationMetasedimentaryrocksMeta-igneousrocksHiltaba granite

Mafic dikes

Magnetite-formingfluidHematite-formingfluidExchangedmeteoric waters

18Ofluid 12-13

18Ofluid 11-12

18Ofluid 8-12

18Ofluid ~10

18Ofluid

carrying highly 34S-enriched sulfate, recorded in hypogenechalcocite mineralization. This is not observed in the subeco-nomic deposits. The major sources of copper and other met-als in the Olympic Dam deposit were at least partly externalto the host Roxby Downs granite and involved mantle-de-rived rocks or magmas (Haynes et al., 1995; Johnson and Mc-Culloch, 1995). In contrast, new data for subeconomicprospects in the Olympic Dam district presented in this studyand by Skirrow et al. (2007) indicate recycling of metals andsulfur from the local host rocks, with limited or no input frommantle-derived sources.

AcknowledgmentsThe authors wish to thank our colleagues in Geoscience

AustraliaPatrick Lyons, Nick Williams, Peter Milligan,Geoff Fraser, Peter Southgate, and Chris Pigramfor theirsupport and valuable discussions on many and varied aspectsof this work. The study was conducted under the auspices ofthe National Geoscience Agreement, as part of the GawlerProject in collaboration with the Department of Primary In-dustries and Resources South Australia. We thank MichaelSchwarz, Paul Heithersay, Sue Daly, and Brian Logan fortheir ongoing support in this collaboration. Andy Barnicoatand Jon Clarke reviewed an earlier version of the manuscriptand suggested significant improvements. Thorough reviewsby Geordie Mark and James Cleverley improved the manu-script, for which we are thankful. Editorial handling by GarryDavidson and Mark Hannington is gratefully acknowledged.We thank Kevin Faure for oxygen and hydrogen stable iso-tope analyses at the Institute of Geological and Nuclear Sci-ences, New Zealand. Geoscience Australia authors publishwith permission of the Chief Executive Officer.

REFERENCESBarton, M.D., and Johnson, D.A., 2004, Footprints of Fe-oxide(-Cu-Au) sys-

tems [ext. abs.]: Centre for Global Metallogeny, University of Western Aus-tralia Publication 33, p. 112-116.

Bastrakov, E., Shvarov, Y., Girvan, S., Cleverley, J., and Wyborn, L., 2004,FreeGs: web-enabled thermodynamic database for modeling of geochemi-cal processes [abs]: Australian Geological Convention, 17th, Geological So-ciety of Australia, Abstracts, no. 73, p. 52.

Belperio, A.P., 2002, The Prominent Hill copper-gold discovery: Explorationstrategies for world-class iron oxide Cu-Au deposits [abs]: Australian Geo-logical Convention, 16th, Geological Society of Australia, Abstracts, no.67,p. 257.

Campbell, A., Rye, D., and Peterson, U., 1984, A hydrogen and oxygen iso-tope study of the San Cristobal mine, Peru: Implications of the role ofwater to rock ratio for the genesis of wolframite deposits: ECONOMIC GE-OLOGY, v. 79, p. 18181832.

Clayton, R.N., and Kieffer, S.W., 1991, Oxygen isotopic thermometer cali-brations, in Taylor, H.P.J., ONeil, J.R., and Kaplan, I., eds., Stable isotopegeochemistry: A tribute to Samuel Epstein: University Park, PA, Geo-chemical Society, p. 310.

Cole, D.R., Horita, J., Polyakov, V.B., Valley, J.W., Spicuzza, M.J., and Coffey,D.W., 2004, An experimental and theoretical determination of oxygen iso-tope fractionation in the system magnetite-H2O from 300 to 800C:Geochimica et Cosmochimica Acta, v. 68, p. 35693585.

Conor, C.H.H., 1995, Moonta-Wallaroo region: An interpretation of the ge-ology of the Maitland and Wallaroo 1:100 000 sheet areas: Mines and En-ergy South Australia Open File Envelope 8886, DME 588/93.

Creaser, R.A., 1989, The geology and petrology of Middle Proterozoic felsicmagmatism of the Stuart Shelf, South Australia: Unpublished Ph.D. thesis,Melbourne, La Trobe University, 434 p.

Creaser, R.A., and Cooper, J.A., 1993, U-Pb geochronology of middle Pro-terozoic felsic magmatism surrounding the Olympic Dam Cu-U-Au-Ag and

Moonta Cu-Au-Ag deposits, South Australia: ECONOMIC GEOLOGY, v. 88, p.186197.

Curtis, J.L., Vanderstelt, B.J., and Parker, A.J., 1993, Pre-Adelaidean base-ment to the Stuart Shelf, South Australia: Drill hole database and prelimi-nary geological interpretation: Department of Mines and Energy, Geologi-cal Survey, South Australia, Report Book 93/95, p. 28.

Daly, S.J., Fanning, C.M., and Fairclough, M.C., 1998, Tectonic evolutionand exploration potential of the Gawler craton: AGSO Journal of AustralianGeology and Geophysics, v. 17, p. 145168.

Davidson, G.J., and Paterson, H.L., 1993, Oak Dam East: A prodigious, ura-nium-bearing, massive iron oxide body on the Stuart Shelf [abs.]: Geologi-cal Society of Australia Abstracts, v. 34, p. 1819.

Davidson, G.J., Paterson, H., Meffre, S., and Berry, R.F., 2007, Characteris-tics and origin of the Oak Dam East breccia-hosted, iron oxide-Cu-U-(Au)deposit: Olympic Dam region, Gawler craton, South Australia: ECONOMICGEOLOGY, v. 102, p. 14711498.

Direen, N.G., and Lyons, P., 2002, Geophysical interpretation of the centralOlympic Cu- Direen, N.G., 20