©2016 Society of Economic Geologists, Inc.Economic Geology, v. 111, pp. 1073–1098

Structural Constraints and Localization of Gold Mineralization in Leather Jacket Lodes, Ballarat, Victoria, Australia

Christopher J. L. Wilson,1,† Darren J. Osborne,2 Jamie A. Robinson,3 and J. Mcl. Miller4

1 School of Earth, Atmosphere and Environment, Monash University, Clayton, Victoria, Australia 38002 Evolution Mining, Pajingo Mine, PO Box 1435, Charters Towers, Queensland, Australia 4820

3 Geological Survey of New South Wales, PO Box 344, Hunter Region Mail Centre, New South Wales, Australia 23104 CSIRO Earth Science and Resource Engineering, Australian Resources Research Centre, Bentley, Western Australia, Australia 6102

AbstractGold mineralization at the Ballarat East deposit, central Victoria, Australia, is hosted in lodes that are histori-cally known as “leather jackets.” These are quartz-dominant vein arrays related to low-displacement, W-dip-ping faults (≤45°) that transect the core and/or eastern limb of tight, asymmetric N-S-trending anticlines. The leather jacket lodes typically have dip extents from 5 to 65 m, widths of ≤20 m, and strike lengths up to hundreds of meters, but their along-strike continuity is disrupted by oblique, low-displacement listric faults known as “cross courses.” The gold lodes are characterized by distinct phases of sulfide paragenesis with minor gold + arsenopyrite + pyrite defining the early sulfide stage. Late-stage coarse gold was precipitated with galena + sphalerite ± pyrrhotite ± chalcopyrite (late pyrite also occurs). The gold mineralization events are linked to low-strain mineralized fracture networks, which are closely related to the final deformation stages and the amplification of the major folds enclosing the lodes. This amplification produced domal fold culminations, with plunges ≤30°, and localized minor parasitic folds with shallower plunges (≤10°). A network of dilation sites, on the W-dipping faults, preferentially developed in the cores of anticlines, particularly in zones where there are changes in strike of bedding or fault bifurcation and refraction through contrasting sandstone and interbedded packages of sandstone and shale.

Numerical three-dimensional simulations were undertaken to test our geologic observations and replicate conditions controlling the emplacement of the leather jacket lodes. Two different scenarios were investigated: first, to determine how changes in the local stress field orientation influences dilation and fluid infiltration; secondly, to test variations in fault geometry during the last stages of deformation—that is, within the final 2% of shortening, when most of the mineralized sites were created. Results show that strain and fluid flow local-ized along refracted sections of faults and around changes in dip, specifically on the shallower dipping sections within subvertical sandstone units. This is consistent with the observation that high-grade gold-bearing quartz is associated with localized changes in fault dip in thicker sandstone and sandstone-shale packages. There was also a component of strike-slip motion and near-field NW-to-SE or N-to-S stress fields, which can be attributed to the development of a component of out-of-plane motion during the development of fold culminations. The preferred model for the distribution of the high-grade auriferous vein arrays defining the leather jacket lodes is one of fold amplification and extension parallel to the fold axes, which produced an increasing out-of-plane relaxation. The main fluid conduits responsible for the leather jacket style of mineralization involve infiltration along steep bedding-parallel faults and veins that link up with the arrays of low-displacement W-dipping faults.

IntroductionThe gold lodes at Ballarat East present particular challenges for both mining and exploration targets because they repre-sent a very discontinuous style of quartz vein development in comparison to many other orogenic gold deposits in central Victoria, such as those found in the Bendigo goldfield (Will-man, 2007; Wilson et al., 2013) or the Castlemaine goldfield (Cox et al., 1991; Willman, 2007). In common with the other gold lodes in central Victoria, the Ballarat East gold deposit is dominated by narrow quartz veins with disseminated gold particles (Dominy et al., 2008), and it shares similarities in terms of relative timing and being hosted by Ordovician tur-bidites within the Bendigo zone (Fig. 1). These host rocks are compressed by an initial deformation (D1) into upright, N-S-trending folds (Willman, 2007) in response to an E-W shortening during the Late Ordovician to Early Silurian

Benambran orogeny (455–440 Ma). In spite of these simi-larities, Ballarat poses particular problems because of inher-ent geometrical, geologic, and grade complexities along strike. At the Bendigo deposits the auriferous quartz reefs are closely associated with open, upright, chevron-shaped folds (wavelengths ~400 m) and steep-dipping (~70°) limb thrusts (Leader and Wilson, 2013). These reefs have along-strike extents of many kilometers and a regular repetition in the hinges of anticlines across the gold-producing district (Willman, 2007). In contrast, the lodes in the Ballarat East deposit are confined to a set of closely spaced (wavelengths ≤200 m) asymmetric, tight to isoclinal anticlines with over-turned eastern limbs. Within these folds there are quartz-rich en échelon vein array networks linked to vertically stacked W-dipping faults (≤45°), which have limited along-strike continuity (≤200 m). Furthermore, the final stages of mineralization at Ballarat are related to a N-S extensional event, which disrupts the areas of earlier mineralization. In contrast, the reefs at Bendigo are overprinted by an E-W

0361-0128/16/4411/1073-26 1073Submitted: July 6, 2015

Accepted: February 29, 2016

†Corresponding author: e-mail, Chris.Wilson@monash.edu

1074 WILSON ET AL.

crenulation cleavage related to N-S compressional event, which produces dilatant jogs and controls the areas of high-grade mineralization (Leader and Wilson, 2013).

The majority of the orogenic gold deposits in central Vic-toria are intimately associated with faults, which accompany the major folding events and provide fluid pathways that facilitate gold deposition (Willman et al., 2010; Wilson and Leader, 2014). Fluid infiltration and deposition of the coarse

gold observed at Ballarat East (Dominy, 2006) are intimately related to such faulting events. Faults in the Victorian crust occur in two distinct regimes: a deep system accompanied by slow convective flow in the basement (Willman et al., 2010), and a shallow fault system with faster convection and flow into an overlying turbidites, such as at Ballarat East (Wil-son and Leader, 2014). Although there are conflicting mod-els proposed to explain the origin of the gold (e.g., Bierlein

BW

BE

NG

A

C D FE G

STAWELL ZONE

37˚S

144˚E 145˚E 146˚E

Melbourne

25 km

Ballarat

BENDIGO ZONE

Stawell Castlemaine

Bendigo

N

Cambrian volcanicrocks

Turbidites

Early Devonian pluton

Late Devonian pluton

Fault

SS

GN MN

S0 n=713 S1 n=137 Joints n=432 Faults n=703 Quartz Veins n=372

Contours of poles to planes (per 1% area)

Fig. 1. Geologic map and representative structural data from the Ballarat East goldfield. (A) The Ballarat East (BE), Ballarat West (BW), and Nerrina (NG) deposits are located within domal culminations of the Ballarat anticlinoria, in the hanging wall of the Williamson Creek fault (modified after Taylor et al., 1996). The position of the Canadian and Eureka gullies has been suggested to correspond with the surface trace of the postulated Blue Whale fault, which is interpreted to be a shallow (≤40°) W-dipping fault zone. The Late Devonian Mount Egerton Granite intrudes the zone between the Blue Whale and William-son Creek faults. The Williamson Creek fault defines a boundary between the basal Lancefieldian and overlying Bendigonian section of the lower Ordovician Castlemaine Group. (B). Location of the Bendigo zone in Victoria. (C-G). Equal-area lower hemisphere projections of structural data shows that parasitic folds are doubly plunging asymmetric E-verging folds that are cut by generally W-dipping faults, with a range of fault, joint, and vein geometries discussed in text.

GOLD MINERALIZATION, BALLARAT, VICTORIA, AUSTRALIA 1075

and McKnight, 2005; Large et al., 2009; Phillips and Pow-ell, 2010), the timing of the gold mineralization in any of these models involves metamorphic devolatilization from deeper crustal sources, which occurs syn- to postpeak meta-morphism (Wilson et al., 2009). A recent model proposed for the formation of the Victorian deposits suggests that the gold is sourced from the surrounding turbidites (Large et al., 2009). This model requires orogenesis to redistribute gold and other metals from the local sedimentary rocks into quartz veins. Other workers have more recently argued that the gold is externally derived from deeply sourced aurifer-ous metamorphic fluids that have been focused along faults into a restricted vein network (Fairmaid et al., 2011; Wilson et al. 2013).

The Ballarat East deposit (56 t Au) is the largest of three deposits within the Ballarat region (Fig. 1), with extensive mining occurring during the later 19th and early 20th centu-ries, before closing down in 1918, with renewed gold pro-duction commencing in 2005 at the Woolshed Gully mine (Williams and Sykes, 2008), where 5.7 t Au averaging about 6.2 g/t Au has been mined between 2005 to 2015, and is still ongoing. Historically, mine exploration on the Ballarat East line was accomplished by tracking thin pyritic horizons, termed “indicators,” to the intersection with predominantly W-dipping (≤45°) quartz lodes known as leather jackets (Fig. 2A, B). Indicator beds were believed to indicate enrichment of gold at the intersection with quartz veins (Lidgey, 1894). A long-held belief, based on the proximity and coarseness of nuggetty gold, was that the indicator beds, which con-tain pyrite and carbonaceous material (Baragwanath, 1923), may be the source of gold (Gregory, 1907). Subsequently d’Avergne (1990) inferred that indicator beds represent bed-ding-parallel cataclasites, equating to slip surfaces or faults, on the limbs of the isoclinally folded sequence, which acted as a conduit for mineralization that hosts the quartz lodes. The historic subdivision of quartz lodes and veins in the Ballarat goldfield (e.g., Baragwanath, 1923) has been based on their orientation and relationship to bedding: (1) fissure lodes (bedded lodes); (2) verticals (breached hinges or axial planar faults); (3) leather jackets (Fig. 2; namely, bedding oblique, W-dipping reverse faults); (4) spurs or subhorizon-tal tension vein arrays, and (5) stockworks usually associated with thin bedded faults or shallow-dipping veins in the hinge of folds. However, many of these vein types are transitional or a combination of different stages of vein development (Fig. 3).

In this paper we provide a robust structural framework based on orientation and timing relationships between folds, faults, and the mineralized veins to understand the origin of the leather jacket style of mineralization. Using detailed underground mapping, drilling information, and surface out-crops we were able to develop a coherent chronology of the deformation, vein development, and mineralization events for the Ballarat East deposit (Table 1). Observations from the vein and fault movements are used to develop a model for gold-bearing vein formation with respect to the struc-tural evolution of the deposit. By using explicit modeling we demonstrate how our understanding of the local geometries of faults and stress field controls the localization of the gold mineralization.

Local Geology

Stratigraphic setting

The host rocks to the Ballarat goldfield belong to the basal Lancefieldian succession of the Early Ordovician, Castlemaine Group (VandenBerg et al., 2000), which consists of thin- to medium-bedded packages of sandstone, mudstone, and black shale. These beds contain lithofacies associations inferred to have been deposited as channelized deposits in a turbidite fan (Boucher et al., 2008). At a mine scale these units have histori-cally been named on the basis of their physical properties. For example, the three major sandstone units are defined as “Big” (≥10 m wide), “Little,” and “Amalgamated,” which are all important hosts of quartz veining and gold mineralization. Sim-ilarly, the shales, which have a prominent cleavage, are referred to by their prospective empirical widths, the “Four-Foot slate” and the “Twelve-Foot slate,” while those shales known to have the potential for gold mineralization are termed the “Mundic (pyritic) slate” and “Indicator slate” (Osborne, 2008). The old-est shale stratigraphic unit identified at Ballarat East is the Baroness shale (Fig. 4) while the Duchess shale unit is located higher up in the sequence; however, these units are difficult to correlate along strike and change dramatically at a mine scale with additional complexity in correlation introduced because of folding and later faulting. At the mine scale there is a change in lithology (Boucher et. al., 2008) from sand-dominated beds predominantly in the south to shale-dominated units in the north. Packages of interbedded shale and sandstone units occur in the central region of the goldfield (Fig. 4).

Structural and metamorphic setting

The Ballarat East goldfield is dominated by N-trending tight folds and cleavage development, and lies on the eastern limb of a regional anticlinorium known as the Ballarat anticline (Taylor et al., 1996). A striking feature of most folds is their strong E-verging asymmetry, noncylindrical nature and sub-vertical or overturned eastern limbs (Fig. 3A). Wavelength and amplitudes range between ~50 to 200 m with interlimb angles ranging from ~0° to 30°. Bedding (S0) and cleavage (S1) relationships suggest that fold axial surfaces are typically inclined to the west (Figs. 1C, D, 3A) and their intersection produces a prominent lineation (Fig. 5A), which enables fold plunges to be readily identified. The main deformation, pro-ducing the folds and steep W-dipping limb thrusts (>80°), is recognized as the D1 stage of deformation (Table 1). With further compression, there is a progressive transition into an episode of fold amplification and brittle fracturing (D2) accompanied by the evolution of discrete sets of mineralized veins (D3), whose timing relationship depends on the geom-etry of the earlier folds.

The major folds identified in the Woolshed Gully mine from W to E are known as the Norwegian, Sulieman, Scandi-navian, First Chance, Oregon, and Exchange anticlines (Fig. 3A). However, their persistence with depth is sometimes dif-ficult to ascertain because of the truncation by W-dipping faults. In the mine area there are major fold plunge reversals; in the southern area of the mine the folds are S plunging at ≤30°, with northerly plunges in the north (Fig. 4).

The folds are cut by linked faults that locally may be (1) bedding-parallel, (2) cleavage-parallel, or (3) multiple sets of

1076 WILSON ET AL.

Fig. 2. Interpreted cross sections of the main Ballarat East goldfield. (A). Historic section of mine workings that illustrates the W-dipping leather jacket lodes (from Brown and Hogan, 1932). (B). Distribution of the major W-dipping leather jacket faults with respect to the main sandstone horizons (modified after Boucher et al., 2008). (C). Schematic cross section, illustrating the different vein-forming environments and distribution of alteration and slip surfaces seen in units such as the Amalgam-ated sandstone.

Gammy Fault

Gammy yyyyyyyyyyyyyy Fa

B C

E

50 mA

W

C

GOLD MINERALIZATION, BALLARAT, VICTORIA, AUSTRALIA 1077

stacked W-dipping reverse faults (~45°, Fig. 2B, C), and the very occasional E-dipping fault, all of which have developed at different times and predominantly affect fold crests and upper limbs. The dominant W-dipping faults (Fig. 1F) are related to

the leather jacket lodes and tend to be best developed in the core and eastern limbs of the chevron folds (Fig. 2), but die out both downdip and along strike, and may be widely vari-able in their strike and dip. These include a subpopulation

West

Bask

ing fa

ultMak

o f

ault

faults

Mineralized quartz

CBA

Quartz veining inN-plunging anticline inAmalgamated sandstone

seric

itic sl

ate

reverse fault

N-plunginganticline

radialtensionfractures

Bedding/cleavage lineation

Faultrelatedslickenlines

20 m0

D E F G

fault with lineation

Fig. 3. Illustration of the varying structural styles of quartz veining (in red) encountered in the Ballarat East goldfield. (A). Cross section (36890 mN) in Sovereign compartment, showing prominent anticlines (A/C) and leather jacket faults that trun-cate straigraphic marker units. The uncolored units represent packets of interbedded shale and sandstone. (C). Level plan (99550 RL) of Sovereign 478 level, brown unit = Amalgamated sandstone and location of section A-A'. (D). Cross section A-A (37090 mN), illustrating the style of a leather jacket lode with E-dipping arrays of quartz veins containing arsenopyrite + gold. (E). Section 37075 mN, illustrating intense stockwork veining in hinge of folded Amalgamated sandstone. A bedding-parallel fault occurs on margin of overlying shale and a subhorizontal fault terminates at this boundary. (F). Section 37085 mN, showing subvertical and steep E-dipping V2-V3 veins related to the flat-lying W-dipping reverse fault (same fault as in D) breaching the limb of the same hinge. (G). Section 37120 mN of N-plunging fold in which the hinge zone is offset by a reverse fault that parallels the W-dipping bedding. (G). Schematic structural image of movement on bedding-parallel fault related to noncylindrical N-plunging First Chance Minor anticline observed in 478 level.

1078 WILSON ET AL.

of faults that dip at ~80° toward an azimuth of 220° (Fig. 1F). The folds are also offset by several major NE- to ENE-trending faults that have dextral/normal displacement of up to

several hundred meters (Fig. 4). These offsetting faults were historically known as “cross courses.”

The gold-bearing lodes at Ballarat East are located within and adjacent to the W-dipping faults (Fig.1F) as steep E-dip-ping quartz vein arrays (Figs. 1G, 2). Their along-strike con-tinuity is short (≤200 m) and confined to what are known as “compartments,” which are bounded by major ENE-striking cross-course faults (Fig. 4). These faults have an important control on the number of ounces/vertical meters mined as they offset and control the distribution of many of the min-eralized quartz veins. Cross-course faults range from small joint-like cracks with <10-cm vertical displacement to damage zones up to several meters in thickness with >100 m of lat-eral displacement (Table 2). The smaller faults generally have NE or NW strikes, with ankerite infilling, are cut by E-W-trending joints (Fig. 1E), and overprint an early set of dikes (Fig. 5D). In general, the NE-trending cross-course faults dip either NW or SE with a movement sense that is predomi-nantly right-lateral (dextral) with minor (<3 m) reverse dip-slip movement. The NW-trending cross-course faults dip SW and display left-lateral (sinistral) strike-slip displacement with minor dip-slip movement.

The amount of movement on several of the larger ENE-trending faults can be constrained by offset of the more con-tiguous fold hinges, stratigraphic units, and the W-dipping faults. Piercing point analysis (Ragan, 1973) of several faults (Table 2) shows that movement on these faults is dominated by dip- rather than strike-slip offset. In each case the north-ern side of the fault has gone down relative to the southern side so that progressively shallower parts of the stratigraphy are present at the same level toward the northern end of the

Table 1. Summary of the Deformation Stages and Veining Events Recorded in the Woolshed Gully Mine, Ballarat

D1 Chevron folding with S1 cleavage, which overprints an earlier locally developed S* bedding parallel fabric, possibly a result of syndepositional mass movement; thrust faults developed on limbs of N-S-trending folds during an E-W-directed compression

V1 Vein development and thin bedding- and cleavage-parallel (S0//S1) quartz veins during peak regional metamorphism (greenschist facies) and subhorizontal E-dipping veins; overprinted by later hydrothermal events

D2 Reactivation of D1 faults and accompanied by extension in S0 and S1 plane, with deposition of sulfides and minor gold in shale-rich horizons during a transition from E-W shortening into a W-to-E extension during an amplification of the anticlines; initiation of W-dipping faults related to leather jacket lodes

V2 Quartz veins associated with arsenopyrite and strong sericite and sulfide alteration proximal to earlier steep W-dipping reverse faults; veins may be oblique to bedding or occur as tension veins perpendicular to bedding

D3 Reactivation of steep-dipping faults during extreme tightening and amplification of chevron folds and propagation of ~45° W-dipping faults, with a component of out-of-plane N-S motion related to extensional strain and development of cross-course faults

V3 Extensional veins, rich in Au, between W-dipping faults and overprinting the V2 veins

V4 E-W-trending shallow-dipping (<45°) thin quartz-carbonate sphalerite-rich veins and steep-dipping transverse fractures with normal movement

Fig. 4. Geologic plan showing distribution of stratigraphic marker units and major structures. Sedimentary units, folding and faulting in plan view of the Ballarat East goldfield at the 10000RL level of the mine. Noncylindrical folding is indicated by the numerous plunge reversals within the N-S-trending sedimentary sequence. Compartments created by offsetting ENE-trending cross-course faults are named after the historic mines in the sector. A relatively large W-dipping fault marks the eastern side of the deposit in the Carlton, Golden Point, and Llanberris compartments. Based on historic data each compart-ment has different ounces/vertical meter (OVM) of contained gold.

GOLD MINERALIZATION, BALLARAT, VICTORIA, AUSTRALIA 1079

deposit (Fig. 4). Mineralized vein arrays at shallower levels toward the southern end of the deposit are located at greater depths toward the northern end of the deposit. These larger faults can comprise heavily fractured quartz vein material and country rock in a puggy clay matrix.

At the base of the current mine workings it has been postu-lated that there is a shallow-dipping (~40°) Blue Whale fault, developed within the hanging wall of the major crustal-scale Williamson Creek fault (Fig. 1A). However, an identification of the Blue Whale fault has not been conclusively established with current diamond drilling.

Metamorphic grade in the Castlemaine Group in the imme-diate vicinity of the underground workings is greenschist

facies and lies within the epizonal field (Wilson et al., 2009, whereas, geochemical studies suggest that mineralized fluids were derived from a deep-seated source generated by meta-morphic devolatilization (Fairmaid et al., 2011) and not in equilibrium with the host rocks. Previous studies at Ballarat have suggested that the gold mineralization events are broadly constrained between 460 and 400 Ma (Bierlein et al., 1999).

Intrusive events

At least three phases of dike emplacement are associated with areas of high-grade gold mineralization. Most dikes are recog-nized by their hydrothermal assemblages that include sericite, chlorite, kaolinite, and carbonate. The youngest generation

L1

Cross-course faultsDikeCross-course faultBedding So

3 m0BA

C D

i

Fig. 5. Photographs of selected folds, faults, and dike observed adjacent to the Woolshed Gully mine. (A). Bedding cleavage intersection plunging 20° → 355°; (WH189). (B). Parasitic folds west of Sulieman anticline in the Big sandstone unit adjacent to the run-of-mine (ROM) pad. (C). Cross-course faults in southern wall adjacent to tailings storage facility, box shows loca-tion of (D). (D). Early dike crosscutting bedding (see also enlargement) being offset by cross-course faults, showing apparent dextral movement.

1080 WILSON ET AL.

of dikes was derived from a mafic source (Brock, 2009) and postdates the main phase of mineralization and intrudes pre-existing structures. The earlier dike sets can be divided into an older set overprinting D1 quartz veins, and a younger unal-tered generation that may be related to the intrusion of the Mount Egerton Granite (Brock, 2009), which crops out ~5 km east of the Ballarat East goldfield (Fig. 1). A 368 ± 4 U/Pb zircon age from these latter dikes (Arne et al., 1998) places a youngest estimate on the age of mineralization, whereas the results of 40Ar/39Ar dating of micas associated with the miner-alization range between 460 and 400 Ma (Bierlein et al., 1999).

Lode Distribution in Relationship to Folds and Faults

Lode distribution related to strike of folds

The along-strike distribution of quartz-bearing lodes at Bal-larat East is highly variable in comparison to the uniform quartz-bearing reefs at Bendigo (Leader et al., 2013). The style of the leather jacket lodes at Ballarat East, named after the tassels that hang from the sleeve of a cowboy’s leather jacket, are very discontinuous along strike and are closely controlled by the geometry of folds and their interaction with faults (Fig. 3A-C). The First Chance and Suliemen N-plunging (≤30º) fold pairs can be traced throughout the northern two thirds of the deposit. Southward, from the Golden Point compartment (Fig. 4), there is an increase in the number and complexity of parasitic folds developed, which include the Scandinavian anticline and First Chance Minor folds (Figs. 3A, 5B). The parasitic folds have shallower plunges (≤10º N) and have lim-ited lateral extent (<300 m). Along the steep-dipping eastern limbs of each fold there are bedding-parallel reverse faults containing slickenside striations that pitch ~90°. Faults trend parallel to the N-S hinge line, and then progressively turn NNE (Fig. 3B) as they truncate the hinge zone. In hinge regions of major folds, where there is an initiation of parasitic folds, there are subtle strike changes in bedding of <10° and refraction of the bedding-parallel faults with the development of the leather jacket zones of mineralization (Fig. 2C). In plan view, the translation of the hinge is typically horizontal and right-handed, that is, a fold hinge is refracted to the right or in a dextral sense (Fig. 3B). These strike changes of the folds and faults coincide with increases in the width of the mineralized zones (Fig. 3C). Similar width increase of mineralized zones is also observed around fold hinges that refract to the left, or in a sinistral sense.

Leather jacket fault-vein networks (Fig. 3C), with an apparent thrust component of motion and offsets of up to ≤20 m, are generally located in a damage zone between multiple ~45° W-dipping reverse faults and en échelon vein arrays (Fig. 6). Fault separation is no more than 10 to 20 m with unclear cumulative offsets. Rather than narrow veins bounded by discrete faults, the leather jackets are dominated by steep E-dipping quartz veins, closely linked to subtle plunge reversals associated with the parasitic folds. An example of this is observed at the intersection of the Mako fault/Big sandstone (Fig. 6C) where a parasitic fold, known as the First Chance Minor anticline, plunges ~10° N, while the adjacent First Chance anticline plunges ~30° S. Here a mod-erately N-plunging lode horizon occurs in the Big sandstone and parallels the intersection between a fault splay and Mako fault, which has a 10- to 15-m offset and looses displacement along strike to the north.

The area containing the Mako leather jacket lode (Fig. 6A-E) is within an elliptical lens with up to four NW-dipping fault splays striking ~030°, and located east of the N-striking Mako fault (Fig. 6C). Within this there is a damage zone, ~10  m wide and 20 m in height, containing an array of subvertical or steep E-dipping quartz veins (5–20 cm wide) hosting coarse free gold. Vertical veins, which may be parallel to cleavage, are best developed where faults are proximal to the bedded fault with a horizontal separation of no more than 10 m (Fig. 6). In the marginal regions of subvertical vein development, veins are thinner, more laminated, have higher concentrations of arsenopyrite, and have W-dipping fault splays that preserve slickenside striations. In such areas the apparent displace-ment varies from left-lateral, through zero to right-lateral.

Lodes observed in cross section

In cross section, en échelon extension veins characterize the leather jacket lodes, and they are discordant to bedding and intimately related to the W-dipping faults (Fig. 6). Across fault planes, offsets of the extensional veins cannot be corre-lated with amount of fault displacement, as there is generally a major component of out-of-plane movement, as evidenced from the orientation of shallow-plunging slickenside linea-tions (Fig. 6A, B). Large irregular quartz masses or stockworks can be identified in the hanging wall of the faults (Fig. 6F), but along strike these grade into en échelon vein arrays (Fig. 6G). The majority of these lodes are preferentially located on the subvertical eastern limb of the folds, but depending on the intensity of the W-dipping faulting, bedding may locally become warped and display a variety of steep E or W dips (Fig. 6G).

Lodes that conform to a stockwork style of mineralization in cross section (Figs. 3D-F, 6F) are also intimately associated with parasitic folds. An example is the First Chance Minor anticline, which folds the Amalgamated sandstone, defining a tight N-plunging fold (10°–20°; Fig. 3D). Here the ore shoots and W-dipping vein arrays rapidly change style along strike, as follows:

1. Where a fault breaches the anticline (Fig. 3F).2. Within the hinge of the folded Amalgamated sandstone

(Fig. 3D). In the First Chance Minor anticline area there is an obvious lack of throughgoing fault, with veins forming

Table 2. Piercing Point Analysis (Ragan, 1973) of Movements on Major Cross-Course Faults Identified in the Woolshed Gully Mine

(location of faults is shown in Fig. 4)

Cross-coursefault Dip-slip Strike-slip Net slip

Seven Stars >74 m N side down >32 m dextral >81 m along (right-lateral) 65° → 100°

Charlie Napier 68-99 m N side down 11-16 m dextral 72–103 m along (right-lateral) 74° → 012°

Bull and Mouth 35 m N side down. 4 m sinistral 36 m along (left-lateral) 84° → 257°

GOLD MINERALIZATION, BALLARAT, VICTORIA, AUSTRALIA 1081

a distinctive radial pattern emanating from the overlying black slate horizon. This radial arrangement of veins was also related historically to signify gold-rich veins (Baragwa-nath, 1923), and we attribute these to tensile fractures.

3. As subvertical veins originating from W-dipping reverse faults, which may develop as a throughgoing structure within the Amalgamated sandstone (Fig. 3E).

4. Veining occurs in an overlying thinly bedded package (<2 m wide) of sericite-carbonated altered black slate and sand-stone, where V2 vein arrays (Table 1) are typically E dipping and extend from a western limb fault or anticlinal fold axis.

Bedding-cleavage intersection lineations (Fig 5A) indicate the First Chance Minor fold axis has a 10° to 20° northerly

GF

D1 fault leather jacket faults

spur vein

Mako fault

W

E

. ..

S0 V2-V3 veins

dip ofleatherjacketfault

EW

.

WH386

WH399

Mineralized quartz

Big sandstone

Altered sericitic unit(in 12 foot slate)

0 10 mWH386

5 m0

A

BC

D

E

faul

t s

play

Mak

o fa

ult

10

WH399

WH386

0 5m

30

MineN

Firs

t Cha

ce M

inor

A/C

Firs

t Cha

ce A

/C

Fig. 6. Vein geometries associated with leather jacket faults. The red dot on fault stereonet indicates orientation of slicken-lines, which are shallow plunging and indicate out-of-plane movement on what appear to be reverse faults. (A). Mako lodes on WH386 level adjacent to quartz splays (white color) from the Mako fault that truncates interbedded sandstones (dark gray color). An E-dipping spur vein array is also related to a minor E-dipping fault. (B). Geometry and average orientation of quartz veins on WH399 684 level, which are related to bedding and leather jacket fault. (C). Level plan showing location of Mako lodes in relationship to Big sandstone (yellow) and to folds in Little sandstone, truncated by Bignalls Union cross-course fault. (D-E). Interpreted cross section of Mako lodes within the Woah Hawp decline, WH399 and above WH386 level, where splays from the Mako fault offset the Big sandstone and adjacent 12-foot slate. The slate is locally replaced by sericite + carbonate alteration. (F). Stockwork vein grading into en échelon veins in hanging wall of Tiger fault that crosscuts overturned bedding on the First Chance anticline (width of photo 5 m). (G). Multiple en échelon vein arrays, which transi-tion into fissure lodes (bedded lodes), associated with minor W-dipping fault splays that warp subvertical bedding (LLBD638 south ore drive; width of photo 5 m).

1082 WILSON ET AL.

plunge (Fig. 3G). There is a difference in the rake of the fold plunge and fault movement where there is a change from breached versus nonbreached fold hinges. Stockwork vein development occurs in sites of fault refraction in the vicinity of the fold hinge (Fig. 3D, E). Zones of faulting are related to a greater areal extent of alteration, for example, alteration observed in the slate that overlies the Amalgamated sandstone on the western limb of the First Chance Minor anticline (Fig. 3F). This slate has a high concentration of chlorite, graphite, siderite veinlets and hosts a distinctive bedding-parallel lami-nated quartz veins with up to 30% arsenopyrite and minor gold (Table 3).

Distribution of lodes along ore drives

In longitudinal sections, subparallel to fold axes, the quartz-rich lodes have the appearance of flat or gentle-dipping quartz sheets (Fig. 7A). In the PR385 ore drive (Fig. 7B) the quartz lodes are dominated by flat-lying V2 veins with visible native gold overprinted by W-dipping faults and shallow dipping V3 and V4 veins (Table 1). The V2 veins are also offset by SW-dipping cross-course faults. The footwall of the listric-shaped cross-course faults host a proliferation of mineralized V3-V4 veins and the presence of brecciated carbonate veins. These fault-vein relationships indicate that there is a component of extension with normal faulting of the variably dipping V2 veins (Fig. 7B) to produce both dextral and sinistral separation in plan view.

Within the lodes, striations are found along surfaces of the W-dipping faults. These striations are characterized by car-bonate crystal fibers or grooves with thin coatings of graphitic

fault gouge. Within the majority of the steeper dipping faults there may be two lineation sets; the first is a steep-pitching lin-eation consistent with flexural slip during E-W compression. The second set of shallow plunging striations indicates local N-S or NW-SE shortening. The magnitude of displacements on late brittle faults, which includes normal faults, is usually small (≤5 m). The majority of these faults have dips that are identical to bedding or older veins (Fig. 8), and hanging-wall transport directions can be defined using a combination of offset bedding, extensional veins and fault structures (Miller and Wilson, 2004). Hanging-wall transport on these faults shows two dominant directions either top-to-the N or top-to-the-S and varies considerably as they are localized on preexit-ing structures within the variably plunging folds. This variable N-S sense of transport is particularly obvious on brittle faults related to V3/V4 sphalerite + gold-bearing quartz-carbonate veins (Fig. 8). Contemporaneous with N-S transport is the development of conjugate sets of listric-shaped cross-course faults (Fig. 7B). Movement indicators such as slickenlines on cross-course faults are scarce, but using offset veins dextral and sinistral movements can be recognized. The shortening direction inferred from taking the acute separation angle between the dextral and sinistral fault planes can be attrib-uted to an E-W bulk shortening (Fig. 8G).

Vein Chronology Related to Deformation HistoryTo establish a rigid vein chronology is difficult as there is repeated overprinting, a complex variation of geometries in both two and three dimensions, with deposition of succes-sive phases of quartz-carbonate and dike emplacement along

Table 3. Vein Characteristics of Ballarat East

Mineralogy Structural setting andVein Morphology distinctive features shortening direction Grade

V1 Width <5-cm cleavage-parallel Qtz ± Ank (sometimes Either— Nil or-perpendicular tension boudinaged) (1) Flexural slip on bedding veins (mode 1) strike N-S, accompanying D1 chevron folds subvertical or subhorizontal (2) Vertical extension on refracted E-dipping fractures in sandstone East to west shortening

V2 Width <5 to 50 cm (mode 2 Qtz ± Asp, Py, Ank, Chl; Fault and shear zones involving Nil unless overprinted by veins-normal offset), subvertical may show Chl being bedding-slip during flexural slip later vein event to moderately E-dipping replaced by Qtz West to east shortening

V3 Width <1-m tension veins Qtz ± Chl, Ank, Sid, Mus, Kao, Reactivates and extends margins of Average where they (mode 2 ~ variable normal and Sph, Gn, Cpy, Po, and Au; V2 veins especially where proximal to overprint V2 veins reverse offset), subhorizontal to “Bucky” quartz gradational strike-slip on early fault and/or High grade predominantly moderately N- or S-dipping; to breccia veins with typical footwall of cross-course faults; in V3 veins rollover tips and tension gashes fresh green coloration, possibly responsible for fissure Variable 5–200 g/t on V2, fissure lodes weathers to pale white clay lode formation at shallow levels East to west shortening and Northeast + southwest extension

V4 Width <5-cm tension veins Ank ± Qtz ± Sph, Gn and Au Reactivates margins of V2 veins High grade (modes 1 and 2), subhorizontal especially where proximal to Variable 20–200 g/t to moderately N fault-dipping; strike-slip on D1 fault and / or cut across V3 and V3 merge into footwall of cross-course fault internal stylolites of V4 veins Northwest to southeast + north-south shortening

Mineral abbreviations: Au = gold, Ank = ankerite, Asp = arsenopyrite, Chl = chlorite, Cpy = chalcopyrite, Gn = galena, Kao = kaolin, Mus = muscovite, Po = pyrrhotite, Qtz = quartz, Sid = siderite, Sph = sphalerite

GOLD MINERALIZATION, BALLARAT, VICTORIA, AUSTRALIA 1083

preexisting structures. It is possible to broadly subdivide the gold-related veins into two stages. The first stage was related to an early, weak gold mineralization event that formed V1/V2 quartz veins (Fig. 9A-C) containing gold and arsenopyrite (Table 3). A structurally later event, but related to the main gold-bearing set of quartz and carbonate V3-V4 veins, repre-sents the second stage. These V3-V4 veins host gold, sphal-erite-pyrite, minor galena, and preserve abundant stylolites. The important structural control on all vein development is their association with faults that propagate across fold hinges and are either associated with bedding-parallel veins that link on opposing limbs of a fold or are veins that link with cross-course faults (Fig. 7B).

V1 veins

These veins are the earliest identifiable vein set and are either bedding-, cleavage-parallel, or tensional veins with tension fractures infilling perpendicular to bedding (Fig. 3G). These are attributed to “mode 1” or tensile fractures (Atkinson, 1987), related to tight E-W folding of the strati-graphy. On fold limbs these may be laminated (Fig. 10A-C) and accompanied with steep (>80°) bedding-parallel faults, generally characterized by graphite-rich gouge in zones of <1 m wide, sometimes referred to as “indicator beds” that may also contain finely laminated and folded quartz-carbon-ate veins and areas of sericite-chlorite-pyrite-arsenopyrite

alteration. Veins of this generation may also occur as thin, vertical, cleavage-parallel and are unmineralized (Baragwa-nath, 1923), except where they are crosscut by later faults or vein sets. There are also arrays of subhorizontal or E-dipping V1 extensional quartz veins referred to as “spur veins” that are perpendicular to the margins of the steep W-dipping beds (>80°) (Fig. 2C). Offsets on these veins along known “indicator” units (e.g. Four-foot and Seven-foot shale) may be up to a meter or they may become folded (referred to as “rollovers”; Osborne, 2008) with local occurrences of visible gold.

The V1 veins are generally laminated and are found on fold limbs (Fig. 10A), and around fold hinges (Fig. 10B, C). The laminated character of these veins suggests that an anker-ite- and quartz-rich fluid was introduced initially along bed-ding and was episodic, comprising several quartz/carbonate generations formed by successive opening events (Fig. 10C). Arsenopyrite is the most abundant sulfide within V1 veins and adjoining wall rock, as euhedral and sometimes twinned grains. Arsenopyrite occurs within and on the margins of quartz veins as massive clusters and crude laminations and is also a prominent alteration product in wall rocks, particu-larly sandstones. Pyrite is nearly as abundant as arsenopyrite and has a similar distribution in and around quartz veins but is most prominent in altered shales, which are sometimes referred as indicator beds.

cross-course fault

leather jacket fault

leather jacketfault

cross-course faultwraps into S0

V3 breccia veins green color ± fine sphalerite

fracture/joint

V2 veins highly fractured

S0 S1

V2 V3V3cross-course fault V2

V4 carbonate veins V3

main V2 quartz veins joints

V3 splaysfrom V2 vein

thin V4 quartz veins

thin V3 quartzsplays fromV2 vein

thin V4 carbonate

veins

thin V3 quartz veins from east-dippingleather jacket fault and V2 vein

3631

0N

3632

0N

3633

0N

3634

0N

3635

0N

3636

0N

V2 vein Auin stylolites

A - FC478

S0S1 S1S0 S1S0 S0 S0S1

Bedding-cleavagelineationN

S0S1S0

Eastwall

Bedding-cleavagelineation

S0S1 dextral cross-course fault

sinistral cross-course faults

Bedding-cleavagelineation

N

B - PR385

0 1 2 5m

Fig. 7. Longitudinal sections and lower hemisphere equal-area spherical projections showing structural data. Quartz veins (red) and V3 breccias veins ± fine sphalerite (green). (A). Sketch of west wall of 478 level north ore drive 1. In this region F1 folds plunge N and V2-V3 veins (red color) dip S. Stereonets illustrate geometry of individual bedding (S0), cleavage (S1) inter-sections, and cross-course faults. (B). West wall of south ore drive (PR385) where there is a subhorizontal bedding-cleavage intersection (S1/ S0) (or fold axis) on an E-facing, W-dipping overturned east limb of the Scandinavian anticline.

1084 WILSON ET AL.

Normal faultEarlyextension veins

V1

V1

V1

V2V4

V4

V3

Fault and V4extensionveins

Fault andextension vein

Fault and V2 extension veins

Faults and V3-V4 extension veins associated with carbonate + sphalerite

Fault and V4carbonate+ sph veins

Cleavage

V4

N

Cross-course faults

BA

C

D

E F G

NW

SE

N S N S

NWSE

NW SE

NWSE

Sandstone Shale

Fig. 8. Extensional structures with corresponding sense of movement superimposed on preexisting V1 and V2 veins. Stereo-graphic nets are equal-area lower hemisphere projections and show the hanging-wall transport direction at pole to fault. Faint lines depict veins or cleavage, thicker lines are faults. (A-B). Late N-dipping carbonate-quartz extension veins (V4) associated with gold and sphalerite splaying from southerly dipping faults (thick lines) that parallel V1 veins (WH189, pen for scale). (C). SE-dipping fault overprinting V1 and V2 veins on WH218 level (width of rock bolt platens 20 cm). This is offset by a bedding-parallel normal fault. (D). Enlargement of area with V1 and V2 veins crosscut by V3 carbonate-rich veins that lie on the faulted margin of a V1 vein. (E). Margin of V1 vein reactivated by fault that produces dilation sites in which extensional V4 veins are deposited. (F). Massive V1 vein in sandstone offset by normal fault that juxtaposes a shale unit containing a laminated V2 vein (WH218 level). (G). Orientation of cross-course faults identified in PR385 and 405 levels with sense of movement determined from offset bedding and V1-V2 veins. The E-W-shortening direction is inferred from their conjugate relationship.

GOLD MINERALIZATION, BALLARAT, VICTORIA, AUSTRALIA 1085

V2 veins

These veins crosscut fold hinges and the S1 cleavage (Fig. 11A, B) and are generally quartz-filled mode 2 fractures (Atkinson, 1987) in which there is either minor (<1 m) normal or reverse displacements, and they carry little gold (Fig. 10D). The V2 veins offset V1 veins, and on overturned W-dipping bedding, dip at moderate and steep (45°–81°) angles and have a normal shear component with displacements of 5 to 15 cm (Fig. 8C). They are also spatially related to steep faults in altered shale (Fig. 8F) and generally refracted in sandstone units (Figs. 8C, 9C). V2 veins typically range from 5 to 50 cm in width and may persist along strike for up to 20 m as simple planar veins, but they are significantly modified by V3 and V4 vein development where they pass into cross-course fault zones. At Black Hill (Fig. 1), V2 veins coalesce into V3/V4 veins at a small gold-bearing cross-course fault (Fig. 10E). This cross-course fault at Black Hill splays from bedding with a curviplanar geometry to form a lateral ramp that truncates both bedding and quartz veins with ~10 cm of normal movement. Prominent bedding-cleavage intersection lineations at Black Hill indicate that the cross-course fault originates where the local folds plunge ~30° to the north (Fig. 10E).

V3 veins

Crosscutting the V2 veins are narrow, flat S- and E-dipping tension gashes, which are designated as V3 veins (Table 3). The V3-stage vein development appears to coincide with a major fault reorganization that was accompanied by the

development of extensional dilation sites (Fig. 11C), two sets of slickenside striations, and the onset of a reverse and left-lateral (sinistral) motion. Moderately (10°–30°) E dipping V3 veins and major stockwork vein networks in hinge regions of preexisting folds dominate these vein types. The dip of these veins and their termination on bedding contacts suggest they formed as extensional fractures. Many of these veins can be interpreted as small-scale faults, but they are not due to shear failure, rather the result from shear displacements during opening and closing of tensile fractures.

The quartz in the V3 veins is composed of blocky or euhedral crystals, and superimposed on this quartz are also mineralized stylolites, which provide evidence for later fluid activity. In the center of some of quartz-rich areas are euhedral crystals with native gold that indicate open-space growth in free fluid. The sulfide assemblage of galena and sphalerite, although occur-ring less frequently than arsenopyrite, is evidently coeval with most of the gold within the V3 veins. Rutile, distinguished by an acicular habit, is also associated with sphalerite + galena ± pyrite. A further sulfide association of pyrrhotite and chal-copyrite occurs as massive zoned growths in association with sphalerite and gold.

Gold grades in the leather jacket lodes vary greatly both across the width and along strike from low, ~20 to 30 g/t to high >50 g/t (Table 3). Assays on niche samples along rep-resentative examples of leather jacket lodes indicate that the original V2 veins contain up to 57 g/t of gold, whereas over-printed V2/V3 veins recorded assays up to 225 g/t, but also sev-eral results of <1 g/t (e.g., PR385 ore drive; Fig. 7B). In the

AA

B

D

F

F

CE

CB ED

VEINV2VEINV1 VEINSV4VEINSV3

Fig. 9. Vein morphologies and representative geometries. (A). Vein features truncated by W-dipping fault (F-F) in north wall of WH217 crosscut. Ellipses show location of images (B-E) (width of face 5 m). (B). Steep bedding- and cleavage-parallel V1 extension veins with orientation of S1 and steep slickenline lineation. (C). Steep E-dipping V2 veins with normal offset. (D). Moderate S-dipping quartz-carbonate V3 veins. (E). Shallow N-dipping, quartz-carbonate-rich V4 veins with stylolites. V3/V4 are associated with sphalerite, galena, and gold.

1086 WILSON ET AL.

B

VEINV2VEINV3 VEINV4 FAULTFAULTS0 S0

south

northLateral ramp

E

A B

CD

Fig. 10. Early vein development at Ballarat East. (A). Bedding-parallel quartz veins and initiation of en échelon vein arrays on small displacement W-dipping faults (LLBD596 NOD1 38125 mN; width of photo 4 m). (B). Drill core sample of folded early laminated quartz-ankerite vein, where S0 is at an angle to and deformed by S1. (C). Cross-polarized photomicrograph showing detail of folded veins in (B) composed of quartz and ankerite. (D). Flat-lying laminated V2 vein with arsenopyrite (WH237 level; scale on ruler is centimeters). (E). Geometry of veins and faults superimposed on N-plunging fold at Black Hill. Red box highlights region of lateral ramp, where cross-course fault splays from bedding.

GOLD MINERALIZATION, BALLARAT, VICTORIA, AUSTRALIA 1087

W

N SW E

WEE E

F

F

F

J

V

V

2

3

4

44

V

V

V4

2

A B

C D

FEFig. 11. Timing relationships of veins and faults in shale sequences. (A). Quartz veins (V2) transecting hinge of fold with verti-cal cleavage in slate unit (PR385 south ore drive). The margins of the quartz veins are overprinted by V4 carbonates, sulfides, and visible gold. Transecting the cleavage are low-angle W-dipping normal faults, between which are pull-aparts with siderite, gold, and sulfides infillings (scale on ruler is centimeters). (B). V2 quartz vein on limb of fold in a cleaved slate. The margins of the quartz vein are faulted and overprinted by later arsenopyrite, gold, and siderite. (C). Boudinaged sericitic slate with a concentration of quartz with late kaolin + chlorite (green) in the boudin neck and crosscut by V2-V3 veins. The V2 veins are rotated into slate with a normal sense of shear. Thin V4 veins with siderite infilling crosscut earlier vein sets (near 25 cm ruler). Within the sandstone is a W-dipping fault with apparent normal movement (F-F) that is refracted into and becomes parallel to altered slate. The fault has N-over-S sense of transport to produce an apparent normal offset. (D). East-dipping carbonate-rich V4 veins overprinting earlier V2 quartz, whose margin contains a minor cross-course fault with a sinistral movement com-ponent, and crosscut by late E-W-trending vertical (J) extensional joints (Sovereign 469 south ore drive). The overall shear sense in this area is south to north. (E). V4 stylolite seams containing gold and sphalerite overprinting earlier quartz-rich vein. (F). Slickenside striations associated with V4 veins (mesh spacing 10 cm).

1088 WILSON ET AL.

more mineralized regions, the bedding-fault intersection has a shallow plunge with bedding trending N-S. In contrast, in regions of the lode where gold grades are low, fault dips are more variable. Gold grades can also be high (>50 g/t) adja-cent to cross-course faults where there are late veins or in the vicinity of altered dikes (Fig. 12).

V4 veins

Veins and stylolites propagate at dip changes and in subhori-zontal sections of preexisting veins as thin (<5-cm) carbonate veinlets or thin quartz veins (Fig. 9E). These veins host gold and sphalerite and are defined as V4 veins (Fig. 11D). These veins generally dip 10° to 30°, with both N and S components of shear (Figs. 8A, B, 11D), and are most pronounced within shale units. Their distribution is sporadic and largely coinci-dent with the oblique fault splays and minor cross courses. The V4 veins on intersecting V2/V3 veins coalesce into crusti-form and stylolitic carbonate laminations or fractures that preserve slickenlines (Fig. 11E, F). In most instances V4 veins are defined by a distinctive mineral assemblage containing carbonate + pyrite + sphalerite + galena + coarse gold (Fig. 11E). The V3/V4 veins may also be associated with lenses of kaolin + chlorite, recognized by their distinct green color (Fig. 11C).

In some cases, V4 veins have developed as linked pull-apart structures between larger E-dipping V2 veins (Fig. 11A, B). There also appears to be a genetic relationship to the nor-mal movement on cross-course faults (Fig. 7), with brecciated

V4 zones distributed as a series of mineralized extensional fractures in the footwall to the fault and with the same dip direction. Many of these V4 veins also have abrupt strike changes and have dip reversals along strike, but all have dip-slip movement vectors. Superimposed on these veins there is also a later overprint associated with distinct extensional E-W-trending fractures or joints (Fig. 11D) that are barren of any mineralization.

Early dikes and relationship to vein networks

Within the Ballarat East deposit there are puggy seams (≤5 cm wide), sometimes considered to be fault gouge or misin-terpreted as indicator beds. These are early Mg- and Fe-rich dikes that parallel the S1 cleavage or exist in V1 to V3 quartz veins. They are generally altered to a mixture of kaolin, chlo-rite, ankerite, goethite, and minor biotite (Brock, 2009). They are generally boudinaged and overprinted by gold-bearing quartz veins (Fig. 12A, B) that predate the V4 vein networks. They are frequently parallel to flat-lying V3 veins or associated with steep-dipping laminated veins and are overprinted by a cleavage. Along their contacts there is an abundance of sul-fides that include arsenopyrite, pyrite, and minor amounts of galena and chalcopyrite and vermicular chlorite intergrowths with quartz and occasional gold grains (Fig. 12C, D).

Ore-stage assemblages

The uncertainties of previous genetic models proposed for the Ballarat gold deposits (Lidgey, 1894; Gregory, 1907;

ii

A B

DC

Au

Au

Fig. 12. Early dikes and relationship to mineralization. (A). Early dike in Prince ore drive 388 enclosed by V2 quartz vein. (B). Gold at margin between dike and quartz vein (PR388). (C-D). Photomicrographs (reflected light) of gold and arsenopy-rite (asp) within dike.

GOLD MINERALIZATION, BALLARAT, VICTORIA, AUSTRALIA 1089

Baragwanath, 1923) stem from a lack of understanding of the stratigraphy, structure, and timing of quartz veining. This study shows that the main mineralization overprints the D1 deformation features that include an initial episode of quartz veining (Table 1). Quartz appears to have been precipitated in V1 veins prior to most of the gold, sulfides, hydrous silicate, and carbonate minerals, which have been recognized in V2 to V3 veins, as grain boundaries between the earlier quartz, and other phases are not in equilibrium (Fig. 13).

The majority of the laminated quartz veins are also over-printed by later quartz veins, which may be accompanied by crustiform carbonate or sericitic alteration, sulfides, and gold

within stylolites (Fig. 13A-D). This quartz displays undulose extinction and serrated grain boundaries when associated with stylolites (Fig. 13D, E). Coarse gold (>300 µm) occurs as both native gold and as gold inclusions in fracture networks (Fig. 13F-I) within sulfides (Dominy et al., 2008), in V2 to V4 veins and overprints earlier laminated veins. Although each vein type has a dominant sulfide and alteration association (Table 3), more than one association may be present in any particular vein type. Gold generally overprints arsenopyrite and pyrite and may be in equilibrium with galena and sphal-erite, but is typically disseminated as coarse visible gold in quartz.

BA C

D E F

G H IAuAspAu

GalGal

Au

Asp

Au

AspAsp

Qtz

Cal

Qtz

Ser

200µm 200µm 200µm

200µm200µm200µm

5mm 10mm 5mm

S1

Fig. 13. Representative drill core (A-C), transmitted (D), and reflective light (E-I) images, showing assemblages and tex-tural features associated with gold-bearing quartz veins. (A). Laminated veins overprinted by later quartz veins and intense crustiform siderite and calcite alteration. (B). Ankerite and crustiform calcite confined to the margin of earlier quartz vein. (C). Laminated vein oblique to S1 cleavage overprinted by stylolites containing galena and sphalerite. (D). Cross-polarized photomicrograph showing subgrain development in vein quartz and hydrothermal sericite precipitated on grain boundar-ies. (E-F). Comparative cross-polarized and reflected light photomicrographs showing spatial association of quartz, calcite, arsenopyrite, and gold. (G). Reflected light image showing galena and gold in equilibrium, from a V3 vein. (H). Galena and gold within fractures and overprinting an earlier arsenopyrite grain. (I). Zoned arsenopyrite in quartz matrix overprinted by a stylolite that contains carbonate and gold. Abbreviations: Asp = arsenopyrite, Cal = calcite, Gal = galena, Qtz = quartz, Ser = sericite.

1090 WILSON ET AL.

Integrated Model for Au Mineralization at Ballarat East

A model that incorporates the structural and vein histories (Tables 1, 3) observed in ore drives can be intimately linked to flexural slip folding occurring during the deformation history (Fig.14). As the noncylindrical folds change profile (Fig. 14) they amplify and trigger the development of small-scale faults in the fold cores and within the overturned limbs. The folds inevitably become what are known as fault-cored folds (Shack-leton and Cooke, 2007). In a flexural slip model (Ramsay, 1974) fault development in this type of system is a function of acute fold tightening during E-W contraction. With ongoing contrac-tion, on either side of the pinned portions of the folds, there is a component of out-of-plane motion (Fig. 14B), where displace-ment vectors on the faults deviate from the E-W contraction (Shackelton and Cooke, 2007). This type of system can explain key components of the overall kinematic development in the

Ballarat East gold deposit, which we have deduced from the sense of slip associated with mineralized veins and on faults.

Our observations indicate that the W-dipping, subvertical eastern limbs of these folds (Fig. 14A) are regions where fluid pulses were able to permeate, as evidenced by an abundance of bedding-parallel laminated veins, sericitic- and sulfidic-bearing altered shales (termed historically indicator beds). The interaction of indicator-type slate units with mineralizing fluid facilitated much of the local gold deposition. The system absorbed the initial phase of deformation linked to folding via flexural slip (D1) with bedding-parallel faults and low-angle extensional V1 veins forming on the limbs of folds (Fig. 15, Table 1). Bedding-parallel movement horizons are almost entirely restricted to shale or siltstone lithologies, often at or near lithologic boundaries (Figs. 8, 10A).

The extreme tightening and amplification of the near-isocli-nal, reclined folds (D2-3) initiated failure in the hinge zones of

C

Flexural sliplineation

Fault relatedslickenside

Fault offset

Bedding/cleavagelineation

Slip directionPlunging fold axis

Flexural slipdirection

Pin

A

B

N

cross-coursefaults

late foldshape

Lateral fold propagation

V2 and V3 vein sets within dilation zones developed during fold amplification V4 fractureinitiation

E

early fold shape

late foldshape

Lateral fold propagation

ExtensionalfracturepatternV3/V4

D

Fluid flow path alongbedding-parallel faults

12211% 1% 1%1%1

2

Big sandstoneAmalgamated sandstonePotential fault location

Fig. 14. Schematic diagrams illustrating amplification of the tight, reclined folds at Ballarat East, and a kinematic model for the development of the V3/V4 vein sets and cross-course faults. (A). During an E-W shortening a localized N-to-S sense of shear occurs in the plunging sections of the fold culminations. Adjacent to the pinned regions movement is orthogonal to the fold axes and parallel to the applied shortening. The crosscutting fault jogs on fold limbs, comprise three separate domains that are developed progressively and have different senses of movement depending on fold plunge and are linked to a fluid-flow path along vertical faults. Fault-related slickensides with different movement senses are seen on crosscutting and bedding-parallel faults on either side of domal culmination. (B). A horizontal section through a domal culmination. Arrow arrays indicate relative motion and out-of-plane movement near the fold hinge, which would also be propagated through the folded sequence and into the W-dipping faults. (C). Possible form surfaces to the Big and Amalgamated sandstones and potential location of leather jacket faults prior to and during an increment of 2% shortening. This induces amplification of the beds and a component of reverse-slip on any fault. Although not sketched here for simplicity, the fault surface also deforms in an out-of-section direction that is coupled with the lateral fold propagation. (D). Section along fold axis showing the inflec-tion point in the profile where the curvature is zero and earlier and later fold profiles. The gray dashed arrows show potential displacement paths of the bed during lateral fold propagation (after Savage et al., 2010). Associated with the lateral fold propagation would be the development of an extensional fracture pattern that would initially develop V2 extensional veins. (E). With progressive fold amplification and propagation there would be the initiation of cross-course faults and V3/V4 veins superimposed on the V2 veins.

GOLD MINERALIZATION, BALLARAT, VICTORIA, AUSTRALIA 1091

the competent sandstone beds and sandstone-shale packages, resulting in the formation of W-dipping reverse faults (Figs. 14C, 15A), which are related to the leather jacket lodes. The faults that accompany the leather jacket lodes are generally initiated on subvertical fold limbs and ramp up to become crosscutting in the hinge zone or splay into a set of curved faults (Fig. 6), which terminate in a set of en échelon quartz veins (Figs. 3C, 6). Fault bifurcation is particularly prevalent near the hinge zone of parasitic folds, which coincides with the development of dilation jogs. The related dilational sites, filled by the mineralized veins, are created by space problems that arise in the hinge zone, where parasitic folds plunge at

a shallower angle or opposite directions to the geometry of larger folds. By this stage local strain distribution is no longer coaxial, with simple shear and cataclastic flow accompanied by strike-slip movements playing an important role on the fold limbs (Fig. 14C).

During the D2-3 deformation stages (Table 1), V1 veins are offset by step-like displacements parallel to bedding (Figs. 8C, 15B) with a movement sense consistent for flexural slip (Tanner, 1989). However, much of the strain was confined to the shale units (or indicator beds) with further extensional fractures accompanied by quartz stockworks (Figs. 3D, 6F). V2 veining is nucleated in the late stages of folding, when the

CA B

D E F

Indicatorbed

Indicatorbed

E-W shortening transitions into W-E extension (weak gold)

E-W shortening and NE-SW extension transitions into NW-SE to N-S hanging wall transport (high grade Au)

West East

Tension

Fault

S0

Legend

321443

G

Fig. 15. Interpretation of the sequence of stages related to gold and vein development and examples of paleostress determi-nations. (A). Thick sandstone unit (yellow) and indicator bed (blue) on the overturned eastern fold limb, with V1 vein devel-opment as either flat or steep bedding-parallel veins initiated during the E-W shortening. (B). V2 veins related to extension during the initiation of E-directed thrusting. The arrows indicate relative motion during vein formation. (C). Weakly mineral-ized V3 veins are emplaced in the waning stages of an extension event. (D-E). Main episode of V3 vein development related to the formation of leather jacket lodes. The veins are mainly S-dipping extension or composite veins with components of extension + shear. (F). V4 veins related to preferential reactivation, during a low-strain event involving N-S movement compo-nents, of preexisting faults, including cross-course faults, and earlier veins. (G). Legend and paleostress determinations using slickenline striations on fault planes (blue) associated with leather jacket lodes. Source of data: (1) WH386, (2) SH489, (3) WH218, (4) PR385, where WH = Woah Hawp and PR = Prince declines (depth from surface in meters is shown after prefix).

1092 WILSON ET AL.

operation of flexural slip on bedding gave rise to area changes and extensional strain in the limbs. The competence con-trast between multilayers of shale and sandstone was high, as indicated by the presence of faults and veins. In summary, continued flexural slip is interpreted to have resulted in the formation of further bedding-parallel V2 veins and steep dip-ping V2 veins oblique to bedding with an en échelon pattern related to flexural slip veins (Fig. 15B).

The field data also indicate that there was then a phase of extensional movement on the W-dipping faults with the for-mation of V3 veins and cross-course faults (Fig. 14 D, E), which have geometries that cannot be linked to a purely two-dimensional flexural slip interpretation (Fig. 15C). Earlier veins are overprinted by later low-angle V3 and V4 veins that reflect a switch back to both reverse and extensional move-ments on the W-dipping faults (Fig. 15D-F). The onset of this later stage of oblique reverse motion on W-dipping faults, with a component of predominantly sinistral transpression, is identified from the NW-SE and N-S movement and paleo-stress reconstructions (Figs. 8, 15G). This is interpreted to be coeval with a redistribution of strain occurring in the over-all fold and fault system. This redistribution of stain was also accompanied by the development of dilatational sites suitable for deposition of gold, particularly in V3/V4 vein arrays, from overpressured mineralized fluids.

The sets of conjugate cross-course faults (Figs. 7, 8G) are intimately linked to the last stage of E-W shortening (D3, Table 1) and the episode of V3/V4 vein development (Table 3). Their intersection with bedding-parallel movement horizons is parallel to late movement direction and associated linea-tions, observed on V3/V4 veins and cross-course faults, and with the same sense of displacement. These structures can be considered to be equivalent to lateral ramps (Horne and Culshaw, 2001) and vary from centimeter-scale (Fig. 11D) to linked systems (Fig. 10E) to throughgoing faults that create the compartmentalization of the Ballarat East goldfield (Figs. 4, 14). Therefore, throughout the evolution of the Ballarat East mineralized system there is an apparent E-W compres-sive event (Fig. 14), with no evidence of the N-S compression observed at Bendigo (Leader et al., 2013). This hypothesis is tested in the following section with a series of numerical mod-els that simulate variations in the distribution of strain and fluid flow associated with the changes in strike of the bedding-parallel faults as they cross fold limbs.

Numerical Modeling Methods and DesignIn order to understand the importance of early-formed bedding-parallel faults and the role they play in controlling the initiation of the W-dipping leather jacket system of lodes, we have used an explicit modeling technique. This allows us to constrain and test the influence of changes to the stress field, and evaluate how the variations in fault geometry influence the development of dilatant sites and fluid flow (Fig. 16). As described above, the leather jacket systems are characterized by a combination of four structural features: (1) in plan view, fault-bedding relationships suggest that strike and dip of pre-existing bedding-parallel faults change during the late stages of deformation (Fig. 17A, B); (2) most leather jacket vein sys-tems occur in thick sandstone units (Fig. 2B, e.g., Big or Amal-gamated sandstone); (3) leather jacket systems are related to

~45° W-dipping faults that are refracted as they transect thick sandstone units (Fig. 17D); (4) in a N-S-oriented long section the strike change occurs close to shallower plunging folds (e.g. First Chance Minor fold) adjacent to an anticlinal culmination defined by the double-plunging N-S-striking folds (Fig. 4).

In the numerical simulations, the geometric relationship between bedding and fault refraction was examined to quan-tify the optimal structural relationships for focusing dilation in the leather jackets during gold mineralization (Robinson et al., 2009). The process involved two stages of modeling. First, we constrained the orientation of stress fields using a static fault geometry and known gold distribution (as a proxy to dilation and fluid flow). This reverse engineering phase aimed to pro-duce fault kinematics consistent with those inferred from struc-tural observations and localize positive volume strain and fluid consistent with gold distribution (Fig. 17C). In a second phase of modeling, a wide range of fault-bedding relationships were evaluated under optimal stress field conditions identified in the first phase of modeling. Incremental changes to fault-bedding relationships were used to identify optimal geometric condi-tions for producing dilation, and sensitivity of dilation and fluid focusing to changes in fault attitude (Fig. 17E, F).

Model technique

FLAC3D (Fast Lagrangian Analysis of Continua; Itasca, 2001) was used to simulate deformation and fluid flow. In this code, deformation is coupled with fluid flow through the influence of fluid pressure on effective stress and the effect of deformation on permeability. FLAC3D uses the finite differ-ence method to solve partial differential equations and thus simulate the behavior and evaluate the response of the physi-cal materials according to the elastic-plastic Mohr-Coulomb constitutive law (Vermeer and de Borst, 1984; Zhang and Sanderson, 2002). As an explicit modeling method, a rectan-gular mesh of hexahedral elements represents each surface. In contrast, an implicit method is where the surface is defined by using direct interpolation of input scalar datasets (Vollgger et al., 2015). Physical properties such as density, permeabil-ity, and porosity, and mechanical properties that determine how a material responds to elastic and plastic deformation are assigned to the hexahedra in order to represent the nature of the materials that are being modeled. The rate-independent elastic-plastic constitutive Mohr-Coulomb law governs the response of materials to deformation.

Deformation in the FLAC3D models involves the execu-tion of a series of numerical time steps. One time step repre-sents one revolution of a calculation cycle in which equations of motion are used to derive new velocities and displacements from stresses and forces that were derived from the execu-tion of the previous cycle. The new velocities are then used to derive new strain rates, which are then used to calculate new stresses (Itasca, 2001). In FLAC3D fluid flow obeys Darcy’s Law, described by Bear and Verruijt (1987), and is driven by changes in the hydraulic head brought about by changes in pore pressure that result from positive volumetric (dilation) changes due to deformation.

Model design, boundary conditions, and material properties

In the first set of models a static geometry was tested under widely varying stress field conditions. The model consists of

GOLD MINERALIZATION, BALLARAT, VICTORIA, AUSTRALIA 1093

a single fault that changes dip and strike as it crosscuts an upright sequence of slate and sandstone that also varys in strike (Fig. 17C). The model is based directly on the PR405 leather jacket vein system (Fig. 16) exposed at the 385/405 level. In the model, a single fault crosscuts a central sandstone unit at an angle of ~45° and dips toward 260°. Within the adja-cent shale and sandstone units the fault dips more steeply and toward 235° to 245°. The models are placed at mean depths of 5 to 9 km (Robinson et al., 2009), which is consistent with the

estimated depth of greenschist metamorphism in the Lachlan orogen (Wilson et al., 2009) and the estimated depths where fluids can reach near-lithostatic pressures (Sibson, 2000). Vari-ous stress field types were applied to the model (compression, sinistral transpression, dextral transpression) with incremen-tal variation in compression directions through 180°. The sec-ond set of models uses a parametrized fault refracting through a sandstone unit where the angular relationship between the fault and bedding is systematically changed while the stress

B

C

D

East West

A

Bedd

ing

Faul

t

N

N

N

N

N

N

N

N

N

N N

N

N

I

I

F

F

Fig. 16. Geometric changes in bedding and fault orientations affecting quartz vein development associated with a leather jacket lode (PR405). (A). Plan view, looking south of bedding and fault traces observed in south ore drive and structural ori-entation data for bedding (S0), cleavage (S1), and fault (in blue). The stereonet data plotted on equal-area lower hemisphere projections is related to N-oriented coordinates. The lodes are offset by minor ~SE-NW-trending cross-course faults (broken orange line). The paleostress and sense of movement on these faults can be related to N-to-S transport. (B-D). Photographs of vein development in mining faces (looking south) along the leather jacket structure with quartz veins following the steep-dipping bedding and offset by W-dipping faults (F-F in B). The orientation data show the relationship between bedding and faults, which intersect with a southerly plunge parallel to lode orientation. In (D) an apparent normal movement is observed on the W-dipping fault as a result of late reactivation.

1094 WILSON ET AL.

field is kept constant. The model is based on variations in dip and strike of the fault as it crosses from an interbedded sand-stone-shale packages, refracts through a thick sandstone unit, and then passes back out into another sandstone-shale pack-age (Fig. 17E-F). The magnitude of bulk shortening applied in these models was 2%, which is consistent with structures associated with mineralization (Fig. 8), where late fault reac-tivation records relatively small displacements (Figs. 14, 15).

The fluid pressure conditions in the models were initialized at 75% of lithostatic pressure to ensure upward flow of fluid along the faults, which is consistent with mineralization mod-els that propose fluids were sourced at depth (Fairmaid et al., 2011, Wilson and Leader, 2014). Furthermore, the depth of mineralization and the tensional vein arrays suggest that purely hydrostatic fluid pressure conditions were unlikely to have existed. The lower and vertical boundaries were impervi-ous to fluid while fluid was free to flow out of the model at the

upper boundary. The material properties applied to the rock units (Table 4) were adjusted to a point where the models best simulate what is observed in the field and are based upon val-ues used in previous models for rock units in central Victoria.

Numerical Model ResultsA summary of key results of the models is displayed (Figs. 18, 19) as a geometrical representation of shear strain, vol-ume strain, and fluid-flow plots after 2% shortening, and is described below:

Model set 1

In assessing the potential stress field it is only those models producing a NW-over-SE hanging-wall transport that local-ize shear and volume strain that would be responsible for the development of dilatant sites around the central, low-angle, W-dipping fault segment. In particular, where a deformation

MN TN

385/405 level

central

fault

outer fault

a

aa

cb

b

b

b

c

c

strike of central faultin sandstone

dip of central faultin sandstone

dip central faultin sandstone-shale unit

c = central faultb = outer sandstone-shale unita = sandstone unit

stress �eldcompressiondirection

BAC

D E FFig. 17. Numeric model design and geometry. (A). The model is based on the leather jacket exposed on the 385/405 level. (B). The key inputs were the orientation of bedding and faults within and outside of the leather jacket. (C). Three-dimen-sional mesh geometry in FLAC with bounding sandstone-shale package removed. Blue = fault, green = sandstone, and red = shale separated by siltstone (uncolored). (D). Two-dimensional illustration of geometries of faults that host the leather jacket vein arrays as they pass through steep-dipping sandstone. These faults may be either planar or refracted where they encoun-ter the adjacent interbedded sandstone-shale package. (E). Geometry used in second set of models where the fault refracts through sandstone (yellow) from adjacent sandstone-shale package (unit b). (F). Convention used in second set of models. In each individual figure the compression direction and central fault through sandstone is kept constant. The dip of central fault in sandstone is also kept constant, whereas the dip of fault through the sandstone-shale package is varied.

GOLD MINERALIZATION, BALLARAT, VICTORIA, AUSTRALIA 1095

involves: (1) compression at 110° to 120° (Fig. 18A, B); (2) sinistral transpression, involving a 90° compression compo-nent with a 1:1 compression to strike/slip ratio (Fig. 18C). Under these conditions, strain and fluid flow localize along N-NW (~170° strike) sections of the fault and adjacent to the

changes in dip when the effective transport direction is NW-over-SE in the range of 110° to 145°. With other orientations or with dextral transpression the dilation or fluid flow is not localized in the region where fault transects the central sand-stone unit.

Model set 2

These generic models illustrate the effects of how geometric variance in the fault may have an impact on the localization of strain and fluid flow (Fig. 19D, F). In these simulations the same fault-bedding geometry was used, however, the dip and strike of different sections of the fault are varied while the stress field is kept constant A total of 260 models were run (Robinson et al., 2009), the results of which were assessed based on identifying fault-bedding relationships that maximize dilation and fluid flow. The best results (Fig. 19) highlight the optimal fault orientations that localize fluid flow within the central sandstone. In order of relative importance these are (1) dip of the fault within the outer sandstone-shale packages would have an optimal dip of 40° to 50°; (2) dip of fault in central sandstone would have optimal dip of 25° to 35°; (3) dip direction of fault within sandstone sequence would be 250° to 270°.

Table 4. Mechanical and Physical Properties Compatible with Field Observations that are Applied to the Models

Psammopelite (sandstone Sandstone Shale and shale) Fault

Bulk modulus (Pa) 3.77e10 5.58e10 3.2e10 1.5e10Shear modulus 3.98e10 3.32e10 1.92e10 1.0e10Density (Kg/m3) 2500 2600 2500 2400Cohesion (Pa) 7.1e7 6.1e7 4.0e7 3.0e7Tensile strength (Pa) 5.1e6 6.1e6 4.0e6 3.0e6Friction angle (°) 30 30 30 30Dilation angle (°) 3 3 4 10Permeability (m2) 1.0e-14 1.0e-15 1.0e-14 1.0e-14Porosity (m2) 0.2 0.2 0.2 0.2

Notes: The properties are based on those used by Schaubs and Zhao (2002), Robinson et al. (2006), Leader et al. (2013), and Wilson and Leader (2014)

A

B

C

Volumetric strain Fluid fluxShear strain

Fig. 18. Results from assessing stress field after 2% bulk shortening, showing distribution of volumetric strain (dilation); shear strain and fluid flux plots. A green star means a good result based upon the localization of strain and fluid flux around the central, low-angle fault segment. (A-B). Strike of bed (black) and fault (blue) are kept constant, but compression direction was rotated from E-W, with best result at 120°. (C). Sinistral transpression with 1:1 strike slip to compression ratio, with an effective transport direction of ~135° produces a good result.

1096 WILSON ET AL.

The results from the models are largely consistent with observed relationships between structures and gold dis-tribution. Results in terms of fluid localization (i.e., around the change in fault dip) are also consistent with the observa-tions made in the mine. With a small component of sinistral transpression, greater shear strain, volume strain (dilation), and fluid flux are localized on the bedding contacts. This is consistent with the observation of late subhorizontal slip on bedding-parallel fault planes being associated with the main phase of mineralized fluid infiltration that accompanies the V3 to V4 vein formation.

Discussion

Role of fold amplification

The major folds associated with mineralization at Ballarat East are double plunging with curvilinear fold hinges. This is particularly common throughout the central Victorian

goldfields (Schaubs et al., 2006; Leader et al., 2013), where fold interlimb angles average 30° to 40° (Cox et al., 1991; Willman, 2007). However, the Ballarat East folds are tight to isoclinal (interlimb angles 0°–30°), which suggests there was significantly greater strain, and the slip directions within fold culminations would have become more complex because of the additional strain (e.g., Tanner, 1989; Horne and Culshaw, 2001). This complexity is seen in the core of these folds where low-displacement W-dipping faults, hosting the majority of mineralization, are concentrated toward the vertical or over-turned eastern limb.

The formation of the cross-course faults is also consistent with fold amplification and the development of double plung-ing folds. Deformation parallel to the fold axes, between the cross-course faults, is interpreted to have been one of distrib-uted extension (Fig. 14C, D). This extension was accommo-dated by initiation of the cross-course faults and sets of V3/V4 veins at a moderate angle to the plunging fold axis (Fig. 14C,

Fluid fluxShear strainSection view(vary fault refraction)

A

B

C

Plan View(vary fault orientation)

Fig. 19. Variations in shear strain and fluid flux identified in generic fault models after 2% bulk shortening related to dip and dip directions of the central fault. The legend for the conventions used is illustrated in Figure 17F. Changes in fault dip are shown in the cross-section sketches down the left-hand side of the figure. (A). Central fault in sandstone strikes 270° and dips 25°, while central fault in adjoining sandstone-shale package dips 50°. (B). Fluid flux decreases with increase in dip in central fault to 60° in adjoining sandstone-shale package. (C). With central fault changing strike to 260° in sandstone and a dip of 35° there would be a further decrease in fluid flux.

GOLD MINERALIZATION, BALLARAT, VICTORIA, AUSTRALIA 1097

D). The extensional event created the dilation sites where the mineralized fluids could reside to form the V3/V4 networks. Few of these extensional veins show any signs of deforma-tion and they are interpreted to have formed very late in the deformation history. These veins represent transient episodes of embrittlement, and the oscillation in the pressure of the fluid phase as seen by the presence of localized stylolites (Fig. 11E).

The fluid system

The distributions of bulk gold grades (Fig. 4) suggest that there was structural controlled fluid compartmentalization separated by major ENE cross-course faults. Between com-partments the fault-cored folds and fractured hinges acted as along-strike fluid conduits with relatively good hydraulic con-ductivity and porosity. The less fractured rocks on the limbs were isolated from these structural fluid pathways. The distri-bution of quartz veins and gold grades also suggest that cross-course faults may have reduced fluid flow along strike.

As well as facilitating folding via flexural slip, the limbs of the large amplitude folds host bedding-parallel veins, hydro-thermally altered early dikes (Fig. 12), and altered indicator beds. This suggests that the strata on the eastern limb also acted as an important vertical fluid conduit. The modeling and field observations indicate there is little to no cross-stratal flow until units are intersected by W-dipping leather jacket faults, which is an effective means of concentrating buoyant overpressured fluids near the crests of the anticlines. Fold kinematic effects (Fig. 14) lead to important spatial or tempo-ral variations in connectivity and permeability and affect the movement, storage, and release of fluids in and adjacent to the faults. The evolving mechanical properties of the rocks not only control the number and distribution of fractures, but also control when these structures form during the folding process.

The numerical modeling results described here also dem-onstrate a geometric dependence on the localization of dila-tion and shear strain and, thus the potential localization for fluid flow and consequently mineralization. In all models, strain and fluid flow are localized around refracted faults (Figs. 18, 19), particularly at the point of inflection between the steeper and the shallower fault sections. With fault strike changes, there was increasing strain and fluid fluxes occurring toward the northern end of a NNW-trending fault segment. However, when the SW-over-NE transport is applied, as a consequence of the out-of-plane movement (Fig. 14), strain and fluid flow concentrates at the southern end of the fault.

In the model we propose for the Ballarat East deposit, the fluid infiltration is time dependent (Fig. 14). A necessary pre-cursor is the tightening of the anticlines before there was a temporal distribution of fluids into the fault-cored folds. The source of the auriferous fluids is probably overpressured fluid reservoirs lying at deeper crustal levels (Willman et al., 2010; Wilson and Leader, 2014). Propagation of these high pore-pressure fluids into the folded structures is facilitated by the fracture system, as it would provide the permeability enhance-ment necessary for the hydraulic connectivity between fluid reservoir and deposition site. It is the out-of-plane extension in the fault-cored folds that has enhanced the permeability. The development of the accompanying extensional fracture

network would be out of equilibrium with peak stress condi-tions responsible for the folding. However, the network pro-vided a pathway for fluid migration along brittle fractures and a direct connectivity for fluids migrating along fold limbs into fold cores. On a deposit scale, fluid migration is inferred to have been episodic, which led to localization of high-grade gold deposition and vein development within the dilating extensional features.

ConclusionsIn the Ballarat East deposit, different structural and aurif-erous compartments, separated by cross-course faults, are affected in different ways. Where there are no parasitic folds there is a small amount of mineralization and the area is domi-nated by V2 vein formation. Where there are changes in fold plunges involving parasitic folding, there is an abundance of V3/V4 veins, with high gold grades associated with late oblique strike-slip movement on the leather jacket faults and a NW-SE and N-S motion. The evolution of strike-slip movement on the leather jacket faults and distribution of high-grade min-eralization appears to be associated with the tightening of the folds. Differing complexities in the form surface of the com-petent units facilitated a different near-field stress regime, which had the potential to focus fluids and mineralization. Despite the apparent structural complexities, changes in the form of the faults and veining are consistent with variation in the geometry of bedding that is associated with strike changes and plunge of the fold axis.

Although the W-dipping faults are historically reported to have reverse offsets of several meters, many faults have mini-mal reverse movement (<0.5 m), some even with an apparent normal sense. The normal movement we interpreted as a sig-nificant relaxation phase where faults had dilated in response to an extensional episode associated with the out-of-plane movement during the final stage of an E-W contraction. This last stage of movement is evidenced by lineation orientations and displacement on bedded faults, early veins, and offsets on V3/V4 veins and cross-course faults, in an inclined N-over-S manner. This sense of movement is inconsistent with the expected sense of transport during flexural slip folding; occur-ring where folds are tight and where bedding on the eastern limb is subvertical. Accompanying this is incremental vein formation in which cyclic fluctuations of fluid pressure con-trolled episodic vein dilation and gold deposition.

AcknowledgmentsThis research has been funded by Australian Research Coun-cil Linkage Grants (LP 0882157 and LP150100717), Bal-larat Goldfields Pty. Ltd. through Lihir Gold Ltd. (LGL) and Castlemaine Goldfields Ltd. We thank the staff at Bal-larat Goldfields and our colleagues, in particular, Steve Olsen, Charles Carnie, Brad Cox, Allison Dugdale, Glen Phillips, Luke Graefe, Troy Fuller, Andrew Allibone, Rod Boucher, Wess Edgar, Matt Herman, and Stefan Valle, for their on-site assistance and advice during our mapping of the underground drives and numerous discussions about Ballarat geology. The numerical modeling was undertaken with the support of CSIRO Earth and Science Resource Engineering and we thank Weronika Gorczyk and Peter Schaubs for their help. We thank June Hill and Julie Rowland for constructive comment

1098 WILSON ET AL.

and review of this work. Larry Meinert is thanked for his con-structive editorial overview and comments.

REFERENCESArne, D.C., Bierlein, F.P., McNaughton, N., Wilson, C.J.L., and Morand, V.J.,

1998, Timing of gold mineralisation in western and central Victoria, Austra-lia: New constraints from SHRIMP II analysis of zircon grains from felsic intrusive rocks: Ore Geology Reviews, v. 13, p. 251–273.

Atkinson, B.K., 1987, Introduction to fracture mechanics and its geophysical applications, in Atkinson, B.K., ed., Fracture mechanics of rock: London, Academic Press, p. 1–26.

Baragwanath, W., 1923, The Ballarat Goldfield: Melbourne, Geological Sur-vey of Victoria Memoirs 14, Melbourne, 257 p.

Bear, J., and Verruijt, A., 1987, Modelling groundwater flow and pollution: with computer programs for sample cases: Dordrecht, D. Reidel, 414 p.

Bierlein, F.P., and McKnight, S., 2005, Possible intrusion-related gold sys-tems in the western Lachlan orogen, southeast Australia: Economic Geol-ogy, v. 100, p. 385– 398.

Bierlein, F.P., Foster, D.A., Mcknight, S., and Arne, D.C., 1999, Timing of gold mineralisation in the Ballarat goldfields, central Victoria: Constraints from 40Ar/39Ar results: Australian Journal of Earth Sciences, v. 46, p. 301–309.

Boucher, R.K., Osborne, D.J., and d’Avergne, P.B., 2008, Stratigraphic con-trols on structures and mineralisation in central Victoria 2: Ballarat East: AIG News 92 p. 6–8.

Brock, N., 2009, An investigation of the mafic-intermediate dykes in the Bal-larat East Goldfield, Victoria, Australia: B.Sc. Honors thesis, Melbourne, University of Melbourne, 50 p.

Brown, G., and Hogan, E.J., 1932, Victoria: Gold and minerals: Melbourne, Mines Department, 100 p.

Cox, S.F., Wall, V.J., Etheridge, M.A., and Potter, T.F., 1991, Deformational and metamorphic processes in the formation of mesothermal vein-hosted gold deposits—examples from the Lachlan fold belt in central Victoria, Australia: Ore Geology Reviews, v. 6, p. 391–423.

d’Auvergne, P.B., 1990, Ballarat east gold deposits, in Hughes, F.E., ed., Geology of the mineral deposits of Australia and Papua New Guinea: Mel-bourne, Victoria, Australian Institute of Mining and Metallurgy, v. 2, p. 1277–1278.

Dominy, S.C., 2006, Inferred mineral resources and gold veins—discussion and case studies: International Mining Conference, 6th, Darwin, NT, 21–23 August 2006, Australasian Institute of Mining and Metallurgy, Melbourne, Proceedings, p. 71–83.

Dominy, S.C., Xie, Y., and Platten, I.M., 2008, Gold particle characteris-tics in narrow vein deposits—implications for evaluation and metallurgy, in Dominy, S., ed., Proceedings of Narrow Vein Mining Conference: Melbourne, Victoria, Australiasian Institute of Mining and Metallurgy, p. 91–104.

Fairmaid, A.M., Kendrick, M.A., Phillips, D., and Fu, B., 2011, The origin and evolution of mineralizing fluids in a sediment-hosted orogenic gold deposit, Ballarat East, southeastern Australia: Economic Geology, v. 106, p. 653–666.

Gregory, J.W., 1907, The Ballarat East Gold-field: Geological Survey of Vic-toria Memoir 4, p. 52.

Horne, R., and Culshaw, N., 2001, Flexural-slip folding in the Meguma Group, Nova Scotia, Canada: Journal of Structural Geology, v. 23, p. 1631–1652.

Itasca, 2001, FLAC: Fast Lagrangian Analysis of Continua, theory and back-ground: Minneapolis, Itasca Consulting Group, Inc., 200 p.

Large, R.R., Danyushevsky, L., Hollit, C., Maslennikov, V., Meffre, S., Gil-bert, S., Bull, S., Scott, R., Emsbo, P., Thomas, H., Singh, B., and Foster, J., 2009, Gold and trace element zonation in pyrite using a laser imaging technique: Implications for the timing of gold in orogenic and Carlin-style sediment-hosted deposits: Economic Geology, v. 104, p. 635–668.

Leader, L.D., Wilson, C.J.L., and Robinson, J.A., 2013, Structural constraints and numerical simulation of strain localization in the Bendigo goldfield, Victoria, Australia: Economic Geology, v. 108, p. 279–307.

Lidgey, E., 1894, Report on the Ballarat East goldfield: Victoria, Department of Mines Special Report.

Miller, J. McL., and Wilson, C.J.L., 2004, Application of structural analysis to faults associated with a heterogeneous stress history: the reconstruction of a dismembered gold deposit, Stawell, western Lachlan fold belt, Australia: Journal of Structural Geology, v. 26, p. 1231–1256.

Osborne, D.J., 2008, The Ballarat goldfield: New insights on an old model, in Dominy, S., ed., Proceedings of Narrow Vein Mining Conference: Mel-bourne, Victoria, Australiasian Institute of Mining and Metallurgy, p. 59–70.

Phillips, G.N., and Powell, R., 2010, Formation of gold deposits: A meta-morphic devolatilization model: Journal of Metamorphic Geology, v. 28, p. 689–718.

Ragan, D.M., 1973, Structural geology; An introduction to geometrical tech-niques, 2nd ed.: New York, John Wiley and Sons, 208 p.

Ramsay, J.G., 1974, Development of chevron folds: Geological Society of America Bulletin, v. 85, p. 1741–1754.

Robinson, J.A., Wilson, C.J.L., and Rawling, T.J., 2006, Numerical model-ling of an evolving gold-lode system: Structural and lithological controls on ore-shoot formation in the Magdala gold mine, western Victoria: Australian Journal of Earth Sciences v. 53, p. 799–823.

Robinson, J.A., Gorczyk, W., and Schaubs, P.M., 2009, Ballaral goldfields: Leather jacket reverse engineering and predictive modelling: CSIRO Report P2009/1522 to Ballarat Goldfields, 108 p.

Savage, H.M., Shackleton, J.R, Cooke, M.L., and Riedel, J.J., 2010, Insights into fold growth using fold-related joint patterns and mechanical strati-graphy: Journal of Structural Geology, v. 32, p. 1466–1475.

Schaubs, P.M., and Zhao, C., 2002, Numerical models of gold deposit for-mation in the Bendigo-Ballarat zone, Victoria: Australian Journal of Earth Sciences, v. 49, p. 1077–1096.

Schaubs, P.M., Rawling, T.J., Dugdale, L.J., and Wilson, C.J.L., 2006, Factors controlling the location of gold mineralisation around basalt domes in the Stawell corridor: Insights from coupled 3D deformation-fluid-flow numeri-cal models: Australian Journal of Earth Sciences, v. 53, p. 841–862.

Shackelton J.R., and Cooke, M.L., 2007, Is plane strain a valid assumption in non-cylindrical fault-cored folds?: Journal of Structural Geology, v. 29, p. 1229–1240.

Sibson, R.H., 2000, A brittle failure mode plot defining conditions for high-flux flow: Economic Geology, v. 95, p. 41–48.

Tanner, G.P.W., 1989, The flexural-slip mechanism: Journal of Structural Geology, v. 11, p. 635–655.

Taylor, D.H., Whitehead, M.L., Olshina, A., and Leonard, J.G., 1996, Ballarat 1: 100 000 map geological report: Geological Survey Report 101, 89 p.

VandenBerg, A.H.M., Willman, C.E., Maher, S., Simons, B.A., Taylor, D.H., Morand, V.J., and Radojkovic, A., 2000, The Tasman fold belt system in Victoria: East Melbourne, Victoria, Geological Survey of Victoria Special Publication, 462 p.

Vermeer, P.A., and de Borst, R., 1984, Non-associated plasticity for soils con-crete and rock: Heron, v. 29, p. 1–62.

Vollgger, S.A., Cruden, A.R., Ailleres, L., and Cowan, E.J., 2015, Regional dome evolution and its control on ore-grade distribution: Insights from 3D implicit modelling of the Navachab gold deposit, Namibia: Ore Geology Reviews, v. 69, p. 268–284.

Williams, K., and Sykes, M., 2008, Historical challenges, sodern Solutions at Ballarat East, in Dominy, S., ed., Proceedings of Narrow Vein Mining Conference: Melbourne, Victoria, Australiasian Institute of Mining and Metallurgy, p. 229–235.

Willman, C.E., 2007, Regional structural controls of gold mineralization, Bendigo and Castlemaine goldfields, Central Victoria, Australia: Minera-lium Deposita, v. 42, p. 449–463.

Willman, C.E., Korsch, R.J., Moore, D.H., Cayley, R.A., Lisitsin, V.A., Rawl-ing, T.J., Morand, V.J., and O’Shea, P.J., 2010, Crustal-scale fluid pathways and source rocks in the Victorian Gold province, Australia: Insights from deep seismic reflection profiles: Economic Geology, v. 105, p. 895–915.

Wilson, C.J.L., and Leader, L.D., 2014, Modeling 3D crustal structure in Lachlan orogen, Victoria, Australia: Implications for gold deposition: Jour-nal of Structural Geology, v. 67, p. 235–252.

Wilson, C.J.L., McKnight, S.W., Dugdale, A.L., Rawling, T.J., Farrar, A.D., McKenzie, M.J., and Melling, W.D., 2009, Illite crystallinity and the b-spacing values of white micas and their implications for gold mineraliza-tion in the Lachlan orogen: Australian Journal of Earth Sciences, v. 56, p. 1143–1164.

Wilson, C.J.L., Schaubs, P.M., and Leader, L.D., 2013, Mineral precipitation in the quartz reefs of the Bendigo gold deposit, Victoria, Australia: Eco-nomic Geology, v. 108, p. 259–278.

Zhang, X., and Sanderson, D.J., 2002, Numerical modelling and analysis of fluid flow and deformation of fractured rock masses: Permagon, Elsevier Science, 288 p.