Ore Geology Reviews 94 (2018) 326–350 Contents lists available at ScienceDirect Ore Geology Reviews journal homepage: www.elsevier.com/locate/oregeorev Entering an immature exploration search space: Assessment of the potential orogenic gold endowment of the Sandstone Greenstone Belt, Yilgarn Craton, by application of Zipf’s law and comparison with the adjacent Agnew Goldfield T Rhys S. Daviesa,b, , David I. Grovesc, Allan Trencha,d,e, John Sykesa,d,e,f, Jonathan G. Standingg ⁎ a Centre for Exploration Targeting, School of Earth and Environment, The University of Western Australia, 35 Stirling Highway, Crawley, WA 6009, Australia Alto Metals Ltd., 9/12-14 Thelma St, West Perth 6005, Australia Orebusters Pty Ltd, Gwelup, WA 6018, Australia d Business School, The University of Western Australia, 35 Stirling Highway, Crawley, WA 6009, Australia e Minex Consulting (Perth Office), 10/7 Centro Avenue, Subiaco, WA 6008, Australia f Greenfields Research Ltd, Hunters Chase, Highfield Farm, Stripe Lane, Hartwith, Harrogate, North Yorkshire HG3 3HA, United Kingdom g Model Earth Pty. Ltd., 2/80 Colin St, West Perth 6005, Australia b c A R T I C L E I N F O A B S T R A C T Keywords: Archean orogenic gold Mineral Systems Concept Zipf’s law Sandstone Greenstone Belt Yilgarn Craton Western Australia The Sandstone Greenstone Belt is an exploration-immature, regolith-covered, approximately 1000 sq. km belt, in the Southern Cross Domain of the Yilgarn Craton. In order to estimate potential endowment, historical gold production and deposit resource estimates are required to be quantitatively analysed for calculation of natural and residual gold endowment. The total residual gold endowment within the oxide zone of the Sandstone Greenstone Belt is estimated, by application of a Zipf’s law statistical assessment, to be 2.3 Moz. This mineralisation is most likely contained in extensions of known deposits and several undiscovered deposits. The fresh rock of the Sandstone Greenstone Belt remains poorly explored. However, a conceptual endowment estimate can be made, based on a minerals system comparison between the exploration-immature Sandstone Greenstone Belt and the well-explored, geologically-similar Agnew Greenstone Belt, 100 km to the east. It is possible that natural endowment at Sandstone could total 21.3 Moz, with nine undiscovered deposits of > 0.5 Moz. Application of such a minerals-system integrated endowment assessment represents an effective motivator to embark on a well-resourced gold exploration campaign in the Sandstone Greenstone Belt, a currently immature exploration search space. 1. Introduction The Sandstone Greenstone Belt (SSGB) covers an area of approximately 1000 sq. km in the Southern Cross Domain, Yilgarn Craton. Since the initial discovery of gold in the late 1800s, the belt has produced over 1.2 Moz gold from surface, open pit and shallow underground workings, far less gold per unit area than most other Archean greenstone belts (Yun, 2000). Unlike neighbouring greenstone belts, there is limited research on gold mineralisation in the SSGB, so it is unclear whether significant residual primary mineralisation remains beneath workings in oxidised regolith profiles and in lesser explored parts of the belt. Whether future exploration presents an attractive costbenefit opportunity requires an assessment of potential gold endowment. This paper considers the economic gold potential and represents ⁎ a first attempt to systematically describe the geology and gold deposits of the SSGB. It incorporates empirical, conceptual and quantitative methods to assess its mineral endowment. Pre-existing and new critical datasets are compiled for the SSGB and interrogated using a mineral systems framework to define critical elements of SSGB gold mineralising systems. The exploration database provides an overview of the maturity of the exploration search space and distribution of mineralisation. Field mapping and geophysical datasets are interpreted to outline geological domains and develop a greater understanding of the structure of the belt, thus defining preferential conduits for the transport of auriferous fluids from gold source to site of deposition. Deposit structure and geology provide insights into deposit-scale features controlling deposition and style and timing of mineralisation. Corresponding author at: Centre for Exploration Targeting, School of Earth and Environment, The University of Western Australia, 35 Stirling Highway, Crawley, WA 6009, Australia. E-mail address: rhyssamuel.davies@research.uwa.edu.au (R.S. Davies). https://doi.org/10.1016/j.oregeorev.2018.01.020 Received 2 January 2017; Received in revised form 15 December 2017; Accepted 18 January 2018 Available online 31 January 2018 0169-1368/ © 2018 Elsevier B.V. All rights reserved. Ore Geology Reviews 94 (2018) 326–350 R.S. Davies et al. 2. Research methodology geodynamic setting of an active- or paleo-convergent margin (Robert et al., 2005; Wyman et al., 2008; Bierlein et al., 2009). This suggests that the critical elements of a gold system are shared by the majority of deposit styles (e.g. Hronsky et al., 2012). From this concept, encompassing mineralising system models have been defined. These models consider the deposit scale variations for each style of gold mineralisation, and then define broader links between each of the diverse mineral systems. This concept has helped to resolve confusion brought about by conflicting deposit-scale models (Wyman et al., 2016). Orogenic gold deposits (Goldfarb et al. (2005) and extensive references therein) have a common set of characteristics. These are summarised as a spatial relationship to regional trans-lithospheric structures; structural control of mineralisation at deposit scale; low salinity aqueous-carbonic fluid compositions; similar alteration assemblages which vary systematically with metamorphic setting and host rocks; and formation at syn- to post-peak metamorphic temperatures along a crustal continuum between 180C and 1 kbar and > 700C and 5 kbar (e.g. Colvine, 1989; Groves, 1993). However, the ore-fluid source is still under debate, with several competing models (Fig. 1), including the late stages of regional metamorphism driving auriferous metamorphic fluids (Goldfarb and Groves, 2015), fluids sourced from the devolatilisation of pyritic sediments above a subducting oceanic slab (Groves and Santosh, 2016), and oxidised intrusion-related ore fluids (Witt et al., 2016). The key characteristics of a mineralising system (Hronsky et al., 2012) include: a regional gold-fertile source and geodynamic fluid migration event; primary structural conduit for propagation of auriferous fluids from a deep-seated source; permeable zones acting as fluid conduits in the mid- to upper-crust; fluids concentrated by localised structurally complex zones, with potential chemically favourable host rock; and preservation of primary depositional zone (Table 2). In this paper, the authors define an exploration search space as a region in which a specific mineralisation type has formed under similar geological conditions. This occurs within a discrete mineral deposit system, in which the ore deposits sought have formed through competition for essential mineralising components. Judging the undiscovered resource endowment of such a search space represents a challenging but vital step in the mineral exploration process. An estimate of residual endowment is key to defining the potential economic cost-benefit of selected ground. Traditional approaches to the prediction of residual endowment rely on the subjective, albeit expert, judgement of a group of geologists to identify the presence of critical geological conditions in order to predict probable grade/tonnes and the number of undiscovered deposits in an area. This approach is open to biases based on the experience and knowledge of the geologists, often resulting in over-optimistic predictions of residual mineral endowment (Fallon et al., 2010). In this study, the mineral systems framework (McCuaig and Hronsky, 2014; Hagemann et al., 2016) has been applied to the poorlyunderstood Sandstone Greenstone Belt (SSGB) to provide a greater understanding of the greenstone belt in order to aid an endowment assessment of the region. Pre-existing datasets and new critical data for the SSGB are compiled to systematically define regional- to depositscale characteristics of its gold mineralisation, and provide an initial conceptual assessment of gold endowment. Along with most Yilgarn Archean greenstone belts, the SSGB hosts significant gold deposits with > 1.2 Moz of gold produced to date. However, the SSGB represents an immature exploration search space, because exploration and mining have been conducted almost exclusively in regolith with scant information available on fresh bedrock. Hence, geological and mineral system characteristics of the SSGB are compared to the geologically similar Agnew Greenstone Belt (AGB), a more mature exploration search space, to provide an initial empirical assessment of potential gold endowment. Finally, a Zipf’s law statistical assessment of known deposits is conducted (Guj et al., 2011) to define the total (natural) and unknown (residual) gold endowment of the SSGB. Through critique of several different models, gaps in geological understanding and critical datasets are outlined. The most cost-effective ways to fill these gaps provide guidance for future exploration strategies in the region. This study provides a framework within which to conduct an integrated qualitative and quantitative endowment assessment of an immature exploration search space. 3. Geological framework of the Sandstone Greenstone Belt 3.1. Regional setting The Yilgarn Craton is divided into three Terranes and one Superterrane (Fig. 2), based on distinct crustal histories and volcanic rock geochemistry (Cassidy, 2006; Czarnota et al., 2010). The Youanmi Terrane and Eastern Goldfields Superterrane (EGS) consist of northtrending greenstone belts with extensive granitoids (e.g. Cassidy, 2006). The Archean tectonic evolution of the Yilgarn Craton remains highly controversial. The Yilgarn craton is widely thought to have evolved through Neoarchean subduction, arc magmatism and accretion of outboard terranes on to a tectonic foreland; the isotopically distinct Youanmi Terrane (summarised by Krapez and Barley (2008), and references therein). An alternative model suggests granitoid plutons were derived from a mantle plume which melted older continental crust with ascending plutons deforming the greenstone belts independently from any plate-tectonic processes (Campbell and Hill, 1988). Van Kranendonk et al. (2013) provided evidence that the Youanmi Terrane and the EGS share a similar Neoarchean history and evolved together as a single crustal element affected by Neoarchean plume-related magmatism and rifting (Table 3). Across the Yilgarn Craton, orogenic gold mineralisation (Groves et al., 1998) occurred late in the structural history, accompanying the development of anastomosing NNW- to NNE- striking brittle-ductile shear zones between 2.66 and 2.63 Ga (Bateman and Hagemann, 2004; Robert et al., 2005; Blewett and Czarnota, 2007; Van Kranendonk et al., 2013; Vielreicher et al., 2015). These trans-lithospheric structures are defined as D4, D3 and D2/3 in the Murchison, Southern Cross and EGS, respectively (Table 4). 2.1. Mineral systems concept The modern mineral systems concept has evolved because deposit attributes, such as size and frequency, follow power-law distributions, much like earthquakes (Hobbs and Ord, 2015). Earthquakes are rare events, considered to be the result of large-scale energy flux because of self-organised critical behaviour, where energy is introduced into a system but then prevented from escaping until a critical point is reached where the stored energy overcomes its barrier to dispersion (Bak, 1996). Based on this, Hronsky (2011) proposed that ore-forming systems are examples of such self-organised critical systems. Since Wyborne et al. (1994) proposed six critical elements for ore formation, based on those developed by the oil industry (cf Magoon and Dow, 1994), several authors have constrained the critical elements of the mineral systems concept and applied the principles to different mineral deposits including Archean lode-hosted gold systems (e.g. McCuaig and Hronsky, 2014; Hagemann et al., 2016: Table 1). The mineral systems concept is currently being used to improve understanding of gold mineralising systems (e.g. Groves et al., 2016; Wyman et al., 2016). It is widely accepted that gold mineralisation can occur within a diverse range of deposit styles, commonly proximal in space and time (Huston et al., 2016), and generally within the same 327 Ore Geology Reviews 94 (2018) 326–350 R.S. Davies et al. Table 1 Critical elements of mineral system; as defined by McCuaig and Hronsky (2014) and Hagemann et al. (2016). restricted to the south-eastern limb of the SSGB. Complex fold interference patterns within the BIF marker units indicate at least two phases of folding. The Mafic Domain appears overlain by the Ultramafic Domain, comprising a poorly-exposed succession of ultramafic (both aluminium depleted and un-depleted komatiites) and high-magnesium mafic volcanic rocks and interflow oxide-facies BIF. The Ultramafic Domain is characterised by extensive folds, fold interference patterns and fault dislocation of stratigraphy, clearly representing three stages of deformation. The Sedimentary Domain, at the norther apex of the belt hosts shale and siltstone, intercalated with BIF and chert, as well as minor, deeply weathered ultramafic rocks. It displays megascopic isoclinal folds and appears to lie unconformably above the succession of the Mafic Domain. Several structures have been proposed to explain relationships between these poorly exposed domains (Fig. 7). The Mafic-Ultramafic Domain bounding Sandstone Decollement, which initiates and terminates at either end of the Dandaraga Fault, together with the northern Sedimentary Domain-bounding Fault appear to be most important. Internally, a variety of felsic rocks intrude the SSGB, the majority within the Ultramafic Domain. There are predominantly strongly-foliated porphyritic monzogranites, gneissic granite, and granitic gneiss along granite-greenstone contacts and regional-scale ductile shear zones. Major granitoid plutons associated with the Diemals Dome intrude the southern margin of the belt. Rocks within the belt are characterised by low-strain, greenschistfacies metamorphism, typical for greenstones away from sheared margins (Ahmat, 1986). Peak metamorphism across the Southern Cross 3.2. Sandstone Greenstone Belt The Sandstone Greenstone Belt (SSGB) lies in the central-northern part of the Southern Cross Domain (Fig. 3), as part of the Youanmi Terrane (Chen, 2006). The geomorphology of the SSGB is complex, with considerable recent cover and deep palaeochannels between upland regions of low BIF ridges. Based on high-resolution aeromagnetic data and seismic line 10GAYU2 (Zibra et al., 2014: Fig. 4), several major structures (Chen et al., 2004) are interpreted to control the geometry of the SSGB. On the western edge of the SSGB lies the north-eastern section of the domainbounding NE-striking Youanmi Shear Zone. On the eastern side of the SSGB, the Edale Shear Zone is continuous to the southeast but merges with the Youanmi Shear Zone at the northern tip of the SSGB (Fig. 3). The crustal-scale structural similarities between the SSGB and Agnew Greenstone Belt (AGB) are striking (Fig. 4). The lithostratigraphy and regional setting of the SSGB (Chen, 2005: Fig. 5) are similar to other Yilgarn greenstone belts, including the AGB, but correlation is difficult (Chen et al., 2006). In particular, there are only geochronological data for a porphyritic microgranite intrusion into ultramafic rocks in the Bulchina pit: a SHRIMP U-Pb zircon age of 2731 ± 14 Ma and a SHRIMP U-Pb monazite age of 2731 ± 3 Ma, representing the minimum deposition age for the belt (Chen, 2005). The SSGB can be subdivided into three lithological domains (Fig. 6). Its central-northern and outer flanks comprise a Mafic Domain dominated by massive to foliated basalts, intercalated with subordinate tremolite-chlorite-talc schist, BIF, chert, and clastic sedimentary rocks, locally intruded by gabbroic sills. This Domain is characterised by strike continuity of the main marker units, and macroscopic folds are 328 Ore Geology Reviews 94 (2018) 326–350 R.S. Davies et al. Fig. 1. Summary of all intra-crustal orogenic gold models (after Groves and Santosh, 2016). downtime during the First and Second World Wars. Early significant discoveries included Hacks, Oroya and Hancock’s/Bull Oak. Modern open pit mining commenced in 1993. The need to sustain an operating mill resulted in several new discoveries and further production at existing operations. Between 1993 and 1999, production continued from Oroya, Hancock’s/Bull Oak, Goat Farm, Shillington, Two Mile Hill and Plum Pudding. From 1999 to 2010, several new open pits were exploited, including Bulchina, Lord Henry, Lord Nelson and Eureka (Fig. 8). Conventional mining and production ceased in the SSGB in 2010 with the closure of the Eureka pit. A comparison of the geological characteristics of deposits within the SSGB (Table 6) shows that they exhibit many common features, including: strong structural controls; similar hydrothermal alteration; greenschist facies metamorphism; high grade ‘free’ gold associated with thin quartz veins; relatively late timing of mineralisation; and highly effective, deep weathering processes, resulting in significant supergene enrichment. Domain is broadly contemporaneous with widespread granitoid intrusion after c. 2685 Ma (Wyche et al., 2004). Small zones of amphibolite to mid-greenschist facies metamorphism, characterised by high-strain, occur along the margins of the SSGB in close proximity to major shear zones and granitoid intrusions. The first three phases of deformation experienced by the SSGB align closely with those recorded within greenstone belts across the Youanmi Terrane. Several post-D3 deformation phases are recognised (Table 5). Although these late events represent only localised variations in the regional stress field, they have a close relationship to gold mineralisation. 3.2.1. Gold deposits Since initial discovery of gold at the end of the 19th century, the SSGB has produced about 1.2 Moz Au from alluvial, shallow underground and open-pit operations. Small-scale mining operations on vein and alluvial deposits have been ongoing in the SSGB, with periods of 329 Ore Geology Reviews 94 (2018) 326–350 3.2.2. Structural controls of mineralisation All the deposits of the SSGB exhibit strong structural controls indicative of sub-horizontal east-west compression that induced thrust displacement along north-south faults. Lord Henry, Lord Nelson and Maninga Marley are hosted by major shear zones. Bulchina is situated at the intersection of two regional shear zones. North-south fault corridors located throughout the centre of the belt host several deposits; including Twin Shafts on fault 722, Hack’s on the Hack’s Creek Fault and Hancock’s/Bulloak on fault 731. Inverted retroarc rift; preferably developed at a continental margin, or margin of deep mantle lithosphere root; long lived “vertically accretive” structure Major subcontinental scale lineament (representing long-lived zone of transverse dislocation within accretionary orogeny); long lived “vertically accretive” structure 3.2.3. Quartz veins High-grade gold mineralisation in SSGB deposits is associated with thin quartz veins, stacked or sheeted quartz vein arrays, or stockworks. Mineralisation is generally ‘free’ gold within quartz veins, with only refractory ore, hosted by sulfidic shale, at Bellchambers. Hydrothermal alteration is strongly controlled by structures, such as brittle-ductile or ductile shear zones and their associated vein systems. Sericite and carbonate alteration halos are common around the quartz veins. Hydrothermal alteration in the Lord Henry open pit exhibits a strong arsenic and base-metal association, whereas other deposits show relatively simple sulfide mineralogy. Wallrock alteration is characterised by selveges of sericite (+ fuchsite) + carbonate + sulfide minerals; most commonly pyrite and arsenopyrite, but in some deposits also sphalerite, chalcopyrite, galena and molybdenite. The gold content of the hydrothermally altered wallrock tends to be low (< 1.0 g/t Au) and restricted to alteration zones. 3.2.4. Metamorphic grade Apart from the Lord Nelson pit, which is characterised by biotite alteration (upper greenschist conditions), deposits across the SSGB exhibit sericite-carbonate-chlorite alteration, broadly compatible in pressure-temperature with low- to very-low grade greenschist-facies metamorphism. N/A at this scale N/A at this scale Discreet au-enriched upper lithospheric domain, particularly near its margins; potentially mantle lithosphere enriched by small volume partial melts prior to termination of orogeny Currently unclear but the occurrence of the western US gold superprovince suggests that some control at this scale exists Deposit Camp Province 3.2.5. Timing of gold mineralisation Based on previous deposit-scale studies (e.g. Standing, 2000), it is inferred that gold mineralisation was late within the evolution of the SSGB, during an E-W compressive regime. However, the exact timing is not clear and represents an important area for further research. 3.2.6. Supergene enrichment The SSGB displays an extensive weathering profile, with many drill holes not intersecting fresh rock, which is generally around 60–70 m depth. Lord Henry provides a notable exception, with near-surface fresh rock and the explored resource of almost entirely primary-zone mineralisation. As a result, in most deposits, mineralisation is considered open at depth, with potential for primary-zone mineralisation. 3.2.7. Different deposit features Gold has been mined from all stratigraphic domains and most lithological units of the SSGB, with the majority of economic deposits situated closer to the centre of the belt, although this may be a function of exploration focus. At Bulchina, gold-bearing veins are locally developed within a felsic complex intruded into the major N-S shear zone, and bounded by ultramafic rock to the east and mafic rock to the west. The Lords pits are characterised by gold mineralisation hosted in sheared granodiorite, proximal to a footwall contact with ultramafic rocks. Mineralisation at Hack’s, Havilah and Maninga Marley is hosted by hydraulically-fractured and sheared dolerite, in contact with ultramafic rocks. In Bull Oak and Shillington, high-grade gold is associated with fault intersections within BIF. At Two Mile Hill, a near-vertical tonalite stock is pervasively mineralised, with high-grade mineralisation along contacts with BIF. The variety of host rocks is similar to that in other Archean cratons and generally corresponds to high-degrees of competency contrast, providing selective fracturing and shearing. Certain lithologies, such as Scale Continental N/A at this scale Ore-shoot N/A at this scale Period of low active tectonic strain, e.g. stress switch causing transient neutral stress state causing fluid system to self-organise; areas of greatest uplift favoured (provides stress switch and high rates of energy and mass transfer) Terminal phase of synore orogenic event (e.g. the transition to insipient extension associated with the termination of collision and locus of subduction retreating oceanward) A major collisional orogenic event within the historic of an evolving accretionary orogen; the major collision that actually terminates a long-lived (> 200 Ma) is most prospective and usually associated with a peak of supercontinent formation Major heterogeneity (e.g. cross-structure intersection) along trend of inverted rift-axial (or rift-marginal) fault with associated physical seal (e.g. antiformal culmination or unconformity) 2nd order – pressure drops 3rd order – favourable substrate 1st order – upper 10 km of crust at time of the mineralising event where fluid pressure (+T, X) gradients are greatest; preserved through multiple orogenic cycles Localised dilatant zone in conduit-hosting structure Pipe-like rock volume more favourable for fluid exit pulse (either local structural complexity or pipe of more competent rock) Primary depositional zone Favourable architecture Favourable geodynamics Fertility Critical elements Table 2 Critical and constituent elements in the formation of gold deposits (after Hronsky et al., 2012). Notes: Note the scale-dependent nature of the constituent elements; these elements require translation into features that can be mapped directly in existing or obtainable geoscience datasets at the appropriate scale to generate targets (e.g. McCuaig et al., 2010). R.S. Davies et al. 330 Ore Geology Reviews 94 (2018) 326–350 R.S. Davies et al. Fig. 2. Tectonic setting of the Sandstone and Agnew Greenstone Belts of the Yilgarn Craton (modified from Chen et al., 2005). dextral strike-slip Waroonga Shear Zone, and lies just east of the domain-bounding Ida Fault. The centre of the AGB is dominated by the Lawlers Anticline, which was formed during D2 compression and plunges 50–60° N (Platt et al., 1978). The Lawlers batholith, a granite intrusion emplaced at 2666 ± 3 Ma, lies at the core of the Lawlers Anticline (Fletcher et al., 1998, 2001). The lithostratigraphy of the AGB is similar to that of the Kalgoorlie Domain, consisting of mafic-ultramafic sequences and sedimentary basins metamorphosed largely to amphibolite facies (McCluskey, 1996; Squire et al., 2010). The Lawlers Greensone Sequence consists of the mafic-ultramafic succession of the Lawlers Basalt, Agnew Ultramafic unit and basaltic Redeemer Formation, with a maximum depositional age of 2692 ± 3 Ma (Squire et al., 2010). A 1500-m-thick sedimentary sequence, lying unconformably on the Lawlers Greenstone Sequence, consists of volcaniclastic conglomerate, overlain by quartzo-feldspathic sandstones of the Scotty Creek Formation (Perriam, 1996). A maximum the Bulchina quartz porphyry, are suited to brittle hydrofracturing, resulting in vein stockworks. Iron-rich tholeiitic basalts, differentiated dolerite sills and BIF together with ferruginous shales act as distinct chemical traps, promoting sulfidation and deposition of gold. The key elements of the mineral system within the belt, along with targeting elements and proxy datasets, are summarised in Table 7. 3.3. Agnew goldfield The Agnew goldfield (AGF) lies within the AGB, which covers an area of approximately 1000 sq. km in the southwest corner of the Agnew-Wiluna belt (Fig. 9), in the Eastern Goldfields Superterrane (Fig. 2). The AGF and AGB have been reviewed in detail (e.g. Thébaud et al., 2012; Jowitt et al., 2014; Voute and Thébaud, 2015) and are only summarised here. The belt is tightly folded, bounded to the west by the 2 km wide Table 3 Summary of features of the terranes of the Yilgarn Craton (modified from Cassidy, 2006). Terrane Age of initial crust formation (By) Depositional ages of greenstones (By) Emplacement ages of granites and gneisses (By) Age of deformation and metamorphism (By) Narryer Terrane 3.8–3.4 ?3.73 c. 3.3, ?2.75, 2.68–2.62 South West Terrane Youanmi Terrane ?3.5–3.0 ?3.4–2.9 ?3.0, ?2.8, 2.7 3.01–2.92, 2.81, 2.76–2.72 EGS Kalgoorlie Terrane ?3.0–2.8 ?2.94, ?2.81, 2.74–2.66 Kurnalpi Terrane (western domains) Kurnalpi Terrane (eastern domains) Burtville Terrane 2.9–2.8 2.71–2.68 3.73–3.6, 3.48, 3.3, 3.0, 2.75, 2.68–2.62 ?3.2, 2.8, 2.68–2.62 3.01–2.92, 2.81, 2.76–2.68, 266–2.62 2.81, 2.75–2.74, 2.68–2.66, 2.65–2.63 2.7, 2.68–2.66, 2.65–2.63 2.67–2.63 ?3.1–2.9 > 2.8, 2.72–2.68 ?2.95, 2.71, 2.68–2.66, 2.65–2.63 2.67–2.63 ?3.0–2.8 2.81, 2.77–2.66 ?2.95, 2.8–2.77, 2.69–2.63 ?2.67–2.63 331 ?2.68–2.62 > 2.74–2.68, 2.66–2.63 2.67–2.63 Ore Geology Reviews 94 (2018) 326–350 R.S. Davies et al. Table 4 Summary of deformation events of the Yilgarn Craton (modified from Chen, 2003). Murchison Domain Southern Cross Domain Eastern Goldfields Superterrane D1 layer parallel fabrics, isoclinal and recumbent folds D2 N-S compression: E-trending folds D3 E-W compression: upright folds, foliation, and gneissic banding D4 E-W compression: NE- and NNE-trending dextral shear zones D1 N-S compression: layer-parallel foliation, thrusts, tight to isoclinal folds in the lower greenstone succession D2 E-W compression: upright folds, axial planar foliation, gneissic banding D3 E-W compression: NE-trending dextral and NW-trending sinistral shear zones, arcuate structures Time-scale Ca. 3.0 Ga D1 N-S compression: thrusts, recumbent folds D2: ENE-WSW shortening: upright folds, thrusts, foliation D3 E-W transpression: NNE-trending sinistral shear zones Ca. 2.7 Ga Ca. 2.67 Ga 3.4. Comparison between the Agnew Greenstone Belt and Sandstone Greenstone Belt depositional age of 2664 ± 5 Ma was obtained from U-Pb SHRIMP dating on detrital zircons (Fletcher et al., 2001). Although recent studies have identified up to twelve local deformation events, the widely accepted structural history of the belt incorporates the four principal deformation events recorded across the EGS (Table 4). This indicates a shared tectonic history with the Youanmi Terrane from 2.72 Ga, including gold mineralisation at 2.66–2.63 Ga (Vielreicher et al., 2015). The gross architecture and mineralisation of the belt is controlled by broadly E-W compression, with orogenic gold deposits associated with N-S trending shear zones such as the Emu and Table Hill Shear Zones (Voute and Thébaud, 2015). Gold deposits are hosted by a wide range of rock types, have relatively simple sulfide alteration mineralogy, an association with regional scale structures, and form within zones of localised structural complexity (e.g. Thébaud et al., 2012; Jowitt et al., 2014; Voute and Thébaud, 2015). The AGB represents a mature exploration search space with significant gold mineralisation discovered within both supergene-enriched oxide and hard-rock primary ore zones: for example, the Turret deposit (Voute and Thébaud, 2015). Importantly, the mature Waroonga deposit hosts significant mineralisation to depths exceeding 1000 m (Sander, 2014). In comparison, the nature of mineralisation within the SSGB is poorly understood, because drilling coverage is mainly restricted to the oxide zone (Fig. 10A), with the bedrock of the SSGB representing an immature search space. From a total of over 14,000 holes drilled in the SSGB, only ∼800 (6%) exceed a depth of 100 m (Fig. 10B) and only about 100 (< 1%) exceed 200 m (Fig. 10C). The gold endowment of the SSGB is similar in scale to the oxide endowment of the AGB in 1996 (Table 8). At that time, the six major deposits of the AGB had produced 1.17 Moz from open pit and 0.74 Moz from underground operations (Fig. 11: Aoukar and Whelan, 1990; Broome et al., 1998; Inwood, 1998). To 2016, the AGB has produced over 13 Moz of gold with > Fig. 3. Interpreted bedrock geological map of the regional setting of the SSGB (using GSWA 1:100,000 interpreted bedrock geology map series). 332 Ore Geology Reviews 94 (2018) 326–350 R.S. Davies et al. Fig. 4. Interpreted sections of the 10GA-YU2 seismic traverse line (Ivanic et al., 2014) showing: a. migrated seismic section; and b. interpreted geological section. The SSGB and AGB are outlined by the boxes (after Zibra et al., 2014). Note the similar crustal sections for the SSGB and AGB. Granites similar to the Lawlers Tonalite are present in the SSGB but not shown on this section. contemporaneous Youanmi Shear Zone and Ida Fault, respectively. The distribution of gold mineralisation within both belts is predominantly controlled by second-order, N-S trending belt-scale structures. These generally propagate along lines of weakness, such as contacts between lithologies with high competency contrasts. Mineralisation at the deposit scale is focussed by third-order, physical throttles and host rocks that act as chemical traps, such as antiforms, intersecting shear zones, cross-cutting fault structures, and fractured iron-rich host rocks. Therefore, it is possible that the gold endowment of the explorationimmature SSGB is broadly equivalent to that of the much more mature AGB, but has not been realised due to lack of targeted exploration and drilling outside the more obvious near-surface gold anomalies under limited regolith cover. There are a number of caveats involved in any comparison between terranes, even though there appears to be a broadly synchronous craton-wide source of gold (Hronsky et al., 2012). At the terrane scale, the AGB lies within the more primitive (in terms of Nd model age) Eastern Goldfields Superterrane (Fig. 2) with a greater number of world-class gold deposits than the Youanmi Supererrane that hosts the SSGB (McCuaig et al., 2010: Fig. 3). However, the Eastern Goldfields Supererrane has approximately an order of magnitude greater abundance of greenstone belts than the Youanmi Superterrane. Further, these greenstone belts are largely at greenschist facies, where conditions close to the brittle-ductile transition favour development of structural permeability and consequent greater fluid flux, compared to the more ductile, dominantly amphibolite-facies domains of the Youanmi Superterrane, particularly the Southern Cross Terrane. Importantly, most gold deposits in the AGB lie close to the greenstone belt margins in amphibolite-facies and lie at the lower end of the scale in terms of gold endowment than deposits in greenschist-facies domains of the Eastern Goldfields Supererrane. Hence, a broad comparison between the SSGB and the AGB isreasonable. Other factors that need to be 3.5 Moz of reserves yet to be mined: an increase of over 680% in 20 years. An endowment of ∼13,000 oz per sq. km for the mature AGB, contrasts markedly with the ∼1200 oz per sq. km of gold produced from the immature SSGB. The minerals systems that led to gold mineralisation within the SSGB and AGB share several important regional to district-scale geological characteristics (Table 9). The belts are a similar size (approx. 1000 sq. km) and have similar geometry due to regional scale folding from E-W compression and the complex nature of intrusion of granitic batholiths into their southern margins (Fig. 12). They display similar mafic-ultramafic lithostratigraphic assemblages, although the AGB contains more sedimentary units. Peak metamorphism occurred at approximately the same time across the belts, with the margins of the SSGB and the entirety of the AGB reaching amphibolite-facies metamorphic conditions. Their tectonic and gold mineralisation histories are shared from 2.72 Ga. Although the gold-deposit forming processes across the Yilgarn Craton remain under discussion, gold mineralisation is developed within greenstone belts across each tectonic subdivision. In agreement with Hronsky et al. (2012), this suggests a shared craton-wide source of gold, where a deep-seated, gold-enriched reservoir can be accessed in a diverse range of tectono-magmatic settings. Outside of the Witwatersrand Basin, South Africa, most gold deposits form in convergent settings. Studies of the tectonic history and timing of gold mineralisation across the Yilgarn Craton indicate a shared geodynamic driver of a ubiquitous auriferous fluid source and related gold dispersion. However, the most plausible model for an ultimate gold source in the Yilgarn remains a topic of discussion. Trans-lithospheric structures are highlighted as pathways for auriferous fluid migration from a deep-seated, fertile gold source (e.g. Groves et al., 2016). Both the SSGB and AGB lie in close proximity to such deep-seated, domain-bounding structures; the broadly 333 Ore Geology Reviews 94 (2018) 326–350 R.S. Davies et al. three-part approach in the assessment of orogenic gold endowment for the Bendigo and Stawell Zones in Victoria and the Mossman Orogen in Queensland (Lisitsin et al., 2010, 2014; Lisitsin, 2016). Many natural and socio-economic systems (e.g. atmospheric, environmental, economic, linguistic) appear to abide to power laws. The size of mineral deposits also follows such a distribution (e.g. Figs. 13A and 13B Guj et al., 2011; Hronsky and Groves, 2008). Mineralisation is the combined result of several critical elements (Table 1). To determine the overall efficacy of these combined elements to produce mineralisation, each element (in terms of their significance and ‘size’) should be quantified and then multiplied together (see multiplication rule: e.g. Megill, 1988). By considering a simple example: the calculation of an area from x, y and z values, through the multiplication of normally distributed variables, results in a lognormal distribution, defined by a power law. Zipf’s law is one such statistical law (Zipf, 1949), applied initially to the use of words in languages, income distribution, immune-system responses and city populations, then later to petroleum (e.g. Schuenemeyer and Drew, 1983; Merriam et al., 2004) and mineral exploration (e.g. Folinsbee, 1977; Paliwal et al., 1986). Zipf’s law is a discrete expression of the continuous Pareto distribution (Pareto, 1927). It is more representative of underlying endowment in immature search spaces than the fractal-related Pareto distribution, as it provides more conservative estimates of total endowment (Guj et al., 2011). Using Zipf’s law, the size of a deposit is expressed mathematically as Sr = S1 r −k , where r is the rank of the deposit in a sequence of deposits arranged in the order of descending sizes, S1 is the size of the largest deposit in the sequence, Sr is the size of the rth deposit, and k is a constant. Therefore, the total endowment of a search space (E) is exn pressed as E = ∑r = 1 Er r k , where Er is the endowment of the rth deposit in a decreasing rank-ordered sequence and n is the number of deposits considered in the analysis, which can be defined by setting a minimum individual deposit endowment considered in an analysis. Using Zipf’s law, the estimated endowment of all deposits in a given search space is dependent on two parameters; the largest deposit in the region and the value of the power coefficient k. The largest deposit in an area is normally discovered early on in exploration history (Hronsky and Groves, 2008), but the deposit’s resources are commonly not fully delineated until much later in its mine life (Guj et al., 2011). This makes Zipf’s law very sensitive to the accuracy of the resource inventory of the largest deposit. Therefore, Zipf’s law can be used to provide minimum endowment or infer possible existence of a largest deposit, but at any particular time the assessment of residual endowment is implicitly conservative (Guj et al., 2011). Under stable conditions of geologic equilibrium, k is equal to approximately −1 (e.g. Paliwal et al., 1986), so that the 2nd largest deposit is half the size of the 1st, the 3rd deposit is one third the size of the 1st, and so on. This is known as the standard Zipf endowment model, and has been shown to apply to mature search spaces. Higher k factors are likely to represent immature exploration systems, where the largest deposit is not yet discovered or not fully delineated (e.g. Mamuse and Guj, 2011; Lisitsin, 2016). Controversy exists around the application of Zipf’s law, especially regarding the assumption that k = −1 is universally applicable, and the fact that the plot commonly overstates the number of smaller deposits (Guj et al., 2011). However, the standard Zipf model has accurately predicted size and number of orogenic gold deposits yet to be discovered in the Yilgarn Craton using historical data from 1973, 1989, 2003 and 2008 (Fig. 13C: Guj et al., 2011), as well as displaying an almost perfect correlation to the 12 largest nickel deposits of the Kambalda Dome (Mamuse and Guj, 2011). Fig. 5. Detailed lithostratigraphic column of the SSGB (after Murdie et al., 2015). considered in future research are the greater regional attenuation of the AGB, implicating higher strain, the greater surface area of sedimentary units in the AGB, and the overall higher metamorphic grade of the AGB. Fewer sedimentary rocks in the SSGB may decrease the potential for fluid aquicludes, potentially decreasing gold prospectivity, whereas the lower metamorphic grade potentially increases prospectivity (Groves et al., 2000). The effect of relatively lower strain in the SSGB is an unknown factor, but probably affects the precise location of the gold deposits. 4. Quantitative endowment assessment 4.1. Zipf’s law Several quantitative methods have been applied to assess mineral endowment (e.g. Allais, 1957; Griffiths, 1978; Harris, 1984; McCammon and Kork, 1992; Drew, 1997), many of which use the threepart quantitative mineral resource assessment (Lisitsin et al., 2010, 2014 and references therein). Zipf’s law provides an easy to implement assessment approach, and has been shown to be comparable to the 4.1.1. Applying Zipf’s law to the Sandstone Greenstone Belt Zipf’s law is applied as a statistical assessment of known deposits, in order to define the natural and residual gold endowment of the SSGB. Testing various plausible models on whether the largest deposit in the region has been discovered and its endowment adequately evaluated, 334 Ore Geology Reviews 94 (2018) 326–350 R.S. Davies et al. Fig. 6. Regional-scale interpreted bedrock geological map of the SSGB, outlining lithological domains (using GSWA 1:100,000 interpreted bedrock geology map series). Fig. 7. Regional-scale interpreted bedrock geological map of the SSGB, outlining major structures (using GSWA 1:100,000 interpreted bedrock geology map series). 335 Ore Geology Reviews 94 (2018) 326–350 R.S. Davies et al. Table 5 Recognised regional and localised deformation phases for the SSGB (after Standing, 2000). Phase of deformation Stress regime Evidence Timing D1 Regional-scale N-S compression Ca. 3.0 Ga D2 Progressive E-W and SE-NW constriction D3 Progressive (from D2) inhomogeneous E-W shortening D4 Localised strain heterogeneity due to diapiric buildup of small volume granitoid pluton D5 Localised strain heterogeneity due to E-W directed stress field during waning stage of D4 D6 Continued E-W compression during cratonisation at the end of the Archaean Layer-parallel D1 foliation and D1 thrusts in tremolite-chlorite (-talc) schist, along with an east-striking, belt-scale D1 syncline, which have been overprinted by later structures (e.g. northerly-trending F2 Sandstone Syncline) to produce type-2 and type3-fold interface patterns. Initiation of the Sandstone Decollement was most likley a product of D1 thrusting Macroscopic folds with NW-striking axial planes and north-plunging fold axes. Refolded D1 folds in the northern half of the Ultramafic Domain, produced northerly trending gneissic banding, and potentially formed periclinal Sandstone Syncline to portray boxfold geometry. Constriction of the northern Sedimentary Domain may have been a product of D2 deformation Kilometers wide, east-trending, sinistral Edale Shear Zone and the northeast-trending, dextral Youanmi Shear Zone. The initiation of the WNW-striking Dandaraga Fault as a sinistral shear (although the majority of its eveolution is likley to have occurred during subsequent deformation events) Open to closed antiforms and tight to isoclinal synforms along the southern margin of the belt, with NNE-striking subvertical axial planes and shallow north-plunging fold axes. Transcurrent flextural-slip deformation along the Sandstone Decollement in response to differential strain-rates between the Mafic and Ultramafic Domains Major fault corridor along N-S and WNW orientations, many of which are intruded by small-scale felsic intrusions. The WNW-striking Dandarage Fault system would have continued to deform in a sinistral manner and it is inferred that the N-S striking, gently-dipping, laminated reef-style gold deposits and the steeply-dipping, shear zonehosted lode-gold deposits were deposited during this event Fissue-like eruptions of doleritic magma to form the E-W trending dolerite dykes evident in the aeromagnetic data Ca. 2.7 Ga Ca. 2.67 Ga Unknown Unknown ? provided in Table 10. Deposits are ranked according to their gold endowment, measured in ounces of contained gold and calculated from historical production as well as current measured and indicated resources. Resource estimates are ideally measured with a common and consistently low cut-off grade. However, cut-off grades are not everywhere readily available, production can in places be aggregated into provides some measure of uncertainty of assessment results. Here, the hypothesised Zipf models follow the methodology employed by previous researchers, including Lisitsin (2016), and are based on the conceptual and empirical assessments of gold endowment for the SSGB outlined above. Historical production and gold resources of the SSGB deposits are Fig. 8. Regional-scale interpreted bedrock geological map of SSGB, outlining major deposits and prospects (using GSWA 1:100,000 interpreted bedrock geology map series). 336 R.S. Davies et al. Table 6 Key geological characteristics of gold mines in the Sandstone Greenstone Belt. Abbreviations: diss = disseminated; sulf = sulfur (after Davies et al., 2017). Deposit Host Structure Strike Dip Local Structural Complexity Host Rock Mineralisation Alteration Ore Assemblage Lord Henry ENE trending Trafalgar Shear Zone NNW trending Trafalgar Shear Zone NW trending Trafalgar(?) Shear Zone ENE-WSW 30° N White mica- chlorite 50° W Sulfide and gold in shear veins 70°N Actinolite-tremolitechlorite-biotite White mica-carbonate Pyrite-galena-arsenopyritesphalerite-chalcopyrite-gold Pyrite-hematite-gold E-W Granodiorite Ultramafic footwall Granodiorite/basalt mix Ultramafic footwall Talc-chlorite carbonate schist Ultramafic footwall Dolerite hangingwall Sulfide and gold in shear veins NNW-SSE Bull Oak N-S trending 731 Fault Zone NW-SE 30–40° NE NW trending cross-cutting faults Centre of NNE trending syncline W limb of NNW trending anticline NE trending cross-cutting faults and granite/migmatite intrusions SW limb of NW trending syncline Cross-cutting BIF units Sulfide and gold in shear veins White mica-carbonate Pyrite-gold Oroya Black Range N-S trending shear zone? N-S trending Hack’s Creek Fault Zone N-S 30–45° W Granodiorite contained by BIF units Basalt footwall and hangingwall Dolerite White mica-carbonate Pyrite-gold N-S 30–45° W Sulfide and gold in shear veins and vein arrays in host rock Sulfide and gold in shear veins White mica-carbonate Pyrite-gold N-S trending 731 Fault Zone NNE trending Bulchina Shear Zone NNE-SSW 45° W? White mica-fuchsite Pyrite-gold NNE-SSW 30° W Shear veins and disseminated sulfide and gold in host rock Sulfide and gold in shear veins and vein arrays in host rock Goethite-white micafuchsite-carbonate Pyrite-gold N-S trending 722 Fault Zone N-S trending 722 Fault Zone Subvertical stock 30° NW NW-SE 30-60° NE Pyrite-galena-molybdenitechalcopyrite-gold Pyrite-gold NE trending shear zone NE-SW Subvertical Lord Nelson Maninga Marley 337 Hack’s Reef Black Range Lady Hamilton Bulchina Two Mile Hill Shillington NE trending cross-cutting faults and BIF units Intersection of two regional shear zones Centre of NNE trending anticline Cross-cutting BIF units Intersection of N-S, NE and NW trending faults N flank of Nungarra dome N-S trending cross-cutting structures Dolerite contained by ferruginised graphitic shale units Basalt and ultramafic rock Quartz porphyry Ultramafic footwall Dolerite hangingwall Tonalite cross-cutting BIF and basalt BIF flanked by dolerite wallrock Vein arrays and disseminated sulfide and gold in host rock Shear veins and disseminated sulfide and gold in host rock White mica-carbonate Black shale and basalt Shear veins and disseminated sulfide and gold in host rock White mica-carbonate? White mica-chloritecarbonate-magnetite Pyrite-gold Pyrite-gold? Ore Geology Reviews 94 (2018) 326–350 Bellchambers E-W trending cross-cutting faults E-W trending cross-cutting faults and shale units Sulfide and gold in shear veins and vein arrays in host rock Ore Geology Reviews 94 (2018) 326–350 R.S. Davies et al. Table 7 Camp-scale targeting parameters for the SSGB. Critical element Process Targeting element Fertility Geodynamic driver 1st order: trans-lithospheric conduit 2nd order: belt-scale conduits Unknown Undetermined Structure that propagates to base of crust No indication at scale of SSGB No indication at scale of SSGB Deep feature D3 structure Pre-D5 structure Intersects 1st order structureExtensive quartz veining Metal anomalism along structure 3rd order: physical throttle Threshold barriers and breaching zones Chemical traps Rocks that promote gold precipitation from fluids Zones of high fluid permeability Lines along 2nd order structure Fault intersections Competency contractAnticlines Cross-cutting lithologies BIF Ferruginous chert Dolerite Proxies Seismic Structure interpretation Seismic Structure interpretation High gravity anomalies Drilling/geochemistry assays Geological mapping (veins) AeromagneticsStructure interpretation High gravity anomalies delineate anticlines Geological mapping Aeromagnetics Drilling/geochemistry assays Geological mapping are shared mineralising processes acting across a relatively short timeperiod, making it highly unlikely that separate systems are responsible for the known gold endowment of the belt. Systematic over-estimation of gold endowment for the larger deposits of the SSGB is highly unlikely. The estimates are predominantly calculated from historical production and the lodes in many deposits are not fully delineated, remaining open at depth. For most deposits, actual pre-mining gold endowment is almost certainly higher than that indicated. As the deposits of the SSGB have developed by the same mineralising system, and subsequent modifying processes are insignificant, it is safe to assume that the size distribution of the deposits of the SSGB should follow the standard model of Zipf’s law where k is equal to -1. Acceptance of the applicability of the standard Zipf’s endowment model leaves the presence of large undiscovered, or poorly delineated, deposits as the only reasonable explanation for the deviation from the standard model, as commonly accepted in recent applications of Zipf’s law (e.g. Mamuse and Guj, 2011; Yigit, 2012). Although the largest deposits are commonly discovered early (Hronsky and Groves, 2008), their true endowment commonly exceeds initial estimates by a significant factor (e.g. Guj et al., 2011). Previous assessments using Zipf’s law (e.g. Quirk and Ruthrauff, 2006; Lisitsin, 2016) have also concluded that there tends to be an over estimation of the number of smaller deposits. For this reason, the size of the smallest deposits for the following Zipf models has been truncated at 0.03 Moz, below which they are unlikely to be of material economic significance. camps rather than specific deposits, and rarely individual deposits have appeared under different names. Nonetheless, considerable effort was made to ensure that SSGB data were consistent and comparable over time. 4.1.2. Zipf model of known mineralisation The deposits of the SSGB are shown in Fig. 14 as a natural-scale plot of size vs rank. The largest deposit is given the rank of 1, second largest is rank 2, and so on. The bars represent known deposits and plot as a concave slope. Similar to previous studies, the size-rank distribution of known gold deposits for the SSGB can be characterised by a best-fit negative power function, due to the presence of a large number of small deposits and fewer large deposits. However, the observed power coefficient of −1.92 for the best fit power function is significantly lower than the standard k value of −1, typically used to characterise deposit size distribution by Zipf’s law. The best-fit power function does not represent the distribution well, with an R2 = 0.8678. The deposits of the SSGB are graphically presented in Fig. 15 using the log-log approach of Folinsbee (1977). Here, the Zipf’s law power coefficient is equal to -0.506, much lower than the standard Zipf’s model, and R2 = 0.8232. The application of Zipf’s law to the known deposits of the SSGB is based on the following assumptions: a). the deposits of the SSGB have formed as a product of the same mineralising system, representing a single population for analysis; b). Two Mile Hill is the largest deposit in the SSGB; c). the endowment of Two Mile Hill has been adequately estimated; and d). the total gold endowment of the SSGB can be adequately characterised by Zipf’s law with k = −1. This provides an impossible negative value for residual endowment of approx. −0.1 Moz for the SSGB, due to many of the significant deposits being over-endowed in comparison to their rank-corresponding Zipf estimate. It is apparent that one or more of the assumptions made above are incorrect. Possible explanations include: a). the presence of mixed populations within the deposits of the SSGB (i.e. different mineralising systems); b). over-estimation of gold endowment for larger deposits; c). the distribution of natural endowment for the SSGB, which does not conform to the standard Zipf function, is characterised by a significantly steeper Zipf function (suggesting that the two largest deposits of the SSGB contain an additional 1.5 Moz); d). the distribution of natural endowment for the SSGB conforms to the standard Zipf function and these observations represent an immature search space, containing a number of large undiscovered or poorly delineated deposits; and e). a combination of the explanations given above. The presence of mixed populations could indicate multiple mineralising systems, with differing constituent processes. However, the mineral systems analysis, conducted as part of this study, suggests there 4.1.3. Estimate of oxide zone endowment The known deposits of the SSGB lie almost exclusively within the oxide zone. As such, a Zipf plot of all existing deposits, except Three Mile Hill, represent an assessment of residual gold endowment for the oxide zone only. Accordingly, a conservative approach to resolving the suggested under-endowment of the larger deposits of the SSGB is to assume that the largest deposit has been discovered but remains insufficiently delineated, so that the current largest deposits of the SSGB contain significant undiscovered resources. A Zipf’s law model in Fig. 16A, assumes that: a). the largest deposit of the SSGB is approximately 0.9 Moz (over three times that of Oroya); and b). the total gold endowment of the SSGB can be adequately characterised by the standard Zipf’s model (k = −1). Gaps in the plot represent deposits not yet discovered, and discrepancies between plotted and Zipf estimated endowment represent a current over- or under-estimation of deposit endowment. It is worth noting that the under-estimation of the three largest deposits does not guarantee that they will remain in their respective positions as additional resources are defined. For example, it 338 Ore Geology Reviews 94 (2018) 326–350 R.S. Davies et al. Fig. 9. Regional geology of the AGB (modified after Jowitt et al., 2014). workings to extract primary mineralisation from fresh bedrock. This model assumes that: a). the largest deposit of the AGB is approximately 0.6 Moz, in order to provide a good fit for the deposit data; and b). the total gold endowment of the AGB can be adequately characterised by the standard Zipf’s model (k = −1). Given these assumptions, the Zipf model suggests that the three largest existing deposits contain a total of approximately 1.1 Moz, with approximately 0.3 Moz remaining to be delineated. Two additional deposits of 0.12 and 0.15 Moz, and eight deposits of between 0.3 and 0.75 Moz remain to be discovered. These represent a total natural endowment of 2.1 Moz and a total residual endowment of 1.1 Moz Moz. may transpire that Hacks represents the first ranked deposit, containing more oxide mineralisation than Oroya or Lord Nelson. Given these assumptions, the Zipf model suggests that the three largest existing deposits contain a total of approximately 1.65 Moz, with approximately 0.9 Moz remaining to be delineated. Three additional deposits of 0.12–0.18 Moz, eight deposits of 0.5–0.1 Moz and eleven deposits of < 0.05 Moz remain to be discovered. These represent a total natural endowment of 3.6 Moz and a total residual endowment of 2.3 Moz. In comparison with the oxide endowment of the SSGB, the known oxide deposits of the AGB in 1990 are plotted in Fig. 16B. These represent the total known mineralisation (past production, resources and reserves) prior to the commencement of significant underground 339 Ore Geology Reviews 94 (2018) 326–350 R.S. Davies et al. Fig. 10A. Drillholes exceeding 50 m of vertical depth in the SSGB (using GSWA 1:100,000 interpreted bedrock geology map series). Fig. 10B. Drillholes exceeding 100 m of vertical depth in the SSGB (using GSWA 1:100,000 interpreted bedrock geology map series). 340 Ore Geology Reviews 94 (2018) 326–350 R.S. Davies et al. Fig. 10C. Drillholes exceeding 200 m of vertical depth in the SSGB. 2011), it is suggested that the size of the largest deposit inventory remains underestimated and that approximately doubling the size of the largest deposit provides a better fit for the deposit data (Lisitsin, 2016). A Zipf’s law model in Fig. 17A, assumes that: a). the largest deposit of the SSGB is approximately 1.15 Moz (over two times that of Three Mile Hill); and b). total gold endowment of the SSGB can be adequately characterised by the standard Zipf’s model (k = −1). Again, it is important to note that the under-estimation of the four largest deposits does not guarantee that they will remain in their respective positions as additional resources are defined. Given these assumptions, the Zipf model suggests that the four largest existing deposits contain a total of approximately 2.4 Moz, with approximately 1.2 Moz remaining to be delineated. Four additional deposits of 0.1–0.2 Moz, eleven deposits of 0.5–0.1 Moz and fourteen deposits of < 0.05 Moz remain to be discovered. These represent a total natural endowment of 5.1 Moz and residual endowment of 3.3 Moz. Arguably, the previous Zipf model for total endowment represents a conservative estimate, as there is a significant likelihood that the largest fresh rock deposit within the SSGB remains undiscovered. In the assumption that the largest deposit remains undiscovered, the known deposits of the SSGB can be made to fit the Zipf model by: a). including a currently undiscovered rank-1 deposit of 1.4 Moz; and b). still assuming the total gold endowment of the SSGB can be adequately characterised by Zipf’s law with k = −1. This provides a better fit for the known deposit data, although endowment of the four largest known deposits of the SSGB remain under-estimated. The Zipf model (Fig. 17B) suggests that, aside from the undiscovered 1.4 Moz deposit, the four largest deposits contain an additional 0.5 Moz, six undiscovered ore bodies contain between 0.1 and 0.2 Moz, 12 deposits contain between 0.05 and 0.1 Moz and there are 19 deposits of < 0.05 Moz: a total natural endowment of 6.2 Moz and residual endowment of 4.4 Moz. Table 8 Historical production of AGF deposits up to 1996 (after Aoukar and Whelan, 1990; Broome et al., 1998; Inwood, 1998). Deposit Ore treated (t) Recovered gold grade (g/t) Gold produced (oz) Emu/Agnew Open pit Underground 4,957,160 72,086 4.08 2.78 714,765 7069 Redeemer Open pit Underground 2,100,000 4,300,000 3.53 3.90 261,486 605,301 Cox-Crusader Open pit Underground 177,077 640,000 5.57 5.90 23,002 133,194 New Holland Open pit 1,168,889 2.86 117,921 Genesis Open pit 491,903 3.03 47,919 Deliverer Open pit 126,000 4.50 20,000 4.1.4. Estimate of total natural endowment – based on SSGB deposits only Having completed an assessment of undiscovered oxide endowment, the next step is to estimate total natural endowment for the SSGB by including oxide and primary mineralisation. The only addition to this assessment is the inclusion of Three Mile Hill, as the other existing deposits of the SSGB lie almost exclusively within the oxide zone. As with the oxide zone assessment, a conservative approach to resolving the suggested under-endowment of the larger deposits is to assume that the largest deposit has been discovered but remains insufficiently delineated. Bearing in mind that the critical elements for significant mineralisation are present within the SSGB and that the actual endowment of larger deposits commonly exceeds initial estimates (e.g. Guj et al., 341 Ore Geology Reviews 94 (2018) 326–350 R.S. Davies et al. Fig. 11. Histogram of historical production of AGF deposits up to 1996, showing relative ratio of oxide (open pit) vs fresh rock (underground) mineralisation. in the SSGB. This model indicates a total natural endowment of approximately 21.3 Moz and residual endowment of approximately 19.6 Moz. Assuming the oxide-endowment estimate of 4.4 Moz (Fig. 17), the remaining primary-zone mineralisation endowment of the SSGB is equal to 12.1 Moz. This model suggests an oxide to primary-zone mineralisation ratio of 1:2.75, where 2.75 times the total oxide-zone mineralisation remains undiscovered as primary-zone mineralisation in fresh rock. In this model, several of the largest deposits have not yet been delineated, casting an element of doubt regarding the realism of the model. As such, these findings should be considered a first step toward defining prospectivity of an immature search space, as opposed to a definitive value for gold endowment. 4.1.5. Estimate of total natural endowment – by comparisons with AGB An interesting conceptual experiment can be conducted to test the potential natural (including fresh rock) gold endowment of the SSGB. Various assumptions are required for this model. However, they are supported by comparisons made with the AGB, as part of the empirical assessment of gold endowment for the SSGB. From comparison with the AGB, it appears plausible that the total gold endowment of these belts is broadly equivalent. The largest deposit of the AGB is the Agnew/Emu deposit, where historical production and reserve estimates (Sander, 2014) provide a total endowment of 4.76 Moz. Based on this comparison, it is possible to define residual endowment for the SSGB by plotting Zipf’s law using Agnew/Emu as the number 1 deposit. A Zipf’s law model is plotted in Fig. 18 by assuming: a). the total natural endowment of the SSGB and AGB are equivalent, and therefore the largest deposit of the SSGB is approximately 4.76 Moz; b). the total gold endowment of the SSGB can be adequately characterised by Zipf’s law with k = −1; and c). the smallest deposits of economic interest have a cut-off endowment of 0.1 Moz. This model suggests that nine deposits of > 0.5 Moz, seven deposits of 0.25–0.5 Moz, 26 deposits of 0.1–0.25 Moz, and many deposits of < 0.1 Moz remain undiscovered or poorly delineated 4.1.6. Reality checks There is a high likelihood that the SSGB hosts a number of large, undiscovered deposits and that significant mineralisation extends beneath known deposits. Orogenic gold deposits generally have small footprints, with narrow veins following complex controlling structures and alteration of host rock rarely extending further than tens of metres Table 9 Comparison of critical mineral system elements between SSGB and AGF. Scale Critical Element Sandstone Greenstone Belt Agnew Gold Camp Belt Fertility Appears gold-fertile but at immature exploration stage Terminal phase of E-W compressive tectonic event Youanmi Shear Zone represents vertically accretive, translithospheric structure proximal to belt Ida Fault represents vertically accretive, trans-lithospheric structure proximal to belt Camp Favourable Geodynamics Favourable Architecture 1st order – translithospheric conduit Favourable Geodynamics Favourable Architecture 2nd order – belt-scale conduits Preservation Highly gold-fertile with total gold production of over 13 Moz and current resource of over 3.5 Moz Terminal phase of E-W compressive tectonic event Extensional/transpressional low active tectonic strain, during D3 and later (i.e. 2.67 Ga onwards) Major heterogeniety along regional scale shear zones and N-S fault corridors (e.g. 722, 731 Faults and Trafalgar Shear) Extensional/transpressional low active tectonic strain, during D3 and later (i.e. 2.67 Ga onwards) Major heterogeniety along regional scale shear zones and N-S fault corridors (e.g. Emu and Table Hill Shear Zones) Supergene enrichment within oxide zone, primary zone mineralisation preserved to unknown depth below oxide zone Significant structural heterogeniety; high competency contrast host rocks; intersecting shear zones; cross-cutting structures; sheared host rock; and anticlinoria Iron-rich host rocks present favourable chemical substrate for gold precipitation Supergene enrichment within oxide zone, primary zone mineralisation preserved to great depth below oxide zone Significant structural heterogeniety; high competency contrast host rocks; intersecting shear zones; structures cross-cutting lithologies; sheared host rock; and anticlinoria Iron-rich host rocks present favourable chemical substrate for gold precipitation Deposit Favourable Architecture 3rd order – physical throttles Chemical Traps 342 Ore Geology Reviews 94 (2018) 326–350 R.S. Davies et al. Fig. 12. Comparison between AGB and SSGB showing deposit locations and drilling below a vertical depth of 100 m, insets show drilling density within close proximity of major deposits (using GSWA 1:100,000 interpreted bedrock geology map series). only. With much of the belt covered by transported material, shallow drilling commonly fails to reach in situ weathered rocks and very few drill-holes extend into fresh bedrock. By reviewing drilling data from 0 to 50 m for the SSGB (Fig. 10A), it is evident that large sections of the (e.g. Goldfarb et al., 2005). These characteristics mean that orogenic gold deposits represent difficult targets for discovery and delineation during exploration. Many of the previous drilling campaigns in the SSGB have consisted of shallow RAB and AC drilling, sampling for gold 343 Ore Geology Reviews 94 (2018) 326–350 R.S. Davies et al. 5. Discussion and conclusions The results of this study represent a first attempt to systematically describe the geology and gold deposits of the SSGB, and utilise empirical, conceptual and quantitative methods to assess residual gold endowment, and aid exploration ‘re-targeting’ within a partially explored region. From a conceptual endowment analysis of the SSGB using the mineral-systems framework (Table 1), it is suggested that: a) regional fertility and geodynamic drivers for gold distribution are evident, but poorly understood, requiring further research; b) the Youanmi Shear Zone represents a major structural conduit for propagation of auriferous fluids from a deep-seated source; c) close to palaeosurface, fluids propagate into secondary, belt-scale structures, such as N-S trending shear zones and fault corridors; d) fluids encounter localised structurally complex zones, with potential chemically favourable host rock; and e) concentrated gold precipitation leads to the formation of economic deposits. In general, the gold deposits of the SSGB can be interpreted to belong to the orogenic gold deposit class of Groves et al. (1998) or the orogenic gold system of Hagemann and Cassidy (2000). Gold mineralisation is characterised by: a) a spatial relationship to regional translithospheric structures; b) structural control of mineralisation at deposit scale; c) low-salinity aqueous-carbonic fluid compositions; d) similar alteration assemblages which vary with host rock; and e) formation at post-peak metamorphic conditions. A comparison of geological and mineral-system characteristics with the mature exploration search-space of the AGB, provides an empirical assessment of endowment for the immature SSGB. The respective belts display similar geometry, geology, belt-scale structures, deposit styles, a shared tectonic history from 2.72 Ga, and similar timing of gold mineralisation, between 2.66 and 2.63 Ga. The mineralising systems that led to their endowment also share several important characteristics (Table 9), such as: a) proximal trans-lithospheric structures; b) secondorder N-S trending belt-scale structures, along zones of high competency contrast, controlling the distribution of mineralisation; c) similar third-order physical throttles, including antiforms, intersecting shear zones and cross-cutting fault structures; and d) iron-rich host rocks acting as chemical traps. Based on this comparison, it is proposed that the total natural gold endowment of the SSGB may be roughly equivalent to that of the more comprehensively explored AGB. It is likely that geological difference between the belts in terms of degree of strain (probably a neutral factor), abundance of sedimentary rocks (more positive for the AGB), and metamorphic grade (more positive for the SSGB) cancel each other out in terms of potential endowment. A power law distribution normally defines the frequency distributions of mineral deposit sizes, despite limitations in data quality and geological understanding of underlying mineral system processes. Zipf’s law is applied as a quantitative assessment to define the natural and residual gold endowment of the SSGB. The total natural gold endowment within the oxide zone is estimated to be 3.6 Moz (Fig. 16A), and is comparable to that of the AGB in 1990 prior to the commencement of significant underground mining of primary fresh rock mineralisation. Residual oxide mineralisation is proposed to be contained in extensions of known deposits and several undiscovered deposits, ranging in size up to 0.9 Moz. The bedrock of the SSGB remains poorly explored. Based on a Zipfian distribution of existing SSGB deposits, the combined oxide and primary mineral endowment of the SSGB is estimated to be between 5.1 and 6.4 Moz, depending on whether the largest deposit has already been discovered. In comparison with the modern-day AGB, and, assuming a natural oxide-zone endowment of 3.6 Moz, the total primaryzone mineralisation of the SSGB is estimated to be 17.7 Moz. This consists of nine undiscovered deposits of > 0.5 Moz (Fig. 18). In this model, the oxide to primary zone mineralisation ratio is approximately 1:5. With 1.2 Moz of gold produced from the SSGB, and a proposed natural endowment of 21.3 Moz, the ratio between these is estimated to Fig. 13A. Size-distribution of gold camps in the Yilgarn (in Moz Au production, resources and reserves, 2008), adapted from Guj et al. (2011). Fig. 13B. Size-distribution of gold deposits at St Ives, WA (in koz Au production and reserves, 1999), adapted from Hronsky and Groves (2008). belt beneath shallow cover remain untested. This accepts the prediction that deposits of the estimated sizes in the oxide zone Zipf models (Fig. 18 and Fig. 19, respectively) could occur near the surface. Drilling data below 100 and 200 m (Fig. 10B and C, respectively) show that most of the SSGB has been inadequately tested and significant room exists to accommodate the number and sizes of primary-zone deposits predicted by the natural endowment Zipf’s model. Figs. 19 and 20 plot cumulative probability for residual endowment estimates for fresh rock endowment models, where the largest deposit is assumed to have not yet been discovered, by plotting known deposits and then comparison with the AGB respectively. The lines plot between the 1st and 99th percentile of the deposit size distribution for residual endowment. These lines indicate the probability of a discovery smaller than a given size. Table 11 summarises the key information from these plots. According to Rose (2000), the ratio of the 10th and 90th percentile should be between 10 and 120. In the case of the SSGB, the ratio of P10/90 for the oxide zone model is equal to 23.5, and for the primary zone model it is equal to 65.2. It is also stated that the Swanson mean size should fall between the 17th and 33rd percentile. The Swanson mean lies within the 33rd percentile for the oxide zone model, and the 30th percentile for natural endowment model. Therefore, the estimated residual endowment for each Zipf model passes both geographical and statistical reality checks. 344 Ore Geology Reviews 94 (2018) 326–350 R.S. Davies et al. Fig. 13C. Evolution of Zipf’s law for three deposits in the Yilgarn Craton from 1989 to 2003, adapted from Guj et al. (2011). 1. The Golden Mile represents the largest deposit and increases in size between these dates as additional resources are delineated. 2. Resources at Boddington increase dramatically post-discovery, moving it from the 11th to 2nd in the size-rank distribution. 3. No further resource was delineated at Laverton post-1989, and, as such, it drops from 20th to 29th as additional deposits are discovered or increase in size. be ∼1:18. Table 10 Deposits of the SSGB, showing historical production and resources (measured in ounces). Data pre-1954 was calculated from historical records (Appendix 1) and post-1954 was sourced from Troy Resources internal reports (Lowe and Ross, 2007; Maddocks et al., 2009). Deposit Historical Production Resources (Inferred) Resources (Indicated) Total Two Mile Hill Oroya Lord Nelson Hacks/Black Range Bulchina Lord Henry Bulloak/ Hancock’s Havilah Tigermoth Shillington Maninga Marley Twin Shafts Goat Farm Vanguard Bellchambers Plum Pudding Ladybird Wiraminna Sandstone North 55 North Piper Wanderie Eureka Bulletin Total 20,000 253,000 207,000 236,000 452,000 10,000 5000 0 18,000 0 37,000 0 490,000 263,000 249,000 236,000 230,000 57,000 80,000 0 2000 18,000 0 58,000 0 230,000 117,000 98,000 34,000 0 23,000 13,000 20,000 19,000 0 1500 10,000 0 0 0 4000 0 3000 2000 1000 1,213,500 2000 31,000 0 8000 0 0 16,000 14,000 2500 3000 10,000 5000 0 0 0 0 0 578,500 15,000 0 6000 0 0 0 0 0 0 9000 0 0 0 4000 0 0 0 147,000 51,000 31,000 29,000 21,000 20,000 19,000 16,000 15,500 12,500 12,000 10,000 5000 4000 4000 3000 2000 1000 1,939,000 5.1. Cost-benefit analysis of additional data collection The ratio of natural to residual endowment defines the maturity of a search space and the probability of making an economic discovery, an important step in justifying investment in exploration efforts. Several models have been proposed, providing estimates for residual endowment of the immature oxide zone and the virtually unexplored fresh rock of the SSGB. Further data collection, leading to an increased geological understanding of the mineralising systems, is required to improve the accuracy of these estimates. Two main factors affecting gold mineralisation in the SSGB are poorly understood: a) variations in fertility on a continental scale, influenced by characteristics of gold and fluid sources, and whether the source for the SSGB and AGB is shared or separate; and b) the degree to which mineralisation extends at depth, influenced by structural and chemical factors controlling gold precipitation, as well as post-depositional preservation of ore bodies. Considering the economic cost-benefit of collecting data to resolve areas of limited geological understanding, it is advised that exploration companies, operating in areas such as the SSGB, focus resources on acquiring comprehensive datasets for the proxies outlined in Table 7. These will aid exploration by providing a better understanding of localised factors affecting camp- to deposit-scale mineralisation and represent mappable targeting proxies for GIS prospectivity analyses. Understanding variations in regional fertility requires the acquisition of datasets covering areas vastly greater than that typically available during the exploration of a single greenstone belt. These datasets are normally beyond the budget of lower- to mid-tier exploration companies. As such, regional fertility should remain the focus of governmental and academic research projects, with collaboration or support from industry sponsors. 345 Ore Geology Reviews 94 (2018) 326–350 R.S. Davies et al. Fig. 14. Deposit size-distribution plot showing best fit power function and Zipf endowment model assuming Two Mile Hill is the largest deposit in the belt and contains approx. 0.5 Moz of gold. framework with which to conduct exploration in the region. The quantitative endowment assessment provides evidence of potential undiscovered gold deposits. The Zipf curve has no spatial connotations and, as such, does not identify the location of undiscovered deposits. The next stage for assessment of the SSGB is to link the estimated residual endowment of the belt to a prospectivity model. Previous exploration has utilised 2D models which, strictly speaking, only provide surface prospectivity estimates. In order to conduct exploration of bedrock, additional data collection and interpretation should include the integration of geological, geophysical and petrophysical datasets to map the subsurface, developing a detailed model for exploration targeting in the SSGB. Integration of an endowment analysis and mineral prospectivity model provides a ranking of the geological merit of targets and the attribution of an order of magnitude, risk-adjusted dollar value to prospective domains, representing an integration of both geological and expected financial value of the belt. Continued application of Zipf’s law to mature exploration search spaces increases confidence in the assumption that the size distribution of mineral deposits fits the standard Zipf model (i.e. k = −1). Mineral systems have been recognised to act in a self-organising manner, where various processes operate and interact at different scales (e.g. Hronsky, Fig. 15. The Zipf curve analysis of mineral deposit sizes can also be graphically presented using the log-log approach, where the x-axis displays the logarithm of the rank of deposits, while the y-axis displays the logarithm of their respective sizes. 5.2. Future research An understanding of the gold-mineralising system of the SSGB provides direct implications for future gold-deposit targeting and a Fig. 16A. Deposit size-distribution plot for oxide zone endowment of SSGB, showing hypothetical Zipf endowment model with a cut-off of 0.03 Moz and assuming the largest deposit has been discovered but not fully delineated; estimated to contain approx. 0.9 Moz of gold. 346 Ore Geology Reviews 94 (2018) 326–350 R.S. Davies et al. Fig. 16B. Deposit size-distribution plot for oxide zone endowment of AGB in 1990, showing hypothetical Zipf endowment model with a cut-off of 0.03 Moz and assuming the largest deposit has been discovered but not fully delineated; estimated to contain approx. 0.6 Moz of gold. Fig. 17A. Deposit size-distribution plot for total natural endowment of SSGB, showing hypothetical Zipf endowment model with a cut off of 0.03 Moz and assuming the largest deposit has been discovered but not fully delineated; estimated to contain approx. 1.15 Moz of gold. Fig. 17B. Deposit size-distribution plot for total natural endowment of SSGB, showing hypothetical Zipf endowment model with a cut-off of 0.03 Moz and assuming the largest deposit remains undiscovered; estimated tocontain approx. 1.4 Moz of gold. 347 Ore Geology Reviews 94 (2018) 326–350 R.S. Davies et al. Fig. 18. Deposit size-distribution plot for total natural endowment of SSGB, showing hypothetical Zipf endowment model with a cut-off of 0.1 Moz and assuming the largest undiscovered deposit is equivalent to current AGB; containing 4.76 Moz of gold. Fig. 19. Cumulative probability curve of residual orogenic gold endowment for oxide zone endowment (liberal), showing P10, 50, 90 and Swanson’s Mean. assumption of k = −1 represents an over-simplification. The varied geological backgrounds and scales at which these systems operate may result in a greater diversity of log-normal type statistical distributions. A study of the ratio between the gold endowment of oxide- and primary 2011). Due to the fractal nature of these systems, power-law type sizedistributions are recognised at various scales from craton through to individual geological domains (e.g. Figs. 13A and 13B). However, with additional studies into endowment, it may become apparent that the Fig. 20. Cumulative probability curve of residual orogenic gold endowment for total natural endowment, showing P10, 50, 90 and Swanson’s Mean. 348 Ore Geology Reviews 94 (2018) 326–350 R.S. Davies et al. Table 11 Probabilistic assessment of residual orogenic gold endowment in the SSGB. Scenario Scenario 2 Gold (Moz) Scenario 3 Gold (Moz) Cumulative Probability (%) Swanson’s Mean Probability of mean or more (%) P10 P50 P90 0.990 0.205 0.042 0.390 33 3.000 0.390 0.046 1.050 30 Chen, S.F., Morris, P.A., Pirajno, F., 2005. Occurrence of komatiites in the Sandstone greenstone belt, north-central Yilgarn Craton. Aust. J. Earth Sci. 52 (6), 959–963. Chen, S. F., Morris, P. A., Pirajno, F., 2006. Komatiites in the Sandstone greenstone belt, north-central Yilgarn Craton, in: GSWA 2006 extended abstracts: promoting the prospectivity of Western Australia: Geological Survey of Western Australia Record 2006/3, pp. 20–21. Colvine, A.C., 1989. An empirical model for the formation of Archean gold deposits. Products of final cratonization of the Superior Province, Canada. In: In: Keays, R.R., Ramsay, W.R.H., Groves, D.I. (Eds.), The Geology of Gold Deposits: The Perspective in 1988, vol. 6. Economic Geology Monograph, pp. 37–53. Czarnota, K., Blewett, R.S., Goscombe, B., 2010. Predictive mineral discovery in the eastern Yilgarn Craton, Western Australia: an example of district scale targeting of an orogenic gold mineral system. Precambrian Res. 183, 356–377. Davies, R.S., Ryan, D., Groves, D.I., Trench, A., Sykes, J.P., Standing, J.G., Jia, C., Robertson, W., 2017. Sandstone Goldfield. In: Phillips, G.N. (Ed.), Australian Ore Deposits. The Australasian Institute of Mining and Metallurgy, Melbourne. Drew, L.J., 1997. Undiscovered Petroleum and Mineral Resources: Assessment and Controversy. Plenum Press, New York. Fallon, M., Porwal, A., Guj, P., 2010. Prospectivity analysis of the Plutonic Marymia Greenstone Belt, Western Australia. Ore Geol. Rev. 38, 208–218. Fletcher, I.R., Mikucki, J.A., McNaughton, N.J., Mikucki, E.J., Groves, D.I., 1998. The age of felsic magmatism and lode-gold mineralisation events in the Lawlers area, Yilgarn Craton, Western Australia. Geol. Soc. Aust. 49, 146. Fletcher, I.R., Dunphy, J., Cassidy, K., Champion, D.C., 2001. Compilation of SHRIMP UPb geochronological data, Yilgarn Craton, Western Australia, 2000–2001. Geoscience Australia. Record 47, 111. Folinsbee, R.E., 1977. World’s view-from Alph to Zipf. Bull. Geol. Soc. Am. 88, 897–907. Goldfarb, R.J., Groves, D.I., 2015. Orogenic gold: common or evolving fluid and metal sources through time. Lithos 233, 2–26. Goldfarb, R.J., Baker, T., Dubé, D., Groves, D.I., Hart, C.J.R., Gosselin, P., 2005. Distribution, character, and genesis of gold deposits in metamorphic terranes. Econ. Geol. 100th Anniversary Volume, 407–450. Griffiths, J.C., 1978. Mineral resource assessment using the unit regional value concept. Math. Geol. 10, 441–472. Groves, D.I., 1993. The crustal continuum model for late-Archaean lode-gold deposits of the Yilgarn Block, Western Australia. Mineral. Deposita 28, 366–374. Groves, D.I., Santosh, M., 2016. The giant Jiaodong gold province: the key to a unified model for orogenic gold deposits? Geosci. Frontiers 7 (3), 409–417. Groves, D.I., Goldfarb, R.J., Gebre-Mariam, M., Hagemann, S.G., Robert, F., 1998. Orogenic gold deposits: a proposed classification in the context of their crustal distribution and relationship to other gold deposit types. Ore Geol. Rev. 13 (1–5), 7–27. Groves, D.I., Goldfarb, R.J., Knox-Robinson, C.M., Ojala, J., Gardoll, S., Yun, G., Holyland, P., 2000. Late-kinematic timing of orogenic gold deposits and significance for computer-based exploration techniques with emphasis on the Yilgarn Block, Western Australia. Ore Geol. Rev. 17 (1), 1–38. Groves, D.I., Goldfarb, R.J., Santosh, M., 2016. The conjunction of factors that lead to formation of giant gold provinces and deposits in non-arc settings. Geosci. Frontiers 7 (3), 303–314. Guj, P., Fallon, M., McCuaig, T.C., Fagan, R., 2011. A timeseries audit of Zipf's law as a measure of terrane endowment and maturity in mineral exploration. Econ. Geol. 106, 241–259. Hagemann, S.G., Cassidy, K.F., 2000. Archean orogenic lode gold deposits. In: Hagemann, S.G., Brown, P.E. (Eds.), Gold in 2000. Society of Economic Geologists Reviews in Econ. Geol. 9–68. Hagemann, S.G., Lisitsin, V.A., Huston, D.L., 2016. Mineral system analysis: Quo Vadis. Ore Geol. Rev. 76, 504–522. Harris, D.P., 1984. Mineral Resources Appraisal—Mineral Endowment, Resources and Potential Supply: Concepts, Methods, and Cases. Oxford University Press, Oxford. Hobbs, B.E., Ord, A., 2015. Structural Geology. The Mechanics of Deforming Metamorphic Rocks, first ed. Elsevier, Amsterdam. Hronsky, J.M.A., 2011. Self-organized critical systems and ore formation: The key to spatial targeting? Society of Economic Geology Newsletter 84, 14–16. Hronsky, J.M.A., Groves, D.I., 2008. Science of targeting: definition, strategies, targeting and performance measurement. Aust. J. Earth Sci. 55 (1), 3–12. Hronsky, J.M.A., Groves, D.I., Loucks, R.R., Begg, G.C., 2012. A unified model for gold mineralisation in accretionary orogens and implications for regional-scale exploration targeting methods. Mineral. Deposita 47 (4), 339–358. Huston, D.L., Mernagh, T.P., Hagemann, S.G., Doublier, M.P., Fiorentini, M., Champion, D.C., Jaques, A.L., Czarnota, K., Cayley, R., Bastrakov, R.S., 2016. Tectono-metallogenic systems – The place of mineral systems within tectonic evolution, with an emphasis on Australian examples. Ore Geol. Rev. 76, 168–210. Inwood, N.A., 1998. New Holland, New Holland South and Genesis gold deposits, Lawlers. In: Hughes, F. (Ed.), Geology of Australian & Papua New Guinean Mineral Deposits. The Australian Institute of Mining and Metallurgy, Melbourne. Ivanic, T.J., Wingate, M.T.D., Korsch, R.J., Blewett, R.S., Jones, L.E.A., Wyche, S., Zibra, I., Doublier, M.P., Romano, S..S., Pawley, M.J., Van Kranendonk, M.J., Gessner, K., Hall, C.E., Chen, S.F., Patison, N., Costelloe, R.D., 2014. Preliminary interpretation of the Youanmi deep seismic reflection lines for Proterozoic intrusive rocks, in: Youanmi and Southern Carnarvon seismic and magnetotelluric (MT) workshop 2013 compiled by Wyche, S., Ivanic, T.J., Zibra, I. Geological Survey of Western Australia, Record 2013/6, pp. 81–85. Jowitt, S.M., Cooper, K., Squire, R.J., Thebaud, N., Fisher, L.A., Cas, R.A.F., Pegg, I., 2014. Geology, mineralogy, and geochemistry of magnetite-associated Au mineralization of the ultramafic-basalt greenstone hosted Crusader Complex, Agnew Gold Camp, Eastern Yilgarn Craton, Western Australia; a late Archean intrusion-related Au deposit? Ore Geol. Rev. 56, 53–72. zones across the Yilgarn Craton would provide an additional reality check for the proposed Zipf models. Funding This work was supported by Alto Metals Limited and the Centre for Exploration Targeting, University of Western Australia. Acknowledgements The sponsorship of this project by Alto Metals Ltd. is gratefully acknowledged, particularly the encouragement and support provided by Dermot Ryan. Thanks are also due to Michael Dentith for his feedback and supervision. The project benefited from constructive comments and open discussion by colleagues Changshun Jia, Bill Robertson, Oliver Robertson, Keith Goode, Caroline Johnson, Mark Munro and Mike Kammermann. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.oregeorev.2018.01.020. References Ahmat, A.L., 1986. Metamorphic patterns in the greenstone belts of the Southern Cross Province, Western Australia. In Professional papers for 1984: Geological Survey of Western Australia. Report 19, 1–21. Allais, M., 1957. Method of appraising economic prospects of mining exploration over large territories—Algerian Sahara case study. Manage. Sci. 3, 285–347. Aoukar, N., Whelan, P., 1990. EMU gold deposit, Agnew. In: Hughes, F.E. (Ed.), Geology of the Mineral Deposits of Australian & Papua New Guinea. The Australian Institute of Mining and Metallurgy, Melbourne. Bak, P., 1996. How Nature Works: The Science of Self-Organized Criticality. SpringerVerlag, Copernicus, New York, pp. 212. Bateman, R., Hagemann, S., 2004. Gold mineralization throughout about 45 Ma of Archaean orogenesis: protracted flux of gold in the Golden Mile, Yilgarn craton, Western Australia. Mineral. Deposita. 39, 536–559. Bierlein, F.P., Groves, D.I., Cawood, P.A., 2009. Metallogeny of accretionary orogens—the connection between lithospheric processes and metal endowment. Ore Geol. Rev. 36, 282–292. Blewett, R., Czarnota, K., 2007. The Y1–P763 Project final report November 2005. Module 3-terrane structure: tectonostratigraphic architecture and uplift history of the Eastern Yilgarn Craton. Geoscience Australia record 2007/15. Broome, J., Journeaux, T., Simpson, C., Dodunski, N., Hosken, J., DeVitry, C., Pilapil, L., 1998. Agnew gold deposits. In: Berkman, D.A., Mackenzie, D.H. (Eds.), Geology of Australian & Papua New Guinean Mineral Deposits. The Australian Institute of Mining and Metallurgy, Melbourne. Campbell, I.H., Hill, R.I., 1988. A two-stage model for the formation of the granitegreenstone terrains of the Kalgoorlie-Norseman area, Western Australia. Earth Planet. Sci. Lett. 90, 11–25. Cassidy, K.F., 2006. Geological Evolution of the Eastern Yilgarn Craton (EYC) and Terrane, Domain and Fault System Nomenclature. Geoscience Australia Record 2006/5. Chen, S.F., 2003. Atley, WA Sheet 2741: Geological Survey of Western Australia, 1:100 000 Geological Series. Chen, S.F., 2005. Geology of the Atley, Rays Rocks, and southern Sandstone 1:100 000 sheets: Geological Survey of Western Australia, 1:100 000 Geological Series Explanatory Notes. p. 42. Chen, S.F., Libby, J.W., Wyche, S., Riganti, A., 2004. Kinematic nature and origin of regional-scale ductile shear zones in the central Yilgarn Craton, Western Australia. Tectonophysics 394, 139–153. 349 Ore Geology Reviews 94 (2018) 326–350 R.S. Davies et al. Schuenemeyer, J.H., Drew, L.J., 1983. A procedure to estimate the parent population of the size of oil and gas fields as revealed by a study of economic truncation. Math. Geol. 15, 145–161. Squire, R.J., Allen, C.M., Cas, R.A.F., Campbell, I.H., Blewett, R.S., Nemchin, A.A., 2010. Two cycles of voluminous pyroclastic volcanism and sedimentation related to episodic granite emplacement during the late Archean: Eastern Yilgarn Craton, Western Australia. Precambrian Res. 183, 251–274. Standing, J.G., 2000. The Geological and Tectonic Setting of the Sandstone Greenstone Belt: Implications for Exploration Targeting. Troy Resources. Thébaud, N., Miller, J.M., McCuaig, T.C., Hilliard, P., Pegg, I., Fisher, L.A., 2012. Structural and mineralization evolution of the Agnew Camp. In: Venrcombe, J. (Ed.), Extended Abstracts of Structural Geology and Resources Symposium. The Australian Institute of Geoscientists, Kalgoorlie, WA. Van Kranendonk, M.J., Ivanic, T.J., Wingate, M.T.D., Kirkland, C.L., Wyche, S., 2013. Long-lived, autochthonous development of the Archean Murchison Domain, and implications for Yilgarn Craton tectonics. Precambrian Res. 229, 49–92. Vielreicher, N., Groves, D.I., McNaughton, N.J., Fletcher, I., 2015. The timing of gold mineralization across the eastern Yilgarn craton using U-Pb geochronology of hydrothermal phosphate minerals. Mineral. Deposita 50, 391–428. Voute, F., Thébaud, N., 2015. Structural, mineralogical and geochemical constraints on the atypical komatiite-hosted Turret deposit in the Agnew–Mt. White district, Western Australia. Mineral. Deposita 50, 697–716. Witt, W., Hagemann, S., Villanes, C., Vennemann, T., Zwingmann, H., Laukamp, C., Spangenberg, J.E., 2016. Multiple gold mineralizing styles in the Northern Pataz District, Peru. Econ. Geol. 111, 355–394. Wyborn, L.A.I., Heinrich, C.A., Jaques, A.L., 1994. Australian Proterozoic Mineral Systems: Essential Ingredients and Mappable Criteria. Australian Institute of Mining and Metallurgy Annual Conference, Melbourne pp. 109–115. Wyche, S., Nelson, C.S., Riganti, A., 2004. 4350–3130 Ma detrital zircons in the Southern Cross granite-greenstone terrane, Western Australia: implications for the early evolution of the Yilgarn Craton. Aust. J. Earth Sci. 51, 31–45. Wyman, D.A., O’Neill, C.O., Ayer, J.A., 2008. Evidence for modern-style subduction to 3.1Ga: a plateau-adakite-gold (diamond) association. Geol. Soc. Am. Spl. Publ. 440, 129–148. Wyman, D.A., Cassidy, K.F., Hollings, P., 2016. Orogenic gold and the mineral systems approach: resolving fact, fiction and fantasy. Ore Geol. Rev. 78, 322–335. Yigit, O., 2012. Discovered and undiscovered gold endowment of Turkey: A quantitative mineral resource assessment using GIS and rank statistical analysis. Mineral. Deposita 47, 521–534. Yun, Y.G., 2000. Controls on orogenic (mesothermal) gold deposits: a craton to provincescale study within a GIS environment. PhD thesis, University of Western Australia, Perth (unpubl.). Zibra, I., Gessner, K., Pawley, M.J., Wyche, S., Chen, S.F., Korsch, R.J., Blewett, R.S., Jones, T., Milligan, P., Jones, L.E.A., Doublier, M.P., Hall, C.E., Romano, S.S., Ivanic, T.J., Patison, N., Kennett, B.L.N., Van Kranendonk, M.J., 2014. Preliminary interpretation of deep seismic line 10GA-YU2: Youanmi Terrane and western Kalgoorlie Terrane, in: Younami and Southern Carnarvon seismic and magnetotelluric (MT) workshop 2013 compiled by Wyche, S., Ivanic, T.J., Zibra, I. Geological Survey of Western Australia, Record 2013/6, 87–96. Zipf, G.K., 1949. Human Behaviour and the Principle of Least Effort: Reading. AddisonWesley, Massachusetts. Krapez, B., Barley, M.E., 2008. Late Archaean synorogenic basins of the Eastern Goldfields Superterrane, Yilgarn Craton, Western Australia. Part III. Signatures of tectonic escape in an arc-continent collision zone. Precambrian Res. 161, 183–199. Lisitsin, V.A., 2016. Rank-Size Statistical assessments of undiscovered gold endowment in the Bendigo and Stawell Zones (Victoria) and the Mossman Orogen (Queensland), Australia: Comparison with three-part assessment results. Natural Resources Res. 25, 269–282. Lisitsin, V.A., Moore, D.H., Olshina, A., Willman, C.E., 2010. Undiscovered orogenic gold endowment in Northern Victoria, Australia. Ore Geol. Rev. 38, 251–269. Lisitsin, V.A., Dhnaram, C., Donchak, P., Greenwood, M., 2014. Mossman orogenic gold province in north Queensland, Australia: regional metallogenic controls and undiscovered gold endowment. Mineral. Deposita 49, 313–333. Lowe, K., Ross, A. F., 2007. Troy Resources NL: Sandstone Project, Mid-West Region Western Australia. Perth. Maddocks, R., Otterman, D., Doyle, P., 2009. Troy Resources NL: Sandstone Project, MidWest Region Western Australia. Perth. Magoon, L.B., Dow, W.G., 1994. The petroleum system. AAPG Memoirs 60, 3–24. Mamuse, A., Guj, P., 2011. Rank statistical analysis of nickel sulphide resources of the Norseman-Wiluna Greenstone Belt, Western Australia. Mineral. Deposita 46, 305–318. McCammon, R.B., Kork, J.O., 1992. One-level prediction – a numerical method for estimating undiscovered metal endowment. Nonrenew. Resour. 2, 139–147. McCluskey, J.B., 1996. The Geology and Geochemical Discrimination of Mafic Rocks of the Southern Portion of the Agnew-Wiluna Greenstone Belt. University of New England, Armidale. McCuaig, T.C., Hronsky, J.M.A., 2014. The mineral system concept: the key to exploration targeting. SEG 2014: Building Exploration Capability for the 21st Century, pp. 153–175. McCuaig, T.C., Beresford, S., Hronsky, J., 2010. Translating the mineral systems approach into an effective exploration targeting system. Ore Geol. Rev. 38, 128–138. Megill, R.E., 1988. Exploration Economics. PennWell, Tulsa p. 238. Merriam, D.F., Drew, L.J., Schuenemeyer, J.H., 2004. Zip's law: a viable geological paradigm? Nat. Resources Res. 13, 265–271. Murdie, R.E., Gessener, K., Chen, S.F., 2015. The sandstone greenstone belt, northern central Yilgarn Craton: 3D modelling using geological and geophysical constraints. Geological Survey of Western Australia. Record 2015/11, 33. Paliwal, H.V., Bhatnagar, S.N., Haldar, S.K., 1986. Lead-zinc resource prediction in India: An application of Zipf's Law. Math. Geol. 18, 539–549. Pareto, V., 1927. Manuel d'economie Politique, 2nd ed. Paris. Perriam, R.P.A., 1996. The geology and mineralization of the Agnew-Wiluna Greenstone Belt. Agnew Gold Operation, p. 39. Platt, J.P., Allchurch, P., Rutland, R., 1978. Archaean tectonics in the Agnew supracrustal belt, Western Australia. Precambrian Res. 3–30. Quirk, D.G., Ruthrauff, R., 2006. Analysis of reserves discovered in petroleum exploration. J. Petroleum Geol. 29, 125–146. Robert, F., Poulsen, K.H., Cassidy, K.F., Hodgson, C.J., 2005. Gold metallogeny of the Superior and Yilgarn craton. Econ. Geol. 100th Anniversary Volume, 1001–1033. Rose, P.R., 2000. Risk analysis and management of petroleum exploration ventures. American Association of Petroleum Geologists, AAPG Methods in Exploration. Series 12, 164. Sander, J., 2014. Gold Fields Australia Site Visit: Agnew/Lawlers Gold Mine. Gold Fields. 350