Origin, geochemistry, stratigraphic and structural setting of the Archean Invincible gold deposit, St Ives gold camp, Yilgarn Craton, Western Australia

The Invincible orogenic gold deposit is located in the world-class St Ives gold field, Yilgarn Craton, Western Australia. The St Ives gold field has produced over 400 t of gold since 1980. The ~1.5 million-ounce Invincible deposit is hosted on the western limb of the Kambalda Anticline, in metamorphosed and hydrothermally altered mudstone and sandstone-conglomerate units of the Upper Black Flag Group. Rare ore veins also occur in the overlying Lower Merougil Group. Prior to the discovery of the Invincible deposit, these young volcano-sedimentary successions were considered less prospective for gold than the mafic-ultramafic stratigraphy, including dolerite sills that intrude near the base of the Lower Black Flag Group. The Black Flag and Merougil groups comprise almost 60% of the Archean basement at St Ives gold field and host newly discovered gold deposits but their internal stratigraphy is relatively poorly understood, hindering exploration at a camp-scale. This thesis documents the lithofacies, provenance and zircon ages of the Black Flag and Merougil groups and uses the results to propose new subdivisions and detailed maps of their distribution. The Lower Black Flag Group is divided into the lower and upper Newtown Felsic formations that comprise: a ~1200 m sequence of thick coarse-grained turbidites and debris flow deposits that were fed by intense subaerial explosive volcanism and intrabasinal dacitic lava domes; and an overlying ~1300 m sequence of juvenile feldspar crystal and lithic-rich turbidites containing locally derived basaltic clasts from the older mafic stratigraphy in addition to dacitic and rhyolitic clasts, respectively. The upper and lower Newtown Felsic formations contained a strong single population of detrital zircon ages with a weighted mean age U-Pb of 2685 ¬± 2 Ma. The overlying Upper Black Flag Group is divided into the Speedway Andesite formation that overlies the upper Newtown Felsic formation across an angular unconformity and the overlying Morgans Island formation. The ~700 m thick Speedway Andesite formation is composed of intrabasinal, submarine andesite lava and lava dome complexes with variably resedimented andesitic hyaloclastite mass-flow breccia intercalated with turbidites consisting of subaerially derived pyroclastic debris. Coherent andesite contained an igneous mean zircon U-Pb age of 2674 ¬± 8 Ma. The ~150 m thick Morgans Island formation contains thick volcanicquartz rich turbidites and debris flow conglomerates and has a mean detrital zircon age of 2670 ¬± 6 Ma. The Morgans Island formation contains a small population of zircon ages similar to the Lower Merougil Group suggesting they were sourced from similarly aged explosive felsic volcanic events. The Lower Merougil Group variably unconformably overlies the Upper Black Flag Group and is divided into the basal Zeta Island and the overlying Triangle Island formations. Both formations have a strong single population of detrital zircon ages with a weighted mean age of 2666 ¬± 4 Ma. The ~150 m thick Zeta Island formation contains extensive sheet-like matrix supported conglomerates and massive volcanic quartz dominated sandstone; whereas the Triangle Island formation is composed of abundant trough cross-bedded sandstone and restricted lenses of conglomerate. This thesis agrees with previous research suggesting the Black Flag Group was deposited into a deep water and represent submarine fan sequences. However, we contend that the variably unconformable contact between the Upper Black Flag and Lower Merougil groups, that does not have a significant or currently distinguishable time gap suggests that the change from deposition in a subaqueous to a subaerial environment was more transitional than previously interpreted. The turbidite sequences of the Morgans Island Formation overlain by the conglomerate sequences of Zeta Island formation are interpreted to represent a fan-delta prograding into water, that is in turn overlain by subaerial alluvial fan ‚- braid-plain successions represented by the Triangle Island formation. The Morgans Island Formation is further divided locally into the Invincible Mudstone and Sandstone members that host the Invincible deposit, referred to as 'mudstone' and 'footwall sandstone' respectively. Three distinct stages of predominantly extension veins and breccias that trend with the mudstone northwest, are recognised at Invincible. These veins reflect at least two gold mineralisation events separated by a period of significant deformation. The widespread extension veins are consistent with formation during protracted northeast-southwest compression. Stage 1 veins are variably steeply dipping parallel extension veins that develop adjacent to bedding-parallel shears in the mudstone. These veins are associated with low-grade gold mineralisation and were folded prior to being cut by stage 2 veins. Stage 2 veins are steeply dipping and commonly boudinaged thick extension and breccia veins that cut folded bedding and S1 shear fabric. These veins and breccias are best developed in areas where the bedding had been folded subparallel to the principal shortening direction. Their geometry suggests they developed when the stratigraphy was less folded and were subsequently rotated during interseismic creep and shear failure. Stage 2 veins are surrounded by zoned alteration facies that extend from the vein margin up to 100 m into the wallrock. These alteration facies included a proximal albite-pyrite, followed by a biotite-carbonate and distal sericite-chlorite. Locally intense phengite alteration that is generally restricted to 20 cm from the vein also occurs around stage 2 veins. Intense phengite alteration is most common around the periphery of large albite-pyrite altered ore zones. Stage 3 veins developed during the waning stages of top-to-NE shear along the bedding parallel fault zone, after the stratigraphy was steepened to near its current orientation. These are undeformed gently dipping extension veins that cut folded and boudinaged stage 1 and 2 veins and are most abundant in the footwall sandstone. Stage 3 veins that formed through magnetite-bearing beds in the footwall sandstone developed a characteristic bright red hematitic alteration halo. The approximately 90¬∞ difference in the present orientation of stage 1-2 and 3 extensional veins is evidence of least two discrete mineralising events separated by a period of significant deformation. Rarely, deformed stage 2 ore veins with albite-pyrite alteration halos occur in the Lower Merougil Group, indicating that gold deposition was younger than c. 2666 Ma. Newly obtained U-Pb ages for hydrothermal monazite associated with gold-bearing stage 2 veins and altered wall rock indicate the main mineralising event occurred at 2632 ¬± 14 Ma. These age constraints and structural features indicate that the onset of gold mineralisation occurred after deposition of the Lower Merougil Group during NE-SW compression associated with D4 (in the generally accepted regional deformation scheme), when most other deposits in the region are interpreted to have formed. Wall rock reaction and gold deposition processes around stage 2 and 3 veins were investigated using a simplified isocon method for determining element mobility and mass-change in altered rocks. In the 30 years since the isocon diagram was proposed to depict mass-changes, there have been many modifications and alternative methods proposed, such that there is no longer a standard approach to mass balance. Many alternative methods are arguably more rigorous, but also more complex and time consuming, and do not appear to have been widely adopted. We propose a modified isocon method that is a quick and simple data-driven approach to the identification of immobile elements and mass-balance calculations that does not rely on scaling or assumption. A simple column graph depicting the component concentration ratios arranged from greatest to least ('ranked ratio' plot) is the fastest and simplest way of identifying the largest group of elements with similar concentration ratios in the altered and least altered rock, which in most cases will be the immobile elements. This method addresses the effect of protolith heterogeneity on mass-balance calculation by comparing each sample of least altered rock to an average of multiple analyses to define a range outside of which mass gains and losses in altered samples can confidently be reported. Identification of immobile elements in weakly to moderately altered rock is typically accomplished by finding the largest group of elements with the same (very similar) concentration ratios that form a plateau on the ranked ratio plot. Mass-changes of each measured component are then depicted as a function of their original concentration in the protolith rather than as a proportion of the total rock mass as in previous methods. Mass balance calculations for large datasets of spatially related samples can be quickly processed using this method, elucidating geochemical gradients from hydrothermal alteration. The new approach is demonstrated with reference to samples from the shale-hosted Invincible (orogenic) gold deposit of the Eastern Goldfields Province, Western Australia, and the sedimenthosted Kansanshi copper-gold deposit of the Copperbelt Province, Zambia. The results of geochemical mass-balance and mineral paragenesis studies indicate gold mineralisation at Invincible occurred partially due to sulfidation of iron-bearing minerals, dominantly biotite and magnetite in the wall rocks. Albite-pyrite wall rock alteration around stage 2 veins overprints least altered mudstone that contained abundant biotite, plagioclase, quartz, muscovite and ilmenite. Fluctuating stability of albite, zoned Ba-rich K-feldspar and sulfate minerals including barite in stage 2 veins indicates varying hydrothermal conditions during mineralisation that may have been caused by fluid mixing....