This study defines a comprehensive chemostratigraphic framework for the 2820-2738 Ma magmatic rocks of the northwestern Youanmi Terrane of the Yilgarn Craton. Sixteen chemostratigraphically distinct units can be correlated between greenstone belts that are separated by later voluminous granitic plutons. On a regional scale, the chemostratigraphic units define discontinuous, broadly NNE striking zones up to 350 km long and 50 km wide (including later granitic plutons). Neodymium isotope and trace element systematics of magmatic rocks require an increasing component of metasomatised mantle from c. 2820 to 2738 Ma and variable recycling of a 3050-2920 Ma crustal component. We interpret the first magmatic episode at 2820-2782 Ma to reflect lithospheric thinning and asthenosphere upwelling in a back-arc-like setting and attribute the metasomatism of sources for multiple mafic units to subduction of 3050-2920 Ma, predominantly isotopically juvenile, mafic crust. Following an approximate 16 Ma hiatus, a second magmatic episode at 2766-2738 Ma is dominated by crustal-derived transitional TTGs with lower volumes of mantle-derived shoshonitic and sanukitoid-like rocks, and LILE-LREE enriched tholeiitic mafic to felsic rocks. The compositions of the transitional TTG and shoshonitic rocks in this study area are similar to potassic granites and shoshonitic rocks (including shoshonitic lamprophyres) from post-orogenic continental collision settings, such as the Mesozoic Tibetan Plateau. We interpret this later episode as reflecting post-orogenic melting of crustal and upper (modified) mantle sources caused by upwelling mantle related to delamination of arc-crust or slab breakoff. The northwestern Youanmi Terrane preserves an extensive package of Archean magmatic rocks that can be confidently attributed to subductionmagmatism and provides a reference point to show that some form of subduction occurred in the MesoNeoarchean.

A striking feature of the Yilgarn craton at the current erosional level is an abundance of late K-rich granites with radiogenic heat production elevated far above global crustal averages. Extrapolated back in time, the total thickness and contribution to crustal heat production and heat flow from these granites were greater, implying that the deeper crustal sources must also have had elevated radiogenic heat production. Through back-calculated and time-integrated one-dimensional thermal modeling underpinned by geological and geochemical constraints for the model crustal columns, we find that elevated radiogenic heat production provided a significant internal driver for prolonged crustal melting and eventual cratonization of the Yilgarn craton. Our results show that elevated thermal gradients driven by high heat production thermally primed the midand deep crust at or above the threshold for large-volume partial melting over long periods of time, as evidenced in the magmatic rock record. This would have been amplified by any additional heat that may have been provided by the mantle melting processes that punctuated the geological history. Over time, advective movement of progressively more radiogenic heat production to the shallower crust would have resulted in two complementary outcomes: progressively refractory deep crust and long-term cooling. The widespread granite "bloom" at 2650-2600 Ma records the final time at which the crust was fertile enough to melt in large volumes and the thermal gradient was hot enough to intersect the solidus. The magnitude of radiogenic heat production in the Yilgarn craton has been underestimated in previous studies, resulting in an underappreciation of the importance of its contribution to internal drivers of magmatism and ultimately cratonization.

Molybdenum isotopes (reported as delta Mo-98 relative to NIST-3134) show resolvable isotope differences in igneous rocks with the continental crust being markedly heavier in isotope composition than mid-ocean ridge lavas, lunar basalts or the Earth's mantle. The tholeiitic differentiation series at the intra-plate Hekla volcano (Iceland) shows no resolvable Mo isotope differences from basaltic to rhyolitic compositions. In contrast, convergent margin lavas show a transition from isotopically lighter mantle to heavy continental crust, suggesting that subduction processes drive continental crust towards heavier values. Archean komatiitic lavas, presumed probes of the Archean mantle, have Mo isotope values identical to modern depleted mantle, raising the questions if and how the Mo isotope crust-mantle disparity developed so early in Earth's history. Here we present new Mo isotope data for a set of cumulate rocks from the Upper Zone of late Archean (2.8 Ga) Windimurra Igneous Complex, a mafic/ ultramafic layered intrusion. The intrusion is not subduction related and contains no apparent primary hydrous minerals. We tested the effect of crystal fractionation on Mo isotopes in relatively dry melt along a tholeiitic liquid line of descent by using the cumulate effect of normally anhydrous minerals in the layered intrusion. Near mono-mineralic olivine-pyroxene-rich, feldspar-rich and Fe-Ti-rich oxides show small variations (similar to 0.15%0) in Mo isotope signatures. This is consequently to predominantly isotopically light Fe-Ti-oxide-rich and isotopically heavier feldspar-rich rocks, respectively. This is suggesting minor Mo isotope fractionation, even in dry, tholeiitic systems, which however, counterbalance each other and thus potentially remain undetected. On average, the Windi-murra mantle source is indistinguishable, or slightly isotopically lighter than the Mo isotope signature of komatiites. This is reinforcing an isotopically light Mo isotope signature of Archean mantle sources of high-degree mantle melts and is extending these signatures to predominantly mafic Archean crust. It remains to be tested if Archean felsic crust resembles modern continental crust in its heavy isotope values and to which extend the mantle was already isotopically depleted in Mo isotopes at Mesoarchean time.

Layered mafic intrusions (LMI) are stratified igneous bodies, with many examples that are economically significant for a number of metals including chromium, platinum group elements, vanadium, titanium, and iron. The Windimurra Igneous Complex (WIC) of the central Murchison Domain, Western Australia, the largest layered mafic intrusion in Australia, comprises a giant (2500 km2) LMI, with an interpreted thickness of approximately 11 km. The Upper Zone is free of primary hydrous minerals and contains ore-grade Fe-enrichment discoveries up to 66.6 wt% FeOt. Elevated Ti, V and Cr in magnetite at the base of the Upper Zone, and high Cr in the Middle Zone, combined with structural observations in the field, are consistent with injection of two separate magmatic pulses. In this study, four deep diamond drill cores and 92 reverse circulation drill holes from the Upper Zone and uppermost Middle Zone of the WIC sample a near continuous section of 820 m in the Upper Zone, including numerous magnetite-rich horizons. An increase in the whole-rock TiO2/V2O5 upward in the Upper Zone, a decrease in the frequency of magnetite-rich horizons, and an increase upwards in P content, have been used in combination with airborne magnetic surveys, to develop a magnetic stratigraphy for the Upper Zone. The stratigraphy reflects the prospectivity for magnetite, vanadium and chromium, all of which are more abundant at or near the base of the Upper Zone. Such exploration vectors can be used to better target local iron, vanadium as well as chromium mineralization, much of which appear spatially associated with the Upper Zone's basal contact.

The formation of stable, evolved (silica-rich) crust was essential in constructing Earth’s first cratons, the ancient nuclei of continents. Eoarchaean (4000–3600 million years ago, Ma) evolved crust occurs on most continents, yet evidence for older, Hadean evolved crust is mostly limited to rare Hadean zircons recycled into younger rocks. Resolving why the preserved volume of evolved crust increased in the Eoarchaean is key to understanding how the first cratons stabilised. Here we report new zircon uranium-lead and hafnium isotope data from the Yilgarn Craton, Australia, which provides an extensive record of Hadean–Eoarchaean evolved magmatism. These data reveal that the first stable, evolved rocks in the Yilgarn Craton formed during an influx of juvenile (recently extracted from the mantle) magmatic source material into the craton. The concurrent shift to juvenile sources and onset of crustal preservation links craton stabilisation to the accumulation of enduring rafts of buoyant, melt-depleted mantle.

Australia's and China's resources (e.g. Olympic Dam Cu-U-Au-Ag and Bayan Obo REE deposits) highlight how discovery and mining of iron oxide copper-gold (IOCG), iron oxide±apatite (IOA) and affiliated primary critical metal deposits in metasomatic iron and alkali-calcic (MIAC) mineral systems can secure a long-term supply of critical metals for Canada and its partners. In Canada, MIAC systems comprise a wide range of undeveloped primary critical metal deposits (e.g. NWT NICO Au-Co-Bi-Cu and Québec HREE-rich Josette deposits). Underexplored settings are parts of metallogenic belts that extend into Australia and the USA. Some settings, such as the Camsell River district explored by the Dene First Nations in the NWT, have infrastructures and 100s of km of historic drill cores. Yet vocabularies for mapping MIAC systems are scanty. Ability to identify metasomatic vectors to ore is fledging. Deposit models based on host rock types, structural controls or metal associations underpin the identification of MIAC-affinities, assessment of systems' full mineral potential and development of robust mineral exploration strategies. This workshop presentation reviews public geoscience research and tools developed by the Targeted Geoscience Initiative to establish the MIAC frameworks of prospective Canadian settings and global mining districts and help de-risk exploration for IOCG, IOA and affiliated primary critical metal deposits. The knowledge also supports fundamental research, environmental baseline assessment and societal decisions. It fulfills objectives of the Canadian Mineral and Metal Plan and the Critical Mineral Mapping Initiative among others. The GSC-led MIAC research team comprises members of the academic, private and public sectors from Canada, Australia, Europe, USA, China and Dene First Nations. The team's novel alteration mapping protocols, geological, mineralogical, geochemical and geophysical framework tools, and holistic mineral systems and petrophysics models mitigate and solve some of the exploration and geosciences challenges posed by the intricacies of MIAC systems. The group pioneers the use of discriminant alteration diagrams and barcodes, the assembly of a vocab for mapping and core logging, and the provision of field short courses, atlas, photo collections and system-scale field, geochemical, rock physical properties and geophysical datasets are in progress to synthesize shared signatures of Canadian settings and global MIAC mining districts. Research on a metamorphosed MIAC system and metamorphic phase equilibria modelling of alteration facies will provide a foundation for framework mapping and exploration of high-grade metamorphic terranes where surface and near surface resources are still to be discovered and mined as are those of non-metamorphosed MIAC systems.

The Archean cratons represent the earliest fragments of the felsic continental crust, which is predominantly composed of sodic granitoids known as Tonalite-Trondhjemite-Granodiorites (TTGs). Here, we analyse a large compilation of CI normalized rare-earth-element (REEN) compositions of TTGs worldwide to elucidate how crust-forming processes may have varied in the early Earth. REENs are best suited for this purpose. But, comparing the REEN patterns of TTGs does not give meaningful results as they have similar shapes. Hence, we use the polynomial approach [1] that quantifies the linear slope and curvature of a REEN pattern using two parameters-l1 and l2, respectively. We calculate these parameters for TTG compositions from different cratons and plot them in a l-diagram (l1 vs l2). Results show that overall, TTGs are distinctly different than Phanerozoic granitoids. TTGs of different cratons exhibit two distinct trends in the l-diagram: (1) a horizontal-trend describing variations in l1-values as compared to l2-values (e.g. Pilbara craton), and (2) an inclined-trend describing a concomitant decrease in l2-values with increasing l1-values (e.g., Superior craton). Given the l1 (slope) and l2 values (curvature) represent relative enrichment of LREEs and MREEs respectively [1], the horizontal-trend signifies large variation in the LREE enrichment pattern while the inclinedtrend implies a coeval change in LREE and MREE concentrations. Trace element modelling suggests that the inclined-trend could be reproduced by decreasing the amphibole:garnet ratio in the residue along a geotherm, i.e. if the residual assemblage changes from (garnet-)amphibolite to (garnet-)granulite with progressive melting. The horizontal-trend could be due to variations in source rock compositions and represents partial melting in presence of amphibolite residues. These different residual assemblages suggest that the P-T conditions of TTG formation varied significantly among the cratons and thus, have implications for interpreting the global Hadean-Archean geodynamics.

Layered mafic-ultramafic intrusions (LMI) are among the largest igneous bodies on Earth, and represent aggregations of large volumes of mantle- and some crustal-derived melts. Melts are emplaced over time-intervals of less than 1 million years, predominantly through multiple pulses of injections into pre-existing melt-crystal slurries. The dynamic interaction of physical processes, including density-driven separation and mixing of different components, within a solidifying magma chamber leads to such extreme chemical diversity between cumulate rock units that no unified model currently explains all aspects of the genesis of these intrusions. Here we present whole-rock stable Fe isotope data (expressed in ‰ variations as δ57Fe relative to IRMM-014) for samples of drill core taken from the stratified paleo-magma chamber of the Upper Zone of the late-Archean Windimurra Igneous Complex, Western Australia. Variations from near chondritic (δ57Fe ∼ + 0‰) to heavy (δ57Fe∼ + 0.2‰) values show a co-variation with initial radiogenic Hf isotope data that is unique to the Windimurra Upper Zone. The systematic isotopic variations from the roof to the base of the Upper Zone are best explained by an intricate sequence of events that included fractional crystallization and physical mixing. We propose that melt freshly sourced from the mantle was injected into and inflated a pre-existing crystal-melt mush comprising the upper Middle Zone. Re-establishment of crystal layering after replenishment introduced a chemical stratification with the formation of what became the Upper Zone and a crystal-interstitial melt ratio decreasing from roof to base. Basal, vanadiferous magnetitite horizons crystallized from Fe out of the liquid-in-residence. Variable degrees of perturbation and chaotic stirring of crystals with imperfect mixing of new and old components was followed by rapid crystal settling and subsequent cumulate stratification. Such melt rejuvenation, proposed here to be the cause for a newly established Upper Zone, leaves no unique petrologic fingerprint and forges a range in hybrid magma compositions throughout the Upper Zone rather than one single parental melt. This incremental growth of the intrusion can explain not only the observed coupled Fe-Hf isotope systematics, but also mineral disequilibria and cryptic layering in layered intrusions. In the context of incremental pluton assembly, two-component mixing paired with source heterogeneity can, at least in parts, potentially explain zircon-hosted Hf isotope heterogeneity so often observed in larger magmatic bodies.

In the Yilgarn Craton, the Mougooderra Formation is an extensive sedimentary sequence that preserves low-variance metamorphic mineral assemblages and unconformably overlies older greenstone sequences within the Youanmi Terrane. Here we present garnet Lu-Hf and Sm-Nd geochronology in combination with pressure-temperature (P-T) estimates of the Mougooderra Formation and the underlying Polelle Group to give insights into the geodynamics of this region at the time of metamorphism. Metamorphic rocks from the Mougooderra Formation have variable metamorphic mineral assemblages that reflect bulk rock concentrations of Fe, Mg, Ti, Al, and K at constant P-T. Based on these variations, the metamorphic rocks of this formation can be separated into a high Fe-Mg group, a high Al group and an average pelitic group that each preserve low-pressure mineral assemblages. Thermobarometry and phase equilibrium modelling indicate that the Mougooderra Formation experienced P-T conditions of 545-580 degrees C and similar to 2-3.5 kbar, while the Polelle Group reached 610-650 degrees C at similar to 4-7.2 kbar. Garnet Lu-Hf and Sm-Nd geochronology returned respective ages of 2686 +/- 18 Ma and 2611 +/- 35 Ma for the Mougooderra Formation and ages of 2685 +/- 15 Ma and 2590 +/- 21 Ma for the Polelle Group. Lu-Hf garnet ages are interpreted to record the timing of late prograde metamorphism and indicate simultaneous metamorphism, whereas the geological significance of the Sm-Nd ages is unclear. Garnet Lu-Hf ages obtained in this study are similar to emplacement ages of syntectonic granites in the region. Although these distal contemporaneous granites may have contributed to elevated heat flow, the high apparent thermal gradients calculated here, as well as synchronous craton-wide magmatism, reflect a regional thermal event at this time. The ages, P-T conditions, and high apparent thermal gradients from the study area suggest metamorphism occurred as the result of high heat flow into thinned crust of the Youanmi Terrane and indicates a complex thermal history within this region of the Yilgarn Craton. These results are consistent with the hypothesis that metamorphism during the Archean was more closely related to heat from mantle-derived magmatism.

Evidence for a multi-stage history of uplift and metasomatism is preserved in a suite of highly chromian, garnet-rich peridotites found as xenoliths in Newlands and Bobbejaan kimberlites, South Africa.A high proportion of the garnets and chromites in these rocks plot in the diamond facies fields on Cr 2 O 3 -CaO and Cr 2 O 3 -MgO wt.% plots respectively.Petrographic evidence suggests that the earliest known mineralogy is a garnet-olivine rich rock with granuloblastic texture.A down-P-T event caused exsolution of spinel and pyroxenes in garnet and these inclusion textures have subsequently been partially annealed in most samples.Harzburgitic samples have garnets with inclusions of serpentine ± chromite and in lherzolitic samples, cpx is also present.Most of the garnets have strongly developed, diffusion-controlled, major element zonation patterns which are a result of: (1) External reequilibration where diffusion is towards the matrix.(2) Internal re-equilibration where diffusion is towards inclusions.(3) Metasomatic zonation between garnet core and its metasomatic rim.The compositional trajectories associated with (1) and ( 2) may be closely modelled by means of sliding, garnet-spinel transition reactions involving differing bulk pyroxene Ca compositions and they conform to a decompression event.The diffusing cations for P-T re-equilibration show a decrease in Cr/Al in the vast majority of samples.Diffusion timescales for re-equilibration zonation are generally 0.3-2 Ma with the shorter timescales belonging to internal and the longer timescales belonging to external re-equilibration zonations.Few samples, however, display evidence for relatively long lived (5-10 Ma) diffusion involving primarily an increase in Ca/Mg and Ti towards the garnet rim, which is attributed to metasomatic fluid that percolated though the matrix.Modal reconstructions reveal that the pre-exsolution bulk compositions are garnets which plot as extensions of the zonation trends at higher Cr.Therefore a substantial change in P-T (involving approx.10kb decompression) is proposed as an explanation for exsolution and the P-T reequilibration zonations.Based on the extent and duration of this event, we have recognised a possible correlation with Archaean crustal events identified along the Colesberg magnetic lineament at 2.7-2.9Ga.

Rocks with chemical compositions similar to Cenozoic boninites occur in many Archean cratons (boninite-like rocks), but they are rarely well-preserved, well-sampled, or presented within chronoand chemo-stratigraphic context. This study provides a detailed description of the most extensive and well-preserved Archean boninite-like rocks reported to date. Within the 2820 to 2740 Ma magmatic suites of the northwest Youanmi Terrane, Yilgarn Craton, boninite-like rocks occur as two distinct units. The first boninite-like unit is thinner (several 10 s of m thick), occurs close to the base of the 2820-2800 Ma Norie Group and includes both volcanic flows and subvolcanic intrusions. The second boninite-like unit is thicker (locally several 100sm), occurs near the base of the 2800-2740 Ma Polelle Group and consists of mainly fine-grained volcanic flows with local cumulate units. On average, major and trace element compositions for Youanmi Terrane boninite-like rocks are marginal between basalt, picrite and boninite and they have asymmetrically concave REE patterns, and Th-, Zr-Hf enrichments, similar to many Phanerozoic low-Si boninite suites, but at generally higher MREE-HREE contents. We report over 300 new whole-rock geochemical analyses, and 16 new Sm-Nd isotopic analyses, and associated petrographic evidence, including representative mineral compositions, which we support with published geochemical analyses and several decades of fieldwork in our study area. Comparison between Archean boninite-like rocks and Cenozoic boninites shows that most Archean examples had less depleted sources. We consider two possible petrogenetic models for the Youanmi Terrain examples: (1) they reflect variably contaminated komatiites, or (2) they reflect melts of metasomatised refractory mantle, analogous to Phanerozoic boninites. Trace element modelling indicates that crustal contamination could potentially produce rocks with boninite-like compositions, but requires an Al-enriched komatiitic parent liquid, for which there is no field evidence in our study area. Initial epsilon(NdT) values in pre-2800 Ma rocks (epsilon(NdT )-0.4 to +1.2) are on average slightly higher than those in 2800-2733 Ma examples (epsilon(NdT) -3.2 to +1.2), compatible with increasing mantle metasomatism involving recycling of >= 2950 Ma crust. Integration of trace element and Nd isotopic data demonstrates that significant direct crustal assimilation was restricted to felsic magmas. The Th-Nb and Ba-Th systematics of mafic-intermediate rocks reflect fluid- and sediment-derived processes in the mantle, with boninite-like examples being linked primarily to fluid metasomatism. We compare the well-preserved igneous textures and mineralogy of Youanmi Terrane boninite-like rocks with those of their Phanerozoic counterparts, and based on studies of the latter, suggest that former had similarly hot, H2O-rich parent magmas. The association of boninite-like rocks in the Norie and Polelle Groups with coeval high-Mg andesites, sanukitoids and hydrous mafic intrusions of the Narndee Igneous Complex strongly suggests a metasomatised mantle source and subduction operating in the Yilgarn between 2820 and 2730 Ma.

This study is the first to constrain the absolute timing of hypogene iron mineralization in Archean BIF located in the Yilgarn Craton. In situ SHRIMP U–Th–Pb geochronology on xenotime and monazite grains has been used to constrain the age of hypogene magnetite replacement ores at the Beebyn deposit and hypogene magnetite vein ores at the Madoonga deposit in the Weld Range study area. The Beebyn magnetite replacement ores (c. 2627 Ma) are younger than the published maximum depositional age of the BIF hosts of the Wilgie Mia Formation (2792 ± 9 Ma) and partially overlaps crystallization ages of granitic rocks (2757–2606 Ma) that intrude supracrustal rocks throughout the study area. These plutons, and their probable subvolcanic expressions, are considered to be likely sources of energy and fluids responsible for magnetite replacement ores. In contrast, Madoonga magnetite veins record multiple dates: the first at 2857 ± 41 Ma, followed by events at c. 2775 Ma through 1812 Ma. The two oldest monazite dates of 2857 ± 41 and 2832 ± 51 Ma are interpreted to be mineralization ages for magnetite veins, possibly related to subseafloor volcanism and base metal VMS systems; whereas the younger dates (i.e. 2775–1812 Ma) probably represent multiple episodes of phosphate mineral precipitation related to reactivation of structures and overprinting by discrete pulses of hydrothermal fluids. The Madoonga BIFs are genetically distinct from the Beebyn BIFs and are the oldest dated BIFs at Weld Range, being older than c. 2857 Ma, but younger than the c. 2970 Ma felsic volcanic rocks in the stratigraphic footwall to the Madoonga BIFs. Hydrothermal events at the Madoonga deposit (2215–2120 Ma) coincide with published dates for rifting, mafic–ultramafic magmatism, and basin development along the northern margin of the Yilgarn Craton (2215–2145 Ma), whereas the younger phosphate mineral dates correspond with the Glenburgh (2002–1947 Ma) and Capricorn Orogenies (1817–1772 Ma). Tectonic activity along the northern margin of the Yilgarn Craton coincides with reported ages for hematite mineralization in BIF-hosted iron-ore deposits in the Hamersley Basin and Pilbara Craton, suggesting the far-reaching effects of tectonism along paleocratonic boundaries as drivers for iron mineralization in BIF.

Figure 1 : Schematic illustration of the generation of humped REE profiles in garnet by disequilibrium processes.Modified from Shimizu (1999).A is the initial garnet composition, B the final garnet composition, C the melt in equilibrium with the initial garnet and D the metasomatic fluid.Intermediate lines between A and B are diffusion-controlled profiles 1-4 representing progressive degrees of diffusion.

Garnets with depleted oxygen isotopes were found among garnet peridotites from Udachnaya kimberlite pipe (Yakutia, Russia). Oxygen isotope analysis of mineral grains was performed at Far East Geological Institute by the fluorination method using BrF 5 and infrared continuous Nd-YAG laser ( λ =1.064 μ m, CW, 100W), coupled with Finnigan MAT 252 spectrometer. Accuracy was 0.1‰ (at 1 S.D. and n =5) for NBS-28, NBS-30 standards. Studied garnet peridotites have a medium temperature calculated range (900-1140 ºC), corresponding to a depth interval from 100 to 160 km of lithospheric Garnets of two peridotites (samples are plotted in sub-cal с ic area of harzburgitic paragenesis (G10 type) and the other garnets are plotted in area of lherzolitic paragenesis

Applying the Th/Yb–Nb/Yb plot of Pearce (2008) to the well-studied Archean greenstone sequences of Western Australia shows that individual volcanic sequences evolved through one of two distinct processes reflecting different modes of crust–mantle interaction. In the Yilgarn Craton, the volcanic stratigraphy of the 2.99–2.71 Ga Youanmi Terrane mainly evolved through processes leading to Th/Yb–Nb/Yb trends with a narrow range of Th/Nb (‘constant-Th/Nb’ greenstones). In contrast, the 2.71–2.66 Ga volcanic stratigraphy of the Eastern Goldfields Superterrane evolved through processes leading to Th/Yb–Nb/Yb trends showing a continuous range in Th/Nb (‘variable-Th/Nb’ greenstones). Greenstone sequences of the Pilbara Craton show a similar evolution, with constant-Th/Nb greenstone evolution between 3.13 and 2.95 Ga and variable-Th/Nb greenstone evolution between 3.49 and 3.23 Ga and between 2.77 and 2.68 Ga. The variable-Th/Nb trends dominate greenstone sequences in Australia and worldwide, and are temporally associated with peaks in granite magmatism, which promoted crustal preservation. The increasing Th/Nb in basalts correlates with decreasing εNd, reflecting variable amounts of crustal assimilation during emplacement of mantle-derived magmas. These greenstones are typically accompanied in the early stages by komatiite, and can probably be linked to mantle plume activity. Thus, regions such as the Eastern Goldfields Superterrane simply developed as plume-related rifts over existing granite–greenstone crust – in this case the Youanmi Terrane. Their Th/Nb trends are difficult to reconcile with modern-style subduction processes. The constant-Th/Nb trends may reflect derivation from a mantle source already with a high and constant Th/Nb ratio. This, and a lithological association including boninite-like lavas, basalts, and calc-alkaline andesites, all within a narrow Th/Nb range, resembles compositions typical of modern-style subduction settings. These greenstones are very rare, and were probably only preserved when fortuitously stabilised by granitic magmatism related to the evolution of later variable-Th/Nb greenstones. The rarity of constant-Th/Nb trends suggests that either processes forming them never dominated Archean greenstone evolution, or that such greenstones simply were rarely preserved. Metamorphic mobility of Th renders the Th/Yb–Nb/Yb plot inappropriate for interpreting Eoarchean greenstone units worldwide. Nevertheless, such sequences appear dominated by volcanic rocks that, in modern settings, reflect only the embryonic or initiation stages of subduction. They probably record subduction failure rather than anything resembling modern-style subduction.

The Archean Windimurra Igneous Complex consists of distinct components, including a thick layered series, with a cumulate mineral stratigraphy similar to the zones identified in the well-studied Bushveld Complex, South Africa. The complex is part of the plume-related and laterally extensive 2.81 Ga Meeline Suite, the intrusive component of a large igneous province. It is an anhydrous tholeiitic suite consisting of five layered mafic-ultramafic intrusions 25-85 km in the long dimension. These intrusions host significant V-Ti mineralization in their fractionated, Fe-rich upper zones. Recent mapping, combined with aeromagnetic, gravity and seismic surveys, has provided unparalleled three-dimensional constraints on the largest of these intrusions. The results of three-dimensional modelling show that it is thicker than previously recognized. At c. 11 km, it is the thickest layered mafic-ultramafic intrusion identified globally and one of the largest such intrusions volumetrically. The mineral zone stratigraphy and many other features associated with this complex share similarities with the c. 800 myr younger Bushveld Complex. On a large scale, three discordant units are delineated geometrically, providing fundamental constraints on a multi-stage genetic model for magma emplacement. The indication of a thick, subsurface Ultramafic Zone provides a potential target for Ni-Cr-platinum group element mineralization.

New evidence has emerged for a different type of platy spinifex texture that has not previously been documented in the existing literature, in this case from 2.8 Ga high-Mg basalts in the Murchison Domain of the Yilgarn Craton, where petrographic and geochemical evidence shows that the dominant platy mineral is pyroxene, rather than olivine. In our samples, two scales of plates are evident. Larger plates have lengths and widths that are approximately equal and range from similar to 1000 to 15 000 mm, with thicknesses typically less than or similar to 120 mu m. These plates have <= 25 mu m thick augite rims, and cores that are now a mixture of low-temperature hydrous alteration minerals. They occur in sets of similarly oriented crystals, and typically intersect other sets of crystals at oblique angles. A second population of smaller augite-only plates occur within the interstices of the larger plates; they have lengths and widths that range from 200 to 1500 mm, and thicknesses that are typically less than or similar to 50 mu m. Pyroxene dendrites are also a typical component of this texture and represent a third scale of crystal growth, which probably crystallized shortly before the remaining liquid quenched to glass. All scales of pyroxene contained within this texture exhibit skeletal features and are considered to have crystallized rapidly. We discuss possible conditions that led to the crystallization of platy habits instead of the typical acicular ones. The exposed volcanic sequence in our study area is volcanologically similar to other Archean komatiites, such as those from the 2.7 Ga Abitibi greenstone belt, for example, and has probably experienced a similar cooling history; however, apart from having similar textures, we cannot demonstrate a komatiitic association. Liquid compositions, estimated from chilled flow margins, are distinctly lower in MgO (14.4-15.8wt %) and higher in SiO2 ( 50.9-52.1wt %) than those for most platy olivine spinifex-textured komatiites; from these compositions, we calculate dry liquidus temperatures of 1312-1342 degrees C and mantle potential temperatures of 1440-1480 degrees C. On the basis of these temperatures we question whether a mantle plume is a necessary element of their petrogenesis. `Platy olivine spinifex' is an igneous texture that characterizes komatiites and its observation in outcrops or drill core (typically prior to, or in lieu of chemical analysis) leads geologists to classify a rock as a komatiite. Field descriptions may therefore drive assumptions and interpretations surrounding the prevailing tectonic or geo-dynamic setting at the time of emplacement. We emphasize the importance of careful discrimination between a variety of spinifex textures within a local volcanological framework and caution against the habit of making direct interpretations of rock type based on the existence of spinifex textures alone.