Further evidence for ∼8 kbar amphibolite facies metamorphism in the Marymia Inlier, Western Australia

Pressure estimates for amphibolite-facies metamorphism at Plutonic Gold Mine (Plutonic), Marymia Inlier, Western Australia, were recently revised significantly upwards from ∼4 ± 2 kbar/550–600°C to ≥8 kbar/∼600°C, based on the calculated stability fields for mineral assemblages in garnet-free mafic rocks. These conditions are anomalous in the context of the Yilgarn Craton. Here, we present new mineral equilibria calculations for rare garnet-bearing rock types from Plutonic that confirm those higher pressure estimates, and provide confidence that the determinations of metamorphic conditions based only on results from metamorphosed mafic rocks are robust and reliable. Taken together, the new estimates (7.3–8.2 kbar/580–590°C) from the garnet-bearing rocks, and the existing results from the mafic rocks, provide evidence that, most probably during the late Archean, rocks now exposed along the northern margin of the Yilgarn Craton underwent substantial increases in pressure, which was likely followed by rapid exhumation.


INTRODUCTION
The Marymia Inlier is an Archean basement high located between the Pilbara and Yilgarn cratons in Western Australia, and surrounded by Proterozoic basins (Figure 1). Plutonic Gold Mine (Plutonic) is located in the Plutonic Well Greenstone Belt (PWGB; Figure 1), a northeastÀsouthwest-trending, »50 km-long and »10 km-wide graniteÀgreenstone terrane within the Marymia Inlier. The early history of the belt is comparable with that of other greenstone sequences in the Yilgarn Craton. Later features, including uplifting to form the basement high, are attributed to suturing events during the Paleoproterozoic Capricorn Orogeny (ca 1830À1770 Ma) that brought together the Pilbara and Yilgarn cratons to form the West Australian Craton (e.g. Tyler & Thorne 1990;Bagas 1999;Cawood & Tyler 2004;Pirajno et al. 2004).
Peak metamorphism at Plutonic is dated between 2660 and 2630 Ma (Vielreicher & McNaughton 2002;Vickery 2004). Metamorphic pressure and temperature (PÀT) conditions at this time in the wider Yilgarn Craton are generally interpreted to lie along anticlockwise PÀT paths with peak pressures of »4 kbar (Goscombe et al. 2009), based on data from rocks located »1000 km to the south of the Marymia Inlier. Gazley et al. (2011a) presented PÀT results from garnet-free mafic rocks at Plutonic that indicated that the rocks in this part of the Marymia Inlier were affected by considerably higher pressures. Here we present PÀT estimates from two rare garnet-bearing rocks from Plutonic that support and even more tightly constrain the estimates obtained from the garnet-free mafic rocks, and in combination suggest that the northern margin of the Yilgarn Craton was affected by a relatively high-pressure metamorphic episode, most probably during the late Archean.

GEOLOGICAL SETTING
Comprehensive reviews of the geological setting of the local Plutonic area and the deposit geology have recently been presented in Gazley et al. (2011aGazley et al. ( ,b, 2012Gazley et al. ( , 2014a and Duclaux et al. (2011Duclaux et al. ( , 2012Duclaux et al. ( , 2013   and ca 1720 Ma) and not associated with amphibolite-facies conditions (e.g. Gazley 2011;Duclaux et al. 2012Duclaux et al. , 2013. Until recently, peak metamorphism was based on interpreted qualitative mineral assemblage associations (P max <3 kbar;  and results of conventional geothermobarometry (P max »4 § 2 kbar; Vickery 2004) to have taken place at relatively low pressures at approximately 550 C . Metamorphism was considered to have involved little overall change in crustal thickness (Bagas 1999). Although garnet-bearing assemblages have traditionally been preferred for pressureÀtemperature studies, recent advances in activityÀcomposition models for amphiboles (e.g. Diener et al. 2007;Bhadra & Bhattacharya 2007;Diener & Powell 2012) have provided increased scope for interpreting PÀT records using garnet-free mafic rocks such as those that make up nearly all of the Mine Mafic Package. This paper presents an examination of the volumetrically insignificant garnetbearing assemblages that was undertaken to assess whether it is possible to more precisely constrain the peak pressure conditions.

ANALYTICAL METHODS
Electron probe microanalyses (EPMA) were obtained using wavelength-dispersive methods at Victoria University of Wellington, New Zealand, using a JEOL JXA-733. The accelerating voltage was 15 kV, and the sample current was 12 nA. Spot sizes of 1À20 mm were used depending on the volatile and/or alkali content of the mineral being analysed. Photographs and back-scattered electron (BSE) images were used to locate and record the sites analysed during EPMA. The BSE images presented in this paper were obtained using a Phillips XL40 scanning electron microscope at the Australian Resources Research Centre, Perth, Australia.
Major-element analyses of whole-rock samples of MG027 and MG327 were carried out by X-ray fluorescence spectrometry (XRF) at SpectraChem Analytical Ltd, Wellington, New Zealand. Determinations of Fe 2C were conducted by titration at Albert-Ludwigs-Universit€ at Freiburg, Germany, following the techniques outlined in Heinrichs & Herrmann (1990). Further details of both whole-rock and titration analyses are available in Gazley et al. (2011a).
PressureÀtemperature pseudosections for the two garnetiferous rock samples were calculated using Theriak/Domino (de Capitani & Brown 1987;de Capitani & Petrakakis 2010) with the internally consistent thermodynamic dataset 5.5 of Holland & Powell (1998, updated 22 November 2003. Mineral stability and the topology of the pseudosections were confirmed using THERMOCALC v. 3.37 (Powell & Holland 1988

PETROGRAPHY AND MINERAL CHEMISTRY
The two samples selected provide a rare opportunity to compare interpretations gained from pseudosection studies of garnet-bearing and garnet-free rocks from Plutonic. It is unlikely that any of the samples were affected by mineralising fluids, as Au mineralisation at Plutonic is associated with very narrow alteration haloes (e.g. Gazley 2011;Gazley et al. 2011b), and both samples were collected from sites distal to Au mineralisation. MG027 is an amphibole-rich garnet-bearing rock that is broadly mafic in composition. MG327 is a thin sedimentary layer that is intercalated with the mafic rocks of the Mine Mafic Package. The samples were collected from the main underground mine at Plutonic: MG027 from diamond drill hole UDD8252 at 7.1 m; and MG327 from diamond drill hole UDD1886 from 178.2 to 179.1 m. Representative BSE images for these two samples are shown in Figure 2 with bulk composition data presented in Table 1.
Sample MG027 is a coarse-grained garnetÀ hornblende rock that lacks plagioclase and contains latestage chlorite replacing the garnet. Garnet crystals can be >1 cm across and the mole fraction of almandine X alm is uniformly »0.62 with X spss [Mn / (Fe T C Mg C Ca C Mn)] ranging from 0.16 to 0.19, and X grs [Ca / (Fe T C Mg C Ca C Mn)] ranging from 0.05À0.06. Quartz is present in the matrix of the rock, as is epidote, which has radioactive haloes in surrounding amphibole. Pyrrhotite is the dominant opaque phase in the matrix, with lesser ilmenite. The garnet contains inclusions of epidote, ilmenite and a tschermakitic hornblende, which is interpreted to be the stable amphibole throughout the highest-grade parts of the metamorphic history. Large amphibole crystals in the rock are sharply zoned with tschermakitic hornblende rims and subcalcic monoclinic amphibole (cummingtonite) cores (Figure 2a). Ilmenite contains 1À2 wt% MnO.
Sample MG327 is a metasediment that contains small, subhedral garnet porphyroblasts (typically <3 mm) in a fine-grained matrix (typically <100 mm crystals) of quartz, plagioclase, biotite, epidote, calcite, pyrrhotite and ilmenite. A weak metamorphic foliation defined by the platy minerals wraps the garnet. The garnet contains inclusions of chlorite, quartz, plagioclase, K-feldspar and rare pyrrhotite, with chlorite more abundant in the garnet interiors, and quartz more abundant nearer the garnet rims. The chlorite inclusions are volumetrically insignificant and are inferred to be inherited from early in the metamorphic history, rather than being part of the equilibrium peak assemblage. The scattered individual plagioclase inclusions in garnet have a wide range of compositions, from X an D 0.10 to 0.86, and there is no systematic pattern of changing plagioclase composition from garnet core to rim. The simplest inference is that this wide compositional range has been inherited from the sedimentary protolith. Occasional grains of calcite are associated with late-stage veining.

PÀT PSEUDOSECTIONS
PressureÀtemperature pseudosections for MG027 and MG327 are presented in Figures 3 and 4, respectively. The input bulk compositions are based on XRF data and are presented in Table 1 along with bulk composition data for metabasaltic samples MG053 and MG313 from Gazley et al. (2011a) for comparison. The growth of garnet and absence of plagioclase in MG027 can be attributed to the higher MnO, and lower Na 2 O and CaO contents compared with MG053 and MG313, despite overall similar mafic bulk compositions (Table 1). The Al 2 O 3 concentration of MG327 (Table 1) is consistent with that of mafic sediment.
The peak metamorphic mineral assemblage in sample MG027 consists of garnetÀhornblendeÀilmenite (Figure 2a; Figure 3, shaded field) without plagioclase, hedenbergite, cummingtonite, chlorite or tremolite. This field lies between approximately 560 and 640 C and between approximately 7.5 and >12.0 kbar. The calculated abundance of garnet (»31 vol%) is greater than the amount present in the rock (»20 vol%) because the garnet has been partially replaced along grain margins and fractures by chlorite. The presence of cummingtonite in the cores of many of the hornblende grains in MG027 (Figure 2a) attests to the early prograde growth of amphibole at »3 kbar and 530 C. Under these conditions calculated abundances of cummingtonite and hornblende are approximately equal. Overgrowth of cummingtonite cores by hornblende during peak metamorphism led to the zoned amphiboles that are preserved in MG027.
The peak metamorphic assemblage in sample MG327 that consists of biotiteÀepidoteÀgarnetÀplagioclaseÀK-feldsparÀilmeniteÀquartz, best corresponds to the field shaded in Figure 4. A calculated abundance of K-feldspar at »4 vol% in this field is slightly higher than is present in the rock and could result from minor K 2 O mobility during metamorphism and the sensitivity of calculated K-feldspar abundances to K 2 O. Decreasing the K 2 O content from 1.66 to 1.41 wt% results in a calculated K-feldspar abundance of »1À2 vol% that more closely matches the concentration in the rock but does not alter the location of the field boundaries significantly. Taken by itself, the narrow peak mineral stability field in MG327, which has a positive slope in PÀT space, extends from approximately 450 to 650 C and from approximately 2.5 to 9.5 kbar but does not provide very useful PÀT constraints. However, using the previously determined peak temperature of 600 C from hornblendeÀplagioclase thermometry (using the calibration of Holland & Blundy 1994) on this field (Gazley et al. 2011a), pressure estimates of 7.4À8.3 kbar are obtained. The fine plagioclase grainsize and the wide variation in plagioclase composition rule out using traditional geothermobarometry methods to obtain estimates to compare with the PÀT estimates derived from pseudosections. Peak metamorphic estimates can be further constrained by considering the overlapping peak assemblage of multiple PÀT pseudosections from rocks that experienced the same metamorphic conditions. Gazley et al. (2011a) used this approach to constrain the peak metamorphic conditions at Plutonic to »600 § 50 C and 8 § 2 kbar. The two new PÀT pseudosections presented here for garnet-bearing rocks   constrain the peak PÀT conditions to be 580À590 C at 7.3À8.2 kbar ( Figure 5) with errors derived from PÀT pseudosections of §50 C and §1 kbar, in line with the recommendations of Powell & Holland (2008). The derived peak metamorphic conditions are also consistent with mineral compositions for garnet calculated in Theriak/ Domino for these samples; MG027 calculated X alm D »0.64, cf. measured X alm D »0.62; MG327 calculated X alm D 0.42, cf. measured X alm D 0.52À0.67. A better match between the calculated X alm content in garnet in MG327 and the measured content can be achieved by decreasing the K 2 O content by a small amount, for example removing 0.25 wt% K 2 O results in a X alm content in garnet of »0.52, further highlighting the sensitivity of this bulk composition to variations in K 2 O concentration. These new constraints for peak metamorphic conditions are entirely consistent with those determined by Gazley et al. (2011a). Importantly our new PÀT pseudosections remove the need to rely on a conventional (hornblendeÀplagioclase) geothermometer to constrain peak pressure conditions at »8 kbar, in a peak assemblage field that Gazley et al. (2011a) defined as extending from 5 to 9 kbar and from 500 to 630 C.

DISCUSSION AND CONCLUSION
The new results presented here confirm the interpretation of Gazley et al. (2011a)   conditions (»600 C/»8 kbar, double the previously accepted pressure estimates) prevailed at the southern end of the Marymia Inlier, most probably during the late Archean. More significantly, we have increased the lower pressure constraints using the pseudosection data errors on pressure determinations of § 1 kbar (Powell & Holland 2008). Our new PÀT determinations are consistent with the interpretation that the Marymia Inlier records an event that occurred along the northern margin of the Yilgarn Craton during the late Archean. However, the nature of this event remains uncertain; in large part because it remains unclear whether the Marymia Inlier is connected to the Yilgarn Craton at depth (Bagas 1999). Without such fundamental data, tectonic interpretations involving the Marymia Inlier can only be conjectural. Gazley et al. (2011a) argued for a steep pressure increase from »3 to 4 kbar at »500 C to 8 kbar at »600 C. Based on our thermodynamic modelling, the peak conditions were »600 C and »8 kbar. Early prograde conditions are constrained at »3À4 kbar and »500 C, as early cummingtonite is overgrown by hornblende in sample MG027 ( Figure 3) and equates to a geothermal gradient between 35 and 45 C per km, consistent with the geothermal gradient derived for metamorphism across a large part of the Yilgarn at this time (Czarnota et al. 2010). Peak metamorphic conditions at Plutonic were at a significantly higher pressure but without a corresponding increase in temperature. This gives rise to an apparent geothermal gradient of only 20 C per km and therefore does not represent an equilibrium geotherm that would have resulted in peak temperatures in excess of 850 C. The lack of evidence that these rocks reached such high temperatures indicates that the rocks may have been exhumed rapidly following the high-pressure metamorphism.