Introduction
The correlation of Rodinian crustal domains relies upon a thorough knowledge of the tectonothermal history of those vestiges that are preserved in the present day continents. The Bunger Hills and Highjump Archipelago (HJA), together with the Windmill Islands, comprise one of a few outcrops of the East Antarctic Craton in Wilkes Land, East Antarctica, and the only exposure that can be directly correlated with the Musgrave–Albany–Fraser Orogen (MAFO), which is the Australian equivalent in Rodinian supercontinent reconstructions (Fig. 1; Clarke et al. Reference Clarke, Sun and White1995, Harris Reference Harris1995, Nelson et al. Reference Nelson, Myers and Nutman1995, Clark et al. Reference Clark, Hensen and Kinny2000, Fitzsimons Reference Fitzsimons2000, Reference Fitzsimons2003, Duebendorfer Reference Duebendorfer2002, Aitken & Betts Reference Aitken and Betts2008, Boger Reference Boger2011). Furthermore, the Bunger Hills and HJA occupy a focal position at the very western periphery of this system.
Location of the Bunger Hills and Highjump Archipelago (HJA) in Wilkes Land, East Antarctica. a. The approximate location of the Musgrave–Albany–Fraser Orogen in Australia and Antarctic. The position of the two continents reflects their relative positioning in Gondwanan reconstructions. b. Wilkes Land coastline between Casey and Mirny stations showing the approximate location of the Bunger Hills, Obruchev Hills and Windmill Islands (figure modified from Boger Reference Boger2011). The location of b. is represented by the boxed region on the map of Australia and Antarctica in a. c. Simplified regional geology of the Bunger Hills and HJA. The location of samples used in this study are indicated (modified from Sheraton et al. Reference Sheraton, Tingey, Oliver and Black1995).
Intercontinental stitching points are invoked between the MAFO and the Bunger Hills–HJA region on the basis of Gondwanan reconstructions (e.g. White et al. Reference White, Gibson and Lister2013 and references therein) that juxtapose these two continents as well as apparent similarities in the timing of metamorphism, magmatism and deformation. However, the precise configuration of the Australian plate is a source of controversy between the various tectonic reconstructions. Furthermore, the HJA and Bunger Hills are located within a region of disputed overlap between the different suggested fits and close to a critical piercing point. While the Australian part of the MAFO has been intensely studied (e.g. Clark et al. Reference Clark, Hensen and Kinny2000, Kirkland et al. Reference Kirkland, Spaggiari, Pawley, Wingate, Smithies, Howard, Tyler, Belousova and Poujol2011, Reference Kirkland, Smithies, Woodhouse, Howard, Wingate, Belousova, Cliff, Murphy and Spaggiari2013, Reference Kirkland, Smithies and Spaggiari2015, Smithies et al. Reference Smithies, Howard, Evins, Kirkland, Kelsey, Hand, Wingate, Collins and Belousova2011, Reference Smithies, Kirkland, Korhonen, Aitken, Howard, Maier, Wingate, de Gromard and Gessner2015, Smits et al. Reference Smits, Collins, Hand, Dutch and Payne2014, Spaggiari & Tyler Reference Spaggiari and Tyler2014, Howard et al. Reference Howard, Smithies, Kirkland, Kelsey, Aitken, Wingate, De Gromard, Spaggiari and Maier2015, Spaggiari et al. Reference Spaggiari, Kirkland, Smithies, Wingate and Belousova2015, Tucker et al. Reference Tucker, Hand, Kelsey and Dutch2015, Walsh et al. Reference Walsh, Kelsey, Kirkland, Hand, Smithies, Clark and Howard2015), prior geological knowledge on the Bunger Hills and HJA is very limited with interpretations based on few reconnaissance geochronological and metamorphic studies.
This study uses in situ U-Pb monazite geochronology and quantitative phase equilibria modelling to place new constraints on the metamorphic evolution of the largely undescribed HJA. We utilize thin sections from samples collected during the 1986 Australian Antarctic expedition to the Bunger Hills, Obruchev Hills and Denman Glacier region (Sheraton et al. Reference Sheraton, Tingey, Oliver and Black1995). The geographical remoteness and thus logistical challenges in accessing the HJA means that legacy samples provide one of the few enduring records of the original rock characteristics. Nevertheless, these samples yield substantial and useful information regarding the timing and pressure–temperature (P–T) conditions of metamorphism in the HJA.
Geological setting
The HJA comprise a narrow belt of small, exposed rocky islands situated north of the Bunger Hills in East Antarctica (~100°E, 66°S; Fig. 1). The HJA spans ~93 km, varies from 9–23 km in width and extends north-eastwards from the Taylor Islands to the Southern Ocean. Few studies document this region with previous geological investigations based on only two Antarctic expeditions (1956–57 and 1986), as primarily recorded by Ravich et al. (Reference Ravich, Klimov and Solovʹev1968) and Sheraton et al. (Reference Sheraton, Tingey, Oliver and Black1995).
The rocks of the Bunger Hills and HJA consist of predominantly tonalitic–granitic orthogneiss with lesser mafic granulite, interlayered orthogneiss–pelite and migmatitic garnet–sillimanite–cordierite-bearing metasedimentary rocks (Sheraton et al. Reference Sheraton, Tingey, Black and Oliver1993, Reference Sheraton, Tingey, Oliver and Black1995). Sheraton et al. (Reference Sheraton, Black and Tindle1992) date the age of the igneous protoliths to felsic orthogneiss from the south-western Bunger Hills and Thomas Island to 1699±15 and 1521±29 Ma, respectively. Tonalitic orthogneiss from the Obruchev Hills, south-west of the Bunger Hills, is dated to the late Archean (c. 2640 Ma; Sheraton et al. Reference Sheraton, Black and Tindle1992). Within the Bunger Hills, three large intrusive bodies, the Paz Cove batholith, the Algae Lake pluton and the Booth Peninsula batholith, also crop out (Sheraton et al. Reference Sheraton, Black and Tindle1992, Reference Sheraton, Tingey, Black and Oliver1993, Reference Sheraton, Tingey, Oliver and Black1995). These voluminous plutonic rocks comprise gabbro, monzogabbro, monzodiorite and granite, and were emplaced between c. 1170–1150 (Sheraton et al. Reference Sheraton, Black and Tindle1992). Dolerite dykes were emplaced at c. 1140 Ma, involving at least four chemically distinct suites, and alkaline mafic dykes intruded at c. 500 Ma (Sheraton et al. Reference Sheraton, Black, McCulloch and Oliver1990).
The definitive structural history of the Bunger Hills–HJA region is debated; however, it is considered to have undergone at least three deformation events (Stüwe & Wilson Reference Stüwe and Wilson1990, Ding & James Reference Ding and James1991, Sheraton et al. Reference Sheraton, Tingey, Black and Oliver1993, Reference Sheraton, Tingey, Oliver and Black1995, Harris Reference Harris1995). The earliest recognized event is characterized by a layer-parallel S1 gneissic fabric (Sheraton et al. Reference Sheraton, Black and Tindle1992) ascribed to crustal extension (Stüwe & Wilson Reference Stüwe and Wilson1990). D2 is characterized by pervasive, tight to isoclinal and asymmetric F2 folding (Ding & James Reference Ding and James1991, Sheraton et al. Reference Sheraton, Black and Tindle1992) interpreted to reflect compression (Stüwe & Wilson Reference Stüwe and Wilson1990). D2 structures were reoriented during regional scale D3 open and upright folding (Ding & James Reference Ding and James1991, Sheraton et al. Reference Sheraton, Black and Tindle1992). The age of emplacement of the major plutonic bodies relative to deformation is the main source of contention. Though generally, intrusion of the Paz Cove batholith and Algae Lake pluton are considered synchronous with the last major stage of deformation (Ding & James Reference Ding and James1991, Sheraton et al. Reference Sheraton, Black and Tindle1992). The Booth Peninsula batholith cross cuts D3 structures and is thus considered to have intruded post-D3 tectonism (Sheraton et al. Reference Sheraton, Black and Tindle1992). Subsequent, localized formation of shear zones, some coeval with mafic dyke emplacement, is also proposed (Sheraton et al. Reference Sheraton, Black and Tindle1992).
Previous studies utilizing conventional geothermobarometry estimated that peak metamorphic conditions were ~750–800°C at 5–6 kbar in the Bunger Hills (Stüwe & Powell Reference Stüwe and Powell1989). Slightly higher pressure conditions were proposed for the HJA (~700–750°C, 7–9 kbar) consistent with the observation of garnet within mafic granulite (Stüwe & Powell Reference Stüwe and Powell1989, Sheraton et al. Reference Sheraton, Tingey, Oliver and Black1995). The timing of granulite facies metamorphism in the Bunger Hills is dated at 1190±15 Ma from metamorphic zircon in one sample of tonalitic orthogneiss from Thomas Island (Sheraton et al. Reference Sheraton, Black and Tindle1992) and is suggested to have accompanied D1 deformation (Stüwe & Powell Reference Stüwe and Powell1989).
Despite the currently limited geochronological dataset from the Bunger Hills, many studies readily identify parallels between the timing of metamorphism and the emplacement age of Mesoproterozoic intrusive rocks in the Bunger Hills–HJA region with rocks in the Windmill Islands in East Antarctica and Stage 2 of the Albany–Fraser Orogeny in south-west Australia (e.g. Sheraton et al. Reference Sheraton, Tingey, Black and Oliver1993, Clarke et al. Reference Clarke, Sun and White1995, Harris Reference Harris1995, Clark et al. Reference Clark, Hensen and Kinny2000, Fitzsimons Reference Fitzsimons2000, Reference Fitzsimons2003, Boger Reference Boger2011). The Mesoproterozoic Albany–Fraser Orogeny is largely interpreted to record the final stages of assembly between the North and West Australian cratons with the combined South Australian Craton and East Antarctic Craton (Clark et al. Reference Clark, Hensen and Kinny2000, Betts & Giles Reference Betts and Giles2006, Cawood & Korsch Reference Cawood and Korsch2008). Specifically, the Albany–Fraser Orogeny has been characterized by coeval felsic and mafic magmatism and high temperature, high pressure metamorphism pertaining to continental collision between 1345–1260 Ma (Stage 1) followed by intracratonic reactivation involving craton-vergent thrusting, high temperature metamorphism and predominantly felsic magmatism between 1215–1140 Ma (Stage 2; Nelson et al. Reference Nelson, Myers and Nutman1995, Clark et al. Reference Clark, Hensen and Kinny2000, Bodorkos & Clark Reference Bodorkos and Clark2004, Kirkland et al. Reference Kirkland, Spaggiari, Pawley, Wingate, Smithies, Howard, Tyler, Belousova and Poujol2011). More recently, however, the Albany–Fraser Orogeny has been suggested to have involved oceanic arc accretion onto the West Australian Craton (Spaggiari et al. Reference Spaggiari, Kirkland, Smithies, Wingate and Belousova2015) and back-arc extensional tectonics (Clark et al. Reference Clark, Kirkland, Spaggiari, Oorschot, Wingate and Taylor2014, Spaggiari et al. Reference Spaggiari, Kirkland, Smithies, Wingate and Belousova2015), rather than simply continental collision and intracratonic orogenesis. Stages 1 and 2 of the Albany–Fraser Orogeny in south-western Australia are also expressed as the Mount West Orogeny (c. 1345–1293 Ma) and Musgrave Orogeny (c. 1220–1150 Ma), respectively, in the Musgrave Inlier within central Australia (Smithies et al. Reference Smithies, Howard, Evins, Kirkland, Kelsey, Hand, Wingate, Collins and Belousova2011, Kirkland et al. Reference Kirkland, Smithies, Woodhouse, Howard, Wingate, Belousova, Cliff, Murphy and Spaggiari2013, Howard et al. Reference Howard, Smithies, Kirkland, Kelsey, Aitken, Wingate, De Gromard, Spaggiari and Maier2015).
Sample descriptions
Samples utilized in this study comprise part of a suite of migmatized metasedimentary rocks collected during the 1986 Australian Antarctic research expedition to the Bunger Hills, Obruchev Hills and Denman Glacier region of East Antarctica and are archived at Geoscience Australia, Canberra. The three samples chosen for this study were collected from north of Raketa Island, Zabytyy Island and Geografov/Currituck Island in the HJA, directly north of the Bunger Hills (Fig. 1, Table I). These three samples are representative of the metapelitic lithologies that are interlayered with predominantly tonalitic orthogneiss within this region of the HJA. Whole rock geochemical analyses (reproduced from Sheraton et al. Reference Sheraton, Tingey, Oliver and Black1995; Table II) show a range from typical pelitic (sample 5607) to more felsic compositions (samples 6264 and 6251) that are more suggestive of impure sandstone–siltstone protoliths and/or sampling of leucocratic-rich domains of a migmatite. Sample names are referred to by their last four digits throughout the text. Sample location grid references are provided in Table I. The mineralogy, age and peak metamorphic conditions of each sample are also summarized in Table I.
Summary of samples.
Mineral abbreviations from Holland & Powell (Reference Holland and Powell1998).
Bulk compositions used for phase equilibria modelling.
* Reproduced from Sheraton et al. (Reference Sheraton, Tingey, Oliver and Black1995)
Sample 6264 (north Raketa Island)
Sample 6264 contains garnet, sillimanite, quartz, plagioclase, K-feldspar, ilmenite, and minor biotite and rutile. Garnet is coarse grained (up to 5 mm in length,~5% abundance), typically elongate and contains partially to completely enclosed inclusions of sillimanite (Fig. 2a). Included sillimanite crystals are aligned with the external fabric as also defined by aggregates of sillimanite. Cordierite is often located adjacent to and separating garnet grains (e.g. Fig. 2a & b) and exhibits relatively sharp grain boundaries (Fig. 2a). Sometimes small grains of garnet are completely enclosed within the cordierite. Sillimanite is often also wholly included within cordierite occurring adjacent to garnet (e.g. Fig. 2c). Cordierite is never included within garnet. In places, fine grained (lamellae <50 µm wide) cordierite–quartz symplectitic intergrowths also occur adjacent to garnet. The spatial relationship between garnet, cordierite and sillimanite is suggestive of the increased growth of cordierite following inferred peak growth of garnet and sillimanite. Rare, blocky rutile grains (~1 mm in length) occur in contact with sillimanite and are aligned with the fabric, but do not contain inclusions of sillimanite. Biotite is sparsely distributed and occurs as randomly oriented, anhedral grains (~200–1000 µm in length) along quartz–feldspar grain boundaries. The inferred peak mineral assemblage for this sample is garnet–sillimanite–ilmenite–rutile–plagioclase–K-feldspar–melt–quartz, with the post-peak development of cordierite after garnet and the inferred retrograde growth of biotite.
Representative thin section photomicrographs. a. Sample 6264: coarse, anhedral garnet grains are aligned with acicular sillimanite that defines the foliation. Cordierite grains (with included monazite showing yellow pleochroic halos) exhibit sharp grain boundaries and only occur adjacent to garnet. b. Sample 6264: anhedral cordierite bridging two garnet grains. Sillimanite is included within garnet and cordierite. c. Sample 6264: sillimanite completely hosted within cordierite. d. Sample 5607: coarse grained garnet is surrounded by idoblastic sillimanite, plagioclase and quartz. Ilmenite and rutile occur within the matrix and garnet grain edge. e. Sample 5607: anhedral ilmenite partially enclosing rutile. f. Sample 6251: cordierite occurs adjacent to coarse, anhedral garnet and in places completely surrounds smaller, relict garnet grains. Locally, cordierite occurs as symplectitic intergrowths with quartz. Minor spinel (<200 µm) is included within both the garnet and cordierite. Mineral abbreviations from Holland & Powell (Reference Holland and Powell1998).
Sample 5607 (Geografov Island)
Sample 5607 contains sillimanite, garnet, rutile, ilmenite, quartz, plagioclase and rare biotite. Garnet is abundant (~25%) as very coarse grained (up to 1 cm in diameter), anhedral poikioblasts which host idioblastic sillimanite, coarse rutile (up to 500 µm), ilmenite and quartz (Fig. 2d). The modal abundance of rutile increases toward the garnet grain edge with many grains only occurring as partial inclusions within the garnet. In general, rutile occurring within garnet is coarser grained than rutile within the matrix (typically <200 µm). Ilmenite is present throughout, but is typically more common external to garnet. Sometimes, ilmenite appears to form partial anhedral overgrowths about rutile (Fig. 2e). Sillimanite is ubiquitous and weakly aligned to define a coarse foliation. Rutile and ilmenite occur in direct contact with sillimanite, but are generally not included within individual sillimanite crystals. Biotite is rare, commonly occurring within fractures in garnet, and is accordingly inferred as being of retrograde origin. The peak metamorphic mineral assemblage for this sample is inferred to be garnet–sillimanite–ilmenite–rutile–plagioclase–melt–quartz.
Sample 6251 (Zabytyy Island)
Sample 6251 contains garnet, spinel, cordierite, quartz, K-feldspar, plagioclase, and minor biotite, sillimanite and ilmenite. Garnet grains are moderately coarse grained (up to 2 mm diameter,~5% abundance) and contain inclusions of minor amounts of dark green spinel (~100–200 µm in size,~1% abundance) and rarely sillimanite (up to 250 µm length,<1% abundance). Spinel grains are also sometimes hosted within cordierite. Commonly, cordierite is anhedral and occurs adjacent to garnet with sharp grain boundaries (Fig. 2f) and/or sometimes as symplectitic intergrowths with quartz (e.g. Fig. 3). Garnet grains are also sometimes wholly included within cordierite (Fig. 2f). Similar to sample 6264, the observed microstructural relationships between garnet and cordierite are suggestive of the increased growth of cordierite after garnet and that some garnet grains are now potentially partially replaced by cordierite. Minor aggregates of subhedral cordierite grains also occur away from garnet and are aligned with the coarse foliation and considered to be preserved prograde–peak grains. Plagioclase is moderately abundant, occurring with quartz and minor K-feldspar as the main matrix phases. Biotite is subhedral (grains up to 1 mm in length) with no clear preferred orientation. Grains are sparsely distributed and typically occur adjacent to ilmenite and at the grain boundaries of quartz and feldspar. The inferred peak mineral assemblage for this sample is garnet–cordierite–ilmenite–plagioclase–K-feldspar–melt–quartz. Post-peak mineral development is limited to the occurrence of randomly oriented biotite and the additional growth of cordierite after garnet. Spinel and sillimanite are interpreted to be potentially part of the prograde, but not peak, metamorphic mineral assemblage.
Qualitative compositional maps of aluminium (Al), iron (Fe), magnesium (Mg) and silica (Si) which highlight the symplectitic intergrowth between cordierite and quartz adjacent to garnet in sample 6251. Dark colours represent low elemental abundance; warm colours represent high abundance. The black scale bar represents 1 mm. Mineral abbreviations from Holland & Powell (Reference Holland and Powell1998).
Analytical methods
Pressure–temperature conditions
Chemical compositions of minerals from all samples were obtained using a Cameca SX51 electron microprobe at Adelaide Microscopy, The University of Adelaide, using a beam current of 20 nA and an accelerating voltage of 15 kV. Elemental X-ray maps of sample 6251 were obtained using the same instrumentation, a 15 kV accelerating voltage and a beam current of 200 nA.
Calculations for P–T were completed using THERMOCALC version 3.37 with the internally consistent updated thermodynamic dataset, ds62, of Holland & Powell (Reference Holland and Powell2011), for the geologically realistic system of NCKFMASHTO (Na2O–CaO–K2O–FeO–MgO–Al2O3–SiO2–H2O–TiO2–Fe2O3), and using the activity relationships of White et al. (Reference White, Powell, Holland, Johnson and Green2014). The bulk compositions used for P–T pseudosection calculations are based on the whole rock XRF geochemical analyses (Table II) of each sample from Sheraton et al. (Reference Sheraton, Tingey, Oliver and Black1995) recalculated to molar oxide percent.
A variety of published methods in metamorphic phase diagram literature are employed to estimate the H2O content in bulk compositions used in phase equilibria modelling (e.g. Kelsey & Hand Reference Kelsey and Hand2015 and references therein). In this study, the H2O content for each sample was approximated from the abundance and chemical compositions of hydrous minerals in the observed mineral assemblages. The sensitivity of the chosen H2O content on the phase equilibria, namely the influence of H2O on the temperature of the solidus, and the range of H2O contents required to develop the observed peak mineral assemblage, is also evaluated for each sample using a calculated Temperature–MH2O diagram (Fig. S1 found at http://dx.doi.org/10.1017/S095410201600033X). The Fe2O3 content for each sample was selected by an approximation of the abundance and stoichiometrically recast (Droop Reference Droop1987) electron microprobe compositions of Fe3+-bearing minerals in each sample. A Temperature–XFe2O3 (XFe2O3=Fe2O3/(FeO + Fe2O3)) diagram is also presented for each sample in Fig. S2 (found at http://dx.doi.org/10.1017/S095410201600033X). This allows for assessment of the change in stability of the preserved mineral assemblages with variation of the Fe oxidation state from all Fe existing as FeO to all Fe as Fe2O3, and a comparison with the selected Fe2O3 content estimated from mineral compositions and abundance. The bulk compositions used for P–T pseudosection calculations are provided in Table II. Compositional isopleths and modal proportions of minerals were calculated using software TCInvestigator (version 1; Pearce et al. Reference Pearce, White and Gazley2015; Fig. S3 found at http://dx.doi.org/10.1017/S095410201600033X) to further constrain the P–T conditions.
Laser ablation inductively coupled plasma mass spectrometry monazite geochronology
Monazite grains were imaged using a back-scattered electron (BSE) detector on a Phillips XL30 SEM. In situ monazite geochronology was undertaken using laser ablation inductively coupled plasma mass spectrometry (LA–ICP–MS) at Adelaide Microscopy, The University of Adelaide, following the methods of Payne et al. (Reference Payne, Hand, Barovich and Wade2008). U-Pb analyses were acquired using a New Wave 213 nm Nd–YAG laser coupled with an Agilent 7500cs ICP–MS under a He-ablation atmosphere. Analyses involved 40 s of background measurement, including 10 s of the laser firing with the shutter closed to allow for beam stabilization, followed by 40 s of ablation. Ablation was performed with a frequency of 5 Hz and a spot size of 12 µm. Isotopes 204Pb, 206Pb, 207Pb and 238U were measured for 10 ms, 15 ms, 30 ms and 15 ms, respectively.
The primary monazite standard MAdel (TIMS normalization data: 207Pb/206Pa=490.0 Ma, 206Pb/238U=518.4 Ma and 207Pb/235U=513.1 Ma; Payne et al. Reference Payne, Hand, Barovich and Wade2008, updated with additional TIMS data) was used to correct for elemental fractionation and mass bias of the monazite data using software ‘Glitter’ (Griffin et al. Reference Griffin, Powell, Pearson and O’reilly2008). Accuracy was monitored by repeat analysis of the in-house standard 94-222/Bruna NW (c. 450 Ma, Payne et al. Reference Payne, Hand, Barovich and Wade2008). Over the duration of the study, the MAdel analyses yielded a weighted mean age of 207Pb/206Pb=490.5±9.2 Ma, 206Pb/238U=517.2±3.2 Ma and 207Pb/235U=512.3±2.6 Ma (n=32), and secondary standard 94-222/Bruna NW yielded weighted mean ages of 207Pb/206Pb=451±15 Ma, 206Pb/238U=460±12 Ma and 207Pb/235U=459±12 Ma (n=12). Instrument drift was corrected for by bracketing every ten unknowns with standard analyses and applying a linear correction. Due to unresolvable interference of 204Hg on 204Pb, a common Pb correction was not applied; however, 204Pb was monitored to assess the common lead of each analysis and analyses were rejected if anomalous 204Pb was observed relative to background levels.
Results
Mineral chemistry
A summary of mineral compositions from all samples is provided in Table III. Full representative analyses for each mineral are presented in Table S1 (found at http://dx.doi.org/10.1017/S095410201600033X). The chemistry of select minerals is discussed as follows.
Summary of mineral chemistry.
Abbreviations: X alm=Fe2+/(Fe2++Mg+Ca+Mn), X py=Mg/(Fe2++Mg+Ca+Mn), X gr=Ca/(Fe2++Mg+Ca+Mn), X spss=Mn/(Fe2++Mg+Ca+Mn), X Fe=Fe2+/(Fe2++Mg), X An =Ca/(Ca+Na+K), X Or=K/(Ca+Na+K), X gah=Zn/(Zn+Mg+Al+Ti), X sp=Mg/(Fe2++Mg).
Garnet grains from sample 6251 show slight core–rim enrichment in X alm (Fe2+/(Fe2++Ca + Mg + Mn)), and generally, depletion in X py (Mg/(Fe2++Ca + Mg + Mn)) and X gr (Ca/(Fe2++Ca + Mg + Mn)) with values of 0.55–0.57, 0.36–0.39 and 0.025–0.038 obtained from garnet cores and values of 0.57–0.63, 0.31–0.37 and 0.029–0.033 from garnet rims, respectively. Garnet from samples 6264 and 5607 do not exhibit any appreciable core–rim compositional zoning. In sample 6264, X alm ranges between 0.54–0.57, X py between 0.39–0.42, X gr between 0.029–0.033 and X spss (Mn/(Fe2++Ca + Mg + Mn)) between 0.01–0.015. In sample 5607, X alm ranges between 0.60–0.65, X py between 0.30–0.35, X gr between 0.039–0.062 and X spss between 0.08–0.1.
Biotite in all samples is magnesian with X Fe (Fe2+/(Fe2++Mg)) values ranging between 0.21–0.25 in sample 6264, 0.22–0.24 in sample 5607 and 0.27–0.38 in sample 6251. Throughout all samples, TiO2 values are moderately high and range between 3.12–4.85 wt%. Biotite in samples 5607 and 6251 exhibit relatively low F contents, between 0.53–0.57 wt% and 0.56–0.96 wt%, respectively; F values range between 2.37–2.76 wt% in sample 6264.
K-feldspar is found in samples 6264 and 6251. K-feldspar is K-rich and Ca-poor with X an (Ca/(Ca + Na + K)) of 0–0.01 and X or (K/(Ca + Na + K)) of 0.80–0.92 in both samples. Plagioclase is present in all three samples and is Na-rich with X an values between 0.38–0.40, 0.41–0.51 and 0.27–0.33 in samples 6264, 5607 and 6251, respectively. X or values in plagioclase range from 0.01–0.06.
Cordierite occurs in samples 6264 and 6251. Cordierite is magnesian in composition with X Fe values between 0.15–0.20 and no apparent variation in composition between cores or rims.
Ilmenite has a low MnO content in all samples, ranging from 0–0.65 wt%.
Spinel occurs in sample 6251. Spinel contains a moderate amount of ZnO, ranging from 0.97–1.17 wt% and with a X gah (Zn/(Zn + Fe2++Mg + Mn)) value of 0.02. X sp (Mg/(Zn + Fe2++Mg + Mn)) ranges from 0.31–0.43 wt%, Cr2O3 varies between 0.19–1.04 wt% and MnO varies between 0.01–0.1 wt%.
Pressure–temperature conditions
T–XFe2O3 diagrams used for the determination of ferric iron in the bulk composition of each sample are shown in Fig. S2. Calculated P–T pseudosections for each sample are presented in Fig. 4a–c and are contoured for X gr and X Fe in garnet and modal proportions of relevant phases in Fig. S3. Superimposed pseudosection peak mineral assemblage stability fields for all samples are shown in Fig. 4d for comparison.
Pressure–temperature (P–T) pseudosections constructed for a. sample 6264, b. sample 5607 and c. sample 6251. Bulk compositions used for the calculation of each pseudosection are provided in Table II. The solidus is shown as a black dashed line. The stability field of the inferred peak mineral assemblage in each sample is outlined in bold. The approximate range of peak P–T conditions on the basis of X Fe and X gr compositional isopleths in garnet and garnet modal abundance is indicated by dashed white circles in b. and c. In a., the peak P–T conditions constrained from garnet compositional isopleths are shown as a white dashed circle; P–T conditions corresponding to observed garnet modal proportions are represented by the grey dashed circle. White arrows reflect the inferred P–T path constrained from microstructural mineral relationships. d. Summary of P–T conditions from all samples. The inferred peak mineral assemblage stability fields corresponding to each sample (shaded regions), P–T constraints from garnet compositional isopleths (solid open circles) and modal proportions where different (open dashed circle for sample 6264) are superimposed. Inferred P–T paths for each sample are shown. Cordierite-in lines from a.–c. are also shown as cordierite is used as a constraint on the P–T path. Sample 6264 is represented in yellow, sample 5607 in blue and sample 6251 in purple. Mineral abbreviations are from Holland & Powell (Reference Holland and Powell1998).
In sample 6264 (Fig. 4a), the garnet stability field is broad, extending to pressures above ~5 kbar over the range 700–1000°C. Decreasing garnet abundance with decreasing pressure is associated with an increase in X Fe in garnet and consistently low X gr (Fig. S3). Cordierite is present at pressures lower than ~4.5 kbar at 700°C and the range of stable pressures increases with temperature. The occurrence of rutile is restricted to pressures higher than ~7 kbar and temperatures >850°C. The interpreted peak metamorphic assemblage garnet–sillimanite–ilmenite–rutile–plagioclase–K-feldspar–melt–quartz for this sample is stable at >900°C and >7 kbar. These P–T conditions are in agreement with calculated garnet modal proportions and petrographic observations, which suggest ~5% garnet abundance in this sample. However, chemical compositions of X Fe and X gr from unzoned garnet cores suggest P–T conditions of ~850–950°C at ~6–7 kbar, occurring within the lower pressure field garnet–sillimanite–cordierite–ilmenite–plagioclase–K-feldspar–melt–quartz. This P–T estimate is inconsistent with the presence of rutile and the inferred absence of peak cordierite in this sample. Cordierite is inferred to have grown during post-peak metamorphism at the expense of garnet ± sillimanite (e.g. Fig. 2a & b). This relationship corresponds to a simultaneous increase in the modal proportion of cordierite and a decrease in garnet and sillimanite; thus it suggests that this rock evolved through a broadly decompressional retrograde P–T path (Figs 4a & S3).
In sample 5607 (Fig. 4b), rutile-bearing assemblages occur at pressures >7 kbar at 700°C with rutile stability contracting to higher pressures with increasing temperature. Ilmenite is present below 7.5 kbar at 700°C and occurs at higher pressures with increasing temperature, coexisting with rutile only over a narrow pressure range (~1 kbar). Garnet is present throughout the majority of the pseudosection. The inferred peak mineral assemblage garnet–sillimanite–ilmenite–rutile–plagioclase–melt–quartz for this sample exists over a narrow P–T range extending from ~7.5–8.5 kbar at 850°C, to >9.5 kbar at 1000°C. Modal proportions of garnet and X gr and X Fe compositions from unzoned garnet cores are in agreement with this P–T estimate. In this sample, a lower abundance of rutile is observed within the matrix relative to rutile included within garnet, and anhedral ilmenite overgrowths occur on some rutile grains (e.g. Fig. 2e). This reflects a decreasing modal proportion of rutile and an increase in the abundance of ilmenite following growth of peak metamorphic garnet. Comparison with calculated mineral modal proportions therefore also suggests a decompressional post-peak evolution for this rock.
In sample 6251 (Fig. 4c), garnet exists at pressures >3 kbar at 700°C and contracts to higher pressures with increasing temperature. Cordierite stability occurs below 7 kbar while sillimanite is restricted to pressures greater than ~5.5–6.5 kbar. Spinel is present at ~900°C at 5.5 kbar, with the range of stable pressures widening to between ~4–7 kbar at 1000°C. The inferred peak mineral assemblage garnet–cordierite–ilmenite–plagioclase–K-feldspar–melt–quartz exists over a small P–T range between ~830–900°C and 5.2–5.8 kbar. However, chemical compositions of X Fe and X gr from garnet suggest that peak conditions were ~850–950°C and 5.4–6.6 kbar, occurring at a slightly higher pressure than the inferred peak assemblage field, within the garnet–sillimanite–cordierite–ilmenite–plagioclase–K-feldspar–melt–quartz-bearing field or adjacent spinel-bearing, sillimanite–absent field. Calculated modal proportions and estimated abundance of garnet and peak cordierite in this sample are in agreement with the peak conditions estimated from the garnet compositional isopleths. Similar to sample 6264, the inferred growth of cordierite after garnet (e.g. Figs 2f & 3) is suggestive of a broadly decompressional retrograde P–T path for this sample. The absence of magnetite in this sample constrains this down-temperature P–T path to have tracked along a relatively shallow P–T trajectory, above ~5 kbar.
Monazite U-Pb LA–ICP–MS geochronology
In situ monazite LA–ICP–MS was undertaken on samples 6251, 6264 and 5607 (Table IV, Fig. 5a–d). Monazite is sparse in all samples, and when present, typically occurs within and along the grain boundaries of matrix minerals. Rarely, monazite is partially hosted within garnet. Representative BSE images of analysed monazites are given in Fig. 5e–j. A few grains exhibit indistinct internal core–rim style zonation; however, the majority of grains appear unzoned in BSE imagery. The monazite grains do not show any relationship between their microstructural location and age. Weighted average calculations use the 207Pb/206Pb age. Concordia diagrams for each sample and a probability density plot of the combined age data are presented in Fig. 5a–d. Collectively, analyses that are ≤5% discordant yield a weighted mean age of 1183±8 Ma (n=32/34, mean square weighted deviation (MSWD)=0.95).
In situ LA-ICP-MS monazite U-Pb geochronology. Data are presented on U-Pb concordia diagrams: a. sample 6264, b. sample 6251 and c. sample 5607. Red ellipses represent analyses from monazite included within garnet; black ellipses represent analyses from matrix monazite. Red and grey ellipses shown with a dashed outline represent analyses that are >5% discordant from monazite hosted in garnet and matrix monazite, respectively. Weighted mean 207Pb/206Pb ages are given for the combined concordant analyses from each sample. d. Histogram and probability density distribution plot for monazite age data from all three samples. Analyses from monazite grains included within garnet are shown in red; analyses from monazite included within matrix minerals are shown in blue. Collectively, the age data define a single population which yields a weighted mean 207Pb/206Pb age of 1183±8 Ma. e.–j. Representative BSE images of analysed monazite grains from all samples. Open black circles represent the approximate location of U-Pb analyses. Corresponding 207Pb/206Pb U-Pb ages are shown. Age uncertainties are given at the 1σ level.
Laser ablation inductively coupled plasma mass spectrometry U-Pb in situ monazite geochronology.
Abbreviations: g=monazite hosted within garnet, m=matrix monazite
In sample 6264, 15 analyses were done on nine grains within the matrix and two analyses from two grains hosted within garnet. Grains are ~40–100 µm in size and sub-rounded to elongate in shape. Some grains display subtle, patchy internal zonation visible in BSE imagery. Five analyses that are >5% discordant are excluded from weighted mean calculations. The remaining 12 analyses range between 1221±22 Ma and 1153±22 Ma and give a 207Pb/206Pb weighted average age of 1188±12 Ma (MSWD=1.13; Fig. 5a).
In sample 6251, 15 analyses were obtained from five grains located partially and completely within garnet and three analyses from three grains within the foliated matrix. Grains are commonly ~30–100 µm in diameter, sub-rounded, generally unzoned in BSE imagery and are commonly pitted or cracked. Two analyses that are >5% discordant (analysis 6251.5.1 and 6251.8.1) and one outlying, older concordant analysis (analysis 6251.6.1, 1276±23 Ma) are excluded from weighted mean calculations. The remaining analyses range between 1240±24 Ma and 1156±21 Ma (n=15) and yield a 207Pb/206Pb weighted mean age of 1186±11 Ma (MSWD=0.94; Fig. 5b).
In sample 5607, eight analyses were collected from five grains hosted within garnet and four analyses from one grain within the matrix. Grains are ~40–60 µm in size, largely unzoned in BSE imagery, anhedral in shape and are commonly pitted and fractured. Four analyses that are >5% discordant (analysis 5607.3.2, 3.4, 5.1 and 7.1) and one anomalously young, concordant analysis (analysis 5607.6.1, 1086±25 Ma) are excluded from weighted mean age calculations. The remaining seven analyses range between 1188±27 Ma and 1125±25 Ma and give a 207Pb/206Pb weighted average age of 1161±18 Ma (MSWD=0.73; Fig. 5c).
Discussion
Age of metamorphism
In situ U-Pb monazite ages obtained in this study range predominantly between c. 1240–1150 Ma and give a weighted average 207Pb/206Pb age of 1183±8 Ma. This mean age is within error of the age of metamorphism previously determined for the Bunger Hills from metamorphic zircon in tonalitic orthogneiss (Sheraton et al. Reference Sheraton, Black and Tindle1992). The timing of metamorphism within this study is also similar to the age of Mesoproterozoic metamorphism and magmatism recorded elsewhere within the MAFO (e.g. Clark et al. Reference Clark, Hensen and Kinny2000, Kirkland et al. Reference Kirkland, Spaggiari, Pawley, Wingate, Smithies, Howard, Tyler, Belousova and Poujol2011, Reference Kirkland, Smithies, Woodhouse, Howard, Wingate, Belousova, Cliff, Murphy and Spaggiari2013, Smithies et al. Reference Smithies, Howard, Evins, Kirkland, Kelsey, Hand, Wingate, Collins and Belousova2011, Tucker et al. Reference Tucker, Hand, Kelsey and Dutch2015, Walsh et al. Reference Walsh, Kelsey, Kirkland, Hand, Smithies, Clark and Howard2015). Previous studies have utilized SHRIMP zircon, monazite and rutile geochronology to ascertain that, specifically, the Mesoproterozoic Albany–Fraser Orogeny in south-western Australia was characterized by two discrete thermotectonic pulses at c. 1345–1260 Ma (Stage 1) and c. 1214–1140 Ma (Stage 2; Nelson et al. Reference Nelson, Myers and Nutman1995, Clark et al. Reference Clark, Hensen and Kinny2000, Kirkland et al. Reference Kirkland, Spaggiari, Pawley, Wingate, Smithies, Howard, Tyler, Belousova and Poujol2011). Tectonism of a similar age is recorded by felsic gneiss and charnokite (c. 1380–1180 Ma) from the Windmill Islands in East Antarctica (Post Reference Post2000, Zhang et al. Reference Zhang, Zhao, Liu, Liu, Hou, Li and Ye2012). Evidence for widespread felsic magmatism at c. 1345–1293 Ma (Mount West Orogeny) and c. 1220–1150 Ma (Musgrave Orogeny) also occurs within the Musgrave Inlier (Smithies et al. Reference Smithies, Howard, Evins, Kirkland, Kelsey, Hand, Wingate, Collins and Belousova2011, Kirkland et al. Reference Kirkland, Smithies, Woodhouse, Howard, Wingate, Belousova, Cliff, Murphy and Spaggiari2013, Howard et al. Reference Howard, Smithies, Kirkland, Kelsey, Aitken, Wingate, De Gromard, Spaggiari and Maier2015), with the latter also linked with pervasive high temperature to ultra-high temperature metamorphism (Tucker et al. Reference Tucker, Hand, Kelsey and Dutch2015, Walsh et al. Reference Walsh, Kelsey, Kirkland, Hand, Smithies, Clark and Howard2015).
The Bunger Hills, HJA and Windmill Islands in East Antarctica have been correlated along strike of the Nornalup Complex of the MAFO in Gondwanan tectonic reconstructions (e.g. Clarke et al. Reference Clarke, Sun and White1995, Harris Reference Harris1995, Fitzsimons Reference Fitzsimons2003). Stage 1 tectonism in the Nornalup Complex is recorded by the Malcolm Gneiss (Clark et al. Reference Clark, Hensen and Kinny2000) and the earliest evidence for Recherche Supersuite magmatism (1330±14 Ma; Nelson et al. Reference Nelson, Myers and Nutman1995). The Nornalup Complex is otherwise dominated by Stage 2 tectonism as evidenced by the emplacement of the Esperance Supersuite, amphibolite facies metamorphism of sedimentary rocks from Mount Ragged (1154±15 Ma; Clark et al. Reference Clark, Hensen and Kinny2000) and granulite facies metamorphism of the garnet–cordierite-bearing Salisbury Gneiss (c. 1215 Ma and c. 1180 Ma; Clark et al. Reference Clark, Hensen and Kinny2000). Geometric considerations suggest that the Stage 1 tectonothermal events recognized in the Windmill Islands and Nornalup Complex, therefore, possibly also affected the Bunger Hills and HJA. However, the results from this study and those obtained previously by Sheraton et al. (Reference Sheraton, Black and Tindle1992) suggest that the Bunger Hills records only the second stage of the Albany–Fraser Orogeny, with magmatism and metamorphism occurring between c. 1190–1150 Ma. More geochronological data is required to determine conclusively whether the apparent absence of c. 1300 Ma tectonism in the Bunger Hills–HJA region is geologically important or if it is merely a reflection of a current paucity in age data.
New metamorphic constraints for the HighJump Archipelago
The rocks of the HJA experienced intermediate pressure, granulite facies metamorphism (~850–950°C, 6–9 kbar; Fig. 4d). The highest pressure conditions are recorded by a sample of garnet–sillimanite–rutile-bearing gneiss (sample 5607) from Geografov Island (~850–950°C, 8–10 kbar). The metamorphic conditions recorded by samples 6264 and 6251 from north of Raketa Island and Zabytyy Island, respectively, are inferred to have occurred at a slightly lower pressure relative to sample 5607 (Fig. 4d). However, the exact peak metamorphic conditions of these two samples are less assuredly constrained due to some inconsistencies between the calculated P–T pseudosections, measured garnet compositions and petrographic observations. The inferred peak mineral assemblage in sample 6264 is stable at P–T conditions >900°C and >7 kbar, with estimated garnet abundance (~5% abundance) further constraining P–T conditions to the lower pressure region of this field (~7–8 kbar). Considering that some garnet grains are potentially now partially replaced by cordierite, this P–T estimate is in effect a minimum with peak P–T conditions probably occurring at an even higher pressure. Therefore, on the basis of garnet modal proportions, the lower pressure boundary of the inferred peak mineral assemblage field defines the minimum limit on the peak metamorphic conditions. In contrast, measured X Fe and X gr garnet chemical compositions from sample 6264 suggest that peak metamorphic conditions were 850–950°C and 6–7 kbar, within the adjacent cordierite-present field. Given the relatively small size of garnet grains in this sample (typically 1–5 mm) and the ability of garnet to be compositionally reset under even short-lived granulite facies metamorphism (e.g. Spear Reference Spear1991, Caddick et al. Reference Caddick, Konopásek and Thompson2010), it is probable that these garnets now record a modified composition resulting from diffusion during cooling from slightly higher pressure, peak metamorphic conditions. This interpretation is consistent with the stability of the observed (cordierite-absent) peak mineral assemblage occurring at higher pressure.
In sample 6251, measured X Fe and X gr compositions in garnet suggest that peak metamorphic conditions were ~850–950°C and 5.4–6.6 kbar. However, these P–T conditions correspond to garnet–cordierite–spinel-bearing and/or garnet–sillimanite–cordierite-bearing assemblages. This is inconsistent with the inference that the observed spinel and sillimanite do not form part of the peak mineral assemblage. Spinel is present as a minor phase (~1% abundance) and always occurs as inclusions within garnet or hosted within cordierite that is inferred to have replaced former peak metamorphic garnet grains. Therefore, spinel is interpreted to have potentially formed part of the prograde mineral assemblage. However, as this study does not account for the possibility that the rocks of the HJA underwent melt loss through melt reintegration modelling (e.g. Korhonen et al. Reference Korhonen, Saito, Brown and Siddoway2010, Anderson et al. Reference Anderson, Kelsey, Hand and Collins2013), the presence or absence of spinel cannot be used reliably to make inferences about the prograde P–T path. Furthermore, electron microprobe mineral chemistry indicates that spinel contains a moderate amount of ZnO (0.97–1.17 wt%; Tables III & S1). Components such as zinc are known to enhance spinel stability to lower temperatures and higher pressures (e.g. Nichols et al. Reference Nichols, Berry and Green1992, Tajčmanová et al. Reference Tajčmanová, Konopásek and Košler2009). However, the current absence of thermodynamic models for Zn-bearing solid solutions limits the phase equilibria modelling approach because this P–T shift cannot be effectively quantified. Thus, the presence or absence of spinel cannot be reliably used as a constraint on the P–T conditions for this sample.
Although sillimanite is not interpreted to be part of the peak metamorphic assemblage, the calculated abundance of sillimanite is relatively low within the P–T region estimated from compositional isopleths (~1–2% abundance). Therefore, at these conditions, any peak sillimanite would probably not be readily observed in a thin section. Sillimanite does occur as rare inclusions within garnet, but not as a noticeable matrix phase. This microstructural relationship is suggestive that sillimanite grew as part of an earlier and probably higher pressure, prograde assemblage, and therefore, supports the notion that this sample evolved along a broadly clockwise P–T evolution. However, as for spinel, the presence of sillimanite cannot be used to definitively constrain the absolute prograde P–T trajectory for this sample.
The new P–T determinations presented in this study suggest that overall the HJA attained slightly higher temperature and pressure conditions than previously obtained by Stüwe & Powell (Reference Stüwe and Powell1989). Previous P–T estimates constrained using conventional thermobarometry were ~750–800°C at 5–6 kbar for the Bunger Hills and ~700–750°C at 7–9 kbar for the HJA (Stüwe & Powell Reference Stüwe and Powell1989, Sheraton et al. Reference Sheraton, Tingey, Oliver and Black1995). The slightly higher pressure conditions proposed for the HJA compared with the Bunger Hills are consistent with the observation of garnet in mafic granulite in the north-eastern HJA (Sheraton et al. Reference Sheraton, Tingey, Oliver and Black1995). However, the previous constraint on peak temperature for the HJA is likely to be unreliable as the garnet utilized for those P–T calculations occurred as a coronae about orthopyroxene and was accordingly interpreted to have been of retrograde origin (Sheraton et al. Reference Sheraton, Tingey, Oliver and Black1995). The P–T calculations done using conventional thermobarometry rely heavily upon individual, measured mineral compositions, which, as in this example, may not be truly indicative of the compositions that were formed at the peak of metamorphism if the minerals underwent diffusion as is typical at granulite facies.
Stüwe & Powell (Reference Stüwe and Powell1989) initially proposed that rocks of the Bunger Hills underwent an anticlockwise P–T evolution. This assertion was made on the basis that relict spinel was commonly surrounded by a cordierite, garnet and/or sillimanite coronae which ostensibly reflected the reaction between a formerly more widespread, low pressure (~800°C, 4 kbar) spinel–quartz-bearing assemblage to produce a higher pressure, garnet–sillimanite–cordierite-bearing assemblage (~750°C, 6–7 kbar; Stüwe & Powell Reference Stüwe and Powell1989). However, Sheraton et al. (Reference Sheraton, Tingey, Oliver and Black1995) observed in some metapelitic rocks that spinel and quartz commonly formed a stable relationship, often occurring in contact with one another. Although only volumetrically minor (~1% abundance), Sheraton et al. (Reference Sheraton, Tingey, Oliver and Black1995) also noted the pervasiveness of spinel throughout the metapelitic rocks and thus question the validity of the notion of subsequent overprinting by a high pressure assemblage. Furthermore, the significance of spinel on P–T determinations relies heavily on its ZnO content; as the Zn content in spinel is unreported by previous workers, earlier interpretations on the metamorphic evolution made on the basis of the presence of spinel cannot be considered overly robust.
Cordierite is present in samples 6264 and 6251. Typically, cordierite occurs adjacent to and bridges anhedral garnet grains (e.g. Fig. 2a & b) and sometimes completely isolates relict garnet grains (e.g. Fig. 2f). Cordierite occurring near garnet also commonly hosts sillimanite in sample 6264 (e.g. Fig. 2b & c). Sillimanite is also often included within garnet, but garnet never hosts cordierite. Therefore, the spatial relationship between garnet and cordierite (and sillimanite in sample 6264) suggests that the abundance of cordierite increased following the growth of garnet and sillimanite at the inferred peak metamorphic conditions. Decreasing garnet and sillimanite modal abundance and increasing cordierite modal abundance occurs along a down-pressure, down-temperature trajectory (e.g. Fig. S3). Furthermore, rare cordierite–quartz symplectites also occur adjacent to garnet (e.g. Fig. 3). Previous studies have ascribed cordierite–quartz symplectites to garnet breakdown and/or the melt assisted breakdown of feldspar in the presence of a Fe–Mg phase (Henry Reference Henry1974, Mohan & Windley Reference Mohan and Windley1993, Nandakumar & Harley Reference Nandakumar and Harley2000); however, the occurrence of these intergrowths typically adjacent to garnet in this instance is suggestive of the former interpretation. In this study, evidence for cordierite growth after garnet in samples 6264 and 6251, and rare growth of ilmenite about rutile, suggests that, in contrast to the initial interpretations of Stüwe & Powell (Reference Stüwe and Powell1989), the rocks of the HJA record evidence for a clockwise (decompressional) P–T path. The absence of magnetite in sample 6251 further suggests that the trajectory of this down-temperature P–T path may have been relatively shallow. This conclusion is supported by another sample of metapelite located north of Raketa Island (sample 86286263) reported by Sheraton et al. (Reference Sheraton, Tingey, Oliver and Black1995) to show garnet partially replaced by orthopyroxene–cordierite–plagioclase symplectites, a microstructural relationship that has been previously ascribed to a decompressional P–T evolution (e.g. Brandt et al. Reference Brandt, Klemd and Okrusch2003, Tong et al. Reference Tong, Liu, Wang and Liang2014). Sheraton et al. (Reference Sheraton, Tingey, Oliver and Black1995) also report on garnet occurring as symplectitic rims about orthopyroxene in granite from the Paz Cove batholith and orthogneiss from the north-east HJA (sample 86285979), inferred to have developed as a result of near-isobaric cooling.
Re-examination of the HJA P–T conditions in this study suggests that the metamorphic evolution of these rocks is incompatible with the tectonic model proposed initially by Stüwe & Powell (Reference Stüwe and Powell1989) involving extension followed by compressional deformation, although Harris (Reference Harris1995) note that this initial interpretation was also subsequently re-evaluated. If the rocks underwent thickening post-extension then an anticlockwise P–T path would be expected (e.g. Brown Reference Brown1993). In contrast, the retrograde evolution of the HJA presented here is characterized by a pressure decrease during cooling from a metamorphic peak that occurred at a high temperature and medium pressure. This scenario appears more consistent with the notion of thinning of over-thickened crust. However, there is also currently little documented structural evidence for extensional thinning in the Bunger Hills and HJA, and therefore, the drivers for metamorphism and the down-pressure retrograde evolution proposed in this study remain unclear.
A two-stage tectonothermal history has been proposed for the Windmill Islands involving upper amphibolite facies metamorphism (>750°C, 4 kbar) at c. 1340–1300 Ma, and subsequently granulite facies metamorphism (~850–900°C, 5–7 kbar) at c. 1240–1140 Ma (Post Reference Post2000). The latter phase, coincident with Stage 2 of the Albany–Fraser Orogeny and high temperature metamorphism in the HJA, is inferred to have involved a post-peak clockwise P–T–time path, interpreted to reflect exhumation and unloading of the terrane following crustal thickening during continental collision (Post Reference Post2000). Similarly, peak granulite facies metamorphism (800°C,>5 kbar) is also recorded by the Salisbury Gneiss in the Nornalup Complex at c. 1214±8 Ma with zircon rims dating to 1182±13 Ma, suggestive of decompression from peak metamorphic conditions by exhumation along thrust zones (Clark et al. Reference Clark, Hensen and Kinny2000). Although peak metamorphism of the HJA occurred during Stage 2 of the Albany–Fraser Orogeny, the P–T conditions attained are also similar to those recorded at c. 1290 Ma by the Fraser Zone within south-western Australia (~850°C and 7–9 kbar; Clark et al. Reference Clark, Kirkland, Spaggiari, Oorschot, Wingate and Taylor2014). These rocks are interpreted to reflect the mid-crustal emplacement of mafic magmatism into a back-arc environment followed by near-isobaric cooling at depth (~9 kbar,~27 km depth; Clark et al. Reference Clark, Kirkland, Spaggiari, Oorschot, Wingate and Taylor2014). A similar scenario for the HJA would justify the elevated temperatures and medium pressures attained during peak metamorphism but does not necessarily explain the inferred clockwise P–T evolution. Geochronological data from the Bunger Hills currently suggests that magmatism also occurred following the main phase of metamorphism (Sheraton et al. Reference Sheraton, Black and Tindle1992).
Conclusions
Legacy thin section samples provide a valuable record of the geological characteristics of rocks of the HJA that are located in a pivotal yet largely inaccessible part of the MAFO. Metasedimentary rocks from the HJA suggest that this region underwent granulite facies metamorphism (~850–950°C and 6–9 kbar) at c. 1183 Ma. The inferred growth of cordierite after garnet further suggests that these rocks cooled along a decompressional P–T trajectory following peak metamorphic conditions. The timing of metamorphism determined in this study confirms the age previously constrained for the Bunger Hills–HJA region. However, the new P–T determinations suggest that peak metamorphic conditions within the HJA attained slightly higher temperatures and pressures than previously proposed. In contrast to the initial interpretations from reconnaissance P–T work, the results of this study also suggest that the rocks evolved along a clockwise rather than anticlockwise P–T path.
Acknowledgements
This work forms part of Australian Antarctic Science Project 4191. I. Fitzsimons and F. Korhonen are thanked for their thorough and constructive reviews of the manuscript. We gratefully acknowledge Geoscience Australia, and in particular Chris Carson, for the use of the samples reported in this study. B. Wade and A. McFadden, Adelaide Microscopy, are thanked for their support with analytical instrumentation. NMT acknowledges the support of an Australian Postgraduate Award and Playford Trust PhD Scholarship.
Author contribution statement
NMT was primarily responsible for data collection, data analysis and preparation of the manuscript. Both authors were involved in the original paper concept and design, and subsequent drafting and revision of the manuscript.
Supplementary Material
Three supplemental figures and a table will be found at http://dx.doi.org/10.1017/S095410201600033X.
