Metamorphism of metachert from the Southern Alps, New Zealand. A thermodynamic forward modelling study

Scattered, scarce occurrences of garnet- and quartz-rich metamorphic rock, probably derived from Mn- and Fe-rich chert, occur within metamorphosed greywacke sequences worldwide. The metamorphism of such garnetiferous metacherts has not previously been investigated using modern thermodynamic forward modelling techniques due to the lack of appropriate, internally-consistent activity-composition (a–x) models for Mn-bearing minerals. The present study applies thermodynamic forward modelling using the recently-proposed a–x models of White et al. (2014) to investigate the metamorphism of garnetiferous metachert samples from the Southern Alps, New Zealand. Pressure-temperature (P–T) pseudosections are used in combination with results from petrography, element composition mapping using micro X-ray fluorescence (µXRF) and scanning electron microscope (SEM) methods, and garnet composition data from analytical transects by electron probe microanalysis (EPMA), to study metachert metamorphism. All the samples are compositionally layered, so the possibility exists that an input bulk rock composition might not match the effective bulk composition at the site of garnet growth. If a mineral assemblage stability field in a calculated P–T pseudosection matched the mineral assemblage in the rock, this was taken as an initial indication of a permissible input bulk rock composition. In that case, refined constraints on the P–T conditions were sought by comparing calculated and measured garnet compositions. The studied rocks include samples that are carbonate-bearing, which require consideration of the effects of fluid composition in mixed H₂O–CO₂ fluids, as well as a sample in which the garnet is strongly zoned, texturally-complex, and inferred to be of polymetamorphic origin. The effects of element fractionation by that garnet were investigated by recalculating the P–T pseudosection using a new bulk rock composition with the garnet core content removed. In none of the samples did the calculated and observed composition isopleths for the garnet cores match, suggesting that initial garnet nucleation in these Mn-rich rocks was locally controlled. For most samples in which the calculated and observed mineral assemblages matched, successful estimates of the peak metamorphic conditions were obtained. A garnet chert (A12E) from the mylonite zone of the Alpine Fault at Vine Creek, near Hokitika, gave a tight intersection of composition isopleths, indicating peak metamorphic conditions of 510 °C/5.5 kbar, after recalculation to correct for element fractionation by the strongly-zoned garnet. This tight, modern constraint is within error of previously-reported results from traditional geothermobarometry (420–600 °C/5.9–13 kbar) and Raman spectroscopy of carbonaceous material (RSCM T = 556 °C) from nearby sites. A peak metamorphic estimate of 520–550 °C/7–10 kbar was obtained from a dolomite-bearing sample from the garnet zone near Fox Glacier (J34), in good comparison with published temperatures from Raman spectroscopy of carbonaceous material in nearby metagreywacke samples (526–546 °C). The prograde metamorphic P–T path was probably steep, based on growth of the garnet core at ~475535 °C/5–9 kbar. The successful results for these garnet chert samples show that the new a-x models for Mn-bearing minerals extend the range of rock types that are amenable to pseudosection modelling. Results obtained in this study also serve to highlight several possible concerns: a) garnet nucleation and initial growth in very Mn-rich rocks may be subject to local compositional or kinetic controls; b) bulk rock compositions may not always mimic the effective bulk composition; c) the existing a–x models for Mn-bearing minerals and white micas may need refining; and d) some rocks may simply be ill-suited to thermodynamic forward modelling. Items a) and b) may be indicated by the common observation of a mismatch between predicted and measured garnet composition isopleths for garnet cores, and by a mismatch between garnet composition isopleths and the appropriate mineral assemblage field for sample AMS01, from the mylonite zone, Hari Hari, Southern Alps. For item c) every P–T pseudosection calculated using the new a–x models for Mn-bearing minerals showed garnet stable to very low temperatures below 300 °C. In addition, the P–T pseudosection for an oligoclase-zone metachporphyroblasts of Fe-Ti oxides (magnetitert (Sample J36) from Hari Mare stream, Franz Josef - Fox Glacier, indicated that the white mica margarite should be present instead of plagioclase (oligoclase), for a rock in which oligoclase is present and margarite is absent, a problem previously noted elsewhere. Item d) is exemplified by a very garnet-rich ferruginous metachert sample (J35, garnet zone, headwater region, Moeraki River, South Westland) which proved impossible to model successfully due to its complex mineral growth and deformation history. This sample contained multiple generations of carbonate with differing compositions, amphibole (not incorporated for modelling with the new a–x models for Mn-bearing minerals), large e associated with smaller, possibly later-formed ilmenite), and the garnet bands were offset by late deformation. The garnetiferous metachert samples studied here preserve in their textures and compositions clues to their growth mechanism and metamorphic history. The textures in at least two of the samples are consistent with the diffusion controlled nucleation and growth model for garnet. This research has successfully used state of the art thermodynamic modelling techniques in combination with the latest internally consistent a-x models on Mn-rich metachert, for the first time, extracting P–T conditions of the metamorphism of garnetiferous metachert from the Southern Alps.