A model for garnet and plagioclase growth in pelitic schists: implications for thermobarometry and P‐T path determinations

Abstract

Numerical models of the progressive evolution of pelitic schists in the NCMnKFMASH system with the assemblage garnet + biotite + chlorite ± staurolite + plagioclase + muscovite + quartz + H2O are presented with the goal of predicting compositional changes in garnet and plagioclase along different P‐T paths. The numerical models support several conclusions that should prove useful for interpreting the P‐T paths of natural parageneses: (i) Garnet may grow along P‐T vectors ranging from heating with decompression to cooling with compression. P‐T paths deduced from garnet zoning that are inconsistent with these growth vectors are self‐contradictory. (ii) There is a systematic relation between garnet and plagioclase composition and growth such that for most P‐T paths, garnet growth requires plagioclase consumption. Furthermore, mass balance in a closed system requires that as plagioclase is consumed the remaining plagioclase becomes increasingly albitic. Inclusions of plagioclase in the core of garnet should be more anorthitic than those near the rim and zoned matrix plagioclase should have rims that are more albitic than the cores. Complex plagioclase textures may arise from the local variability of growth and precipitation kinetics. (iii) A decrease of Fe/(Fe + Mg) in a garnet zoning profile is a reliable indicator of increasing temperature for the assemblage modelled. However, there is no single reliable ΔP monitor and inferences about ΔP can only be made by considering plagioclase and garnet together. (iv) Consumption of garnet during the production of staurolite removes material from the outer shell of a garnet and may make recovery of peak metamorphic compositions and P‐T conditions impossible. Low ‘peak’temperatures typically recorded by staurolite‐bearing assemblages may reflect this phenomenon. (v) Diffusional homogenization of garnet affects the computed P‐T path and results in a clockwise rotation of the computed P‐T vector relative to the true P‐T path.