P-T PATHS FROM ZONED GARNETS - SOME MINIMUM CRITERIA

There are at least two processes that can lead to Ca zoning in garnets: changes in P or T and fractionation of Ca between garnet and rock matrix. Ordinarily, these are linked through equilibrium fractionations between garnet and other minerals(s) in the progress of particular reactions. Computer simulation of garnet growth in the KNCFMASH reaction: staurolite + muscovite + (Ca, Na) plagioclase + quartz = (Ca, Mg, Fe) garnet + biotite + kyanite + Na-richer plagioclase + H2O shows that with very limited diffusion of Fe, Mg, and Ca, garnets zoned in Ca can grow by isothermal, isobaric processes through fractionation of Ca into earlier garnet cores. Our "synthetic" model garnets have Ca-richer cores and Ca-poorer rims, and their simulated growth is accompanied by depletion of Ca from the matrix reservoir, plagioclase. Zoning patterns displayed by these garnets mimic those found in natural garnets inferred to have grown in the above reaction during an episode of decompression. Unless the contribution of fractionation to garnet zoning can be constrained, the zoning pattern cannot be used to determine uniquely the P-T history during garnet growth. Three minimal conditions must be met before the zoning pattern of a garnet can be used for relative thermobarometry: (1) the assemblage or assemblages present at all stages of garnet growth must be known; (2) the changes in composition of phases participating in the garnet-producing reaction, particularly biotite and plagioclase, must be known; and (3) any modification of growth zoning by post-crystallization diffusive processes must be understood. In addition, any P-T changes indicated by garnet zoning alone should be viewed with caution and corroborated by textural or structural evidence, if possible.