A comparison between phase diagram modelling of metamafic rocks and experimental and independent thermobarometric data

The use of equilibrium phase diagrams such as pseudosections has become widespread in recent years, but few studies tested the phase relations predicted by these models. This work aims at comparing predicted phase relations with mineral assemblages observed in experimental products and in natural samples, in this latter case using as reference metamorphic conditions pressures and temperatures determined for metapelites metamorphosed under the same conditions as the mafic rocks. Pseudosections were calculated with THERMOCALC for six mafic bulk-rock compositions, in the range 425-700 degrees C and 2-8 kbar. The resulting diagrams can be divided into two groups: one in which a classic sequence of mineral assemblages occur and another where quartz is unstable and diopside is stable throughout. The results compare well to the independent data in three of the six studied cases. The main inconsistencies found in the pseudosections are underestimation of the stability field of garnet and overestimation of the stability fields of diopside and hornblende, which is at odds not only with data specific to the rocks studied here, but with the literature on metamafic rocks in general. In the case of garnet, the most likely case of discrepancy is the absence of Mn in the activity-composition models available. On the other hand, the issue with diopside and hornblende is much harder to evaluate, as it may be related to any of the activity-composition models, or to relationships between two or more of them. Nonetheless, we can affirm that the overestimation of the diopside stability field is related to SiO2 content, although there are other factors that influence it. (C) 2019 Elsevier B.V. All rights reserved. (AU)