a question - H2O in the MELTS caculation

[xuyingkui]

Hi, I came to a question when I used windows adiabat_1ph_3-0. The question is that: I input initial H2O=2.5 in the melts file (andesite.melts), after caculation(magma fractionation process), and I find the new produced file Liquid_comp_tbl.txt, open it and the HO2 has been changed to 1.78 while the temperature is high than the magma liquidus at that pressure, while in the bulk_comp_tbl, H2O is still 2.5. I was wandering if it is ture that adiabat can caculate the solubility of H2O at T and P automatically, if the initial H2O seted higher than the solubility, adiabat will throw the portion away directly.
       Thanks for your reply!   -xuyingkui

[Paula]

Hi xuyingkui,

Yes! Adiabat_1ph should not have actually thrown away the excess H2O but rather exsolved a vapour phase ('water'). In the MELTS model, the 'solid' phases are everything that is not silicate melt, including the pure water phase.  So you should find the missing 0.72 H2O in the Solids_main_tbl.txt (this file is written before any fractionation processes).  For the purposes of fractional crystallisation, water is treated separately from the other solid phases: ADIABAT_FRATIONATE_SOLIDS removes all phases except water and ADIABAT_FRACTIONATE_WATER removes any water phase, but not dissolved H2O. So setting ADIABAT_FRACTIONATE_WATER with a deliberately high initial H2O content is a quick and dirty way to calculate the maximum solubility of H2O in the liquid for a given temperature. (Indeed this is the recommended way to perform water saturated 'amoeba' calculations as the aH2O buffer and amoeba routine do not work together very well.)

Paula