Melt migration vs. isentropic decompression melting, more or less

Abstract

Decompression melting of the asthenospheric mantle is routinely modelled as an isentropic process, but as thermodynamic models evolve to the point where this constraint can be rigorously (and perhaps accurately) imposed, it becomes important to test the quality of this assumption. In particular, as McKenzie (1984) recognised, when melt migration occurs the upwelling is neither adiabatic nor reversible due to advection of heat by the melt and gravitational dissipation. McKenzie estimated that these effects were negligible relative to the rather large uncertainties in his melting parameters, but concluded that melt migration leads to excess melting above that generated in the isentropic case, by at most 60%. However, relative movement of melt and solids has another first-order effect - the possibility of introducing chemical disequilibrium, or something akin to fractional melting. It is well known that fractional melting processes are less productive in general than equilibrium melting, so this raises the possibility that melt migration leads to substantially less melting and crustal production than the isentropic equilibrium case. In order to settle which of these effects is dominant and to bound the magnitude of both effects, I construct an energy equation similar to McKenzie's that allows for chemical and thermal disequilibrium between migrating melts and residues and perform a series of one-dimensional MELTS calculations that show the net melt production of various cases.

Additional details