A model for disequilibrium mantle melting incorporating melt transport by porous and channel flows

Abstract

ATTEMPTS to understand melt transport in the Earth's mantle have focused on the two mechanisms of porous flow along grain boundaries1–4 and channel flow in fractures5–10. To elucidate the contributions of these segregation processes to the chemical evolution of melt and solid, I have developed a disequilibrium melting model for a one-dimensional upwelling mantle in which melt is produced, transported by porous flow and sucked into chemically isolated channels. The model is based on recent disequilibrium melting models11,12, in which the residual melt in the porous flow system need not equilibrate with the residual solid. By changing the parameter values such as the rates of chemical equilibration and melt removal into channels, this model predicts trace element concentrations for various melting conditions, including disequilibrium effects which cannot be assessed quantitatively using previous models for dynamic mantle melting1,6,13–15. Comparison between the model results and compositions of abyssal peridot!tes16 supports the idea that, if equilibrium melting is achieved, imperfect fractional melting (corresponding to 80% melt removal) occurs beneath mid-ocean ridges17,18. If disequilibrium melting is important, however, more efficient melt removal (in other words, chemical isolation from the residue) is required to explain the peridotite data.