H2O-fluxed melting buffers crustal temperatures and stabilizes magmatic mushes

Many conceptual models in igneous petrology and volcanology involve the protracted presence of large volumes of magmatic mushes in the crust. We used 1D numerical simulation to explore how the inflow of aqueous fluids into a section of crust facilitates melting and stabilizes columns of mush.

Inflow of H₂O into a crustal section whose temperatures are above the water-saturated solidus results in the following chain of processes: it induces melting; melting consumes latent heat; latent heat consumption lowers temperatures; reduced temperatures cause an increase in heat flow into the melting region. In detail, the behaviour of the upward migration of the water-fluxed melting front depends on the relative ratios between heat and H₂O diffusivities.  The upwards flow of H₂O is accompanied by melting until the H₂O front reaches the water-saturated solidus isotherm.  If the transfer of H₂O through chemical diffusion enhanced by advection (effective diffusivity) is slower than the transfer of heat, the melting front progress smoothly upwards. If the transfer of H₂O, aided by advection, is faster than the transfer of heat, then the depth of the melting front oscillates up and down, resulting in parts of the crust going through more than one episode of melting.

Our results show that H₂O-fluxed melting of an haplogranite crust produces columns of mush that are more vertically extensive and more long-lived than dehydration melting of a biotite-gneiss, with the amount of melt depending on both the quantity of H₂O in the system and how fast it diffuses relative to heat. Thus, the net effect of the inflow of H₂O into a hot crust is to cause cooling and lowering of the heat flow through the crust to the surface while increasing the heat content stored in the crust, in the form of a low-temperature mush column.