A Physical Model for Three‐Phase Compaction in Silicic Magma Reservoirs

Abstract We develop a model for phase separation in magma reservoirs containing a mixture of silicate melt, crystals, and fluids (exsolved volatiles). The interplay between the three phases controls the dynamics of phase separation and consequently the chemical and physical evolution of magma reservoirs. The model we propose is based on the two‐phase damage theory approach of Bercovici et al. (2001, https://doi.org/10.1029/2000JB900430 ) and Bercovici and Ricard (2003, https://doi.org/10.1046/j.1365-246X.2003.01854.x ) because it offers the leverage of considering interface (in the macroscopic limit) between phases that can deform depending on the mechanical work and phase changes taking place locally in the magma. Damage models also offer the advantage that pressure is defined uniquely to each phase and does not need to be equal among phases, which will enable us to consider, in future studies, the large capillary pressure at which fluids are mobilized in mature, crystal‐rich, magma bodies. In this first analysis of three‐phase compaction, we solve the three‐phase compaction equations numerically for a simple 1‐D problem where we focus on the effect of fluids on the efficiency of melt‐crystal separation considering the competition between viscous and buoyancy stresses only. We contrast three sets of simulations to explore the behavior of three‐phase compaction, a melt‐crystal reference compaction scenario (two‐phase compaction), a three‐phase scenario without phase changes, and finally a three‐phase scenario with a parameterized second boiling (crystallization‐induced exsolution). The simulations show a dramatic difference between two‐phase (melt crystals) and three‐phase (melt‐crystals‐exsolved volatiles) compaction‐driven phase separation. We find that the presence of a lighter, significantly less viscous fluid hinders melt‐crystal separation.