Phase separation in magmatic reservoirs at intermediate crystallinities: experimental and modeling constraints

Before large volumes of crystal poor, high-SiO2 rhyolites are mobilized as melt, they are, in most cases, first extracted through the reduction of pore space within their corresponding crystal matrix (compaction). The timescales associated with such extraction processes have important ramifications for hazards, however, it is unclear whether compaction is suitable for a continuum description, or if granular models are instead required. To explore this issue, we develop and apply a 2-phase continuum model of compaction to phase separation experiments conducted on mixtures of glass particles and corn syrup. We characterize the ability of the crystal matrix to resist porosity change (effective matrix viscosity) using a suite of descriptions, including laws informed by diffusion creep and laws rooted in detailed parameterizations of granular phenomena. We explore the closure models for matrix rheology that can reproduce the experimental data. The experiments span a range of intermediate crystallinities (~0.4-0.6), where the reduction in matrix porosity is accommodated by the rotation and translation of individual grains, rather than their internal deformation, such as those that act at high crystallinities (>0.8). These conditions are believed to be applicable to melt extraction in long-lived silicic magma systems, however, a closure model for the effective matrix viscosity remains elusive at intermediate crystallinities. By applying the two-phase model to the experimental data, we find that descriptions of matrix viscosity that include particle-particle friction and/or describe the existence of jamming states are more suited to model compaction at intermediate crystallinities. Importantly, outside nearly instantaneous force chain disruption events, melt loss is continuous, and the two-phase dynamics can be solved at the continuum scale with an effective matrix viscosity. Further work, however, must be done to determine the effect of force chains when the domain size is several orders of magnitude larger than the effective particle size.