Search

Abstract In a deforming partially molten rock, melt concentrates into a grain‐scale melt pocket aligned at a preferred orientation (melt‐preferred orientation, or MPO). However, observing this texture alone provides limited information on the 3D orientation and geometry of these melt pockets, which are critical parameters for estimating permeability. Here, we modeled the MPO of experimentally deformed peridotites by simulating melt streaks arising from melt pockets of various shapes and 3D orientations. The model aims to identify 3D distribution and characteristics of melt pockets that could account for the observed length, thickness, and the probability of melt streaks. Results show that melt pockets at preferred orientation exhibit greater length, thickness, and number density compared to those perpendicular. These results can be incorporated into the simulation of melt flow through individual melt pockets, which allows us to estimate the permeability corresponding to the observed MPO. We found that the permeability of vertically compressed peridotites increases with increasing compressive strain and a more elongated and thickened shape for melt pocket aligned at preferred orientation. The vertical permeability in the sample with 30% compressive strain is at least 40 times larger than that of an undeformed sample. For peridotites deformed under simple shear, the permeability exhibits an anisotropy of at least three. Such anisotropic permeability, coupled with the formation of melt‐rich bands and other melt channels, is believed to cause lateral melt focusing beneath mid‐ocean ridges.

Abstract Large silicic eruptions result from rapid evacuation of large, upper crustal reservoirs of silicic magmas. These silicic magmas are thought to be replenished by melt extracted from underlying crystal‐rich source mushes, but the timescales and mechanisms of such melt extraction are unclear. Geochemical observations suggest that the replenishing melt is often more primitive than the eruptives and must thus cool and crystallize to generate the highly silicic magmas that eventually lead to large eruptions. Motivated by these observations, we use thermal models to explore the replenishment conditions capable of building an eruptible silicic reservoir to generate a large eruption. Results show that the minimum melt replenishment rate required for a silicic reservoir to start growing increases with the effective thermal diffusivity of the overlying crust and decreases with the depth of the reservoir. For an eruptible reservoir at 6 km depth to grow, the initial replenishment rate needs to be greater than 2 × 10−9 m/s. High replenishment rates are required to provide enough advected heat to counterbalance rapid heat loss and consequent freezing that would prevent the eruptible reservoir from growing. However, these high initial replenishment rates must then subside over time for the eruptible reservoir to cool, crystallize, and evolve to highly silicic melts. Thermal histories of some natural systems suggest assembly of large eruptible reservoirs in