Abstract

Crystal zonations provide valuable snapshot of the dynamic changes within a magma reservoir. However, there can be inconsistencies between crystals and their zonations, such that deciphering their signatures becomes convoluted. Crystals grow and shrink with temperature, pressure, and composition changes, which can change spatially and temporally as crystals transverse and evolve in the volcanic plumbing system. In this manuscript, we investigate how and at what spatial and temporal scales zonations record cooling processes in a magma lens after injection of fresh magma from depth. We simulate the cooling interface in either hot basaltic or dacitic magmas after their injection into a cold magma reservoir. We couple fluid dynamics to thermodynamics by allowing crystals with constant density and size to form and dissolve based on ambient melt properties and we resolve flow at the scale of each individual crystal. We infer zonations in our simulated crystals by tracking the temperatures they sample over time, but we do not grow or partially dissolve crystals. Our results show that when thermodynamics and fluid dynamics are coupled, a self-sustaining instability arises, because crystals transport the cooler-than-ambient melt in which they formed, creating conditions favorable to crystal formations where the crystal fraction is already high. Our results show that oscillatory zonation patterns can result from the self-sustaining instability, potentially overprinting larger, system-scale trends. Also, our results show that many of the crystals in the instability dissolve and lose their record of the instability. Many of the crystals that persist to the end of the simulation are shorter lived than the entire simulations, under estimating how long the domain remained at high temperatures. Our results suggest that zonations and their heterogeneity can be indicative of local instead of system scale processes, highlighting the importance of having multiple indicators to decipher a system scale process. Thus, we propose multiple indicators for the self-sustaining instability that we describe here.