Modeling decompression paths in a basaltic andesite magma using the nucleation and growth of plagioclase microlites

Abstract Plagioclase microlites in a magma nucleate and grow in response to melt supersaturation (Δϕplag). The resultant frozen plagioclase crystal size distribution (CSD) preserves the history of decompression pathways (dP/dt). SNGPlag is a numerical model that calculates the equilibrium composition of a decompressing magma and nucleates and grows plagioclase in response to an imposed Δϕplag. Here, we test a new version of SNGPlag calibrated for use with basaltic andesite magmas and model dP/dt for the ca. 12.6 ka Curacautín eruption of Llaima volcano, Chile. Instantaneous nucleation (Nplag) and growth (Gplag) rates of plagioclase were computed using the experimental results of Shea and Hammer (J Volcanol Geotherm Res 260:127–145, 10.1016/j.jvolgeores.2013.04.018, 2013) and used for SNGPlag modeling of basaltic andesite composition. Maximum Nplag of 6.1 × 105 cm h−1 is achieved at a Δϕplag of 44% and the maximum Gplag of 27.4 μm h−1 is achieved at a Δϕplag of 29%. Our modeled log dP/dtavg range from 2.69 ± 0.09 to 6.89 ± 0.96 MPa h−1 (1σ) with an average duration of decompression from 0.87 ± 0.25 to 16.13 ± 0.29 h assuming a starting pressure Pi of 110–150 MPa. These rates are similar to those derived from mafic decompression experiments for other explosive eruptions. Using assumptions for lithostatic pressure gradients (dP/dz), we calculate ascent rates of < 1–6 m s−1. We conducted a second set of Monte Carlo simulations using Pi of 15–30 MPa to investigate the influence of shallower decompression, resulting in log dP/dtavg from 2.86 ± 0.49 to 6.00 ± 0.86 MPa h−1. The dP/dt modeled here is two orders of magnitude lower than those calculated by Valdivia et al. (Bull Volcanol, 10.1007/s00445-021-01514-8, 2022) for the same eruption using a bubble number density meter, and suggests homogeneous nucleation raises dP/dt by orders of magnitude in the shallow conduit. Our modeling further supports the rapid-ascent hypothesis for driving highly explosive mafic eruptions.