Modelling volcanic deformation from coupled magmatic and hydrothermal systems; application to Soufrière Hills Volcano, Montserrat

Soufrière Hills is an active dome building volcano on the island of Montserrat, part of the Eastern Caribbean, that has been in a state of ongoing eruption since 1995. Multi-parametric monitoring is conducted by the Montserrat Volcano Observatory, including an island-wide ground deformation GNSS network operating for nearly three decades. The ground displacement timeseries has been key to modelling the subsurface processes and pressure changes causing them, often using a pressurized cavity or, in more recent models, a poroelastic body in an elastic medium. However, a purely magmatic deformation source has thus far been unable to fully account for the observed deformation signal across the island, leading to significant residuals between simulated and observed geodetic data, particularly at sites closest to the vent. In this study, we will investigate the influence of the Soufrière Hills hydrothermal system on the deformation field. Fumarolic fields and heated springs suggest the presence of an active hydrothermal system at high elevations near the volcanic vent. In the southwest, a more distal geothermal upwelling, as well as anomalies in seismic tomography and gravity data, suggests the presence of a deeper accumulation of hydrothermal fluids, hypothesised to have formed due to the intersection of a number of regional faults and zones of weakness.

In this study we compare magmatic, hydrothermal, and combined deformation source simulations to investigate how different causal mechanisms influence the modelled surface displacement field across Montserrat. We use observed deformation from Montserrat between 2010 and 2022 via GNSS records from 14 continuous monitoring stations to validate our models. Two different model setups are tested: a homogeneous model as a computationally inexpensive baseline, and a heterogeneous model containing seismically defined low permeability andesitic cores in the north of the island, faults in the southwest, and a clay capped region of high permeability in the region of the inferred hydrothermal aquifer. Deviating from traditional volcano-deformation models, our models include a seismically inferred magma reservoir geometry in a poroelastic model domain in an effort to better simulate observed deformation at near-vent GNSS stations. The results from this study will assist volcanic hazard assessment and contribute to the investigation of on-island geothermal resources.