A conceptual model of the December 2018 submarine eruption at Ambrym volcano (Vanuatu). The inset is a zoom on Ambrym’s tectonic setting, emphasizing that Ambrym’s rift zone orientation is sub-parallel to the regional maximum compressive stress, allowing us to interpret Ambrym as a large tension fracture.

At Ambrym volcano (Vanuatu) in 2015, a reservoir (blue spheroid) at depths of 4–5 km b.s.l. fed a dike intrusion and intracaldera fissure eruption. An extraction of less than 7% of magma from the reservoir was sufficient to unclamp the western portion of Ambrym's caldera ring-fault (light green rectangle). We hypothesize that the fault was prestressed by a previous event associated with a depressurizing reservoir (due to an eruption, degassing, etc.), and the reservoir deflation in 2015 brought the fault to failure. 

A schematic showing a theoretical model used to explain the end of subsidence at Ambrym volcano in 2017, after 2 years of deformation (modified after Girona et al. (2014)). This model assumes a reservoir is connected to the surface by an open conduit, and that the pressure in the reservoir equilibrates the weight of the magma in the conduit. Consequently, mass changes within the system (either in the conduit or reservoir) due to degassing result in reservoir depressurization, causing ground subsidence.

An interpretation of Ambrym's magmatic plumbing system after the December 2018 eruption. The annotated sketch includes variables from a theoretical model discussed in this study. Using geodetic modelling and a data assimilation approach, we estimate values of pressure change over time for two separate reservoirs. Furthermore, we apply a theoretical model to demonstrate that the reservoirs may not currently be hydraulically connected, despite evidence of physical mixing of magma derived from each reservoir during the December 2018 eruption. 

This current work applies the finite element method to investigate how changes in surface temperature relate to pore pressure conditions beneath and within a quiescent lava dome. Preliminary results find that high surface temperatures alone do not necessarily indicate high overpressures and pre-eruptive conditions, and that distributed regions of moderately to highly elevated temperatures may indicate increased dome overpressure. Time-dependent models also indicate that a dome permeability decrease may cause a local and temporary decrease in surface temperature.