2nd Year PGR Talks series - Examining the relationship between seismic swarms and ground deformation during volcanic unrest.

Identifying patterns in the relationship between streams of monitoring data during precursory volcanic unrest may aid forecasting over month to year timescales, and help anticipate the location of surface volcanism. Here we examine the relationship between seismicity and ground deformation during unrest at an active volcanic setting (Eyjafjallajökull, Iceland). Using monitoring data prior to the well-studied 2010 eruption, we identify a step-change in ratio between cumulative seismic moment and radial GPS ground movement. This is interpreted as a change in magmatic source between two sills, given the sudden increase in seismic energy release relative to deformation, and changes in the direction of displacement. The change in ratio is thought of as an increase in seismic efficiency, suggesting a decrease in aseismic inflation without frictional slip between sills. Locations of seismic foci and further changes in speed and direction of GPS movement allow us to identify a third source; a dyke connecting both sills to ground surface and feeding an effusive eruption. By examining changes in seismic efficiency within each source and relating these to focal locations, we see differences in how the inferred inflation generates seismicity. The decrease in aseismic inflation for the shallower of the two sills relates to magmatic ‘lobes’ separated by seismogenic fault zones. Combining these observations with relationships between seismic moment and modelled volume (by combining radial and vertical GPS data with source models in literature) for the different sources, we compare evolution of the magmatic complex of Eyjafjallajökull to previous studies comparing seismicity against volume, and find linear trends similar to other empirical examples. This case study aims to provide insight into subsurface processes prior to eruptive activity, by relating changes in relationship between precursory seismicity and deformation to volcano characteristics, with the hope that trends observed may improve forecasting for less well-monitored volcanoes.