Seismic response to volcanic processes at Mount Etna: coupling thermomechanical simulations with seismic wave-equation modelling

Mount Etna, located in the central Mediterranean on the north-eastern coast of Sicily (Italy) is one of the most active and hazardous strato-volcanoes in the world whose south-eastern flank is sliding towards the Ionian Sea. Flank collapses can be caused by gravitational spreading and tectonic forces, leading to crustal fracturing and possible opening of magma pathways. To investigate their effect on the sliding flank at Etna, we created a 3D geodynamic model of the volcano, containing the primary geological parameters and produced ground deformation with the 3D thermomechanical finite differences code LaMEM (Lithosphere and Mantle Evolution Model). To achieve a sufficient model resolution, simulations were run in parallel on an Open Computing Cluster for Advanced data Manipulation (OCCAM). The geometry of the model was built with the GeophysicalModelGenerator.jl package, also including real topography through the GMT.jl package by using the Julia programming language. Implemented geometries, which are based on geological maps and interpretation of seismic velocity data, were assigned rock parameters from laboratory-scale values and computational simplicity.

The model includes most importantly an upper brittle crust with a south-eastern flank that is decoupled by surrounding objects of weak rock characteristics relative to the flank, represented by two bounding fault systems (Pernicana in the north and Mascalucia-Tremestieri, Fiandaca-Pennisi and Gravina in the south) and an underlying Apennine-Maghrebian Chain. Additionally, the effect of a visco-elastic basement, a solidified high velocity intrusive body (HVB) and a supercritical fluid volume (LVZ) with strong and weak rock parameter values, respectively, were tested. 4 types of low-velocity zones (LVZ) and 2 types of high-velocity bodies (HVB) were implemented, with geometries deriving from seismic imaging. 

We observed the impact of each implemented geometry and its sensitivity to parameters by comparing the model results to GPS observations quantitatively, estimating the misfit between the model and data using the coefficient of determination. To quantitatively compare the geodynamic model results, deformation data from time periods (2004-2006), covering inflation and deflation phases of low intensity (Bonforte et al. 2008), were used to isolate the sliding of Etna’s southeastern flank through gravitational spreading. The greatest effect could be observed when adding the visco-elastic basement to the brittle upper part of the model, which also increased the fit to GPS observations. HVB presence seems to not have a significant effect on geodynamic model, but contributes to its stabilization, while a sphere shape geometry is the best choice for fitting the LVZ