Imaging volcanic structures by seismic tomography

The knowledge of the distribution of masses (densities) and elastic modules in a volcanic edifice is needed to understand its mechanical behaviour and to perform the modelling of its deformation. In this frame we set up an earthquake arrival-time tomographic inversion method, using an accurate numerical scheme for the computation of theoretical travel times and a bayesian regularization of the inverse problem. This method allows us to get a stable and accurate spatial distribution of the seismic velocity. It has been used at Kilauea (Monteiller et al., 2005 ; Got et al., 2008), Piton de la Fournaise (2008) and Popocatepetl (Berger et al., 2011) volcanoes.

Figure 1 : Spatial distribution of the P-wave seismic velocity at 6-km depth, computed by tomographic inversion of earthquake arrival-times. The map shows the main structural features, the seismic stations (black triangles) and the earthquakes (white dots) used in the tomographic inversion. The model is the simplest that best fits the data (optimal a priori information for velocity parameters : 5-km correlation length, 1-km/s standard deviation) (from Got et al., 2008).