Identification of internal structures in an active volcano is mandatory to quantify the physical processes preceding eruptions. We propose a fully unsupervised Bayesian inversion method that uses the point compound dislocation model as a complex source of deformation, to dynamically identify the substructures activated during magma migration. We applied this method at Piton de la Fournaise. Using 7-day moving trends of Global Navigation Satellite System (GNSS) data preceding the June 2014

more »

... on, we compute a total of 15 inversion models of 2.5 million forward problems each, without a priori information. Obtained source shapes (dikes, prolate ellipsoids, or pipes) show magma migration from 7-8 km depth to the surface, drawing a mechanical "tomography" of the magma pathway. Our results also suggest source geometries compatible with observed eruptive fissures and seismicity distribution. In case of finite magma volume involved in final dike injection, source volume estimates from this method allow forecasting volumes of erupted lava. Plain Language Summary Imaging the interior of an active volcano and estimating volumes of magma present at depth are major challenges of eruption anticipation. In this work we propose an effective method of data processing that combines a mathematical model of the potential source at depth and standard ground deformation measurements at the surface in a fully automated process that has been implemented as a real-time monitoring tool to anticipate eruptions at Piton de la Fournaise volcano. The method is sensitive to magma migration, highlighting the magma pathway, like a scanner that displays a 3-D image of the volcano plumbing system. In specific circumstances, this method can be used also to forecast volumes of erupted lava.