Probabilistic modelling of lava flow hazard at Mount Etna

This thesis aims at presenting a groundbreaking methodology for lava flow hazard assessment that represents a challenging topic in modern volcanology. Specifically, the methodology is based on three main steps: (i) the construction of a spatiotemporal probability map for the future opening of new eruptive vents; (ii) the estimation of occurrence probabilities associated to classes of expected eruptions; and (iii) the overlapping of a large number of lava flows simulated using the MAGFLOW model. The results from these steps are processed in order to obtain a hazard map showing, for a given area, the probability of being affected by at least one lava flow inundation during the time interval considered. The preferred scenario for this study has been Mount Etna being one of the most intensively monitored volcanoes in the world, thereby offering large amounts of input data for the application of the proposed methodology.