IMPROVING MODELING AND SIMULATION OF RAINFALL-INDUCED LANDSLIDES: FROM PREDICTION TO POST-FAILURE DYNAMICS

Landslides are a natural hazard that cause impactful effects not only on the earth’s natural environment (i.e. morphology, flora, and fauna) but also on people and infrastructures. Direct and indirect damage to property and human settlements, as well as many casualties all over the world are unfortunately consequences of landslide events. Hence, the prediction and modeling of such events is of high interest to scholars aiming at understanding both the triggering mechanism and the post-failure dynamics, either by empirical or physically-based approaches. The present thesis addresses this key problem by identifying innovative approaches to one of the most key aspects of landslide hazard assessments, namely the spatio-temporal prediction of landslide occurrence. In particular, the main goals of the work can be summarized as follows: i) the improvement of the spatial prediction of landslides through physically-based models, with a special focus on the real advantages and disadvantages of 1D vs. 3D slope stability analysis at the catchment scale; ii) the analysis of the post-failure dynamics as well as the assessment of the influence of the geotechnical and rheological parameters influencing the front celerity and the impact force of the sliding mass; iii) the improvement of performances of empirical rainfall thresholds through the integration of the reanalysis soil moisture information within a hydro-meteorological framework.