Resumen
Propagation models can study the runout and deposit of potential flow-like landslides only if a reliable estimate of the shape and size of the volumes involved in the phenomenon is available. This aspect becomes critical when a collapse has not yet occurred and the estimation of the unstable volume is not uniquely predictable. This work proposes a strategy to overcome this problem, using two established analysis methods in sequence; first, a Strength Reduction Method (SRM)-based 3D FEM allows the estimate of the instable volume; then, this data becomes an input for a Smoothed Particle Hydrodynamics (SPH)-based model. This strategy is applied to predict the possible evolution of Sant?Andrea landslide (North-Eastern Italian Alps). Such a complex landslide, which affects anhydrite?gypsum rocks and is strongly subject to rainfall triggering, can be considered as a prototype for the use of this procedure. In this case, the FEM?SRM model is adopted, which calibrates using mapping, monitoring, geophysical and geotechnical data to estimate the volume involved in the potential detachment. This volume is subsequently used as the input of the SPH model. In this second phase, a sensitivity analysis is also performed to complete the evaluation of the most reliable final soil deposits. The performed analyses allow a satisfactory prediction of the post-collapse landslide evolution, delivering a reliable estimate of the volumes involved in the collapse and a reliable forecast of the landslide runout.