Valutazione e mitigazione dei rischi geo-idrologici derivanti da eventi catastrofici nell'area del Parco Nazionale delle Cinque Terre

This research is included in the field of studies for the evaluation and mitigation of geo-hydrological risks deriving from catastrophic rainfall events conducted in the Cinque Terre National Park area (Liguria) by the Department of Earth Sciences of the Environment and Life of University of Genoa (Italy). This activity made it possible to create and test a mathematical simulation model, in a physically based GIS environment, capable of carrying out 3D slope stability analysis (extended to a spatial domain) according to the pluviometric conditions assumed in the calculations. The experimentation first highlighted the limits of application of the "infinte slope" limit equilibrium method (LEM) used for the spatially 3D slope stability analysis in a GIS computerized geographical environment, also indicating improvement solutions for a more correct use. The input parameters of the soils (geotechnical, hydrogeological and geometric) entering the above mathematical expression were then examined. Their spatial distribution and their relationship with the local geo-morphometric-environmental control factors, on a basin and regional scale, were assessed, as well as their influence on the results was tested by comparing them with real phenomena and geomorphological processes observed in the chosen pilot area (i.e. Vernazza catchment affected on 25/10/2011 by a catastrophic rainfall event). All this experimentation led to the formulation of innovative calculation procedures and methods which were then coded in the Python programming language and developed in the open source QGIS application system in the form of executable "scripts". The predictive ability of the new model was tested in the aforementioned basin, comparing the spatial output data derived from the mathematical simulation with the real distribution of the surface landslides obtained from the analysis of aerial photos and site surveys. In summary, the system showed excellent potential for the preventive localization of soil mobilization phenomena on the slopes of a territorial domain where, in a neighborhood of 25 m away from the focal landslide areas, it was able to identify the 84% of the phenomena that actually occurred with a probability of 93% of the total of the aforementioned areas (with only 7% of false alarms).