Characterization of hillslope deposits for physically-based landslide forecasting models

Physically-based models employed for landslide forecasting are extremely sensitive to the use of geological information and a standard, universally accepted method to input maps containing information of geological interest into the models still has never been established. In this study, we used the information contained in a geo-database aimed to characterize the geotechnical and hydrological parameters of the hillslopes deposits in Tuscany, to find out how to organize and group the measurements to spatially create classes that mirror the distribution of the various types of bedrock lithology. Despite the deposits analysed are mainly consisting of well sorted silty sands, statistical analyses carried out on geotechnical and hydrological parameters highlighted that it is not possible to define a typical range of values with relation to the main mapped lithologies, because soil characteristics are not simply dependent on the bedrock typology from which the deposits originated. Instead, the analysis of the relationship of soil parameters with morphometric parameters (slope angle, profile curvature, planar curvature) shows that the highest correlation between the soil grain size class type (USCS classification) and morphometric attributes is with slope curvature, both profile and planar.