Methods for Shallow Landslides Susceptibility Mapping: An Example in Oltrepo Pavese

The principal aim of this study is to compare different landslide susceptibility zonation models for predicting areas prone to shallow landsliding using both physically distributed landslide models and artificial neural networks. Necessary geotechnical and hydrological parameters were obtained coupling sample laboratory analysis and in situ measures; soil thickness was estimated using an empirical model while distribution of rainfall intensity was analyzed by performing a spatial interpolation. The predictive capabilities of these models were finally evaluated using a threshold-independent quantitative method (the ROC plot).