Rainfall thresholds for the activation of shallow landslides in the Italian Alps: The role of environmental conditioning factors

The aim of the present work is to investigate the role exerted by selected environmental factors in the activation of rainfall-triggered shallow landslides and to identify site-specific rainfall thresholds. The study concerns the Italian Alps. The region is exposed to widespread slope instability phenomena due to its geological, morphological and climatic features. Furthermore, the high level of anthropization that characterizes wide portions of the territory increases the associated risk. Hence, the analysis of potential predisposing factors influencing landslides triggering is worthwhile to improve the current prediction skills and to enhance the preparedness and the response to these natural hazards. During the last years, the Italian National Research Council's Research Institute for Hydrogeological Protection (CNR-IRPI) has contributed to the analysis of triggering conditions for rainfall-induced landslides in the framework of a national project. The project, funded by the National Department for Civil Protection (DPC), focuses on the identification of the empirical rainfall thresholds for the activation of shallow landslides in Italy. The first outcomes of the project reveal a certain variability of the pluviometric conditions responsible for the mass movements activation, when different environmental settings are compared. This variability is probably related to the action of local environmental factors, such as lithology, climatic regime or soil characteristics. Based on this hypothesis, the present study aims to identify separated domains within the Italian Alps, where different triggering conditions exist and different countermeasures are needed for risk prevention. For this purpose, we collected information concerning landslides activated in the period 2000-2012 and econstructed 453 rainfall events supposed to be responsible for the activations. Then, we selected a set of thematic maps to represent the hypothesised landslide conditioning factors and to identify the supposed homogeneous domains within the study area. We employed an existing statistical method for the definition of the cumulated event rainfall vs. rainfall duration (ED) thresholds, for both the entire catalogue of rainfall events and for the events falling in the separated domains. The obtained results contribute to a better understanding of the role exerted by geological, pedological and climatic factors in landslides activation and help identifying separated domains where different risk managing strategies should be adopted. The proposed methodology can be a valid support for risk reduction strategies planning at regional scale.