Physically-based assessment and redefinition of existing rainfall thresholds in territorial warning systems for shallow landslides

Rainfall thresholds adopted in territorial landslide early warning systems (Te-LEWS) typically rely on statistical analyses of precipitation without any physical correlation with the infiltration dynamics into the soil. On the other hand, the use of physically-based modeling that feeds procedures warning for landslides is typically limited at systems operational at local scale, as they require spatially distributed geotechnical and hydrological parameters and high computational demand for real-time operation. This paper proposes an original approach integrating geotechnical modeling and data-driven analysis to interpret empirical rainfall thresholds used in Te-LEWS and to define new physically-based rainfall thresholds for the initiation of shallow landslides. Physically-based thresholds differ from the traditional empirical ones as they account for the actual physical processes governing soil behavior. To this aim, the proposed procedure exploits relationships between rainfall characteristics, geotechnical parameters, infiltration processes, and slope stability conditions. The framework is structured into four key modules: geotechnical modeling; equalization time matrix generation; thresholds interpretation; thresholds definition. The procedure is proposed and tested in Campania region, southern Italy, although it is structured to be general and can be easily adapted to other territorial warning models employing empirical rainfall thresholds. The main innovation of this study is the development of a novel framework to integrate empirical and physically-based approaches for landslide early warning at regional scale, considering them synergic rather than alternatives. • Empirical rainfall thresholds physically interpreted through infiltration dynamics. • Modular framework integrating geotechnical modeling and data-driven analysis. • Approach bridging empirical and physical methods for Te-LEWS applications. • More reliable warnings for Te-LEWS. • Model designed and tested in an area of Campania region, southern Italy.