Novel Rainfall Thresholds for Shallow Slip Prediction Based on Field Monitoring: Case Study of the Lattari Mountains, Italy

Flowlike landslides (debris flows, debris avalanches, and flowslides) often initiate as shallow slip and, due to their propagation downslope, pose significant threats to infrastructure and human lives worldwide. Such landslides are typically triggered by heavy rainfall infiltrating granular and partially saturated slopes. Field monitoring provides valuable insights into slope responses to hydraulic loads and strongly improves the forecasting of slip initiation. However, current strategies mostly rely on rainfall thresholds derived from empirical and statistical approaches, and the integration of field monitoring into prediction procedures remains a challenge. This paper focuses on the development of physically based rainfall thresholds for predicting shallow slip occurrence. The proposed approach relies on field monitoring designed to capture the short-term response of unsaturated slopes to individual rainstorms on an hourly scale. First, data from automatic field monitoring collected over two years on a shallow pyroclastic slope are reported. The test site is located on Mt. Faito in the Lattari Mountains/ Sorrento Peninsula (Campania, Southern Italy), an area historically prone to flowlike landslides. Field monitoring was designed to detect the short-term response of the pyroclastic slope to rainstorms by collecting automatic readings of meteorological data, matric suction, and volumetric water content at hourly time intervals. These observations prove that the measured parameters can be effectively regarded as precursors of shallow slip occurrence. Then, the analyzed field data were used to validate a numerical model predicting the slope’s hydrological response to rainstorms and to develop physically based rainfall thresholds for shallow slip occurrence. These thresholds are defined in terms of mean rainfall intensity, rainfall duration, and subsoil suction preceding the rainstorm and validated against homologous data on historical landslides within the reference geological context.