Geotechnical and geophysical property models of soil-covered slopes prone to landsliding. The case study of the Ischia Island (southern Italy)

The physical and geotechnical characterization of surficial geological systems forming slopes and the definition of soil hydrological conditions leading to rainfall-induced shallow landslides are key factors for the effective landslide hazard assessment and setting of Landslide Early Warning Systems (LEWS). Indeed, poor knowledge of the thicknesses, local stratigraphic features and geotechnical properties of soil/regolith coverings, as well as their soil hydrological status, can lead to oversimplified assumptions resulting in high uncertainties in the prediction of landslide hazard in space and time. This work proposes the integration of geotechnical and geophysical approaches to advance the reconstruction of surficial engineering geological models and the assessment of the hydrological status of soil-covered slopes prone to landsliding, to be used for the setting of physical-based slope models. A combined field and laboratory geotechnical and geophysical study of a sector of Mount di Vezzi (Ischia Island, southern Italy), known for the recurrent shallow rainfall-induced landslides affecting settlements in the foothills, was carried out to test our approach. As a principal result, the integration of both approaches advances the reconstruction of a comprehensive physical model of the potentially unstable slopes. Significantly, the coupling of geophysical and geotechnical properties, such as electrical resistivity, volumetric water content and soil water pressure head, allows the use of resistivity as a proxy for the hydrological status of the soil mantled slopes at a volume scale consistent with that of the initial stage of shallow landslides, ranging from 101 to 103 m3. The results obtained are expected to be applicable for the definition of the parameters of physical-based slope models to be used for supporting LEWS, built on rainfall thresholds and hydrological measurements, thus for the estimation of the spatial and temporal variation of shallow rainfall-induced landslide hazard.