Integration of geophysical and geotechnical modelling to define hydrogeological conditions for unsaturated soils prone to shallow flowslides

Flowslides, which occur mainly in shallow granular soils resting on bedrock, are among the most dangerous natural hazards to humans and utilities in both mountainous and volcanic areas. These phenomena are strongly controlled by the stratigraphic and topographic characteristics of the slope and the groundwater regime, which are commonly recognized as predisposing factors for the occurrence of flow-like landslides. Therefore, the study of the spatial variability of the local geological setting and hydrogeological conditions in partially saturated slopes is of fundamental importance for the prediction of flowslide events. In this framework, we propose a procedure based on 3D time-lapse electrical resistivity tomography imaging of the slope, integrated with geotechnical numerical modelling of hydraulic phenomena affecting the land cover, to analyse the effects of the stratigraphic variability in terms of geometry, continuity, and thickness of the soil horizons on the groundwater regime over time. The goal of the proposed approach is to set up an effective tool for predicting debris-flow landslides occurrence at the slope scale, thereby increasing the predictive capacity of early warning systems. The proposed multidisciplinary study was applied to the test site on Mt. Faito, in the northernmost part of the Lattari Mountains (Naples, Southern Italy), where loose pyroclastic deposits from the explosive eruptions of the nearby Somma-Vesuvius volcano cover a karst-fractured carbonate bedrock with a natural slope more than 30° steep. Specifically, seasonally repeated 3D electrical resistivity tomography surveys, suitably complemented with geological and geotechnical investigations, were carried out to determine the electro-stratigraphic and geological setting of the pyroclastic cover, the local morphology and physical conditions of the underlying carbonate bedrock, and the saturation degree distribution on a seasonal time scale. The latter was estimated through the resistivity vs. water content characteristic curves of the different soil horizons, which were obtained from laboratory measurements on specimens sampled in the survey area. The maps of the water content distribution within the pyroclastic cover, determined by the repeated field resistivity surveys, were validated by comparison with those obtained from 2D geotechnical numerical modelling aimed at simulating hydraulic phenomena affecting the soil cover. As main findings, the integrated approach showed that i)  the buried bedrock morphology heavily influences the pore water distribution in the soil cover and ii) ashy material fills the upper karst portion of the bedrock, providing a hydraulic connection of the water flow infiltrating from the topsoil downward.