An approach for large-scale soil characterization for the application of non-structural landslide risk mitigation

This paper describes the methodology used to create a database of the physical and mechanical properties of the soils within the Perugia Province, for the application of probabilistic predictive models at large scale. Starting from an extensive data collection from previous geotechnical campaigns, provided by the Civil Protection and Structural Control Office of the Perugia Province, the geostatistic Kriging technique was applied to obtain: (i) a spatial distribution of the mechanical characteristics of the soil within the selected study area, taking into account the collected measurement points and their spatial correlation; (ii) an evaluation of data reliability on the basis of the computed experimental variograms for the main types of soil identified. After this large-scale characterization, an application of the probabilistic physically-based model PG-TRIGRS [SALCIARINI et al., 2017] for rainfall-induced shallow landslide assessment over a selected study area in the Perugia Province is presented, in order to demonstrate the importance of the availability of quantitative and geo-referenced information concerning the mechanical properties of soils to apply predictive tools.