A data-driven approach to physics-based risk models for deep-seated landslides

This study addresses the modeling of deep-seated landslides, focusing on the El Forn landslide in Andorra, using remote sensing and data-driven approaches to create risk maps. A temperature-based model is adjusted with data from an instrumented borehole to determine material properties and conditions. The calibrated model is compared to Interferometric Synthetic Aperture Radar (InSAR) data, using the data for spatial analysis and creating a correlation map through kriging. This map leads to a physics-informed risk map indicating areas of instability. An uncertainty analysis of the model highlights its limitations but underscores the utility of such maps for policy and planning in areas prone to landslides. This approach provides a novel tool for assessing landslide risks, combining in-situ and remote sensing data for effective risk management.