Assessing the Impacts of Climate Change on Landslide Susceptibility in Northwestern Alps

The impacts of climate change are increasingly visible through the rising frequency and intensity of landslides. Heavy rainfall and temperature changes in mountainous areas are major contributors to landslide events, with per-mafrost thawing linked to the latter. This study assesses how climate change may influence landslide proneness in the Italian Northwestern Alps by using Extreme Gradient Boosting (XGBoost) to model and map susceptibility un-der both current and projected climate conditions. Key static and dynamic variables were gathered across the study area to forecast landslide susceptibility. For model training and validation, a dataset of 728 points, including land-slide and non-landslide occurrences, was used.Monthly susceptibility maps were generated; however, only two months (May and February) are presented in this paper. May represents the month with the most landslides, while February represents the month with the fewest. To account for climate change effects, downscaled data from the Coupled Model Intercomparison Project Phase 6 (CMIP6) was applied, incorporating projections from global cli-mate models based on shared socio-economic pathways (SSPs). The high-emission scenario (SSP585) was em-ployed to estimate precipitation and temperature impacts on landslide susceptibility for the period 2021-2040. Model performance for each month was assessed using the area under the curve (AUC) metric, showing high pre-dictive accuracy with values surpassing0.96. The comparison of current and future landslide susceptibility reveals a marked increase in landslide risk due to climate change, with February experiencing the greatest impact.This study paves the way for future research on how landslides may impact regional infrastructure.

Keywords: Landslides, Susceptibility mapping, Climate change, Machine learning, CMIP6, GIS.