Integration of machine learning model and CMIP6 analysis for climate change impact-led landslide susceptibility and population exposure assessments in the Nepal Himalaya

Abstract Background Landslides are one of the most frequent and damaging disasters in the Nepal Himalaya region, triggered primarily by intense rainfall events. With the increasing impacts of climate change, extreme precipitation is expected to rise in both frequency and intensity, potentially exacerbating slope instability. Traditional landslide susceptibility assessments have typically relied on static environmental variables, limiting their ability to reflect future climatic conditions. This study aims to address this gap by integrating a machine-learning-based susceptibility model with precipitation projections from the Coupled Model Intercomparison Project Phase 6 (CMIP6) to assess current and future landslide susceptibility and population exposure across the Nepal Himalaya region. Results A Random Forest (RF) classification model was developed using 13,912 historical landslide records and 15 geo-environmental predictors, including topographic, geological, anthropogenic, and hydrological variables. The model demonstrated strong predictive performance (AUC = 0.93) and was applied to simulate future landslide susceptibility under two Shared Socioeconomic Pathways: SSP2-4.5 (intermediate emissions) and SSP5-8.5 (high emissions), for the periods centered on 2040 and 2085. CMIP6 projections indicate moderate increases in mean annual precipitation of 3–9% under SSP2-4.5, with substantially larger increases of up to 24% under SSP5-8.5. These changes result in spatial redistribution of landslide susceptibility, particularly across Nepal’s mid-hill regions. Under baseline conditions, approximately 21.9% of Nepal’s land falls within High to Very High susceptibility classes. Population exposure analysis indicates an increase of up to 4.1% within High-susceptibility zones under future climate scenarios. Conclusions The integration of machine learning with climate projections provides a robust framework for climate-sensitive landslide risk assessment. The model successfully captures both spatial variability and temporal changes in landslide susceptibility under future climate conditions. Although land use changes were not considered, the study offers valuable insights for long-term disaster preparedness and infrastructure planning in Nepal’s mountainous regions. The results highlight the importance of climate-informed susceptibility and exposure modeling for supporting resilient development in hazard-prone environments.