Answers are grounded in this paper and cite their sources.
With increasing rainfall variability, rapid urbanisation in hilly terrains, and more frequent landslides, there is an urgent need for forward-looking landslide susceptibility mapping (LSM). However, most existing studies focus on historical or current events, with limited assessment of future susceptibility under evolving climatic and land-use conditions. This study projects landslide susceptibility across the Western Ghats, India, for 2021–2100 under four Shared Socioeconomic Pathways (SSPs). Multiple machine learning models were evaluated, and the best-performing XGBoost model, combined with SHapley Additive Explanations (SHAP), was employed for interpretability. The study uniquely integrates future climate projections (CMIP6), projected land-use dynamics, and explainable artificial intelligence (XAI) to produce transparent and scenario-based LSM outputs. Results reveal an overall rise in susceptibility across all SSPs, with the most pronounced increase under SSP5-8.5 during 2051–2100, where 6.38% of the area falls within the very high susceptibility category. Three hotspot zones exhibit marked upward shifts in susceptibility linked to maximum extreme rainfall events. The Kerala region of the Western Ghats shows the most severe deterioration under SSP5-8.5, with nearly two-thirds of hilly areas becoming highly susceptible by 2100. The central and southern zones also show greater susceptibility increases compared to the northern region. By integrating climate projections, land-use dynamics, and XAI, this study provides a novel, transparent framework for future landslide risk assessment and region-specific adaptation planning, offering critical insights to enhance disaster preparedness under changing environmental and socio-economic scenarios.
No takes yet. Share an insight, caveat, or question.
Kushwaha et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: