Predictive analysis of landslide susceptibility under hydrological aspects of climate change in Kegalle district, Sri Lanka

Abstract

Climate change is widely recognized as a global phenomenon that has far-reaching consequences, including an increase in the frequency and severity of landslides. In response to this global concern, this study narrows its focus to predict landslide susceptibility in the Kegalle District in Sri Lanka, specifically, emphasizing the hydrological aspects influenced by climate change within this region. Potential changes in future rainfall regimes were projected using the HadGEM3-GC31-LL model from the Coupled Model Intercomparison Project Phase 6 (CMIP6). Coarse-resolution rainfall from this model was statistically downscaled to the meteorological station level using the Long Ashton Research Station Weather Generator, known as LARS-WG. In addition to the rainfall influence, the effects of watershed slope, elevation, soil type, land use and land cover type, and the distance from the river were also considered in the Analytical Hierarchy Process (AHP). By analyzing historical landslide occurrences, results demonstrated a concordance of 50%- 70% between observed landslide occurrences and rainfall effects. The incorporation of other landslide-triggering factors improved the accuracy by up to 86%. Subsequently, future landslide maps were generated. The output will aid the decision-makers in prioritizing mitigation efforts and disaster-resilient urban planning.