Data Driven Based Spatio-Temporal Quantification and Prediction of Landslide Susceptibility for the Himalayan Region

Landslides are the most common natural hazard in the hilly regions of the world. In India, high landslide susceptibility zones are mainly found in the Himalayas, the Western Ghats, and North-east India. The use of landslide susceptibility and hazard maps for land use planning has increased significantly during the last few decades. Landslide Susceptibility (LS) mapping is an essential step in mitigation measures for planning and recognising the regions needing protective measurements. Many studies have performed these mapping measures; however, they all lack consistency in selecting landslide-causing factors for the susceptibility analysis and mapping. The variability in choosing factors for the same region by different researchers has made it challenging to compare the models' prediction accuracies. In the present study, a scientific method was adopted for identification of significant landslide causing factors for LS analysis. The chosen combination was also tested on two test sites with similar terrain conditions. Further, dynamic factors such as land use land cover and climate variables were adopted to predict future projections of LS analysis. The result shows that chosen 11 significant factor model has highest prediction accuracy of 0.93 area under curve (AUC) value. The Analytical Hierarchy Process (AHP) based LS mapping for Chamba, Bhuntar, and Tehri regions achieved a prediction accuracy of 0.86, 0.82, and 0.89 AUC values. Also, the results show a promising increment in the built-up area and agriculture field and reduced forest area in LULC projection of 2050. Further, the results also indicate that the zones of very high landslide susceptibility class will increase by 2%, 4%, 7%, and 9% under projected LULC and climate scenarios of Shared Socioeconomic Pathway (SSP) 1-2.6, SSP 2-4.5, SSP 3-7.0, and SSP 5-8.5 respectively.

Keywords: Future landslide susceptibility mapping, Land use land cover projections, Climate projections.