An integrated approach of machine learning and remote sensing for evaluating landslide hazards and risk hotspots, NW Himalaya

Abstract

Landslides have become more frequent in the last decade in the NW Himalayan region, threatening people and damaging infrastructure. The study presented here aims to identify landslide hazards and risk hotspots in the NW Himalaya. The area is strategically important due the many holy pilgrimage sites and major hydropower projects. A Multilayer Perceptron (MLP) was used to generate the landslide susceptibility map, which was combined with the rainfall intensity map to create the hazard map. To determine the spatial landslide vulnerability, land use and land cover components were taken into account with their respective monetary values. The landslide risk map is the combination of landslide hazard and vulnerability maps, showing that ∼5% of the area falls in high and very high-risk zones, ∼6% in moderate, ∼47% in low, and ∼42% in very low landslide risk zones. High to very high landslide risk zones are mainly confined to Uttarkashi and its surroundings in the southwestern part of the study area, as well as the Tehri, Karanprayag, and Pithoragarh regions in the southern area. Societal risk was also analyzed and revealed that ∼53% of the human population resides in high to very high landslide risk-prone areas. The findings of this study will be beneficial for promoting sustainable development and safe urbanization in the Himalayan region, if used for planning.