Modelling landslides in the Lesser Himalaya region using geospatial and numerical simulation techniques

The Himalayan region has always been hard hit by landslides, posing a serious threat to local communities and development activities. Taking cues from empirical and ancillary data from the deadliest landslide in recent times, the present study used an integrated approach to model landslide hazard in the Lesser Himalayan region, especially the Ukhimath tehsil of the Rudraprayag District, Uttarakhand, India. This holistic study involved mapping of a detailed landslide inventory, a hazard and risk analysis, and numerical simulations to model debris flow. Using the detailed landslide inventory (~ 505 landslides), this study produced a landslide hazard map using a bivariate statistical technique (frequency ratio) with 12 data layers: slope, aspect, relative relief, distance to linear features, geology, soil depth and soil erosion, geomorphology, vegetation, and land use/land cover (LULC). This was subsequently validated with receiver operating characteristic curve (ROC) technique which revealed that 89% of the observed landslides occurred in the predicted high-hazard zones which comprised 10% of the total, implying good accuracy. Despite the ubiquitous risk of landslides across the entire area, the final hazard map revealed that only 24% of the area was in a high-to-very-high-hazard zone, whereas a risk analysis revealed that 11% of the total area was located in the very-high-risk zone. Targeting a high hazard zone, four basic parameters: height, velocity, pressure, and momentum of a debris flow were derived using numerical modelling, incorporating the Voellmy friction law in a physics-based model. This study could provide valuable insights for structural and non-structural measures to protect life and property and strengthen landslide mitigation.