GIS-based statistical analysis of rainfall-induced landslides: A case study of the 2018 Kerala event

Abstract

This study presents a comprehensive spatial–statistical analysis of landslide occurrences in Kerala, India, focusing on the relationship between landslide frequency, size, and various predisposing factors. The analysis reveals a nonlinear correlation between slope angle and both landslide frequency and size: The highest frequency of events occurs on moderate slopes (20°–35°), whereas larger landslides are associated with steeper slopes (>40°). Although 44% of events occurred in regions receiving <440 mm of rainfall, these areas showed higher probability densities, indicating frequent, smaller landslides. In contrast, lower landslide frequencies but larger events were observed in zones exceeding 440 mm rainfall and at higher elevations. Western and southwestern slope aspects account for the majority of landslides, with probability density peaking on west-facing slopes. Additionally, convex and concave slope curvatures were most susceptible to frequent failures, though curvature exerted minimal influence on landslide size, while terrain ruggedness index (TRI) values between 0.4 and 0.58 are strongly correlated with both frequency and size. Proximity to roads demonstrates an inverse relationship with frequency and size; while most landslides occurred 2001–6000 m from roads, the largest events and highest probability densities were recorded beyond 6000 m. These findings highlight the intricate interplay between topographic, hydrological, and anthropogenic variables in controlling landslide behaviour and offer critical insights for probabilistic landslide susceptibility modelling and hazard mitigation strategies in Kerala. © Indian Academy of Sciences 2025.