Time stamped landslide inventory and its causal factors in Rudraprayag, India

The Rudraprayag district in the Uttarakhand Himalayas, India, is highly prone to landslides, exacerbated by a combination of natural and anthropogenic factors. This study employs a Random Forest classification algorithm to create a time-stamped landslide inventory using Sentinel-2 satellite images (2019–2023) and ancillary datasets, including ALOS PALSAR DEM. Landslide locations were validated through visual interpretation of high-resolution Google Earth imagery and field visits. The results identify 196 confirmed landslide locations, with most occurrences concentrated near road networks and influenced by rainfall and anthropogenic activities.

Topographic metrics such as elevation, slope, aspect, and ruggedness emerged as significant predictors of landslides, while other features like Topographic Wetness Index and curvature had minimal influence. Rainfall analysis revealed no statistically significant correlation with landslide occurrence timing, though extreme rainfall events, such as in July 2023, contributed to gradual landslide expansions. Seismic analysis showed a weak correlation with landslides, suggesting the need for denser seismic monitoring networks for further exploration.

This inventory supports the development of susceptibility maps and disaster management strategies. The study underscores the importance of integrating geological, hydrological, and anthropogenic factors for comprehensive landslide risk assessments, with implications for expanding such analyses across the broader Himalayas.