Scientists at IIT Mandi integrate satellite data and AI to predict landslides across Indian Himalaya

The project began with the extraction of roughly 26,000 historic landslide events from the Geological Survey of India database, which formed the basis for a high‑resolution susceptibility map. The team then incorporated multiple triggering factors—such as slope angle, soil type and land cover—into an ensemble of machine‑learning algorithms. By feeding NASA’s Global Landslide Catalog and IMERG rainfall datasets into a probabilistic model, the researchers derived a rain‑induced landslide probability index that updates in real time. This AI‑driven approach is expected to improve emergency response times and foster better coordination among state disaster agencies.