Advancing landslide early warning in the Indian Himalayas: SIGMA-based rainfall thresholds for Chamoli district

Abstract

This study develops a regional landslide early warning system (LEWS) for Chamoli district, Uttarakhand (India), using a SIGMA-based decision algorithm. This statistical model defines critical rainfall thresholds as multiples of the standard deviation () from historical data and compares real-time rainfall with these statistically calibrated thresholds. The system employs a dual time-scale framework with seasonally adjustable accumulation windows to reflect distinct triggering mechanisms: short-term rainfall (1–3 days) is used to detect rapid, shallow landslides induced by intense monsoon events, while long-term rainfall accumulation captures deep-seated failures linked to slow pore-pressure build-up in low-permeability formations. Thresholds were derived from 166 historical landslides and rainfall data (2006–2021) and spatially calibrated to account for the geological diversity of the district, ranging from fractured schists to clay-rich phyllites. This spatial adjustment significantly reduced false alarms without compromising sensitivity. Validation against an independent set of 41 landslide events (2022–2024) confirmed strong predictive performance, with the model correctly identifying 88% of failures, generating fewer than 5% false negatives, and achieving a negative predictive value exceeding 99%. The 3-day accumulation window proved most effective for shallow landslide detection, while deeper failures required longer accumulation periods, which also carried a higher number of false alerts due to delayed and nonlinear hydrological responses. The model prioritizes safety through a conservative alerting strategy, offering a balance between early detection and operational usability. By aligning rainfall thresholds with local terrain behaviour and dynamically adjusting them to seasonal variability, this study establishes a process-informed framework for early warning. The system presents a scalable, efficient solution for reducing landslide risk across Himalayan regions increasingly affected by climate-driven rainfall extremes. © The Author(s), under exclusive licence to Springer Nature B.V. 2026.