Glacial Lake outburst floods in the Himalayas: A comprehensive hazard and risk analysis for disaster preparedness

Abstract

Glacial Lake Outburst Floods (GLOFs) have emerged as one of the most significant and escalating threats to high-altitude Himalayan regions, posing serious risks to downstream communities, infrastructure, and hydropower installations. These hazards are increasingly exacerbated by climate-induced glacier retreat and the intensification of extreme precipitation events, driven by global warming. Against this backdrop, the present study offers a comprehensive, multi-dimensional assessment of GLOF risk through an integrated approach that combines susceptibility mapping, spatiotemporal analysis of glacial lakes, extreme flood event reconstruction, and modelling of future flood scenarios. The overarching aim is to provide a robust framework for hazard forecasting, risk mitigation, and sustainable development in glacier-fed river basins. The study begins by employing a Glacial Lake Outburst Flood Susceptibility Mapping (GLOFSM) framework for the Himalayan region using a hybrid Frequency Ratio–Analytical Hierarchy Process (FR-AHP) model. By synthesizing qualitative expert judgment with quantitative, data-driven analysis, the approach minimizes subjectivity and enhances reliability. A total of 851 glacial lakes were analyzed against twelve weighted factors derived from 25 historical GLOF events—such as lake area, glacier proximity, slope, avalanche risk, and precipitation levels. The model classified 324 lakes as highly hazardous and 50 as very highly hazardous, achieving a predictive accuracy of 86.65% based on Area Under the Curve (AUC) validation. This susceptibility mapping provides a critical tool for prioritizing GLOF potential lakes and planning targeted mitigation strategies. To capture the evolution of glacial lakes over time, a spatiotemporal analysis was conducted for the Sikkim Himalaya using Landsat satellite data from 1990 to 2020. The region witnessed a marked increase in glacial lakes—from 309 to 440 in number, and from 22.83± 0.78 km2 to 30.71±1.07 km2 in total area—highlighting the accelerating pace of glacier melt and hydrological transformation. An enhanced hazard assessment framework was proposed by integrating qualitative assessments (e.g., ice avalanches, landslides) with a Fuzzy AHP (FAHP) approach and stochastic inundation modelling using the Monte Carlo Least Cost Path (MC-LCP) method. Among the 51 lakes flagged for potential GLOF activity, 13 were categorized as either high or highly hazardous, informing region-specific monitoring and emergency preparedness efforts. Recognizing the potential for cascading lake failures, the study further explores worst-case multi-lake GLOF scenarios through two-dimensional hydrodynamic simulations in HEC-RAS. Focusing on three critically vulnerable lakes—South Lhonak, Gurudongmar, and Shako Cho—seven breach scenarios were modeled, ranging from individual lake outbursts to simultaneous failures. The simulations revealed flow depths between 5 and 20 m and velocities up to 4 m/s, with the most extreme event producing a peak discharge of approximately 10000 m³/s at Chungthang, a key downstream settlement. These findings underscore the immense destructive potential of multi-lake outbursts and the pressing need for integrated, basin-wide risk assessments. To better understand real-world events, the 2023 South Lhonak GLOF—triggered by a combination of lake breach and prolonged rainfall—was reconstructed using a coupled HEC-RAS and HEC-HMS modelling framework. Simulated peak discharges reached ~7355 m³/s±845 at Chungthang and ~8282±952 m³/s at Gazoldoba, closely matching observed streamflow data with an accuracy of 78%. Additionally, three future GLOF scenarios were projected, estimating discharges of 8826±1015 m³/s, 10817±1244 m³/s, and 12385±1424 m³/s. These projections provide essential insights for developing early warning systems and long-term risk forecasting. Expanding the spatial scope, the study also evaluates compound hazard scenarios in the Beas and Sutlej River basins by coupling glacial lake outburst modelling with Probable Maximum Flood (PMF) estimation. Using the Hershfield method, Probable Maximum Precipitation (PMP) values were estimated at 650.12 mm for the Beas Basin and 530.68 mm for the Sutlej Basin. Rainfall depth analysis using the Gumbel method further supported the derivation of PMF discharges, which were estimated at 17086 m³/s for the Pandoh Dam and 23478±2699 m³/s for the Bhakra Dam. Under worst-case conditions involving both GLOFs and extreme precipitation, peak discharges rose to 17356±1995. m³/s and 24723±2843 m³/s, respectively. Inundation maps identified widespread risks to build infrastructure, affecting 94 critical assets in the Beas Basin and more than 588 structures—including major hydropower and industrial installations—in the Sutlej Basin. In conclusion, this study presents a robust, integrative framework for GLOF hazard assessment in the Himalayan region, leveraging satellite-based observations, advanced statistical modelling, and coupled hydrological-hydrodynamic simulations. The findings reveal not only the accelerating vulnerability of glaciated catchments but also the compounded threat posed by cascading failures and extreme weather events. The study highlights the need for continuous glacial lake monitoring, improved early warning systems, and infrastructure adaptation policies. Ultimately, this research contributes to a deeper understanding of climate-sensitive hazards in mountainous regions and provides a scientific foundation for disaster-resilient planning and sustainable watershed management in the Himalayas.