Enhanced glacial lake activity threatens numerous communities and infrastructure in the Third Pole

The Third Pole, encompassing the Tibetan Plateau and surrounding mountain ranges, is a crucial water tower. Climate warming has led to glacier retreat and the formation of numerous glacial lakes. While these lakes offer potential for hydropower, they also pose a significant threat due to GLOFs. GLOFs occur when glacial lakes are destabilized, causing sudden and destructive water releases. Previous studies on glacial lake changes and GLOFs in the Third Pole have been fragmented and lacked detail. This study aims to address these gaps by providing a comprehensive assessment of GLOF risks, integrating high-resolution satellite imagery, high-precision GLOF modeling, and detailed downstream exposure data to better understand the current and future threats posed by glacial lakes.

Existing research on glacial lake changes and GLOF activity in the Third Pole has revealed inconsistencies in glacial lake inventories due to varying definitions and methodologies. Discrepancies in the number of glacial lakes reported (ranging from 10,000 to 30,000) hinder detailed analysis. Incomplete GLOF datasets also make it challenging to accurately analyze trends, magnitudes, and drivers. Previous studies on GLOF frequency have yielded conflicting results, with some reporting a decrease or stability, while others suggest an increase. Risk assessments have often been qualitative or semi-quantitative, relying on simple models for flood path simulation and rough downstream exposure data. These limitations highlight the need for an integrated approach incorporating improved data and advanced modeling techniques to accurately assess GLOF risks.

This study integrated various datasets to map glacial lakes and compile a GLOF dataset. Sentinel-2A/B images (2018, 2020, 2022) were used to map glacial lakes ≥0.02 km² primarily fed by glacier meltwater. Lakes were classified into five types based on their topological positions relative to glaciers: proglacial, periglacial, extraglacial, supraglacial, and ice-dammed. Glacial lake area and volume changes were calculated. A GLOF inventory since 1900 was compiled from various regional datasets, focusing on moraine-dammed lakes. GLOF triggers were identified. A quantitative hazard assessment was developed for glacial lakes (excluding supraglacial and ice-dammed lakes), using a conceptual model and six indicators (mean slope of parent glacier, mass movement potential, mean slope of moraine dam, watershed area, lake perimeter, and horizontal distance between glacier terminus and lake) to evaluate GLOF susceptibility. The HEC-RAS hydraulic model was used to simulate potential GLOF impacts for high-risk lakes. Downstream exposure was assessed by combining GLOF simulation results with data on buildings, hydropower projects, farmland, roads, and bridges. GLOF probability was defined to indicate the threat at specific locations. The potential disaster intensity was defined as the exposure level per unit of inundated area. Early warning time was calculated based on distance to the nearest downstream community.

In 2022, 5894 glacial lakes were mapped in the Third Pole, totaling 748.79 km². Proglacial lakes showed the most significant expansion (3.03% relative area change between 2018 and 2022). The overall expansion rate decreased from 1990-2018 to 2018-2022, but the expansion rate of proglacial lakes increased. A total of 145 GLOFs from 122 lakes since 1900 were identified, with a significant increase in frequency since the 1980s. Southeastern Tibet and the China-Nepal border showed intensified GLOF activity. Ice avalanches were the main GLOF triggers. Regional GLOF activity correlated with glacial lake abundance and area change rate. 379 glacial lakes were classified as very high hazard and 1120 as high hazard. GLOF simulations indicated potential inundation of ~6353 km², posing threats to significant infrastructure and populations, particularly in the Eastern Himalayas, Southeastern Tibet, and Inner Tibet. China had the highest potential inundation area. The study identifies 28 valleys with high GLOF probability (>0.24), many in transboundary basins, highlighting transboundary risks. 55,808 buildings, 105 hydropower projects, 194 km² of farmland, 5005 km of roads, and 4038 bridges are threatened. Early warning times were shorter in Southeastern Tibet and the Eastern Himalayas.

This study's findings confirm the intensification of GLOF activity in the Third Pole, particularly in Southeastern Tibet and the China-Nepal border area. The correlation between GLOF activity and glacial lake expansion highlights the importance of monitoring glacial lake changes. The high accuracy of the hazard assessment (93%) lends confidence to the identification of potentially dangerous glacial lakes. The comprehensive risk assessment provides valuable insights into the spatial distribution of GLOF impacts and identifies critical infrastructure and populations at risk. Transboundary risks are significant, necessitating regional cooperation. While the western Third Pole currently shows less GLOF activity, its potential for future increases warrants attention. The study's findings provide a crucial basis for developing targeted disaster prevention and mitigation strategies.

This study provides a comprehensive assessment of GLOF hazards and risks in the Third Pole using a novel integrated approach. The findings underscore the urgent need for disaster mitigation strategies, including lake drainage, dam reinforcement, monitoring, and early warning systems. Regional cooperation is crucial given the transboundary nature of many GLOF risks. Future research should focus on improving data availability, particularly in under-represented areas, and refining GLOF modeling to better capture uncertainties and account for climate change impacts.

The study's accuracy is limited by data availability, particularly for downstream exposure elements in some regions. The GLOF simulations represent maximum potential scenarios, and actual impacts could vary. The study focuses primarily on moraine-dammed lakes, and the assessment of ice-dammed lakes requires a separate approach. Uncertainty in the drainage volume estimations in the GLOF simulations could influence results.