Glacial lake outburst floods (GLOFs) are among the most devastating natural disasters, posing significant socioeconomic risks and endangering communities and infrastructure in mountainous areas. To prevent these events, it is crucial to understand their mechanisms and monitor glacial lakes accurately. However, current mapping techniques often lack recent inventories, and many GLOF events remain unrecorded or lack essential data. Our team has created a new framework that leverages satellite optical imagery, Synthetic Aperture Radar (SAR), field surveys, and relevant documents to map lakes, identify past GLOFs, and fill data gaps. Our findings demonstrate that deep-learning algorithms using remote sensing can automatically and reliably detect regional glacial lakes, and that combining multiple satellite datasets significantly advances GLOF research. We have mapped lakes in the Hindu Kush, Karakoram, and the Himalayas, documenting nearly 300 GLOF events in High Mountain Asia from 1900 to 2022 and showing a trend of increasing frequency and distribution. Some GLOFs have damaged infrastructure in Asia; recent examples include floods from Shisper Lake and Purepu Lakes, underscoring the severity of these events. The rapid expansion of infrastructure and more frequent GLOFs heighten risks for populations and assets in High Mountain Asia. GLOF risk assessments enable the identification of high-risk lakes for targeted monitoring and intervention, including early warning systems and engineering approaches. Additionally, we developed a system that integrates on-site and remote sensing data to enable real-time GLOF risk analysis and support decision-making. Climate change and accelerating glacial melt pose further threats to Asia's economic development. Hazard assessments, continuous monitoring, and early warning systems for glacial lakes are vital for improving disaster response and management in these mountainous regions.

GLOFs; Deep learning; Climate change; Monitoring and warning system