GUWAHATI: Researchers at IIT Guwahati have created a new way to predict where glacial lakes are likely to form in the Eastern Himalayas. These lakes can cause sudden floods called glacial lake outburst floods (GLOFs), which have caused severe damage and loss of life. The study appears in Nature Scientific Reports. The team includes Ajay Dashora, Anushka Vashistha, and Afroz Ahmad Shah. GLOFs happen when natural dams holding glacial lakes break, releasing large amounts of water and debris. A recent GLOF in Sikkim in October 2023 killed 94 people and damaged homes, roads, and a big hydropower dam. The researchers found that rising temperatures are making glaciers retreat faster, causing more and larger glacial lakes. Instead of just using climate data, this study focuses on the land shapes, like bowl-shaped cirques and U-shaped valleys, where meltwater can collect and form lakes. Using high-resolution satellite images and digital maps, the researchers studied over 12,000 sites in the Eastern Himalayas. They used machine learning, especially a Bayesian neural network, to predict where lakes might form and how certain these predictions are. Their maps highlight high-risk zones often near retreating glaciers and gentle slopes where water pools. Some of these areas are close to villages and infrastructure, which increases the danger. According to Prof. Dashora, "By pinpointing high-risk areas, the framework can guide early-warning systems for GLOFs, help plan safer locations for roads, hydropower projects, and settlements, and support long-term water-resource management." The research also shows that land shapes play a bigger role than climate alone in deciding where lakes form. This method can also be used in other mountainous regions worldwide to help fight climate risks and disasters.