Study maps future glacial lakes

GUWAHATI: A team of researchers from the Indian Institute of Technology Guwahati (IIT-G) has developed a new way to identify where glacial lakes are likely to form in the Himalayas, offering a potential breakthrough in disaster preparedness for mountain regions facing rapid climate change.

Their study focuses on the Eastern Himalayas, which has recorded the highest frequency of glacial lake outburst floods, or GLOFs, in the entire mountain range. These sudden floods occur when natural dams holding back glacial lakes collapse, releasing vast volumes of water, debris and sediment downstream.

The findings of this research have been published in Nature Scientific Reports journal. The paper has been co-authored by Ajay Dashora of IIT-G’s Department of Civil Engineering, along with his research scholar Anushka Vashistha and Afroz Ahmad Shah of the Universiti of Brunei Darussalam.

GLOFs have caused extensive loss of life and damage to infrastructure in the past, including roads, bridges, hydropower projects, and agricultural land. The last major disaster caused by a GLOF was in October 2023 in Sikkim, which killed 94 people, damaged 1,835 houses, displaced 2,563 people, and swept a 1,200-megawatt dam away.

According to the researchers, the number and size of glacial lakes are increasing as glaciers retreat faster due to rising temperatures, heightening the risk to downstream communities.

While earlier studies largely relied on climate data to assess glacial lake hazards, the new research takes a different approach. It places geomorphology—the physical shape and structure of the land—at the centre of prediction.

The study examines how specific landforms influence where meltwater accumulates. These include cirques, which are bowl-shaped depressions carved by glaciers, U-shaped valleys, meltwater flow channels, retreating glacier fronts and neighbouring lakes. Together, these features determine whether a landscape can trap water long enough for a glacial lake to form.

Grid locations analysed

Using high-resolution satellite imagery and digital elevation models, the researchers analysed more than 12,000 grid locations across the Eastern Himalayas. They then applied machine learning techniques to estimate the probability of glacial lake formation at each site.

Among the tested models, a Bayesian neural network produced the most reliable results. Unlike conventional models, this approach not only predicts where lakes are likely to form but also quantifies uncertainty in those predictions. This is especially important in remote mountain regions, where field data are limited, and terrain conditions vary sharply over short distances.

The resulting probability maps highlight zones with a high likelihood of future lake formation. Many of these areas coincide with regions of active glacier retreat and gentle slopes that favour water accumulation. Some are located upstream of existing settlements and infrastructure, underscoring the potential risk.

“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,” Prof. Dashora said.

“Beyond hazard management, the method can help understand how water systems may change as glaciers continue to retreat. Importantly, the framework is adaptable to other glaciated mountain regions around the world, making it a valuable tool for climate-resilient planning and disaster-risk reduction globally,” he said.

The study also challenges the assumption that climate alone determines glacial lake formation. While temperature controls the meltwater supply, the research shows that landforms largely dictate where that water ultimately collects. EOM