Climate change is causing Himalayan glaciers to shrink rapidly and natural hazards to increase, while downstream exposure is growing. Glacier shrinkage promotes the formation of glacial lakes, which can suddenly drain and produce glacier lake outburst floods (GLOFs). Bhutan is one of the most vulnerable countries globally to these hazards. Here we use remotely sensed imagery to quantify changes in supraglacial water storage on Tshojo Glacier, Bhutan, where previous supraglacial pond drainage events have necessitated downstream evacuation. Results showed a doubling of both total ponded area (104 529 m2 to 213 943 m2) and its std dev. (64 808 m2 to 158 550 m2) between the periods 1987–2003 and 2007–2020, which was predominantly driven by increases in the areas of the biggest ponds. These ponds drained regularly and have occupied the same location since at least 1967. Tshojo Glacier has remained in the first stage of proglacial lake development for 53 years, which we attribute to its moderate slopes and ice velocities. Numerical modelling shows that pond outbursts can reach between ~6 and 47 km downstream, impacting the remote settlement of Lunana. Our results highlight the need to better quantify variability in supraglacial water storage and its potential to generate GLOFs, as climate warms.

The Sichuan-Tibet Railway, China's second inland railway to Tibet, is currently being constructed and will run between Chengdu and Lhasa. It will cross the southeastern Tibetan Plateau and be surrounded by glacial lakes, some of which may pose a threat of glacial lake outburst flood (GLOF) events. Both the specific location and the outburst susceptibility of these glacial lakes are largely unknown. In this study, we mapped the glacial lakes using declassified Corona KH-4 and Hexagon KH-9 from the 1960s and Sentinel-2 imagery from 2020 and assessed their spatio-temporal changes. GLOF-susceptibility criteria were established based on historical GLOF events. The results show that the total area (number) of glacial lakes has increased by 22% (20%) from 126.1 ± 2.4 km2 (1662 lakes) in the 1960s to 153.6 ± 11.1 km2 (1994 lakes) in 2020. Of these lakes, this study identified 38 very high and 85 high GLOF-susceptibility lakes; mainly distributed along the Bomi-Nyingchi railway section in the Parlung Zangbo River basin. Four of the very high GLOF-susceptibility glacial lakes may pose a threat to the railway and will require monitoring. The insights from this study can be used to mitigate the risk of GLOFs during the construction and maintenance of the Sichuan-Tibet Railway.