High frequency of moraine-dammed lake outburst floods driven by global warming

Article https://doi.org/10.1038/s41467-025-67650-3 High frequency of moraine-dammed lake outburst floods driven by global warming Received: 14 February 2025 Accepted: 4 December 2025 1234567890():,; 1234567890():,; Check for updates Taigang Zhang 1,2, Weicai Wang 1 , Ioannis Kougkoulos3, Simon J. Cook Sihan Li6, Pablo Iribarren-Anacona7, C. Scott Watson 8, Baosheng An1 & Tandong Yao1 4,5 , Glacial lake outburst floods (GLOFs) represent a major hazard in mountain regions, yet considerable uncertainty persists regarding whether their frequency has increased in recent decades and to what extent this trend is linked to climate change. Here, we developed a new inventory of GLOFs from moraine-dammed lakes, analyzing 609 events worldwide between 1900 and 2020. Insights from historical reports and geomorphological evidence presented a low but fluctuating increase in the global frequency of reported GLOFs prior to the 1970s. However, a marked acceleration occurred after the 1980s, with the annual frequency increasing from 5.2 GLOFs during 1981–1990 to 15.2 GLOFs during 2011–2020. Overall, the long-term trajectory of reported GLOF frequency closely parallels variations in global air temperature, exhibiting a lag-correlated pattern on timescales of approximately 20 years. The concept of GLOF response time was employed to explain this delayed reaction, which is attributed to warming-induced glacier recession, glacial lake expansion, and slope destabilization surrounding such lakes, ultimately triggering GLOFs. Glacial lake outburst floods (GLOFs) stemming from moraine-dammed lakes represent a major hazard in mountainous regions1. The related cascading processes often span tens to hundreds of kilometers and lead to profound societal and geomorphological consequences2,3. For example, in 2023, in the Sikkim Himalaya, the South Lhonak GLOF killed 55 people, with a further 74 reported missing, and caused devastation as far as >160 km downstream, destroying ~30 bridges, 2000 buildings, and three hydropower plants4,5. Other catastrophic GLOFs have been well-documented globally, including in British Columbia, Canada6, Cordillera Blanca, Peru7, and the Patagonian Andes8,9, and are garnering increasing scientific and policy interest10. The occurrence of GLOFs is typically linked to glacier recession (i.e., debuttressing of valley slopes by shrinking glaciers, and ice avalanches into lakes from unstable ice fronts), permafrost degradation (rockfalls, landslides, and melting of dead ice in the moraine dam), and extreme weather events (extreme rainfall and melt events), all of which are exacerbated by global warming11–13. Nevertheless, the link between GLOFs and climate change remains contentious and is often biased by potential omissions and recording errors14. Analysis of existing GLOF inventories has so far shown a decrease or stagnation in frequency at both regional and global scales since the 1980s15–17, which seems at odds with the intuitive expectation that, given recent intense warming18 and glacial lake proliferation19,20, GLOF occurrence should have increased. Constantly updated and revised GLOF inventories are instrumental in addressing these discrepancies. They enable analyses of GLOF frequencies and climate linkages while providing information for 1 State Key Laboratory of Tibetan Plateau Earth System, Environment and Resources (TPESER), Institute of Tibetan Plateau Research, Chinese Academy of Sciences, Beijing, China. 2College of Earth and Environmental Sciences, Lanzhou University, Lanzhou, China. 3Department of Science and Mathematics, Deree —The American College of Greece, Athens, Greece. 4Division of Energy, Environment and Society, University of Dundee, Dundee, UK. 5UNESCO Centre for Water Law, Policy and Science, University of Dundee, Dundee, United Kingdom. 6School of Geography and Planning, University of Sheffield, Sheffield, United Kingdom. 7Instituto de Ciencias de la Tierra, Facultad de Ciencias, Universidad Austral de Chile, Valdivia, Chile. 8School of Geography and water@leeds, e-mail: University of Leeds, Leeds, UK. Nature Communications | (2025)16:11173 1 Article https://doi.org/10.1038/s41467-025-67650-3 identification of unreported GLOFs (see Methods) to improve the dataset quality and support robust assessments of GLOF characteristics across different spatiotemporal scales. Key findings indicate a phased increase in GLOF frequency since the 20th century, largely driven by intensified global warming. broader applications, such as validating the reliability of GLOF susceptibility assessments21, identifying lakes for GLOF scenario modeling22–24, and informing GLOF risk mitigation and policy development25,26. A precise GLOF chronology can be constructed using two approaches. The first involves compiling events reported in documented sources, including journal papers, news reports, and local administrative records27. The second integrates post-failure geomorphic analysis, leveraging GLOF traces such as decreases in lake levels, breached dams, outwash fans, and downstream devastation, to identify previously unreported events28,29. Employing manual or computer-automated analysis of images from diverse sources can substantially enrich inventories. Research in the Himalaya17, tropical Andes7, and Southern Andes30 indicates that relying solely on passively reported information underestimates the quantity of morainedammed GLOFs by 0.5–2 times. Thus, constructing a global GLOF inventory necessitates continuous updating of reported events, as well as proactive identification of unreported historic GLOFs via geomorphic assessments. Moreover, for ambiguous cases, GLOF traces can confirm their reliability31, while extensive satellite imagery archives such as Landsat and Sentinel can be used to further constrain GLOF timing for those events with otherwise large time-of-outburst windows. The latest global inventory (version 4.1) includes 463 morainedammed GLOFs that occurred between 1900 and 202032 but still requires improvements in completeness, continuity, and accuracy. In this work, we present an updated global inventory of morainedammed GLOFs that enables systematic analysis of their distributions, frequencies, and climatic connections. We apply a three-step framework consisting of dataset synthesis, outburst timing calibration, and a Results GLOF distribution and frequency By synthesizing diverse regional inventories and conducting systematic geomorphic assessments, we identified an additional 178 GLOFs from documented sources and detected an additional 88 GLOFs, which were incorporated into the latest global inventory. The reliability of each GLOF was validated by assessing residual geomorphic traces in satellite imagery. In total, these efforts led to the compilation of 609 GLOFs that originated from 512 different moraine-dammed lakes globally during 1900–2020 (Fig. 1a). This represents a 285% increase compared to the first dedicated GLOF inventory published in 2 (...truncated)