Rasuwa Disaster Was a Rock-Ice Avalanche, Not a GLOF, and Needs Seismic Warning Networks

The news

Rasuwa, Nepal, and Tibet. The disaster that struck across the Nepal-Tibet border began with a rock-ice avalanche that turned into a debris flow, not with a glacial lake outburst, argues Stuart Dunning, professor of applied geomorphology at Newcastle University, UK, in The Indian Express. A rock-ice avalanche is a sudden collapse of rock and glacier ice from a steep mountain slope; a debris flow is a fast-moving slurry of water, mud and boulders that can travel far down a valley. A glacial lake outburst flood (GLOF) is the sudden release of water from a lake held back by a glacier or by loose glacial debris. Dunning says the same sequence caused the Dharali disaster in 2025, and that a rock-ice avalanche from Ronti Peak in 2021 became the debris flow behind the Chamoli disaster. None of these, he writes, is “natural”: they become disasters only when they hit lives and property. First reports from Nepal suggested an earthquake and possibly a GLOF. GLOFs are a real risk: his group once calculated that 15 million people globally face GLOF danger, over 9 million of them in High Mountain Asia and nearly 3 million in India alone. But the “earthquake” was the force of the avalanche hitting the valley floor, and there was no glacial lake in the valley. The Rasuwa event was therefore harder to plan for: there was no lake to model floods from, to drain or to monitor. Such avalanches could start from many hundreds of thousands of steep, rapidly deglaciating slopes with thawing permafrost (ground that stays frozen for years) in the Himalaya. In both Chamoli and Rasuwa, a significant number of the dead or missing were workers at hydropower infrastructure in the path of the flows. Existing systems gave enough warning downstream but not upstream, and warning depends on detection, communication and how fast people act. A denser seismic station network could detect and locate the landslide-generated tremor and the continuous seismic “noise” that floods make, but this has not been done because it needs cross-border cooperation across the Himalaya and tens of millions of dollars. The science, he argues, must be embedded in communities that trust the warnings and know what to do; even then, “we may save lives… but we do not save the places”. The syllabus link is GS1 geomorphology and GS3 disaster management.

Static syllabus linkage

1. A GLOF and a rock-ice avalanche are different hazards with different warning logic. A glacial lake outburst flood occurs when a lake dammed by a glacier or by a moraine, the loose ridge of debris a glacier leaves behind, breaches suddenly, often after an avalanche into the lake, heavy rain or an earthquake. Because the source is a known lake, it can be mapped, modelled, partly drained and fitted with sensors. The South Lhonak lake GLOF in Sikkim in October 2023 destroyed the Chungthang dam of the Teesta-III hydropower project, a textbook case of this type. A rock-ice avalanche has no fixed source: it can come from any steep slope where rock and ice are weakening, which is why Dunning calls it harder to plan for.
2. Deglaciation and permafrost thaw weaken mountain slopes. Permafrost is ground that stays at or below 0°C for at least two consecutive years; in high mountains it includes ice in the cracks of bedrock that acts like cement. As temperatures rise, this ice thaws and the rock mass loses strength. Retreating glaciers also remove the support they once gave to valley walls, a process called debuttressing, leaving steep slopes prone to collapse. The Himalaya, often called the Third Pole for its stock of ice, is warming faster than the global average, so these slope failures are expected to become more frequent.
3. India’s disaster law and NDMA guidance treat GLOFs as a distinct hazard. The Disaster Management Act, 2005 set up the National Disaster Management Authority (NDMA), chaired by the Prime Minister, with State and District authorities below it. In 2020 NDMA issued national guidelines on the management of GLOFs, covering the mapping of glacial lakes, early warning systems, lowering lake levels where needed, and land-use regulation in flood paths. In the regional sphere, the International Centre for Integrated Mountain Development (ICIMOD), based in Kathmandu, brings together eight Hindu Kush Himalayan countries, including India, Nepal and China, for research on mountain hazards. Chamoli in February 2021 showed the limits of this framework, since its flood struck the Rishiganga and Tapovan-Vishnugad hydropower projects without any lake burst.
4. The Sendai Framework makes early warning a global target. The Sendai Framework for Disaster Risk Reduction 2015-2030 was adopted in March 2015 at the Third UN World Conference on Disaster Risk Reduction in Sendai, Japan, succeeding the Hyogo Framework for Action (2005-2015). It sets seven global targets and four priorities: understanding disaster risk, strengthening disaster risk governance, investing in risk reduction for resilience, and enhancing preparedness with ‘Build Back Better’ in recovery. Its Target G seeks to increase the availability of and access to multi-hazard early warning systems. India launched the Coalition for Disaster Resilient Infrastructure (CDRI) in 2019, headquartered in New Delhi, which focuses on making infrastructure such as power plants resilient to disasters.

Why UPSC loves this

1. Disaster management and geophysical phenomena are fixed syllabus items. GS1 covers important geophysical phenomena and GS3 covers disaster and disaster management. Mains has asked directly about the Sendai Framework and how it differs from the Hyogo Framework, and Himalayan disasters such as Kedarnath and Chamoli have prompted questions on causes and preparedness. This op-ed gives the latest science: that the dominant hazard may not be the one the policy is built around.
2. Prelims tests hazard definitions and institutions. Definitions of GLOF, permafrost and debris flow, and facts about NDMA, CDRI, ICIMOD and the Sendai Framework, are regular Prelims material. The distinction between a lake-sourced flood and a slope-sourced avalanche is exactly the kind of conceptual trap an examiner can set.
3. Hydropower in fragile zones links environment, energy and security. Questions on Himalayan hydropower now combine clean-energy benefits with disaster risk and cross-border water relations with Nepal and China. The fact that hydropower workers made up many of the victims in Chamoli and Rasuwa gives a human dimension candidates can use in GS3 and in Essay.

Analysis

1. Calling every mountain flood a GLOF leads to the wrong fix. Policy has concentrated on glacial lakes because they are visible, countable and can be engineered: lowered, siphoned or fitted with sensors. If the Rasuwa, Chamoli and Dharali disasters all began on slopes rather than in lakes, then a lake-centred programme protects against only part of the risk. Dunning’s point that there are hundreds of thousands of deglaciating slopes means that the hazard cannot be managed site by site. The policy vocabulary needs to widen from ‘GLOF risk’ to ‘cascading high-mountain hazards’, with separate tools for each. That does not make GLOF work wasteful, since the South Lhonak flood in Sikkim was a genuine lake burst.
2. Warnings protect those downstream; the people upstream are often hydropower workers. Dunning notes that existing systems gave enough warning downstream but not upstream, and that hydropower workers were a significant share of the dead in both Chamoli and Rasuwa. Dams, intakes and tunnels are built high in valleys precisely where the warning time is shortest. This makes worker safety a matter of project design: evacuation routes out of tunnels, sirens at work sites and rules that stop work when upstream sensors trigger. Environmental clearance for such projects should require an assessment of cascading hazards, not only a flood design standard. The counter-view is that hydropower gives firm, low-carbon power and revenue to Himalayan States, so the answer is safer siting and operation rather than a blanket halt.
3. The seismic network is cheap; the hard part is sharing data across borders. Tens of millions of dollars is small compared with the losses from a single disaster and the value of the hydropower assets at risk. The real obstacle is that a useful network must span India, Nepal, Bhutan and China, since avalanches in Tibet can flow into Nepal and floods can cross into India. India–China arrangements for flood-season hydrological data on the Brahmaputra and Sutlej have been intermittent, which shows how politics can interrupt science. India could lead by offering seismic data from its side openly and by funding stations in Nepal and Bhutan through bodies such as ICIMOD or CDRI. Shared warnings are also a low-cost form of neighbourhood diplomacy.
4. Trust is a governance variable, not a soft extra. Dunning’s insistence that science be embedded in communities who trust the warnings reflects hard experience. Warnings that are too frequent and wrong lead people to ignore them, while warnings that come without instructions leave people unsure where to go. Odisha’s cyclone record shows the difference: after the 1999 super cyclone, community shelters, drills and local volunteers cut deaths dramatically in later cyclones. Mountain villages need the same: marked safe zones, rehearsed routes, local volunteers trained to relay alerts, and feedback so that false alarms are explained. Detection without this last mile is incomplete.
5. Saving lives is possible; saving places requires land-use decisions. Dunning closes by saying that warnings may save lives but not places, which raises the uncomfortable question of where people can and cannot live. Settlements and hotels on river terraces and debris fans, such as those hit at Dharali, sit on landforms built by past floods. The Joshimath subsidence of 2023 showed how unregulated construction in the mountains creates risk. Hazard zonation that restricts new construction on such landforms is politically difficult because it limits livelihoods and tourism. Yet without it, every warning system only buys time for people to flee buildings that should not have been there.

Administrator's brainstorm