Srinagar: Mohan Bahadur Chand, a Himalayan glaciologist with Kathmandu University, was in a discussion on glaciers, permafrost, snow and mountain hazards on the morning of 26 August when reports of the disaster arrived.
A torrent of water, mud, rock and debris was barrelling down the Lhende Khola valley along the mountainous Nepal-Tibet border, killing thousands as it tore through three major Nepali districts spanning roughly 100km.
Early accounts suggested a glacial lake outburst flood, or GLOF, in which a glacial lake suddenly releases a large volume of stored water. But as Chand and his colleagues reconstructed the event, a different sequence emerged. There had been no lake at the point of failure. A massive ice-rock the size of six football fields had broken loose from a mountain and transformed into a destructive debris flood downstream.
“We had no clue how it happened, so we coordinated with Chinese scientists across the border to identify the cause and location of the ice-rock failure,” Chand, a scientist with Kathmandu University’s Himalayan Cryosphere, Climate and Disaster Research Center, told Mint.
The collapse generated a seismic signal equivalent to a magnitude 5.2 earthquake. On 5 September, Nepal’s National Disaster Risk Reduction and Management Authority said 1,344 people had been killed and 5,531 remained missing.
The Himalayas have always been unstable, shaped by steep slopes, active tectonics that can cause high-magnitude earthquakes and landslides, and glaciers accompanied by repeated freezing and thawing. What is changing is the environment in which these operate.
The Nepal disaster showed how a mountain failure can begin without the warning signs associated with a glacial lake and change character downstream. For India, whose Himalayan region is home to almost 33,000 glaciers and stretches nearly 2,500 km across Jammu and Kashmir, Ladakh, Himachal Pradesh, Uttarakhand and Sikkim, it raises a difficult question: are existing monitoring systems watching the right risks?
Lurking dangers
The Nepal flood exposed a monitoring gap because the failure did not begin with a lake that could be watched as it expanded.
“This was an unprecedented and unexpected disaster, and there was simply no detection and no time to issue a warning,” said Farooq Azam, senior cryosphere specialist at the Kathmandu-based International Centre for Integrated Mountain Development (ICIMOD), during a webinar last week. (While a glaciologist focuses on moving bodies of ice like glaciers, a cryosphere specialist studies the entire frozen global system and how it interacts with the atmosphere, oceans, and land.)
Azam said the 2021 Chamoli disaster in Uttarakhand, when an ice-rock mass detached from Ronti Peak, had at least one detectable precursor, or a measurable warning signal—surface fissures were visible in high-resolution satellite imagery. In Nepal, the initial rupture had developed beneath the rock surface, making the failure much harder to detect.
Satellites can track retreating glaciers and expanding glacial lakes, but are less effective at detecting instability beneath rock, ice or frozen ground. India’s Central Water Commission tracks 2,843 glacial lakes and water bodies larger than 10 hectares across Himalayan river basins through satellite remote sensing. Such monitoring identifies visible changes but cannot establish the internal stability of a fractured rock face or hanging ice mass.
The wider cryosphere is changing in ways that can change this risk. Tariq Abdullah, a geo-informatics researcher based in Kashmir, has analysed 12,243 glaciers in the Upper Indus Basin—which spans India, Pakistan, Afghanistan, and China—using satellite-derived elevation data. Between 2000 and 2012, the glaciers thinned by about 0.35 metres a year, representing an estimated cumulative mass loss of roughly 70 gigatonnes, or 700 billion tonnes, of ice.
Such retreating glaciers can prove disastrous as the thinning results in reduced support for high-altitude rock faces, while repeated freezing and thawing can cause fractures.
Cascading mountain hazard
In the Indian Himalayas, the pace of glacier melting has nearly doubled over the last decade, contributing to the expansion of glacial lakes and increasing the potential for GLOFs, according to Shakil Ahmad Romshoo, a renowned earth scientist and vice chancellor of the Islamic University of Science and Technology, Kashmir.
Snow adds another dimension. Across the Hindu Kush Himalayan region, for example, seasonal snow cover from November 2025 to March 2026 was nearly 28% below its long-term average, according to ICIMOD.
Black carbon is another atmospheric influence. Abdullah said the particles, produced largely by incomplete combustion of fossil fuels and biomass, darken snow and reduce its reflectivity, potentially accelerating snowmelt.
A National Hazards Review in the Journal of Glaciology earlier this year mapped 155 glacial lakes across the Kashmir Himalaya and found that ice-contact proglacial lakes expanded 26% between 1992 and 2024.
Proglacial lakes are water bodies formed immediately in front of a glacier or ice sheet, directly fed by its melting ice. Five such lakes, Bramsar, Chirsar, Nundkol, Gangbal and Bhagsar, were classified as having very high GLOF susceptibility. A preliminary spatial assessment placed 2,704 buildings, 15 major bridges, roads and a hydropower project within their potential downstream impact areas.
Riyaz Ahmad Mir, scientist-in-charge at the National Institute of Hydrology’s Western Himalayan Regional Centre in Jammu, identified Chenab Valley, particularly the Kishtwar, Doda and Ramban districts, as requiring detailed geological assessment. Ladakh faces exposure to glacier and rock collapses, while the Kashmir Valley faces exposure to landslides, flash floods, avalanches and other glacier-related hazards.
“The vulnerability assessment has to consider the entire system,” Mir said, referring to the ecosystem of glaciers, rock slopes, permafrost (ground frozen for at least two consecutive years), water, river channels and downstream areas.
A collapse high on a mountain can become a debris flow kilometres downstream. A landslide can block a river, create a temporary lake and later breach, sending water and sediment towards settlements. The point where a hazard begins can therefore be far from where damage is greatest.
Yasir Altaf, assistant professor in the Department of Environment, Sustainability and Climate Change at the Islamic University of Science and Technology, Kashmir, described this as a cascading or compound mountain hazard. “One event can trigger another, creating a cascading or compound mountain hazard,” Altaf said.
Himalayan disasters
The 2013 Kedarnath disaster exposed the consequences of extreme rainfall and flooding in a densely occupied temple valley in Uttarakhand’s Rudraprayag district. The 2021 Chamoli ice-rock avalanche destroyed the 13.2 MW Rishiganga hydropower project and damaged NTPC Ltd’s under-construction Tapovan-Vishnugad project, killing more than 200 people. In October 2023, South Lhonak lake failed in Sikkim, sending a flood down the Teesta river and destroying the 1,200 MW Teesta III dam at Chungthang.
Climate writer Hridayesh Joshi said during last week’s webinar that records showed 681 avalanches in the high mountains of Asia over five decades, causing around 3,100 deaths. About 60 of those avalanches occurred in Nepal and India, accounting for roughly 1,400 deaths. “The problem is not only the hazards themselves, but also how these areas are administered and how many people are living in them,” said Joshi.
He added: “Over the past 50 years, the population in these areas also increased by nearly 600%. In 2000, around 8,000 sq.m. of land was under habitation, but this is projected to increase to 150,000 sq.m. by 2030. That means many more people and settlements are being exposed to disaster risks.”
The key policy question is not whether the Himalayas are becoming a death trap. Mountain communities have adapted to these conditions for generations. The concern is the growing concentration of hotels, roads, hydropower projects and other infrastructure in hazard-prone areas, which can turn natural events into major disasters.
For instance, an analysis using India Meteorological Department data documented 2,863 extreme weather events in J&K between 2010 and 2022, associated with 552 deaths.
Himachal Pradesh shows how changing mountain conditions can create risks beyond glacial lake outbursts. The Ghepang Ghat glacial lake in Lahaul-Spiti grew from about 85 hectares in 2015 to 110 hectares in 2025, holding an estimated 24.7 million cubic metres of water, according to Nisha Rani, professor of environmental science at Himachal Pradesh University. The lake lies upstream of the Sissu tourism area, National Highway 3 and the Atal Tunnel corridor.
Rani said the state monitors glaciers, snow cover, glacial lakes, extreme rainfall, slope deformation and landslides, while maintaining hazard and vulnerability assessments and landslide inventories. The main task, however, is converting that information into decisions about specific locations.
“Extreme rainfall, rock instability, glacier melt, avalanches, debris flows, river obstruction and sudden breaching can interact and amplify one another,” Rani said. “The response needs to combine real-time rainfall and river-level monitoring with slope monitoring, satellite observations, early-warning networks and clear public communication.”
The infrastructure trap
The Hindu Kush Himalayan arc—which stretches roughly 3,500 km across eight countries, from Afghanistan in the west through Pakistan, India, China, Nepal, Bhutan, and Bangladesh, to Myanmar in the east—acts as a water tower for more than 240 million mountain residents, while 1.65 billion people in downstream river basins depend on its runoff.
The Intergovernmental Panel on Climate Change, a United Nations body, estimates that 671 million people lived in high-mountain zones globally in 2010, a population projected to rise to 736 million–844 million by 2050.
In India, tourism and pilgrimage add seasonal populations to narrow valleys where flat land is scarce. Roads have been widened, tunnels driven through mountains, hydropower projects built along rivers, and hotels and settlements expanded around transport corridors. This changes the consequences of a mountain failure even when its frequency remains uncertain.
A 2016 study published in UK-based IOPscience’s Environmental Research Letters found that 66% of 257 sampled Himalayan hydropower projects were on potential GLOF pathways. This does not mean the projects are destined to fail. But it raises questions on whether planning assumptions remain valid as glaciers, lakes, slopes and rainfall patterns change.
In Chamoli, the 2021 avalanche destroyed the Rishiganga hydropower project and severely damaged NTPC’s Tapovan-Vishnugad project, causing losses estimated at more than ₹1,500 crore, according to the Union ministry of power. In Sikkim, the South Lhonak outburst destroyed the 1,200MW Teesta III hydropower plant and damaged infrastructure hundreds of kilometres downstream.
Nadeem Qadri, an environmental lawyer and amicus curiae before the High Court of Jammu and Kashmir and Ladakh, said the problem also lies in how projects are assessed. “Major projects such as highways, tunnels and dams require rigorous environmental scrutiny because they can permanently change fragile mountain systems,” Qadri said. “The weakness is rarely the absence of environmental laws, but their ground implementation.”
From maps to action
While India has expanded its disaster-management architecture, the main question is what happens after a hazard is mapped.
Earlier this month, following the Nepal floods, Union home secretary Govind Mohan convened a review with chief secretaries and disaster-management officials from J&K, Himachal Pradesh, Uttarakhand, Sikkim and Arunachal Pradesh, alongside the relevant departments. The review examined vulnerable glacial lakes, avalanche zones, downstream flow paths and warning systems.
The Centre also directed states to expedite GLOF mitigation projects sanctioned in 2024 under the ₹150-crore National GLOF Risk Mitigation Project.
Sikkim has moved further since the South Lhonak disaster. A 13-member commission on glacial hazards in 2024 identified 40 high-risk glacial lakes, including 16 in Category A (highest risk), while the state expanded water-level monitoring, automatic weather stations, bathymetric surveys to map the contours of water bodies such as lakes and rivers, and geophysical investigations.
J&K has also built a substantial risk database. Under the World Bank-funded Jhelum and Tawi Flood Recovery Project, authorities have prepared a multi-hazard risk assessment and a GIS-based decision support system for the disaster management department.
The difficulty, however, lies in turning those systems into decisions on the ground.
In J&K, bathymetric surveys of five high-susceptibility lakes identified in the Kashmir Himalaya have faced logistical delays, while real-time ground monitoring remains a challenge. In Himachal Pradesh, Ghepang Ghat is being monitored even as major transport infrastructure remains downstream. Uttarakhand faces the same challenge in valleys where roads, hydropower, settlements and tourism infrastructure are concentrated. And Sikkim’s experience shows the problem from another angle.
“Scientific monitoring has improved considerably since the 2023 disaster,” said Nischal Wanjari, associate professor of geology at Sikkim University. “The far more difficult challenge is moving people away from active river corridors. Towns such as Singtam, Majhitar and Rangpo are tied to highways, hotels and local trade.”
Relocation, Wanjari said, involves land rights, commercial livelihoods and political capital. “People are often unwilling to move away from riverbeds because that is where their economic survival is rooted.”
An early-warning system is useful only if people can act on it. A sensor can detect a change, but it cannot move a settlement built inside a projected flood or debris-flow path.
- 2,843 | Glacial lakes and water bodies that India’s Central Water Commission currently monitors across Himalayan river basins.
- 26% | Expansion of ice-contact proglacial lakes in the Kashmir Himalaya between 1992 and 2024, highlighting growing GLOF risk.
- 70 gigatonnes | Estimated cumulative glacier mass loss in the Upper Indus Basin between 2000 and 2012