select search filters
briefings
roundups & rapid reactions
Fiona fox's blog
Scientists comment on flash floods in Nepal.
On the barrier lake:
Prof Maria Shahgedanova, Professor of Climate Science at the Department of Geography and Environmental Science, University of Reading, said:
“We now believe that the Rasuwa flood was triggered by a large rock and ice avalanche at around 5,200m above sea level, when a huge section of glacier and the underlying bedrock, estimated to be between 1 and 1.3km wide, detached from the Lirung massif.
“The avalanche rapidly carried large volumes of ice, rock and sediment downslope. On reaching the valley floor, it interacted with deposits from a previous flood, the 2025 Purepu glacial lake outburst flood, temporarily damming the river. The failure of this temporary blockage released water and debris as a powerful flood wave.
“Researchers from the HiRisk project, which investigates cryosphere-related hazards in High Mountain Asia, suggest that longer-term glacier retreat and permafrost degradation may have contributed to slope destabilisation. The failed glacier retreated by approximately 450m between 1990 and 2020, potentially reducing the mechanical support provided by the glacier to the underlying rock slope. Permafrost degradation may have further contributed to instability. The warmer than average temperatures observed in the area this summer might have contributed to the glacier collapse too, although observations at these elevations remain limited.
“The event demonstrates how climate-driven changes in the cryosphere can change conditions that make hazards in high-mountain regions even more risky for people in the area. Such events cannot be prevented, but their impacts can be reduced through systematic hazard assessment and monitoring. High-resolution satellite observations and InSAR, a type of satellite-based radar that detects small changes in the earth’s surface, can identify changes in glaciers and slope stability and contribute to early-warning systems, particularly in remote mountain regions where conventional monitoring is limited. While the initial avalanche at Rasuwa was too rapid to provide substantial warning, longer warning times downstream could potentially have reduced loss of life.”
Prof Reza Ahmadian, Professor of Water and Environmental Engineering, Cardiff University, said:
1. What is a barrier lake, how has it formed and where has it formed?
“A barrier lake which is also referred to as landslide dammed lake or debris dammed lake, forms when material suddenly blocks the natural flow of a river or stream. Unlike a conventional reservoir, the dam has not been designed, built and engineered. It consists of naturally deposited material, and its geometry, internal structure and stability is therefore highly uncertain.
“In this particular event, the evidence currently indicates that a major glacial collapse and associated ice and rock avalanche occurred on 26 August in the high mountains near the Nepal China border. The USGS has located the source on a glaciated mountain cliff on the northern side of Langtang Lirung. The collapsing ice and rock moved downslope, entraining water, sediment and additional debris and producing the destructive debris flow and flood that subsequently travelled through the river system.
“Part of this material blocked or substantially restricted river channels, allowing water to accumulate behind the debris and creating natural barrier lakes.
“The lake that has received most attention is reported near the confluence of the Chhochen Khola and Purepu Tsangpo rivers, upstream of the Bhotekoshi river system near the Nepal China border. These rivers ultimately feed into the Bhotekoshi and then the Trishuli river system.
“One important point is that this should not simply be described as an ordinary glacial lake. It is a newly formed barrier lake associated with the deposition of ice, rock, sediment and debris following the collapse. Its size can increase rapidly as it has blocked natural flow and its stability is uncertain.
2. What is the likelihood that it will overflow or breach its banks?
“As far as I understand, there are reports that indicate that overflow has already occurred, and there have also been reports of a breach or partial release.
“There is clearly a significant risk, but at present it would not be scientifically responsible to attach a precise probability to a breach. These natural dams are made from an irregular mixture of rock, sediment and ice rather than engineered materials, and their internal structure is not well known. The risk depends on how quickly water is entering the lake, how quickly it can escape, the geometry and composition of the blockage, erosion of the overflow channel and whether further landslides or ice and rock falls occur.
“It should be noted that overflow does not necessarily mean catastrophic failure. Water can establish a relatively stable outlet and progressively lower the lake. Conversely, overflow can erode unconsolidated debris, deepen the outlet and produce a rapidly enlarging breach. That distinction is critical.”
3. What could happen if the barrier lake breaches its banks?
“The principal concern is a sudden release of stored water, producing another flood wave downstream.
“The severity would depend on the volume of water released, how rapidly the breach develops, the size and geometry of the opening, the amount of sediment and debris entrained, and the shape and gradient of the downstream valley.
“A rapid breach could generate a high velocity flood containing water, sediment, boulders and woody or structural debris. In a steep confined Himalayan valley, such a flow can be particularly destructive. It could threaten settlements and people along the river corridor and damage roads, bridges, hydropower infrastructure and other structures.
“The areas of immediate concern are along the Bhotekoshi and subsequently the Trishuli river system. Nepalese authorities have already reported increases in Bhotekoshi water levels associated with releases from the lake.
“There is also the possibility of a cascading hazard. A flood wave can erode riverbanks and valley sides, mobilise additional sediment, destabilise slopes and create further temporary blockages downstream. So the hazard is not necessarily limited to the volume of water initially stored in the lake.
“The scale should nevertheless not be exaggerated. A breach of the present barrier lake would not automatically reproduce the extraordinary event of 26 August. The original event involved a major glacial collapse followed by a debris flow and flood that the USGS estimates travelled approximately 100 kilometres. The consequences of a new breach depend on the actual volume released and the way in which the natural dam fails.”
4. Is it possible to predict when that may happen?
“Not with precision. It is not currently possible to predict the exact time at which a natural barrier might fail. These barriers are composed of irregular mixtures of rock, sediment, ice and other debris, and their behaviour can change quite rapidly as water levels rise and water begins to flow through or over the material.
“What we can do, however, is continuously monitor the conditions that could indicate an increasing likelihood of failure. This includes changes in lake level, inflow and outflow, deformation or erosion of the barrier, and changes in downstream river levels. Satellite observations, aerial surveys, cameras, drones and river monitoring stations can all contribute to this. Nepalese authorities are currently monitoring two lakes and are planning additional cameras and sensors downstream. One lake seems to be draining, while the second, larger lake seemed to be growing and had no visible outlet in the latest satellite observations.
“Hydraulic modelling is also extremely important. We may not be able to say precisely when the barrier will fail, but we can model a range of possible failure scenarios and estimate how much water could be released, the resulting peak discharge, flood depths and velocities, the areas that could be inundated, and critically, how long the flood wave could take to reach different communities downstream.
“This information should feed directly into evacuation planning. Rather than waiting until we know that the barrier is about to fail, which may provide very little warning, authorities can identify vulnerable areas in advance, determine which communities and critical facilities need to be evacuated, identify safe areas on higher ground, establish evacuation routes and define the warning thresholds that would trigger evacuation.
“An important consideration in this particular event is that roads and bridges have already been severely damaged. Therefore, evacuation planning cannot assume that normal transport routes will remain available. Alternative routes, safe assembly points and arrangements for people who may require assistance all need to be considered. Rescue and relief personnel working close to the river are themselves also exposed and need to be incorporated into the evacuation plan.
“There is also a strong case for precautionary evacuation where the potential consequences are severe, and the warning time is uncertain. Authorities do not necessarily need to wait for confirmation that a complete breach has begun. If monitoring indicates a significant increase in risk, moving people away from the river corridor and onto higher ground can substantially reduce the potential loss of life.
“The key point is that although we cannot reliably predict the exact time of failure, we can predict possible consequences and approximate flood travel times. That allows emergency planners to work backwards from the time available and establish when warnings and evacuations need to take place.
“In a situation such as this, good forecasting is therefore not simply about predicting the flood. It is about converting monitoring and modelling information into decisions early enough for people downstream to reach safety.
Prof Liz Stephens, Professor in Climate Risks and Resilience, Department of Meteorology, University of Reading, said:
“The barrier lake has formed behind a pile of debris that was left behind by the initial flood of water, rock and debris, with this debris dam blocking the normal flow of the river.
“Chinese engineers are monitoring the levels of the lake. The higher the water levels get the higher the pressure that builds up behind it, increasing the likelihood that it could fail.
“This is not an uncommon engineering problem in this part of the world. Chinese and Nepalese engineers know what to do, but they will be challenged by the inaccessibility of the landscape high up in the mountains. This is a very new debris dam so it will be very unstable in itself. It will be a difficult and dangerous environment to work in for the engineers on site in the mountains.
“It appears this morning that water levels have started falling. This fall in lake levels suggests some water is now escaping naturally through the dam. That’s very promising, it reduces the imminent threat and may reduce the need to intervene with an engineering solution, to drain the lake artificially. However, this does not mean the risk is over, the debris dam will remain unstable, and the authorities will need to be on the lookout for rain upstream, which could add to the challenge.
“It is important to be aware that this might not be the only lake that forms. The initial flood was so big that it will have had enormous erosional power. It may well have undercut the steep slopes as it went down through the valley into Nepal, leaving behind a much more unstable landscape. Any monsoon rains could threaten that stability further. Monitoring for further landslides is critical, the risks to people in the area, including those involved in rescue and recovery efforts, are definitely not over yet.”
Dr Matt Westoby, Associate Professor of Physical Geography, University of Plymouth, said:
“High Mountain Asia is one of the fastest warming regions globally. It hosts rapidly melting glaciers, growing meltwater lakes and some of the world’s steepest mountains. As the climate warms, shrinking glaciers and thawing frozen ground can destabilise mountain slopes, increasing the risk of rockfalls, landslides and avalanches. These can trigger devastating ‘hazard cascades’ – essentially chain reactions – releasing huge volumes of water, mud and debris into valleys where vulnerable communities and vital infrastructure are often highly exposed, creating a high degree of risk.
“Climate change is making many high-mountain areas around the world more unstable, with increasing reports of landslides, ice and rock avalanches, glacier collapses and resulting hazard chains. In places like High Mountain Asia, more people and infrastructure, such as hydropower, are moving into mountain areas, increasing the potential impacts. However, extreme events are rare and records are limited, so it is still difficult to say exactly how much climate change is increasing the frequency or size of individual hazards.
“A future flood on the scale of what we’ve seen in Nepal/Tibet is possible, so improving early warning and evacuation plans is essential. Seismic and other monitoring systems could quickly detect that something has happened, while mobile-phone networks can help get reliable warnings to people rapidly. People also need to know exactly what to do and, most importantly, where to go to stay safe, with evacuation routes, drills and safe places prepared in advance; this is easier for known glacial lakes, where flood arrival times and likely size can be predicted. Monitoring unstable glaciers and slopes may also help identify potential problems before they happen, although predicting exactly when a slope or glacier will fail remains extremely difficult.”
Prof Hugh Sinclair, Personal Chair in Surface Geodynamics, University of Edinburgh, said:
“A big question is whether these events are increasing in frequency and magnitude due to warming of the high mountains. No single event can be attributed to climate change, but the trend is worrying. Melting is driving collapse of the high Himalayan glacial valleys, and this is likely to accelerate with global warming.”
“The Flood wave travelled at over 100 mph when it hit the Nepal/China border control. Where did the water come from to feed this event? Some suggest damming of the river, but the event hit the border so quickly, that I find that hard to believe. There was no evidence of damming with the initial event, but the topography that was left behind by the initial event has resulted in ponding and the formation of a lake, and this could still collapse and drive further floods in the coming days.”
“Early warning systems can be improved in the high Himalaya with new monitoring techniques, but it needs investment in high altitude monitoring of environmental conditions, and of the initiation of the event itself.”
Prof Geoffrey Boulton, Regius Professor of Geology Emeritus, University of Edinburgh, said:
“Glaciers worldwide reached their greatest recent extents at the beginning of the 20th century. As the whole earth has warmed rapidly in recent decades their retreat has accelerated. Under these conditions there are four processes which either individually or in combination can lead to catastrophic events in glaciated valleys and which are accentuated by rapid retreat.
Thus, climate warming can produce damaging events that tend not to occur as climate cools and glaciers’ advance.”
Prof Mikael Attal, Personal Chair of Geomorphology, University of Edinburgh, said:
“Over the past decade, I have seen numerous simulations of outburst floods. I just couldn’t comprehend the speed and the scale involved. Depths of tens of metres. Impacts hundreds of km downstream. The videos of the Nepal disaster brought these numbers to life, in the most nightmarish manner.
“The exact causes of the disaster are still being investigated but we may still ask questions:
“Can we prevent or control such events? No. No engineering can control such vast amounts of water once they are put in motion. Landslides damming valleys and proglacial lakes are a common feature of mountainous areas, and the latter are becoming an increasing concern, as they are growing in size and number.
“Can we predict such events? Unfortunately, not all lake outburst floods have precursor signs. High risk sites may be monitored, and action can be taken in some circumstances, such as when a landslide dam was breached with diggers following the 2008 Wenchuan earthquake to prevent the lake from filling up to its highest level. However, many of these sites are inaccessible. Remote sensing can help, and this is probably where most progress can be made. But we need to be able to identify the trigger, for example, an unstable glacier or a failing slope. Once the trigger has been pulled, there is very little time: communities close to the outburst site will be affected within second or minutes.
“Further downstream, early warning may save countless lives, but communities may still have only tens of minutes to reach higher ground which, in some cases, could mean 30-40 metres above the river! Fast, clear and efficient communication, collaboration, trust, clear escape routes and shelters are essential.
“I have heard that communities in Nepal received some warning; this will have saved many lives. This is progress brought by technology. In 1786, an earthquake caused a large landslide that dammed the Dadu River in Sichuan, China. Behind the dam, a lake filled up for 9 days, then failed, causing 100,000 deaths across cities downstream, according to historical records. I am confident that such a disastrous outcome is unlikely in our current connected age. Outburst floods still represent a major hazard for mountainous communities though, and the Nepal disaster is a stark reminder of how quickly such a disaster can unfold. My thoughts are with all those affected.”
Dr Fatima M. Pillosu, hydrometeorologist and flash floods researcher from the Department of Meteorology, University of Reading, said:
“A barrier lake forms when a landslide or flood dumps enough debris into a valley blocking a river. Hence, water ponds behind a natural dam of loose rock and mud.
“Unlike an engineered dam, this barrier has no spillway and no solid core. So, dams of this kind often fail, many within days or weeks of forming. The most common type of failure is due to rising water spilling over the crest (overtopping) and the flow begins to cut down through the loose debris. If that erosion accelerates, much of the stored water is released as a single flood wave that carries mud and boulders down the valley at high speed and gives people very little time to react.
“The exact moment of failure cannot be predicted, because it depends on the internal structure of a dam. However, close monitoring of the water flowing in, the lake level and the seepage through the barrier can show whether conditions are stabilising or deteriorating, and can give hours of warning if a breach begins.”
Dr Jeff Da Costa, Researcher in Hydrometeorological Hazards, Early Warning & Disaster Risk, University of Reading, said:
“The immediate challenge is managing uncertainty. We may not be able to predict exactly when or how much water will be released. Monitoring can show how conditions are changing, modelling can estimate where the water could go and how quickly, and engineers can assess options to reduce the risk. That information can then be used to warn communities, move people away from exposed areas and keep rescue teams out of danger.”
“The first disaster has changed the risk downstream. Roads and bridges have been destroyed, communications disrupted and rescue operations are still underway. An evacuation route that was available before the flood may no longer exist. A second flood does not have to be as large as the first to have serious consequences, so warnings now need to reflect what has changed on the ground.”
“This is a transboundary river system, and information about the barrier lake is being shared between China and Nepal. Observations upstream can give people downstream valuable time, but they become useful when they are connected to modelling, warnings and practical decisions on the ground. With a rapidly developing flood, even a short period of warning can make a real difference.”
“We still do not know the full warning timeline for the original disaster, and we should be careful about judging the warning system from the death toll. Warnings and precautionary action may have saved many lives. We need to establish what was detected, when information was shared, who received it and what action followed. That reconstruction matters because it tells us what worked as well as what can be improved.”
Comments sent out Thursday 27th August:
Prof Hannah Cloke OBE, Regius Professor of Meteorology and Climate Science, University of Reading, said:
“When a disaster unfolds on this scale, keeping people safe from harm is complex task, and forecasting the hazard is only part of the challenge. The chain of tragic events along this river show that there were three distinct zones of risk, each with its own lessons for early warning systems.
“Highest up the valley near the source of the flood, the floodwaters arrived within minutes. In these environments, there is often very little time between the triggering event and the impact, making warnings exceptionally difficult, no matter how advanced the technology.
“Further downstream, many miles away, the situation was different. While the floodwaters caused severe damage to property, there were several hours between the first warning of a flood and the arrival of the torrent. According to Nepal’s forecasting authorities, those warnings helped save many hundreds of lives, demonstrating the value of effective forecasting and communication.
“The greatest challenge lay in between. This was a flood that arrived beneath clear skies. What came down the valley was not simply water, but a devastating torrent of water, mud and debris. In these circumstances, very often many people struggle to believe the severity of a risk because there was no local rainfall. For many, a flood is something that follows a storm. When a warning describes a threat that does not match people’s mental picture of danger, it can be easy to dismiss.
“This highlights a critical but sometimes overlooked weakness in early warning systems. Success is not determined solely by the accuracy of forecasts, the quality of sensors, or the reach of communication networks. It also depends on how people interpret, trust and act upon the information they receive.
“This imagination gap can come between people who are in danger hearing a warning and truly understanding what it means. If people cannot visualise a danger they have never experienced before, they may underestimate the risk, even when the warning is accurate and timely.
“As forecasting science and technology continue to improve, we will become increasingly capable of detecting and predicting extreme events. But better predictions alone will not eliminate disaster risk. Early warning systems must be designed around human behaviour as well as physical hazards. They need to anticipate how people perceive risk, how they make decisions under uncertainty, and how they respond to unfamiliar threats.
“The most effective warning is not one that is simply heard. It is one that is understood, believed and acted upon. Unless we close the imagination gap between issuing a warning and inspiring a response, we will continue to lose lives even when the information needed to prevent those tragedies is available.”
On reports of link with Glacier Collapse:
Dr Richard Waller, Senior Lecturer in Physical Geography, Keele University, said:
Do we know that glacial collapse caused this? How is that established? What are the signs?
“This is what is emerging from repeat satellite imagery that can provide before and after images. Here’s one of the best posts I’ve seen the demonstrates the area affected and the loss of a substantial part of a glacier terminus [1]”.
What is glacier collapse and what does it do?
“It involves the catastrophic failure of a mountainside that features a glacier. In some cases, glaciers themselves have failed catastrophically [2]. But in this case it looks like a steep flank of the mountain collapsed taking a large part of the overlying glacier with it. The subsequent melting of the glacier ice provided the meltwater for the flood / debris flow which it seems travelled for over a 100 km down the catchment which is remarkable. The Kolka-Karmadon event in the Caucasus Mountains in 2002 provides an earlier example in which a rockfall hit a glacier causing it to shear off the mountainside resulting in a similar if not quite so catastrophic event. [3]
What’s the climate link with glacier (in)stability?
“The progressive loss of snow and glacier ice results in the surface being less reflective, so it absorbs more solar energy and heats up. In combination with atmospheric warming, this is leading to the melting of mountain permafrost. Think of the high mountains like this as shattered bedrock glued together by ice-filled joints. As the permafrost and the ice-filled joints melt, then there’s the potential for these types of catastrophic failure.
Where in the world are glaciers most vulnerable? What’s the likely consequence of rising temperatures?
“Any high mountain areas with glaciers and permafrost – e.g. European Alps, Caucasus, Himalayas, Andes etc. In terms of impacts, this is one of the more catastrophic impacts of rising temperatures in high mountain areas. Others include the failure of ice or moraine-dammed lakes in what are called “glacier outburst flood events” (GLOFs). E.g. [4] “
1 – https://lnkd.in/p/eXdYbHnW
2 – https://tc.copernicus.org/articles/15/1751/2021/tc-15-1751-2021.html
3 – https://science.nasa.gov/earth/earth-observatory/collapse-of-the-kolka-glacier/
4 – https://www.nature.com/articles/s44304-026-00168-w
Prof Sarah Boulton, Director of the Centre for Research in Natural Hazards and Risk Reduction, University of Plymouth, said:
“The devastating event in Nepal is an evolving situation. But at the present time it appears that part of a glacier may have broken away and failed in an avalanche which has then travelled downslope, picking up more material as it went and transforming into an extremely large and devastating flood. This failure caused the seismic energy that was first identified and ascribed to an earthquake, but it was subsequently found to be that the failure itself has led to this seismic signal.
“It is an event that appears to be similar to the Chamoli event that took place in neighbouring India in 2021, where a large part of glacier also broke away and transformed into an avalanche. The subsequent debris flow there also caused significant damage in the river catchment, though it is likely that the event that occurred yesterday will be found to be larger than the 2021 event.
“These catastrophic flood events are not unknown and have occurred in other mountainous regions, including in the Himalayas, in the past. But it is likely that the climate emergency is causing these events to become more frequent and/or more extreme owing to rising temperatures and melting glaciers in high mountain regions.
“This area of Nepal is characterised by extremely steep slopes, high reliefs, and frequent land sliding during the monsoon season (which runs from June to September each year). The region is remote and access is difficult, factors that make preparedness and rescue efforts in the area difficult.”
Prof Jeffrey Kargel, Senior Scientist, Planetary Science Institute, said:
“We know that a glacier close to Langtang Lirung peak and a mass of bedrock detached from the upper north slope. It is a small “hanging glacier” that lost its lower part. The mass descended the first 22 km at an average speed of 193 km per hour, traversing that first 22 km in just 6 minutes and 50 seconds. The ice melted on the way down by “friction,” you may call it (conversion of potential energy into kinetic energy and then thermal energy, which melted the ice). The initial detected rock and ice mass is somewhere above 50 million cubic meters, maybe 200 million cubic meters–even more by one estimate, according to the M5.2 seismic energy release (earlier estimates of M4.4 were based on an assumed tectonic earthquake, but that proved incorrect so a revision of the energy was needed. The debris mass then picked up sediment along the way (deposited earlier by an unrelated glacial lake outburst flood), and then sediment plus water when it reached the Bhote Koshi River. So, instead of diffusing down, this event grew in magnitude as it flowed. It eventually did mix with enough water and reached lower gradients so that sediment could begin to settle. In any case, the flood reached the India border by yesterday evening, and now its remains are in India.
“This event is indirectly related to climate change, especially if the preceding glacial lake outburst flood produced sediment deposits on yesterday’s flood pathway that then were remobilized, making this recent event even worse. Furthermore, there is a possibility that a warming climate had produced some amount of freeze-thaw activity, which could wedge open crack in the rocky mass, but this has to be worked out quantitatively to see if it was a likely factor. Perhaps more specific to the climate angle is that the glacier has retreated due to climate warming, and this retreat yanked physical supporting mass out from the lower parts of the glacier, leaving it “hanging” and susceptible to this collapse event.”
Dr Hamish Pritchard, Glaciologist, BAS, said:
“Reports suggest that this disaster was caused by a large ice and snow avalanche from the north slope of Langtang Lirung, a 7000 m mountain. This happened in the middle of the monsoon season when heavy snow and rain fall in this region, making this the peak season for avalanches. I have temperature sensors in and next to a lake lying at 5100 m about 10 km away and they show that today has the highest lake temperature of the year, and two days ago was the highest ground temperature in the last two years, so this is the warmest time of year in the high mountains. These high temperatures would have weakened the snowpack, filled crevasses with water and thawed the bonds between ice and rock that hold these glaciers in place. The heavy rain would have saturated the soil and filled the rivers downstream, setting the scene for such a disaster, and these conditions are made much more likely by climate change that is causing rapid retreat of Himalayan glaciers. A similar, smaller flood hit the same river last July, and the 2015 Nepal earthquake released a major avalanche down the south face of the same mountain, which destroyed Langtang village, with many fatalities. More avalanches and more floods are inevitable, particularly as lakes grow in front of the retreating glaciers, meaning that early warning systems are urgently needed to protect the people living in these valleys.”
Prof Dame Sarah Springman DBE FREng, Principal of St Hilda’s College Oxford and a Fellow of the Royal Academy of Engineering, said:
“When large volumes of rock or soil are destabilised either by ice thawing in rock joints or water infiltrating into steep soil slopes through rainfall or snowmelt, they exchange potential energy like elevation for velocity. If sufficient water is present to cause a debris flow, then the flowing mass behaves as a heavy fluid and is very mobile. It then erodes more debris from the sides and base of the valley it is flowing through. Eventually, it will drain, and the solids will settle forming debris flow cones – this becomes more complicated when the debris flow meets infrastructure.
“These events are difficult to predict so early warning systems are essential. One can either monitor unstable slopes, which is almost impossible in a country like Nepal, or measure water heights in the streams feeding the rivers with alarm values to mobilise the local authorities. Then procedures for evacuation and seeking higher ground can take place, but the scale in this case was greater than what we have seen before and the impact has become more like a mega event such as in a Richter scale 8 earthquake.”
Prof Tristram Hales, Professor of Environmental Hazards, School of Earth and Environmental Sciences, University of Cardiff, said:
Do we know that glacial collapse caused this? How is that established? What are the signs?
“The evidence for this being a glacial collapse comes from satellite imagery. Colleagues of mine utilised a satellite called Planet to compare images of the glacier above this area before and after the event. Most of the glaciers in this region are brown and covered in dirt and rocks. In the imagery on the 26th of August, a large blue streak appears on the glacier, showing where the glacier has collapsed and clean ice has been exposed. This evidence supports a potential glacial collapse origin for this event. There was a small earthquake recorded as well, but it is not clear whether the collapse itself generated what was recorded on the seismometers as an earthquake. It is very early, so there are still a lot of questions about what exactly caused this event.”
What is glacier collapse and what does it do?
“Glaciers are largely composed of ice. They can be tens of metres to several kilometres thick in this region. At the front of the glacier the slope of the glacier is steep, putting strain on the ice and causing it to break and fracture (these create cracks called crevasses). Additionally, as it is warmer near the front of the glacier, there is often significant meltwater present that can further melt in these cracks and apply pressure to the cracks that can cause them to break more. These processes can combine at glacier fronts to cause a catastrophic collapse, something like an ice landslide, which we call a glacier collapse.”
What’s the climate link with glacier (in)stability?
“Warmer temperatures can enhance the potential instability of glaciers in two ways. Firstly, the melting of the glaciers increases the amount of meltwater which can enter the cracks and cause greater instability at the base of the glaciers. Secondly, there is more water present in the rivers downstream of the glacier. While we still do not know the exact mechanism of this particular event. Previous, similar events in the Himalaya have sucked water from the rivers into the flow, making them bigger. Simply put, if the climate was cooler, more of that water would be locked in ice and it would be less likely for this event to become as large as it did.”
Where in the world are glaciers most vulnerable? What’s the likely consequence of rising temperatures?
“The short answer is everywhere in the world. We have seen a dramatic increase in the number of extremely hazardous flows such as this globally in the past decade or so. In the past five years in the Himalayas, we have seen major events at Chamoli, where a glacier collapses damaged hydroelectric power infrastructure and created a dangerous flood, and at Melamchi where a flood inundated a town and resulted in a number of fatalities. Two years ago, a glacier collapse in the Swiss town of Blatten caused one fatality and significant long-term damage. The reason for an increase in the number of damaging flows is twofold. Firstly, increased glacial melting is increasing the frequency of glacial collapses and other similar phenomena. Secondly, the amount of infrastructure that is vulnerable to these flows is increasing. We see more people in harm’s way and more important infrastructure being built in the potential pathways of these flows.”
Dr Ella Gilbert, Climate Scientist, British Antarctic Survey (BAS), said:
“Early analysis suggests that the tragic disaster in Nepal was triggered by a chunk of ice falling away from an upstream glacier, dragging mud, ice, and rock downstream where it entered the Lhende Khola river and caused the flash flood and mudslide that claimed hundreds of lives.
“We can’t say definitively yet what the links to climate change are, but we know that climate heating is destabilising glaciers in the Himalaya, with the rate of loss doubling since 2000. When glacier stability is undermined, sudden failures can occur, leading to horrible incidents like this one.
“Mountain glaciers worldwide are under threat from climate heating – particularly those in tropical regions. The higher temperatures rise, the greater the risk of sudden and devastating collapse.”
Prof Bethan Davies, Chair in Glaciology, Newcastle University, said:
“Firstly, let me say that I am appalled by the footage and imagery coming out of Nepal and Tibet, and the sight of people running for their lives.
“The exact drivers and triggers of this event are still unfolding, and it will take a few days to comprehensively understand them. However, it seems that a mixture of ice and rock collapsed in the glaciers north of Langtang-Lirung mountain. This ice and rock then flowed into a river, turning into the devastating debris flow and flash flood that we see in the CCTV footage at the border. This caused the complex multi-hazard cascade and was responsible for the power of the flood. Ice, rock, debris flow and flooding interacted to make the event so powerful. The earthquake initially reported seems to be a consequence of the glacier collapse.
“Ice detachments, or glacier collapse, events such as these have previously occurred in this and in other high mountain environments, and they are a significant hazard. The glaciers of Langtang are currently retreating rapidly, as shown by Gunjan Silwal’s PhD research [1]. Glacier loss rates have been accelerating, especially after 2000, with rising snowlines increasingly intersecting with areas of steep ice, causing glacier recession, fragmentation, and warming. From 2000-2023, the catchment lost ice at a rate 1.5 times faster than the rate from 1964-2000.
“Glacier collapse events have occurred regularly in recent years, for example, Birch Glacier in Switzerland and Chamoli glacier in 2021 in Indian Himalaya, or the 1970 Huascaran avalanche in Peru, which cost 6000 lives. early signs suggest that this may have been a similar event to the collapse at Chamoli glacier. The area has steep slopes, high topographic relief, and seismic activity that make it prone to mass movements.
“A warming environment can make these kinds of events more common or likely. Melting permafrost can make mass movements onto glacier surfaces more likely (as seen in the Blatten landslide). Glacier detachments can occur when ice previously frozen to the bed reaches the ice-bed interface, and allows the ice to start sliding. On steep slopes, that can contribute to glacier detachments and the catastrophic type of events seen here.”
1 – Silwal et al., 2026, https://www.sciencedirect.com/science/article/pii/S0921818126002766).
Prof Liz Stephens, Professor in Climate Risks and Resilience, Department of Meteorology, University of Reading, said:
“Most people think of flash floods as being caused by heavy rainfall, but in Nepal and Tibet, as in many other high-mountain regions, they can also result from complex chains of hazards, including landslides, avalanches and glacial lake outburst floods. This complexity makes early warning especially difficult.
“Initial analysis of before and after satellite images now suggests that part of a glacier collapsed high in the mountains. This huge ice-rock avalanche appears to have blocked the river, and the failure of that debris dam led to catastrophic flooding downstream. This avalanche of debris will have picked up more material as it travelled down the mountain.
“The resulting flooding is likely to have caused considerable erosion, undercutting slopes as it travelled downstream, picking up more debris and leaving behind a very unstable landscape. Scientists in the region will be concerned that the initial event could lead to more dangerous landslides and debris dams over the coming days.
“The exact trigger for the glacier collapse and the precise chain of events may not be known for some time. In-situ assessments will be extremely difficult, and what we can see from satellites is hampered by monsoon clouds.
“Climate change is rapidly altering high-mountain environments. The IPCC report has high confidence that glacier retreat and permafrost thaw are reducing the stability of mountain slopes and increasing the potential for complex cascading events such as this. However, the exact contribution of climate change to this particular event may be difficult to determine.”
Dr Wouter Buytaert, Reader in Hydrology and Water Resources, Imperial College London, said:
“While scientists are still studying the exact causes of this disaster, it is exactly the type of event that is triggered by climate change. Global warming shrinks glaciers, making it more likely for them to break apart. Land degradation, due to increasing wildfires, extreme weather, and unsustainable land use, destabilizes valley slopes and mobilizes rocks and sediment. In mountain regions with steep slopes and narrow valleys, this creates the perfect conditions for events such as this.”
Previous comments:
Dr Fatima M. Pillosu, hydrometeorologist and flash floods researcher from the Department of Meteorology, University of Reading, said:
“Everyone in a mountain valley faces the same hazards, but tourists are more vulnerable because they may not recognise the danger or understand local warnings. Visitors from regions where such events are rare may not appreciate how fast a mountain river can rise, nor that these valleys also face avalanches, landslides and floods from melting glaciers, even in dry weather. Knowing in advance what can happen, and what to do if it does, is one of the most useful protections.
“A flood thick with mud and rock can be even more lethal than a rain-fed flash flood. Both can strike with great force and speed and carry boulders and debris along. Fast flowing water can easily pick up anything in its way and carry it off downstream. Mud simply buries people. The moving mass is much heavier than water, and it grows heavier as it travels, scouring soil, gravel and boulders from the riverbed and banks and carrying them along.
“The boulders it carries can destroy buildings and bridges that a water flood might have left standing. When the flow stops, the mud and rock settle into a dense mass that is hard to dig through, making rescue even slower and more difficult.”
Dr Jeff Da Costa, Researcher in Hydrometeorological Hazards, Early Warning & Disaster Risk, University of Reading, said:
“For early warning, you do not always need to know exactly what has happened upstream before taking action downstream. A sudden and unusual change in the river can be enough to warn communities while the cause is being established.
“This is particularly important because the same river corridor experienced a major flood originating in Tibet in July 2025. After that event, Nepalese flood forecasting officials said they had received no advance warning from China and that there was no mechanism in place for sharing that information. Nepal subsequently committed to strengthening coordination with China. We do not yet know what information was available before today’s flood, when it became available or how it moved across the border. Establishing that timeline will be important, including what changed after the 2025 event and how much time communities had to act.
“Early warning in transboundary river basins depends on international cooperation. Monitoring can take place in one country while communities at risk are across the border. Rapid information sharing and trusted communication between neighbouring countries are part of the warning system. This has wider relevance. Many communities around the world depend on environmental information from across a border. Where relations are unstable or cooperation is limited, maintaining those channels becomes critical.
“Events like this can be extremely difficult to predict precisely. Warning systems need to work under that uncertainty by recognising danger quickly, sharing information and having agreed actions that can begin before the exact cause is known.”
Prof Liz Stephens, Professor in Climate Risks and Resilience, Department of Meteorology, University of Reading, said:
“Most people think of flash floods as being caused by heavy rainfall, but in Nepal and Tibet, as in many other high-mountain regions, they can also result from complex chains of hazards, including landslides, avalanches and glacial lake outburst floods. This complexity makes early warning especially difficult.
“Initial reports from Nepal suggest that an earthquake occurred around the same time as a landslide or avalanche, which then led to the catastrophic flooding. One way this can happen in steep mountain valleys is when a landslide blocks a river, allowing water to build up behind it. If that landslide dam, then fails, it can cause a sudden, devastating release of water and sediment downstream.”
From the NZ SMC:
Associate Professor in River Science Jon Tunnicliffe, University of Auckland, said:
“First, our thoughts are with the families of those killed and still missing, and with the Nepali and Chinese agencies and communities carrying out the search. The toll includes local residents, workers, pilgrims and travellers from many countries, and it will take time to establish the full scale of the disaster.
“Although this occurred at the height of the monsoon season, significant rainfall does not appear to have been the trigger. Preliminary investigations instead point to a large collapse of glacier ice and rock high above the Lhende Khola. Satellite imagery indicates that the material entered the valley from around 5,200 metres elevation, and investigators are examining whether debris temporarily blocked the river before releasing a much larger surge downstream. River levels at Galchhi reportedly rose by as much as nine metres within thirty minutes.
“That sequence is what geomorphologists call a hazard cascade: instability high on a glacierised slope becomes an ice–rock avalanche; that can block a river, entrain water and sediment, and transform into a debris-laden flood travelling tens of kilometres downstream. Each process is reasonably familiar. The danger lies in the connections between them. A localised failure at more than 5,000 metres can very rapidly become a catastrophe for people far down the valley.
“There is an important climate-change context here, but we need to be careful about attribution. We do not yet know why this particular glacier failed. What we do know is that conditions across the Hindu Kush Himalaya are changing rapidly. Glaciers are retreating, permafrost is degrading, new lakes are forming and steep slopes are being exposed and destabilised. Climate change can amplify hazards that already exist by altering the thresholds at which one process triggers another.
“Sediment is a crucial part of why these cascades are so destructive. Once a surge begins to entrain rock, gravel and boulders, it can grow enormously downstream. It is no longer simply a flood of water, but a fast-moving mixture capable of destroying bridges and buildings and radically reshaping the river channel. Narrow mountain valleys concentrate that energy, while roads, settlements and hydropower infrastructure are often concentrated on those same valley floors.
“The danger also does not necessarily end when the initial surge passes. Fresh deposition can block tributaries, raise riverbeds, divert channels and create new temporary dams. A major event changes the landscape in ways that can generate further hazards over the following days, weeks and sometimes years.
“The broader lesson is that we increasingly need to manage hazard chains rather than individual hazards. Satellite monitoring can identify changing glaciers, lakes and unstable slopes and quickly establish what has happened after an event. That needs to be combined with seismic monitoring, river gauges and warning systems capable of detecting a cascade once it begins.
“Hazard models likewise need to route sediment as well as water, because runout, erosion, deposition and channel switching often determine where the damage occurs. And planning matters just as much as prediction. We will never monitor every unstable slope, so communities need mapped runout corridors, evacuation plans, appropriate land-use controls, and resilient roads, bridges, power and communications.
“Finally, these are international problems. Mountain catchments cross borders, and observations and warnings need to do the same. New Zealand lives with its own versions of cascading mountain hazards — slope failures, landslide dams, debris floods and large sediment pulses moving through river systems. The settings and scales differ, but the underlying lesson is the same: the hazard is often not the first event. It is what that event sets in motion.”
Dr Lauren Vargo, Glaciologist, Antarctic Research Centre, Victoria University of Wellington, said:
“A glacial lake outburst flood is the release of water when the dam of a glacial lake fails, which could be triggered from ice breaking off and falling or avalanching down a valley. It’s unclear if this was the cause of the flood.
“From what I’ve read, scientists who have looked at aerial imagery can see that it looks like a chunk of ice fell from the mountains to the valley floor.
“The location where the seismic activity was recorded is where there are lots of glaciers, but no glacial lake downstream. However, there are other glacier lakes nearby, so without seeing the recent satellite imagery to know exactly where the ice fell from, it’s hard to say.
“We can look at satellite imagery taken before and after the event, and try and see what has changed. Essentially looking for, what is missing from the mountains in the ‘after’ imagery. If we can see if there is ice missing that was there in the ‘before’ imagery, that tells us it was likely ice (likely with some water and rock with it).
“I’m not sure what kind, if any, warning systems exist in this region. My understanding of how this works is that once a breach or surge is detected, signals are sent downstream. The closer communities and infrastructure are to the source of the flood, the less time they would have to respond.
“Theoretically, warming temperatures could melt snow and ice more, making these events more likely. However, I’m not sure if research has been done to formally link increasing temperature and an increase in likelihood of these events. Downstream warning systems can help communities prepare for events like this. We can also monitor the slopes – my understanding is that this is how a community in Blatten, Switzerland, knew there was increased rock/ice avalanche risk in 2025, and evacuated the community downstream before the avalanche occurred. The biggest way to mitigate the risk would be to reduce our greenhouse gas emissions to limit warming.”
Dr Tom Robinson, School of Earth and Environment, University of Canterbury, said:
“The images and videos of massive flash flooding in Tibet and Nepal are horrific, demonstrating the power of such natural hazards. At this stage, there is still much uncertainty over what has happened. From the reports we currently have, it appears that a large landslide has occurred somewhere in Tibet causing part of a glacier to collapse and cascade down the valley for many tens of kilometres. The landslide and glacier collapse itself must have been extremely large – seismic networks in the region recorded the landslide impacting the valley floor as a magnitude 4 earthquake.
“Unfortunately, these kinds of events are not uncommon in the Himalaya. We’ve seen similar massive landslides and floods in the region before: in 2021 two separate events caused huge damage in Melamchi, Nepal and Chamoli, India. Linking these disasters directly to climate change is difficult, however we know that the Himalayan region is already experiencing dramatic effects from our warming world. Glaciers in the region are rapidly melting and destabilizing and when large landslides, which are common in the Himalaya, fall onto them, they can trigger huge rock and ice avalanches that can travel massive distances.
“The first order impacts from these types of events are obvious from the images and videos – tragically we should be prepared for large numbers of fatalities, with early reports already indicating several hundred people have been killed or are missing. There are also likely to be longer term effects as the flood has wiped out critical infrastructure in the valley, including roads, bridges and hydropower stations. Getting access and basic amenities to the affected region will be a significant challenge and it’s clear that this requires a major humanitarian response in one of the poorest countries in the world.”
From the Aus SMC:
Dr Soniya Rijal is a postdoctoral research associate in the School of Project Management at the University of Sydney, said:
“I’ve heard people were notifying neighbours to leave before the flooding hit, but because there was no rain leading to the flash floods, many of these warnings weren’t taken seriously. What makes this disaster unusual is that it was so sudden, coming without warning.
“The immediate priority should be to search for missing people, and local communities can help do this. Nepal is a very collective society, and locals will have a better idea of how to locate missing people than those without local knowledge.
“Shelter and food will be essential to meet immediate needs. Local communities can work with agencies to oversee the logistical hurdles to deliver essential goods and services. They will also be key to mapping the location of available shelters and storage facilities, e.g. schools, gumbas (Buddhist monasteries) in the immediate aftermath of this disaster. These facilities will be important distribution centres for essential goods and services, and serve as community hubs.”
Associate Professor Nader Naderpajouh leads the Organising Risk and Resilience research lab and is Head of the School of Project Management at the University of Sydney, said:
“This is a very unique disaster, in terms of intensity and speed, which had a distinct blueprint of impact on communities, and activities around immediate response and recovery. The reports suggest the flood and landslide were caused by a buildup of water in the glacier, which was triggered by seismic activity.
“This unique event may trigger the implementation of future warning systems, which is feasible with current state-of-the-art technology. At least the buildup of water can be an early warning sign for downstream communities, a form of seasonal warning for preparation. Such communications and preparations can save lives and reduce the impact at scale.
“With global warming, there is an expectation that similar patterns of disasters will recur. We can benefit from the implementation of a warning system, similar to those triggered for tsunami warnings after 2004’s tsunami.
“The long-term recovery can focus on connecting communities isolated by this disaster. In essence, the priority is to provide baseline infrastructure and connections, so the close-knit communities Soniya mentioned can self-organise their response.”
Mr Andrew Gissing is the CEO at Natural Hazards Research Australia, said:
“The flash flood disaster in Nepal is tragic. Though the exact cause of the event requires further investigation, the steep and narrow valleys in the region would contribute to its devastating nature of deep, fast flowing and debris laden water.
“The Himalayan Region has had a history of deadly flash floods. The event appears to have occurred without warning, catching those downstream unaware. Research shows that early warning saves lives and there are efforts globally to improve early warning systems.
“Attention of authorities will now shift to downstream search and rescue and relief for communities impacted.”
Dr Adrian McCallum is Discipline Lead in Engineering at the University of the Sunshine Coast, and a Glaciologist and Polar Engineer
“I’m yet to fully appraise myself of the situation, but these floods are probably ‘glacial lake outburst floods’ (GLOFs); these are increasingly common in alpine regions globally as our climate changes.
“As glaciers melt and recede, ice turns to water, the water is stored (temporarily) behind natural moraine dams, but when those dams breach, tremendous amounts of water can be released down a valley, taking out everything in its path.
“I was meant to be in the Indian Himalaya now, studying GLOFs, with Indian glaciological colleagues.”
Declared interests
Prof Maria Shahgedanova: “no conflicts to declare”
Prof Reza Ahmadian: “no conflicts of interest”
Dr Matt Westoby: “no conflict of interest declared”
Prof Mikael Attal: “no conflicts of interest”
Prof Hannah Cloke: “Hannah Cloke advises the Environment Agency, the European Centre for Medium-range Weather Forecasts, the Copernicus Emergency Management Service, local and national governments and humanitarian agencies on extreme weather and the forecasting and warning of natural hazards. She is a fellow of the European Centre for Medium-range Weather Forecasts. Her research is funded by the UKRI Engineering & Physical Sciences Research Council, the UKRI Natural Environment Research Council, the Foreign, Commonwealth & Development Office and the European Commission.”
Prof Sarah Boulton: “no conflicts of interest declared.”
Dr Hamish Pritchard: “I have not received any industry funding for my research. I am the named inventor for a patent-protected method to measure snowfall which I am using at my study site in Nepal, though this IP is owned by the Natural Environment Research Council and has not yet generated income. It is not an early warning system for avalanches or floods.”
Prof Dame Sarah Springman: “I am an emeritus professor from ETH and a former member of the Platform for Natural Hazards PLANAT in Switzerland and am now Principal of St Hilda’s College and a Royal Commissioner of the Royal Exhibition of 1851. No conflicts of interest regarding research and funding.”
Prof Tristram Hales: “No interests to declare.”
Dr Ella Gilbert: “freelancer, currently working with Royal Meteorological Society, honorary researcher at British Antarctic Survey”
Prof Bethan Davies: “I don’t have any declarations of interest to declare.”
Dr Jeff Da Costa: “I have no conflicts of interest to declare. “
Prof Liz Stephens: “No conflicts. Liz also works for the Red Cross Red Crescent Climate Centre.”
Dr Fatima M. Pillosu: “No conflicts of interest”
Associate Professor Jon Tunnicliffe: “Not yet received.”
Dr Lauren Vargo: “None.”
Dr Tom Robinson: “No conflicts to report.”
Dr Soniya Rijal: “Soniya has not declared any conflicts of interest.”
Associate Professor Nader Naderpajouh: “Nader has not declared any conflicts of interest.”
Mr Andrew Gissing: “Andrew has not declared any conflicts of interest.”
Dr Adrian McCallum: “Adrian has declared no conflicts of interest.”
For all other experts, no reply to our request for DOIs was received.