An international analysis published three weeks after the Nepal collapse links warming, thawing permafrost and thinning ice to the mountain’s instability. It stops well short of saying that climate change caused the disaster, and the difference is the point.
Within days of the 26 August collapse, the question that follows every large mountain disaster was being asked: was it climate change? The answer published on 17 September by an international team, led by Imperial College London and hosted by World Weather Attribution, is more precise than a yes or no, a distinction easily lost in headlines. It describes climate change as “a destabilising factor acting on a pre-existing geological predisposition, rather than the fundamental cause of the failure”.
Cause, contributor, predisposition
Three words carry the argument. A cause is something without which the event would not have happened. A contributor raises the probability or severity of an event that had other origins. A predisposition is a standing condition, here the steep, fractured rock of the north face of Langtang Lirung, which made a failure possible in the first place. The analysis places climate in the second category, acting on the third.
The authors are explicit about the limit of what they did. They write that they “have not assessed whether this specific rock-ice avalanche would have occurred in the absence of human-induced climate change”. Nor did they conduct a conventional single-event attribution, the kind used for heatwaves or extreme rainfall. They integrated evidence from glaciology, mountain hydrology, climate science, seismology and social science, with researchers from Nepal, Pakistan, the United Kingdom, Ireland, Sweden, Denmark, Norway, the United States, New Zealand and the Netherlands, and partners including the Dutch meteorological institute KNMI and the Red Cross Red Crescent Climate Centre.
What the analysis measured
On temperature, it estimates that human-caused warming increased July and August temperatures in the region by about 1.5°C, comparable to the level of global warming, and annual temperatures by about 2°C. In individual winter months the increase reached as much as 3°C. The freezing level has risen by roughly 100 metres a decade in the monsoon and post-monsoon seasons, and a greater fraction of precipitation now falls as rain rather than snow at higher elevations.
The freezing level, the altitude above which air temperature falls below 0°C, marks roughly where precipitation switches between rain and snow. Its rise of about 100 metres a decade shifts that line higher on the mountain. The 1.5°C figure refers to the temperature increase attributable to human-caused warming in the region, not to a measured rise in the temperature of the rock itself.
Why single events are hard to attribute
Conventional attribution studies compare how likely an event is in today’s climate with how likely it would have been in a world without human-caused warming, using long observational records and large model ensembles. That works for heatwaves and heavy rain, which recur often enough to be counted. A rock-ice avalanche of this size has no comparable record, the physics of thaw and fracture inside a rock wall is poorly represented in climate models, and the outcome depends on the exact geometry of one slope. The analysis therefore takes a different route, asking which climate-driven changes plausibly acted on the mountain and how large they were.
On the ice, the Langtang Lirung glacier has thinned by more than half a metre a year over decades, and its retreat has accelerated from about 0.5 per cent a year across the previous two centuries to between 1 and 2.3 per cent a year over the past 16 years. The months before the collapse were unusual: exceptionally high precipitation in October 2025, then a warm 12 months from September 2025 to August 2026, with July and August the warmest two of the year.
An independent caution
Other scientists have urged restraint on the same evidence. The climate researcher Zeke Hausfather wrote that formal attribution of this specific collapse was premature, noting that single-event attribution of rock-ice avalanches typically needs detailed forensic work over months or years. He pointed to precedents that complicate a simple reading: the 2015 Langtang avalanche was triggered by an earthquake, and the 1895 collapse of the Altels glacier in Switzerland, involving about four million cubic metres, occurred before significant industrial warming.
The point is not that climate is irrelevant. Hausfather’s argument, like the analysis, is that a warming mountain makes such failures more likely, and that confident claims in either direction outrun the evidence. ICIMOD has likewise cautioned that it is premature to state climate change’s specific role in the August event.
Why the distinction matters
For readers, the difference between “climate change caused this” and “climate change raised the odds” is not pedantry. It determines what can be prevented. If warming is a destabilising factor acting on rock that was already vulnerable, then the risk is a combination of the mountain and the climate, and reducing it involves both.
The practical rule for readers and reporters is to ask what a source is actually claiming: a probability, a contribution or a cause.
The analysis draws a further conclusion about adaptation. It finds that the event was beyond the design and predictive limits of existing risk-reduction measures in Nepal, and that no existing early warning system could have provided enough lead time or prevented the scale of impacts. Its authors add a policy view, that minimising future risk requires a rapid transition away from fossil fuels and delivery of climate finance commitments, which is an argument distinct from the scientific findings above. What remains open is quantitative. More detailed glacier and permafrost modelling could say how much of the risk on this particular slope is due to human-caused warming. Until it does, the accurate description is the analysts’ own: a destabiliser, not a proven cause.
G S Ganesh
