This summer the same story keeps appearing with a different cast. The normal route up Mont Blanc is shut, the Zermatt guides have cancelled every Matterhorn ascent, a block of rock crashes down at Oeschinensee and tourists run from the shore. Add the routes that had to be moved: on Piz Palü above Diavolezza, on the Lagginhorn above Saas-Grund.
Nearly every report contains the same word, usually in a subordinate clause: permafrost. Rarely does it say what that actually is, or why its thawing takes a mountain apart.
Here is an attempt at the explanation, with the numbers that exist. And with the part the reports usually leave out: not every rockfall is permafrost.
What permafrost is
Permafrost is neither ice nor soil. It is a temperature condition, defined by duration: ground that stays at or below zero degrees for at least two consecutive years.
It can be loose material, but in the Alps it is usually rock. Ice sits in its joints and fissures. That ice is the subject.
In the Swiss Alps the lower limit sits roughly between 2400 and 2700 metres depending on aspect. North faces are colder, so permafrost reaches further down; south faces need more altitude. Around five per cent of the country counts as permafrost terrain, and a good share of it is exactly where climbers go and where lift companies have anchored their pylons.
Why the ice holds the mountain together
A rock summit is not a monolith. It is a stack of blocks shot through with joints. What holds it together is partly friction between blocks and partly the ice between them. The ice does two different jobs, and this is the part you have to grasp:
First it glues. Ice in a joint has measurable shear strength. It holds the block above the joint in place, like an adhesive that sets when cold.
Second it seals. An ice-filled joint is watertight. Rain and meltwater run off the outside instead of entering the mountain.
When the ice thaws, both go at once. The glue is gone and the mountain becomes permeable. The second is often the more dangerous: water that enters a joint and cannot drain builds pressure. This joint water pressure acts as a wedge and can lever out blocks that held for centuries.
The municipality of Kandersteg puts it exactly that way for the Spitze Stei above Oeschinensee:
Advancing permafrost degradation, which increases the permeability of the rock to water, is likely to be an important reason for the destabilisation of the mountain flank.
Note what comes first in that sentence. Not the missing glue, but the permeability.
What is measured, and what comes out
Switzerland has monitored its permafrost in a coordinated way for about 25 years through the PERMOS network and its 23 borehole sites. It records temperatures at various depths, the thickness of the active layer, and the speed of rock glaciers.
The hydrological year 2024, October 2023 to September 2024, produced new record values at practically every site.
Two further findings are more concrete than any temperature curve:
The active layer is growing. That is the top layer which thaws in summer and refreezes in winter. Over the past two decades it has thickened by anything from a few decimetres to several metres, depending on site and ice content.
At Schilthorn, in winter 2024, it failed to refreeze for the first time at any PERMOS site. That is the finding to hang the whole story on. The winter was no longer enough to undo what the summer had thawed. Anyone standing on Schilthorn in winter is standing on ground that no longer freezes through.
Rock glaciers are accelerating. A rock glacier is an ice-rich mass of debris creeping downhill. In the early 1990s they moved a few decimetres per year. By 2024 it was several metres per year. That is not a doubling, it is an order of magnitude.
Why the danger lags behind the warming
This is the part that is least understood, and the part that makes the situation so awkward.
Heat needs time to penetrate rock. The 0.8 degrees at ten metres depth are the result of the last ten summers, not the last one. At twenty metres only half of that has arrived. And failures often occur not in the top, fast-reacting layer but along deep joint surfaces that are only now warming up.
Three things follow:
- Today’s rockfall belongs to the warming of the year before last. What breaks off this summer was not destabilised this summer.
- A cool summer ends nothing. The heat is already inside the mountain and keeps travelling downward regardless of current weather.
- It will increase even if warming stopped tomorrow. The heat already stored has still to arrive.
At the same time the current heat acts immediately, by a second path: it thaws the active layer deeper and delivers meltwater into the joints. That is why events cluster during heatwaves even though the underlying cause is years old. The slow process loads the weapon; the heatwave pulls the trigger.
Where this hits ski areas
For a ski area permafrost is not an abstract topic but a foundation problem. Top stations, pylons and anchors sit at altitudes where the ground ought to be frozen. When it is not, the anchoring loses its bearing capacity and the ground can creep.
This summer that was on display in Zermatt: summer skiing on the Theodul glacier stopped in late July because meltwater was threatening the stability of the pylons on Plateau Rosa. The problem was not the piste but what the installation sits in.
Two more contact points this summer lie directly above ski areas: on Piz Palü above Diavolezza the normal route had to be shifted west, and on the Lagginhorn above Saas-Grund the route was changed to avoid rockfall zones.
And now the caveat
It would be convenient to file every Alpine rockfall under permafrost. It would also be wrong, and Switzerland’s largest recent case shows why.
In the Blatten landslide of May 2025, the village in the Lötschental was largely destroyed by a debris and ice avalanche. The chain of events began with rockfalls from the Kleines Nesthorn. But that steep face lies more than 500 metres above the permafrost limit and faces north. How much permafrost thaw contributed there is unresolved to this day.
A note on the ski area in the same valley, since people ask: Lauchernalp lies down-valley from Blatten, and the lift company states plainly that access and ski operations are unaffected and never were.
Similarly at Oeschinensee: the Spitze Stei has been under observation since 2018, and the event of 8 August 2026 was small at around 1500 cubic metres. For comparison, roughly 10,000 cubic metres came down the same flank in August 2022, and the exclusion zone is dimensioned for failures up to 200,000 cubic metres. The mountain did not start this summer.
What holds is the trend. What does not hold is the shortcut “rockfall, therefore permafrost, therefore climate change” applied to every individual case. Mountains have always shed stone. What is new is the frequency, the altitude and the size, and that faces long considered reliable are now affected.
What it means in practice
For mountaineers the rule of thumb has shifted: the cold early season matters more, high summer is increasingly unusable. Classic high routes that used to work in July and August are moving into June, or into autumn after the first cold snap.
For hikers the risk stays small but not zero, and it concentrates below steep rock faces above 2000 metres. Closed paths there are not bureaucracy but the result of measurement: at the Spitze Stei the block that fell on 8 August had been tracked moving about six centimetres a day, and its collapse was announced in the situation report the day before.
For skiers little changes in winter. A piste sits on snow, and snow does not care about the permafrost beneath. What changes is the infrastructure: installations in permafrost terrain need more elaborate foundations, more frequent inspection and eventually replacement. That is a cost factor which weighs more heavily on small resorts than on large ones, and it appears on no day pass.
Sources
- PERMOS, Swiss Permafrost Monitoring Network: 23 borehole sites, 2025 assessment. Warming of more than 0.8 °C at 10 m depth over 2014 to 2025, about half of that at 20 m. Record values in the hydrological year 2024. Active layer at Schilthorn failed to refreeze for the first time in winter 2024. Rock glaciers from a few decimetres per year in the 1990s to several metres per year in 2024.
- Municipality of Kandersteg, Spitze Stei situation reports, August 2026. Failure of 8 August 2026 at 13:30, around 1500 m³; the block concerned moving about 6 cm per day, the flank as a whole about 1 cm per day; exclusion zone dimensioned for failures up to 200,000 m³; altitude 2200 to 2900 m.
- Blick, 24 August 2022, on the earlier failure of around 10,000 m³ on the same flank.
- On the role of permafrost in the Blatten landslide of May 2025: the affected face on the Kleines Nesthorn lies more than 500 m above the permafrost limit and faces north; the contribution of permafrost thaw is unresolved.
- On the 2026 closures and route changes: statements from the guides’ offices and reporting by SRF, NZZ and the German Alpine Club.
Where figures conflict, we use the authority’s number. The 10,000 cubic metres circulating in some reports about Oeschinensee belong to the 2022 event, not to this August.
