Long record

Since 1950: how natural snow has changed at 44 ski resorts

Modelled natural snow at every resort Cirque covers, winter by winter since 1950, from one consistent weather record. The biggest losses are at the villages.

The short answer

In Cirque’s model, the village of 39 of the 43 resorts with a snowy village had fewer days with 30 cm or more of natural snow in recent winters (1995/96–2024/25) than in 1961/62–1990/91, typically 20 fewer a winter. At mid-mountain the change was smaller, typically 4 days, because most mid-stations are still deep in snow for almost the whole winter.

These are modelled figures. Checked against snow measured at Swiss weather stations, the same model got the direction of change right at 18 of 19 stations but overstated its size, by about 41%. At 25 stations in the western United States it got the direction right at 22 and overstated the size roughly twofold. Treat the changes below as upper estimates, and the direction as the solid part. The model is also better at the overall picture than at telling neighbouring resorts apart, so small differences between resorts in the lists below mean little.

How it is worked out

Cirque runs its snow model on ERA5, the European reanalysis of the world’s weather, at each resort’s village and mid-mountain height for every winter since 1950. ERA5 is one consistent record for the whole period, so the change is not an artefact of switching sources. A figure is the days in a winter (November to April in the north, May to October in the south) with 30 cm or more of natural snow, averaged over thirty winters.

Villages: the biggest losses

1. Cortina d’Ampezzo, Italy (1,224 m): 83 days, then 22, 60 fewer.

2. Hemsedal, Norway (625 m): 93 days, then 38, 56 fewer.

3. Åre, Sweden (380 m): 80 days, then 31, 49 fewer.

4. Val Gardena, Italy (1,236 m): 52 days, then 6, 46 fewer.

5. Whistler Blackcomb, Canada (675 m): 78 days, then 35, 43 fewer.

6. Les Deux Alpes, France (1,300 m): 88 days, then 49, 39 fewer.

7. Ischgl, Austria (1,377 m): 126 days, then 87, 39 fewer.

8. St. Anton, Austria (1,304 m): 134 days, then 96, 38 fewer.

9. Chamonix, France (1,035 m): 82 days, then 46, 36 fewer.

10. Whakapapa (Mt Ruapehu), New Zealand (1,630 m): 76 days, then 40, 36 fewer.

11. Kitzbühel, Austria (800 m): 57 days, then 22, 35 fewer.

12. Mt Buller, Australia (1,375 m): 48 days, then 14, 34 fewer.

13. Revelstoke, Canada (512 m): 38 days, then 5, 33 fewer.

14. Falls Creek, Australia (1,500 m): 79 days, then 49, 30 fewer.

15. Obergurgl, Austria (1,793 m): 155 days, then 126, 29 fewer.

Left out because the village has too little natural snow in either period to measure a change: Cerro Catedral.

Mid-mountain

1. Mt Buller, Australia (1,590 m): 110 days, then 73, 37 fewer.

2. Mt Hotham, Australia (1,648 m): 93 days, then 64, 29 fewer.

3. Val Gardena, Italy (1,877 m): 163 days, then 135, 28 fewer.

4. Falls Creek, Australia (1,640 m): 114 days, then 88, 27 fewer.

5. Thredbo, Australia (1,701 m): 101 days, then 76, 26 fewer.

6. Perisher, Australia (1,830 m): 79 days, then 56, 23 fewer.

7. Kitzbühel, Austria (1,400 m): 159 days, then 142, 17 fewer.

8. Hemsedal, Norway (1,061 m): 164 days, then 150, 13 fewer.

9. Mt Hutt, New Zealand (1,746 m): 161 days, then 150, 11 fewer.

10. Cortina d’Ampezzo, Italy (2,078 m): 173 days, then 162, 10 fewer.

Where little has changed

High mid-stations in the Alps, the Rockies and the Andes, and the snowy coasts of Japan, show little change in the model: they still have 30 cm or more on almost every winter day. That is the same pattern the Swiss measurements show above about 2,100 m.

Questions

Has natural snow at ski resorts declined since 1950?

In Cirque’s model, mostly yes at village level: 39 of 43 resort villages with a snowy village lost days with 30 cm or more, typically 20 a winter. The model overstates the size of the change compared with measurements, so treat the figures as upper estimates.

Why use a model and not measurements?

Long snow measurements exist for few ski resorts. Where they do, as at Swiss weather stations, Cirque publishes them, and uses them to check the model.

Does this include snowmaking?

No. These are natural snow figures only. Many resorts now make snow on most of their pistes, which keeps runs open when natural snow is thin.

Snow is modelled from eleven winters of ERA5 weather data, natural snow only. It ignores snowmaking, grooming and wind, so use it to compare places, dates and heights, not as a guarantee. Always check local avalanche bulletins before going off-piste.

Weather and live snowfall: Open-Meteo (ERA5, Copernicus). Terrain: AWS Terrain Tiles. Basemap: OpenFreeMap, © OpenMapTiles. Ski areas, pistes and lifts: OpenSkiMap, © OpenStreetMap contributors (ODbL). Satellite view: Sentinel-2 cloudless 2016 by EOX (CC BY 4.0).

© 2026 Methodis Ltd