How it works

Where the numbers come from

PowPlanet turns a global weather model into a snowfall estimate at each resort. It is a guide to where snow is likely, not a promise.

The data

Forecasts use the ECMWF IFS HRES model, published as open data under CC BY 4.0. We use the 00Z and 12Z runs and rebuild the site about ten hours after each run, so the numbers refresh twice a day.

The model works on a 0.25 degree grid, roughly 28 km across. For each resort we take the nearest grid point and read its snowfall (as water equivalent) and 2 m temperature.

From model to resort

Temperature is adjusted for the resort's base and summit elevation using a standard cooling rate of 6.5 C per km. Snowfall water equivalent is converted to estimated new snow depth with a ratio that depends on temperature: colder air gives lighter, deeper snow. Totals are shown for the next 24 hours, 72 hours and 7 days at base and summit. The snow quality score is a simple guide to how light and dry the new snow is likely to be.

Limits you should know

  • A 28 km grid smooths out mountains. Storms can hit one side of a range harder than the model shows.
  • It estimates new snowfall. It does not measure snow depth, base, grooming, wind, lift status or avalanche danger.
  • Past values on the globe are archived forecasts, not observations.
  • Forecasts beyond a few days are less certain.

Always check the resort's own report before you travel, and local avalanche information before you leave the marked runs.

Resort list and credits

Resorts come from OpenSkiData and OpenStreetMap: named, operating downhill areas with at least three mapped lifts. The resort database is published under ODbL 1.0 at /data/resorts.json. Coastlines and borders come from Natural Earth (public domain).