Snow Forecast Accuracy: What 'Expected 40 cm' Really Means

    Almost every ski site shows a forecast. Very few are honest about how good those forecasts actually are.

    This page explains where the numbers come from, when to trust them, and why a 7-day total is more useful than a single-day prediction.

    Why mountain forecasts are harder than city forecasts

    A weather model is a grid. Each grid cell is typically 9 km wide and assumes one elevation. A real ski resort spans 1,200 m of vertical and three or four exposed faces, often inside one model cell.

    That means the model under-predicts orographic snowfall on the windward side and over-predicts on the leeward side. Anyone who tells you their 7-day total is accurate to the centimetre is bluffing.

    What we publish — and what we deliberately do not

    We publish 24-hour and 7-day forecast totals for upper mountain elevation. We deliberately do not publish hour-by-hour precipitation graphs, because at 9 km grid resolution they are noise dressed up as signal.

    • 24h total — useful for 'should I book a powder day?' decisions.
    • 7-day total — best signal-to-noise ratio for trip-planning a week out.
    • Powder alerts — triggered at ≥20 cm in 24 h or ≥30 cm in 7 days, with a 5-day cool-down per resort to avoid spam.

    Altitude corrections we apply

    Most public forecasts give snow at the model's reference elevation. A village at 1,400 m and a peak at 2,800 m do not get the same amount. We apply a per-resort altitude correction using the upper-mountain elevation we hold in our catalogue.

    This is the same reason a generic 'Tignes' forecast on a consumer app under-predicts what falls on the Grande Motte glacier above.

    How to read our forecast pages

    Compare 7-day totals across nearby resorts before fixating on absolute numbers. The relative ranking — 'Engelberg is forecast 80 cm, Andermatt 40 cm' — is far more reliable than the exact value at either site.

    Then cross-check with current snow depth. A resort forecast to get 40 cm on top of a 220 cm base is a much safer bet than the same forecast on a 60 cm base.

    Frequently asked questions

    How far ahead is a snow forecast reliable?

    Day 1–3 totals are usually within 30% of what falls. Day 4–7 is directional — useful for ranking resorts against each other, not for pinpoint planning. Beyond 7 days, take any number with a large pinch of salt.

    Why do two sites show different forecast totals for the same resort?

    Different underlying models (ECMWF, GFS, ICON), different altitudes assumed for the resort, and different aggregation windows. None of them are 'right.' Compare ranges, not single values.

    Do you forecast for the village or the peak?

    Peak / upper-mountain elevation, because that is what determines whether new snow turns into skiable powder rather than rain or slush.