Verified forecast accuracy
Which weather model predicts snow best?
Across SnowSure's verification window, Nbm is the most accurate snowfall-forecast model in the Alps (accuracy 87.4/100).
SnowSure scores every model against observed snowfall, by region and horizon — primary-source verification, recomputed weekly. Machine-readable: /api/v1/model-accuracy.
Forecast Trust — Verified, not guessed — corpus scale, blend vs best model, and the full record.
Verification corpus
10,561,992 verified rows · 592 resorts
2024-01-01 → 2026-10-29 · Ground truth: SNOTEL, resort-reported totals, ERA5
As of 2026-10-30 · n = 10,561,992
Most accurate model by region
Per-model accuracy (overall)
| Region | Model | Accuracy | MAE (cm) | Samples |
|---|---|---|---|---|
| the Alps | Nbm | 87.5/100 | 2.51 | 126 |
| the Alps | Météo-France | 85.9/100 | 2.82 | 628 |
| the Alps | ICON | 84.5/100 | 3.11 | 1,270 |
| the Alps | JMA | 83.4/100 | 3.31 | 2,149 |
| the Alps | GEM | 82.6/100 | 3.47 | 1,882 |
| the Alps | ECMWF | 82.2/100 | 3.57 | 2,709 |
| the Alps | GFS | 81.5/100 | 3.7 | 2,709 |
| the Alps | Icon2i | 81.2/100 | 3.76 | 437 |
| the Alps | UKMO | 81.2/100 | 3.76 | 1,513 |
| the Rockies | Météo-France | 78.5/100 | 4.3 | 3,165 |
| the Alps | Meteoswiss | 77.7/100 | 4.46 | 914 |
| the Rockies | Nbm | 76.7/100 | 4.65 | 7,137 |
| the Rockies | ECMWF | 75.4/100 | 4.91 | 13,041 |
| the Rockies | ICON | 75.1/100 | 4.97 | 6,190 |
| the Rockies | GEM | 74.9/100 | 5.01 | 9,047 |
| the Rockies | UKMO | 74.7/100 | 5.07 | 4,933 |
| the Rockies | JMA | 74.5/100 | 5.09 | 10,327 |
| the Rockies | Nam | 74.2/100 | 5.15 | 799 |
| the Rockies | GFS | 73.1/100 | 5.37 | 13,319 |
| Cascades | UKMO | 68.7/100 | 6.26 | 377 |
| Cascades | ICON | 64.8/100 | 7.05 | 380 |
| Cascades | GEM | 64.7/100 | 7.06 | 553 |
| Cascades | Météo-France | 64.7/100 | 7.06 | 197 |
| Cascades | JMA | 62.6/100 | 7.49 | 642 |
| Cascades | ECMWF | 62.1/100 | 7.59 | 828 |
| Cascades | Nam | 59.9/100 | 8.01 | 61 |
| Cascades | Nbm | 57.8/100 | 8.43 | 567 |
| Cascades | GFS | 57.2/100 | 8.57 | 832 |
| the Andes | Météo-France | 56.9/100 | 8.62 | 143 |
| the Andes | ECMWF | 55/100 | 9.01 | 607 |
| the Andes | ICON | 54.4/100 | 9.11 | 286 |
| Sierra | Nbm | 48/100 | 10.4 | 63 |
| Sierra | ICON | 43.8/100 | 11.24 | 54 |
| the Andes | GFS | 42.4/100 | 11.51 | 607 |
| the Andes | JMA | 40.1/100 | 11.98 | 474 |
| Sierra | JMA | 39.9/100 | 12.02 | 90 |
| Sierra | GEM | 38.8/100 | 12.24 | 80 |
| Sierra | GFS | 36/100 | 12.8 | 117 |
| Sierra | UKMO | 35.3/100 | 12.93 | 38 |
| the Andes | GEM | 34.3/100 | 13.13 | 412 |
| Sierra | ECMWF | 33.1/100 | 13.39 | 117 |
Lower MAE = closer to observed snowfall. Accuracy is the inverse-error score (0–100). Source: SnowSure forecast verification, updated 2026-10-30.
See also: does blending models beat the best single one? · Forecast Trust
