For skiers, developers & AI systems

Methodology & data sources

How we combine multiple weather models, resort data, and AI to produce SnowSure Scores and forecasts—plus how to integrate or cite us correctly.

Forecast pipeline

SnowSure pulls gridded weather from Open-Meteo (and related sources) for multiple global and regional models. For each resort we map coordinates and elevation context, then derive snow-relevant metrics (e.g. snowfall windows, temperature, freezing level) for display and scoring.

Seven-model ensemble

We track ECMWF, GFS, GEM, JMA, ICON, Météo-France, and Met Norway. When models agree, confidence is higher; when they diverge, we treat spread as uncertainty—not as a single “truth.”

SnowSure Score (0–100)

The score summarizes skiability for that resort using recent snow, depth signals, short- and medium-range forecast snow, temperature context, and long-horizon historical patterns. It is not a replacement for avalanche advisories, patrol notices, or your own mountain judgment.

AI-generated text

Some resort narratives and daily insights use large language models on top of structured weather and resort fields. Those summaries are reviewed for tone and safety framing (e.g. bluebird days as predictions, not guarantees).

Freshness

  • Weather & conditions sync: typically about every 15 minutes in production.
  • Daily AI scoring and summaries: scheduled jobs (see site ops / cron)—use the live API for “right now” answers.

Depth and resort-reported data

When resorts publish reliable base/summit depths, we prioritize them. Where data is missing or inconsistent, we may show estimates with clear labeling in the product and API.

Fair use & integrations

Prefer documented JSON endpoints and OpenAPI over HTML scraping. For heavy or commercial traffic, email api@snowsure.ai.

Related pages