ECMWF
ECMWF Integrated Forecast SystemEuropean Centre for Medium-Range Weather Forecasts · 0.4° (OpenData) · 15 days · Twice daily
What This Model Is Showing
Interpretation temporarily unavailable.
The European model — widely considered the most accurate global forecast model. Free OpenData access provides reduced-resolution products.
Highest-skill global forecast, hurricane tracking, comparison with GFS
500mb Heights — Atmospheric Pattern
ECMWF · 20 framesSea Level Pressure — Surface Systems
ECMWF · 20 framesHow to Read ECMWF Charts
The "Euro" Model
The ECMWF IFS (often called "the Euro") is widely considered the most accurate global forecast model in the world. It consistently outperforms the American GFS in verification scores, especially for medium-range (3-10 day) forecasts.
European weather agencies pooled resources to build it, and it runs at higher resolution with more sophisticated physics than GFS. When GFS and ECMWF disagree, experienced forecasters typically lean toward the Euro.
When to trust ECMWF over GFS: - Storm track forecasting (hurricane paths, nor'easters) - 5-10 day pattern changes - Precipitation type and amounts
When GFS might be better: - Very short range (0-24h) in the continental US - When ECMWF data is from the older 12Z run and GFS has a fresh 18Z run
We pull ECMWF data from their free OpenData service at 0.4° resolution — slightly coarser than the operational model, but the same physics and the same skill.
Comparing GFS and ECMWF
The most valuable thing you can do with model data is compare. Pull up the GFS 500mb F072 and the ECMWF 500mb F072 side by side.
If they agree: High confidence. The pattern is well-determined and you can plan around it.
If they disagree on position: The truth is usually between them, but closer to ECMWF. A trough that GFS puts over Ohio and ECMWF puts over Virginia will probably end up somewhere in between, but lean Virginia.
If they disagree on existence: One model has a feature the other doesn't? Check GEFS and ECMWF EPS ensemble spread in that area. If the ensembles show high spread, the feature is uncertain regardless of what the deterministic runs show.
For equestrians: This is how you become a better forecaster over time. Watch where the models disagree, make a prediction (use the Forecast Challenge!), and see who was right. After a few months, you'll develop intuition for which model handles your region better.