Which weather model is most accurate in San Diego?
San Diego, California · daily high temperature in °C · trailing 30 days · refreshed at most every 6 hours
Every day, The Spread archives what each major weather model predicted for San Diego at 1-, 3-, and 7-day leads, then scores those forecasts against the SAN observations once the day has happened. This page covers the trailing 30 days (2026-08-04 to 2026-09-02), for daily high temperature.
One day out, too close to call; at three days, too close to call; at seven days, too close to call.
The largest standing bias belongs to ECMWF IFS, which has run warm by about 5.31°C at 3-day lead over this window.
The Spread's own bias-corrected blend scored a 0.69°C typical miss at one day. Its corrections come from the same window it is scored on (in-sample), so read it as a ceiling, not a promise.
The scoreboard
Typical miss (MAE) in °C with the signed bias in parentheses — positive bias means the model ran warm. Days = scored days behind the 1-day number.
| Model | 1-day | 3-day | 7-day | Days |
|---|---|---|---|---|
| NOAA GFS | 0.8° (-0.5°) | 1.1° (-0.3°) | 1.3° (-0.5°) | 30 |
| GEM (Canada) | 0.9° (+0.7°) | 0.8° (+0.6°) | 1.2° (+1.0°) | 30 |
| DWD ICON | 0.9° (+0.1°) | 1.1° (+0.4°) | — | 30 |
| ECMWF AIFS (AI) | 1.8° (+1.8°) | 2.1° (+2.1°) | 2.6° (+2.5°) | 30 |
| NOAA AIGFS (AI) | 1.8° (+1.8°) | 1.5° (+1.4°) | 1.5° (+1.1°) | 30 |
| UKMO (UK) | 3.6° (+3.6°) | 4.0° (+4.0°) | — | 30 |
| ECMWF IFS | 5.0° (+5.0°) | 5.3° (+5.3°) | 5.3° (+5.3°) | 30 |
| NBM (US blend) | 0.7° (-0.3°) | 0.9° (-0.4°) | 1.2° (-0.8°) | 30 |
| Blended (bias-corrected) | 0.7° (+0.0°) | 0.7° (+0.0°) | 0.9° (-0.0°) | 30 |
Method, honestly
The Spread's San Diego point is the San Diego Lindbergh Field station (SAN), 2.8 km from downtown, used for both forecasts and verifications. Forecasts come from each model's actual archived runs via Open-Meteo's Previous Runs API — what the model said at the time, not a reconstruction. Truth is SAN observations — the station's own daily highs, a thermometer rather than a grid, and no correction applies: the thermometer is the truth outright. Rankings cover the models themselves; the two consensus products at the bottom of the table — NBM, which is itself a blend of models, and The Spread's own bias-corrected blend — are shown for comparison, never ranked or named "most accurate" here. The blend's corrections come from the same trailing window it is scored on (in-sample), so treat its score as a ceiling. A rolling 30-day window means these numbers move — that is the point.
Receipts — was the forecast right?
One page per settled date: what every model said as the day approached, what happened, and who was closest.
- Tue Sep 1 7 days out the models said 28–30; it hit 26. GFS led (off 0.1°C); ECMWF ran 4°C hot.
- Mon Aug 31 6 days out the models said 26–30; it hit 26. GFS led (off 0.4°C); ECMWF ran 5°C hot.
- Sun Aug 30 5 days out the models said 27–33; it hit 27. ICON led (off 0.4°C); ECMWF ran 4°C hot.
- Sat Aug 29 4 days out the models said 29–38; it hit 31. ICON led (off 0.3°C); ECMWF ran 6°C hot.
- Fri Aug 28 3 days out the models said 29–39; it hit 29. ICON led (off 1.2°C); ECMWF ran 7°C hot.
Other cities
Atlanta, Georgia Berlin, State of Berlin Boston, Massachusetts Buffalo, New York Charlotte, North Carolina Chicago, Illinois Dallas, Texas Denver, Colorado Detroit, Michigan Houston, Texas London, England Los Angeles, California Miami, Florida Milan, Lombardy Minneapolis, Minnesota New Orleans, Louisiana New York, New York Oklahoma City, Oklahoma Orlando, Florida Paris, Île-de-France Region Philadelphia, Pennsylvania Phoenix, Arizona Salt Lake City, Utah San Antonio, Texas San Francisco, California Seattle, Washington St. Louis, Missouri Tampa, Florida Washington, District of Columbia
The Spread — the forecast spread — is in closed beta. These verification pages are public while the rest is gated.