GFS vs ECMWF Demand-Hub Outlook: 2026-09-11 00:00 UTC

Executive Summary

Models show significant disagreement over the 15-day outlook, driven by conflicting solutions for a late-period cool down. The ECMWF is materially colder, forecasting stronger early-season heating demand, particularly in the Midwest. The GFS maintains more warmth, especially in the South, implying a stronger tail to cooling demand. The GFS also saw a significant colder shift from its prior run, though a large gap with the ECMWF remains, and overall forecast confidence is low.

Hub-Level Weather

The primary drivers of the national forecast are the Midwest and East, where models diverge on the strength of a cool-off after September 20th.

• Midwest (Chicago, 35% weight): This is the main point of contention. The ECMWF forecasts a sharp increase in heating demand late in the period, accumulating 37.4 HDDs versus only 13.2 HDDs in the GFS.
• East (New York City, 30% weight): A similar, though less extreme, pattern is visible here. The ECMWF is colder with 27.8 HDDs compared to the GFS's 13.3 HDDs.
• South Central (Houston, 15% weight): The GFS is significantly hotter in this key cooling market, forecasting 295.3 CDDs against the ECMWF's 246.1 CDDs. This GFS heat is concentrated in the 6-10 day period.
• Northeast (Boston, 10% weight): In a reversal of the broader theme, the GFS is colder in Boston, projecting 61.7 HDDs versus 31.2 HDDs from the ECMWF.

Demand Implications (Natural Gas)

The conflicting weather forecasts imply different outcomes for Natural Gas demand as the market transitions from cooling to heating season.

• Total Demand-Weighted Outlook: The ECMWF is more bullish for NG prices, with a total of 25.3 weighted HDDs and 83.6 weighted CDDs. The GFS is more bearish, with only 15.9 weighted HDDs and a higher 91.2 weighted CDDs.
• Run-over-Run Shift: The latest GFS run shifted significantly colder versus its prior cycle, adding +5.9 weighted HDDs and shedding -22.8 weighted CDDs. The ECMWF was far more stable, adding just +0.9 weighted HDDs. This suggests the GFS is trending toward the ECMWF's colder solution, but the large gap warrants caution.

Model Agreement & Caveats

Overall forecast confidence is low due to poor model agreement and recent GFS volatility.

• Convergence Score: The model agreement score is very low at 0.39, highlighting the significant divergence in the 8-15 day period.
• Primary Disagreement: The models differ on the timing and intensity of a trough moving into the central and eastern U.S. The ECMWF shows a colder, more impactful event, while the GFS holds onto more warmth for longer. Given the GFS's large run-to-run shift, the ECMWF solution currently appears more stable.
• Source Limitations: This analysis is based on Open-Meteo's processing of GFS and ECMWF model data for specific city hubs. It is not an official government forecast and may not fully capture all nuances of regional demand.