GFS vs ECMWF Demand-Hub Outlook: 2026-06-08 00Z

Executive Summary

The 00Z forecast cycle reveals a significant bullish shift in the 15-day outlook, led by the GFS model, which added +27.9 weighted cooling degree days (CDDs) versus its prior run. The GFS is now materially hotter than the ECMWF, projecting 149.2 total weighted CDDs versus 135.9 for the ECMWF. This divergence is concentrated in the 8-15 day period, primarily driven by a much hotter GFS outlook for key demand hubs in the East and South Central regions.

Hub-Level Weather

The primary drivers of the national-level model disagreement are the East and South Central demand hubs.

• East (New York City): The GFS projects 151.3 CDDs, a notable 18.0 CDDs more than the ECMWF's 133.3. This difference is most pronounced late in the forecast, for example on June 19th, where the GFS mean temperature is 18°F warmer than the ECMWF.
• South Central (Houston): A significant divergence exists here, with the GFS forecasting 285.1 CDDs against the ECMWF's 252.5. This 32.6 CDD differential is a key contributor to the warmer national picture.
• Midwest (Chicago): The models show a notable timing difference. The GFS forecasts a significant cool-down mid-period, introducing 12.7 HDDs between June 14-16 where the ECMWF shows none. However, the GFS rebounds to be much warmer than the ECMWF in the latter part of the forecast.

Demand Implications (Natural Gas)

The updated GFS forecast implies a period of strong, above-normal cooling demand, which would increase natural gas consumption for power generation. The magnitude of the run-over-run GFS shift (+27.9 weighted CDDs) is decidedly bullish and points to higher weather-adjusted power burn if the forecast holds. The ECMWF solution (135.9 weighted CDDs) is more moderate but its own +5.7 CDD run-over-run increase lends some support to the overall warmer trend. The GFS-specific forecast for a mid-period cool shot in the Midwest could temporarily suppress demand there before stronger, more widespread heat builds late in the period.

Model Agreement & Caveats

Overall model agreement is moderate-to-low, reflected in a convergence score of 0.71. Confidence in the forecast is tempered by two factors: 1. The significant divergence between the GFS and ECMWF models beyond day 7, particularly in the eastern half of the country. 2. The large magnitude of the GFS's run-over-run change, which suggests potential for volatility in subsequent model runs.

The market will likely wait for confirmation from the next few forecast cycles before fully pricing in the hotter GFS solution. This analysis is based on Open-Meteo point forecasts for major hubs with a 65°F degree day base.