GFS vs ECMWF Demand-Hub Outlook: 2026-06-23T00:00Z

Executive Summary

The 15-day outlook shows a significant spread between models, with the GFS forecasting a much stronger cooling demand scenario (+19.8 weighted CDDs) than the ECMWF. This divergence is almost entirely driven by the GFS projecting a major heat event across the Eastern U.S., particularly in the New York City hub. While overall model agreement has decreased since the prior cycle, a major bullish shift in the ECMWF (+23.9 CDDs) lends some support to the GFS's hotter outlook.

Hub-Level Weather

The primary drivers of the national demand forecast are the Midwest (Chicago, 35% weight) and the East (New York City, 30% weight).

• Midwest (Chicago): The models are in close agreement, showing a cumulative difference of only 4 CDDs over the 15-day period (GFS: 202.1 CDDs, ECMWF: 206.3 CDDs). The GFS is slightly cooler in the near term but diverges towards a much cooler solution than the ECMWF in the final days of the forecast.
• East (New York City): This hub is the source of the largest model disagreement. The GFS is forecasting a cumulative 191.3 CDDs, a substantial 59 CDDs more than the ECMWF's 132.3 CDDs. The divergence begins around day 5 (June 27) and becomes extreme in the Day 8-15 period.
• South/South Central (Atlanta/Houston): The GFS is consistently hotter in these regions as well, showing a +25 CDD surplus in Atlanta and a +19 CDD surplus in Houston compared to the ECMWF.

Demand Implications (Natural Gas)

Both models forecast strong, above-normal cooling demand consistent with the late June/early July timeframe, which will drive significant natural gas consumption for power generation (power burn).

• The GFS solution (207.7 weighted CDDs) implies a significantly tighter supply/demand balance than the ECMWF (187.9 weighted CDDs). The ~20 national weighted CDD difference is a material spread for a 15-day forecast.
• The key market risk is centered on the Eastern U.S. Should the GFS verify, power burn in the East would be substantially higher than the ECMWF currently suggests.
• Minor heating demand (HDDs) appears in the GFS for the Midwest in the near term but is negligible to the overall national demand picture.

Model Agreement & Caveats

Confidence in this forecast is mixed due to model volatility and specific regional disagreements.

• Convergence: The overall convergence score is a moderate 0.88, but this has decreased by 0.07 since the prior 00Z cycle, indicating that the models are moving further apart and uncertainty is increasing.
• Run-over-Run Shifts: The ECMWF made a significant bullish change, adding +23.9 weighted CDDs from its prior run. In contrast, the GFS trended slightly cooler, shedding -4.0 CDDs. The large shift in the ECMWF suggests it is beginning to resolve a warmer pattern, lending some credence to the hotter GFS solution despite the current gap.
• Caveats: The most significant model disagreement occurs in the less reliable Day 8-15 period. This analysis is based on a limited number of demand hubs and uses data from Open-Meteo's API, which may not capture all complexities of the full model suites.