GFS vs ECMWF Demand-Hub Outlook: 2026-07-18T00:00Z

Executive Summary

The 15-day outlook shows a material disagreement between models, with the ECMWF forecasting significantly higher national cooling demand than the GFS (225.5 vs. 201.9 weighted CDDs). This divergence is driven by the ECMWF's much hotter outlook for the high-demand Midwest and East Coast hubs. Confidence in the forecast is low, underscored by a major run-over-run bullish shift in the ECMWF (+32.1 weighted CDDs), signaling significant model volatility.

Hub-Level Weather

The national forecast masks significant regional differences between the two models:

• Midwest & East (Hotter ECMWF): The ECMWF is substantially hotter in the most heavily weighted demand hubs.

  • Midwest (Chicago): ECMWF forecasts 235.0 cumulative CDDs, a stark contrast to the GFS's 173.9 CDDs. The models diverge significantly after July 23rd.
  • East (New York City): ECMWF projects 206.0 CDDs versus 174.5 CDDs from the GFS, with most of the separation occurring in the Day 8-12 period.

• South (Hotter GFS): The GFS shows more intense and sustained heat across the South, partially offsetting the coolness elsewhere.

  • South Central (Houston): GFS is the hotter model, with 351.7 cumulative CDDs compared to 311.5 for the ECMWF.
  • South (Atlanta): The GFS is significantly hotter, forecasting 280.4 CDDs against the ECMWF's 209.8 CDDs.

Demand Implications (Natural Gas)

The conflicting outlooks suggest different regional impacts on natural gas demand for power generation (power burn):

• The ECMWF scenario implies robust cooling demand and higher-than-expected power burn across the Midwest and Northeast, which are key natural gas consuming regions.
• The GFS scenario points to more moderate demand in the northern U.S. but suggests stronger power burn across the Southeast and Texas.
• Overall, the ECMWF presents a more bullish national demand picture due to the heat being concentrated in population-dense, heavily weighted hubs.

Model Agreement & Caveats

Confidence in this 15-day forecast should be low due to the sharp regional disagreements and model instability.

• Run-over-Run Shift: The ECMWF added a significant +32.1 weighted CDDs compared to its prior 12z cycle, while the GFS was largely stable (-0.2 weighted CDDs). This large shift in the ECMWF reduces confidence in its current solution.
• Misleading Convergence: The convergence_score of 1.0 is based on Heating Degree Days (HDDs), which are zero for all hubs in this summer forecast. It does not reflect the substantial model disagreement on Cooling Degree Days (CDDs) and should be disregarded.
• Source: Weather model data is from the Open-Meteo API. Degree days are calculated against a 65°F base.