GFS vs ECMWF Demand-Hub Outlook: 2026-07-22T00:00Z

Executive Summary

The 15-day outlook shows a significant divergence between models, with the GFS forecasting a considerably hotter and more demand-intensive period than the ECMWF. The GFS is +10.9 weighted CDDs hotter overall, driven by intense heat in the South and East, particularly in the 8-15 day range. This view is further amplified by a major +20.4 TDD hotter shift on the GFS run-over-run, signaling a bullish trend but reducing forecast confidence due to model disagreement.

Hub-Level Weather

The national picture is a composite of strong regional disagreements between the GFS and ECMWF models.

• Midwest (Chicago, 35% weight): This is the only region where the ECMWF is materially hotter, forecasting 182.3 cumulative CDDs compared to the GFS's 154.2 CDDs. This +28.1 CDD difference is most pronounced from July 26th to July 31st.
• East (New York City, 30% weight): The GFS is notably hotter in the East, forecasting 190.4 CDDs versus the ECMWF's 170.1 CDDs. The GFS projects a much stronger heat event during the final week of the forecast period (Aug 1-5).
• South Central (Houston, 15% weight) & South (Atlanta, 10% weight): The GFS is aggressively hotter across the South. In Houston, it forecasts 362.7 CDDs to the ECMWF's 308.1 (+54.6 CDDs). The divergence is even more extreme in Atlanta, with the GFS at 310.5 CDDs against the ECMWF's 222.3 (+88.2 CDDs). This GFS heat is concentrated in the 6-15 day outlook.

Demand Implications (Natural Gas)

Both models indicate strong and sustained cooling demand, consistent with late July seasonality. However, the GFS solution implies a significantly higher call on natural gas for power generation.

• The total demand-weighted forecast is 205.8 CDDs (GFS) vs. 194.9 CDDs (ECMWF).
• The GFS's significant hotter trend (+20.4 TDDs since the prior cycle) points to a potentially robust period of power burn, especially if its forecast for intense heat in the South and East materializes.
• The ECMWF's outlook, while still strong, is more moderate and has trended slightly cooler (-1.4 TDDs). The key difference lies in its less extreme heat forecast for the southern and eastern demand centers in early August.

Model Agreement & Caveats

Overall forecast confidence is low due to the significant regional disagreements and the large run-over-run GFS shift.

• Divergence: The models have diverged significantly since the last cycle, with the GFS adding substantial heat while the ECMWF remained flat-to-cooler. This increases uncertainty, particularly for the 8-15 day period where the temperature differences are largest.
• Confidence Signal: Despite a convergence_score of 1.0, the material TDD differences at both the hub and national-weighted levels indicate poor model agreement. The perfect score is likely an artifact of the calculation method in a summer period with zero HDDs. The primary signal is one of uncertainty.
• Source Limitations: This analysis is based on a limited set of demand hubs from Open-Meteo's API and may not capture the full complexity of U.S. natural gas demand patterns.