GFS vs ECMWF Demand-Hub Outlook: 2026-06-13T00:00Z

Executive Summary

The 15-day outlook shows a national cooling demand forecast that is stronger in the GFS model than the ECMWF. The GFS projects 126.5 demand-weighted Cooling Degree Days (CDDs), while the ECMWF forecasts a milder 118.5. This divergence is driven by the GFS forecasting significantly more heat in the South, which outweighs a warmer East Coast signal from the ECMWF. Compared to the prior cycle, the GFS has trended hotter (+2.9 CDDs) while the ECMWF has trended cooler (-0.9 CDDs).

Hub-Level Weather

The primary drivers of the national forecast are the key demand hubs in the Midwest and East, with significant regional disagreements between the models.

• Midwest (Chicago, 35% weight): The models are in good agreement, with the GFS forecasting 46.8 cumulative CDDs and the ECMWF at 43.7.
• East (New York City, 30% weight): The ECMWF is notably warmer, projecting 133.4 CDDs versus the GFS's 125.7. The main divergence occurs in the day 8-12 period (June 21-25).
• South Central (Houston, 15% weight): This hub shows the largest disagreement. The GFS is significantly hotter with 298.3 CDDs, a substantial 41.2 CDDs more than the ECMWF's 257.1.
• South (Atlanta, 10% weight): Similar to Houston, the GFS is much hotter, forecasting 205.6 CDDs compared to the ECMWF's 156.1.

Demand Implications (Natural Gas)

The forecast period is dominated by cooling demand, which drives natural gas consumption for power generation (power burn).

• The GFS solution implies a stronger call on natural gas for power generation nationally, driven by its forecast for sustained heat across the South and South Central regions.
• The ECMWF's cooler profile in the South suggests more moderate power burn, though its warmer forecast for the East Coast would provide some regional support.
• The run-over-run trend shows models diverging. The GFS is adding to an already strong cooling demand forecast, while the ECMWF is moderating its outlook. This suggests increasing forecast uncertainty.

Model Agreement & Caveats

Confidence in the overall forecast should be tempered by significant regional disagreements, despite a high headline agreement score.

• Convergence: The overall convergence score is a relatively high 0.8369, indicating good agreement on the day-to-day temperature patterns. This score has improved by 0.062 since the last model cycle.
• Caveats: The high convergence score masks critical regional differences. The primary uncertainty is the intensity of heat in the southern hubs (Houston, Atlanta), where the GFS and ECMWF have materially different outcomes. The opposing trends in the run-over-run deltas (GFS hotter, ECMWF cooler) warrant caution.
• Source: This analysis is based on GFS and ECMWF data from Open-Meteo, using a predefined, limited set of demand-weighted hubs. This does not represent a full CONUS-wide grid analysis.