GFS vs ECMWF Demand-Hub Outlook: 2026-03-30 00z

Executive Summary

The 15-day outlook points to a period of declining heating demand and emerging cooling demand. On a national weighted basis, the GFS is the warmer model, forecasting fewer HDDs and more CDDs than the ECMWF. Both models trended significantly warmer versus their prior runs, led by a substantial drop in forecast HDDs in the GFS, a bearish signal for near-term natural gas demand.

Hub-Level Weather

The largest demand hubs, the Midwest and the East, drive the national forecast. While the models show strong agreement in the Midwest (Chicago), there are material disagreements in the East and South.

• East (New York City, 30% weight): This is a key area of divergence. The GFS is substantially warmer, forecasting only 116.3 cumulative HDDs compared to 148.2 HDDs in the ECMWF.
• Northeast (Boston, 10% weight): Conversely, the GFS is much colder in Boston, with 275.2 HDDs versus the ECMWF's 217.5 HDDs. The higher weight of NYC means the GFS is warmer overall in the populated Northeast corridor.
• South Central (Houston, 15% weight): The GFS is consistently warmer, driving a higher cooling demand forecast of 110.1 cumulative CDDs versus 88.0 CDDs in the ECMWF.

Demand Implications (Natural Gas)

The total demand-weighted outlook is for 134.0 HDDs (GFS) and 138.8 HDDs (ECMWF) over the next 15 days. This suggests a transition to a shoulder season pattern with below-average heating demand. The key takeaways are:

• Heating Demand: The overall HDD forecast is modest. The significant run-over-run warming trend, particularly the GFS losing 29.4 weighted HDDs from its prior forecast, signals a sharp drop in expected residential/commercial heating load.
• Cooling Demand: The GFS forecasts a stronger start to cooling season, with 25.1 weighted CDDs versus the ECMWF's 17.9. This demand, centered in the South, will begin to support power sector gas burn for air conditioning.

Model Agreement & Caveats

Confidence in the overall warmer trend is reasonably high due to good model agreement on the national scale.

• Convergence: The models show solid agreement, reflected in a low convergence score of 0.24 and a weighted mean absolute daily HDD difference of 3.16. Furthermore, the models moved into better agreement compared to the previous cycle.
• Caveats: While the national totals are close, the underlying regional differences, especially in the East, introduce uncertainty. The significant bearish shift in the GFS is the primary driver of the change in the 15-day outlook; if it were to revert colder in subsequent runs, the demand picture would change materially.
• Source: This analysis is based on public GFS and ECMWF model data from Open-Meteo. The demand-weighting is a simplified representation of actual natural gas consumption patterns.