GFS vs ECMWF Demand-Hub Outlook: 2026-06-03 00Z

Executive Summary

The 15-day outlook is dominated by cooling demand, with both models forecasting a significant build in national CDDs. The GFS model is net hotter than the ECMWF, driven by a substantially warmer forecast for the South (Houston, Atlanta), though this is offset by GFS coolness in the Northeast. A major bullish shift in the latest GFS run (+26.9 weighted CDDs) has moved it closer to the ECMWF's solution, increasing overall forecast confidence despite persistent regional differences.

Hub-Level Weather

The national forecast is primarily driven by the high-weighted Midwest and East hubs, but the key model disagreements lie in the South and Northeast.

• Key Drivers: The Midwest (Chicago, 35% weight) and East (NYC, 30% weight) are the largest demand hubs. Both models show similar cumulative cooling demand in these regions, though they differ on the day-to-day timing of heat.
• South (Houston/Atlanta): This is the area of greatest bullish divergence for the GFS. The GFS is significantly hotter, forecasting 284.9 CDDs in Houston and 201.1 CDDs in Atlanta, compared to the ECMWF's much cooler 238.3 CDDs and 130.0 CDDs, respectively.
• Northeast (Boston): The models show a stark disagreement. The GFS is much cooler, projecting 23.7 HDDs and only 41.6 CDDs. In contrast, the ECMWF is significantly warmer with only 15.6 HDDs and 89.0 CDDs.
• Midwest (Chicago): While cumulative CDDs are similar (GFS 153.6 vs ECMWF 156.1), the models disagree on timing. The ECMWF is warmer in the 6-10 day period, while the GFS shows a much hotter pattern in the 11-15 day range.

Demand Implications (Natural Gas)

The forecast implies robust cooling demand, which will support natural gas power burn. The GFS presents a more bullish case due to its intensity in the South.

• Cooling Demand: The GFS projects a total of 168.4 weighted CDDs, slightly higher than the ECMWF's 161.4. This difference is almost entirely due to the GFS's stronger and more persistent heat across the southern demand hubs.
• Heating Demand: Heating demand is minimal and not a significant market driver. The GFS forecasts a national total of 2.9 weighted HDDs, primarily from its cooler outlook in the Northeast, compared to just 1.6 HDDs from the ECMWF.
• Run-over-Run Trend: The most significant signal from this cycle is the GFS model's large bullish shift. It added +26.9 weighted CDDs compared to its prior run, while the ECMWF was nearly flat (+1.8 CDDs). This indicates a trend towards a hotter consensus.

Model Agreement & Caveats

Overall model agreement is fair, but this masks important regional divergences that warrant caution.

• Convergence: The national convergence score is 0.864, indicating fair agreement on the total population-weighted temperature outlook. This score improved by +0.0822 from the previous cycle as the GFS trended hotter, moving closer to the ECMWF solution.
• Caveats: While the national CDD totals are reasonably close, the significant disagreements in the South and Northeast lower the confidence in the specific regional distribution of demand. The GFS's substantially hotter southern forecast, if it verifies, would represent a more bullish outcome for natural gas demand than the ECMWF currently suggests.
• Source Limitations: This analysis is based on deterministic GFS and ECMWF model runs provided by Open-Meteo. It does not incorporate ensemble data, which would offer a more complete view of forecast uncertainty and the range of potential outcomes.