GFS vs ECMWF Demand-Hub Outlook: 2026-05-11 00Z

Executive Summary

The 15-day outlook shows a transition from shoulder season heating to early summer cooling, but with significant disagreement between models. The GFS is substantially warmer than the ECMWF, projecting 20.4 more total weighted CDDs and 4.5 fewer weighted HDDs. This divergence is driven by a much hotter GFS forecast in southern cooling hubs and a stronger late-period warm-up in the Midwest and East, a view reinforced by a major +11.3 weighted CDD shift warmer in the latest GFS cycle.

Hub-Level Weather

The primary drivers of the national demand forecast are Chicago (35% weight) and New York City (30% weight), with southern hubs like Houston (15%) becoming critical for cooling demand.

• South/South Central (Houston & Atlanta): This is the region of greatest disagreement. The GFS is significantly hotter, forecasting 63 more CDDs in Houston and 48 more in Atlanta over the 15-day period compared to the ECMWF. This points to a much stronger power burn scenario in the GFS.
• East (New York City): The models diverge significantly on heating demand. The ECMWF is colder, projecting 72.6 cumulative HDDs, while the GFS forecasts only 54.5 HDDs. The GFS sees more late-period warmth, adding 10.4 more CDDs than the ECMWF.
• Midwest (Chicago): The models show a complex difference. While total HDDs are similar (GFS: 79.6, ECMWF: 74.9), the timing differs. The GFS is much warmer in the back half of the forecast (May 22-25), leading to a higher cooling demand forecast of 47.6 CDDs versus the ECMWF's 38.1 CDDs.

Demand Implications (Natural Gas)

The conflicting model guidance presents two distinct gas demand scenarios.

• GFS Scenario: Implies a rapid end to heating demand and a robust start to cooling season. The strong heat in the South would drive a significant call on natural gas for power generation, creating a bullish demand outlook. Total weighted cooling demand is high for this time of year at 76.3 CDDs.
• ECMWF Scenario: Presents a more moderate, bearish outlook. It retains more heating demand in the East (72.6 HDDs in NYC) and projects a much tamer 55.9 total weighted CDDs. This would result in significantly lower power sector gas consumption compared to the GFS.

The GFS model's latest run trended sharply hotter (+11.3 weighted CDDs), while the ECMWF made a more modest warmer shift (-1.7 weighted HDDs). This large GFS move amplifies its bullish demand signal.

Model Agreement & Caveats

Confidence in this forecast is low, particularly beyond the first week.

• Convergence: The model agreement score is very low at 0.38, indicating substantial disagreement on the 15-day pattern. A score of 1.0 represents perfect agreement.
• Forecast Confidence: The significant divergence, especially in key southern hubs and in the day 8-15 period for northern hubs, warrants caution. The large run-over-run shift in the GFS, while notable, can sometimes be a sign of model instability and may revert in subsequent cycles.
• Source Limitations: This analysis is based on deterministic runs from Open-Meteo (gfs_seamless, ecmwf_ifs025). A review of ensemble model guidance would be necessary to better quantify the range of potential outcomes and overall forecast uncertainty.