GFS vs ECMWF Demand-Hub Outlook: 2026-05-31 00Z

Executive Summary

The 00Z model suite presents a high-demand forecast from the GFS, which projects significantly more total degree days than the ECMWF over the next 15 days. This divergence is driven by the GFS forecasting both more heating demand in the Northeast and substantially more cooling demand in the Midwest and South. Confidence in this outlook is low, as reflected by a poor model convergence score of 0.52 and run-over-run trends showing the models moving in opposite directions.

Hub-Level Weather

The national forecast is shaped by major disagreements in the highest-weighted demand hubs, particularly in the latter half of the 15-day period.

• Midwest (Chicago, 35% weight): This is a key area of divergence. The GFS is much hotter, forecasting 136.8 cumulative CDDs versus just 94.2 from the ECMWF. The disagreement is most pronounced late in the period; for June 14, the GFS forecasts a mean temperature of 84.2°F while the ECMWF forecasts a chilly 53.6°F.
• East (NYC, 30% weight): The models are in closer alignment here, though the ECMWF is warmer overall, projecting 120.5 CDDs to the GFS's 111.6 CDDs.
• Northeast (Boston, 10% weight): The models show opposing scenarios. The GFS is significantly colder, with 62.3 HDDs and only 28.3 CDDs. In contrast, the ECMWF is much milder and forecasts a net cooling load with 36.2 HDDs and 84.2 CDDs.
• South (Houston & Atlanta): Both models show consistent cooling demand, but the GFS is materially hotter across the region, projecting 279.6 CDDs in Houston and 227.2 CDDs in Atlanta, compared to the ECMWF's 240.8 and 149.6 CDDs, respectively.

Demand Implications (Natural Gas)

The model disagreement translates into two distinct natural gas demand scenarios.

• GFS View (Bullish): The GFS outlook, with 148.8 weighted CDDs and 13.0 weighted HDDs, implies strong early-summer power burn demand supplemented by lingering heating demand in the Northeast. This combination would be supportive of natural gas prices.
• ECMWF View (Moderate): The ECMWF projects a more moderate demand picture with 128.6 weighted CDDs and 10.8 weighted HDDs. This suggests a smoother transition into summer, with cooling demand closer to seasonal norms.
• Run-over-Run Shift: The GFS trended less supportive compared to its prior run, shedding 2.8 weighted CDDs. Conversely, the ECMWF trended more supportive, adding 1.8 weighted CDDs. This opposing movement between cycles further clouds the forecast.

Model Agreement & Caveats

Confidence in the 15-day forecast is low due to significant model disparity.

• Convergence: The model agreement score is low at 0.516 (where 1.0 is perfect agreement), indicating poor consensus. This score worsened slightly from the prior cycle (delta_convergence_score: -0.025), suggesting increasing forecast uncertainty.
• Divergence: Disagreement is most acute in the Day 8-15 range, which is typical but particularly extreme in this cycle. The large temperature spreads in key hubs like Chicago reduce the reliability of the long-range outlook.
• Source Limitations: This analysis is based on a demand-weighted average of five specific city hubs from Open-Meteo's data feed. It serves as a strong proxy for national weather-driven gas demand but does not represent a full CONUS-gridded analysis.