GFS vs ECMWF Demand-Hub Outlook: 00z 2026-06-15

Executive Summary

The 15-day outlook shows a significant divergence between the GFS and ECMWF models, leading to low forecast confidence. The ECMWF is materially more constructive for natural gas demand, forecasting a total of 130.8 weighted Cooling Degree Days (CDDs), driven by strong warmth in the high-demand Midwest and East hubs. In stark contrast, the GFS is far more bearish, showing only 105.9 weighted CDDs and even projecting minor heating demand in northern hubs.

Hub-Level Weather

The primary driver of the national forecast difference is the Midwest and East. * Midwest (Chicago, 35% weight): The models are in sharp disagreement. The ECMWF forecasts significant summer cooling demand with 75.7 cumulative CDDs. The GFS, however, shows a much cooler, unseasonable pattern with only 24.7 CDDs and 11.9 Heating Degree Days (HDDs) over the period. The divergence grows substantially in the 6-15 day range (June 21-29). * East (New York City, 30% weight): A similar, though less extreme, pattern exists. The ECMWF is consistently warmer, accumulating 139.9 CDDs versus the GFS's 97.4 CDDs. The difference is most pronounced from June 22nd onward. * South Central (Houston, 15% weight): Interestingly, the models are inverted in the south. The GFS is the hotter of the two models here, forecasting 285.5 CDDs compared to the ECMWF's 244.3 CDDs.

Demand Implications (Natural Gas)

The conflicting model guidance presents two very different scenarios for natural gas power burn over the next two weeks. * ECMWF Scenario: The 130.8 weighted CDDs suggest robust and above-normal cooling demand, particularly in population centers that drive national load. This outlook would be supportive of natural gas prices. * GFS Scenario: The 105.9 weighted CDDs imply significantly weaker power burn. The presence of GFS-forecasted HDDs in the Midwest (Chicago) and Northeast (Boston) in late June is exceptionally bearish, indicating periods where weather-driven gas demand would be minimal. * Run-over-Run Shift: Confidence in the GFS solution is further weakened by a major bearish shift from its prior cycle, losing 28.9 weighted CDDs (delta_weighted_cumulative_cdd_gfs: -28.87). The ECMWF remained much more stable, with a minor cooling adjustment of -1.6 CDDs (delta_weighted_cumulative_cdd_ecmwf: -1.61).

Model Agreement & Caveats

Forecast confidence is low due to significant model disagreement. * Convergence: The national convergence score is 0.76, indicating poor agreement between the GFS and ECMWF solutions. This score has slightly degraded since the previous model cycle. * Caveats: The primary source of disagreement is a major pattern difference across the Midwest and Northeast beginning around day 6 (June 21st). Given the GFS's large bearish shift and the ECMWF's relative stability, the market may be inclined to lean toward the warmer ECMWF forecast, but the low overall confidence warrants caution. * Source Limitations: Analysis is based on public GFS and ECMWF model data from Open-Meteo and may not reflect proprietary or ensemble model versions. The hub weighting is a simplified representation of national demand.