GFS vs ECMWF Demand-Hub Outlook: 2026-08-06 00Z

Executive Summary

The 15-day outlook indicates strong, persistent cooling demand, though models diverge on the overall intensity. The GFS model is notably more bullish, projecting 197.8 total weighted CDDs versus 186.4 on the ECMWF. This GFS heat is a recent development, with the model adding 13.1 weighted CDDs since its prior run, while the ECMWF remained stable, widening the inter-model spread.

Hub-Level Weather

The national-level disagreement is driven by significant regional differences, primarily concentrated in the southern United States.

• South Central (Houston): The GFS is substantially hotter, forecasting 350.1 cumulative CDDs compared to the ECMWF's 300.3. This ~50 CDD difference in a region with a 15% demand weight is a primary driver of the GFS's warmer national outlook.
• South (Atlanta): A similar pattern exists here, with the GFS projecting 258.2 CDDs against the ECMWF's 210.4.
• Midwest (Chicago): In the largest demand hub (35% weight), the ECMWF is warmer, forecasting 124.8 CDDs versus 109.6 for the GFS. This partially offsets the GFS's heat elsewhere.
• East (New York City): The GFS shows more heat in this key hub, projecting 220.0 CDDs to the ECMWF's 199.1.

Demand Implications (Natural Gas)

Both models forecast a robust period for power burn to meet air conditioning load, which should provide strong underlying support for Natural Gas consumption.

• The GFS scenario, with its 11.3 additional weighted CDDs, points to a significantly tighter supply/demand balance than the ECMWF. The significant run-over-run trend hotter in the GFS (+13.1 CDDs) introduces a bullish element to the forecast narrative.
• The ECMWF, while still showing strong cooling demand, presents a less extreme scenario that would imply a more moderate call on gas-fired generation.
• The market may treat the GFS's newfound heat with some caution until subsequent runs show consistency, given the ECMWF's stability. The significant model divergence implies higher uncertainty for NG futures pricing in the medium term.

Model Agreement & Caveats

Confidence in this forecast is moderate to low due to material model disagreement on cooling demand.

• The provided convergence_score is a misleading 1.0. This score is based on the difference in Heating Degree Days (HDDs), of which there are none in this summer forecast.
• In reality, the models show significant divergence on Cooling Degree Days (CDDs), with a national difference of 11.3 weighted CDDs over the 15-day period.
• This analysis is based on publicly available GFS and ECMWF data via the Open-Meteo API. The interpretation is limited to the provided demand hubs and weighting scheme.