EIA's November outlook trims winter Henry Hub gas to $2.87 per MMBtu, 16% below its October forecast. The report points to a warm November and to supply. It never puts a number on El Niño.
The U.S. Energy Information Administration lowered its forecast for the winter Henry Hub natural gas price to $2.87 per million Btu on 10 November. That is 16% below its October forecast and about 36% below last winter. The tempting reading is that a mild El Niño winter is making gas cheap. EIA's own report does not say that. It points to a warm November and to supply, and it never puts a number on El Niño.
This post separates the two pieces EIA gives us, the weather contribution and the supply contribution, shows what the report supports and what it leaves unsaid, and explains why a low seasonal average still leaves room for an expensive cold spell.
The winter average is the headline, but the revision runs wider than the winter. EIA also trimmed its annual averages for 2026 and 2027, and the 2027 cut is the larger of the two.
| Henry Hub, per MMBtu | Previous forecast | November forecast | Change |
|---|---|---|---|
| Winter, November to March | about $3.42 (implied) | $2.87 | −16% (EIA) |
| Calendar 2026 | $3.48 | $3.37 | −3.2% (our arithmetic) |
| Calendar 2027 | $3.16 | $2.88 | −8.9% (our arithmetic) |
The $3.42 is not an EIA figure. It is what $2.87 implies if the cut is exactly 16%, so treat it as approximate. The point of the table is the pattern: a revision that reaches into 2027 is hard to credit to one warm month, although storage carried into spring links the two.
What the report does not do is decompose the revision. EIA does not say how many cents came from weather and how many from supply. Anyone who offers you a clean split is supplying their own model, not EIA's.
EIA's weather input is specific. For November, it assumes heating degree days 13% below the 10-year average. That is a single month. For the winter as a whole, it assumes 1% fewer heating degree days than last winter. Within that, January is 10% lower than a year earlier and March is 31% higher.
Two details deserve attention. First, EIA completed this forecast on 5 November, so most of the month it calls 13% mild had not yet happened. That is our inference from the dates, and it means the November assumption can still be wrong in either direction. Second, residential and commercial demand is forecast at 36.4 Bcf/d, down 3.8% from last winter. Those percentages alone do not establish how sensitive the price is to weather, and EIA's report does not explain the gap between a 1% heating degree day difference and a 3.8% demand difference.
For a winter-wide comparison against the 10-year average, the best number on hand is older. EIA's October outlook put November to March at about 3,150 heating degree days, 3% fewer than the previous 10-winter average, as we cited in our Northwest hydropower piece. The November report does not republish a winter-wide comparison to the 10-year average, so we do not.
Supply gives EIA more to work with, and the numbers are blunter. It forecasts end-of-November working gas in storage at 3,840 Bcf, which is 67 Bcf above what it expected a month earlier. A fuller tank going into winter tends to lower the price needed to balance the market, whatever the weather does later.
Production is the larger structural story. EIA forecasts production of 113.01 Bcf/d in 2026, against 107.64 Bcf/d in 2025, a rise of 5.37 Bcf/d or about 5% (our arithmetic). LNG exports rise from 15.1 Bcf/d to 17.6 Bcf/d, an increase of 2.5 Bcf/d, or about 17% (our arithmetic). Production growth is therefore roughly twice the export pull, although production and exports are not the only items in the balance.
That is annual background, not the October-to-November revision, and it does not explain why the forecast fell last month. EIA's own comparison of its two most recent outlooks shows what changed. Forecast dry gas production rose from about 112.2 to 113.0 Bcf/d for 2026 and from 116.1 to 117.6 Bcf/d for 2027 (rounded figures, EIA comparison table, 10 November outlook against 6 October).
| Dry gas production, Bcf/d (rounded) | October forecast | November forecast |
|---|---|---|
| 2026 | 112.2 | 113.0 |
| 2027 | 116.1 | 117.6 |
| Item | 2025 | 2026 forecast | Change |
|---|---|---|---|
| Production, Bcf/d | 107.64 | 113.01 | +5.37 (our arithmetic) |
| LNG exports, Bcf/d | 15.1 | 17.6 | +2.5 (our arithmetic) |
El Niño reaches a gas forecast through one important route: a temperature outlook, which becomes heating degree days, which becomes demand. NOAA's Climate Prediction Center (CPC) said on 15 October that El Niño conditions are strongly in place and gave an 80% probability that at least a strong event continues through January to March 2027. Its seasonal outlooks favor above-normal temperatures across the northern tier of the country.
That is a probabilistic outlook, not a forecast of any one month. And the warm signal is not national. In our heating-oil analysis, the 1997-98 and 1982-83 analog winters ran colder than normal in Atlanta (21.3% and 13.1% more heating degree days) and Dallas (8.6% and 10.4% more), even as the Northeast ran warm. That warm-north, colder-south pattern comes from analog winters, not from a forecast.
So the honest accounting is this. EIA cites a warm November, and EIA does not tie that warmth to El Niño. The table lays out what each source supports and what it leaves open.
| Possible driver | What the sources support | What they do not |
|---|---|---|
| Warm November | EIA assumes heating degree days 13% below the 10-year average. | How much of the roughly $0.55 implied revision it explains. |
| Storage | End-November stocks forecast 67 Bcf above last month's forecast. | How much is weather-driven and how much is supply-driven. |
| Production and LNG | Forecast 2026 production raised to 113.0 from 112.2 Bcf/d, 2027 to 117.6 from 116.1; LNG exports up 2.5 Bcf/d in 2026 versus 2025. | How much of the price revision this explains. |
| El Niño | CPC: strongly in place, 80% chance of at least a strong event through January to March 2027. | Any EIA quantification of its role in this revision. |
The $2.87 is an average of monthly averages across five months. It describes the center of EIA's forecast, not the tails. EIA itself warns that severe cold in January or February could lift demand above its forecast. Reduced November consumption can preserve storage for later months, but cannot prevent a January cold spell.
March has 31% more heating degree days than March 2026; that does not establish below-normal temperatures. That year-over-year increase does not tell us whether March will be warmer or colder than normal. Separately, EIA warns that severe winter cold could cause temporary price spikes. Prices can respond to demand on the coldest days and to how much gas is left in storage afterward, and neither shows up in a seasonal mean.
That is the practical reading for a household or a trader. A 16% cut to the seasonal forecast is a statement about the center of the distribution. It is not a statement that cold spells have become cheap.
Our claim: EIA identifies warm November weather and growing supply as downward pressures, but does not quantify their separate contributions or El Niño's share.
Next check: In December, compare EIA's revised weather, production, storage and price forecasts. Report its explanation without assigning numerical contributions it does not publish.
Next checkpoints: CPC's ENSO diagnostic discussion on 12 November 2026, EIA's weekly natural gas storage report (normally Thursdays at 10:30 a.m. Eastern, with holiday exceptions on Friday 13 November at 10:30 a.m. and Wednesday 25 November at 12:00 p.m. Eastern), CPC's seasonal outlook on 19 November 2026, and EIA's Short-Term Energy Outlook on 8 December 2026.