A strong El Niño winter usually heats fewer homes. Not always.
Heating demand is counted in heating degree days (HDD): each day, how far the average temperature falls below 65°F, added up over the season. A strong El Niño stretches the Pacific jet stream east across the southern U.S. and keeps the Arctic air of the polar jet farther north, so much of the northern U.S. runs milder than usual, so fewer degree days, less gas and heating oil burned, and more gas left in storage in spring. In the 1997–98 super El Niño, Changnon (1999) found big cuts in northern natural-gas and heating-oil spending. This route shows how often that has happened, region by region, and how far it varies.
Nov–Mar heating degree days by census division
9 strong El Niño winters · trend removedHow each region's winters have gone
Select a row to chart it below. "Strong vs neutral" is the average detrended difference from ENSO-neutral winters, with its 95% bootstrap interval. "Lower third" counts strong winters in the mildest third of the detrended 1991–2020 winters. The 2026–27 column places each strong winter's departure on today's trend level.
Month by month: strong minus neutral
Detrended HDD difference for the selected region, with the 95% interval of resampled winters. Orange bars: fewer degree days (milder).
Reading it as a gas trader
The mild signal lives in the north and east, where most gas heat is burned: New England, the Middle Atlantic, the Great Lakes and the northern Plains. The South and the Mountain West split both ways, because a wetter, cloudier southern tier can run cooler days. The winter's biggest single risk is a sudden stratospheric warming that dumps Arctic air late in the season: see the stratosphere route. For the odds at your own city, see U.S. Winter Odds; the Season Book's station HDD lines open on Oct 15 (see how the book was tested).
Degree days here are weighted by population, a close stand-in for residential and commercial heating load. They are not gas-weighted: a region that heats with electricity or oil counts the same.
What history implies for 2026–27
the historical spread, not a forecastNov–Mar HDD if 2026–27 repeated each strong El Niño winter
Each dot is one strong winter's detrended departure placed on the 2026 trend level, shown against the 1991–2020 normal. The open diamond is the trend alone: where an average winter now sits because the climate has warmed. The gap between the diamond and the dots is what El Niño has added.
Gas storage at the end of March
the memory reservoirEnd-of-March working gas vs the average of the five winters before
A mild winter leaves storage high going into the injection season, and a cold winter after it can draw it hard: storage carries a winter's weather into the next one.
How these numbers are made
NOAA's nClimDiv publishes monthly heating degree days for every state and the nation, each weighted by where people live (2010 Census). We add each state into its census division using its 2020 Census population from NCEI's climate-division population file. For each region, Nov–Mar HDD in every winter since 1950–51 has a straight-line trend removed, fitted on the ENSO-neutral winters only so that El Niño years do not bend their own baseline. The detrended 1991–2020 winters set the thirds. Strong El Niño winters are the ONI "strong" and "very strong" classes. Hit counts carry Wilson 95% intervals; differences carry bootstrap intervals that resample whole winters, not months.
Nine winters are a small sample. A count of 4 in the mild third where 3 would be expected by chance is weak evidence on its own; the strong-minus-neutral difference, which uses every winter, is the stronger test. Where the two disagree, the page shows both.