NOAA GHCND · LIVE
How Weather Records Are Counted · Methodology

How Accurate Is the NOAA Winter Outlook? We Checked It Against the Record

The NOAA winter outlook is reliably better than a coin flip for temperature and roughly a coin flip for precipitation — and it is at its best in strong El Niño and La Niña winters, exactly like the one shaping up for 2026–27. That’s the honest answer. Here’s what NOAA’s own scorecard says, and what we found when we tested the last two strong El Niño winters against our station data.

By the Weather On This Day research team||Skill scores: NOAA CPC. Verification: our NOAA station analysis

Every October, NOAA’s Climate Prediction Center publishes a map of the coming winter — reds for “leaning warmer,” blues for “leaning cooler,” greens for “leaning wetter” — and every October the same question goes around: can you actually trust it? The useful answer is that the outlook carries real but modest skill, that it is far more skillful for temperature than for precipitation, and that its skill is highest in exactly the years when a strong El Niño or La Niña is steering the atmosphere.

NOAA doesn’t hide the scorecard — the Climate Prediction Center grades its own outlooks publicly. The trick to reading them is understanding that these are odds, not forecasts, and that “skill” here means beating a blind guess, not being right every time.


What NOAA’s Own Scorecard Says

CPC scores its seasonal outlooks with the Heidke skill score, a measure built so that 0 means the forecast did no better than chance, 100 would be perfect, and −50 is the worst possible. On that scale, CPC’s seasonal temperature outlooks run consistently positive — a long-run average skill around the low 20s in the areas where the outlook actually leans away from “equal chances,” and lower (roughly the low teens) once the many “equal-chances” areas are folded in. The seasonal precipitation outlooks score distinctly lower — only weakly positive — because rain and snow are so much harder to pin down a season in advance.

Those numbers bounce around from season to season — in a strong-signal winter the scores run well above the long-run average, and in a quiet, “equal-chances” year there’s little to score at all. But the shape never changes: temperature is reliably better than chance; precipitation is only a little better than chance; and both improve when El Niño or La Niña is strong. That last part is the key to 2026–27.


We Tested the Last Two Strong El Niño Winters

The 2026–27 winter outlook rests on a strong-to-very-strong El Niño. Only two comparably strong El Niño winters sit inside our station archive: 1997–98 and 2015–16, both with peak ocean warming above the “very strong” threshold. The classic strong-El-Niño outlook calls for a warmer northern tier and a wetter southern storm track. So we pulled the December–February numbers at representative stations and checked each signal against each station’s own long-term normal.

The temperature call: verified both times

StationRegion1997–98 vs normal2015–16 vs normal
MinneapolisNorthern tier+7.4°F+6.0°F
SeattlePacific NW+1.2°F+2.8°F

The “warmer North” signal landed in both winters, and emphatically. Minneapolis ran 7.4°F above its own winter normal in 1997–98 and 6.0°F above in 2015–16 — enormous seasonal departures, and exactly what the outlook favored. That’s the temperature outlook earning its keep.

The precipitation call: textbook once, busted once

Station (southern tier)1997–98 (% of normal)2015–16 (% of normal)
Los Angeles277%61%
Phoenix176%65%
Tampa, FL426%127%
Atlanta132%184%

This is where the precipitation outlook’s lower skill becomes visible. In 1997–98 the “wetter South” signal was flawless: Los Angeles took 277% of its normal winter rain, Tampa a staggering 426%, Phoenix 176%. Then came 2015–16 — an El Niño just as strong, famously forecast to drench California — and Los Angeles got only 61% of normal and Phoenix 65%. The “Godzilla El Niño” largely skipped Southern California. The Gulf and Southeast still ran wet (Atlanta 184%, Tampa 127%), so the signal wasn’t wrong everywhere — but a forecast that nailed the South in one strong El Niño half-missed it in the next. The temperature call, by contrast, verified both times.

December–February values computed November 9, 2026 from our Neon copy of NOAA GHCN-Daily, for airport stations at each city, versus each station’s own ~56-winter DJF mean, with a quality-control filter applied. These are representative stations, not a full regional average; they illustrate the outlook’s signal, they don’t grade the official outlook map.


What That Means for the 2026–27 Outlook

The 2026–27 winter outlook is built on a strong-to-very-strong El Niño — the kind of setup where the outlook is at its most trustworthy. Read it the way the evidence suggests: lean into the temperature tilt (the warmer-North signal is the call most likely to verify, as it did in both 1997–98 and 2015–16), and hold the precipitation tilt more loosely — treat “wetter South” as a real but not guaranteed lean, the kind of signal that was perfect in one strong El Niño and patchy in the next.

See our full breakdown of the 2026–27 winter forecast for what the outlook favors state by state, dig into the specific question of whether El Niño actually means more snow, and compare the seasonal outlook’s skill with how accurate day-to-day weather forecasts are by lead time. If you want to see how a non-government forecast stacks up, we ran the same honest test on the Farmers’ Almanac. And when this winter is over, check our page back here — we’ll grade the 2026–27 outlook against what our stations actually recorded.


Frequently Asked Questions

Is the NOAA winter outlook reliable?

It is reliably better than chance for temperature and much weaker for precipitation. NOAA’s Climate Prediction Center publishes its own verification: on the long-run Heidke skill score (where 0 is no better than a coin flip and 100 is perfect), seasonal temperature outlooks average positive skill in the low-to-mid 20s, while precipitation outlooks score far lower. That is modest but genuinely useful — it is a nudge in the right direction, not a guarantee, and it is strongest in strong El Niño or La Niña winters.

What does "40% chance of a warmer-than-normal winter" actually mean?

The NOAA winter outlook is probabilistic, not a prediction of a specific temperature. A "40% chance of above-normal" means above-normal is the most likely of the three categories (above, near, below), each of which would otherwise have a 33% baseline chance. It does not mean a warm winter is guaranteed, and a 33–40% tilt still leaves a real chance of a near-normal or even colder season. The map shows where the odds lean, not what will certainly happen.

Why is the precipitation outlook less accurate than temperature?

Temperature is spatially smooth and strongly nudged by large-scale drivers like El Niño, so a seasonal signal shows up across whole regions. Precipitation is far noisier — it depends on exactly where individual storms track and stall, which the atmosphere does not commit to months ahead. Our own test shows it: in the strong El Niño of 1997-98 the "wetter South" signal was textbook, but in the equally strong 2015-16 El Niño, Southern California stayed dry even as the Gulf Coast ran wet. The temperature call verified both winters; the precipitation call only half-verified in one of them.

How far ahead can forecasters predict a whole season?

CPC issues seasonal outlooks months in advance, but they are odds, not day-by-day forecasts. The skill comes almost entirely from slow-moving boundary conditions — above all the state of El Niño/La Niña in the tropical Pacific — not from predicting individual storms. That is why a strong El Niño winter like 2026-27 is a better-than-average setup for the outlook: the signal driving the odds is large and well understood.


Sources & Method

The skill-score framing is from NOAA’s Climate Prediction Center seasonal verification pages and NOAA Climate.gov’s seasonal-verification explainers, which describe the Heidke skill score (0 = chance, 100 = perfect) and the consistently higher skill of temperature outlooks over precipitation outlooks, checked November 9, 2026. The December–February temperature departures and precipitation percentages for 1997–98 and 2015–16 were computed this run from our Neon copy of NOAA’s Global Historical Climatology Network–Daily dataset, for airport stations at each city, against each station’s own long-term DJF mean, with a quality-control filter applied. The El Niño strength classifications are from NOAA’s Oceanic Niño Index; both 1997–98 and 2015–16 rank among the strongest El Niño events on record.