Andy Garcia · freelance dev
FRENES
Back to the blog
Climate & Numbers

NOAA Gives 83% Odds That October to December Tops Every El Niño Since 1950. July to September: +1.7

October 19, 2026 7 min read

On Thursday, October 8, NOAA's Climate Prediction Center published its monthly ENSO Diagnostic Discussion with this sentence: "Chances remain high for a historic event that would exceed the strength of previous El Niño events dating back to 1950 (+2.5°C or more for a 3-month RONI value): the chance is 54% for September-November, 83% for October-December, and 70% for November-January."

On Monday, October 12, the Copernicus Climate Change Service reported that the monthly sea surface temperature in the Niño 3.4 region reached an all-time high of 29.70°C in September, above the peaks of the very strong El Niño events of 1982/1983, 1997/1998 and 2015/2016. On October 15, NASA published satellite maps of the same event: the sea surface raised, the water warmer, the ocean bluer.

By the time the number reached NASA's article, it read: "there's an 83% chance that the strongest El Niño since 1950 is currently unfolding." Close. But the window, the index and the threshold are gone.

I build BidForge, an AI that reads French public tender files for small businesses, and I fact-check my own posts with Jev, a judge model that returns probabilities. In both, the job is the same: a number has to travel with its conditions. So I read the NOAA pages behind the sentence, the Copernicus bulletin and the NASA maps. Here is what they say, what they do not, and what I would do with them.

What the 83% is a probability of

RONI is NOAA's Relative Oceanic Niño Index. It takes the three-month running mean of sea surface temperature anomalies in the Niño 3.4 box (5°N to 5°S, 120°W to 170°W), subtracts the average anomaly across the tropics (20°N to 20°S), and rescales the result. The base period is 1991-2020. The subtraction is the point: if the whole tropical ocean warms, that warming does not count as a stronger El Niño.

On NOAA's historical RONI table, the highest value since 1950 is 2.4, reached in late 1982 and early 1983. The 1997 and 2015 events peaked at 2.3. The +2.5°C line sits one tenth above the record. The latest value posted as of October 19, for July to September 2026, is +1.7, and NOAA says the most recent values "should be considered an estimate" that can change for up to two months. The discussion also cites +2.1°C for Niño-3.4 over the past month: one month, not the three-month RONI the 83% refers to.

So the full sentence is: an 83% chance that the October-to-December average of one index reaches +2.5°C or more, from a July-to-September estimate of +1.7. That is a forecast, not an observation. NOAA's outlook puts the October-to-December median at 2.74, with a 90% range from 2.25 to 3.23. Even the low end is above 2.0, the line for NOAA's top strength category, and the strength table issued the same month gives a value of 2.0 or more a 100% chance for October to December.

Then the second 83%. The synopsis at the top of the same bulletin says: "a strong-to-very strong El Niño likely through January-March 2027 (remaining greater than an 83% chance)." That is a different claim, with a lower bar (1.5°C, where "strong" starts) and a later window. On the strength table, January to March gets 37% for strong and 46% for very strong, which adds up to 83%. Same number, two statements, one page. A summary that keeps only "83%" cannot tell you which one it means.

The table also forecasts the end: by April to June 2027, neutral conditions are the most likely category, at 54%.

Three measurements, one direction

What makes me take the event itself seriously is not one number. It is three kinds of measurement pointing the same way.

Copernicus works from its ERA5 dataset, whose sea temperatures are estimates at about 10 meters depth. Beyond the Niño 3.4 record, it reports the September average sea surface temperature between 60°S and 60°N at 21.05°C, the highest for the month, above the 20.92°C of 2023. Global surface air temperature was 16.27°C, 0.82°C above the 1991-2020 average and 1.63°C above pre-industrial: the second-warmest September, 0.11°C below September 2023.

NASA's sea level satellites measure the height of the ocean, a record that runs unbroken since 1993 and continues with Sentinel-6. NASA says a Kelvin wave, a slow pulse of warm water, was spotted near Micronesia in January 2026 and had reached the eastern Pacific by early April. Its maps show local sea levels as much as 30 to 40 centimeters higher in some areas, and 15 to 20 centimeters lower than normal around Indonesia and northern Australia.

NASA's PACE satellite measures ocean color. Equatorial Pacific waters have become measurably bluer over the last several months, a sign of fewer phytoplankton as weaker trade winds bring up fewer nutrients. This is the first El Niño since PACE launched in 2024.

Temperature, height and color are different physics, measured in different ways. When they agree, I trust the signal. They also explain an apparent contradiction. The Copernicus record is an absolute temperature, 29.70°C. NOAA's index is relative: it strips out the tropical average. A record on the first, in a warmer ocean, is not the same claim as a record on the second, and both pages can be right at the same time. Copernicus made the same distinction to AFP: it "cannot yet say that the current El Nino is the strongest ever recorded," AFP reported on October 12, and a new index it had adopted, which accounts for the warming of tropical oceans, suggested the event was not yet stronger than previous ones.

What the evidence does not show

It does not show impacts. NOAA's strength page is blunt: "The strength of El Niño or La Niña does not necessarily correspond with the strength of the influence or expected impact." The discussion says the chances of impacts consistent with El Niño "are larger, though not guaranteed," and points readers to separate seasonal outlooks for temperature and rain.

It does not explain Europe. The Copernicus bulletin reports the third-warmest September on record over European land, at 16.26°C, 1.54°C above average, and the joint-warmest September for western Europe, tied with 2023 at 2.51°C above average. It reports record-low September river flows in parts of the Rhine, Danube, Vistula and Garonne-Dordogne, with disruption to navigation, water supply and agriculture in several countries. It does not attribute any of that to El Niño.

And two months above 1.5°C are not the Paris limit crossed. Carlo Buontempo, director of the Copernicus Climate Change Service, said the new run of months above 1.5°C is likely to last into 2027 and that crossing the Paris Agreement's long-term limit "is only a matter of time." Only a matter of time means not yet.

What I would do with it

If you run a business that depends on weather, harvests or river freight, or you advise one, four habits.

Copy the whole sentence, not the percentage. Probability, threshold, index, window, issuer and date: "83% chance that NOAA's October-to-December RONI reaches +2.5°C or more, CPC, October 8." If any part is missing in a deck or a memo, it is a different claim.

Sort every input into past, forecast and attributed. The low rivers are past: September flows, as simulated by the European Flood Awareness System's river model. The historic El Niño is a forecast. A link between the two is not claimed by any of the sources I read. A plan should treat those three differently.

Put the update dates in the calendar. The next ENSO Diagnostic Discussion is scheduled for November 12. The strength tables update on the second Thursday of each month, and recent RONI values can be revised for two months. A plan built on October's 83% should be reread then.

Plan on the impact you can absorb, not on the event's rank. Whether this El Niño peaks at 2.4 or 2.6 on one index matters to the record books. What matters to a small business is the outlook for its own region and its own months, from the national weather service where it operates.

And if you build software that reads long documents for people, as I do, here is a cheap test I would add before shipping: give the summarizer NOAA's paragraph and check whether the window and the threshold come out the other side. In NASA's sentence, the number survived and the conditions did not.

The ocean has been measured three ways, and it agrees. The forecast is a number with conditions attached. Keep them attached.

Sources

A project like this one?

I design and deploy products like this. Let's talk.

Let's talk