If you spend any time reading the briefings, you will notice that the climate forecasting community is in a quiet, almost embarrassed mood this year. A developing El Niño was confidently predicted by the major international models last spring; some even hinted at a “Super” event, the kind that the media refers to as Godzilla. A lot of that didn’t work out as the headlines suggested twelve months later. Oceanographers, who are typically cautious with their language, have begun to use a word that they hardly ever use in public. Incorrect.
The change in tone is difficult to ignore. The spring 2025 forecasts, according to one researcher at a recent NOAA briefing, are “overconfident in a way we should have caught earlier.” Scientists don’t say things like that lightly. It turns out that the models relied too much on tropical Pacific signals, such as trade winds and sea surface temperatures, while ignoring influences that drifted in from areas they weren’t trained to monitor as closely. In particular, the tropical Atlantic has been behaving peculiarly.
| Subject | Details |
|---|---|
| Phenomenon | El Niño-Southern Oscillation (ENSO) |
| Primary Region | Tropical Pacific Ocean (with Atlantic teleconnections) |
| Defining Threshold | Eastern Pacific SST anomaly of +2°C above the long-term average |
| Key Forecasting Body | Australian Bureau of Meteorology |
| Notable 2025 Failure | Spring Predictability Barrier — high-confidence forecasts that didn’t verify |
| Lead Researchers Cited | Aaron Levine (University of Washington); Michelle L’Heureux & Caihong Wen (NOAA CPC) |
| Index Now in Use | Relative Niño Index (adjusts for global background warming) |
| Average Response Time to Atlantic SST | Roughly 6–9 months (post-1980s), down from ~20 months |
| Funding Source for Recent Study | NOAA’s Climate Variability & Predictability (CVP) program |
| Current Forecast Status (April 2026) | Officially uncertain — “maybe” |
In a recent paper published in npj Climate and Atmospheric Science, Aaron Levine of the University of Washington, Michelle L’Heureux, and Caihong Wen of the Climate Prediction Center examined the so-called spring predictability barrier, which is the frustrating window from March to May when forecasts veer off course. They came to an uncomfortable conclusion. It wasn’t just bad luck for the models. The entire forecast chain was subtly distorted by their structural bias toward the incorrect signals, which mimicked rainfall in the western Pacific. Maps that were given to decision-makers, such as farmers in Queensland, water authorities in California, and commodities traders in Chicago, turned out to be derived from insufficient data.
And then there’s the Atlantic issue, which is what really caught people off guard. The dominant mode of tropical Atlantic SST variability has been moving northward since the mid-1980s, according to research published this year, which shortens the time it takes for Atlantic warming to push the Pacific.
The response took about twenty months prior to that change. It’s getting closer to six or nine now. Older models, which were based on presumptions from a different climate, were just not made to account for this amazing shift. Some researchers believe that the basins are now communicating with one another more quickly than the math predicted.
The message has shifted outside the Bureau of Meteorology in Melbourne, where forecasters discreetly released one of the most cautious 2026 outlooks in recent memory. Give up focusing on just one number. Give up trying to get the “Super” label. The Bureau now uses a Relative Niño Index, a sort of inflation correction for sea surface temperature, to account for the fact that the entire ocean is warmer than it once was. Although it sounds bureaucratic, it is important. Without it, every year begins to resemble an El Niño year, which tells you very little.
What remains is a slower, more truthful form of forecasting. Fewer Godzillas, less drama, and more “maybe.” Although some continue to pursue the loudest models, investors and insurers appear to be making adjustments. The construction and dissemination of seasonal forecasts may be quietly reevaluated in the coming years. As you watch this happen, it seems like science is finally realizing how complex the ocean has become.
