The ocean remembers its own weather. Each month, surface temperatures form a vast map, and if you find a similar map in the past, you can see what happened next. That is exactly how the new forecasting method for El Niño works: a warm current that, every few years, changes the climate of half the planet.

Previously, forecasts were built on linear models. They work well when changes are smooth and proportional: slightly warmer water leads to slightly stronger winds. But El Niño does not behave that way. Its birth involves nonlinear interactions that a linear model cannot even see. Hence, its accuracy drops after six months, and beyond that the forecast becomes useless.

The new method belongs to the family of nonparametric approaches: it does not try to guess the formula governing ocean changes. Instead, it uses observational history from 1870 to 1997 and then tests itself on the following two decades. The result: a useful forecast persists for 9–14 months, whereas the linear model loses skill by 6–7 months. For the historical El Niño of 2015/16, a forecast made in June 2015 predicted a strong event peaking in December, missing the amplitude by only 0.4°C. Now farmers, energy planners, and insurers get more reliable risk assessments.