Two conductors stand before one orchestra. One waves the baton quickly, setting the tempo every few bars. The second is almost motionless, but his slow breathing changes the overall sound throughout the symphony. So it is in the tropical Pacific: one part of the wind governs the fast oscillations of El Niño, the other governs decadal climate shifts.

Until now, these two roles were blended. The wind blowing over the ocean seemed like a single force that waxed and waned. The authors of the new study decomposed this force into two components, much as light splits into colors through a prism. The first component is divergent: it pushes water to spread from some places and pile up in others, rapidly changing surface temperature. The second is rotational: it swirls the water without changing its volume, and acts much more slowly.

These slow waves are like deep breaths of the ocean. They depend weakly on initial conditions and are therefore predictable decades ahead. When such a wave reaches the central Pacific, it strengthens or weakens El Niño. Reconstructing the surface temperature index from the fast part alone gave a correlation of 0.83 with observations, but missed the extreme years. Adding the slow waves markedly improved the match.

Ocean stratification, meanwhile, is not static. The reduced gravity, a measure of how strongly the warm upper layer "floats" over the cold lower one, grew linearly at a rate of about 0.66·10⁻⁴ m/s² per year. It is as if the spring connecting the two layers gradually became stiffer. The stiffer the spring, the stronger the response to the slow rotational wind forcing.

Now the two conductors are distinguishable. The fast one sets the rhythm of individual notes, the interannual oscillations we feel as warm or cold winters. The slow one changes the score itself, the decadal regimes in which extreme El Niños become more or less frequent. By separating them, climatologists gain a chance to better predict both.