The tropical Pacific breathes in two rhythms. One lasts several years: the water off the coast of Peru warms and cools, which fishermen call El Niño. The second rhythm is fast: every month or two, a giant cloud wave is born over the warm waters of the Indian and Pacific Oceans, creeping eastward at a runner's pace and dying out near South America. These two rhythms seem independent, but in fact one adjusts to the other.

Until now, climatologists could not combine both rhythms in a single simple model. Complex computer simulations either described the fast cloud waves too crudely or gave incorrect statistics for slow warmings. A group of scientists built a bridge: they took an ocean of three connected basins and an atmosphere of three traveling moisture waves. Between them they laid two communication channels. Winds from the cloud waves push the water, and heat from the water feeds the clouds.

The secret to accuracy turned out to be noise. In the real climate, random wind gusts and evaporation bursts are not uniform: their strength depends on what is happening right now. When the water is already warm, random kicks more often accelerate warming than dampen it. The authors added such "smart noise" to the model, and the statistics matched observations: moderate and extreme warmings, multi-year droughts, and seasonal locking to winter.

Now climatologists have a cheap sandbox. In it, they can test how the two rhythms are connected without running a supercomputer for a month. They can generate long data series for training neural networks that predict weather a season ahead. The bridge between ocean and atmosphere rhythms is built, and it is already walkable.