In 1982, a warm current off the coast of Peru grew so large that it altered weather across the entire planet. That year, India saw drought, Australia wildfires, and California torrential rains. The world learned that ocean and sky are linked by invisible threads: an event in one hemisphere echoes as rainfall in another. These threads are called teleconnections, and one of the most important runs from El Niño to the Indian summer monsoon.
The connection between El Niño and the monsoon has long been known: when the ocean warms, Indian rains typically weaken. But does this connection change over time? To find out, scientists ran the same climate model 63 times with slightly different initial conditions. The result is an ensemble—like 63 parallel worlds where climate evolves under the same laws but with different random fluctuations. If the connection strengthens consistently across all worlds, that points to a systematic trend, not mere chance.
And here is what they found: the connection is indeed strengthening, both in the past century and in the future under strong warming. But the real surprise lies elsewhere. The strengthening is not because El Niño itself is becoming more powerful. In fact, its variability may even decline. Rather, the monsoon is becoming more responsive: for each degree of ocean warming, rainfall changes more dramatically. It is as if the rope linking ocean and sky were growing stronger, but woven unevenly—thicker in some places, thinner in others.
The scientists tested whether the strengthening could be described by a simple linear formula. It could not. The connection tightens in jumps, and the rate of strengthening varies from decade to decade. It is like measuring the strength of a rope at different points and getting different values. This behavior is called non-ergodic, and it means that forecasts based solely on past observations can be misleading.
This study is the first to test the non-ergodicity of teleconnections using a large model ensemble. The results are important for monsoon forecasting: if the connection strengthens nonlinearly, models tuned to historical data may underestimate future droughts or floods. And the accuracy of such forecasts determines how much rice India will harvest and how many people will suffer from extreme weather. A rope woven unevenly demands closer inspection.
The non-ergodicity test showed that even if the correlation changes linearly over time, biases can arise if mean values or standard deviations change. In this case, these trivial contributions turned out to be insignificant, confirming the true nonlinearity of the forced response.