A Heatwave Builds Faster Than It Fades, and the Gap Widens Farther North

A heatwave arrives faster than it leaves. Anyone who has sat through one knows the shape of it: the air thickens over a day or two, then sits there for a week before grudgingly letting go. Climate science has mostly not measured that. Extreme heat gets cataloged by its extremes: the peak temperature, the count of days above a threshold, the total heat accumulated. Those describe an event the way a mountain's height describes a mountain. They say nothing about the slope.

Shiyue Zhang and Deliang Chen, of Tsinghua University and Nanjing University of Information Science and Technology, went after the slope. In a paper published on Aug. 28, 2026, in Environmental Research Letters, they followed heatwaves across East Asia through observations running from 1979 to 2024, then repeated the exercise inside CMIP6, the coordinated set of global model runs the field leans on for projections. Each event got two numbers: how fast the temperature climbed from the start to the peak, and how fast it came back down. Heatwaves in the region consistently intensify faster than they dissipate, and the imbalance is stronger at higher latitudes.

Measuring a heat event by its rise and fall is not a new idea; it came from the ocean. Marine heatwaves have been described by onset and decline rates for years. In 2021, Claire Spillman and colleagues gave the two halves names that have stuck: the reaction window, from the start of an event to its peak, and the coping window, from the peak to the end. One is how long you have to do something. The other is how long you have to sit with it. Their global survey found the trend in onset rates outpacing the trend in decline rates at sea. The East Asian result runs the same way, which makes it the direction a specialist would have expected rather than a surprise.

The paper's second finding is the one with teeth, and it is about method rather than weather. To count a heatwave at all, you need a definition of extreme, and in practice that means a temperature threshold set high in the local climate's own distribution. The question this study puts is: the local climate of when?

Pin the threshold to a historical reference period and the answer is tidy. The world warms, peaks climb, and both onset and decay rates climb with them. That rise is driven mainly by the rising peaks themselves. Measured against the climate of decades ago, East Asian heat events look like they are coming on harder and letting go faster.

Let the threshold move instead: recompute it as the baseline climate shifts, so that extreme always means extreme for the present day. The authors find those increases shrink substantially across the mid-latitudes.

The two framings are not rival measurements of one thing. They answer different questions. A fixed baseline asks how heat events are changing in absolute terms, which is what a body or a crop actually meets. A moving baseline asks whether the atmosphere is generating more unusual heat than it used to, on top of the general warming, which is a question about the weather machinery rather than about the thermometer. Most of the fixed-baseline increase, on this reading, is the warming trend being carried along underneath the event. What is left once you subtract it is smaller, and the authors suggest it may involve changes in climate variability and in atmospheric circulation.

The choice cannot be sidestepped. Every projection of future extremes rests on one reference frame or the other, and the paper's point is that the two can disagree about how much a trend is growing while describing the very same model runs. The reference state is not background; it is doing part of the work usually credited to the climate.