The Year's First Heatwave Is Arriving Earlier, and the Last One Later

Ask when the hot season begins where you live and most people can name a month. Across much of the world that month has been sliding, and now it has a number on it.

A team led by Wenfang Xu and Juxiu Liu at the South China Botanical Garden in Guangzhou, part of the Chinese Academy of Sciences, set out to measure the slide. Philippe Ciais, of France's Laboratoire des Sciences du Climat et de l'Environnement, is among their co-authors. Their analysis appeared Sept. 23 in Nature Climate Change. It asks a question most heatwave research has passed over: not how hot, and not how long, but when in the year.

Averaged over the world's land from 1979 to 2023, the year's first heatwave arrived about 3.3 days earlier with each decade that passed. The last one hung on about 5.4 days later. The stretch between the two, the heat season, grew by roughly 8.7 days a decade. The season's close has drifted further than its opening.

Those are global averages, and the map under them is patchy. The team reports the heat season lengthening in the same direction across 92.1% of the world's land. But a trend that points one way is not the same as a trend that can be pinned down. The lengthening is statistically significant, meaning unlikely to be chance, on 34.4% of land. The same gap runs through the earlier onset and the later ending, and for those two the confirmed share is smaller still. If you want a single number for how much of the planet's land is demonstrably running a longer heat season, it is about a third, not nine-tenths.

The gap is not a flaw in the work. Heatwave dates jump around from year to year, and four and a half decades is a short record to test against that much noise. A trend can be real and still not yet be separable from the swings around it.

An independent reading is available, from a different group using different data. Catherine C. Ivanovich of NASA's Goddard Institute for Space Studies and two colleagues published their own account of shifting heat seasons on Sept. 10 in AGU Advances. They fixed each place's baseline heat season as the three months that held most of its extreme heat in the 1980s, then measured how far extremes now spill outside it. Seasons for extreme dry heat and for extreme humid heat have significantly expanded over about half the world's land.

That half does not contradict the larger figure above. Ivanovich's team counted only expansion that clears a significance test, which is why its result sits nearer this paper's confirmed third than its near-global direction. The two studies also use different definitions, different source data and different windows.

That points to the thing to hold loosely here. All three rates depend on this study's own definition of a heatwave, and the field has no agreed one. Count days above a fixed temperature and you get one answer. Count days above a local threshold, so that a hot day in Helsinki and a hot day in Hyderabad each qualify as hot for their own place, and you get another. Rebecca Emerton and three colleagues took a different route in the same journal on June 22. They measured heat stress by what they call "feels-like" temperature, and also found the heat stress season extending. Their numbers and these are not interchangeable and were never meant to be.

The team ran the same analysis on three separate global temperature records: ERA5, Berkeley Earth and Japan's JRA-3Q. All three are public, and the group posted its heatwave detection code as well. It also reports that dry regions have shifted more than humid ones. And it records what the authors describe as a recent move toward heatwaves that set in faster.

The argument at the end of all this is about the calendar rather than the thermometer. Heat alerts, outdoor work rules, hospital staffing and the school year are built around when heat is expected, and a season that opens in what used to be spring arrives before any of that is in place. Ivanovich's team names a second reason early heat is harder on people than the same heat at midsummer. Nobody has acclimatized yet, and the cooling nobody expected to need is not switched on. Heatwave research has spent years on how hot and how long. The case these authors make is that when belongs beside them, in early-warning systems and adaptation plans.