The heatwave season is stretching at both ends

The first heatwave can arrive sooner and the last one can linger later, leaving a wider stretch of the calendar in which exceptional heat may disrupt ordinary plans. That is the change researchers found across global land in a study of 1979 to 2023. Their result comes with a distinction: a longer heatwave season does not mean every day between its endpoints was hot.

The researchers found that the first detected heatwave moved earlier by an average of 3.29 days per decade, while the last moved later by 5.41 days per decade. The span between those two events lengthened by 8.71 days per decade. The two endpoint shifts add to 8.70 days; the slight difference is rounding.

Those figures describe heatwaves detected during specified warm-season windows: May through September in the Northern Hemisphere and November through March in the Southern Hemisphere, excluding Antarctica. Within each window, the researchers marked the first and last detected heatwaves and measured the distance between them. They were not searching every day of the year for a new start or finish date.

A longer span can contain cooler days

Imagine two heatwaves separated by several weeks. If the first comes three days earlier and the last five days later, the span grows by eight days even if the weather between them stays much the same. That simple calendar comparison is the article’s central measurement. It answers a different question from how many days actually met the study’s heatwave definition.

The distinction is visible in the paper’s supplementary analysis: it reports 1.52 more actual heatwave days per decade, compared with 8.71 more days in the span from first to last. Both are global trends calculated under the study’s warm-season windows. Neither says that a particular place gained exactly that many hot days.

To examine changes across decades and locations, the researchers used temperature data including ERA5. ERA5 is a reanalysis: it combines observations with a weather model to produce consistent estimates on a global grid. That gives researchers a way to compare broad patterns, though a grid estimate is not the same thing as a thermometer reading on one street.

A broad pattern with local variation

In the study’s maps, trends toward a longer heatwave season covered 92.1% of land, while trends judged statistically significant covered 34.4%. The researchers also found more pronounced timing shifts in drylands than in humid regions. The global average therefore describes a widespread direction of change, not a uniform shift at every location.

The study reports a recent move toward faster heatwave onset as well. Here, “faster” concerns how quickly an event develops toward its peak temperature. It is separate from an earlier onset, which concerns where the first event falls on the calendar.

For someone planning around heat, the finding explains why counting hot days alone can miss part of the story: the first and last episodes may be spreading farther apart. It cannot predict the date of next summer’s first heatwave in any town. That requires a local forecast, while this study measures a long-term change across many places and years.