Europe’s record heat stress: what the thermometer misses

Western Europe’s record summer averaged 21.7°C, but an afternoon outdoors can put much more strain on the body than that number suggests. A seasonal average smooths together hot days, cooler nights and different places. Humidity, wind and sunshine help determine how readily a person can lose heat.

Those distinctions matter in Copernicus’s October 2 assessment. June–August 2026 was western Europe’s warmest summer on record, surpassing 2003, while Europe overall ranked third. A record 52% of European land had experienced at least very strong heat stress by summer’s end. Reuters reported those findings from Copernicus.

Here is what the three headline numbers measure:

The temperature comparison uses the same season in the 1991–2020 reference period. It does not mean every western European town was 2.5°C warmer, or that the region stayed at 21.7°C throughout the summer. These are land-area averages for Copernicus’s defined regions, rather than population averages or figures confined to the European Union.

The underlying weather dataset, ERA5, is a reanalysis: observations combined with weather modelling to reconstruct conditions across a grid. Its near-surface air temperature represents air about two metres above the ground. That supplies a consistent climate record; it does not describe every sunlit pavement or shaded garden. Copernicus’s methods page explains the measurements.

For the body outdoors, researchers use another measure: the Universal Thermal Climate Index, or UTCI. It combines air temperature, humidity, wind and radiation—the energy reaching and leaving the body, including sunshine.

The physics is familiar. Sweat cools as it evaporates, so humid air can make that route less effective. Still air and strong radiant heat can also impede cooling or add to the heat burden. The World Health Organization explains how those conditions affect the body’s ability to dispose of heat.

UTCI translates the combination into an equivalent temperature: the air temperature in a reference environment that would produce the same response in a physiological model. The paper describing its development sets out that principle. A UTCI of 38°C therefore need not mean the actual air temperature was 38°C. Nor is it a measurement of anyone’s internal body temperature.

The 52% figure adds geography to this modelled stress. In the summer review, the summer map shows each location’s most severe June–August category. The cumulative curve instead keeps a running tally through the year: once land crosses the 38°C threshold, it remains counted even after cooling.

Imagine a simplified map of 100 equal-area squares. Thirty cross the threshold during one hot spell; later, 22 previously unaffected squares cross it. The cumulative total is 52%, even if the first 30 have cooled. Heating the same squares again does not add new area.

Duration answers another question. The review counts summer threshold-crossing days locally, then averages those counts over land. Western Europe averaged 13 days with at least very strong heat stress. That does not identify 13 dates when the entire region experienced it together.

None of these figures measures how many people became ill or died. WHO says vulnerability also depends on age, health, exertion, clothing, exposure duration and access to cooling. A land-area percentage cannot capture those differences, and the index should not be treated as a personal safety boundary.

Water systems were strained too: Copernicus reported the most severe low-flow summer in its river-flow record, which begins in 1992, with disruption to cargo transport and pressure on irrigation and power generation.

For an afternoon outdoors, WHO’s practical advice is to follow official heat warnings, seek shade and avoid strenuous activity during the hottest part of the day.