The Temperature the Body Reads

This essay opens a new page on globaia.org — /wbgt/, an editorial atlas of wet-bulb globe temperature, with an interactive map that lets you switch projections, rotate the centre, and read the value under any pixel. Both rest on a single dataset, recently released by Qinqin Kong and Matthew Huber at Purdue, which I think belongs in the planetary conversation alongside the better-known indicators of warming.

The thermometer on the wall is a stranger to the body. It records the temperature of the air, which is a single property of a single fluid; it knows nothing of humidity, wind, or sun. The human body, by contrast, is an engine that runs at 37 °C and must shed about a hundred watts of waste heat every hour simply to stay alive. There is only one mechanism it relies on at high ambient temperatures: the evaporation of sweat. And evaporation is governed not by how hot the air is, but by how much water the air can still accept. Once the surrounding air is saturated — once it can take no more vapour — sweat sits on the skin, the body’s core temperature climbs, and within a few hours a fit person resting in the shade is in mortal danger. This is the threshold past which heat ceases to be discomfort and becomes a physical limit.

Wet-bulb globe temperature (WBGT) is the standard, century-old way to compress that limit into a single number. Developed for the United States Marine Corps in the 1950s and adopted by athletic, military, and occupational-safety bodies ever since, it weights four ingredients — air temperature, the temperature a wet thermometer reaches as water evaporates from it, the temperature of a matte-black sphere absorbing sunlight, and the cooling effect of wind — into one operational reading¹. The U.S. National Institute for Occupational Safety and Health uses it to draw five bands, from safe through extreme, that tell a healthy worker how many minutes per hour they can keep working before they must rest. A theoretical upper limit of 35 °C wet-bulb has long been treated as the boundary of human tolerance²; recent human-subject experiments suggest the real ceiling, at rest, is closer to 30.5 °C, and lower still under exertion³. Either way, the metric maps cleanly to physiology, which is rare among climate indicators.

What it has lacked, until now, is a clean projection. Most published WBGT futures rely on simplified formulas (Stull’s empirical fit overestimates by more than 1 °C at high heat⁴) applied to raw climate-model output, which itself carries systematic biases of several degrees over the regions that matter most. Kong and Huber’s 2025 dataset⁵ addresses both problems at once. They reimplement the gold-standard Liljegren algorithm — the one derived from first-principles heat and mass transfer around an actual WBGT sensor — and apply it at three-hour resolution to a 15-model CMIP6 ensemble, then bias-correct the whole thing against ERA5 reanalysis at a quarter-degree grid. The result is the first global, high-resolution, physiologically meaningful projection of heat stress at warming targets aligned with the Paris Agreement: today’s climate, +1.5 °C, +2 °C, +3 °C, +4 °C. The corrections matter: raw models overestimate the hottest 5 % of WBGT hours by 4 °C across the Indo-Gangetic Plain — exactly the place where a billion people farm under the sun.

The reason this dataset belongs in public view, and not only in the supplementary materials of impact studies, is that the conversation about a warming world is still dominated by the wrong number. Global mean surface temperature is a useful summary; it is also a thermometer on a wall. It tells you nothing about whether the rice paddies of the Indus Valley will be workable in July, whether the brick kilns of Sindh will remain operable, whether the fishermen of the Mekong will be able to launch their boats at noon, whether the construction crews of Doha or Houston can keep building. WBGT does. It is also the number that exposes one of the most counter-intuitive feedbacks in climate adaptation: irrigation cools dry-bulb temperature but raises humidity, and can increase heat stress even as it lowers the reading on the wall⁶. Mitigation that ignores this distinction will misallocate effort.

There is, finally, a question of moral arithmetic. Heat stress is already the leading weather-related cause of death worldwide⁷, and its burden falls overwhelmingly on outdoor labourers in the tropics — the people who feed cities, build infrastructure, harvest fish, and have the fewest options to refuse the noon shift. A high-resolution map of where the body’s cooling system fails, at the warming levels we are actually heading toward, is therefore not a technical instrument. It is a piece of public information. The maps on /wbgt/ are an attempt to surface it.

References

1. Kong, Q. & Huber, M. (2025). A global high-resolution and bias-corrected dataset of CMIP6 projected heat stress metrics. Scientific Data, 12, 246. DOI: 10.1038/s41597-025-04527-6
2. Sherwood, S.C. & Huber, M. (2010). An adaptability limit to climate change due to heat stress. PNAS, 107(21), 9552–9555. DOI: 10.1073/pnas.0913352107
3. Vecellio, D.J., Wolf, S.T., Cottle, R.M., & Kenney, W.L. (2022). Evaluating the 35 °C wet-bulb temperature adaptability threshold for young, healthy subjects (PSU HEAT Project). Journal of Applied Physiology, 132(2), 340–345. DOI: 10.1152/japplphysiol.00738.2021
4. Liljegren, J.C., Carhart, R.A., Lawday, P., Tschopp, S., & Sharp, R. (2008). Modeling the wet-bulb globe temperature using standard meteorological measurements. Journal of Occupational and Environmental Hygiene, 5(10), 645–655. DOI: 10.1080/15459620802310770
5. Mishra, V., Ambika, A.K., Asoka, A., et al. (2020). Moist heat stress extremes in India enhanced by irrigation. Nature Geoscience, 13, 722–728. DOI: 10.1038/s41561-020-00650-8