Indices of human heat stress in times of climate change

Body temperature is one of the cardinal regulated variables in human physiology, along with blood gasses, pH, and osmolality. Pathological deviations of body temperature from normal, some potentially lethal, are becoming more likely with climate change. Sherwood and Huber¹ famously used a critical wet-bulb temperature that would cause pathology to project areas of the world that would become uninhabitable in a climate-changed future. Many other indices that incorporate the wet-bulb temperature have been advanced to predict human heat stress. As pointed out by Maloney,² and recently confirmed empirically,³ those indices underestimate the impact of climate change on human thermoregulation, especially at lower humidity when physiological, rather than environmental, factors limit evaporative cooling (Figure 1).

Because the human body exchanges heat with the environment by four routes (conduction, convection, radiation, and evaporation), each impacted by different environmental variables, no single number can quantify that heat exchange accurately.⁵ But single numbers have nevertheless been proposed as indices of human heat stress. Because the dry-bulb temperature was poor at predicting the thermoregulatory responses of humans to different environments, in 1905 the English physiologist John Scott Haldane proposed the wet-bulb temperature as an alternative. The wet-bulb temperature is measured by placing a wetted sleeve over the bulb of a normal thermometer. Evaporation from the sleeve lowers the reading on the thermometer; the lower the humidity, the lower the reading of the wet-bulb thermometer below that of the normal thermometer. Wet-bulb temperature has been incorporated in many of the more than 150 indices of human thermal stress that have been developed over the past century.⁵ As well as underestimating the likelihood of pathology in some conditions, many of those indices also ignore the impact of air movement on both human heat exchange and on the wet-bulb temperature.

We all have been comforted by a breeze on a hot day, if we have been sweating. That effect has been quantified for acclimated women walking on a treadmill by measuring the upper limits of the prescriptive zone (ULPZ; that range of conditions in which core body temperature is affected by the level of metabolic heat production but not by the environment). In still air, the women could achieve heat balance in the conditions indicated by red shading below the dotted line in Figure 1. Conditions above that dotted line were above ULPZ, and the women became hyperthermic. At 1 ms⁻¹ of forced air movement, their ULPZ increased to include the conditions indicated by dark-yellow shading below the solid line in Figure 1, so the ability of those women to avoid pathology improved. That improvement would not have been predicted by an index based on wet-bulb temperature. Why? Because of boundary layers.

Boundary layers are the layers of fluids that are adjacent to surfaces, and they are fundamental to heat exchange and mass transport. When water evaporates from a surface, water vapor is added to the boundary layer, increasing water vapor pressure above that of the surrounding atmosphere. This increase reduces the drive for evaporation. Forced convection (wind or movement of the surface) disturbs that boundary layer, facilitating evaporation. A wet-bulb thermometer also has a boundary layer, a phenomenon often overlooked in analyses of heat stress. How fast that boundary layer is removed determines whether a measured wet-bulb temperature is the natural wet-bulb temperature or the ventilated wet bulb (also called the psychrometric or the thermodynamic wet bulb) temperature.

Air movement over the wet-bulb thermometer displaces its boundary layer and lowers the temperature registered on the thermometer below that without air movement. Haldane knew that, noting “… the wet-bulb thermometer read about 2° higher when left stationary.” A wet-bulb thermometer exposed to the prevailing atmospheric air movement measures the natural wet-bulb temperature. To measure the ventilated wet-bulb temperature the thermometer is whirled or aspirated, to expose the wetted bulb to more than 3 ms⁻¹ of wind, the speed above which evaporative cooling of the bulb plateaus. Thus, as environmental wind speed increases, the natural wet-bulb temperature moves ever closer to the ventilated wet-bulb temperature, and the two temperatures converge to the same asymptote above about 3 ms⁻¹ of wind.

The ventilated wet-bulb temperature is related, by thermodynamic principles, to the dry-bulb temperature and to relative humidity, with those principles expressed mathematically in many different algorithms. Those algorithms are often used to derive the wet-bulb temperature from standard meteorological measurements, to avoid the tedious direct measurement of ventilated wet-bulb temperature. Because of the plateau, the ventilated wet-bulb temperature so calculated always is unaffected by environmental wind speed.

The wet-bulb temperature that the human body experiences, when cooling evaporatively, is not the ventilated wet-bulb temperature. The thermometer should be exposed to the same air movement as the human body. That is the natural wet-bulb temperature. So indices of human heat stress that incorporate the wet-bulb temperature, like wet-bulb globe temperature, require the natural wet-bulb temperature.⁵ It is theoretically possible to derive the natural wet-bulb temperature from the dry-bulb temperature, the relative humidity, and the wind speed, but there is a potential for large errors.⁵ Often inadvertently, it is the convenient ventilated wet-bulb temperature that ends up being employed in heat stress indices, so inevitably generating an artifact.

It would be feeble to condemn the use of the wet-bulb temperature as an index of human heat stress without offering an alternative. The more accurate the alternative, the more it will depart from simplicity. The most accurate alternatives accept the complexity and quantify the exchange of heat between a human and the environment by all routes (conduction, convection, radiation, and evaporation) at a given level of activity (metabolic heat production and induced air movement), clothing (which determines the resistance to heat flow and evaporation), and the state of heat acclimation (which determines the maximum sweat rate).⁵ Such models exist, for example, the open-source MANMO⁶ and the proprietary Fiala Thermal Physiology and Comfort model.⁷ The former provided a good approximation of the empirical ULPZ.⁸ The latter has been converted into an index to reduce computational requirements, the Universal Thermal Climate Index, and validated against empirical datasets.⁹ The algorithms are complex but, with modern computing power, become feasible with standard meteorological data to better predict the responses of humans with no need to rely on indices.

Good predictions need good models. We do ourselves a disservice when we simplify complex phenomena into messages that can be disproved with evidence, such as advocating wet-bulb temperature as an index of heat stress.

Shane K. Maloney: Conceptualization; validation; writing – original draft; methodology. Michael R. Kearney: Conceptualization; validation; writing – review and editing; methodology. Duncan Mitchell: Conceptualization; validation; writing – review and editing; methodology.

No funding was received for the work included in this editorial.

The authors declare no conflict of interest.

No patient consent was required for the work included in this editorial.

No material from other sources is included in this editorial.

This editorial is not based on a clinical trial.