Closer limits to human tolerance of global heat

Heat waves, a deadly hazard, are becoming increasingly severe due to anthropogenic global warming (1). Can we keep adapting? As recently as the 2007 Intergovernmental Panel on Climate change (IPCC) Impacts and Adaptation report (2), there was no identified limit to heat tolerance, very little literature on heat health and climate change (barely two pages in the report), and it was widely assumed, at least by economists, that we humans could live on a planet up to 12 °C warmer if we had to, albeit with large costs (3). Despite significant scientific advances since then, the upper limits of human climate tolerance have remained unclear and even their existence is not widely understood. In this issue, Vecellio et al. (4), joining another new study (5), show that the onset of heat stress beyond human tolerance is much closer than previously thought. With rapid warming and narrow safety margins, must we consider the possibility that insufficiently mitigated global warming could render parts of the planet uninhabitable?

Heat stress depends on temperature and humidity, both of which are increasing. The amount of water vapor in the atmosphere is growing at roughly 6 to 7% per degree of warming because of the Clausius–Clapeyron effect. The importance of this humidification is substantial and overlooked: it accounts for well over half of the increased total heat content of air near the tropical surface (6), yet was only considered as a heat stress component in the most recent IPCC reports and is still omitted from standard heat wave metrics.

Sherwood and Huber (7) first proposed an upper limit of human heat tolerance based on an environmental wet-bulb temperature (T_w), a metric incorporating both temperature and humidity, of 35 °C, at which maintenance of a normal core body temperature of 37 °C would become thermodynamically impossible. Though this threshold was widely referred to as the survival limit, the authors pointed out that it represented an upper bound of compensability based on optimal conditions of shade, hydration, and rest, with realistic tolerances certain to be lower–but without evaluating these. More recently a physiological study of healthy human subjects confirmed that uncompensable heat stress–where thermoregulation is overwhelmed–occurs much sooner, at T_w of 31 °C or lower (8). Environmental T_w tends to hit a ceiling, currently near 30 °C, due to a thermostat-like heat-transport mechanism (7). Nonetheless, T_w of 35 °C has been recorded a few times, and T_w surpasses 31 °C occasionally in a few regions for very short periods of time (9). Thus, in many parts of the world, particularly in the densely populated tropics and subtropics, we are already bumping up against the limits of heat tolerance. While it is possible to exceed these thresholds for short periods, our safety margin is narrowing and will be exceeded more frequently with every increment of warming (4).

Vecellio et al. (4) quantify the emergence of T_w beyond human thermal tolerance using the new and lower empirical limits, reaching broadly similar conclusions to another recent study (5). In comparison to previous estimates using the 35 °C idealized limit (7, 10), the empirical thresholds are exceeded with less warming and with a vastly expanded footprint (Fig. 1), although projected hotspots of high exposure remain the same. Extended periods of uncompensable heat emerge at less than 2 °C global-mean warming across sub-Saharan Africa, east China, the Indus River Valley, and the Persian Gulf. With every increment of warming above 2 °C, substantially more people are exposed to uncompensable heat for at least 1 wk of the year and in some regions for months. At 4 °C, exposure is widespread, extending across much of North and South America, the Middle East, South, and east Asia and into northern Australia.

While occasional exceedances of critical limits have already been recorded, and are already increasing in frequency (5), associated large-scale mortality has not. Whether this indicates physiological adaptation beyond the empirical thresholds measured so far (8) remains unclear. Elevated core body temperature resulting from thermoregulatory failure is not necessarily lethal if it is not too severe or long-lasting but, especially when chronic, such exposures can result in long-term organ damage (11). Laboratory experiments to intentionally push subjects closer to death are clearly out of the question, and epidemiological datasets required to estimate heat-mortality relationships often lack humidity data and are heavily biased toward temperate, high-income nations (12) leaving regions with the highest T_w severely underrepresented. That includes the new empirical limit, obtained from subjects in the United States (8). Some physiological adaptation to T_w closer to 35 °C is therefore possible, though individual differences in age, health, and fitness are likely to outweigh any further population-level adaptation (11). Thus, for most individuals, critical limits are likely to be even lower than those applied in the new studies (4, 5) and additionally depend on the level of physical activity and other climatic factors such as sunshine and wind (6).

Survival in a world regularly exposed to uncompensable heat would require large-scale behavioral and technological adaptation, with severe impairment to quality of life and economic productivity. Imagine millions unable to undertake the annual Muslim pilgrimage to Makkah because it is too hot (13) or the disruption to economic growth and global supply chains if outdoor work is restricted to a few safe hours a day (14). Widespread reliance on air-conditioning is neither a sustainable nor equitable option (15) and is likely to remain out of reach for most residents of the poorest and most vulnerable countries which comprise the majority of high-exposure regions. Moreover, those living and working in predominantly air-conditioned environments become maladapted to prevailing ambient heat (11, 16) potentially resulting in worse health outcomes when faced with inevitable exposure. Air conditioning would, however, be essential to protect life during the worst extremes (e.g., public heat shelters), although blackouts driven by high electricity demand during heat waves can have deadly consequences (6). Such extremes would also be disastrous for many wildlife and livestock without air-conditioned homes or heat shelters.

Processes not captured at the global scale or resolved in climate models mean that local T_w exposure is underestimated in some places. In cities, for example, urban “heat islands” occur due to nonnatural surfaces and waste heat (generated, in part, by air-conditioners). Although these heat islands are usually (by design) missed by meteorological station records, daytime urban T_w in wet climates is on average 0.17 °C higher than nonurban surrounds (17) and exposure in densely urbanized locales can be worse (18). In agricultural regions, including T_w hotspots across South Asia, irrigation can increase humidity, worsening heat stress (19). Even these relatively small extra increments in T_w can mean significantly more frequent exposure given the proximity of critical limits.

The majority of future uncompensable heat exposure will be driven by increases in humidity rather than temperature (4). Humidity clearly needs more attention in the context of extreme heat. Among the reasons it has been overlooked (12) is that for vulnerable populations in wealthy countries that experience relatively mild heat extremes, temperature dominates the apparent risk, probably because of dehydration. This can be addressed through behavioral change, but physical limits to heat compensability (7) cannot. Extrapolation of relationships seen in these conditions is not a guide to a hot and humid future.

These studies clearly refute any notion of humanity thriving on a planet with many degrees of warming, and collectively are a wakeup call that heat extremes are rapidly becoming more dangerous, due to the combined effect of higher temperatures and humidity and the scarce leeway between current extremes and tolerance limits (4, 5). The likely symptoms of increasingly intolerable heat during peak events will be human migration away from the most vulnerable regions toward cooler ones but also mortality of those who stay put, on much larger scales than seen so far. The impending heat-stress crisis makes it even more important to remain under the Paris 2 °C target or overshoot by as little as possible. The new studies (4, 5) show that this would keep heat within tolerable (if not comfortable) levels.