High Heat Is Preferentially Killing the Young, Not the Old, New Research Finds by Kevin Krajick, December 6, 2024, Columbia Climate School
Many recent studies assume that elderly people are at particular risk of dying from extreme heat as the planet warms. A new study of mortality in Mexico turns this assumption on its head: it shows that 75% of heat-related deaths are occurring among people under 35–a large percentage of them ages 18 to 35, or the very group that one might expect to be most resistant to heat.
“It’s a surprise. These are physiologically the most robust people in the population,” said study coauthor Jeffrey Shrader of the Center for Environmental Economics and Policy, an affiliate of Columbia University’s Climate School. “I would love to know why this is so.” The research appears this week in the journal Science Advances.
The researchers chose Mexico for the study because it collects highly granular geographical data on both mortality and daily temperatures. The researchers reached their conclusions by correlating excess mortality–that is, the number of deaths above or below the average–with temperatures on the so-called wet-bulb scale, which measures the magnified effects of heat when combined with humidity.
The analysis found that from 1998 to 2019, the country suffered about 3,300 heat-related deaths per year. Of these, nearly a third occurred in people ages 18 to 35–a figure far out of proportion with the numbers in that age bracket. Also highly vulnerable: children under 5, especially infants. Surprisingly, people 50 to 70 suffered the least amount of heat-related mortality.
Based on this, “we project, as the climate warms, heat-related deaths are going to go up, and the young will suffer the most,” said the study’s co-lead author, R. Daniel Bressler, a PhD. candidate in Columbia’s Sustainable Development program.
The researchers say several factors may be at work. Young adults are more likely to be engaged in outdoor labor including farming and construction, and thus more exposed to dehydration and heat stroke. The same goes for indoor manufacturing in spaces that lack air conditioning. “These are the more junior people, low on the totem pole, who probably do the lion’s share of hard work, with inflexible work arrangements,” said Shrader. Young adults are also more likely to participate in strenuous outdoor sports, the researchers point out. A previous separate analysis by Mexican researchers showed that death certificates of working-age men were more likely to list extreme weather as a cause than those of other groups.
The vulnerability of infants and small children came as somewhat less of a surprise. It is already known that their bodies absorb heat quickly, and their ability to sweat, and therefore cool off, is not fully developed. Their immune systems are also still developing, which can make them prey to ailments that become more common with humid heat, including vector-borne and diarrheal diseases.
Wet bulb temperatures are often converted by popular media into “real-feel” heat indexes on the Fahrenheit scale, where numbers can vary depending on the exact combination of heat and humidity. According to the study, wet-bulb temperatures of around 13 C (equivalent to 71 F with 40% humidity) are ideal for young people; in this range, they suffer minimum mortality. Previous research has suggested that workers begin to struggle when wet-bulb temperatures reach about 27 degrees C, which would equate to 86 to 105 F, depending on humidity. However, the new study found that the largest number of deaths occurred at wet-bulb temperatures of just 23 or 24 C, in part because those temperatures occurred far more frequently than higher ones, and thus cumulatively exposed more people to dangerous conditions.
Using the same daily temperature and mortality data, the researchers found that elderly people died predominantly not from heat, but rather modest cold. (Mexico is mainly tropical and subtropical, but has many climate zones including high-elevation areas that can get relatively chilly.) Among other things, older people tend to have lower core temperatures, making them more sensitive to cold. In response, they may be prone to staying indoors, where infectious diseases spread more easily.
Despite all the attention given to the dangers of global warming, extensive research has revealed that cold, not heat, is currently the world’s number one cause of temperature-related mortality, including in Mexico. However, the proportion of heat-related deaths has been climbing since at least 2000, and this trend is expected to continue.
The new study has global implications, say the researchers. Mexico is a middle-income country; by share of population under 35, it is about average, and some 15% of workers are employed in agriculture. By contrast, many poorer, hot countries, mainly in Africa and Asia, have much younger populations that work in manual labor at much higher percentages. Thus, if Mexico is any indicator, heat-related mortality in those nations could be massive. A study published last year showed that farmworkers in many poor countries are already planting and harvesting amid increasingly oppressive heat and humidity.
Bressler said the team is now looking to firm up its conclusions by expanding its research into other countries, including the United States and Brazil.
The study was co-led by Andrew Wilson of Stanford University. Coauthors include Cascade Tuholske of Montana State University; Colin Raymond of the University of California, Los Angeles; Patrick Kinney of Boston University, Teresa Cavazos of the Centro de Investigación Científica y de Educación Superior de Ensenada, Baja California; and Catherine Ivanovich, Radley Horton and Adam Sobel of the Columbia Climate School.
Heat disproportionately kills young people: Evidence from wet-bulb temperature in Mexico by Andrew J. Wilson, R. Daniel Bressler, Catherine Ivanovich, Cascade Tuholske, Colin Raymond, Radley M. Horton, Adam Sobel, Patrick Kinney, Tereza Cavazos and Jeffrey G. Shrader, Dec 6, 2024, Science Advances, Vol 10, Issue 49
Abstract
Recent studies project that temperature-related mortality will be the largest source of damage from climate change, with particular concern for the elderly whom it is believed bear the largest heat-related mortality risk. We study heat and mortality in Mexico, a country that exhibits a unique combination of universal mortality microdata and among the most extreme levels of humid heat. Combining detailed measurements of wet-bulb temperature with age-specific mortality data, we find that younger people who are particularly vulnerable to heat: People under 35 years old account for 75% of recent heat-related deaths and 87% of heat-related lost life years, while those 50 and older account for 96% of cold-related deaths and 80% of cold-related lost life years. We develop high-resolution projections of humid heat and associated mortality and find that under the end-of-century SSP 3–7.0 emissions scenario, temperature-related deaths shift from older to younger people. Deaths among under-35-year-olds increase 32% while decreasing by 33% among other age groups.
INTRODUCTION
Historically, temperature exposure has caused a large number of premature deaths (1–3). Heat-related mortality is expected to increase under climate change (4–27). As the evidence base has grown, multiple studies have found that the elderly are especially vulnerable to heat (6, 11, 14, 17, 18, 28, 29). Furthermore, many other studies have expressed particular concern for joint heat and humidity extremes, given the importance of perspiration for human thermoregulation (30–36).
In this study, we explore the relationship between humid heat and mortality in Mexico, a country that exhibits a unique combination of rich, age-specific, universal mortality microdata and among the most extreme historical humid heat exposures. We find that historically, the majority of heat-related mortality in Mexico has been concentrated among younger people: 75% of heat-related deaths and 87% of heat-related lost life years occur among those under 35 years old. By contrast, the vast majority of cold-related mortality is concentrated among older people: 98% of cold-related deaths and 90% of cold-related lost life years occur among those over 35, with the majority of cold-related deaths occurring among individuals older than 70 years. We then develop projections of humid heat and associated outcomes to assess the future implications of these findings. As in other studies, we find that climate change is expected to increase heat-related mortality while decreasing cold-related mortality. However, we uncover an important source of future climate-driven inequality: The disproportionate impact of heat and cold across age groups reallocates the temperature-related mortality burden from the elderly (who are more affected by cold) to the young (who are more affected by heat). This has important implications for understanding the distributional impacts of climate change and for developing effective policies to adapt to these impacts.
…
DISCUSSION
The unique combination of elements in this study—station-level wet-bulb temperature estimates, granular mortality data from across the entire age distribution in a country with a wide diversity of climatic conditions, a statistical method that captures age-specific heterogeneity in temperature vulnerability, and high-resolution projections of humid heat—deepens our understanding of multiple aspects of the impact of temperature on mortality. By focusing on granular, age-specific temperature-mortality impacts, our study contributes to the existing literature that has usually focused on mortality irrespective of age (1, 8, 52), across broader age groups (6, 42), or on the elderly alone (12, 13). In particular, in our setting, we find that while individuals 35 and older suffer the vast majority of the cold-related mortality burden, those younger than 35 suffer most of the heat-related mortality burden. In addition, we identify a source of climate-driven inequality that has not been identified in previous studies: Across all future emissions scenarios, we find that climate change causes the temperature-related mortality burden to shift away from the elderly toward the young. Given that temperature-related mortality is projected to be the largest single source of climate damages (53, 54), the disproportionate burden of this impact on the young is likely an important source of future climate-driven inequality.
Prior research has discussed multiple reasons that older individuals are vulnerable to cold temperatures. These reasons are physiological, behavioral, and social. First, the elderly exhibit lower shivering temperature thresholds (55) and have substantially lower levels of brown adipose tissue (key for nonshivering thermogenesis) (56). Second, a relatively large proportion of elderly individuals have preexisting medical conditions or attendant respiratory illnesses that can be contributing factors in cold-related mortality (57). Third, elderly individuals are increasingly living alone, making it more difficult for them to access public health resources during extreme weather events, and they experience higher rates of loneliness, which is correlated with worse cardiovascular health (58). Fourth, energy poverty—spending a large fraction of income on energy—can be particularly acute for elderly individuals. Mexico has both a high rate of energy poverty and a high prevalence of credit constraints that might prevent adoption of protective but energy-intensive home heating (59). In this study, we indeed find that the elderly are, in terms of absolute mortality impacts, far more vulnerable to cold than other age groups (fig. S1). We find that the vast majority of cold-related mortality is concentrated in those 50 and older, as shown in the top panels of Fig. 2.
However, we find that young people are particularly vulnerable to heat: The majority of heat-related deaths are concentrated in those under 35, and those under 35 are overrepresented in heat-related deaths relative to their fraction of the population despite their far lower background crude death rate (fig. S5). Our finding that children younger than 5 years old are especially vulnerable to heat (Fig. 1) is directionally consistent with some prior work, although we find particularly acute effects. Multiple potential mechanisms may contribute to this result. First, infants have a higher body surface area-to-body weight ratio than adults, which means that they gain heat more rapidly and are more susceptible to overheating; infants also have a less developed thermoregulatory system (exhibiting reduced sweating), which means that they are not as efficient at regulating their body temperature (60). Second, very young children have less well-developed immune systems, making them more vulnerable to climate-related infectious diseases including vector-borne diseases and diarrheal diseases that might be especially affected by humid heat (61, 62). Last, both infants and young children have less freedom of movement than adults and may not be able to express their discomfort or distress as easily as adults, making it more difficult for caregivers—the primary providers of child adaptation to heat exposure—to recognize and respond to their heat stress (63).
We also find that heat disproportionately affects those 18 to 34 years old. Younger adults are more physiologically robust to heat, but multiple behavioral, social, and economic factors can contribute to higher heat-related mortality among this age group (41). Younger individuals are exposed to ambient heat through sports and other recreational activities (41). Households with older household heads are more likely to have an air conditioner (64). One important channel may be occupational heat exposure: Young adults are more likely than older adults to work in outdoor occupations with minimal flexibility for precautionary action (65). An analysis of death certificates in Mexico shows that men of working age are more likely to have extreme weather events listed as a cause of death (28), though we note that death certificates typically do not capture all deaths due to extreme weather (66). Relatedly, we find that individuals who live in regions with higher income (itself correlated with the amount of weather-exposed occupations) are less sensitive to heat (fig. S8). Occupational exposure is likely to be an important mechanism in other countries as well given that Mexico is not out of the ordinary in terms of occupational exposure to heat. For example, during our sample period, 15% of the workforce in Mexico was used in agriculture. This is lower than the rate for other middle-income countries (30% in 2018) and all countries globally (27% in 2018) (67). If occupational heat exposure is indeed a driver of mortality among younger individuals, then this highlights the importance of occupational heat exposure standards for workers (68).
Our finding that young people in Mexico are especially vulnerable to heat may have global implications because hotter and lower-income countries—which are expected to be the most adversely affected by climate change—have among the youngest populations in the world currently and over the coming century (69). Figure S12 shows the current global pattern of age and wet-bulb temperature exposure. The map in the top panel breaks down countries by their most extreme wet-bulb temperatures and fraction of population younger than 35 years of age (70). The youngest and hottest locations in the world are concentrated in Africa, Central America, the Middle East, and portions of South and Southeast Asia. The bottom panel of fig. S12 situates Mexico in the context of the rest of the world. Mexico is near the middle of the global distribution of countries by share of population under 35, and its extreme wet-bulb temperatures are essentially only surpassed by countries in Asia. The figure also shows that historical exposure to hot wet-bulb temperature is positively correlated with the fraction of the population under 35. If our age-specific results in this study hold for other countries around the world that are younger and hotter, then existing estimates of temperature-related mortality impacts in these countries—which neither fully capture age-specific heterogeneity in the temperature-mortality relationship nor account for the impact of humid heat—may be incorrect. In past work, the lack of age-specific mortality data has been a limiting factor in exploring the age-specific temperature-mortality relationship across a large number of countries (6, 52), which underscores the need for improvements in vital statistics systems, especially in the places most vulnerable to climate change.
We conclude by highlighting a few important caveats and also point to potential areas of focus for future work. Recent work has pioneered the use of both temperature and humidity for constrained joint projections (38, 71, 72). While regional and global climate models are our best tools for assessments of future heat stress risk, the relatively coarse time resolution of most model output limits the ability to project extreme values. The NASA Earth Exchange Global Daily Downscaled Projections (NEX-GDDP) dataset used in this study reports variables at a daily resolution, like many other climate model products. Given the misalignment of the diurnal cycles of temperature and humidity, using available daily mean values to calculate heat stress metrics such as wet-bulb temperature limits the accuracy of daily mean projections and is virtually impossible for daily maximum projections. These data challenges relating to the subdaily fluctuations in individual variables are even more pronounced for heat stress metrics such as wet-bulb globe temperature that incorporate additional variables relevant to the physiology of heat stress (e.g., solar insolation and wind speed) (73). These limitations underpin efforts to increase the temporal resolution of model data output available to end users to better represent the most extreme heat stress conditions of the future.
Our projections assume that our estimated temperature-mortality relationships will remain unchanged under future warming. There are opposing reasons why the wet-bulb temperature exposure-response functions may become either more or less severe in the future. Research on the US shows that mortality vulnerability to nonoptimal dry-bulb temperatures has decreased historically (74, 75). Recent work has shown that locations with different long-run climates show different patterns of consumption responses to weather shocks (76). Figure S8 shows that a similar pattern holds for mortality in Mexico. However, as wet-bulb temperatures approach uncompensable levels with substantially greater frequency (34, 51)—exposures of this degree are almost nonexistent in the historical record—we may learn that mortality associated with a given level of humid heat exposure is higher than existing estimates. Furthermore, our projections hold socioeconomic conditions fixed. Recently published subnational population projections for Mexico would allow future work to relax this assumption (77). These projections could yield higher estimates of mortality if population is trending younger in areas that are warming, or these projections could yield lower mortality estimates if the population is becoming older over time. Further estimation of the effect of income and occupational exposure could also enrich these projections and help shed light on the role of adaptation in mediating temperature-related mortality. We leave the exploration of these questions to future work.
Last, our conclusions further underscore the importance of ethical choices around monetizing the cost of premature deaths.
We find that climate change is expected to shift the mortality burden away from older individuals (more affected by cold) to younger individuals (more affected by heat).
Thus, the choice of whether to value life years—where premature deaths among younger individuals are considered more costly than premature deaths among old individuals—or to value all premature deaths the same becomes especially important. The US tends to value all premature deaths the same in its benefit-cost analysis (78), whereas UK guidance suggests that analysts can value either lives or life years (79). Although we do not take a stance on this difficult ethical choice, our findings further emphasize the importance of this debate for evaluations of the impact of climate change, given that we are finding that climate change is expected to shift the temperature-related mortality burden toward the young.
…
Data and materials availability: All data needed to evaluate the conclusions in the paper are present in the paper and/or the Supplementary Materials. A complete replication package can be found at https://doi.org/10.5281/zenodo.14182717.
***
