https://orcid.org/0000-0001-8866-0609
Figures
Abstract
Over the past century, extreme heat events (EHEs) have become more frequent and intense, resulting in significant health impacts and economic challenges worldwide. In the United States, extreme heat is the leading weather-related cause of death, claiming more lives annually than hurricanes, floods, and tornadoes combined. However, the characteristics of extreme heat events can vary widely across events and over time. Even events perceived as similarly severe can result in vastly different health and societal outcomes—differences that remain largely understudied. In this paper, we explore regional trends in heat severity and mortality rates across the conterminous United States from 1981-2022 and provide a regional examination of how specific EHE characteristics impact heat mortality. We find that the number of extreme heat days has the strongest influence on heat related mortality. We observe increasing trends in heat-related mortality in every climate region throughout the U.S., except for the Western North Central region. These increases—likely tied to rising counts of annual EHE days—signal a structural shift to a new, elevated baseline of heat-related mortality in the U.S. Further, in the Southwest and Southeast regions, heat-related mortality is increasing at a higher rate than heat severity, suggesting potential for modification by community and individual level social vulnerability. Future heat mortality models should be holistic in their approach, incorporating not only multiple characteristics of heat but also measures of vulnerability to fully capture the complex dynamics of risk and exposure.
Citation: Narayanan A, Keellings D (2025) Rise in heat related mortality in the United States. PLOS Clim 4(8): e0000610. https://doi.org/10.1371/journal.pclm.0000610
Editor: Steven L. Forman,, Baylor University, UNITED STATES OF AMERICA
Received: March 24, 2025; Accepted: July 23, 2025; Published: August 27, 2025
Copyright: © 2025 Narayanan, Keellings. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: The data and code used in this study are openly available to facilitate transparency, reproducibility, and further research. The underlying datasets, including climate, heatwave severity, and mortality data, have been sourced from publicly accessible datasets and government agencies. Specifically, meteorological data to calculate heat severity were obtained from the Parameter-elevation Regressions on Independent Slopes Model (PRISM; available at https://prism.oregonstate.edu/) while heat mortality data were accessed through the Centers for Disease Control and Prevention (CDC) WONDER database (https://wonder.cdc.gov/).
Funding: This research was supported by the National Science Foundation under Awards (BCS-2203235 and BCS-1853775 to DK) and from the University of Florida College of Liberal Arts and Sciences (to DK). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
1. Introduction
Over the past century, extreme heat events (EHEs) have increased in frequency, intensity, and duration across much of the globe, posing a growing threat to public health, infrastructure, and economic stability [1–5]. Notable disasters such as the 2003 European heatwave— which led to more than 70,000 deaths [6] —and the 2010 heat episodes across the Northern Hemisphere (~56,000 deaths) [7] demonstrate the deadly potential of sustained high temperatures. Recent global estimates attribute approximately 489,000 excess deaths to heat exposure between 2000 and 2019 alone [8], with the true burden likely even higher due to underreporting and attribution challenges.
In the United States, EHEs have emerged as the leading cause of weather-related mortality, exceeding annual fatalities from hurricanes, floods, and tornadoes combined [9,10]. From 1999 to 2023, over 20,000 heat-related deaths were officially reported in the U.S. [11]. However, this figure likely underrepresents the full extent of heat’s impact, as many deaths linked to cardiovascular, respiratory, or renal complications are not explicitly attributed to heat exposure. Heat-related mortality is shaped by a complex interplay of meteorological, physiological, and social factors. Synoptic-scale climate patterns—such as persistent high-pressure systems and anticyclonic circulations—can intensify EHEs by trapping stagnant air masses, limiting ventilation, and amplifying surface heat accumulation [12–14].
These same atmospheric conditions also degrade air quality by preventing the dispersion of pollutants, resulting in elevated concentrations of ozone and fine particulate matter (PM₂.₅) [15–17]. Such air quality deterioration during heat waves is linked to increased cardiopulmonary hospitalizations and mortality, especially among vulnerable populations including older adults, children, and those with pre-existing heart or respiratory conditions [18–26]. In older adults, PM₂.₅ exposure has been associated with heightened cardiovascular mortality and stroke, likely due to oxidative stress and inflammation [27,28]. One study of individuals aged 50–71 found that every 10 μg/m³ increase in PM₂.₅ was linked to a 16% increase in ischemic heart disease mortality and a 14% rise in stroke mortality [29]. Among pregnant women, exposure to fine particulate matter has been tied to increased risk of preterm birth [30,31], while in children, it has been linked to higher risks of asthma and impaired lung development [32–34]. However, because these environmental stressors often overlap, the full impact of heat on mortality is likely underestimated—particularly when deaths from cardiac arrest, stroke, asthma, or COPD are not directly attributed to heat exposure on death certificates.
Despite the well-documented physical and environmental impacts of extreme heat, the risk of dying during an EHE is also shaped by social and structural vulnerabilities. Disparities in income, housing quality, social cohesion, healthcare access, and occupational exposure all contribute to unequal health outcomes during extreme heat events [35–40]. For instance, during the 1980 United States Heatwave, only 43% of American homes had air conditioning [41], limiting people’s ability to escape the oppressive heat and contributing to more than 1,200 deaths [42]. During the 1995 Chicago Heatwave, eleven of the communities with the highest death rates were predominantly African American; three were marked by infrastructure that amplified heat exposure, and one was home to elderly Polish residents who were culturally and linguistically isolated from the surrounding, increasingly Latino neighborhood [39]. Similarly, in the 2021 Western North American Heatwave, most fatalities occurred among older adults (aged 60+), individuals living alone, and those without any form of air conditioning [40].
Although EHEs affect all regions of the U.S., the severity of health impacts and the underlying pathways of vulnerability vary by geography, social context, and event characteristics such as duration, intensity, and areal extent. Recent findings from Jones et al. [43] suggest that regions unaccustomed to frequent heat extremes—such as urban areas in the U.S. Midwest and Northeast—exhibit a higher mortality response to heat exposure compared to regions with more routine heat exposure. These findings are consistent with earlier work by Anderson and Bell [44], who observed that the health impacts of EHEs—particularly those characterized by higher intensity and longer duration—were more pronounced in the traditionally cooler (seasonally) Midwest and Northeast compared to the warmer, more heat-acclimatized South.
Although research has significantly advanced our understanding of heat-related mortality, important gaps remain—particularly regarding how distinct meteorological characteristics of EHEs influence mortality across U.S. regions. Specifically, the comparative roles of event duration, intensity, and spatial extent—and how these factors interact—have not been systematically examined at the national scale. Existing studies tend to focus either on localized case studies (e.g., the U.S. state of Alabama, [45] and the Eastern U.S., [46]) or rely on data from major metropolitan areas to generalize findings to the entire U.S. population [44,47,48]. This often excludes rural, sparsely populated, or climatically unique regions from meaningful analysis. For instance, while Anderson and Bell [44] provided an in-depth assessment of EHEs in the Northeast, Midwest, and South, their analysis offered limited insight into trends in the Northwest and Western North Central regions—likely due to insufficient observations during their study period (1987–2005). Similarly, Shindell et al. [48] applied heat-response models derived from just 10 major cities to project future mortality across the entire conterminous U.S. While these studies offer valuable contributions, they also underscore common limitations in geographic coverage and regional representativeness. Anderson and Bell’s study reflects challenges stemming from sparse data in less-populated areas [44], while Shindell et al.’s approach risks overlooking critical variations in climate, demographics, and infrastructure by generalizing urban-based models [48]. These gaps constrain our ability to fully understand how heat exposure affects diverse populations across the country.
In this paper, we analyze trends in heat severity and mortality rates from 1981-2022 across nine regions of the conterminous U.S., examining specific characteristics (size, intensity, and days of exposure) of EHEs and their association with heat-related mortality. Building on prior research that suggests the impacts of heat cannot be fully explained by variations in single event characteristics [49–51], we compare the influence of individual characteristics to the cumulative effect of multiple characteristics (total heat severity). Additionally, we investigate differences in predictive capabilities when using individual versus collective heat characteristics to model heat-related mortality.
2. Materials and methods
2.1. Measuring total heat severity
To examine the impact of individual versus combined heat characteristics on mortality, we calculated the total annual heat severity of EHEs occurring during the summer months (May through September) from 1981 to 2022. We define annual heat severity as a cumulation of three measurable heat characteristics: size (areal extent), intensity (exceedance above the 95th percentile), and total number of extreme heat days and measure it using a modified version of the Heat Severity and Coverage Index (HSCI) [52]. The HSCI was developed to perform holistic assessments and comparisons of EHEs, accounting for intensity, duration, and areal extent [52]. A humidity modified version of the HSCI was later introduced to account for humid conditions during EHEs (HSCIH) [53] and is used in this study to measure total heat severity year-to-year (Equation 1).
Here, mi denotes the average magnitude of heat index temperature exceedance above a predefined threshold, measured in degrees Celsius, and ai represents the proportion of the total area affected by the EHE relative to the NOAA Climatically Consistent Region [54] where the event predominantly occurs. Each component, mi and ai is calculated daily throughout the duration of the event, designated by n days.
EHEs are defined as 2 or more days of hot-humid temperatures above the historical 95th percentile (1981–2022) occurring during the summertime period (May through September) of each year. For each day of each extreme event, the HSCIH value is calculated and summed to create event scores using daily gridded temperature and dew point temperature data from the Parameter-elevation Regressions on Independent Slopes Model (PRISM; available at https://prism.oregonstate.edu/). Annual Total Heat Severity are assessed by summing daily HSCIH values for each EHE occurring during the summertime period of each year. Heat assessments are performed at the climate region level, defined using NOAA Climatically Consistent Climate Regions [54].
2.2. Individual component analysis
The relationship between the individual characteristics of extreme events—intensity, total number of extreme heat days, and areal extent—and mortality is analyzed using two methods: Pearson Correlations to assess linear relationships and multiple linear regression to evaluate the impact of each characteristic on mortality. To identify the most influential characteristic of heat related to mortality, standardized beta coefficients and their associated p-values are compared for each characteristic and climate region. Since the exposure and outcomes (mortality) are measured cumulatively (annually) rather than for each individual event, our calculations compare the annual total event intensity, average areal extent, and total number of extreme heat days (exposure) to the crude heat mortality rate for each year. The annual total event intensity is the sum of the degrees (°C) above the 95th percentile for each event throughout the year. The average areal extent represents the mean event size annually. Lastly, the total number of extreme heat days counts all days classified as part of extreme heat events during the summertime period of each year.
To further evaluate the effectiveness of different predictors of heat-related mortality across various climate regions, a series of regression analyses are conducted, fitting and comparing multiple models for each region. Four distinct models were specified: a comprehensive model including event size, event intensity, and total heat days; and three simpler models, each focusing on one of these characteristics individually. Each model is fitted using Ordinary Least Squares (OLS) regression. We then use the Akaike Information Criterion (AIC) to assess the relative quality of these models, identifying the model with the lowest AIC as the best fit for each region. Additionally, we perform a Likelihood Ratio Test (LRT) to compare the comprehensive model with each of the simpler models, evaluating whether the more complex model provided a significantly better fit. To account for multiple comparisons and to avoid inflating the Type I error rate, we applied the Benjamini–Hochberg (B-H) procedure [55] to control the false discovery rate at 5% within each family of related tests. The original and B-H adjusted p-values are reported for each statistical test to provide transparency and allow assessment of the robustness of the findings.
2.3. Heat mortality analysis
Annual summertime mortality data from 1981 through 2022 was collected from the Center for Disease Control (CDC) WONDER (Wide‐Ranging OnLine Data for Epidemiologic Research) database [56]. For the long-term trend analysis between heat severity and heat-related mortality, underlying (primary) causeses of death were filtered by International Classification of Diseases Ninth (ICD-9) and Tenth Revision (ICD-10) codes for hyperthermia (1981–1998 ICD-9: E900.0; 1999–2022 ICD-10: X30) and aggregated by climate region.
While this definition enables a consistent and specific identification of heat-related deaths, it does not account for fatalities where extreme heat may have exacerbated pre-existing health conditions—such as cardiovascular disease (e.g., heart attacks, strokes [21,22,51,57]) and respiratory illness (e.g., asthma and chronic obstructive pulmonary disease (COPD) [18–20,58]—that ultimately led to death. These indirect or contributing causes are not captured in this analysis due to data limitations. This is especially important for vulnerable populations, including older adults and young children with cardiopulmonary conditions, who are disproportionately affected by extreme heat exposure [22–24]. As such, the exclusive reliance on hyperthermia codes likely underestimates the full burden of heat-related mortality [59]. Underreporting may be particularly pronounced in rural regions, where limited healthcare access and variability in death attribution practices can hinder the accurate classification of heat-related deaths. Accordingly, CDC-reported mortality figures should be interpreted as minimum estimates of the public health impacts of extreme heat.
Ideally, mortality data would be available at a monthly resolution for all years, enabling a more precise analysis of heat-related mortality during the summer months (May through September). However, due to limitations in data availability, the analysis of mortality was conducted on an annual basis. Through CDC WONDER (http://wonder.cdc.gov), mortality data from 1999-2022 can be filtered by both year and month, however, data prior to 1999 is only available on an annual basis. While previous investigations by Vaidyanathan et al. [59] indicates that 90% of heat-related deaths occur from May through September, it is important to consider that heat-related deaths outside these months may still be present in the dataset for the years 1981–1998 [58,60].
For year-to-year statistical comparison, mortality data is normalized to crude death rates per one million people (deaths 10⁶ yr⁻¹). Annual population data for normalization are annual Census Bureau estimates provided by CDC WONDER. Correlations between total heat severity, as measured by the HSCIH, and crude mortality rates are assessed using a Pearson Correlation. Additionally, trends in heat severity and crude mortality rates are examined using the Mann-Kendall test [60], and the magnitudes of these trends are quantified using Sen’s Slope [61]. To maintain the privacy of individuals, mortality data from the CDC WONDER database is not reported for deaths totaling nine or fewer during any specified period. Therefore, not all climate regions have mortality data for each year during the 1981–2022 period.
To evaluate the impact of various heat event characteristics on mortality rates across different U.S. climate regions, a multiple regression-based scenario analysis was conducted. The analysis focused on understanding how different combinations of Event Size, Intensity, and Exposure affect mortality under five key scenarios: (1) High Event Size, Low Intensity and Exposure, (2) High Intensity, Low Event Size and Exposure, (3) High Exposure, Low Event Size and Intensity, (4) All Characteristics High, and (5) All Characteristics Low. For each predictor variable, the 1st quartile (Q1) and 3rd quartile (Q3) values were calculated across the dataset to define “low” and “high” levels used in the scenario analysis. These quartiles represent data-driven thresholds for what constitutes low and high levels of the predictors. Before fitting the regression models, Min-Max (0–1) standardization was applied to each variable. This scaling ensured that all predictors were brought to the same range, allowing their contributions to the model to be directly comparable and preventing any variable with a larger numerical range from disproportionately influencing the results.
3. Results
3.1. Trends and regional relationships in heat severity and heat related mortality
Between 1981 and 2022, heat-related mortality rates increased significantly across all U.S. climate regions except the Western North Central (Table 1). After adjusting for multiple comparisons with the Benjamini–Hochberg adjusted p-values, 7 of 8 heat mortality temporal trends (Table 1) remained statistically significant at the 0.05 level. The Southwest exhibited the most pronounced rise, with a Sen’s Slope of 0.14, followed by the Northwest (0.05) and South (0.04). In these regions, total heat severity also increased significantly at the 95% confidence level. In contrast, the Central, Eastern North Central, Northeast, Southeast, and West regions experienced more moderate mortality increases, without corresponding significant changes in heat severity (Table 1, Fig A in S1 Text). Applying an exponential smoothing function (α = 0.3), following Keellings and Moradkhani [52] and Narayanan et al. [53] confirmed these regional disparities in both mortality and heat severity trends, as depicted in Fig 1.
For each year, mortality is only assessed if at least 9 cases of heat mortality (hyperthermia) are reported and therefore, trends in regions with less than 42 years should be viewed with some caution.
Focusing on regions with a consistent record of high heat-related mortality (defined as having more than 30 years of mortality data), several regions appeared to exhibit concurrent increases in both mortality and heat severity between 1995 and 2005. To identify significant shifts in temporal trends, we applied Pettitt’s test [62], a non-parametric, rank-based method commonly used to detect a single abrupt change point in a time series. Results revealed statistically significant (p < 0.05) upward shifts in heat-related mortality between 1993 and 2005 in the Central, Northeast, South, Southeast, Southwest, and West regions (Table A in S1 Text) [64]
To evaluate whether these shifts were followed by sustained trends, we conducted a Mann–Kendall test on the post-change point data for each region. In all regions except the West, the Mann–Kendall test revealed no significant monotonic trends (p < 0.05), suggesting that heat-related mortality stabilized following the initial step increase. In contrast, the West region exhibited a significant post-change upward trend (M.K. Stat = 119, p = 0.002, S.E. = 37.86), indicating a continued rise in mortality beyond the initial shift. These results point to a potential regime shift in heat-related risk conditions in most regions, where mortality levels increased abruptly and then plateaued, while the West continued on an upward trajectory.
To assess changes in mortality relative to heat severity, we calculated the Sen’s Slope Ratio (SSR) by dividing the Sen’s Slope of crude heat-mortality rate (deaths 10⁶ yr⁻¹; Table 1) by that of total heat severity (HSCIH). An SSR above 1 indicates a greater relative increase in mortality compared to heat severity, while an SSR below 1 suggests a lesser relative increase in mortality compared to heat severity. The Southeast exhibited the largest relative increase in mortality, with an SSR of 5, reflecting a disproportionate rise in deaths despite a minimal change in heat severity over the 42-year period. In contrast, the Southwest showed a more robust association between heat severity and mortality, with an SSR of 1.4 and significant trends in both variables. The Northwest, however, displayed the smallest relative mortality increase (SSR = 0.06), as its substantial heat severity rise (Sen’s Slope = 0.83) was paired with a modest mortality increase (Sen’s Slope = 0.05), suggesting potential mitigating factors in this region.
3.2. Individual characteristic analysis results
During the study period, individual heat event characteristics showed few consistent trends across U.S. climate regions (Table B and Fig B in S1 Text), consistent with prior observations by Keellings and Moradkhani [52]. However, in years with high heat-related mortality (>9 annual deaths per region), all U.S. climate regions except the Western North Central showed significant increases in heat exposure (Table 2). Applying the Benjamini–Hochberg correction, 8 of 8 temporal trends in exposure and 1 of 3 temporal trends in Event Size (Table 2) remained statistically significant at the 0.05 level. Analysis of each characteristic’s influence on mortality revealed distinct regional patterns: exposure, defined as the total number of extreme heat days, consistently drove mortality across most regions, except in the Eastern North Central, where intensity was more influential, and the Southwest, where event size played a larger role (Table 3, Fig C-E in S1 Text). In contrast, event size had a limited impact nationally, with significant effects confined to the West and Southwest. Total event intensity, measuring the degree by which temperatures exceeded 95th percentile thresholds, proved less predictive of mortality than exposure duration, suggesting that the persistence of extreme heat outweighs its intensity in driving regional death rates.
Although all three characteristics—event size, intensity, and exposure—were included in the regression models, total heat exposure (number of extreme heat days) consistently emerged as the strongest individual predictor of heat-related mortality across most U.S. climate regions (Table 4). Using a ΔAIC > 2 to identify significant changes in model performance, the simplest model using exposure alone often performed as well as or better than more complex models that incorporated additional predictors, in seven out of nine regions. The full model, which includes all three predictors, only outperformed (ΔAIC > 2) simpler models in a limited number of cases, such as in the Southwest and Eastern North Central region (Table 4). Likelihood ratio tests supported the use of the full model in some regions, but not uniformly. These findings support the dominant role of prolonged heat exposure in driving mortality and suggest that, in some regions, simpler models focusing on exposure may be sufficient for predicting heat-related health impacts.
3.3. Scenario analysis results
Scenario analysis of heat event characteristics revealed distinct regional patterns in predicted annual crude mortality rates (deaths 10⁶ yr⁻) across U.S. climate regions (Table 5). The scenario emphasizing high exposure, with low event size and intensity, consistently produced the highest mortality rates, particularly in the Northwest (7.27 deaths 10⁶ yr⁻¹), Southwest (3.35 deaths 10⁶ yr⁻¹), and South (1.93 deaths 10⁶ yr⁻¹), underscoring their vulnerability to prolonged heat exposure. The West, Central, Southeast, and Northeast exhibited more moderate increases, while the Eastern North Central showed the least sensitivity to extended heat duration, with a rate of 0.16 deaths 10⁶ yr⁻¹. In contrast, the scenario focusing on high total intensity, with low event size and exposure, yielded generally lower mortality rates, though the Southwest (2.13 deaths 10⁶ yr⁻¹) and Eastern North Central (0.55 deaths 10⁶ yr⁻¹) displayed greater responsiveness to temperature intensity. The scenario highlighting high average event size, with low intensity and exposure, identified the Southwest as most sensitive (4.25 deaths 10⁶ yr⁻¹), followed by the South (1.71 deaths 10⁶ yr⁻¹) and West (1.20 deaths 10⁶ yr⁻¹), while other regions showed moderate to low responses; the Northwest’s negative predicted rate (-0.40 deaths 10⁶ yr⁻¹) in this scenario suggests potential model limitations. The All Characteristics High and All Characteristics Low scenarios, serving as reference bounds, further contextualized the range of impacts, highlighting the varied influence of heat event characteristics on regional mortality risks.
4. Discussion
While previous studies have observed a significant decrease in U.S. heat-related mortality prior to the mid-2000s [63,64], we find that both heat-related mortality and exposure has increased throughout much of the U.S from 1981 to 2022 (Table 1). This rise in EHE days found throughout much of the U.S. mirrors a global pattern of increasing EHE days [1,3,65] and carries serious negative implications for public health. Previous studies have highlighted that both morbidity and mortality risks tend to rise on EHE days [44,66,67]. For instance, Khatana et al. [66] found that each additional day of extreme heat per month was associated with an increase of 0.07 deaths per 100,000 adults. Our study, focusing exclusively on mortality directly attributed to heat exposure, suggests that the rising trend in EHE days in regions like the Northwest, Southwest, West, South, and Southeast could lead to higher heat-related mortality in the future if these patterns persist (Table 2).
Although this analysis primarily relies on long-term trends assessed through linear regression, such methods assume gradual and continuous change and may overlook abrupt shifts in climate and health dynamics. The step changes in heat-related mortality identified by Pettitt’s test correspond with well-documented spikes in average U.S. temperatures during the late 20th century [68,69]. Notably, significant change points in total heat days were also detected in nearly all the same regions around 1999–2000, suggesting that the observed increases in mortality may reflect genuine intensification of heat exposure during this period [64] (Table A in S1 Text). While some of these changes may coincide with the 1999 transition from ICD-9 to ICD-10 coding standards, there is little evidence to suggest a significant shift in the classification of hyperthermic deaths between the two systems [70], strengthening the likelihood that these mortality increases reflect true changes in heat exposure.
Following these identified change points, Mann–Kendall trend analyses revealed no significant monotonic trends in most regions, indicating that both mortality and heat severity metrics plateaued after the sharp rise in the late 1990s to early 2000s. The West was a notable exception, with a significant post-change upward trend (p = 0.002) suggesting that heat-related risks continued to increase beyond the initial shift. Despite this exception, the broader pattern aligns with the widely discussed “hiatus” or slowdown in global surface warming from the late 1990s into the early 2010s [71]. Rather than indicating continued escalation, these trends suggest a structural transition to a new, elevated baseline of heat-related risk. However, this stabilization should not be interpreted as a return to safety: as shown in Fig 1, many of the most severe heat wave years in most regions have occurred since 2000, underscoring the ongoing intensification of extreme heat events within this new baseline.
4.1. The geography of heat
In our analysis, the three heat characteristics—total days of heat exposure, total intensity, and average event size—were significantly correlated with mortality in most regions (Table C in S1 Text), though the influence of each characteristic varied across regions (Table 3). In the Northeast, West, Western North Central, Northwest, Central, South, and Southeast regions, the total number of EHE days was the most significant predictor of annual heat mortality. This suggests that increased exposure to heat in these regions plays a central role in raising mortality risk, regardless of event size or intensity. In contrast, in the Eastern North Central and Southwest regions, other characteristics proved more influential: temperature intensity was a key driver in the Eastern North Central region, while event size had the greatest impact in the Southwest. These findings align with those of Anderson and Bell [44], who also observed regional differences in the impacts of heat characteristics. While they focused on event duration (consecutive days of heat), their results similarly emphasized the role of prolonged exposure in regions like the Northeast and Midwest, whereas other regions were more sensitive to intensity. Although our study measures cumulative rather than consecutive exposure, both analyses point to extended heat exposure as a central factor in heat-related mortality.
Regional differences in heat sensitivity may also be shaped by geographic and demographic patterns. In areas like the Southwest and West, which encompass vast expanses of sparsely populated land, heat waves can cover large geographic areas without necessarily affecting densely populated zones. Consequently, even modest increases in event size may substantially elevate the number of people exposed to hazardous heat conditions. This uneven population distribution may help explain the strong association between event size and mortality observed in the Southwest (β = 0.43, p < 0.01, B-H p = 0.02), and West (β = 0.29, p = 0.04, B-H p = 0.16; Table 3), where high heat severity may coincide with population centers. Future work should explore the integration of high-resolution gridded population and heat wave footprint data to assess exposure-adjusted event metrics that more directly capture population-level risk.
These patterns reinforce the importance of tailoring model complexity to regional dynamics. In some areas, simpler models based on a single dominant heat characteristic may perform as well as, or better than, more complex models that include multiple predictors (Table 4). However, this does not negate the value of multi-characteristic models, particularly in regions where complex interactions among variables may influence outcomes. Since heat-related mortality is generally higher when events are longer, more intense, and involve more frequent exposure (Table 5), the cumulative burden of these factors supports a holistic modeling approach—even when individual variables show the strongest associations in isolation (Table C in S1 Text). However, where data or computational capacity is limited, focusing on the most influential characteristic for a given region may still offer a practical and efficient approach to improving heat risk assessment and early warning strategies.
4.2. The role of social vulnerability and community resilience
While much attention is often placed on the characteristics of extreme heat events—such as temperature thresholds, humidity, and duration—the role of social vulnerability must also be considered when assessing health risks. Research on hazards and vulnerability has long emphasized the relationship between social conditions and recovery outcomes [72–74], leading many studies to explore how vulnerability interacts with exposure to shape overall risk [75–78].
For example, the 1995 Chicago Heatwave—which resulted in over 700 fatalities—demonstrated that social and demographic factors such as race, poverty, and social cohesion play a critical role in heat-related mortality. An ethnographic investigation of the event found that in fifteen Chicago community areas with the highest death rates, ten had populations that were between 94% and 99% African American [39]. The remaining areas were characterized by high concentrations of seniors in public housing, heat-retaining infrastructure, and the “cultural and linguistic isolation” of elderly Polish and Latino populations [39]. Notably, mortality among Latino communities was significantly lower—despite comparable levels of poverty and exposure—due in part to stronger social connectivity, intergenerational household structures, and frequent social interaction, which helped ensure that vulnerable individuals received assistance during the event [39]. These observations support this notion that social isolation, rather than age or income alone, was a key determinant of risk during this event [35,80]
Following this event, cities like Chicago implemented a wide range of social and institutional interventions—including wellness checks on isolated individuals, expanded public messaging, and the development of cooling centers and heat warning systems—to reduce population-level vulnerability [79,80]. As a result, these interventions have yielded tangible improvements [79]. During a similar heatwave in 1999, Chicago experienced nearly 80% fewer deaths compared to the 1995 event, a decline widely attributed to improved institutional coordination & emergency response systems, community outreach, and access to cooling centers [79,80]. Research has also shown that the mere presence of social connections can serve as a protective factor: individuals with strong community ties are more likely to receive timely information and assistance, especially when physically or socially isolated [35,39,81]. These observations highlight that socially cohesive neighborhoods, when combined with heat-adaptive infrastructure such as cooling centers, exhibited significantly lower mortality rates during extreme heat—even after accounting for other socioeconomic variables [39]. Further, these findings demonstrate that declines in heat mortality in some regions may not solely be the result of meteorological variation, but instead may reflect potential proactive public health measures, investments in adaptive infrastructure, and increased social capital within communities [84]
In contrast, regions where social vulnerabilities remain high—due to persistent poverty, lack of social cohesion, or limited access to adaptive infrastructure—may not experience similar improvements. For example, in this study, the Southeast and Southwest exhibited the greatest increases in mortality relative to changes in heat severity (Fig 1, Table 1). These regions are also marked by high social vulnerability, comparatively lower resilience [73,82,83], and limited social capital and economic connectivity [84,85]. Even with widespread air conditioning access [86]—typically a protective factor—these areas continue to experience rising mortality rates, indicating that technological solutions alone may be insufficient to offset broader structural and social vulnerabilities.
Ultimately, the observed divergence in heat-related mortality trends across the U.S. points to the critical role of social infrastructure, institutional preparedness, and local investment in shaping population resilience. Future heat adaptation strategies should prioritize not only physical interventions (e.g., cooling centers, improved housing) but also relational dimensions of resilience—such as social connectivity, neighborhood cohesion, and access to trusted public services—which can significantly reduce mortality risk during extreme heat events.
4.3. Urban vs. rural heat
While this study examines mortality on a regional scale, heat-related mortality is frequently concentrated in and studied within densely populated urban areas. Factors such as the urban heat island (UHI) effect [87,88] and socioeconomic disparities [38,39] increase the risk of heat-related illness and death, while strong surveillance infrastructure enhances the detection and reporting of these events. In contrast, more rural regions—such as the Western U.S., characterized by dispersed populations and mid-sized urban centers (<500,000 residents)—are often underrepresented in heat risk assessments. This is largely due to a historical emphasis on urban-centric heat dynamics, such as the UHI effect and high population density, which have guided both research and policy priorities [89,90]. However, despite receiving less policy attention, the health risks to residents in rural areas are substantial. A study on the outcomes of heat related emergency medical services (EMS) calls in the U.S. found that in the Western U.S., rural EMS calls were notably associated with worse patient outcomes compared to urban areas [89]. Specifically, the odds of rural EMS patients ending up with positive outcomes to a heat related emergency call were 54% lower than urban patients [89]. Further, rural residents are more likely to suffer from pre-existing health conditions such as heart disease and diabetes, which can increase sensitivity to extreme heat [91]. Recent investigations suggests that the relative risk of heat-related mortality is approximately 3.3% greater in rural settings than in urban settings [92]. However, despite these indicators of rural vulnerability, the reporting of heat morbidities and mortalities is suspected to be underreported [93] as a result of limited healthcare access [94,95], reduced health insurance coverage [95,96], and non-standardized methods for counting heat related deaths [97,98]. As such, interpretations of regional vulnerability in this study should be made cautiously, recognizing that rural mortality may be underestimated due to gaps in surveillance and attribution.
4.4. EHEs and air quality
Along with exacerbating underlying health conditions such as cardiovascular disease and diabetes, EHEs are also associated with increased rates of respiratory illness through the elevation of fine particulate matter (PM₂.₅) and ground-level ozone concentrations in the lower atmosphere. High temperatures can intensify atmospheric stagnation and facilitate thermal inversions, which trap pollutants near the surface and enhance inhalation exposure. These conditions are especially pronounced in urban areas where persistent anticyclonic circulation during prolonged heat events serves to retain and concentrate urban-based emissions, including PM₂.₅, PM₁₀, and ozone. This stagnant air mass reduces atmospheric mixing, creating a feedback loop where pollution levels rise as the heat event continues.
Several studies have documented that heat-driven stagnation significantly elevates concentrations of PM₂.₅ and ozone [16,25], exacerbating respiratory conditions. As a result, during heat waves, spikes in PM₂.₅ and ozone are associated with increased emergency department visits for asthma [99], higher incidence of heart attacks [57], and elevated all-cause mortality [26,100]. Therefore, relying solely on meteorological indicators of heat severity may underestimate the total health burden of EHEs, particularly in urban environments where pollution and heat synergistically interact to magnify health risks. The significant association observed in this study between heat event duration and mortality—compared to intensity alone—may reflect this compounding effect. Longer heat events provide more time for pollutants to accumulate, thereby increasing the likelihood of adverse health outcomes independent of temperature thresholds. Future studies and early warning systems should consider incorporating air quality indices (e.g., Air Quality Index (AQI), ozone levels, PM₂.₅) as co-exposures when modeling heat-related health impacts, especially in densely populated urban settings where the convergence of heat and pollution is most severe.
4.5. Study limitations
While this study provides valuable insights into the regional dynamics of heat-related mortality across the U.S., it is not without limitations, particularly regarding the definition of heat-related deaths. Within this study, heat-related mortality is defined as deaths listed with hyperthermia as the underlying (primary) cause of death. The standards for defining and reporting heat-related mortality can vary across regions and time periods, which could influence the accuracy of the data [101]. These variations may result in either underestimations or overestimations of heat-related deaths in comparison to earlier decades. Additionally, while hyperthermia is commonly used to identify heat-related mortality, it may fail to capture deaths that occur indirectly due to the exacerbation of pre-existing conditions such as cardiovascular and respiratory diseases. The availability of Underlying Cause of Death data from 1968 to the present via CDC WONDER is beneficial; however, historical Multiple Cause of Death data, which could provide deeper insights into the contributions of heat, is only available from 1999 onward. While organizations like the National Archive of Computerized Data on Aging (NACDA) and the National Bureau of Economic Research (NBER) offer access to simplified versions of this data, grouping various heat-related illnesses under general categories such as “Accidents and adverse effects (E800-E949)” or “All other external causes (E980-E999)”, this limits our ability to conduct long-term, nuanced analyses. Expanding the availability of historical data to align with post-1999 standards would significantly enhance future research in this area.
Despite these limitations, this research contributes to our understanding of heat-related risks and stresses the importance of a multifaced approach in heat risk assessments and emergency planning. Future research should work to explore linkages between human vulnerability and relationships with specific characteristics of EHEs. Further, integrating finer-scale demographic data and more detailed climate event loggings may enhance predictions and the development of mitigation strategies for the impacts of heat on vulnerable populations.
Conclusion
This study reveals a significant increase in both heat-related mortality and the number of extreme heat event (EHE) days across much of the U.S. from 1981 to 2022, consistent with global trends of intensifying heat extremes. Step changes in heat mortality—identified around the late 1990s to early 2000s—signal a structural shift toward a new baseline of elevated heat-related risk. Among the heat characteristics examined, the total number of EHE days emerged as the most consistent and influential predictor of mortality across regions, supporting the critical role of prolonged exposure in shaping public health outcomes. These findings highlight the importance of developing regionally tailored risk models, with simpler frameworks often providing effective solutions in contexts with limited data or modeling capacity.
It is important to note, however, that this analysis captures only those deaths where heat is listed as the underlying cause. As a result, the findings likely represent a conservative estimate of the full mortality burden. Deaths where heat was a contributing factor—such as those attributed to cardiovascular, renal, or respiratory conditions—are excluded, despite strong evidence that extreme heat exacerbates these underlying health issues. Consequently, the true public health burden of extreme heat is likely substantially higher than reported here.
While this study demonstrates that heat characteristics such as duration and intensity significantly influence mortality, it also affirms the role of broader social and environmental factors. Persistent social vulnerabilities, limited access to cooling infrastructure, and worsening urban air quality compound the health risks of extreme heat. To more effectively address these growing challenges, future research should advance region-specific heat risk models that integrate meteorological variables with social vulnerability metrics, measures of social connectivity, and characteristics of the built environment.
Supporting information
S1 Text.
Fig A. Regional correlations between annual crude death rate (deaths per 1,000,000) by hyperthermia and total heat severity (measured by the HSCIH), 1981–2022. Fig B. Smoothed Event Characteristic trend plots for each climate region (1981–2022; α = 0.3). Data within these plots are reported only for years with corresponding heat mortality data (only assessed if at least 9 cases of heat mortality (hyperthermia) are reported and therefore, trends in regions with less than 42 years should be viewed with some caution. Fig C. Regional correlations between annual crude death rate (deaths per 1,000,000) by hyperthermia and heat exposure days, 1981–2022. Fig D. Regional correlations between annual crude death rate (deaths per 1,000,000) by hyperthermia and total event intensity, 1981–2022. Fig E. Regional correlations between annual crude death rate (deaths per 1,000,000) by hyperthermia and average event size, 1981–2022. Table A. Significant Pettitt’s Test Results. Regions denoted with an asterisk (*) are with relatively fewer years of heat mortality data (less than 30 years of data out of 42 years total) and thus should be interpreted with extreme caution. Table B. Trends in humid heat characteristics and severity. Trends in characteristics are calculated across the 1981–2022 period. Table C. Pearson Correlation Coefficient (R) of Individual Characteristics (size, intensity, and number of days) and Total Heat Severity (HSCIH) vs. Crude Mortality Rate (Crude Death Rate per 1,000,000).
https://doi.org/10.1371/journal.pclm.0000610.s001
(DOCX)
Acknowledgments
The authors would like to express their gratitude to Dr. Kaitlyn Lawrence of the National Institute of Environmental Health (NIEH) for her valuable feedback on the methodology and manuscript. They would like to express their appreciation for the valuable feedback provided by the editor and anonymous reviewers, which significantly improved the quality of this paper.
References
- 1. Perkins SE, Alexander LV, Nairn JR. Increasing frequency, intensity and duration of observed global heatwaves and warm spells. Geophys Res Lett. 2012;39(20).
- 2. Perkins-Kirkpatrick SE, Gibson PB. Changes in regional heatwave characteristics as a function of increasing global temperature. Sci Rep. 2017;7(1):12256. pmid:28947762
- 3. Perkins-Kirkpatrick SE, Lewis SC. Increasing trends in regional heatwaves. Nat Commun. 2020;11(1):3357. pmid:32620857
- 4. Mora C, Dousset B, Caldwell IR, Powell FE, Geronimo RC, Bielecki CR, et al. Global risk of deadly heat. Nature Clim Change. 2017;7(7):501–6.
- 5. Callahan CW, Mankin JS. Globally unequal effect of extreme heat on economic growth. Sci Adv. 2022;8(43):eadd3726. pmid:36306351
- 6. Robine J-M, Cheung SLK, Le Roy S, Van Oyen H, Griffiths C, Michel J-P, et al. Death toll exceeded 70,000 in Europe during the summer of 2003. C R Biol. 2008;331(2):171–8. pmid:18241810
- 7.
United Nations Office for Disaster Risk Reduction (UNDRR). The Human Cost of Disasters 2000-2019: An Overview of the Last 20 Years. 2020.
- 8. Zhao Q, Guo Y, Ye T, Gasparrini A, Tong S, Overcenco A, et al. Global, regional, and national burden of mortality associated with non-optimal ambient temperatures from 2000 to 2019: a three-stage modelling study. Lancet Planet Health. 2021;5(7):e415–25. pmid:34245712
- 9. Luber G, McGeehin M. Climate change and extreme heat events. Am J Prev Med. 2008;35(5):429–35. pmid:18929969
- 10.
National Oceanic and Atmospheric Administration - National Centers for Environmental Information. Billion-Dollar Weather and Climate Disasters. [cited 20 Jul 2025]. Available from: https://www.ncei.noaa.gov/access/billions/events
- 11. Howard JT, Androne N, Alcover KC, Santos-Lozada AR. Trends of Heat-Related Deaths in the US, 1999-2023. JAMA. 2024;332: 1203.
- 12. Wang R, Bei N, Hu B, Wu J, Liu S, Li X, et al. The relationship between the intensified heat waves and deteriorated summertime ozone pollution in the Beijing-Tianjin-Hebei region, China, during 2013-2017. Environ Pollut. 2022;314:120256. pmid:36152720
- 13. Fortelli A, Scafetta N, Mazzarella A. Influence of synoptic and local atmospheric patterns on PM10 air pollution levels: a model application to Naples (Italy). Atmos Environ. 2016;143:218–28.
- 14. Wang Y, Mahowald N, Hess P, Sun W, Chen G. The relationship between PM2.5 and anticyclonic wave activity during summer over the United States. Atmos Chem Phys. 2022;22(11):7575–92.
- 15. Bloomer BJ, Stehr JW, Piety CA, Salawitch RJ, Dickerson RR. Observed relationships of ozone air pollution with temperature and emissions. Geophys Res Lett. 2009;36(9).
- 16. Sun W, Hess P, Liu C. The impact of meteorological persistence on the distribution and extremes of ozone. Geophys Res Lett. 2017;44(3):1545–53.
- 17. Tajudin MABA, Kubo R, Ng CFS, Hashizume M, Seposo X, Kim Y, et al. The effect modification of PM2.5 and ozone on the short-term associations between temperature and mortality across the urban areas of Japan. Environ Health Prev Med. 2024;29:57. pmid:39462582
- 18. Zhu Z, Ji B, Tian J, Yin P. Heat exposure and respiratory diseases health outcomes: An umbrella review. Sci Total Environ. 2025;970:179052. pmid:40056553
- 19. Witt C, Schubert AJ, Jehn M, Holzgreve A, Liebers U, Endlicher W, et al. The Effects of Climate Change on Patients With Chronic Lung Disease. A Systematic Literature Review. Dtsch Arztebl Int. 2015;112(51–52):878–83. pmid:26900154
- 20. Zhu Z, Deng T, Pan X. Global climate change: The dangers of heatwaves for chronic obstructive pulmonary disease patients cannot be ignored. J Glob Health. 2024;14:03032. pmid:39238356
- 21. Yin J, Wang S, Deng J, Yang N, Zhou Z, Zhou H, et al. Associations of heatwaves and their characteristics with ischaemic stroke hospital admissions. Sci Rep. 2025;15(1):4929. pmid:39929981
- 22. Souilla L, Amedro P, Morrison SA. Children With Cardiac Disease and Heat Exposure: Catastrophic Converging Consequences? Pediatr Exerc Sci. 2024;36(3):118–22. pmid:38171350
- 23. Meade RD, Akerman AP, Notley SR, McGinn R, Poirier P, Gosselin P, et al. Physiological factors characterizing heat-vulnerable older adults: A narrative review. Environ Int. 2020;144:105909. pmid:32919284
- 24. Kenny GP, Yardley J, Brown C, Sigal RJ, Jay O. Heat stress in older individuals and patients with common chronic diseases. CMAJ. 2010;182(10):1053–60. pmid:19703915
- 25. Jacob DJ, Winner DA. Effect of climate change on air quality. Atmospheric Environment. 2009;43(1):51–63.
- 26. Di Q, Dai L, Wang Y, Zanobetti A, Choirat C, Schwartz JD, et al. Association of Short-term Exposure to Air Pollution With Mortality in Older Adults. JAMA. 2017;318(24):2446–56. pmid:29279932
- 27. Simo L. The effects of PM2.5 air pollution on human health: A narrative review with a focus on cerebrovascular diseases. Environm Dis. 2024;9(4):75–8.
- 28.
Anderson RN, Miniño AM, Hoyert DL, Rosenberg HM. Clinical Outcomes Related to Particulate Matter Exposure and Cardiovascular Disease. In: The United States Environmental Protection Agency [Internet]. 1 Jul 2024 [cited 10 Jul 2025]. Available from: https://www.epa.gov/air-research/clinical-outcomes-related-particulate-matter-exposure-and-cardiovascular-disease#:~:text=inflammation%20and%20oxidative%20stress,heart%20disease%20and%20heart%20failure
- 29. Hayes RB, Lim C, Zhang Y, Cromar K, Shao Y, Reynolds HR, et al. PM2.5 air pollution and cause-specific cardiovascular disease mortality. Int J Epidemiol. 2020;49(1):25–35. pmid:31289812
- 30. Kwag Y, Kim M-H, Oh J, Shah S, Ye S, Ha E-H. Effect of heat waves and fine particulate matter on preterm births in Korea from 2010 to 2016. Environ Int. 2021;147:106239. pmid:33341584
- 31. Sheridan P, Ilango S, Bruckner TA, Wang Q, Basu R, Benmarhnia T. Ambient Fine Particulate Matter and Preterm Birth in California: Identification of Critical Exposure Windows. Am J Epidemiol. 2019;188(9):1608–15. pmid:31107509
- 32. Bignier C, Havet L, Brisoux M, Omeiche C, Misra S, Gonsard A, et al. Climate change and children’s respiratory health. Paediatr Respir Rev. 2025;53:64–73. pmid:39107182
- 33. Li S, Cao S, Duan X, Zhang Y, Gong J, Xu X, et al. Long-term exposure to PM2.5 and Children’s lung function: a dose-based association analysis. J Thorac Dis. 2020;12(10):6379–95. pmid:33209476
- 34. Zanobetti A, Ryan PH, Coull BA, Luttmann-Gibson H, Datta S, Blossom J, et al. Early-Life Exposure to Air Pollution and Childhood Asthma Cumulative Incidence in the ECHO CREW Consortium. JAMA Netw Open. 2024;7(2):e240535. pmid:38416497
- 35. Kim Y-O, Lee W, Kim H, Cho Y. Social isolation and vulnerability to heatwave-related mortality in the urban elderly population: A time-series multi-community study in Korea. Environ Int. 2020;142:105868. pmid:32593050
- 36. Lehnert EA, Wilt G, Flanagan B, Hallisey E. Spatial exploration of the CDC’s Social Vulnerability Index and heat-related health outcomes in Georgia. Int J Disaster Risk Reduct. 2020;46:101517. pmid:35923219
- 37. Manware M, Dubrow R, Carrión D, Ma Y, Chen K. Residential and Race/Ethnicity Disparities in Heat Vulnerability in the United States. Geohealth. 2022;6(12):e2022GH000695. pmid:36518814
- 38. Harlan SL, Declet-Barreto JH, Stefanov WL, Petitti DB. Neighborhood effects on heat deaths: social and environmental predictors of vulnerability in Maricopa County, Arizona. Environ Health Perspect. 2013;121(2):197–204. pmid:23164621
- 39.
Klinenberg E. Heat Wave: A Social Autopsy of Disaster in Chicago. Second edition. Chicago: University of Chicago Press; 2015.
- 40.
Multnomah County Health Department. Health Impacts from Excessive Heat Events in Multnomah County, Oregon, 2021. 2022 Jun. Available from: https://www.opb.org/pdf/multco-heat-report-final-06262022_1656296951051.pdf
- 41.
Siegel J. Change Is in the Air as the 80s Unfold. Air Conditioning Heating Refrigeration News. 2001.
- 42.
Myatt K. A silent killer, the heat wave of 1980 remains among the nation’s deadliest weather events. In: Washington Post [Internet]. 11 Jul 2020 [cited 10 Jul 2025]. Available from: https://www.washingtonpost.com/weather/2020/07/11/heat-wave-1980/.
- 43. Jones B, Dunn G, Balk D. Extreme Heat Related Mortality: Spatial Patterns and Determinants in the United States, 1979–2011. Spat Demogr. 2021;9(1):107–29.
- 44. Anderson GB, Bell ML. Heat waves in the United States: mortality risk during heat waves and effect modification by heat wave characteristics in 43 U.S. communities. Environ Health Perspect. 2011;119(2):210–8. pmid:21084239
- 45. Kent ST, McClure LA, Zaitchik BF, Smith TT, Gohlke JM. Heat waves and health outcomes in Alabama (USA): the importance of heat wave definition. Environ Health Perspect. 2014;122(2):151–8. pmid:24273236
- 46. Curriero FC, Heiner KS, Samet JM, Zeger SL, Strug L, Patz JA. Temperature and mortality in 11 cities of the eastern United States. Am J Epidemiol. 2002;155(1):80–7. pmid:11772788
- 47. Weinberger KR, Haykin L, Eliot MN, Schwartz JD, Gasparrini A, Wellenius GA. Projected temperature-related deaths in ten large U.S. metropolitan areas under different climate change scenarios. Environ Int. 2017;107:196–204. pmid:28750225
- 48. Shindell D, Zhang Y, Scott M, Ru M, Stark K, Ebi KL. The Effects of Heat Exposure on Human Mortality Throughout the United States. Geohealth. 2020;4(4):e2019GH000234. pmid:32258942
- 49. Mazdiyasni O, Sadegh M, Chiang F, AghaKouchak A. Heat wave Intensity Duration Frequency Curve: A Multivariate Approach for Hazard and Attribution Analysis. Sci Rep. 2019;9(1):14117. pmid:31575944
- 50. Wanyama D, Bunting EL, Weil N, Keellings D. Delineating and characterizing changes in heat wave events across the United States climate regions. Clim Change. 2023;176(2).
- 51. Yin Q, Wang J. The association between consecutive days’ heat wave and cardiovascular disease mortality in Beijing, China. BMC Public Health. 2017;17(1):223. pmid:28228117
- 52. Keellings D, Moradkhani H. Spatiotemporal Evolution of Heat Wave Severity and Coverage Across the United States. Geophys Res Lett. 2020;47(9).
- 53. Narayanan A, Rezaali M, Bunting EL, Keellings D. It’s getting hot in here: Spatial impact of humidity on heat wave severity in the U.S. Sci Total Environ. 2025;963:178397. pmid:39826213
- 54.
Karl T, Kloss WJ. Regional and national monthly, seasonal, and annual temperature weighted by area, 1895-1983. 1984.
- 55. Benjamini Y, Hochberg Y. Controlling the False Discovery Rate: A Practical and Powerful Approach to Multiple Testing. J Royal Stat Soc Series B Stat Methodol. 1995;57(1):289–300.
- 56. Friede A, Reid JA, Ory HW. CDC WONDER: a comprehensive on-line public health information system of the Centers for Disease Control and Prevention. Am J Public Health. 1993;83(9):1289–94. pmid:8395776
- 57. Xu R, Huang S, Shi C, Wang R, Liu T, Li Y, et al. Extreme Temperature Events, Fine Particulate Matter, and Myocardial Infarction Mortality. Circulation. 2023;148(4):312–23. pmid:37486993
- 58. Hayes D Jr, Collins PB, Khosravi M, Lin R-L, Lee L-Y. Bronchoconstriction triggered by breathing hot humid air in patients with asthma: role of cholinergic reflex. Am J Respir Crit Care Med. 2012;185(11):1190–6. pmid:22505744
- 59. Vaidyanathan A, Malilay J, Schramm P, Saha S. Heat-Related Deaths - United States, 2004-2018. MMWR Morb Mortal Wkly Rep. 2020;69(24):729–34. pmid:32555133
- 60. Mann HB. Nonparametric Tests Against Trend. Econometrica. 1945;13(3):245.
- 61. Sen PK. Estimates of the Regression Coefficient Based on Kendall’s Tau. J Am Stat Assoc. 1968;63(324):1379–89.
- 62. Pettitt AN. A non-parametric approach to the change-point problem. J Royal Stat Soc Series C (Applied Statistics). 1979;28:126–35.
- 63. Sheridan SC, Kalkstein AJ, Kalkstein LS. Trends in heat-related mortality in the United States, 1975–2004. Nat Hazards. 2008;50(1):145–60.
- 64. Bobb JF, Peng RD, Bell ML, Dominici F. Heat-related mortality and adaptation to heat in the United States. Environ Health Perspect. 2014;122(8):811–6. pmid:24780880
- 65. Zhao Q, Li S, Ye T, Wu Y, Gasparrini A, Tong S, et al. Global, regional, and national burden of heatwave-related mortality from 1990 to 2019: A three-stage modelling study. PLoS Med. 2024;21(5):e1004364. pmid:38743771
- 66. Khatana SAM, Werner RM, Groeneveld PW. Association of Extreme Heat With All-Cause Mortality in the Contiguous US, 2008-2017. JAMA Netw Open. 2022;5(5):e2212957. pmid:35587347
- 67. Arsad FS, Hod R, Ahmad N, Ismail R, Mohamed N, Baharom M, et al. The Impact of Heatwaves on Mortality and Morbidity and the Associated Vulnerability Factors: A Systematic Review. Int J Environ Res Public Health. 2022;19(23):16356. pmid:36498428
- 68.
NOAA National Centers for Environmental Information. Climate at a Glance: National Time Series,. In: National Centers for Environmental Information [Internet]. Jul 2025 [cited 13 Jul 2025]. Available from: https://www.ncei.noaa.gov/access/monitoring/climate-at-a-glance/national/time-series
- 69.
NOAA National Centers for Environmental Information. Monthly National Climate Report for Annual 1998. 1999 Jan. Available from: https://www.ncei.noaa.gov/access/metadata/landing-page/bin/iso?id=gov.noaa.ncdc:C00674
- 70. Anderson RN, Miniño AM, Hoyert DL, Rosenberg HM. Comparability of Cause of Death Between ICD-9 and ICD-10: Preliminary Estimates. NCHS Data Brief. 2011.
- 71. Wei M, Song Z, Shu QI, Yang X, Song Y, Qiao F. Revisiting the existence of the global warming slowdown during the early twenty-first century. J Clim. 2022.
- 72. Bergstrand K, Mayer B, Brumback B, Zhang Y. Assessing the Relationship Between Social Vulnerability and Community Resilience to Hazards. Soc Indic Res. 2015;122(2):391–409. pmid:29276330
- 73. Flanagan BE, Gregory EW, Hallisey EJ, Heitgerd JL, Lewis B. A Social Vulnerability Index for Disaster Management. J Homel Secur Emerg Manag. 2011;8(1).
- 74.
Tapsell S, Mccarthy S, Faulkner H, Alexander M. Social vulnerability to natural hazards Report Number WP4 Location London Deliverable Number D4.1. 2010. Available from: http://caphaz-net.org/outcomes-results/CapHaz
- 75. Bjarnadottir S, Li Y, Stewart MG. Social vulnerability index for coastal communities at risk to hurricane hazard and a changing climate. Nat Hazards. 2011;59(2):1055–75.
- 76. Pauline EL, Knox JA, Seymour L, Grundstein AJ. Revising NCEI’s Climate Extremes Index and the CDC’s Social Vulnerability Index to Analyze Climate Extremes Vulnerability across the United States. Bull Am Meteorol Soc. 2021;102(1):E84–98.
- 77. Narayanan A, Peter BG, Keellings D. A Climate Extremes Resilience Index for the Conterminous United States. Weather, Climate, and Society. 2024;16(1):87–103.
- 78.
National Risk Index Technical Documentation. 2020.
- 79.
Grossman E, Nayak SG, Roach M. The Use of Cooling Centers to Prevent Heat-Related Illness: Summary of Evidence and Strategies for Implementation Climate and Health Technical Report Series Climate. 2017. https://doi.org/10.13140/RG.2.2.32267.59688
- 80. Palecki MA, Changnon SA, Kunkel KE. The Nature and Impacts of the July 1999 Heat Wave in the Midwestern United States: Learning from the Lessons of 1995. Bull Amer Meteor Soc. 2001;82(7):1353–67.
- 81. Semenza JC, Rubin CH, Falter KH, Selanikio JD, Flanders WD, Howe HL, et al. Heat-related deaths during the July 1995 heat wave in Chicago. N Engl J Med. 1996;335(2):84–90. pmid:8649494
- 82. Cutter SL, Derakhshan S. Temporal and spatial change in disaster resilience in US counties, 2010–2015. Environmental Hazards. 2018;19(1):10–29.
- 83. Cutter SL, Finch C. Temporal and spatial changes in social vulnerability to natural hazards. Proc Natl Acad Sci U S A. 2008;105(7):2301–6. pmid:18268336
- 84.
The Geography of Social Capital in America. 2018 Apr. Available from: https://www.jec.senate.gov/public/_cache/files/e86f09f7-522a-469a-aa89-1e6d7c75628c/1-18-geography-of-social-capital.pdf
- 85. Chetty R, Jackson MO, Kuchler T, Stroebel J, Hendren N, Fluegge RB, et al. Social capital I: measurement and associations with economic mobility. Nature. 2022;608(7921):108–21. pmid:35915342
- 86.
Highlights for air conditioning in U.S. homes by state, 2020. 2023 Mar. Available from: https://www.eia.gov/consumption/residential/data/2020/state/pdf/State%20Air%20Conditioning.pdf
- 87. Li D, Bou-Zeid E. Synergistic Interactions between Urban Heat Islands and Heat Waves: The Impact in Cities Is Larger than the Sum of Its Parts*. Am Meteorol Soc. 2013;52.
- 88. Oke TR. City size and the urban heat island. Atmos Environ. 1973;7:769–79.
- 89. Ahn M, Keith L, Brown HE. Rural heat health disparities: Evidence from the U.S. National Emergency Medical Services Information System (NEMSIS). J Clim Change Health. 2025;22:100432.
- 90. Keith L, Meerow S, Wagner T. Planning for Extreme Heat: A Review. J of Extr Even. 2019;06(03n04):2050003.
- 91.
Wickerson G, Preston T, Hernandez P. Extreme heat is not just an urban threat: The risk to rural communities. In: Headwaters Economics. 13 May 2025.
- 92. Comparing Urban and Rural Vulnerability to Heat-Related Mortality: A Systematic Review and Meta-analysis. GEEH. 2017;:9–15.
- 93. Sugg MM, Konrad CE 2nd, Fuhrmann CM. Relationships between maximum temperature and heat-related illness across North Carolina, USA. Int J Biometeorol. 2016;60(5):663–75. pmid:26364040
- 94.
CCRH WP - 2025 Rural health state of the state_021125. 2025.
- 95. Maganty A, Byrnes ME, Hamm M, Wasilko R, Sabik LM, Davies BJ, et al. Barriers to rural health care from the provider perspective. Rural Remote Health. 2023;23(2):7769. pmid:37196993
- 96.
The uninsured in rural America. 2003.
- 97. Wittenberg A, Harvey C. U.S. Deaths from Heat Are Dangerously Undercounted. Scientific American. 2024.
- 98.
Snow A, LaFleur K. Inconsistent methods for counting U.S. heat deaths stymie public health efforts. In: PBS. 13 Aug 2023.
- 99. Mirabelli MC, Vaidyanathan A, Flanders WD, Qin X, Garbe P. Outdoor PM2.5, Ambient Air Temperature, and Asthma Symptoms in the Past 14 Days among Adults with Active Asthma. Environ Health Perspect. 2016;124(12):1882–90. pmid:27385358
- 100. Rahman MM, McConnell R, Schlaerth H, Ko J, Silva S, Lurmann FW, et al. The Effects of Coexposure to Extremes of Heat and Particulate Air Pollution on Mortality in California: Implications for Climate Change. Am J Respir Crit Care Med. 2022;206(9):1117–27. pmid:35727303
- 101. Ellis FP. Mortality from heat illness and heat-aggravated illness in the United States. Environ Res. 1972;5(1):1–58. pmid:4555874