How do urban green spaces influence heat-related mortality in elderly? A realist synthesis

BACKGROUND

An important consequence of climate change for urban health is heat-related mortality. Vulnerable groups, especially elderly, will be the most affected. A solution put forward in many reports and policy documents is the introduction or expansion of urban green spaces. While they have a proven effect in decreasing the ambient temperature and reducing heat related mortality, the causal pathways are far from clear. Moreover, results vary for different contexts, population types and characteristics of green spaces as they are 'complex systems thrusted into complex systems'. To our knowledge, there is no systematic synthesis of the literature that examines the mechanisms by which and the circumstances under which green spaces work to decrease heat-related mortality for elderly.

METHODS

We performed a realist synthesis- a theory-driven review method- to develop a complexity- and context-sensitive program theory. As a first step, a causal loop diagram was constructed which describes the possible pathways through which urban green spaces influence heat-related mortality in elderly. In a second step, one of the pathways - how they may lead to a reduction of heat-related mortality by increasing social capital - was further explored for underlying mechanisms, the context in which they work and the differentiated patterns of outcomes they generate. Literature was searched for evidence supporting or contradicting the initial programme theory, resulting in a refined theory.

RESULTS

Results show how urban green space can impact on heat-related mortality in elderly by its influence on their exposure to outdoor and indoor heat, by improving their resilience as well as by affecting their access to treatment. Urban green spaces and their interactions with social capital affect the access to health information, social support, and the capacity for effective lobbying. Several mechanisms help to explain these observed demi-regularities, among others perceived behavioural control, perceived usefulness, receptiveness, ontological security, and self-interest. If and how they are triggered depends on the characteristics of the urban green space, the population, and other contextual factors.

CONCLUSION

Looking into the impact of urban green spaces on heat-related mortality in elderly, researchers and policy makers should take interest in the role of social capital.

Avoidable heat risk under scenarios of carbon neutrality by mid-century

To prevent anthropogenic warming of the climate system above dangerous thresholds, governments are required by the Paris Agreement to peak global anthropogenic CO₂ emissions and to reach a net zero CO₂ emissions level (also known as carbon neutrality). Growing concerns are being expressed about the increasing heat stress caused by the interaction of changes in temperature and humidity in the context of global warming. Although much effort has been made to examine future changes in heat stress and associated risks, gaps remain in understanding the quantitative benefits of heat-risk avoidance from carbon-neutral policies, limited by the traditional climate projections from the Coupled Model Intercomparison Project Phase 6 (CMIP6). Here we quantify the avoided heat risk during 2040-2049 under two scenarios of global carbon neutrality by 2060 and 2050, i.e., moderate green (MODGREEN) and strong green (STRGREEN) recovery scenarios, relative to the baseline scenario (FOSSIL), based on multi-model large ensemble climate projections from a new climate model intercomparison project (CovidMIP) that endorsed by CMIP6. We show that global population exposure to extreme heat stress increases by approximately four times its current level during 2040-2049 under the FOSSIL scenario, whereas the heat exposure could be reduced by as much as 12 % and 23 % under the MODGREEN and STRGREEN scenarios, respectively. Moreover, global mean heat-related mortality risk is mitigated by 14 % (24 %) under the MODGREEN (STRGREEN) scenario during 2040-2049 relative to the FOSSIL scenario. Additionally, the aggravating heat risk could be mitigated by around a tenth by achieving carbon neutrality 10 years earlier (2050 versus 2060). In terms of spatial pattern, this heat-risk avoidance from low-carbon policies is typically greater in low-income countries. Our findings assist governments in advancing early climate change mitigation policy-making.

Human tolerance to extreme heat: evidence from a desert climate population

BACKGROUND

Ambient temperatures exceeding 40 °C are projected to become common in many temperate climatic zones due to global warming. Therefore, understanding the health effects of continuous exposure to high ambient temperatures on populations living in hot climatic regions can help identify the limits of human tolerance.

OBJECTIVE

We studied the relationship between ambient temperature and non-accidental mortality in the hot desert city of Mecca, Saudi Arabia, between 2006 and 2015.

METHODS

We used a distributed lag nonlinear model to estimate the mortality-temperature association over 25 days of lag. We determined the minimum mortality temperature (MMT) and the deaths that are attributable to heat and cold.

RESULTS

We analyzed 37,178 non-accidental deaths reported in the ten-year study period among Mecca residents. The median average daily temperature was 32 °C (19-42 °C) during the same study period. We observed a U-shaped relationship between daily temperature and mortality with an MMT of 31.8 °C. The total temperature-attributable mortality of Mecca residents was 6.9% (-3.2; 14.8) without reaching statistical significance. However, extreme heat, higher than 38 °C, was significantly associated with increased risk of mortality. The lag structure effect of the temperature showed an immediate impact, followed by a decline in mortality over many days of heat. No effect of cold on mortality was observed.

IMPACT STATEMENT

High ambient temperatures are projected to become future norms in temperate climates. Studying populations familiar with desert climates for generations with access to air-conditioning would inform on the mitigation measures to protect other populations from heat and on the limits of human tolerance to extreme temperatures. We studied the relationship between ambient temperature and all-cause mortality in the hot desert city of Mecca. We found that Mecca population is adapted to high temperatures, although there was a limit to tolerance to extreme heat. This implies that mitigation measures should be directed to accelerate individual adaptation to heat and societal reorganization.

Heat strain and mortality effects of prolonged central European heat wave-an example of June 2019 in Poland

The occurrence of long-lasting severe heat stress, such as in July-August 2003, July 2010, or in April-May 2018 has been one of the biggest meteorological threats in Europe in recent years. The paper focuses on the biometeorological and mortality effects of the hot June that was observed in Central Europe in 2019. The basis of the study was hourly and daily Universal Thermal Climate Index (UTCI) values at meteorological stations in Poland for June 2019. The average monthly air temperature and UTCI values from 1951 to 2018 were analysed as background. Grosswetterlagen calendar of atmospheric circulation was used to assess synoptic conditions of heat wave. Several heat strain measures were applied : net heat storage (S), modelled heart rate (HR), sultriness (HSI), and UTCI index. Actual total mortality (TM) and modelled strong heat-related mortality (SHRM) were taken as indicators of biometeorological consequences of the hot June in 2019. The results indicate that prolonged persistence of unusually warm weather in June 2019 was determined by the synoptic conditions occurring over the European region and causing advection of tropical air. They led to the emergence of heat waves causing 10% increase in TM and 5 times bigger SHRM then in preceding 10 years. Such increase in SHRM was an effect of overheating and overload of circulatory system of human organism.

Human health outcomes at the neighbourhood scale implications: Elderly's heat-related cardiorespiratory mortality and its influencing factors

The excessively warm weather, especially in cities, can lead to several adverse impacts, including heat-related mortality, becoming an increasingly important public health issue. Previous studies on heat-related mortality have assessed risk factors at the municipal scale, missing the intra-urban variability in heat risk and vulnerability. The knowledge of the spatial intra-variability can help to design spatially targeted measures to better protect citizens' health. Through hot spot analysis, we identified the neighbourhood-scale spatial pattern of heat-related cardiorespiratory mortality in the elderly, during the yearly warmest five months of a three years period. Potential associations between spatial variability in heat-related mortality and several independent factors in each neighbourhood were investigated and their predictions. Two approaches were adopted: one is eminently statistical, using Generalized Linear Models (GLM) and another using Geographically Weighted Regression (GWR). This new recent regression technique is increasing in international attention on spatial modelling. The spatial model explains about 60% of the spatial variations in elderly's heat-related cardiorespiratory mortality. The two-analyses produced an overlapping set of predictor variables, with emphasis on the elderly, vegetation cover and employment. The results also show that the areas where heat-related mortality is high, are also the areas where the number of deaths is higher than expected. These neighbourhoods should be considered as the most vulnerable to heat-related mortality. We concluded that studying human health outcomes at neighbourhood-scale is relevant for public health heat-related plans. Essential suggestions are provided to decision-making support and city planners designing strategies to reduce heat-related mortality.

A simple high-resolution heat-stress forecast for Seoul, Korea: coupling climate information with an operational numerical weather prediction model

To provide a simple high-resolution heat-stress forecast for Seoul, Korea, we coupled a high-resolution climate simulation (25 m grid spacing) for an average heat day with the operational forecasting model (5 km grid spacing). Thereby, we accounted for the meso-scale weather conditions and local-scale air temperature induced by land cover and the urban heat island effect. Moreover, we estimated the impacts of heat events using heat-related mortality rate. Applying the simple high-resolution heat-stress forecast for July and August 2016, we detected a substantial spatial variability in maximum air temperature and heat-related mortality rate in Seoul. The evaluation of simulated maximum air temperature compared to observations revealed a small deviation (MB = 0.11 K, RMSD = 1.40 K). Despite the limitation of using average conditions, it was an efficient way to identify particularly affected areas, neighbourhoods, and districts for releasing more location-specific heat-stress warnings.

Temporal changes in years of life lost associated with heat waves in the Czech Republic

Seniors constitute the population group generally most at risk of mortality due to heat stress. As life expectancy increases and health conditions of elderly people improve over time, vulnerability of the population to heat changes as well. We employed the years-of-life-lost (YLL) approach, considering life expectancy at the time of each death, to investigate how population ageing affects temporal changes in heat-related mortality in the Czech Republic. Using an updated gridded meteorological database, we identified heat waves during 1994-2017, and analysed temporal changes in their impacts on YLL and mortality. The mean impact of a heat-wave day on relative excess mortality and YLL had declined by approximately 2-3% per decade. That decline abated in the current decade, however, and the decreasing trend in mean excess mortality as well as YLL vanished when the short-term mortality displacement effect was considered. Moreover, the cumulative number of excess deaths and YLL during heat waves rose due to increasing frequency and intensity of heat waves during the examined period. The results show that in studies of temporal changes it is important to differentiate between mean effects of heat waves on mortality and the overall death burden associated with heat waves. Analysis of the average ratio of excess YLL/death per heat-wave day indicated that the major heat-vulnerable population group shifted towards older age (70+ years among males and 75+ years among females). Our findings highlight the importance of focusing heat-protection measures especially upon the elderly population, which is most heat-vulnerable and whose numbers are rising.

The predictability of heat-related mortality in Prague, Czech Republic, during summer 2015-a comparison of selected thermal indices

We compared selected thermal indices in their ability to predict heat-related mortality in Prague, Czech Republic, during the extraordinary summer 2015. Relatively, novel thermal indices-Universal Thermal Climate Index and Excess Heat Factor (EHF)-were compared with more traditional ones (apparent temperature, simplified wet-bulb globe temperature (WBGT), and physiologically equivalent temperature). The relationships between thermal indices and all-cause relative mortality deviations from the baseline (excess mortality) were estimated by generalized additive models for the extended summer season (May-September) during 1994-2014. The resulting models were applied to predict excess mortality in 2015 based on observed meteorology, and the mortality estimates by different indices were compared. Although all predictors showed a clear association between thermal conditions and excess mortality, we found important variability in their performance. The EHF formula performed best in estimating the intensity of heat waves and magnitude of heat-impacts on excess mortality on the most extreme days. Afternoon WBGT, on the other hand, was most precise in the selection of heat-alert days during the extended summer season, mainly due to a relatively small number of "false alerts" compared to other predictors. Since the main purpose of heat warning systems is identification of days with an increased risk of heat-related death rather than prediction of exact magnitude of the excess mortality, WBGT seemed to be a slightly favorable predictor for such a system.

Quantification and evaluation of intra-urban heat-stress variability in Seoul, Korea

This study quantifies heat-stress hazard (air temperature), vulnerability (heat vulnerability index and age score), and risk (heat-related mortality) on the district scale in Seoul, Korea, for a comprehensive heat-stress impact assessment. Moreover, the heat-stress impact assessment is evaluated by checking the spatial consistency between heat-stress hazard, vulnerability, and risk, which was rarely done before. We applied numerical and geo-empirical models to simulate the spatial pattern of heat-stress hazard. For heat-stress vulnerability, we used demographic and socioeconomic factors. Heat-related mortality was estimated based on an event-based heat-stress risk analysis. Results are that heat-stress hazard, vulnerability, and risk are spatially variable in Seoul. The highest heat-stress hazard was detected in the districts Mapo, Yeongdeungpo, and Yangcheon, the highest vulnerability in Jongno and the highest risk in Jongno and Yangcheon. The different components (heat-stress hazard, vulnerability, and risk) and variables (heat vulnerability index and percentage of seniors) showed different spatial patterns. Knowledge about the causes of higher heat-stress risk, either the hazard or vulnerability, is helpful to design tailored adaptation measures that focus on the reduction of thermal loads or on the preparation of the vulnerable population. The evaluation showed that heat-stress vulnerability and hazard explain the spatial pattern of risk only partly. This highlights the need to evaluate heat-stress impact assessment systems to produce reliable urban heat-stress maps.

Cold-related mortality vs heat-related mortality in a changing climate: A case study in Vilnius (Lithuania)

INTRODUCTION

Direct health effects of extreme temperatures are a significant environmental health problem in Lithuania, and could worsen further under climate change. This paper attempts to describe the change in environmental temperature conditions that the urban population of Vilnius could experience under climate change, and the effects such change could have on excess heat-related and cold-related mortality in two future periods within the 21st century.

METHODS

We modelled the urban climate of Vilnius for the summer and winter seasons during a sample period (2009-2015) and projected summertime and wintertime daily temperatures for two prospective periods, one in the near (2030-2045) and one in the far future (2085-2100), under the Representative Concentration Pathway (RCP) 8.5. We then analysed the historical relationship between temperature and mortality for the period 2009-2015, and estimated the projected mortality in the near future and far future periods under a changing climate and population, assuming alternatively no acclimatisation and acclimatisation to heat and cold based on a constant-percentile threshold temperature.

RESULTS

During the sample period 2009-2015 in summertime we observed an increase in daily mortality from a maximum daily temperature of 30 °C (the 96th percentile of the series), with an average of around 7 deaths per year. Under a no acclimatisation scenario, annual average heat-related mortality would rise to 24 deaths/year (95% CI: 8.4-38.4) in the near future and to 46 deaths/year (95% CI: 16.4-74.4) in the far future. Under a heat acclimatisation scenario, mortality would not increase significantly in the near or in the far future. Regarding wintertime cold-related mortality in the sample period 2009-2015, we observed increased mortality on days on which the minimum daily temperature fell below - 12 °C (the 7th percentile of the series), with an average of around 10 deaths a year. Keeping the threshold temperature constant, annual average cold-related mortality would decrease markedly in the near future, to 5 deaths/year (95% CI: 0.8-7.9) and even more in the far future, down to 0.44 deaths/year (95% C: 0.1-0.8). Assuming a "middle ground" between the acclimatisation and non-acclimatisation scenarios, the decrease in cold-related mortality will not compensate the increase in heat-related mortality.

CONCLUSION

Thermal extremes, both heat and cold, constitute a serious public health threat in Vilnius, and in a changing climate the decrease in mortality attributable to cold will not compensate for the increase in mortality attributable to heat. Study results reinforce the notion that public health prevention against thermal extremes should be designed as a dynamic, adaptive process from the inception.

Projecting future climate change impacts on heat-related mortality in large urban areas in China

Global climate change is anticipated to raise overall temperatures and has the potential to increase future mortality attributable to heat. Urban areas are particularly vulnerable to heat because of high concentrations of susceptible people. As the world's largest developing country, China has experienced noticeable changes in climate, partially evidenced by frequent occurrence of extreme heat in urban areas, which could expose millions of residents to summer heat stress that may result in increased health risk, including mortality. While there is a growing literature on future impacts of extreme temperatures on public health, projecting changes in future health outcomes associated with climate warming remains challenging and underexplored, particularly in developing countries. This is an exploratory study aimed at projecting future heat-related mortality risk in major urban areas in China. We focus on the 51 largest Chinese cities that include about one third of the total population in China, and project the potential changes in heat-related mortality based on 19 different global-scale climate models and three Representative Concentration Pathways (RCPs). City-specific risk estimates for high temperature and all-cause mortality were used to estimate annual heat-related mortality over two future twenty-year time periods. We estimated that for the 20-year period in Mid-21st century (2041-2060) relative to 1970-2000, incidence of excess heat-related mortality in the 51 cities to be approximately 37,800 (95% CI: 31,300-43,500), 31,700 (95% CI: 26,200-36,600) and 25,800 (95% CI: 21,300-29,800) deaths per year under RCP8.5, RCP4.5 and RCP2.6, respectively. Slowing climate change through the most stringent emission control scenario RCP2.6, relative to RCP8.5, was estimated to avoid 12,900 (95% CI: 10,800-14,800) deaths per year in the 51 cities in the 2050s, and 35,100 (95% CI: 29,200-40,100) deaths per year in the 2070s. The highest mortality risk is primarily in cities located in the North, East and Central regions of China. Population adaptation to heat is likely to reduce excess heat mortality, but the extent of adaptation is still unclear. Future heat mortality risk attributable to exposure to elevated warm season temperature is likely to be considerable in China's urban centers, with substantial geographic variations. Climate mitigation and heat risk management are needed to reduce such risk and produce substantial public health benefits.

Heat and health in Antwerp under climate change: Projected impacts and implications for prevention

BACKGROUND

Excessive summer heat is a serious environmental health problem in several European cities. Heat-related mortality and morbidity is likely to increase under climate change scenarios without adequate prevention based on locally relevant evidence.

METHODS

We modelled the urban climate of Antwerp for the summer season during the period 1986-2015, and projected summer daily temperatures for two periods, one in the near (2026-2045) and one in the far future (2081-2100), under the Representative Concentration Pathway (RCP) 8.5. We then analysed the relationship between temperature and mortality, as well as with hospital admissions for the period 2009-2013, and estimated the projected mortality in the near future and far future periods under changing climate and population, assuming alternatively no acclimatization and acclimatization based on a constant threshold percentile temperature.

RESULTS

During the sample period 2009-2013 we observed an increase in daily mortality from a maximum daily temperature of 26°C, or the 89th percentile of the maximum daily temperature series. The annual average heat-related mortality in this period was 13.4 persons (95% CI: 3.8-23.4). No effect of heat was observed in the case of hospital admissions due to cardiorespiratory causes. Under a no acclimatization scenario, annual average heat-related mortality is multiplied by a factor of 1.7 in the near future (24.1deaths/year CI 95%: 6.78-41.94) and by a factor of 4.5 in the far future (60.38deaths/year CI 95%: 17.00-105.11). Under a heat acclimatization scenario, mortality does not increase significantly in the near or in the far future.

CONCLUSION

These results highlight the importance of a long-term perspective in the public health prevention of heat exposure, particularly in the context of a changing climate, and the calibration of existing prevention activities in light of locally relevant evidence.

Small-area spatiotemporal analysis of heatwave impacts on elderly mortality in Paris: A cluster analysis approach

BACKGROUND

Heat-waves have a substantial public health burden. Understanding spatial heterogeneity at a fine spatial scale in relation to heat and related mortality is central to target interventions towards vulnerable communities.

OBJECTIVES

To determine the spatial variability of heat-wave-related mortality risk among elderly in Paris, France at the census block level. We also aimed to assess area-level social and environmental determinants of high mortality risk within Paris.

METHODS

We used daily mortality data from 2004 to 2009 among people aged >65 at the French census block level within Paris. We used two heat wave days' definitions that were compared to non-heat wave days. A Bernoulli cluster analysis method was applied to identify high risk clusters of mortality during heat waves. We performed random effects meta-regression analyses to investigate factors associated with the magnitude of the mortality risk.

RESULTS

The spatial approach revealed a spatial aggregation of death cases during heat wave days. We found that small scale chronic PM₁₀ exposure was associated with a 0.02 (95% CI: 0.001; 0.045) increase of the risk of dying during a heat wave episode. We also found a positive association with the percentage of foreigners and the percentage of labor force, while the proportion of elderly people living in the neighborhood was negatively associated. We also found that green space density had a protective effect and inversely that the density of constructed feature increased the risk of dying during a heat wave episode.

CONCLUSION

We showed that a spatial variation in terms of heat-related vulnerability exists within Paris and that it can be explained by some contextual factors. This study can be useful for designing interventions targeting more vulnerable areas and reduce the burden of heat waves.

Projected heat-related mortality under climate change in the metropolitan area of Skopje

BACKGROUND

Excessive summer heat is a serious environmental health problem in Skopje, the capital and largest city of the former Yugoslav Republic of Macedonia. This paper attempts to forecast the impact of heat on mortality in Skopje in two future periods under climate change and compare it with a historical baseline period.

METHODS

After ascertaining the relationship between daily mean ambient air temperature and daily mortality in Skopje, we modelled the evolution of ambient temperatures in the city under a Representative Concentration Pathway scenario (RCP8.5) and the evolution of the city population in two future time periods: 2026-2045 and 2081-2100, and in a past time period (1986-2005) to serve as baseline for comparison. We then calculated the projected average annual mortality attributable to heat in the absence of adaptation or acclimatization during those time windows, and evaluated the contribution of each source of uncertainty on the final impact.

RESULTS

Our estimates suggest that, compared to the baseline period (1986-2005), heat-related mortality in Skopje would more than double in 2026-2045, and more than quadruple in 2081-2100. When considering the impact in 2081-2100, sampling variability around the heat-mortality relationship and climate model explained 40.3 and 46.6 % of total variability.

CONCLUSION

These results highlight the importance of a long-term perspective in the public health prevention of heat exposure, particularly in the context of a changing climate.

The application of the European heat wave of 2003 to Korean cities to analyze impacts on heat-related mortality

The goal of this research is to transpose the unprecedented 2003 European excessive heat event to six Korean cities and to develop meteorological analogs for each. Since this heat episode is not a model but an actual event, we can use a plausible analog to assess the risk of increasing heat on these cities instead of an analog that is dependent on general circulation (GCM) modeling or the development of arbitrary scenarios. Initially, the 2003 summer meteorological conditions from Paris are characterized statistically and these characteristics are transferred to the Korean cites. Next, the new meteorological dataset for each Korean city is converted into a daily air mass calendar. We can then determine the frequency and character of "offensive" air masses in the Korean cities that are historically associated with elevated heat-related mortality. One unexpected result is the comparative severity of the very hot summer of 1994 in Korea, which actually eclipsed the 2003 analog. The persistence of the offensive air masses is considerably greater for the summer of 1994, as were dew point temperatures for a majority of the Korean cities. For all the Korean cities but one, the summer of 1994 is associated with more heat-related deaths than the analog summer, in some cases yielding a sixfold increase over deaths in an average summer. The Korean cities appear less sensitive to heat-related mortality problems during very hot summers than do large eastern and Midwestern US cities, possibly due to a lesser summer climate variation and efficient social services available during extreme heat episodes.

Intra-urban vulnerability to heat-related mortality in New York City, 1997-2006

The health impacts of exposure to summertime heat are a significant problem in New York City (NYC) and for many cities and are expected to increase with a warming climate. Most studies on heat-related mortality have examined risk factors at the municipal or regional scale and may have missed the intra-urban variation of vulnerability that might inform prevention strategies. We evaluated whether place-based characteristics (socioeconomic/demographic and health factors, as well as the built and biophysical environment) may be associated with greater risk of heat-related mortality for seniors during heat events in NYC. As a measure of relative vulnerability to heat, we used the natural cause mortality rate ratio among those aged 65 and over (MRR₆₅₊), comparing extremely hot days (maximum heat index 100 °F+) to all warm season days, across 1997–2006 for NYC's 59 Community Districts and 42 United Hospital Fund neighborhoods. Significant positive associations were found between the MRR₆₅₊ and neighborhood-level characteristics: poverty, poor housing conditions, lower rates of access to air-conditioning, impervious land cover, surface temperatures aggregated to the area-level, and seniors’ hypertension. Percent Black/African American and household poverty were strong negative predictors of seniors’ air conditioning access in multivariate regression analysis.

Challenges associated with projecting urbanization-induced heat-related mortality

Maricopa County, Arizona, anchor to the fastest growing megapolitan area in the United States, is located in a hot desert climate where extreme temperatures are associated with elevated risk of mortality. Continued urbanization in the region will impact atmospheric temperatures and, as a result, potentially affect human health. We aimed to quantify the number of excess deaths attributable to heat in Maricopa County based on three future urbanization and adaptation scenarios and multiple exposure variables. Two scenarios (low and high growth projections) represent the maximum possible uncertainty range associated with urbanization in central Arizona, and a third represents the adaptation of high-albedo cool roof technology. Using a Poisson regression model, we related temperature to mortality using data spanning 1983-2007. Regional climate model simulations based on 2050-projected urbanization scenarios for Maricopa County generated distributions of temperature change, and from these predicted changes future excess heat-related mortality was estimated. Subject to urbanization scenario and exposure variable utilized, projections of heat-related mortality ranged from a decrease of 46 deaths per year (-95%) to an increase of 339 deaths per year (+359%). Projections based on minimum temperature showed the greatest increase for all expansion and adaptation scenarios and were substantially higher than those for daily mean temperature. Projections based on maximum temperature were largely associated with declining mortality. Low-growth and adaptation scenarios led to the smallest increase in predicted heat-related mortality based on mean temperature projections. Use of only one exposure variable to project future heat-related deaths may therefore be misrepresentative in terms of direction of change and magnitude of effects. Because urbanization-induced impacts can vary across the diurnal cycle, projections of heat-related health outcomes that do not consider place-based, time-varying urban heat island effects are neglecting essential elements for policy relevant decision-making.