Next Article in Journal
Climate-Conscious Medication Therapy Management: Perspectives of Canadian Primary Care Pharmacists
Previous Article in Journal
Under the Heat of Tradition: Thermal Comfort During Summer Correfocs in Catalonia (1950–2023)
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
Climate
Volume 14
Issue 1
10.3390/cli14010016
climate-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

share Share announcement Help format_quote Cite question_answer Discuss in SciProfiles
first_page
settings
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Determinants of Spatial Variation in Vulnerability to Extreme Temperatures in Austria from 1970 to 2020

by
Hanns Moshammer
1,2,*,
Martin Jury
3,
Hans-Peter Hutter
1 and
Peter Wallner
1
1
Department for Environmental Health, Center for Public Health, Medical University of Vienna, 1090 Vienna, Austria
2
Karakalpakstan Medical Institute, Nukus 230100, Uzbekistan
3
Wegener Center for Climate and Global Change, University of Graz, 8010 Graz, Austria
*
Author to whom correspondence should be addressed.
Climate 2026, 14(1), 16; https://doi.org/10.3390/cli14010016
Submission received: 3 December 2025 / Revised: 12 January 2026 / Accepted: 12 January 2026 / Published: 13 January 2026
(This article belongs to the Section Weather, Events and Impacts)
Browse Figures
Review Reports Versions Notes

Abstract

Vulnerability to heat and cold is influenced by many characteristics. This study analyzed determinants of vulnerability at the district level in the whole of Austria. Daily deaths (1970–2020) and daily temperatures per district were entered into time series models using negative binomial General Additive Models controlling for long-term and seasonal trends and for the day of the week. District-wise effect estimates of 111 districts in total were entered into linear meta-regression models seeking determinants of inter-district variation in heat and cold vulnerability. Generally, temperature effects on the daily number of deaths were highly significant in all districts, with higher death rates occurring when the same-day temperature exceeded a clear threshold and higher death rates with declining temperature averaged over the previous 14 days, in that case not showing any clear threshold effect. A higher heat vulnerability was observed for more densely populated areas, especially for the city of Vienna, for districts with a higher percentage of singles, of homeless people, of unemployed, and of migrants. Surprisingly, a higher percentage of outdoor workers seemed to be protective. Higher cold vulnerability was found for an increasingly autochthonous population, for districts with a higher employment rate, with more commuters, more agricultural workers, and more green spaces.

Graphical Abstract

1. Introduction

Social vulnerability and resilience are complex issues, and multiple definitions, applications, and indices of vulnerability exist [1,2]. Within the context of climate change, the Intergovernmental Panel on Climate Change [3] states in the 6th Assessment Report that vulnerability “encompasses a variety of concepts and elements, including sensitivity or susceptibility to harm and lack of capacity to cope and adapt.” Therefore, studies on social vulnerability and extreme temperatures must address socioecological and sociodemographic complexities [4,5,6].
Persons with low socioeconomic status, the elderly, children, outdoor workers, migrants, socially isolated, and people with medical conditions are known to be more heat vulnerable [7,8,9,10]. Less is known about which groups are especially vulnerable to cold. The literature primarily mentions elderly people, people living in deprived areas, and persons with certain pre-existing medical conditions, especially cardiovascular and respiratory diseases [11,12,13,14].
Socioeconomic variables are interconnected in many ways and are associated with characteristics that influence heat- and cold-related health risks, such as living in neighborhoods with less green and higher temperatures or in insufficiently insulated homes in rural regions with poor medical infrastructure [15,16].
Vulnerability factors must be considered on an individual basis, as well as at group and community levels. Partially, communities with a higher percentage of vulnerable individuals display greater vulnerability at the group level [17,18]. However, even individuals who themselves do not display signs of individual vulnerability but live among vulnerable communities could be affected [19].
Different cities may have different sets of heat vulnerability variables [20]. In a study from Switzerland, socioeconomic variables were important determinants for heat-mortality risks in peri-urban and rural areas, while for cold, they influenced vulnerability across all types of areas [6].
In Chinese cities, illiterate people were shown to be more heat- and cold-vulnerable [13,21], and unemployment also increased vulnerability to temperature extremes [8,22].
In addition to older people, children, especially babies, are also a risk group for thermal stress [8,10,23,24]. In some studies, younger age groups were shown to be (more) vulnerable [25,26,27,28,29,30]. Vulnerability in younger workers, as demonstrated by heat illness and injuries, may be explained by a lack of experience and awareness [28,31,32]. In the Czech Republic, cold effects were most pronounced in middle-aged men; this was attributed to working outdoors in winter [33].
Karthick et al. [34] state in a review that older construction workers (over 55 years) are more affected by heat stress. Petiti et al. [35] reported that older workers (≥65) in agricultural occupations were at high risk of heat-associated death. Workers aged ≥55 years in agriculture, forestry, and fishing, in the electricity, gas, and water industry, and tradespersons had a higher risk of injury and heat illness during heatwaves in Adelaide [36].
Not only urban but also rural populations may be vulnerable to heat [24,37,38], potentially due to factors like agricultural labor, worse medical infrastructure, and the aging population [4,39,40]. During heatwaves in Western Australia, increased health service usage (hospitalization and ED attendance) was attributed to remoteness [41]. A rather complex role of urbanization and rurality in heat vulnerability was found in Korea: the association between population density and heat-mortality risk was U-shaped; rural populations showed the highest risk [42]. The effects of heat and cold on cardiovascular mortality were higher in suburban Beijing than in the urban areas [43]. In urban areas, heat-mortality risk increases with population density. This association has been found in many studies [8].
Future temperature increases are predicted to dominate heat-related deaths at high latitudes and high altitudes in China, while in other regions population changes will likely have a greater impact than the temperature [44].
Mortality during heatwaves was greater in districts with a low proportion of green spaces [19,45]. Urban green and blue spaces reduced heat-related mortality in the elderly population of Lisbon [46], and persons living in greener areas of New York City were less likely to die during heatwaves [47].
Poor housing conditions led to stronger heat and cold effects [48,49,50,51]. Lack of thermal insulation and living on the top floor were risk factors in the devastating 2003 heatwave in France [52], whilst having additional rooms reduced the risk. Furthermore, a systematic review identified lack of access to air conditioning as a vulnerability factor [24]. Future heat health risks in Beijing will be reduced slightly by improved urbanization; however, there will be a considerable increase in risk overall because of an aging population [53].
Ethnic/racial minorities and migrants are other vulnerable groups [5,8,10,26,48]. Apart from socioeconomic reasons, occupation, social isolation, and linguistic isolation/language barriers [8,54,55] may play a role.
Further risk factors for heat vulnerability include living alone, being widowed, divorced, separated or never married [56,57,58], or being a tourist [54], while homeless individuals are vulnerable to both heat and cold [50,59,60,61].
With regard to gender, the majority of studies highlighted that females are at higher risk during heatwaves [62,63,64]. However, Ngarambe et al. [65], Kathana et al. [66], and Salvador et al. [67] reported the opposite results. Varghese et al. [28], McInnes et al. [32], and Adam-Poupart et al. [31] reported that male workers are at higher risk. Mashhoodi [68] reported that women were exposed to higher surface temperatures in 2400 Dutch residential zones and argued that this phenomenon is more likely to occur in moderate climates with their high variations in terrestrial surface temperatures. Men seem to be more vulnerable to cold than women [69].
Therefore, this study aimed at analyzing determinants of vulnerability to extreme temperatures at the district level in the whole of Austria. This was performed within the Austrian project DISCC-AT, and parts of this paper (literature review, first results of the meta-analysis) have already been published on the project website (https://wegcwp.uni-graz.at/discc-at/, accessed 11 January 2026).

2. Materials and Methods

Investigating district-specific indicators of heat and cold vulnerability was performed in two steps. In the first step, temperature effects on mortality were calculated for each district of Austria separately. In the second step, in a meta-regression, the associations between the coefficients of effect and district-wide parameters were assessed.

2.1. Data Collection: Mortality and Temperature Data

Since the cause of death for temperature-related mortality is not very specific, we decided to analyze daily total mortality (and not mortality from different causes separately) per district in Austria. This decision was also motivated by the fact that for districts with fewer inhabitants, daily deaths even from main causes are often rare and are therefore difficult to analyze precisely. To increase statistical power and generate more precise effect estimates, we obtained daily deaths (per district of the home address) for a long time period, namely 1970 to 2020.
We used daily temperature data (mean, maximum, minimum) and derived so-called Tropical Nights (minimum temperature > 20 °C), Summer Days (maximum temperature > 25 °C), Hot Days (maximum temperature > 30 °C), and Heat Episodes (minimum temperature > 18 °C, maximum temperature > 30 °C, for at least 3 days) from the SPARTACUS dataset [70] on a 1 km grid covering the whole area of Austria. Within DISCC-AT, Geosphere Austria granted access to a new version (v3.0), which additionally includes dew-point temperature; this allowed us to additionally derive Heat Index, Humidex, and Wet Bulb Global Temperature for both the daily mean and daily maximum because dew-point temperatures remain relatively constant throughout the day [71].
To determine the best measure of temperature, a stepwise approach was used for the cities of Vienna, Graz, Linz, and Klagenfurt and the district of Wolfsberg. First, the different temperature measures were each averaged across all grid points within each district and compared in a simple model that just added same-day temperature measure, the average of the past 14 days’ temperature, and the square of the same-day temperature to the base model. That model has already been validated in a previous paper about temperature effects in the city of Vienna [72] and was therefore used as a starting point to select the best temperature metric according to the Akaike Information Criterion. Second, the optimal approach to represent district-wide temperature based on 1 km grid data was chosen: unweighted or population weighted average or the maximal value within each district. The latter method was applied because, at least in mountainous districts, the majority of the population is located in valleys where the temperature is usually highest. Also, it was reasoned that even people living outside the main residential area of the district (e.g., higher up the mountains in the mountain districts) would likely spend time in the district center for work, shopping, etc. Still, it was expected a priori that the population weighted average would be the best descriptor of the district-wide exposure and thus should provide the best fit.

2.2. Study Area

Austria, with a population of about 8 million, consists of 116 districts as the smallest political units. In some instances, district boundaries had changed over the course of the study period, while meteorological data were provided for the current districts. In Lower Austria, the district “Wien Umgebung” was suspended in 2016, and parts of the district were allocated to the neighboring districts Tulln, Korneuburg, Bruck an der Leitha, and St. Pölten Land. Therefore, for these four districts, a continuous time series could not be established. In Burgenland, the small town of Rust is, for historical reasons, a district of its own. As it lies embedded in the larger district Eisenstadt Umgebung, the daily deaths in Rust were added to those in the larger district, and the meteorological data for Eisenstadt Umgebung were used in the models for the combined districts. In Styria, around 2012, four new districts were created out of two old districts each. Thus, the new districts could be treated as if they existed since 1970 by simply adding the daily deaths of the 2 old districts each. In total, this approach allowed for the analysis of 111 districts. Some of the district characteristics were taken from a previous study [73] that excluded the 23 districts of Vienna. Thus, depending on the determinants analyzed, a different number of districts were available.

2.3. Statistical Analysis and Choice of Models

In building the base model, we controlled for long-term trends in the average number of deaths (due to changes in population number and structure), seasonal variation, and day of the week. Long-term and seasonal variations can be controlled for through a natural spline. In accordance with the APHEA protocol, we optimized the number of knots for that natural spline by minimizing the sum of the partial autocorrelation [74] and included day of the week as a dummy variable. We selected the optimal number of knots based on daily deaths in all of Austria. In this dataset with large daily numbers, distribution of daily deaths approximated a Poisson distribution well enough to allow for a Poisson regression. The number of knots (320 in total or approximately 6 knots per year) was then also tested on four cities of different sizes (Vienna, Graz, Linz, and Klagenfurt) and on a smaller rural district (Wolfsberg in Carinthia). While the number of knots also worked well with most of these districts, the assumption of a Poisson distribution no longer held true for all the districts. Therefore, to ensure the same model was used for every district, for the per-district analysis, a negative binomial regression analysis (General Additive Model (GAM) with family negative binominal) was performed. In the first step, this base model was calculated for every district (using R version 4.0.3):
GAM (Deaths ~ s (date, k = 320) + as.factor (Day_of_Week), data = 1970_2020, family = nb),
with s standing for the (cubic) spline, k for the number of knots, and as.factor describing a nominal variable. The residuals from each district were stored for further use.
In a study on temperature effects in Vienna [72], two parametric models represented a fair approximation of the more complex temperature-effect association as modeled with a distributed lag non-linear model (DLNM) [75,76]. Again, for the current study, these two models were compared to the more detailed DLNM for the above-mentioned districts. Visual inspection of the temperature-effect curves and model fit examined by the Akaike Information Criterion (AIC) were used to assess the validity of the simplified models in comparison. This comparison is described in more detail in Appendix A. This simplification was deemed necessary to arrive at unique (linear) effect values per district to be entered into the meta-regression of the second step. Both models controlled for the (almost) linear adverse effect of cold temperature over the previous 14 days (T14). The same-day effects of temperature were not linear and were therefore modeled either by a second-degree polynomial (Temp and Temp_squared) or by a threshold model. For the former model, the base model (Formula (1)) was extended by adding T14, Temp, and Temp_squared. The coefficient of Temp_squared signifies the increase in mortality risk at high temperatures, while the coefficient of Temp defines the temperature with the minimal daily deaths (“optimal temperature”): the first derivative of the formula y = a x2 + bx + c gives 2 ax + b. When we set dy/dx to zero, the optimal temperature is −b/2a. Since the coefficient of Temp_squared (a) is expected to be positive, a negative coefficient of Temp (b) will shift the optimal temperature above zero.
The threshold model was built using STATA (Vers. 17), where a non-linear model (nl) was applied to the residuals from the base model:
nl (residual = {coeff1} × T14 + (Temp > {threshold = 10}) × (Temp−{threshold}) × {coeff2} + {const}),
where coeff1 is the coefficient of the linear effect of the 14-day average temperature and coeff2 is the coefficient of the linear effect of same-day temperature above a certain threshold (threshold).

2.4. Characteristics of Districts as Possible Determinants of Vulnerability

District-wide characteristics were primarily calculated from data from Statistik Austria (https://www.statistik.at/, accessed 11 January 2026): population density (both based on area of settlement and permanent area of settlement), percentage of persons aged ≥75 years, average age of inhabitants (estimated based on broader age groups), percentage of persons with Austrian citizenship, and percentage of persons born in Austria were based on data from 2022. If percentage of persons with Austrian citizenship turned out to be predictive, percentage of persons with citizenship in the new EU countries and percentage of persons with non-EU European citizenship (including Turkey) were also assessed. Percentage of persons with only primary education and of persons with a university degree, as well as percentage of homeless people and percentage of people with certain occupations (as a percentage of all working persons) were based on 2020 data according to the Statistical Classification of Economic Activities in the European Community, namely, agriculture and forestry, mining and quarrying, manufacturing, electricity, gas, steam and air-conditioning supply, water supply, sewerage, waste management and remediation, and construction,. Also, 2020 data were used for the percentage of homeless people and of people living in single households, as well as working people and unemployed people as a percentage of all persons seeking to work. Average household income data were from 2019.
“Urbanity” was defined as follows: most districts were considered rural (0). With districts where the rural part is an independent district and the large central town is a separate district, the central town district was considered “urban” (1). Some of these central towns are also capitals of federal states (Klagenfurt, St. Pölten, Linz, Salzburg, Graz and Innsbruck) and are larger (more and more densely populated) than the normal “urban” districts (2). The city of Vienna is also a separate federal state and consists of 23 districts, which were considered even more urbanized (3).
Altitude above sea level was taken from the altitude of the main town of each district. The percentage of (daily) smokers was derived from the last large Austrian Health Survey from 1986 [77]. This survey included 71,961 respondents, which was considerably more than that included in the later Austrian parts of the European Health Interview Surveys (ATHIS). For example, ATHIS 2019 had 15,461 respondents only [78]. Nonetheless, even the 1986 survey lacked the power to estimate smoking prevalence per district sufficiently; therefore, only smoking prevalence per Eurostat NUTS-3 region was reported. Austria consists of 35 NUTS-3 regions; therefore, each region consists of several districts. It was assumed that every district within the same NUTS-3 region had the same smoking prevalence. The percentage of green area per district was calculated based on regional information of the Federal Office of Metrology and Surveying [79].
We used the percentage of commuters (i.e., people working outside of the district they live in) and the number of tourist nights divided by population from our COVID-19 study [71]. This study did not include the individual districts of the city of Vienna.

2.5. Meta-Regression

Because of the small number of data points (111 districts or 88 districts excluding the districts of Vienna), only univariate linear regressions were run between single characteristics of the districts and the coefficients of the temperature models. Only on rare occasions when an association was very likely confounded by another variable was a model with two independent variables run. Altitude and urbanity (because many factors most likely differ between urban and rural areas) were chosen as possible confounders.
Only results with a p-value < 0.1 were reported. Even more simplified, the graphical abstract only reports effects that could be demonstrated through both models.

3. Results

Among all temperature parameters, the average temperature worked best in all tested districts according to the Akaike criterion. As expected, using average temperature at each grid point, calculating the population weighted average of all grid points performed best in nearly all tested districts. However, for heat episode days, which also include information on the duration of a heatwave (3 days at least), all the heat indices did not improve the models, and the gain from adding heat episode days was negligible. Also, using combined heat and humidity metrics did not yield better results than using the bare temperature metric. Therefore, the average daily temperature population weighted across all grid points of each district was chosen as the temperature metric of choice.
As the non-linear effect of temperature can be demonstrated using natural splines, this is exemplified for Vienna in Figure 1. It demonstrates that the chronic effect can be represented by a linear association, while the acute effect can be approximated by either a linear threshold model or a quadratic function (temperature and temperature squared). In the example districts, this simple model was not inferior to the third-degree polynomial distributed lag model according to the Akaike Information Criterion. The results of the third-degree polynomial distributed lag model are presented in Figure 2 for the city of Vienna again.
With the simple model, the coefficients for T14, Temp, and Temp_squared were calculated for every district. As expected, the 14-day average always had a negative and usually highly significant coefficient. The squared temperature was positive in nearly all districts and was usually significantly different from zero. The same-day temperature had values below or above zero and reached significance only in some districts, indicating that the minimum of the quadratic curve for the same-day temperature (the optimal temperature after controlling for 14-day average temperature and seasonal variation) was around zero degree Celsius. As the coefficient for T14 is the best indicator of cold vulnerability and the coefficient for Temp_squared is the best indicator of heat vulnerability, the geographic distributions of these two coefficients are depicted in Figure 3. For the sake of Figure 3, the coefficients for the four districts of Lower Austria that were excluded from the meta-regression are also shown. These coefficients were calculated by adding a dummy variable (0/1) to signify before/after the border change. Table 1 also presents the (population weighted) averages of the coefficients of all districts.
Next, the threshold model run on the residuals of the base model was also applied to each district, delivering a coefficient for the effect of the 14-day running average, a coefficient for the effect of the same-day temperature above a threshold, and a coefficient for the threshold itself, again providing an always significant negative coefficient for T14, an always positive and almost always significant coefficient for Temp, and an always positive and almost always significant threshold for the latter. Only for seven and eight (usually smaller) districts with a broader confidence interval did the coefficient and the threshold, respectively, not reach significance (p < 0.05).
The effects of district characteristics on the coefficients calculated by linear meta-regression are displayed in Table 1. For the sake of clarity, only those effects that at least showed a trend (p < 0.1) were included. These are the results of univariate linear meta-regressions weighted by the 2022 population number per district plus some selected results from models controlling for either urbanity or altitude. The coefficients of temperature in the quadratic regression and the thresholds in the threshold meta-regression model should have (and indeed usually have) opposite signs. Interpretation of these factors is not so straightforward because a combination of threshold and slope thereafter should be assessed.
Since the average temperature over the last 14 days (T14) displayed a negative slope in all districts (higher mortality rates after a colder fortnight), a negative coefficient in the meta-regression indicates an even steeper slope and thus higher vulnerability to cold. This was also observed for an increasing autochthonous population (either per birthplace or per citizenship) and for districts with a higher employment rate, more commuters, more agricultural workers, and more green spaces.
As same-day T-squared (T2) in the GAM and same-day temperature (T) above the threshold in the threshold model had a positive slope in nearly all districts, a positive coefficient in the meta-regression indicated a steeper slope and thus greater vulnerability to heat. This was observed for more densely populated areas and especially for the city of Vienna, for districts with a higher percentage of single, homeless, and unemployed people and of people either not born in Austria or without Austrian citizenship. Surprisingly, a higher percentage of outdoor workers seemed protective, and “richer” districts (average net income) and more educated districts (higher percentage of people with a university degree) displayed higher vulnerability. The latter lost significance when controlling for urbanity. With same-day temperature (T) in the (quadratic) GAM, a higher coefficient indicates a lower optimal temperature, which translates into a lower threshold in the threshold model. Altitude, which is usually associated with lower average temperatures, leads to lower thresholds or optimal temperatures.

4. Discussion

This is the first study of the effects of district-wide socioeconomic, demographic, and socioecological characteristics on heat and cold vulnerability in Austria. We have applied several different metrices of temperature, i.e., mean, minimum, and maximum temperature, plus a variety of combined indicators of temperature and humidity. We have also examined three different ways of estimating district-wide temperature exposure using temperature data on a 1 km grid. We have also tested a DLNM and two simplified models. All approaches demonstrated the expected effects of extreme temperatures.
Extreme temperatures cause physiological stress as the human body tries to stabilize its core temperature. While effects of high temperatures have been the subject of more recent studies [53,80] due to the imminent threat of climate change, effects of cold temperatures also are well established [81]. Individual vulnerability factors clearly shape the risk of mortality during extreme temperatures. Personal vulnerability factors include frailty due to chronic disease, old and very young age, poor housing conditions, and other socioeconomic factors, as well as lack of social support as in the case of members of minority groups. There can be a combination of vulnerability factors, like living in an uninsulated home and being exposed to extreme temperatures at the workplace [5].

4.1. Spatial Differences Between Districts

For the meta-regression approach, we compared effect estimates of two different models. Only findings that were significant in both models were considered relevant. It is plausible that population density [82,83], percentage of immigrants [5,8,10], percentage of homeless people [50,59], unemployed [8,22], singles [56,58], and urban characteristics [10,47,84] influence the association between heat and mortality in the direction of higher vulnerability. Several studies have demonstrated that people in densely populated urban areas are more vulnerable to heatwaves [85,86]. During such periods, temperatures in densely built areas are even higher [87]. This is caused by the so-called heat island effect that is due to human-made surfaces (concrete, asphalt) absorbing and re-emitting solar heat, reduced vegetation for cooling (transpiration/shade), and waste heat from buildings and vehicles. Our study demonstrates that even when comparing the same temperatures, urban populations (especially in Vienna) are more vulnerable to heat than rural ones.
The role of urbanization and rurality in hot and cold vulnerability may be rather complex [42,43], and the impact of heat in rural regions especially is often underestimated. For example, older adults who are still working in agriculture or forestry may be exposed to high temperatures without sufficient protection or hydration [40].
It is difficult to explain our finding that longer education can lead to greater vulnerability. However, in an Italian study conducted in Turin [88], men with higher education levels were more vulnerable to heat. In Barcelona (Spain), significant heat effects were found for women of all educational levels, not only for the low-education groups [89]. With regard to 66 communities in China, heatwave effects on mortality were stronger in populations with higher education [85]. Nevertheless, the greater vulnerability of populations with a higher percentage of university graduates in our study was demonstrated only by the quadratic model and not by the threshold model. Even in the quadratic model, any effect was lost after controlling for urbanity.
Altitude is likely a proxy for the average temperature and thus for temperature adaptation, leading to a lower threshold level or a lower optimal temperature as previously shown by us for the temporal trends in Vienna [72]. Because mountainous districts usually have smaller populations and only rarely experience very high temperatures, altitude impacts on effect estimates for high temperatures (T above the threshold or T2 in the quadratic model) are less dependable.
Certain occupations that are linked to higher heat stress seem to have a protective effect. The latter might be explained by imprecise classifications and/or certain characteristics of the districts with higher percentages of these occupations. For example, even in districts with a relatively high percentage of agricultural workers, the percentage is still low in absolute terms. All the inhabitants not engaged in agriculture might benefit from the greenness of the agricultural district and lower night-time temperatures. However, even if the percentage of green area is the same, a higher percentage of agricultural workers would indicate a tendency toward more labor-intensive, likely more small-scale farming. Such farming practices might improve the recreational value of the landscape because of increased biodiversity [90].
It is not exactly clear if and how smoking prevalence is linked to temperature-related vulnerability. Greater vulnerability may be explained by the lower socioeconomic status of smokers and the negative influence of smoking on the cardiovascular system and the respiratory tract. Persons with cardiovascular and respiratory diseases are at higher risk during heatwaves and cold spells [11,91,92,93]. All these issues would require a more complex multivariate analysis, and even then, it is unclear if all these issues could be solved. It should be noted that a seemingly protective effect of smoker prevalence against cold was only seen in the quadratic model but not in the threshold model.
A higher cold vulnerability was found for an increasingly autochthonous population for districts with a higher employment rate, more commuters, more agricultural workers, and more green spaces. This might be explained in part by characteristics of these districts, which lead to higher cold exposure than expected from temperature measurements.
Certainly, not all characteristics of districts that displayed a significant effect on any of the coefficients of the temperature–mortality association had a causal impact. For example, tourism lost its significant protection against heat when combined with altitude. A more sophisticated meta-regression controlling for possible confounders was outside the scope and power of this study. Effect estimates per district came with substantial uncertainty and, hence, a linear regression with only 111 data points at maximum lacked the necessary power. It was surprising that despite the poor power, many of the coefficients were highly significant. Even when significant, some characteristics or factors served only as proxies for more complex (socioeconomic, etc.) conditions.

4.2. Limitations

This analysis aimed for rather simple models returning a small set of coefficients that can be interpreted with ease. Therefore, distributed lag effects were replaced by an acute effect of same-day temperature and a subacute effect of the running 14-day average. Also, non-linear associations were approximated either by a quadratic polynomial or by a linear threshold model. These approximations worked fairly well but were still approximations only. In addition, only the daily temperature was included in the model. More complex indicators of perceived temperature that also included humidity did not provide a better model fit. Nevertheless, other meteorological variables would likely also have some impact on health (and wellbeing).
Extreme temperatures act as stressors to the human body, thereby reducing physical and mental performance and increasing morbidity and mortality risk. Also, fast changes in temperature have such an effect [94,95,96]. A more sophisticated model would therefore also include measures of temperature change, either between consecutive days or within a day, e.g., the difference between the maximum and minimum temperatures.
We have already demonstrated in a previous study, using data from the city of Vienna alone [72], that there was a shift in heat vulnerability over time. We interpreted this as an indication of an adaptation to rising temperatures. Adaptation, on the one hand, occurred due to public measures (information campaigns and heat alerts, better preparedness of hospitals and care facilities, opening of cooling rooms for, inter alia, homeless people) [97], mostly introduced in Austria after the 2003 heatwave [98] and mostly in urban areas; on the other hand, this was also due to measures by individuals like installing cooling devices in their homes. But we also interpreted our previous observations from Vienna as a sign of an ongoing selection process: vulnerable persons are more likely to die in a heatwave, rendering the remining population less vulnerable to future heatwaves. We could only analyze shorter time intervals when looking at a larger population, in this instance the whole of Vienna, not single districts. Especially for rural districts (with fewer inhabitants and thus fewer daily deaths), effect estimates are not precise enough when only shorter time intervals are examined [40]. Since we were interested in spatial differences, we opted to neglect temporal change. Also, some of the district indicators were not available with the same definition at different points in time. All this considered, a temporal analysis of changes in vulnerability was out of the scope of our paper but would be an important issue for future research.
The actual effect estimates for temperatures might not be accurate; nevertheless, this would not invalidate the relative differences in the effect estimates between districts, which were the main objective of this study. For this reason, we ran the same models for every district, even though for some districts, depending mainly on population size, different and possibly even simpler models would have worked quite as good or, in some instances, even slightly better.
The use of smoking prevalence data from 1986 is another constraint. Even with these old data, precision was only deemed sufficient on the NUTS-3 level. Over the years, the percentage of smokers in Austria overall has declined, but, according to later Health Interview Surveys that reported smoking prevalence on the level of federal countries only, the rank order with higher smoking prevalence in the urban area of Vienna, on the one hand, and the western provinces, on the other hand, remained stable. Nevertheless, effects of smoking prevalence should be interpreted with care.
This study analyzed spatial differences only. To allow for sufficiently precise estimates regarding smaller districts, a long time interval had to be considered. That rendered the analysis of temporal variation impossible. While we considered a change in vulnerability, especially a reduced vulnerability towards heat with increasing temperatures likely [72], this issue was outside of the scope of this study.

5. Conclusions

In conclusion, this district-wise analysis and meta-regression provides insights into vulnerability factors at the district level that affect the impact of extreme temperature on mortality risk in Austria. Our study could shed light on previously unrecognized risk groups. There is ongoing debate whether in temperate climate the adverse health effects of hotter days are already outweighing the beneficial effects of less cold days. This study might aid in health impact, with greater spatial resolution than previous ones [76].
Even when comparing the same temperatures, urban populations are more vulnerable to heat. Districts with, inter alia, more agricultural workers and more green spaces were more vulnerable towards cold. Measures to reduce population vulnerability should target multiple aspects including a reduction in emissions both of greenhouse gases and air pollutants [44,99,100,101], mitigating socioeconomic inequalities. We recommend targeted awareness rising campaigns and protective programs for, among others, the elderly and chronically ill, migrants and rural populations, better medical infrastructure, (organized) community networks, and more cooling rooms. Our analysis does provide a first indication for Austrian policymakers of which districts should be prioritized when considering those measures.

Author Contributions

Conceptualization, H.M. and P.W.; methodology, H.M.; formal analysis, H.M.; resources, M.J.; writing—original draft preparation, H.M. and P.W.; writing—review and editing, H.-P.H. All authors have read and agreed to the published version of the manuscript.

Funding

This paper was prepared as part of the Austrian ACRP 14/2022 project DISCC-AT (project number: KR21KB0K00001) funded by the Austrian “Klima- und Energie-Fonds” (Climate and Energy Fund).

Institutional Review Board Statement

Because this study only used anonymous and aggregated human register data, ethics committee approval was deemed not necessary.

Data Availability Statement

The raw daily death and daily temperature data were obtained from Statistics Austria and Geosphere Austria, respectively, under the condition that these data are not distributed to third parties. However, these data are likely available from the same sources for a reasonable fee. The characteristics of the districts are freely available on the web. Upon request, an Excel file summarizing the data used, as well as the coefficients and thresholds calculated for each district, can be provided by the corresponding author.

Acknowledgments

We want to express our great thanks to Kristina Standeven for critical reading and language review of the manuscript.

Conflicts of Interest

The authors declare no conflicts of interest.

Abbreviations

The following abbreviations are used in this manuscript:
ACRPAustrian climate research program
AICAkaike Information Criterion
DLNMDistributed lag non-linear model
GAMGeneral Additive Model
NUTS-3Nomenclature of territorial units for statistics, small region

Appendix A

The appendix describes the comparisons of the distributed lag non-linear models (DLNMs) with the simplified model only applying a linear term for the 14-day average temperature and a linear and quadratic term for same-day temperature. At first, DLNMs were used to select the best temperature metric. For all tested regions, regarding the temporal course of the temperature, the daily mean temperature performed slightly better than the minimal and maximal daily temperature. Regarding the spatial distribution of temperatures (over a 1 × 1 km grid), the population weighted average over the grid points performed better than the unweighted average or the maximal temperature. The larger and more populous the region, the higher the explained deviance. So, for the city of Vienna (about 2 million inhabitants, 23 districts), the different temperature metrices explained between 71.9 and 72.4% of the deviance. The better fit was observed in models also including various indices in addition to the temperature. But this small gain was usually offset by the loss of degrees of freedom. So, the AIC was usually slightly larger for the models including an additional index variable. Only regarding the Kyseli index, that is defined over a series of days rather than for single days, the AIC (129,151) was slightly smaller than for the mean temperature alone (129,270.5). But the difference was small and statistically not significant.
Replacing the DLNM by the simplified model reduced the explained deviance to 70%. But the gain in degrees of freedom translated even into a further reduced AIC of 129,231.6 (again, not significantly different).
For single districts with substantially smaller population and, hence, also fewer cases per day, the explained deviance was much smaller, e.g., about 20–25% for Graz, 5–7% for Linz, or about 2% for Klagenfurt. In these smaller regions, even according to the AIC, the DLNMs did perform better (Graz: 89,593.21 for the DLNM and 89,734.15 for the simplified model, Linz: 85,262.32 versus 85,335.18, respectively). Again, this difference was not deemed sufficient to stick to the DLNMs, but the simplified models were used because they allowed for easier interpretation of the effect estimates to be used in the meta-regression.

References

  1. Biswas, S.; Nautiyal, S. A review of socio-economic vulnerability: The emergence of its theoretical concepts, models and methodologies. Nat. Hazards Res. 2023, 3, 563–571. [Google Scholar] [CrossRef]
  2. Mah, J.C.; Penwarden, J.L.; Pott, H.; Theou, O.; Andrew, M.K. Social vulnerability indices: A scoping review. BMC Public Health 2023, 23, 1253. [Google Scholar] [CrossRef]
  3. IPCC. Climate Change. Impacts, Adaptation, and Vulnerability; Contribution of Working Group II to the Sixth Assessment Report of the Intergovernmental Panel on Climate Change; Pörtner, H.O., Roberts, D.C., Tignor, M., Poloczanska, E.S., Mintenbeck, K., Alegría, A., Craig, M., Langsdorf, S., Löschke, S., Möller, V., et al., Eds.; Cambridge University Press: Cambridge, UK; New York, NY, USA, 2022. [Google Scholar]
  4. Harlan, S.L.; Declet-Barreto, J.H.; Stefanov, W.L.; Petitti, D.B. Neighborhood effects on heat deaths: Social and environmental predictors of vulnerability in Maricopa County, Arizona. Environ. Health Perspect. 2013, 121, 197–204. [Google Scholar] [CrossRef]
  5. Gronlund, C.J. Racial and socioeconomic disparities in heat-related health effects and their mechanisms: A review. Curr. Epidemiol. Rep. 2014, 1, 165–173. [Google Scholar] [CrossRef]
  6. de Schrijver, E.; Royé, D.; Gasparrini, A.; Franco, O.H.; Vicedo-Cabrera, A.M. Exploring vulnerability to heat and cold across urban and rural populations in Switzerland. Environ. Res. Health 2023, 1, 025003. [Google Scholar] [CrossRef]
  7. Benmarhnia, T.; Deguen, S.; Kaufman, J.S.; Smargiassi, A. Vulnerability to heat-related mortality: A systematic review, meta-analysis, and meta-regression analysis. Epidemiology 2015, 26, 781–793. [Google Scholar] [CrossRef]
  8. Ellena, M.; Breil, M.; Soriani, S. The heat-health nexus in the urban context: A systematic literature review exploring the socio-economic vulnerabilities and built environment characteristics. Urban Clim. 2020, 34, 100676. [Google Scholar] [CrossRef]
  9. Hansen, J. Experts discuss how extreme heat affects vulnerable populations—And how communities and health care systems can respond. Stanford Report, 14 August 2023. Available online: https://news.stanford.edu/stories/2023/08/heat-affects-vulnerable (accessed on 11 January 2026).
  10. van Daalen, K.R.; Tonne, C.; Semenza, J.C.; Rocklöv, J.; Markandya, A.; Dasandi, N.; Jankin, S.; Achebak, H.; Ballester, J.; Bechara, H.; et al. The 2024 Europe report of the Lancet Countdown on health and climate change: Unprecedented warming demands unprecedented action. Lancet Public Health 2024, 9, e495–e522. [Google Scholar] [CrossRef]
  11. Ryti, N.R.; Guo, Y.; Jaakkola, J.J. Global association of cold spells and adverse health effects: A systematic review and meta-analysis. Environ. Health Perspect. 2016, 124, 12–22. [Google Scholar] [CrossRef]
  12. Wan, K.; Feng, Z.; Hajat, S.; Doherty, R.M. Temperature-related mortality and associated vulnerabilities: Evidence from Scotland using extended time-series datasets. Environ. Health 2022, 21, 99. [Google Scholar] [CrossRef]
  13. Yang, J.; Zhou, M.; Guo, C.; Zhu, S.; Sakhvidi, M.J.Z.; Requia, W.J.; Sun, Q.; Tong, S.; Li, M.; Liu, Q. Drivers of associations between daytime-nighttime compound temperature extremes and mortality in China. Commun. Med. 2024, 4, 125. [Google Scholar] [CrossRef] [PubMed]
  14. Gasparrini, A.; Masselot, P.; Scortichini, M.; Schneider, R.; Mistry, M.N.; Sera, F.; Macintyre, H.L.; Phalkey, R.; Vicedo-Cabrera, A.M. Small-area assessment of temperature-related mortality risks in England and Wales: A case time series analysis. Lancet Planet. Health 2022, 6, e557–e564. [Google Scholar] [CrossRef]
  15. Park, J.; Bangalore, M.; Hallegatte, S.; Sandhoefner, E. Households and heat stress: Estimating the distributional consequences of climate change. Environ. Dev. Econ. 2018, 23, 349–368. [Google Scholar] [CrossRef]
  16. Government UK; Department for Energy Security and Net Zero. Household Energy Efficiency Statistical Release 30 March 2023; Department for Energy Security and Net Zero: London, UK, 2023. Available online: https://assets.publishing.service.gov.uk/media/64230bbd3d885d000cdadd20/HEE_Stats_Detailed_Release_-_Mar_23.pdf (accessed on 11 January 2026).
  17. Hajat, S.; Kosatky, T. Heat-related mortality: A review and exploration of heterogeneity. J. Epidemiol. Community Health 2010, 64, 753–760. [Google Scholar] [CrossRef] [PubMed]
  18. Reid, C.E.; O’Neill, M.S.; Gronlund, C.J.; Brines, S.J.; Brown, D.G.; Diez-Roux, A.V.; Schwartz, J. Mapping community determinants of heat vulnerability. Environ. Health Perspect. 2009, 117, 1730–1736. [Google Scholar] [CrossRef]
  19. Kim, E.J.; Kim, H. Effect modification of individual- and regional-scale characteristics on heat wave-related mortality rates between 2009 and 2012 in Seoul, South Korea. Sci. Total Environ. 2017, 595, 141–148. [Google Scholar] [CrossRef]
  20. Nanda, L.; Chakraborty, S.; Mishra, S.K.; Dutta, A.; Rathi, S.K. Characteristics of households’ vulnerability to extreme heat: An analytical cross-sectional Study from India. Int. J. Environ. Res. Public Health 2022, 19, 15334. [Google Scholar] [CrossRef]
  21. Yang, J.; Yin, P.; Sun, J.; Wang, B.; Zhou, M.; Li, M.; Tong, S.; Meng, B.; Guo, Y.; Liu, Q. Heatwave and mortality in 31 major Chinese cities: Definition, vulnerability and implications. Sci. Total Environ. 2019, 649, 695–702. [Google Scholar] [CrossRef] [PubMed]
  22. Ho, H.C.; Knudby, A.; Walker, B.B.; Henderson, S.B. Delineation of spatial variability in the temperature-mortality relationship on extremely hot days in Greater Vancouver, Canada. Environ. Health Perspect. 2017, 125, 66–75. [Google Scholar] [PubMed]
  23. Sheffield, P.E.; Herrera, M.T.; Kinnee, E.J.; Clougherty, J.E. Not so little differences: Variation in hot weather risk to young children in New York City. Public Health 2018, 161, 119–126. [Google Scholar] [CrossRef]
  24. Arsad, F.S.; Hod, R.; Ahmad, N.; Ismail, R.; Mohamed, N.; Baharom, M.; Osman, Y.; Radi, M.F.M.; Tangang, F. The impact of heatwaves on mortality and morbidity and the associated vulnerability factors: A systematic review. Int. J. Environ. Res. Public Health 2022, 19, 16356. [Google Scholar] [CrossRef]
  25. Leone, M.; D’Ippoliti, D.; De Sario, M.; Analitis, A.; Menne, B.; Katsouyanni, K.; de’ Donato, F.K.; Basagana, X.; Salah, A.B.; Casimiro, E.; et al. A time series study on the effects of heat on mortality and evaluation of heterogeneity into European and Eastern-Southern Mediterranean cities: Results of EU CIRCE project. Environ. Health 2013, 12, 55. [Google Scholar] [CrossRef]
  26. Hondula, D.M.; Davis, R.E.; Saha, M.V.; Wegner, C.R.; Veazey, L.M. Geographic dimensions of heat-related mortality in seven U.S. cities. Environ. Res. 2015, 138, 439–452. [Google Scholar] [CrossRef] [PubMed]
  27. Joe, L.; Hoshiko, S.; Dobraca, D.; Jackson, R.; Smorodinsky, S.; Smith, D.; Harnly, M. Mortality during a large-scale heat wave by place, demographic group, internal and external causes of death, and building climate zone. Int. J. Environ. Res. Public Health 2016, 13, 299. [Google Scholar] [CrossRef]
  28. Varghese, B.M.; Barnett, A.G.; Hansen, A.L.; Bi, P.; Nairn, J.; Rowett, S.; Nitschke, M.; Hanson-Easey, S.; Heyworth, J.S.; Sim, M.R.; et al. Characterising the impact of heatwaves on work-related injuries and illnesses in three Australian cities using a standard heatwave definition—Excess Heat Factor (EHF). J. Expo. Sci. Environ. Epidemiol. 2019, 29, 821–830. [Google Scholar] [CrossRef]
  29. Sandholz, S.; Sett, D.; Greco, A.; Wannewitz, M.; Garschagen, M. Rethinking urban heat stress: Assessing risk and adaptation options across socioeconomic groups in Bonn, Germany. Urban Clim. 2021, 37, 100857. [Google Scholar] [CrossRef]
  30. Astone, R.; Vaalavuo, M. Climate change and health: Consequences of high temperatures among vulnerable groups in Finland. Int. J. Soc. Determ. Health Health Serv. 2022, 53, 94–111. [Google Scholar] [CrossRef]
  31. Adam-Poupart, A.; Smargiassi, A.; Busque, M.A.; Duguay, P.; Fournier, M.; Zayed, J.; Labrèche, F. Summer outdoor temperature and occupational heat-related illnesses in Quebec (Canada). Environ. Res. 2014, 134, 339–344. [Google Scholar] [CrossRef]
  32. McInnes, J.A.; Akram, M.; Macfarlane, E.M.; Keegel, T.; Sim, M.R.; Smith, P. Association between high ambient temperature and acute work-related injury: A case-crossover analysis using workers’ compensation claims data. Scand. J. Work Environ. Health 2017, 43, 86–94. [Google Scholar] [CrossRef]
  33. Kysely, J.; Pokorna, L.; Kyncl, J.; Kriz, B. Excess cardiovascular mortality associated with cold spells in the Czech Republic. BMC Public Health 2009, 9, 19. [Google Scholar] [CrossRef] [PubMed]
  34. Karthick, S.; Kermanshachi, S.; Pamidimukkala, A.; Namian, M. A review of construction workforce health challenges and strategies in extreme weather conditions. Int. J. Occup. Saf. Ergon. 2023, 29, 773–784. [Google Scholar] [CrossRef]
  35. Petitti, D.B.; Harlan, S.L.; Chowell-Puente, G.; Ruddell, D. Occupation and environmental heat-associated deaths in Maricopa county, Arizona: A case–control study. PLoS ONE 2013, 8, e62596. [Google Scholar] [CrossRef]
  36. Xiang, J.; Bi, P.; Pisaniello, D.; Hansen, A. The impact of heatwaves on workers’ health and safety in Adelaide, South Australia. Environ. Res. 2014, 133, 90–95. [Google Scholar] [CrossRef]
  37. Kovach, M.M.; Konrad, C.E.; Fuhrmann, C.M. Area-level risk factors for heat-related illness in rural and urban locations across North Carolina, USA. Appl. Geogr. 2015, 60, 175–183. [Google Scholar] [CrossRef]
  38. Odame, E.A.; Li, Y.; Zheng, S.; Vaidyanathan, A.; Silver, K. Assessing heat-related mortality risks among rural populations: A systematic review and meta-analysis of epidemiological evidence. Int. J. Environ. Res. Public Health 2018, 15, 1597. [Google Scholar] [CrossRef] [PubMed]
  39. Wu, P.C.; Lin, C.Y.; Lung, S.C.; Guo, H.R.; Chou, C.H.; Su, H.J. Cardiovascular mortality during heat and cold events: Determinants of regional vulnerability in Taiwan. Occup. Environ. Med. 2011, 68, 525–530. [Google Scholar] [CrossRef]
  40. Fastl, C.; Arnberger, A.; Gallistl, V.; Stein, V.K.; Dorner, T.E. Heat vulnerability: Health impacts of heat on older people in urban and rural areas in Europe. Wien. Klin. Wochenschr 2024, 136, 507–514. [Google Scholar] [CrossRef]
  41. Xiao, J.; Spicer, T.; Jian, L.; Yun, G.Y.; Shao, C.; Nairn, J.; Fawcett, R.J.B.; Robertson, A.; Weeramanthri, T.S. Variation in population vulnerability to heat wave in Western Australia. Front. Public Health 2017, 5, 64. [Google Scholar] [CrossRef]
  42. Lee, W.; Choi, M.; Bell, M.L.; Kang, C.; Jang, J.; Song, I.; Kim, Y.-O.; Ebi, K.; Kim, H. Effects of urbanization on vulnerability to heat-related mortality in urban and rural areas in South Korea: A nationwide district-level time-series study. Int. J. Epidem. 2022, 51, 111–121. [Google Scholar] [CrossRef] [PubMed]
  43. Xing, Q.; Sun, Z.; Tao, Y.; Zhang, X.; Miao, S.; Zheng, C.; Tong, S. Impacts of urbanization on the temperature-cardiovascular mortality relationship in Beijing, China. Environ. Res. 2020, 191, 110234. [Google Scholar] [CrossRef] [PubMed]
  44. Zhang, G.; Sun, Z.; Han, L.; Iyakaremye, V.; Xu, Z.; Miao, S.; Tong, S. Avoidable heat-related mortality in China during the 21st century. npj Clim. Atmos. Sci. 2023, 6, 81. [Google Scholar] [CrossRef]
  45. Bao, J.; Li, X.; Yu, C. The construction and validation of the heat vulnerability index, a review. Int. J. Environ. Res. Public Health 2015, 12, 7220–7234. [Google Scholar] [CrossRef]
  46. Burkart, K.; Meier, F.; Schneider, A.; Breitner, S.; Canario, P.; Alcoforado, M.J.; Scherer, D.; Endlicher, W. Modification of heat-related mortality in an elderly urban population by vegetation (urban green) and proximity to water (urban blue): Evidence from Lisbon, Portugal. Environ. Health Perspect. 2016, 124, 927–934. [Google Scholar] [CrossRef]
  47. Madrigano, J.; Ito, K.; Johnson, S.; Kinney, P.L.; Matte, T. A case-only study of vulnerability to heat wave–related mortality in New York City (2000–2011). Environ. Health Perspect. 2015, 123, 672–678. [Google Scholar] [CrossRef]
  48. Tanner, L.M.; Moffatt, S.; Milne, E.M.; Mills, S.D.; White, M. Socioeconomic and behavioural risk factors for adverse winter health and social outcomes in economically developed countries: A systematic review of quantitative observational studies. J. Epidemiol. Community Health 2013, 67, 1061–1067. [Google Scholar] [CrossRef]
  49. Klein Rosenthal, J.; Kinney, P.L.; Metzger, K.B. Intra-urban vulnerability to heat-related mortality in New York City, 1997–2006. Health Place 2014, 30, 45–60. [Google Scholar] [CrossRef] [PubMed]
  50. Bezgrebelna, M.; McKenzie, K.; Wells, S.; Ravindran, A.; Kral, M.; Christensen, J.; Stergiopoulos, V.; Gaetz, S.; Kidd, S.A. Climate change, weather, housing precarity, and homelessness: A systematic review of reviews. Int. J. Environ. Res. Public Health 2021, 18, 5812. [Google Scholar] [CrossRef] [PubMed]
  51. López-Bueno, J.A.; Navas-Martín, M.A.; Díaz, J.; Mirón, I.J.; Luna, M.Y.; Sánchez-Martínez, G.; Culqui, D.; Linares, C. aAnalysis of vulnerability to heat in rural and urban areas in Spain: What factors explain Heat’s geographic behavior? Environ. Res. 2022, 207, 112213. [Google Scholar] [CrossRef]
  52. Vandentorren, S.; Bretin, P.; Zeghnoun, A.; Mandereau-Bruno, L.; Croisier, A.; Cochet, C.; Ribéron, J.; Siberan, I.; Declercq, B.; Ledrans, M. August 2003 heat wave in France: Risk factors for death of elderly people living at home. Eur. J. Public Health 2006, 16, 583–591. [Google Scholar] [CrossRef]
  53. Xing, Q.; Sun, Z.; Tao, Y.; Shang, J.; Miao, S.; Xiao, C.; Zheng, C. Projections of future temperature-related cardiovascular mortality under climate change, urbanization and population aging in Beijing, China. Environ. Int. 2022, 163, 107231. [Google Scholar] [CrossRef] [PubMed]
  54. Hansen, A.; Bi, L.; Saniotis, A.; Nitschke, M. Vulnerability to extreme heat and climate change: Is ethnicity a factor? Glob. Health Action. 2013, 6, 21364. [Google Scholar] [CrossRef]
  55. Nayak, S.G.; Shrestha, S.; Kinney, P.L.; Ross, Z.; Sheridan, S.C.; Pantea, C.I.; Hsu, W.H.; Muscatiello, N.; Hwang, S.A. Development of a heat vulnerability index for New York State. Public Health 2018, 161, 127–137. [Google Scholar] [CrossRef]
  56. Stafoggia, M.; Forastiere, F.; Agostini, D.; Caranci, N.; de’Donato, F.; Demaria, M.; Michelozzi, P.; Miglio, R.; Rognoni, M.; Russo, A.; et al. Factors affecting in-hospital heat-related mortality: A multicity case-crossover analysis. J. Epidemiol. Community Health 2008, 62, 209–215. [Google Scholar] [CrossRef] [PubMed]
  57. Wong, M.S.; Peng, F.; Zou, B.; Shi, W.Z.; Wilson, G.J. Spatially analyzing the inequity of the Hong Kong Urban Heat Island by socio-demographic characteristics. Int. J. Environ. Res. Public Health 2016, 13, 17. [Google Scholar] [CrossRef]
  58. Xia, Y.; Shi, C.; Li, Y.; Jiang, X.; Ruan, S.; Gao, X.; Chen, Y.; Huang, W.; Li, M.; Xue, R.; et al. Effects of ambient temperature on mortality among elderly residents of Chengdu city in Southwest China, 2016–2020: A distributed-lag non-linear time series analysis. BMC Public Health 2023, 23, 149. [Google Scholar] [CrossRef]
  59. Ramin, B.; Svoboda, T. Health of the homeless and climate change. J. Urban Health 2009, 86, 654–664. [Google Scholar] [CrossRef]
  60. Romaszko, J.; Cymes, I.; Dragańska, E.; Kuchta, R.; Glińska-Lewczuk, K. Mortality among the homeless: Causes and meteorological relationships. PLoS ONE 2017, 12, e0189938. [Google Scholar] [CrossRef]
  61. Schwarz, L.; Castillo, E.M.; Chan, T.C.; Brennan, J.J.; Sbiroli, E.S.; Carrasco-Escobar, G.; Nguyen, A.; Clemesha, A.; Gershunov, E. S.; Benmarhnia, T. Heat waves and emergency department visits among the homeless, San Diego, 2012–2019. Am. J. Public Health 2022, 112, 98–106. [Google Scholar] [CrossRef]
  62. van Steen, Y.; Ntarladima, A.M.; Grobbee, R.; Karssenberg, D.; Vaartjes, I. Sex differences in mortality after heat waves: Are elderly women at higher risk? Int. Arch. Occup. Environ. Health 2019, 92, 37–48. [Google Scholar] [CrossRef]
  63. Paterson, S.K.; Godsmark, C.N. Heat-health vulnerability in temperate climates: Lessons and response options from Ireland. Glob. Health 2020, 16, 29. [Google Scholar] [CrossRef]
  64. Liu, J.; Varghese, B.M.; Hansen, A.; Zhang, Y.; Driscoll, T.; Morgan, G.; Dear, K.; Gourley, M.; Capon, A.; Bi, P. Heat exposure and cardiovascular health outcomes: A systematic review and meta-analysis. Lancet Planet. Health 2022, 6, e484–e495. [Google Scholar] [CrossRef]
  65. Ngarambe, J.; Santamouris, M.; Yun, G.Y. The impact of urban warming on the mortality of vulnerable populations in Seoul. Sustainability 2022, 14, 13452. [Google Scholar] [CrossRef]
  66. Khatana, S.A.M.; Werner, R.M.; Groeneveld, P.W. Association of extreme heat and cardiovascular mortality in the United States: A county-level longitudinal analysis from 2008 to 2017. Circulation 2022, 146, 249–261. [Google Scholar] [CrossRef]
  67. Salvador, C.; Gullón, P.; Franco, M.; Vicedo-Cabrera, A.M. Heat-related first cardiovascular event incidence in the city of Madrid (Spain): Vulnerability assessment by demographic, socioeconomic, and health indicators. Environ. Res. 2023, 226, 115698. [Google Scholar] [CrossRef] [PubMed]
  68. Mashhoodi, B. Feminization of surface temperature: Environmental justice and gender inequality among socioeconomic groups. Urban Clim. 2021, 40, 101004. [Google Scholar] [CrossRef]
  69. Rocklöv, J.; Forsberg, B.; Ebi, K.; Bellander, T. Susceptibility to mortality related to temperature and heat and cold wave duration in the population of Stockholm County, Sweden. Glob. Health Action 2014, 7, 22737. [Google Scholar] [CrossRef]
  70. Hiebl, J.; Frei, C. Daily Temperature grids for Austria since 1961—Concept, creation and applicability. Theor. Appl. Climatol. 2016, 124, 161–178. [Google Scholar] [CrossRef]
  71. Casanueva, A.; Kotlarski, S.; Herrera, S.; Fischer, A.M.; Kjellstrom, T.; Schwierz, C. Climate projections of a multivariate heat stress index: The role of downscaling and bias correction. Geosci. Model. Dev. 2019, 12, 3419–3438. [Google Scholar] [CrossRef]
  72. Weitensfelder, L.; Moshammer, H. Evidence of adaptation to increasing temperatures. Int. J. Environ. Res. Public Health 2019, 17, 97. [Google Scholar] [CrossRef]
  73. Moshammer, H.; Poteser, M.; Weitensfelder, L. COVID-19: Regional differences in Austria. Int. J. Environ. Res. Public Health 2022, 19, 1644. [Google Scholar] [CrossRef]
  74. Samoli, E.; Schwartz, J.; Wojtyniak, B.; Touloumi, G.; Spix, C.; Balducci, F.; Medina, S.; Rossi, G.; Sunyer, J.; Bacharova, L.; et al. Investigating regional differences in short-term effects of air pollution on daily mortality in the APHEA project: A sensitivity analysis for controlling long-term trends and seasonality. Environ. Health Perspect. 2001, 109, 349–353. [Google Scholar] [CrossRef]
  75. Gasparrini, A.; Armstrong, B.; Kenward, M.G. Distributed lag non-linear models. Stat. Med. 2010, 29, 2224–2234. [Google Scholar] [CrossRef]
  76. Martínez-Solanas, È.; Quijal-Zamorano, M.; Achebak, H.; Petrova, D.; Robine, J.M.; Herrmann, F.R.; Rodó, X.; Ballester, J. Projections of temperature-attributable mortality in Europe: A time series analysis of 147 contiguous regions in 16 countries. Lancet Planet Health 2021, 5, e446–e454. [Google Scholar] [CrossRef]
  77. Österreichisches Statistisches Zentralamt. Mikrozensus Rauchgewohnheiten der Österreichischen Bevölkerung; Österreichisches Statistisches Zentralamt: Vienna, Austria, 1986. [Google Scholar]
  78. Statistik Austria. Österreichische Gesundheitsbefragung 2019. Hauptergebnisse des Austrian Health Interview Survey (ATHIS) und Methodische Dokumentation; Bundesministerium für Soziales, Gesundheit, Pflege und Konsumentenschutz: Vienna, Austria, 2020.
  79. Federal Office of Metrology and Surveying (Bundesamt für Eich- und Vermessungswesen). 2023. Available online: https://data.bev.gv.at/geonetwork/srv/ger/catalog.search#/metadata/2d298415-e6b5-4631-8354-df558b91a4c8 (accessed on 11 January 2026).
  80. Sun, Z.; Zhang, X.; Li, Z.; Liang, Y.; An, X.; Zhao, Y.; Miao, S.; Han, L.; Li, D. Heat exposure assessment based on high-resolution spatio-temporal data of population dynamics and temperature variations. J. Environ. Manag. 2024, 349, 119576. [Google Scholar] [CrossRef] [PubMed]
  81. Gasparrini, A.; Guo, Y.; Hashizume, M.; Lavigne, E.; Zanobetti, A.; Schwartz, J.; Tobias, A.; Tong, S.; Rocklöv, J.; Forsberg, B.; et al. Mortality risk attributable to high and low ambient temperature: A multicountry observational study. Lancet 2015, 386, 369–375. [Google Scholar] [CrossRef]
  82. Lee, W.; Ebi, K.L.; Kim, Y.; Hashizume, M.; Honda, Y.; Hideki, H.; Choi, H.M.; Choi, M.; Kim, H. Heat-mortality risk and the population concentration of metropolitan areas in Japan: A nationwide time-series study. Int. J. Epidemiol. 2021, 50, 602–612. [Google Scholar] [CrossRef] [PubMed]
  83. de Visser, M.; Kunst, A.E.; Fleischmann, M. Geographic and socioeconomic differences in heat-related mortality among the Dutch population: A time series analysis. BMJ Open 2022, 12, e058185. [Google Scholar] [CrossRef]
  84. Tong, S.; Prior, J.; McGregor, G.; Shi, X.; Kinney, P. Urban heat: An increasing threat to global health. BMJ 2021, 375, n2467. [Google Scholar] [CrossRef] [PubMed]
  85. Ma, W.; Zeng, W.; Zhou, M.; Wang, L.; Rutherford, S.; Lin, H.; Liu, T.; Zhang, Y.; Xiao, J.; Zhang, Y.; et al. The short-term effect of heat waves on mortality and its modifiers in China: An analysis from 66 communities. Environ. Int. 2015, 75, 103–109. [Google Scholar] [CrossRef]
  86. López-Bueno, J.A.; Navas-Martín, M.A.; Linares, C.; Mirón, I.J.; Luna, M.Y.; Sánchez-Martínez, G.; Culqui, D.; Díaz, J. Analysis of the impact of heat waves on daily mortality in urban and rural areas in Madrid. Environ. Res. 2021, 195, 110892. [Google Scholar] [CrossRef]
  87. Chen, L.; Yu, B.; Yang, F.; Mayer, H. Intra-urban differences of mean radiant temperature in different urban settings in Shanghai and implications for heat stress under heat waves: A GIS-based approach. Energy Build. 2016, 130, 829–842. [Google Scholar] [CrossRef]
  88. Ellena, M.; Ballester, J.; Mercogliano, P.; Ferracin, E.; Barbato, G.; Costa, G.; Ingole, V. Social inequalities in heat-attributable mortality in the city of Turin, northwest of Italy: A time series analysis from 1982 to 2018. Environ. Health 2020, 19, 116. [Google Scholar] [CrossRef]
  89. Marí-Dell’Olmo, M.; Tobías, A.; Gómez-Gutiérrez, A.; Rodríguez-Sanz, M.; García de Olalla, P.; Camprubí, E.; Gasparrini, A.; Borrell, C. Social inequalities in the association between temperature and mortality in a South European context. Int. J. Public Health 2019, 64, 27–37. [Google Scholar] [CrossRef] [PubMed]
  90. Ricciardi, V.; Mehrabi, Z.; Wittman, H.; James, D.; Ramankutty, N. Higher yields and more biodiversity on smaller farms. Nat. Sustain. 2021, 4, 651–657. [Google Scholar] [CrossRef]
  91. Li, M.; Gu, S.; Bi, P.; Yang, J.; Liu, Q. Heat waves and morbidity: Current knowledge and further direction-a comprehensive literature review. Int. J. Environ. Res. Public Health 2015, 12, 5256–5283. [Google Scholar] [CrossRef] [PubMed]
  92. Qiu, H.; Tian, L.; Ho, K.F.; Yu, I.T.; Thach, T.Q.; Wong, C.M. Who is more vulnerable to death from extremely cold temperatures? A case-only approach in Hong Kong with a temperate climate. Int. J. Biometeorol. 2016, 60, 711–717. [Google Scholar] [CrossRef]
  93. Desai, Y.; Khraishah, H.; Alahmad, B. Heat and the Heart. Yale J. Biol. Med. 2023, 96, 197–203. [Google Scholar] [CrossRef]
  94. Neuberger, M.; Rabczenko, D.; Moshammer, H. Extended effects of air pollution on cardiopulmonary mortality in Vienna. Atmos. Environ. 2007, 41, 8549–8556. [Google Scholar] [CrossRef]
  95. Neuberger, M.; Moshammer, H.; Rabczenko, D. Acute and subacute effects of urban air pollution on cardiopulmonary emergencies and mortality: Time series studies in Austrian cities. Int. J. Environ. Res. Public Health 2013, 10, 4728–4751. [Google Scholar] [CrossRef]
  96. Wen, B.; Wu, Y.; Guo, Y.; Gasparrini, A.; Tong, S.; Overcenco, A.; Urban, A.; Schneider, A.; Entezari, A.; Vicedo-Cabrera, A.M.; et al. Comparison for the effects of different components of temperature variability on mortality: A multicountry time-series study. Environ. Int. 2024, 187, 108712. [Google Scholar] [CrossRef]
  97. Bundesministerium für Arbeit, Soziales, Gesundheit, Pflege und Konsumentenschutz (Ed.) Nationaler Hitzeschutzplan Österreich; Bundesministerium für Arbeit, Soziales, Gesundheit, Pflege und Konsumentenschutz: Vienna, Austria, 2025. Available online: https://www.sozialministerium.gv.at/dam/jcr:3e224ed1-b320-461d-8d5e-2f3fdf643bde/Nationaler%20Hitzeschutzplan%20(Juni%202025).pdf (accessed on 11 January 2026).
  98. Hutter, H.-P.; Moshammer, H.; Wallner, P.; Leitner, B.; Kundi, M. Heatwaves in Vienna: Effects on mortality. Wien. Klin. Wochenschr 2007, 119, 223–227. [Google Scholar] [CrossRef]
  99. Han, L.; Qin, T.; Sun, Z.; Ren, H.; Zhao, N.; An, X.; Wang, Z. Influence of Urbanization on the Spatial Distribution of Associations Between Air Pollution and Mortality in Beijing, China. Geohealth 2023, 7, e2022GH000749. [Google Scholar] [CrossRef] [PubMed]
  100. Anenberg, S.C.; Haines, S.; Wang, E.; Nassikas, N.; Kinney, P.L. Synergistic health effects of air pollution, temperature, and pollen exposure: A systematic review of epidemiological evidence. Environ. Health 2020, 19, 130. [Google Scholar] [CrossRef]
  101. Moshammer, H.; Mayer, M.; Rieder, H.; Schmidt, C.; Bednar-Friedl, B.; Wallner, P.; Hutter, H.-P. Attributable deaths in Austria due to ozone under different climate scenarios. Eur. J. Public Health 2024, 34, 1015–1020. [Google Scholar] [CrossRef] [PubMed]
Climate 14 00016 g001
Figure 1. The non-linear effect of same-day temperature (left) and the fairly linear effect of 14-day average temperature (right) using natural splines for the city of Vienna as an example (effect estimates and 95% confidence intervals displayed).
Figure 1. The non-linear effect of same-day temperature (left) and the fairly linear effect of 14-day average temperature (right) using natural splines for the city of Vienna as an example (effect estimates and 95% confidence intervals displayed).
Climate 14 00016 g001
Climate 14 00016 g002
Figure 2. The results of the third-degree polynomial distributed lag model presented for the city of Vienna. (a) Each curve represents the temperature–mortality association on a separate lag (from lag0 meaning the same day, lag1 the previous day temperature, and so on until lag13). (b) The same data are presented in a three-dimensional surface (temperature from −10 to +30 degrees from left to right, lags from 0 to 13 from back to front, effect estimates from bottom to top).
Figure 2. The results of the third-degree polynomial distributed lag model presented for the city of Vienna. (a) Each curve represents the temperature–mortality association on a separate lag (from lag0 meaning the same day, lag1 the previous day temperature, and so on until lag13). (b) The same data are presented in a three-dimensional surface (temperature from −10 to +30 degrees from left to right, lags from 0 to 13 from back to front, effect estimates from bottom to top).
Climate 14 00016 g002
Climate 14 00016 g003
Figure 3. Geographic distribution of the coefficients for T14 (top) and T2 (bottom) in Austria.
Figure 3. Geographic distribution of the coefficients for T14 (top) and T2 (bottom) in Austria.
Climate 14 00016 g003
Table 1. Coefficients from the meta-regression on two models (bold: p < 0.05).
Table 1. Coefficients from the meta-regression on two models (bold: p < 0.05).
CoefficientsIn the General Additive ModelIn the Threshold Model
For T14For TFor T2For T14ThresholdFor T
Population weighted mean−0.01658250.00258670.0002777−0.017426213.57440.0302054
Minimum−0.03047−0.0068376−0.0000382−0.02888941.2711220.0050144
Maximum−0.00741450.019160.0005028−0.005549421.700160.0822177
CharacteristicFor T14For TFor T2For T14ThresholdFor T
Density SA −1.29 × 10−74.31 × 10−9 0.00008493.91 × 10−7
Density PSA −2.01 × 10−76.72 × 10−9 0.00013216.33 × 10−7
Age > 75 * 0.0003342 −0.0021118
Mean age 0.0003659−8.97 × 10−6 −0.0018286
(corr. urban) 9.57 × 10−6
Single −0.00025476.09 × 10−6
AT citizen−0.00005730.0001222−2.58 × 10−6 −0.0684247−0.0004073
(corr. urban)−0.0002029 5.46 × 10−6
Non-EU Europe0.000135−0.00030047.01 × 10−6 0.24258110.0013124
(corr. urban)0.0002923 −6.15 × 10−6−0.0002847 0.0011408
EU_new −0.00041780.0000123 0.35443830.0018426
(corr. urban) 0.0017414
AT born−0.00004830.0001114−2.56 × 10−6 −0.0667124−0.0003726
Basic educ. 0.0006108
University * −0.00012094.05 × 10−6
Homeless −0.00646020.0001433 8.0029410.0257093
Average income 9.24 × 10−6
Capital city −0.0014314
Vienna −0.00306310.0001199 2.2473480.0107399
Urban −0.00097520.0000348 0.68779490.0033279
Altitude2.49 × 10−65.79 × 10−6−3.63 × 10−73.86 × 10−6−0.0117487−0.0000404
Daily smoker0.0232675−0.0228118
(corr. urban)0.0251933−0.011098
(corr. altitude)0.0245164−0.0204694
Working 0.0003687−0.0000113 −0.3406221−0.0011915
Unemployed −0.00056720.0000174 0.41258320.0023906
Commuters−0.0000340.0000311
(corr. urban)−0.0000767 3.96 × 10−6−0.00010910.19555280.0008155
Green space−0.00003450.0000417
Tourism0.0000319 −8.93 × 10−70.0000359−0.0406086−0.0001016
(corr. altitude)0.0000381−0.0000193 0.0000347
% Jobs in:
Agriculture−0.00034550.0005917−0.0000107 −0.0016686
(corr. urban)−0.00061180.00028550.000011
Mining * 0.002492−0.0001113 −4.055816−0.0146685
Production * 0.0001526−5.69 × 10−6 −0.1181437−0.0005262
Energy0.0028134 −0.00014810.0040567 −0.0128922
(corr. urban)0.0039258 −0.00007970.0052977
Water/Waste 0.002235
(corr. urban) 0.0000891 4.9011240.0143599
Construction * 0.0003196−8.14 × 10−6
* After controlling for urbanity there are no longer any relevant trends. SA = settlement area, PSA = permanent settlement area; Capital city: Klagenfurt, St. Pölten, Linz, Salzburg, Graz, and Innsbruck compared to rural; Vienna: compared to rural; Urban: linear trend from rural–urban–capital–Vienna; Daily smoker: % daily smoker in last Austrian health micro-census (NUTS-3 data).
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Moshammer, H.; Jury, M.; Hutter, H.-P.; Wallner, P. Determinants of Spatial Variation in Vulnerability to Extreme Temperatures in Austria from 1970 to 2020. Climate 2026, 14, 16. https://doi.org/10.3390/cli14010016

AMA Style

Moshammer H, Jury M, Hutter H-P, Wallner P. Determinants of Spatial Variation in Vulnerability to Extreme Temperatures in Austria from 1970 to 2020. Climate. 2026; 14(1):16. https://doi.org/10.3390/cli14010016

Chicago/Turabian Style

Moshammer, Hanns, Martin Jury, Hans-Peter Hutter, and Peter Wallner. 2026. "Determinants of Spatial Variation in Vulnerability to Extreme Temperatures in Austria from 1970 to 2020" Climate 14, no. 1: 16. https://doi.org/10.3390/cli14010016

APA Style

Moshammer, H., Jury, M., Hutter, H.-P., & Wallner, P. (2026). Determinants of Spatial Variation in Vulnerability to Extreme Temperatures in Austria from 1970 to 2020. Climate, 14(1), 16. https://doi.org/10.3390/cli14010016

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop