Spatiotemporal Dynamics of Heat-Related Health Risks of Elderly Citizens in Nanchang, China, Under Rapid Urbanization

Abstract

Heatwaves intensified by climate change increasingly threaten urban populations, especially the elderly. However, most existing studies have concentrated on short-term or single-scale analyses, lacking a comprehensive understanding of how land cover changes and urbanization affect the vulnerability of the elderly to extreme heat. This study aims to investigate the spatiotemporal distribution patterns of heat-related health risks among the elderly in Nanchang City and to identify their key driving factors within the context of rapid urbanization. This study employs Crichton’s risk triangle framework to the heat-related health risks for the elderly in Nanchang, China, from 2002 to 2020 by integrating meteorological records, land surface temperature, land cover data, and socioeconomic indicators. The model captures the spatiotemporal dynamics of heat hazards, exposure, and vulnerability and identifies the key drivers shaping these patterns. The results show that the heat health risk index has increased significantly over time, with notably higher levels in the urban core compared to those in suburban areas. A 1% rise in impervious surface area corresponds to a 0.31–1.19 increase in the risk index, while a 1% increase in green space leads to a 0.21–1.39 reduction. Vulnerability is particularly high in economically disadvantaged, medically under-served peripheral zones. These findings highlight the need to optimize the spatial distribution of urban green space and control the expansion of impervious surfaces to mitigate urban heat risks. In high-vulnerability areas, improving infrastructure, expanding medical resources, and establishing targeted heat health monitoring and early warning systems are essential to protecting elderly populations. Overall, this study provides a comprehensive framework for assessing urban heat health risks and offers actionable insights into enhancing climate resilience and health risk management in rapidly urbanizing regions.

1. Introduction

Heatwaves are generally defined as prolonged periods of excessively high temperatures within a specific region. However, a universally accepted definition remains lacking in the academic community, largely due to variations in geographic, climatic, and disciplinary contexts [1]. In this study, we adopt the widely accepted definition used by the China Meteorological Administration, which characterizes a heatwave as a weather event during which the daily maximum temperature reaches or exceeds 35 °C for at least three consecutive days. This definition aligns with the operational requirements of China’s meteorological warning systems and public health interventions while also enabling robust spatiotemporal assessments of heat-related health risks at the urban scale [2,3]. The frequency, intensity, and duration of heatwaves have risen sharply worldwide, posing escalating threats to public health, particularly within urban settings [4,5,6]. Robust epidemiological evidence links heatwaves to significantly increased morbidity and mortality from chronic diseases such as cardiovascular, cerebrovascular, and respiratory disorders [7,8]. This intensifying health burden is especially acute in mid- to high-latitude regions and developing countries, positioning heatwaves as a paramount global health security challenge [9,10]. Consequently, there is an urgent need for systematic, spatiotemporally explicit assessments of the heatwave-related health risks in urban populations, alongside identification of their key driving factors. Such investigations are crucial to revealing spatial and temporal vulnerabilities in cities; informing targeted adaptation strategies; optimizing the allocation of limited resources; and underpinning the development of evidence-based public health policy.

Crichton’s risk triangle framework provides a robust model for assessing heat-related health risks by conceptualizing risk as the interplay of heat hazards, human exposure, and vulnerability [11,12]. This framework supports detailed analyses of the spatiotemporal distribution of heat risks across diverse urban settings and populations [13]. Advances in remote sensing and Geographic Information Systems (GISs) have refined high-resolution environmental metrics, such as land surface temperatures, impervious surface fractions, and green space coverage, enabling precise urban heat risk assessments [14,15]. However, most studies employ static, cross-sectional methods, often focusing on a single time point, thus failing to capture the temporal dynamics of heat-related health risks [16]. Moreover, quantitative insights into the relationship between land use changes and heatwave-induced health risks remain limited, and the potential of green space optimization as a sustainable mitigation strategy has largely been unexplored [17,18]. These gaps underscore the urgent need for longitudinal, multi-dimensional studies integrating temporal land cover dynamics with health vulnerability analyses to guide effective urban heat risk management.

Although numerous studies have explored the relationship between land use changes and heat-related health risks, most have been limited to single-scale, single-variable analyses and have predominantly concentrated on developed urban contexts, such as major cities in Europe and the United States [19,20,21]. Systematic investigations addressing medium-sized cities undergoing rapid urbanization remain scarce, particularly in developing countries. In such contexts, land use changes can indirectly influence heat health risks by altering population density, infrastructure distribution, and the allocation of healthcare and environmental resources [22,23]. Urban cores, characterized by a high building density and limited green coverage, often experience elevated heat exposure, while peri-urban and fringe areas tend to suffer from inadequate infrastructure and a reduced adaptive capacity, resulting in heightened vulnerability to heat-related health impacts [24,25,26]. These spatial disparities underscore the importance of evaluating both the direct and indirect effects of land use changes on heat health risks. As a result, integrated research that considers the heterogeneous impacts of urban form and land cover dynamics has emerged as a critical direction in advancing the field of health risk assessments in urban climates.

The elderly, a highly thermally vulnerable demographic, face significantly elevated health risks during heatwaves [27,28]. Aging impairs thermoregulatory capacity, physiological resilience, and immune function, increasing susceptibility to heat-related conditions such as heatstroke, dehydration, and electrolyte imbalances [29]. Moreover, older adults exhibit a higher prevalence of chronic diseases, including hypertension, diabetes, and congestive heart failure, which are exacerbated by extreme heat, elevating the risks of severe outcomes like cardiovascular events and respiratory failure [22]. Beyond physiological factors, socioeconomic and environmental constraints—such as limited income, substandard housing, and weak social support—heighten vulnerability further by restricting access to critical heat mitigation measures, including air conditioning, public cooling centers, and timely medical care [30,31]. These complex, intersecting challenges highlight the urgent need to integrate age-specific vulnerabilities into urban heat risk assessments and develop targeted adaptation strategies to protect this population during extreme heat events.

Nanchang, a rapidly urbanizing city in central China with a growing elderly population, has experienced significant land use transformation over recent decades [32]. The extensive conversion of agricultural land, water bodies, and forests into urban construction zones has increased impervious surfaces and reduced green spaces, intensifying the urban heat island (UHI) effect and elevating the heat-related health risks for elderly residents during extreme heat events [33,34,35]. Spatial disparities in urban form, infrastructure, and medical resource distribution exacerbate the uneven heat vulnerability among elderly populations across neighborhoods further [36]. Given these dynamics, Nanchang provides a critical case for studying the interplay between land use changes and heat-related health risks in aging urban populations. This study integrates multi-source data, including meteorological records, land surface temperature, land use dynamics, and socioeconomic indicators within Crichton’s risk triangle framework, to quantify the spatiotemporal distribution and evolution of heat health risks among the elderly. It also examines the mechanisms through which land use changes influence these risks. Specifically, this study addresses three questions: (1) What are the spatiotemporal patterns of heat-related health risks for Nanchang’s elderly? (2) How do sensitivity and adaptive capacity vary among elderly populations across urban areas? (3) How do land use changes, such as the expansion of impervious surfaces and alterations in green space, affect heat health risks in the elderly?

2. Methods

2.1. The Study Area

Nanchang, located in the central part of Jiangxi province (115°27′–116°35′ E, 28°10′–29°11′ N), covers an area of 7195 km² (Figure 1). The city’s resident population was 6.54 million, with an urbanization rate of 56.38%, as of 2022 [37]. The city’s topography is predominantly flat, characterized by low mountains, hills, and river valley plains, with the overall elevation increasing from south to north. Nanchang experiences a subtropical monsoon climate, marked by significant interannual temperature variability. Historical meteorological records indicate that extreme summer temperatures can reach 40.9 °C, while winter temperatures can drop to −15.2 °C. Due to its monsoonal climate, this region faces an uneven rainfall distribution and is prone to extreme weather events such as heatwaves, droughts, and heavy rainfall-induced flooding, all of which significantly impact the health of its elderly population.

2.2. The Conceptual Framework for Heat Risk Assessment

This study employs Crichton’s risk triangle framework to assess heat-related health risks in the elderly, conceptualizing risk as the interplay of three components: the Heat Hazard Index (HHI), the Human Exposure Index (HEI), and the heat vulnerability index (HVI) [11]. The elevation-adjusted ambient human settlement index (EAHSI), which correlates strongly with the population distribution, is used to represent the grid-scale HEI, aligned with the spatial distribution of heat hazards [38]. To ensure a comprehensive and robust analysis, data from four temporal snapshots (2002, 2008, 2014, and 2020) were analyzed. The selection of these years was guided by three principal considerations: (1) each year recorded significant heatwave events during the 2002–2020 period, thereby serving as representative samples; (2) comprehensive data on extreme heatwaves and corresponding socioeconomic statistics for these years were readily accessible; and (3) these years closely coincide with China’s national census years (2000, 2010, and 2020), which improves the accuracy and consistency of exposure and vulnerability assessments. Standardized indicators were applied across multiple dimensions to ensure consistency and comparability (Figure S1). The composite heat risk index (HRI) was constructed by standardizing, weighting, and aggregating the dimension scores (hazard, exposure, and vulnerability). Z-score standardization was applied to normalizing each dimension’s scores, ensuring comparability across indicators. The weights for each dimension were determined based on expert evaluations and the results of a principal component analysis (PCA). All indicators were normalized using the principal components extracted via the PCA prior to weighting, ensuring comparability across different dimensions and indicators. This integrated assessment model improves the accuracy of heat-related health risk evaluations and provides a scientific framework for public health responses to heatwaves.

The formulas for HRI calculation are provided below.

$$HRI=0.4\times HHI+0.3\times HEI+0.3\times HVI$$

(1)

Land cover data were obtained from the European Space Agency’s World Cover product at a 10 m resolution (https://worldcover2020.esa.int/, accessed on 16 June 2024). Land surface temperature data were derived from the MODIS 1 km resolution LST product (https://modis.gsfc.nasa.gov/data/dataprod/, accessed on 17 June 2024). Daily maximum temperature data were obtained from the NOAA National Centers for Environmental Information (https://www.ncei.noaa.gov/, accessed on 19 June 2024). To ensure compatibility with the MODIS land surface temperature datasets, the original meteorological station observations were spatially interpolated to a 1 km × 1 km resolution. This preprocessing step facilitated the subsequent raster-based analysis of heat hazards. Air quality data were provided by the Jiangxi Provincial Ecological and Environmental Monitoring Center. Nighttime light data were acquired from the NOAA National Geophysical Data Center (https://ngdc.noaa.gov/eog/, accessed on 16 June 2024). Digital Elevation Model (DEM) data were obtained from China’s National Geographic Information Public Service Platform (http://www.ngcc.cn/, accessed on 20 June 2024). The Enhanced Vegetation Index (EVI) was derived from the MODIS MOD13Q1 dataset (https://modis.gsfc.nasa.gov/data/dataprod/mod13.php, accessed on 21 June 2024), and water body data were sourced from the USGS Global Surface Water product (https://global-surface-water.appspot.com/ accessed on 22 June 2024). Socioeconomic data, including the distribution of the elderly population, the prevalence of disability among the elderly and elderly individuals living alone (without caregivers or cohabitation with children), road density, housing conditions, per capita GDP, and healthcare resource availability, were collected from the Nanchang census.

2.3. Heat Hazards

This study analyzed daily maximum air temperature, air quality, and daytime and nighttime land surface temperature (LST) data for heatwaves occurring in 2002, 2008, 2014, and 2020. To minimize cloud contamination and noise, MODIS LST data were processed by removing cloud-affected pixels and filling gaps via interpolation (Figure S2a–d). The daytime and nighttime LST values were averaged separately to capture diurnal temperature variations. Additionally, the NOAA meteorological station data were spatially interpolated to a 1 km × 1 km resolution to ensure spatial consistency with the MODIS datasets.

The heat hazard indicators (daily maximum air temperature, air quality, daytime LST, and nighttime LST) were normalized using Z-score standardization. A Pearson’s correlation analysis, performed in SPSS, assessed the indicator relationships, eliminating redundant variables. The ArcGIS 10.8 raster calculation tool applied weighted summation and normalization based on the indicator weights, generating the Heat Hazard Index (HHI; 0–100) for 2002, 2008, 2014, and 2020. The HHI was classified into five levels using interval classification.

2.4. Human Exposure

Human exposure in Nanchang was evaluated using four indicators: total elderly population, elderly individuals receiving subsistence allowances, disabled elderly, and those living alone (i.e., without caregivers or cohabitation with children). The Pearson’s correlation analysis revealed that all pairwise correlation coefficients among these indicators exceeded 0.714 (p < 0.001) (Table S2). To reduce dimensionality and eliminate redundancy, a principal component analysis (PCA) was conducted using SPSS 29.0 software. The PCA reduces multicollinearity by extracting the principal components that capture the maximum variance and assigns corresponding weights to each indicator. A composite exposure index was then derived by combining the four indicators through a variance-weighted method, utilizing the rates of contribution of the first four principal components to the variance as weights (Appendix S1).

In the calculation of the elevation-adjusted human settlement index (EAHSI), this study utilized the approach proposed by Yang et al. [39], integrating NTL images, EVI data, and DEM data to generate EAHSI images, which greatly reduced the average relative error in the population estimation. The formulas are as follows:

$$EAHSI={\displaystyle \frac{(1-EV{I}_{\max })+OL{S}_{nor}}{(1-OL{S}_{nor})+EV{I}_{\max }+OL{S}_{nor}\times EV{I}_{\max }}}\times DE{M}^{-0.314}\times P_{\mathrm{e}lder}$$

(2)

$$OL{S}_{nor}=(OLS-OL{S}_{\min })/(OL{S}_{\max }-OL{S}_{\min })$$

(3)

where EVI_max is the maximum value for the EVI data in 2002, 2008, 2014, and 2020; DEM is the elevation data for Nanchang; P_elder is the percentage of the elderly population in each county/district; and OLS_nor is the standardized nighttime light data for four years. OLS_max and OLS_min correspond to the year’s maximum and minimum values in the nighttime light data. Finally, the EAHSI was divided into five grades using the interval classification method and normalized using ArcGIS 10.8 software from 0 to 100.

2.5. Heat Vulnerability

This study established a heat vulnerability indicator system for Nanchang comprising 11 indicators: proximity to vegetation, proximity to water bodies, slope, relief degree of the land surface, annual per capita GDP, average number of rooms per dwelling, per capita dwelling area, number of households with bathing facilities, number of households with toilets, total number of beds in health institutions, and highway density. To address multicollinearity, collinearity diagnostics were applied, leading to adjustments of the initial indicators (Appendix S2, Table S3). Variance inflation factors (VIFs) were calculated for each variable, and those exhibiting high VIF values were excluded to ensure that the final set of indicators had low multicollinearity. This approach effectively minimized the influence of redundant information on the heat vulnerability index.

To validate the principal component weights derived from the PCA further, the weight distribution for each indicator was confirmed based on the variance explained by each principal component. This validation ensured that the final heat vulnerability index accurately reflected the relative importance of different variables and reduced the risk of overfitting. Spatial data processing was performed using ArcGIS 10.8, where pixel-level statistics for vegetation, water bodies, slope, and degree of relief of the land surface were extracted from the Digital Elevation Model (DEM). Socioeconomic indicators were standardized and spatially aligned with administrative boundary data. Pearson’s correlation analysis and collinearity diagnostics were employed to identify and remove redundant or irrelevant indicators. The remaining variables were normalized in ArcGIS 10.8 to generate a heat vulnerability index (HVI) ranging from 0 to 100, which was subsequently classified into five levels using interval classification.

2.6. Validation

To validate the heat risk assessment, monthly all-cause mortality data for Nanchang’s towns and subdistricts during the summers (June–August) of 2014 and 2020 were analyzed, with the total population per administrative unit as a control variable. A generalized linear mixed model (GLMM) and polynomial regression were used to evaluate the relationship between the heat risk index (HRI) and summer mortality. Due to the absence of heat-specific mortality data, all-cause summer mortality served as a proxy, acknowledging its limitations. The average HRI values for each town and subdistrict in 2014 and 2020 were calculated using ArcGIS 10.8. The GLMM’s results revealed a statistically significant, albeit weak, correlation between the HRI and all-cause summer mortality (r = 0.136, p < 0.0001; Table S1), supporting the validity of the heat risk assessment model. Polynomial regression yielded a coefficient of determination R² = 0.3171 (p < 0.0001; Figure 2), further confirming the HRI’s reliability in assessing heat risk.

2.7. Data Statistics and Analysis

Ecosystem services were evaluated using the coefficient table developed by Xie et al. [40], based on land cover data from 2002, 2008, 2014, and 2020. Land cover types were reclassified in ArcGIS 10.8, and the corresponding ecosystem service values were calculated (Table S4). Correlation analyses were performed to examine the relationships between the ecosystem service values and the heat risk index (HRI) for each year, with positive and negative correlations indicated by positive and negative values, respectively. A 5 × 5 m grid was generated in ArcGIS 10.8 to extract the proportions of impervious surfaces and green spaces per grid cell. Zonal statistics were used to compute the HRI values for each grid. Linear regression analyses, conducted in R, assessed the relationships between the proportion of impervious surfaces, green space coverage, and the HRI.

3. Results

3.1. The Spatial Distribution of Heat Hazards in Nanchang from 2002 to 2020

Between 2002 and 2020, Nanchang’s Heat Hazard Index (HHI) exhibited a distinct north–south gradient, with significant increases over time, particularly in the central and southern districts (Figure 3a–d). In contrast, the northern and northwestern districts consistently showed lower HHI values, with minimal annual variation. The descriptive statistics (Figure 3e) highlight substantial fluctuations in the central urban area. For example, the median HHI in Qingshanhu district rose from 51.72 in 2002 to 72.54 in 2020, while Qingyunpu district reached 74.68 in 2020. Xihu, Donghu, and Honggutan districts also recorded high values, averaging 69.24, 68.35, and 71.46, respectively, in 2020. Conversely, Xinjian and Wanli districts maintained HHI values below 40 throughout the study period, indicating lower heat hazard levels. Notably, areas such as Anyi County and Donghu district exhibited considerable variability in 2014, reflecting pronounced intra-city disparities. These spatial patterns likely stem from varying urbanization levels and uneven green space distribution.

3.2. The Spatial Distribution of Human Exposure in Nanchang from 2002 to 2020

From 2002 to 2020, Nanchang’s Human Exposure Index (HEI) increased steadily, with pronounced spatial clustering toward the urban core (Figure 4a–d). In 2002, high-exposure zones were concentrated in the central urban area, while the peripheral regions exhibited low exposure levels. The exposure intensity rose progressively, with high-exposure zones expanding significantly after 2014. By 2020, extremely high-exposure zones encompassed nearly the entire central urban area and parts of surrounding regions, reflecting population concentration driven by urbanization. District-level statistics (Figure 4e) show that in 2020, the median HEI values for Qingyunpu and Xihu districts were 65.23 and 68.64, respectively, compared to values below 40 in Xinjian and Wanli districts. The interquartile range of the HEI in the central urban area widened over time, indicating increasing disparities in population density. The linear regression analysis (R² = 0.92; Figure S6) revealed a strong correlation between the total population and the Living Environment Index, with higher HEI values in densely populated areas.

3.3. The Spatial Distribution of Heat Vulnerability in Nanchang from 2002 to 2020

From 2002 to 2020, Nanchang’s heat vulnerability index displayed a general decline, with higher values in the eastern region and lower values in the central and western areas, revealing significant regional disparities (Figure 5a–d). In 2002, the high-vulnerability zones were concentrated in the eastern and central parts of Nanchang, but this pattern weakened over time. By 2020, high-vulnerability zones were primarily located along the eastern periphery, while western areas exhibited predominantly low vulnerability. The district-level statistics (Figure 5e) indicate that Jinxian County’s median HVI consistently exceeded 58.65 from 2002 to 2020, reflecting high citywide vulnerability, whereas Qingyunpu and Wanli districts maintained low HVI values below 30. By 2020, the vulnerability decreased significantly across most areas; however, disparities persisted in the central urban zone, suggesting an improved but unevenly distributed adaptive capacity.

3.4. The Spatial Distribution of Heat Risk in Nanchang from 2002 to 2020

From 2002 to 2020, Nanchang’s heat risk index (HRI) increased consistently, with its spatial distribution shifting from dispersed to concentrated in the central urban area (Figure 6a–d). High-risk zones were primarily located in Donghu, Xihu, Qingshanhu, and Qingyunpu districts, whereas peripheral areas, including Anyi, Wanli, and Jinxian Counties, exhibited lower risk. The district-level statistics (Figure 6e) show that in 2002, the median HRI values in high-risk areas ranged from 40 to 65, compared to values below 20 in low-risk areas. By 2014, Qingyunpu district’s HRI peaked at 75.46, and by 2020, the central urban risk had intensified, while the peripheral areas remained below 30. This rising HRI trend correlates strongly with population aging. For example, Donghu district’s proportion of the elderly population increased from 7.19% in 2000 to 14.74% in 2020. High-risk areas generally aligned with regions with a high elderly population density, highlighting demographic structure as a key driver of spatial disparities in heat risk.

3.5. The Correlation Between Land Use Changes and the Heat Risk Index in Nanchang from 2002 to 2020

From 2002 to 2020, Nanchang’s land cover underwent significant changes, with a marked decline in cropland and forests and a substantial increase in urban construction land, particularly in the central urban areas and adjacent regions (Figure S8). This loss of cropland and forests reduced the natural cooling capacity, intensifying heatwave effects. High-risk zones, characterized by a high proportion of built-up land, were concentrated in the urban cores, while low-risk zones, with greater forest and water body coverage, were primarily located in the peripheral areas (Figure 7).

The linear regression analyses (Figure 8a) revealed a positive correlation between the proportion of impervious surfaces and the heat risk index (HRI) from 2002 to 2020 (red line: slope = 0.31, R² = 0.1286; blue line: slope = 1.19, R² = 0.3639), indicating that urbanization-driven increases in impervious surfaces exacerbate heat risks. Conversely, a negative correlation was found between the proportion of green space and the HRI (Figure 8b; red line: slope = −1.39, R² = 0.2798; blue line: slope = −0.21, R² = 0.1093), suggesting that higher green space coverage mitigates heat risks, while its reduction amplifies the risk during urbanization.

4. Discussion

Applying Crichton’s risk triangle framework, this study quantifies the spatiotemporal dynamics of heatwave-related health risks among Nanchang’s elderly population from 2002 to 2020 and examines the influence of land use changes. The results reveal that heat risks are most pronounced in the urban core, with a significant increase over the study period. A strong positive correlation between the proportion of impervious surfaces and the heat risk index (HRI) underscores urbanization as a primary driver of heightened heat risks. Conversely, a significant negative correlation between green space coverage and the HRI highlights the critical role of green spaces in mitigating heat-related health risks. These findings confirm urbanization’s exacerbating effect on heat risks and emphasize the importance of strategic planning of green space to reduce the health impacts of heatwaves on vulnerable populations.

4.1. The Spatiotemporal Distribution and Evolution of Health Risks from Heatwaves

This study reveals that the highest heatwave-related health risks in Nanchang are concentrated in the urban core, with a notable increase observed between 2002 and 2020. This trend is primarily driven by factors such as a high population density, expanding coverage of impervious surfaces, and a reduction in green spaces within the city center. These findings are consistent with previous research conducted in Zhejiang [38] and Chongqing [41], which similarly attribute elevated heatwave risks to urbanization-induced land use changes. Compared to these regions, Nanchang exhibits more pronounced fragmentation of green spaces in its central urban area, compounded by a lack of systematic green space planning, potentially exacerbating the urban heat island effect. Moreover, the spatial distribution of heat risk closely corresponds with the HHI and the HEI, suggesting that these indices are primary contributors to high-risk zones. This observation aligns with global studies on urban heat islands. The interplay between urban expansion, land use changes, and the HEI has intensified the spatial concentration of heat-related health risks in Nanchang, mirroring the patterns observed in other global cities [42,43]. This supports the conceptual framework of “urban development–land use change–heightened heat-related health risk”. Mitigating future urban heat risks requires not only enhanced protection for vulnerable populations but also optimized urban planning and green space distributions. Consequently, rapidly urbanizing cities like Nanchang should integrate environmental and health considerations into their urban expansion and adopt comprehensive climate adaptation strategies to alleviate heatwave-associated health threats.

4.2. The Effects of Land Use Changes on Heat Risk

This study reveals that land use changes in Nanchang significantly influence the HRI, with the proportion of impervious surfaces positively correlated and the proportion of green spaces negatively correlated with the HRI. The expansion of impervious surfaces exacerbates the urban heat island effect, raising local temperatures and heat risks, consistent with findings in Southeast Asia [44]. Although the overall area of green space has increased, its uneven spatial distribution—especially in the urban core—limits the effectiveness of cooling. Unlike studies that have considered only the total area of green space [45,46], our results emphasize the importance of its spatial configuration to mitigating heat risks. Furthermore, lag effects may delay the impact of changes in impervious surfaces on heat risks, while interactions with population density and aging amplify vulnerability. Areas with a high population density and aging populations experience greater heat risks, whereas younger populations may exhibit a stronger adaptive capacity. This interaction warrants further investigation. Therefore, urban planning should prioritize optimizing the layout of green spaces over simply expanding their area; focusing on high-risk and densely populated zones; and establishing ecological buffer zones in peripheral towns to enhance the overall cooling benefits.

4.3. Spatial Disparities in Heat Vulnerability and Human Exposure

This study identifies that heat vulnerability in Nanchang is predominantly concentrated in peripheral towns, where lower socioeconomic development and limited infrastructure heighten the susceptibility to heatwave impacts. This finding aligns with Hu et al. [38]’s in Zhejiang province but contrasts with Reid et al. [17]’s in U.S. cities, where higher vulnerability was observed in densely populated urban centers. These differences likely reflect distinct urbanization patterns and socioeconomic contexts between China and the United States. In Nanchang, the relatively higher socioeconomic status and better healthcare infrastructure in the urban core reduce vulnerability, whereas resource scarcity and inadequate medical facilities exacerbate risks in peripheral towns (Figures S10 and S11). Despite overall urban improvements, peripheral areas continue to face high heat vulnerability driven by factors including low income and education levels; limited awareness and preventive behaviors; and imbalanced policy support, leading to insufficient healthcare and emergency services. Therefore, targeted investments to enhance the infrastructure, healthcare, and social security in peripheral towns, especially for emergency responses to heatwaves, are essential to mitigating heat vulnerability.

4.4. Mitigation Measures for Heatwave Health Risks

This study highlights that the elderly population in Nanchang faces elevated health risks from heatwaves, driven by rapid urbanization, increased impervious surfaces, and unequal access to green spaces and healthcare services. While urban centers experience intensified heat hazards due to dense built environments and limited vegetation, peripheral regions exhibit greater vulnerability owing to inadequate infrastructure and lower socioeconomic status. These findings underscore the critical need to integrate land use planning with health risk mitigation strategies that address both environmental and social determinants of heat vulnerability. A comprehensive approach is necessary, emphasizing the expansion and equitable distribution of green spaces—including pocket parks, rooftop gardens, and ecological buffer zones—to alleviate urban heat stress, particularly in high-density and under-served areas [47,48,49]. Nonetheless, challenges such as resource limitations and land ownership complexities require attention, with local governments encouraged to leverage public–private partnerships (PPPs) to overcome these barriers [50].

To effectively reduce heat-related health risks among the elderly, these interventions should be embedded within local governance frameworks. Policy priorities must focus on promoting green infrastructure in high-risk zones, fostering collaboration among urban planning, public health, and community organizations to ensure coordinated implementation [51,52]. Establishing cooling centers in vulnerable neighborhoods is vital, complemented by temporary cooling stations in resource-constrained public spaces [53]. Furthermore, mobile healthcare services should be enhanced in peripheral areas through on-site units and improved critical utilities, including water and electricity infrastructure. Collaboration with meteorological and public health agencies is also essential to develop targeted early warning systems for the elderly, utilizing SMS alerts, media broadcasts, and community outreach to deliver timely information [54,55,56]. Collectively, these efforts will strengthen urban resilience and protect vulnerable populations from increasing climate-induced heat stress. Continuous monitoring and feedback mechanisms are imperative for evaluating the effectiveness of interventions and facilitating adaptive policy adjustments.

4.5. Limitations and Perspectives

While this study provides a comprehensive analysis of the spatiotemporal patterns in and key drivers of heatwave-related health risks among the elderly in Nanchang, several limitations warrant consideration. The county-level spatial resolution, constrained by the data availability, may obscure finer intra-urban variations; future research should leverage higher-resolution socioeconomic and remote sensing data alongside field surveys to address this gap. The use of four temporal snapshots limits the ability to capture dynamic factors such as urban expansion and population aging. Moreover, the validation based on summer all-cause mortality data is limited, as it cannot specifically attribute deaths to heat exposure; thus, future studies should incorporate more direct health data sources, including heat-related mortality records and individual health monitoring, complemented by long-term health tracking and detailed surveys, to strengthen causal inference. The uncertainties in remote-sensing-derived land surface temperatures may have affected the exposure estimates, which could be improved by integrating ground-based meteorological data and refining the correction methods. Lastly, although the roles of green space and impervious surfaces were analyzed, interactions between thermal environments and social vulnerability were not explored. A socio-ecological system approach that integrates natural conditions, infrastructure, socioeconomic disparities, and healthcare access is recommended to elucidate the complex mechanisms driving heat-related health risks in the elderly better.

5. Conclusions

This study applied Crichton’s risk triangle framework to examining the spatiotemporal evolution of heatwave-related health risks among the elderly in Nanchang from 2002 to 2020. The results indicate that the highest risks are concentrated in the urban center, driven by elevated heat hazards and dense elderly populations, while peripheral rural areas experience greater vulnerability due to their socioeconomic disadvantages and limited access to healthcare. Rapid urbanization has intensified these risks by increasing impervious surfaces, whereas the expansion of green space has mitigated heat impacts to some extent; however, fragmented and unevenly distributed green spaces limit their effectiveness. These findings underscore the need for integrated land use planning and targeted green infrastructure strategies to reduce urban heat health risks.

Based on these insights, urban management and policy should prioritize vulnerable populations in high-risk areas. In urban centers, enhancing small-scale green spaces and improving green network connectivity can increase the heat buffering capacity. Peripheral areas require strengthened medical resource allocation and expanded green space coverage to establish essential ecological and service support systems. Future land use planning should incorporate the coupling between the layout of green spaces, population density, and heat risk as key indicators to improve climate adaptation. Further research is needed to explore the heat health risk responses among different socioeconomic groups and quantify the long-term effectiveness of adaptation strategies through simulations. Establishing interdisciplinary collaboration across urban planning, public health, and climate adaptation sectors is recommended to develop dynamic heat risk monitoring and early warning systems, thereby enhancing urban health resilience and protecting elderly populations better.