Finer-scale urban health risk assessment based on the interaction perspective of thermal radiation, human, activity, and space

Abstract

Urban heat stress profoundly affects the health of residents. However, current research primarily focuses on quantifying the risk of urban heat based on 1ST, T_a, etc., overlooking the crucial and intimate influence of continuous intense solar radiation on human thermal comfort and health. Simultaneously, there is a lack of smaller units to support more precise planning. This study utilized the radiant heat stress intensity (RHSI) metric concentrating on the intensity and duration of thermal radiation, to develop a thermal-radiation induced health risk (TIHR) assessment system. Leveraging technologies such as the SOLWEIG model, Python, BERT, and GIS enables precise calculations of 12 spatial indices, including RHSI and Weibo heat. This facilitates a more accurate assessment of health risks at the smallest urban units (blocks) and directly guides planning. The application of this workflow in the case of Suoyuwan, Dalian, China, confirms its value, as it can be used to determine which blocks should be prioritized for specific aspects of risk prevention and control. The results show that some blocks exhibited differences in TIHR even within close proximity, with disaster-causing factors varying according to locations. This study proposes a novel assessment framework based on the interactive perspective of thermal radiation-human-activity-space.

Headnote

Received 17 November 2023; received in revised form 7 January 2024; accepted 1 February 2024

KEYWORDS

Health risk assessment;

Thermal radiation;

Finer-scale;

Block;

Planning;

Human-centered

Abstract Urban heat stress profoundly affects the health of residents. However, current research primarily focuses on quantifying the risk of urban heat based on 1ST, T_a, etc., overlooking the crucial and intimate influence of continuous intense solar radiation on human thermal comfort and health. Simultaneously, there is a lack of smaller units to support more precise planning. This study utilized the radiant heat stress intensity (RHSI) metric concentrating on the intensity and duration of thermal radiation, to develop a thermal-radiation induced health risk (TIHR) assessment system. Leveraging technologies such as the SOLWEIG model, Python, BERT, and GIS enables precise calculations of 12 spatial indices, including RHSI and Weibo heat. This facilitates a more accurate assessment of health risks at the smallest urban units (blocks) and directly guides planning. The application of this workflow in the case of Suoyuwan, Dalian, China, confirms its value, as it can be used to determine which blocks should be prioritized for specific aspects of risk prevention and control. The results show that some blocks exhibited differences in TIHR even within close proximity, with disaster-causing factors varying according to locations. This study proposes a novel assessment framework based on the interactive perspective of thermal radiation-human-activity-space.

(ProQuest: ... denotes formulae omitted.)

1. Introduction

Increasing global warming and urbanization have exacerbated urban heat stress, affecting sustainable urban development, as well as the safety and health of residents. An editorial published in The Lancet (2022) suggests that the health effects of climate change have been significantly underestimated. Heat stress can lead to fatigue, heat stroke, cardiovascular disease, and even pose fatal threats (Fang et al., 2021; He et al., 2022a, 2022b).

Heat risk assessment based on high urban temperatures has received widespread attention. Most studies have used land surface temperature (1ST) (Chen et al., 2023; Hondula et al., 2015; Kotharkar et al., 2019; Russo et al., 2017) and air temperature (T_a) (Hajat et al., 2010; Jänicke et al., 2019) to characterise the hazard intensity because of their accessible. Thermal radiation caused by sensible and latent heat flux has a more direct impact on human health and thermal comfort under strong sunlight (Amani-Beni et al., 2023; Chen et al., 2016; Huang et al., 2022), yet current research falls short in its examination of thermal radiation risk assessment from a human-centric standpoint. The mean radiant temperature (T_mrt) more accurately reflects the actual heat load perceived by the human body in exposed environments and is more effective in guiding health protection measures in thermal environments than considering T_a and LST alone. T_mrt is affected by complex urban geometries and surface materials, resulting in significant spatial heterogeneity, and has been widely used in human bio-meteorological studies (Chen et al., 2016). Research by Lau et al. (2016) suggests that health studies should consider T_mrt rather than T_a because T_mrt more accurately determines thresholds for heat-related mortality. When T_mrt exceeds 59.4 °C, the risk of mortality increases significantly. In addition, the intensity and duration of thermal hazard occurrences together influence the level of heat stress in cities (Hajat et al., 2010). Current research metrics only reflect hazard intensity and lack comprehensive assessment that explicitly reflects both (Dong et al., 2020). Chen et al. (2016) proposed a threshold of T_mrt = 60 °C for the occurrence of heat stress and cumulative hourly difference in T_mrt above 60 °C for all hours of sunshine in a day to define the Radiant Heat Stress Intensity (RHSI) Index.

Highly accurate and spatially detailed data can help capture the distribution of heat risks and their corresponding health impacts at a finer level, providing a more practical basis for decision-makers (Dong et al., 2020; Schmidtlein et al., 2008). However, the majority of the current studies operate at resolutions ranging from a few tens of meters to kilometers and are often based on a corresponding raster size or on cities and districts at the administrative level as the study unit (Dong et al., 2020; Estoque et al., 2020; Kotharkar et al., 2019; Russo et al., 2017; Schmidtlein et al., 2008). This approach, unfortunately, leads to a notable oversight in capturing the finer details of the spatial thermal environment where activities take place, making it difficult to accurately portray the spatial heterogeneity of the urban thermal environment.

In addition, the dimensions from which the thermal radiation-induced health risk (TIHR) system is constructed are important prerequisites for proposing effective prevention and control decisions (Prudhomme et al., 2013). Crichton's Risk Triangle and the Heat Vulnerability Index (HVI) stand as two widely adopted assessment frameworks (Chen et al., 2018; Dong et al., 2020; Wang et al., 2023). The Crichton's Risk Triangle framework integrates hazard, exposure, and vulnerability (HEV) into a mathematical model, enabling a quantitative analysis of heat risk assessment. Conversely, some scholars define risk as a function associated with exposure, sensitivity, and adaptability, illustrating the degree of threat posed by climate to individuals and cities, the responsiveness of these entities to climate-related impacts, and their capability to minimize resultant losses (Reid et al., 2009; Wilhelmi and Hayden, 2010). The HVI framework revolves around these aspects. Exposure and sensitivity, similar to the heat hazard index and high-temperature exposure in Crichton's model, complement the dimension of adaptability, which has the opposite meaning to vulnerability (Yang et al., 2010; Zheng et al., 2020). Taking cues from the systematic and complementary nature of these widely used frameworks, exposure, hazard (indicating the level of hazard and likelihood of disaster occurrence), system vulnerability, and adaptability are integrated into a more comprehensive Hazard-Exposure-Vulnerability-Adaptability (HEVA) framework (Wu et al., 2019). However, to the best of our knowledge, there is a dearth of studies on human health risks that utilize this framework (Morabito et al., 2015; Tomlinson et al., 2011), particularly with high-precision spatial data.

In order to provide risk assessment information closely related to human health and finer planning guidelines, we focus on evaluating health risks from a human-centered interaction perspective of thermal radiation, human, activity, and space. Based on the theory of HEVA, the block is used as the basic research unit to provide high-precision risk assessment results for typical summer days. This approach aims to achieve finer-scale urban thermal radiation-induced health risks (TIHR) assessment. Specifically, 1) to select and determine the TIHR assessment indices and to construct the TIHR system, in order to establish a scientific and replicable TIHR assessment process; 2) to identify the medium and high-risk areas and analyze the inducers of the risk, which is taken the typical area of Dalian, China as an example; 3) to propose the targeted prevention and control measures by combining the types of construction land.

2. Methodology

2.1. Study area

Dalian is located in a cold region of northeastern China with high summer temperatures and abundant sunshine. Urban heat stress has become increasingly serious in recent years. Extreme heat events in Northeast Asia during the summer of 2018 led to a dramatic increase in the number of patients with heatstroke in Dalian, as compared to previous years (Guo et al., 2023b; Ren et al., 2020). People in cold regions are more sensitive to hot weather (Wang et al., 2021). Based on this, the Suoyuwan area in Dalian was selected for risk assessment in this paper. The total area of this region is about 43.5 km2, with a land area of 28.7 km2 (Fig. 1). The city centre is located south of Shugang Road, which is known for being the most economically developed and densely populated. The eastern part of the area is adjacent to Dalian Suoyuwan, with heavy industrial zones along the bay and new urban areas to the north. With diverse land use types, road network patterns and building forms, urban heat stress varies significantly at both fine spatial and temporal levels. Therefore, conducting a scientific risk assessment at finer scale is an important measure for effectively mitigating urban heat risks.

2.2. Data sources

This study utilizes multiple sources of data to quantify the assessment of health risks. The data sources are presented in Table 1.

Figure 2 shows a research framework for this study. First, applying the NEVA theory, we emphasized the interactive perspective of "thermal radiation-human-activityspace" to conduct a health-oriented assessment of heat risks. Second, correlation analysis was used to filter the indices to construct the TIHR assessment index system. The assessment indices were calculated using the visual processing platforms of ArcGIS and QGIS. Lastly, the TIHR map of the Dalian Suoyuwan area was obtained using the CRITIC (Criteria Importance Through Intercriteria Correlation) weighting method and the function model method. Risk management measures for the study area are proposed through the dominant risk factors.

2.3. Methods of analysis

2.3.1. Construction of the index system

Climate risk assessment is crucial for revealing the spatial heterogeneity caused by differences in the physical environment, land use, and socio-economic activities within urban areas. Identifying vulnerable spatial locations aids in proposing scientifically supported mitigation decisions (Abrar et al., 2022; Chen et al., 2020; Dong et al., 2020; Lim and Skidmore, 2020; Xiang et al., 2022). Extensive research highlights that climate risk assessment depends on four crucial components: (1) Hazard: this aspect refers to the direct impact of disasters on localities, with temperature being a dominant factor in assessing heat-related risks. (2) Exposure: it pertains to the number or value of elements exposed to the impact of catastrophic factors (such as population, roads, infrastructure). (3) Vulnerability: signifying the extent of susceptibility or resistance of a system to damage or harm, this facet notably includes the proportion of vulnerable populations. (4) Adaptation: this includes both mitigation (such as the layout of medical facilities and cooling facilities) and adaptation (related to the level of economic development), both being equally vital aspects. These aspects, which are formed by multiple meteorological or social indices, constitute the guideline dimensions.

Previous research has mainly concentrated on the correlation between the constructed environment and heat exposure at the urban or regional level. Insufficient attention has been given to the block that accommodates people's activities, and there is a lack of fine optimization of the spatial environment from the perspectives of human health and social activities in the face of climate risks. Therefore, this study is based on the HEVA theoretical framework, and the dimensions focus on the interaction system of thermal radiation, human, activity and space.

Combining the meaning of the HEVA in the context of thermal radiation-human-activity-space interaction, RHSI is a comprehensive index that characterizes the intensity and duration of human exposure to thermal radiation, reflecting the interaction between thermal radiation and humans. Exposure is driven by a combination of human spontaneity and social activities. Existing studies primarily focused on population concentration, infrastructure levels, and transportation (Abrar et al., 2022; Dong et al., 2023b; Pouke et al., 2016; Xiang et al., 2022). Vulnerable populations are often concentrated in specific spaces in cities, such as older urban districts (Diaz et al., 2015; Medina-Ramón et al., 2006; Royé et al., 2020). Elderly people and children are often targeted (Medina-Ramón et al., 2006). Information on real-time body heat perception and long-term suffered by outdoor workers or pedestrians in space under sunlight can help characterise broader population vulnerability, which is seldom used in risk assessment due to the lack of easy access (Wang et al., 2021; Xiang et al., 2022). Semantic analysis methods can directly extract information from accessed social media data to capture representations of "hot" topics generated by subjective perceptions (Royé et al., 2020). Adaptability is closely linked to the level of social activity and spatial resilience. Hazard mitigation capacity, including medical care and cooling facilities, and adaptive capacity, such as GDP and land value, were emphasized (Chen et al., 2020; Dong et al., 2020). Based on this initial selection of indices, specific details are presented in Table 2.

To ensure independence between the indices, correlation analysis was used for selecting indices. Since POI data contains more comprehensive information than Weibo check-in data, and older people are more sensitive to heat than women (Thorsson et al., 2014), the Weibo check-in data and female distribution data were mentioned and excluded. Following the principles of representativeness, systematicity, accessibility, and quantifiability, we preserved the widely recognized indices (e.g., population density, population density of older adults and children, GDP) for the exposure, vulnerability, and adaptation dimensions in established studies related to heat risk assessment. Expanded indices of demographic characteristics of vulnerable populations (Weibo heat). We replaced the previous 1ST or T_a for heat risk with the biometeorological index (RHSI) to represent the impact of urban heat stress more accurately on human health. Twelve indices were finally selected for the TIHR assessment system (Table 3).

2.3.2. Calculation of indices

The 3 m pixel level can capture small but significant changes in heat risk, and all indices were harmonized for ease of calculation. A clear and cloudless daytime on July 26, 2021 was selected as a typical scenario for a summer day, which is widely accepted in similar studies (Chen et al., 2016; Lau et al., 2016). The spatial distribution of the mean radiant temperature (T_mrt) in the study area was modeled using SOLWEIG (Solar and Long Wave Environmental Irradiance Geometry). It captures changes in 3D radiative fluxes at a finer spatial and temporal resolution, and effectively evaluates the impacts of urban building geometry and trees on T_mrt. The accuracy of the model has been extensively verified (Aminipouri et al., 2019; Chen et al., 2016; Guo et al., 2023a; Lindberg et al., 2008; Thom et al., 2016). Therefore, the simulation parameter settings, simulation periods, and intervals were referred to the validation study conducted in Dalian (Guo et al., 2023a). Due to the limited number of rasters that can be simulated by SOLWEIG, we adopted the method of Jänicke (2016) to divide all rasters of study areas into six blocks of tiles. Each edge was overlapped by 200 m to avoid boundary effects, resulting in large-scale, highresolution simulations. The spatial variation of T_mrt over time can be simulated by inputting T_a, relative humidity (RH), and global radiation (G) as time series data. The RHSI (CJ can then be calculated using the formulas referenced in Chen et al. (2016).

C₂, C₆, C₇ were all derived from age-specific population data obtained from WorldPop, which were corrected against the Seventh Population Census data to ensure accuracy. C₃, C₆, and C_11-12 were calculated using the POI dataset. C₆ referred to the distribution of kindergartens, primary schools, and secondary schools therein. C₄ came from geospatial data downloaded from Bigemap. Due to the presence of spatial spillover and neighborhood effects in C_2-7, C₁₁, and C₁₂, they were quantified in ArcGIS using the kernel density tool (Liu et al., 2019), while indices C₈, C₉, and C₁₀ made use of spatial interpolation. C₅ was derived from extracting text entries related to heat from massive data, referred to as Weibo heat in this study. To efficiently process extensive text data, we employed Bidirectional Encoder Representations from Transformers (BERT), a language encoder developed by Google in 2018 (Wang et al., 2021). BERT translates input sentences or paragraphs into relevant semantic features, allowing for the accurate extraction of Weibo entries of interest.

2.3.3. Assessment of TIHR

(1) CRITIC Method for Weight Determination

Criteria Importance Though Intercrieria Correlation (CRITIC) is an improvement of the entropy weight method. The core of it lies in considering the contrast and contradiction between indices, calculating the information content contained in each index, and subsequently obtaining the weights (Diakoulaki et al., 1995; Fan et al., 2022). Compared to common assessment methods such as layer stacking and principal component analysis (Aubrecht and Özceylan, 2013; Guo et al., 2023; Johnson et al., 2012; Reid et al., 2009), the CRITIC method provides a clear consideration of the relative importance among various factors. It also better preserves the original data characteristics, exhibiting higher flexibility and precision, and is more suitable for addressing complex and multifaceted issues. The specific steps are as follows:

® Index contrast

The data are standardised using the range method with the following formulas for the positive and negative indices:

... (1)

... (2)

The standard deviation aj is used to indicate the comparability of the j th index:

... (3)

... (4)

@ Index contradiction

Assuming the contradiction magnitude between index j and the remaining indices is denoted as f p

... (5)

... (6)

@ Information carrying capacity

Assuming the information carrying capacity of index j is denoted as C7:

... (7)

Assuming the weighting of index j is denoted as wp

... (8)

(2) Modeling the TIHR assessment system

It is necessary to determine the comprehensive assessment index G, of each NEVA dimensions to construct the TIHR assessment system. The specific steps are as follows:

... (9)

Both addition and subtraction as well as multiplication and division have been widely utilized in existing risk assessment studies (El-Zein and Tonmoy, 2015; Frazier et al., 2014). In this paper, the addition and subtraction method is chosen for calculation. The TIHR formula is as follows:

... (10)

(3) Identify the dominant risk factors in blocks

Based on the risk assessment, the areas are delineated and categorized by dimension stacking through combinatorial relationships for the grade difference of the dominant risk factors (H, E, V, A). This is done in order to guide specific risk prevention strategies (Chang et al., 2013). In instances of high adaptability, if a dimension or multidimensions have a higher level (level V to VII), they are categorized as factor-induced or multifactor-induced for that dimension. If all other dimensions have a lower level, then it is non-existent dominant dimension. Areas with low adaptability, where the remaining dimensions are also low (level I to IV), are classified as adaptive factor-induced. Conversely, areas where all other dimensions have high levels are deemed integrated.

3. Results

3.1. Quantitative values of health risk assessment indices

The results of the calculations for the 12 indices are shown in Fig. 3. The RHSI (C^ for a typical summer day in Suoyuwan ranged from 0 to 113.91 °C h. The wide range of thresholds suggested that the combined effects of heat stress intensity and duration vary greatly across different areas. In detail, the RHSI between short distances were very different, mainly due to the spatial geometry that leads to changes of Tmrt. The Zhongshan Square area in the southeast had fewer areas with ultra-high RHSI compared to other areas due to its large floor area ratio and relatively favorable microclimate. Ultra-high radiative heat stress areas above 84.88 °Ch occurred in the low and medium density urban areas in the western part of the study area, and were more pronounced on roads, squares, and the new urban area, as well as on the south side of buildings.

The trends in population density (C2) and the elderly population density (C7) exhibited consistency. Densely populated areas were predominantly concentrated within the BSHG-lndustrial Park, XAL-trade area, and the economic hub centered around Zhongshan Square. Higher values of Points of Interest density (C3), disaster shelter aggregation (Сц), and medical facilities aggregation (C12) were predominantly situated along economically thriving and accessible areas, such as Huabei-Donglian Road, Southwest Road, XAL-trade area, Anshan Road, Zhongshan Square serving as the core with a diminishing gradient in all directions. The road network density (C4) reached its zenith, attributed to the Xianglujiao Overpass area, a pivotal transportation nexus in the city. The southern part of the study area exhibited superior accessibility in contrast to its northern counterpart on the whole. Weibo heat (C5) exhibited notable fluctuations within a confined range, underscoring the heightened sensitivity of crowd-perceived heat on social media. Its elevated values were principally concentrated around Dingshan Park, XAL-trade area, and adjacent established residential areas. The southeast area registered markedly lower Weibo heat, indicating a high degree of urban resilience within the central city. Child population density (C6) was most pronounced in the vicinity of Anshan Road, an area characterized by dense habitation and excellent educational resources. Construction year (C8) displayed a juxtaposition of both historic and contemporary architectural structures. The spatial distribution of land value (C9) exhibits significant differences, with generally higher values observed near the XAL-Trade area, to the west of Laodong Park, along Anshan Road, and in the northeast part of Diamond Bay. GDP (C10) demonstrated significant variations between the northern and southern areas.

3.2. Quantification of H-E-V-A dimensions

3.2.1. Hazard

Figure 4(a) shows the results of the hazard assessment of 421 blocks, which were classified from low to high in ascending order from level I to level VII using Natural Breaks classification method. The number of blocks counted for each risk level was 46 (level I), 59 (level II), 97 (level III), 78 (level IV), 55 (level V), 46 (level VI), and 40 (level VII), respectively. 33.5% of the blocks were at a medium-high risk. In terms of spatial pattern, there was a significant difference in TIHR between blocks, even among neighboring blocks. This difference can be attributed to the high sensitivity of the selected indicator RHSI, proving the superiority of this index in reflecting heat risk from the human-centered.

3.2.2. Exposure

The spatial distribution of exposure is shown in Fig. 4(b), which demonstrated a general increase from the coast to inland and from north to south. The number of blocks, classified from low to high exposure risk levels, were as follows: 36 blocks (level I), 56 blocks (level II), 47 blocks (level III), 97 blocks (level IV), 81 blocks (level V), 77 blocks (level VI), and 27 blocks (level VII). Among them, the highly exposed area (level VI) and ultra-highly exposed area (level VII) accounted for 27.1% of all the lots, which were mainly located in the XAL-trade area and Zhongshan Square area with dense population activities, welldeveloped infrastructures, and convenient traffic. And they formed a continuous whole south of Anshan Road, indicating that the two areas with the highest urban vitality had a promoting influence on the exposure of the surrounding blocks.

3.2.3. Vulnerability

The overall vulnerability distribution was high in the south and low in the north (Fig. 4(c)). The percentage of vulnerability blocks in ascending order of risk levels were 9.6% (level I), 8% (level II), 12.2% (level III), 28% (level IV), 23.8% (level V), 14.8% (level VI), and 3.6% (level VII). The proportion of high vulnerability areas above level V was 42.2%. These areas were mainly located near Donglian Road and south of Shugang Road, where there was a dense population of vulnerable people. This is highly related to the aging population issue in Dalian.

3.2.4. Adaptability

The adaptability grading is shown in Fig. 4(d). The proportion of blocks, categorized from low to high adaptability risk levels, was outlined as follows: 10.7% (level I), 14.5% (level II), 20.2% (level III), 17.1% (level IV), 17.6% (level V), 12.8% (level VI), and 7.1% (level VII). The overall adaptability was decreasing from the core areas of Zhongshan Square and Donglian Road to the surrounding areas. The strong disaster prevention capability and urban resilience were attributed to the high level of economic development and good medical facilities.

3.3. Thermal-radiation induced health risk

The spatial distribution of TIHR was mapped by integrating hazard, exposure, vulnerability, and adaptability (Fig. 5). The risk decreased from the main urban core built-up area in the south to the sub-core built-up area in the center, and finally to the new urban area in the north. 44.2% of the blocks were at a medium-high risk. The risk level in the south was 1.25 times higher than that in the north, as defined by Northwest Road-Shugang Road, suggesting that there were significant regional differences in the health risks caused by thermal radiation. XAL-trade area had the highest risk value of 1.72, while a newly constructed block in the northern part of the new area had the lowest risk value of 0.56, with the highest value more than three times greater than the lowest value. Comparison of neighboring blocks revealed significant differences in TIHR over short distances. For instance, consider two adjacent blocks: subway station and business district, exhibiting risk values of 1.5 and 1.2, respectively. The higher value is 1.3 times greater than that of the lower value block. That is, individuals traveling through the area would encounter a substantial shift in risk.

Overall, the dominant risk factors were spatially heterogeneous and vary over short distances (Fig. 6). The research area was categorized based on the number of blocks, with the following order: double dimensions (51.1%), triple dimensions (24.8%), single dimensions (17.1%), and combined dimensions (5.2%) (Table 4). Based on the percentage of area was double dimensions (60.8%), single dimensions (19.6%), triple dimensions (16.3%), and combined dimensions (1.5%). According to the area and number of blocks with the dominant risk factor, HA had the largest area (10.2 km2, 35.6%), while EV had the highest number of blocks (101 blocks, 24%).

HA and EV represent "high hazard-low exposure-low vulnerability-low adaptability" and "low hazard-high exposure-high vulnerability-high adaptability", tively, which for 53.1% of the total area and 42.5% of the total number of blocks. It can be assumed that exposure, vulnerability, and adaptability are positively correlated in approximately half of the blocks in the research area, while they are negatively correlated with hazardousness. Figure 6 shows that the southern part of Shuanggang Road was characterized by clusters of EV areas. These areas have high building densities, abundant green spaces, and permeable ground surfaces, which help to reduce heat stress. Moreover, these areas were located in the core area of the city, benefiting from a favorable geographical location, a high economic level, and welldeveloped infrastructures, which contributed to their strong disaster response capacity. However, due to the high population concentration and the frequency of residents' activities, the exposure and vulnerability were high. HA was mainly located in the north of Shuanggang Road. These blocks consisted mainly of residential land, industrial land, and urban renewal land, the area had uncompact buildings, impermeable surfaces, and few trees and other shading facilities, resulting in a high level of heat stress. Additionally, the lack of disaster shelters and medical facilities and the relatively low level of economic development in the area led to high hazard and low adaptability. 4.

Discussion

4.1. Governance strategies based on risk assessment

A comprehensive assessment of TIHR can pinpoint blocks with elevated risk levels, discern their primary causes, and offer precise guidance for implementing refined risk prevention and control measures. However, the map obtained by synthesizing the complex indices cannot directly guide planning and design. Based on the main types of urban construction land in the research area, statistics for TIHR grading were conducted (Fig. 7). It can be found that 67.6%, 59.3%, 53.7%, and 35.4% of the blocks of commercial and service industry facilities land (B), green space and square land (G), public management and public service facilities land (A), and residential land (R) are located in higher-risk areas (Level 5 to Level 7). TIHR associated with industrial land (M) is generally very low and emptiness despite their large land area. Targeted and differentiated measures are proposed according to the characteristics of urban land use types and their dominant risk factors.

Statistics on the causes of higher risk formation in the major construction land types B, G, A, and R show that В has 57.4% of the dominant risk factors as EV and 19.1% as EVA (Fig. 8). They are primarily situated in the Zhongshan Square area and XAL-trade area and must decrease their exposure and vulnerability. Mitigating the pressure of crowded city centers is a proven approach (Estoque et ak, 2020). If deemed essential, the government can achieve population evacuation by decongesting the area and creating a polycentric city. In addition, the concentration of thermal-sensitive outdoor activists or workers leads to a high vulnerability. However, due to a subjective lack of awareness regarding heat-related health risks, coupled with objectively imposed work demands, these individuals exhibit a heightened frequency of activity. Therefore, on one hand, it is necessary to strengthen the efforts in publicizing and educating people about related knowledge. On the other hand, in order to provide pedestrians with a crucial space for energy regulation and heat stress recovery, various protective measures can be implemented. These measures may include the addition of medical facilities and cooling facilities, as well as the installation of awnings, porch shelters, sprinkler systems, air-cooling systems, misters, and greenery.

Urban green spaces offer a multitude of social and ecological benefits, promoting physical activity, social interaction, and recreational pursuits across different age groups. Consequently, they exhibit high exposure and vulnerability characteristics (Fig. 8) (Amani Beni et al., 2023; Khalilnezhad et al., 2023). Simultaneously, they offer extensive ecosystem services, including climate regulation, alleviation of urban heat islands, and enhancement of air quality, making them an indispensable and valuable resource within urban settings (Dehghanifarsani et al., 2023; Dong et al., 2022, 2023a). However, at the human scale, the health risk to individuals is more pronounced due to the form of green spaces, where green areas are more open, and the ground receives relatively direct radiation. 33.3% of the G dominant risk type is HEV, mainly in urban squares and production protection green spaces. Studies have shown that prolonged exposure to outdoor thermal radiation poses significant health risks for pedestrians (Jänicke et ak, 2016; Yang et ak, 2024). Therefore, it is recommended to implement shade measures or necessary tree planting at key nodes for urban squares (Guo et ak, 2023a; Lindberg and Grimmond, 2011; Thorsson et ak, 2011). Approximately 25.9% of G is attributed to EVA, primarily comprising parks or small microgreen spaces. Owing to its size or geographical location, the facility exhibits suboptimal functionality, incomplete supporting amenities, and lower adaptability (Dong et al., 2023a). As a result, more shadow can be added and the service quality of green infrastructure can be upgraded.

The dominant types of medium and high risk for A are EV and HEV, predominantly in hospitals and schools. The clustering of vulnerability is more clearly characterized. For this type of situation, it is possible to arrange the attendance of both older and young people at the clinic or sports activities by scheduling them in different time slots or by setting up specialized venues to avoid heat as much as possible through effective management and coordination among the relevant departments. School playgrounds can enhance ventilation and heat mitigation capabilities by incorporating heat-insulating materials and installing mechanical ventilation devices, etc. R presents a potential risk primarily due to EV and EVA. These blocks are characterized by high population densities and high floor area ratios. Supporting medical assistance facilities and disaster shelters should be provided according to the size of the neighborhood. For neighborhoods populated by the elderly and children, it is important to prioritize the design and modification of outdoor activity areas in order to minimize the risk of discomfort for vulnerable populations.

4.2. Construction of a finer scale TIHR assessment system

In this paper, we developed an index system based on the interactions between "thermal radiation-human-activityspace" with a focus on human-centered orientation, and formed a complete and feasible set of workflows for risk assessment grading and identification of dominant factors in TIHR. Although heat risk assessments have been widely studied, they are not sufficiently closely related to human health, the ongoing impacts of risk occurrence are not considered, and the precision of the data is difficult to guide more detailed planning (Dong et al., 2020; Lim and Skidmore, 2020; Xiang et al., 2022). The RHSI can reflect the risk level of thermal hazards that are closely related to human health in a continuous strong radiation environment. Not only does it fulfill the objective requirement that the RHSI reflects the intensity of disaster occurrence, but its persistence is also interpreted. Firstly, Tmrt depends on the surrounding surface temperature and is good at capturing the spatial differences in the intensity of thermal disasters. Studies have shown that the Tmrt difference between areas in direct sunlight and nearby shaded areas can be as high as 30 °C in the afternoon, while the difference in air temperature is less than 3 °C (Mayer and Höppe, 1987). Secondly, this study quantifies the intensity of heat stress and takes into account the duration of hazardous intensity. RHSI is proposed by Chen et al. (2016), which visually represents the impacts of persistent heat stress on the area and effectively explains the direct relationship between heat and the risk to human health and mortality. Lastly, SOLWEIG has the capability to simulate high-resolution Tmrt using real 2.5-dimensional geographic and meteorological data, which can accurately reflect the influence of spatial patterns and microclimate environment on the risk. Compared to the existing studies, the accuracy and resolution of the data have been greatly improved. It is more convincing to reflect the effects of urban heat stress on people during hot summers (Guo et al., 2023a; Thorsson et al., 2014).

Most of the current research characterizes vulnerable populations based on census data of different age groups, such as the young and the elderly (Medina-Ramón et al., 2006). However, these data ignore differences in perceptions among populations outside of these. The study by Wang et al. (2021) demonstrated that analyzing social media data can reveal the spatial heterogeneity of heat stress caused by heat waves in terms of frequency, intensity, and duration. The real-time social media semantic analysis collected "hot"-related messages posted on Weibo, effectively characterizing the real-time human perception of hot among outdoor workers or pedestrians under heat stress, providing a key vulnerability index for evaluating the health risks. Importantly, this study takes the smallest city block as the research unit, which not only can accurately quantify the TIHR at a human scale, but also helps to provide scientific insights and precise planning guidelines for urban policy makers, designers, and developers.

4.3. Limitations and future research directions

This assessment system is applicable to cities in various climatic contexts. Substitutions and additions to the indices, as well as adjustments to the weightings, are allowed according to climatic characteristics, social environment, urban backdrop, etc., to ensure flexibility and applicability of the workflow in various contexts, taking into account the availability of data. The results of the assessment are valuable for implementing strategies that leverage the attributes and characteristics of the block to enhance the prosperity and liveliness of the neighborhood, as well as to encourage pedestrians to utilize outdoor space appropriately.

There are still some limitations that require further exploration in future studies. Primarily, owing to the complex nature of the risk environment, it is crucial to have a more detailed understanding of the varying degrees of human health impacts from thermal radiation at various time periods. Future studies should contemplate employing indices with time-series variation properties to comprehensively address this aspect. To clarify these distinctions, the variation in hourly heat stress intensity can be illustrated. This method aims to improve the characterization of heat stress dynamics with a more detailed temporal resolution. This requires consideration of the fact that indices associated with other guideline dimensions also have time-series attributes. For instance, mobile signaling data, demographic heat map, activity tracking data, etc., can substitute existing exposure indices, while social media data per unit of time can be counted to reflect vulnerability characteristics. Through these methods, the process of substituting and refining indices is achieved. Additionally, as technology advances and data accessibility increases, the inclusion of richer time-series attribute indices will enhance the scientific aspect of spatio-temporal risk assessments to guide risk response strategies. Secondarily, the simulation scale and efficiency of the SOLWEIG model is somewhat compromised due to the limited number of grids available for a single simulation. This highlights the necessity for future exploration into optimizing algorithmic models, exploring parallel computing techniques, and investigating alternative computational methods. These efforts aim to improve computational performance and efficiency, addressing the requirements for studying the spatial and temporal distribution of thermal radiation across large-scale areas.

4.4. Practical implications

Firstly, our TIHR assessment outcomes provide detailed evidence about the specific locations, spatial distribution, and risk levels within the urban thermal environment that urgently necessitate improvement. This contributes to informed planning and management decisions by relevant government departments and planning agencies. Secondly, the disparities observed in the various dominant risk factors in areas with moderate to high risk highlight the importance of holistic risk mitigation strategies beyond just addressing hazards by implementing cooling measures. Strategies focusing on reducing vulnerability, minimizing exposure, and enhancing adaptability become imperative. This comprehensive perspective yields benefits in tackling the inevitable challenges posed by global warming and urbanization. Furthermore, our assessment framework has implications for other types of disaster risk assessment. The systematic nature of the framework in this study allows the effects of climate, social activities, and physical space on human health risks to be measured in the interconnected domains of hazard, exposure, vulnerability, and adaptation. Lastly, we emphasize the importance of increasing knowledge dissemination about solar radiation risks (Adams et al., 2022). This initiative aims to raise awareness among outdoor workers, planners, vulnerable groups such as the elderly, children, individuals sensitive to heat, and the wider public, enabling timely interventions to mitigate health risks related to radiation.

5. Conclusion

This study offers crucial empirical evidence through practice in the Suoyuwan, Dalian to support finer thermal risk assessment and decision-making, as well as to facilitate a more harmonious interface with planning and management. The study shows that:

(1) The TIHR in Suoyuwan shows a generally declining trend, decreasing from the main urban core in the southern area to the sub-core in the central area and further to the new urban area in the northeast. And the TIHR between blocks differed significantly within a short distance range.

(2) 53.1% of the area and 42.5% of the number of blocks are in НА-EV. It shows that exposure, vulnerability, and adaptability are positively correlated in approximately half of the blocks in the study area, while they are negatively correlated with hazard.

(3) 67.6%, 59.3%, 53.7%, and 35.4% of the blocks of commercial and service industry facilities land (B), green space and square land (G), public management and public service facilities land (A), and residential land (R) are located in higher-risk areas (Level 5 to Level 7). The analysis of urban land use types and their dominant risk factors offers valuable reference for implementing targeted and tailored measures.

Our study focuses on the health risks to pedestrians caused by solar radiation. From the thermal radiationhuman-activity-space interaction perspective, we have developed an assessment system, broadening the current common understanding of heat risk assessment beyond surface or atmospheric temperature. Meanwhile, the identification of dominant risk factors provide valuable insights for developing comprehensive disaster prevention and mitigation strategies from the viewpoints of government management, urban design, and public engagement.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgments

This study was supported by the National Natural Science Foundation of China (Grant No. 52208045).

Sidebar

Peer review under responsibility of Southeast University.

Footnote

* Corresponding author.

*· Corresponding author.

E-mail addresses: [email protected] (F. Guo), [email protected] (J. Dong).

References

References

Abrar, R., Sarkar, S.K., Nishtha, K.T., Talukdar, S., Shahfahad Rahman, A., Islam, A.R.M.T., Mosavi, A., 2022. Assessing the spatial mapping of heat vulnerability under urban heat island (UHI) effect in the dhaka metropolitan area. Sustainability 14, 4945. https://doi.org/10.3390/su14094945.

Adams, R.M., Evans, C., Wolkin, A., Thomas, T., Peek, L., 2022. Social vulnerability and disasters: development and evaluation of a CONVERGE training module for researchers and practitioners. Disaster Prev. Manag. 31, 13-29. https://doi.org/ 10.1108/DPM-04-2021-0131.

Amani Beni, M., Khalilnezhad, M.R., Shen, Z.W., 2023. Iranian citizens' behaviors and attitudes on visitations and appreciation of the urban green space during the COVID-19 epidemic. Acta Hortic. 85-92. https://doi.org/10.17660/ActaHortic.2023.1374.11.

Amani-Beni, M., Tabatabaei Malazı, M., Dehghanian, K., Dehghanifarsani, L., 2023. Investigating the effects of wind loading on three dimensional tree models using numerical simulation with implications for urban design. Sci. Rep. 13, 7277. https://doi.org/10.1038/s41598-023-34071-5.

Aminipouri, M., Knudby, A.J., Krayenhoff, E.S., Zickfeld, K., Middel, A., 2019. Modelling the impact of increased street tree cover on mean radiant temperature across Vancouver's local climate zones. Urban For. Urban Green. 39, 9-17. https: //doi.org/10.1016/j.ufug.2019.01.016.

Aubrecht, C., Özceylan, D., 2013. Identification of heat risk patterns in the U.S. National Capital Region by integrating heat stress and related vulnerability. Environ. Int. 56, 65-77. https: //doi.org/10.1016/j.envint.2013.03.005.

Chang, Q., Liu, D., Liu, X., 2013. Ecological risk assessment and spatial prevention tactic of land destruction in mining city. Trans. Chin. Soc. Agric. Eng. 29, 245-254.

Chen, L., Yu, B., Yang, F., Mayer, H., 2016. 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. 130, 829-842. https: //doi.org/10.1016/j.enbuild.2016.09.014.

Chen, Q., Ding, M., Yang, X., Hu, K., Qi, J., 2018. Spatially explicit assessment of heat health risk by using multi-sensor remote sensing images and socioeconomic data in Yangtze River Delta, China. Int. J. Health Geogr. 17, 1-15. https://doi.org/10.1186/ s12942-018-0135-y.

Chen, S., Zhuang, D., Zhang, H., 2020. GIS-based spatial autocorrelation analysis of housing prices oriented towards a view of spatiotemporal homogeneity and nonstationarity: a case study of guangzhou, China. Complexity 2020, 1-16. https: / /doi.org/10.1155/2020/1079024.

Chen, Y., Amani-Beni, M., Chen, C., Liang, Y., Li, J., Yang, L., 2023. Projection of urban land surface temperature: an inter- and intra-annual modeling approach. Urban Clim. 51, 101637. https: //doi.org/10.1016/j.uclim.2023.101637.

Chong, H., 2012. Building vintage and electricity use: old homes use less electricity in hot weather. European Economic Review, Green Building, the Economy, and Public Policy 56, 906-930. https://doi.Org/10.1016/j.euroecorev.2012.02.009.

Dehghanifarsani, L., Khalilnezhad, M.R., Amani Beni, M., 2023. Birkhauser: Basel, Switzerland 978-3035625240; elSBN: 9783035625271. Land 12, 1483. In: 2023. Book Review: prominski et al. River.Space. Design. Planning Strategies, Methods and Projects for Urban Rivers, 3rd and Enlarged ed. https: //doi.org/10.3390/land12081483

Deng, Y., Liu, J., Liu, Y., Luo, A., 2019. Detecting urban polycentric structure from POI data. ISPRS Int. J. Geo-Inf. 8, 283. https: //doi.org/10.3390/ijgi8060283.

Diakoulaki, D., Mavrotas, G., Papayannakis, L., 1995. Determining objective weights in multiple criteria problems: the critic method. Comput. Oper. Res. 22, 763-770. https: / /doi.org/10.1016/0305-0548(94)00059-H.

Díaz, J., Carmona, R., Mirón, I.J., Ortiz, C., León, I., Linares, C. , 2015. Geographical variation in relative risks associated with heat: update of Spain's Heat Wave Prevention Plan. Environ. Int. 85, 273-283. https://doi.Org/10.1016/j.envint. 2015.09.022.

Dong, J., Guo, F., Lin, M., Zhang, H., Zhu, P., 2022. Optimization of green infrastructure networks based on potential green roof integration in a high-density urban area-a case study of beijing, China. Sei. Total Environ. 834. https://doi.org/10.1016/ j. scitotenv. 2022.155307.

Dong, J., Guo, R., Guo, F., Guo, X., Zhang, Z., 2023a. Pocket parksa a literature review. Environ. Res. Lett. 18, 083003. https: / /doi.org/10.1088/1748-9326/ace7e2.

Dong, Jing, Lin, M., Zuo, J., Lin, T., Liu, J., Sun, C., Luo, J., 2020. Quantitative study on the cooling effect of green roofs in a high-density urban Area-a case study of Xiamen, China. J. Clean. Prod. 255, 120152. https://doi.Org/10.1016/j.jclepro.2020. 120152.

Dong, Jianquan, Peng, J., He, X., Corcoran, J., Qiu, S., Wang, X., 2020. Heatwave-induced human health risk assessment in megacities based on heat stress-social vulnerability-human exposure framework. Landsc. Urban Plann. 203, 103907. https: //doi.org/10.1016/j.landurbplan.2020.103907.

Dong, J., Zhang, B., Guo, F., Guo, R., Zhang, H., 2023b. Potential evaluation for compound use of urban municipal infrastructure land in high-density cities: a case study in Shenzhen, China. IEEE J. Sei. Top. Appl. Earth Obs. Rem. Sens. 1-14. https: //doi.org/10.1109/JSTARS.2023.3312512.

El-Zein, A., Tonmoy, F.N., 2015. Assessment of vulnerability to climate change using a multi-criteria outranking approach with application to heat stress in Sydney. Ecol. Indicat. 48, 207-217. https://doi.Org/10.1016/j.ecolind.2014.08.012.

Estoque, R., Ooba, M., Seposo, X., Togawa, T., Hijioka, Y., Takahashi, К., Nakamura, S., 2020. Heat health risk assessment in Philippine cities using remotely sensed data and social-ecological indicators. Nat. Commun. 11. https://doi.org/ 10.1038/S41467-020-15218-8.

Fan, W., Xu, Z., Wu, В., He, Y., Zhang, Z., 2022. Structural multi-objective topology optimization and application based on the criteria importance through intercriteria correlation method. Eng. Optim. 54, 830-846. https://doi.org/10.1080/0305215X. 2021.1901087.

Fang, Z., Tang, T., Zheng, Z., Zhou, X., Liu, W., Zhang, Y., 2021. Thermal responses of workers during summer: an outdoor investigation of construction sites in South China. Sustain. Cities Soc. 66, 102705. https://doi.Org/10.1016/j.scs.2020.102705.

Frazier, T.G., Thompson, C.M., Dezzani, R.J., 2014. A framework for the development of the SERV model: a Spatially Explicit Resilience-Vulnerability model. Appl. Geogr. 51, 158-172. https://doi.Org/10.1016/j.apgeog.2014.04.004.

Guo, F., Guo, R., Zhang, H., Dong, J., Zhao, J., 2023a. A canopy shading-based approach to heat exposure risk mitigation in small squares. Urban Clim. 49, 101495. https://doi.org/ 10.1016/j.uclim. 2023.101495.

Guo, F., Zhao, J., Zhang, H., Dong, J., Zhu, P., Lau, S.S.Y., 2023b. Effects of urban form on sea cooling capacity under the heatwave. Sustain. Cities Soc. 88, 104271. https://doi.org/10.1016/ j.scs.2022.104271.

Guo, S., Huang, Y., Wang, Z., Wang, Y., Zhang, Y., 2023. Comprehensive Evaluation of Spray Arrangement Strategies for SprayLocal Exhaust Ventilation. Energy and Built Environment. https://doi.org/10.2139/ssrn.4503690.

Hajat, S., Sheridan, S.C., Allen, M.J., Pascal, M., Laaidi, K., Yagouti, A., Bickis, U., Tobias, A., Bourque, D., Armstrong, B.G., Kosatsky, T., 2010. Heat-health warning systems: a comparison of the predictive capacity of different approaches to identifying dangerously hot days. Am. J. Publ. Health 100,1137-1144. https: //doi.org/10.2105/AJPH. 2009.169748.

He, B.-J., Zhao, D., Dong, X., Xiong, K., Feng, C., Qi, Q., Darko, A., Sharifi, A., Pathak, M., 2022a. Perception, physiological and psychological impacts, adaptive awareness and knowledge, and climate justice under urban heat: a study in extremely hothumid Chongqing, China. Sustain. Cities Soc. 79, 103685. https://doi.Org/10.1016/j.scs.2022.103685.

He, B.-J., Zhao, D., Dong, X., Zhao, Z., Li, L., Duo, L., Li, J., 2022b. Will individuals visit hospitals when suffering heat-related illnesses? Yes, but. Build. Environ. 208, 108587. https: / /doi.org/10.1016/j. buildenv.2021.108587.

Hondula, D.M., Davis, R.E., Saha, M.V., Wegner, C.R., Veazey, L.M., 2015. Geographic dimensions of heat-related mortality in seven U.S. cities. Environ. Res. 138, 439-452. https: //doi.org/10.1016/j.envres.2015.02.033.

Huang, Y., Guo, S., Gao, H., Wang, Y., Li, W., Zhang, Y., Wang, Z., 2022. Flow-field characteristics and ventilation performance of the high-temperature buoyant jet controlled by spray-local exhaust ventilation. Build. Environ. 225, 109644. https: / /doi.org/10.1016/j. buildenv.2022.109644.

Jänicke, B., Holtmann, A., Kim, K.R., Kang, M., Fehrenbach, U., Scherer, D., 2019. Quantification and evaluation of intra-urban heat-stress variability in Seoul, Korea. Int. J. Biometeorol. 63, 1-12. https://doi.org/10.1007/s00484-018-1631-2.

Jänicke, B., Meier, F., Lindberg, F., Schubert, S., Scherer, D., 2016. Towards city-wide, building-resolving analysis of mean radiant temperature. Urban Clim. 15, 83-98. https://doi.org/10.1016/ j.uclim.2015.11.003.

Johnson, D.P., Stanforth, A., Lulla, V., Luber, G., 2012. Developing an applied extreme heat vulnerability index utilizing socioeconomic and environmental data. Appl. Geogr. 35. https: //doi.org/W.W16/j.apgeog.2012.04.006.

Khalilnezhad, M.R., Amani Beni, M., Shen, Z.W., 2023. Accessibility, visibility and connectivity between urbanites and edible landscape in the Persian gardens. Acta Hortic. 109-116. https: / /doi.org/10.17660/ActaHortic.2023.1374.14.

Kotharkar, R., Bagade, A., Ramesh, A., 2019. Assessing urban drivers of canopy layer urban heat island: a numerical modeling approach. Landsc. Urban Plann. 190, 103586. https://doi.org/ W.W16/j.landurbplan.2019.05.017.

Lancet, T., 2022. Global heating: an urgent call for action to protect health. Lancet 400, 1557. https://doi.org/10.1016/S01406736(22)02028-1.

Lau, K.K.-L., Ren, C., Ho, J., Ng, E., 2016. Numerical modelling of mean radiant temperature in high-density sub-tropical urban environment. Energy and Buildings, SI: Countermeasures to Urban Heat Island 114, 80-86. https://doi.org/10.1016/ j.enbuild.2015.06.035.

Lim, J., Skidmore, M., 2020. Heat vulnerability and heat island mitigation in the United States. Atmosphere 11, 558. https: / /doi.org/10.3390/atmos11060558.

Lindberg, F., Grimmond, C.S.B., 2011. Nature of vegetation and building morphology characteristics across a city: influence on shadow patterns and mean radiant temperatures in London. Urban Ecosyst. 14, 617-634. https://doi.org/10.1007/s11252011-0184-5.

Lindberg, F., Holmer, B., Thorsson, S., 2008. Solweig 1.0 - modelling spatial variations of 3D radiant fluxes and mean radiant temperature in complex urban settings. Int. J. Biometeorol. 52, 697-713. https: //doi.org/10.1007/s00484-008-0162-7.

Liu, Y., Zhu, A.-X., Wang, J., Li, W., Hu, G., Hu, Y., 2019. Land-use decision support in brownfield redevelopment for urban renewal based on crowdsourced data and a presence-andbackground learning (PBL) method. Land Use Pol. 88, 104188. https://doi.Org/10.1016/j.landusepol.2019.104188.

Martin, A., Julián, A.B.A., Cos-Gayón, F., 2019. Analysis of Twitter messages using big data tools to evaluate and locate the activity in the city of Valencia (Spain). Cities 86, 37-50. https: //doi.org/W.1016/j.cities.2018.12.014.

Mayer, H., Höppe, P., 1987. Thermal comfort of man in different urban environments. Theor. Appl. Climatol. 38, 43-49. https: / /doi.org/10.1007/ BF00866252.

Medina-Ramón, M., Zanobetti, A., Cavanagh, D.P., Schwartz, J.D., 2006. Extreme temperatures and mortality: assessing effect modification by personal characteristics and specific cause of death in a multi-city case-only analysis. Environmental Health Perspectives, https://doi.org/10.1289/ehp.9074.

Morabito, M., Crisci, A., Gioli, B., Gualtieri, G., Toscano, P., Stefano, V.D., Orlandini, S., Gensini, G.F., 2015. Urban-hazard risk analysis: mapping of heat-related risks in the elderly in major Italian cities. PLoS One 10, eO127277. https: //doi.org/10.1371 /journal, pone.0127277.

Pouke, M., Goncalves, J., Ferreira, D., Kostakos, V., 2016. Practical simulation of virtual crowds using points of interest. Comput. Environ. Urban Syst. 57, 118-129. https://doi.org/10.1016/ j.compenvurbsys.2016.02.004.

Prudhomme, C., Crooks, S., Kay, A.L., Reynard, N., 2013. Climate change and river flooding: part 1 classifying the sensitivity of British catchments. Climatic Change 119, 933-948. https: / /doi.org/10.1007/S10584-013-0748-x.

Reíd, C.E., O'Neill, M.S., Gronlund, C.J., Brines, S.J., Brown, D.G., Diez-Roux, A.V., Schwartz, J., 2009. Mapping community determinants of heat vulnerability. Environ. Health Perspect. 117, 1730-1736. https: //doi.org/10.1289/ehp.0900683.

Ren, L., Zhou, T., Zhang, W., 2020. Attribution of the recordbreaking heat event over Northeast Asia in summer 2018: the role of circulation. Environ. Res. Lett. 15, 054018. https: / /doi.org/10.1088/1748-9326/ab8032.

Royé, D., Codesido, R., Tobías, A., Taracido, M., 2020. Heat wave intensity and daily mortality in four of the largest cities of Spain. Environ. Res. 182, 109027. https://doi.Org/10.1016/j.envres. 2019.109027.

Russo, S., Sillmann, J., Sterl, A., 2017. Humid heat waves at different warming levels. Sci. Rep. 7, 7477. https: //doi.org/10.1038/s41598-017-07536-7.

Schmidtlein, M.C., Deutsch, R.C., Piegorsch, W.W., Cutter, S.L., 2008. A sensitivity analysis of the social vulnerability index. Risk Anal. 28, 1099-1114. https://doi.Org/10.1111/j.1539-6924. 2008.01072.x.

Thom, J.K., Coutts, A.M., Broadbent, A.M., Tapper, N.J., 2016. The influence of increasing tree cover on mean radiant temperature across a mixed development suburb in Adelaide, Australia. Urban For. Urban Green. 20, 233-242. https://doi.org/ 10.1016/j.ufug.2016.08.016.

Thorsson, S., Lindberg, F., Björklund, J., Holmer, B., Rayner, D., 2011. Potential changes in outdoor thermal comfort conditions in Gothenburg, Sweden due to climate change: the influence of urban geometry. Int. J. Climatol. 31, 324-335. https: //doi.org/W. 1002/joc.2231.

Thorsson, S., Rocklöv, J., Konarska, J., Lindberg, F., Holmer, B., Dousset, B., Rayner, D., 2014. Mean radiant temperature - a predictor of heat related mortality. Urban Clim. 10, 332-345. https://doi.Org/10.1016/j.uclim.2014.01.004.

Tomlinson, C.J., Chapman, L., Thornes, J.E., Baker, C.J., 2011. Including the urban heat island in spatial heat health risk assessment strategies: a case study for Birmingham, UK. Int. J. Health Geogr. 10,42. https: / /doi .org/10.1186/1476-072X-10-42.

Wang, J., Chen, Y., Tett, S.F.B., Yan, Z., Zhai, P., Feng, J., Xia, J., 2020. Anthropogenically-driven increases in the risks of summertime compound hot extremes. Nat. Commun. 11, 528. https://doi.org/10.1038/s41467-019-14233-8.

Wang, J., Meng, B., Pei, T., Du, Y., Zhang, J., Chen, S., Tian, B., Zhi, G., 2021. Mapping the exposure and sensitivity to heat wave events in China's megacities. Sei. Total Environ. 755, 142734. https://doi.Org/W.W16/j.scitotenv. 2020.142734.

Wang, S., Sun, Q.C., Huang, X., Tao, Y., Dong, C., Das, S., Liu, Y., 2023. Health-integrated heat risk assessment in Australian cities. Environ. Impact Assess. Rev. 102, 107176. https: //doi.org/W. 1016/j.eiar.2023.107176.

Wilhelmi, O.V., Hayden, M.H., 2010. Connecting people and place: a new framework for reducing urban vulnerability to extreme heat. Environ. Res. Lett. 5, 014021. https: //doi.org/W. 1088/1748-9326/5/1/014021.

Wu, X., Liu, Q., Liu, G., Huang, C., Li, H., 2019. Risk assessment of heat waves: a review. Journal of Geo- information Science 21, 1029-1039. https://doi.org/10.12082/dqxxkx.2019.180544.

Xiang, Z., Qin, H., He, B.-J., Han, G., Chen, M., 2022. Heat vulnerability caused by physical and social conditions in a mountainous megacity of Chongqing, China. Sustain. Cities Soc. 80, 103792. https://doi.Org/W.W16/j.scs.2022.W3792.

Yang, H., Xu, Y., Tao, S., Pan, J., Liu, K., Wu, M., 2010. Advances in research on vulnerability and adaptability of warm heat waves. Science and technology guide 28, 98-102.

Yang, L., Yang, H., Yu, B., Lu, Y., Cui, J., Lin, D., 2024. Exploring non-linear and synergistic effects of green spaces on active travel using crowdsourced data and interpretable machine learning. Travel Behaviour and Society 34, 100673. https: //doi.org/W. W16/j.tbs.2O23.100673.

Zhai, W., Bai, X., Shi, Y., Han, Y., Peng, Z.-R., Gu, C., 2019. Beyond Word2vec: an approach forurban functional region extraction and identification by combining Place2vecand POIs. Comput. Environ. Urban Syst. 74, 1-12. https://doi.org/10.1016/j.compenvurbsys.2018.11.008.

Zheng, M., Zhang, J., Shi, L., Zhang, D., Sharma, T.P.P., Prodhan, F.A., 2020. Mapping heat-related risks in northern jiangxi province of China based on two spatial assessment frameworks approaches. IJERPH 17, 1-24. https: / /doi.org/10.3390/ijerph17186584.

Zhou, S., Xie, M., Kwan, M.-P., 2015. Ageing in place and ageing with migration in the transitional context of urban China: a case study of ageing communities in Guangzhou. Habitat Int. 49, 177-186. https://doi.Org/10.1016/j.habitatint.2015.05.022.