Nothing Special   »   [go: up one dir, main page]
Next Article in Journal
Dynamic Response Study of Coral Reef Revetment Project Under Extreme Wave Action
Previous Article in Journal
Strengthening Sustainable Higher Education with Digital Technologies: Development and Validation of a Digital Competence Scale for University Teachers (DCS-UT)
You seem to have javascript disabled. Please note that many of the page functionalities won't work as expected without javascript enabled.
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
Sustainability
Volume 16
Issue 22
10.3390/su16229938
sustainability-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

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

How Urban Street Spatial Composition Affects Land Surface Temperature in Areas with Different Population Densities: A Case Study of Zhengzhou, China

by
Mengze Fu
1,2,
Kangjia Ban
1,
Li Jin
1 and
Di Wu
1,2,*
1
School of Architecture, Zhengzhou University, Zhengzhou 450001, China
2
Henan International Joint Laboratory of Eco-Community & Innovative Technology, Zhengzhou 450001, China
*
Author to whom correspondence should be addressed.
Sustainability 2024, 16(22), 9938; https://doi.org/10.3390/su16229938
Submission received: 29 August 2024 / Revised: 3 November 2024 / Accepted: 12 November 2024 / Published: 14 November 2024
Browse Figures
Figure 1
<p>Study area: (<b>a</b>) location of Zhengzhou in China, (<b>b</b>) locations of the four ring roads in Zhengzhou.</p> "> Figure 2
<p>Study view sampling point.</p> "> Figure 3
<p>Population density.</p> "> Figure 4
<p>Research framework.</p> "> Figure 5
<p>Indicators of the street environment. (<b>a</b>) The green space index refers to the proportion of plant pixels (trees, flowers, grass, etc.) in the image. (<b>b</b>) The tree canopy index refers to the proportion of tree canopy pixels among the plant pixels, representing the vertical structure of greenery along the streets. (<b>c</b>) The sky open index indicates the proportion of sky pixels in the image. (<b>d</b>) The spatial enclosure index is the sum of the proportions of pixels representing buildings, walls, fences, pillars, and other similar elements in the image; appropriate enclosure contributes to ventilation and provides a comfortable feeling. (<b>e</b>) The road width index represents the proportion of pixels for road surfaces, including vehicle lanes and sidewalks, reflecting the relative width of the street. (<b>f</b>) The street walking index refers to the proportion of sidewalk pixels among the road width-related pixels, indicating the relative width of sidewalks in the street.</p> "> Figure 6
<p>Spearman’s correlation analysis results.</p> "> Figure 7
<p>Results of the GAM analysis.</p> "> Figure 8
<p>Distribution of beta values in the GWR model. (<b>a</b>) GSI; (<b>b</b>) TCI; (<b>c</b>) SOI; (<b>d</b>) SEI; (<b>e</b>) RWI; (<b>f</b>) SWI.</p> ">
Versions Notes

Abstract

:
The arrangement and design of urban streets have a profound impact on the thermal conditions within cities, including the mitigation of excessive street land surface temperatures (LSTs). However, previous research has mainly addressed the linear relationships between the physical spatial elements of streets and LST. There has been limited exploration of potential nonlinear relationships and the influence of population density variations. This study explores multi-dimensional street composition indicators obtained from street-view imagery and applies generalized additive models (GAMs) and geographically weighted regression (GWR) to evaluate the indicators’ impact on LST in areas with various population densities. The results indicate the following: (1) The six indicators—green space index (GSI), tree canopy index (TCI), sky open index (SOI), spatial enclosure index (SEI), road width index (RWI), and street walking index (SWI)—all have significant nonlinear effects on summer daytime LST. (2) Among all categories, the GSI negatively affects LST. Moreover, the TCI’s impact on LST shifts from negative to positive as its value increases. The SOI and SWI positively affect LST in all categories. The SEI’s effect on LST changes from negative to positive in the total and high-population (HP) categories, and it remains negative in the low-population (LP) category. The RWI positively affects LST in the total category, shifts from negative to positive in the LP category, and remains negative in the HP category. (3) The influence ranking is GSI > SEI > SWI > SOI > TCI > RWI, with GSI being the most significant factor. These findings provide key insights for mitigating street LSTs through design interventions, contributing to sustainable urban development.

1. Introduction

As the process of global urbanization accelerates, the replacement of natural, permeable surfaces with artificial, impermeable ones has led to a shift in energy flow, precipitating a series of changes in the urban climate [1,2]. The urban heat island (UHI) effect has become a public issue that hinders the sustainable development of cities, posing a severe threat to the ecological environment, human health, and urban energy consumption. For example, heat-related illnesses claimed the lives of 60,000 Europeans in 2022 [3]. Such heatwaves have a profound impact on public health, increasing incidence and mortality rates, with a disproportionate burden on vulnerable societal groups [4,5]. In China, more than 90% of major cities experience diurnal surface urban heat islands (SUHIs), and all major cities exhibit nocturnal urban heat islands [6]. Therefore, it is crucial to rapidly and effectively assess the urban thermal environment and develop scientific mitigation measures for human well-being and the promotion of sustainable urban development.
The urban land surface temperature (LST) is a significant observational indicator of the urban heat island effect and actively influences the formation of this phenomenon [7]. In addition, studies have shown that streets experiencing medical emergencies during high temperatures are correlated with higher surface temperatures [8]. Historically, LST research has transitioned from single- to multi-indicator analyses, moving from two-dimensional (2D) to three-dimensional (3D) morphological considerations [9,10,11,12]. A multitude of studies have corroborated the impact of a 2D spatial configuration on urban climate, such as the cooling effect of vegetation and water body areas on LST [13], as well as the significant influence of land use and land cover (LULC) on LST [14,15,16,17]. Building upon this foundation, additional research incorporating factors such as population density [17,18] and air quality [19] has further elucidated the pivotal role of urban land use in alleviating LST [20,21]. However, due to urban heterogeneity, 3D morphological indicators sometimes provide more compelling insights than 2D indicators [22,23,24]. Specifically, building height is significantly and negatively correlated with surface temperature [25], and semi-enclosed spaces coupled with dense tree cover can mitigate the heat island effect in summer [26]. As vital public spaces within urban areas, streets account for 10% to 20% of a city’s land area [27]. Emissions from the vehicular traffic on these streets are recognized as a key contributor to the urban heat island effect [28]. Meanwhile, as streets represent the most frequently used outdoor environment for residents’ daily activities [29], the LST and thermal conditions of the streets directly affect residents’ health and travel experience. Therefore, regulating the LST and thermal environment of streets holds significant implications for enhancing urban sustainability and mitigating urban heat islands.
Streets form canyons with diverse morphological elements. Previous studies have often focused on certain characteristic morphological elements of streets. For instance, studies have indicated that the sky view factor (SVF) of a street is positively correlated with temperature [30,31]; furthermore the cooling benefits of different leaf area indices of street trees also vary [32,33], and extensive street areas do not significantly contribute to urban heat [34]. A multitude of studies have underscored the potential of urban street design to mitigate heat and augmenting thermal comfort through various measures, such as enhancing the reflectivity of building facades; optimizing street geometry, including orientation and width-to-height ratios; fostering three-dimensional greening; and increasing vegetation along thoroughfares [33,35,36,37,38,39,40]. However, there is a need for a more comprehensive discussion on how street morphology, particularly from the perspective of the street cross-section, influences surface temperature. Despite the use of field surveys, localized imagery, and simulation data, including longitudinal satellite data, research in this field lacks a comprehensive analysis of horizontal street cross-sections using large-sample datasets to further elucidate the physical spatial elements’ impact on LST. As the new data environment transcends the limitations of traditional field measurement collection, the practice of utilizing big data street-view imagery for research is increasingly emerging. Street-view imagery has rapidly ascended as an important data source for geospatial data collection and urban analytics, providing insights and supporting informed decisions [41]. Previous studies have employed algorithms such as the region-based convolutional neural network (R-CNN) [42], the pyramid scene parsing network (PSPNet) [43], and DeepLabV3+ with a deep convolutional neural network [44] to validate the correlation between street-view image measurements and actual data, thereby confirming the usability of street-view data [45]. For example, using street-view images to survey street trees not only simplifies the collection of street tree data but also lays the foundation for urban environmental managers to formulate strategies [46].
Previous research has relied on simple linear correlation analyses to discuss the impact of various urban factors on surface temperature [47,48,49,50,51,52], such as linear regression models [49], principal component analysis [50,52], and Pearson correlation analysis [51]. However, due to the differences in urban substrate materials [40,53], there are marginal effects of vegetation and water bodies [54]. Recent studies have found that the relationship between urban morphological factors and land surface temperature (LST) is nonlinear [54,55]. In the face of the complex relationship between urban morphological factors and LST, there is a need to seek a nonlinear regression modeling technique. The generalized additive model (GAM) can flexibly handle highly nonlinear relationships between response variables and a set of predictors while also having linear prediction capabilities [56]. It combines multiple functions to discover the relationship between LST and predictive factors (environmental and human factors) [57].
This study focuses on the roads within the fourth ring road of Zhengzhou City. Considering the large urban scale, street-view images cover extensive areas of regional data collected at different time points. This temporal discrepancy may lead to spatial heterogeneity in the relationship between street spatial indicators and LST at different locations. Therefore, this study primarily aims to achieve two objectives: (1) The first objective is to explore the nonlinear response relationship between street spatial indicators and surface temperature. By interpreting street-view images, we construct a holistic set of street cross-sectional morphological indicators and include population density as a control variable for grouped observations. (2) The second objective is to use geographically weighted regression (GWR) to explore spatial heterogeneity, thereby more accurately revealing the local relationships at different spatial locations, providing a deeper perspective for understanding the complex relationship between street spatial indicators and surface temperature. The findings of this study are expected to offer insights for city managers and planners to improve the thermal environment of streets through spatial design interventions.

2. Methodology

2.1. Study Area

Situated in the north–central region of Henan Province, Zhengzhou stands as a national central city, as designated by the National Development and Reform Commission. Characterized by a temperate continental monsoon climate, the city experiences dry, chilly winters and rainy, sweltering summers, with an average annual temperature of 22 °C. According to the data published by the Zhengzhou Municipal Bureau of Statistics for the year 2023, the city has a permanent resident population of 12.828 million and an urbanization rate of 79.4% among its permanent residents. As a burgeoning metropolis, Zhengzhou’s urban sprawl and construction growth have been incessantly expanding, a trend that is often intricately linked to the exacerbation of the urban heat island effect [58]. Notably, during the summer of 2019, the city witnessed 19 days of extreme heat in July. Studies have found that the central areas of the city exhibit a more pronounced heat island effect and higher surface temperatures than the peripheral regions [59]. Consequently, the present study delineates the core area of Zhengzhou City, bounded by the fourth ring road (34°39′ N~34°53′ N, 113°31′ E~113°49′ E), as the focal region of investigation. This area encompasses several municipal districts, including Jinshui, Huiji, Erqi, and Zhongyuan (refer to Figure 1).

2.2. Data Sources

To obtain data on the land surface temperature, this study primarily utilized imagery from the Landsat 8 satellite, which has a high spatial resolution of 30 m, suitable for a detailed analysis of street spaces. In contrast, MODIS satellite imagery, with a spatial resolution of 1000 m, is less detailed. The Landsat 8 OLI/TIRS data were sourced from the official USGS website (https://www.usgs.gov). To ensure the stability and reliability of the land surface temperature results, two cloud-free summer images (July to September) were employed, captured on 7 July 2019 and 26 August 2020, both around 11 AM Beijing time. This selection of images under clear sky conditions was intended to provide accurate temperature readings for the analysis.
The road network data for the sampling points were sourced from OpenStreetMap (OSM), with the coordinates converted to the WG1984 geodetic system (refer to Figure 2). By utilizing deep learning techniques in conjunction with street-view imagery (SVI), we constructed descriptive indicators of street spaces. Given the extensive coverage of urban areas in China by Baidu Street View, we employed the Baidu API through Python to retrieve panoramic static images of the study area. By setting parameters such as the field of view, image size, and geographic coordinates, we could obtain street-view images that aligned with our research objectives. The time span of the SVI data ranged from 2017 to 2022, and the data were accessed in June 2023. At each street viewpoint, we collected four orthographic images (0°, 90°, 180°, and 270°). After excluding images from tunnel locations and those not taken during the summer months, we obtained a total of 28,116 images for the study. We employed the PSPNet model to perform computational segmentation on the SVI dataset, from which we extracted six descriptive indices of street spatial characteristics: the green space index (GSI), tree canopy index (TCI), sky open index (SOI), spatial enclosure index (SEI), road width index (RWI), and street walking index (SWI).
The population density data, which serve as an indicator of human activity, were sourced from WorldPop (https://www.worldpop.org/, accessed on 9 June 2023), with a resolution of 100 m (refer to Figure 3). The sources and descriptions of the data utilized in this study are presented in Table 1.

2.3. Research Flow

First, this study employs ENVI 5.3 to invert remote sensing imagery, calculating the two-year average land surface temperature (LST) for the research area. Second, using ArcGIS 10.6, streets within the fourth ring road of Zhengzhou City are sampled every 100 m, yielding a total of 7030 sampling points after excluding locations such as tunnels and underpasses. Concurrently, street-view image data at the sampling points are obtained from the Baidu Map API using Python. Subsequently, semantic segmentation techniques are applied to compute the street morphological parameter values at the sampling points. Additionally, population density is utilized as a control variable, examining the nonlinear relationship between street spatial indicators and LST in a grouped study. Finally, geographically weighted regression (GWR) is employed to explore the local relationships and spatial heterogeneity between street spatial indicators and LST. The overall process is depicted in Figure 4.

2.3.1. Land Surface Temperature Retrieval

Land surface temperature retrieval is conducted based on the radiative transfer equation algorithm, which is widely used in satellite land surface temperature inversion tasks. This algorithm has a solid physical foundation and a high degree of inversion accuracy [14,60,61,62]. L1TP-level images have already undergone geometric correction, so the images obtained can be directly subjected to radiometric calibration and atmospheric correction using ENVI 5.3 software. For the relevant formulas, please refer to [13].

2.3.2. Extraction of Street Space Indicators

The pyramid scene parsing network (PSPNet) is an enhanced semantic segmentation network based on the fully convolutional network (FCN) and is widely used as one of the classic semantic segmentation models [63,64]. When comparing the Cityscapes and ADE20K datasets, which are suitable for a variety of scenes, it was found that using ADE20K as the training set allows for the segmentation images to be converted into 150 foreground labels (such as buildings, trees, and sky) [65,66], enabling a more detailed discrimination of street environmental elements. This study referenced three relatively mature morphological indicators: the green view factor, sky view factor, and tree canopy cover ratio [67,68,69]. Considering existing research that has found that variations in the openness of urban residential areas and public spaces affect plot heat dissipation and ventilation [26,70], we constructed morphological parameters for street spatial enclosure. Vehicle exhaust emissions are a notable contributor to the urban heat island effect [28]. The morphological parameter of road width, reflecting traffic volume to an extent, can thus be utilized to evaluate the influence of vehicular traffic on LST. Aligning with the guidelines for constructing child-friendly cities, we considered the comprehensiveness of the slow traffic system and developed a morphological parameter to measure the proportion of sidewalk space. Finally, six descriptive indicators of street space composition were extracted from the results: the green space index, tree canopy index, sky open index, spatial enclosure index, road width index, and street walking index (Figure 5). The formulas are as follows:
G r e e n   s p a c e   i n d e x = ∑ i = 1 4 G r e e n e r y   p i x e l s i ∑ i = 1 4 T o t a l   p i x e l s i
The “green space index” refers to the average proportion of plant pixels in the images of the sampling points; “i” denotes the number of images, with each sampling point comprising four images; and “greenery pixels” indicates the pixels corresponding to trees, grass, flowers, and other vegetation within a single image.
T r e e   c a n o p y   i n d e x = ∑ i = 1 2 T r e e   p i x e l s i ∑ i = 1 2 G r e e n e r y   p i x e l s i
The “tree canopy index” refers to the average proportion of plant pixels in the images at the sampling points; “i” represents the number of images, with each sampling point comprising two images taken perpendicular to the street orientation (90° and 270°); “tree pixels” denotes the pixels of street tree canopies on both sides of the street within the images; and “greenery pixels” encompasses the pixels of trees, grass, flowers, and other plant life in a single image.
S k y   o p e n   i n d e x = ∑ i = 1 4 S k y   p i x e l s i ∑ i = 1 4 T o t a l   p i x e l s i
The “sky open index” represents the average proportion of sky pixels in the images at the sampling points; “i” indicates the number of images, with each sampling point consisting of four images; and “sky pixels” refers to the pixels depicting the sky in the images.
S p a t i a l   e n c l o s u r e   i n d e x = ∑ i = 1 4 E n c l o s u r e   p i x e l s i ∑ i = 1 4 T o t a l   p i x e l s i
The “spatial enclosure index” denotes the average proportion of pixels that contribute to spatial enclosure within the images at the sampling points; “i” represents the number of images, with each sampling point comprising four images; and “enclosure pixels” refers to the collective pixels in the images that represent elements such as walls, buildings, fences, skyscrapers, poles, and streetlights.
R o a d   w i d t h   i n d e x = ∑ i = 1 2 R o a d   p i x e l s i ∑ i = 1 2 T o t a l   p i x e l s i
The “road width index” refers to the average proportion of pixels representing the width of the road in the images at the sampling points; “i” signifies the number of images, with each sampling point capturing two images parallel to the street orientation (0° and 180°); and “road pixels” encompasses the pixels within the images that pertain to roads, cars, and sidewalks.
S t r e e t   w a l k i n g   i n d e x = ∑ i = 1 2 S i d e w a l k   p i x e l s i ∑ i = 1 2 R o a d   p i x e l s i
The “street walking index” denotes the average proportion of sidewalk pixels in the images at the sampling points; “i” represents the number of images, with two images taken per sampling point, aligned parallel to the street orientation (0° and 180°); and “sidewalk pixels” refers specifically to the pixels within these images that correspond to sidewalk areas.

2.3.3. Data Check

During the initial phase, a descriptive statistical analysis was performed, revealing that the numerical data, with the exception of land surface temperature, did not conform to a normal distribution. The mean and standard deviation were employed to characterize these non-normal datasets. The population density values were sorted in ascending order, with the first 50% designated as the low-population (LP) group and the last 50% designated as the high-population (HP) group. Student’s t-test and the Mann–Whitney U test were applied to assess whether there were significant differences between the groups. Spearman’s rank correlation analysis was utilized to explore the correlation among independent variables. In the literature on the relationship between the built environment and land surface temperature, linear relationships have been extensively discussed using models such as linear regression and multiple linear regression [71,72,73]. Consequently, to investigate the presence of nonlinear relationships between the dependent and independent variables, generalized additive models (GAMs) were employed. These models were computed using the “mgcv” and “ggplot2” packages in R language version 4.3.1.
In the study of urban land surface temperature, the geographical and spatial characteristics are indispensable factors to consider. Therefore, we employed geographically weighted regression (GWR) to specifically investigate the impact of the independent variables on the dependent variable, as well as the spatial heterogeneity. The GWR is a method that uses regression principles to study the quantitative relationships between two or more variables with spatial distribution characteristics. Its distinctive feature is the assumption within the linear regression model that the regression coefficients are dependent on the geographical location of the observation points. This approach allows for a more nuanced understanding of how relationships between variables may vary across space, accounting for the spatial structure and heterogeneity inherent in geographical data.

3. Results

3.1. Preliminary Testing Results of Data

From Supplementary Materials SA, it is evident that the street space indicators do not conform to a normal distribution, log-normal distribution, exponential distribution, or gamma distribution. As shown in Table 2, normally distributed data are described using the mean and standard deviation (SD), while non-normally distributed data are characterized by the median and interquartile range. For normally distributed data, the t-test is applied, and for non-normally distributed data, the Mann–Whitney U test is used. The results indicate that there is no significant difference in the green space index between the low-population (LP) and high-population (HP) groups. However, variables such as the land surface temperature, tree canopy index, sky open index, spatial enclosure index, road width index, and street walking index show significant differences (p < 0.001) between the two groups. The mean of the LST values in the LP group is lower than that in the HP group. This phenomenon demonstrates that the higher the population density, the greater the LST, indicating a positive impact of population density on LST. In summary, population density plays a significant role in the differentiation of the relationships between street composition elements and land surface temperature.
Spearman’s correlation analysis was performed to assess the correlation between all pairs of independent variables. As indicated in Figure 6, there is a significant, strong negative correlation between the SOI and the SEI among the independent variables. The other independent variables do not show multicollinearity, with correlation coefficients below 0.7. Moreover, the correlations between the SOI and the RWI, as well as between the POP and the GSI, are not statistically significant.

3.2. Nonlinear Relationships Between Indices and LST

In Figure 7, the horizontal axis represents the levels of various street composition elements, while the vertical axis indicates the changes in the mean surface temperature, with the shaded areas representing the 95% confidence intervals of the estimated functions. “Total-LST” refers to the relationship between the street composition elements and surface temperature across all 7029 sampling points. The overall results show that the green space index (GSI) and canopy vegetation index (CVI) have a nonlinear negative effect on surface temperature; the sky open index (SOI), road width index (RWI), and street walking index (SWI) exhibit a nonlinear positive effect on surface temperature; and the spatial enclosure index (SEI) has a negative effect on surface temperature when it is less than 0.5 and a positive effect when it is greater than 0.5.
After controlling for the confounding factor of population density, significant nonlinear associations between street composition elements and surface temperature are still observed in both the low-population (LP) and high-population (HP) groups. In the LP group, the GSI, CVI, and SEI all have a nonlinear negative effect on surface temperature, with the SEI’s influence on surface temperature being minimal within the range of 37 °C to 38 °C. In the HP group, the GSI, CVI, and RWI all have a significant nonlinear negative effect on surface temperature. The relationship of the SOI and SWI with surface temperature continues to be primarily positive. The influence trend of the SEI on surface temperature is consistent with the total group, but the surface temperature range that it responds to is reduced to 38 °C to 39 °C.
Generalized additive regression was used to analyze the impact of different indicators, and the results are detailed in Table 3. The p-values for the nonlinear relationships between street morphology elements and LST are less than 0.001, indicating significance. However, the strongest explanatory power of the GSI is only 16.6%, suggesting a weak correlation. This result demonstrates that street morphology parameters have a significant impact on LST, and, from the perspective of overall human well-being, optimizing street space design is beneficial. However, due to the large spatial span of the study area and the collection of street-view images at different time points, geographically weighted regression (GWR) is needed to explore the local relationships and spatial heterogeneity between street morphology parameters and LST at different locations.

3.3. Highest-Impact Index on LST

We conducted a global Moran’s I analysis using ArcGIS 10.6 to explore the spatial autocorrelation. The global Moran’s I value ranges from 1 to −1, where a value greater than 0 indicates a positive spatial autocorrelation, a value less than 0 indicates a negative spatial autocorrelation, and a value of 0 indicates no spatial autocorrelation. As shown in Table 4, both land surface temperature and street descriptive indices exhibit spatial autocorrelations (p < 0.001), with all showing a positive spatial autocorrelation. Among them, the spatial enclosure index has the highest positive spatial autocorrelation, with a value of 0.516342, while the street walking index has the lowest spatial correlation, with a value of 0.185312.
This indicates that there is a significant spatial clustering of factors affecting street space, necessitating the establishment of a spatial model to accurately reflect the impact of each variable on LST. Additionally, due to the large spatial span of the study area and the collection of street-view images at different time points, to further explore the local influence of street shape parameters on LST, we employ the GWR model for analysis. Furthermore, Moran’s I can also be utilized to verify the spatial distribution of residuals in the GWR model, testing whether GWR accurately reflects the relationship between LST and explanatory variables, as well as the characteristics of local spatial non-stationarity. For details, see Supplementary Materials SB.
Table 3 shows that, compared with the GAM results, geographically weighted regression can reduce the time error caused by the spatial span. The green space index exerts the most significant influence on surface temperature, accounting for 72.04% of the variance. This is followed by the spatial enclosure index at 64.44%, while the road width index has the least influence, explaining 41.43% of the variance. The spatial display of the individual factors indicates that the influence on surface temperature decreases in the following order: GSI > SEI > SWI > SOI > TCT > RWI.
Figure 8 displays the spatial distribution of the regression coefficients for the influencing factors in the GWR model, with positive estimates colored red and negative estimates colored blue (for a detailed distribution of values, refer to Supplementary Materials SB Figure S2). When examining Figure 3, it becomes evident that the regression coefficients for the indicators show a positive effect in the central region, suggesting a spatial correlation with areas of higher population density. More specifically, within the street space indicators, the GSI, TCI, and SEI generally exhibit significant cooling effects, yet they exert a positive effect on regions with dense populations (as shown in Figure 8a,b,d). The SOI consistently demonstrates a positive impact on LST (Figure 8c). Both the RWI and SWI, which generally have a positive influence on LST, are associated with better infrastructure in densely populated central areas, and thus, they also contribute to cooling effects (Figure 8e,f).

4. Discussion

4.1. Impact of Street Composition Indicators on LST

This study employs generalized additive models (GAMs) and reveals that the relationship between street composition indices and land surface temperature (LST) exhibits significant nonlinearity, transcending the limitations of traditional linear associations [9]. The findings reveal that, among the various morphological elements of streets, the green space index (GSI) has the most significant potential for reducing surface temperatures. The nonlinear relationship between the GSI and surface temperature indicates that as the GSI value increases, the cooling effect on surface temperature diminishes gradually, highlighting the marginal cooling effect of vegetation on LST [13]. There is a threshold effect for the cooling impact of the tree canopy index (TCI) and sky enclosure index (SEI) on surface temperature; when their values surpass 0.75 and 0.5, respectively, they can lead to an increase in LST. Some studies have noted that semi-enclosed spaces and dense tree cover can create a comfortable thermal environment in summer [26], a result that is closely linked to the cooling effects of wind [74]. Under such circumstances, increasing the TCI and SEI values might be ineffective or could even negatively impact the thermal environment [53,59,75]. Taking into account the impact of human activity intensity [11,76,77], overall, the contribution of the RWI to LST lies in the absorption of solar radiation by the impermeable surfaces of roads [23,78]. In areas with less human activity, roads tend to be narrower, and associated facilities, such as green spaces, which have a cooling effect on LST, gradually improve with an increasing road width [79]. Beyond a certain threshold, however, the heat absorbed by these facilities cannot cool the large surface area of wider roads, leading to faster temperature increases with an increasing road width. Yet areas with a high population density often require higher-grade roads to provide better traffic services [80,81]. Concrete materials with high albedo are cooler than other impervious material surfaces [34]. The SOI and SWI have a significant positive effect on surface temperature, aligning with the majority of studies. During the daytime in summer, solar radiation is the primary heat source, and as the SOI and SWI values increase, the amount of solar radiation received by streets increases, leading to a rise in LST [82].
What makes a street cooler in summer? Studies have shown that increasing the density of visible greenery has a more significant cooling effect on LST than enhancing the configuration of tree canopies [79]. Moreover, enhancing street greenery not only provides physical cooling but also offers a sense of psychological refreshment [35]. Therefore, it can be deduced that creating semi-enclosed streets with high tree coverage and multi-layered greening is the most comfortable option for summer. Naturally, in densely populated areas, using road surfacing materials with high reflectivity is particularly beneficial for summer cooling.

4.2. Further Research Possibilities

First, this study utilizes street-view imagery for data collection, the availability of which has been validated. However, the time span of street-view image data is relatively large compared with that of on-site measurements, which also explains why the generalized additive model (GAM) analysis shows a weaker correlation between street spatial elements and LST in the three control experiments. When exploring the spatial heterogeneity of the street spatial elements’ impact on LST using geographically weighted regression (GWR), the explanatory power of each morphological element indicator is significantly enhanced. Direct measurement data will be leveraged in future research for a more nuanced exploration of the relationship between street spatial elements and LST, with the goal of identifying precise threshold effects. This will inform the development of credible variables for street design guidelines.
Second, this study only uses population density as the control variable for grouping. However, besides the impact of human activities, urban land surface temperature is also strongly affected by different climate zones and land use [21,83]. Future research can consider building multiple scenarios to explore the relationship between street morphology and LST.

5. Conclusions

This study utilizes street-view imagery to construct indicators of the street environment and analyzes the impact of these indicators on land surface temperature in areas with different population densities. Roads within the fourth ring road of Zhengzhou City are selected as samples, and surface temperature datasets from 2019 and 2020 are collected. Through comparisons and statistical analyses, this study reaches the following conclusions:
  • The generalized additive model reveals that the relationship between the composition indicators of the street environment and LST across different groups is significantly nonlinear. The GSI has a negative effect on LST. The influence of the TCI and SEI on land surface temperature follows a “J”-shaped curve pattern, with a longer left side and a shorter right side. Specifically, inflection points emerge when the TCI and SEI values are approximately 0.75 and 0.5, respectively. This suggests that excessively high TCI and SEI values can hinder street ventilation, subsequently shifting their cooling effect on LST to a heating one. Conversely, the SOI and RWI demonstrate a “J”-shaped curve trend with a longer right side, accelerating the temperature rise effect on LST once their values surpass the inflection points. Notably, in the HP group, the RWI exhibits a cooling effect, which is attributed to the improved state of green facilities supporting the streets. Lastly, the SWI exhibits a unidirectional positive effect on LST.
  • Based on the generalized additive model and geographically weighted regression, the impact of various constituent elements on land surface temperature (LST) is ranked as follows: GSI > SEI > SWI > SOI > TCI > RWI. The GSI has the strongest impact on LST at 72.04%, while the RWI has the weakest at 41.43%, indicating that the proportion of green plant space is key to optimizing the land surface temperature of urban roads. Overall, the composition indicators of the street environment contribute significantly to the LST of streets.
Obviously, the influence mechanism of different variables on LST is different. Based on this, the construction of street microclimate models enables researchers to understand the heat status of urban streets more comprehensively, being a basic tool for formulating urban planning strategies. Various factors can be implemented in the design of urban streets to better cope with the survival challenges of an extreme climate; for example, the reasonable enclosure of space can be used to optimize air circulation and heat exchange. Through research and analyses of the relationship between street composition indices and surface temperature, street design can better meet people’s actual needs and later use in the face of rising temperatures.

Supplementary Materials

The following supporting information can be downloaded at: https://www.mdpi.com/article/10.3390/su16229938/s1, Figure S1: Standardized residual results of GWR; Figure S2: Distribution of regression coefficients for various indicators; Table S1: Residual spatial autocorrelation of GWR.

Author Contributions

Conceptualization, M.F.; methodology, K.B.; data curation, L.J. and K.B.; writing—original draft preparation, K.B.; writing—review and editing, M.F. and D.W.; supervision, D.W. All authors have read and agreed to the published version of the manuscript.

Funding

This work was supported by the National Natural Science Foundation of China [grant number 52308084], the China Postdoctoral Science Foundation [grant number 2022M712877], and the Key R&D and Promotion Projects of Henan Province [grant number 222102110125; 232102321078].

Institutional Review Board Statement

Not applicable.

Informed Consent Statement

Not applicable.

Data Availability Statement

The data that support the findings of this study are available from the author upon reasonable request.

Acknowledgments

We sincerely thank the editor and anonymous reviewers for their valuable comments and suggestions to improve the quality of this paper.

Conflicts of Interest

The authors declare no conflicts of interest.

References

  1. Li, X.; Zhou, Y.; Asrar, G.R.; Imhoff, M.; Li, X. The surface urban heat island response to urban expansion: A panel analysis for the conterminous United States. Sci. Total Environ. 2017, 605–606, 426–435. [Google Scholar] [CrossRef] [PubMed]
  2. Schneider, A.; Mertes, C.M. Expansion and growth in Chinese cities, 1978–2010. Environ. Res. Lett. 2014, 9, 024008. [Google Scholar] [CrossRef]
  3. La, V. Hot days or heat waves: A split over how to count heat deaths. Science 2024, 385, 922. [Google Scholar] [CrossRef] [PubMed]
  4. Liu, J.; Qi, J.; Yin, P.; Liu, W.; He, C.; Gao, Y.; Zhou, L.; Zhu, Y.; Kan, H.; Chen, R.; et al. Rising cause-specific mortality risk and burden of compound heatwaves amid climate change. Nat. Clim. Chang. 2024, 14, 1201–1209. [Google Scholar] [CrossRef]
  5. Alho, A.M.; Oliveira, A.P.; Viegas, S.; Nogueira, P. Effect of heatwaves on daily hospital admissions in Portugal, 2000–2018: An observational study. Lancet Planet. Health 2024, 8, e318–e326. [Google Scholar] [CrossRef]
  6. He, B.-J.; Wang, J.; Zhu, J.; Qi, J. Beating the urban heat: Situation, background, impacts and the way forward in China. Renew. Sustain. Energy Rev. 2022, 161, 112350. [Google Scholar] [CrossRef]
  7. Man, S.; Jian, C.; Di, W.; Bangyong, Q.; Shengyang, L. Remote Sensing Monitoring and Analysis of Surface Thermal Environment at City District Scale. Infrared Technol. 2019, 41, 84–91. [Google Scholar]
  8. O’Brien, D.T.; Gridley Msui, B.; Trlica, A.; Wang, J.A.; Shrivastava, A. Urban Heat Islets: Street Segments, Land Surface Temperatures, and Medical Emergencies During Heat Advisories. Am. J. Public Health 2020, 110, e1–e8. [Google Scholar] [CrossRef]
  9. Ge, Z.; Zhen, T.; Qi, W.; Yongqing, S.; Yan, Y.; Haitao, W. Study on the influence of urban form on land surface temperature based on machine learning. Environ. Sci. Technol. 2022, 45, 214–227. [Google Scholar] [CrossRef]
  10. Zhang, Y.; Middel, A.; Turner, B.L. Evaluating the effect of 3D urban form on neighborhood land surface temperature using Google Street View and geographically weighted regression. Landsc. Ecol. 2019, 34, 681–697. [Google Scholar] [CrossRef]
  11. Xingyu, L.; Yaolin, L.; Chaohui, Y.; Zhide, T. Infulence of Urban Physical Form and Human Activities on Thermal Environment at Community Scale. J. Wuhan Univ. Technol. 2020, 42, 72–79. [Google Scholar]
  12. Yuting, C.; Zhaohua, L.; Linlin, L.; Shibiao, L.; Qingting, L. Spatio-temporal Dynamics of Surface Urban Heat Island in Coastal Mega Cities along the Belt and Road from Remote Sensing Data. Remote Sens. Technol. Appl. 2020, 35, 1197–1205. [Google Scholar]
  13. Yang, L.; Yu, K.; Ai, J.; Liu, Y.; Yang, W.; Liu, J. Dominant Factors and Spatial Heterogeneity of Land Surface Temperatures in Urban Areas: A Case Study in Fuzhou, China. Remote Sens. 2022, 14, 1266. [Google Scholar] [CrossRef]
  14. Liu, F.; Zhang, X.; Murayama, Y.; Morimoto, T. Impacts of Land Cover/Use on the Urban Thermal Environment: A Comparative Study of 10 Megacities in China. Remote Sens. 2020, 12, 307. [Google Scholar] [CrossRef]
  15. Zhang, Y.; Sun, L. Spatial-temporal impacts of urban land use land cover on land surface temperature: Case studies of two Canadian urban areas. Int. J. Appl. Earth Obs. Geoinf. 2019, 75, 171–181. [Google Scholar] [CrossRef]
  16. Nurwanda, A.; Honjo, T. The prediction of city expansion and land surface temperature in Bogor City, Indonesia. Sustain. Cities Soc. 2020, 52, 101772. [Google Scholar] [CrossRef]
  17. Li, L.; Tan, Y.; Ying, S.; Yu, Z.; Li, Z.; Lan, H. Impact of land cover and population density on land surface temperature: Case study in Wuhan, China. J. Appl. Remote Sens. 2014, 8, 084993. [Google Scholar] [CrossRef]
  18. Ramírez-Aguilar, E.A.; Lucas Souza, L.C. Urban form and population density: Influences on Urban Heat Island intensities in Bogotá, Colombia. Urban Clim. 2019, 29, 100497. [Google Scholar] [CrossRef]
  19. Talukdar, S.; Ratnam, M.V. Mutual response hypothesis between surface temperature and aerosol (BC and non-BC) mass concentration observed in an urban environment. Sci. Total Environ. 2023, 887, 164048. [Google Scholar] [CrossRef]
  20. Zhou, W.; Huang, G.; Cadenasso, M.L. Does spatial configuration matter? Understanding the effects of land cover pattern on land surface temperature in urban landscapes. Landsc. Urban Plan. 2011, 102, 54–63. [Google Scholar] [CrossRef]
  21. Dong, R.; Wurm, M.; Taubenbock, H. Seasonal and Diurnal Variation of Land Surface Temperature Distribution and Its Relation to Land Use/Land Cover Patterns. Int. J. Environ. Res. Public Health 2022, 19, 12738. [Google Scholar] [CrossRef] [PubMed]
  22. Wu, W.-B.; Yu, Z.-W.; Ma, J.; Zhao, B. Quantifying the influence of 2D and 3D urban morphology on the thermal environment across climatic zones. Landsc. Urban Plan. 2022, 226, 104499. [Google Scholar] [CrossRef]
  23. Luo, P.; Yu, B.; Li, P.; Liang, P.; Liang, Y.; Yang, L. How 2D and 3D built environments impact urban surface temperature under extreme heat: A study in Chengdu, China. Build. Environ. 2023, 231, 110035. [Google Scholar] [CrossRef]
  24. Chen, J.; Zhan, W.; Jin, S.; Han, W.; Du, P.; Xia, J.; Lai, J.; Li, J.; Liu, Z.; Li, L.; et al. Separate and combined impacts of building and tree on urban thermal environment from two- and three-dimensional perspectives. Build. Environ. 2021, 194, 107650. [Google Scholar] [CrossRef]
  25. Liu, J.; Zhang, S.; Yin, L.; Li, B.; Zhang, B. Exploring the relationship between seasonal variations of land surface temperature and urban morphological factors in complex urban areas. Environ. Sci. Pollut. Res. 2023, 30, 59861–59876. [Google Scholar] [CrossRef]
  26. Yang, F.; Lau, S.S.Y.; Qian, F. Summertime heat island intensities in three high-rise housing quarters in inner-city Shanghai China: Building layout, density and greenery. Build. Environ. 2010, 45, 115–134. [Google Scholar] [CrossRef]
  27. Tingting, Y.; Huiyun, P.; Shaohua, T.; Mingying, Z. From Urban Stress to Neuro-Urbanization: Review and Framing of Healthy Street Research from an Interdisciplinary Perspective. Urban Plan. Int. 2023, 38, 29–38. [Google Scholar] [CrossRef]
  28. Shikwambana, L.; Kganyago, M.; Mhangara, P. Temporal Analysis of Changes in Anthropogenic Emissions and Urban Heat Islands during COVID-19 Restrictions in Gauteng Province, South Africa. Aerosol Air Qual. Res. 2021, 21, 200437. [Google Scholar] [CrossRef]
  29. Xiaohui, G.; Zhiyi, B.; Fan, W.; Shimin, Y.; Hai, Y. Influence of Street View Factors on Microclimate and Thermal Comfort of Urban Street Canyons in Summer Afternoon. Chin. Landsc. Archit. 2021, 37, 71–76. [Google Scholar] [CrossRef]
  30. Jiang, Y.; Han, X.; Shi, T.; Song, D. Microclimatic Impact Analysis of Multi-Dimensional Indicators of Streetscape Fabric in the Medium Spatial Zone. Int. J. Environ. Res. Public Health 2019, 16, 952. [Google Scholar] [CrossRef]
  31. Li, G.; Ren, Z.; Zhan, C. Sky View Factor-based correlation of landscape morphology and the thermal environment of street canyons: A case study of Harbin, China. Build. Environ. 2020, 169, 106587. [Google Scholar] [CrossRef]
  32. Sanusi, R.; Johnstone, D.; May, P.; Livesley, S.J. Microclimate benefits that different street tree species provide to sidewalk pedestrians relate to differences in Plant Area Index. Landsc. Urban Plan. 2017, 157, 502–511. [Google Scholar] [CrossRef]
  33. Coutts, A.M.; White, E.C.; Tapper, N.J.; Beringer, J.; Livesley, S.J. Temperature and human thermal comfort effects of street trees across three contrasting street canyon environments. Theor. Appl. Climatol. 2015, 124, 55–68. [Google Scholar] [CrossRef]
  34. Engel, R.A.; Millard-Ball, A.; Turner, V.K. Contributions of roads to surface temperature: Evidence from Southern California. Environ. Res. Commun. 2023, 5, 015004. [Google Scholar] [CrossRef]
  35. Klemm, W.; Heusinkveld, B.G.; Lenzholzer, S.; van Hove, B. Street greenery and its physical and psychological impact on thermal comfort. Landsc. Urban Plan. 2015, 138, 87–98. [Google Scholar] [CrossRef]
  36. Segura, R.; Krayenhoff, E.S.; Martilli, A.; Badia, A.; Estruch, C.; Ventura, S.; Villalba, G. How do street trees affect urban temperatures and radiation exchange? Observations and numerical evaluation in a highly compact city. Urban Clim. 2022, 46, 101288. [Google Scholar] [CrossRef]
  37. Tan, Z.; Dong, J.; Xiao, Y.; Tu, J. A numerical study of diurnally varying surface temperature on flow patterns and pollutant dispersion in street canyons. Atmos. Environ. 2015, 104, 217–227. [Google Scholar] [CrossRef]
  38. Gillner, S.; Vogt, J.; Tharang, A.; Dettmann, S.; Roloff, A. Role of street trees in mitigating effects of heat and drought at highly sealed urban sites. Landsc. Urban Plan. 2015, 143, 33–42. [Google Scholar] [CrossRef]
  39. Qin, Y.; Wei, P.; Wang, S.J.; Tan, K. Experimental study on the thermal characteristics of urban mockups with different paved streets. Environ. Sci. Pollut. Res. Int. 2021, 28, 63008–63016. [Google Scholar] [CrossRef]
  40. Zhang, Y.; Wei, P.; Wang, L.; Qin, Y. Temperature of Paved Streets in Urban Mockups and Its Implication of Reflective Cool Pavements. Atmosphere 2021, 12, 560. [Google Scholar] [CrossRef]
  41. Biljecki, F.; Ito, K. Street view imagery in urban analytics and GIS: A review. Landsc. Urban Plan. 2021, 215, 104217. [Google Scholar] [CrossRef]
  42. Xu, Z.; Zhang, F.; Wu, Y.; Yang, Y.; Wu, Y. Building height calculation for an urban area based on street view images and deep learning. Comput.-Aided Civ. Infrastruct. Eng. 2022, 38, 892–906. [Google Scholar] [CrossRef]
  43. Gong, F.-Y.; Zeng, Z.-C.; Zhang, F.; Li, X.; Ng, E.; Norford, L.K. Mapping sky, tree, and building view factors of street canyons in a high-density urban environment. Build. Environ. 2018, 134, 155–167. [Google Scholar] [CrossRef]
  44. Deng, M.; Yang, W.; Chen, C.; Wu, Z.; Liu, Y.; Xiang, C. Street-level solar radiation mapping and patterns profiling using Baidu Street View images. Sustain. Cities Soc. 2021, 75, 103289. [Google Scholar] [CrossRef]
  45. Hou, Y.; Biljecki, F. A comprehensive framework for evaluating the quality of street view imagery. Int. J. Appl. Earth Obs. Geoinf. 2022, 115, 103094. [Google Scholar] [CrossRef]
  46. Hu, Y.; Wang, H.; Yan, H.; Han, Q.; Nan, X.; Zhao, K.; Bao, Z. Alternative scenarios for urban tree surveys: Investigating the species, structures, and diversities of street trees using street view imagery. Sci. Total Environ. 2023, 895, 165157. [Google Scholar] [CrossRef]
  47. Rahman, M.T. Examining and Modelling the Determinants of the Rising Land Surface Temperatures in Arabian Desert Cities: An Example from Riyadh, Saudi Arabia. J. Settl. Spat. Plan. 2018, 9, 1–10. [Google Scholar] [CrossRef]
  48. Peng, J.; Jia, J.; Liu, Y.; Li, H.; Wu, J. Seasonal contrast of the dominant factors for spatial distribution of land surface temperature in urban areas. Remote Sens. Environ. 2018, 215, 255–267. [Google Scholar] [CrossRef]
  49. Park, J.; Kim, J.-H.; Lee, D.K.; Park, C.Y.; Jeong, S.G. The influence of small green space type and structure at the street level on urban heat island mitigation. Urban For. Urban Green. 2017, 21, 203–212. [Google Scholar] [CrossRef]
  50. Firozjaei, M.K.; Alavipanah, S.K.; Liu, H.; Sedighi, A.; Mijani, N.; Kiavarz, M.; Weng, Q. A PCA–OLS Model for Assessing the Impact of Surface Biophysical Parameters on Land Surface Temperature Variations. Remote Sens. 2019, 11, 2094. [Google Scholar] [CrossRef]
  51. Lu, L.; Weng, Q.; Xiao, D.; Guo, H.; Li, Q.; Hui, W. Spatiotemporal Variation of Surface Urban Heat Islands in Relation to Land Cover Composition and Configuration: A Multi-Scale Case Study of Xi’an, China. Remote Sens. 2020, 12, 2713. [Google Scholar] [CrossRef]
  52. Soltanifard, H.; Kashki, A.; Karami, M. Analysis of spatially varying relationships between urban environment factors and land surface temperature in Mashhad city, Iran. Egypt. J. Remote Sens. Space Sci. 2022, 25, 987–999. [Google Scholar] [CrossRef]
  53. Siqi, J.; Yuhong, W.; Ling, C.; Xiaowen, B. A novel approach to estimating urban land surface temperature by the combination of geographically weighted regression and deep neural network models. Urban Clim. 2023, 47, 101390. [Google Scholar] [CrossRef]
  54. Wu, Z.; Qiao, R.; Zhao, S.; Liu, X.; Gao, S.; Liu, Z.; Ao, X.; Zhou, S.; Wang, Z.; Jiang, Q. Nonlinear forces in urban thermal environment using Bayesian optimization-based ensemble learning. Sci. Total Environ. 2022, 838, 156348. [Google Scholar] [CrossRef] [PubMed]
  55. Gu, X.; Wu, Z.; Liu, X.; Qiao, R.; Jiang, Q. Exploring the Nonlinear Interplay between Urban Morphology and Nighttime Thermal Environment. Sustain. Cities Soc. 2024, 101, 105176. [Google Scholar] [CrossRef]
  56. Liu, X.; Huang, J.; Zhang, Y.; Li, H. Analysis of influencing factors on forest soil organic matter in Guangdong Province based on GAM model. Chin. J. Ecol. 2022, 41, 2278–2288. [Google Scholar]
  57. Shafizadeh-Moghadam, H.; Weng, Q.; Liu, H.; Valavi, R. Modeling the spatial variation of urban land surface temperature in relation to environmental and anthropogenic factors: A case study of Tehran, Iran. GIScience Remote Sens. 2020, 57, 483–496. [Google Scholar] [CrossRef]
  58. Yu, W.; Shi, J.; Fang, Y.; Xiang, A.; Li, X.; Hu, C.; Ma, M. Exploration of urbanization characteristics and their effect on the urban thermal environment in Chengdu, China. Build. Environ. 2022, 219, 109150. [Google Scholar] [CrossRef]
  59. Gao, Y.; Zhao, J.; Han, L. Exploring the spatial heterogeneity of urban heat island effect and its relationship to block morphology with the geographically weighted regression model. Sustain. Cities Soc. 2022, 76, 103431. [Google Scholar] [CrossRef]
  60. Duan, S.; Ru, C.; Li, Z.; Wang, M.; Xu, H.; Li, H.; Wu, P.; Zhan, W.; Zhou, J.; Zhao, W.; et al. Reviews of methods for land surface temperature retrieval from Landsat thermal infrared data. Natl. Remote Sens. Bull. 2021, 25, 1591–1617. [Google Scholar] [CrossRef]
  61. Yu, X.; Guo, X.; Wu, Z. Land Surface Temperature Retrieval from Landsat 8 TIRS—Comparison between Radiative Transfer Equation-Based Method, Split Window Algorithm and Single Channel Method. Remote Sens. 2014, 6, 9829–9852. [Google Scholar] [CrossRef]
  62. Yingshu, W.; Peiji, S.; Wusheng, Z.; Ziyang, W.; Xiaoyi, X. Spatial-temporal characteristics and influencing factors of thermal environment in Lanzhou. Ecol. Sci. 2022, 41, 59–65. [Google Scholar] [CrossRef]
  63. Zhou, J.; Hao, M.; Zhang, D.; Zou, P.; Zhang, W. Fusion PSPnet Image Segmentation Based Method for Multi-Focus Image Fusion. IEEE Photonics J. 2019, 11, 6501412. [Google Scholar] [CrossRef]
  64. Yuan, W.; Wang, J.; Xu, W. Shift Pooling PSPNet: Rethinking PSPNet for Building Extraction in Remote Sensing Images from Entire Local Feature Pooling. Remote Sens. 2022, 14, 4889. [Google Scholar] [CrossRef]
  65. Zhou, B.; Zhao, H.; Puig, X.; Xiao, T.; Fidler, S.; Barriuso, A.; Torralba, A. Semantic Understanding of Scenes Through the ADE20K Dataset. Int. J. Comput. Vis. 2018, 127, 302–321. [Google Scholar] [CrossRef]
  66. He, D.; Miao, J.; Lu, Y.; Song, Y.; Chen, L.; Liu, Y. Urban greenery mitigates the negative effect of urban density on older adults’ life satisfaction: Evidence from Shanghai, China. Cities 2022, 124, 103607. [Google Scholar] [CrossRef]
  67. Yan, H.; Wu, F.; Nan, X.; Han, Q.; Shao, F.; Bao, Z. Influence of view factors on intra-urban air temperature and thermal comfort variability in a temperate city. Sci. Total Environ. 2022, 841, 156720. [Google Scholar] [CrossRef]
  68. Erdem Okumus, D.; Terzi, F. Evaluating the role of urban fabric on surface urban heat island: The case of Istanbul. Sustain. Cities Soc. 2021, 73, 103128. [Google Scholar] [CrossRef]
  69. Hu, Y.; Qian, F.; Yan, H.; Middel, A.; Wu, R.; Zhu, M.; Han, Q.; Zhao, K.; Wang, H.; Shao, F.; et al. Which street is hotter? Street morphology may hold clues-thermal environment mapping based on street view imagery. Build. Environ. 2024, 262, 111838. [Google Scholar] [CrossRef]
  70. Jin, Y.; Jin, H.; Kang, J.; Wong Nyuk, H. Field Study of Thermal Environment and Thermal Comfort in Public Spaces in Urban Residential Areas of Severe Cold Regions. Build. Sci. 2021, 37, 29–37. [Google Scholar]
  71. Sun, Y.; Gao, C.; Li, J.; Li, W.; Ma, R. Examining urban thermal environment dynamics and relations to biophysical composition and configuration and socio-economic factors: A case study of the Shanghai metropolitan region. Sustain. Cities Soc. 2018, 40, 284–295. [Google Scholar] [CrossRef]
  72. Feng, X.; Myint, S.W. Exploring the effect of neighboring land cover pattern on land surface temperature of central building objects. Build. Environ. 2016, 95, 346–354. [Google Scholar] [CrossRef]
  73. Jenerette, G.D.; Harlan, S.L.; Buyantuev, A.; Stefanov, W.L.; Declet-Barreto, J.; Ruddell, B.L.; Myint, S.W.; Kaplan, S.; Li, X. Micro-scale urban surface temperatures are related to land-cover features and residential heat related health impacts in Phoenix, AZ USA. Landsc. Ecol. 2015, 31, 745–760. [Google Scholar] [CrossRef]
  74. He, B.-J.; Ding, L.; Prasad, D. Outdoor thermal environment of an open space under sea breeze: A mobile experience in a coastal city of Sydney, Australia. Urban Clim. 2020, 31, 100567. [Google Scholar] [CrossRef]
  75. Guo, F.; Zhao, J.; Zhang, H.; Dong, J.; Zhu, P.; Lau, S.S.Y. Effects of urban form on sea cooling capacity under the heatwave. Sustain. Cities Soc. 2023, 88, 104271. [Google Scholar] [CrossRef]
  76. Xinbo, S.; Gen, L.; Dongpo, L.; He, H.; Jun, G. Relationship between urban spatial morphology factors and land surface temperature in summer: A case of the central district of Tianjin. Sci. Geogr. Sin. 2023, 43, 360–369. [Google Scholar] [CrossRef]
  77. Chen, Y.; Yang, J.; Yu, W.; Ren, J.; Xiao, X.; Xia, J.C. Relationship between urban spatial form and seasonal land surface temperature under different grid scales. Sustain. Cities Soc. 2023, 89, 104374. [Google Scholar] [CrossRef]
  78. Neog, R.; Acharjee, S.; Hazarika, J. Spatiotemporal analysis of road surface temperature (RST) and building wall temperature (BWT) and its relation to the traffic volume at Jorhat urban environment, India. Environ. Dev. Sustain. 2020, 23, 10080–10092. [Google Scholar] [CrossRef]
  79. Bartesaghi-Koc, C.; Osmond, P.; Peters, A. Innovative use of spatial regression models to predict the effects of green infrastructure on land surface temperatures. Energy Build. 2022, 254, 111564. [Google Scholar] [CrossRef]
  80. Gennian, S. Statistics Analysis on the Relationship of Highway Net Density with Population Density and Per Capita GNP in National Region. J. Southwest Jiaotong Univ. 2000, 2, 220–223. [Google Scholar]
  81. Fengbo, L.; Kewei, Z. Study of Spatial Pattern and Interaction Relationship betweenTraffic Dominance and Population Mobility. Highway 2023, 68, 243–252. [Google Scholar]
  82. Feng, Y.; Chen, L.; He, X. Sky view factor calculation based on Baidu street view images and its application in urban heat island study. J. Geo-Inf. Sci. 2021, 23, 1998–2012. [Google Scholar]
  83. Zhang, Y.; Li, D.; Liu, L.; Liang, Z.; Shen, J.; Wei, F.; Li, S. Spatiotemporal Characteristics of the Surface Urban Heat Island and Its Driving Factors Based on Local Climate Zones and Population in Beijing, China. Atmosphere 2021, 12, 1271. [Google Scholar] [CrossRef]
Sustainability 16 09938 g001
Figure 1. Study area: (a) location of Zhengzhou in China, (b) locations of the four ring roads in Zhengzhou.
Figure 1. Study area: (a) location of Zhengzhou in China, (b) locations of the four ring roads in Zhengzhou.
Sustainability 16 09938 g001
Sustainability 16 09938 g002
Figure 2. Study view sampling point.
Figure 2. Study view sampling point.
Sustainability 16 09938 g002
Sustainability 16 09938 g003
Figure 3. Population density.
Figure 3. Population density.
Sustainability 16 09938 g003
Sustainability 16 09938 g004
Figure 4. Research framework.
Figure 4. Research framework.
Sustainability 16 09938 g004
Sustainability 16 09938 g005
Figure 5. Indicators of the street environment. (a) The green space index refers to the proportion of plant pixels (trees, flowers, grass, etc.) in the image. (b) The tree canopy index refers to the proportion of tree canopy pixels among the plant pixels, representing the vertical structure of greenery along the streets. (c) The sky open index indicates the proportion of sky pixels in the image. (d) The spatial enclosure index is the sum of the proportions of pixels representing buildings, walls, fences, pillars, and other similar elements in the image; appropriate enclosure contributes to ventilation and provides a comfortable feeling. (e) The road width index represents the proportion of pixels for road surfaces, including vehicle lanes and sidewalks, reflecting the relative width of the street. (f) The street walking index refers to the proportion of sidewalk pixels among the road width-related pixels, indicating the relative width of sidewalks in the street.
Figure 5. Indicators of the street environment. (a) The green space index refers to the proportion of plant pixels (trees, flowers, grass, etc.) in the image. (b) The tree canopy index refers to the proportion of tree canopy pixels among the plant pixels, representing the vertical structure of greenery along the streets. (c) The sky open index indicates the proportion of sky pixels in the image. (d) The spatial enclosure index is the sum of the proportions of pixels representing buildings, walls, fences, pillars, and other similar elements in the image; appropriate enclosure contributes to ventilation and provides a comfortable feeling. (e) The road width index represents the proportion of pixels for road surfaces, including vehicle lanes and sidewalks, reflecting the relative width of the street. (f) The street walking index refers to the proportion of sidewalk pixels among the road width-related pixels, indicating the relative width of sidewalks in the street.
Sustainability 16 09938 g005
Sustainability 16 09938 g006
Figure 6. Spearman’s correlation analysis results.
Figure 6. Spearman’s correlation analysis results.
Sustainability 16 09938 g006
Sustainability 16 09938 g007
Figure 7. Results of the GAM analysis.
Figure 7. Results of the GAM analysis.
Sustainability 16 09938 g007
Sustainability 16 09938 g008
Figure 8. Distribution of beta values in the GWR model. (a) GSI; (b) TCI; (c) SOI; (d) SEI; (e) RWI; (f) SWI.
Figure 8. Distribution of beta values in the GWR model. (a) GSI; (b) TCI; (c) SOI; (d) SEI; (e) RWI; (f) SWI.
Sustainability 16 09938 g008
Table 1. Data source and descriptions.
Table 1. Data source and descriptions.
Date SourceDateSpatial ResolutionDescription
Landsat8 OLI/TIRS7 July 201930 mData for retrieving LST
26 August 2020
Street-view imagery2020100 mVariable extraction
Population density2020100 mStudy area population extraction
Open street map2020-Study point extraction
Table 2. Distribution of selected variables in the study area.
Table 2. Distribution of selected variables in the study area.
VariablesLP (n = 3515)HP (n = 3514)Total (n = 7029)p-Values
LST (°C, mean ± SD)37.838, 1.9838.267, 1.84938.053, 1.927<0.001 a
Green space index (median, IQR)0.1582, 0.22610.1684, 0.250.1639, 0.2390.925
Tree canopy index (median, IQR)0.7828, 0.360.8394, 0.43240.8048, 0.4014<0.001 b
Sky open index (median, IQR)0.4157, 0.23290.2909, 0.27140.3519, 0.2688<0.001 b
Spatial enclosure index (median, IQR)0.2592, 0.21790.3797, 0.27960.3158, 0.264<0.001 b
Road width index (median, IQR)0.2925, 0.07990.3007, 0.07080.2967, 0.0742<0.001 b
Street walking index (median, IQR)0.0147, 0.0950.0384, 0.13640.0266, 0.1167<0.001 b
a Student’s t-test was used to compare normally distributed continuous variables among the groups. b The Mann–Whitney U test was used to compare non-normally distributed variables among the groups.
Table 3. Geographically weighted regression.
Table 3. Geographically weighted regression.
VariablesGWR-EVGAM-EV
Green space index72.04%16.60%
Tree canopy index52.12%9.32%
Sky open index55.45%1.56%
Spatial enclosure index64.44%2.52%
Road width index41.43%0.58%
Street walking index58.06%0.5%
Note: GWR-EV and GAM-EV denote the explained variance of each parameter in the model.
Table 4. Global Moran’s I.
Table 4. Global Moran’s I.
VariablesMoran’s IZ-Scorep-Values
LST0.485145.1657<0.01
Green space index0.3556106.4326<0.01
Tree canopy index0.204161.1253<0.01
Sky open index0.4726141.4299<0.01
Spatial enclosure index0.5163154.5318<0.01
Road width index0.236770.8661<0.01
Street walking index0.185355.5262<0.01
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Fu, M.; Ban, K.; Jin, L.; Wu, D. How Urban Street Spatial Composition Affects Land Surface Temperature in Areas with Different Population Densities: A Case Study of Zhengzhou, China. Sustainability 2024, 16, 9938. https://doi.org/10.3390/su16229938

AMA Style

Fu M, Ban K, Jin L, Wu D. How Urban Street Spatial Composition Affects Land Surface Temperature in Areas with Different Population Densities: A Case Study of Zhengzhou, China. Sustainability. 2024; 16(22):9938. https://doi.org/10.3390/su16229938

Chicago/Turabian Style

Fu, Mengze, Kangjia Ban, Li Jin, and Di Wu. 2024. "How Urban Street Spatial Composition Affects Land Surface Temperature in Areas with Different Population Densities: A Case Study of Zhengzhou, China" Sustainability 16, no. 22: 9938. https://doi.org/10.3390/su16229938

APA Style

Fu, M., Ban, K., Jin, L., & Wu, D. (2024). How Urban Street Spatial Composition Affects Land Surface Temperature in Areas with Different Population Densities: A Case Study of Zhengzhou, China. Sustainability, 16(22), 9938. https://doi.org/10.3390/su16229938

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

Article Metrics

Back to TopTop