1. Introduction
As the global urbanization process continues to accelerate, a large number of people are flocking to cities. Urban areas are facing increasing challenges due to rapid urbanization and climate change. Urbanization has significantly transformed natural surfaces into impervious surfaces, which alters the materials, energy, radiation, and composition of the atmospheric structure in the near-surface layer [
1]. Changes in underlying urban surfaces and frequent human activities have an impact on the processes and services of natural ecosystems, resulting in a series of social and ecological problems [
2,
3]. The urban heat island (UHI) effect is one of the best-documented urbanization-related environmental problems and presents a variety of challenges for people, including decreasing environmental quality, injuring human health, increasing energy consumption, and jeopardizing the region’s sustainable development [
4,
5,
6]. Consequently, alleviating the UHI effect has become one of the key research topics at present, and mitigating associated negative effects in natural ways is considered an environmentally friendly and politically and socially acceptable approach [
7,
8].
Urban Blue Infrastructure (UBI) and Urban Green Infrastructure (UGI) are all critical natural landscapes in mitigating the UHI effect [
9,
10,
11,
12]. However, most of the research focuses on UGI’s cooling effect [
13,
14]. UBI includes artificial or natural water bodies, such as rushing water bodies (e.g., rivers, creeks, and canals) and stilling water bodies (e.g., lakes, ponds, wetlands, reservoirs, and seashore lines) [
15]. The cooling effect of UBI is a result of its innate characteristics and its interaction with the state of the surrounding environment [
14]. Compared with other natural surfaces, water bodies are an ideal thermal radiation absorber, as it has high thermal inertia and storage, low heat conductivity, and brightness [
16]. Notably, UBI can provide a remarkable cooling effect on areas close to the water [
17,
18,
19], and compared to the seasonal variation of UGI, the cooling effect of UBI is more stable [
20,
21]. Furthermore, compared to green spaces, water bodies can deliver a greater cooling effect [
13,
14,
15,
16,
17,
18,
19,
20]. Therefore, a deeper understanding of the cooling effect of water bodies and its impact factors can enhance the current insights into the adaptation strategies for UHI effects [
13,
22,
23].
The cooling range and cooling intensity of water bodies in different regions have been investigated and identified [
9,
18,
21,
24,
25,
26]. Additionally, there is increasing attention focusing on the influencing factors of water bodies cooling effect [
18,
27,
28,
29,
30]. Statistical methods, such as principal component analysis [
31,
32], linear regression analyses [
33,
34,
35], stepwise multiple-linear regression [
36,
37], and ordinary least-squares regression (OLS) [
38], have been widely used in the investigation of influencing factors using multiple variables. However, most of the statistical approaches are based on the linear-based hypothesis, which cannot be used to adequately measure factors influencing the spatial heterogeneity of the UBI cooling effect in a nonlinear way. Considering that the cooling effect of UBI is usually influenced by multiple influencing factors, the above methods are based on the linear assumption, and difficult to deal with the combined effects. Based on spatial hierarchical heterogeneity, the Geodetector model is a statistical approach, and its core idea is based on the assumption that if an independent variable has an important influence on a dependent variable, the spatial distribution of the independent variable and the dependent variable should be similar. The principle is to analyze the relationship between the within-strata variance and the total variance of each factor and use spatial stratified heterogeneity to detect the influencing force of each factor on the dependent variable. The key advantages of this model are that it has fewer assumptions than the statistical methods mentioned above, and the results are not impacted by the collinearity of several variables. Second, this model is also capable of detecting the comprehensive contribution of two factors interacting with the dependent variable [
39,
40]. The Geodetector has been extensively utilized to examine the driving force in public health [
39,
40,
41], regional economics [
42,
43], meteorology [
44,
45], and land use fields [
46,
47]. Considering that the cooling effect of water body is affected by multiple factors and the interaction between them is extremely complex, it is difficult to fully measure the internal mechanism by using the linear hypothesis. Therefore, this research attempts to explore the application of the Geodetector model to investigate the driving mechanisms of the cooling effect of UBI.
This paper takes Hefei, the capital city of Anhui province, as a case study region. Hefei, an essence city in central China, has experienced a pronounced summertime UHI effect in recent years [
48]. The study aims to investigate the relationship between UBI and land surface temperature (LST), estimate the cooling effects of UBI on the thermal environment in summer, and investigate the influence of natural and socioeconomic factors on UBI cooling effects. This study focuses on the following two questions: (1) What are the cooling intensity and cooling range of UBI? (2) What are the crucial influencing factors of the UBI cooling effect and their interactions? This study is helpful in offering a theoretical reference and practical guidance for the protection, planning, and design of UBI for moderating the UHI effect.
4. Discussion
4.1. The Effects of UBI on LST
In this study, we found that there were obvious spatial differences in UHI, and the LST decreased gradually from the central city to the suburbs of Hefei (
Figure 6). This is due to the strong correlation between the urban thermal environment and the urban land use types [
73,
74,
75], among which UBI was one of the key factors in mitigating the UHI effect. Based on the spatial superposition analysis of water bodies and LST, we found that the average water LST corresponding to the core area, inner ring area, central area, and outer ring area were 29.88 °C, 30.06 °C, 29.28 °C, and 27.76 °C, which were 5.18 °C, 5.18 °C, 4.93 °C, and 3.19 °C, lower than the mean LST of each area, respectively. This indicated that the UBI can effectively slow down the rise of the urban LST and become a ‘cold island’ in the urban thermal environment. Compared to urban green infrastructure, the cooling effect of UBI are steadier because it is the best absorber of radiation [
30], and the UBI are able to create a stronger cooling effect than urban green infrastructure [
60,
76]. Meanwhile, due to the high heat capacity and low thermal conductivity, water bodies have the ‘thermostatic effect’ to preserve their original temperature through the cooling effect when the ambient temperature rises [
59]. Generally, the larger the water area, the better the cooling effect [
77]. In this research, the size of the water body had the strongest explanatory power for cooling intensity except for road density, indicating that the area of urban water played a significant role in the mitigation of the thermal atmosphere, which is consistent with earlier studies [
27,
30,
77].
4.2. Implications for UBI Landscape Planning and Management
With the rapid urbanization process, UHI has become one of the most important environmental issues in the world. The nature-based solutions to mitigate UHI are considered to be the best way [
78]. Our study indicated that UBI has an obvious cooling effect on surrounding areas. Urban water areas, especially small-sized water bodies, may be preferentially transformed into other land uses to meet the need for the urban socioeconomic development [
79,
80,
81]. As the key natural landscape type in urban areas, better planning and management of UBI may have significant implications for improving UHI.
Previous studies have shown that socioeconomic development influenced by urbanization was the main influencing factor of UHI [
60,
72]. Our research indicated the socioeconomic development variables all had strong explanatory power to the cooling effect of UBI and were positively correlated with the
WCI, and the RD had the greatest explanatory power. The UBI planning and management should consider not only the old downtown area with dense roads but also the organic combination of road construction and UBI in the new development area. Related studies have proved that the cooling intensity of UBI had a close relationship with the size and shape of water bodies [
21,
54,
76,
82]. Our results also showed that patch size and landscape shape index of water bodies had a strong explanatory power on
WCI variation. Some researchers argued that it is unrealistic and unsustainable to increase the size of water bodies due to over-urbanized areas often lacking sufficient space [
13,
54,
59,
60,
83]. Therefore, water bodies with more complicated shapes should be designed during UBI planning, which is of great significance for UHI mitigation.
The interaction between water patch size (WPS) and road density (RD) and nighttime light (NTL) greatly enhanced the interpretation of WCI variation, respectively, which means planning should consider both patch size and local socioeconomic development to improve WCI and mitigate UHI effects. Our research results found that slope, the distance of the surrounding water system, and water connectivity had a great explanatory power, which emphasizes the need to consider the natural environment factors around UBI. Therefore, the UBI planning should consider both water characteristics, such as water size and shape index, regional socioeconomic growth, and the natural environment to efficiently reduce the UHI effect.
Compared with other studies [
29,
61], our research also found that the LSI of water was the only potential driving factor on
WCR. The interactive detection results showed that interaction between water slope (
Ws) and landscape shape index (LSI), water slope (
Ws), and average building height (
AHb) greatly enhanced the interpretation power of
WCR respectively, probably because these influencing factors can change the wind speed and trend around the water bodies, thus affecting the
WCR. Therefore, the spatial arrangement and geographical location of UBI should be considered to improve the
WCR of water bodies.
4.3. Limitations of this Study and Research Directions in the Future
In this paper, we selected the potential influencing factors of the cooling effect of UBI from three aspects, but the internal mechanism of the cooling effect is extremely complex, with numerous influencing factors. Therefore, more factors need to be explored in the future, such as water deepness, volume, and types (rushing water or still water). Furthermore, some exploratory findings on the interaction of the UBI cooling effect have also been summarized in accordance with the interaction detector model; however, the Geodetector model is difficult to be used to finish the calculation of the interaction among three or more factors. Therefore, how to simulate the interactions in increasingly complicated circumstances deserves to be developed in further research.
5. Conclusions
Based on GIS spatial statistics and remote sensing technology, this paper analyzed the cooling effect of UBI in Hefei city using the indices of WCI and WCR. Meanwhile, the Geodetector models were adopted to identify the influencing factors of the cooling effect of blue infrastructure from multiple aspects. The results are as follows:
- (1)
The surface thermal environment of the built-up area of Hefei presented obvious spatial differentiation characteristics. The high-temperature area was mainly concentrated in the core and inner ring area, while the low-temperature area was mainly distributed in the outer ring area and several large reservoirs and forest parks.
- (2)
Nine factors have a significant influence on WCI, including DIST, Ws, Wc, WPS, LSI, Ai, AHb, RD, and NTL, among which road density had the highest explanatory power for WCI variation. In contrast, only the landscape shape index had a significant impact on WCR variation.
- (3)
The cooling effect of UBI is the result of the comprehensive effects of environmental characteristics, water body characteristics, and socioeconomic development characteristics. The interaction of the three type factors had a significant effect on WCI and WCR, and the interaction relationship between the influencing factors was mutually enhanced.
Our results indicate that several factors, such as socioeconomic development and natural environment, have an impact on the cooling effect of UBI. On this basis, it is suggested that future studies should consider more types of UBI and other influencing factors. In conclusion, this study extends our understanding and perception of the cooling effects of UBI, especially the factors that potentially influence the cooling effects of water bodies. These findings can provide suggestions on how to design UBI geographically and spatially to achieve better cooling effects and improve the urban thermal environment, particularly in cities that are undergoing rapid urbanization.