Elsevier

Urban Climate

Volume 41, January 2022, 101074
Urban Climate

Reexamining the relationship between surface urban heat island intensity and annual precipitation: Effects of reference rural land cover

https://doi.org/10.1016/j.uclim.2021.101074Get rights and content

Highlights

  • Relations between SUHII and mean annual precipitation (MAP) at the global scale are reexamined.

  • Impact of variability of reference rural land cover on the SUHII–MAP correlations is investigated.

  • The significant SUHII–MAP correlations are dependent on variability of reference land cover.

  • The significant SUHII–MAP correlations disappear when keeping reference land cover type fixed.

  • Respecting SUHII as a relative measure is of great importance to interpret SUHI.

Abstract

Previous studies have shown that surface urban heat island intensity (SUHII) across cities is significantly positively correlated with mean annual precipitation (MAP), including linear and nonlinear correlations. Different explanations for these SUHII–MAP relations have been suggested, yet a systematic examination of the impact of variability in reference rural land cover on geographic variations of daytime SUHII and the relation with MAP is still lacking. In this study, the previously proposed SUHII–MAP relations are reexamined by investigating 60 cities across North America and 346 cities across the globe, respectively. The focus of the study is on the role of variability of reference land cover types in the SUHII–MAP relations. The 10-yr time series of satellite-observed land surface temperature, precipitation and solar radiation data, in conjunction with global land cover data, are used for analysis. Our results confirm the reproducibility of both the linear and nonlinear SUHII–MAP relations; nevertheless, the significant positive correlations between the daytime SUHII and the MAP are found to be dependent mainly on the variability of the reference land cover types associated with MAP levels (i.e., varying from desert to tree cover). In contrast, when a constant land cover type is taken as the reference, no significant correlations between the daytime SUHII and the MAP are observed. The finding highlights that respecting SUHI intensity as a relative measure and accounting for variability of reference land cover is of crucial importance for proper interpretation and understanding of SUHI.

Introduction

Conversion of non-urban areas to urban land alters the surface energy balance, resulting in many urban climate effects. A well-documented phenomenon is the urban heat island (UHI) effect, referring to a tendency for urban areas to have higher temperatures than their rural surroundings. The UHI effect has profound impacts on the ecology (Kabano et al., 2021; Grimm et al., 2008), economy (Estrada et al., 2017), and human health (Laaidi et al., 2012). Proper interpretation and understanding of the UHI effect is a cornerstone for developing urban heat mitigation strategies.

The intensity of UHI is a difference in temperature between urban areas and their surroundings, which can be based on temperature differences at the surface, in the air (urban canopy layer or urban boundary layer), and in the substrate (Oke et al., 2017). The focus of this study is urban–rural differences in surface temperature, known as the surface urban heat island (SUHI). The SUHI has been observed and described for a few decades. With the proliferation and easy access of satellite remotely sensed land surface temperature (LST) data, an exponentially increasing trend of SUHI research since 2005 was found (Zhou et al., 2019). The progress, challenges, and perspectives of SUHI research have been systematically reviewed (Voogt and Oke, 2003; Zhou et al., 2019; Stewart et al., 2021; Kim and Brown, 2021). Most SUHI studies have been conducted from an urban perspective, focusing on spatiotemporal dynamics of SUHIs and their relationships with urban factors such as morphology, materials, and population (e.g., Peng et al., 2011; Clinton and Gong, 2013; Yao et al., 2018; Zhou et al., 2017). Besides, local background climate is also thought to have strong influences on SUHIs (Zhou et al., 2014; Zhou et al., 2018; Zhao et al., 2014; Manoli et al., 2019). Some studies have demonstrated that the intensity of SUHIs (SUHII) across cities varies with mean annual precipitation (MAP). An analysis on 1449 Chinese cities showed that larger annual mean daytime SUHIIs occur in humid regions while the smallest occur in arid and semiarid regions (Li et al., 2019a). A study of 65 cities across North America found that the mean annual daytime SUHII increases linearly with the MAP (Zhao et al., 2014). They suggested that the linear SUHII–MAP relation could be largely explained by variations in convection efficiency among cities, rather than those in evapotranspiration. For example, for cities in dry climates, urban areas are aerodynamically rougher than their barren or desert surroundings and therefore dissipate heat more efficiently, resulting in a cooling effect (smaller SUHIIs), while the opposite occurs in humid regions. Another study of 60 cities across North America reported a similar linear correlation between the daytime SUHII and the MAP (Li et al., 2019a). However, they indicated that such a linear relation is controlled by variations in the capacity of urban and rural areas to evaporate water instead of variations in the convection efficiency. An analysis of 32 major Chinese cities showed that the daytime SUHII is positively correlated with the MAP in a quadratic form, which implies that there may be a MAP threshold above which SUHII is insensitive to precipitation changes (Zhou et al., 2016). Manoli et al. (2019) analyzed the relationship between the summertime SUHII and the MAP on a global scale and found a nonlinear increase in SUHII with MAP, which is similar to the prior finding from Chinese cities (Zhou et al., 2016). They suggested that such a nonlinear SUHII–MAP relation is modulated by both changes in convection efficiency and evapotranspiration across different climate regions. In short, although these previous studies have indicated that SUHII increases with MAP, there remains debate regarding the patterns (i.e., linear or nonlinear) and main determinants (i.e., evapotranspiration or convection efficiency or both) of the relationship between SUHII and MAP. Consequently, different interpretations and suggestions with respect to these SUHII–MAP relations were given in previous studies (e.g., Zhao et al., 2014; Li et al., 2019b; Manoli et al., 2019). There is still a lack of general understanding and interpretation of the correlation between SUHII and MAP.

The intensity of SUHIs is a relative measure, which is dependent on the surface temperature levels of both the urban and rural areas. This is a fundamental fact that has to be kept in mind when understanding and interpreting the SUHII and its relationships with various factors. Regarding the topic of this study, variations of the SUHIIs across cities can be caused by variations of both the urban LSTs and the rural reference LSTs and it is critical to identify which is dominant. The importance of this point can be illustrated by the hypothesis proposed by Martilli et al. (2020b): two cities with similar structure, cover, materials and population may experience very different UHI intensities, solely owing to the differing features of their rural surroundings. This is because differences in reference rural temperature lead to different UHI intensities due to the relative nature of UHI intensity. In this case, the different UHI intensities of two cities do not mean that there are contrasting urban temperature regimes between them. A few studies have paid attention to the role of characteristics of the rural surroundings in determining SUHIIs. A comparison of two cities in the northeast United States (US) revealed that larger SUHIIs were observed in the city with a higher tree coverage in the rural area. This is attributed to the lower surface temperatures associated with trees (Zhang et al., 2012). The relation between ecological context and SUHI has been investigated in the continental US (Imhoff et al., 2010) and on a global scale (Zhang et al., 2010). They found that cities in forests biomes exhibit strong SUHIs, while cities in semi-arid and arid biomes are associated with weak SUHIs or even surface urban cool islands. A regional-scale study in northern Italy demonstrated that rural LST regimes, depending on the composition of the rural landscape, lead to variations in SUHII (Heinl et al., 2015). These studies have demonstrated the crucial role played by the surrounding rural landscape in determining the SUHIIs, highlighting the fact that the intensity of SUHIs is sensitive to the surface cover of the rural surroundings (Oke et al., 2017). However, the role of the surrounding rural landscape is often overlooked in SUHI studies. Regarding the previous studies on the relationship between SUHII and MAP, there is a need to systematically examine the influence of the land covers of the rural surroundings. Given that these SUHII–MAP relation studies were carried out at large geographic scales (national, continental, and global), there is a possibility that the variation in SUHII among cities is dominant by variation of the rural reference LSTs rather than that of the urban LSTs. This is because variation of the rural LSTs across a broad geographic scale may be quite large and closely associated with land cover variability that is largely controlled by annual precipitation levels. Additionally, the SUHI intensity in previous studies was commonly defined as the mean LST difference between the whole city and its entire surrounding area (which is typically a buffer zone around the outline of the urban extent). However, for real cities, both the urban and surrounding rural areas are usually characterized by mixtures of many land cover types. Consequently, using the mean LSTs of the entire urban and surrounding rural areas to quantify SUHII inevitably involves uncertainties associated with variations in land cover compositions, and probably masks the role of land covers in the SUHIIs.

This study aims to improve our understanding of geographic variations of SUHII and the relation with MAP, focusing particularly on the role of land covers of the rural surroundings. The relationships between daytime SUHII and MAP on the continental and global scales are reexamined, using a 10-yr time series of MODIS LST data in conjunction with the global precipitation, solar radiation, and land cover data. In the study, in order to reduce the uncertainties associated with variability in land cover compositions, the SUHII is calculated as the LST difference between the core area of a city and the selected rural reference area with an unmixed land cover. Our specific objective is to systematically examine how variability in reference rural land covers affects geographic variations of daytime SUHII and the relation with MAP.

Section snippets

Data

For the objective of this study, the following datasets are used:

  • Global land surface temperature: The LST observations are extracted from the MODIS data product MYD11A2 Version 6 (Wan et al., 2015), which has a spatial resolution of 1 km per pixel. Each pixel value is an average of clear-sky LSTs during an eight-day period, with an error of less than 2 °C according to the product quality control flags. The daily overpass times of the satellite are approximately 13:30 and 1:30 local time. Here,

Results for the 60 cities across North America

Fig. 4a demonstrates a positive and significant correlation between the mean annual daytime SUHII and the MAP when using local undisturbed natural land cover as reference, and the linear regression line is fairly consistent with the correlation determined in Zhao et al. (2014). This confirms the reproducibility of the prior finding by Zhao et al. (2014). But surprisingly, when rainfed cropland is used as a constant reference for all the cities, the daytime SUHII exhibits a statistically

Further implications of our results

Our analysis indicates that geographic variations of daytime SUHIIs are closely associated with variability in the types of reference rural land covers, which plays a crucially important role in determining SUHII and the SUHII-MAP relations. The results also demonstrate that SUHII quantification can be significantly affected by uncertainties associated with the reference areas, as pointed out in the literature (Zhou et al., 2019; Schwarz et al., 2011; Stewart, 2011). This study highlights that

Conclusions

Previous studies have shown that surface urban heat island intensities (SUHIIs) across cities are significantly positively correlated with mean annual precipitation (MAP) and different explanations have been suggested. However, there is still a lack of a systematic examination of the impact of variability in rural reference landscape on geographic variations of daytime SUHIIs and the relation with MAP. In this study, we reexamine the previously proposed linear and nonlinear SUHII–MAP relations

Author contribution statement

Xiaoshan Yang: Funding acquisition, Conceptualization, Methodology, Investigation, Formal analysis, Visualization, Writing - original draft, review & editing.

Lingye Yao: Funding acquisition, Investigation, Data curation.

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 manuscript.

Acknowledgement

This study is supported by the Natural Science Foundation of Jiangsu Province under grant no. BK20201360, the State Key Laboratory of Subtropical Building Science, South China University of Technology under grant no. 2021ZB01, and the Department of Housing and Urban-Rural Development of Jiangsu Province under the Evaluation of Comprehensive Benefits of Green District project.

References (44)

  • J.A. Voogt et al.

    Thermal remote sensing of urban climates

    Remote Sens. Environ.

    (2003)
  • J. Wang et al.

    Attenuating the surface urban Heat Island within the local thermal zones through land surface modification

    J. Environ. Manag.

    (2017)
  • R. Yao et al.

    Interannual variations in surface urban heat island intensity and associated drivers in China

    J. Environ. Manag.

    (2018)
  • D. Zhou et al.

    Surface urban heat island in China’s 32 major cities: spatial patterns and drivers

    Remote Sens. Environ.

    (2014)
  • D. Zhou et al.

    Remote sensing of the urban heat island effect in a highly populated urban agglomeration area in East China

    Sci. Total Environ.

    (2018)
  • Center for International Earth Science Information Network - CIESIN - Columbia University, CUNY Institute for...
  • J. Ching

    WUDAPT: an urban weather, climate, and environmental modeling infrastructure for the anthropocene

    B. Am. Meteorol. Soc.

    (2018)
  • M. Demuzere et al.

    LCZ generator: A web application to create local climate zone maps

  • C. Dian et al.

    Annual and monthly analysis of surface urban heat island intensity with respect to the local climate zones in Budapest

    Urban Clim.

    (2020)
  • F. Estrada et al.

    A global economic assessment of city policies to reduce climate change impacts

    Nat. Clim. Chang.

    (2017)
  • N.B. Grimm

    Global change and the ecology of cities

    Science

    (2008)
  • K. Laaidi

    The impact of heat islands on mortality in Paris during the august 2003 heat wave

    Environ. Health Perspect.

    (2012)
  • Cited by (13)

    • Natural-anthropogenic environment interactively causes the surface urban heat island intensity variations in global climate zones

      2022, Environment International
      Citation Excerpt :

      2) The recovery time of Bowen ratio and albedo after rain events is less than a day, which is shorter than that of SUHI (Hong and Hong, 2016; Ward et al., 2013), and recovery of stored heat inside the buildings and roads is longer than the drying time of surface skin due to evaporation (Hong et al., 2019). Nevertheless, the significant positive correlations between the SUHI and the PREC are found to be dependent mainly on the variability of the reference land cover types associated with mean annual precipitation levels (Yang and Yao, 2022). Tetali et al. (2022) studied that the US showed a paradoxical SUHI phenomenon: LST decreases with increasing vegetation and decreasing with built-up surfaces, but NDVI and LST correlations were weak in India.

    • Mitigating urban heat with optimal distribution of vegetation and buildings

      2022, Urban Climate
      Citation Excerpt :

      We expect that this will help city planners and policymakers to optimize sustainable solutions for the built environment and urban infrastructures. The model sets revenue maximization as the objective function while maintaining a limit for the UHI intensity, which is the difference in temperature between an urban site and a surrounding rural or natural area (Levermore et al., 2018; Yang and Yao, 2022). We proceed as follows: first, we present the optimization model and its underlying assumptions, model variables and parameters, and a formal model description and verbal elaboration.

    View all citing articles on Scopus
    View full text