Spatially-heterogeneous impacts of surface characteristics on urban thermal environment, a case of the Guangdong-Hong Kong-Macau Greater Bay Area

Highlights

• The impacts of urban form on thermal comfort differ among WBGT, AT, and UTCI.

• Large low-rise LCZ type is critical in future urban design for thermal mitigation.

• Low plants LCZ type can be a better thermal mitigation strategy than water LCZ type.

• The relationships between thermal indices and urban form are locally varied.

• Variations in the percentages of LCZs have greater impacts on UTCI than WBGT and AT.

Abstract

One-size-fits-all approach is common in climate-sensitive urban design due to neglecting spatial heterogeneities in urban form and urban climate. This study explores a spatially-varied climate-sensitive urban design based on the Guangdong-Hong Kong-Macau Greater Bay Area (GBA). Three thermal indices, the Wet Bulb Globe Temperature (WBGT), the Apparent Temperature (AT), and the Universal Thermal Climate Index (UTCI) are used to assess the outdoor thermal environments. The local climate zone (LCZ) classification system is used to map urban form including built and land-cover types. Moreover, incorporating spatial effects, geographically weighted regression (GWR) models are used to account for spatially-varied thermal variations due to urban form changes. Our findings indicate that the large low-rise type (LCZ 8) needs more attention in built-up planning for thermal mitigation, and urban low plants type (LCZ D) should be a more effective nature-based climate mitigation strategy compared with the water bodies (LCZ G). The GWR results show a stronger consistency between UTCI and LCZ 8 and LCZ D, compared with WBGT and AT. UTCI is thus suggested for application in future urban climate studies. More importantly, the spatially-varied relationship between UTCI and urban form specifies the strategies and appropriate locations for thermal mitigation in climate-sensitive urban design.

Introduction

Rapid regional development affects thermal conditions due to shifting biophysical and ecological processes of land uses with extensive uncontrolled land conversions degrading regional thermal environments in the process of urbanization (Watkins et al., 2007). Shifting climatic and thermal environments related to land use and land cover changes undermine human health and social well-being (Eliasson et al., 2007; Kuchcik, 2020; Dimitrova et al., 2021). Ongoing changes in the regional sprawls and landscape covers affect the thermal conditions and even exacerbate thermal stress (Kalnay and Cai, 2003; Wang et al., 2021a, Wang et al., 2021b; Li et al., 2021). Therefore, urban design for thermal mitigation is becoming increasingly important, for which empirical information of outdoor thermal conditions and urban form is crucial.

Urban forms in terms of land-cover type and urban morphology have influences on urban climate (Zhao et al., 2011). Several individual climate variables are related to urban form to guide the climate-sensitive design in urban settings (e.g., Lau et al., 2015). Air temperature (T_a) is the most commonly used variable to describe urban climate conditions and to assess the cooling effects of vegetation (Ng et al., 2012). Apart from T_a, mean radiation temperature (T_mrt) based on solar radiation is another important indicator of urban heat load (Brown et al., 2015), which is directly affected by the urban form, such as building density (Thorsson et al., 2014). Differences in relative humidity (RH) are observed among urban forms (Yang et al., 2020a). Also, increased wind speed (V) contributes to heat loss and T_a reduction through ventilation (Ng et al., 2011; Yang et al., 2019; He et al., 2020a; He et al., 2020b). An air ventilation assessment project for improving the wind environment was thus initiated by the Hong Kong Government for example (Ng, 2009). Furthermore, combining several climatic variables, thermal indices were established and frequently used to identify biometeorological conditions (Parsons, 2014). For example, the wet bulb globe temperature (WBGT) is a widely used thermal index generally calculated using air temperature and humidity (Budd, 2008). Apparent temperature (AT) is another metric to indicate people's averaged ‘feeling’ based on physical climatic factors including air temperature, wind speed, and humidity (Jacobs et al., 2013). Moreover, universal thermal climate index (UTCI) is the thermal index that has applications in climate research related to human well-being because it involves the simulations of human physiological thermal reactions (Jendritzky et al., 2012). In comparison to AT, solar radiation is also considered in the UTCI quantifications. Relating to urban form, Yeo et al. (2021) compared the thermal behavior of impervious surfaces and tree-covered surfaces based on their values of thermal indices. Tapias and Schmitt (2014) used thermal indices to analyze how the building layout affects urban microclimate. Multiple thermal indices were adopted in the aforementioned research. However, it is unsure which thermal index is more desirable in the field of urban design for thermal mitigation. We assume that the changes in urban form have varying degrees of positive or negative influences on these thermal indices.

Urban design decisions for thermal mitigation may be ill-informed due to the lack of standardized definitions of urban form. For example, many studies claimed a lower daytime temperature in a dense urban setting (e.g., Jamei et al., 2016), while their conclusions are not comparable with each other due to having different criteria of dense urban settings. Similarly, some studies concluded that increasing the number of street trees can help to cool the air and lower radiation fluxes (Klemm et al., 2015; Paschalis et al., 2021). In contrast, other studies claimed that increasing tree density may have negative influences on thermal comfort because of the reduced airflow (e.g., Gartland, 2012). Hence, to avoid ambiguous urban planning, a standardized urban form classification for urban climate studies is required, which can be provided by the local climate zone (LCZ) classification system (Stewart and Oke, 2012). LCZ classification categorizes local scale (10² to 10⁴m) urban surface structures and covers into 17 standard types, including 10 built-up types (LCZs 1–10) and seven land-cover types (LCZs A-G). While built-up types (LCZs 1–10) exhibit more ‘urban’ landscape features related to building and tree layouts, land-cover types (LCZs A-G) present more ‘rural’ landscapes, such as forests and water bodies. Cities have unique sets of thermal properties due to their various built-up and land-cover types, and inter-city urban climate can be compared (Ng and Ren, 2015). Therefore, the LCZ classification system is increasingly used in urban climate and thermal studies to demonstrate the spatial characteristics of urban form (Kotharkar et al., 2019; Kwok et al., 2019; Shi et al., 2018). On a regional basis, however, the LCZ classification is less common.

Using thermal indices and LCZs, the relationships between urban thermal environments and urban form have been greatly explored. However, many of them focused on the spatially stationary relations between LCZs and the thermal indices (Geletič et al., 2018; Jacobs et al., 2019). For example, disparities of thermal indices over LCZs were confirmed based on specific points of time, instead of continuous-time periods (Das and Das, 2020; Das et al., 2020). Moreover, the impacts of urban morphological indicators of LCZs on thermal indices have been quantified in some studies (Liu et al., 2018; Das and Das, 2020), however only averaged and homogeneous correlations were investigated by using a global regression model. Additionally, previous studies mainly used observation data derived from station measurements or field surveys, which cannot cover whole research areas with a coarse spatial resolution (Liu et al., 2018; Kotharkar et al., 2021; Yang et al., 2020b). In this study, based on continuous thermal variations from 2009 to 2019, we use a geospatial model to depict the local patterns of the impacts of urban form on thermal environments. Comparing multiple thermal indices that include different combinations of climatic variables, we can specify the thermal index with the greatest associations with the urban form for future climate-sensitive urban design. The overall objectives of this study are 1) to measure the temporal and spatial variations in urban form based on LCZs types and thermal stress based on multiple indices; 2) to quantify the spatially-variant correlations between percentage changes of LCZs and thermal variations; 3) to identify the thermal index that has the greatest correlations with variations in the percentages of LCZs; and 4) to suggest the urban design strategies for thermal mitigation based on local spatial information.