Elsevier

Building and Environment

Volume 206, December 2021, 108354
Building and Environment

3D building configuration as the driver of diurnal and nocturnal land surface temperatures: Application in Beijing's old city

https://doi.org/10.1016/j.buildenv.2021.108354Get rights and content

Highlights

  • 3D landscape metrics firstly applied to associate urban buildings and LST.

  • Identification of how buildings affect DLST and NLST over four seasons.

  • Relative importance of 3D landscape metrics in describing the variations of LST.

  • Suggestions for the old city protection and modification to mitigate urban heat.

Abstract

Urbanization has produced extremely diverse structures of buildings, including horizontal sprawl, vertical growth, and a transition from traditional to modern architecture. Although the influence of urban morphology on urban heat formation is unquestioned, previous research has relied just on the 2D building composition and its influence on diurnal land surface temperatures (DLSTs). However, it is not well known that the 3D building configuration affects nocturnal land surface temperatures (NLSTs) and seasonal variations. In a new approach, a set of 3D landscape metrics, based on both aspects of composition and configuration, is here proposed and tested for spatiotemporal associations to land surface temperatures (LSTs) in Beijing's old city. The combination of classical and modern architecture styles makes this region an ideal laboratory for LST studies in highly different urban structures. Major findings include: 1) 3D landscape metrics effectively and suitably describe the diversity, irregularity and spatial arrangement of buildings; 2) Denser and more compact building patterns result in higher DLSTs, whereas highest NLSTs occur around modern high-rise buildings; 3) 3D landscape metrics have sensitive correlations to DLSTs, but in general NLSTs are closer associated with composition metrics rather than configuration metrics; 4) Both DLST and NLST are most importantly affected by building numbers and nearest distances between buildings; 5) The association between urban morphology and LSTs is fairly stable over all four seasons; with the variation that the summer relationship was relatively lower due to stronger solar radiation and evapotranspiration of urban vegetation.

Introduction

Urbanization is one of the most significant human activities since the 20th century, and lots of buildings have been built, remodeled and enlarged, which affects the urban heat environment significantly [[1], [2], [3]]. Buildings can alter the reflection and absorption of solar radiation, as well as the proliferation of heat in urban area [4]. The surface roughness and irregularity caused by different height, arrangement and density of buildings lead to location dependent and time dependent temperature variations [[5], [6], [7]]. Elevated urban temperatures can threaten the health of city dwellers and the living condition of flora and fauna [[8], [9], [10]]. Therefore, it is important for future urban planning and management to determine how building patterns influence temperatures in cities.

Satellite remote sensing provides up-to-date and spatially explicit land surface temperatures (LSTs) with higher spatial coverage than in situ observations that are limited by low-density monitoring networks and uncertain observation accuracy [11,12]. Remote sensing imagery is increasingly used in the literature to identify the spatiotemporal influences of urban buildings on urban heat [[5], [6], [7]]. A weakness with these studies is that they mainly focused on the diurnal influence of buildings on LSTs on a specific date, while the nocturnal relationship and seasonal variations were rarely considered. The nocturnal temperature is highly related to the human comfort, and might arouse more power consumption for cooling, which in turn, raises air pollution and greenhouse gas emissions [13]. Despite the undisputed importance of 3D spatial structures on LSTs, a comprehensive understanding and explanation are still lacking. Previous studies mostly relied on the two-dimensional (2D) features and three-dimensional (3D) vertical features of buildings (e.g., buildings height, volume, and 3D surface area), rather than 3D spatial configuration. In this study, the characteristics of spatial configuration mainly refer to the compactness and arrangement irregularity of buildings. A compact building structure is usually designed to meet the basic housing requirements for increasing urban population and ease up the conflicts between built-up land and other land use as much as possible [2,14]. Building irregularity is originated from the complexity of single and multiple buildings, and the combination type of building arrangements is directly related to the heat accumulation or heat removal by affecting the urban ventilation, radiation balance schemes, and sunshine conditions [2]. Compared with the composition characteristics, the method for measuring configuration of urban building patterns in 3D space is less targeted and systematic, and particularly lacks a complexity evaluation of building arrangements [15,16].

Metrics for pattern recognition have been widely applied to provide more accurate ecological interpretations for the influence of land use/cover changes on LSTs during past decades [7,10,11,17]. Traditional landscape metrics are usually calculated in 2D space without 3D vertical information, while the urban buildings actually refer to a 2.5D or 3D representation [18,19]. Thanks to the progress in 3D information extraction technology (e.g., SAR, LiDAR, and oblique photogrammetry), several 3D landscape pattern metrics have been introduced by combining traditional 2D landscape metrics with 3D vertical features [[19], [20], [21]]. 3D landscape analysis has the advantage of incorporating internal heterogeneity within patches into the calculation and avoids the shortcoming of considering a patch as totally homogeneous in 2D space [18,20]. However, these new metrics are rarely considered in describing the spatial configuration and composition of urban building patterns, and their efficiency and suitability are also uncertain. Before applying these new metrics to associate the urban buildings with LST, the following scientific questions need to be solved: 1) How to define the concept ‘patch’ and ‘class’ in building patterns, considering that landscape metrics are usually calculated based on a patch-mosaic model? 2) How to interpret the ecological significance of these new metrics and what building characteristics can they reflect? 3) Which landscape metrics are more sensitive to the correlation between urban building morphology and LSTs?

The originality of our approach is the incorporation of 3D urban morphology (composition and configuration of buildings) into studies of LSTs. During daytime, the landscape characteristics [22,23] in a high-rise building region might lead to less sky visibility and less direct solar radiation, which is conducive to a mitigation of high LSTs [4]. At night, the buildings replace the sun in warming the surrounding areas. 3D features of buildings might affect the intensity and spatial variations of heat [24], and a 3D analysis of the built landscape can quantify and compare the heat release at night and heat storage during daytime.

This paper aims to investigate the relationships between the 3D structure of buildings and LSTs and the main objectives include:

  • An evaluation of the effectiveness and suitability of a 3D landscape analysis for studying the spatial heterogeneity of urban building patterns, and its relevance to variations of LSTs.

  • The identification of diurnal and nocturnal impacts of urban buildings on the urban thermal environment during the four seasons.

To this end, the experimental setup of this study included 14 remote sensing images over four seasons for extracting land surface temperatures, and 3D geographical data of Beijing's old city. The results can contribute to a deeper understanding of the influence of urban morphology on urban heat and provide suggestions for the management and conservation of traditional buildings and the old city from the perspective of urban heat management.

Section snippets

Study area

Beijing is one of the largest cities in the world, covers approximately 16000 km2 with more than 20 million urban permanent populations. Beijing has been built as a city 3000 years ago, and taken as the national political center for 800 years. Beijing has a humid continental monsoon climate with severe, dry winters, hot summers and strong seasonality (Köppen-Geiger climate class Dwa = humid boreal climate, 593 mm yearly precipitation, 11.9 °C annual mean temperature, maximum temperature up to

Methods

The data processing consists of three steps (Fig. 3): 1) 3D landscape metrics were calculated based on the footprint and height characteristics of buildings; 2) The DLST and NLST were retrieved using Landsat 8 OLI/TIRS and Terra ASTER remote sensing images, respectively; 3) the diurnal and nocturnal associations between 3D landscape metrics and LSTs were calculated using Pearson correlation coefficient, and their relative importance on affecting the LSTs was evaluated by the random forest

3D landscape metrics for measuring the building patterns in the samples

Significant differences in the building diversity, irregularity and compactness between selected regions were revealed by 3D landscape metrics. In the composition metrics (Table 3), the HMN in Sample 3 (14.301 m) exceeded that in Sample 1 (3.399 m) and Sample 2 (9.707 m) as expected. The NP (1681) and PD (0.005) in Sample 1 were both highest, but the RD (5.977) and volume (1137393 m3) were lowest, because traditional dwellings-courtyard houses account for the largest proportion in this region,

The efficiency of 3D landscape metrics measuring building patterns

Knowledge about how to measure building patterns in 3D space is a key for further monitoring of urban dynamics and its relationship with urban heat [5,15,16]. Traditional methods mainly relied on the building's height, footprint and volume to describe the 3D building characteristics [4,10,34]. These parameters aim at reflecting the horizontal and vertical information of single buildings, which are useful and simple but not sufficient. The influence of buildings on the urban heat environment is

Conclusions

Much research was been conducted over the past decade to understand how urban morphology affects urban heat, mostly focused on diurnal relationships and 2D building characteristics. Involving the urban 3D morphology, this study proposed a set of 3D landscape metrics that characterize the composition and configuration of urban buildings and were calculated from thermal infrared remote sensing images. From our results we conclude:

  • The suggested 3D landscape metrics measure and systematically

Declaration of competing interest

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

Acknowledgement

The first author would like to express his gratitude for the research support from China Scholarship Council under Grant No. 202008080124.

References (36)

Cited by (31)

View all citing articles on Scopus
View full text