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

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.