Road surface temperature evaluated with streetview-derived parameters in a hot and humid megacity

Abstract

This study explores the potential of a novel method that uses streetview-derived parameters to explain road surface temperature (RST) variations. The relationships between RST and streetview-derived parameters, including view factors (VFs), sunlit/shaded status, and sunlit hours, are investigated in an urban central district in a hot and humid climate. Thermal mapping of RST was conducted in 14 traverses at different times on a hot summer day and on both sides of the street canyon. The performances of streetview-derived parameters were also compared with commonly used urban morphological parameters. Results show that the correlations between RST and streetview-derived sky view factor (SVF) and green view factor (GVF) can reach 0.92 and − 0.93, respectively, at around 11:25. Building view factor (BVF) shows a weak correlation with RST and exhibits different directions of influence at different times of the day and in different types of routes. Correlations between streetview-derived VFs and RST are scale-dependent, varying by time of day and showing different patterns on the two sides of one canyon. The correlations between RST and traditional morphological parameters (BCR, FAR and meanH) are weaker than streetview-derived ones (sunlit hours, screen status, SVF and GVF). Multiple linear regressions show that streetview-derived view factors can explain 59% ∼ 82% of the spatial variation in daily maximum RST. The study highlights the potential of using streetview-derived parameters to predict RST.