Nonlinear effects of urban and industrial forms on surface urban heat island: Evidence from 162 Chinese prefecture-level cities

Abstract

Urban and industrial forms were widely concerned as influential factors of surface urban heat island (SUHI) at the macroscale. However, there is still a lack of systematic understanding of urban and industrial forms’ nonlinear and threshold effects on SUHI. This study aimed to disclose their nonlinear effects on daytime and nighttime SUHI based on the random forest (RF) model, using the case of 162 prefecture-level cities in China. The RF results showed that industrial factors collectively contributed 43.38% to daytime SUHI variance, whereas urban factors contributed 20.44% to nighttime SUHI variance. Industrial diversity and urban contiguity were the most critical non-natural factors influencing daytime and nighttime SUHI, respectively. Most factors positively affected daytime and nighttime SUHI, except for the urban elongation. Urban boundary complexity had a distinct effect on daytime and nighttime SUHI. Urban compactness/contiguity and industrial agglomeration/diversity had nonlinear and threshold effects on SUHI. Based on the findings, we have a further understanding of the effects of urban and industrial forms on SUHI in the effective ranges.