Influences of urban spatial factors on surface urban heat island effect and its spatial heterogeneity: A case study of Xi'an

Optimizing the urban spatial form to mitigate the urban heat island (UHI) effect is considered an efficient technical approach. This study focused on district-scale UHIs in typical high-density cities in northern China. The urban spatial characteristics were extracted and quantified from multiple data sources. Utilizing remote sensing inversion and spatial analysis methods, we calculated the surface urban heat island (SUHI) intensity within each block and the values of 15 urban spatial factors. Multi-scale geographically weighted regression (MGWR) was employed to explore the impact of various factors on the SUHI and their heterogeneity in spatial relationships. Additionally, the random forest model was utilized to assess the contributions of selected factors to the SUHI. The results indicate that the building height and the ratio of permeable surfaces are the dominant factors influencing the SUHI, with contribution rates of 12.0% and 10.3%, respectively. The topographic relief has the least influence on the SUHI, and it can only explain 2.5% of variance. The local variable water area ratio exhibits a distinct cooling effect, while the industrial area ratio demonstrates a noticeable warming effect. The impact of most factors on the SUHI intensity exhibits significant spatial heterogeneity, which proves that the MGWR can better elucidate the extent of spatial variations in the coefficients of factors compared with the global model. The findings help reveal the causes of the spatial heterogeneity of the SUHI and provide a theoretical basis for addressing UHI issues.