Impact of urban morphology and landscape characteristics on spatiotemporal heterogeneity of land surface temperature

Urban microclimates have undergone extreme changes against the current backdrop of global warming and rapid urbanization. Although several studies on the spatiotemporal characteristics and influencing factors of land surface temperatures (LSTs) have been reported, only a few of them describe the key factors and specific spatial relationships that affect LSTs from time series. Therefore, it is of great significance to explore the spatial heterogeneity of LST from multiple time series and analyze its relationship with various influencing factors. A case study of the Ganjingzi District (GJZ) of Dalian City, China, was performed, with multi-source data and spatial analysis methods being jointly used to explore the spatiotemporal characteristics and driving factors of multi temporal LSTs. The results indicate the following: (i) the GJZ LSTs gradually increased, with the average increase being 0.55 degrees C; (ii) there was a strong and positive correlation among the spatial distribution of LSTs, and the Moran's I value for global spatial autocorrelation exceeded 0.85; and (iii) the fitting degree between the LST and various factors (Adjust-R-2) for the geographically weighted regression (GWR) model was better than that for the ordinary least squares approach. NDBI and LST are highly correlated and have good spatial stability. From the spatial distribution of Local-R-2, it can be seen that with the expansion of urban space, the influence of urban areas on LSTs increases. Therefore, reasonable planning of urban land is of great significance for mitigating heat islands.