Characterizing the spatial heterogeneity and controlling factors of land surface temperature clusters: A case study in Beijing

摘要

The spatial variation in land surface temperature (LST) is a measure of the exchange and interaction of energy flux between the land surface and near-surface atmosphere. Understanding the connections between LST and complex landscape patterns is vital for selecting proper mitigation strategies to improve the urban thermal environment. However, the high spatial heterogeneity of the urban landscape makes it tricky to accurately characterizing the quantitative contributions from man-made or natural objects. In this study, we first estimated LST with Landsat-8 TIRS by applying a split-window algorithm. We then grouped the urban landscape into 10 urban function zones (UFZs) and classified these UFZs as LST hot or cool spots. We used stepwise regression to examine the relationships between surface temperatures in hot and cool spots and 7 climate-sensitive parameters. We obtained some interesting results: (1) Urban thermal environment is highly heterogeneous at the scale of UFZ. Every UFZ can be partitioned into both hot and cool spots. (2) The prediction accuracy of LST is found to be dependent on a combination of different parameters. The 2-dimensional parameters were more effective in predicting LST than the 3-dimensional parameters according to the stepwise regression analyses. (3) A new 3-dimensional parameter (sun-view factor-SunVF) adopted in this study, provides an alternative option in characterizing the variation of LST of various landscape structures. The results of this study provide insights into how the current spatial pattern of LST formed and into which mitigation strategies can be applied to improve the urban thermal environment.