Local climate zone framework: seasonal dynamics of surface urban heat island and its influencing factors in three Chinese urban agglomerations

The Local Climate Zone (LCZ) framework offers a standardized and widely applicable approach that facilitates detailed investigations of Surface Urban Heat Islands (SUHI) within cities. Based on the LCZ framework, this study selected three urban agglomerations with distinct climate backgrounds: the Guangdong-Hong Kong-Macao Greater Bay Area (GBA), the Liaozhongnan urban agglomeration (LZS), and the Ningxia Yellow River urban agglomeration (NXY). Using a geographically weighted regression model, Moderate Resolution Imaging Spectroradiometer land surface temperature data were downscaled to a resolution of 100 meters. The study used the extreme gradient boosting model and the Shapley Additive Explanations (SHAP) method to analyze the single-factor effects and their interaction effects influencing seasonal SUHI. The results revealed that significant seasonal variations in SUHI exist within the same LCZ type in the three urban agglomerations with different climate backgrounds. In the GBA urban agglomeration, persistent precipitation effectively mitigates the SUHI effect within built-up LCZ types (LCZ 1 to LCZ 10) during summer, resulting in average SUHI intensities that are 0.341 K and 0.132 K lower than in autumn and winter, respectively. Overall, natural factors play a dominant role in influencing SUHI in all four seasons, with latitude and distance from the sea factors being the most crucial factors in all seasons and all three urban agglomerations. However, water bodies smaller than 0.004 km2 fail to provide sufficient cooling effects to significantly alleviate SUHI. Moreover, apart from the proportion of water bodies, the absolute SHAP values of interactions among natural factors are significantly higher than those among human factors or between human and natural factors. This study enhances the refined understanding of the seasonal variations in SUHI within different LCZ types and their crucial influencing factors.