Contrasting Daytime and Nighttime Characteristics and Driving Mechanisms of Surface Urban Heat Islands Across Local Climate Zones—Beijing as a Case Study

The urban heat island (UHI) effect poses critical challenges to urban sustainability, yet the driving mechanisms of the daytime and nighttime surface urban heat island intensity (SUHII) remain insufficiently understood. This study employs an integrated approach combining remote sensing and GIS to produce a high-accuracy Local Climate Zone (LCZ) classification in Beijing, and utilizes satellite-derived land surface temperature data to analyze the spatial distribution of SUHII across different LCZ types. The XGBoost model, combined with SHAP values, is used to quantify the contributions of nine urban morphological parameters to SUHII and to assess the spatiotemporal differences in these impacts. The results reveal substantial variations in daytime and nighttime SUHII across different LCZs. The highest daytime SUHII occurs in LCZ 8 with an average of 8.48°C, while the highest nighttime SUHII occurs in LCZ 1 and LCZ G, with averages of 3.68°C and 3.64°C, respectively. Crucially, different parameters dominate SUHII at different times: building surface fraction and average building volume are key drivers of daytime SUHII, whereas sky view factor and distance to water play dominant roles at night. Moreover, some parameters, such as average building height and its standard deviation, exhibit contrasting effects between day and night, highlighting the necessity of separate assessments for daytime and nighttime conditions. These findings emphasize the spatiotemporal heterogeneity of SUHII driving mechanisms, underscore the value of LCZ-based classification in disentangling UHI complexity, and demonstrate the importance of quantifying daytime–nighttime differences for designing targeted urban heat mitigation strategies.