Seasonal and Regional Variations in Land Surface Temperature Responses to Joint PM2.5 and Anthropogenic CO2 Emissions Pressures Across Urban Agglomerations in China

With urbanization advancing, urban land surface temperatures (LST) are rising, and air pollution is intensifying. LST is shaped by multiple atmospheric and surface processes, yet the combined association of PM2.5 and CO2 emissions with LST remains insufficiently understood at the urban agglomeration scale. This study focuses on the Beijing–Tianjin–Hebei urban agglomeration (BTHUA) in northern China and the Chengdu–Chongqing urban agglomeration (CCUA) in southern China. It quantitatively examines the spatial correspondence of PM2.5 and anthropogenic CO2 emissions, and their seasonal associations with LST. The XGBoost-SHAP model is used to characterize model-based attribution across six management-relevant groups (urban ventilation, blue infrastructure, green infrastructure, surface permeability, low-carbon transition, and air quality management). The results show that (1) cold seasons exhibit a stronger PM2.5 and CO2 emissions compound pressure index (CPI)-LST association, with significant positive CPI-LST relationships; (2) across different categories, areas with high pollution and high anthropogenic CO2 emissions typically have higher LST. In summer, high–high categories reached mean LST of 33.5 °C in BTHUA and 31.1 °C in CCUA, whereas the low–low category showed the lowest winter mean LST values of −3.5 °C and 6.9 °C, respectively; and (3) in BTHUA, blue infrastructure accounted for the largest baseline-model share in category-level model attribution. Within CCUA, air quality management accounted for the largest baseline-model share in category-level model attribution, except during summer. These results provide model-based evidence for seasonally differentiated heat mitigation strategies from an urban agglomeration perspective.