The dual impacts of built environment on extreme heat and urban heat resilience: A comparative study in Beijing

Mitigating urban heat requires both reducing extreme land surface temperatures (ELST) and enhancing urban heat resilience (UHR). However, these two goals are often studied in isolation. This study provides an integrated analysis of ELST and UHR across Beijing's core urban area. We define UHR as an index of thermal resistance—ability to resist temperature increases during extreme hot days. Using Geographically Weighted Random Forest (GWRF) with SHapley Additive exPlanations (SHAP), we analyze how built environment attributes simultaneously influence both ELST and UHR. Our findings reveal that: 1) A critical 94.3 % spatial mismatch between areas with ELST and UHR, indicating that two metrics capture related but distinct aspects of urban thermal environment; 2) While water bodies were the most influential factor in reducing ELST, greenery had the greatest impact on enhancing UHR; 3) The effects of all built environment attributes on ELST and UHR were spatially heterogeneous; 4) The synergistic benefits of greenery and water were context-dependent, reducing ELST in 27.6 % of study area and enhancing UHR in 30.7 %; and 5) The GWRF model significantly outperformed non-spatial approaches (R2 = 0.825 for ELST, 0.903 for UHR). This research underscores the necessity of tailoring strategies that distinctly target temperature reduction and resilience enhancement.