Understanding the spatiotemporal patterns and driving mechanisms of cooling effects from urban blue–green spaces (BGS) is important for mitigating heat risks in high-density cities. However, many existing studies rely on geometrically defined cooling zones, which may have limited physical interpretability, and insufficiently consider the nonlinear interactions between internal BGS attributes and surrounding urban morphology. To address these gaps, this study examines BGS cooling effects within Beijing’s Sixth Ring Road. A physically interpretable Land Surface Temperature Parameterization Method (LSTPM) was used to delineate BGS cooling zones from land surface temperature data. An Extreme Gradient Boosting (XGBoost) model combined with Shapley Additive Explanations (SHAP) was then applied to assess the effects of internal BGS characteristics, surrounding built-environment attributes, socioeconomic factors, and macro-scale spatial context. The results show that BGS cooling effects exhibit clear seasonal variation and urban–rural gradients, with stronger cooling generally observed in the high-density urban core during summer. Internal attributes, particularly patch size and vegetation coverage, were the dominant drivers of cooling performance. Surrounding urban morphology, especially building density, also contributed to the spatial heterogeneity of cooling effects. SHAP analysis revealed nonlinear responses and threshold effects, including diminishing marginal returns for patch size and tipping points associated with surrounding building density and peripheral greenness. These findings provide planning-relevant evidence for optimizing BGS configuration and surrounding urban form to support heat-mitigation strategies in high-density cities.
Zhang et al. (2026) studied this question.