Synergizing climate, energy, and air quality: Uncovering urban heat island, energy consumption, carbon emission, and air pollution nexus linkages

Urban environmental stressors such as urban heat island, energy consumption, carbon emission, and air pollution often share common anthropogenic sources, forming a complex heat–energy–carbon–pollution (HECP) nexus. This study developed an integrated analytical framework that combines a structural equation model (SEM) and interpretable machine learning via a coupled extreme gradient boosting and SHapley Additive exPlanations (XGBoost–SHAP) model to systematically identify the indirect drivers, nonlinear interactions, and governance pathways of HECP across 269 Chinese cities from 2005 to 2020. The SEM results revealed that socioeconomic factors, including gross domestic product (GDP), traffic pressure, fixed asset investment, government intervention, industrial structure, and industrialization, exert indirect effects on HECP via mediators such as the normalized difference vegetation index (NDVI), population density, and energy structure. XGBoost–SHAP further quantified the relative importance and interactive effects of these drivers, with GDP emerging as the most influential factor (average Shapley value >0.1). Notably, this analysis revealed complex nonlinear interactions among indirect and direct drivers such as GDP–NDVI and traffic pressure–population density, which differentially impact urban heat island intensity, energy consumption, and concentrations of carbon dioxide, fine particulate matter, and ozone. These findings highlight that effective HECP governance cannot rely on single-factor interventions, but instead requires systemic, multi-factorial strategies integrating economic restructuring, land use optimization, and clean energy transition. This study not only advances methodological approaches to synergistic environmental governance but also provides a robust scientific basis for designing adaptive, regionally tailored policy interventions in rapidly urbanizing regions. • An SEM–XGBoost-SHAP framework identifies HECP synergistic governance drivers. • Indirect pathways are identified via NDVI, population density, and energy structure. • Nonlinear interactions exist among key HECP governance drivers and moderators. • Six indirect socio-economic drivers significantly influence HECP governance.