<p>Urban form exerts asymmetric influences on environmental outcomes, posing complex trade-offs for spatial governance. This study identifies and conceptualizes a “Pollutant’s Dilemma” within the Beijing–Tianjin–Hebei urban agglomeration, wherein urban concentration simultaneously intensifies PM<sub>2.5</sub> emissions through agglomeration effects and reduces CO<sub>2</sub> emissions via intensification. Using 5-km grid data for three structurally significant time points—2014 (policy baseline), 2018 (structural restructuring phase), and 2022 (transition endpoint)—we develop a novel analytical framework that couples Stacking ensemble learning with Shapley Additive Explanations (SHAP) interpretation to uncover both dominant and interaction effects of urban form. The analysis yields three principal findings. First, urban concentration exerts opposing effects on the two pollutants—amplifying PM<sub>2.5</sub> via agglomeration while suppressing CO<sub>2</sub> via intensification—with a population density threshold of approximately 600 persons·km⁻² marking the critical governance inflection point. Second, PM<sub>2.5</sub> drivers follow a coherent three-stage evolutionary trajectory (from “ecological dependency” through “scale-ecology trade-off” to “efficiency-balancing”), whereas CO<sub>2</sub> drivers exhibit persistent structural lock-in dominated by economic and built-up scale throughout. Third, High-High clusters embody the Dilemma in its mature form—with antagonistic multi-factor PM<sub>2.5</sub> regulation co-existing alongside unidirectional scale-driven CO<sub>2</sub> promotion—while Low-Low clusters represent a “pre-dilemma” state governed by morphological and ecological mechanisms. These findings challenge the pro-density planning paradigm and provide a stage-sensitive empirical basis for integrating spatial planning with climate mitigation and public health objectives in rapidly urbanizing megaregions.</p>

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

Unraveling urban environmental asymmetry: divergent pathways of PM2.5 and CO₂ shaped by urban form in agglomerations evidence from stacking–SHAP analysis in Beijing–Tianjin–Hebei

  • Yixuan Wang,
  • Yingzhi Lu,
  • Tian Chen,
  • Linghao Wang,
  • Gaoyuan Wang

摘要

Urban form exerts asymmetric influences on environmental outcomes, posing complex trade-offs for spatial governance. This study identifies and conceptualizes a “Pollutant’s Dilemma” within the Beijing–Tianjin–Hebei urban agglomeration, wherein urban concentration simultaneously intensifies PM2.5 emissions through agglomeration effects and reduces CO2 emissions via intensification. Using 5-km grid data for three structurally significant time points—2014 (policy baseline), 2018 (structural restructuring phase), and 2022 (transition endpoint)—we develop a novel analytical framework that couples Stacking ensemble learning with Shapley Additive Explanations (SHAP) interpretation to uncover both dominant and interaction effects of urban form. The analysis yields three principal findings. First, urban concentration exerts opposing effects on the two pollutants—amplifying PM2.5 via agglomeration while suppressing CO2 via intensification—with a population density threshold of approximately 600 persons·km⁻² marking the critical governance inflection point. Second, PM2.5 drivers follow a coherent three-stage evolutionary trajectory (from “ecological dependency” through “scale-ecology trade-off” to “efficiency-balancing”), whereas CO2 drivers exhibit persistent structural lock-in dominated by economic and built-up scale throughout. Third, High-High clusters embody the Dilemma in its mature form—with antagonistic multi-factor PM2.5 regulation co-existing alongside unidirectional scale-driven CO2 promotion—while Low-Low clusters represent a “pre-dilemma” state governed by morphological and ecological mechanisms. These findings challenge the pro-density planning paradigm and provide a stage-sensitive empirical basis for integrating spatial planning with climate mitigation and public health objectives in rapidly urbanizing megaregions.