Revealing the local heterogeneity of urban development’s impact on PM2.5 pollution across Chinese cities: An interpretable spatially aware machine learning approach

Junjia Liu

doi:10.1016/j.apr.2026.103114

Revealing the local heterogeneity of urban development’s impact on PM2.5 pollution across Chinese cities: An interpretable spatially aware machine learning approach

2026-06-18

0

PRE

AI

J

Junjia Liu

B

Baoxin Zhai *

Q

Qiannan Duan

Z

Zhuoyi Xu

DOI：10.1016/j.apr.2026.103114

Original

Original request for help

Save

Abstract

En 中文

• GXGB outperforms other spatial models, effectively eliminating spatial residual bias. • Core PM2.5 drivers shifted from industrial structure to population and economic density. • Associations of population and built-up land with PM2.5 reverse beyond critical thresholds. • An integrated "Source-Sink-Regulation" theoretical framework is established. • Cities are classified into seven policy archetypes for targeted PM2.5 control.

Revealing the local heterogeneity of urban development’s impact on PM2.5 pollution across Chinese cities: An interpretable spatially aware machine learning approach

Abstract

Journal

Organization

Cited Papers

Citing Papers