Understanding the Relationship Between Urban Green Infrastructure and PM 2.5 Based on an Explainable Machine Learning Model: Evidence From 288 Cities in China

Abstract Urban green infrastructure (UGI) is critical for mitigating fine particulate matter (PM 2.5 ) pollution, a major obstacle to sustainable urban development. However, the morphological spatial patterns of UGI and their impact on PM 2.5 remain largely unexplored, as most related studies have focused solely on case studies. This study employed morphological spatial pattern analysis to document the national scale spatial distribution of seven UGI morphology space patterns (MSPs) across 288 Chinese cities. It verified the disparities of each MSP under varying geographic conditions and scalar categories. Using advanced interpretable machine learning methods that account for aggregated contribution of location features, the study confirmed the positive role of UGI proportion in mitigating PM 2.5 levels. Significantly, the findings revealed that smaller non‐core UGI areas, such as perforation and islet, exert a more pronounced positive impact on reducing PM 2.5 . Furthermore, the study explored the potential PM 2.5 risks facing Chinese cities due to temporal changes of UGI. The study results not only fill the gap in UGI research, but also contributes a feasible urban planning method and provide a basis for reducing PM 2.5 to promote sustainable urban development.