生态与农村环境学报2025，Vol.41Issue(9)：1170-1179,10.DOI:10.19741/j.issn.1673-4831.2025.0441
AI驱动生活垃圾填埋场邻苯二甲酸(2-乙基己基)酯(DEHP)迁移的因果识别研究
AI-Driven Causal Identification of Di(2-ethylhexyl)Phthalate(DEHP)Migration in Municipal Solid Waste Landfills
摘要
Abstract
With the continued expansion of domestic waste landfilling and the pervasive use of plastic products in China,the occurrence and migration of emerging pollutants such as di(2-ethylhexyl)phthalate(DEHP)in landfill environments have become increasingly severe,particularly at informal landfills,where the absence of engineered measures(e.g.,anti-seepage systems)leads to more concealed and complex pollution risks.To address the limitations of traditional correlation-based methods in uncovering pollution-driving mechanisms,this study introduces a causal forest model to establish a causal inference framework.The model is used to systematically evaluate the causal effects and stratified heterogeneity of multiple environmental factors,including landfill characteristics,soil physicochemical properties,heavy metal contamination,and waste composition on DEHP migration.The results show that the migration process of DEHP exhibited significant nonlinear response characteristics and stratification sensitivity.The landfill age strongly promoted migration in the younger group(average treatment effect ATE=4.32),while the effect turned negative in the older group(ATE=-0.16).The inhibitory effect is strongest when the pH is in a high range(ATE=-5.66).The heavy metals Cd and Hg exhibited significant syn-ergistic migration potential in the bottom soil layers,with ATE values as high as 49.49 and 54.80,respectively.In con-trast,factors such as landfill depth and organic matter have weaker effects or threshold changes;Rubber and plastic com-ponents weakly promote migration at a moderate proportion(ATE=0.15),while ash and soil components continuously in-hibit migration(ATE=-1.61--0.41).This study demonstrates the capacity of causal machine learning to identify pollu-tion drivers in complex systems,offering a novel analytical tool for pollution control,risk zoning,and source intervention in informal landfills.It supports a paradigm shift from correlation-based attribution to mechanism-driven causal regulation,providing theoretical and methodological guidance for achieving precise and targeted solid waste pollution management.关键词
DEHP/AI驱动/因果识别/迁移机制/垃圾填埋Key words
DEHP/AI-driven/causal identification/migration mechanism/landfill分类
资源环境引用本文复制引用
齐亚平,梁丽琛,张后虎,许元顺,卜元卿..AI驱动生活垃圾填埋场邻苯二甲酸(2-乙基己基)酯(DEHP)迁移的因果识别研究[J].生态与农村环境学报,2025,41(9):1170-1179,10.基金项目
中央级公益性科研院所基本科研业务费专项(GYZX240201) （GYZX240201）
生态环境部事业费项目(历史遗留大宗固废专项排查整治) （历史遗留大宗固废专项排查整治）
生态环境部事业费项目(重点区域危险废物环境风险防控技术支持) （重点区域危险废物环境风险防控技术支持）
