Identifying heavy metal sources and health risks in soil-vegetable systems of fragmented vegetable fields based on machine learning, positive matrix factorization model and Monte Carlo simulation

Jiacheng Shi,Yu Yang,Zhijie Shen,Yuding Lin,Nan Mei,Chengzhong Luo,Yongmin Wang,Cheng Zhang,Dingyong Wang
DOI: https://doi.org/10.1016/j.jhazmat.2024.135481
2024-08-10
Abstract:Urban fragmented vegetable fields offer fresh produce but pose a potential risk of heavy metal (HM) exposure. Thus, this study investigated HM sources and health risks in the soil-vegetable systems of Chongqing's central urban area. Results indicated that Cd was the primary pollutant, with 28.33 % of soil samples exceeding the screening value. Amaranth was particularly problematic, exceeding thresholds for Cd, Hg, and Cr, and both amaranth and celery showed significantly higher HM accumulation (p < 0.05). The HM pollution level in the soil-vegetable system was moderate or above. The sources of HMs identified via Positive matrix factorization (PMF) model included agricultural activities (18.19 %), natural soil parent material (25.88 %), mixed metal smelting and transportation (30.72 %), and coal combustion (25.21 %). Furthermore, evaluations using the Random Forest (RF) model revealed an intricate interaction of factors influencing the presence of HMs, where enterprise density, population density, and road density played significant roles in HMs accumulation. Monte Carlo assessments revealed higher non-carcinogenic risks for children (Pb, As) and greater carcinogenic risks for adults (Cd). Therefore, the issue of HM pollution in soils and vegetables from fragmented fields in industrial urban areas need attention, given the potential for elevated health risks with long-term vegetable consumption.
What problem does this paper attempt to address?