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 ( < 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.

中文翻译：

---

基于机器学习、正矩阵分解模型和蒙特卡罗模拟识别破碎菜地土壤-蔬菜系统重金属来源及健康风险


城市分散的菜地提供新鲜农产品，但存在重金属（HM）暴露的潜在风险。因此，本研究调查了重庆中心城区土壤-蔬菜系统中重金属的来源和健康风险。结果表明，Cd为首要污染物，28.33%的土壤样品超标。苋菜的问题尤其严重，超过了 Cd、Hg 和 Cr 的阈值，并且苋菜和芹菜都显示出显着较高的 HM 积累 (< 0.05)。土壤-蔬菜系统重金属污染程度为中度以上。通过正矩阵分解（PMF）模型识别的重金属来源包括农业活动（18.19%）、天然土壤母质（25.88%）、混合金属冶炼和运输（30.72%）和煤炭燃烧（25.21%）。此外，使用随机森林（RF）模型的评估揭示了影响HMs存在的因素之间复杂的相互作用，其中企业密度、人口密度和道路密度在HMs积累中发挥着重要作用。蒙特卡罗评估显示儿童的非致癌风险较高（Pb、As），成人的致癌风险较高（Cd）。因此，考虑到长期食用蔬菜可能增加健康风险，工业城市地区碎片化农田土壤和蔬菜的重金属污染问题值得关注。