Machine learning-assisted source identification and probabilistic ecological-health risk assessment of heavy metal(loid)s in urban park soils

Abstract The accumulation of heavy metal(loid)s (HMs) in the soils of urban parks in industrial cities has raised global concerns because of their environmental and health impacts. However, traditional deterministic assessments commonly overlook uncertainties in pollution evaluation, failing to accu...

Full description

Saved in:
Bibliographic Details
Main Authors: Jun Li, Xu Li, Xi-Sheng Tai, Xin-Ying Tuo, Fa-Yuan Zhou, Yi-Jing Rong, Fei Zang
Format: Article
Language:English
Published: Nature Portfolio 2025-05-01
Series:Scientific Reports
Subjects:
Online Access:https://doi.org/10.1038/s41598-025-02307-1
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849731830094233600
author Jun Li
Xu Li
Xi-Sheng Tai
Xin-Ying Tuo
Fa-Yuan Zhou
Yi-Jing Rong
Fei Zang
author_facet Jun Li
Xu Li
Xi-Sheng Tai
Xin-Ying Tuo
Fa-Yuan Zhou
Yi-Jing Rong
Fei Zang
author_sort Jun Li
collection DOAJ
description Abstract The accumulation of heavy metal(loid)s (HMs) in the soils of urban parks in industrial cities has raised global concerns because of their environmental and health impacts. However, traditional deterministic assessments commonly overlook uncertainties in pollution evaluation, failing to accurately quantify source-specific contributions and associated risks. This study combines multivariate statistical methods, machine learning techniques, and positive matrix factorization (PMF) with Monte Carlo simulation to investigate HM sources, probabilistic pollution levels, source-based ecological risks, and population-specific health hazards in seven urban parks in a nickel-based mining city in China. Results showed that average concentrations of Cd (0.53 mg/kg), Cr (77.72 mg/kg), Cu (171.15 mg/kg), Hg (0.03 mg/kg), Ni (125.42 mg/kg), Pb (27.13 mg/kg), and Zn (81.97 mg/kg) exceeded their background values, except for As (11.85 mg/kg), particularly for Cd, Cu, and Ni, with exceedance rates of 98.46%, 100.00%, and 100.00%, respectively. Probabilistic assessments revealed that pollution levels were particularly high due to Cd, Cu, and Ni. Source apportionment using PMF, correlation analysis, hierarchical cluster analysis, and super-clustering of self-organizing maps identified fertilizers and pesticides (19.33%), industrial atmospheric deposition (21.13%), mining and agrochemicals (16.41%), and mining and transport activities (43.13%) as the major pollution sources. Probabilistic ecological risk assessments showed significant risks from Cd, Hg, and Cu. Non-carcinogenic risks were negligible, while carcinogenic risks were cautionary, especially for children. Mining and transportation activities were the main contributors to ecological risks, while fertilizers, pesticides, and Ni were the primary health risk factors. This study provides a robust framework to improve the accuracy of risk evaluation and offers valuable guidance for targeted interventions and sustainable management of urban soils.
format Article
id doaj-art-dc8bf0777f244cb1bc8899740f954a3e
institution DOAJ
issn 2045-2322
language English
publishDate 2025-05-01
publisher Nature Portfolio
record_format Article
series Scientific Reports
spelling doaj-art-dc8bf0777f244cb1bc8899740f954a3e2025-08-20T03:08:25ZengNature PortfolioScientific Reports2045-23222025-05-0115111510.1038/s41598-025-02307-1Machine learning-assisted source identification and probabilistic ecological-health risk assessment of heavy metal(loid)s in urban park soilsJun Li0Xu Li1Xi-Sheng Tai2Xin-Ying Tuo3Fa-Yuan Zhou4Yi-Jing Rong5Fei Zang6School of Environment and Urban Construction, Lanzhou City UniversitySchool of Environment and Urban Construction, Lanzhou City UniversitySchool of Environment and Urban Construction, Lanzhou City UniversitySchool of Environment and Urban Construction, Lanzhou City UniversitySchool of Environment and Urban Construction, Lanzhou City UniversitySchool of Environment and Urban Construction, Lanzhou City UniversityCollege of Pastoral Agriculture Science and Technology, Lanzhou UniversityAbstract The accumulation of heavy metal(loid)s (HMs) in the soils of urban parks in industrial cities has raised global concerns because of their environmental and health impacts. However, traditional deterministic assessments commonly overlook uncertainties in pollution evaluation, failing to accurately quantify source-specific contributions and associated risks. This study combines multivariate statistical methods, machine learning techniques, and positive matrix factorization (PMF) with Monte Carlo simulation to investigate HM sources, probabilistic pollution levels, source-based ecological risks, and population-specific health hazards in seven urban parks in a nickel-based mining city in China. Results showed that average concentrations of Cd (0.53 mg/kg), Cr (77.72 mg/kg), Cu (171.15 mg/kg), Hg (0.03 mg/kg), Ni (125.42 mg/kg), Pb (27.13 mg/kg), and Zn (81.97 mg/kg) exceeded their background values, except for As (11.85 mg/kg), particularly for Cd, Cu, and Ni, with exceedance rates of 98.46%, 100.00%, and 100.00%, respectively. Probabilistic assessments revealed that pollution levels were particularly high due to Cd, Cu, and Ni. Source apportionment using PMF, correlation analysis, hierarchical cluster analysis, and super-clustering of self-organizing maps identified fertilizers and pesticides (19.33%), industrial atmospheric deposition (21.13%), mining and agrochemicals (16.41%), and mining and transport activities (43.13%) as the major pollution sources. Probabilistic ecological risk assessments showed significant risks from Cd, Hg, and Cu. Non-carcinogenic risks were negligible, while carcinogenic risks were cautionary, especially for children. Mining and transportation activities were the main contributors to ecological risks, while fertilizers, pesticides, and Ni were the primary health risk factors. This study provides a robust framework to improve the accuracy of risk evaluation and offers valuable guidance for targeted interventions and sustainable management of urban soils.https://doi.org/10.1038/s41598-025-02307-1Green spacesHeavy metal(loid)s pollutionEcological-health riskSource apportionmentMonte Carlo simulation
spellingShingle Jun Li
Xu Li
Xi-Sheng Tai
Xin-Ying Tuo
Fa-Yuan Zhou
Yi-Jing Rong
Fei Zang
Machine learning-assisted source identification and probabilistic ecological-health risk assessment of heavy metal(loid)s in urban park soils
Scientific Reports
Green spaces
Heavy metal(loid)s pollution
Ecological-health risk
Source apportionment
Monte Carlo simulation
title Machine learning-assisted source identification and probabilistic ecological-health risk assessment of heavy metal(loid)s in urban park soils
title_full Machine learning-assisted source identification and probabilistic ecological-health risk assessment of heavy metal(loid)s in urban park soils
title_fullStr Machine learning-assisted source identification and probabilistic ecological-health risk assessment of heavy metal(loid)s in urban park soils
title_full_unstemmed Machine learning-assisted source identification and probabilistic ecological-health risk assessment of heavy metal(loid)s in urban park soils
title_short Machine learning-assisted source identification and probabilistic ecological-health risk assessment of heavy metal(loid)s in urban park soils
title_sort machine learning assisted source identification and probabilistic ecological health risk assessment of heavy metal loid s in urban park soils
topic Green spaces
Heavy metal(loid)s pollution
Ecological-health risk
Source apportionment
Monte Carlo simulation
url https://doi.org/10.1038/s41598-025-02307-1
work_keys_str_mv AT junli machinelearningassistedsourceidentificationandprobabilisticecologicalhealthriskassessmentofheavymetalloidsinurbanparksoils
AT xuli machinelearningassistedsourceidentificationandprobabilisticecologicalhealthriskassessmentofheavymetalloidsinurbanparksoils
AT xishengtai machinelearningassistedsourceidentificationandprobabilisticecologicalhealthriskassessmentofheavymetalloidsinurbanparksoils
AT xinyingtuo machinelearningassistedsourceidentificationandprobabilisticecologicalhealthriskassessmentofheavymetalloidsinurbanparksoils
AT fayuanzhou machinelearningassistedsourceidentificationandprobabilisticecologicalhealthriskassessmentofheavymetalloidsinurbanparksoils
AT yijingrong machinelearningassistedsourceidentificationandprobabilisticecologicalhealthriskassessmentofheavymetalloidsinurbanparksoils
AT feizang machinelearningassistedsourceidentificationandprobabilisticecologicalhealthriskassessmentofheavymetalloidsinurbanparksoils