Precise source tracing and risk quantification of multi-source heavy metals (HMs) contamination in soil pose significant challenges for regional environmental management. This study proposed an innovative dynamic coupling model that combines a self-organizing map (SOM) and positive matrix factorization (PMF), overcoming the limitations of single models. The geo-accumulation index (Igeo) and the initial ecological risk index (RI) were integrated with human health risks to form a comprehensive evaluation chain of “pollution source-exposure pathway-risk contribution.” The results showed that the average concentrations of eight HMs in soil exceeded background levels, with Cd exhibiting the highest exceedance ratio (0.67 mg/kg, 4.17 × background value). Cd, Zn, and As are influenced by human activities, as evidenced by variation coefficients greater than 70%. The SOM-PMF model successfully identified five pollution sources and their respective contribution rates (contribution/main elements): agricultural sources (8.48%, As), traffic sources (30.24%, Pb, As, Cr), natural background sources (23.04%, Ni, Cu, Cr), industrial sources (18.34%, Cd, Pb, Zn), and smelting sources (19.91%, Mn). Ecological risk assessments have indicated that Cd represents the most significant environmental cumulative risk (<InlineEquation ID="IEq1"> <InlineMediaObject> <ImageObject Color="BlackWhite" FileRef="10661_2025_14310_Article_IEq1.gif" Format="GIF" Height="17" Rendition="HTML" Resolution="72" Type="Linedraw" Width="12" /> </InlineMediaObject> <EquationSource Format="TEX">$\overline{I}$</EquationSource> <EquationSource Format="MATHML"><math> <mover> <mi>I</mi> <mo>¯</mo> </mover> </math></EquationSource> </InlineEquation> geo = 1.52) and potential ecological risk (<InlineEquation ID="IEq2"> <InlineMediaObject> <ImageObject Color="BlackWhite" FileRef="10661_2025_14310_Article_IEq2.gif" Format="GIF" Height="17" Rendition="HTML" Resolution="72" Type="Linedraw" Width="25" /> </InlineMediaObject> <EquationSource Format="TEX">$\overline{RI}$</EquationSource> <EquationSource Format="MATHML"><math> <mover> <mrow> <mi mathvariant="italic">RI</mi> </mrow> <mo>¯</mo> </mover> </math></EquationSource> </InlineEquation> microorganism = 0.34). However, source-risk coupling analysis reveals that agricultural activities (47.3% of As) contribute 24% to the biological risk, with As replacing Cd as the priority element for control. Traffic sources are the primary contributors to health risks, with Pb and As being the main elements (with proportions of 65.18% and 45.52%, respectively), followed by agricultural sources primarily associated with As (47.3%). These sources contribute over 42% to both carcinogenic and non-carcinogenic risks. As was the major contributor to health risks in the study area, with carcinogenic and non-carcinogenic risks in children significantly higher than in adults (3.28 and 4.24 times, respectively). This study provides a scientific basis and technical support for the zonal management and collaborative control of multi-source soil HM pollution. Graphical abstract

Source apportionment and risk quantification of soil heavy metals using SOM-PMF model: implications for ecological and human health management

期刊论文
发表日期：

2025年7月5日

查看全文

Xinghao Lv,
Chao Li,
Huanhuan Wang,
Shufeng Yang,
Shiqi Wang,
Chao Yang,
Fei Teng,
Jing Zhao,
Xuesheng Xu,
Hao Zhang,
Jianbo Shang

摘要

Precise source tracing and risk quantification of multi-source heavy metals (HMs) contamination in soil pose significant challenges for regional environmental management. This study proposed an innovative dynamic coupling model that combines a self-organizing map (SOM) and positive matrix factorization (PMF), overcoming the limitations of single models. The geo-accumulation index (I_geo) and the initial ecological risk index (RI) were integrated with human health risks to form a comprehensive evaluation chain of “pollution source-exposure pathway-risk contribution.” The results showed that the average concentrations of eight HMs in soil exceeded background levels, with Cd exhibiting the highest exceedance ratio (0.67 mg/kg, 4.17 × background value). Cd, Zn, and As are influenced by human activities, as evidenced by variation coefficients greater than 70%. The SOM-PMF model successfully identified five pollution sources and their respective contribution rates (contribution/main elements): agricultural sources (8.48%, As), traffic sources (30.24%, Pb, As, Cr), natural background sources (23.04%, Ni, Cu, Cr), industrial sources (18.34%, Cd, Pb, Zn), and smelting sources (19.91%, Mn). Ecological risk assessments have indicated that Cd represents the most significant environmental cumulative risk ( $\overline{I}$ $\bar{I}$ _geo = 1.52) and potential ecological risk ( $\overline{RI}$ $\bar{RI}$ _{microorganism} = 0.34). However, source-risk coupling analysis reveals that agricultural activities (47.3% of As) contribute 24% to the biological risk, with As replacing Cd as the priority element for control. Traffic sources are the primary contributors to health risks, with Pb and As being the main elements (with proportions of 65.18% and 45.52%, respectively), followed by agricultural sources primarily associated with As (47.3%). These sources contribute over 42% to both carcinogenic and non-carcinogenic risks. As was the major contributor to health risks in the study area, with carcinogenic and non-carcinogenic risks in children significantly higher than in adults (3.28 and 4.24 times, respectively). This study provides a scientific basis and technical support for the zonal management and collaborative control of multi-source soil HM pollution.

Graphical abstract