Quantitative assessment of Pb sources in urban–rural river sediments based on Pb isotopes and PMF and MixSIAR models

Abstract

The complex land use patterns in urban–rural rivers and the presence of diverse point and non-point source pollution pose significant challenges for tracing heavy metal(loid) sources in river sediments. This study employed a combined approach using lead (Pb) stable isotopes, positive matrix factorization (PMF), and a Bayesian mixture model (MixSIAR) to determine the concentrations of Cr, As, Cd, Mn, Cu, Zn, Ni, and Pb along with Pb isotope distribution characteristics in sediments from a typical urban–rural river (Yinghe River). Our investigation enabled the quantitative identification of heavy metal(loid) sources and revealed the contribution patterns of multi-source Pb pollution. The results showed that mean concentrations of all heavy metal(loid)s except Cr and Mn exceeded local soil background values. PMF analysis identified four potential sources: natural sources (19.6%) contributing primarily Cr and Mn; industrial sources (32.1%) associated with Cd, Pb, and Ni; agricultural sources (28.0%) linked to Pb, As, and Zn; and traffic sources (20.3%) related to Cu and Zn. Furthermore, by combining Pb stable isotopes with MixSIAR, the contributions of different Pb pollution sources were quantified as agricultural sources (32.1%), industrial sources (30.5%), traffic sources (27.2%), and natural sources (10.3%). The less-than-10% difference in contribution rates between PMF and MixSIAR for Pb source apportionment demonstrated model reliability. Based on the significant correlation between Pb pollution and land use patterns in the Yinghe River, corresponding pollution prevention strategies were proposed. These findings provide a novel perspective for quantitative source identification of heavy metal(loid) pollution in urban–rural river sediments, offering valuable support for river management and heavy metal(loid) pollution control.

---

Environmental significance

This study provides a new perspective for identifying the anthropogenic effects and quantitative sources of Pb pollution in urban–rural rivers, and is expected to be an important basis for guiding the decision-making of urban–rural river management and pollution prevention and control.

1 Introduction

As the main reservoir of river pollutants, sediments accumulate nutrients, persistent organic pollutants, and heavy metal(loid)s.^1–4 These pollutants are mainly derived from urban dust, agricultural runoff, atmospheric deposition, and direct industrial emissions.⁵ Among them, heavy metal(loid)s have become the core of sediment pollution research due to their high toxicity, persistence, and bioenrichment.^6–8 During river transport, the low solubility of heavy metal(loid)s allows them to quickly adsorb to suspended particles and settle to form bottom sediments.^9–11 Under certain conditions, heavy metal(loid)s in sediments can be released into water bodies and transmitted through the food chain, posing a threat to human health.¹² This issue has attracted wide attention at home and abroad.^13–16 Therefore, the accumulation of heavy metal(loid)s in sediments is regarded as an important environmental indicator that can be used to assess regional pollution levels.^17–19

Traditional source analysis of soil heavy metal(loid) pollution mainly relies on statistical models based on concentration data. For example, multivariate statistical methods (including principal component analysis and cluster analysis), positive matrix factorization (PMF), and absolute factor scores–multiple linear regression (APCS-MLR).^20,21 These methods identify pollution source types by analyzing the correlation between elements or performing matrix factorization, and quantify the contribution rate of each pollution source based on statistical assumptions.²² However, in complex pollution environments, traditional methods often struggle to provide accurate and reliable pollution source analysis results. In recent years, Pb isotope tracing technology has gradually become an important tool for soil heavy metal(loid) pollution source analysis due to its unique “fingerprint” characteristics and high accuracy.²³ By analyzing the differences in Pb isotope ratios of different pollution sources, this technology can accurately distinguish Pb pollution from different sources and quantitatively assess source contributions.²⁴ Compared with traditional methods, Pb isotope tracing technology has higher anti-interference ability and requires a smaller sampling volume.²⁵

The sources of heavy metal(loid) pollution in urban–rural rivers are extensive and complex, including industrial emissions, agricultural activities, domestic sewage, and traffic-related pollution.^26–28 The differences in land use patterns and the intensity of human activities in different regions further aggravate the spatial differentiation of the occurrence forms and migration behaviors of heavy metal(loid)s in the environment.²⁹ As a typical urban–rural river, the Yinghe River is dominated by point-source emissions from mineral mining and metal(loid) processing in the upstream industrial agglomeration areas, domestic sewage discharge and traffic pollution diffusion in the middle reaches of urbanized areas, and non-point source input pressure such as fertilizer and pesticide application in the downstream agricultural areas. The spatial and temporal superposition effects of industrial emissions, agricultural activities, and urban pollution sources, as well as the systematic differences in the intensity of human activities and land use patterns across the urban–rural gradient, lead to significant challenges in the interpretation of pollutant sources.

Given the complex source apportionment context in the Yinghe River, this study employed an integrated methodology to identify and quantify the sources and contributions of Pb pollution, thereby addressing the limitations of single-method approaches. The specific objectives were as follows: (1) to utilize principal component analysis (PCA) and cluster analysis (CA) for processing the concentrations of eight heavy metal(loid)s (Cr, Mn, Zn, Pb, Cu, Ni, Cd, and As) in sediments to qualitatively identify potential pollution sources; (2) to apply the PMF model for preliminary identification and quantification of major pollution sources; and (3) to leverage the high objectivity of Pb isotope tracing to further refine the quantification of source contributions, achieving a more accurate analysis of Pb pollution origins. This study provides a novel framework for identifying anthropogenic influences and quantitatively apportioning Pb sources in urban–rural rivers, offering a scientific basis for informed river management and pollution control strategies.

2 Materials and methods

2.1 Study area

The Shaying River basin is the largest tributary of the Huai River and one of its most polluted tributaries. It originates from Funiu Mountain in western Henan Province, crosses Henan and Anhui provinces, and flows through more than 40 cities and counties such as Pingdingshan, Luohe, Xuchang, Zhoukou and Fuyang, with a total length of 624 km. The Shaying River is called the Yinghe River in Anhui, and its main tributary is the Fenquan River. It starts from Jieshou City in Fuyang, flows through Taihe County, Yingquan County and Yingdong County, and finally merges into the Huai River through the mouth of the Mohe River in Yingshang County.³⁰ Fuyang City is located in the north of Anhui Province, China, in the middle of Jianghuai Plain, with a total land area of 11 346 square kilometers and a permanent population of about 2.3 million. The city has a temperate monsoon climate with four distinct seasons, an average annual temperature of 14–16 °C, and an average annual precipitation of 800–1100 mm.³¹ The land use types along the Yinghe River are rich and diverse, covering planting land, agricultural production land, urban construction and residential land, and other types of land dominated by industrial activities (Fig. 1). At the same time, this area is rich in coal resources, and with its resource advantages, it has become one of the important energy bases in China.³²

Fig. 1 Study area and sampling site locations.

2.2 Sample collection

As the main river in Fuyang City, the Yinghe River has significant spatial heterogeneity in the area through which its main stream flows. In this study, 27 sediment sampling points (Y1–Y27) were arranged along the main stream of Yinghe River, as shown in Fig. 1. The sample at each sampling point consisted of four to five subsamples at a collection depth of 0–10 cm collected using a Petersen grab. The locations of sampling sites were determined using a handheld global positioning system (GPS). In addition, in order to further study the potential sources of human influence, 13 samples of potential anthropogenic sources were collected from the surrounding area, including 3 samples of coal, 3 samples of automobile exhaust dust, 3 samples of metallurgical dust, 2 samples of compound fertilizer, and 2 samples of livestock manure. Three additional deep soil samples were collected and used as natural origin samples. All samples were stored in clean polyethylene zipper bags after collection and quickly transferred to the laboratory. The samples were air-dried and passed through a 2 mm screen to remove stones and plant debris. Samples were then ground and sieved through a 200-mesh nylon screen, and all samples were stored at 4 °C in the dark before analysis.³³

2.3 Test and analysis

2.3.1 Reagents and standards. The chemical reagents used in this study were Guaranteed Reagent (GR) provided by Guotai Junan Chemical Reagent Co., LTD (Shanghai, China), which could be used directly without further purification. The experimental water was ultrapure water (resistivity 18.2 MΩ cm, Sichuan Youlu Pure Technology Co., LTD, Sichuan, China), and all solutions were prepared using this ultrapure water. The multi-element standard stock solution (100 mg L⁻¹, containing 21 target elements, National Center for Analysis and Testing of Nonferrous Metals and Electronic Materials, Beijing, China) was stored in cold storage at 4 °C, and the working standard solution was prepared by stepwise dilution with 2% (v/v) nitric acid solution before daily experiments. Sediment reference materials GBW07311 (GSD-11) and GBW07366 (GSD-23) were purchased from the National Reference Materials Resource Sharing Platform (Beijing, China).

2.3.2 Analysis of sediment samples. According to the Ecological Environment Standard of the People's Republic of China, No. HJ1315-2023, the sediment samples were treated by the microwave digestion method, and the concentrations of Pb, Cr, Cu, Ni, Zn, Mn, As and Cd were determined. The operational procedure was conducted as follows. Approximately 0.1–0.5 g of sediment sample was weighed and moistened with a small amount of experimental water. Then, 9 mL of HNO₃ (68%) and 3 mL of HCl (37%) were added to the mixture. After homogenization, the digestion vessel was sealed and placed into a microwave digestion system (CEM, Mars 6). The sealed vessel was subjected to a two-hour heating program following the instrument's digestion protocol. After cooling and pressure release, the resulting digest was transferred to a crucible. The digestion vessel and its lid were rinsed thoroughly, and the rinsate was combined with the digest in the crucible. Subsequently, 2 mL of HF (40%) was added to the crucible, which was then heated to 120–140 °C to remove silica, maintaining the sample in a viscous state throughout this process. Following this, 1 mL of HClO₄ (70%) was introduced, and the temperature was raised to 160–180 °C. Heating continued until the evolution of white fumes nearly ceased. After the crucible had cooled, the residue was rinsed with a solution of HNO₃ (2%). The resulting solution was quantitatively transferred into a 50 mL volumetric flask. The container was washed with HNO₃ (2%), and the washings were added to the flask, which was then diluted to the mark with the same acid solution. The mixture was shaken thoroughly to ensure homogeneity and stored appropriately prior to analysis. Finally, the concentrations of heavy metal(loid) elements in the prepared samples were determined using an Agilent 7900 inductively coupled plasma mass spectrometer (ICP-MS).

All the analyzed data underwent strict quality assurance and quality control. Instrument calibration was performed daily using reference materials to ensure the accuracy of the analytical results. The certified reference materials GBW07311 (GSD-11) and GBW07366 (GSD-23) were used to validate the precision and accuracy of the analytical method. The recovery rates of the certified reference materials for sediments GBW07311 (GSD-11) and GBW07366 (GSD-23) are presented in Table S1. The detection limits of Pb, Cr, Cu, Ni, Zn, Mn, As and Cd were 1 mg kg⁻¹, 2 mg kg⁻¹, 0.7 mg kg⁻¹, 2 mg kg⁻¹, 5 mg kg⁻¹, 2 mg kg⁻¹, 0.2 mg kg⁻¹ and 0.03 mg kg⁻¹, respectively. In each batch of sample processing, both procedural blanks and instrument blanks were analyzed simultaneously to monitor contamination. The concentrations of all target elements in the blanks were below the method detection limits (MDLs).

2.3.3 Analytical procedures for determination of Pb isotope ratios. Approximately 0.1 g (accurate to 0.0001 g) of freeze-dried sediment sample was weighed into a Teflon digestion vessel. Subsequently, 8 mL of HNO₃ (68%), 2 mL of HF (40%), and 1 mL of HClO₄ (70%) were added. The mixture was digested at a controlled temperature using a microwave digestion system (CEM, Mars 6). The resulting digest was then evaporated to near dryness, followed by redissolution in 3 mL of a 2 mol L⁻¹ HBr solution for subsequent analysis. A micro-column with a bed volume of 2 mL was prepared using AG1-X8 anion-exchange resin (200–400 mesh, Bio-Rad). The resin was sequentially conditioned with ultrapure water, 6 mol L⁻¹ HCl, and 0.5 mol L⁻¹ HBr. After loading the sample solution onto the micro-column, the matrix elements were eluted stepwise with 1 mL of 0.5 mol L⁻¹ HBr. Finally, the Pb fraction was collected by eluting with 2 mL of 6 mol L⁻¹ HCl.³⁴ The eluate was evaporated to complete dryness and reconstituted in 1 mL of 2% HNO₃.

Pb isotope ratios were determined using a multi-collector inductively coupled plasma mass spectrometer (MC-ICP-MS, Nu Plasma) at the State Key Laboratory of Environmental Geochemistry, Institute of Geochemistry, Chinese Academy of Sciences, with an external reproducibility (1σ) of ± 0.01%. The internal standard thallium was used to calibrate instrument mass discrimination. Measurement accuracy was checked by analysis of certified reference material (NIST, SRM-981). The measured ²⁰⁶Pb/²⁰⁴Pb and ²⁰⁷Pb/²⁰⁴Pb ratios were 16.9347 ± 0.0007 and 15.4925 ± 0.0004 (n = 20), respectively, which are in good agreement with the values reported in previous studies.^35,36

2.4 Data analysis

Basic statistical analyses were conducted using Microsoft Excel 2021 (Microsoft Corp, USA) software. The geo-accumulation index (I_geo) was adopted to evaluate the heavy metal(loid) pollution in sediments. To initially identify the potential sources of Pb pollution in sediments, principal component analysis (PCA) and factor analysis (CA) were performed using SPSS 27.0 (IBM Corp., USA), statistical software for social sciences. The Positive Matrix Factorization model (PMF 5.0, U.S. EPA) was applied to quantify the potential sources and contributions of Pb in the sediments. Pb stable isotope data and physicochemical data were processed using Origin 2022 (OriginLab Corp., Northampton, MA, USA), and relevant graphs were generated. The MixSIAR model (v1.2.4, University of California, Davis, USA) was employed to quantitatively apportion Pb pollution sources. Land use type analysis was conducted using ArcGIS 10.4 (Environmental Systems Research Institute, Redlands, CA, USA).

2.4.1 PMF model. Positive definite matrix factorization (Matrix, Factorization, PMF) is a source apportionment technique based on the receptor model and it deconstructs the covariance matrix of a multivariate dataset through a dimension–reduction algorithm, and it can quantitatively analyze the multi-source contributions of chemicals at the water–soil–air interface.³⁷ The core algorithm uses non-negative constraint matrix factorization (X = GF + E) to extract the fingerprint spectrum of potential pollution sources while retaining physical significance, and an embedded uncertainty weighting function (Q = ∑(X_ij − ∑G_ikF_kj)²/U_ij²) is used to adaptively deal with missing data and measurement errors.³⁸

2.4.2 MixSIAR model. MixSIAR is a tool for the quantitative assessment of mixture source contributions and is available through R Studio 2022 under the framework of R 4.1.3 version. When applying Bayesian statistical methods to estimate the contribution of different sources to a mixture, the stable isotope values of each source are assumed to follow a normal distribution.²³ A significant advantage of MixSIAR is its ability to combine fixed and random effects as covariates, thus accounting for variability in mixing proportions. By introducing the Markov chain Monte Carlo (MCMC) method, the model can effectively deal with the posterior distribution problem and reduce the error of isotope separation. Therefore, MCMC has been widely used in related studies such as Pb pollution source identification.³⁴ The MixSIAR model employed in this study is a dual isotope ratio model, which was run in the concentration-independent mode. The long iteration version was used for calculations, with the number of iterations set to 100 000. Convergence statistics show that all Gelman–Rubin statistics are less than 1.01, indicating that the model met the convergence criteria. A total of 4 pollution sources were incorporated into the model, and 15 sediment samples were used as mixture data. The model expressions are as follows:

S_k ∼ N(µ_k,ω_k²)

C_k ∼ N(λ_k,τ_k²)

ε_k ∼ N(0,σ²)

Here δ^206/207Pb_i and Δ^208/206Pb_i represent the isotope ratio (δ^206/207Pb and Δ^208/206Pb) of sediment sample i, where i = 1, 2, 3…, N; k refers to the potential Pb sources, k = 1, 2, 3, 4 representing industrial sources, agricultural sources, traffic sources, and natural sources, respectively in this study; P_k is the proportional contribution of Pb source k, which is estimated using the Bayesian mixing model; S_k is the isotopic composition of Pb source k, which is normally distributed with mean µ_k and standard deviation ω_k²; C_k represents the isotope fractionation factor of Pb source k, which is normally distributed with mean λ_k and standard deviation τ_k²; and ε_k is the residual error representing the additional unquantified variation between individual mixtures, which follows a normal distribution with a mean of 0 and standard deviation σ². The Bayesian mixing model takes CSV data from Microsoft Excel of δ^206/207Pb and Δ^208/206Pb, S_k and C_k as inputs. Table 1 lists the average isotope ratios corresponding to the input data of each pollution source and their respective standard deviations.

Table 1 Stable Pb isotope ratios of pollution sources (mean ± standard deviation)

Type of pollution source	^206Pb/^207Pb (mean ± SD)	^208Pb/^206Pb (mean ± SD)	Number of samples	Data source
Industrial source	1.174 ± 0.008	2.110 ± 0.008	6	Metallurgical dust and industrial coal combustion
Agricultural source	1.183 ± 0.004	2.090 ± 0.007	4	Livestock manure and compound fertilizer
Traffic source	1.160 ± 0.005	2.108 ± 0.004	3	Automobile exhaust deposited particles
Natural source	1.196 ± 0.005	2.082 ± 0.003	3	Deep soil in the study area

---

2.4.3 The geological accumulation index assessment. The enrichment of heavy metal(loid)s in soil can be determined by comparing the current values with the background values (Müller et al., 1969; Ji et al., 2008; Zhao et al., 2022). The geological accumulation index (I_geo) was calculated using the following equation:

I_geo = log₂(C_i/1.5S_i)

where I_geo is the geological accumulation index of heavy metal(loid) in soil; C_i is the measured content value of heavy metal(loid), mg kg⁻¹; S_i is the background value. The evaluation criteria for I_geo are shown in Table S2.

3 Results

3.1 Composition of heavy metal(loid)s in sediments

Based on the land use types and spatial distribution characteristics, the main stream of the Yinghe River can be divided into three sections: IM (Y1–Y9) in the upstream section was dominated by industrial mixed land, UM (Y10–Y18) in the middle stream was dominated by urban mixed land, and AM (Y19–Y27) in the downstream section was mainly agricultural mixed land. Table 2 lists the concentrations of 8 elements in the sediments. The results showed that the mean concentrations (range) of Mn, Zn, Cr, Pb, Cu, Ni, As, and Cd were 511.6 (304.3–731.7), 73.79 (57.72–94.72), 63.04 (42.64–89.92), 29.10 (19.56–45.96), 28.76 (21.31–44.42), 28.74 (22.23–41.63), 11.43 (6.35–23.74), and 0.31 (0.20–0.42) mg kg⁻¹, respectively. River sediments are a mixture of surface weathering products in the watershed, so the element concentrations in sediments are closely related to the soil environment of the watershed.³⁹ The I_geo analysis results showed that Cd in all sediment samples was moderately contaminated or unpolluted to moderately contaminated; As in 7 samples was unpolluted to moderately contaminated; Zn in 5 samples was unpolluted to moderately contaminated; Pb in 4 samples was unpolluted to moderately contaminated; Cu and Ni in 2 samples were unpolluted to moderately contaminated; while Mn and Cr in all samples were almost uncontaminated. The order of average I_geo values in all samples was: Cd > Zn > As > Pb > Ni > Cu > Mn > Cr.

Table 2 Statistical Summary of heavy metal(loid) concentrations in sediments (mg kg⁻¹)^a

Land use types		Mn	Zn	Cr	Pb	Cu	Ni	As	Cd
a IM, industrial mixed land; UM, urban mixed land; AM, agricultural mixed land.
IM (Y1∼Y9)	Max	731.69	94.72	89.92	43.86	44.42	39.33	23.74	0.42
	Min	403.89	64.52	55.22	23.56	23.12	26.03	10.64	0.24
	Mean	532.83	78.09	63.54	31.35	31.19	29.69	14.23	0.33
	SD	103.39	7.97	10.34	7.68	5.43	4.21	3.95	0.05
UM (Y10∼Y18)	Max	78.64	75.92	78.62	45.96	34.62	34.63	13.34	0.37
	Min	42.62	57.72	42.62	19.56	21.31	22.23	6.84	0.20
	Mean	490.11	69.13	62.90	28.17	26.57	27.35	10.06	0.28
	SD	72.31	5.64	9.83	8.67	3.87	2.64	1.94	0.04
AM (Y19∼Y27)	Max	685.39	89.72	82.22	39.66	40.12	41.63	15.74	0.40
	Min	304.30	66.92	47.82	22.76	23.82	25.73	6.35	0.25
	Mean	511.86	74.17	62.68	27.79	28.54	29.19	10.04	0.32
	SD	101.05	7.24	9.03	4.75	4.99	4.58	3.32	0.05
Total	Max	731.69	94.72	89.92	45.96	44.42	41.63	23.74	0.42
	Min	304.30	57.72	42.64	19.56	21.31	22.23	6.35	0.20
	Mean	511.60	73.79	63.04	29.10	28.76	28.74	11.43	0.31
	SD	94.94	7.82	9.75	7.13	5.17	4.02	3.74	0.05

---

3.2 PMF results

Four principal components were identified in the sediments through principal component analysis (PCA). The variance explained by each component and its factor contributions to the variables were further assessed using positive matrix factorization (PMF), as shown in Fig. 2.

Fig. 2 Results of correlation analysis, PCA and PMF of heavy metal(loid)s in sediments.

F1 accounted for 19.6% of the four components, mainly including Mn (43.9%) and Cr (41.3%). F2 accounted for 28% of the four components, mainly including As (61.2%), Pb (33.1%) and Zn (24.6%). F3 accounted for 32.1% of the four components, mainly including Cd (48.2%), Ni (37.7%) and Pb (33.6%). F4 accounted for 20.3% of the four components, mainly including Cu (38.8%) and Zn (31.4%).

3.3 Pb isotopic composition

Pb pollution in sediments mainly originates from human activities, including non-ferrous metal(loid) mining and smelting, coal burning, traffic-related sources, and the use of Pb-containing compound fertilizers, livestock manure, and pesticides in agricultural activities.^22,40–42 In the present study, three categories of representative anthropogenic source samples were systematically collected from the research area, including livestock manure and compound fertilizers (representing agricultural sources), metallurgical dust and coal combustion residues (representing industrial sources), and automobile exhaust deposition particles (representing traffic sources). Concurrently, deep soil samples from the vicinity of the study area were collected as natural source samples. Pb isotope analyses were subsequently performed on 15 sediment samples, 13 potential anthropogenic source samples, and 3 natural source samples. Fig. 3 shows the relationships between the ²⁰⁶Pb/²⁰⁷Pb and ²⁰⁸Pb/²⁰⁶Pb ratios in sediment samples, as well as the relationships between 1/Pb and the ²⁰⁶Pb/²⁰⁷Pb and ²⁰⁸Pb/²⁰⁶Pb ratios, respectively.

Fig. 3 The ratios of (a)²⁰⁶Pb/²⁰⁷Pb and ²⁰⁸Pb/²⁰⁶Pb, 1/Pb and (b)²⁰⁶Pb/²⁰⁷Pb, and (c) ²⁰⁸Pb/²⁰⁶Pb in sediments. IM: industrial mixed land; UM: urban mixed land; AM: mixed agricultural land.

In general, natural sources exhibit a relatively higher ²⁰⁶Pb/²⁰⁷Pb ratio.³⁴ The ²⁰⁶Pb/²⁰⁷Pb and ²⁰⁸Pb/²⁰⁶Pb ratios in the deep soil samples from the study area are 1.1965 ± 0.0045 and 2.0815 ± 0.0025, respectively. In comparison, anthropogenic sources show a relatively low ²⁰⁶Pb/²⁰⁷Pb ratio. The ²⁰⁶Pb/²⁰⁷Pb and ²⁰⁸Pb/²⁰⁶Pb ratios of anthropogenic sources are as follows: agricultural sources (1.182 ± 0.004; 2.090 ± 0.007), industrial sources (1.173 ± 0.009; 2.111 ± 0.010), and traffic sources (1.1605 ± 0.0045; 2.1075 ± 0.0035). The lead isotope composition of sediment samples shows significant spatial variations, and such variations are closely related to the land use pattern of the river basin (Fig. 3a). The lead isotope composition (²⁰⁶Pb/²⁰⁷Pb and ²⁰⁸Pb/²⁰⁶Pb) in the upstream industrial area (IM) are 1.1776 ± 0.0030 and 2.1158 ± 0.0040, respectively; those in the midstream urban area (UM) are 1.1628 ± 0.0030 and 2.102 ± 0.0045, respectively; and those in the downstream agricultural area (AM) are 1.1806 ± 0.0025 and 2.09 ± 0.0055, respectively. In this study, the MixSIAR model was used to analyze the lead isotope ratios of 15 sediment samples, and the results were presented in the form of boxplots. In the boxplots, the lower quartile, upper quartile, and median of the box correspond to the values of the MixSIAR model variables at the 25%, 75%, and 50% probability levels, respectively. The lower and upper limits of the whiskers correspond to the values of the model variables at the 5% and 95% probability levels, respectively; the whiskers typically represent the confidence interval, so as to reflect the uncertainty range of the variable values. Additionally, in this study, the mean value from the MixSIAR model run was used to characterize source contribution, as this mean value can serve as a comprehensive estimate of the “most probable values” of the parameters. The source analysis results based on the MixSIAR model showed that the contribution rates of Pb in sediments were as follows: agricultural source (32.1%) > industrial source (30.5%) > traffic source (27.2%) > natural source (10.3%) (Fig. 4a).

Fig. 4 Contribution of Pb in sediments from potential sources based on MixSIAR(a and b) and PMF (a). The mean value represents the source contribution.

4 Discussion

4.1 Assessment of heavy metal(loid) pollution and source contribution

4.1.1 Assessment using the geoaccumulation index. The I_geo analysis results indicated that heavy metal(loid)s have caused varying degrees of pollution to the water environment of the Yinghe River, with Cd contamination being the most severe in particular. This result is consistent with previous research.³⁰ Heavy metal(loid)s were significantly enriched in the surface sediments of the Yinghe River, with a wide range of concentrations. In particular, the coefficient of variation of As was as high as 0.33, indicating significant spatial variation in heavy metal(loid) content within the study area. This high degree of variability may result from the diversity of pollution sources within the region or the presence of point source pollution.⁴³

In the upper IM reach, the average I_geo values of Zn and As are greater than 0 and close to the level of moderate contamination, indicating significant enrichment of Zn and As. In the middle UM reach, except for Cd, the average I_geo values of all other elements are less than 0, suggesting almost no heavy metal(loid) contamination. This variation may be related to the effective interception of upstream pollutants by urban sewage treatment facilities and the management of urban point source pollution.⁴⁴ In the lower AM reach, the average I_geo values of Pb and Cu increase significantly and are both close to the level of moderate contamination, which may be potentially affected by agricultural non-point source pollution. The average I_geo values of Cd in all three reaches are close to or exceed 1, indicating that Cd contamination is severe and affected by multiple factors, including not only industrial waste discharge from the upper reaches but also external pollution inputs from agricultural production and urbanization processes.⁴⁵

4.1.2 PMF analysis. Cr and Mn were strongly correlated, and their average contents were lower than the local soil background values, which indicates that they were not affected by obvious human activities and that their sources may be closely related to the normal weathering and erosion process of geological layers.⁴⁶ The content of Cr in the soil is affected by the parent material, and Mn is one of the most abundant elements in the lithosphere and a main component in the soil; so F1 is considered to be of natural origin.^47,48

There is a positive correlation between As and Pb. As is an important compound fertilizer and an organic fertilizer component, and is often used together with Pb as a livestock feed additive.⁴⁹ This mode of use, which is used to promote animal growth, may lead to significantly higher concentrations of arsenic and Pb in livestock manure.⁵⁰ In addition, Zn, as a common component in agricultural pesticides, contributes about 1200 tons of Zn input to agricultural soils in China each year through pesticide application.⁵¹ Considering the large agricultural runoff in the study area, F2 can be judged as an agricultural pollution source.

Cd and Ni were strongly and positively correlated. Numerous studies have shown that Cd and Ni originate from industrial activities, such as perennial mining, including ore smelting, coal consumption, steel production, and metal(loid) processing.^52,53 Fuyang City has the largest green recycling and utilization base of regenerated Pb in China, and Pb, Cd and Ni can also accumulate in the environment due to metal(loid) smelting and coal combustion. Therefore, F3 is considered to be of industrial origin.

Cu and Zn are positively correlated, and both Cu and Zn are important components of key parts such as vehicle brake systems, wheel bearing, and transmission systems. Due to the friction generated during vehicle braking, dust containing Cu and Zn enters the surface soil together with dust in the atmosphere.^54,55 In addition, modern vehicles are usually supplemented with copper containing lubricants and fuel additives during use. These substances are emitted through the exhaust during vehicle driving, and Zn is also gradually released into the environment through tire friction.⁵⁶ Therefore, F4 is considered to be the source of traffic.

4.1.3 MiXSIAR analysis. The lower ²⁰⁶Pb/²⁰⁷Pb ratios in sediments compared to natural sources indicate a significant anthropogenic influence on sedimentary Pb. Fig. 3b and c show that there is no significant correlation between 1/Pb and ²⁰⁶Pb/²⁰⁷Pb, and ²⁰⁸Pb/²⁰⁶Pb ratios in sediments, which suggests that a simple binary mixing model cannot be used to evaluate the relative contributions of natural and anthropogenic sources. Therefore, this study employs a Bayesian mixing model (MixSIAR) for multi-endmember quantitative analysis. This model effectively overcomes the linear assumption limitation of traditional models through the Markov chain Monte Carlo algorithm.⁵⁷

The weak contribution of natural sources confirmed that human activities were the dominant factor influencing sediment Pb content. The source of Pb in sediments is significantly coupled with the land use pattern in the study area, as shown in Fig. 4b. In the IM section, the contribution of industrial sources to Pb in sediments was the most significant, while the contribution of the other three Pb sources was much lower than that of industrial sources, which was closely related to the spatial distribution of non-ferrous metal(loid) smelting industrial clusters such as recycled Pb industrial parks and coal-fired power plants in this section.⁵⁸ In the UM section, the contribution of traffic sources to Pb in sediments was the most prominent, while the contribution of natural sources was relatively minimal, which might be attributed to the changes of topography in the process of urbanization and the emission of Pb-containing particles caused by high-intensity traffic activities.⁵⁹ In the AM section dominated by agricultural activities, the contribution of agricultural sources to Pb in sediments was the most significant, while the contributions of industrial sources and natural sources were similar. It may be related to the application of compound fertilizer in intensive agricultural areas, the long-term accumulation of waste from livestock and poultry breeding, and the input of compound pollution caused by the dense distribution of rural settlements.^60,61

4.1.4 Comparative analysis between PMF and MixSIAR. The PMF and MixSIAR models provided mutually validating results for the source apportionment of Pb pollution in sediments. PMF identified agricultural, industrial, transportation, and natural sources through matrix factorization of element concentrations, while MixSIAR quantified the contribution rates of each source using Pb isotope ratios. Specifically, the contribution rates of agricultural sources were 33.1% (PMF) and 32.1% (MixSIAR), industrial sources were 33.6% (PMF) and 30.5% (MixSIAR), traffic sources were 17.7% (PMF) and 27.2% (MixSIAR) and natural sources were 15.6% (PMF) and 10.3% (MixSIAR), respectively. Despite the different approaches of the two models, the results were highly consistent with each other, with differences of less than 10%, which verified the reliability of the models. The integration of multiple models overcomes the limitations of any single method, provides highly credible results for the accurate analysis of heavy metal(loid) pollution sources, and offers a solid basis for pollution prevention and control decision-making.

4.2 Suggestions for prevention and control of Pb pollution

As shown in Fig. 5, the source apportionment results derived from the MixSIAR model revealed pronounced spatial heterogeneity in the contribution of Pb pollution sources across different land use patterns. This indicates that human activities—specifically industrial, agricultural, and traffic sources—are the primary contributors to Pb pollution, necessitating differentiated control strategies.⁶² Watershed ecosystem restoration and pollution load control are the core approaches to ensure river water quality and reduce Pb input.³⁸ Among them, industrial wastewater discharge containing Pb and other toxic heavy metal(loid)s should be the key focus of control, as such point source pollution poses direct hazards to water, soil and river ecosystems.⁶³ In high-intensity industrial activity areas represented by the IM section, source reduction should be achieved by strictly formulating pollutant discharge standards and promoting clean production technologies.⁶⁴ For agricultural-dominated areas represented by the AM section, it is necessary to optimize fertilization patterns, adopt precision fertilization technologies, and promote bio-based fertilizers to reduce non-point source pollution caused by agricultural activities.⁶⁵ In traffic source-dominated areas represented by the UM section, dust control technologies can be used to prevent Pb-containing dust (such as automobile exhaust particles and tire wear debris) from entering rivers through precipitation runoff.⁶⁶ Systematic governance must strengthen the collaborative control of multiple sources and establish a Pb pollution prevention and control system based on the spatial heterogeneity of the river basin.

Fig. 5 Pb sources and land use changes in sediments.

5 Conclusions

In this study, PMF was used to analyze the source and contribution of heavy metal(loid) pollution in sediments, and combined with MixSIAR and Pb stable isotope tracing technology, the contribution characteristics of multi-source Pb pollution were systematically revealed. The results indicated that Zn, Pb, Cu, Ni, and As all exhibited varying degrees of contamination in the sediments, while the average concentrations of Cr and Mn remained below the local background values. Four major pollution sources were identified by PMF analysis: industrial, agricultural, traffic, and natural sources. Among them, Cd, Pb and Ni were mainly from industrial sources, Pb, As, and Zn were closely related to agricultural sources, Cu and Zn were closely related to traffic sources, and Cr and Mn were closely related to natural sources. The analysis results based on MixSIAR and Pb isotope tracing technology showed that the difference in contribution rates between the MixSIAR model and PMF model was less than 10%, indicating that the analytical results of the two models were highly consistent, which further enhanced the credibility of source analysis. There was a significant correlation between Pb pollution and land use patterns. According to this, pollution prevention and control strategies were proposed for different land use patterns.

Conflicts of interest

There are no conflicts to declare.

Data availability

The data supporting this study's findings are available from the corresponding author upon reasonable request.

Supplementary information (SI) is available. See DOI: https://doi.org/10.1039/d5em00509d.

Acknowledgements

This work was supported by the Anhui Provincial Quality Infrastructure Standardization Special Project (2023MKS16), the Natural Science Research Project of the Anhui Provincial Education Commission (2023AH040040, 2024AH040046), and the Ecological Environment Science and Technology Project of Anhui Province (2024hb011).

References

1. M. M. Ali, M. L. Ali, M. S. Islam and M. Z. Rahman, Preliminary assessment of heavy metals in water and sediment of Karnaphuli River, Bangladesh, Environ. Nanotechnol., Monit. Manage., 2016, 5, 27–35,  DOI:10.1016/j.enmm.2016.01.002.
2. M. Ao, T. Deng, S. Sun, M. Li, J. Li, T. Liu, B. Yan, W.-S. Liu, G. Wang, D. Jing, Y. Chao, Y. Tang, R. Qiu and S. Wang, Increasing soil Mn abundance promotes the dissolution and oxidation of Cr (III) and increases the accumulation of Cr in rice grains, Environ. Int., 2023, 175, 107939,  DOI:10.1016/j.envint.2023.107939.
3. S. Audry, J. Schäfer, G. Blanc and J.-M. Jouanneau, Fifty-year sedimentary record of heavy metal pollution (Cd, Zn, Cu, Pb) in the Lot River reservoirs (France), Environ. Pollut., 2004, 132(3), 413–426,  DOI:10.1016/j.envpol.2004.05.025.
4. J. Bai, B. Cui, B. Chen, K. Zhang, W. Deng, H. Gao and R. Xiao, Spatial distribution and ecological risk assessment of heavy metals in surface sediments from a typical plateau lake wetland, China, Ecol. Modell., 2011, 222(2), 301–306,  DOI:10.1016/j.ecolmodel.2009.12.002.
5. A. J.-M. Bory, W. Abouchami, S. J. G. Galer, A. Svensson, J. N. Christensen and P. E. Biscaye, A Chinese imprint in insoluble pollutants recently deposited in central Greenland as indicated by lead isotopes, Environ. Sci. Technol., 2014, 48(3), 1451–1457,  DOI:10.1021/es4035655.
6. L. Chai, Y. Wang, X. Wang, L. Ma, Z. Cheng, L. Su and M. Liu, Quantitative source apportionment of heavy metals in cultivated soil and associated model uncertainty, Ecotoxicol. Environ. Saf., 2021, 215, 112150,  DOI:10.1016/j.ecoenv.2021.112150.
7. T. Chen, Q. Chang, J. Liu, J. G. P. W. Clevers and L. Kooistra, Identification of soil heavy metal sources and improvement in spatial mapping based on soil spectral information: A case study in northwest China, Sci. Total Environ., 2016, 565, 155–164,  DOI:10.1016/j.scitotenv.2016.04.163.
8. H. Cheng and Y. Hu, Lead (Pb) isotopic fingerprinting and its applications in lead pollution studies in China: A review, Environ. Pollut., 2010, 158(5), 1134–1146,  DOI:10.1016/j.envpol.2009.12.028.
9. F. Derakhshan-Babaei, K. Nosrati, D. Tikhomirov, M. Christl, H. Sadough and M. Egli, Relating the spatial variability of chemical weathering and erosion to geological and topographical zones, Geomorphology, 2020, 363, 107235,  DOI:10.1016/j.geomorph.2020.107235.
10. T.-T. Ding, S.-L. Du, Z.-Y. Huang, J.-C. Wu and D.-H. Yan, Water quality criteria and ecological risk assessment for ammonia in the Shaying River basin, China, Ecotoxicol. Environ. Saf., 2021, 215, 112141,  DOI:10.1016/j.ecoenv.2021.112141.
11. X. Gao and P. Li, Concentration and fractionation of trace metals in surface sediments of intertidal Bohai Bay, China, Mar. Pollut. Bull., 2012, 64(8), 1529–1536,  DOI:10.1016/j.marpolbul.2012.04.026.
12. Q. Guan, F. Wang, C. Xu, N. Pan, J. Lin, R. Zhao, Y. Yang and H. Luo, Source apportionment of heavy metals in agricultural soil based on PMF: A case study in Hexi Corridor, northwest China, Chemosphere, 2018, 193, 189–197,  DOI:10.1016/j.chemosphere.2017.10.151.
13. S. Guo, Y. Zhang, J. Xiao, Q. Zhang, J. Ling, B. Chang and G. Zhao, Assessment of heavy metal content, distribution, and sources in Nansi Lake sediments, China, Environ. Sci. Pollut. Res., 2021, 28(24), 30929–30942,  DOI:10.1007/s11356-021-12729-9.
14. L. Han, B. Gao, H. Hao, H. Zhou, J. Lu, J. Sun and K. Sun, Lead contamination in sediments in the past 20 years: A challenge for China, Sci. Total Environ., 2018, 640–641, 746–756,  DOI:10.1016/j.scitotenv.2018.05.330.
15. C. K. Jain, H. Gupta and G. J. Chakrapani, Enrichment and fractionation of heavy metals in bed sediments of River Narmada, India, Environ. Monit. Assess., 2008, 141(1–3), 35–47,  DOI:10.1007/s10661-007-9876-y.
16. H.-H. Jiang, L.-M. Cai, H.-H. Wen, G.-C. Hu, L.-G. Chen and J. Luo, An integrated approach to quantifying ecological and human health risks from different sources of soil heavy metals, Sci. Total Environ., 2020, 701, 134466,  DOI:10.1016/j.scitotenv.2019.134466.
17. X. Ke, S. Gui, H. Huang, H. Zhang, C. Wang and W. Guo, Ecological risk assessment and source identification for heavy metals in surface sediment from the Liaohe River protected area, China, Chemosphere, 2017, 175, 473–481,  DOI:10.1016/j.chemosphere.2017.02.029.
18. P. Kelderman and A. A. Osman, Effect of redox potential on heavy metal binding forms in polluted canal sediments in Delft (The Netherlands), Water Res., 2007, 41(18), 4251–4261,  DOI:10.1016/j.watres.2007.05.058.
19. C. Li, M. Sun, C. Song, P. Tao, Y. Yin and M. Shao, Assessment of heavy metal contamination in the sediments of the Shuangtaizi Estuary using multivariate statistical techniques, Soil Sediment Contam., 2017, 26(1), 45–58,  DOI:10.1080/15320383.2017.1245710.
20. F. Li, J. Huang, G. Zeng, X. Yuan, X. Li, J. Liang, X. Wang and X. Tang, Spatial risk assessment and sources identification of heavy metals in surface sediments from the Dongting Lake, middle China, J. Geochem. Explor., 2013, 132, 75–83,  DOI:10.1016/j.gexplo.2013.05.007.
21. X. Li, H. Liu, W. Meng, N. Liu, P. Wu and B. Bai, Accumulation and source apportionment of heavy metal(loid)s in agricultural soils based on GIS, SOM and PMF: A case study in superposition areas of geochemical anomalies and zinc smelting, southwest China, Process Saf. Environ. Prot., 2022, 159, 964–977,  DOI:10.1016/j.psep.2022.01.072.
22. H. Liu, Q. Zhang, Y. Wang, Z. Cai and G. Jiang, Occurrence of polychlorinated dibenzo-p-dioxins, dibenzofurans and biphenyls pollution in sediments from the Haihe River and Dagu Drainage River in Tianjin City, China, Chemosphere, 2007, 68(9), 1772–1778,  DOI:10.1016/j.chemosphere.2007.03.061.
23. J. Liu, Y. Li, B. Zhang, J. Cao, Z. Cao and J. Domagalski, Ecological risk of heavy metals in sediments of the Luan River source water, Ecotoxicology, 2009, 18(6), 748–758,  DOI:10.1007/s10646-009-0345-y.
24. J. Liu, P. Yin, B. Chen, F. Gao, H. Song and M. Li, Distribution and contamination assessment of heavy metals in surface sediments of the Luanhe River estuary, northwest of the Bohai Sea, Mar. Pollut. Bull., 2016, 109(1), 633–639,  DOI:10.1016/j.marpolbul.2016.05.020.
25. M. Liu, J. Chen, X. Sun, Z. Hu and D. Fan, Accumulation and transformation of heavy metals in surface sediments from the Yangtze River estuary to the East China Sea shelf, Environ. Pollut., 2019, 245, 111–121,  DOI:10.1016/j.envpol.2018.10.128.
26. Y. Liu, Y. Yuan, L. Zhang and T. Du, Exploring the differences of moisture traceability methods based on MixSIAR model under different nitrogen applications of wheat in the arid region of northwest China, Agric. Water Manag., 2024, 294, 108716,  DOI:10.1016/j.agwat.2024.108716.
27. M. Mariyanto, M. F. Amir, W. Utama, A. M. Hamdan, S. Bijaksana, A. Pratama, R. Yunginger and S. Sudarningsih, Heavy metal contents and magnetic properties of surface sediments in volcanic and tropical environment from Brantas River, Jawa Timur Province, Indonesia, Sci. Total Environ., 2019, 675, 632–641,  DOI:10.1016/j.scitotenv.2019.04.244.
28. A. Müller, H. Österlund, J. Marsalek and M. Viklander, The pollution conveyed by urban runoff: A review of sources, Sci. Total Environ., 2020, 709, 136125,  DOI:10.1016/j.scitotenv.2019.136125.
29. B. Parinet, A. Lhote and B. Legube, Principal component analysis: An appropriate tool for water quality evaluation and management—application to a tropical lake system, Ecol. Modell., 2004, 178(3–4), 295–311,  DOI:10.1016/j.ecolmodel.2004.03.007.
30. E. Resongles, C. Casiot, R. Freydier, L. Dezileau, J. Viers and F. Elbaz-Poulichet, Persisting impact of historical mining activity to metal (Pb, Zn, Cd, Tl, Hg) and metalloid (As, Sb) enrichment in sediments of the Gardon River, southern France, Sci. Total Environ., 2014, 481, 509–521,  DOI:10.1016/j.scitotenv.2014.02.078.
31. N. Reutova, A. Seleznev, I. Yarmoshenko, M. Chervyakovskaya and M. Streletskaya, Stable lead isotopic ratios as indicator of urban geochemical processes, IOP Conf. Ser. Earth Environ. Sci., 2021, 906(1), 012098,  DOI:10.1088/1755-1315/906/1/012098.
32. C. Rezza, S. Albanese, R. Ayuso, A. Lima, J. Sorvari and B. De Vivo, Geochemical and Pb isotopic characterization of soil, groundwater, human hair, and corn samples from the Domizio Flegreo and Agro Aversano area (Campania region, Italy), J. Geochem. Explor., 2018, 184, 318–332,  DOI:10.1016/j.gexplo.2017.01.007.
33. R. Setia, S. S. Dhaliwal, V. Kumar, R. Singh, S. S. Kukal and B. Pateriya, Assessment of environmental and toxicity impacts and potential health hazards of heavy metals pollution of agricultural drainage adjacent to industrial zones in Egypt, Chemosphere, 2020, 318, 137872,  DOI:10.1016/j.chemosphere.2023.137872.
34. B. H. Shomar, Trace elements in major solid-pesticides used in the Gaza Strip, Chemosphere, 2006, 65(5), 898–905,  DOI:10.1016/j.chemosphere.2006.03.004.
35. K. P. Singh, D. Mohan, V. K. Singh and A. Malik, Studies on distribution and fractionation of heavy metals in Gomti River sediments—A tributary of the Ganges, India, J. Hydrol., 2005, 312(1–4), 14–27,  DOI:10.1016/j.jhydrol.2005.01.021.
36. I. D. C. Souza, H. P. Arrivabene, C.-A. Craig, A. J. Midwood, B. Thornton, S. T. Matsumoto, M. Elliott, D. A. Wunderlin, M. V. Monferrán and M. N. Fernandes, Interrogating pollution sources in a mangrove food web using multiple stable isotopes, Sci. Total Environ., 2018, 640–641, 501–511,  DOI:10.1016/j.scitotenv.2018.05.302.
37. Z. Tao, G. Peng, F. Chen, Q. Guo, R. Wei, J. Liu, Y. Li, H. Jiang, Y. Liao and K. Pan, Isotopic signatures unveil the lead sources and migration in surface mangrove sediments, Sci. Total Environ., 2023, 892, 164765,  DOI:10.1016/j.scitotenv.2023.164765.
38. M. Thévenot, J.-F. Brunel, F. Brunel, M. Briatte, Y. Desplanques and P. Dufrénoy, Identification of particle emission mechanisms during dry friction with thermomechanical operando monitoring, Wear, 2025, 205974,  DOI:10.1016/j.wear.2025.205974.
39. M. Tsangas, I. Papamichael, D. Banti, P. Samaras and A. A. Zorpas, LCA of municipal wastewater treatment, Chemosphere, 2023, 341, 139952,  DOI:10.1016/j.chemosphere.2023.139952.
40. M. Varol, Assessment of heavy metal contamination in sediments of the Tigris River (Turkey) using pollution indices and multivariate statistical techniques, J. Hazard. Mater., 2011, 195, 355–364,  DOI:10.1016/j.jhazmat.2011.08.051.
41. B. Wang, D. Xia, Y. Yu, J. Jia, Y. Nie and X. Wang, Detecting the sensitivity of magnetic response on different pollution sources – a case study from typical mining cities in northwestern China, Environ. Pollut., 2015, 207, 288–298,  DOI:10.1016/j.envpol.2015.08.041.
42. L. Wang, S. Gao, X. Yin, X. Yu and L. Luan, Arsenic accumulation, distribution and source analysis of rice in a typical growing area in north China, Ecotoxicol. Environ. Saf., 2019, 167, 429–434,  DOI:10.1016/j.ecoenv.2018.10.015.
43. S.-L. Wang, X.-R. Xu, Y.-X. Sun, J.-L. Liu and H.-B. Li, Heavy metal pollution in coastal areas of south China: A review, Mar. Pollut. Bull., 2013, 76(1–2), 7–15,  DOI:10.1016/j.marpolbul.2013.08.025.
44. R. Wei, F. Ge, S. Huang, M. Chen and R. Wang, Occurrence of veterinary antibiotics in animal wastewater and surface water around farms in Jiangsu province, China, Chemosphere, 2011, 82(10), 1408–1414,  DOI:10.1016/j.chemosphere.2010.11.067.
45. X. Yao, Z. Wang, D. Li, H. Sun, C. Ren, Y. Yu, F. Pei and Y. Li, Distribution, mobilization, risk assessment and source identification of heavy metals and surface sediments of three urban-rural rivers after long-term water pollution treatment, Sci. Total Environ., 2024, 932, 172894,  DOI:10.1016/j.scitotenv.2024.172894.
46. L. Yu, T. Zheng, X. Zheng, Y. Yuan and R. Yuan, Nitrate source apportionment in groundwater using Bayesian isotope mixing model based on nitrogen isotope fractionation, Sci. Total Environ., 2020, 718, 137242,  DOI:10.1016/j.scitotenv.2020.137242.
47. C. Zhang, Q. Qiao, J. D. A. Piper and B. Huang, Assessment of heavy metal pollution from a Fe-smelting plant in urban river sediments using environmental magnetic and geochemical methods, Environ. Pollut., 2011, 159(10), 3057–3070,  DOI:10.1016/j.envpol.2011.04.006.
48. C. Zhang, W. Rao, Z. Wu, F. Zheng, T. Li, C. Li, X. Lei, H. Xie and X. Chu, Anthropogenic impacts and quantitative sources of nitrate in a rural-urban canal using a combined PMF, δ15N/δ18O–NO3-, and MixSIAR approach, Environ. Res., 2024, 251, 118587,  DOI:10.1016/j.envres.2024.118587.
49. Z. Zhang, Y. Lu, H. Li, Y. Tu, B. Liu and Z. Yang, Assessment of heavy metal contamination, distribution and source identification in the sediments from the Zijiang River, China, Sci. Total Environ., 2018, 645, 235–243,  DOI:10.1016/j.scitotenv.2018.07.026.
50. K. Zhao, K. Bao, Y. Yan, Y. Zhang, H. Huang, L. Li and W. Jia, Spatial distribution of potentially harmful trace elements and ecological risk assessment in Zhanjiang mangrove wetland, south China, Mar. Pollut. Bull., 2022, 182, 114033,  DOI:10.1016/j.marpolbul.2022.114033.
51. Z. Zhao, M. Zhang, Y. Chen, Y. Yu, Y. Wang, F. Li, J. Liu, L. Zhang and W. Zhai, Traceability of nitrate polluted hotspots in plain river networks of the Yangtze River delta by nitrogen and oxygen isotopes coupling Bayesian model, Environ. Pollut., 2022, 315, 120438,  DOI:10.1016/j.envpol.2022.120438.
52. Z.-Q. Zhao, W. Zhang, X.-D. Li, Z. Yang, H.-Y. Zheng, H. Ding, Q.-L. Wang, J. Xiao and P.-Q. Fu, Atmospheric lead in urban Guiyang, southwest China: Isotopic source signatures, Atmos. Environ., 2015, 115, 163–169,  DOI:10.1016/j.atmosenv.2015.05.049.
53. W. Zhou, R. Yu, F. Guo, C. Shen, Y. Liu and Y. Huang, Source apportionment and risk assessment of soil heavy metals in the Huangshui River basin using a hybrid model, Ecol. Indic., 2024, 160, 111906,  DOI:10.1016/j.ecolind.2024.111906.
54. A. Zhu, J. Liu, S. Qiao and H. Zhang, Distribution and assessment of heavy metals in surface sediments from the Bohai Sea of China, Mar. Pollut. Bull., 2020, 153, 110901,  DOI:10.1016/j.marpolbul.2020.110901.
55. Y. B. Guan, C. F. Li, X. M. Cheng, Y. Ba and L. X. Cui, Pollution status and ecological risk assessment of a section of the Shaying River basin, J. Environ. Occup. Med., 2013, 30(11), 825–828,  DOI:10.13213/j.cnki.jeom.2013.11.002.
56. M. Zhong, Y. Wan, A. Wan, J. Hu, S. An and X. Leng, Characteristics and ecological risk assessment of heavy metal pollution in sediments of the Shaying River basin, Chin. J. Ecol., 2016, 35(7), 1857–1864,  DOI:10.13292/j.1000-4890.201607.008.
57. G. Müller, Index of geoaccumulation in sediments of the Rhine River, GeoJ., 1969, 2, 108–118.
58. Y. Q. Ji, Using geoaccumulation index to study source profiles of soil dust in China, J. Environ. Sci., 2008, 20(5), 571–578.
59. H. Zhao, Y. Wu, X. Lan, Y. Yang, W. Zhang, S. Xie and L. Wu, Comprehensive assessment of harmful heavy metals in contaminated soil in order to score pollution level, Sci. Rep., 2022, 12, 3552,  DOI:10.1038/s41598-022-07602-9.
60. R. S. Ahmed, M. E. Abuarab, M. M. Ibrahim, M. Baioumy and A. Mokhtar, Assessment of environmental and toxicity impacts and potential health hazards of heavy metals pollution of agricultural drainage adjacent to industrial zones in Egypt, Chemosphere, 2023, 318, 137872,  DOI:10.1016/j.chemosphere.2023.137872.
61. M. Anas, W. A. Khattak, W. Sarfraz, N. Khalid, N. Naeem, M. Farid, M. H. Saleem, K. A. Khan, A. Khalid, T. Ahmed and S. Fahad, EDTA-enhanced phytoextraction: Leveraging Ricinus communis for the decontamination of Pb and Cd in agricultural soils, Plant Growth Regul., 2025, 105(1), 215–228,  DOI:10.1007/s10725-024-01270-7.
62. H. Dong, Y. Xun, M. Yoneda and L. Feng, Risk assessment of complex contaminated sites based on the fractional transformation of metals: Model development and a case study, Sci. Total Environ., 2021, 786, 147509,  DOI:10.1016/j.scitotenv.2021.147509.
63. Z. Iannuzzi, M. C. Casado-Martinez, T. De Oliveira, D. Lebouil, B. J. D. Ferrari, J. Gasperi and B. Mourier, Tire road wear particles, microplastics and metals in sediment of stormwater detention basins: Co-occurrence and ecological risk assessment, Environ. Int., 2025, 202, 109676,  DOI:10.1016/j.envint.2025.109676.
64. H. Luo, L. Yang, C. Zhang, X. Xiao and X. Lyu, Early warning of heavy metals contamination in agricultural soils: Spatio-temporal distribution and future trends in the Hexi Corridor, Ecol. Indic., 2024, 160, 111908,  DOI:10.1016/j.ecolind.2024.111908.
65. Y. Pan, L. Ding, S. Xie, M. Zeng, J. Zhang and H. Peng, Spatiotemporal simulation, early warning, and policy recommendations of the soil heavy metal environmental capacity of the agricultural land in a typical industrial city in China: Case of Zhongshan City, J. Cleaner Prod., 2021, 285, 124849,  DOI:10.1016/j.jclepro.2020.124849.
66. S. A. Shetaia, M. Li, J. Chen, Y. Wang, M. S. Elhebiry, A. Abdelaal, H. Su, Z. Chen and A. Salem, Comprehensive assessment of remarkable heavy metals contamination in road dust in Greater Cairo: Priority control factors based on source apportionment and risk assessment, J. Environ. Manage., 2025, 394, 127607,  DOI:10.1016/j.jenvman.2025.127607.

---