Spatial Distribution and Source Apportionment of Potentially Toxic Elements in Soils Across a Full Lead–Zinc Mining–Beneficiation–Smelting–Tailings System

Abstract

Potentially toxic elements (PTE) pollution from lead–zinc (Pb–Zn) production poses significant ecological risks, requiring systematic assessment across the industrial chain. This study investigated soil, surface water, and sediments near a Pb–Zn mining area, integrating pollution indices (I_geo, NIPI, RI) with human health risk models. A spatial analysis framework was established by combining proportional symbol mapping and Thiessen polygons to analyze contamination patterns under small-sample conditions. Results showed a clear pollution hierarchy: smelting > beneficiation > tailings ≈ mining. Smelting and beneficiation zones exhibited multi-element pollution; Hazard Index (HI) exceedance probabilities reached 89% and 95%, respectively, while carcinogenic risk (CR) exceedance approached 100% across all zones. Cd was the dominant ecological risk factor, particularly in mining and tailings zones, where risk was mainly driven by a single element. Source apportionment identified two industrial groups—smelting-related (Pb, Hg, Zn, Se) and ore-associated (As, Cd, Cu, Sb)—whereas Cr, Ni, Co, and V were mainly derived from natural sources. These results indicate the need for coordinated management of beneficiation and smelting processes and provide a spatial analysis approach for small-sample assessment.

1. Introduction

Lead–zinc (Pb–Zn) mining activities represent a globally recognized source of pollution by potentially toxic elements (PTEs). The complete industrial chain—from spanning mining, mineral beneficiation, and smelting to tailings disposal—establishes multiple pathways for the release of PTEs, leading to their accumulation and dispersal in environmental media. Typical PTEs, such as Pb, Zn, and their associated elements cadmium (Cd) and arsenic (As), are characterized by their non-degradability, bioaccumulation, environmental persistence, and inherent toxicity [1]. Once sequestered in soils, sediments, and aquatic systems, these elements are highly recalcitrant, posing enduring risks to both ecological integrity and human health.

While the deleterious environmental impacts of lead–zinc mining are well-documented, extant research has predominantly concentrated on isolated industrial segments, such as smelters or tailings ponds [2,3]. There remains a conspicuous lack of systematic investigation into the entire “mining-beneficiation-smelting-tailings” industrial chain, leading to a fragmented understanding of the pollution contributions from different production stages. In particular, the beneficiation (ore dressing) stage represents a “gray zone”; its source identification and risk contribution are frequently overlooked or conflated with other processes, resulting in contentious source apportionment. Some studies suggest that soil contamination at beneficiation sites stems primarily from internal production activities—such as concentrate processing and product stockpiling [4]—while others categorize these sites under the broad umbrella of “industrial and mining sources,” failing to distinguish their contributions from upstream mining or peripheral smelting activities [5]. This lack of clarity significantly hampers the development of targeted management and control strategies for the beneficiation stage.

To quantitatively evaluate PTE contamination levels and ecological health risks, a series of established assessment methods has been established, including the Geo-accumulation Index (I_geo), Pollution Index (PI), Pollution Load Index (PLI), Nemerow Integrated Pollution Index (NIPI), and Potential Ecological Risk Index (PERI), alongside human health risk assessment models [6,7,8]. Regarding source apportionment, multivariate statistical techniques such as Principal Component Analysis (PCA), Cluster Analysis (CA), and correlation matrices are widely employed to identify PTE origins and their respective contributions [9]. For spatial characterization, geostatistical interpolation—most notably Kriging—is frequently used to delineate distribution patterns [10]. However, Kriging’s efficacy is often constrained by its rigorous requirements for sample size, making it less reliable for small-scale datasets.

To address this limitation, proportional symbol maps, introduced by Bølviken et al. and Björklund & Gustavsson (1987), link symbol diameter to metal concentration through continuous functions, enabling intuitive visualization of PTE concentrations at sampling points [11,12]. Furthermore, Thiessen polygons can help illustrate the theoretical spatial domain and relationships among sampling points [13,14]. The integration of these two methods clearly delineates the relative scale and distribution of spatial units without the need for numerical prediction in unsampled areas, thereby voiding the uncertainties associated with interpolation for sparse datasets. Despite the maturity of individual evaluation frameworks, systematic research across the entire “mining–beneficiation–smelting–tailings” chain remains scarce, with most studies focusing on isolated components like tailings ponds or smelters [15,16]. This lack of differential analysis obscures distinct pollution drivers across production stages, hindering the formulation of targeted management strategies based on zoning and classification. Notably, the combined application of proportional symbol maps and Thiessen polygons remains underexplored in PTE research, necessitating exploratory application to enhance spatial assessment precision.

Consequently, this study focuses on a representative lead–zinc mining cluster to conduct a systematic, chain-wide comparative analysis. This study aims to address the complexities of source superposition and functional differentiation that isolated studies frequently fail to resolve. The specific objectives are as follows:

(1)

To quantify the disparities in soil PTE contamination across four functional zones (mining, beneficiation, smelting, and tailings) and reveal their spatial patterns and heterogeneity.

(2)

To assess the adult occupational health risks (both carcinogenic and non-carcinogenic) associated with exceeding PTE levels in soils in the beneficiation and smelting zones and identify the primary exposure pathways.

(3)

To qualitatively and quantitatively apportion soil PTE sources and analyze their correlation with soil physicochemical properties to elucidate the critical factors governing environmental behavior.

(4)

To formulate differentiated risk management strategies based on functional zoning.

This study tests the following hypotheses:

(1)

Significant differences exist in the pollution levels, source compositions, and health risks of PTEs across different functional zones of the industrial chain.

(2)

Adult occupational health risks associated with the beneficiation stage have long been underestimated.

(3)

In scenarios with limited sample sizes, the integration of proportional symbol maps and Thiessen polygons can effectively reveal spatial distribution patterns of PTEs without the need for interpolation.

The expected outcomes of this research include:

At the regional scale, it will deepen the understanding of the spatial patterns of multi-metal composite pollution. At the industrial scale, it aims to promote a shift in the mineral processing sector from “neglect” to “integrated management”. Methodologically, it provides a visualization approach without interpolation (proportional symbol maps combined with Thiessen polygons) for analyzing heavy metal spatial distribution under small-sample constraints.

2. Materials and Methods

2.1. Study Area and Sampling

The study area is situated in the southern portion of Feng County, nestled within the middle-to-low mountainous terrain of the southern Qinling Mountains. This region is characterized by a warm-temperate humid continental monsoon climate, with an average annual temperature of 10.4 °C. The mean annual precipitation is 647 mm, distributed over an average of 115.5 rainy days per year. The study area is located in the southern Qinling orogenic belt, where the exposed strata are mainly carbonate rocks (limestone and dolomite) interbedded with clastic rocks. The primary ore minerals are galena and sphalerite, with gangue minerals dominated by calcite and dolomite. This geological setting provides a natural source for Pb, Zn, and related potentially toxic elements (PTEs) in the soils. The sampling campaign was conducted in November 2022. At that time, the mining zone and tailings zone were under active operation, while the beneficiation zone and smelting zone were no longer in operation. Due to limited historical records, the exact starting or cessation years of these activities cannot be provided. Sampling sites were strategically positioned beyond the peripheral boundaries of four distinct functional zones to cover the entire industrial chain. A total of 22 soil samples were collected, comprising 7 from the mining zone, 7 from the beneficiation zone, 5 from the smelting zone, and 3 from the tailings zone. Additionally, two water samples and two sediment samples were collected from the natural river located near the boundary between the beneficiation and smelting zones (as shown in Figure 1d), not from any runoff discharge outlet. The two sampling points for water and sediment were approximately 3 m apart, which can be considered as parallel samples taken at essentially the same location on the scale of water flow and sediment distribution; therefore, they are not differentiated in the figure. Due to topographical constraints (steep slopes and cemented surfaces), soil sampling points were placed in as many directions as possible. Surface soil samples (0–20 cm) were collected at each sampling site using a plum-blossom (five-point) method, and all soil, water, and sediment samples were immediately stored at 4 °C during transport and prior to analysis.

2.2. Sample Analysis and Quality Control

2.2.1. Physicochemical Analysis of Soil and Sediments

The pH of soil and sediment samples was determined potentiometrically using a pH/mV/conductivity meter (Model SX723, Shanghai Sanxin Instrument Factory, Shanghai, China). The cation exchange capacity (CEC) was measured via the hexamminecobalt(III) chloride extraction-spectrophotometric method using a visible spectrophotometer (UV2600A, Shanghai Sunny Hengping Scientific Instrument Co., Ltd., Shanghai, China). Total organic carbon (TOC) was quantified through the combustion oxidation-non-dispersive infrared method (Elementar Vario TOC, Elementar Analysensysteme GmbH, Langenselbold, Germany). A total of thirteen PTEs were analyzed in this study. Pb and Cd were determined by graphite furnace atomic absorption spectrophotometry (GFAAS, ZEEnit 700p, Analytik Jena AG, Jena, Germany). Chromium (Cr), copper (Cu), Zn, and nickel (Ni) were measured using flame atomic absorption spectrophotometry (FAAS, iCE 3500, Thermo Fisher Scientific, Waltham, MA, USA). Hexavalent chromium (Cr(VI)) was extracted with an alkaline solution and subsequently analyzed by FAAS. Mercury (Hg), As, selenium (Se), antimony (Sb), and bismuth (Bi) were determined by atomic fluorescence spectrometry (AFS-9130, Beijing Jitian Instruments Co., Beijing, China) following microwave digestion, and the LOD for Sb was 0.2 μg/L. Cobalt (Co) and vanadium (V) concentrations were analyzed via inductively coupled plasma mass spectrometry (ICP-MS, Agilent 7800, Agilent Technologies, Santa Clara, CA, USA) after aqua regia extraction. To ensure data reliability, all chemical reagents used were of analytical grade and met national standards. Quality assurance and quality control (QA/QC) were strictly maintained using certified reference materials (GBW07405 [17]). The relative standard deviation (RSD) for all analytical results ranged from 3% to 5%.

2.2.2. Surface Water Analysis

The pH of surface water was measured potentiometrically. Concentrations of Pb and Cd were determined by inductively coupled plasma mass spectrometry (ICP-MS, Agilent 7800, Agilent Technologies, Santa Clara, CA, USA). Cr, Cu, and Zn were analyzed using inductively coupled plasma optical emission spectrometry (ICP-OES, iCAP 7600, Thermo Fisher Scientific, Waltham, MA, USA). Hg, As, and Sb were quantified via atomic fluorescence spectrometry (AFS-9130, Beijing Jitian Instruments Co., Ltd., Beijing, China). For (VI), the 1,5-diphenylcarbazide spectrophotometric method was employed, utilizing a multi-parameter spectrophotometer (DR6000, Hach, Loveland, CO, USA).

2.3. Evaluation Methods and Models

2.3.1. Assessment of Contamination Levels

Geo-Accumulation Index and Pollution Load Index

The I_geo is a widely utilized metric for evaluating the degree of soil PTE contamination and its historical accumulation effects. Complementarily, the PLI is employed to characterize the overall pollution load and intensity of soil in a given region. The calculation formulas for I_geo and PLI are presented in Equations (1)–(3) [7,18].

$${\mathrm{I}}_{\mathrm{g}\mathrm{e}\mathrm{o}}={\log }_2\left({\displaystyle \frac{C_n}{{kB}_n}}\right)$$

(1)

$$CF={\displaystyle \frac{C_n}{B_n}}$$

(2)

$$PLI=\sqrt[\mathrm{n}]{{CF}_1\times {CF}_2\times \cdots \times {CF}_n}$$

(3)

In these equations, C_n represents the measured concentration of PTE n in the soil; B_n denotes the corresponding geochemical background value [19]; and CF is the contamination factor. The constant k is typically assigned a value of 1.5 to account for potential variations in background values caused by lithogenic effects. The classification criteria for I_geo and PLI are presented in Table S1 and Table S2, respectively.

Single-Factor Pollution and Nemerow Integrated Pollution Indices

The PI is a parameter utilized to evaluate the contamination degree of environmental media by a specific pollutant. To complement the insights from I_geo and PLI, this study integrates PLI and NIPI for a holistic assessment. This combined approach allows for a balanced consideration of the overall pollution level and provides an early warning for “hotspots” characterized by exceptionally high concentrations. The calculation formulas for PI and NIPI are defined in Equations (4) and (5), respectively [18].

$$\mathrm{P}\mathrm{I}={\displaystyle \frac{{\mathrm{C}}_{\mathrm{n}}}{C_{0,\mathrm{n}}}}$$

(4)

$$\mathrm{N}\mathrm{I}\mathrm{P}\mathrm{I}=\sqrt{{\displaystyle \frac{{{PI}_{\mathrm{A}\mathrm{v}\mathrm{g}}}^2+{{PI}_{\mathrm{M}\mathrm{a}\mathrm{x}}}^2}{2}}}$$

(5)

where C_n represents the measured concentration of PTE n, and C_0,n denotes the soil environmental standard value, which is derived from the Chinese national guidelines [20]. The exceedance rates for PI were calculated based on the Risk Screening Values (RSV) and Risk Intervention Values (RIV) for Category II development land, as stipulated in the Soil Environmental Quality Control Standard for Soil Pollution of Development Land (GB 36600-2018) [20]. PI_Avg and PI_Max represent the average and maximum PI values, respectively. The classification thresholds for PI and NIPI are detailed in Supplementary Tables S3 and S4.

2.3.2. Potential Ecological Risk Assessment

The PERI was employed to evaluate the ecological hazards posed by PTEs in both soil and sediment [21]. This method, originally proposed by Hakanson [21], incorporates both the toxicity and environmental sensitivity of pollutants. The formula is as follows:

$$\mathrm{P}\mathrm{E}\mathrm{R}\mathrm{I}={\displaystyle \sum _{i=1}^n}{E}_{\mathrm{r}}^i={\displaystyle \sum _{i=1}^n}{(T}_{\mathrm{r}}^i{\times C}_{\mathrm{f}}^i)={\displaystyle \sum _{i=1}^n}{(T}_{\mathrm{r}}^i\times {\displaystyle \frac{C^i}{C_{\mathrm{n}}^i}})$$

(6)

where Cⁱ represents the measured concentration of the PTE; C_nⁱ is the geochemical background value; C_fⁱ is the contamination factor; and T_rⁱ denotes the toxic response factor. Following the frameworks established by Hakanson and Ma et al. [21,22], the toxic response factors (T_rⁱ) were assigned as follows: Hg = 40; Cd = 30; As = 10; Se = 10 (analogous to As in toxicity); Sb = 10 (congeneric with As); Pb = Cu = Ni = Co = 5; Bi = 3 (positioned between Cr and Pb); Cr = V = 2; and Zn = 1.

2.3.3. Occupational Health Risk Assessment

This study utilized the human health risk assessment (HHRA) model recommended by the United States Environmental Protection Agency (USEPA) [23] to evaluate the non-carcinogenic and carcinogenic risks posed by soil PTEs to adults. The assessment is based on the Hazard Index (HI) and Carcinogenic Risk (CR) metrics. The USEPA model accounts for three primary exposure pathways: incidental oral ingestion (ADD_ing), dermal contact (ADD_derm), and inhalation of soil particles (ADD_inh). The governing equations are as follows [23]:

$${ADD}_{\mathrm{i}\mathrm{n}\mathrm{g}}={\displaystyle \frac{C_{\mathrm{s}}\times \mathrm{I}\mathrm{n}\mathrm{g}\mathrm{R}\times \mathrm{E}\mathrm{F}\times \mathrm{E}\mathrm{D}}{\mathrm{B}\mathrm{W}\times \mathrm{A}\mathrm{T}\times {10}^6}}$$

(7)

$${ADD}_{\mathrm{d}\mathrm{e}\mathrm{r}\mathrm{m}}={\displaystyle \frac{C_{\mathrm{s}}\times \mathrm{S}\mathrm{A}\times \mathrm{A}\mathrm{F}\times \mathrm{A}\mathrm{B}\mathrm{S}\times \mathrm{E}\mathrm{F}\times \mathrm{E}\mathrm{D}}{\mathrm{B}\mathrm{W}\times \mathrm{A}\mathrm{T}\times 10}}$$

(8)

$${ADD}_{\mathrm{i}\mathrm{n}\mathrm{h}}={\displaystyle \frac{C_{\mathrm{s}}\times \mathrm{I}\mathrm{n}\mathrm{h}\mathrm{R}\times \mathrm{E}\mathrm{F}\times \mathrm{E}\mathrm{D}}{\mathrm{B}\mathrm{W}\times \mathrm{A}\mathrm{T}\times \mathrm{P}\mathrm{E}\mathrm{F}}}$$

(9)

The parameters used in the equations are summarized in

Table 1. Parameters for calculating average daily dose (ADD) via ingestion, dermal contact, and inhalation for adults.

Parameter	Symbol	Unit	Value in This Study	References
Soil/dust ingestion rate	IngR	mg·day−1	100	[24]
Inhalation rate	InhR	m3·day−1	14.5	[25]
Exposure frequency	EF	days/year−1	350
Exposure duration	ED	years	25
Body weight	BW	kg	61.2	[26]
Averaging time	AT	days	Carcinogens: 25,550 (70 years × 365); Non-carcinogens: 9125 (365 × ED)	[24]
Particulate emission factor	PEF	m3·kg−1	1,360,000,000
Exposed skin surface area	SA	cm2	5700
Skin adherence factor	AF/SAF	mg·cm−2	0.2	[25]
Dermal absorption fraction	ABS	Dimensionless	As: 0.03; Other metals: 0.001	[24,27]

. C_s represents the measured concentration of PTEs in the soil (mg·kg⁻¹). To assess health risks under a Reasonable Maximum Exposure (RME) scenario, the 95% Upper Confidence Limit (UCL) of the soil PTE concentration was adopted as the Exposure Point Concentration (EPC). These values were calculated using USEPA ProUCL 5.2 software, which selected the optimal 95% UCL estimators based on the log-normal distribution characteristics of the dataset.

The Hazard Quotient (HQ) was utilized to characterize the non-carcinogenic risk posed by an individual pollutant via a single exposure route, as defined by the following equation [23]:

$$HQ={\displaystyle \frac{ADD}{\mathrm{R}\mathrm{f}\mathrm{D}}}$$

(10)

where RfD denotes the reference dose for a specific element via each exposure route (mg·(mg·(kg·d)⁻¹)). To account for the cumulative non-carcinogenic risk arising from multiple pollutants and diverse exposure pathways, the HI was calculated according to the following equation [23]:

$$HI={\sum }_{i=1}^{n}H{Q}_i={\sum }_{i=1}^n{\displaystyle \frac{AD{D}_i}{\mathrm{R}\mathrm{f}{\mathrm{D}}_i}}$$

(11)

An HQ or HI exceeding unity (>1) signifies a significant potential for adverse non-carcinogenic health effects, whereas values below this threshold suggest no significant risk. CR represents the incremental probability of an individual developing cancer over a lifetime of exposure to potential carcinogens, which is quantified as follows [23]:

$$CR=ADD\times SF$$

(12)

where SF denotes the carcinogenic slope factor for each element via its respective exposure pathway. The specific toxicological parameters for both RfD and SF used in these equations are summarized in

Table 2. Reference dose (RfD) and slope factor (SF) parameters for different exposure routes in the health risk assessment model.

Element	Exposure Pathway	RfD (mg·kg−1·d−1)	References	SF (kg·d·mg−1)	References
As	ingestion	6.00 × 10−5	[28]	32.0	[28]
	inhalation	4.29 × 10−6	[29]	18.1	[30]
	dermal contact	6.00 × 10−5	[28]	32.0	[28]
Pb	ingestion	3.50 × 10−3	[28]	-	-
	inhalation	3.52 × 10−3	[29]	6.30 *	[29]
	dermal contact	3.50 × 10−3	[28]	-	-

* Given that the USEPA IRIS database ceased to provide direct Inhalation Slope Factors in 1991, only the Inhalation Unit Risk (IUR, 4.30 × 10⁻³) was available. This necessitated the conversion of IUR to SF using the relationship SF = (IUR × BW × 1000)/InhR, where SF is expressed in kg·day·mg⁻¹.

.

2.3.4. Spatial Analysis

To visualize the spatial distribution of PTEs, proportional symbol maps, with symbol area proportional to concentration, were integrated with Thiessen polygons to partition the study area. This combined mapping approach effectively illustrates both point-specific pollution intensity and broader regional differentiation. Distance gradient regression analysis was utilized to quantitatively model the attenuation of pollutant concentrations as a function of spatial distance, thereby elucidating regional spatial patterns. Due to the limited sample size within the tailings zone, this area was excluded from quantitative distance-decay modeling and was represented solely through spatial visualization. Prior to characterizing the regional distance-decay patterns, outliers were identified using Tukey’s boxplot method [31,32], with the upper inner fence defined as Q₃ + 1.5 × IQR. Based on the dataset, the Q₃ of the NIPI was 3.16, and the IQR was 2.64, yielding an upper threshold of 7.12. A single sampling point (NIPI = 17.2) significantly exceeded this threshold and was identified as an outlier. This point was retained in the broader analysis, as its extreme value reflects a localized “pollution hotspot”; the remaining data points were used for distance-based regression.

Given the pervasive influence of mining activities, undisturbed background sampling sites were not available. Consequently, a “relative background value” was established using the mean NIPI (1.05 ± 0.23) of all sampling points situated more than 500 m from the beneficiation and smelting facilities (n = 12). It should be noted that this relative baseline may be slightly higher than the actual natural background; thus, the estimated influence range in this study should be interpreted as the “maximum impact distance.”

2.3.5. Source Apportionment of PTEs

Spearman’s rank correlation coefficients were calculated to assess the co-variation between elements, visualized through correlation heatmaps. Hierarchical Cluster Analysis (HCA) was performed using Ward’s method based on correlation distances to generate dendrograms. PCA was conducted using a correlation matrix to identify primary pollution sources and their spatial characteristics. PCA loading and score plots were utilized to characterize these features, with score plots categorized by functional zones and overlaid with 95% confidence ellipses to distinguish regional clusters.

2.4. Statistical Analysis and Quality Assurance

Descriptive statistics, PCA, and correlation analysis were performed using Origin 2026 (OriginLab, Northampton, MA, USA). Statistical significance was defined by a two-tailed p-value < 0.05. Map generation, including sampling site distribution and spatial distribution mapping, was conducted in ArcGIS 10.5 (ESRI, Redlands, CA, USA). Data processing was managed via Microsoft Excel 2010. Kruskal–Wallis H tests followed by Dunn’s multiple comparison tests were executed using GraphPad Prism 9.5 (GraphPad Software, San Diego, CA, USA). Additionally, statistical distributions were verified using ProUCL 5.1 (US EPA, Washington, DC, USA) with a significance level of α = 0.05. For samples with concentrations below the Method Limit of Detection (LOD), the following data imputation and exclusion strategies were adopted based on the analytical objectives:

(1)

For descriptive statistics, potential ecological risk assessments, and spatial analysis, values below the LOD were replaced with 1/2 LOD.

(2)

For pollution indexing, Cr, Zn, Se, and Bi were excluded from PI and NIPI calculations due to the absence of corresponding thresholds in soil risk standards for development land. Samples with non-detected Se and Bi were excluded from I_geo, CF, and PLI calculations. Furthermore, samples with non-applicable coefficients of variation (CV) were omitted from the bubble plots.

(3)

For multivariate statistics (correlation, HCA, and PCA), Se, which had a detection rate exceeding 50%, was included after 1/2 LOD imputation. In contrast, Bi was excluded from these analyses due to its detection rate being below 50%. All multivariate analyses were performed on natural logarithm-transformed values to approximate a normal distribution and ensure the robustness of the statistical outputs.

3. Results

3.1. PTE Concentrations in Soil, Sediment, and Surface Water, and Soil Physicochemical Properties

Soil samples within the resource development areas of the investigated Pb–Zn mining cluster exhibit pronounced accumulation characteristics of PTEs. The descriptive statistics for soil PTE concentrations across the different functional zones are detailed in Table S6. Notably, Cr(VI) concentrations at all sampling sites were below the LOD and were thus excluded from subsequent analysis and evaluation. Compared to national background values, Cd, Pb, Zn, and Hg emerged as the primary contaminants in the study area. In the smelting zone, their mean concentrations surpassed the national background levels by factors of 434.5, 205.3, 21.2, and 20.0, respectively. Similarly, in the beneficiation zone, these elements were 305.2, 101.7, 18.7, and 16.2 times higher than the baseline. Apart from Hg, Se, and Bi in the mining and tailings zones, which remained within local background levels, PTE concentrations across all functional zones consistently exceeded the national background values.

Regarding national regulatory thresholds, the average Pb concentrations in the smelting and beneficiation zones were 5337 and 2644 mg·kg⁻¹, respectively, exceeding the RIV by factors of 2.1 and 1.1. Similarly, the mean As concentration in the smelting zone (59.33 mg·kg⁻¹) surpassed the RSV by 1.0-fold. Globally, a meta-analysis of Pb–Zn smelting areas reported that Pb and As are the main contaminants, with mean Pb concentrations often exceeding 5000 mg·kg⁻¹ in smelting zones [33], which is comparable to our results (5337 mg·kg⁻¹). Notably, while the maximum Cd concentration in the smelting zone reached 92.19 mg·kg⁻¹ (1.42 times the RSV), its average concentration (42.15 mg·kg⁻¹) remained below the screening threshold. The mean concentrations of other PTEs did not exceed their respective RSVs across any functional zones. High spatial variability was observed, as evidenced by the standard deviation (SD) exceeding 30% of the mean for several elements: Pb, Hg, and Se across all four zones; Cu, Zn, As, Sb, and Bi in the smelting and beneficiation zones; and Cd specifically in the smelting zone. Such significant variance indicates that the PTE distribution within these functional areas is likely driven by pronounced anthropogenic disturbances or spatial heterogeneity.

In summary, the pollution intensity across the functional zones follows the order of: smelting zone > beneficiation zone > tailings zone ≈ mining zone. Pb and As were identified as the representative PTEs in the study area, characterized by a “dual-exceedance” pattern whereby they surpass both the national geochemical background levels and the regulatory thresholds.

Regarding the aquatic environment, the average pH values for sediments and surface water were 8.60 and 8.02, respectively, indicating an alkaline nature. In surface water, concentrations of Cd, Pb, Cr, Cu, Zn, Hg, and As remained below the method quantification limits. Detected levels of Sb (3.45 μg·L⁻¹) and Cr (VI) (<0.004 mg·L⁻¹) were very low, suggesting that water quality has not been significantly impaired by PTEs. In the sediment, Pb, Zn, and V exhibited relatively higher concentrations (means: 89.4, 248.5, and 86.1 mg·kg⁻¹, respectively), while Cr, Ni, Co, Cu, As, and Sb were present at moderate levels (means: 51, 31, 12.15, 24, 22.6, and 1.97 mg·kg⁻¹, respectively). Conversely, Cd, Se, and Hg were found in lower concentrations (means: 0.935, 0.07, and 0.054 mg·kg⁻¹, respectively), and both Cr(VI) and Bi were below detection limits. Notably, PTE concentrations in both sediment and surface water remained within regulatory limits. This indicates that sediment and surface water pollution represent a subsequent stage in the evolution of soil pollution. The findings suggest that PTE pollution in this study area is currently centered within the soil mantle and has not yet deteriorated to the stage of widespread aquatic degradation.

Statistical results for soil physicochemical properties are summarized in Table S7. The smelting zone exhibited the highest pH (mean: 8.33), characterized as alkaline, while the mining zone showed the lowest values (mean: 7.78), indicating a slightly alkaline nature. This alkaline pH is consistent with other Pb–Zn smelting sites in China, such as Zhuzhou (pH 6.47–8.50) and Gansu (alkaline) [34,35]. This alkaline condition is primarily attributed to the carbonate-rich parent material, with further elevation in the smelting zone likely due to intensive industrial activities (e.g., dust deposition and alkaline waste seepage). Regarding CEC, the smelting zone recorded the lowest mean (9.7 cmol(+)·kg⁻¹), whereas the tailings zone reached the highest (24.93 cmol(+)·kg⁻¹). TOC was markedly higher in the smelting zone (mean: 3.71%) compared to the relatively uniform levels observed in the mining (1.77%), beneficiation (1.83%), and tailings (1.98%) zones. This indicates that TOC is unevenly distributed across the four functional zones, and the smelting zone may have enrichment of organically bound PTEs. Furthermore, the combination of low CEC and severe contamination in the smelting zone indicates an exceptionally high source intensity and rapid input rates, implying that PTEs may exist in forms independent of cation exchange. Given the weak sequestration capacity of smelting zone soils, these areas are highly susceptible to pollution, indicating high environmental risk.

3.2. Pollution Levels

3.2.1. Pollution Assessment Based on I_geo and PLI

Overall, a positive correlation was observed between the mean I_geo values and their corresponding SDs, suggesting that higher contamination levels are generally accompanied by greater spatial variability. However, some elements deviated from this trend, reflecting distinct spatial patterns driven by varying pollution sources. As shown in Table S8 and Figure 2a, while most data points align with this positive mean–SD correlation, Pb in the beneficiation zone (mean: 5.82; SD: 1.01) stands out as a significant outlier within the “high mean-low SD” quadrant. This indicates that Pb contamination in this area has reached an extremely polluted level while remaining spatially uniform. Such a pattern is characteristic of homogenized non-point source pollution, likely resulting from the extensive atmospheric deposition of mineral dust across the beneficiation area. Furthermore, in this zone, the pollution intensities for Zn and Hg reached Class 4, while As and Sb were categorized as Class 2. Cd consistently was located in the “high mean-low SD” quadrant across all four functional zones, exhibiting low coefficients of variation (CV < 30%). This indicates a highly uniform spatial distribution of Cd, with its contamination intensity following a progressive increment along the industrial chain: mining < tailings < beneficiation < smelting. This gradient suggests that Cd is influenced by a dual mechanism: a regional background source and point-source emissions associated with smelting activities.

In the smelting zone, Pb (mean: 5.56; SD: 2.77) and Bi (mean: 5.20; SD: 4.32) occupied the “high mean-high SD” quadrant, reflecting an uneven deposition gradient primarily driven by point-source emissions from smelting stacks. Within this zone, pollution intensities for Zn, Sb, and Bi reached Class 3, while Cu, As, and Se were categorized as Class 2. In sharp contrast, Pb in the mining zone (mean: 0.33; SD: 0.71) was situated in the “low mean-low SD” region. This reveals a widespread but low-intensity anthropogenic influence, where concentrations are uniformly and marginally elevated without yet reaching the “polluted” threshold (I_geo < 0). These three distinct spatial patterns underscore the differential roles of industrial chain stages in shaping the regional pollution landscape.

The PLI boxplots (Figure 2b) further elucidate the spatial differentiation of multi-element composite pollution from a holistic perspective. Descriptive statistics of PLI indicate that the smelting zone exhibited the highest pollution load (median: 7.33), followed by the beneficiation zone (5.25), the tailings zone (2.00), and the mining zone (1.93). Regarding dispersion, the smelting zone displayed the longest interquartile range (IQR) and the widest whisker span, with PLI values ranging from approximately 2.10 to 17.47. This high degree of internal heterogeneity aligns with the high SD values observed for multiple elements in Figure 1a. Conversely, the significantly shorter boxes for the mining, beneficiation, and tailings zones indicate more concentrated PLI distributions, reflecting a relatively uniform spatial allocation of the overall pollution load.

The Kruskal–Wallis H test revealed highly significant differences in PTE concentrations across the four functional zones (H = 13.95, p = 0.0030). Post hoc Dunn’s tests further identified significant disparities specifically between the smelting and mining zones (p = 0.0218) and between the beneficiation and mining zones (p = 0.0092). No significant differences were detected between the tailings zone and any other group (all p > 0.05).

3.2.2. Pollution Assessment Based on PI and NIPI

The PI provides the most intuitive determination of regulatory compliance and serves as the direct basis for implementing national environmental management standards. Specifically, exceedance rates relative to the RSV indicate “potential risks,” whereas those exceeding the RIV define the “mandatory remediation red line.” As tabulated in Table S9 and illustrated in Figure 2c, the smelting zone exhibited mean Pb and As values of 6.672 (exceedance rate: 6) and 0.989 (exceedance rate: 1), respectively. In the beneficiation zone, Pb and As means were 3.306 (exceedance rate: 4) and 0.644 (exceedance rate: 1). Pb in both the smelting and beneficiation zones, as well as As in the smelting zone, surpassed the RIV, with specific exceedance counts of 3, 2, and 1, respectively. In contrast, no Pb or As exceedances were detected in the mining or tailings zones. These findings imply that soil Pb and As contamination in the smelting and beneficiation areas has reached a critical threshold, necessitating the initiation of stringent risk management or ecological remediation protocols. Furthermore, although the mean Cd concentration in the smelting zone was 0.648 ± 0.441, one sampling site recorded a PI of 1.41, representing a localized RSV exceedance rate of 20.0%. This suggests a localized potential Cd risk that warrants attention, although it has not yet crossed the legal threshold for mandatory restoration. For all other PTEs across the four functional zones, the mean PI exceedance rates were 0%, indicating generally high soil quality unaffected by large-scale anthropogenic pollution. From a variability perspective, the high standard deviations of Pb in the smelting (9.943) and beneficiation (1.891) zones relative to their means further support the non-uniform distribution of contaminated “hotspots.” Conversely, the generally low standard deviations for other elements indicate a high degree of spatial homogeneity in pollution levels within their respective zones.

The NIPI results indicated that the smelting zone exhibited the highest pollution intensity (4.80 ± 7.06), categorized as “heavily polluted,” followed by the beneficiation zone (2.37 ± 1.35), categorized as “moderately polluted.” In contrast, the tailings (0.29 ± 0.04) and mining zones (0.26 ± 0.03) were classified as uncontaminated. These findings are consistent with the results of the PLI assessment, highlighting the smelting zone as the area of greatest contamination and environmental risk. As shown in the boxplots in Figure 2d, the Kruskal–Wallis H test revealed highly significant differences in NIPI across the four functional zones (H = 13.39, p = 0.0039, denoted by **). Dunn’s post hoc comparisons further showed a significant disparity specifically between the beneficiation and mining zones (p = 0.0051). However, the difference between the smelting and mining zones did not reach statistical significance (p = 0.0572 > 0.05), and no significant differences were observed between the tailings zone and the other three groups (all p > 0.05). Interestingly, while both PLI and NIPI identified significant differences between the beneficiation and mining zones, a divergence occurred regarding the smelting zone: PLI showed significant differences, whereas NIPI only approached significance. This discrepancy suggests that different assessment indices possess varying sensitivities and discriminatory powers in identifying the environmental impacts of smelting activities.

3.3. Spatial Distribution Characteristics of PTEs

3.3.1. Visualization of Spatial Distribution

The NIPI values across the four functional zones exhibit pronounced spatial differentiation. The topography of the study area slopes downward from north to south, with the predominant wind direction being East-Northeast (ENE), as derived from wind frequency data (2001–2020) from the Feng County Meteorological Station. The proportional symbol map (Figure 3a) uses colored circles of varying sizes to represent NIPI values. High-value zones (NIPI > 3), shown as large colored circles, are concentrated southwest of the smelting area (downwind) and throughout the southern portion of the beneficiation zone (lower elevation). This spatial pattern aligns with the mechanisms of wind-borne transport and downwind deposition, combined with topographic accumulation. Typical stack heights for such Pb–Zn smelting facilities are 30–50 m, which allows airborne pollutants to be transported several hundred meters to a few kilometers downwind, consistent with the observed spatial pattern of high NIPI values southwest of the smelting area. These findings suggest that the spatial distribution of PTEs is driven by both atmospheric dry/wet deposition and surface runoff. Notably, NIPI values drop significantly after crossing the river, which indicates that the river serves as an effective geochemical barrier. Similarly, the low NIPI values at the westernmost sites, distal to the smelting center, conform to the distance-decay law. This evidence confirms that PTE contamination is primarily confined to the near-source areas of smelting and beneficiation facilities without trans-river dispersion. Such results provide a robust scientific basis for delineating risk management zones using river boundaries and distance-dependent attenuation. Furthermore, the consistently low NIPI in the tailings zone suggests that containment and drainage infrastructures are well-maintained and effectively sealed, preventing this area from acting as a contemporary pollution source.

3.3.2. Evaluation of Spatial Representativeness of Sampling Sites

As illustrated in Figure 3b, PTE contamination exhibits a clustered distribution concentrated near the core industrial chain (beneficiation–smelting). The beneficiation–smelting area displays a dual-core spatial configuration, indicating distinct pollution origins rather than a contiguous high-pollution plume; this underscores the multi-source nature of PTEs in the study area. The mean area of the Voronoi polygons across the three functional zones was 0.426 km², with a CV of 0.95. Applying the representativeness criteria proposed by Duyckaerts and Godefroy [36], the sampling sites in the beneficiation zone (CV = 125%) were identified as “extremely clustered,” representing only localized, fine-scale environments. In the smelting zone (CV = 73%), the distribution was “clustered,” representing general local areas. Conversely, the tailings zone (CV = 12%) exhibited a “uniform” distribution, effectively representing regional-scale characteristics. Notably, two high-value Thiessen polygons were identified: one with a value of 17.29 located immediately west of the smelting area, and another with a value of 4.17 southeast of the beneficiation area. The former is attributed to downwind deposition of smelting emissions under the prevailing southwesterly wind, while the latter is likely due to dust emissions from beneficiation activities, low-lying topography acting as a pollutant sink, and proximity to tailings transport routes. From the perspective of source identification and high-risk area assessment, the sampling strategy, characterized by the preferential allocation of sampling resources to the beneficiation and smelting zones, is scientifically justified. Although the non-uniform distribution in these core areas limits regional generalization, it facilitates the fine-grained characterization of localized “pollution hotspots.” In contrast, while the tailings zone maintains high spatial uniformity, the limited number of sampling points means that it is mainly suitable for descriptive statistics rather than spatial modeling.

3.3.3. Distance Gradient Analysis in the Beneficiation–Smelting Area

To further assess the spatial extent of contamination, distance gradient analysis was employed to quantify the influence of the beneficiation–smelting cluster on the surrounding soil. Figure 4 illustrates the regression results with and without the identified outlier. When the outlier was included, the R² was negligible (0.00003) and the slope was non-significant (p > 0.05), indicating no linear relationship between distance and NIPI, and demonstrating how localized anomalies can mask broader distance-decay patterns. Conversely, upon the removal of this outlier, the R² improved significantly to 0.422, with a significant negative correlation emerging (p < 0.01) and a slope of −0.00265. While the specific origin of this outlier remains undetermined, its presence indicates a localized “pollution hotspot.” The refined regression model following outlier exclusion is expressed as NIPI = 2.62 − 0.00265 × D, where D represents the minimum distance from the sampling site to the nearest beneficiation or smelting facility.

These results indicate a significant negative correlation between the NIPI and distance, demonstrating a clear distance-decay pattern with increasing distance from the beneficiation–smelting area. The NIPI decreased to the background level (1.05) at a distance of approximately 592 m. Consequently, the primary impact radius of PTE contamination originating from the beneficiation–smelting facilities is estimated to be approximately 600 m.

3.4. Potential Ecological Risk Summary

To evaluate the influence of Bi on the overall assessment, we compared the potential ecological risk coefficient (E_r) and the comprehensive potential ecological risk index (RI) both with and without Bi. Due to the low toxic response factor of Bi (T_r = 3), the cumulative toxicity factors (∑T_r) for the two scenarios were 128 and 125, respectively. Analysis using the classification method proposed by Ma et al. [22] revealed no discrepancies in the E_r and RI risk grading between the two scenarios (Table S5), confirming that the inclusion of Bi does not alter the final risk determination.

For a clearer visualization of the relative E_r contributions, the original values were log₁₀ transformed. For elements with E_r < 1 (e.g., Se and Bi), which would yield negative values after transformation, the values were set to zero for plotting purposes, reflecting their negligible ecological risk.

As illustrated by the cumulative bar heights in Figure 5, the integrated ecological risks in the smelting and beneficiation zones were significantly higher than those in the mining and tailings zones. Cd was the dominant risk factor; even its minimum Er value (3.78) exceeded the maximum values of other metals, and its contribution was consistent across the four zones. In contrast, Pb, Hg, As, Se, and V showed higher contributions in the smelting and beneficiation zones. Pb and As were relatively uniform in the beneficiation zone but varied considerably in the smelting zone. Cu, Zn, Ni, Cr, Sb, Bi, and Co contributed less to the overall risk. The log10 transformation reduced the masking effect of Cd, allowing clearer differentiation of other elements. Overall, the ecological risk in the smelting and beneficiation zones was mainly driven by Cd, Pb, Hg, and As, while other elements had limited contributions.

Figure 6a presents the integrated potential ecological risk index across the functional zones, while Figure 6b illustrates the RI following the exclusion of Cd to evaluate the relative contributions of dominant elements. After removing Cd, the smelting and beneficiation zones still belonged to the “very high risk” category, whereas the mining and tailings zones decreased to “low” and “moderate” levels, respectively. These results indicate that ecological risk in the mining and tailings zones is almost entirely driven by Cd, while the smelting and beneficiation zones show characteristics of multi-element pollution.

The RI boxplots in Figure 6c,d show that the smelting zone exhibited strong heterogeneity, whereas the other zones were relatively uniform. Inter-zonal comparisons indicated that with the inclusion of Cd, significant differences were observed between the smelting and mining zones (p = 0.0290) as well as between the beneficiation and mining zones (p = 0.0121). After excluding Cd, these differences remained statistically significant, with the p-values decreasing to 0.0189 and 0.0021, respectively. These findings indicate that the ecological risk is highest in the smelting zone, followed by the beneficiation zone, while the mining and tailings zones pose low risks.

Figure S1 presents the Spearman rank correlation matrices (panels a and c) and Pearson correlation matrices (panels b and d) derived from the potential ecological risk factors (E_r) of the PTEs. Strong positive correlations (rs > 0.95) were identified between Cd–As (0.99), Cd–Cu (0.98), Cu–As (0.98), Pb–Cu (0.97), Pb–Se (0.97), and Sb–Bi (0.96). Co was negatively correlated with most PTEs (−0.47 to −0.53), while V showed negative correlations with Cd, Cu, and Zn (−0.46 to −0.64), yet was strongly positively correlated with Cr (0.85) and Ni (0.91). These results were also supported by Pearson correlation coefficients. The high consistency between the Spearman and Pearson tests, before and after Cd exclusion, indicates reliability of the source identification and risk assessment.

3.5. Human Health Risk Assessment

Given that the industrial zones investigated function primarily as occupational environments for adults rather than residential areas, health risks for children were not assessed. Furthermore, while Cd presented a prominent ecological risk, its mean concentration remained below the national regulatory thresholds [20]. Following the relevant technical guidelines [25], Cd was excluded from the human health risk assessment, which focused on Pb and As. The assessment results are summarized in

Table 3. Summary of probabilistic human health risk assessment results across different functional zones.

Functional Zone	P(HI > 1)	P(CR > 1 × 10−4)	P(CR > 1 × 10−6)	Mean HI	95th Percentile HI	Mean CR	95th Percentile CR
Mining Zone	0.0861	1	1	0.66	1.09	6.05 × 10−4	9.59 × 10−4
Tailings Zone	0.1491	1	1	0.73	1.23	5.77 × 10−4	9.35 × 10−4
Beneficiation Zone	0.9527	1	1	2.62	5.29	1.16 × 10−3	2.26 × 10−3
Smelting Zone (Primary UCL) a	0.8901	0.999	1	5.35	19.68	1.77 × 10−3	4.37 × 10−3
Smelting Zone (Sensitivity UCL) b	0.8970	0.999	1	4.36	13.5	1.77 × 10−3	4.69 × 10−3

P denotes the exceedance probability. ^a Based on the 95% Adjusted Gamma UCL. ^b Based on the 95% Chebyshev UCL.

.

Monte Carlo simulations demonstrated convergence, with the difference between results from 5000 and 10,000 iterations restricted to approximately 9%, indicating that the 10,000-iteration results are reliable. Probabilistic results for the HI indicated that mean values in both the beneficiation and smelting zones exceeded the unity threshold (HI > 1), with exceedance probabilities reaching 95% and 89%, respectively. This signifies substantial non-carcinogenic health threats in these functional areas. Similarly, CR values across all functional zones surpassed the 10⁻⁴ threshold with nearly 100% exceedance probability, indicating unacceptable carcinogenic risks associated with the study area.

The Pb concentration data in the smelting zone exhibited a pronounced skewed distribution. To ensure statistical rigor, health risks were evaluated under two EPC scenarios: the 95% Adjusted Gamma UCL (8318 mg·kg⁻¹) and the 95% Chebyshev UCL (20,843 mg·kg⁻¹). The results showed that the probability of HI exceeding 1 (P(HI > 1)) remained consistent at 89.01% and 89.70%, respectively, while the probability of CR exceeding 10⁻⁴ (P(CR > 10⁻⁴)) was near 100% in both scenarios. The negligible difference (<1%) between these estimates confirms that the conclusion of significant health risk in the smelting zone is robust and independent of the statistical method selected for EPC estimation.

3.6. Source Apportionment Results

3.6.1. Spearman Correlation Analysis

To investigate the co-variation among PTEs, Spearman rank correlation analysis was performed on the log-transformed concentrations of 12 elements, with the results illustrated in Figure 7a. A distinct group of eight elements—As, Pb, Cd, Cu, Zn, Hg, Se, and Sb—exhibited significant and strong positive correlations (ρ ranging from 0.72 to 0.96, p < 0.001), suggesting a high degree of homology. These elements are the primary components (Pb and Zn) or associated elements (As, Cd, Cu, Hg, Se, and Sb) of lead–zinc sulfide ores.

In addition, Cr, Ni, Co, and V also showed strong positive correlations (e.g., Cr–Ni: ρ = 0.85, p < 0.001). However, this group showed weak or even negative correlations with the aforementioned eight elements (e.g., Cr–Pb: ρ = −0.31, p < 0.05). This indicates that these elements are mainly derived from natural parent materials and are less affected by anthropogenic activities.

3.6.2. Hierarchical Cluster Analysis

The results of the HCA are illustrated in the dendrogram in Figure 7b, which partitions the elements investigated into three primary clusters.

Cluster I (Pb, Hg, Zn, and Se) is characterized by the preferential linkage of Pb and Hg at a distance of 0.072, followed by the sequential integration of Zn (0.112) and Se (0.216). This cluster represents the main pollution characteristics of Pb–Zn mining and smelting, with Hg and Se as associated elements. Cluster II (As, Cd, Cu, and Sb) exhibits a tight aggregation of As and Cd (0.024) and Cu and Sb (0.047), with these two sub-groups merging at a distance of 0.147. This group represents pollution associated with beneficiation–smelting processes.

Cluster III (Cr, Ni, Co, and V) shows Cr and Ni merging at 0.053, while Co and V link at 0.124. These elements are mainly derived from natural parent materials and may be affected by minor industrial inputs (e.g., equipment abrasion), indicating a mixed natural–industrial source.

Clusters I and II converge at a distance of approximately 0.4, representing anthropogenic inputs. In contrast, Cluster III merges with the anthropogenic groups only at a significantly higher distance of 7.5—far exceeding the intra-group distances—which indicates its natural background origin.

3.6.3. Principal Component Analysis

PCA identified two primary components, accounting for a cumulative variance of 88.1%. As depicted in the loading plot (Figure 7c), PC1 (73.7% of the variance) exhibited strong positive loadings (0.15–0.28) for Pb, Sb, Hg, As, Cu, Se, and Cd, while showing negative loadings (−0.25 to −0.10) for Ni, Cr, V, and Co. Conversely, PC2 (14.4% of the variance) was characterized by moderate positive loadings (0.25–0.48) for the Ni–Cr–V–Co group.

The score plot (Figure 7d) revealed distinct spatial configurations along the PC1 axis. Samples from the smelting zone were localized entirely within the positive region (PC1 scores: 0.74–8.26). Most samples from the beneficiation zone also occupied the positive region (PC1 scores: −0.25 to 5.01), although several points clustered near the origin. In contrast, samples from the mining and tailings zones were situated exclusively within the negative region (PC1 scores: −4.13 to −1.65). Along the PC2, mining samples exhibited high dispersion, whereas beneficiation samples were predominantly on the positive side. Tailings samples clustered tightly near the origin, indicating a uniform composition.

Although the 95% prediction ellipses for the smelting zone partially overlapped with certain mining samples, one-way ANOVA confirmed highly significant disparities in PC1 scores across the four functional zones (F(3,18) = 12.37, p = 0.0001). Post hoc Tukey’s HSD tests indicated no significant difference between the smelting and beneficiation zones (p = 0.5553), suggesting comparable intensities of industrial source strength. Conversely, PC1 scores for the mining and tailings zones were significantly lower (p < 0.05), showing minimal anthropogenic influence and closer proximity to the natural background.

Overall, the PCA results are consistent with the loading analysis and HCA results, providing supporting evidence for source apportionment.

4. Discussion

4.1. Transition of Pollution Patterns: From Monoelemental Dominance to Multielemental Synergy

The ecological risk profiles in the study area exhibit a clear transition in driving mechanisms across the industrial chain. In the mining and tailings zones, the ecological risk is uniquely dominated by Cd, with risks from other elements being negligible, representing a “monoelemental dominance” pattern. Conversely, the risks in the beneficiation and smelting zones are collectively driven by Cd, Pb, Hg, and As. This shift toward a “multielemental synergistic” pattern is further corroborated by the NIPI and PLI results, which point to composite contamination. This divergence is rooted in the distinct physicochemical environments and source characteristics of each zone. The mining and tailings zones are characterized by lower TOC (≈1.83%) and slightly lower pH value. Mining activities primarily expose raw ores, such as sphalerite and galena, which are naturally enriched with Cd. The tailings consist of residual solids following mineral extraction; therefore, the release of elements in these areas is governed by the natural weathering of primary minerals. Due to its higher geochemical mobility and low natural background values, Cd acts as the main driver of ecological risk in these environments.

In contrast, beneficiation processes such as crushing and flotation redistribute multiple elements. In the smelting zone, higher TOC (3.71%) and alkaline pH (8.33) promote the complexation and stabilization of multiple PTEs, enhancing the co-enrichment of Cd, Pb, As, and Hg. In addition, smelting is a high-temperature process in which Pb, As, Cd, and Hg undergo volatilization–condensation–deposition, resulting in multi-element composite pollution.

4.2. Drivers of Integrated Pollution Load and Spatial Congruence

As the geometric mean of contamination factors (CF), the PLI is sensitive to the simultaneous increase in multiple elements but less affected by extreme values of a single element. This explains why, despite the high Pb enrichment in the beneficiation zone, its PLI did not exceed that of the smelting zone due to the limited contributions from other PTEs. Soil properties further modulated these patterns: the tailings zone exhibited the highest CEC (24.93 cmol(+)·kg⁻¹) yet the lowest PLI due to minimal pollutant input. Conversely, the smelting zone—characterized by the lowest CEC (9.7 cmol(+)·kg⁻¹) and the highest TOC (3.71%)—showed widespread, high-intensity enrichment of multiple PTEs (e.g., Pb, Cd, and Bi). These findings indicate that source intensity dominates the integrated pollution load. In the smelting zone, nearly all investigated elements significantly exceeded their respective background levels, resulting in the highest geometric mean and identifying smelting as a systemic source of synchronized multi-element accumulation.

A remarkable congruence was observed between the quantitative indices (ecological risk, integrated pollution, and health risk) and the spatial analysis results (proportional symbol mapping, Voronoi partitioning, and distance gradient regression). All methodologies yielded a consistent hierarchy of contamination intensity: smelting ≥ beneficiation > tailings ≈ mining. Notably, health risks in the smelting zone have exceeded acceptable levels, indicating the need for urgent management measures.

4.3. Composite Source Characteristics of Pollution from Beneficiation Activities

The characterization of PTE contamination within the beneficiation zone as a composite source is supported by four distinct lines of evidence:

• First item: Following the exclusion of Cd (a characteristic smelting marker), the disparity in ecological risk between the beneficiation and mining zones became more pronounced (p-value decreased from 0.0121 to 0.0021), indicating Cd enrichment in the beneficiation zone independent of smelting.
• Second item: The RI boxplots revealed an elongated interquartile range with prominent outliers (local “hotspots”) in the smelting zone, whereas the beneficiation zone exhibited a more compact and uniform distribution, indicating relatively homogeneous non-point source inputs (e.g., waste rock piles and dust).
• Third item: PCA results show that the beneficiation and smelting zones have similar industrial source intensity, indicating that beneficiation processes can generate PTE pollution comparable to smelting. Meanwhile, the tailings zone showed no significant difference in PC1 scores compared with the beneficiation zone, supporting the contribution of beneficiation activities.
• Fourth item: The beneficiation zone is located approximately 60 m east (upwind) of the smelting zone. Given the ENE prevailing wind direction, upwind transport is difficult. In addition, the high-value area in the southern beneficiation zone is spatially separated from the southwestern downwind high-value area of the smelting zone, with a low-value transition zone between them, indicating a dual-core pattern and independent pollution sources.

In summary, PTE contamination in the beneficiation zone shows composite source characteristics, driven by both intrinsic ore-dressing processes and localized spatial factors.

4.4. Source Apportionment and Dominant Factors

By synthesizing the potential Er, log-transformed Spearman correlations, HCA, and PCA, a distinct binary source configuration for soil PTEs was elucidated:

• Processing or smelting activities: Eight elements (As, Pb, Cd, Cu, Zn, Hg, Se, and Sb) exhibited strong positive inter-correlations, high PC1 loadings, and tight hierarchical clustering. Collectively, these elements characterize the contamination footprint released across the entire “mining-beneficiation-smelting” industrial cycle. HCA further subdivided this group into a smelting-related signature (Pb–Hg–Zn–Se) and an ore-associated signature (As–Cd–Cu–Sb), reflecting differences in elemental release among production stages.
• Natural Source (Pedogenic and Lithogenic Background): In contrast, the Cr–Ni–Co–V group showed strong internal correlations but negative PC1 loadings and was negatively correlated with the anthropogenic tracers. They cluster independently at large distances, indicating control by geological background and parent materials with limited anthropogenic influence. This interpretation of a natural lithogenic source is supported by studies of carbonate-hosted Pb–Zn deposits [37].
• Functional Zone Gradient: The spatial hierarchy of PC1 scores (smelting > beneficiation > mining ≈ tailings) is perfectly congruent with the patterns revealed by the multi-index evaluation (PLI, NIPI, RI) and health risk levels. This alignment confirms the smelting zone as the epicenter of anthropogenic contamination and the primary driver of multielemental composite pollution and subsequent health risks.

Multivariate statistical results are consistent with spatial analysis. PC1 (representing mining-related sources) and clustering patterns are consistent with the pollution gradients shown in the proportional symbol map, indicating that beneficiation and smelting areas are pollution sources rather than sinks. Pollution intensity decreases with distance from the industrial core. Although the tailings area shows some contamination, its lower PC1 score may be related to soil cover or weathering dilution. These results provide a basis for zoned management.

4.5. Risk Discrepancies and Management Implications

Although the alkaline pH and elevated TOC in the smelting zone theoretically enhance the sequestration of PTEs and reduce their bioavailability, the health risks for Pb and As—calculated based on total concentrations—remain extremely high (P(HI) ≈ 90%; P(CR) ≈ 100%), indicating that source intensity dominates overall risk. While the high CEC in the tailings zone acts as a potential natural geochemical barrier, management priorities must center on the active reduction in smelting emissions. Cd shows extremely high ecological risk but does not exceed national standards, reflecting differences in assessment criteria. Therefore, a tiered evaluation framework is recommended:

• Health risk standards for occupational areas;
• Ecological risk for surrounding environments;
• Long-term monitoring of sensitive elements such as Cd.

Spatial management strategies: Smelting zone—focus on emission control and soil remediation; beneficiation zone—apply a similar control level as smelting, with emphasis on dust, leakage, and runoff management; tailings zone—strengthen anti-seepage systems and ecological restoration; mining zone—standardize waste rock management to prevent leachate pollution.

Health risk control: personal protection—use protective masks to reduce dust inhalation and soil ingestion; monitoring—regularly monitor Pb, As, and Cd in soil and air; health surveillance—conduct regular health examinations.

4.6. Limitations and Future Perspectives

Despite these findings, several limitations of this study should be noted. Firstly, due to monitoring constraints, the simultaneous measurement of atmospheric heavy metal concentrations was not conducted. Therefore, the direct contribution of smelting emissions to the beneficiation zone via atmospheric dispersion could not be quantified. Future research will focus on integrating dispersion modeling to calculate deposition fluxes more accurately. Secondly, as this study was a cross-sectional investigation, it lacks time-series data and vertical soil profile analysis. This limits the ability to reconstruct the historical evolution of pollution or assess the risk of migration to groundwater. Establishing long-term monitoring plots will be essential to address these gaps. Finally, this study focused on a single representative mining cluster. Future research will expand the geographic scope to encompass additional Pb–Zn mining regions. This expansion is essential to verify the universality of the identified pollution gradients. Overall, these limitations do not affect the robustness of the core conclusions.

5. Conclusions

This study focused on the environmental media surrounding a typical Pb–Zn mining cluster in the Qinling Mountains to evaluate the impacts of the full industrial chain. The main conclusions are as follows:

• Functional zone gradients: A clear hierarchy of pollution intensity was identified: Smelting Zone (multi-elemental composite pollution, high health risk) > Beneficiation Zone (extreme Pb enrichment, moderate integrated risk) > Mining and Tailings Zones (Cd-dominated, low integrated risk). Surface water and sediments remained uncontaminated.
• Priority contaminants: Pb and As are the main pollutants in the study area. In hotspot areas, they reached high pollution levels, with As posing a carcinogenic risk and Pb posing a non-carcinogenic risk at unacceptable levels, indicating the need for priority control.
• Binary source structure: Anthropogenic industrial activities are the fundamental cause of the composite pollution of Pb, Cd, As, Zn, Hg, and Sb. Conversely, natural pedogenic processes govern the distribution of Cr, Ni, Co, and V. Anthropogenic inputs strongly superimpose on the natural background.
• Spatial patterns: The pollution footprint follows a centric-radial pattern centered on the smelting–beneficiation areas. While smelting is the epicenter of multi-elemental accumulation, the environmental contribution of the beneficiation stage is significant and non-negligible.

Overall, this research not only characterizes the “smelting-dominant, beneficiation-superimposed” risk profile of Pb–Zn mining areas but also demonstrates the regional necessity for full-chain systematic assessments. Methodologically, it establishes a novel spatial analysis combination for small-sample constraints. These findings provide a basis for pollution source identification and risk management in similar polymetallic mining areas. Despite these findings, this study has several limitations, including the use of conventional source apportionment methods (without isotope tracing), a lack of local parameter optimization for health risk assessment, limited tailings samples (n = 3), and no time-series data, cross-sectional design, topographic sampling constraints, and unknown stack heights. Additional limitations include the absence of PTE speciation analysis, insufficient discussion of soil–water–sediment transport, and control measures lacking implementation details and cost–benefit analysis. These limitations do not undermine the main conclusions but highlight directions for further research. Future studies should integrate isotopic tracers (e.g., Pb, Sr, Cd), local exposure surveys, increased tailings sampling, long-term monitoring, speciation analysis, transport process investigation, and detailed cost-effective management strategies.