Occupational Exposure to Heavy Metal(loid)-Contaminated Soil from Mining Operations: A Case Study of the Majdanpek Site, Serbia

Abstract

This study investigated the occupational hazard effects of heavy metal(loid)s (HMs) from soil in several critical mining activity areas at the Majdanpek copper mine in Serbia. Soil contamination and associated ecological and health risks to workers were evaluated through an apportionment of sources and a quantitative evaluation of ecological and health risks. The majority of soil samples had increased concentrations of Cd, Cu, Pb, Zn, Hg, As, Mo, and Sb. The results of the multivariate statistical analysis suggested the existence of multiple sources. The positive matrix factorization further explained these associations between HMs and defined three main pollution sources: natural (Factor 1), mixed source (Factor 2), and mining pollution (Factor 3). According to the RI, the average value was 1215, with more than half of the samples (57.4%) showing very high pollution levels, while 3.3% of the samples had an RI lower than 150. The ecological risk was dominated by Cd, Cu, and Hg, with Factor 3 contributing the most to the RI values. Assessment of worker exposure to soil revealed that outdoor workers had a higher potential for adverse health effects, with mean HI and TCR being 0.18 and 2.9 × 10⁻⁵, respectively. The identified sources had similar impacts on non-carcinogenic and carcinogenic risks, with a decreasing trend: Factor 3 > Factor 2 > Factor 1. Indoor workers were exposed to neither non-carcinogenic or carcinogenic risks, whereas outdoor workers suffered from possible health issues regarding TCR. Source-specific health risk assessment indicated mining pollution as the only risk contributing factor. A Monte Carlo simulation of risks revealed that the probability of developing carcinogenic issues for outdoor workers was within the safety threshold (TCR < 10⁻⁴). The findings of this study emphasize the need for regulation and control strategies for worker health risks from HM-contaminated soil in mining areas.

1. Introduction

Soil is an essential part of the environment, and its contamination with heavy metal(loid)s (HMs) has drastically increased in recent decades [1,2]. Rapid human development, including mining, transportation, intensive agriculture, urbanization, industrialization, and other anthropogenic activities, has led to the excessive release of heavy metal(loid)s (HMs) into the environment [3]. Crop development, food safety, and human health are among the environmental impacts of soil degradation. Thus, soil pollution is considered a worldwide problem, with HMs recognized as priority pollutants [4,5]. They are characterized by high toxicity, non-biodegradability, and strong accumulation in the soil. As a result, HMs may negatively impact different environmental media and deteriorate human health [6,7]. Heavy metal(loid)s can enter the food chain through soil contamination and have serious consequences on human health. Even at low concentrations, the accumulation of HMs can result in various diseases [8,9]. Although HMs may naturally originate from geogenic sources and parent materials, anthropogenic sources are the primary cause of soil contamination [10]. Identifying the potential sources of soil HMs is important for environmental protection worldwide.

Mining activities are one of the major sources of HMs accumulation in soil. Mining regions are considered one of the most contaminated areas [11]. High levels of As, Cd, Cu, Pb, and Zn are released into soil during the extraction and processing of ores through tailings, wastewater discharge, and atmospheric deposition [12,13,14]. As a result, the soil in the surrounding areas may suffer serious deterioration. Many studies have confirmed that soil in mining regions is severely polluted with HMs [15,16,17]. Therefore, it is crucial to identify the pollution sources in mining regions to develop policies and regulations for risk mitigation.

Identification of pollution sources and implementation of risk assessment methods in mining regions have become vital for the management of soil quality [18]. Multivariate statistical methods and receptor models are the main techniques used for source analyses [19,20]. Multivariate statistics can initially detect sources of HMs by evaluating their relationships. However, receptor models are more efficient and can quantify source contributions for each HM [19]. One of the most popular methods, positive matrix factorization (PMF), has been widely applied for the quantification of HM sources. This model enables the identification of the primary sources of HMs and quantifies their contribution to the concentration of each pollutant [21,22]. The ecological risk assessment determines the environmental impact of HM concentrations in soil, whereas the health risk model evaluates the possible harmful effects of HMs on the human body through non-carcinogenic and carcinogenic risks [23,24]. The conventional deterministic approach uses single-point values, which may cause unrealistic representation of results and lead to underestimation or overestimation of health risks. However, Monte Carlo simulation overcomes this limitation by including parameter uncertainty in the assessment [25]. Compared with the deterministic model, this probabilistic approach offers a more precise risk interpretation. Numerous studies have incorporated probabilistic model in their research to achieve a better understanding of HMs’ similarities, pollution sources, and potential hazards [26,27]. The best distinction between the impact of natural and anthropogenic sources on risk is achieved by combining source apportionment models with ecological and health risks. The implementation of source-specific analysis allows the determination of risk levels related to each source [28,29,30].

In diverse mining and mineral-processing contexts, modern occupational risk assessments increasingly integrate probabilistic frameworks and soil testing to better quantify multi-pathway HM exposures and attendant uncertainties [31,32]. Such integrated approaches have revealed substantial health risks to the workers. Previous studies have reported that ingestion of arsenic- and lead-contaminated soil in a major Zn–Pb mining district produced hazard indices exceeding safe thresholds [33]. Likewise, artisanal gold miners chronically exposed to As, Cd, Hg, and Pb exhibited cumulative carcinogenic risk above the 1 × 10⁻⁶ benchmark [34]. Furthermore, in industrial smelting and informal recycling settings, worker inhalation and dermal contact with metal-rich particulates have been linked to hazard quotients above one and significantly higher cancer odds ratios relative to unexposed populations [35,36].

This study investigated the ecological and occupational health impacts of HM contamination in the soil of a copper mine located in Majdanpek (eastern Serbia). Continuous ore exploitation in the Majdanpek area has caused an increase in HM concentrations in the soil. Therefore, this area was examined to assess the effects of mining on the soil contamination and worker health. The objectives were to: (1) investigate the content of HMs in soil; (2) identify the main pollution sources using multivariate statistical analysis and PMF; (3) apply an integrated source-specific ecological risk to evaluate soil pollution; and (4) combine source-specific health risk assessment with Monte Carlo simulation to evaluate the potential danger for workers arising from different pollution sources. The findings will expand the current knowledge of soil HM contamination monitoring at mining sites and for developing preventive strategies for workers.

2. Materials and Methods

2.1. Study Area and Sample Collection

The Majdanpek copper mining area in eastern Serbia was studied (Figure 1). This area has a long mining history, and its copper reserves make it one of the country’s most significant mines. Ongoing operations are separated into two primary sections: the North Revir and the South Revir. The North Revir primarily comprises lead and zinc ores with a smaller amount of porphyry Cu. However, the South Revir is characterized by significant porphyry Cu deposits together with Cu-Au sulfide mineralization [37]. The Majdanpek mining area is linked to the Timok Magmatic Complex (TMC), which is part of the Carpatho-Balkan Magmatic Belt. Additionally, the deposit is located at the northernmost point of the TMC and offers excellent transportation routes to the Danube River and another mining region in Bor [38,39].

Prior to sampling, a field investigation was conducted, which was enabled by an examination of the research area. A total of 61 soil samples were collected from crucial locations in the copper mining area: filtration unit, vehicle service area, and mineral processing plant. The samples were collected using a soil sampling probe at a 0–30 cm depth of the surface soil, including the active soil layer that most strongly interacts with plant roots. Each soil sample weighed approximately 1 kg and consisted of five subsamples. The latitude and longitude of each sample were tracked using a handheld Global Positioning System (GPS). The samples were numbered, preserved in polyethylene zip-lock bags, and then transported to the laboratory.

2.2. Sample Preparation and Instrumental Measurements

In the laboratory, the samples were air-dried at 25 °C and large stones and other debris were removed. The samples were then crushed and homogenized to pass through a 2 mm sieve. The clay content, organic matter, and concentrations of 15 HMs (Cd, Cr, Cu, Ni, Pb, Zn, Hg, As, Ba, Co, Mo, Sb, Be, V, and B) were analyzed in the prepared samples.

The Walkley-Black procedure was used to measure the organic matter content, while the clay percentage was determined by the pipette method. In addition, the soil samples were digested in a CEM Mars 6 iWAVE microwave oven (CEM Corporation, Matthews, NC, USA) for HM analysis. Approximately 1.0 g of each sample was digested using an acidic mixture of HCl, HNO₃, and HClO₄ (3:1:1). The digested soil solution was cooled, filtered through a 0.45 μm syringe filter, and adjusted to 50 mL. The total concentration of the 15 HMs was determined by inductively coupled plasma mass spectrometry (Thermo iCap Q ICP-MS instrument, Thermo Scientific, Waltham, MA, USA).

Quality assurance and control of the procedure were ensured by including reagent blanks, sample replicates, and a standard reference material (NIST SRM 2711a—Montana II soil). The recovery rates varied between 93% and 114%, and the relative standard deviation was less than 9%.

2.3. Source Apportionment

Multivariate statistical analysis is important for the preliminary characterization of potential pollution sources. However, source apportionment methods are crucial for measuring individual source contributions [40]. This study employed Pearson’s correlation analysis and hierarchical cluster analysis (HCA) to identify patterns among variables and distinguish potential sources of contamination. Correlation analysis was performed to identify statistically significant relationships between the parameters, whereas HCA was applied to group and classify the parameters based on their mutual similarity [41,42]. Additionally, the PMF receptor model was selected for precise quantification of the pollution sources. This model, developed by USEPA, represents a mathematical model that precisely differentiates and quantifies contamination sources by minimizing the objective function Q [43]. By processing the original dataset (X), this model generates two matrices: factor contribution and factor profile matrix [41,44]. The following formulas describe the PMF model:

$${\mathrm{x}}_{\mathrm{i}\mathrm{j}}={\sum }_{\mathrm{k}=1}^{\mathrm{p}}{\mathrm{g}}_{\mathrm{i}\mathrm{k}}\times {\mathrm{f}}_{\mathrm{k}\mathrm{j}}+{\mathrm{e}}_{\mathrm{i}\mathrm{j}}$$

(1)

$$\mathrm{Q}={\sum }_{\mathrm{i}=1}^{\mathrm{n}}{\sum }_{\mathrm{j}=1}^{\mathrm{m}}{\left[{\displaystyle \frac{{\mathrm{x}}_{\mathrm{ij}}-\sum _{\mathrm{k}=1}^{\mathrm{p}}{\mathrm{g}}_{\mathrm{i}\mathrm{k}}\times {\mathrm{f}}_{\mathrm{k}\mathrm{j}}}{{\mathrm{u}}_{\mathrm{ij}}}}\right]}^2$$

(2)

$${\mathrm{u}}_{\mathrm{ij}}=\left\{\begin{array}{r}{\displaystyle \frac{5}{6}}\times \mathrm{M}\mathrm{D}\mathrm{L}, {\mathrm{x}}_{\mathrm{i}\mathrm{j}}\le \mathrm{M}\mathrm{D}\mathrm{L}\\ \sqrt{{(\mathrm{e}\mathrm{r}\mathrm{r}\mathrm{o}\mathrm{r} \mathrm{f}\mathrm{r}\mathrm{a}\mathrm{c}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n}\times {\mathrm{x}}_{\mathrm{i}\mathrm{j}})}^2+{(0.5\times \mathrm{M}\mathrm{D}\mathrm{L})}^2}, {\mathrm{x}}_{\mathrm{i}\mathrm{j}}>\mathrm{M}\mathrm{D}\mathrm{L}\end{array}\right.$$

(3)

where i represents the sample, j is the HM, k denotes the factor, e_ij is the residual matrix, u_ij is the uncertainty, and MDL is the method detection limit.

2.4. Risk Assessment

2.4.1. Ecological Risk Assessment

The potential ecological risk index (RI) introduced by Hankanson (1980) is widely used to evaluate and classify soil pollution [45,46]. The RI combines the concentrations, background values, and toxicological properties of heavy metal(loid)s, providing a comprehensive evaluation of their impact on the environment. This study assessed the ecological risk associated with each specific source by integrating RI and PMF results. Consequently, it is feasible to quantify the environmental impacts of individual pollution [10]. The source-specific RI was calculated using the following equation:

$${\mathrm{R}\mathrm{I}}_{\mathrm{i}}^{\mathrm{k}}=\sum {{\mathrm{E}}_{\mathrm{r}}}_{\mathrm{i}\mathrm{j}}^{\mathrm{k}}=\sum {{\mathrm{T}}_{\mathrm{r}}}_{\mathrm{j}}\times {{\mathrm{C}}_{\mathrm{f}}}_{\mathrm{i}\mathrm{j}}^{\mathrm{k}}$$

(4)

where RI represents a potential ecological risk for all heavy metal(loid)s in sample i from source k; Er denotes the potential ecological risk of heavy metal(loid) j in sample i from source k; C_f is the contamination factor of HM j in sample i from source k; and T_r is the toxic response factor for each heavy metal(loid), with the following values: 30, 2, 5, 5, 5, 1, 40, and 10 for Cd, Cr, Cu, Ni, Pb, Zn, Hg, and As, respectively [45]. The background concentrations of HMs were obtained from the upper continental crust [47]. The RI results were classified into four levels based on Hakanson’s categorization: RI < 150, low ecological risk; 150 ≤ RI < 300, moderate ecological risk; 300 ≤ RI < 600, considerable ecological risk; and RI ≥ 600, very high ecological risk [48,49].

2.4.2. Source-Specific Worker Health Risk Assessment

The human health risk model introduced by the USEPA was employed to assess the possible hazardous impacts of HMs on worker health [50]. To estimate the risk posed by each source, the PMF and health risk assessment (HRA) models were combined. This comprehensive approach enables the identification of pollution sources with the greatest impact on both carcinogenic and non-carcinogenic risks [51]. Worker HRA considered three main exposure pathways (ingestion, inhalation, and dermal contact) for two groups: indoor workers and outdoor workers. The hazard index (HI) and total carcinogenic risk (TCR) were used to characterize the non-carcinogenic and carcinogenic risks, respectively [10,17].

The average daily dose (mg/kg-day) for each exposure pathway was calculated according to the following equations:

$${\mathrm{A}\mathrm{D}\mathrm{D}}_{\mathrm{i}\mathrm{j}\mathrm{i}\mathrm{n}\mathrm{g}}^{\mathrm{k}}={\displaystyle \frac{{\mathrm{C}}_{\mathrm{i}\mathrm{j}}^{\mathrm{k}}\times \mathrm{C}\mathrm{F}\times {\mathrm{R}\mathrm{B}\mathrm{A}}_{\mathrm{i}\mathrm{j}}^{\mathrm{k}}\times \mathrm{I}\mathrm{n}\mathrm{g}\mathrm{R}\times \mathrm{E}\mathrm{F}\times \mathrm{E}\mathrm{D}}{\mathrm{B}\mathrm{W}\times \mathrm{A}\mathrm{T}}}$$

(5)

$${\mathrm{A}\mathrm{D}\mathrm{D}}_{\mathrm{i}\mathrm{j}\mathrm{i}\mathrm{n}\mathrm{h}}^{\mathrm{k}}={\displaystyle \frac{{\mathrm{C}}_{\mathrm{i}\mathrm{j}}^{\mathrm{k}}\times \mathrm{E}\mathrm{F}\times \mathrm{E}\mathrm{D}\times (\mathrm{E}\mathrm{T}/24)\times (\frac{1}{{\mathrm{V}\mathrm{F}}_{\mathrm{i}\mathrm{j}}^{\mathrm{k}}}+\frac{1}{{\mathrm{P}\mathrm{E}\mathrm{F}}_{\mathrm{i}\mathrm{j}}^{\mathrm{k}}})}{\mathrm{A}\mathrm{T}}}$$

(6)

$${\mathrm{A}\mathrm{D}\mathrm{D}}_{\mathrm{i}\mathrm{j}\mathrm{d}\mathrm{e}\mathrm{r}\mathrm{m}}^{\mathrm{k}}={\displaystyle \frac{{\mathrm{C}}_{\mathrm{i}\mathrm{j}}^{\mathrm{k}}\times \mathrm{C}\mathrm{F}\times \mathrm{S}\mathrm{A}\times \mathrm{A}\mathrm{F}\times \mathrm{A}\mathrm{B}\mathrm{S}\times \mathrm{E}\mathrm{F}\times \mathrm{E}\mathrm{D}}{\mathrm{B}\mathrm{W}\times \mathrm{A}\mathrm{T}}}$$

(7)

The HI was calculated based on the hazard quotient (HQ), while the TCR was determined using the carcinogenic risk (CR):

$${\mathrm{H}\mathrm{I}}_{\mathrm{i}}^{\mathrm{k}}=\sum {\displaystyle \frac{{\mathrm{A}\mathrm{D}\mathrm{D}}_{\mathrm{i}\mathrm{j},\mathrm{p}}^{\mathrm{k}}}{{\mathrm{R}\mathrm{f}\mathrm{D}}_{\mathrm{j},\mathrm{p}}}}$$

(8)

$${\mathrm{T}\mathrm{C}\mathrm{R}}_{\mathrm{i}}^{\mathrm{k}}=\sum {\mathrm{A}\mathrm{D}\mathrm{D}}_{\mathrm{i}\mathrm{j},\mathrm{p}}^{\mathrm{k}}\times {\mathrm{C}\mathrm{S}\mathrm{F}}_{\mathrm{j},\mathrm{p}}$$

(9)

where i is the sample, j is the heavy metal(loid), and k is the source. HQ and HI below one indicate no harmful non-carcinogenic effects, while values exceeding this threshold indicate the presence of a health risk. In parallel, TCR values within the range of 1 × 10⁻⁶ to 1 × 10⁻⁴ are considered to pose an acceptable risk, whereas values exceeding 1 × 10⁻⁴ reflect a significant health risk [23,52]. The calculation procedure used in this study was adopted from Miletić et al. [53], whereas the parameter values were obtained from the USEPA (Table S1). The reference dose (RfD) and cancer slope factor (CSF) are presented in Table S2.

2.4.3. Monte Carlo Simulation

The Monte Carlo simulation is used to obtain probability distributions and quantify the uncertainty in risk assessment models. The traditional deterministic health risk approach relies on fixed values in the calculation process [54]. However, various factors such as social features, lifestyle differences, and geographical characteristics may cause an increase in uncertainty in such models. Therefore, the conventional HRA model is not fully capable of accurately determining the risks to human health caused by heavy metal(loid)s in soil [55,56]. This limitation can be avoided by employing probabilistic risk that incorporates a set of data rather than individual point values. A probabilistic health risk assessment (Crystal Ball 11.1.24, Oracle, Redwood City, CA, USA) was performed to quantify the non-carcinogenic and carcinogenic health risks for each HM and pollution source. A total of 10,000 Monte Carlo iterations were performed at the 95% confidence level to ensure accurate results. The distribution types for variables C, IngR (ingestion rate), InhR (inhalation rate), EF (exposure frequency), AF (adherence factor), and BW (body weight) were adopted from Miletić et al., 2023 [57].

3. Results and Discussion

3.1. Organic Matter, Clay and Heavy Metal(loid)s in Mining Area Soil

The descriptive statistics of the 15 HMs, clay, and organic matter are summarized in

Table 1. Descriptive statistics of HMs (mg/kg), clay (%), and organic matter (%) in soil at the Majdanpek mining area (n = 61).

Parameter	Mean	Median	Skew	Kurt	Stdev	Max	Min	UCC *
Cd	2.42	1.40	2.38	4.97	3.36	13.5	0.08	0.09
Cr	66.4	39.0	2.59	5.73	82.8	346	14.0	92
Cu	1,056	334	1.78	2.10	1414	5239	26.0	28
Ni	20.6	16.0	2.97	9.56	16.8	88.0	6.0	47
Pb	100.7	51.0	1.35	0.65	106.3	362	10.0	17
Zn	529	254	1.95	2.39	698	2424	38.0	67
Hg	0.14	0.13	1.28	2.21	0.06	0.32	0.03	0.05
As	33.0	30.0	0.81	−0.12	23.7	84.0	2.60	4.8
Ba	94.2	91.0	1.88	4.75	48.4	265	37.0	624
Co	11.3	11.0	−0.49	0.11	3.25	17.0	3.00	17.3
Mo	11.5	9.10	0.76	−0.26	8.95	35.0	0.70	1.1
Sb	6.5	4.90	2.55	5.75	7.27	30.0	1.20	0.4
Be	0.23	0.20	0.35	−0.65	0.09	0.40	0.10	2.1
V	54.5	24.0	2.29	3.82	82.7	315	2.90	97
B	3.33	2.90	1.26	1.15	2.13	9.50	0.90	17
Clay	9.95	10.1	0.42	−0.33	4.97	23.0	2.40	-
OM	5.24	5.20	0.51	−0.43	1.32	8.30	3.20	-

* The concentration of HMs reported in the upper continental crust (UCC) was used as the background value.

and Figure S1. The clay was not dominant in some soil samples, with a range of 2.4–23% and an average value of 9.9%. The percentage of organic matter varied from 3.2% to 8.3%, with a mean value of 5.2%. All analyzed HMs, except Be, showed a wide range of concentrations. The average concentration of HMs (mg/kg) followed a decreasing order: Cu (1056) > Zn (529) > Pb (101) > Ba (94.2) > Cr (66.4) > V (54.5) > As (33.0) > Ni (20.6) > Mo (11.5) > Co (11.3) > Sb (6.5) > B (3.33) > Cd (2.42) > Be (0.23) > Hg (0.14). The average concentrations of Cd, Cu, Pb, Zn, Hg, As, Mo, and Sb were 26.9, 37.7, 5.9, 7.9, 2.8, 6.9, 10.5, and 16.2 times greater than the Upper continental crust (UCC) background values [47]. Antimon (Sb) had a 100% proportion of samples exceeding the UCC values, while Cd, As, Cu, and Hg exhibited proportions greater than 90%. Moreover, in addition to the maximum concentrations of these elements, the maximum contents of Cr, Ni, and V were higher than their corresponding UCC values. These elevated concentrations suggest that the soil was contaminated with HMs from various anthropogenic sources [27]. Evidently, higher concentrations of HMs originate from mining operations, which causes the emission of various contaminants into the environment [1,58]. The average levels of Ba, Co, Be, and B were below the reference values, indicating their natural origin.

3.2. Identification of Soil Heavy Metal(loid) Sources

Pearson correlation analysis and hierarchical cluster analysis were applied to investigate the mutual relationships and similarities among heavy metal(loid)s, while the PMF model was used for further identification and quantification of their potential sources [22,51,59].

The strong association among these elements implies similar sources of origin [41].

Table 2. The Pearson correlation analysis of 17 parameters investigated in the soil of the study area.

	Cd	Cr	Cu	Ni	Pb	Zn	Hg	As	Ba	Co	Mo	Sb	Be	V	B	Clay
Cr	0.446
Cu	0.774	0.631
Ni	0.442	0.776	0.581
Pb	0.793	0.609	0.905	0.547
Zn	0.838	0.654	0.931	0.649	0.937
Hg	0.723	0.542	0.889	0.329	0.780	0.825
As	0.869	0.484	0.745	0.566	0.697	0.729	0.624
Ba	0.158	0.455	0.323	0.653	0.271	0.397	0.287	0.259
Co	0.223	0.622	0.387	0.549	0.388	0.341	0.342	0.460	0.624
Mo	0.576	0.295	0.539	0.163	0.550	0.559	0.541	0.471	0.163	0.362
Sb	0.611	0.846	0.771	0.687	0.791	0.810	0.681	0.639	0.447	0.551	0.322
Be	−0.173	0.102	−0.323	−0.077	−0.283	−0.264	−0.111	−0.106	0.166	0.336	0.034	−0.105
V	−0.063	0.709	0.087	0.283	0.077	0.098	0.148	−0.034	0.094	0.450	0.153	0.354	0.478
B	−0.335	−0.258	−0.255	−0.053	−0.352	−0.348	−0.209	−0.178	0.328	0.030	−0.406	−0.228	0.098	−0.315
Clay	−0.042	−0.236	−0.067	−0.219	−0.208	−0.121	0.179	−0.003	0.287	0.143	−0.111	−0.201	0.271	−0.182	0.289
OM	−0.033	0.012	−0.193	0.331	−0.139	−0.032	−0.226	0.020	0.330	−0.012	−0.333	−0.054	0.123	−0.050	0.397	0.158

Colors: blue (r > 0.7), red (0.5 < r < 0.7)

shows the results of the Pearson’s correlation analysis for the 17 parameters measured in the soil samples. A strong positive correlation (r > 0.7; blue color) was observed between Cd, Cu, Pb, Zn, Hg, As, and Sb (r = 0.723–0.937). Additionally, Mo was moderately (0.5 < r < 0.7; red color) correlated with this group of HMs, demonstrating that they could originate from similar sources. Correlation analysis revealed significant correlations between Cr and Ni (r = 0.776), Sb (r = 0.846), and V (r = 0.709) and moderate correlations between Ni and Pb (r = 0.547), Zn (r = 0.649), As (r = 0.566), Ba (r = 0.653), Co (r = 0.549), and Sb (r = 0.687). Negative correlations were observed between Be, V, B, clay, OM, and all the other HMs. These findings confirm that these elements may originate from natural and geogenic sources, in contrast to other HMs, which are predominantly affected by anthropogenic activities.

Hierarchical cluster analysis classifies the soil samples and parameters into distinct clusters based on their similarity (Figure 2) [60]. The obtained polar dendrogram indicated that the HMs, OM, and clay were classified into two main clusters. The first cluster connected Co, Ba, Sb, Cr, Ni, V, Be, OM, B, and clay, while Zn, Pb, Cu, Hg, As, Cd, and Mo were located in the second cluster. Clay minerals are characterized by a high surface area and negative charge, which affect the adsorption of HMs. Therefore, the type and abundance of clay minerals influence the concentration and availability of heavy metal(loid)s in soil. This clustering supports the findings of correlation analysis. The grouping of Zn, Pb, Cu, Hg, As, Cd, and Mo was related to their similar anthropogenic origins, particularly mining activities. The behavior and variation of HMs can also be analyzed using a dendrogram of soil samples. Additionally, five different sample clusters were identified, which can further help in the interpretation of HMs’ differences. Compared to the other four clusters, the first cluster had significantly higher concentrations of Sb, Cr, V, Zn, Pb, Cu, Hg, As, and Cd. Significant soil enrichment was detected near the filtration unit, suggesting that mining activities have a significant impact on soil pollution. Samples 49, 50, and 51 in the second cluster showed elevated concentrations of Ba and Ni. However, the third cluster was identified by the low contents of Co, Ba, Sb, Cr, Ni, V, and Be. These results may reflect the reduced anthropogenic impacts and dominant contribution of geological sources. The moderate concentrations of HMs in the fourth cluster can be explained by the similar impacts of the natural and anthropogenic sources. Samples 47, 48, 45, 46, 16, 17, 24, 23, and 25 exhibited significantly lower concentrations of Cd, As, Hg, Cu, Pb, and Zn and were located in the fifth cluster. These samples were collected mainly from open fields near critical spots in the Majdanpek mine and were less exposed to mining operations.

Quantification of the main pollution sources in this study was performed using the positive matrix factorization model set to a maximum of 20 runs, with the iteration starting from a randomly selected position. The number of factors was varied from three to five, until the minimum Q value was obtained. The best PMF solution was established when the number of factors was three. Based on the R² values for all parameters, the PMF fitting was considered accurate and reliable. The signal-to-noise ratios were classified as strong with a value of 10 for each species. Figure 3 and Table S3 show the results of PMF identification of the three main pollution sources.

Factor 1 explained 26.6% of the total contribution and was primarily characterized by Ba (38.9%), Be (51.4%), B (51.5%), clay (86.8%), and OM (44%). The results of the statistical analysis of the HMs indicated that these elements most likely originated from natural sources. Minimal human influence was observed, while the average concentrations of Ba, Be, and B were below background levels [47,61]. These HMs, which are primarily produced by geogenic processes and parent rock weathering, can be considered naturally occurring soil components. Therefore, Factor 1 can be regarded as a natural source.

Factor 2 exhibited the highest loadings for Cr (50.9%), Ni (50.8%), Co (42.5%), and V (53.9%). Similarly to the previous study, this factor contributed 25.9% of the total contribution. In this study, HMs analysis suggested that a smaller number of samples were contaminated with Cr, Ni, and V, while all samples had Co concentrations below the corresponding UCC values. The HCA suggested that these elements were more likely to be connected to Be, Ba, and B, whose sources were classified as natural. However, the Pearson correlation analysis showed a strong correlation between Cr, Ni, Co, and other HMs: Cr-Cu, Cr-Pb, Cr-Zn, Cr-Hg, Cr-Sb, Ni-Pb, Ni-Zn, Ni-As, Ni-Sb, and Co-Sb. These findings imply that the accumulation of Cr, Ni, Co, and V is closely related to both natural and anthropogenic sources. The concentrations of Cr and Ni are primarily influenced by the soil parent material and pedogenesis [21,28]. However, anthropogenic activities, such as mining, can impact and increase their accumulation in the soil [23]. Previous studies have shown that mining-related activities lead to higher concentrations of Co and V in the surrounding soil [11,46,62], but they can also be controlled by pedogenesis [63]. Hence, factor 2 can be explained as a combination of natural and anthropogenic sources.

Among the three main pollution sources, Factor 3 had the highest contribution of 47.5%. It was predominantly composed of Cd (85.2%), Cu (84.3%), Pb (79.5%), Zn (81.7%), Hg (48.5%), As (74.2%), Mo (80.1%), and Sb (63.9%). The strong positive correlation (r > 0.7) between these HMs also suggests that they have similar origin. Descriptive statistics showed that the soil in the study area was contaminated with Cd, Cu, Pb, Zn, Hg, As, Mo, and Sb. Considering that their concentrations in the soil were several times higher than the background values, it can be concluded that these HMs are related to mining operations. Large amounts of Cu are released into the surrounding soil as a result of mining in this area. Copper ore extraction, flotation operations, tailings, sludge, and process wastewater are some of the major activities contributing to copper enrichment in the soil [5]. During ore extraction, Mo and Sb can be released and accumulated in the surrounding soil [13,22]. Mineral exploitation is also related to the accumulation of Cd, Pb, and As. Subsequently, the combustion of fossil fuels and transportation (tires, engines, and brake pads) release these HMs into the environment [21,37,51]. The degradation of vehicle components, such as tires, contributes to the accumulation of Zn in soil [10]. The elevated Hg concentrations in the study area were most likely concentrated in atmospheric dust released through exhaust emissions [27,28]. In addition to primary mining activities, traffic significantly contributes to the soil contamination process. Intensive heavy vehicle transport at these locations emits exhaust gases containing HMs that are released into the air and deposited in the soil. Moreover, numerous studies have shown that elevated concentrations of these HMs are associated with mining activities [14,23,26]. Thus, Factor 3 can be explained as a source of mining pollution, including ore extraction, fuel combustion, and transportation.

This study focused only on HM sources related to mining activities. They were the most influential on HM concentrations because of the nature of the selected research area. However, other sources such as agricultural or industrial activities in the vicinity may also contribute to overall HM pollution.

3.3. Source-Specific Ecological Risks

The implementation of ecological risk is important for managing pollution control in the soil environment [64]. The potential ecological risk index (RI) incorporates the impact of multiple heavy metal(loid)s [41,65]. Therefore, ecological risk assessment was performed by combining the RI and PMF models. Such a comprehensive method allows the distinction of pollution originating from various sources. Owing to the lack of toxic response factor data for Ba, Co, Mo, Sb, Be, V, and B, these elements were excluded from the RI calculation.

Table 3. Summary statistics of RI and contribution of sources to ecological risk.

	Total RI	Factor 1	Factor 2	Factor 3
mean	1215	185	47.2	982
max	5840	832	167	4853
min	99.8	22.8	8.0	67.5
Classification (%)
RI < 150	3.3	60.7	91.8	18
150 < RI < 300	21.3	26.2	8.2	11.5
300 < RI < 600	18	4.9	0	21.3
RI > 600	57.4	8.2	0	49.2

and Table S4, and Figure S2 summarize the results of the source-specific ecological risk assessment.

RI varied between 99.8 and 5840, with an average value of 1215, indicating strong pollution in the study area. Of the 61 samples, 96.7% showed some degree of contamination, with 57.4% having a very high pollution level (RI > 600). The main contributors to the RI were Cd, Hg, and Cu, accounting for 55.8%, 18%, and 13.9%, respectively. Regarding the potential ecological risk of single elements (Er), the investigated soil demonstrated moderate to high contamination by Cd, Cu, As, and Hg. In contrast, Cr, Ni, and Zn had negligible impacts on ecological risk, with all samples in the low-risk category. The contribution of each source to RI was as follows: Factor 3 (80.8%) > Factor 1 (15.2%) > Factor 2 (3.9%). As expected, Factor 3 (mining-related activities) had the greatest impact and contributed the most to the RI. The corresponding average RI values for these three factors were 185 (F1), 47.2 (F2), and 982 (F3), with ranges of 22.8–832, 8.0–167, and 67.5–4853, respectively. The majority of the samples (91.8%) had low risk, considering the impact of Factor 2. However, for Factor 3, only 18% of the samples had a low risk and nearly 50% fell into the category of very high pollution. For all factors, the highest RI values were recorded for samples 37, 38, 39, 40, and 41, which were collected near the filtration unit. The Er results showed that the soil samples were primarily contaminated by Cd for Factor 1 and Cu for Factor 2, while Factor 3 had the combined contribution of both HMs. According to Er’s classification, Factor 3 exhibited a very high Cd risk in 54% of samples. The results showed that the soil was uncontaminated with Cr, Ni, and Zn for all factors (Er < 40).

In addition to elevated HM concentrations, these findings can be explained by the toxic response factors that influence the final RI value. Owing to its high toxicity, even at low concentrations, Cd significantly contributes to the potential ecological risk index. RI highlighted the pollution with HMs in the investigated copper mining area, implying a huge anthropogenic impact on soil quality.

3.4. Source-Specific Health Risks in Workers

To identify the sources with the greatest impact on non-carcinogenic and carcinogenic risks, the HRA model included both deterministic and probabilistic source-oriented evaluations. The results of the total HRA for the indoor workers and outdoor workers are presented in Table S5.

Regarding non-carcinogenic risk, indoor workers were less exposed to HMs compared to outdoor workers. Average HI values were 0.09 for indoor workers (HI_iw) and 0.18 for outdoor workers (HI_ow), respectively. The HI_iw ranged from 0.02 to 0.32, with all samples below the threshold value of 1. Similarly, 100% of the outdoor workers showed no elevated non-carcinogenic risk, with a maximum HI_ow value of 0.62.

The primary contributor to HI for both groups of workers was As, accounting for 50% of the share (Figure 4). In addition to As, the contributions of Cr, Pb, and Cu affected the non-carcinogenic risk results, with shares of 12.1%, 11.6%, and 8.4% for outdoor workers, respectively.

Similarly, the carcinogenic risk affected outdoor workers more, with average values of 1.5 × 10⁻⁵ (TCR_iw) and 2.8 × 10⁻⁵ (TCR_ow) for indoor workers and outdoor workers, respectively. The TCR_iw ranged from 2.8 × 10⁻⁶ to 5.8 × 10⁻⁵, with none of the samples above the critical limit of 1 × 10⁻⁴. However, 8% of the samples had a high carcinogenic risk for outdoor workers (TCR_ow > 1 × 10⁻⁴), with TCR values varying between 5.1 × 10⁻⁶ and 1.1 × 10⁻⁴. Consistent with the non-carcinogenic risk results, both worker groups showed a similar trend of HM contribution to TCR. The elevated TCR values originated primarily from As and Cr, which affected the carcinogenic risk by nearly 52% and 35%, respectively. Ingestion was the predominant exposure route for HI and TCR, followed by dermal contact [17,21]. This study confirmed that outdoor workers are more exposed to adverse health effects due to their specific behavioral characteristics.

The non-carcinogenic and carcinogenic risks in this area were significant. High HI and TCR values indicate that mining activities are strongly related to health risks. Therefore, it is important to develop a source-specific HRA model to determine the primary anthropogenic contributors. Furthermore, this study can help establish regulatory measures to reduce soil pollution.

An analysis of the health risks originating from the three identified sources was illustrated in Figure 5 and Table S6. In terms of non-carcinogenic risk, the impact of sources was identical for indoor workers and outdoor workers and decreased in the following order: Factor 3 > Factor 2 > Factor 1. The indoor/outdoor workers HI contribution rates were 68.7% and 69.2% for mining pollution, 18.7% and 18.8% for mixed sources, and 12.5% and 12.0% for natural sources, respectively. For each source, the average HI values were 0.04 (F1), 0.06 (F2), and 0.22 (F3) for indoor workers and 0.35 (F1), 0.55 (F2), and 2.02 (F3) for outdoor workers. According to F1, As and V had the greatest contribution rates to HI, with shares of 69.3%/70.2% and 19.4%/18.4% for indoor workers and outdoor workers, respectively. The HMs with the highest impact on F2 were Cr (28.5%/27.7%), As (24%/24.6%), and Pb (12.9%/13.2%), while As (54.7%/55.1%), Pb (13.5%/13.6%), and Cu (10.4%/10.5%) contributed to the increased values of HI in Factor 3 for indoor workers and outdoor workers. For all three factors, HI_iw did not exceed the threshold value for any sample. However, the HI_ow values increased, especially in Factor 3, where 8.2% of the samples surpassed the value of 1. The contribution of different sources to carcinogenic risk was in agreement with the non-carcinogenic risk results. The contribution rates for TCR of indoor workers/outdoor workers were 66.4% and 66.7% for mining pollution, 23.1% and 22.6% for mixed sources, and 10.4% and 10.7% for natural sources, respectively. The average TCR_iw and TCR_ow values for each source were below 1 × 10⁻⁴. Nevertheless, Factor 3 showed that several samples exceeded this value for outdoor workers, whereas Factors 1 and 2 did not affect the indoor and outdoor workers’ health. Arsenic was the dominant HM in Factors 1 and 3, with contribution percentages of 84 and 56, respectively. For Factor 2, the heavy metal(loid) with the highest influence on the TCR was Cr.

Overall, these results imply that operations in copper mine cause notable health risks to workers. The results showed that the samples collected near the filtration unit (37, 38, 39, 40, and 41) had the greatest influence on the health risk results. By analyzing the risks derived from different sources, it is possible to develop specific strategies for pollution control [28,66]. The main HMs that pose health risks, primarily As, Cd, Cr, Pb, and Cu, are derived from ore exploitation and processing. Their concentrations need to be managed through various control measures, as they are highly toxic and may pose a significant health risk in the area. The primary measures in the Majdanpek mining area should control waste disposal, wastewater treatment, and the application of remediation techniques in cases of increased soil contamination. Monitoring soil pollution is necessary in this area.

3.5. Monte Carlo Simulation of Health Risks

Estimating health risks in workers in areas with high anthropogenic influences is quite complex. Because HMs may originate from various sources, a deterministic approach is unable to assess health risks precisely [27,67,68]. Therefore, a Monte Carlo simulation that considered uncertainties was used to estimate the risk associated with the three main pollution sources. The probabilities of outdoor workers developing health problems due to HM hazards are shown in Figure 6 and Table S7. According to the results, the mean total HI_ow (0.18) was far below the threshold value of one, with the 90th percentile of 0.20. Similarly, indoor workers had a 100% probability of not being at high non-carcinogenic risk. Including the source contributions in MCS, it has recognized mining activities as the primary driving factor to health risks. Outdoor workers were a more vulnerable group, although not in a critical non-carcinogenic status, even for Factor 3. Factors 1 and 2 showed no risk probability for either of the worker groups. Similar findings were obtained for carcinogenic risk, implying that outdoor workers were in an acceptable risk range with a mean TCR value of 2.9 × 10⁻⁵, and a TCR value at the 90th percentile of 3.3 × 10⁻⁵. These findings confirmed that mining activities were the major contributors to human exposure to heavy metal(loid)s; however, the risks were within safety limits.

The key parameters that significantly affected the health risk results were determined through sensitivity analysis. Modifying the assumed parameter values, such as HMs concentration, EF, IngR, and BW, leads to variations in HI and TCR results [69,70,71]. Identical patterns were observed for indoor workers and outdoor workers, and the results are presented using tornado graphs in Figure 7. Among all the variables, EF was the most significant factor for both non-carcinogenic and carcinogenic risks, contributing to a total variance of 58.4% and 56.8%, respectively. The second and third most influential parameters were BW and As, contributing 28.5% and 6.7% to HI and 27.7% and 6.5% to TCR, respectively. Deterministic health risk assessment also confirmed that As was the primary contributor among all the heavy metal(loid)s. IngR had a relatively low impact, while the last influential parameter was Pb for non-carcinogenic risk and Cr for carcinogenic risk. These findings suggest that greater attention should be paid to the elevated concentrations of As, Pb, and Cr in the vicinity of mining areas.

4. Conclusions

Severe soil contamination was confirmed through source-specific ecological risks, particularly near the filtration unit area.

Cadmium was the HM that contributed the most to the ecological risk. A concentration-based health risk analysis showed a noteworthy health risk regarding outdoor workers. The non-carcinogenic risk for both worker groups was generally low, and the estimated carcinogenic risk remained below the threshold for all samples. Ingestion was the predominant route of non-carcinogenic and carcinogenic exposure in indoor workers and outdoor workers. The combination of PMF and HRA revealed a high impact of mining activities (Factor 3) on overall health exposure, with As being the most contributing element. These results were confirmed by a Monte Carlo simulation, where the highest HI and TCR values were obtained for Factor 3. The findings of this study highlight the necessity for regulatory and waste control management in mining regions., to protect the environment and workers from HM contamination.

However, this study has several limitations, including the inherent uncertainty of the PMF source apportionment model and the limited spatial and temporal coverage of sampling, which may constrain the generalizability of the findings. Additionally, by focusing primarily on mining-related soil HM emissions, other potential pollutants and environmental media such as air and water were not examined. Future research will be expanded to include longitudinal data collection and the integration of additional environmental variables to capture HM contamination dynamics more comprehensively.