Characterizing the source contributions and spatial controls of heavy metals in stream sediments in northwestern China’s mining areas

Heavy metal pollution in stream sediments of the Qaidam Basin, northwestern China, has attracted widespread attention due to its environmental and ecological implications. Estimate of the sources of heavy metals in sediments and their driving factors is a pre-requisite for the management and control of heavy metal pollution. In this study, a large number of stream sediment samples (n = 8,280) from typical alluvial fans in the Qaidam Basin were collected for heavy metal analysis. Positive matrix factorization (PMF) was employed to explore their sources and spatial contributions. The findings indicate that mineralization (39.1%) and soil parent material (34.3%) are the predominant sources of heavy metals in sediments, accounting for over 70% of the total. The contributions of industrial activities (19.1%) and atmospheric deposition (7.4%) are significant and cannot be disregarded. Furthermore, the presence of heavy metals in stream sediments demonstrates significant spatial variability (p < 0.05). The spatial distribution of natural sources is subject to the influence of geological conditions, exhibiting congruence with the spatial distribution of strata and structures. This study provides a foundation for implementing site-specific heavy metal risk mitigation strategies in mining areas.

1 Introduction

Stream sediments function as integrative products of material breakdown and transport within a drainage basin (Zhou et al., 2024; Szabó et al., 2025). Their composition reflects the combined effects of physical weathering, erosion, and subsequent mobilization by fluvial, aeolian, or glacial processes. The sediment supply can derive from multiple sources, including bedrock, soil layers, wind-blown deposits, and glacial residues (Nwagbara et al., 2025). Under certain geomorphic and climatic settings, chemical weathering and hydromorphic dispersion of mobile elements in groundwater or surface water can exert significant influence, whereas in a limited number of environments, precipitates or organic matter may play a dominant role (Kara, 2025). Chemically mobilized elements may subsequently precipitate within drainage sediments, organic matrices, or Fe–Mn–Al oxides present in the sediment. Stream sediments may be enriched in heavy minerals and/or in chemically transported elements that adsorb onto Fe–Al–Mn oxides or organic matter coating the streambed (Ren et al., 2024; Ye et al., 2024). Heavy metals have been observed to accumulate in stream sediments as a result of atmospheric deposition or surface water, and may undergo secondary migration in response to changes in hydraulic conditions (Zhang et al., 2025). Heavy metals are characterized by high toxicity, bioaccumulation, and resistance to degradation, thereby posing a considerable threat to ecosystems and human health (Liang et al., 2017; Jiang et al., 2023). The presence of elevated concentrations of heavy metals in sediments has the potential to enter the human and animal body through inhalation, skin absorption, or indirect intake by crops, posing a serious threat to human and animal health (Li et al., 2020; Al-Tamimi et al., 2025). In regions characterized by recurrent mining and agricultural activities, the accumulation of heavy metals in stream sediments is frequently influenced by a multitude of sources of pollution (Astatkie et al., 2021; Budianta, 2021). The Jinshuikou (JSK) region is located at the southern edge of the Qaidam Basin and is characterized by a typical riverine alluvial landscape (Nuralykyzy et al., 2021; He et al., 2022). The region is abundant in metal mineral resources, with multiple metal mines distributed in the southern mountainous area and frequent agricultural activities in the northern plains (Zhu et al., 2024). The presence of heavy metals in stream sediments, attributable to both anthropogenic and natural sources, has significant ramifications for the ecological environment and the safety of drinking water for residents downstream. Consequently, a comprehensive analysis of the sources and spatial distribution of heavy metals in stream sediments within the JSK region is imperative for the effective control and remediation of these substances.

The spatial distribution and geochemical behavior of heavy metals in stream sediments arise from the integrated influences of tectonic activity, hydrogeological conditions, and watershed-scale land-use patterns (Dung et al., 2013; Liu et al., 2024). Metamorphic strata, as the primary targets of magmatic and tectonic alteration, provide material sources for acidic to ultramafic magmas originating from deep within the crust. Over prolonged tectonic evolution, repeated episodes of intense deformation and magmatism have driven hydrothermal metasomatic mineralization processes, thereby establishing the fundamental geochemical baseline (Zhang et al., 2014; 2019). Subsequent weathering, erosion intensity, and sediment transport pathways determine the initial release and redistribution of metals from mineralized zones. Superimposed hydrological processes further regulate the downstream migration, fractionation, and spatial redistribution of heavy metals within stream sediments. In addition, human activities are also an important factor in heavy metal pollution that cannot be ignored (Sabti et al., 2000; Kim et al., 2007). The production of chemical fertilizers, pesticides, and metal mines can lead to increased concentrations of heavy metals in stream sediments, while metal smelting and coal combustion can cause associated metals (such as Hg) to diffuse into the atmosphere and migrate again (Liang et al., 2017).

However, the complex topography, tectonic activity, and lithology of the JSK region render it challenging to quantitatively identify the sources of heavy metals in stream sediments and their spatial variability. The interaction among these geological, hydrological, mineralogical, and human-driven factors forms the framework shaping the spatial distribution of heavy metals in stream sediments. A variety of quantitative models for analyzing pollutants have been utilized to ascertain the sources of heavy metals in sediments (Deng et al., 2018; Vestenius et al., 2021). These include the chemical mass balance method (CMB), principal component analysis (PCA), and positive matrix factorization (PMF) (Badol et al., 2008; Sun et al., 2015; Hu et al., 2020). Among these, PMF has been shown to efficiently process large sample data and obtain a non-negative source contribution with a relatively realistic representation, and it is able to incorporate analytical uncertainties into the calculation process (Wang et al., 2022; Sun et al., 2025; Xia et al., 2025). Furthermore, PMF accommodates non-additive and heterogeneous source behaviors by allowing variability within factor compositions, thereby capturing natural geochemical diversity that conventional receptor models may oversimplify (Huang et al., 2021). Consequently, the PMF model is currently employed extensively in determining air and sediment heavy metal pollutants. Therefore, this study combines the PMF model to provide a detailed analysis of the sources and spatial distribution of heavy metals in regional stream sediments.

This study aims to: 1. identify the influencing factors of heavy metals in stream sediments based on the PMF model and quantify their contributions; 2. natural mineralization is generally regarded as the primary control on heavy metal distribution in stream sediments rather than anthropogenic inputs, this study investigates the contributions of these factors and characterizes their spatial variability.

2 Materials and methods

2.1 Study area

The study area, Jinshuikou (JSK), is rich in mineral resources and an important metal mineral base in north-west China. It is located on the southern edge of the Qaidam Basin and belongs to the Kunlun Mountains in Qinghai Province (Figure 1). JSK is situated between 96°15′–96°26′E longitude and 36°00′–36°18′N latitude. JSK has a typical continental highland desert climate, characterized by arid and cold conditions with an annual average precipitation of 50–200 mm. Evaporation far exceeds precipitation, and the highlands have permafrost year-round. There is little distinction between the seasons, with a warm season lasting from May to September featuring frequent rain, snow and hail, and a cold season from October to April marked by dry, cold and windy conditions. The average temperature in January, the coldest month, is −26 °C, while in July, the hottest month, it is 9.7 °C. The annual average temperature is −4 °C.

Figure 1

Figure 1. Geological map and sample distribution of the study area.

The study area is characterized by steep mountainous terrain and a highly developed drainage network dominated by dense, dendritic patterns. Most channels are seasonal, with the Nomuhong River being the only perennial watercourse. In sectors with gentler and more open terrain, drainage development is comparatively weaker, displaying comb-like patterns, while minor gully systems occur locally within floodplain settings. Streamflow varies markedly in summer in response to snowmelt dynamics and episodic flood pulses during the rainy season. Hydrologically, the river systems are primarily sustained by glacial and snowmelt runoff, generating substantial elevation gradients and strong erosive energy. These vigorous hydrodynamic conditions, together with intense physical weathering, promote well-developed sediment production and deposition across major channels, where sediment particle sizes generally fall below 4 mm. Soils are mainly desert soils and weakly developed young soils. Additionally, aeolian inputs transported by cyclonic airflows from the Qaidam Basin and the interior plateau accumulate predominantly in the northern mountainous belt, forming loess mantles along slopes and riverbanks. The thickness and continuity of these loess covers gradually decrease from the mountain front toward the basin interior.

JSK’s topography is higher in the west and south, and lower in the east and north. As shown in Figure 2, the mountains are generally steep, with significant differences in elevation. The study area is at an altitude of 3,200–4,842 m. The northern part of the study area is mainly covered by thick Quaternary loose sediments, which gradually become thinner from the foot of the mountain to the mountainous area, while the southern mountainous areas mainly expose the underlying Triassic and Permian strata (Figure 1). The main strata in the study area include, from top to bottom, Quaternary sandstone, Triassic conglomerate, Permian diorite, and Ordovician conglomerate (Figure 2). The study area is located in the Qin-Qi-Kun Orogenic Belt, with the North Kunlun Magmatic Belt to the north and the East Kunlun South Slope Subduction Zone to the south. Over the course of its long geological history, it has undergone tectonic evolution such as convergence and collision, resulting in a complex internal structure with frequent tectonic and magmatic activity. The Triassic strata in the study area are intruded by extensive intermediate to acidic granitic bodies, which are predominantly distributed along the eastern and northern margins. The eastern intrusive body is the most prominent, with an exposed area of approximately 69 km². Numerous angular xenoliths of the host rock are developed within the inner contact zone between the intrusion and the surrounding strata. The outer contact zone is characterized by pronounced thermal metamorphism and contact-baking alteration, forming skarn belts that are typically 10–50 m in width.

Figure 2

Figure 2. Topographic map of the study area.

The study area and its surrounding regions have a relatively developed mining economy and are rich in various mineral resources. The mineral resources that have been discovered so far mainly include coal, iron, manganese, copper, antimony, lithium, zinc, lead, boron, gold, graphite, silicate, and more than 40 other types of minerals, with a total of 15 metal mining areas. Specifically, the study area contains one large deposit (Hongshui River iron deposit), three medium-sized deposits (Wulonggou gold deposit, Hongqigou gold deposit, and Qingshui River iron deposit), and eleven small-scale non-ferrous metal deposits. The most significant mineral resources in this region are gold, iron, copper, and antimony (Figure 3). Iron ore and non-ferrous metal mines in the southern mountainous regions are currently in the active phase of initial exploitation. Benefiting from abundant surface water resources, there are large agricultural areas in the northern part of the study area, and the agricultural economy is well developed.

Figure 3

Figure 3. Land use patterns in the study area.

Agriculture in the Qaidam Basin is dominated by oasis-irrigated farming, with a total cultivated area of approximately 470 km². These agricultural activities are primarily concentrated near cities in the southeastern river-valley region of the basin (e.g., Golmud and Delingha). Zongjia Town contains China’s largest contiguous goji-berry cultivation base, characterized by extensive planting areas (Figure 3). Phosphate and organic fertilizers are widely applied throughout the agricultural zones. These fertilizers may contain residual mercury derived from fertilizer production processes or fungicide additives, and long-term application can lead to mercury accumulation in farmland soils.

2.2 Sampling and testing

The present study collated a total of 8,280 stream sediment samples during the course of 2018. The basic sampling unit was defined as 0.25 km² and further subdivided into four smaller grid cells, each containing one sampling point (0.0625 km² per point). Sampling locations were selected based on drainage morphology and variations in topography and geomorphology. In addition, sites with wind-deposited sand, abundant organic matter were avoided to ensure that the collected sediments accurately reflected the geochemical characteristics of the exposed bedrock. Due to the steep terrain and significant safety hazards in the northeastern part of the study area, this region was designated as an inaccessible sampling zone. The excluded northeastern polygon area is relatively small (<5 km²) and accounts for less than 3% of the study area’s total area. It lies within a minor tributary system and does not contain known mining sites, major agricultural zones, or distinct lithological units differing from the sampling domain. Future field surveys should prioritize coverage of this area to eliminate residual uncertainties caused by inaccessible sampling zone. After excluding the inaccessible zone, the actual sampling area was 415.05 km², with an average sampling point density of 20 points per km².

All samples were placed in polyethylene bags and sealed after collection. The sampling depth is mostly between 30 and 50 cm. During sampling, the sampling tool is directly inserted through the humus layer. Each sample weighed more than 500 g and was dried at room temperature before being screened through a 2 mm sieve. The subsamples are then treated with 6% hydrogen peroxide to oxidize and eliminate residual organic matter, followed by digestion and instrumental analysis. The sample testing was conducted by the Xining Mineral Resources Supervision and Testing Center of the Ministry of Land and Resources. A total of 8,280 samples were tested for 16 elements: Au, Ag, As, Sb, Hg, Cu, Pb, Zn, Nb, Cr, Mn, Bi, Co, La, Mo and Ni. The concentrations of Au, Bi, Co, La, Mo, and Ni were analyzed using an inductively coupled plasma mass spectrometer (ICP-MS), the relative standard deviation (RSD) values for repeated measurements were <5%. The concentration of Ag was analyzed using an emission spectrometer (ES), the RSD value for repeated measurements was <8%. The concentrations of As, Sb and Hg were analyzed using an atomic fluorescence spectrometer (AFS), the RSD values for repeated measurements were <5%. X-ray fluorescence spectrometry (XRF) was used to analyze the concentrations of Cu, Pb, Zn, Nb, Cr, and Mn, the RSD values for repeated measurements were <10%. To evaluate the accuracy of the test method, four standard samples were repeatedly inserted into each batch of 100 samples, for a total of 356 samples. The detection limits for Au, Ag, As, Sb, Hg, Cu, Pb, Zn, Nb, Cr, Mn, Bi, Co, La, Mo, and Ni were 0.00025, 0.02, 1, 0.1, 0.0046, 0.9, 1.2, 1, 1, 3, 1.9, 0.025, 0.92, 3.5, 0.11, and 1.5 mg/kg, respectively (Table 1). Based on detection limits and instrument performance, following standard laboratory allocation protocols, certain method variability may exist between different testing methods. All instruments (ICP-MS, ES, AFS, and XRF) were operated with full QA/QC procedures. For ICP-MS and ES, multi-point calibration curves (5–7 points) were prepared using certified standard solutions, with R² values consistently >0.999. Certified reference material was analyzed at a frequency of four per 100 samples; measured values fell within 93%–105% of the certified concentrations. Instrument drift was monitored using calibration check standards every 10 samples. Spike-recovery tests (10% spike level) yielded recoveries of 88%–110%. Procedural blanks were consistently below the detection limits. Because each element was analyzed using only one instrument, potential inter-method variability among instruments could not be independently assessed.

Table 1

Table 1. Sample testing methods and detection limits.

2.3 Positive matrix factorization

The PMF model has been widely used to identify and quantify potential source contributions in pollutants. The PMF model can be used to solve the chemical mass balance problem between heavy metal concentrations and source distribution using the following Equations 1, 2 (Liang et al., 2017; Norris and Duvall, 2014).

$x_{ij}={\displaystyle \sum _{k=1}^p}{g_{ik}f}_{kj}+e_{ij}(1)$

$Q={\displaystyle \sum _{i=1}^n}{\displaystyle \sum _{j=1}^m}{\left(\frac{x_{ij}-{\displaystyle {\sum }_{k=1}^p}{g_{ik}f}_{kj}}{u_{ij}}\right)}^2(2)$

where $x_{ij}$ denotes the concentration of the $j_{th}$ chemical species in the $i_{th}$ sample. $g_{ik}$ represents the fractional contribution of source k to sample $i_{th}$, and $f_{kj}$ signifies the concentration of species $j_{th}$ within the profile of source k. The residual term $e_{ij}$ captures the unexplained variance between the modeled and observed concentrations.

Moreover, PMF uses both sample concentration and uncertainty associated with the sample data to weight individual points, which could account for the confidence in the measurement. The Uncertainties of individual data points are estimated as follows: when the concentration of each element is less than or equal to the species-specific method detection limit (MDL), the value of uncertainty is Equation 3 (Li J. et al., 2021):

$u_{ij}=\frac{5}{6}MDL(3)$

When the concentration of each element is more than the MDL, the value of uncertainty is Equation 4 (Li W. et al., 2021):

$u_{ij}=\sqrt{{\left(ErrorFraction\times x_{ij}\right)}^2+{\left(0.5\times MDL\right)}^2}(4)$

where $u_{ij}$ is the degree of uncertainty for the $j_{th}$ chemical component in the $i_{th}$ sample. MDL is the method detection limit for element. Error fraction is the percentage of measurement uncertainty.

This study additionally incorporated an uncertainty factor of 10% to account for uncertainties associated with field sampling operation such as sample collection, transport, and storage in this study (Zanotti et al., 2019). While PMF model residuals generally diminish with increasing dimensionality, excessive factors risk fragmenting meaningful sources or processes into artifactual subcomponents, compromising interpretability (Yan et al., 2016; Zhang et al., 2020). Consequently, the optimal dimensionality in this analysis was determined by prioritizing factor interpretability within the environmental context, Q-values, and residual distribution characteristics (Brown et al., 2015). More details of PMF principle can be found in the EPA PMF User Guide (Mao et al., 2023). Considering the processes including anthropogenic pollution and natural processes, and the effects of pollution on natural processes, the 16 variables (Au, Ag, As, Sb, Hg, Cu, Pb, Zn, Nb, Cr, Mn, Bi, Co, La, Mo, and Ni) were selected in the PMF model, the resulting PMF factors should be interpreted as composite geochemical patterns rather than purely independent sources (Huang et al., 2021; Rui et al., 2025).

3 Results and discussion

3.1 Concentrations of heavy metals in stream sediments

Table 2 presents the basic statistical characteristics of heavy metal concentrations in stream sediment samples. Soil heavy metal analysis reveals Mn had the highest mean concentration (516.3 mg/kg), followed sequentially by Zn (57.8 mg/kg), Cr (38.8 mg/kg), La (31.4 mg/kg), Pb (25.9 mg/kg), Hg (22.9 mg/kg), Cu (18.2 mg/kg), Ni (17.9 mg/kg), Co (10.7 mg/kg), As (7.7 mg/kg), Nb (12.4 mg/kg), Mo (1.0 mg/kg), Sb (0.7 mg/kg), and Bi (0.4 mg/kg), while Ag (53.5 μg/kg) and Au (0.8 μg/kg) show minimal concentrations (Figure 4). Extreme spatial variability characterizes Hg (785.7%), Sb (382.1%), and As (488.3%), contrasting sharply with stable elements Nb (33.5%), La (47.2%), and Mn (47.8%). Furthermore, these elements exhibit extreme maximum values (Hg: 13,147 mg/kg, As: 1,963 mg/kg), indicating severe localized pollution. Normality was evaluated for all 16 heavy-metal concentration variables (n = 8,280) using skewness–kurtosis diagnostics and the Kolmogorov–Smirnov (K–S) test. The results indicate that the dataset exhibits a moderate degree of right-skewness, reflecting the typical non-normal distribution commonly observed in environmental geochemical data. Consequently, a logarithmic transformation was applied in subsequent analyses, which substantially improved the distributional properties of the data. Compared with the metal concentrations in sedimentary sediments across the province, elevated concentrations of elements (e.g., Pb, Nb, Co, Mo, Zn) in the JSK area exceeded provincial background levels (p < 0.01), whereas concentrations of elements (e.g., Cu, Zn, Ag, As, Sb) were comparable to these reference levels. Concentrations of elements (e.g., Au, Cr, Ni, Mn) were depressed relative to background values, with Au exhibiting marked depletion (p < 0.01). This distribution pattern likely results from the predominance of igneous rocks (constituting ≈ 70% of the lithology), leading to depressed elemental concentrations.

Table 2

Table 2. Descriptive statistics of heavy metal concentrations in stream sediments. The units for Ag and Au are ug/kg, while the units for other elements are mg/kg.

Figure 4

Figure 4. Box plot of heavy metal concentrations. The units for Ag and Au are ug/kg, while the units for other elements are mg/kg.

Table 3 presents the basic statistical characteristics of heavy metal concentrations in stream sediments across the major geological units of the study area. The Triassic strata, extensively exposed in the central and northern regions (Figure 1), exhibit average concentrations of Co, Cr, Cu, Hg, Mn, Mo, and Ni that are lower than the study-area average, whereas all other elements exceed the study-area average. In particular, Au, La, Sb, and Ag show markedly elevated concentrations relative to the study-area average, indicating clear enrichment (Table 2). Ag, Au, and Pb display relatively high coefficients of variation (2.45, 2.67, and 1.69, respectively), likely reflecting the influence of intense low-temperature hydrothermal mineralization within the Triassic strata, which facilitated the mobilization and redistribution of these elements along the contact zones between intrusive and metamorphic rocks. The Devonian strata, mainly distributed in the southern mountainous areas, exhibit average concentrations for all elements that are close to or slightly below the study-area average. Permian strata, which are widely exposed in the central and northern regions, show average concentrations for most elements comparable to study-area levels, except for Co, Cr, and Cu. These three elements exhibit relatively low coefficients of variation (0.43, 0.64, and 0.45, respectively), indicating a more homogeneous spatial distribution.

Table 3

Table 3. Descriptive statistics of heavy metal concentrations in major geological units. The units for Ag and Au are ug/kg, while the units for other elements are mg/kg.

3.2 Source apportionment by the PMF model

Sixteen variables (Au, Ag, As, Sb, Hg, Cu, Pb, Zn, Nb, Cr, Mn, Bi, Co, La, Mo, and Ni) were selected from stream sediment samples as the input variables. Subsequently, the concentration-related uncertainty values were calculated on the basis of the concentration data of heavy metals in 8,280 stream sediment samples. The concentration and uncertainty data were entered into the PMF model 5.0. Three to five factors were applied using a random seed pattern, with 100 runs conducted. The convergence run yielding the Q-value closest to the target Q-value was designated as the baseline run. Subsequently, 100 Bootstrap runs and a Displacement analysis were performed to assess uncertainty in factor loadings and factor scores from the base run (Huston et al., 2012; Brown et al., 2015). Table 4 summarizes the error estimation diagnostics for the PMF analysis results. Among the three-to-five factor models, the four-factor model exhibited the smallest Q_robust/Q_true. When increasing the number of factors from three to four, the Q/Q_expected ratio decreased from 6.59 to 5.76. Further increasing the factors from four to five reduced the Q/Q_expected ratio from 5.76 to 5.00, representing a comparatively smaller improvement. Because Q-values generally decline as the number of factors increases, additional diagnostic evaluations are required to substantiate the appropriateness of the selected factor number. DISP analyses showed no swaps across the three tested configurations. Bootstrap resampling reproduced 100% of the factors for the three- and four-factor solutions, whereas one factor in the five-factor solution was reproduced in less than 80% of Bootstrap resamples (Table 4). Consequently, four factors were determined based on the above analysis, with scaled residuals ranging from −4 to 4 for nearly all samples in this solution (Figure 5). This finding suggests that the product exhibits adequate fitting performance. The contributions of the four factors to each element, as well as their concentration relationships, are demonstrated in Figure 6a. In general, Hg exhibited a notably elevated proportion in factor 1, similar to the patterns observed for As and Pb, which were predominantly loaded in factor 4. This finding suggests that these elements may originate from independent anthropogenic sources, i.e., they are influenced by agriculture, animal husbandry, or mining activities. For Mn, which is abundant in soil and has the highest concentration in stream sediment samples, the main load is on factors 2 and 3, indicating that it is mainly affected by natural factors.

Table 4

Table 4. Summary of PMF error estimation diagnostics.

Figure 5

Figure 5. Scaled residual histogram.

Figure 6

Figure 6. (a) PMF model factor contributions based on soil heavy metal concentrations; (b) Total proportion of each factor; (c) The relative contribution of PMF factors to each element. The units for Ag and Au are ug/kg, while the units for other elements are mg/kg.

The Spearman correlation coefficients for 16 heavy metals are shown in Figure 7. Co was significantly correlated with Cr, Cu, Mn, and Ni; As was significantly correlated with Sb; and Ni was significantly correlated with Cu, Cr, and Co (p < 0.01). Significant correlation coefficients between metal elements in sediments may indicate that these metal elements have the same source or influencing factors. For instance, As and Sb may originate from pollutants released during metal mining operations (e.g., antimony mining) and coal combustion within the study area (Chen et al., 2022). In contrast, the element Hg exhibited no significant correlation with other elements, thereby suggesting the possibility of a distinct origin for Hg (Jeong and Ra, 2023; Ye et al., 2024). Phosphate fertilizers and pesticides have been found to contain significant quantities of mercury, which is released into the atmosphere through soil and plant transpiration. This suggests that the primary source of Hg may be attributable to agricultural activities. The analysis of Spearman’s correlation coefficient is not comprehensive. Consequently, the findings of the PMF model and correlation analysis are amalgamated to examine the factors influencing heavy metal concentrations in stream sediments.

Figure 7

Figure 7. Correlation analysis between different elements and their contributions to each factor. The correlations with p-values < 0.01 is indicated by *.

Factor 1 is predominantly characterized by the presence of Hg, exhibiting minimal correlation with other elements Figure 6c. Typically, mercury is a by-product of anthropogenic activities, including the use of pesticides, fertilizers and industrial production (Liang et al., 2017; Chen et al., 2022). Given that mercury frequently occurs in association with ores in the non-ferrous metal smelting industry, mercury emissions are a consequence of the smelting process, resulting in severe mercury contamination of the surrounding soils (Ghosh et al., 2023). Atmospheric mercury constitutes a significant source of soil mercury, with an estimated annual emission of approximately 500 tons of mercury into the atmospheric cycle attributable to human activities. Mercury is volatilized from soils and vegetation and subsequently released into the atmosphere through transpiration processes. Once emitted, it behaves as a nearly insoluble gaseous species that can be transported over long distances by wind before redeposition. Atmospheric mercury can then be reabsorbed by plant leaves and subsequently transferred to surface environments, where it becomes adsorbed onto colloids in soils and water bodies. These processes collectively enhance the accumulation of mercury in stream sediments (Devai et al., 2007; Yan et al., 2024). The JSK region is characterized by extensive farmland in its downstream areas, where intensive agricultural activities—including the use of organic fertilizers and phosphate fertilizers—can lead to mercury accumulation in agricultural soils (Figure 3). Meanwhile, upstream there are numerous metal mining areas, which provide a substantial source for mercury accumulation in sediments through atmospheric deposition. Therefore, factor 1 is interpreted as atmospheric deposition.

Factor 2 is primarily characterized by elements such as Cr, Cu, Ni, and Co. Previous studies have shown that the sources of elements such as Cr, Cu, and Ni in sediments are primarily influenced by the parent material (Xia et al., 2025; Zhang et al., 2025). The coefficient of variation for Cr, Cu, Ni, and Co in the JSK region is relatively low compared to other elements (Table 2). Additionally, the concentrations of Cr, Cu, Ni, and Co in sediment samples are either at or below background levels, indicating that these elements are primarily influenced by the parent material. As shown in Figure 7, the correlation coefficients between Cr, Cu, Ni, and Co are relatively high (p < 0.01), suggesting a common source. The presence of Co and Ni in sediments may be related to the dissolution of Co and Ni oxides (Yan et al., 2025). Moreover, limonite (Fe₂O₃·nH₂O) often coexists with Co and Ni oxides. Consequently, Ni and Co may be the result of the dissolution of Fe oxides under natural conditions, with the release of Co and Ni, thereby explaining their high correlation. Therefore, factor 2 is attributed to soil parent material.

Factor 3 primarily loads on Ag, Au, La, Nb, Pb, and Zn. Generally, the formation and mineralization of La, Nb, and Au elements are associated with intermediate-acidic intrusive rocks (Darwish, 2017a). In addition, La and Nb show a significant positive correlation (Figure 7, p < 0.01). The JSK region contains a large amount of Triassic intermediate-acidic intrusive rocks (Figure 1), with rock types including granite and diorite. This corresponds to the ring-shaped tectonic structure at the tectonic margin of the region. The metamorphic rock strata in this area have undergone tectonic deformation, which facilitates further element enrichment and mineralization (Darwish, 2017b; Moghaddam et al., 2023). The frequent magmatic activity in this region provides both a heat source and material source for mineralization, while intense tectonic movements provide pathways for element enrichment. Therefore, F3 can be assigned as a mineralization.

Factor 4 is predominantly associated with As and Sb. As and Sb are considered to be fingerprints associated with industrial pollution (Marshall et al., 2010). As and Sb are primarily derived from activities such as mining and the refining of non-ferrous metals (Cao and Guo, 2024). The presence of industrial wastewater in the water bodies surrounding the study area has been demonstrated to result in increased concentrations of As and Sb in sediments. It is noteworthy that As and Sb exhibit a high degree of correlation (Figure 7, p < 0.01), suggesting that these two elements may share a common source or be influenced by similar factors. The JSK region is abundant in mineral resources and has a highly developed mining economy, with a large number of metal mines distributed throughout the region. Furthermore, antimony is one of the primary minerals in the study area, and in natural environments, antimony often occurs in association with arsenic and antimony sulfides. Consequently, the mining and smelting of antimony ores may contribute to elevated concentrations of arsenic and antimony in stream sediments. It can thus be concluded that factor 4 is attributable to industrial activities.

3.3 Spatial difference of source apportionment

The large sample size (n = 8,280) in this study, combined with the PMF model, can reveal the factors affecting heavy metal concentrations in stream sediments more comprehensively. As illustrated in Figure 8, the spatial distribution of weights for each PMF factor has been mapped using Kriging interpolation. Among these, the areas where each PMF factor contributes over 50% of the total area are as follows: Factor 1 (atmospheric deposition) accounts for approximately 7.7%, Factor 2 (soil parent material) accounts for 13%, Factor 3 (mineralization) accounts for 15.6%, and Factor 4 (industrial/mining activities) accounts for 7.6%. On the whole, factor 1 (atmospheric deposition) and factor 4 (industrial activities) are anthropogenic sources, accounting for 7.4% and 19.1% of the total contribution rate, respectively (Figure 6b). Factor 1 (atmospheric deposition) contributes significantly to areas primarily located in the northern part of the study area and in low-altitude river valleys, contributed approximately 61% of the total heavy metal load (Figure 8a). The northern part of the study area is an alluvial plain with a flat terrain, primarily composed of Quaternary loose sediments. The predominant land use in this region is agricultural, characterized by frequent agricultural activities and extensive utilization of fertilizers and pesticides. The utilization of phosphate and organic fertilizer has resulted in elevated mercury (Hg) concentrations in sediments. The southern upstream region of the study area contains numerous metal mines, which release substantial amounts of mercury during smelting processes. Regional cyclonic airflows promote the transport of mercury from the Qaidam Basin and the inland plateau toward the southern mountainous areas. Mercury volatilized from soils and vegetation is released into the atmosphere through transpiration and can be carried over long distances before redeposition. During atmospheric transport, gaseous mercury can be reabsorbed by plant foliage and subsequently transferred to surface environments, ultimately accumulating in sediments. Consequently, the spatial pattern of Factor 1 contributions corresponds closely with vegetation distribution (Figures 3, 8a), suggesting that mercury emitted to the atmosphere through human activities is subsequently intercepted by plant canopies and redistributed within the watershed. In general, elevated Hg concentrations in the northern agricultural plains primarily are most likely associated with the historical applications of Hg-based pesticides and fertilizers, while the southern upwind mining and smelting zones may contribute additional Hg loads through long-range atmospheric transport. The regions exhibiting a substantial contribution of factor 4 (industrial activity) are predominantly concentrated in the southern portion of the study area, accounting for approximately 70% of the total load. A large iron ore deposit and a smaller non-ferrous metal deposit are located in the southern mountainous region and adjacent river valleys (Figure 3), with the latter currently in the early stages of development. The locations of these mining areas align with the spatial distribution of Factor 4 contribution rates. These regions are characterized by a significant presence of metal mines, with industrial wastewater from non-ferrous metal smelting contributing to an increase in As and Sb concentrations in stream sediments. In the future, targeted soil remediation strategies could be implemented in areas exhibiting severe anthropogenic heavy metal contamination (Tanko et al., 2025).

Figure 8

Figure 8. Spatial variability in the contributions of (a) Factor 1, (b) Factor 2, (c) Factor 3, and (d) Factor 4. Colors indicate the proportional contribution of each factor.

Factors 2 (soil parent material) and 3 (mineralization) accounted for 34.3% and 39.1% of the total contribution, respectively. These two factors contributed more than 70% of the total heavy metal contribution, indicating that natural sources are the primary source of heavy metals in stream sediments in the JSK region. The areas most strongly influenced by Factor 2 (soil parent material) are predominantly located in the central and northern parts of the study region, corresponding to the widespread Permian strata. Within these Permian units, concentrations of Co, Cr, and Cu are elevated relative to other strata, whereas the remaining elements remain below the regional average and display relatively uniform spatial distributions. This pattern underscores the dominant control of soil parent material on heavy metal concentrations in stream sediments within these regions. With regard to factor three (mineralization), elevated values are evidently concentrated in areas where Triassic strata are distributed (Figures 8c, 1), contributing approximately 68% of the heavy metal load. Intense magmatic mineralization activity has been observed within the study area, particularly at the intersection of intrusive and metamorphic rock formations, where mineralization is particularly pronounced and widespread. Furthermore, Triassic quartz diorite intrusions are distributed along the margins of the JSK structural zone, and the strata have undergone tectonic overprinting and modification, promoting further enrichment and mineralization of heavy metals. Large-scale fault structures have been demonstrated to provide the driving force and pathways for mineralization and the transport of heavy metals. This led to markedly elevated concentrations of Au, La, Sb, and Ag in the Triassic strata compared with the study area average. The high coefficients of variation observed for Ag, Au, and Pb reflect the influence of intense, localized low-temperature hydrothermal mineralization within the Triassic deposits. Consequently, mineralization is the most significant factor contributing to heavy metals in sediments in the study area. Its spatial distribution is consistent with the distribution of strata and structures, and it dominates the enrichment of elements such as Ag, Au, La, and Nb.

It should be emphasized that this study is based on a single-year sediment dataset. Consequently, the samples primarily reflect the long-term accumulation and geochemical stability of heavy metals in sedimentary environments, limiting their temporal representativeness for assessing pollution source contributions. Therefore, the spatial distribution patterns and source signals inferred via PMF should be interpreted as preliminary indications rather than definitive long-term trends. Sediment–water exchange processes—including adsorption–desorption, redox reactions, and hydrodynamic resuspension—affect the short-term variability and bioavailability of heavy metals. Future studies that integrate sediment data with water column measurements and analyses of suspended particulate matter would enable a more comprehensive understanding of heavy metal migration, transformation mechanisms, and interannual variability within the JSK watershed.

4 Conclusion

This study focused on the Qaidam Basin, identifying and quantifying the controlling factors of heavy metals in stream sediments and exploring the spatial distribution patterns of these sources by combining PMF modeling and geochemical methods. The primary conclusions of this study are as follows.

The distribution of stream sediments in the JSK region exhibits significant spatial variations. PMF was utilized to preliminarily identify and quantify four factors affecting stream sediments: atmospheric deposition, soil parent material, mineralization, and industrial activity. Among these, anthropogenic factors (atmospheric deposition and industrial activity) had a relatively minor impact, accounting for 7.4% and 19.1%, respectively. The areas with the greatest contribution to atmospheric deposition are mainly the Quaternary loose sediments in the northern part of the study area, where agricultural activities are frequent. While the areas affected by industrial activities are concentrated in the south, where there are several metal mining areas. Natural sources (soil parent material and mineralization) dominate the concentration of heavy metals in stream sediments, accounting for 34.3% and 39.1%, respectively. The distribution of natural sources is primarily governed by geological conditions, exhibiting a congruence with the spatial arrangement of strata and structures. These results contribute to the remediation and regional management of heavy metals in stream sediments in this and similar regions. Future efforts could focus on the development of long-term monitoring systems and targeted remediation strategies, such as sediment stabilization and phytoremediation using aquatic plants, tailored to the spatial control factors of heavy metals in water system sediments. It should be emphasized that although this study draws on a large-sample sediment dataset from the water system of a typical alluvial fan mining area, its temporal representativeness is constrained by the fact that the dataset derives from a single year and a single environmental medium. Future work incorporating multi-season sampling, sedimentological characterization, mineral-host analyses, hydrological transport modeling, and independent geochemical tracers will be essential to refine the mechanisms governing these controlling factors.

Data availability statement

The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found in the article/Supplementary Material.

Ethics statement

The manuscript presents research on animals that do not require ethical approval for their study.

Author contributions

SX: Conceptualization, Investigation, Methodology, Visualization, Writing – original draft, Writing – review and editing. CQ: Resources, Supervision, Validation, Writing – review and editing. YZ: Methodology, Writing – review and editing. BJ: Visualization, Writing – original draft. ZA: Software, Writing – original draft. SM: Data curation, Supervision, Validation, Visualization, Writing – review and editing. ZL: Investigation, Writing – original draft. SL: Resources, Writing – original draft. GW: Investigation, Writing – original draft.

Funding

The author(s) declared that financial support was received for this work and/or its publication. This research was funded by Qinghai Provincial Bureau of Geological Mineral Exploration and Development, grant number 2025-40-17.

Conflict of interest

Author YZ was employed by Western Mining Co., Ltd. Author ZL was employed by PowerChina Qinghai Electric Power Engineering Co., Ltd.

The remaining author(s) declared that this work was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Generative AI statement

The author(s) declared that generative AI was not used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Supplementary material

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fenvs.2025.1744400/full#supplementary-material

References

Al-Tamimi, W., Hamzaoui-Azaza, F., Ghanem, M., and Bouhlila, R. (2025). Groundwater contamination with heavy metals: a case study in Hebron, Palestine. Asian J. Water, Environ. Pollut. 22, 51–63. doi:10.36922/AJWEP025040020

CrossRef Full Text | Google Scholar

Astatkie, H., Ambelu, A., and Mengistie, E. (2021). Contamination of stream sediment with heavy metals in the awetu watershed of Southwestern Ethiopia. Front. Earth Sci. 9, 658737. doi:10.3389/feart.2021.658737

CrossRef Full Text | Google Scholar

Badol, C., Locoge, N., and Galloo, J.-C. (2008). Using a source-receptor approach to characterise VOC behaviour in a French urban area influenced by industrial emissions. Sci. Total Environ. 389, 429–440. doi:10.1016/j.scitotenv.2007.09.002

PubMed Abstract | CrossRef Full Text | Google Scholar

Brown, S. G., Eberly, S., Paatero, P., and Norris, G. A. (2015). Methods for estimating uncertainty in PMF solutions: examples with ambient air and water quality data and guidance on reporting PMF results. Sci. Total Environ. 518–519, 626–635. doi:10.1016/j.scitotenv.2015.01.022

PubMed Abstract | CrossRef Full Text | Google Scholar

Budianta, W. (2021). Heavy metal pollution and mobility of sediment in tajum river caused by artisanal gold mining in banyumas, Central java, Indonesia. Environ. Sci. Pollut. Res. 28, 8585–8593. doi:10.1007/s11356-020-11157-5

PubMed Abstract | CrossRef Full Text | Google Scholar

Cao, J., and Guo, Z. (2024). Spatial patterns of toxic elements in stream sediment transportation at a hilly mine area. Sci. Total Environ. 948, 174597. doi:10.1016/j.scitotenv.2024.174597

PubMed Abstract | CrossRef Full Text | Google Scholar

Chen, Z., Ding, Y., Jiang, X., Duan, H., Ruan, X., Li, Z., et al. (2022). Combination of UNMIX, PMF model and Pb-Zn-Cu isotopic compositions for quantitative source apportionment of heavy metals in suburban agricultural soils. Ecotoxicol. Environ. Saf. 234, 113369. doi:10.1016/j.ecoenv.2022.113369

PubMed Abstract | CrossRef Full Text | Google Scholar

Darwish, M. A. G. (2017a). Geochemical stream sediment survey in the wadi umm rilan area, south eastern desert, Egypt: a new occurrence for gold mineralization. Acta Geol. Sinica-English Edition 91, 1041–1062. doi:10.1111/1755-6724.13323

CrossRef Full Text | Google Scholar

Darwish, M. A. G. (2017b). Stream sediment geochemical patterns around an ancient gold mine in the wadi El quleib area of the allaqi region, south eastern desert of Egypt: implications for mineral exploration and environmental studies. J. Geochem. Explor. 175, 156–175. doi:10.1016/j.gexplo.2016.10.010

CrossRef Full Text | Google Scholar

Deng, J., Zhang, Y., Qiu, Y., Zhang, H., Du, W., Xu, L., et al. (2018). Source apportionment of PM 2.5 at the Lin’an regional background site in China with three receptor models. Atmos. Res. 202, 23–32. doi:10.1016/j.atmosres.2017.11.017

CrossRef Full Text | Google Scholar

Devai, I., Delaune, R. D., Devai, G., Aradi, C., Gori, S., Nagy, A. S., et al. (2007). Characterization of Mercury and other heavy metals in sediment of an ecological important backwater area of river tisza (hungary). J. Environ. Sci. Health Part A-Toxic/Hazardous Subst. and Environ. Eng. 42, 859–864. doi:10.1080/10934520701373141

PubMed Abstract | CrossRef Full Text | Google Scholar

Dung, T. T. T., Cappuyns, V., Swennen, R., and Phung, N. K. (2013). From geochemical background determination to pollution assessment of heavy metals in sediments and soils. Rev. Environ. Sci. Biotechnol. 12, 335–353. doi:10.1007/s11157-013-9315-1

CrossRef Full Text | Google Scholar

Ghosh, S., Banerjee, S., Prajapati, J., Mandal, J., Mukherjee, A., and Bhattacharyya, P. (2023). Pollution and health risk assessment of mine tailings contaminated soils in India from toxic elements with statistical approaches. Chemosphere 324, 138267. doi:10.1016/j.chemosphere.2023.138267

PubMed Abstract | CrossRef Full Text | Google Scholar

He, H., Wei, H., Wang, Y., Wang, L., Qin, Z., Li, Q., et al. (2022). Geochemical and statistical analyses of trace elements in Lake sediments from qaidam basin, Qinghai-Tibet Plateau: distribution characteristics and source apportionment. Int. J. Environ. Res. Public Health 19, 2341. doi:10.3390/ijerph19042341

PubMed Abstract | CrossRef Full Text | Google Scholar

Hu, B., Zhou, Y., Jiang, Y., Ji, W., Fu, Z., Shao, S., et al. (2020). Spatio-temporal variation and source changes of potentially toxic elements in soil on a typical plain of the yangtze river Delta, China (2002–2012). J. Environ. Manag. 271, 110943. doi:10.1016/j.jenvman.2020.110943

PubMed Abstract | CrossRef Full Text | Google Scholar

Huang, J., Wu, Y., Sun, J., Li, X., Geng, X., Zhao, M., et al. (2021). Health risk assessment of heavy metal(loid)s in park soils of the largest megacity in China by using monte carlo simulation coupled with positive matrix factorization model. J. Hazard. Mater. 415, 125629. doi:10.1016/j.jhazmat.2021.125629

PubMed Abstract | CrossRef Full Text | Google Scholar

Huston, R., Chan, Y. C., Chapman, H., Gardner, T., and Shaw, G. (2012). Source apportionment of heavy metals and ionic contaminants in rainwater tanks in a subtropical urban area in Australia. Water Res. 46, 1121–1132. doi:10.1016/j.watres.2011.12.008

PubMed Abstract | CrossRef Full Text | Google Scholar

Jeong, H., and Ra, K. (2023). Pollution and ecological risk assessments for heavy metals in coastal, river, and road-deposited sediments from apia city in upolu island, Samoa. Mar. Pollut. Bull. 188, 114596. doi:10.1016/j.marpolbul.2023.114596

PubMed Abstract | CrossRef Full Text | Google Scholar

Jiang, W., Meng, L., Liu, F., Sheng, Y., Chen, S., Yang, J., et al. (2023). Distribution, source investigation, and risk assessment of topsoil heavy metals in areas with intensive anthropogenic activities using the positive matrix factorization (PMF) model coupled with self-organizing map (SOM). Environ. Geochem. Health 45, 6353–6370. doi:10.1007/s10653-023-01587-8

PubMed Abstract | CrossRef Full Text | Google Scholar

Kara, H. (2025). Geochemical assessment of potentially toxic elements in urban stream sediments draining into the keban dam Lake, Turkey. Appl. Sci. 15, 7565. doi:10.3390/app15137565

CrossRef Full Text | Google Scholar

Kim, J. G., Ko, K.-S., Kim, T. H., Lee, G. H., Song, Y., Chon, C.-M., et al. (2007). Effect of mining and geology on the chemistry of stream water and sediment in a small watershed. Geosciences J. 11, 175–183. doi:10.1007/BF02913931

CrossRef Full Text | Google Scholar

Li, W., Lin, S., Wang, W., Huang, Z., Zeng, H., Chen, X., et al. (2020). Assessment of nutrient and heavy metal contamination in surface sediments of the xiashan stream, eastern Guangdong province, China. Environ. Sci. Pollut. Res. 27, 25908–25924. doi:10.1007/s11356-019-06912-2

PubMed Abstract | CrossRef Full Text | Google Scholar

Li, J., Wang, G., Liu, F., Cui, L., and Jiao, Y. (2021a). Source apportionment and ecological-health risks assessment of heavy metals in topsoil near a factory, central China. Expo. Health 13, 79–92. doi:10.1007/s12403-020-00363-8

CrossRef Full Text | Google Scholar

Li, W., Wu, J., Zhou, C., and Nsabimana, A. (2021b). Groundwater pollution source identification and apportionment using PMF and PCA-APCS-MLR receptor models in tongchuan city, China. Archives Environ. Contam. Toxicol. 81, 397–413. doi:10.1007/s00244-021-00877-5

PubMed Abstract | CrossRef Full Text | Google Scholar

Liang, J., Feng, C., Zeng, G., Gao, X., Zhong, M., Li, X., et al. (2017). Spatial distribution and source identification of heavy metals in surface soils in a typical coal mine city, lianyuan, China. Environ. Pollut. 225, 681–690. doi:10.1016/j.envpol.2017.03.057

PubMed Abstract | CrossRef Full Text | Google Scholar

Liu, M., Zhao, L., Lin, L., Zhang, Y., Huang, H., Deng, W., et al. (2024). Distribution characteristics, sources and risk assessment of heavy metals in the surface sediments from the largest tributary of the lancang river in the Tibet Plateau, China. Environ. Geochem Health 46, 414. doi:10.1007/s10653-024-02188-9

PubMed Abstract | CrossRef Full Text | Google Scholar

Mao, H., Wang, G., Liao, F., Shi, Z., Zhang, H., Chen, X., et al. (2023). Spatial variability of source contributions to nitrate in regional groundwater based on the positive matrix factorization and Bayesian model. J. Hazard. Mater. 445, 130569. doi:10.1016/j.jhazmat.2022.130569

PubMed Abstract | CrossRef Full Text | Google Scholar

Marshall, S., Pettigrove, V., Carew, M., and Hoffmann, A. (2010). Isolating the impact of sediment toxicity in urban streams. Environ. Pollut. 158, 1716–1725. doi:10.1016/j.envpol.2009.11.019

PubMed Abstract | CrossRef Full Text | Google Scholar

Moghaddam, M. Z., Saadat, S., and Ariaee, A. A. (2023). Accumulation of lead, antimony and arsenic in soil and water in khunik gold prospecting area, eastern Iran. Iran. J. Earth Sci. 15, 44–53. doi:10.30495/ijes.2022.1950343.1711

CrossRef Full Text | Google Scholar

Norris, G., and Duvall, (2014). EPA positive matrix factorization (PMF) 5.0 fundamentals and user guide.

Google Scholar

Nuralykyzy, B., Wang, P., Deng, X., An, S., and Huang, Y. (2021). Heavy metal contents and assessment of soil contamination in different land-use types in the qaidam basin. Sustainability 13, 12020. doi:10.3390/su132112020

CrossRef Full Text | Google Scholar

Nwagbara, V. U., Naomab, E., Tyobeka, M. B., Onjefu, S. A., Abah, J., Kgabi, N., et al. (2025). Assessment of heavy metal pollution of the stream sediment from the omaruru basin, Erongo region of western central Namibia. All Earth 37, 1–13. doi:10.1080/27669645.2025.2533063

CrossRef Full Text | Google Scholar

Ren, J., He, G., Yang, Y., Yu, M., Deng, Y., Pang, Y., et al. (2024). Ultraselective enrichment of trace elements in seawater by Co-rich ferromanganese nodules. Glob. Planet. Change 239, 104498. doi:10.1016/j.gloplacha.2024.104498

CrossRef Full Text | Google Scholar

Rui, X., Gong, H., Hu, J., Yuan, H., Wang, Y., Yang, L., et al. (2025). Distribution, removal and potential factors affecting antibiotics occurrence in leachate from municipal solid waste incineration plants in China. Water Res. 275, 123187. doi:10.1016/j.watres.2025.123187

PubMed Abstract | CrossRef Full Text | Google Scholar

Sabti, H., Hossain, M., Brooks, R., and Stewart, R. (2000). The current environmental impact of base-metal mining at the tui mine, Te aroha, New Zealand. J. R. Soc. N. Z. 30, 197–207. doi:10.1080/03014223.2000.9517617

CrossRef Full Text | Google Scholar

Sun, H., Zhang, Y., and Xie, J. (2015). Applications of positive matrix factorization (PMF) for source apportionment of PAHs in the environment. Asian J. Ecotoxicol. 10, 25–33.

Google Scholar

Sun, H., Ren, J., Wang, J., Yu, J., Zuo, L., Wu, X., et al. (2025). Spatial distributions, potential sources, and ecological risks of heavy metals in stream sediments in Zambia. J. Geochem. Explor. 270, 107659. doi:10.1016/j.gexplo.2024.107659

CrossRef Full Text | Google Scholar

Szabó, P., Jordan, G., Kardos, L., Šajn, R., and Alijagić, J. (2025). A geochemical modeling approach from receptor modeling to contamination source fingerprinting. Characterization of sediment-contamination processes altered by river regulation in the drava river floodplain impacted by historical mining in the alps mountains. Environ. Earth Sci. 84, 631. doi:10.1007/s12665-025-12570-0

CrossRef Full Text | Google Scholar

Tanko, W. S., Albalwe, F. M., Bakr, A. A., Ali, A. M., Ibrahim, R. I. H., Abdel-Mageed, A. M., et al. (2025). Submerged hydrophytes as a tool for the removal of heavy metals. Asian J. Water, Environ. Pollut. 22, 64–74. doi:10.36922/AJWEP025150105

CrossRef Full Text | Google Scholar

Vestenius, M., Hopke, P. K., Lehtipalo, K., Petäjä, T., Hakola, H., and Hellén, H. (2021). Assessing volatile organic compound sources in a boreal forest using positive matrix factorization (PMF). Atmos. Environ. 259, 118503. doi:10.1016/j.atmosenv.2021.118503

CrossRef Full Text | Google Scholar

Wang, Y., Li, B., Li, J., Wang, Z., Xian, F., and Fan, Y. (2022). Spatial distribution characteristics and pollution levels of heavy metals in surface water and sediments of the heihe cascade reservoir system, China. Water Supply 22, 8353–8371. doi:10.2166/ws.2022.360

CrossRef Full Text | Google Scholar

Xia, J., Gao, L., and Yang, J. (2025). Analysis of heavy metal sources in xutuan mining area based on APCS-MLR and PMF model. Appl. Sci. 15, 4249. doi:10.3390/app15084249

CrossRef Full Text | Google Scholar

Yan, C., Nie, W., Äijälä, M., Rissanen, M., Canagaratna, M., Massoli, P., et al. (2016). Source characterization of highly oxidized multifunctional compounds in a boreal forest environment using positive matrix factorization. Atmos. Chem. Phys. 16, 12715–12731. doi:10.5194/acp-16-12715-2016

CrossRef Full Text | Google Scholar

Yan, X., Chen, X., Zheng, W., Zhang, G., and Dong, A. (2024). Heavy metals in the sediments of dongshan Bay, China: sources, distribution, and controlling factors. Water 16, 905. doi:10.3390/w16060905

CrossRef Full Text | Google Scholar

Yan, Y., Zhang, Y., Sun, Z., Xie, Z., Yao, R., Chen, S., et al. (2025). Using unsupervised machine learning and positive matrix factorization models to drive groundwater chemistry and associated health risks in a coal − mining rural region. J. Hydrology 661, 133691. doi:10.1016/j.jhydrol.2025.133691

CrossRef Full Text | Google Scholar

Ye, Z., Chen, J., Liang, Z., Wu, P., Li, R., and Gopalakrishnan, G. (2024). Contamination, fraction, and source apportionment of heavy metals in sediment of an industrialized urban river in China. Environ. Res. 262, 119936. doi:10.1016/j.envres.2024.119936

PubMed Abstract | CrossRef Full Text | Google Scholar

Zanotti, C., Rotiroti, M., Fumagalli, L., Stefania, G. A., Canonaco, F., Stefenelli, G., et al. (2019). Groundwater and surface water quality characterization through positive matrix factorization combined with GIS approach. Water Res. 159, 122–134. doi:10.1016/j.watres.2019.04.058

PubMed Abstract | CrossRef Full Text | Google Scholar

Zhang, Z., Hong, W., Jiang, Z., Duan, S., Li, F., and Shi, F. (2014). Geological characteristics and metallogenesis of iron deposits in Western Tianshan, China. Ore Geol. Rev. 57, 425–440. doi:10.1016/j.oregeorev.2013.09.012

CrossRef Full Text | Google Scholar

Zhang, Z., Shu, Q., Yang, X., Wu, C., Zheng, C., and Xu, J. (2019). Review on the tectonic terranes associated with metallogenic zones in southeast Asia. J. Earth Sci. 30, 1–19. doi:10.1007/s12583-019-0858-0

CrossRef Full Text | Google Scholar

Zhang, H., Cheng, S., Li, H., Fu, K., and Xu, Y. (2020). Groundwater pollution source identification and apportionment using PMF and PCA-APCA-MLR receptor models in a typical mixed land-use area in Southwestern China. Sci. Total Environ. 741, 140383. doi:10.1016/j.scitotenv.2020.140383

PubMed Abstract | CrossRef Full Text | Google Scholar

Zhang, W., Xin, C., Du, W., and Yu, S. (2025). Health risk assessment of heavy metals based on source analysis and Monte Carlo in the lijiang river basin, China. Ecol. Indic. 176, 113620. doi:10.1016/j.ecolind.2025.113620

CrossRef Full Text | Google Scholar

Zhou, S., Cheng, Q., Weindorf, D. C., Yang, B., Gong, Z., and Yuan, Z. (2024). Multiple approaches for heavy metal contamination characterization and source identification of farmland soils in a metal mine impacted area. Appl. Geochem. 174, 106125. doi:10.1016/j.apgeochem.2024.106125

CrossRef Full Text | Google Scholar

Zhu, M., Yao, Z., Xu, X., Wei, Y., Yan, X., and Xiao, M. (2024). Accumulation, source apportionment, and ecological-health risks assessment of topsoil heavy metals in agricultural and pastoral areas in the eastern qaidam basin, China. Water 16, 3719. doi:10.3390/w16243719

CrossRef Full Text | Google Scholar