Source Identification of Cd and Pb in Typical Farmland Topsoil in the Southwest of China: A Case Study

Abstract

Cd and Pb in farmland topsoil are controlled by many factors. To identify the source of potential toxic metals in the farmland topsoil around Mianyuan River, the chemical analysis and multivariate statistical analysis are performed in this study. The results indicate the following: (1) The concentration of Cd and Pb in soil exceed the background value of Chinese soil elements. (2) Cd is significantly enriched in the whole region and Pb is locally enriched, both of them are more or less influenced by human activities. (3) The contents of Cd and Pb increase significantly following the flow direction of river. (4) $Pb$ isotope analysis indicates that the main source of Pb in the soil include the air dust, coal and phosphate plant, and the contribution of them decreases successively. (5) Linear correlation analysis and principal component analysis show that the main sources of Cd in the soil are mining phosphate rock, air dust, phosphate plant and coal mining.

1. Introduction

The toxicity and bioavailability of potential toxic metals (such as Cd and Pb) in soils have recently attracted increasing levels of public attention [1,2,3]. Cadmium (Cd) is a harmful heavy metal that ranks seventh among all heavy metals in terms of harm to animals and plants [4]. Lead is a priority heavy metal pollutant, as designated by the U.S. Environmental Protection Agency [5]. Generally, Pb enters the natural environment via industrial production (mining and smelting), the use of lead-containing products, the combustion of fossil fuels (coal, leaded gasoline), and the use of mineral fertilizers and sewage sludge [6]. Cd pollution is mainly caused by the soil parent material, soil type, water system [7], dustfall [8,9] and human activities [10]. Previous studies have found that the mining and smelting of metalliferous materials, electroplating, gas exhaust emissions, energy and fuel production, and fertilizer and pesticide application are the main ways that human activities cause Cd pollution [11]. Nowadays, anthropogenic factors are playing an important role in potential toxic metals contamination [12].

Isotope ratios are commonly used in the field of environmental science to determine composition. The mixture of lead from different sources determines the variation in the lead isotope ratio of the soil [13]. Thus, lead isotope ratios combined with concentration data have been successfully used to trace pollution sources and estimate source contribution rates [14]. Multivariate analysis (principal component analysis) and correlation analysis has been widely used to assist the interpretation of environmental data and sources of pollution [15].

The study area, located in the farm belt of Chengdu Plain in the southwest of China, is suffering from severe Cd and Pb pollution due to the development of the social economy. In terms of the environmental quality standards for soils in China (GB15618-1995), the Cd concentrations here generally exceed the Level 2 threshold (0.3 mg/kg) and Pb concentrations generally exceed the Level 1 threshold (35 mg/kg) [16,17]. Level 1 is typical for natural reserves, and Level 2 is used to indicate the threshold value for safe agricultural production and for protecting human health [18]. Cadmium rice seriously endangers the health of local residents and the development of the regional economy. The terrain slopes slowly from northwest to southeast. Holocene alluvial and pluvial deposits cover the surface in the region. Mines (distributed in the Longmen Mountains) and industrial and residential areas (distributed between the study area and the Longmen Mountains) constitute the potential sources of potential toxic metals (Pb, Cd) contamination in the area [19]. Studies have shown that potential toxic metals pollution around Mianyuan River has not changed in recent years [7,20,21,22].

Previous studies on the Mianyuan River Basin have mainly focused on the contamination of stream sediments near mines. The contribution of sources of Cd and Pb pollution have not been made clear because of the complexity of the potential pollution sources. Therefore, an investigation into the sources and topsoils in this region, in order to further clarify the sources of pollution and their proportional contributions, was performed. Various methods are used to identify the sources of Cd and Pb and to calculate contribution rates. $Pb$ isotopes were used to calculate the contribution rate of Pb sources. Multivariate statistical analysis and $Pb$ isotopes were used to calculate the contribution rate of Cd sources. It is expected that This study is helpful for understanding the mechanism of potential toxic metal pollution in the study area and providing a basis for the regional prevention and control of potential toxic metal pollution in the plain.

2. Materials and Methods

2.1. Sampling and Analysis

To investigate the factors influencing Cd and Pb pollution in the Mianzhu region’s farmland topsoil, the study area was divided into several blocks, from which samples were collected. A field investigation was conducted to individually investigate the pollution sources in the vicinity of the study area. Samples of pollution sources were collected from mines, residential areas and industrial areas in the upstream region of the study area.

A series of investigations into farmland topsoil and pollution sources in the area were conducted during 2015. The sampling grid is shown in Figure 1. According to the grid standard of 2 × 2 km, 38 topsoil samples were collected for the whole area. Each soil sample was collected in the 0–20 cm region. In order to reduce differences in soil composition, five subsamples were integrated. The collected samples were air-dried and stored in polypropylene bottles. The reference materials were HJ/T 166-2004. In total, 5 samples from the 38 samples were analyzed for $Pb$ isotopes. Kriging interpolation was used to treat the Pb and Cd contents. Four kinds of pollution source sample were collected. Phosphate ore (W01) was collected at the Qingping Phosphate Mine in the upper region of the Mianyuan River. The coal mine sample (W02) was collected at the Tianchi coal mine near Meizigou, a tributary in the upper region of Mianyuan River. A dustfall sample (W03) was collected from residential gathering areas, and the phosphogypsum sample (W04) was collected from the phosphating plant. In total, 5 samples were randomly taken from 38 kinds of topsoil to test the element content again in 2018. The results, though, showed that there was no significant difference in potential toxic metals concentration between the two sampling periods. Therefore, it can be concluded that the results from the 2015 investigation can guide the current stage of environmental governance.

2.2. Chemical Analysis

The treatment and analysis of the elements in the soil samples followed the GBW07404 and GBW07405. The soil and pollution source samples were dried and crushed using a 0.071 mm sieve. According to the United States Environmental Protection Agency (USEPA) procedure, the samples were digested with HNO₃, HClO₄ and HF (USEPA-3052 1996; USEPA-3050B 1996) [23]. Soil samples of 0.1 g were mixed with 5 mL of HNO₃ in a Teflon vial and heated to dryness at 80 °C. Then, 2.5 mL of HF, 2 mL of HNO₃ and 1 mL of HClO₄ were added and the mixture was heated to 160 °C for digestion and concentration. Totals of 1.25 mL of deionized water, 1.25 mL of HNO₃ and 2.5 mL of H₂O₂ were added to the dry residue for further digestion, and then, 5% (v/v) HNO₃ was added at a constant volume up to 25 mL for preservation. The blank sample was used to ensure the correctness of the detection. The As concentration was determined by atomic fluorescence spectrometry (AFS), and the concentrations of other elements were determined by inductively coupled plasma–optical emission spectroscopy (ICP-OES). $Pb$ isotope analysis was performed using a high-precision inductively coupled plasma mass spectrometer. The ratios of $\frac{{208}_{Pb}}{{207}_{Pb}}$, $\frac{{208}_{Pb}}{{206}_{Pb}}$ and $\frac{{206}_{Pb}}{{207}_{Pb}}$ were obtained, controlling the precision and accuracy with a reference material of $Pb$ isotope and internal standard. The standard error of the isotope measurement was less than 0.5%.

2.3. Multivariate Analysis

Descriptive analysis, Pearson correlation coefficient analysis and principal component analysis are broadly used to identify the relationship between potential toxic metals and possible pollution sources in the environment [24,25,26]. In this study, SPSS for Windows, version 19 (SPSS Inc, USA), was utilized in the multivariate statistical analysis.

Descriptive analysis, including the mean, standard deviation (SD), skewness, coefficient of variation (CV), etc., can be utilized to reflect the dispersion degree for potential toxic metals in the soil, and indirectly indicate the sources of said potential toxic metals.

The enrichment factor (EF) can be used to determine the enrichment of potential toxic metals in the soil and to differentiate between metals originating from human activities and those originating from natural processes. When using enrichment factors, a reference element should be introduced to standardize the tested elements. The reference element that is used is often a conservative one, such as Ni, Mn, Sc, etc. [27]. The enrichment factor (EF) value can be calculated via the formula proposed by Buat-Menard and Chesselet [28]:

$$\mathrm{EF}=\left[C_n\left(sample\right)/C_{ref}\left(sample\right)\right]/\left[B_n\left(baseline\right)/B_{ref}\left(baseline\right)\right]$$

(1)

where $C_n$ (sample) is the concentration of the examined element in the farmland topsoil, $C_{ref}$ (sample) is the concentration of the reference element in the farmland topsoil, $B_n$ (baseline) is the content of the examined element in Chinese soils (here, Chinese background values were selected as the baseline), and $B_{ref}$ (baseline) is the content of the reference element in Chinese soils.

Based on the linear correlation efficient, the relationship among the potential toxic metals was explored [29,30]. Principal component analysis (PCA) is widely used in the examination of soil environmental pollution [31,32] by extracting common factors from variable groups to seek implicit relations [33]. KMO and the Bartlett spherical test are usually used to test the suitability of PCA. A KMO > 0.9 indicates perfect usability; 0.9 > KMO > 0.7 means suitable for use; 0.7 > KMO > 0.6 means generally effective; KMO < 0.5 means not suitable for use [34].

3. Results and Discussion

3.1. Potential Toxic Metals Concentrations

The statistical descriptions are shown in

Table 1. Potential toxic metal concentrations of farmland topsoil around Mianyuan River (mg·kg⁻¹).

Element	Range	Mean	Median	SD	CV 1	Skewness	Reference Value 2	EFs
Cd	0.21–2.2	0.81	0.72	0.40	0.49	1.62	0.08	9.16
Pb	14.3–139	33.32	31.35	18.38	0.55	5.35	30.90	0.97
Cu	21–61.2	39.24	39.90	8.38	0.21	−0.09	31.10	1.13
Zn	67.4–391	125.84	118.50	53.07	0.42	3.46	86.50	1.31
Cr	62.7–122	91.40	88.80	12.67	0.14	0.28	79.00	1.04
Ni	21.2–50.2	36.25	36.65	6.79	0.19	−0.04	32.60	1.00
As	4.89–40.8	10.11	9.22	5.60	0.55	4.62	10.40	0.87
Hg	0.087–0.41	0.21	0.19	0.09	0.42	0.91	0.06	3.04
Se	0.22–1.39	0.70	0.68	0.23	0.33	0.43	0.10	6.58
P	662–3030	1537.03	1405.00	606.95	0.39	0.46	713.53	1.94

¹ Coefficient of variation (CV)—standard deviation (SD)/mean. ² Background values of soil elements in China (1990).

. The average contents of Pb, Cd, Cu, Zn, Cr, Ni, Hg, Se, and P exceeded the reference value (background value of Chinese soil elements), and the level of Cd exceeded the reference value more than 10-fold, which indicates that the potential toxic metals in the farmland topsoil of the study area were enriched. In addition, the CV results of these 10 elements can be split into two categories, that is, the CV values of most of the elements (Cd, Pb, As, Hg, Zn, Se, P) are greater than 0.3, and those of Cu, Ni and Cr are less than 0.2. Potential toxic metals contributed primarily by natural sources have low CV values, while those affected by human activities generally have high CV values [35]. This indicates that the concentrations of Cd, Pb, Zn, Hg and P have significant spatial heterogeneity, and their production may be dominated by human activities [36]. The presence of Ni may be primarily caused by natural factors [37]. The skewness values of Cd, Pb, Zn and As are all greater than 1, which indicates that the concentrations of Cd, Pb, Zn and As are generally high. The presence of Cu and Cr with a large SD and high mean exhibit characteristics of human activities. However, the Cu and Cr have very low CV and skewness values, which are characteristics indications of a natural source.

The content range of Ni is narrower, and the average is closer to the reference value. These factors indicate that Ni is not, or is only slightly, enriched in the farmland topsoil. The CV of Ni is very small, and the skewness of Ni is close to 0. These factors indicate that the distribution characteristics of the Ni in the study area are much more normal, that is, they are the characteristics of natural processes. Combined with the above conclusions, Ni can be selected as the reference element. The results show that the EFs of Cd, P and Se are large, and the EF of Cd is close to 10, suggesting the influence of human activities. Combined with their characteristics of high SD and CV, Cd, P and Se are also widely enriched in the study area. The EFs of Pb, Cu, Cr and Zn are small, indicating characteristics of local enrichment, and they may be determined by both natural and human activities [38,39].

3.2. Spatial Distribution Characteristics of Cd and Pb

The spatial patterns of the Pb and Cd processed by Kriging interpolation are shown in Figure 2. The contents of Pb and Cd in the study area are mostly higher than the background concentration values, which are 30.90 mg/kg for Pb and 0.08 mg/kg for Cd. Spatially, Pb and Cd are highly enriched in the northwest region, and this value gradually decreases towards the southeast region. The contents of Pb and Cd in the northwest region are relatively high, with values of over 33 mg/kg and 1.0 mg/kg, respectively. Correspondingly, the contents of Pb and Cd in the southeastern region are lower than those in the northwest, with values of 30 mg/kg and 0.7 mg/kg, respectively. This indicates that topography affects the spatial distribution of Pb and Cd in farmland topsoil to a certain extent, which may relate to the rapid decline in water flow velocity from the mountains to the plains. Pb and Cd are precipitated in the form of complexes [40,41,42,43]. Moreover, the contents of Pb and Cd tend to decrease as one moves further away from the Mianyuan River. Accordingly, the content of Pb decrease from over 31 mg/kg to 25 mg/kg, and the content of Cd decreases from over 0.58 mg/kg to below 0.4 mg/kg. This distribution also proves that the Cd and Pb concentrations in the cultivated soil are controlled by the transportation of the Mianyuan River and its tributaries. In the northwest of the study area, there are phosphate mines and phosphating plants, both of which are likely to be major sources of cadmium and lead. Deng et al. found that the frequency human activities in developed areas is also an important factor leading to potential toxic metals enrichment [44]. The high concentrations of Cd and Pb in the towns of Mianzhu and Fuxin indicate that the population concentration is also a factor contributing to the accumulation of Cd and Pb in the study area. The Cd and Pb were brought into the study area from the northwestern part of Mianyuan River, and they then migrated into the topsoil through irrigation, and became enriched in it. At the same time, potential toxic metals are released from urban living areas into the topsoil.

3.3. Pb Source Identification Based on Isotopic Ratio

A mixed linear model is established from the $Pb$ isotope ratio in order to trace the sources of Pb pollution and determine their relative contributions [18,45]. According to the results of the principal component analysis (see below), Pb in farmland topsoils can be assessed as having three components, which can be further explored by using the ternary linear model of the $Pb$ isotope. The following formula shows the ternary linear model.

$$\left\{\begin{array}{l}{\left(\frac{{206}_{Pb}}{{207}_{Pb}}\right)}_s=f_1\times {\left(\frac{{206}_{Pb}}{{207}_{Pb}}\right)}_1+f_2\times {\left(\frac{{206}_{Pb}}{{207}_{Pb}}\right)}_2+f_3\times {\left(\frac{{206}_{Pb}}{{207}_{Pb}}\right)}_3\\ {\left(\frac{{208}_{Pb}}{{206}_{Pb}}\right)}_s=f_1\times {\left(\frac{{208}_{Pb}}{{206}_{Pb}}\right)}_1+f_2\times {\left(\frac{{208}_{Pb}}{{206}_{Pb}}\right)}_2+f_3\times {\left(\frac{{208}_{Pb}}{{206}_{Pb}}\right)}_3\\ f_1+f_2+f_3=1\end{array}\right.$$

(2)

In these equations, $\frac{{206}_{Pb}}{{207}_{Pb}}$ represents the ratio of ${206}_{Pb}$ to ${207}_{Pb}$, $\frac{{208}_{Pb}}{{206}_{Pb}}$ represents the ratio of ${208}_{Pb}$ to ${206}_{Pb}$, and $f_1$, $f_2$ and $f_3$ represent the contribution rates of the three pollution sources to the $Pb$ present in the samples, respectively.

Table 2. Isotope ratios and elemental contents of lead sources.

Sample	Description	Pb	Cd	$\frac{{206}_{\mathit{P}\mathit{b}}}{{207}_{\mathit{P}\mathit{b}}}$	$\frac{{208}_{\mathit{P}\mathit{b}}}{{206}_{\mathit{P}\mathit{b}}}$
		mg·kg−1
S1	Agricultural Soil	36.1	0.64	1.190	2.059
S2	Agricultural Soil	14.3	0.73	1.202	2.050
S3	Agricultural Soil	27.9	0.49	1.206	2.062
S4	Agricultural Soil	21.5	0.63	1.213	2.041
S5	Agricultural Soil	23.8	0.57	1.204	2.038
W01	Phosphate Ore	24.3	1.13	1.396	1.743
W02	Coal Ore	22.1	9.94	1.271	1.966
W03	Dustfall	138	4.38	1.163	2.114
W04	Ardealite	35.3	0.46	1.358	1.785

shows that the Cd contents of the four pollution sources range from 0.22 mg/kg to 9.94 mg/kg, which is higher than the natural background value of 0.2 mg/kg. The Pb content primarily ranges from 20 to 35.3 mg/kg, which is close to the natural background value of 35 mg/kg. This shows that these sources have the potential to cause potential toxic metals (Pb and Cd) pollution. The contents of Pb (14.3 to 36.1 mg/kg) and Cd (0.49 to 0.73 mg/kg) have values higher than the natural background value, indicating that the agricultural soils in the study area were polluted.

Although phosphate and coal mining activities, as well as phosphating plants, are likely to be the main sources of lead in the study area, atmospheric coal burning, and automobile exhaust emissions also need to be considered. The emission of automobile exhausts and the burning of coal cause atmospheric depositions [18]. Zhao et al. found that the isotopic composition of lead in automobile exhausts was between 1.150 and 1.162 ($\frac{{206}_{Pb}}{{207}_{Pb}}$) and between 2.106 and 2.115 ($\frac{{208}_{Pb}}{{206}_{Pb}}$) [46]. The value range of the $Pb$ isotopes of the dustfall in the study area ($\frac{{206}_{Pb}}{{207}_{Pb}}$ = 1.163, $\frac{{208}_{Pb}}{{206}_{Pb}}$ = 2.114) is close to the value range of automobile exhaust, but far from that of the coal mine, phosphate ore and phosphogypsum. In addition, the study area has a well-developed transportation infrastructure, with a large volume of traffic and automobile exhaust emissions. This indicates that the source of the dustfall in the study area is dominated by automobile exhaust deposition.

Figure 3 shows that there is a linear relationship between the $\frac{{206}_{Pb}}{{207}_{Pb}}$ and $\frac{{208}_{Pb}}{{206}_{Pb}}$ of the four pollution sources and five agricultural soils. The $Pb$ isotope values of the five agricultural soils are located in the triangular end element region, composed of dustfall, coal ore and phosphogypsum; that is to say, these three layers play a major role in controlling the $Pb$ contents in the agricultural soils of the study area. In Table 2, the $\frac{{206}_{Pb}}{{207}_{Pb}}$ and $\frac{{208}_{Pb}}{{206}_{Pb}}$ values of the agricultural soils, dustfall, coal mine and phosphogypsum are substituted into equation group (1). The contribution of Pb from dustfall to agricultural soils is 74.94%, that from coal mines is 14.87%, and that from phosphogypsum is 10.19%.

The $Pb$ isotope analysis indicates that the Pb in the topsoils of the study area is primarily contributed by vehicle exhaust. Coal mining activities and the production of phosphorus products also contribute to the Pb in topsoils.

3.4. Multivariate Analysis Results

3.4.1. Correlation Analysis

A correlation analysis of 10 elements in the agricultural soils of the study area was carried out. As shown in

Table 3. Pearson correlation matrix for the metal concentrations.

Correlation	Cd	Pb	Cu	Zn	Cr	Ni	As	Hg	Se	P
Cd	1	0.000	0.000	0.000	0.010	0.006	0.000	0.244	0.000	0.000
Pb	0.624 2	1	0.010	0.000	0.930	0.798	0.000	0.581	0.000	0.045
Cu	0.796 2	0.415 2	1	0.000	0.000	0.000	0.178	0.001	0.000	0.000
Zn	0.881 2	0.873 2	0.727 2	1	0.093	0.069	0.000	0.323	0.000	0.000
Cr	0.413 2	0.015	0.594 2	0.276	1	0.001	0.393	0.276	0.265	0.014
Ni	0.441 2	−0.043	0.642 2	0.298	0.503 2	1	0.534	0.043	0.004	0.002
As	0.603 2	0.881 2	0.223	0.756 2	−0.143	−0.104	1	0.493	0.006	0.040
Hg	0.194	0.093	0.529 2	0.165	0.181	0.331 1	−0.115	1	0.189	0.102
Se	0.646 2	0.583 2	0.645 2	0.729 2	0.186	0.454 2	0.441 2	0.218	1	0.000
P	0.856 2	0.328 1	0.779 2	0.696 2	0.395 1	0.483 2	0.335 1	0.270	0.628 2	1

Note: The left lower part is the correlation coefficient; the right upper part is the significance level. ¹ p < 0.05 (2-tailed). ² p <0.1 (2-tailed).

, Pb was strongly correlated with Zn, Cd, As and Se, and the Cd was in correlation with Pb, Zn, Cu, As, Se and P. The correlation coefficients are all above 0.5, which indicates good correlation, and that the pollution sources of these elements are roughly similar. However, the correlation coefficients of each element are different, indicating that the levels of influence of the different pollution sources are different. Moreover, the correlation coefficient between Cd and P is 0.856, which indicates that the Cd may be affected by phosphate mining or processing. Excluding Cu, the linear correlation between Ni and the other elements is not significant. This indicates that Cu is affected by natural processes. However, there is a significant linear correlation between Cu and Cd, Se and P. This indicates that Cu is affected by human activities, related to phosphate mining or processing.

3.4.2. Principal Component Analysis

Given the good correlation of Cd, Pb, Zn, Cu, As, Se and P in the soil of the study area, common factors can be extracted. To further explore the sources of Pb and Cd, the principal component analysis method was used. The results show that the KMO value of the seven elements was 0.763, which means they are suitable for principal component analysis. The Sig index of the Bartlett spherical test was 0, while the approximate chi-square (331.76) and df index (21) were larger, indicating that the Bartlett spherical test achieved significance. There were an obvious linear correlation between the soil samples, and principal component analysis could be carried out.

Based on the results of the $Pb$ isotope analysis, all seven initial eigenvalues extracted from the potential toxic metals concentration values are processed by the maximum variance method. The analysis results are shown in

Table 4. Total variance explained and component matrices illustrated (four factors selected).

Component	Initial Eigenvalues			Extraction Sums of Squared Loadings			Rotation Sums of Squared Loadings
	Total	% of Variance	Cumulative (%)	Total	% of Variance	Cumulative (%)	Total	% of Variance	Cumulative (%)
1	4.89	69.80	69.80	4.89	69.80	69.80	2.57	36.74	36.74
2	1.25	17.83	87.63	1.25	17.83	87.63	1.99	28.43	65.17
3	0.44	6.29	93.93	0.44	6.29	93.93	1.14	16.21	81.39
4	0.28	3.97	97.89	0.28	3.97	97.89	1.08	15.40	96.79
5	0.07	0.99	98.89
6	0.06	0.86	99.75
7	0.02	0.25	100.00

Note: Extraction method: principal component analysis.

. The results show that seven elements were extracted from the agricultural soils, including Pb, Zn, Cd, Cu, As, Se and P. The eigenvalues of the first four were all greater than 1, and the cumulative variance contribution rate was over 95%. The variance contribution rates of the fifth, sixth and seventh factors were very small, and their influences can be neglected.

According to the analysis results, the main characteristics of the elements (variables) were calculated via the criteria of eigenvalue > 1 and cumulative variance contribution > 85%. The results are shown in

Table 5. Rotated component matrix for the data of the topsoil.

Element	Component
	1	2	3	4
Pb	0.915	0.082	0.259	0.227
Cd	0.458	0.718	0.210	0.370
Zn	0.697	0.455	0.320	0.383
Cu	0.123	0.526	0.293	0.788
As	0.956	0.204	0.113	−0.041
Se	0.301	0.321	0.866	0.234
P	0.136	0.916	0.270	0.257
Eigenvalue Percent of Variance	36.744	28.427	16.213	15.404
Cumulative Percent	36.744	65.171	81.385	96.788

Note: Extraction method: principal component analysis. Rotation method: Varimax with Kaiser normalization. Rotation converged after fixed iterations.

.

Combined with the factor load matrix results, the Pb content can be assessed to be mainly controlled by the first main factor, and the contributions of the third factor and the fourth factor decrease. Pb and As showed high similarity in terms of the first major factor, indicating that they have similar provenances. Further, this factor represents the main source of Pb in the agricultural soils of the study area. Combined with the results of the $Pb$ isotope, the first factor can be determined to be dustfall. Previous studies have found that the deposition of vehicle exhaust can lead to an increase in the Pb and Cd concentrations in dustfall [47,48], and that dustfall is an important factor leading to the enrichment of Cd and Pb in the soil of the study area. The Cd is mainly affected by the second main factor. The second main factor is very similar to P. Field investigation shows that the phosphorite ores produced upstream of the study area contain svanbergite, which is the main useful mineral produced in this process. The P in the mineral forces the Cd from the ores to migrate into the water system in large quantities, and from there into the study area. Therefore, the second main factor is determined to be the phosphate mine. Se and the third factor show very high similarity. By comparing the Se contents of several pollution sources, we see that stone from the coal mine is rich in Se. Combined with the results for the $Pb$ isotope, this factor is determined to be the coal mine. Cu shows very high similarity with the fourth factor. From the geochemical affinity of the elements, we can infer Cu’s sulfur-affinity characteristic [49]. According to field observations, phosphate ore production and processing enterprises are present upstream of the study area. A number of sulfur-affinity elements are contained in phosphate ore. Cu mainly migrates through processes of hydrolysis, precipitation, ion exchange and adsorption [50]. In the process of phosphate ore treatment, the Cu moved into the Mianyuan River and migrated via the co-precipitation or adsorption of colloidal Al and Fe sediments. Finally, it entered the agricultural soils of the study area through irrigation channels. Combined with the results for the $Pb$ isotope, it can be determined that the fourth main factor is the phosphogypsum produced by the phosphate plant. The sources of Cd can be ranked, in terms of contribution, as follows: phosphate mine > dustfall > phosphate plant > coal mine.

Our analysis of source contribution indicates that the concentrations and pollution degrees of Cd and Pb in farmland topsoils are related to industrial production, mineral mining activities and land use patterns. The Pb in the farmland topsoils of the study area is primarily contributed by vehicle exhaust. Coal mining activities and the production of phosphorus products also contribute to the Pb levels in the topsoils. Cd is mainly contributed by phosphate mining, dustfall, phosphate plant activities and coal mining, and phosphate mining contributes the most to the Cd levels in the soil. The most severe metallic pollution of the farmland topsoils was found in the northwest of the study area and the residential area, which are priority areas for pollution control. Pb has low chemical mobility during soil weathering, and easily forms stable complexes in topsoils [51]. Cd has greater chemical mobility than Pb, but it can be adsorbed by clay minerals. A series of physical and chemical measures should be taken to immobilize the Cd and Pb in topsoils [52,53].

4. Conclusions

The agricultural topsoil in the study area is polluted by Pb and Cd. Cd displays clear regional enrichment characteristics, while Pb shows local enrichment characteristics. The presence of both of these is mainly determined by human activities. The contents of Pb and Cd decrease from NE to SW and tend to decline as one moves away from the Mianyuan River. The factors influencing topsoil Pb pollution can be ranked as follows: dustfall > coal mine > phosphogypsum. The factors influencing topsoil Cd pollution can be ranked as follows phosphate pine > dust fall > phosphate plant > coal mine.