Excel was used to carry out the statistics and calculations of the heavy metal contents and ground accumulation index. ArcGIS 10.8 was used to produce the overview map of the study area and the spatial distribution maps of the heavy metals. Origin 2024 was used to plot the ground accumulation index. SPSS 26 was used to conduct the principal component and correlation analyses; and EPA PMF 5.0 was used to analyze the heavy metal pollution sources.
The environmental quality of the soil in the study area is presented in Table 1, which shows that the soil in the study area is generally neutral. Eight heavy metal elements were within the medium-high intensity variation range. Hg and Pb were within the high intensity variation range, and the order of the values of the coefficients of variation are as follows: Pb > Hg > Cd > Zn > As > Cu > Cr > Ni. Compared with the background values of the soil heavy metals in Yunnan Province, the average contents of the eight heavy metals were higher than the background values of the soil, i.e., 1.26 to 6.01 times higher. The pollution levels of the eight heavy metals in the soils in the study area were evaluated using the soil pollution screening values in the Soil Environmental Quality Risk Control Standards for Soil Pollution on Agricultural Land (Trial) (GB 15618-2018) as reference values (Table 2). The heavy metal contents of all of the soil samples collected from the study area exceeded the standard values, except for Hg, and their exceedance rates were Cd (80.0%), Cu (52.0%), Cr (36.0%), As (35.5%), Pb (30.0%), Ni (29.5%), and Zn (23.8%).
Inverse distance weighted (IDW) interpolation was used to map the spatial distributions of the heavy metal contents. By analyzing the natural turning points and characteristic points, the heavy metal contents were classified into clusters with similar properties to maximize the differences between the classes (Fig. 3). Cd, Pb, Zn, and Cu exhibited similar piecewise enrichment characteristics in the study area, and the high-value areas were mainly concentrated in the southern part of the study area, demonstrating that human activities had a significant influence on the distributions of the soil heavy metals in the areas close to the industrial concentration areas and near the main transportation routes. Cr and Ni exhibited similar piecewise enrichment, and there was a strong spatial correlation between the distributions of As, Hg, and pH. In the high-value heavy metal content areas, the pH value was correspondingly higher. Based on analysis of the coefficients of variation, the spatial distribution patterns of these elements in the study area are similar, indicating that the distributions of these elements are influenced by similar factors.
The results of the evaluation of the soil geoaccumulation index of the cultivated land in the study area are shown in Fig. 4. The results show that the ranges of the land accumulation indices for Cu, Zn, As, Hg, Cd, Pb, Ni, and Cr were - 1.5-2.8, - 1.6-3.3, - 3.4-2.7, - 2.1-4.5, - 1.0-4.9, - 2.1-6.0, - 1.7-2.4, and - 0.9-2.8, with mean values of 0.6, 0.5, - 0.4, 0.3, 1.5, 0.6, 0.7, and 0.9, respectively. The order of the pollution degrees of the elements was Cd ≥ Cr ≥ Ni ≥ Pb ≥ Cu ≥ Zn ≥ Hg ≥ As. The spatial distributions of land accumulation indices for the eight heavy metals exhibited different degrees of variations (Fig. 5). Among then, Cd and Pb had the highest land accumulation pollution indices, reaching 4.9 and 6.0, respectively, indicating that there was a significant anthropogenic pollution source in localized areas. These areas were mainly concentrated in the southern part of the study area and near the industrial and transportation areas, suggesting that industrial activities and transportation emissions were the key factors influencing the accumulation of these heavy metals. Comparatively speaking, the geoaccumulation index of As was lower, indicating that the degree of accumulated pollution was slight and was closer to the natural background value. The overall level of the soil heavy metal accumulation in the study area was mainly slight to mildly polluted, and it did not reach heavy or serious pollution. The pollution level was greater in the southern part of the study area than in the northern part of the study area.
Pearson correlation analysis was performed on the contents of the eight heavy metal elements. The results are shown in Fig. 6. The results showed that many of the heavy metal elements and the pH exhibited significant correlations, indicating that they had common sources and similar migration and transformation processes. Cu, Zn, Cd, and Pb exhibited significant positive correlations, indicating that these four elements had similar input pathways, such as industrial emissions, mining activities, and/or agricultural inputs. Additionally, the correlations of As with Hg, Cd, and Pb further revealed the contributions of deposition and/or anthropogenic disturbances to the accumulation of these heavy metals. In addition, the independence of Ni and Cr indicated that they originated from natural geochemical processes rather than from significant exogenous pollution inputs. In addition, the pH, as an important environmental factor, profoundly affected the morphology, solubility, and transport behaviors of the heavy metals in the soil.
Multiple linear regression analysis of the eight heavy metal elements revealed that the Bartlett's sphericity test companion probability was 0, which was lower than the significance level of 0.05, and the Kaiser-Meyer-Olkin (KMO) test statistic value was 0.753, indicating that the sample data for the study area was suitable for PCA. Through PCA, three principal components were extracted (Table 3). The first principal component (F1) had a variance contribution rate of 44.78%, and it explained the largest proportion of the variance in soil heavy metal distributions, corresponding to high concentrations or strong symbiosis of the heavy metal fractions. The second principal component (F2) and the third principal component (F3) contributed 24.53% and 11.22%, respectively, with a cumulative variance contribution of 80.53%, which explained most of the information about the heavy metal pollution.
The explanatory degree of F1 was significantly higher than those of the other principal components, and the heavy metals with higher loadings were Hg, Cu, Zn, Cd, and Pb. Lv et al. (2013) and Wang et al. (2021) showed that when Cd, Pb, and Zn were classified into the same principal component, they were mainly affected by anthropogenic sources, and the contribution of Hg was the highest (66.99%). Studies have also shown that Hg is affected by multiple factors, such as soil-forming matrices, industrial emissions, transportation, and agriculture, and that bituminous coal has a significant effect on its accumulation in soil. In addition, transportation exhaust, dust, and particles can all contribute to the increase in soil Hg after atmospheric deposition. According to the spatial distribution characteristics and correlation analysis of the heavy metals, Cu, Hg, Zn, Cd, and Pb were highly enriched and correlated in the southern part of the study area, and the pollution was particularly serious in the industrial parks and near the transportation routes, indicating that in the study area, these metals were greatly influenced by industrial and transportation sources. Based on the high value areas in the spatial distributions of these five elements and the distributions of the industrial zones and transportation routes, it was concluded that F1 was an industrial and transportation source.
The F2 source contributed the most to Cr and Ni, reaching 96.99% and 93.51%, followed by Cd (25.26%). Some studies have shown that Cr and Ni are mainly affected by the soil parent material and soil-forming processes. The ground accumulation indexes and comprehensive pollution indexes of Ni and Cr in the study area did not indicate pollution, and their coefficients of variation were relatively small, further indicating that the Ni and Cr in the soil in the study area were less influenced by exogenous sources and mainly came from the parent material of soil formation processes. Therefore, we concluded that F2 was the parent material source.
Source F3 contributed the most to As, Zn, and Pb, with values of 51.75%, 26.18%, and 22.95%, respectively. According to the preliminary investigation, this area is rich in mineral resources, and there are non-ferrous metal resources such as gold, lead-zinc, and copper ores in the study area, which is an important mineral development zone in Yunnan Province. As is an accompanying element in gold ores. Based on the fact that the areas with high As, Zn, and Pb values overlapped with smelting plant locations, we concluded that F4 was a mining source.
Unknown source F4 mainly contributed to As, Pb, and Hg, with contribution values of 26.49%, 25.52%, and 19.86%, respectively, although these contribution values were not high. Hg and Pb are important components of pesticide fertilizers, can enter the soil through their application and become enriched on the soil surface. The main land use type in the study area is arable land, suggesting that this part of the impact originates from agricultural activities; therefore, we conclude that F4 is an agricultural source.
The APCS-MLR model was used to analyze the eight heavy metals via multiple linear regression (Fig. 7). The predicted values fit well with the measured values, and the results show that the R values of Cu, Zn, As, Hg, Cd, Pb, Ni, and Cr are 0.687, 0.867, 0.841, 0.497, 0.810, 0.838, 0.949, and 0.942, with an average R value of 0.804. The estimated/measured values (E/O) of all of the elements were close to 1, indicating that the APCS-MLR model analysis results have a high credibility. Notably, the R value of Hg was relatively moderate (0.497), which may be attributed to the discontinuous and nonlinear nature of agricultural Hg inputs, such as sporadic application of Hg-containing fertilizers or pesticides. The industrial and transportation activities contributed substantially to the Cu, Zn, Hg, Cd, and Pb pollution. The soil matrices had a significant influence on the Cr, Ni, and Cd pollution and were the main sources. The mining industry also had a large influence on the As, Zn, and Pb pollution.
In order to verify the reliability of the source analysis results of the APCS-MLR model, the PMF model was used to conduct quantitative source analysis of the eight heavy metals. Three to five factors were selected for random iterative operations, and four factors were finally determined to obtain the contributions of the four heavy metal pollution sources to each heavy metal (Fig. 8).
The elements with high loadings for Factor 1 were Pb, Cd, Zn, and Cu, which were mainly found in mining areas, transportation routes, and other areas with intensive human activities. This aligns well with the results of the APCS-MLR model analysis. The results of both models indicated that Factor 1 was an industrial and transportation source. Factor 2 contributed the most to Cr and Ni, which had small coefficients of variation. The APCS-MLR model results were highly consistent with the results of the APCS-MLR model, and both models indicated that Factor 2 was a parent material source. Factor 3 had the highest loading for As, which aligned with the fact that the geologic background of the study area was rich in mineral resources and is an important development zone for non-ferrous metals, such as gold, in Yunnan Province. This is essentially the same as the results from the APCS-MLR model, and both models indicate that Factor 3 is a mining industry source. Factor 4 had high loadings for Hg and Cd. Several studies have shown that fertilizers and pesticides often contain elements, such as Hg, Cd, and Pb, which can be indicative of agricultural activities such as the application of pesticides and fertilizers. Thus, we concluded that Factor 4 was an agricultural source. The PMF model identifies the four sources of contamination of the eight heavy metals as industrial traffic, parent material, mining, and agricultural sources.
The results of the pollution source contribution analysis using the PMF model (Fig. 8) correspond well with the results of the correlation analysis (Fig. 6) and the APCS-MLR model analysis (Fig. 7). The industrial and transportation activities contributed more to the Cu, Zn, Hg, Cd, and Pb pollution. The soil matrix had an important influence on the Cr and Ni, which aligned with the independence of several elements in the correlation analysis and the natural sources corresponding to F2 in the APCS-MLR model. The mining sources had an influence on As, Hg, and Cd, which aligned with the analysis of F3 in the APCS-MLR model, and the fact that agricultural sources had a greater influence on Hg and Cd. This conclusion is reflected by the results presented in Fig. 6, Table 3, and Fig. 7, which show the main composition of the soil heavy metal pollution sources and their relative contributions in the study area.