3 Results and discussion
3.1 Soil HMs contents
The respective HM (Cu, Zn, Pb, Cd, As, Hg, Cr, and Ni) concentration in PC topsoil (0-20 cm) is presented in Table 1. Compared with the soil environmental background values in various provinces and the national standard, the amounts of 8 HMs were increased to varying degrees; the increase was more prominent for Cu, Zn, Pb, Cd, and Hg. In the whole area, soil amounts of Cu, Zn, Pb, and Cd were 2.09, 1.56, 1.21, 4.70, and 1.06 times higher than the national background value, respectively. Cr and Ni were 1.13 and 1.15 times higher, respectively. The corresponding soil samples in provinces showed a similar trend. In Liaoning province, the greenhouse soil concentration of Cu, Zn, Pb, Cd, and Hg reached up to 1.87, 1.77, 1.10, 1.82, and 1.65 times the background value, respectively. Likewise, compared to the background value, in Hebei province, Cu, Zn, Pb, Cd, and Hg were higher by 1.07, 1.40, 1.03, 2.44, and 0.89 times; in Shandong province, the increase was 1.18, 1.95, 0.76, 3.38, and 2.10 times; In Henan province, the increase was 3.33, 2.78, 1.26, 8.28, and 1.82 times; in Shaanxi province, the increase was 1.24, 1.26, 1.04, 1.69, and 6.92 times; in Ningxia Autonomous Region, the increase was 1.14, 1.39, 0.84, 1.55, and 1.57 times; in Jiangsu province, the increase was 0.77, 1.26, 0.76, 0.92, and 1.85 times, and in Yunnan Province, the increase was 2.30, 1.57, 1.73, 5.82, 2.16 times, respectively.
Based on GB15618-2018 standard and the distribution point exceeding the rate of soil HMs, Cd (27.3%) was critically above the over-standard rate, followed by Cu (15.2%), Zn (4.5%), Pb (3.0%), and As (2.3%), while Hg and Cr were within the over-standard rates. The over-standard rate of the 8 HMs in greenhouse soil from respective provinces also showed higher levels of Cd. Yunnan province was most significantly affected (84.0%) followed by Henan (50.0%), Liaoning (16.7%), Shandong (11.8%), and Hebei (6.7%) provinces, respectively. Shaanxi province, Ningxia Autonomous Region, and Jiangsu province had Cd within the standard range. In general, Yunnan province was most significantly affected; apart from Cd, the over-standard rates of Cu, Zn, Pb, As and Ni reached 64.0, 4.0, 16.0, 12.0, and 8.0%, respectively. Next was Henan province with 25.0% over-standard rates for Cu and Zn. Studies showed that HMs enrichment occurs under low pH conditions, mobility, and biological activity (Lu et al., 2020). Notably, PC soil pH and in general soil pH of the whole area was 0.8 units higher than the average background value in China. This can be attributed to soil conditioners or remediation agents used in PC (Hamid et al., 2019; Lebrun et al., 2020; Wu et al., 2016). However, some areas, such as Liaoning, Henan, and Jiangsu, showed acidification characteristics, which could promote soil HMs exposure to the food chain causing health risks.
3.2 Risk assessment of soil HMs
Next, we calculated the SFPI, NCPI (Pnc ), and soil comprehensive quality index (IICQ) to evaluate the cumulative risk of HMs in PC soil in the whole area. The results are shown in Figure 1. We found that the whole area average Pnc and IICQ reached up to 1.03 and 1.71, suggesting a slightly contaminated level. Cd contributed the largest; the average PCd was 1.25 at a slightly contaminated level. The remaining 7 HMs were within the safe limits.
Among the different test regions, the pollution risks are more serious in Henan and Yunnan provinces. Yunnan province is at the highest cumulative risk with the NCPI (Pnc average 2.92) reaching to moderately contaminated level andIICQs (average 6.04) reaching the extremely contaminated level. Cd (PCd average 3.71) and Cu (PCu average 1.52) contributed most with the average SFPI (Pi ) at the heavily and slightly contaminated levels, respectively. The average NCPI (Pnc 1.23) and IICQs (2.10) respectively suggested that Henan Province is at slightly and moderately contaminated levels. Here too, Cd contributed the highest with the SFPI (PCd 1.60) at a slightly contaminated level. Meanwhile, the PC soils in the remaining 6 provinces are at the uncontaminated or safe level.
Excessive soil HMs accumulation, with concerns of phytotoxicity and risk to human health via the food chain, has become one of the prime environmental issues worldwide (Ye et al., 2020; Corguinha et al., 2015; Zhang et al., 2018). Cd is considered redundant for metabolism and other biological function and has been listed in the top 20 hazardous substances by the United States Environmental Protection Agency and the Agency for Toxic Substances and Disease Registry (ATSDR) (Rai et al., 2019; Khalid et al., 2017). On the contrary, although Cu is relatively low-toxic, excessive Cu concentrations may pose a serious health risk, such as liver and kidney damage.
3.3 Sources of soil HMs
Principal component analysis (PCA), a classic multivariate statistical method is popularly used to identify the natural and anthropogenic sources of soil HMs (Chen et al., 2019). Kaiser-Meyer-Olkin (KMO) and Bartlett tests were performed to analyze the concentration data of HMs. The KMO value at 0.76 >0.5 (P <0.05) suggested that the data was suitable for PCA analysis. Based on the principal component analysis, Kaiser standardized orthogonal rotation method, and the maximum variance method, the factor loading matrix was orthogonally rotated. The cumulative contribution rate (76.78%) of the variance of the soil HMs concentration eigenvalues, principal factor 1 (PF1), and principal factor 2 (PF2) with eigenvalues >1 explain the sources of HMs in the PC soil (Table 2). The variance contribution rate of PF1 was 44.35%, and the PF1 factor loadings for Pb, Cd, and As were 0.894, 0.840, and 0.924, respectively. The variance contribution rate of PF2 was 32.43%, and the factor loadings for Cu and Zn were 0.930 and 0.713, respectively. The other three elements, Hg, Cr, and Ni were mainly affected by PF1 and PF2, with factor loadings for Hg reaching 0.756 and 0.433, Cr 0.417 and 0.625, and Ni 0.626 and 0.656, respectively.
Next, we analyzed the Pearson correlation (Table 3) between the HMs concentration and physiochemical indicators (the proportion of sand, powder, clay, soil pH, SOM, and CEC). We found that Cu and Zn showed a significant correlation (R =0.547, P <0.05) with 6 soil physiochemical indices (P <0.01). The correlation coefficient (R -value) between Cu concentration and soil organic matter (SOM) reached 0.482, and between Zn concentration and SOM reached 0.453, respectively. These indicated that soil Cu and Zn concentrations were strongly correlated to SOM. Notably, SOM in the PC soil mainly originates from the return of crop residues and organic fertilizer. Furthermore, PF1 can be attributed to chemical fertilizer, organic fertilizer, and other anthropogenic activities (Geng W et al., 2013; Sager, 2007; Liu et al., 2018; Huang et al., 2017). Soil Pb, Cd, and As exhibited a very significant correlation with soil physical indicators (the proportion of sand, powder, and clay) (P <0.01), but not with soil chemical indicators (pH, SOM, CEC) (P>0.05). This suggests that Pb, Cd, and As are mainly from soil geological backgrounds, such as soil parent material and the PF2 can be attributed to soil parent material and other natural factors (Yang et al., 2020). Some studies showed that livestock manure and chemical fertilizers (especially phosphate fertilizers) can be a source of Pb, Cd, and As (Luo et al., 2009); however, compared with the geological background source factor, the contribution of fertilization was found minimal for Pb, Cd, and As soil contamination in PC. Moreover, Hg, Cr, and Ni contents in PC soils were found affected by anthropogenic agricultural activities and natural factors such as soil parent material.
3.4 Relationship between the soil HMs contents and planting year
HMs have an obvious cumulative effect in the PC soil; the over-standard GB15618-2018 rates of Cu, Zn, Pb, Cd, and As exceeded 2%. Specifically, in the study area, the accumulation of Cu and Cd is at serious environmental risk. Several studies suggest that the available soil concentration is the bioavailable part of the total HMs concentration (Fang et al., 2018; Xia et al., 2016; Wang et al., 2020). Excessive bioavailable HMs in the soil get easily absorbed into the food chain (crops) or leached by irrigation water as agricultural non-point source pollution (Natasha et al., 2021). To further understand the relationship between the available concentration of five HMs (Cu, Zn, Pb, Cd, and As) and the planting years in PC, we did a correlation analysis (Fig. 2). We found that the available Cu (Fig. 2 a), Zn (Fig. 2 b), and As (Fig. 2 e) exhibited a significant (P <0.05) linear correlation with the planting years, while the available Cd (Fig. 2 c) and Pb (Fig. 2 d) did not (P >0.05). This further indicates that Cu and Zn are mainly derived from frequent anthropogenic agricultural activities, and the cumulative effect and potential cumulative risk increase with the extension of planting years in PC. Likewise, available As showed a significant linear relationship (Fig. 2 e) with the planting years; however, the main source of As is the soil parent material. This suggests that frequent anthropogenic agricultural activities in the PC can increase total As conversion into available As (Antoniadis et al., 2019). Available Pb and Cd did not correlate with the extension of planting years, indicating a minor contribution of anthropogenic agricultural activities compared to soil geological background value.