In October 2016, samples were collected from 19 different sites in Wuliangsuhai Lake: S1-S17, A and Z from the overlying water (the inlet and outlet of the lake, respectively) The data sets of Chinese Map and Wuliangsuhai Lake are provided by Data Center for Resources and Environmental Sciences, Chinese Academy of Sciences (http://www.resdc.cn); the date set of Hetao irrigation system was cited from Liu et al.^[6] and modified by Photoshop (Fig.1). Seventeen sediments (S1— S17) were also gathered by a gravity corer. Ten P.pectinatus (S1, S4— S10, S13, and S17) and 8 P. australis samples (S2, S3, S7— S10, S15, and S16) were acquired from where they were growing. 43 freshly caught lake fish (n=10~15 individuals per species) were purchased from the fishermen of Wuliangsuhai Lake fishery, in which four species of fish were crucian, carp, catfish and mullet highly consumed by local residents.

Fig.1

	Figure Option ViewDownloadNew Window
	Fig.1 Map of Wuliangsuhai Lake and the sampling locations

The collected plant, sediment and fish samples were digested with HNO₃-H₂O₂ (guaranteed reagent, GR), HNO₃-H₂O₂-HF (GR) and HNO₃ (GR), respectively. Then, the heavy metals concentrations were measured by using inductively coupled plasma mass (ICP-MS) spectrometry (Agilent Technologies, USA), and the operating parameters were listed in Table 1.

Table 1

Table 1

Table 1 Operating parameters of ICP-MS Parameters Numerical Value RF power/W 1 550 Nebulizer gas flow/(L· min-1) 0.9(0.6~1.0) Auxiliary gas flow/(L· min-1) 0.25(0.3~1.0) Sampling depth/mm 8(7~10) Peristaltic pump speed/(rps) 0.1(0.1~0.3) Sample cone/mm 1.0 Skimmer cone/mm 0.7

Table 1 Operating parameters of ICP-MS

The enrichment factor (EF), bioconcentration factor (BCF) and the biota-sediment accumulation factor (BASF) were used to determine the enrichment characteristics of the HMs, which were referenced from Ref.[8] and [9]. In sediment, the potential ecological risk index (R_I) was used to evaluate the potential ecological risk of metal contaminants. The R_I grading standard was modified by using the weighted average method^[10]. Data analysis was performed using the SPSS 22.0 software package for Windows. Principal component analysis (PCA) was used to determine the relationship among HMs in sediments and their possible sources. The spatial distribution patterns of the R_I value at each site were obtained by using geostatistical analysis in geographic information system (GIS) software (Arc GIS, version 10.2)

In this study, the mean levels of Cr, Ni, Cu, Mn, Pb and Zn mainly followed an order of sediment> P. pectinatus (submerged plants)> P. australis (emergent plants)> fishes> overlying water (Fig.2), but we found that the concentration distribution of As in overlying water was higher than that in P. australis and fishes possibly because As and P are cognate elements with similar physical and chemical properties and that P from agricultural runoff competed with As for adsorption points on the oxide surfaces of solid-liquid interfaces, causing As to desorb and ultimately drain into the channel^[6]. Additionally, it was interesting that the order of the Cd content— P. australis> fishes> P. pectinatus> sediment (Cd was not detected in the overlying water)— was different from that of the other elements, and the content of Cd in P. australis was almost 50 times higher than that in normal plants (0.05 to 0.2 mg kg^{-1 [5]}); in four types of fish, Cd was 3.3 times higher than the permissible threshold standards in China, leading to inhibition of plant growth, excessive accumulation of Cd in plants, and potential hazards to fish and human health via food chain bioaccumulation^[11].

Fig.2

	Figure Option ViewDownloadNew Window
	Fig.2 HM contents in overlying water-sediment-plant-fish in Wuliangsuhai Lake

Water contents (μ g· L^-1) were shown in a line graph; contents (mg· kg^-1) in sediment, P. pectinatus, P. australis and fish were shown in a bar graph; and their respective Chinese national quality values (solid lines) are shown for the collected samples According to the EF grading criteria, As and Cd experienced moderately severe enrichment and might originate from industrial and agricultural nonpoint source pollution [Fig.3(a)]. Based on the BCF and BASF, we found that in P. pectinatus and P. australis [Fig.3(b), (c)], the BCF of all metals was greater than 1, but the BASFs were lower than 1, except Cd, demonstrating that P. pectinatus was more likely to accumulate HMs from overlying water because its leaf surface had a special porous structure that increased the area available for adsorption and promotes the adsorption of metal ions^[11]. Finally, in the crucian, carp, catfish and mullet, the BCFs and BASFs were almost less than 1 (except for the BCFs of Cr and Ni and the BASF of Cd) [Fig.3(d)]. These results indicated that these fish species tended to restrict their uptake of metals from water and sediments to guarantee their growth; however, the bioaccumulation level of metals in fish might not cause any adverse effects in the fish itself but might pose a hazard to their predators, such as piscivorous birds, mammals and human^[1].

Fig.3

Figure Option ViewDownloadNew Window

Fig.3 Box-Whisker plots of the EFs in surface sediments and BCF/BASF in plants and fishes (a): EFs in sediment; (b): BCFs/BASFs in P. pectinatus; (c): BCFs/BASFs in P. australis; (d): BCFs/BASFs in fish. The box shows the 25th and 75th percentiles, and the whiskers represent the smallest and the largest values. The red line represents the values of 2 and 5, the dotted line shows EF=1 in Fig.3(a), and the red lines stand for a value of 1 in Fig.3(b)— (d)

The contents of As, Cd and Pb in sediments were 6.4, 5.8 and 3.1 times higher than the soil background values and exceeded the standard limits of 100%, 18% and 18% at all sites (Fig.2), respectively, demonstrating that they could affect lake sediment quality.

Based on the ecological risk grades of a single metal, the Erivalues of the sediment at S6 were higher than those at the other sites, indicating higher HM pollution (Fig.4). In all sediments, because the Erivalues were lower than 40, Cr, Mn, Ni, Cu, Pb and Zn posed low ecological risk. However, moderate risks posed by As were found in all sediments. At S5, S6 and S17, Cd showed rather higher Erivalues (> 500), indicating very high ecological risk. However, in other sediments, Cd posed considerable risk. R_I represents the sensitivity of various biological communities to toxic substances and illustrates the potential ecological risk posed by HMs^[9]. The results showed that the R_I values at all sites were higher than 105, indicating moderate to considerable ecological risk. Furthermore, the sediment at S6 posed the highest risk compared with the sediments at the other sites. The areas with higher potential ecological risks posed by HMs were mainly distributed in the northwest and southwest areas of the lake (Fig. 4), but the eastern area of the lake had low risk, which was mainly caused by natural conditions and human activities.

Fig.4

	Figure Option ViewDownloadNew Window
	Fig.4 Evaluation on potential ecological risk of HM pollution in sediments of the Wuliangsuhai Lake

The results of principal component analysis (PCA) revealed that there were three principal components (PCs) for the sediment samples, with an explanation of 86.33% (PC1 30.3%; PC2 27.17%; PC3 25.94%) of the total variance (Fig.5). In Fig.5(c) and (d), we found that sites S4, S6, S7 and S17 apparently had a high correlation with PC1, PC2 and PC3, indicating that the HM contents at these sites were high, which was in agreement with the R_I assessment showing a very high risk at S6 and S17. Fig.5(a) and (b) showed that PC1 had high loadings for Pb (r=0.97), Cu (r=0.818), Mn (r=0.781), Ni (r=0.76) and Cr (r=0.743), which are abundant elements in nature. Clay minerals were derived from weathered rock basins and reach lake water by leaching in surface runoff, finally accumulating in sediments^[6]. The sediments in Wuliangsuhai Lake were mainly formed from clay; therefore, Cu, Fe, Mn and Cr were the main components of the sediments and weathered products. The ecological risks posed by Cr, Cu and Pb were all low in this study, indicating that the impacts from human activities on these elements were small. Therefore, PC1 was mainly associated with natural weathering and deposition.

Fig.5

	Figure Option ViewDownloadNew Window
	Fig.5 Loading and score diagram from the PCA of HMs and other factors in Wuliangsuhai Lake (a), (b): The arrows with different colors represent the variables belonging to diverse components; (c), (d): yellow, green and gray dots represent relatively lower, intermediate and relatively higher levels of scores for the sampling sites, respectively

PC2 had high positive loadings for Zn (r=0.847), Cd (r=0.625), and pH (r=0.876), which were deemed to have primarily anthropogenic sources. We found that around Wuliangsuhai Lake, there were dozens of chemical fertilizers and mineral processing enterprises, such as the Huafeng zinc oxide operation, the Tanyaokou pyrite operation, mining plants, the Bameng Ring Farm mine, the Urals fertilizer plants, the chemical plants in Wuyuan and the fertilizer plants in Linhe (Fig.1). These operations were located in drainage channels that empty into the lake, and industrial wastewater and municipal sewage were the main potential sources of pollution by HMs, such as Cd and Zn^[12]. Therefore, PC2 in sediments was mainly an industrial sewage and waste source. In this study, the weakly alkaline conditions (pH 7.8~8.7) were unfavorable for HM release from sediments. In addition, Cd had low stability and was easily released from sediments to water with environmental changes (pH, redox potential, temperature, etc.) to become bioavailable^[13]. The sources, including those of Zn and Cd, and the correlations with pH were mainly characterized by PC2.

PC3 had high positive loadings for As (r=0.84), organic matter (OM) (r=0.867), and TOC (r=0.811). The heavy use and low efficiency of chemical fertilizers have resulted in large amounts of phosphate residues in agricultural runoff. As mentioned above, phosphorus competes with As for adsorption to sediments; thus, phosphate in agricultural runoff will compete for the adsorption sites on oxide surfaces, causing As to desorb from oxide surfaces and ultimately drain into the channel. Huang et al. (2008)^[14] showed that the average content of TP in the Hetao Irrigation District was 1.99 mg· kg^-1, exceeding Grade V (0.2 mg· kg^-1 in lakes) of the Surface Water Quality guidelines in China (GB 3838— 2002). It was likely that an increasing amount of As in the water and soil would be released into the channel and the lake, that is, the As pollution was partly caused by agricultural nonpoint source pollution and agricultural runoff.

Overall, PC1 was the result of natural weathering and deposition, and PC2 was primarily due to anthropogenic sources, and PC3 was mostly a result of agricultural nonpoint sources.