Elemental Metals in Agricultural Soils Along Two Tributaries of Chaohu Lake: Implication on Soil Erosion Caused Enrichment in Sediments

Agricultural soil as a potential source of input of heavy metals into the lacustrine sediments of Chaohu Lake via soil erosion and weathering has not been well documented. In the present study, rice soil samples along two main tributaries of Chaohu Lake were collected to determine 30 metal elements which were compared with the data on the elements in surface sediments of Chaohu Lake to explore the impacts of soil erosion on the accumulation in the sediments. The results showed that the concentrations of most elements were less than their soil background values of Anhui Province, but Se, Sb, Sn, As, and Ba were greater than the corresponding background values. Furthermore, these elements except for Sc in the soils appeared a low variability in spatial distribution. Enrichment factor (EF) also suggested that most elements were minimally enriched, but Se, Th and As were found with moderate enrichment in the study areas. More than 90% of agricultural soils were not expected with significant contamination by the residual elements based upon the consideration of pollution level index (PLI) combining probabilistic density function. Compared with the residual elements in the surface sediments of Chaohu Lake, most elements (Ti, Mn, Zr, Sr, Zn, Ga, Hf, U, Se, Sn, Nb, Th, V, Cd) appeared depletion in agricultural soils but enrichment in sediments. Specially for Cd, it was not detectable in all soils samples, but it existed at an extreme average content in the lacustrine sediments. These results possibly implied that elements in agricultural soils in Chaohu Lake were depleted because of soil erosion and weathering during long-term agricultural practice, resulting in enrichment in sediments of the lake. The depletion of elements in soils was predominately due to the fact that these metals existed in the soil with a large mobile fraction. While As, Th and Se moderately accumulated in the soils due to their insusceptibility of transport potentials. Because Sr and U are easier to be leached from soils than Rb and Th, respectively, a greater ratio of Rb/Sr and a lower value of U/Th was found in soil samples than in sediments. Therefore, our study implied agricultural soils could play an important role in the input of heavy metals into sediments via soil erosion and weathering.


INTRODUCTION
As well known, some elements are essential to human body and play irreplaceable roles in the functioning of body organs, and all living organisms require varying amounts of metals (Lane and Morel, 2000). It will be toxic at higher concentrations, and certain metals can even lead to a wide range of toxic effects, such as carcinogenicity (Qu, et al., 2012b, Thomas, et al., 2009. Chronic exposure to those heavy metals and metalloids even at relatively low levels has adverse effects on human body health and other animals (Kamunda, et al., 2016, Qu, et al., 2012b, Tsai and Lee, 2013. Therefore, widespread attention on the heavy metals pollution and related controls strategies has been attracted around the world (Davis, et al., 2009, Sun, et al., 2013. Elemental Metals in Agricultural Soils Along Two Tributaries of Chaohu Lake: Implication on Soil Erosion Caused Enrichment in Sediments Soil contamination with heavy metals not only threats the food safety but also damages to the human body health (Breivik, et al., 2004, Chen, 1999, Cheng, 2003, Tsuchiya, 1977, Wang, 2001. However, intake of heavy metals via the soil-crop system has been considered as the predominant pathway of human exposure to these toxic elements (Luo, et al., 2011, Luo, et al., 2012, Luo, et al., 2006, Wei and Yang, 2010. Heavy metals can enter the human body through the chronic low-level intake of soil metals including inhalation of dust, direct ingestion, dermal contact, and consumption of infected animals, farm products which take up heavy metals by absorbing them from contaminated soils or from atmospheric deposition by polluted air (Qu, et al., 2012a, Tsai andLee, 2013). The rapid urban and industrial development and increasing reliance on agrochemicals over the past decades have been verified as a vital factor on the accumulation of heavy metals in agricultural soils (Chen, et al., 2015, Wong, 2002, Wu, et al., 2016. However, the metals can be removed from soil by crop adsorption, leaching soil erosion and weathering (Quinton and Catt, 2007, Salt, et al., 1995, Sun, et al., 2001. The transport potential of each element varied largely depending on the speciation in soil (Fanfani, et al., 1997).
Soil erosion and weathering are the important environmental processes which can transport heavy metals from agricultural soils into the sediment of receiving water system (Jiao, et al., 2014, Quinton and Catt, 2007, Tang, et al., 2010. However, this issue has not been well documented for Chaohu Lake watershed (Tang, et al., 2010). Therefore, the objectives of this study are to (1) investigate the accumulation and spatial variability of 30 metal elements in the rice soils along two main tributaries of Chaohu Lake; (2) evaluate the accumulation status by pollution indices, (3) explore the potential impacts of metals in agricultural soils on sediments of Chaohu Lake.

Study area and sample collection
The Chaohu Lake which is located in the middle of Anhui province, between Yangtze River and Huaihe River watershed, with longitude from E117°16′54″ to E117°51′46″ and latitude from N31°43′28″ to N31°25′28″ (Zhang, 2008). It is one of five largest fresh water lakes in China, and is surrounded by Chaohu, Hefei, Lujiang, Feidong, and Feixi cities. This study was conducted around two main tributaries (i.e. Tongyang River and Hangbu River, which situated at Northeast and Southwest of Chaohu Lake, respectively). Both areas are characterized by long-term intensive cultivation, and rice is the predominant crop. These both rivers have been found with significant impacts by the non-point source inputs from agricultural activities (Wang, et al., 2016b). In the present study, a total of twenty-two surface soil samples were collected in the rice fields. Ten samples were corrected along Tongyang River (defined as N1-N10) and the remaining twelve soils were obtained along Hangbu River (referred as S1-S12) using a plastic sampler in August, 2015 ( Fig. 1). Surface soil samples were almost completely comprised by silt. A global positioning system was used to determine the geographical coordinate of all the sampling sites. After sampling, all soil samples were placed in polyethylene zip bags, and immediately transported into the laboratory. Before storage in a refrigerator until laboratory analysis, the samples were subsequently treated by air drying, manual homogenization, trituration in a ceramic mortar to pass through a 100 mesh stainless steel sieve.

Analytic methods
The analytic methods of the surface soil can be found in the previous study (Liu and Shen, 2014) and [GB/T 17138-1997, HJ 491-2009; GB/T 17141-1997]. In the present study, accurately weighted freeze-dried soil samples (0.2000±0.0005g) were digested by a mixture of acids including 10 mL HCl, 5 mL HF 5 mL HClO 4 and 3 mL HNO 3 in Teflon beakers until the solid residue disappeared and the solution turned into white or light yellow-greenish pasta-like material (Zhang, et al., 2009). After digestion, the samples were then diluted to 50 mL with ultrapure water. The concentrations of 30 metal elements including Cd, Se, Sb, Hf, U, Dy, Gd, Sc, Sn, Pr, Cs, Th, Co, Nb, Ga, Cu, Ni, La, Pb, Li, Ce, As, Sr, Cr, Zr, V, Zn, Mn, Ba, and Ti in the final solutions were measured by an Agilent 7700X inductively coupled plasma-mass spectrometer (ICP-MS), and analytic data were quantified with an internal standards method (1 mg/L Rh and In as internal standards).
A quality control procedure was also applied throughout the diff erent steps from sampling preparation of analysis. The analytic precision, determined by using two sets of parallel samples while three replicate comprises one set, both two sets were randomly selected from those samples. Furthermore, blanks and standards were regularly run with each series of analyses in order to check the accuracy of data. For every twelve samples, a laboratory blank and a standard was incorporated into the analytic procedure (Zan, et al., 2011).

Statistical analysis
All element concentrations were reported as ug/g dry weight. Normal distribution of individual element was checked using Kolmogorov-Smirnov test (with a signifi cance level at 0.05). A principal component analysis (PCA) was performed in SPSS software package (SPSS 16.0) to assess the possible origins of metals. All statistical analyses were functioned using the SPSS 20.0 software (Chicago, IL, USA). To obtain the pollution degree of elements in agricultural area in Chaohu Lake watershed, enrichment factor (EF) and pollution level index(PLI) were calculated. EF values are used to assess the degree of enrichment of a given element from anthropogenic activities: (4) where: C i (soil) and C i (background) -the concentration and background value of i th element; CR(soil) and CR(background) -the concentrations in the soil samples and background value of the reference element.
In the present study, The soil background values for all selected elements in Anhui Province  were adopted in the present study. And Ti was selected as the reference element because of the low residual concentration in the soil samples (3890 ± 478 µg/g for soils in Hangbu River Valley, 4070 ± 313 µg/g for soils in Tongyang River Valley, respectively) and which was closed to the background concentration (3800 µg/g). EF i value less than 1.0 indicates no enrichment; 1 < EF < 3 means minor enrichment, 3 < EF < 5 means moderate enrichment, 5 < EF < 10 means moderately severe enrichment, 10 < EF < 25 means severe enrichment; 25 < EF < 50 means very severe enrichment and EF < 50 means extremely severe enrichment, respectively (Lu, et al., 2014, Tasdemir and Kural, 2005, Zhang and Liu, 2002. PLI value was developed to assess the overall pollution level by all observed metal element for a certain given sample: (2) When PLI leass than 1.0 means no metal pollution, but it greater than 1.0 do suggests contamination (Varol, 2011).
The coeffi cient of variation (CV) is usually used to understand the degree of variability for the concentrations of metal in the soil. If CV less than 20% generally suggests low variability, 20%<CV≤50% means moderate variability, while higher than 50% indicates high variability (Karimi Nezhad, et al., 2015), In the present study, only Sc was characterized by high variability, elements of Ti, Rb, Zn V, Zr, Cr, Pb, Ga, and Nb were showed low variability (Fig. 4), and the remaining elements were moderate variability. These results suggested Ti, Rb, Zn V, Zr, Cr, Pb, Ga, and Nb were less infl uenced by human activities or were impacted by intensive weathering processes (Jiao, et al., 2014).

Evaluation of enrichment
EF of individual element in soils was calculated to distinguish the metals derived from anthropogenic activities and those from natural origin.    in the study areas. The EF values of Pb, Li and Ba in the soils along Hangbu River were significantly greater than those along Tongyang River (p<0.05), suggesting that Pb, Li and Ba were signifi cantly enriched in soils along Hangbu River. PLI value was calculated to assess multiple metals pollution. PLI values for all sampling sites varied 0.405 to 1.08 with a mean of 0.78 ± 0.15 (Fig. 5). Two sampling sites (site 7 and 9 in Hangbu River valley) had PLI greater than 1.0. The distribution of PLI value was tested using Monte Carlo simulations (10000 times). Results showed PLI value followed a normal distribution. The probability of PLI higher than 1.0 could be estimated to predict the contamination of the study areas based upon the probabilistic density function. Results showed that 8.2% of sampling areas had PLI greater than 1.0. Thus, most surface soils (representing 92% of study areas) were not signifi cantly contaminated by mixture of elements, approximate 8.2% of the areas was predicted to be polluted by the mixture of metals.

Source evaluation
PCA was performed to reduce the high dimension of the variable space, and further to identify the potential sources of elements in the present study. Totally, fi ve factors were obtained (Table 2), which could explain a total variance of 88.6%. Of them, only the fi fth factor (F5) appears to be associated with natural source, because this factor was dominated by Hf, U, and Zr, whose levels were lower than the background values (Fig. 2). First factor (F1) which explained 46.5% of total variance was highly loaded by Se, Sc, Dy, Gd, Pr, Th, La and Ce. Of them, Th was highly enriched in the soils of the study areas, indicating anthropogenic origins. Sc was considered as a marker of agroindustry such as sugar and alcohol production (Fernandes, 1993). As a member of lanthanide family, Ga is usually found the coexistence with Sc in the environment (Py, et al., 2011). The second factor (F2) which might suggest fertilizer-associated source (Quinton and Catt, 2007). Some phosphate fertilizer has potentially toxic As, Cd, Cr Hg, Ni, and V (Mortvedt, 1996). Both sampling areas were located in agricultural fi elds, industrial and mining sources of Ni, Li, Cu, and Cr could be ignored. For the third factor which was characterized by high loading of Cu, Pb and As, it might imply poultry farming related sources (Cang, et al., 2004). The forth factor (F4) was highly loaded by Sb, Sr and Ba. Sb can be used to trace the emission of coal burning (Cao, et al., 2014, Tang, et al., 2017, and Ba is widely used in manufactured materials such as tiles, brick and glass (McBride, et al., 2014). In the study areas, little manufactures of tiles, brick and glass were generally operated in the villages or towns which were surrounded by agricultural land. Therefore, this factor possibly implied the infl uence from little manufacture in rural areas.

A comparison between soil and sediment: implication on soil erosion and weathering
Elements in arable soils are generally associated with organic matter or other charged material (Tipping, et al., 2003). Thus soil weathering and  erosion are frequently expected as the important pathways to transport metals from soil to receiving water (Quinton and Catt, 2007). Sediment will be a key sink of these pollutants derived from soil particles. Under the supposition that chemical components in a given sediment sample have a strong correspondence to the chemical composition of donor soil. A comparison of elements in soils from Chaohu Lake and in sediments in the lake was made (Fig. 6). Results clearly showed that As, Cu, Cr and Ba tended to accumulate in soil rather than in sediment. The mobility of these elements is predominantly responsible for the differences on the accumulation potential in soil and sediment. Despite chemical speciation of metals in soils of Chaohu Lake watershed is scarce, Cu, Cr and As in sediment of Chaohu Lake mainly existed in residual fraction Li, 2007, Kong, et al., 2015).
Elements in residual fractions are insusceptible to be lost during chemical weathering or soil erosion, resulting in high accumulation in soils. However, many elements (including Cd, U, Se, Hf. Sn, Th, Zr, Mn, Sr, Ga, Nb) appeared higher accumulation in sediment than in soil. Especially for Cd, it was undetectable in all soils in the present study. Similar study also demonstrated that low concentration of Cd was also observed in soil samples from Chaohu Lake watershed (Zhang, 2007, Zhong, et al., 2015. For example the average value of Cd was 0.226 ± 0.245 µg/g with a range of 0.045-1.478 µg/g for Daxing town, and was 0.145 ±0.075 µg/g with a range of 0.054-0.432 µg/g for Yicheng Town (Zhang, 2007). Zhong et al. reported a mean concentration of Cd was 0.44 ± 0.05 µg/g for the soil samples collected in Chaohu city (Zhong, et al., 2015). Thus soils in Chaohu Lake Fig. 6. Comparison of metal concentrations in the soils from Chaohu Lake watershed to sediments in Chaohu Lake. The data of Cd in soils was obtained from the literature (Zhang, 2007), and the concentrations of metals in sediments of Chaohu Lake were obtained from the literatures (Wang, et al., 2016a, Zhang, et al., 2007 watershed contained Cd content much lower than the lacustrine sediments did (Tang, et al., 2010, Tang, et al., 2014, Wang, et al., 2016a The depletion of metal element in soil was predominantly caused by soil erosion or weathering due to long-term agricultural cultivation (Jiao, et al., 2014). A previous study on the metal accumulation in sediments caused by soil erosion also indicated that the concentrations of selected elements were much lower in soils than in sediment (Quinton and Catt, 2007), which in part supported our statement on the accumulation of elements in the sediment due to the soil erosion. In order to further verify that soil erosion and weathering were the vital reasons on the metal depletion in soil and accumulation in sediment in Chaohu Lake watershed, the ratio of Rb to Sr was used because it has been widespreadly proposed to refl ect the degree of pedogenesis and weathering (Jin, et al., 2006, Xu, et al., 2010, Zeng, et al., 2012. Generally Rb coexists with K but Sr tends to be dispersed in Ca -bearing minerals. Weathering can leach Sr much easier than Rb, as a result, Rb tends to be enriched in the parent soil comparing with Sr. Due to the diff erent geochemical behaviors Rb and Sr are easily fractionated, which leads to a diff erent ratio of Rb/Sr. A higher degree of weathering usually is characterized by a greater ratio of Rb/Sr, while a lower degree of weathering is commonly related to a lower ratio value. Furthermore, a low ratio value of Rb/Sr will often be observed in weathering products compared with parent materials. When weathering materials are eventually transported into lake, resulting in low Rb/Sr ratio for sediments. Therefore, the ratio of Rb/Sr has been considered as a useful tool to identify the intensity of soil weathering and erosion. The ratio of Rb/Sr for soil samples varied between 0.78 to 3.5 with a mean of 1.8 ± 0.73 (Fig. 7a), which is signifi cantly greater than that in sediment of Chaohu Lake (between 0.16 to 0.85 with a mean of 0.38 ± 0.23). Furthermore, Rb/Sr ratio in appeared signifi cant correlation against 1/ Sr for soils and sediments in Chaohu Lake watershed (p<0.01). Thus these results indicated sediments in Chaohu Lake were signifi cantly impacted by soil erosion and weathering. Previous studies also demonstrated that Cd was largely associated with mobile fractions of sediments in Chaohu Lake Li, 2007, Xu, et al., 2008), possibly implying Cd was easily depleted in soils and accumulated in sediments duo to intensive soil erosion and weathering after long-term agricultural practices. Non-residual Sr and Cd are easily to be lost during soil erosion and weathering, thus the residual concentrations of both elements should appear similar trends in soils and sediments. Sr and Cd in sediments of Chaohu Lake displayed a signifi cant correlation (Fig. 7b), but the relationship for soils was not investigated because Cd was under detectable in all samples in the present study. Therefore, these results suggested that soil erosion and weathering could be the key reasons on undetectable Cd in soils of study areas.
Rati o of U/Th is proposed to estimate the anthropogenic contribution of U in the agricultural soils (Takeda, et al., 2004). Generally, higher ratio value of U/Th was found in the agricultural soils, but lower value was obtained in the non-agricultural soils. In the present, the ratio of U/Th varied from 0.18 to 0.46 with a mean of 0.27± 0.07 for all soil samples, which was signifi cantly lower than those found in lacustrine sediments of Chaohu Lake (varying from 0.34 to 0.95 with a mean of 0.68 ± 0.12). These results clearly suggested that soils in Chaohu Lake watershed were less impacted by phosphatic fertilizers, while sediments of Chaohu Lake were strongly infl uenced by agrochemicals. This phenomenon could be explained by the fact that U was easier to be leached from soil than Th. As a result, U was depleted in the soils but was enriched in the sediments duo to long-term soil erosion and weathering. Correspondingly, the ratio of U/Th was reduced in soils, but was elevated in sediments.

CONCLUSIONS
Intensive cultivation has been considered as an important factor that causes heavy metal pollution in agroecological system. These heavy metals could be depleted by soil erosion and weathering, resulting in the accumulation in the sediment of receiving aquatic system. In the present study, surface soil samples in the typical agricultural areas along Hangbu River and Tongyang River, the main infl owing tributaries of Chaohu Lake, were collected to investigate 30 metal elements. The results were further compared with previously published data on the heavy metals in surface sediments of Chaohu Lake. For those elements (Cd, Sr, Hf, Sc, U etc.) with large portion of non-residual fractionation, they were found to be depleted in the soils but be accumulated in the sediments. As a result, the ratio of Rb/Sr was signifi cantly higher in the soil than in sediment, but the ratio of U/Th appeared an opposite trend. As, Se and Th moderately accumulated in the soils due to their weak mobility. These results clearly suggested that high degree of soil erosion and weathering could lead to the loss of heavy metals in the agricultural soils but accumulation in the sediments. Thus, negligible of the ecological risk caused by most of the selected heavy metals in the agricultural soils in Chaohu Lake watershed could be predicted.