Environ Sci Pollut Res DOI 10.1007/s11356-015-4255-7
RESEARCH ARTICLE
Modeling the transfer of arsenic from soil to carrot (Daucus carota L.)—a greenhouse and field-based study Changfeng Ding & Fen Zhou & Xiaogang Li & Taolin Zhang & Xingxiang Wang
Received: 12 November 2014 / Accepted: 18 February 2015 # Springer-Verlag Berlin Heidelberg 2015
Abstract Reliable empirical models describing arsenic (As) transfer in soil-plant systems are needed to estimate the human As burden from dietary intake. A greenhouse experiment was conducted in parallel with a field trial located at three sites through China to develop and validate soil-plant transfer models to predict As concentrations in carrot (Daucus carota L.). Stepwise multiple linear regression relationships were based on soil properties and the pseudo total (aqua regia) or available (0.5 M NaHCO3) soil As fractions. Carrot As contents were best predicted by the pseudo total soil As concentrations in combination with soil pH and Fe oxide, with the percentage of variation explained being up to 70 %. The constructed prediction model was further validated and improved to avoid overprotection using data from the field trial. The final obtained model is of great practical relevance to the prediction of As uptake under field conditions.
Keywords Arsenic . Carrot . Soil-plant transfer . Controlling factors . Prediction models . Field validation
Responsible editor: Michael Matthies Electronic supplementary material The online version of this article (doi:10.1007/s11356-015-4255-7) contains supplementary material, which is available to authorized users. C. Ding : F. Zhou : X. Li : T. Zhang : X. Wang (*) Key Laboratory of Soil Environment and Pollution Remediation, Institute of Soil Science, Chinese Academy of Sciences, Nanjing 210008, China e-mail: [email protected] X. Wang Jiangxi Key Laboratory of Ecological Research of Red Soil, Ecological Experimental Station of Red Soil, Chinese Academy of Sciences, Yingtan 335211, China
Introduction The metalloid arsenic (As) was ranked as the most hazardous substance on the 2013 Substance Priority List by the Agency for Toxic Substances and Disease Registry of the USA based on the frequency of occurrence, toxicity, and the potential for human exposure (ATSDR 2013). The transfer of As from soils to the edible plant parts is a key step in the route of As entry into the human food chain. Phytoaccumulation of As depends not only on the plant species but also on the pseudo total As concentration and its availability in the soil (Rosas-Castor et al. 2014). Previous studies have investigated the soil-plant transfer characteristics of As. However, soil physicochemical properties such as pH, texture, redox status, contents of phosphorus, Fe–Al–Mn oxides, and organic matter could dramatically modify As phytoavailability (Huang et al. 2006; Jiang et al. 2014; Seyfferth et al. 2014; Williams et al. 2011). Hence, it is not surprising that measurement of pseudo total As concentrations in soil alone delivers no satisfactory prediction of plant As concentrations (Fitz and Wenzel 2006). Empirical models describing the transfer of cadmium from soil to plants have been extensively studied (Adams et al. 2004; Brus et al. 2009; Römkens et al. 2009), but models for uptake into food crops from soils for As are somewhat rarely reported (Legind and Trapp 2010; McGrath and Zhao 2013). Ideally, soil-plant transfer models should be developed using field soils having a wide range of soil properties together with a range of heavy metal loadings sourced from historic contamination. However, this is seldom the case in practice (McLaughlin et al. 2010). Therefore, the approach of metal salts added to soils is used to overcome this problem (Smolders et al. 2009). It has been well recognized that the amount of metal accumulated in plants grown under controlled pot conditions was relatively higher than that under field conditions, probably because plants explore potted soil more
Environ Sci Pollut Res
extensively (Brunetti et al. 2011). Therefore, it is necessary to validate and improve the derived soil-plant transfer models in field soils to avoid overprotection. Since As is generally retained in plant roots, there is some concern that As concentrations in the edible root may pose a potential risk in root vegetables (Chou et al. 2014; Moreno-Jiménez et al. 2012). Carrot (Daucus carota L.) is a widely cultivated and consumed root vegetable all over the world and is regarded as very tolerant of As toxicity (Adriano 2001). It is possible that As could accumulate in the edible part of the carrot without displaying toxicity symptoms. Therefore, the aims of this study were to (1) investigate the soilcarrot transfer characteristics of As using a range of soils, (2) establish the transfer models based on pseudo total (aqua regia) and available (0.5 M NaHCO3) soil As concentrations while taking into account soil properties, and finally, (3) validate and improve the constructed model by applying it to predict the uptake data from field trials.
Fig. 1 Map of the soil sampling sites and the three field experimental sites in China
Materials and methods Greenhouse study A total of 21 soils were sampled throughout China (Fig. 1) for greenhouse study. The details of sampling strategy can be referred to Ding et al. (2014). The main physicochemical properties of the soils are shown in Table 1. The greenhouse pot experiment was performed in Nanjing, Jiangsu Province, China. Arsenic was added to air-dried soil samples (7 kg) in the form of arsenate (Na3AsO4·12H2O), the predominant form of As in aerobic soils (Moreno-Jiménez et al. 2012), on May 25, 2011. The As limits for soils used for vegetable production are 40, 30, and 25 mg kg−1 for pH 7.5, respectively, according to the soil environmental quality standard of China (State Environmental Protection Administration of China 1995). Therefore, three As dosages were used: the control (CK, no As added to soil), low-As (As1, 15–30 mg kg−1), and high-As (As2, 30– 60 mg kg−1) (Supplementary Materials, Table S1). The soils
Environ Sci Pollut Res Table 1
Selected physicochemical properties of soils used in this study
Soil Location
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21
Coordinates
Guiyang, 26° 26′ N, Guizhou 106° 31′ E Nanning, 22° 36′ N, Guangxi 108° 21′ E Yingtan, Jiangxi 28° 12′ N, 116° 56′ E Chongqing 29° 49′ N, 106° 24′ E Shenyang, 41° 30′ N, Liaoning 123° 28′ E Daye, Hubei 30° 05′ N, 114° 59′ E Nanjing, Jiangsu 32° 06′ N, 119° 00′ E Qiyang, Hunan 28° 16′ N, 112° 42′ E Gongzhuling, 43° 37′ N, Jilin 124° 54′ E Haikou, Hainan 19° 58′ N, 110° 15′ E Tianjin 39° 41′ N, 117° 25′ E Lhasa, Xizang 29° 40′ N, 91° 06′ E Fuzhou, Fujian 26° 12′ N, 118° 51′ E 43° 54′ N, Gongzhuling, Jilin 124° 59′ E Shuangliao, Jilin 43° 25′ N, 123° 31′ E Suzhou, Jiangsu 31° 19′ N, 120° 28′ E Shijiazhuang, 38° 23′ N, Hebei 114° 42′ E Hohhot, Inner 41° 19′ N, Mongolia 109° 56′ E Lanzhou, Gansu 35° 52′ N, 104° 14′ E Xi’an, Shaanxi 36° 35′ N, 109° 35′ E Urumqi, 44° 30′ N, 87° Xinjiang 45′ E
OC (g kg−1)
CEC Clay (%) (cmol kg−1)
FeOX (g kg−1)
AlOX (g kg−1)
Background As (mg kg−1)
Soil type
pH
Argosols
4.67±0.02 20.6±0.18 15.4±0.14
55.8±0.54 86.2±7.34 195±3.02 37.3±1.04
Ferrosols
4.81±0.03 14.6±0.45 7.63±0.07
36.7±0.06 54.7±8.59 115±0.80 21.5±1.01
Ferrosols
4.84±0.01 5.43±0.32 9.31±0.07
45.8±0.14 50.2±7.80 131±4.47 15.0±1.35
Cambosols
4.99±0.02 9.92±0.02 16.9±0.28
20.2±0.94 43.8±0.89 138±0.70 3.50±0.08
Argosols
5.35±0.08 8.81±0.11 15.8±1.11
22.4±0.17 37.9±0.73 132±9.99 9.90±0.07
Ferrosols
5.68±0.08 10.1±0.49 12.3±1.04
29.6±0.03 95.7±11.1
Argosols
6.28±0.02 12.9±0.81 12.1±0.28
15.9±0.23 27.9±0.35 90.8±3.39 7.50±0.10
Anthrosols
6.31±0.05 16.5±0.05 14.0±0.21
31.3±0.60 46.9±1.72 123±5.69 17.2±1.74
Isohumosols 6.52±0.02 14.0±0.10 24.9±0.28
33.6±0.72 29.8±0.82 123±0.51 10.3±0.33
Ferralsols
6.83±0.02 6.06±0.26 4.53±0.56
17.8±0.32 45.9±0.59 87.0±1.07 5.20±0.33
Cambosols
6.93±0.11 9.90±0.42 24.1±0.56
36.5±0.92 40.2±2.45 139±0.99 9.50±0.48
Cambosols
7.01±0.01 12.1±0.15 8.53±0.00
10.1±0.07 36.6±2.50 120±3.52 14.0±0.54
Argosols
7.12±0.01 9.32±0.14 10.2±1.11
20.9±0.03 39.7±0.44 204±2.29 2.80±0.23
Isohumosols 7.30±0.06 15.2±0.12 23.2±1.04
33.4±0.91 33.1±4.60 129±1.10 8.50±0.09
Halosols
7.88±0.06 21.6±0.09 14.4±0.49
6.35±0.73 22.2±6.26 94.9±2.46 5.10±0.17
Anthrosols
8.04±0.04 5.55±0.35 8.18±0.28
13.8±0.97 43.8±0.12 117±0.48 7.70±0.30
Argosols
8.23±0.04 8.58±0.23 9.16±0.00
9.16±2.10 32.1±0.67 122±1.90 7.90±0.07
Isohumosols 8.37±0.02 9.54±0.05 9.70±0.07
8.34±0.16 28.2±0.94 123±3.37 10.2±0.25
Aridosols
8.41±0.13 7.38±0.11 7.63±0.07
11.3±0.72 42.3±0.43 107±0.54 12.7±1.35
Primosols
8.65±0.03 5.88±0.17 6.86±0.28
8.21±0.99 34.0±0.55 107±1.10 12.0±0.91
Aridolsols
8.67±0.25 4.30±0.15 6.65±0.07
10.3±0.01 32.2±0.32 115±0.08 7.50±0.06
170±2.20 15.7±0.20
Soil numbers were sequenced in the order of increasing pH. Soil types were classified according to Chinese Soil Taxonomy (Cooperative Research Group on Chinese Soil Taxonomy 2001). Background As concentration was presented as pseudo total soil As fraction. Mean values±SD (n=3) are shown OC organic carbon, CEC cation exchange capacity (buffered), FeOX total Fe oxide, AlOX total Al oxide
were then left for aging under natural conditions for 3 months (Alexander et al. 2006). During the aging period, soil water content was maintained at 80 % of the maximum water holding capacity in the field. All of the pots were laid out randomly with three replicates. Basal fertilizers consisting of 0.15 g N (in urea), 0.05 g P (in calcium biphosphate), and 0.10 g K (in potassium sulfate) per kilogram soil were applied for each pot after the aging period. Seeds of carrot cultivar New Kuroda were purchased from Nanjing Qiutian Seed Company and were sown to each
pot on August 24, 2011. The number of seedlings was thinned to three per pot after the emergence. The soils were watered to maintain normal growth of carrots, and the accumulation of excess water at the bottom of the pots was avoided. Field study The field study was conducted at three sites with different climate conditions through China in the year 2012. They were located in Yingtan, Jiangxi Province; Nanjing, Jiangsu
Environ Sci Pollut Res
Province; and Haikou, Hainan Province, China, respectively (Fig. 1). The three soil types were the same with soils 3, 7, and 10 in Table 1 as have been used in the greenhouse study. Each site was evenly divided into 1×2 m plots using PVC boards. The surface soil (0–20 cm) of each plot was transferred onto a canvas and thoroughly mixed with the same amount of Na3AsO4·12H2O solution as has been applied in the greenhouse study. Then the mixed soil was filled back into the corresponding plot. A control plot with no As addition was also included. Each plot with three replicates was arranged randomly. After aging for 3 months, carrot plants (cv. New Kuroda) were sown and grown under local farming management style. Soil and plant analysis After the aging period, soil samples were collected, air-dried, ground, and digested with aqua regia for pseudo total As analysis (Lu 2000). Approximately 4 months after sowing, the edible part of the carrot was harvested, and the soil was sampled for available As determination. The details of washing and digestion method of the carrots were similar to that of Hg in our previous study (Ding et al. 2014). The available As fraction was extracted with 0.5 M NaHCO3 (Su et al. 2014). The As concentration in soils and carrots was measured by atomic fluorescence spectrometry (AFS-610D2, Rayleigh, Beijing, China). To ensure the precision of the analytical procedure, carrot and soil certified reference materials (GBW10047 and GBW07450, respectively, National Research Center for Certified Reference Materials, China) were used, and the recoveries were 89~103 % and 95~105 %, respectively. Data analysis The bioconcentration factor (BCF) (Algreen et al. 2014) is used to assess the transfer of As from soil to carrot: BC F ¼ Ascarrot =Assoil
ð1Þ
where Ascarrot and Assoil are As concentrations in carrots and soils, respectively. A path analysis (PA) model (Richards et al. 2012) was applied to examine the causal path of soil properties to the carrot As concentrations. Stepwise multiple linear regression (SMLR) analysis was used to construct empirical models capable of predicting carrot As concentrations. The stepping criteria employed for entry and removal was based on the significance level of 0.05. The quality of the models was evaluated on the basis of the coefficient of determination (R2) and the root mean square error (RMSE) value.
Analyses of the data were performed with the SPSS 18.0 and SigmaPlot 11.0. All data were log-transformed (except for soil pH) prior to analysis to obtain normality.
Results Soil-to-carrot transfer characteristics of As According to the Chinese Food Safety Standard for Contaminants in Foods (GB 2762–2012), the maximum level (ML) for total As in fresh vegetables is 0.5 mg kg−1 (Ministry of Health of China 2013). To compare our results to the ML, Fig. 2a shows the total As concentrations in carrots expressed on a fresh weight basis with measured water content of 87 %. Under low and high As application treatments, the As concentrations in carrots ranged from 0.003 to 0.18 mg kg−1 (average 0.04 mg kg−1) and 0.005 to 0.33 mg kg−1 (average 0.07 mg kg−1) for the 21 soils, respectively, with none of them exceeding the ML. As shown in Fig. 2b, the BCF with As additions generally increased with soil pH. The maximum of BCF was observed in the most alkaline soil 21 (pH 8.67) under both As addition treatments, while the minimum was found in acidic soils 1 (pH 4.67) and 3 (pH 4.84) under low and high As addition treatments, respectively. A s i g n if ic a n t p o s it i v e c o r r e l a t i o n (R 2 = 0 . 3 9 , P < 0.0001) was observed between the log-transformed As concentrations in carrot and the pseudo total soil As fraction when combining the data of the three As treatments (63 observations) (Fig. 3a). A much better correlation (R2 =0.65, P
RESEARCH ARTICLE
Modeling the transfer of arsenic from soil to carrot (Daucus carota L.)—a greenhouse and field-based study Changfeng Ding & Fen Zhou & Xiaogang Li & Taolin Zhang & Xingxiang Wang
Received: 12 November 2014 / Accepted: 18 February 2015 # Springer-Verlag Berlin Heidelberg 2015
Abstract Reliable empirical models describing arsenic (As) transfer in soil-plant systems are needed to estimate the human As burden from dietary intake. A greenhouse experiment was conducted in parallel with a field trial located at three sites through China to develop and validate soil-plant transfer models to predict As concentrations in carrot (Daucus carota L.). Stepwise multiple linear regression relationships were based on soil properties and the pseudo total (aqua regia) or available (0.5 M NaHCO3) soil As fractions. Carrot As contents were best predicted by the pseudo total soil As concentrations in combination with soil pH and Fe oxide, with the percentage of variation explained being up to 70 %. The constructed prediction model was further validated and improved to avoid overprotection using data from the field trial. The final obtained model is of great practical relevance to the prediction of As uptake under field conditions.
Keywords Arsenic . Carrot . Soil-plant transfer . Controlling factors . Prediction models . Field validation
Responsible editor: Michael Matthies Electronic supplementary material The online version of this article (doi:10.1007/s11356-015-4255-7) contains supplementary material, which is available to authorized users. C. Ding : F. Zhou : X. Li : T. Zhang : X. Wang (*) Key Laboratory of Soil Environment and Pollution Remediation, Institute of Soil Science, Chinese Academy of Sciences, Nanjing 210008, China e-mail: [email protected] X. Wang Jiangxi Key Laboratory of Ecological Research of Red Soil, Ecological Experimental Station of Red Soil, Chinese Academy of Sciences, Yingtan 335211, China
Introduction The metalloid arsenic (As) was ranked as the most hazardous substance on the 2013 Substance Priority List by the Agency for Toxic Substances and Disease Registry of the USA based on the frequency of occurrence, toxicity, and the potential for human exposure (ATSDR 2013). The transfer of As from soils to the edible plant parts is a key step in the route of As entry into the human food chain. Phytoaccumulation of As depends not only on the plant species but also on the pseudo total As concentration and its availability in the soil (Rosas-Castor et al. 2014). Previous studies have investigated the soil-plant transfer characteristics of As. However, soil physicochemical properties such as pH, texture, redox status, contents of phosphorus, Fe–Al–Mn oxides, and organic matter could dramatically modify As phytoavailability (Huang et al. 2006; Jiang et al. 2014; Seyfferth et al. 2014; Williams et al. 2011). Hence, it is not surprising that measurement of pseudo total As concentrations in soil alone delivers no satisfactory prediction of plant As concentrations (Fitz and Wenzel 2006). Empirical models describing the transfer of cadmium from soil to plants have been extensively studied (Adams et al. 2004; Brus et al. 2009; Römkens et al. 2009), but models for uptake into food crops from soils for As are somewhat rarely reported (Legind and Trapp 2010; McGrath and Zhao 2013). Ideally, soil-plant transfer models should be developed using field soils having a wide range of soil properties together with a range of heavy metal loadings sourced from historic contamination. However, this is seldom the case in practice (McLaughlin et al. 2010). Therefore, the approach of metal salts added to soils is used to overcome this problem (Smolders et al. 2009). It has been well recognized that the amount of metal accumulated in plants grown under controlled pot conditions was relatively higher than that under field conditions, probably because plants explore potted soil more
Environ Sci Pollut Res
extensively (Brunetti et al. 2011). Therefore, it is necessary to validate and improve the derived soil-plant transfer models in field soils to avoid overprotection. Since As is generally retained in plant roots, there is some concern that As concentrations in the edible root may pose a potential risk in root vegetables (Chou et al. 2014; Moreno-Jiménez et al. 2012). Carrot (Daucus carota L.) is a widely cultivated and consumed root vegetable all over the world and is regarded as very tolerant of As toxicity (Adriano 2001). It is possible that As could accumulate in the edible part of the carrot without displaying toxicity symptoms. Therefore, the aims of this study were to (1) investigate the soilcarrot transfer characteristics of As using a range of soils, (2) establish the transfer models based on pseudo total (aqua regia) and available (0.5 M NaHCO3) soil As concentrations while taking into account soil properties, and finally, (3) validate and improve the constructed model by applying it to predict the uptake data from field trials.
Fig. 1 Map of the soil sampling sites and the three field experimental sites in China
Materials and methods Greenhouse study A total of 21 soils were sampled throughout China (Fig. 1) for greenhouse study. The details of sampling strategy can be referred to Ding et al. (2014). The main physicochemical properties of the soils are shown in Table 1. The greenhouse pot experiment was performed in Nanjing, Jiangsu Province, China. Arsenic was added to air-dried soil samples (7 kg) in the form of arsenate (Na3AsO4·12H2O), the predominant form of As in aerobic soils (Moreno-Jiménez et al. 2012), on May 25, 2011. The As limits for soils used for vegetable production are 40, 30, and 25 mg kg−1 for pH 7.5, respectively, according to the soil environmental quality standard of China (State Environmental Protection Administration of China 1995). Therefore, three As dosages were used: the control (CK, no As added to soil), low-As (As1, 15–30 mg kg−1), and high-As (As2, 30– 60 mg kg−1) (Supplementary Materials, Table S1). The soils
Environ Sci Pollut Res Table 1
Selected physicochemical properties of soils used in this study
Soil Location
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21
Coordinates
Guiyang, 26° 26′ N, Guizhou 106° 31′ E Nanning, 22° 36′ N, Guangxi 108° 21′ E Yingtan, Jiangxi 28° 12′ N, 116° 56′ E Chongqing 29° 49′ N, 106° 24′ E Shenyang, 41° 30′ N, Liaoning 123° 28′ E Daye, Hubei 30° 05′ N, 114° 59′ E Nanjing, Jiangsu 32° 06′ N, 119° 00′ E Qiyang, Hunan 28° 16′ N, 112° 42′ E Gongzhuling, 43° 37′ N, Jilin 124° 54′ E Haikou, Hainan 19° 58′ N, 110° 15′ E Tianjin 39° 41′ N, 117° 25′ E Lhasa, Xizang 29° 40′ N, 91° 06′ E Fuzhou, Fujian 26° 12′ N, 118° 51′ E 43° 54′ N, Gongzhuling, Jilin 124° 59′ E Shuangliao, Jilin 43° 25′ N, 123° 31′ E Suzhou, Jiangsu 31° 19′ N, 120° 28′ E Shijiazhuang, 38° 23′ N, Hebei 114° 42′ E Hohhot, Inner 41° 19′ N, Mongolia 109° 56′ E Lanzhou, Gansu 35° 52′ N, 104° 14′ E Xi’an, Shaanxi 36° 35′ N, 109° 35′ E Urumqi, 44° 30′ N, 87° Xinjiang 45′ E
OC (g kg−1)
CEC Clay (%) (cmol kg−1)
FeOX (g kg−1)
AlOX (g kg−1)
Background As (mg kg−1)
Soil type
pH
Argosols
4.67±0.02 20.6±0.18 15.4±0.14
55.8±0.54 86.2±7.34 195±3.02 37.3±1.04
Ferrosols
4.81±0.03 14.6±0.45 7.63±0.07
36.7±0.06 54.7±8.59 115±0.80 21.5±1.01
Ferrosols
4.84±0.01 5.43±0.32 9.31±0.07
45.8±0.14 50.2±7.80 131±4.47 15.0±1.35
Cambosols
4.99±0.02 9.92±0.02 16.9±0.28
20.2±0.94 43.8±0.89 138±0.70 3.50±0.08
Argosols
5.35±0.08 8.81±0.11 15.8±1.11
22.4±0.17 37.9±0.73 132±9.99 9.90±0.07
Ferrosols
5.68±0.08 10.1±0.49 12.3±1.04
29.6±0.03 95.7±11.1
Argosols
6.28±0.02 12.9±0.81 12.1±0.28
15.9±0.23 27.9±0.35 90.8±3.39 7.50±0.10
Anthrosols
6.31±0.05 16.5±0.05 14.0±0.21
31.3±0.60 46.9±1.72 123±5.69 17.2±1.74
Isohumosols 6.52±0.02 14.0±0.10 24.9±0.28
33.6±0.72 29.8±0.82 123±0.51 10.3±0.33
Ferralsols
6.83±0.02 6.06±0.26 4.53±0.56
17.8±0.32 45.9±0.59 87.0±1.07 5.20±0.33
Cambosols
6.93±0.11 9.90±0.42 24.1±0.56
36.5±0.92 40.2±2.45 139±0.99 9.50±0.48
Cambosols
7.01±0.01 12.1±0.15 8.53±0.00
10.1±0.07 36.6±2.50 120±3.52 14.0±0.54
Argosols
7.12±0.01 9.32±0.14 10.2±1.11
20.9±0.03 39.7±0.44 204±2.29 2.80±0.23
Isohumosols 7.30±0.06 15.2±0.12 23.2±1.04
33.4±0.91 33.1±4.60 129±1.10 8.50±0.09
Halosols
7.88±0.06 21.6±0.09 14.4±0.49
6.35±0.73 22.2±6.26 94.9±2.46 5.10±0.17
Anthrosols
8.04±0.04 5.55±0.35 8.18±0.28
13.8±0.97 43.8±0.12 117±0.48 7.70±0.30
Argosols
8.23±0.04 8.58±0.23 9.16±0.00
9.16±2.10 32.1±0.67 122±1.90 7.90±0.07
Isohumosols 8.37±0.02 9.54±0.05 9.70±0.07
8.34±0.16 28.2±0.94 123±3.37 10.2±0.25
Aridosols
8.41±0.13 7.38±0.11 7.63±0.07
11.3±0.72 42.3±0.43 107±0.54 12.7±1.35
Primosols
8.65±0.03 5.88±0.17 6.86±0.28
8.21±0.99 34.0±0.55 107±1.10 12.0±0.91
Aridolsols
8.67±0.25 4.30±0.15 6.65±0.07
10.3±0.01 32.2±0.32 115±0.08 7.50±0.06
170±2.20 15.7±0.20
Soil numbers were sequenced in the order of increasing pH. Soil types were classified according to Chinese Soil Taxonomy (Cooperative Research Group on Chinese Soil Taxonomy 2001). Background As concentration was presented as pseudo total soil As fraction. Mean values±SD (n=3) are shown OC organic carbon, CEC cation exchange capacity (buffered), FeOX total Fe oxide, AlOX total Al oxide
were then left for aging under natural conditions for 3 months (Alexander et al. 2006). During the aging period, soil water content was maintained at 80 % of the maximum water holding capacity in the field. All of the pots were laid out randomly with three replicates. Basal fertilizers consisting of 0.15 g N (in urea), 0.05 g P (in calcium biphosphate), and 0.10 g K (in potassium sulfate) per kilogram soil were applied for each pot after the aging period. Seeds of carrot cultivar New Kuroda were purchased from Nanjing Qiutian Seed Company and were sown to each
pot on August 24, 2011. The number of seedlings was thinned to three per pot after the emergence. The soils were watered to maintain normal growth of carrots, and the accumulation of excess water at the bottom of the pots was avoided. Field study The field study was conducted at three sites with different climate conditions through China in the year 2012. They were located in Yingtan, Jiangxi Province; Nanjing, Jiangsu
Environ Sci Pollut Res
Province; and Haikou, Hainan Province, China, respectively (Fig. 1). The three soil types were the same with soils 3, 7, and 10 in Table 1 as have been used in the greenhouse study. Each site was evenly divided into 1×2 m plots using PVC boards. The surface soil (0–20 cm) of each plot was transferred onto a canvas and thoroughly mixed with the same amount of Na3AsO4·12H2O solution as has been applied in the greenhouse study. Then the mixed soil was filled back into the corresponding plot. A control plot with no As addition was also included. Each plot with three replicates was arranged randomly. After aging for 3 months, carrot plants (cv. New Kuroda) were sown and grown under local farming management style. Soil and plant analysis After the aging period, soil samples were collected, air-dried, ground, and digested with aqua regia for pseudo total As analysis (Lu 2000). Approximately 4 months after sowing, the edible part of the carrot was harvested, and the soil was sampled for available As determination. The details of washing and digestion method of the carrots were similar to that of Hg in our previous study (Ding et al. 2014). The available As fraction was extracted with 0.5 M NaHCO3 (Su et al. 2014). The As concentration in soils and carrots was measured by atomic fluorescence spectrometry (AFS-610D2, Rayleigh, Beijing, China). To ensure the precision of the analytical procedure, carrot and soil certified reference materials (GBW10047 and GBW07450, respectively, National Research Center for Certified Reference Materials, China) were used, and the recoveries were 89~103 % and 95~105 %, respectively. Data analysis The bioconcentration factor (BCF) (Algreen et al. 2014) is used to assess the transfer of As from soil to carrot: BC F ¼ Ascarrot =Assoil
ð1Þ
where Ascarrot and Assoil are As concentrations in carrots and soils, respectively. A path analysis (PA) model (Richards et al. 2012) was applied to examine the causal path of soil properties to the carrot As concentrations. Stepwise multiple linear regression (SMLR) analysis was used to construct empirical models capable of predicting carrot As concentrations. The stepping criteria employed for entry and removal was based on the significance level of 0.05. The quality of the models was evaluated on the basis of the coefficient of determination (R2) and the root mean square error (RMSE) value.
Analyses of the data were performed with the SPSS 18.0 and SigmaPlot 11.0. All data were log-transformed (except for soil pH) prior to analysis to obtain normality.
Results Soil-to-carrot transfer characteristics of As According to the Chinese Food Safety Standard for Contaminants in Foods (GB 2762–2012), the maximum level (ML) for total As in fresh vegetables is 0.5 mg kg−1 (Ministry of Health of China 2013). To compare our results to the ML, Fig. 2a shows the total As concentrations in carrots expressed on a fresh weight basis with measured water content of 87 %. Under low and high As application treatments, the As concentrations in carrots ranged from 0.003 to 0.18 mg kg−1 (average 0.04 mg kg−1) and 0.005 to 0.33 mg kg−1 (average 0.07 mg kg−1) for the 21 soils, respectively, with none of them exceeding the ML. As shown in Fig. 2b, the BCF with As additions generally increased with soil pH. The maximum of BCF was observed in the most alkaline soil 21 (pH 8.67) under both As addition treatments, while the minimum was found in acidic soils 1 (pH 4.67) and 3 (pH 4.84) under low and high As addition treatments, respectively. A s i g n if ic a n t p o s it i v e c o r r e l a t i o n (R 2 = 0 . 3 9 , P < 0.0001) was observed between the log-transformed As concentrations in carrot and the pseudo total soil As fraction when combining the data of the three As treatments (63 observations) (Fig. 3a). A much better correlation (R2 =0.65, P
