JEQ Grow Your Career With ASA
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Related articles in JEQ
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via ISI Web of Science (21)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Adams, M. L.
Right arrow Articles by Chambers, B. J.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Adams, M. L.
Right arrow Articles by Chambers, B. J.
Agricola
Right arrow Articles by Adams, M. L.
Right arrow Articles by Chambers, B. J.
Related Collections
Right arrow Heavy Metals
Right arrow Soil Pollution
Published in J. Environ. Qual. 33:532-541 (2004).
© ASA, CSSA, SSSA
677 S. Segoe Rd., Madison, WI 53711 USA

TECHNICAL REPORT

Heavy Metals in the Environment

Predicting Cadmium Concentrations in Wheat and Barley Grain Using Soil Properties

M. L. Adamsa, F. J. Zhao*,a, S. P. McGratha, F. A. Nicholsonb and B. J. Chambersb

a Agriculture and Environment Division, Rothamsted Research, Harpenden, Hertfordshire AL5 2JQ, UK
b ADAS Gleadthorpe Research Centre, Meden Vale, Mansfield, Nottinghamshire NG20 9PF, UK

* Corresponding author (Fangjie.Zhao{at}bbsrc.ac.uk).

Received for publication February 24, 2003.

   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
The entry of Cd into the food chain is of concern as it can cause chronic health problems. To investigate the relationship between soil properties and the concentration of Cd in wheat (Triticum aestivum L.) and barley (Hordeum vulgare L.) grain, we analyzed 162 wheat and 215 barley grain samples collected from paired soil and crop surveys in Britain, and wheat and barley samples from two long-term sewage sludge experiments. Cadmium concentrations were much lower in barley grain than in wheat grain under comparable soil conditions. Multiple regression analysis showed that soil total Cd and pH were the significant factors influencing grain Cd concentrations. Significant cultivar differences in Cd uptake were observed for both wheat and barley. Wheat grain Cd concentrations could be predicted reasonably well from soil total Cd and pH using the following model: log(grain Cd) = a + b log(soil Cd) – c(soil pH), with 53% of the variance being accounted for. The coefficients obtained from the data sets of the paired soil and crop surveys and from long-term sewage sludge experiments were similar, suggesting similar controlling factors of Cd bioavailability in sludge-amended or unamended soils. For barley, the model was less satisfactory for predicting grain Cd concentration (22% of variance accounted for). The model can be used to predict the likelihood of wheat grain Cd exceeding the new European Union (EU) foodstuff regulations on the maximum permissible concentration of Cd under different soil conditions, particularly in relation to the existing Directive and the proposed new Directive on land applications of sewage sludge.

Abbreviations: MPC, maximum permissible concentration


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
IN RECENT YEARS, there has been increasing awareness and concern over heavy metal contamination of soils and the effects this may be having on the food chain. High concentrations of heavy metals in agricultural soils can occur naturally or via atmospheric deposition or the application of metal-contaminated sewage sludges, fertilizers, and animal manures (Ryan et al., 1982; Alloway and Steinnes, 1999). With regard to human health, Cd concentrations in agricultural produce are of particular importance, as the consumption of agricultural foodstuffs is thought to contribute significantly to the dietary Cd intake. The entry of Cd into the food chain is of concern as it can cause chronic health problems in humans such as bone disease, lung edema, renal dysfunction, liver damage, anemia, and hypertension (Nordberg, 1974; Nath et al., 1984; Staessen et al., 1999). Due to this, Cd is one of a very small group of metals for which the Food and Agriculture Organization and World Health Organization (1978) have set a limit for the provisional daily intake by humans (70 µg Cd d–1). In the 1980s the U.S. adult population was reported to receive about 20% of the FAO/WHO allowable daily intake of Cd from the consumption of grain and cereal products (Wagner et al., 1984). In contrast, during the same period, grain and cereal products accounted for about 30 to 40% of the daily allowable Cd intake in the European Community (Hutton, 1982). In light of the potential health risk, the European Union has recently introduced legislation defining the maximum permissible concentrations (MPCs) of Cd in foodstuffs (European Commission, 2001). For cereals excluding wheat grain, bran, germ and rice, the MPC for Cd is 0.1 mg kg–1 fresh wt., while the limit for the aforementioned exceptions is 0.2 mg kg–1 fresh wt. Assuming a grain moisture content of 15%, which is common for cereal grain, the MPCs of Cd based on the dry weight basis are 0.118 and 0.235 mg kg–1 for barley and wheat grain, respectively. Analyses of nationally representative samples showed the Cd concentration in wheat grain in the ranges of from <0.002 to 0.21 mg kg–1 dry wt. in the USA (Wolnik et al., 1983), from 0.024 to 0.41 mg kg–1 dry wt. in the Netherlands (Wiersma et al., 1986), from 0.004 to 0.31 mg kg–1 dry wt. in the UK (Chaudri et al., 1995), and from <0.01 to 0.24 mg kg–1 dry wt. in Canada (Gawalko et al., 2001). For barley grain, the range of Cd concentration was reported to be from 0.012 to 0.64 mg kg–1 dry wt. in the Netherlands (Wiersma et al., 1986). Clearly, these national surveys indicate that a proportion of cereal grain produced in different countries would not be compliant with regard to the newly introduced EU MPCs for Cd.

Agricultural management practices that directly affect Cd concentrations and availability in the soil may influence Cd accumulation by crops. Specifically, the addition of sludge or fertilizers having high Cd concentrations to agricultural land may cause significant increases in the uptake of Cd by crops (Grant et al., 1999). A number of soil factors have been shown to affect the availability of Cd to plants. Soil pH is often the most notable factor (e.g., He and Singh, 1993a; Eriksson and Söderström, 1996; Oliver et al., 1996; Wenzel et al., 1996; Gavi et al., 1997; Hooda et al., 1997; Grant et al., 1999). Increasing pH favors the adsorption of Cd to metal binding sites and decreases the partition of Cd to soil solution (McBride et al., 1997; Sauvé et al., 2000). McBride (2002) re-analyzed data on Cd uptake by lettuce (Lactuca sativa L.), Swiss chard (Beta vulgaris L.), and corn (Zea mays L.) leaves from sewage sludge–amended soils, and concluded that a combination of soil pH and total soil Cd was a reasonable predictor of plant tissue Cd concentrations. Other soil properties that can influence Cd availability include the contents of soil organic matter and Fe and Mn oxides (He and Singh, 1993b; Wenzel et al., 1996). The concentration of Zn in soil can affect Cd uptake by plants, presumably due to competition between these two metals for uptake and transport inside the plant (Oliver et al., 1994; Welch et al., 1999). Salinity is another factor that has been shown to influence Cd accumulation by plants (Bingham et al., 1984; Li et al., 1994; McLaughlin et al., 1994; McLaughlin et al., 1999). Chloride forms relatively stable complexes with Cd, and as a result mobilizes soil Cd, hence increasing its availability for plant uptake.

A number of previous studies have investigated the relationships between various soil characteristics and uptake of Cd in cereals under field conditions. For example, Wenzel et al. (1996) showed that Cd accumulation in wheat grain grown at seven experimental sites in Austria was significantly affected by soil chemical characteristics and by cultivar. Multiple linear regression analyses indicated that about 80% of the variation in Cd accumulation was explained by cultivar, total soil Cd, and organic carbon. Garrett et al. (1998) used a nonlinear model to predict the concentration of Cd in durum wheat grain in the Canadian Prairies. They showed that the model, including 0.01 M Na4P2O7–extractable Cd and extractable organic C, explained 74% of the variability of grain Cd (n = 34). Soil pH was not found to be a significant factor in the predictive model, mainly because 0.01 M Na4P2O7–extractable Cd was strongly pH-dependent. Gray et al. (2001) reported that soil total Zn and 0.05 M CaCl2–extractable Cd together explained 59% of the variability of grain Cd in 13 winter wheat samples in New Zealand. The three studies mentioned above included relatively small number of sites (soils), and therefore the results obtained may not represent a wider geographical region. Eriksson and Söderström (1996) included a larger number of sites in their study. They found that the Cd concentration of wheat grain grown on noncalcareous soils in southern Sweden was related positively to soil total Cd and grain yield, and negatively to soil pH, ammonium lactate–extractable P, and HNO3–extractable Zn. However, these parameters explained only about 40% of the variability of grain Cd (n = 121). For wheat grown on calcareous soils (n = 69), the relationship between grain Cd and soil properties was considerably poorer. Norvell et al. (2000) sampled 124 paired samples of soil and grain of durum wheat (Triticum turgidum L.) from a single field in North Dakota, which was affected by soil salinity. They found that DTPA-extractable Cd and water-soluble Cl in soil accounted for 66% of the variability of Cd in grain. A number of extractants have been proposed for the assessment of the phytoavailability of Cd to plants (He and Singh, 1993a; Krishnamurti et al., 1995; Mench et al., 1997; Garrett et al., 1998; Gray et al., 1999). However, none of the proposed extraction methods have gained universal acceptance, thus making the comparison between different studies difficult. In contrast, basic soil properties such as pH, soil organic matter content, and total metal concentrations are routinely measured. Combinations of these basic properties may explain Cd uptake by plants on a species or variety specific basis, as suggested by McBride (2002). Such an approach may provide an attractive alternative to the use of diverse, yet nonstandardized, extraction methods.

In the present study, we sampled paired soil and wheat or barley grain samples from a large number of field sites that encompassed a wide range of soil types and geographic and climatic variations within the UK. Our objective was to investigate the relationship between soil properties and the concentrations of Cd in wheat or barley grain, with the aim of obtaining simple algorithms that could be used to predict grain Cd concentrations from soil properties under field conditions. Because applications of sewage sludge are one of the main factors that can lead to elevated grain Cd, we also investigated the relationship between soil properties and the concentrations of Cd in wheat or barley grain in two long-term sludge experiments. This allowed us to compare the regression equations obtained from the nationwide paired soil and grain survey and from long-term sludge experiments.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Paired Crop and Soil Samples
Paired soil and crop samples from wheat and barley crops were collected across the main cereal-growing areas in Britain shortly before harvest. A total of 34, 61, and 67 paired soil and wheat samples were collected in the 1998, 1999, and 2000 harvest years, respectively. All but 11 of the wheat samples were winter wheat. For barley, 27, 95, and 93 paired samples were collected in the three years. Of the barley samples, 125 were spring barley and 90 winter barley. Sites were selected on commercial farms to represent a range of soil types. In 1999 and 2000, some sites with potentially high concentrations of soil Cd were targeted, including soils of naturally high background levels due to geochemical factors, potential contamination from past sewage sludge or industrial or other waste applications, and proximity to major roads or motorways. Information recorded at the sampling site included details of site and sampling location, soil type, history of manure and/or sewage sludge applications in last 10 yr, fertilizer inputs, crop variety, drilling and harvest dates, and proximity of the sampling location to any major roads or industrial sites.

At each site, a 1-m2 quadrat was placed at random in the field to be sampled. A 0- to 15-cm topsoil sample (approximately 1 kg in weight and comprised of 15 cores) was obtained by manual coring within the quadrat area. A crop sample was collected at the same time by hand-cutting the crop near ground level from all of the marked quadrat area. The whole-crop sample was subsequently threshed and the resulting grain sample retained for analysis.

Long-Term Sewage Sludge Experimental Sites
Long-term sewage sludge experimental sites at Gleadthorpe (Nottinghamshire) and Rosemaund (Herefordshire), England, were sampled to provide further site-specific information on the effects of soil parameters on Cd concentrations in the grain of cereal crops. At Gleadthorpe, a total of 22 sludge treatments were established in 1982 and 1986 to obtain a range of soil heavy metal concentrations. More details of this experiment were given by Bhogal et al. (2003). At Rosemaund, untreated controls and nine sludge treatments were established between 1968 and 1971 that had varying amounts of heavy metals. More details of this experiment were given by Chaudri et al. (2001) and Bhogal et al. (2003). There were two replicates per treatment at Gleadthorpe, and four replicates per treatment at Rosemaund except that the untreated control was replicated in eight plots. In both experiments, winter wheat (cv. Hereward) was grown in the 1998–1999 season and spring barley (cv. Optic) in the 1999–2000 season. Grain and soil samples were taken from 40 experimental plots at Gleadthorpe, and from 44 plots at Rosemaund in both the 1999 and 2000 harvest years. Grain yield per plot was measured by hand cutting the crop at harvest and a cleaned subsample (approximately 500 g) of the threshed grain was taken per plot to determine dry matter content and to provide a dried subsample for heavy metal analysis. Topsoil samples (0–15 cm, approximately 500 g from 10 cores per plot) were taken manually shortly after harvest.

Soil and Grain Analyses
Soil samples were air-dried and ground to pass through a 2-mm sieve. Soil pH was determined using a glass electrode in a soil to water ratio of 1:2.5. Soil organic matter was determined by loss on ignition based on the method described by Ball (1964). Soil chloride was extracted with a 1:5 soil to water ratio, and determined by ion chromatography (DX500; Dionex, Sunnyvale, CA) fitted with an IonPac AS9-SC column and AG9 guard column. Soil Al and Fe oxides and/or hydroxides were extracted using a mixture of ammonium oxalate and oxalic acids (0.114/0.086 M) following the procedure of Janssen et al. (1997), before determination by inductively coupled plasma atomic emission spectrometry (ICP–AES; Fisons ARL Accuris, Ecublens, Switzerland). Subsamples of soils were finely ground to <150 µm in an agate ball mill. Subsamples of 0.25 g were digested with aqua regia (4:1 v/v concentrated HCl to HNO3), following the method of McGrath and Cunliffe (1985). Cadmium concentrations in the digest solutions were determined by graphite-furnace atomic absorption spectrometry (GF–AAS) (4100-ZL; PerkinElmer, Wellesley, MA). Quality control for soil analysis was ensured by the digestion of replicate samples of a Community Bureau of Reference (BCR) Certified Reference Light Sandy Soil CRM142R in each digestion batch of 54 samples.

Samples of wheat and barley grain were ground to <0.5 mm using a Reitch ultracentrifugal stainless steel mill (Glen Mills, Clifton, NJ) and oven-dried at 80°C before analysis. Subsamples (approximately 1 g) of the dried and ground grain were digested in XP1500plus Teflon PFA microwave liners (CEM Corp., Matthews, NC) using 3 mL of Primar ultrapure concentrated nitric acid (70% w/v) (Fisher Scientific, Hampton, NH), 2 mL of Primar 30% w/v hydrogen peroxide (Fisher Scientific), and 7 mL ultrapure water (18 M{Omega} specific resistance; ELGA Maxima, High Wycombe, UK). Digestion was performed at 115°C for 1 min and 175°C for 10 min, both at a maximal pressure of 3.1 MPa, using a CEM Mars X microwave oven. After completion of the heating process the vessels were cooled and made up to 25 mL with ultrapure water before analysis. Cadmium concentrations in the digest solutions were determined by GF–AAS using a palladium matrix modifier. Quality control was ensured by inclusion of the Certified Reference Materials BCR CRM191 brown bread and National Institute of Standards and Technology (NIST) SRM1567a wheat flour in each batch of samples. All glassware and microwave vessels were acid-washed and thoroughly rinsed with deionized and ultrapure water before use. The concentration of Cd in grain was expressed on a dry weight basis.

Statistical Analyses
Statistical analyses, including multiple linear regressions and analysis of variance (ANOVA), were performed using Genstat 5 for Windows (Numerical Algorithms Group, 1998). Where appropriate, variates that showed skewed distributions were log10–transformed before statistical testing to achieve normality and homogeneity of variances. For samples having Cd concentrations below the analytical detection limits (<0.003 mg kg–1 dry weight), concentrations equal to half the value of the detection limit were used in the subsequent statistical analyses.


   RESULTS AND DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
CONCLUSIONS
 REFERENCES
 
Quality Control
Repeated analysis of NIST SRM1567a wheat flour and BCR CRM191 brown bread gave mean Cd concentrations of 0.026 with a standard deviation (SD) of 0.002 mg kg–1 (n = 41) and 0.030 with a SD of 0.003 mg kg–1 (n = 7), respectively. These values were in excellent agreement with their respective certified values of 0.026 ± 0.002 and 0.0284 ± 0.0014 mg kg–1. For the certified soil BCR CRM142R, we obtained a mean value of 0.249 with a SD of 0.02 mg Cd kg–1 (n = 23), compared with the certified value of 0.25 ± 0.01 mg Cd kg–1. The results show that the digestion and analysis procedures used were reproducible and accurate for Cd analysis.

Paired Soil and Crop Samples
Soils collected from the surveys varied widely in their properties (Table 1). Soil Cd concentration varied by more than two orders of magnitude for the barley survey and almost three for the wheat survey. A small number of soils from the wheat survey were highly contaminated with Cd and Zn. In contrast, no soils from the barley survey contained more than 2.5 mg Cd kg–1.


View this table:
[in this window]
[in a new window]
  		Table 1. Summary of soil properties from the paired soil and crop surveys.

 
A wide range of Cd concentrations was found in the wheat grain collected over the three harvest years (1998–2000), with values ranging from 0.010 to 0.620 mg kg–1 dry wt. Median and mean grain Cd concentrations were 0.056 and 0.077 mg kg–1, respectively. The frequency distributions and the box-plots of Cd concentrations in wheat grain were broadly similar in the three years (Fig. 1a and 1b) , with no significant (P > 0.05) difference between the yearly mean concentrations according to ANOVA. No sample in 1998 exceeded the current EU limit for wheat grain of 0.235 mg Cd kg–1 dry wt., but approximately 6% of the grain samples collected in 1999 and 2000 exceeded the limit. The number of wheat grain samples having relatively high concentrations of Cd was larger than that found in the national representative survey in the UK reported previously (Chaudri et al., 1995). This was due to the sampling strategy adopted in this study, which was designed to include some Cd-contaminated soils.



View larger version (34K):
[in this window]
[in a new window]
  		Fig. 1. Cadmium concentrations in wheat and barley grain from the nationwide paired soil and crop surveys. (a) Frequency distribution of wheat grain Cd, (b) box-plot of wheat grain Cd, (c) frequency distribution of barley grain Cd, and (d) box-plot of barley grain Cd.

 
The concentrations of Cd in barley grain varied from less than the analytical detection limit (<0.003 mg kg–1) to 0.100 mg kg–1 dry wt. (Fig. 1c and 1d), with median and mean concentrations of 0.014 and 0.019 mg kg–1 dry wt., respectively. Of the 215 barley samples collected over the three years, no samples exceeded the EU limit for Cd in barley of 0.118 mg kg–1 dry wt., even though the sampling strategy was designed to include some contaminated soils. It is clear that the concentrations of Cd in barley grain were much lower than those in wheat grain. This may be partly due to the fact that the wheat survey included more high Cd sites (Table 1). However, the median Cd concentration for the wheat samples was fourfold higher than that for the barley samples, even though the median concentrations of soil Cd were similar for the wheat and barley surveys (Table 1), suggesting that wheat tends to accumulate more Cd in grain than barley under comparable conditions. A larger accumulation of Cd by wheat than by barley was also observed by Kjellström et al. (1975).

In the paired soil and crop surveys, 18 and 31 cultivars were recorded for wheat and barley, respectively, although only 5 wheat cultivars and 6 barley cultivars had 10 samples or more (Table 2). In both surveys, crop cultivars were found to be significantly different in their mean Cd concentrations according to ANOVA. The wheat cultivars Consort and Rialto appeared to have relatively large concentrations of Cd, whereas the cultivars Charger, Hereward, Malacca, and Spark had relatively small concentrations of Cd (Table 2). In barley, the cultivar Regina had a mean Cd concentration about double or more of the other cultivar's means. In the barley data set, ANOVA showed that there was a significant (P < 0.001) difference between the mean concentrations of Cd in spring barley (n = 125) and winter barley (n = 90), with mean concentrations of 0.012 and 0.018 mg kg–1 dry wt., respectively. However, in the survey data ANOVA cannot separate true cultivar effects from environmental variations, and some of the "apparent difference" between cultivars may be due to differences in soil Cd and other environmental factors. For example, 3 of the 10 samples for the wheat cultivar Consort were from contaminated soil, thus contributing to its overall high cultivar mean (Table 2).


View this table:
[in this window]
[in a new window]
  		Table 2. Mean and standard deviation of Cd concentrations in wheat and barley cultivars having five or more samples in the paired soil and crop surveys.

 
In the wheat data set, there was a significant correlation (r = 0.57) between the log-transformed soil Cd and log-transformed grain Cd concentrations (Fig. 2a) . Six samples from the most contaminated soils appeared to have a large influence on the overall correlation. However, the correlation was still significant (r = 0.40) when these six samples were excluded. In contrast, the correlation between log-transformed soil Cd and log-transformed Cd concentration in barley grain was weak and not significant (r = 0.16; Fig. 2b).



View larger version (26K):
[in this window]
[in a new window]
  		Fig. 2. Relationship between soil total Cd and the concentrations of Cd in (a) wheat grain and (b) barley grain from the nationwide paired soil and crop surveys.

 
Long-Term Sewage Sludge Experiments
The long-term sewage sludge experiments at Gleadthorpe and Rosemaund provided a useful resource for investigating the effects of previous sludge applications on Cd uptake into crops. Table 3 shows a summary of main soil properties in the two experiments. Winter wheat was grown in both experiments in 1999, and spring barley in 2000. The log-transformed Cd concentrations in wheat and barley grain were found to correlate with log-transformed soil total Cd, although the relationships were different between the two experimental sites (Fig. 3) . The more acidic soil at Gleadthorpe tended to produce higher Cd concentrations in grain, particularly in the plots with elevated concentrations of soil Cd. As in the paired soil and crop surveys, wheat grain contained considerably higher Cd than barley grain in the two experiments. None of the wheat and barley grain samples exceeded the current EU limits, although some of the barley samples produced at Gleadthorpe were close to the limits.


View this table:
[in this window]
[in a new window]
  		Table 3. Soil characteristics in two long-term sewage sludge experiments at Gleadthorpe and Rosemaund, England.

 


View larger version (24K):
[in this window]
[in a new window]
  		Fig. 3. Relationship between soil total Cd and the concentrations of Cd in (a) wheat grain and (b) barley grain in two long-term sewage sludge experiments.

 
Soil Factors Influencing Grain Cadmium
It is clear from Fig. 2 and 3 that the total concentration of Cd in soil alone does not reflect Cd bioavailability to plants. Multiple linear regression analysis was therefore used to identify significant soil variables that could be used to fit a model describing grain Cd concentration. Soil variables used in the multiple regression initially included total Cd concentration, pH, organic mater content, the contents of Fe and Al oxides, total Zn concentration, and Zn to Cd ratio. Soil variables (except pH) and grain Cd concentrations were log-transformed before regression to ensure homogeneity of variances. In some cases (see below), crop cultivar was also included as a factor in the regression model.

Wheat
Because there was no statistically significant difference between the concentration of Cd in wheat grain over the three harvest years sampled in the paired soil and crop surveys, data from all three years (n = 162) were pooled for subsequent statistical analysis. Of all soil variables tested in the multiple regression, only soil Cd concentration and pH were found to be significant (P < 0.05) terms. The combination of soil Cd concentration and pH resulted in a highly significant (P < 0.001) regression model that explained 49% of the variance in the concentration of Cd in wheat grain from the paired soil and crop surveys. The model obtained was:

[1]

Figure 4a shows the fit between the observed and fitted grain Cd values from this regression model. In the paired soil and crop survey for wheat, inclusion of soil variables such as organic matter, the contents of Fe and Al oxides, total Zn concentration, and Zn to Cd ratio into the regression model (Eq. [1]) produced little further improvement, suggesting that these variables have relatively small effects on Cd bioavailability to wheat. Cultivar effects on wheat Cd concentrations were found to be significant (P < 0.05) by ANOVA and incorporation of a cultivar term as a factor into the model resulted in an improved fit to the observed data, increasing the percentage of variance accounted for by the model from 49 to 56%. The cultivar term was a significant variable in the regression (P < 0.05), suggesting that there were real differences between cultivars in Cd uptake and/or transport to grain. Figure 4b shows the fit between the observed and fitted grain Cd values using cultivar, soil total Cd, and pH. Cultivar differences in Cd accumulation have been widely reported. For example, Wenzel et al. (1996) also found that a cultivar term was a significant variable in regression models explaining Cd uptake in wheat grown on seven soil types. However, inclusion of the cultivar term in the regression model means that each cultivar is fitted with an individual equation, thus greatly reducing the number of samples and the robustness associated with each equation. In the wheat survey, 13 out of the 18 cultivars recorded had fewer than 10 samples each. While differences between cultivars are noted, establishing cultivar-specific models would require a much larger survey so that sufficient numbers of samples could be collected for each of the 18 cultivars recorded.



View larger version (36K):
[in this window]
[in a new window]
  		Fig. 4. Relationship between observed and fitted values of Cd concentrations in wheat grain. Fitted values were calculated from the regression models including (a) total soil Cd and pH as variables as in Eq. [1] for the paired soil and crop surveys; (b) total soil Cd, pH, and cultivar terms for the paired soil and crop surveys; (c) total soil Cd and pH as variables as in Eq. [2] for the two long-term sewage sludge experiments; and (d) total soil Cd and pH as variables as in Eq. [3] for the combined data set.

 
Several studies modeling the uptake of Cd into crops have identified soil chloride as an important variable in regression modeling (Bingham et al., 1984; Li et al., 1994; McLaughlin et al., 1994, 1999; Smolders et al., 1998; Norvell et al., 2000). Chloride (Cl) concentrations were determined in selected soils from the 2000 harvest year, which had been sampled at either coastal or inland locations. Although Cl concentrations were marginally higher in the coastal soils than the inland ones, all samples had low Cl levels (<5 mg kg–1), which is two orders of magnitude below the concentration where Cl was found to affect Cd uptake by potato tubers (McLaughlin et al., 1994, 1999). These two papers dealt with crops grown in soils containing high Cl concentrations or were irrigated with saline waters in Australia, where soil Cl concentrations reached >1500 mg kg–1. Chloride levels in British soils are therefore not expected to be a significant factor influencing Cd uptake by wheat.

Multiple regression was also applied to the wheat data set of the long-term sewage sludge experiments at Gleadthorpe and Rosemaund. Incorporating both soil pH and soil Cd concentration into a regression model showed that grain from both sites followed the same relationship, accounting for 73% of the variance in grain Cd concentrations (Fig. 4c). The regression equation (Eq. [2]) was similar to that obtained from the paired soil crop survey (Eq. [1]):

[2]

Inclusion of other soil variables did not improve the model fit further. The high degree of fit observed is probably due to the facts that only two soil types were included, and the same wheat cultivar (cv. Hereward) was grown at both sites. The predictive power of equations similar to Eq. [1] and [2] has also been shown for other crops including lettuce, Swiss chard, and corn leaves in several sewage sludge experiments, although coefficients were different for different crops (McBride, 2002).

The similarity in the coefficients between Eq. [1] and [2] indicates that the influence of soil Cd concentration and pH on wheat Cd uptake was similar in the two long-term sludge experiments and in the nationwide paired soil and crop surveys. Finally, data sets from both the paired soil and crop survey and from the long-term sewage sludge experiments were combined and used in multiple regression. This resulted in the following equation (Eq. [3]), which explained 53% of the variance of grain Cd concentration:

[3]

The relationship between measured grain Cd and fitted grain Cd using Eq. [3] is shown in Fig. 4d.

Barley
Combining the data for spring barley from the two long-term sewage sludge experiments in a multiple regression incorporating soil pH and Cd concentration resulted in Eq. [4]:

[4]
which explained 62% of the variance in barley grain Cd concentrations. Inclusion of other soil variables did not improve the model fit significantly. Figure 5a shows measured grain Cd concentrations versus the concentrations calculated from Eq. [4]. The model appeared to fit the Rosemaund data better than the Gleadthorpe data. For the latter, Eq. [4] overestimated Cd availability at the higher end of grain Cd concentrations, but underestimated Cd availability at the lower end of grain Cd concentrations.



View larger version (26K):
[in this window]
[in a new window]
  		Fig. 5. Relationship between observed and fitted values of Cd concentrations in barley grain. Fitted values were calculated from the regression models including (a) total soil Cd and pH as variables as in Eq. [4] for spring barley in the two long-term sewage sludge experiments and (b) total soil Cd, pH, and cultivar terms for winter barley in the paired soil and crop surveys.

 
In the paired soil and crop data set for barley (n = 215), none of the soil variables tested alone or in combination could significantly explain the variation of grain Cd concentration. The data set was then separated into winter and spring barley. In the data subset for winter barley only (n = 90), soil pH was found to be the most significant variable (P < 0.001), which explained 20% of the variance of grain Cd concentrations. Inclusion of soil Cd concentration (P < 0.05) into the regression model improved the fit of the model to a small extent (by increasing the percentage variance accounted for to 22%), and gave the following equation:

[5]

Although Eq. [5] indicates a positive effect of soil Cd concentration and a negative effect of soil pH on grain Cd concentration, which was consistent with Eq. [1–4], the model fit was poor. Inclusion of other soil variables did not improve the model fit. However, the inclusion of a cultivar term into the regression did substantially improve the goodness of fit to 47% (Fig. 5b), again indicating the influence of cultivar on Cd accumulation. As with the wheat survey, the barley survey encompassed a large number of cultivars (31), with many of them (25) having fewer than 10 samples each. Cultivar-specific regression models would not be robust for the cultivars with few samples.

In the data subset for spring barley (n = 125), none of the soil variables tested alone or in combination could significantly explain the variation of grain Cd concentration. This was most probably due to the generally low and narrow range of Cd concentrations in spring barley grain.

European Union Limits for Grain Cadmium and Sewage Sludge Applications
Equation [3] may be used to predict wheat grain Cd concentration under different soil conditions. Figure 6 shows the predicted grain Cd (±95% confidence intervals) as influenced by different combinations of soil total Cd and pH. As expected, the predicted wheat grain Cd increases with soil total Cd and decreases with soil pH. At a soil pH of 5, total soil Cd has to be less than 1.0 mg kg–1 to ensure compliance, with 95% confidence, of wheat grain Cd with the current EU MPC (0.235 mg kg–1 dry wt.). At soil pH of 6 and 7, the corresponding threshold values of soil total Cd are 2.4 and 5.8 mg kg–1, respectively. Currently, the UK regulations for land applications of sewage sludge allow soil Cd to accumulate up to 3 mg kg–1, with no allowance made for the effects of pH on Cd bioavailability (Statutory Instrument, 1989, 1990). Clearly, the UK regulations on sewage sludge applications to soil would not be sufficiently protective in light of the current EU MPC for Cd in grain. With a total soil Cd of 3 mg kg–1, soil pH has to be maintained at 6.3 or above to ensure compliance with the MPC with 95% confidence. The current draft European Commission Working Document on Sludge (European Commission, 2000) proposes revisions to the current soil metal limits. It has proposed maximum permitted soil Cd concentrations for sludge applications that depend on soil pH, so that a lower soil metal concentration is necessary for soils of lower pH. The latest draft of the working document specifies limit values of Cd in soil of 0.5 mg Cd kg–1 for soils with pH in the range of 5 to 6, 1.0 mg Cd kg–1 for soils with pH in the range of 6 to 7, and 1.5 mg Cd kg–1 for soils having pH of ≥7. Figure 6 shows that the proposed soil Cd limits would provide sufficient protection against the possibility of wheat grain Cd in breach of the current EU MPC.



View larger version (22K):
[in this window]
[in a new window]
  		Fig. 6. Predicted Cd concentration in wheat grain as influenced by total soil Cd and pH. Vertical bars represent 95% confidence intervals.

 

   CONCLUSIONS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
REFERENCES
 
Multiple regression analysis showed that soil total Cd and pH were the two significant variables that influenced the concentration of Cd in wheat grain, in both nationwide paired soil and crop surveys and long-term sewage sludge experiments. Wheat grain Cd concentration can be predicted reasonably well from soil total Cd and pH using the following model: log(grain Cd) = a + b log(soil Cd) – c(soil pH), with 53% of the variance being accounted for. The coefficients obtained from the data sets of the paired soil and crop surveys and from long-term sewage sludge experiments were similar, suggesting similar controlling factors of Cd bioavailability in sludge-amended or unamended soils. Inclusion of other soil properties, including soil organic matter, contents of Fe, Al, and Mn oxides, and Zn to Cd ratio, did not significantly improve the fit of the regression model. Crop cultivar was a significant factor for the concentration of Cd in wheat grain. Calculations using the obtained regression model showed that the current UK regulations on sewage sludge applications, allowing an accumulation of soil Cd up to 3 mg kg–1, are not sufficiently protective against the possibility of wheat grain Cd exceeding the current EU MPC. In contrast, the newly proposed EU draft regulations on sewage sludge applications, allowing <0.5 mg Cd kg–1 for soil pH of 5 to 6, <1.0 mg Cd kg–1 for soil pH of 6 to 7, and 1.5 mg Cd kg–1 for soil pH of >7, would provide sufficient protection.

Barley accumulated much lower concentrations of Cd in the grain than wheat under comparable soil conditions. Soil pH and soil total Cd concentration were the two significant factors influencing the concentration of grain Cd for a spring barley cultivar in the two long-term sludge experiments, and for winter barley cultivars in the nationwide paired soil and crop surveys. However, the models obtained were much less satisfactory for predicting the concentration of Cd in barley grain. Results from both the nationwide paired soil and crop surveys and long-term sludge experiments showed that barley grown in Britain under typical field conditions and management regimes is unlikely to exceed the current EU MPC for Cd.


   ACKNOWLEDGMENTS
 
This work was funded by the UK Home-Grown Cereals Authority and the Royal Agricultural Society of England Hill's Bequest. Rothamsted Research receives grant-aided support from the UK Biotechnology and Biological Sciences Research Council. The cooperation of farmers who provided sampling sites for the paired crop and soil samples is gratefully acknowledged.


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 


Related articles in JEQ:

This Issue in Journal of Environmental Quality

JEQ 2004 33: 413-418. [Full Text]  




This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Related articles in JEQ
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via ISI Web of Science (21)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Adams, M. L.
Right arrow Articles by Chambers, B. J.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Adams, M. L.
Right arrow Articles by Chambers, B. J.
Agricola
Right arrow Articles by Adams, M. L.
Right arrow Articles by Chambers, B. J.
Related Collections
Right arrow Heavy Metals
Right arrow Soil Pollution

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
The SCI Journals Agronomy Journal Crop Science
Vadose Zone Journal Journal of Plant Registrations
Journal of Natural Resources
and Life Sciences Education		 Soil Science Society of America Journal