Spatial Variability of Micronutrients in Rice Grain and Paddy Soil

Spatial Variability of Micronutrients in Rice Grain and Paddy Soil

Pedosphere 19(6): 748–755, 2009 ISSN 1002-0160/CN 32-1315/P c 2009 Soil Science Society of China  Published by Elsevier Limited and Science Press Sp...

262KB Sizes 0 Downloads 48 Views

Pedosphere 19(6): 748–755, 2009 ISSN 1002-0160/CN 32-1315/P c 2009 Soil Science Society of China  Published by Elsevier Limited and Science Press

Spatial Variability of Micronutrients in Rice Grain and Paddy Soil∗1 WANG Lin, WU Jia-Ping∗2 , LIU Yan-Xuan, HUANG Hui-Qing and FANG Qian-Fang College of Environment and Natural Resources, Zhejiang University, Hangzhou 310029 (China) (Received September 28, 2008; revised September 24, 2009)

ABSTRACT Consumption of rice is the main source of micronutrients to human in Asia. A paddy field with unknown anthropogenic contamination in Deqing County, Zhejiang Province, China was selected to characterize the spatial variability and distribution of micronutrients in rice grain and soil. A total of 96 paired soil and rice grain samples were collected at harvest. The micronutrients in the soil samples were extracted by diethylenetriamine pentaacetic acid (DTPA). The mean micronutrient concentrations in rice grain were 3.85 μg Cu g−1 , 11.6 μg Fe g−1 , 39.7 μg Mn g−1 , and 26.0 μg Zn g−1 . The mean concentrations were 2.54 μg g−1 for DTPA-Cu, 133.5 μg g−1 for DTPA-Fe, 30.6 μg g−1 for DTPA-Mn, and 0.84 μg g−1 for DTPA-Zn. Semivariograms showed that measured micronutrients in rice grain were moderately dependent, with a range distance of about 110 m. The concentrations of the DTPA-extractable micronutrients all displayed strong spatial dependency, with a range distance of about 60 m. There was some resemblance of spatial structure between soil pH and the grain Cu, Fe, Mn, and Zn. By analogy, similar spatial variation was observed between soil organic matter (SOM) and DTPA-extractable micronutrients in the soil. Kriging estimated maps of the attributes showed the spatial distributions of the variables in the field, which is beneficial for better understanding the spatial variation of micronutrients and for potentially refining agricultural management practices at a field scale. Key Words:

correlation, dependency, kriging, soil properties

Citation: Wang, L., Wu, J. P., Liu, Y. X., Huang, H. Q. and Fang, Q. F. 2009. Spatial variability of micronutrients in rice grain and paddy soil. Pedosphere. 19(6): 748–755.

Micronutrients (Cu, Fe, Mn, and Zn) are essential for human health, and their deficiency can lead to substantial public health problems. Micronutrient deficiency is a critical concern among children and women throughout the world. For example, approximately 66% of the children and women aged 15–44 years in the developing countries have iron deficiency, and 10%–20% of women of childbearing age in developed countries have iron deficiency anemia (Scrimshaw, 1991). But when present at high concentrations, micronutrients can also result in phytotoxicity (Mitchell and Burridge, 1979). Soil is the preeminent source of most biologically active micronutrients (Mitchell and Burridge, 1979), and therefore, it is beneficial to know the relationship between soil characteristics and micronutrient concentrations in soil and plant. Metal concentrations that are present in soil solution and available for plant uptake are the most important in environmental terms (Rieuwerts et al., 2006). In general, micronutrient contents are affected by several soil parameters, such as pH, organic matter (SOM), soil texture, and concentrations of other micronutrients (Norvell et al., 2000; Kabata-Pendias and Pendias, 2001). Among these variables, the effects of pH and SOM on the extractability of metals deserve special attention (Karaca, 2004). Rieuwerts et al. (2006) described that pH was the most important predictor variable for the estimation of extractable heavy metals from soils. SOM has also been of particular interest in studies of heavy metal sorption by soils because of the significant effect on binding of heavy metals in soil and speciation in soil solution (Lo et al., 1992). ∗1 Project

supported by the National Natural Science Foundation of China (No. 30771253) and the Science and Technology Department of Zhejiang Province, China (No. 2006C22026). ∗2 Corresponding author. E-mail: [email protected]



Observations of soil and plant properties are always related to particular locations in space and time. Geostatistics provides the basis for incorporating spatial or temporal coordinates of observations in data processing. It is based on the theory of a regionalized variable (Webster and Oliver, 2001). The spatial correlation function (semivariogram model) allows a quantification of the spatial variability. Considering that property variables vary in space through variogram models, geostatistical interpolation method, such as kriging, provides a solution to make optimal predictions at unsampled locations. Kriging is a form of weighted interpolation with the weights based on the structure of the semivariogram (Webster and Oliver, 2001). Many studies had been performed on the applications of geostatistics method in studying the spatial and temporal variability of soil properties (Wu et al., 2002). This method has also been applied to study the spatial variation of plant properties (Timlin et al., 1998; Wu et al., 2002). However, little information is available about the spatial variability of soil and plant attributes in a paddy field. Unlike other crops, rice production is unique for its completely leveled surface and evenly distributed soil moisture condition throughout an entire paddy field. The major objective of this study was to characterize the magnitudes, spatial variability, and spatial distributions of Cu, Fe, Mn, and Zn in rice grain and paddy soil using statistical and geostatistical methods. MATERIALS AND METHODS This study was conducted in a paddy field located in the Hang-Jia-Hu Plain, which is a main rice production area in Deqing County, Zhejiang Province, Southeast China (Fig. 1). This field is a part of privately owned farm land and used for growing rice with an average yield of 7 500 kg ha−1 . On June 5, 2005, approximately 150 kg ha−1 N-P-K compound fertilizer was added as basal fertilizer three days before seeding. Then, approximately 150 kg ha−1 N fertilizer on June 25, 150 kg ha−1 K fertilizer (KCl) on the July 2, and 150 kg ha−1 N fertilizer again on July 25 were applied. A total of 96 paired topsoil (0–20 cm, about the thickness of cultivated horizon) and rice grain samples were obtained on October 25, 2005 when the rice was ready for harvest (Fig. 1). A grid sampling with grid intervals from about 20, 30, and 40 m was used. In addition, seven sites were located as nests and a nested sampling with intervals of about 1, 2, 3, 5, 8, and 12 m was conducted. This integrated grid and nested sampling approach could reveal the spatial variability at the field level with limited sample numbers (Webster and Oliver, 2001). For the sake of eliminating influence of species difference, all grain samples were from a uniform genotype of Japonica rice (Oryza Staiva L.), Xiushui 09. To avoid any unexpected potential contamination from road, all samples were collected 5 m away from the field boundary.

Fig. 1

General location of the studied paddy field and the sample distribution in the field.

Rice grain samples were dried at 60 ◦ C in an oven overnight. Dried heads were threshed by hand and ground in a mill (MM301, Retsch, Haan, Germany). Subsamples of 0.5 g grain were digested in concentrated nitric acid and then dissolved in deionized water for analysis of micronutrients. Soil samples were air-dried and ground to pass through a 2-mm nylon sieve for laboratory analysis. A 50 g


L. WANG et al.

subsample was ground in an agate mortar and then extracted by diethylenetriamine pentaacetic acid (DTPA) using the method of Lindsay and Norvell (1978). A 1:3 soil/solution ratio (w:v) was used to increase recovery of filtrate and decrease filtration time as described in detail by Norvell et al. (2000). Blanks and certified standard rice flour and soil samples were also digested synchronously in order to check error levels of the above digestion procedure. Analyses for micronutrients in both soil and grain samples were carried out by inductively coupled plasma-mass spectrometry (ICP-MS) (Agilent7500a, Agilent, Santa Clara, CA, USA). Soil pH was measured using a pH meter (Sartorius Basic pH meter PB-10) with a soil:water ratio of 1:2.5, and SOM was measured by Walkley-Black method (Nelson and Sommers, 1982). Univariate statistics and correlation analysis were performed using SPSS version 13.0 for windows (SPSS Inc., Chicago, IL). Kolmogorov-Smirnov test was conducted to determine data normality. For the data with severely skewed distribution, a natural log-transformation was applied. Pearson correlation coefficients were calculated to determine the relationship between micronutrient concentrations and soil properties. Geostatistical analysis was performed with GS+ (version 3.1 for Windows, Gamma Design Software, Plainville, MI). Spatial variability was exhibited by standardized semivariograms, that is, the semivariance was divided by their sample variance. It determines the relative spatial variance independent of the magnitude of the population variance. A spherical model, the most frequently used one in geostatistics, was selected to fit the semivariograms. It is defined using the following equation (Webster and Oliver, 2001):  C0 + C[1.5(h/A) − 0.5(h/A)3 ] for h ≤ A (1) γ(h) = for h > A C0 + C Where γ(h) is the semivariance at lag distance h; C0 is the nugget variance; C0 + C, sill variance; A, range. Using the fitted models, spatial interpolation was accomplished with a block kriging approach. Block kriging estimates average values over an area (i.e., a block) around an interpolation point, as opposed to an estimate for a precise point with point (punctual) kriging (Webster and Oliver, 2001). Geostatistics assumes that a variable satisfies a second-order stationarity, that is, the mean u is not spatially varied but invariant (Webster and Oliver, 2001). However, in some circumstances, the expected mean value is no longer constant but is a function of position, that is, a trend or drift exists. The assumption of second-order stationarity does not hold. In this case, we modeled this trend with a first-order function (Webster and Oliver, 2001). u(x, y) = a + bx + cy


where u(x, y) is the mean value of the modeled trend, x and y are coordinates, and a, b, and c are fitted coefficients. After the removal of the modeled trend from the original data, the variogram was recalculated using the detrended data. If the stationarity assumption was satisfied, kriging was performed on the detrended data. The trend was added to the kriging output to form the final map. RESULTS Summary statistics A statistical summary of the attribute data is presented in Table I. Application of KolmogorovSmirnov test indicated that most attributes were normally distributed, with medians close to their means. The DTPA-extractable Zn (ZnDTPA ) was the only variable with large skewness and kurtosis far from a normal distribution. A natural logarithmic transformation was then applied to ZnDTPA . Besides, natural log-transformation of SOM was also performed to stabilize its variance (Cambardella et al., 1994). There was a wide range of variability for the attributes. The Cu concentration in rice grain (Cugrain )



TABLE I Statistics of micronutrients in rice grain and in paddy soil, soil pH, and soil organic matter (SOM) in the studied field Variable





pH SOM ln(SOM) Micronutrient in grain Cu Fe Mn Zn Micronutrient in soilc) Cu Fe Mn Zn lnZn a) Standard






5.66 15.8 2.7

μg 5.55 14.1 2.6

0.4 4.9 0.3

% 7.1 31.0 11.1

5.01 9.4 2.2

6.79 29.0 3.4

0.99 1.25 0.75

0.48 0.70 −0.15

3.85 11.6 39.7 26.0

3.80 11.1 38.9 22.4

0.86 3.5 7.1 8.5

22.3 30.2 17.9 32.7

2.18 5.5 26.9 14.1

5.82 21.3 59.8 48.5

0.61 0.68 0.72 1.32

0.27 0.24 0.35 0.48

2.54 133.5 30.6 0.84 −0.21

2.50 130.7 28.8 0.78 −0.25

0.45 39.0 8.2 0.24 0.24

17.7 29.2 26.8 28.6 −114.0

1.62 48.3 16.3 0.56 −0.58

3.86 206.5 51.5 2.11 0.75

0.50 −0.02 0.51 2.64 1.45

−0.04 −0.76 −0.32 9.61 3.06

b) Coefficient

of variation;

c) Extracted


Maximum g−1

by diethylenetriamine pentaacetic acid (DTPA).

varied from 2.18 to 5.82 μg g−1 , with a mean of 3.85 μg g−1 . Grain Fe (Fegrain ) ranged from 5.5 to 21.3 μg g−1 , about a four-fold difference, with a mean of 11.6 μg g−1 . The mean concentration of grain Mn (Mngrain ) was 39.7 μg g−1 , with a minimum of 26.9 μg g−1 and a maximum of 59.8 μg g−1 . Grain Zn (Zngrain ) ranged from 14.1 to 48.5 μg g−1 and averaged at 26.0 μg g−1 . The coefficients of variation (CV) for grain attributes varied from 17.9% (Mngrain ) to 32.7% (Zngrain ). The means for DTPA-extractable soil Cu (CuDTPA ), Fe (FeDTPA ), Mn (MnDTPA ) and Zn (ZnDTPA ) were 2.54, 133.5, 30.6 and 0.84 μg g−1 , respectively. They varied substantially from 2.4 (CuDTPA ) to 4.3-fold (FeDTPA ) between maximum and minimum values in the field. The CV values for DTPA-extractable attributes ranged from 17.7% (CuDTPA ) to 29.2% (FeDTPA ). Soil properties also varied widely. Soil pH ranged from a strongly acidic value of 5.01 to neutral value of 6.79, with a mean of 5.66. The SOM had a range of approximately 3-fold between the minimum and maximum values. According to the classification of Wilding (1985), all the variables were ranked into moderate variability. The pH variability was greatly suppressed due to its log-transformation of original hydrogen ion concentration. To examine the relationship between micronutrient concentrations and soil properties, a Pearson correlation analysis was performed. Complicated correlations between soil properties and micronutrient concentrations were observed (Table II). TABLE II Pearson correlation coefficients among the micronutrients in rice grain (Cugrain , Fegrain , Mngrain and Zngrain ) and in paddy soil (CuDTPA , FeDTPA , MnDTPA and ZnDTPA )a) , soil pH and soil organic matter (SOM) Variable








ln(ZnDTPA )


Fegrain Mngrain Zngrain CuDTPA FeDTPA MnDTPA ln(ZnDTPA) pH SOM

−0.10 0.69** 0.13 0.52** −0.53** 0.18 0.13 0.48** −0.36**

−0.15 −0.14 −0.04 0.16 −0.01 0.09 −0.13 0.08

0.23* 0.34** −0.40** 0.33** 0.08 0.31** −0.34**

−0.10 0.07 0.27* 0.01 0.10 −0.13

−0.66** 0.25* 0.13 0.69** −0.53**

−0.03 0.22* −0.64** 0.63**

0.23* 0.10 −0.36**

0.14 0.22*


*, **Significant at P < 0.05 and P < 0.01 levels, respectively. a) Micronutrients in paddy soil were extracted by diethylenetriamine pentaacetic acid (DTPA).


L. WANG et al.

There was difference in the influence of pH on various elements. Significant and positive correlation was observed between soil pH and CuDTPA . A significant and negative relationship was found between soil pH and FeDTPA . As for micronutrients in rice grain, pH was positively correlated with Cugrain and Mngrain . No apparent association between soil pH and other micronutrients existed. SOM had a positive effect on FeDTPA but a negative effect on Cu and Mn in both soil and grain. With regard to the correlation between micronutrients in grain and in soil, Cugrain and Mngrain seemed to be positively associated with CuDTPA and MnDTPA , respectively. Grain Fe and Zn had no significant correlation with the corresponding soil DTPA-extractable micronutrients. Semivariogram analysis To determine the degree of spatial correlation, sample omnidirectional semivariograms were computed. The spherical model (Eq. 1) was used to fit the variograms. Parameters of the model used to describe the semivariograms for the variables are shown in Fig. 2.

Fig. 2 Experimental omnidirectional semivariograms (dots) and fitted semivariograms (lines) for the data obtained in the study. C0 is the nugget effect; C0 + C is sill; A is the range distance of fitted spherical models; C0 /(C0 + C) is the ratio of nugget to sill (%); Cugrain , Fegrain , Mngrain and Zngrain are the Cu, Fe, Mn and Zn in rice grain, respectively; CuDTPA , FeDTPA , MnDTPA , and ZnDTPA are the Cu, Fe, Mn, and Zn in paddy soil, respectively, extracted by diethylenetriamine pentaacetic acid (DTPA); ln(SOM) and ln(ZnDTPA ) are the natural logarithmic transformed soil organic matter and ZnDTPA , respectively.

Semivariograms showed that micronutrients, soil pH, and SOM spatially varied within the field. The calculation of semivariance was set to a maximum lag distance of 150 m, with an interval between two lag classes of 10 m. Each lag distance class contained at least 100 pairs of points in semivariogram calculations. With regard to Zngrain , the second-order stationarity assumption was not satisfied. We removed the general trend from the original data using the Eq. 2. After the removal of the trend, the residual values were used to recalculate the semivariogram. The fitted model reached its maximum semivariance (sill) at the range distance. Range is of conside-



rable importance, marking the distance beyond which samples have no spatial dependence. The ratio between the nugget (C0 ) and sill (C0 + C) characterizes the random component in the whole field spatial variability, providing quantitative measures of spatial dependency for every attribute. The range was about 110 m for all grain micronutrients. In contrast, the range of DTPA-extractable elements was only 60 m. Nugget variance (C0 ) was 0.57 for Cugrain , 0.60 for Fegrain , 0.44 for Mngrain , and 0.39 for Zngrain residual, representing 47.5%, 50.0%, 36.4%, 30.2% of the sill variance. C0 was 0.19 for CuDTPA , 0.23 for FeDTPA , 0.25 for MnDTPA , and 0.25 for ln(ZnDTPA ), representing 20.4%, 20.4%, 22.3%, 22.3% of the sill variance. A much larger nugget variance and nugget to sill ratio and longerrange distance for grain micronutrients indicated that there was a different spatial pattern from that of DTPA-extractable micronutrients. The similarity showing within the group of grain or soil DTPAextractable micronutrients suggested that some common causes existed for each group. Soil pH had a similar spatial structure with grain micronutrients, with C0 , 0.51; range, 110 m; nugget to sill ratio, 44.7%. Natural log-transformed SOM showed similarity to the DTPA-extractable micronutrients, with C0 , 0.12; range, 60 m; nugget to sill ratio 9.4%. Based on the ratio of nugget and sill, the spatial dependency of the data was assessed. Cambardella et al. (1994) defined this ratio of < 25, 25 to 75, and > 75 as categories of strong, moderate, and weak spatial dependence, respectively. According to this classification, DTPA-extractable micronutrients showed a strong spatial dependency, whereas the grain micronutrients were moderately dependent. Soil pH fell in the same class as grain micronutrients, whereas SOM was similar to the group of DTPA-extractable micronutrients. Spatial prediction and mapping Using the fitted models and corresponding parameters, we performed a block kriging with a block size of 1 m × 1 m to obtain interpolated values for all variables throughout the paddy field (Fig. 3). For Zngrain , the block kriging was performed on the detrended data, followed by adding trend surface to kriging values to yield the final map.

Fig. 3 Block kriging maps of the attributes in the paddy field. Cugrain , Fegrain , Mngrain , and Zngrain are Cu, Fe, Mn, Zn in rice grain, respectively; CuDTPA , FeDTPA , MnDTPA , and ZnDTPA are Cu, Fe, Mn, and Zn in paddy soil, respectively, extracted by diethylenetriamine pentaacetic acid (DTPA).

These maps clearly showed that all the attributes were distributed in spatially dependence pattern across the field. For Cu and Mn in both rice grain and soil DTPA extractants, higher values were


L. WANG et al.

found in the west and lower in the east. In contrast, Fe in grain and soil both displayed adverse spatial patterns, higher in the east and lower in the west. Concentrations of Zngrain showed spatial pattern alike that of Mngrain and Cugrain . The ZnDTPA tended to distribute relatively evenly. Large difference between Zngrain and ZnDTPA maps confirmed their weak association. Soil pH map showed some similar characteristics to that of Cugrain , CuDTPA , Mngrain , MnDTPA , and Zngrain , and SOM similar to Fegrain and FeDTPA . DISCUSSION In this study, most of the DTPA-extractable micronutrients in soil fell in the range found by Liu et al. (2004) in the Hang-Jia-Hu Plain on a larger scale. Grain Fe and Zn were normal and consistent with the results of Herawati et al. (1998) and Gregorio et al. (2000). The Cu and Mn concentrations in our rice grain samples were similar to the values reported by Nriagu and Lin (1995) and Chukwuma (1995). The rice grain produced in this field was acceptable for human consumption at both nutritional and toxic levels. The data on yields of the field, with an average rice of 7 500 kg ha−1 , indicated that they were good. However, there may be differences in the nutrient content in rice, which is adequate for obtaining high yields and is essential for human health and which is affected by the amount people consumes daily and by the bioavailability of the micronutrients in rice. This is an important topic but beyond the scope of this article. Complicated relationships between micronutrients in rice grain and that in soil DTPA-extractants were observed as previously reported by Boekhold and Van der Zee (1994), and Berndtsson and Bahri (1995). Inconsistent correlations were noted between various micronutrient concentrations and soil properties. For example, soil pH had positive effects on Cugrain , CuDTPA , and Mngrain , wheras had a negative influence on FeDTPA . A positive correlation between micronutrients and SOM is often mentioned in the literature (Sharma et al., 2004). However, contradictory results have also been reported by Haldar and Mandal (1979). In that study, Cu and Mn both in grain and soil decreased as SOM increased, whereas FeDTPA was positively correlated with SOM. Usually, it is assumed that micronutrients in grain reflect the level of available microelements in soils. However, this assumption can only hold if all other nutrients are available in sufficient quantities and the micronutrient studied is growth-limiting. In this study, pH ranged from acidic to neutral. The measured concentrations of soil micronutrients are all in ranges that unlikely become growth-limiting according to Lindsay and Norvell (1978). Both of the micronutrients in rice grain and paddy soil fall into the ranges typically found in the region, and there have been no reports about any micronutrients deficiency occurring in the area. For these reasons, we doubt about growth limitation from micronutrient deficiencies. Although it is interesting and useful for us to know the underlined mechanism of the spatial variation of microelement concentrations in rice grain and soils, as well as their inter-relationships, this is beyond the scope of this article. Geostatistical analysis indicated that all variables showed obvious spatial dependency. Soil DTPAextractable micronutrients presented strong spatial dependency but presented moderate dependency for grain micronutrients. Smaller range values for DTPA-extractable micronutrients manifested their patchy distribution, which may be the result of extrinsic factors, such as long-term manure, tillage, or other production activities in this field (Cambardella et al., 1994). Grain micronutrients tended to distribute in a more even pattern, showing less sensitivity to soil factors. Interpolated maps of the variables measured in the field revealed the distribution of micronutrients in great detail. The maps of pH, Cugrain , and Mngrain showed spatial similarity to some extent. Organic matter showed similar pattern compared with Fegrain and FeDTPA . There was a general trend in spatial distribution between micronutrients in grain and corresponding elements in the soil, except for Zn. Therefore, spatial patterns of soil DTPA-extractable micronutrients still could indicate those of rice grain in general. The results reported here suggested that it should be necessary to consider spatial variability of soil



parameters and that of micronutrients in crops when selecting sites for crop breeding for micronutrient enrichment of grain. Great variation and distinct spatial variability of micronutrients in crop grain could result from soil heterogeneity, which could potentially lead to faults in crop breeding strategies. Furthermore, knowledge on magnitude and spatial distribution of micronutrients in both soil and rice grain can facilitate future site specific management and fertilization approaches. ACKNOWLEDGEMENTS The authors are grateful to Mr. LI Jian-Guo from the Agricultural Extension Center in Deqing County, China and Prof. LI Ren-An, Drs. FENG Ying and YANG Xiao-E from the College of Environment and Natural Resources, Zhejiang University of China for their assistance in site selection, sampling, and sample analysis. We extend our great appreciation to Dr. C. WANG from the Agriculture and Agri-Food Canada and the anonymous reviewers for their valuable comments and suggestions. REFERENCES Berndtsson, R. and Bahri, A. 1995. Field variability of element concentrations in wheat and soil. Soil Sci. 159: 311–320. Boekhold, A. E. and Van der Zee, S. E. A. T. M. 1994. Field scale variability of cadmium and zinc in soil and barley. Environ. Monit. Assess. 29: 1–15. Cambardella, C. A., Moorman, T. B., Novak, J. M., Parkin, T. B., Karlen, D. L., Turco, R. F. and Konopka, A. E. 1994. Field-scale variability of soil properties in central Iowa soils. Soil Sci. Soc. Am. J. 58: 1501–1511. Chukwuma, C. 1995. Evaluating baseline data for copper, manganese, nickel and zinc in rice, yam, cassava and guinea grass from cultivated soils in Nigeria. Agr. Ecosyst. Environ. 53: 47–61. Gregorio, G. B., Senadhira, D., Htut, H. and Graham, R. D. 2000. Breeding for trace mineral density in rice. Food Nutr. Bull. 21: 382–386. Haldar, M. and Mandal, L. N. 1979. Influence of soil moisture regimes and organic matter application on the extractable Zn and Cu content in rice soils. Plant Soil. 53: 203–213. Herawati, N., Rivai, I. F., Koyama, H. and Suzuki, S. 1998. Zinc levels in rice and in soil according to the soil types of Japan, Indonesia, and China. B. Environ. Contam. Tox. 60: 402–408. Kabata-Pendias, A. and Pendias, H. 2001. Trace Elements in Soils and Plants. 3rd Edition. CRC Press Inc., Florida. Karaca, A. 2004. Effect of organic wastes on the extractability of cadmium, copper, nickel, and zinc in soil. Geoderma. 122: 297–303. Lindsay, W. L. and Norvell, W. A. 1978. Development of a DTPA soil test for zinc, iron, manganese, and copper. Soil Sci. Soc. Am. J. 42: 421–428. Liu, X. M., Xu, J. M., Zhang, M. K., Huang, J. H., Shi, J. C. and Yu, X. F. 2004. Application of geostatistics and GIS technique to characterize spatial variabilities of bioavailable micronutrients in paddy soils. Environ. Geol. 46: 189–194. Lo, K. S. L., Yang, W. F. and Lin, Y. C. 1992. Effects of organic matter on the specific adsorption of heavy metals by soil. Toxicol. Environ. Chem. 34: 139–153. Mitchell, R. L. and Burridge, J. C. 1979. Trace elements in soils and crops. Phil. Trans. R. Sco. Lond. B288: 15–24. Nelson, D. W. and Sommers, L. E. 1982. Total carbon, organic carbon, and organic matter. In Page, A. L. (ed.) Methods of Soil Analysis. Part 2. Chemical and Microbiological Properties. 2nd Edition. ASA and SSSA, Madison, WI. pp. 539–579. Norvell, W. A., Wu, J., Hopkins, D. G. and Welch, R. M. 2000. Association of cadmium in durum wheat grain with soil chloride and chelate-extractable soil cadmium. Soil Sci. Soc. Am. J. 64: 2162–2168. Nriagu, J. O. and Lin, T. S. 1995. Trace metals in wild rice sold in the United States. Sci. Total Environ. 172: 223–228. Rieuwerts, J. S., Ashmore, M. R., Farago, M. E. and Thornton, I. 2006. The influence of soil characteristics on the extractability of Cd, Pb and Zn in upland and moorland soils. Sci. Total Environ. 366: 864–875. Scrimshaw, N. S. 1991. Iron deficiency. Sci. Am. 265: 46–52. Sharma, B. D., Harsh-Arora, Raj-Kumar and Nayyar, V. K. 2004. Relationship between soil characteristics and total and DTPA-extractable micronutrients in Inceptisols of Punjab. Commun. Soil Sci. Plan. 35: 799–818. Timlin, D. J., Pachepsky, Y., Snyder, V. A. and Bryant, R. B. 1998. Spatial and temporal variability of corn grain yield on a hillslope. Soil Sci. Soc. Am. J. 62: 764–773. Webster, R. and Oliver, M. 2001. Geostatistics for Environmental Scientists. John Wiley & Sons, New York. Wilding, L. P. 1985. Spatial variability: Its documentation, accommodation and implication to soil surveys. In Nielsen, D. R. and Bouma, J. (eds.) Soil Spatial Variability. Pudoc, Wageningen, the Netherlands. pp. 166–194. Wu, J., Norvell, W. A., Hopkins, D. G. and Welch, R. M. 2002. Spatial variability of grain cadmium and soil characteristics in a durum wheat field. Soil Sci. Soc. Am. J. 66: 268–275.