Abstract
To evaluate representative elemental composition of agricultural soils in Japan and to investigate the relationship between elemental composition of agricultural soils and their soil type, land use and region in Japan, 180 soil samples were collected from the surface layer of paddy or upland fields in Japan and their total contents of 20 elements were determined. Total concentrations of aluminum (Al), iron (Fe), calcium (Ca), magnesium (Mg), titanium (Ti), phosphorus (P), manganese (Mn), barium (Ba), vanadium (V), strontium (Sr), zinc (Zn), copper (Cu) and nickel (Ni) were determined by ICP-AES and those of potassium (K) and sodium (Na) by atomic absorption spectrophotometry after wet digestion of the finely ground samples with nitric acid (HNO3), hydrofluoric acid (HF) and perchloric acid (HClO4). Total concentration of selenium (Se) was determined by HPLC after treatment of the wet digested solution (HNO3 and HClO4) of the samples with 2,3-diaminonaphthalene and extraction with cyclohexane. Total concentrations of carbon (C) and nitrogen (N) were determined by the dry combustion method, and those of silicon (Si) and oxygen (O) were determined by calculation. The medians of the total concentrations of the 10 major elements were 504 g O kg−1, 291 g Si kg−1, 76.6 g Al kg−1, 36.8 g Fe kg−1, 24.8 g C kg−1, 15.0 g K kg−1, 14.3 g Na kg−1, 11.9 g Ca kg−1, 8.78 g Mg kg−1 and 3.82 g Ti kg−1, which accounted for 98.7% of the total. The medians of the others were 2.15 g N kg−1, 1.43 g P kg−1, 705 mg Mn kg−1, 394 mg Ba kg−1, 140 mg V kg−1, 125 mg Sr kg−1, 90.5 mg Zn kg−1, 24.5 mg Cu kg−1, 14.3 mg Ni kg−1 and 0.42 mg Se kg−1. In terms of soil type, volcanic soils had relatively high Al, Fe, C, N and P contents and alluvial soils had relatively high Si, K and Ba contents, whereas red yellow soils had extremely low Ca, Mg and Na contents. In terms of land use, upland soils had significantly higher Al, Fe, C, Ca, Mg, Ti, N, P, Mn, V and Se contents and lower Si, K and Ba contents than paddy soils. Cluster analysis of elemental composition of agricultural soils grouped regions in Japan into three groups: (1) Okinawa region, (2) Hokkaido, Tohoku, Kanto, Chugoku and Kyushu regions, and 3) Hokuriku-Chubu, Kinki and Shikoku regions, reflecting their soil types. In conclusion, these data can be used as basic information on which development of sustainable agriculture and promotion of environmental conservation is to be established.
Introduction
Surface soils of agricultural fields interface directly with agricultural crops and are linked with the atmosphere, hydrosphere and lithosphere. Accordingly, they directly affect agricultural productivity and at the same time determine a portion of the terrestrial environment in Japan. In this context, therefore, information on the elemental composition of agricultural surface soils is essential not only for the evaluation of inherent soil fertility and soil quality from an agricultural point of view, but also for the understanding of the stock and flow of elements through a variety of natural and anthropogenic processes from an environmental viewpoint.
So far, there have been several reports on the elemental composition of agricultural surface soils in Japan. Kyuma and Kawaguchi (1976) investigated the concentrations of 9 major elements of 155 paddy soils and carried out soil material classification. Kato et al. (Citation2000) investigated the concentrations of 18 elements of 366 alluvial soils, mainly from paddy fields, and evaluated their relationship with the classification of cultivated soils in Japan. On the other hand, Shoji et al. (Citation1993) and Nanzyo et al. (Citation2002) investigated elemental composition of volcanic ash soils in Japan in relation to soil genesis and classification. Their findings were, however, confined to a part of soil groups rather than overall soils in Japan. Uchida et al. (Citation2007a, Citation2007b) evaluated 54 elements of 62 upland soils and 50 paddy soils, but the data were used to investigate soil-to-plant transfer factors of stable elements and naturally occurring radionuclides rather than to investigate elemental composition of soils. Takeda et al. (Citation2004) and Yamasaki et al. (Citation2001) also investigated the concentrations of 57 elements of 514 soils from 78 profiles with variable land use in Japan with reference to soil group and agricultural use, but the data were not confined to surface soils nor agricultural soils according to their experimental design. In these reports, soil samples investigated were generally limited in number, soil type, land use or especially in region, and accordingly, representative information on the elemental composition of agricultural soils in Japan has not been obtained. In addition, regional trends in the elemental composition of agricultural soils in Japan has not been elucidated, which can be compared with spatial data on elemental composition of other environmental components such as stream sediments (Imai et al. Citation2004) and river water (Kobayashi Citation1971) in Japan.
In this study, surface soils of agricultural fields were collected from all over Japan as samples to be investigated. Such a sample set would enable us to elucidate the current situation of agricultural soils in Japan with reference to their elemental composition. The objectives of this research were, therefore, 1) to evaluate representative elemental composition of agricultural soils based on sufficient data from all over Japan, and 2) to investigate the relationship between elemental composition of agricultural soils in Japan and their soil type, land use and region.
Materials and Methods
Soil samples
One hundred eighty soil samples were used, which were collected from the surface layer (0–15 cm) of agricultural fields all over Japan, i.e., from 38 prefectures from Hokkaido to Okinawa (). Ninety-six samples were collected from paddy fields and 84 samples were from upland fields. The number of the samples corresponded to about one per 270 and 250 km2 for paddy and upland fields, respectively. The soils investigated were classified into 16 soil types (Cultivated Soil Classification Committee Citation1995). The numbers of samples for each soil type, land use and region are listed in . The soil samples used in the present study were thought to be adequately representative of agricultural soils in Japan, because the average total N content of 180 samples, 2.51 g kg−1, was very close to the expected 2.62 g kg−1, which was calculated based on 2272 data points of Japanese agricultural soils by the Soil Conservation Project (Oda et al. Citation1987).
Figure 1. Location of the sampling sites. •, volcanic soils; ○, non-volcanic soils.
Table 1. Number of soil samples in relation to soil type, land use and region
Analytical methods
In this study, total concentrations of twenty elements were determined: oxygen (O), silicon (Si), aluminum (Al), iron (Fe), carbon (C), potassium (K), sodium (Na), calcium (Ca), magnesium (Mg), titanium (Ti), nitrogen (N), phosphorus (P), manganese (Mn), barium (Ba), vanadium (V), strontium (Sr), zinc (Zn), copper (Cu), nickel (Ni) and selenium (Se). These elements were selected to cover most of the major elements found in soils.
A 100-mg finely ground soil sample was decomposed with a mixture of nitric acid (HNO3), hydrofluoric acid (HF) and perchloric acid (HClO4) in a teflon beaker on a hot plate, and heated to strong fumes of perchloric acid (about 200°C). To the resultant solution, 1 mL of concentrated HNO3 was added, and the solution was filled up to 50 mL with deionized water. In the wet digestion, all the digestion was carried out in duplicate and one standard soil sample was incorporated in each lot. The concentrations of Al, Fe, Ca, Mg, Ti, P, Mn, Ba, V, Sr, Zn, Cu and Ni in the prepared solution were determined by inductively coupled plasma atomic emission spectrophotometry (ICP-AES, SPS-1500VR; Seiko Instuments, Chiba, Japan) and those of K and Na were determined by atomic absorption spectrophotometry (AA-6200; Shimadzu, Kyoto, Japan).
A 50-mg finely ground soil sample was decomposed with HNO3 and HClO4 and the released Se was reduced to Se(IV) by heating with 2 mL of 6 mol L−1 hydrochloric acid (HCl). The concentration of Se was then determined by high-performance liquid chromatography with a fluorescence detector after treatment with 2,3-diaminonaphthalene and extraction with cyclohexane, as reported by Yamada et al. (Citation2009).
The concentrations of total C and N were determined by the dry combustion method (Sumigraph NC analyzer NC-800, Sumika Chem. Anal. Service, Osaka, Japan) using about 50–100 mg of finely ground soil samples, as reported partly by Sano et al. (Citation2004).
The concentration of Si was calculated as explained by Kato et al. (Citation2000). The SiO2 concentration was calculated by subtracting the sum of contents of Al2O3, Fe2O3, K2O, Na2O, CaO, MgO, TiO2, P2O5 and MnO, total amounts of microelements (Ba, V, Sr, Zn, Cu and Ni) and loss of ignition from the total. The Si concentration was then converted from that of SiO2. Preliminary experiments, i.e., direct measurement of the total Si concentration for 6 samples by fusing with lithium metaborate at 1000°C followed by determination of Si with ICP-AES (Goto et al. Citation1991), indicated that the differences between calculated and measured values were within 2.5% on average, suggesting the reliability of the calculation for the estimation of Si content in soil.
Finally, the concentration of O was calculated by subtracting total concentrations of the 19 elements mentioned above from the total. This calculated value was regarded as the first approximation of the total O concentration in soils, because all the experimental errors and the concentrations of the elements not determined in this experiment such as H and S are included as errors in this calculation, even though these effects were thought to be relatively insignificant compared with the O concentration.
Statistical analysis
Descriptive statistics were calculated for the overall dataset of the total concentrations of 20 elements in soil. For the overall dataset, analysis of variance was carried out with reference to soil type and region, and t-test was carried out with reference to land use. Cluster analysis was further carried out by hierarchical clustering by single linkage method based on Pearson clustering distance to investigate the similarity of elemental composition of soils among regions.
Results and Discussion
Descriptive statistics of the elemental concentrations of agricultural soils in Japan
shows the descriptive statistics, i.e., median, arithmetic mean, geometric mean, minimum and maximum of the total concentrations of the 20 elements. The medians of the total concentrations of the 10 major elements were 504 g O kg−1, 291 g Si kg−1, 76.6 g Al kg−1, 36.8 g Fe kg−1, 24.8 g C kg−1, 15.0 g K kg−1, 14.3 g Na kg−1, 11.9 g Ca kg−1, 8.78 g Mg kg−1 and 3.82 g Ti kg−1, which accounted for 98.7% of the total. The medians of the other elements were 2.15 g N kg−1, 1.43 g P kg−1, 705 mg Mn kg−1, 394 mg Ba kg−1, 140 mg V kg−1, 125 mg Sr kg−1, 90.5 mg Zn kg−1, 24.5 mg Cu kg−1, 14.3 mg Ni kg−1 and 0.42 mg Se kg−1. The histograms indicated that the distributions of O, Si, Al, K, Na, Mg and Ba were similar to normal distributions and accordingly medians and arithmetic means were regarded to be representative values of the dataset. The distributions of Fe, C, Ca, Ti, N, P, Mn, V, Sr, Zn, Cu, Ni and Se were, on the contrary, similar to the log-normal distribution and accordingly medians and geometric means were regarded to be representative values of the dataset. These distribution patterns of the data were thought to reflect the characteristics of the original elemental composition of the parent materials as well as mobility or durability of the elements during soil formation.
Table 2. Median, arithmetic mean, geometric mean, minimum and maximum of the elemental concentrations of agricultural soils in Japan
Table 3. Medians of the elemental concentrations in soils
The medians of the elements obtained here were compared with the reported values for the elemental composition of the soils in Japan (Takeda et al. Citation2004) and soils of the world (Bowen Citation1979), as shown in . In general, the data of this study were within the ordinary range of those observed for Japanese soils and for worldwide soils already reported. There was a tendency, however, that the concentrations of C, K, Na, N and P were relatively higher and those of Fe, Ti, Cu and Ni were relatively lower than the reported values. This would be partly due to the fact that our samples were only surface soils under agricultural activities, whereas samples by Takeda et al. (Citation2004) were from the soil profiles including surface and subsurface soils. Judging from the fact that soil sampling in our study was carried out intensively from plow layers of agricultural fields all over Japan, reflecting relative areas of paddy and upland fields, it would be reasonable to conclude that the median values reported here can be regarded as the first approximation of the representative data of elemental composition of agricultural soils in Japan.
Elemental concentrations of agricultural soils with reference to soil type
shows arithmetic means of the elemental concentrations of agricultural soils in Japan with reference to soil type. Arithmetic means were used instead of medians hereinafter because the former were thought to be more reliable if the number of samples for each group was rather restricted when samples were classified into many groups.
Table 4. Arithmetic mean of the elemental concentrations of agricultural soils in Japan with reference to soil type
Table 5. Arithmetic mean of the elemental concentrations of agricultural soils in Japan with reference to land use (all the samples)
Soil type had a significant effect on the concentration of all the elements investigated except for Zn and Cu, reflecting both the parent materials and degree of weathering of the soils investigated. Peat soils tended to have relatively high C and N concentrations and relatively low Al concentrations, directly reflecting their high organic matter content. Sand-dune Regosols had highest concentrations of Si, K, Na and Sr and lowest concentrations of Al, Fe, Ti, V, Zn and Se. Volcanogeneous Regosols had highest concentrations of Ca, Mg and V and lowest concentrations of O, K and Ba. Wet Andosols, Non-allophanic Andosols and Andosols tended to have relatively high Al, Fe, C, N, P, Cu and Se concentrations and relatively low K and Ba concentrations. These would be due to the mafic nature of the volcanic materials from which these soils were derived. Lowland Paddy soils, Gley Lowland soils, Gray Lowland soils and Brown Lowland soils had relatively higher Si, K and Ba concentrations and relatively lower Fe and Mn concentrations with moderate concentrations of other elements. These would reflect the abundance of light minerals such as feldspar and plagioclase in alluvial deposits. Terrestrial Regosols (Jahgaru in Okinawa) tended to have higher K, Ca, Mg and Ba concentrations and lower C, Na, N and P concentrations. Dark Red soils (Shimajiri-Mahji in Okinawa) had relatively high O, Al, Fe, Ti, Cu and Ni concentrations and relatively low Na, Ca and Mg concentrations, partly reflecting their parent material (limestone) in addition to degree of weathering, and Red soils and Yellow soils had relatively high O, Al, Fe, Ti, Cu and Ni concentrations and extremely low Na, Ca and Mg concentrations, reflecting their high degree of weathering. Brown Forest soils tended to have high O, Al, Fe and Ti concentrations and low Na, Ca, Mg and Sr concentrations. Similar tendencies among soil groups were reported by Takeda et al. (Citation2004), in which comparison of Andosols, Cambisols, Gleysols and Acrisols indicated that “Andosols had relatively higher contents of mafic elements, which might be reflected by the composition of parent rock type, and Acrisols had lower concentrations of many other elements, especially Na, Mg, Ca, and Sr, possibly because of their heavy weathered nature.” Such a trend was thought to regulate the inherent fertility status of the soils investigated.
Elemental concentrations of agricultural soils with reference to land use
Arithmetic means of the elemental concentrations of agricultural soils in Japan with reference to land use are shown in . The t-test indicated that paddy soils had significantly higher Si, K and Ba concentrations whereas upland soils had significantly higher Al, Fe, C, Ca, Mg, Ti, N, P, Mn, V and Se concentrations. On the contrary, comparison of 25 pairs of paddy and upland soil samples collected from similar locations with almost identical soil types (from Hokkaido to Okinawa) indicated that there was no statistical difference between paddy and upland soils for any element (). It can be concluded from these findings that the difference between paddy and upland soils based on 180 data points were mainly due to the difference of soil types composing paddy and upland soils () rather than due to the difference of management between paddy and upland soils. In conclusion, judging from soil fertility status for major essential elements, paddy soils tended to have high inherent fertility for K and upland soils tended to have high inherent fertility for N, P, Ca and Mg.
Table 6. Arithmetic mean of the elemental concentrations of agricultural soils in Japan with reference to land use (paired samples)
Elemental concentrations of agricultural soils with reference to region
shows arithmetic means of the elemental concentrations of agricultural soils in Japan with reference to region. Region had a significant effect on the concentration of all the elements investigated (p < 0.01). Hokkaido, Kanto and Kyushu regions tended to have high Al, Fe, C, Na, Ca, Mg, Ti, N, P, Mn, V, Sr and Se concentrations and low Si, K and Ba concentrations. Tohoku and Chugoku regions had similar but slightly weaker tendencies compared to Hokkaido, Kanto and Kyushu regions. On the contrary, Hokuriku-Chubu, Kinki and Shikoku regions tended to have high Si, K and Ba concentrations and low Al, Fe, C, Ca, Mg, N, P, Mn, V, Sr and Se concentrations. Okinawa region had an extremely different elemental composition from those of other regions: it had highest O, Al, Ti and Ni concentrations and lowest C, Na, Ca and N concentrations. Similar regional trends were reported for alluvial soils (Kato et al. Citation2000) and for stream sediments along rivers (Imai et al. Citation2004) in Japan. In the case of river water, however, a similar trend was observed for Ca, but opposite trends were observed for Si and K (Kobayashi Citation1971), probably reflecting the affinity of the elements for clay minerals in soil, controlling their leaching from soil to hydrosphere. Such a relationship should be investigated in more detail to understand the dynamic flow of elements in terrestrial ecosystems in Japan.
Table 7. Arithmetic mean of the elemental concentrations of agricultural soils in Japan with reference to region
presents a dendrogram showing the relationship among regions in Japan with respect to elemental composition of soils. It directly indicates, first, that the Okinawa region was far different from the others, and, second, the others were clustered into two groups, i.e., Hokuriku-Chubu, Kinki and Shikoku regions and Kanto, Kyusyu, Hokkaido, Tohoku and Chugoku regions. It should be noted that the last group was further clustered into two subgroups: i.e., Kanto, Kyushu and Hokkaido subgroup, and Tohoku and Chugoku subgroup.
Figure 2. Dendrogram showing the relationship among regions in Japan with respect to elemental composition of soils.
The fact that the Okinawa region was considerably different from the others would be mainly due to the high degree of weathering and accelerated decomposition of organic matter derived from its subtropical climate with high temperature and heavy rainfall, and partly due to the characteristic parent material (limestone, etc.) of some samples. The fact that Hokuriku-Chubu, Kinki and Shikoku regions were clustered from the others can be ascribed to the predominance of acidic rocks as parent materials (Kato et al. Citation2000) and the very weak effect of volcanic deposition in these regions (). The last group, i.e., Hokkaido, Tohoku, Kanto, Chugoku and Kyushu regions, would be characterized by the relative abundance of Andosols in these regions (). Subgrouping into Kanto, Kyushu and Hokkaido subgroup and Tohoku and Chugoku subgroup would correspond to the predominance of allophanic Andosols in the former and the predominance of non-allophanic Andosols in the latter, as shown in (Saigusa and Matsuyama Citation1998). In conclusion, agricultural soils in Japan were grouped into 3 groups: Okinawa region, Hokkaido, Tohoku, Kanto, Chugoku and Kyushu regions, and Hokuriku-Chubu, Kinki and Shikoku regions, according to the similarity of their elemental composition. In other words, the inherent fertility status of agricultural soils in Japan showed considerable regional variability, which should be taken into account for rational management of the soils.
Figure 3. Distribution of allophanic Andosols and non-allophanic Andosols in Japan (Saigusa and Matsuyama, 1998).
Conclusion
In this study, the elemental composition of agricultural soils in Japan was evaluated based on 180 soil samples collected from the surface layer of paddy or upland fields all over Japan. The medians of the total concentrations of 20 elements, i.e., O, Si, Al, Fe, C, K, Na, Ca, Mg, Ti, N, P, Mn, Ba, V, Sr, Zn, Cu, Ni and Se, were determined as the representative elemental composition of agricultural soils in Japan. Investigation of the relationship between elemental composition of soils and soil type, land use, and region suggested that elemental composition of the agricultural soils in Japan was mainly related to soil type, which reflected both parent material and degree of weathering of the soils. These findings can be used as basic information on which development of sustainable agriculture and promotion of environmental conservation should be established. Further investigation is also required to determine (1) total analysis of trace and ultra-trace elements which were not analyzed in this study, and (2) chemical speciation and/or availability evaluation of the elements in soils, for the same soil samples used in this research.
Acknowledgments
The authors wish to thank Dr. Shuji Sano, Research Institute of Environment, Agriculture and Fisheries, Osaka Prefecture, for his collaboration in collecting soil samples and for valuable information on the soil samples investigated, and Prof. Masanori Saigusa, Toyohashi University of Technology, for his valuable information on the distribution of allophanic and non-allophanic Andosols in Japan. The authors also wish to thank staff members of many agricultural experimental stations and universities, and many farmers all over Japan, for supplying soil samples and field information.
Related Research Data
References
- Bowen , HJM . (Ed.) 1979: Environmental Chemistry of the Elements. Academic Press Inc. Ltd., London
- Cultivated Soil Classification Committee 1995: Classification of Cultivated Soils in Japan (3rd Approximation). Miscellaneous publication of the National Institute of Agro-Environmental Sciences, No. 17, Tsukuba, Japan
- Goto , I , Muramoto , J and Ninaki , M . 1991 . Application of inductively coupled plasma atomic emission spectrometry (ICP-AES) to soil analysis (Part 3). Total analysis of major elements in soils by lithium metaborate fusion ICP-AES . Jpn. J. Soil Sci. Plant Nutr. , 62 : 521 – 528 . (in Japanese with English summary)
- Imai , N , Terashima , S and Ohta , A . et al. 2004: Geological Map of Japan. Geological Survey of Japan, AIST (in Japanese with English summary)
- Kato , K , Obara , H , Nakai , M and Higashi , T . 2000 . Elemental composition of recent alluvial soils in Japan – regional differences and relation to classification of cultivated soils in Japan . Jpn. J. Soil Sci. Plant Nutr. , 71 : 142 – 153 . (in Japanese with English summary)
- Kobayashi , J . 1971 . Health Examination of Water , Tokyo , , Japan : Iwanami-Shoten .
- Kyuma , K and Kawaguch , K . 1976 . Soil material classification for paddy soils in Japan . Soil Sci. Plant Nutr. , 22 : 111 – 124 .
- Nanzyo , M , Yamasaki , S and Honna , T . 2002 . Changes in content of trace and ultratrace elements with an increase in noncristalline materials in volcanic ash soils of Japan . Clay Sci. , 12 : 25 – 32 .
- Oda , K , Miwa , E and Iwamoto , A . 1987 . Compact data base for soil analysis data in Japan . Jpn. J. Soil Sci. Plant Nutr. , 58 : 112 – 131 . (in Japanese)
- Saigusa , M and Matsuyama , N . 1998 . Distribution of allophonic Andosols and non-allophanic Andosols in Japan . Tohoku J. Agric. Res. , 48 : 75 – 83 .
- Sano , S , Yanai , J and Kosaki , T . 2004 . Evaluation of soil nitrogen status in Japanese agricultural lands with reference to land use and soil types . Soil Sci. Plant Nutr. , 50 : 501 – 510 .
- Shoji , S , Nanzyo , M and Dahlgren , RA . 1993 . Volcanic Ash Soils: Genesis, Properties and Utilization, Developments in Soil Science 21 , Amsterdam : Elsevier .
- Takeda , A , Kimura , K and Yamasaki , S . 2004 . Analysis of 57 elements in Japanese soils, with special reference to soil group and agricultural use . Geoderma , 119 : 291 – 307 .
- Uchida , S , Tagami , K and Hirai , I . 2007a . Soil-to-plant transfer factors of stable elements and naturally occurring radionuclides: (1) Upland field crops collected in Japan . J. Nucl. Sci. Technol. , 44 : 628 – 640 .
- Uchida , S , Tagami , K and Hirai , I . 2007b . Soil-to-plant transfer factors of stable elements and naturally occurring radionuclides: (2) Rice collected in Japan . J. Nucl. Sci. Technol. , 44 : 779 – 790 .
- Yamada , H , Kamada , A , Usuki , M and Yanai , J . 2009 . Total selenium content of agricultural soils in Japan . Soil Sci. Plant Nutr. , 55 : 616 – 622 .
- Yamasaki , S , Takeda , A , Nanzyo , M , Taniyama , I and Nakai , M . 2001 . Background levels of trace and ultra-trace elements in soils of Japan . Soil Sci. Plant Nutr. , 47 : 755 – 765 .