Introduction

In recent years, there has been growing public concern about the contamination of soils with various materials involving potentially toxic elements. Research on soil pollution should therefore focus on the development and improvement of existing soil quality standard systems by setting limit values that take soil properties into consideration or by establishing limit values for element fractions with different solubility (Vermes 2007; Gupta et al. 1996).

In most countries, quality standards are based on the total element concentrations of the soils (e.g. after extraction with concentrated acid) (Acosta et al. 2011). However, the total soil metal content is not sufficient in itself for full-field risk assessment because the total element concentration of the soil has only a weak correlation with soil properties such as clay content, pH, organic matter and plant element uptake (Rieuwerts et al. 1998). For pragmatic purposes, Gupta et al. (1996) suggested three soil element fractions (pseudo total, mobilisable and mobile) that can be used in forming environmental protection policies. These fractions can be distinguished according to their solubility in the soil, as determined using various extraction methods. The pseudo total fraction can be determined by acid digestion (e.g. cc. HNO3 + H2O2), the mobilisable fraction by extraction with chelating agent (e.g. NH4-acetate + EDTA) and the mobile fraction by unbuffered salt (e.g. NH4NO3) extraction. The pseudo total fraction serves only for evaluating the potential risk of contamination. The mobilisable fraction shows the potentially active metal fraction in the soil. The mobile fraction may simulate plant element uptake. In the present work, extraction with 1 M NH4NO3 was used to determine the mobile fraction, an accepted method (DIN 19730 1995; Gryschko et al. 2005; Pueyo et al. 2004; Hall et al. 1998). Although this extraction method provides element concentrations that are good predictors for plant element content, low concentrations are recorded in the extract even in the case of contaminated soils (Aten and Gupta 1996; Song et al. 2004; Gryschko et al. 2005; Symeonides and McRae 1977; He and Singh 1993; Rékási and Filep 2010). Nevertheless, this method is accepted for the determination of soil quality standards. Prueβ (1997) suggested action values (intervention levels) for the soil 1 M NH4NO3-soluble element fraction, and in Baden-Württemberg, the mobile element content is used as a guide line (UM [Umweltministerium Baden-Württemberg Hrsg.] 1993).

The other problem about soil quality standards is the determination of limit values for element concentrations. The limit values are based on soil background concentrations. The background value depends on the given area and the scale of sampling. Thus, background concentrations must always be verified in any assessment of sites for contamination (Salminen and Tarvainen 1997; Fügedi et al. 2006). However, general background values are important because they can be used as guide lines in the judgement of soil contamination in a given location. Reimann and Garrett (2005) published an extensive study on this issue, concluding that ‘a statistically derived background value and associated range cannot be used alone to detect anthropogenic influences on the environment, geochemical mapping at an appropriate scale is essential’. However, statistically derived background values are good tools to put the locally measured element concentrations into context.

Hawkes and Webb (1962) suggested that background concentrations should be described by a range instead of a single value. In the present study, the general background values are given as a range covering 95% of the variance of the representative samples (Dudka 1992). The upper limit of this range can be defined as a threshold value (Garrett 1991). Concentrations above this value do not mean necessarily the anthropogenic contamination of the soil. But then, these soils should be monitored in order to avoid degradation (Micó et al. 2007). These threshold values should be always lower than the contamination limit values defined in environmental policies that values are set as the limit of the soil multifunctionality (Gupta et al. 1996).

In Hungary at the present time, the soil toxic element contamination limit is regulated only by a single pseudo total concentration value (6/2009. IV. 14. 2009), and these limit values are principally based on the regulation of other EU member countries not on Hungarian reference concentrations (Kádár 2007; Bálint 2009). The background concentrations for the pseudo total and mobilisable fractions of Hungarian soils are available from earlier studies, but the concentrations were described using arithmetic mean, median, minimum and maximum values; furthermore, data on the mobile fraction are limited (Kádár 1998; Rékási and Filep 2010). The aim of this study was to establish background values for three different soil element fractions in 193 Hungarian soils and compare them with soil quality standards and to study the geochemical and physico-chemical behaviour of the elements in each fraction.

Materials and methods

Soil samples

The soil samples were obtained from the Hungarian Soil Information and Monitoring System (TIM), which represents the soils of Hungary (Várallyay 1994). The 193 soil samples used in the present study included 159 from agricultural land and 34 from forests (Fig. 1). The sample collection comprised 39 gleysols, 30 luvisols, 27 cambisols, 26 chernozems, 15 fluvisols, 13 leptosols, 12 arenosols, 9 greyzems, 6 phaeozems, 5 podzoluvisols, 5 regosols, 4 solonetz and 2 vertisols. The samples are not supposed to be directly contaminated by any human activities.

Fig. 1
figure 1

Locations of the sampling points

Full size image

Soil analysis

Samples were taken from the top 20 cm of the soils between 15 September and 15 October in 2004. All the samples were air-dried and sieved through a 2-mm stainless steel mesh.

The following soil parameters were taken from the TIM database (measurements were carried out in the accredited laboratories of CAO Plant and Soil Protection Services): pseudo total (cc. HNO3 + H2O2-soluble) element content, mobilisable (NH4-acetate + EDTA-soluble) element content, pH (measured in KCl), soil organic matter and particle size fractions (sand, silt and clay). The mobile (1 M NH4NO3-soluble) element contents were determined in the present study.

According to Várallyay et al. (1995), the soil properties were determined by the following methods. The soil pH was measured according to the Hungarian Standard (MSZ-08-0206/2:1978) in 1:2.5 soil:1 M KCl suspension 12 h after mixing. The organic matter content was determined using the modified Walkley–Black method by digesting the soil organic matter with 5% K2Cr2O7 and cc. H2SO4 and analysing the colour of the suspension, which was related to the organic matter content of the samples, colorimetrically (MSZ-08-0452:1980). Particle size distribution was determined with the pipette method. The soil/water suspension was mixed in a sedimentation cylinder, then sampled with a pipette to collect particles of a given size (MSZ-08-0205:1978). Table 1 gives a statistical summary of selected soil properties.

Table 1 Statistical summary of selected properties for soil samples
Full size table

The determination of soil pseudo total element concentrations was carried out with the inductively coupled plasma-atomic emission spectroscopy (ICP-AES) method after digestion with cc. HNO3 + H2O2 in a microwave teflon bomb (MSZ 21470-50:2006). The mobilisable soil element fraction was measured in 0.5 M NH4-acetate + 0.02 M EDTA extract (Lakanen and Erviö 1971). The element concentrations were measured using the ICP-AES method.

For the present analysis, the mobile element fractions represented by the 1 M NH4NO3-extractable element concentrations were determined in stored soil samples from TIM. The method, outlined in the DIN 19730 (1995), is as follows: 20 g air-dried soil is shaken for 2 h with 50 cm³ 1 M NH4NO3 solution. Then, the suspension is left to stand for 15 min. After which, the supernatant is filtered through a 0.45-μm filter. The first 5 cm³ of the filtrate must be tipped off. Finally, the extract is acidified with one drop of cc. HNO3. The determination of element concentrations was conducted with the ICP-AES method (Jobin-Yvon Ultima 2 sequential instrument). For element analyses, Merck calibration standards were used. The measurement was carried out according to the manual of the instrument. In each measurement session, the extract of a standard soil sample was also measured to control the results. The calibration curves were determined after each 12th sample. Detection limit values were as follows (in micrograms per kilogram): Al (60), As (20), B (2), Cd (1), Co (2), Cr (5), Cu (12), Mn (10), Ni (9), Pb (15), Sr (1) and Zn (15).

Data processing and statistical analysis

Most of the element concentrations showed log-normal distribution (total As, B, Cd, Cu, Mn and Pb; mobilisable Al, As, B, Cd, Co, Cr, Cu, Mn, Ni, Pb and Zn; and mobile Al, B, Cd, Co, Cr, Cu, Mn, Ni, Pb and Zn). The skewness was determined with the Kolmogorov–Smirnov test. The background concentrations of log-normally distributed elements were calculated using the formulae geometric mean (GM)/geometric standard deviation2 (GSD2) and GM × GSD2, which included 95% of the sample population (Dudka 1992; Tume et al. 2006). The concentrations of elements that had approximately normal distribution were represented by arithmetic means and standard deviations (SDs). The 95% expected ranges were taken as values between the 5th and 95th percentiles. In some cases of 1 M NH4NO3 extraction, the concentration values were below the detection limits. For the calculations of the background, these concentrations were replaced with the detection limit value.

Principal component analysis (PCA) was performed for potentially toxic element concentrations using varimax rotation. For PCA, outliers were removed by estimating the z scores and eliminating those with values above 3 and below −3. The concentrations of log-normally distributed elements were transformed to logarithms (base 10) as they had positively skewed frequency distributions. Statistica v.9 (StatSoft Inc.) software was used for all the statistical evaluations. The 1 M NH4NO3-soluble As, Cr and Pb were excluded from the PCA because they were above the detection limit only in 11%, 11% and 24% of the samples, respectively. Thus, these elements could cause the undue distortion of the results.

Results and discussion

Both the arithmetic means and medians of the cc. HNO3 + H2O2-soluble element concentrations in this study are comparable with the results of survey covering the total area of the country indicating that the present dataset is suitable for the representation of Hungarian soils (Table 2). The mean pseudo total As, Co, Cu, Ni, Pb and Zn concentrations were higher than in the soils of areas under environmental protection, denoting that the soils of this study might be affected by human activities in more extent due to the fact that the most of the sampling points were on arable land. The median concentration values that are stable against outliers were lower than mean values but still higher than the mean concentrations in the soils of protected areas. The only exception was Pb that had median value as low as the mean concentration in the soils of protected areas. Nevertheless, the investigated soils can serve as a basis of background calculations since more or less the element concentrations of surface soils are affected by human activities on the whole planet and natural concentrations no longer exist (Reimann and Garrett 2005). In the case of the mobilisable fraction, the differences between the results of the present and earlier studies were more marked (Table 3). This shows that the determination of the mean concentration of more soluble element fractions for a larger area cannot be reproduced as precisely as the pseudo total concentrations. This was confirmed by the SD values of the pseudo total, mobilisable and mobile fractions (Tables 2, 3 and 4). The average SD value for the pseudo total element concentration was 64% of the mean values. This value was 86% and 140%, respectively, for mobilisable and mobile fractions. The differences in SD values may originate from the fact that the total element concentrations in the soil are mostly determined by the parent material, while the volume of the mobilisable and mobile fractions is more closely to the soil properties, especially pH and CEC (Tume et al. 2006; Pérez-Sirvent et al. 2009; Díez et al. 2009; Klassen 1998; Mühlbachová et al. 2005; Antoniadis et al. 2008; Du Laing et al. 2009). The discrepancies between the results of different surveys might be reduced by using median value for comparisons instead of mean. Unfortunately, these data of the earlier surveys were not available in this case.

Table 2 Summary statistics of pseudo total (cc. HNO3 + H2O2-soluble, in milligrams per kilogram) element concentrations of soils in this study and mean values of earlier surveys
Full size table
Table 3 Summary statistics of mobilisable (NH4-acetate + EDTA-soluble, in milligrams per kilogram) element concentrations of soils in this study and mean values of earlier surveys
Full size table
Table 4 Summary statistics of mobile (1 M NH4NO3-soluble, in micrograms per kilogram) element concentrations of the soil samples
Full size table

The order of average element abundances based on the arithmetic mean values in the different fractions was the following: pseudo total: Al > Mn > Zn > Sr > Ni > Cr > Pb > Cu > Co > B > As > Cd; mobilisable: Mn > Al > Sr > Cu > Pb > Ni > Zn > Co > B > As > Cr > Cd; and mobile: Mn > Sr > Al > Zn > Ni > Co > Cu > Pb > B > As > Cr > Cd.

Generally, the most abundant elements in each fraction were Mn and Al. This result was expected since these metals are among the most abundant elements in earth’s crust (Kabata-Pendias and Pendias 1992). In the mobilisable and mobile fractions, Sr was the third and the second in the sequence that was due to its high abundance in the lithosphere and very mobile characteristic (Cortizas et al. 2003). Cadmium had the lowest concentration in each fraction; that result is consistent with the results of previous surveys (Tables 2 and 3). Considering the positions of the elements in the abundance sequences of different fractions, it can be concluded that Cr was the least mobile of the elements investigated.

Element mobility can be calculated as the quotient of the total and mobile element fractions (Kabala and Singh 2001). Based on this ratio, the order of element mobility (where the values in brackets represent the percentage of the mobile fraction in the pseudo total element content) was: Sr (45%) > Mn (6%) > Cd (2.0%) > Co (0.9%) > Zn (0.8%) > B (0.5%) > Ni (0.4%) > As (0.3%) ≥ Cu (0.3%) > Pb (0.2%) > Al (0.06%) > Cr (0.03%).

The results supported the above findings, so Cr had the smallest mobility value, while Sr may be considered as the most mobile element. The second most mobile element was Mn; that result confirms the findings of Mahanta and Bhattacharyya (2011). The sequence above demonstrates that the mobility of Cd is relatively high even on the moderately acidic pH of the investigated soils. This may also justify the strict contamination limit value of cadmium in soils (Table 5).

Table 5 Pseudo total (cc. HNO3 + H2O2-soluble) concentration of elements in soil samples and contamination limit values (in milligrams per kilogram)
Full size table

Tables 5, 6 and 7 show the background values of the pseudo total, mobilisable and mobile fractions. In the soils studied, the pseudo total element concentration background ranges were found to be within the observed range (Table 5). The only exception was B; the highest concentration of which was lower than the upper limit of the calculated background range. According to Tume et al. (2006), these high values may suggest either the contamination of these soils or the influence of pedogenic factors.

Table 6 Mobilisable (NH4-acetate + EDTA-soluble) concentration of elements in soil samples (in milligrams per kilogram)
Full size table
Table 7 Mobile (1 M NH4NO3-soluble) concentration of elements in soil samples (in micrograms per kilogram)
Full size table

Out of the upper end of background values (threshold), that of Cd coincided with the soil quality standards currently in force in Hungary (6/2009. IV. 14. 2009). In the case of As, the threshold was above the value of the quality standard but the threshold of Co, Cr, Cu, Ni and Zn was below the contamination limit. These results show that the contamination limit value of Cd and As could be set higher but the high toxicity and mobility of these elements may explain the strict rules. However, these differences may be the outcome of the application of foreign countries’ limit values. This result may suggest the reconsideration need of the limit values of these elements. This was reinforced by the fact that the upper limit of the observed range of As, Cd, Cu and Ni was also higher than the limit values (in 8, 4, 2 and 22 samples, respectively) lay down by the order. The limit values of EU directive and the Dutch intervention values show that the quality standards could be more permissive (CEC 1986).

The concentration of the mobilisable fraction of a given element is a consequence of its pseudo total concentration and the soil properties. The upper end of the range observed for the mobilisable fractions of B, Co, Mn and Ni was below the upper limit of background concentration in the soils investigated (Table 6). In the case of Cu, the upper end of the observed range almost reached the contamination limit value for pseudo total fraction (75 mg/kg).

The suggested action value (Prueβ 1997), based on the 1 M NH4NO3-extractable Co concentration was exceeded by the upper value of both the observed range and the background (Table 7). In the case of Cr, Cu and Ni, only the upper end of the observed range was above the action value. This may indicate either the contamination of the soil samples in this study with these elements or the need to revise the suggested action values. Furthermore, this discrepancy may demonstrate also the necessity of the establishment of limit values on local reference database. However, the mean mobile concentrations (Table 4) were below the action values for each element. Only the highest observed Mn and Sr concentrations were below the upper limit of the background range.

The results show that there was no element that had lower observed concentrations in each fraction than the calculated highest background values. This may lead to suppose that the concentrations of all investigated elements were affected by human activity or pedogenic processes.

PCA has been used to identify elements with similar geochemical and physico-chemical behaviour in the three fractions. For the pseudo total element concentrations determined after cc. H2O2 + HNO3 digestion, the first four principal components (PC) had eigenvalues greater than 1, accounting for 82.3% of the total variance. Table 8 gives the rotated loadings of the PCs. The loadings of the first component were the greatest for Al, As, Cd, Cr, Cu, Ni, Sr and Zn. Co and Mn had substantial eigenvector values in the second component, while B had the highest eigenvector in the third PC, followed by Sr. The fourth principal component contained only Pb.

Table 8 Principal component analysis with varimax rotation results based on the pseudo total (cc. HNO3 + H2O2-soluble) and mobilisable (NH4-acetate + EDTA-soluble) fractions of 12 potentially toxic elements
Full size table

The interrelationships between the elements in PC1 indicate the same geological behaviour stemming from the same source. Co and Mn are also of geogenic origin and are closely associated because they both occur in Mn oxides (Prakongkep et al. 2008). B is probably also mainly of geogenic origin, but it has a different geochemical source compared to the previous groups. The fourth component may reveal the level of soil pollution; a large proportion of which can be attributed to Pb (Borůvka et al. 2005).

In the case of the mobilisable fraction, there are three principal components explaining 100% of the variation (Table 8). PC 1 contains B, Cd, Co, Mn, Ni and Zn; Al, B, Cr, Cu, Mn, Pb and Sr have the highest eigenvector values in PC2, while As, Cu, Ni and Pb made a considerable contribution to PC3. The grouping of the elements in individual principal components was much harder than for the total element concentrations because this fraction was controlled by the parent material to a lesser extent, and consequently, the NH4-acetate + EDTA-soluble concentration was mainly controlled by the physico-chemical attributes of the soils. The adsorption of metal ions on clay minerals and/or the formation of organic complexes are the main processes controlling the potentially toxic element concentrations in the soil solution (Alloway 1995). Therefore, the principal components extracted from the element concentrations of the mobilisable fraction, which comprised the element fraction in the soil solution, were influenced more by the element properties than geological behaviour. In other words, it explained a much higher proportion of the total variance of element properties than soil geochemical properties.

In the case of 1 M NH4NO3 extraction, the physico-chemical soil properties are supposed to have even more importance on element concentrations than in the case of mobilisable fraction. Three PCs explain 84.2% of the variation of mobile element concentrations (Table 9). The elements Cd, Zn, Co, Ni, Mn and Al constitute PC1. PC2 comprises B and Sr, and in PC3, only Cu can be found. These PCs reinforce that in the mobile fraction the geogenic origin has no effect on the element grouping.

Table 9 Principal component analysis with varimax rotation results of mobile (1 M NH4NO3-soluble) fraction of nine potentially toxic elements
Full size table

Based on the data of Tables 8 and 9, there are element pairs that are always in the same PC in the three fractions. These are Cd–Ni and Co–Mn. The Cd–Ni relation can have geochemical basis since both element has strong affinity for S and Cd compounds are isotopic with the compounds of Ni2+. However, Cd is concentrated in argillaceous rocks, while the Ni is concentrated in ultramafic rocks. As regards the solubility, both elements’ mobility is inversely related with soil pH. The Co–Mn connection is based on the above-mentioned geogenic origin, and their solubility is highly dependent on soil pH and redox potential (Kabata-Pendias and Pendias 1992).

Conclusions

This study established background concentrations for the pseudo total, mobilisable and mobile fractions of Hungarian surface soils. The means of pseudo total concentrations of the investigated samples were comparable with the dataset of covering the total area of Hungary. When the element concentrations on a given location are in the calculated concentration ranges, it may be assumed that the soil is not contaminated. However, the differences between observed element concentrations and the calculated background values indicated anthropogenic or pedogenic impact in each fraction.

The comparison between limit values of soil quality standards and the calculated background and observed concentrations demonstrated that the limit values of soil quality standards should be based on reference values concerning the marking area. The limit values of a certain region or country are not suitable for other areas. This was demonstrated by both the standard limits currently in force and the suggested action values for mobile element concentrations. The limit values of the Hungarian soil quality standard for potentially toxic elements could be more permissive.

The PCA of the potentially toxic elements in the three fractions revealed that there was difference in the pattern of the elements between investigated fractions. It indicates different geochemical and physico-chemical behaviour of the elements in the different fractions; the pseudo total fraction was influenced more by the geological behaviour, while mobilisable and mobile fraction explained a much higher proportion of the total variance of soil physico-chemical properties than soil geochemical properties.