Abstract

Cation exchange capacity (CEC) is one of the most important properties of soils. The NH₄OAc (pH = 7.0) exchange method is usually recommended to determine CEC (CEC₁) of all soils with different pH values, particularly for studies on soil taxonomy. But comparatively the BaCl₂-MgSO₄ forced-exchange method is more authentic in determining CEC (CEC₂) of tropical and subtropical highly-weathered acid soils. But so far little is known about the difference between CEC₁ and CEC₂. In this study, the physiochemical data of 114 acid B horizon soils from 112 soil series of tropical and subtropical China were used, CEC₁ and CEC₂ were determined and compared, the influencing factors were analyzed for the difference between CEC₁ and CEC₂, and then a regression model was established between CEC₁ and CEC₂. The results showed that CEC₂ was significantly lower than CEC₁ (p < 0.01), CEC₂ was 14.76% - 63.31% with a mean of 36.32% of CEC₁. In view of the contribution to CEC from other properties, CEC₂ was mainly determined by pH (45.92%), followed by silt (21.05%), free Fe₂O₃ (17.35%) and clay contents (12.76%), CEC₁ was mainly decided by free Fe₂O₃ content (40.38%), followed by pH (28.39%) and silt content (27.29%; and the difference between CEC₁ and CEC₂ was mainly affected by free Fe₂O₃ (50.92%), followed by silt content (26.46%) and pH (21.80%). The acceptable optimal regression model between CEC₂ and CEC₁ was established as CEC₂ = 2.3114 × CEC₁^1.1496 (R² = 0.410, P < 0.001, RMSE = 0.15). For the studies on soil taxonomy, the BaCl₂-MgSO₄ forced-exchange method is recommended in determining CEC of the highly-weathered acid soils in the tropical and subtropical regions.

Keywords

Acid Soil, CEC Determination, NH₄OAc (pH = 7.0) Exchange Method, BaCl₂-MgSO₄ Forced-Exchange Method

Share and Cite:

1. Introduction

Soil cation exchange capacity (CEC) is one of the most important chemical characteristics of agricultural lands [1], which can influence the stability of soil structure, nutrient availability, soil pH and the soil’s reaction to fertilizers and other ameliorants, provide a buffer against soil acidification [2]. CEC is often used as a measure of soil fertility, nutrient retention capacity [3], and also used as an identification and classification index of soil types in soil taxonomy [4] [5], in which the NH₄OAc (pH = 7.0) exchange method [6] [7] is recommended to determine CEC for all soils with different pH values. However, for highly-weathered acid soils in the tropical and subtropical regions, the BaCl₂-MgSO₄ forced-exchange method [8], which doesn’t adjust pH of soil samples, is recommended to determining CEC. Comparatively, because the buffer salt system (pH = 7.0) in the first method will increase soil pH, thus will increase the charge of soil colloids and result in higher measurement results [9] [10], which may lead to the misjudgment of soil types [11].

But so far, little is known about the difference in CEC values determined by the two methods, thus, in this study the physiochemical data of 114 acid B horizon soils from 112 soil series in the tropical and subtropical regions of south China were used to: 1) disclose the difference in CEC values determined by the two methods, 2) clarify the influencing factors of the difference, and 3) setup the regression model for predicting CEC₂ by CEC₁.

2. Materials and Methods

2.1. Background of Tested Soil Samples

Figure 1 shows the spatial distribution of used 112 soil series in the tropical and subtropical regions of south China [12] - [22]. For a soil sample, the particle size distribution was determined by the pipette method, pH was measured with by the potentiometer method (soil:water = 1:2.5), organic matter was obtained by the Walkley-Black wet oxidation method, free Fe₂O₃ was determined by the phenanthroline colorimetry method, CEC was analyzed by the NH₄OAc (pH = 7.0) exchange method (CEC₁) [6] [7] and the BaCl₂-MgSO₄ forced-exchange method (CEC₂) [8], respectively.

2.2. Data Statistical Analysis

Microsoft Excel 2016 and IBM Statistics SPSS 22.0 software were used for statistical analysis of the data, and Duncan test method (2-tailed) was used for variance analyses and multiple comparisons.

3. Results

3.1. Statistical Results of Soil Physiochemical Properties

Table 1 lists the measured values of soil physiochemical properties, it showed that CEC₁ ranged from 5.12 to 35.41 cmol(+) kg⁻¹ with a mean of 12.40 cmol(+) kg⁻¹, while CEC₂ ranged from 2.22 to 6.60 cmol(+) kg⁻¹ with a mean of 4.16 cmol(+) kg⁻¹. Comparatively, CEC₂ was significantly lower than CEC₁ (p < 0.01), CEC₂ was 14.76% - 63.31% with a mean of 36.32% of CEC₁.

Table 1 also showed that clay content was meanly 412 g·kg⁻¹, while sand content was meanly 281 g·kg⁻¹; meanwhile, free Fe₂O₃ content was meanly 44.01 g·kg⁻¹, which prove further that soils in the tropical and subtropical regions of south China are clayey and rich in free Fe₂O₃ [23].

3.2. Factors Influencing CEC₁, CEC₂ and Their Difference

Table 2 lists the correlation between CEC₁, CEC₂ and the difference between CEC₁ and CEC₂ (ΔCEC, CEC₁-CEC₂) with other properties. It could be found that pH had significant positive correlation with CEC₁ (p < 0.01), CEC₂ (p < 0.01) and ΔCEC (p < 0.05), free Fe₂O₃ had significant positive correlation with CEC₁ and ΔCEC (p < 0.01), sand content had significant negative correlation with CEC₁ and ΔCEC (p < 0.05), silt content had significant positive correlation with CEC₁ (p < 0.05) and CEC₂(p < 0.01), while clay content had significant negative correlation with CEC₂ (p < 0.05).

Note: 1) Sand, silt, clay, SOM and free Fe₂O₃, g·kg⁻¹; CEC₁ and CEC₂, cmol(+) kg⁻¹; 2) CEC₁ and CEC₂, determined by the methods of NH₄OAc (pH = 7.0) and BaCl₂-MgSO₄, respectively. The same below; 3) data of CEC₁ and CEC₂ followed by different capitals are significantly different at p < 0.01 level.

Note: 1) *, **, Correlation is significant at p < 0.05 or 0.01 level (2-tailed)l; 2) ΔCEC = CEC₁ − CEC₂.

The contribution of one property to CEC was calculated as the follows: firstly, all properties were normalized by the Z-score method with IBM Statistics SPSS 20.0 to ensure them with the same magnitude, and then the regression coefficients between each property with CEC was used to indicate their contribution to CEC [24] [25] [26]. The contribution of one property (C_i) to CEC was calculated as C_i = K_i/K_sum, in which K_i is the regression coefficient of the i property, and K_sum is the total sum of all coefficients, the obtained linear regression models of CEC with other properties were listed in Table 3, and the calculated contribution of other properties to CEC were listed in Table 4.

In view of the contribution of other properties to CEC, it can be seen from Table 4 that CEC₁ was mainly decided by free Fe₂O₃ (40.38%), followed by pH and silt content (28.39% and 27.29%, respectively); CEC₂ was mainly determined by pH (45.92%), followed by silt content (21.05%), then followed by free Fe₂O₃ and clay content (17.35% and 12.76%, respectively), and ΔCEC was mainly affected by free Fe₂O₃ (50.92%), followed by silt content and pH (26.46% and 21.80%, respectively).

3.3. CEC₂ Predicting Model Based on CEC₁

The scatter diagram of CEC₂ and CEC₁ are shown in Figure 2, and IBM statistics SPSS 20.0 was used to obtain the optimal regression model between CEC₂ and CEC₁. It could be found from Figure 2 that a significant positive power correlation between CEC₂ and CEC₁, and the optimal regression model was as CEC₂ = 2.3114 × ${\text{CEC}}_1^{1.1496}$ (R² = 0.410**, P < 0.001 F = 77.99, RMSE = 0.15, RMSE/S.D = 0.19).

4. Discussions

4.1. Value Difference CEC Determined by Different Methods

For highly-weathered acid soils in the subtropical and tropical regions, because the buffer salt system (pH = 7.0) could increase soil pH, thus would increase the charge of soil colloids, so CEC determined by the NH₄OAc (pH = 7.0) exchange method (CEC₁) usually is higher than that determined by the BaCl₂-MgSO₄ forced-exchange method (CEC₂) [9] [10]. Our study quantitatively assessed this phenomenon, for the acid B horizon soils in the subtropical and tropical regions of south China, CEC₂ was significantly lower (P < 0.01) than CEC₁, the former meanly 36.32% of the latter (see Table 1).

Our study also disclosed the differences in the influencing factors of CEC₁ and CEC₂, in which pH and silt content were the common factors of CEC₁ and CEC₂, but CEC₁ was also influenced by free Fe₂O₃ and sand content, while CEC₂ was also affected by clay content (see Table 2). Furthermore, our study proved further that the difference between CEC₁ and CEC₂ was mainly decided by free Fe₂O₃ content (the contribution was 50.92%, see Table 4), followed by silt content and pH (the contributions were 26.46% and 21.80%, respectively, see Table 4), while little or no effect from sand and clay contents.

4.2. Influencing Factors of CEC

Table 5 lists the correlation between CEC and other properties of soils found in

some previous studies. pH usually has significant negative correlation with CEC for soils with high pH (for example, higher than 7.0) [27] [28] [29] [30] but has positive correlation with CEC for soils with low pH (for example, lower than 7.0) [31] [32] [33] [34] [35]. Since all soil samples used in our study were acid (pH < 7.0), so significant positive correlation was found in our study between pH and CEC₁ and CEC₂.

SOM usually has significant positive correlation with CEC [27] [28] [30] - [40], but our results showed that SOM had no significant correlation with CEC₁ and CEC₂ (Pearson correlation coefficient was 0.069 and 0.001, respectively, See Table 2; contribution to CEC was 1.42% and 2.93%, respectively, see Table 4), which could be attributed to the low SOM content [28] [37] [38] [41] in B horizon soils in the subtropical and tropical regions of south China (mean SOM content was 8.24 g·kg⁻¹ in our study).

Clay content usually also has significant positive correlation with CEC of humid soils [28] - [40], but our results showed that clay had no significant correlation with CEC₁ (R was 0.060, see Table 2; contribution to CEC was 2.52%, see Table 4) and had weak negative significant correlation with CEC₂ (R was 0.220, p < 0.05, see Table 2; contribution to CEC was 12.76%, see Table 4), which could be attributed to greater microaggregating effect of Fe oxides in highly-weathered soils in the tropical and subtropical regions [42], which enhanced the participation of clay in the microaggregation, reduced the amount of “free” clay particles, thus decreased clay contribution to CEC [40]. Few studies analyzed the correlation between free Fe₂O₃ and CEC because free Fe₂O₃ in subtropical and tropical highly-weathered soils usually exist as clay fraction or strongly cemented with clays [42] [43] [44], so more attentions were paid to the correlation between clay content rather than free Fe₂O₃ with CEC (p < 0.01). However, our studies found that free Fe₂O₃ was significantly correlated with CEC₁, while clay content was significantly correlated with CEC₂ (p < 0.05).

Our study also found that CEC₁ had negative correlation with sand content, which is consist with the previous studies [29] [31] [38] [40] [41], while CEC₂ had significant positive correlation with silt content as found in other studies [32] [34], which could be attributed to that in subtropical and tropical humid climate soils, sand fraction is mainly composed of quartz and iron concretions which present low charge density [45], while the silt fraction is often composed of vermiculite and mica minerals which can hold negative charges [46].

4.3. Recommendation Using CEC₂ Predicting Model for Soil Taxonomy

In Chinese Soil Taxonomy, the LAC-ferric horizon is the diagnostic horizon for Ferrosols, one of its requirements is that CEC₇ < 24 cmol (+) kg⁻¹ clay in partial B horizons (≥10 cm in thickness) [4]. However, CEC_7clay is not directly measured by the extracted clays, it was calculated as: soil CEC₇ × 1000/clay content [4]. Our study shows that for B horizons of the highly-weathered acid soils in the tropical and subtropical regions of south China, CEC determined by the NH₄OAc (pH = 7.0) exchange method is 1.58 - 6.78 times with a mean of 2.96 times of that decided by the BaCl₂-MgSO₄ forced-exchange method. This obvious overestimation of CEC [9] is most likely to lead to some authentic LAC-ferric horizons being misjudged as other diagnostic horizons, thus leading to misjudgment of soil types [10]. However, since the NH₄OAc (pH = 7.0) exchange method was used in almost all previous studies on soil taxonomy, thus, to verify the identification accuracy of soil types in the previous studies, the CEC₂ predicting model established in our study based on CEC₁ is recommended to obtain CEC of highly-weathered acid soils in the tropical and subtropical regions in order to ensure the accurate identification of soil types. Nevertheless, for the future studies, it is recommended to using the BaCl₂-MgSO₄ forced-exchange method for CEC determination of the highly-weathered acid soils in the tropical and subtropical regions.

5. Conclusion

Our study quantitatively proved that for the highly-weathered acid soils in the tropical and subtropical regions of south China, CEC determined by the NH₄OAc (pH = 7.0) exchange method was significantly higher than that determined by the BaCl₂-MgSO₄ forced-exchange method. CEC of the former method was mainly affected by free Fe₂O₃ and pH, followed by silt and sand contents, while CEC of the latter method was mainly affected by pH, followed by silt and clay contents. CEC differences between the two methods were mainly influenced by free Fe₂O₃, followed by sand content and pH. For the studies on soil taxonomy, the BaCl₂-MgSO₄ forced-exchange method is recommended for CEC determination of the highly-weathered acid soils in the tropical and subtropical regions.

Acknowledgements

This study was supported by projects of the National Natural Science Foundation of China (No. 41877008) and the National S&T Basic Special Foundation Project (No. 2014FY110200). We would like to express thanks to the contribution of all colleagues in the data preparation and the establishment of the soil series.

Conflicts of Interest

The authors declare no conflicts of interest regarding the publication of this paper.

References