Estimation of cation-exchange capacity in routine soil testing

The efficiency of the soil testing method used in Finland forpredicting the effective cation-exchangecapacity was studied in a material of 430 topsoil samples. The effective cationexchange capacity was estimated 1)by summation of exchangeableCa, Mg and acidity displaced by unbuffered 1 M KCI and 2) by summation of exchangeable Ca, Mg, K and Na displaced by neutral 1 M ammonium acetate and exchangeable acidity. In soil testing, Ca, Mg and K were extracted by acid ammonium acetate and soil pH measured in water-suspension. The estimates of the effective CEC were highly correlated and dependenton the clay and organic carbon content and pH(CaCl 2) of the soil, the coefficient of multiple determination being over 80 Vo. Exchangeable Ca was the dominating cation. The proportion of Ca of the effective CEC was about 80 Vo. Acid ammonium acetate-extractable Ca together with pH(H 20) explained over 80 % of the variation in the effective CEC. For the whole material consisting of mineral soils with great variations in texture, organic carbon content and properties under evaluation, the regression equation predicting the effective CEC (KCI method) was CEC(mvalZkg) = 309—56.8pH(H2 0) + o.oBsCa(mg/l). Only 16 % of the estimates of the effective CEC calculated with this regression equation deviated more than 15 % from the measured values. Index words: exchangeable Ca, Mg, K, Na, exchangeableacidity, effective cation-exchange capacity, soil pH, soil


Introduction
In Finnish soil testing, lime requirement has been estimated on the basis of pH(H 2 0) and acid ammonium acetate-extractable Ca.Ac- cording to Mäntylahti and Yläranta(l9Bo), however, soil pH alone gives a better estimate.
The content of extractable Ca in soil cannot be used as an index of liming requirements as long as the cation-exchange capacity and base saturation vary.On the other hand, exchangeable calcium contents together with acidity could estimate cation-exchange capacity of soil.The aim of this study was to examine the relationship between estimates of effective cation-exchange capacity of mineral soils and the possibility of predicting them by soil testing.

Material and methods
The material consisted of 430 plough layer (0-25 cm) samples from the agricultural area of the Viikki Experimental Farm.The soil characteristics have been described previously by Jokinen (1983, 1984) and Niskanen and  Jaakkola (1985).The samples were air-dried and ground to pass a 2-mm sieve.On the ba- sis of the particle-size distribution determined by the pipette method (Elonen 1971), the material consisted of 230 clay soils with a clay content > 30 % and 200 coarser soil samples with a clay content < 30 % (Table 1).The pH of the soil was measured in a soil-0.01M CaCl 2 suspension (1:2.5)(Ryti 1965).The mean pH of the coarser soils was 0.3 pH units higher than that of clay soils (Table 1).The organic carbon content of the soil determined by a modified (Graham 1948) Alten's wet combustion method was on the average 1.5 % higher in clay soils (Table 1).
Exchangeable Ca and Mg were extracted from 10 g soil by four successive treatments with 50 ml of neutral 1 M ammonium acetate and 50 ml 1 M KCI and determined by atom- ic absorption spectrophotometry.Exchangeable K and Na were extracted with neutral 1 M ammonium acetate and determined by flame photometry.
The exchangeable acidity was displaced with 1 M KCI and titrated with 0.01 M NaOH (Yuan 1959).Two estimates for the effective cation-exchange capacity (CEC) were deter- mined according to Kaila (1971 a, 1972).The effective CECI was estimated by summation of exchangeable Ca, Mg and acidity displaced by 1 M KCI.The effective CEC2 was esti- mated by summation of exchangeable Ca, Mg, Na and K extracted by neutral 1 M am- monium acetate with exchangeable acidity displaced by unbuffered 1 M KCI.The pH of the soil-H 2 0 suspension (1:2.5 v/v) and Ca, Mg and K extracted with acid ammonium ace- tate (0.5 M acetic acid, 0.5 M ammonium ace- tate, pH 4.65, ratio 1:10 v/v) (Vuorinen and  Mäkitie 1955) were determined at a commer- cial soil testing laboratory (Viljavuuspalvelu Oy).
The relationship between exchangeable acidity and soil pH was curvilinear, but the logarithmic values of exchangeable acidity were linearly correlated with soil pH.With pH(H 2 0) as an independent variable, the re- gression equation for the whole material (n = 430) was log(acidity) = 4.66-0.68pH(H 2 0), R 2 = 0.83.The dependence of the effective CECI on soil properties was studied using clay and silt content (%), organic carbon content (*Vo) and pH(CaCl 2 ) as independent variables in the regression analysis.Clay content, organic car- bon content and pH(CaCl 2 ) together ex- plained 83.4 % of the variation in the effec- tive CECI of the whole material, the regression equation being CECI (mval/kg) = -227 + I.Bsclay-% + 45.6 pH + 11.20rg.C- -%.The silt content was an insignificant explainer.The content of clay fraction explained 67.5 %, organic carbon content 62.9 ®/o and pH(CaCI 2 ) 65.4 % of the variation in the effective CECI, when the effect of the other two independent variables was eliminated.
The values of the effective CECI were about 85 % of the effective CEC2 (Table 1), because exchangeable K and Na were not included in the CECI, and more Ca was extracted with neutral ammonium acetate than with KCI.
The content of clay fraction explained 65.6 %, silt content 19.0 %, pH(CaCl 2 ) 80 % and organic carbon content 14.3 % of the variation in the cation sum, when the effect of the other three independent variables was eliminated.The sum of cations extracted with acid ammonium acetate was not very closely correlated with the estimates of effec- tive CEC.The correlation coefficient with CECI was r = o.6s***(n = 430).
The usability of pH(H 2 0) and Ca, Mg and K extracted by acid ammonium acetate in pre- dicting the effective CEC was tested with clay soils, coarser soils and the whole material.The highest values of the coefficient of determination were obtained by using acid ammonium acetate-extractable Ca instead of cation sum as an independent variable.With Ca and pH(H 2 0) as independent variables explaining the variation in the effective CECI, the coefficients of determination and cor- responding regression equations were as follows: -52.9 pH + I.BBCa(mval/l) The partial correlation coefficients for the CEC2 (2), pH(H 2 Q) (3) and acid ammonium relation between the effective CECI (1) or acetate-extractable Ca (4) were as follows: Omitting pH(H 2 0) from the regression analysis decreased the coefficient of determi- nation considerably.With acid ammonium acetate-extractable Ca as the only independent variable, 50.2 % of the variation in the effec- tive CECI for the whole material was ex- plained.
Theoretically it was justified to test the re- lationship between effective CEC, pH (H 20)   and acid ammonium acetate-extractable Ca also with a regression model in which all variables were logarithmic.However, this model was not superior to the model tested before.According to the logarithmic model, the relationship between the effective CECI, pH(H z O) and Ca in the whole material was described by the equation IogCECI = 1.12 + l.OSlogCa -0.19 pH, R 2 = 0.75.
The coefficients of determination for clay and coarser soil groups obtained with Ca and pH(H 2 Q) as independent variables did not deviate very much from each other.Therefore the examination of the material as a whole was appropriate.The variation in the estimates of the effective CEC was explained equally well.
The measured values of the effective CECI were compared with the values calculated ac- cording to the regression equation CEClfmval/ kg) = 309 -56.8 pH + I.7oCa(mval/l).The relationship between measured and predicted values is presented in Figures 1 and 2. The pro- portion of samples with the predicted value deviating more than 15 % from the measured value was 16 % of the whole material.This proportion included 11 °7o of the clay soil samples and 22.5 % of the coarser soil samples.In clay soils, the deviation did not exceed 30 % of the measured value.Coarser soil samples with the predicted value much below the measured value (No.I -3 ) 1 -3) were Fig. 1.Relationship between measured effective CECI (Ca + Mg + acidity) and its estimate calculated on the basis of soil testing data (Ca + pH) in coarse mineral soils.characterized by high clay and organic carbon content.In clay soils the sample with the greatest underestimation of the measured value (No. 9) was characterized by high or- ganic carbon content.Samples with the pre- dicted value much above the measured value (No. 4-B) were characterized by low or mod- erate clay and organic carbon content.The pH(H 2 0) of these soils were 6.0-7.0 and in two cases (No. 7 -B) the content of acid am- monium acetate-extractable Ca was high, over 160 mval/1.It is possible that some undis- solved calcium carbonate existed in the soil.

Discussion
Although the material was collected from a restricted area, it was characterized by rela- tively great variations in soil properties.Under these circumstances it is justified to consider the material quite suitable for the present examination.
It has been shown earlier that the cationexchange capacity of Finnish mineral soils is largely dependent on clay and organic carbon content (Heinonen 1960, Marttila 1965,  Jokinen 1984) together with soil pH (Kaila  1971 a, b).In the study of Kaila (1971 b), the three variables together explained 83 % of the variation in the effective cation-exchange ca- pacity estimated as the sum of exchangeable Ca, Mg and acidity displaced by unbuffered KCI.The result is in good agreement with the results obtained in this study.However, in the material (n = 230) of Kaila (1971 b), or- ganic carbon content was a lesser explainer as compared with the present material with higher mean content and wider range of or- ganic carbon and smaller range of pH(CaCl 2 ).
The observation of Kaila (1972) that ex- changeable Ca and Mg together saturate on the average 80-90 % of the effective CEC of cultivated soils was manifested also in this study.The proportion of exchangeable acidity was of minor importance in most soils.The average ratio of exchangeable Ca to Mg was about 7 in clay soils and nearly 9 in coarser soils.In the study of Kaila (1972), the ratio in sand and fine sand soils was about 9 and in clay soils (clay-% < 60) about 4. According to Schmid (1965), the ratio of Ca to Mg would be s -B 5 -8 to 1.The average ratio of ex- changeable Mg to K was about 2 both in clay soils and in coarser soils.The average ratio of exchangeable K to Na was higher in coarser soils than in clay soils, the values being 4.6 and 3.9, respectively.In the study of Kaila  (1972), the ratio of Mg to K was 2-B and the ratio of K to Na 2-2.8.
On the basis of soil testing, the cation- exchange capacity of soil was to some extent predictable.The variation in the effective CEC was best explained by pH in soil-water suspension and Ca extracted by acid ammo- nium acetate.Indeed, additional variation can be caused by the fact that cation-exchange ca- pacity was expressed on a weight basis and acid ammonium acetate-extractable Ca on a volume basis.Using the sum of Ca and Mg or Ca, Mg and K extracted by acid ammonium acetate instead of Ca did not increase the coefficient of determination.This was largely due to the fact that K and Mg were only minor components of CEC.
The effective CEC was adequately predictable by a regression equation in which pH(H 2 0) and acid ammonium acetate-ex- tractable Ca were independent variables.The same equation could be applied to all mineral soils.When Ca is expressed as mg/1, the equation takes the form CEC(mvalZkg) = 309 56.8 pH(H 2 0) + o.oßsCa(mg/l).Most of the material was such that the predicted CEC value did not deviate more than 15 % from the measured value.Some characteristics common to the samples with a great deviation were observed.High content of organic carbon, e.g., seemed to be associated with a great underestimation of CEC.The immediate reason in this case might be the smaller bulk density which affects the relationship between values given on weight and volume basis.

Fig. 2 .
Fig. 2. Relationship between measured effective CECI (Ca + Mg + acidity) and its estimate calculated on the basis of soil testing data(Ca+pH) in clay soils.