An evaluation of lime Requirement methods for selected South African soils

[摘要] ENGLISH ABSTRACT: A laboratory study was conducted to evaluate seven widely used methods to predict soil limerequirement (LR) using 20 acidic South African top and sub-soils with a wide range of properties. TheLR methods which were evaluated against a standard CaCO3 incubation LR procedure, included: theoriginal Eksteen method with organic matter correction factor (OMCF) , commonly used in the WesternCape; two modifications of the Eksteen method, namely: (i) Eksteen-KCl method, involving the use of1 M KCl exchangeable acidity instead of titratable acidity at pH 7, and (ii) Modified-Eksteen method,where a correction factor was applied to titratable acidity that was derived from soil data obtained inthis study; the Cedara method, most commonly used in KwaZulu-Natal; the ARC-SGI method,developed primarily for Free State soils by the ARC- Small Grain Institute in Bethlehem; theShoemaker-McLean-Pratt single buffer (SMP-SB) method most commonly used in the North East andNorth Central regions of the USA; the Adams and Evans single buffer (AE-SB) method most commonlyused in the South East and Mid-Atlantic regions of the USA.The original Eksteen method, although highly correlated with incubation LR, was found to be arelatively inaccurate predictor of LR. The Eksteen-KCl and Cedara methods were found to be highlycorrelated with incubation LR, yet consistently underestimated LR. The modified-Eksteen method wasfound to be highly correlated with incubation LR, and was a good predictor of LR. The ARC-SGImethod was a considerably poorer predictor of LR, and tended to grossly overestimate LR. The SMPSBmethod was found to be highly correlated with incubation LR, and was shown to be reasonablyaccurate to achieve a target pHKCl of 5.5. Recalibration of the SMP-SB soil-buffer pH with incubationLR resulted in considerable increases in accuracy. The AE-SB method was found to be highlycorrelated with incubation LR, yet tended to overestimate LR. Recalibration of the AE-SB soil-bufferpH with incubation LR resulted in a sufficient increase in accuracy.A correlation study was conducted to investigate the relationship between soil properties and bothincubation LR and LR methods. It was revealed that soil properties other than soil pH, which are usefulindicators of LR are: soil C > variable charge > CECpH 7 > clay + silt. Soil C was found to be a significantcontributor to LR due to its association with exchangeable Al and due to its high pH dependent acidity.Titratable acidity was found to be the soil property that most strongly related to soil LR. Variablecharge was also shown to exhibit significant relationships with soil parameters that most stronglyinfluence LR. For these reasons, a multiple regression equation was developed that utilised onlytitratable acidity and variable charge. The multiple regression equation was able to predict 96.76% ofthe variation observed for incubation LR, and was 97.86% accurate in predicting the LR for eachspecific soil to obtain a target pHKCl of 5.5. Regarding the relationship between soil properties and LRmethods, it was revealed that the local methods, except the ARC-SGI method, were most stronglyinfluenced by exchangeable acidity and Al, and had significant relationships with soil C. The American direct buffer methods were strongly correlated with essentially all of the soil properties studied. Thisis indicative of the ability of these methods to directly determine the amount of acidity that mayoriginate from various sources in the soil, in order to make a sufficiently accurate LR. It is thereforerecommended that the application of direct buffer methods be further developed for use on SouthAfrican soils in order to further improve the accuracy of LR determination in South Africa. The existingmethod that was found to most accurately predict LR on a wide range of soils was the modified-Eksteenmethod.

[发布机构] Stellenbosch University