A study of the kinetics of the adsorption of phosphorus by selected South African soils

Abstract

Plots of the decline of the concentration of phosphorus in soil suspensions over time were constructed for 23 selected soils from South Africa and, using non-linear regression analysis, various models were fitted to the time curves, including mechanistic first and second order kinetic functions. The model best describing the data was found to be the empirical modified Freundlich kinetic function. The curvature of the Freundlich kinetic model, b, appeared to be independent of the initial concentration of P of the additional input solution. The parameter b is a useful soil-P kinetics index for the selection of P fertiliser strategies and making P fertiliser recommendations, especially in situations where the fertiliser cannot be incorporated into the plough layer annually as is the case with perennial crops or in no till cropping systems. The measurement of b, however, requires the construction of a multiple point time curve, a lengthy and cumbersome process that does not lend it easily to routine determination purposes. A multiple linear regression function is proposed which allows for the rate of the reaction of P adsorption onto soil to be determined from soil properties. These include clay and citrate bicarbonate dithionite-Fe (sesquioxide) content, as well as a single measure of the concentration of P in solution after a 24hour incubation period with the soil. The signs of the coefficients of the function suggest that soils with high clay and low iron sesquioxide contents tend to exhibit less P fixation over the long term than soils with comparatively lower clay and higher iron sesquioxide content. This in turn implies that the residual value of P fertiliser applications in soils high in clay and low in iron sesquioxide lasts longer. The function was not only useful in identifying factors influencing the rate of the adsorption of P onto soils, but can also assist in estimating b, thus bypassing the need to construct a time curve. Implications of these findings for soil testing laboratories involved in making P fertiliser recommendations are discussed.