Prediction of exchangeable sodium percentage in soils by mechanistic and layer models

Abstract

An experiment was conducted in Plexiglass columns ( i d . 5.6 cm) to study depth distribution of adsorbed Na+ in Ca-saturated loamy sand, sandy loam and clay loam soils on percolating different volumes of 0.1M NaCl solution. The exchangeable sodium percentage (ESP) decreased gradually with depth and increased with increase in volume of solution percolated. The depth development of ESP of the soils was computed by using a numerical solution of a one-dimensional convective-dispersive partial differential equation (model I) and a conceptual layer model (model II). To account for the effect of adsorption on cation flow, the slopes of the nonlinear adsorption functions fitted to the normalized adsorption isotherms of Na+ were employed in model I , and the Gapon equation was employed in model II. As the Gapon selectivity coefficient (Kg) is dependent on Na saturation, its values were calculated from the regression equation: Kg = 0.22 + 0.0043SAR (where SAR is the sodium adsorption ratio). Model I overpredicted ESP of loamy sand and sandy loam soils throughout the column depth; the overprediction was more pronounced in the former. The coefficient of determination, calculated on the basis of pooled data of three soils and two solution volumes showed that the accuracy of prediction of model II was considerably higher (r2 = 0.80) than that of model 1 (r2 = 0.54).