Predicting sodium dynamics by mechanistic and layer models in different textured soils.

Abstract

Exchangeable sodium percentage (ESP) buildup in Ca-saturated columns of clay loam, loam, and sandy loam soils on percolation with smaller and bigger volumes of 0.1M NaCl solution was predicted by a mechanistic and a layer model. The mechanistic model was based on the numerical solution of a one-dimensional convective-dispersive partial differential equation, and the layer model was based on Gapons’ exchange equilibria. ESP values predicted by these models were compared with those determined experimentally at a regular-depth interval. The mechanistic model predicted ESP buildup fairly well in deeper layers of clay loam uponthe percolation of a smaller volume of solution. However, predictability was poor throughout the profile upon percolation of a bigger volume. In the case of loam and sandy loam soils, the difference between experimental and predicted values was relatively larger regardless of volumes percolated. The conceptual-layer model predicted ESP buildup through clay loam with great accuracy, regardless of volumes of percolation. However, predictability was fairly good with a smaller volume in the case of loam soil. The layer model could not predict ESP buildup through sandy loam soil with the same accuracy to that of clay loam and loam. Performance of the mechanistic model was poor (R2¼0.72) as compared with the layer model (R2¼0.97).