Solute Dispersion in Soil

Abstract

An experimental study of solute transport during one-dimensional sorption of a solution of KC1 into a uniformly packed horizontal soil column of Akuse clay, (a tropical clay loam from Ghana which is dominantly montmorillonitic) and Brookston clay of the Huron Catena (dominantly illitic) is reported in this thesis. A hydrodynamic equation based on the existing theories of irreversible thermodynamics is developed to incorporate the effect of anion exclusion in the usual hydrodynamic equation for one-dimensional flow of solute in soil. This equation is used to calculate the. longitudinal dispersion coefficient D^ in horizontal infiltration experiments where anion exclusion effects were observed. Also, the analysis of hydrodynamic dispersion during one-dimensional horizontal flow developed by Smiles eit al. (1978) is extended to include gravitational effects present during one-dimensional vertical infiltration by using the power series form of solution developed by Elrick _et a^L. (1979). Simulation methods using computer programs written in system/360 Continuous System Modeling Program (CSMP) are used to solve the equations used in this study. The longitudinal dispersion coefficient is in all the analyses presented in this study, assumed to be independent of the Darcy flux and a function of the water content only. In all experiments, both the soil solution content, the chloride concentration and the potassium concentration preserved similarity in terms of distance divided by square root of time. This confirmed the assumption that the longitudinal dispersion coefficient is effectively independent of the Darcy flux for the conditions of the experiments reported in this study. These results were found to be consistent with the studies of Saffman (1959) , Pfannkuch (1963) and Scotter and Raats (1970). The observed incomplete piston-like displacement of the antecedent water content suggests the presence of stagnant or 'immobile' water fractions in the soil columns considered. The derived longitudinal dispersion coefficient for Cl agreed fairly well with the derived longitudinal dispersion coefficient for K . This is an indication that under the experimental conditions imposed, the equilibrium adsorption isotherm used to describe the adsorption term in the hydrodynamic equation for solutes which interact with soil particles, is adequate. Simulation of water content, chloride concentration and concen- + - 1/2 tration of K as a function of X(=xt ) using derived data D(X), Dg (X) for Cl and Dg (X) for K+ showed very good agreement with experimental data. The simulated c(X) for Cl also showed excellent agreement with calculation of c(X) using a program for the analytical solution. Water content profile and chloride concentration profile obtained for vertical infiltration experiments indicated good agreement with theoretical water content and Cl concentration profile simulated with CSMP. The theoretical chloride concentration profiles showed a progressively dispersed 'front' with long infiltration time periods whereas the theoretical water content profiles for various time periods showed a sharp and abrupt wetting front