Solute dispersion in selected tropical soils

Abstract

The dispersion of chloride, nitrate, potassium, calcium, and phosphate was investigated in three tropical soils represented by three Hawaiian soils through a specially designed flow apparatus. Observed differences in the dispersion of these solutes for the range of velocities tested were attributed to differences in ion adsorption capacity and pore geometry. Tritium showed negligible interaction with these soils and was used to trace fluid dispersion. Tritium breakthrough curves differed because of different dead-end porosities of these soils and were estimated to range from 4 to 30 per cent. Potassium, calcium, and phosphate showed trends in their breakthrough curves that were due to dead-end pores. When dead-end pore volume was large, adsorption was relatively small even when the potential for adsorbing ions was great. Because of dead-end porosity a new dimension is added to the interpretation of adsorption data. This in some cases invalidates predicted adsorption based on traditional equilibration techniques. A theoretical model was found adequate to describe the dispersion of tritium, phosphate, chloride, and nitrate under saturated-flow conditions. The unsaturated breakthrough curves of chloride and nitrate also agreed fairly adequately with theory. The unsaturated and saturated breakthrough curves of chloride and nitrate differed because of the added effect of diffusion and restriction of adsorption sites due to stagnant pores in the unsaturated case. A quantity, Ve, which is a volume measured at the arrival of one-half of the initial concentration appears to be a more useful qualitative (as well as quantitative) index of dispersion and adsorption than the dispersion coefficient, D. The experimentally obtained ratio of Ve/Vo can be interpreted to indicate the following: Ve/Vo > 1, dispersion and adsorption; Ve/Vo = 1, dispersion only; Ve/Vo < 1, dispersion in the presence of stagnant or dead-end pores. Vo is the total pore volume. Since Ve is a function of the type of solute, its concentration and pore velocity, and can be estimated by simple inspection of the breakthrough curve, its usefulness becomes more apparent than D. The dispersion approach to the study of silica-phosphate interrelations yielded results which were not in general agreement with adsorption data obtained by non-flowing, batch equilibration techniques. The adsorption of phosphate as shown by breakthrough curves was enhanced by silica treatment in two soils. Less phosphate adsorption (fixation) reported in the literature which have been attributed to silica treatment was not shown to occur in the test soils. This particular finding suggests that the reasons given for crop response to silicate applications to soil need to be reexamined.