Main content area

Advection-dispersion analysis of solute transport in undisturbed soil monoliths

Jensen, K.H., Destouni, G., Sassner, M.
Ground water 1996 v.34 no.6 pp. 1090-1097
sandy soils, loam soils, chlorides, labeling techniques, equations, soil water movement, hydrodynamic dispersion, solutes, loamy sand soils
The results of an extended analysis of the chloride breakthrough curves (BTCs) obtained from transport experiments on undisturbed soil monoliths previously reported by Sassner et al. (1994) are discussed. Parameter values for different forms of the advection-dispersion equation (ADE) were obtained by curve-fitting using the nonlinear least-squares optimization code CXTFIT developed by Parker and van Genuchten (1984). Good fits to the experimental BTCs from the individual soil monoliths were obtained for three different forms of the ADE, allowing for anion exclusion, immobile water, and both of these processes, respectively. Although the more complex forms of the ADE possess more flexibility to provide a slightly more refined fit to the individual breakthrough curves, larger uncertainty is associated with the fitted parameter values. Due to the limited amount of information manifested in the BTCs it is not possible on the basis of this information to distinguish which of the models is physically more meaningful. Furthermore, when more parameters are included in the optimization procedure, the resulting values are more uncertain. The same ADE models were also fitted with equally good results to a hypothetical large-scale BTC derived by flux-averaging the responses from the individual monoliths. However, regardless of the form of the ADE used, the resulting parameter values for the large-scale transport (i.e., not only the dispersivity value that is expected to increase) were inconsistent with the corresponding parameter values obtained for the individual BTCs. This supports previous indications that ADE models may not be accurate for predicting large-scale transport in heterogeneous soil systems corresponding to e.g. the scale of an agricultural field or a grid element in a numerical catchment model because model parameters are not determinable from independent measurements on a scale corresponding to a practical manageable core size or to standard intrumentation.