INTERPRETATION OF SOLUTE PROFILE DYNAMICS IN IRRIGATED SOILS .2. CONVECTIVE-DISPERSIVE APPROACH

The simple field-oriented convective-dispersive equation based model of Rose et al. (1982a and b) has potential to be used to monitor changes in soil chloride levels on commerical farms. As yet applications of the model have only been reported for detailed studies conducted on coarse textured soils. Here the model was applied to a wider range of soil types using input data likely to be available on commercial farms. The model gave good predictions of measured soil chloride profiles at five sites, which included two spray-irrigated Vertisols. To provide accurate predictions in these two Vertisols, infiltration of rainfall into macropores had to be accounted for (by an approximate means). Values of parameters in the model (including the solute dispersivity coefficient) at these five sites were within the range found in previous studies, even for the Vertisols. Variation in the solute dispersivity coefficient within the extreme values that had been found in this and previous studies, did not greatly affect predicted soil chloride profiles. At another three sites, all flood irrigated Vertisols, predicted soil chloride levels were lower than measured values, regardless of values chosen for model parameters. The effect of depth variation in field capacity and wilting point moisture contents on chloride movement was investigated at these three sites as a possible cause of the inaccurate predictions, and found to have a negligible effect on predicted profiles. Differences between measured and predicted soil chloride levels were attributed to infiltration of irrigation water into cracks and, at one site, to the effects of high watertables. These processes were not represented in the model, explaining the failure in this context. Application of the PEAK model provided useful insights into the root-zone hydrology of all soils studied. © 1990 Springer-Verlag.