PREDICTING SOLUTE DISTRIBUTION WITH CONVECTIVE-DISPERSIVE APPROACH

Increasing concern about the presence of pesticides and other agricultural chemicals as potential groundwater pollutants has led to strict regulation of their use, and in particular on their concentration levels allowable in ground water. Water is nature’s primary vehicle for the transport of dissolved chemicals and methods for analyzing the movement and fate of pesticides in the soil to minimize the risks of ground-water contamination have to be developed. Most of conventional transport models developed are based on the convective-dispersive equation which, in most formulations, assumes that water and solutes follow an average path through the soil-which is to say that a given molecule, starting at the surface, is equally likely to follow any one of a multitude of available paths.

The approach used by many soil scientists to simulate water and solute flow in unsaturated soils is based on a theory derived by L.A. Richards in 1931. In 1956 Wiebe vand der Molen combined aspects of this model with the the theory of dispersive movement to predict the course of desalinization of land in the Netherlands that had been inundated by seawater. The resulting convective-dispersive equation assumes that water and solutes follow an average path through the soil- which is to say that a given molecule, starting at the surface, is equally likely to follow any one of a multitude of available paths.

Typical of models based on the convective-dispersive equation is MOUSE (Model of Underground Solute Evaluation) developed by Steenhuis et. al. in 1987. MOUSE and similar models have proven quite satisfactory for predicting the overall flow of water and the transport of nitrates. In studies conducted on Long Island, the predicted values are acceptably close to values determined by sampling (Figures 1 and 2). But studies of pesticides conducted at similar sites failed to predict the speed with which the contaminants penetrated the soil (Figure 3). This failure occurs because the models are unable to predict spatial and temporal variations of pesticide concentration sometimes found in the groundwater. The models does not consider preferential flow paths through which the solute can quickly bypass the biologically active root zone and thereby reduce the potential for degradation.

Figure 1. Precipitation and recharge

Figure 2. Nitrate concentration

Figure 3. Atrazine concentration

The difference in predictive accuracy between the pesticides, on one hand, and nutrient and water, on the other hand, results, in part, in a wide disparity in the amount of chemicals that will cause health standards to be exceeded. According to limits set by the Environmental Protection Agency, nitrates are tolerable in concentrations up to 10,000 parts per billion, which are not attained unless more than 40% of the amount commonly applied reaches groundwater. In contrast, the maximum permissible pesticide concentration is usually below 10 parts per billion. On Long Island, a concentration of 2.5 parts per billion may be attained if just one quarter of one percent of the amount usually applied reaches groundwater. Since this limit can be exceeded if only a small amount of solute follows a non-average flow path, averaging models based on the convective-dispersive equation are too simplistic. A more sophisticated model must be developed to recognize that water does not infiltrate the soil uniformly, but shows a preference for certain pathways such as wormholes, cracks, structural faults or fractures, animal burrows, sub-surface erosion, root holes, etc; resulting in much faster transport of pesticides, and other contaminants, including nutrients, trace metals, and manurial pathogens.