Modeling Preferential Flow In Agricultural Settings Using a Source-Responsive Approach.

Preferential flow is an important process in agricultural settings where transport of contaminants and aquifer recharge are of concern. Models that simulate variably saturated flow using Richards’ equation require extensive hydraulic property data that are rarely available and therefore estimated using more easily measured properties such as soil texture.  These models often produce unrealistic results where preferential flow is significant yet not explicitly included in the model parameters. Here we present and test a model with a structure specifically designed to estimate unsaturated zone preferential fluxes of water with minimal physical characterization. Direct evidence suggests that fast flow is a response to the water source conditions. The basic concept behind the source-responsive model, which incorporates film-flow theory with conservation of mass, is that preferential flow is most sensitive to boundary conditions such as whether the source is continuous or intermittent. We demonstrate the utility of the source-responsive model for simulating rapid water-table fluctuations at three agricultural sites with relatively long records of precipitation and water levels. One site is sandy in texture with no tile drainage, while the others are finer textured and require drains to keep the water table from inundating the root zone. Results show that this simple approach for quantifying preferential flow can be used successfully, especially where water tables show a rapid response to surface water input. Results also show limits of model applicability related to groundwater level, storm intensity, and antecedent moisture conditions.