The influence of soil hydraulic property estimation on the predictive accuracy of solute transport modeling

Prieksat, Mark Alan, author
Butters, Greg L., advisor
Ahuja, Laj, committee member
Cardon, Grant, committee member
McWhorter, David B., committee member
Journal Title
Journal ISSN
Volume Title
Colorado State University. Libraries
Because of the complexity of processes governing water and chemical movement in the unsaturated zone, numerical models will necessarily play a key role in predicting the fate and transport of chemicals. If models are to fulfill their role as a tool in managing agrichemicals, then clearly models need to be tested for sensitivity to the method used to measure the soil hydraulic properties used in the model. The objectives of this study were to compare hydraulic parameter estimates obtained using alternative methods and to evaluate the effect of the parameter estimation method on the predictive accuracy of the HYDRUS-20 model. Soil hydraulic parameters were determined at multiple depths within a 2 meter deep soil profile using a variety of methods; field methods included the instantaneous profile method (IPM) and tension infiltrometry at two scales (5.08 cm-diameter (4TI) and 20.32 cm-diameter (8TI)); lab methods included determination of saturated hydraulic conductivity using the falling head method and analysis of moisture retention using pressure plate analysis at two scales (5.08 cm-diameter x 5.08 cm-long soil cores (2C) and 10.16 cm-diameter x 5.08 cm-long soil cores (4C)); indirect methods (IND) included determining hydraulic parameters from three literature sources based on soil textural data. In addition, identical parameter estimates were made under two tillage treatments (non-tilled (NT) and tilled (T)). It was found that estimation of soil hydraulic properties was sensitive to the measurement method selected. Estimation of α and Ks parameters was more sensitive to method and scale than was estimation of the n and ɵs parameters. The IPM method showed the least variability in parameter estimates. Tillage introduced significant changes in the hydraulic properties and increased spatial variability in the parameter estimates. Data obtained using the alternative estimation methods was used as input to the HYDRUS-2D model to predict water and solute movement. Water content and solute concentration profiles estimated using the model were compared with observed water content and solute concentration profiles obtained during a field-scale solute transport study using bromide tracer. Most of the parameter estimation methods resulted in simulated water content that were within about 2 to 5 percent of the measured data. The IND method had the largest deviations (about 10 percent) from the measured data. Under T soil conditions, all the methods (except the IND) predicted bromide movement well and simulated bromide concentration data were well correlated to measured data at all times throughout the simulation. Under NT conditions, all methods, except the 8TI, under-predicted the center of mass movement of bromide. Analysis of the simulation data revealed that the hydraulic properties of the surface soil are extremely important in controlling water and solute movement. Results indicate that the IPM is probably the most reliable method, but other estimation methods may result in similar predictions of water and solute movement when made in repetition.
1999 Spring.
Includes bibliographical references (pages 168-182).