From Bureau of Economic Geology, The University of Texas at Austin (www.beg.utexas.edu).
For more information, please contact the author.

International Applied Phytotechnologies Conference, Chicago, March 4, 2003

Intercode Comparisons for Simulating the Water Balance of Engineered Covers in Semiarid Regions

Bridget Scanlon, Marty Christman, Robert C. Reedy, Indrek Porro, Jirka Simunek, and Gerald N. Flerchinger

Abstract:

The water balance of engineered covers is critical for assessing the performance of covers. Because we need to predict cover performance for long-time periods, we rely on numerical models to predict the water balance. The purpose of this study was to compare water-balance simulation results from 7 different codes, HELP, HYDRUS-1D, SHAW, SoilCover, SWIM, UNSAT-H, and VS2DTI, using 1-3 yr water-balance monitoring data from nonvegetated engineered covers (3 m deep) in warm (Texas) and cold (Idaho) desert regions. Simulation results from most codes were similar and reasonably approximated measured water-balance components.

Simulation of infiltration-excess runoff was a problem for all codes, underscoring the difficulties of representing actual precipitation intensities and of measuring hydraulic conductivity of surficial sediments (as shown by the data from Texas). Drainage is the most critical parameter for evaluation of contaminant transport, engineered covers for waste containment, and groundwater recharge. Drainage could be estimated to within ~ ± 64% by most codes. Outliers for the various simulations could be attributed to the following factors:
·• the modeling approach, i.e., water-storage routing versus Richards' equation,
·• the upper boundary condition during precipitation and time discretization of precipitation input,
·• water retention functions (i.e., van Genuchten versus Brooks and Corey), and
·• the lower boundary condition (i.e., unit gradient versus seepage face).

The water storage routing approach does not seem to adequately represent the flow system in semiarid regions. By assuming that gravity is the only driving force and ignoring matric-potential gradients that are often upward in semiarid regions, downward flow is generally overestimated and ultimately results in overestimation of drainage. The approach used to simulate the upper boundary condition during precipitation is crucial when precipitation is input on a daily or larger time step. Setting PE to zero on rain days (VS2DTI) greatly underestimated evaporation and overestimated drainage. Subtracting PE from precipitation and applying net precipitation or net PE on a daily basis (HYDRUS-1D) had a much lesser impact on simulation results. The best approach is to disaggregate daily precipitation and apply it at a specified rate, allowing PE to occur throughout the rest of the day, as shown by the UNSAT-H simulations. The impact of water retention functions was demonstrated at the Idaho site, where increased unsaturated hydraulic conductivity based on the Brooks and Corey functions relative to the van Genuchten functions resulted in overestimation of evaporation and underestimation of drainage. The most appropriate lower boundary condition for simulating wickless lysimeters is a seepage face. Simulations using HYDRUS-1D demonstrated that this boundary condition could be approximated by simulating a thin bottom layer of gravel with a unit gradient boundary condition in codes that use Richards' equation but do not include a seepage face option. However, use of a unit-gradient lower boundary condition alone greatly overestimated drainage. This study demonstrates the usefulness of conducting intercode comparisons to evaluate the reliability of water-balance simulations and to determine important factors controlling water-balance simulation results.

Marty Christman, GeoSyntec Consultants Inc.
Indrek Porro, Idaho National Engineering and Environmental Laboratory
Jirka Simunek, George E. Brown, Jr., Salinity Laboratory
Gerald N. Flerchinger, U.S. Deptartment of Agriculture, Boise, Idaho