Hydrologic Modeling of Protective Barriers: Comparison of Field Data and Simulation Results

Protective barriers, which consist of layers of silt loam over sand and gravel, have been proposed as covers for waste sites located in semiarid south‐central Washington state. The ability of an uncalibrated model to predict water contents, water storage, and drainage in barriers was tested for durations as long as 1.5 yr. Eight nonvegetated lysimeters containing the barrier layering sequence have been monitored since November 1987. The lysimeters were subjected to one of three precipitation treatments: ambient, 2 × average, and breakthrough (i.e., until drainage occurred). Distributions of measured and simulated water contents with depth were similar; maximum differences ranged from 0.023 cm 3 /cm 3 for the ambient treatment to 0.089 cm 3 /cm 3 near the soil‐sand interface for the breakthrough treatment. Simulated storage followed the trend in the measured values, although differences as much as 5 cm were observed at certain times. Generally, the model overpredicted evaporation in the winter and underpredicted it in the summer. Root‐mean‐square errors were 1.47 and 2.21 cm for the ambient and 2 × ‐average treatments, respectively. Sensitivity tests revealed that the hydraulic‐conductivity function, snow cover, and potential evaporation were important to successful modeling of storage in a protective barrier. When the above parameters and processes were adjusted (though not optimized), the root‐mean‐square error for the 2 × ‐average treatment was reduced 63% to 0.81 cm. For the breakthrough treatment, simulated drainage was obtained only by using field‐measured sorption and saturated‐conductivity data. This result indicates that hysteresis is important to successful modeling of drainage through protective barriers.