Modeling unsaturated flow and atmospheric interactions

Four codes used to simulate unsaturated flow with atmospheric interactions are described and evaluated. The codes are similar conceptually and function in a similar manner. However, the codes use different algorithms, and therefore yield different predictions for the same input. A field validation exercise conducted with three of the codes shows that the predictions appear realistic regardless of which code is being used. However, each code provides a different prediction, and none of the predictions are in agreement with field measurements. The differences can be attributed in part to the methods used to implement the atmospheric boundary. More subtle features of the codes also have a significant effect on predictions (e.g., hydraulic properties model, hysteresis, vapor flow, thermally-driven flow, transpiration algorithms, ground freezing, snow melt, etc.). The user should consider each of these features carefully so that realistic, but conservative predictions are obtained.