A COMPARISON OF MODELS PREDICTING GROUND WATER LEVELS ON HILLSIDE SLOPES

ABSTRACT: A comparative study of ground water level predictions on hillside slopes using two models is presented. The models are a simplified mass balance model that has components for evapotran‐spiration, recharge, and drainage; and a two‐dimensional finite difference model that employs kriging to estimate soil parameters and accounts for non‐uniform thickness of the soil layer. These models are representative of a wide range of modeling capabilities and are used to illustrate the sensitivity of ground water level predictions to the sophistication of the modeling techniques. The drainage and recharge components of the two models are evaluated and the importance of unsaturated flow in recharge computations is underscored. Piezometric observations in a small drainage depression on the slope of Kennel Creek Valley in Tongass National Forest, Alaska, were used to evaluate the two models. The results show that, although the predictions differ from the field observations, the simple physically‐based mass balance model predicts the ground water levels as well as the two‐dimensional model. It is suggested that caution should be exercised in using complex models to validate simpler models.