Analytical models used to analyze hydraulic head data from pumping tests incorporate boundary conditions defined along the water table to account for drainage from the overlying vadose zone. Although these boundary conditions do not explicitly describe the details of flow in the vadose zone, they do specify a bulk vadose zone response. We have quantified this response in terms of the cumulative drainage flux through the water table and undrained vadose zone storage. A comparison of bulk vadose zone response predicted by analytical models and inferred from field measurements was performed using hydraulic head data and soil moisture content profiles obtained during a seven day pumping test at CFB Borden, Ontario. Three different analytical models (one assuming instantaneous drainage and two types of delayed drainage using exponential decay functions) were used to predict the bulk vadose zone response from the hydraulic head measurements. The pumping test analyses using these models gave reasonable values for aquifer parameters, and the model predictions of the transient hydraulic head response replicated the observed hydraulic head data. However, there are significant differences between the model predictions and field observations of the bulk vadose zone response. The instantaneous drainage model entirety neglected the substantial undrained storage that was observed. In the vicinity of the pumping well, the delayed drainage models significantly overestimated the cumulative drainage flux and underestimated the undrained storage determined from the field data. The delayed drainage models predicted a relatively rapid dissipation of the undrained storage while the observed undrained storage exhibited little, if any, decay throughout the entire pumping test. Our results indicate that the water table boundary conditions used in these analytical models do not adequately replicate the mechanisms controlling the vadose zone behavior during a pumping test. (c) 2006 Elsevier B.V. All rights reserved.
