The method of Bouwer and Rice (1976) for analyzing slug test data is widely used to estimate hydraulic conductivity (K). Based on steady state flow assumptions, this method is specifically intended to be applicable to unconfined aquifers. Therefore it is of practical value to investigate the limits of accuracy of the K estimates obtained with this method. Accordingly, using a numerical model for transient flow, we evaluate the method from two perspectives. First, we apply the method to synthetic slug test data and study the error in estimated values of K. Second, we analyze the logical basis of the method. Parametric studies helped assess the role of the effective radius parameter, specific storage, screen length, and well radius on the estimated values of K.

The difference between unconfined and confined systems was studied via conditions on the upper boundary of the flow domain. For the cases studied, the Bouwer and Rice analysis was found to give good estimates of K, with errors ranging from 10% to 100%. We found that the estimates of K were consistently superior to those obtained with Hvorslev's (1951) basic time lag method. In general, Bouwer and Rice method tends to underestimate K, the greatest errors occurring in the presence of a damaged zone around the well or when the top of the screen is close to the water table. When the top of the screen is far removed from the upper boundary of the system, no difference is manifest between confined and unconfined conditions. It is reasonable to infer from the simulated results that when the screen is close to the upper boundary, the results of the Bouwer and Rice method agree more closely with a ''confined'' idealization than an ''unconfined'' idealization. In effect, this method treats the aquifer system as an equivalent radial flow permeameter with an effective radius, Re, which is a function of the flow geometry. Our transient simulations suggest that Re varies with time and specific storage. Thus the effective radius may be reasonably viewed as a time-averaged mean value. The fact that the method provides reasonable estimates of hydraulic conductivity suggests that the empirical, electric analog experiments of Bouwer and Rice have yielded shape factors that are better than the shape factors implicit in the Hvorslev method. ¿ American Geophysical Union 1995