Christensen et al. (2009) studied the optimal design of pumping tests whose goal is to estimate hydrogeologic parameters affecting stream depletion caused by pumping in a leaky aquifer. Their analysis relies on the analytical solutions of Zlotnik and Tartakovsky (2008), which are based on the assumptions of negligible drawdown in the source bed of a leaky aquifer and of horizontal flow in an aquifer of infinite extent. We conduct a series of two- and three-dimensional simulations to identify the validity and range of applicability of these assumptions, and to quantify their impact on estimation of aquifer drawdown, stream depletion, their sensitivities to the hydrogeologic parameters, and on selection of optimal locations to observe drawdown during a pumping test. Quantitative criteria for the use of these analytical solutions are formulated in terms of the source bed transmissivity and the distance from the well to prescribed-head boundaries other than the stream.