Fluid cavitation constitutes an expensive computational nuisance in underwater-shock response calculations for structures at or just below the free surface. In order to avoid the use of a large array of cavitating acoustic finite elements (CAFE), various wet-surface approximations have been proposed. This paper examines the performance of two such approximations by comparing results produced by them for 1-D canonical problems with corresponding results produced by more rigorous CAFE computations. It is found that the fundamental limitation of wet-surface approximations is their inability to capture fluid-accretion effects. As an alternative, truncated CAFE fluid meshes with plane-wave radiation boundaries are shown to give good results. In fact, a single layer of CAFE is found to be comparable in accuracy to the better of the wet-surface approximations. The paper concludes with an examination of variations in CAFE modeling.