Output is compared from four methods used to estimate the overtopping rate at seawalls subject to random wave action: two empirical models, a semiempirical model, and a numerical model. The empirical models were developed by fitting dimensionless groups to data derived from physical model tests. The semiempirical model was derived from consideration of the unsteady flow of water over a weir. However, like the empirical models, it was calibrated with the results of physical model tests. In contrast, the numerical model AMAZON is a high-resolution two-dimensional finite volume model based on the nonlinear shallow-water equations.

In this study, we calculated the mean overtopping discharge for a range of seawalls with front slopes from 1 : 1 to 1 : 20 and for incident wave steepnesses from 0.01 to 0.03. The results are considered alongside four sets of data from physical model tests. They show general agreement between the output from the numerical and semiempirical models and the data. Agreement with the empirical models depends principally on the value of the surf similarity parameter. The empirical models substantially overpredict discharges for some conditions.