Numerical simulations of cross‐shore wave transformation and associated wave setup were performed for 100,000 realistic combinations of wave height, period and beach profile. Results of numerical simulations were compared to a widely‐used empirical predictor of wave setup and the outcome was a large variability in setup relative to that calculated by the empirical predictor. The variability, comparable in magnitude to the scatter around the empirical predictor developed with field observations, arose from unaccounted‐for wave energy dissipation over the cross‐shore profile. The beachface slope commonly used in empirical predictors can be a poor descriptor of wave transformation across the beach profile where most of the wave energy dissipation and setup generation occurs. Shoreline setup appears to be far more dependent on accurate characterisation of the offshore profile (e.g. bar location and depth) than the beach face slope.
