STOCHASTIC BOUNDARY UNCERTAINTY IN MEAN WAVE OVERTOPPING RATE ESTIMATES

Resumen

Accurate backshore flood prediction requires quantifying overtopping volumes. Wave overtopping can be predicted through empirical formulas (e.g., EurOtop), neural networks or numerical modeling. The first two methods are easy to implement and provide rapid estimates for engineering design. However, empirical estimates reflect average overtopping rates, cannot resolve impulsive individual wave overtopping volumes or infragravity energy, and may differ from observational or numerical estimates by an order of magnitude (e.g., Gallien, 2016). Numerical models present an attractive alternative, however, predictions in phase-resolving models for spectral boundary conditions (BCs) vary due to the stochastic transformation to the time-domain, leading to significant errors in overtopping estimates when the prediction uncertainty is not accounted for. The purpose of this study it to derive an analytical formula for the coefficient of variation of the mean overtopping rate, applicable to all cases for which overtopping volumes are Weibulldistributed. Theory is used to provide guidance on estimating the uncertainty of numerical modeling predictions given the number of simulated waves.