COMPARING RESPONSE-BASED AND EVENT-BASED OVERTOPPING DESIGN

Abstract

Coastal structure crest elevations are routinely designed to a specific hazard level. In the U.S., for example, levee crest elevations often correspond to the 1 percent annual exceedance probability (AEP) overtopping rate at 90 percent confidence level (CL). Statistical methods to compute coastal structure response range from simply inputting wave and water level forcing conditions at a certain AEP into a response equation (i.e. event-based or frequency-based approach) to a fully stochastic Monte Carlo numerical simulation where thousands of storm responses are sampled and epistemic uncertainties incorporated (i.e. response-based approach). Event-based approaches oversimplify both statistics and physics; however, time, cost, and complexity can limit application of the response-based simulation. Response-based stochastic simulation approaches tend to more realistically characterize structure responses. Herein we compare common frequency-based and response-based stochastic approaches for levee and floodwall overtopping design. For frequency-based approaches, structure responses are computed using wave and water level forcings at a given AEP, and for response-based approaches, a large number of storms are sampled in a Monte Carlo simulation.