Uncertainty Analysis Related to Beach Morphology and Storm Duration for More Reliable Early Warning Systems for Coastal Hazards

Abstract

Early warning systems (EWSs) for coastal erosion are cost-effective instruments for risk reduction. Among other aspects, the selection of the pre-storm beach morphology and the definition of storm characteristics can affect EWS reliability. Here, XBeach simulations were used to assess the uncertainties in beach-dune erosion related to the variability of storm severity and duration and pre-storm morphology. Wave height return periods (from 5 to 50 years) determined the severity and the duration variability was established from confidence intervals after an adjustment with wave height. The variability of steep profiles included different berm morphologies (from fully developed to eroded berms). Three indicators, relative eroded volume, proportional berm retreat and proportional dune retreat, were evaluated. The experiments revealed that: (a) Relative eroded volume uncertainties related to the pre-storm morphology variability were slightly lower (maximum 8%) than the uncertainties related to storm duration (11%–18%). (b) Pre-storm profile variability can induce large uncertainties in the proportional berm retreat (up to 88%) for moderate events such as the 5- and 10-year events. Storm duration variability had less influence on this indicator (maximum 12%). (c) The uncertainties in the proportional dune retreat increased with storm severity and they ranged between 14% and 41% for pre-storm profile variability and between 2% and 40% for storm duration variability. Duration variability even governed the occurrence of dune breaching on eroded berm profiles in the most extreme event. Hence, the uncertainties related to initial/forcing conditions, namely pre-storm morphology and storm duration, must be assessed to develop reliable coastal erosion EWSs.