Process-Based Model Prediction of Storm-Induced Coastal Morphological Changes using Flume Experiment and Field data

Abstract

Coastal areas are subjected to storm-induced flooding and erosion, which are threats to coastal safety in the coastal region. Accurate prediction of storm-induced coastal erosion is of help to reinforce the safety and of importance for coastal dweller, coastal hazard mitigation, and coastal erosion early warning system. Therefore, the process-based model for morphological changes during storm (i.e. XBeach) have been developed and improved for reliable model results. However, key parameters in XBeach remain to be adjusted to simulate morpho-hydrodynamics to a specific area of interest since XBeach with default parameters was intended for the application for storm erosion on dissipative sandy beaches. In other words, the degree to which datasets having different wave conditions influence model performance is still unclear at a given parameter space. In the present study, a process-based model (XBeachX) was tested and calibrated to improve the accuracy of simulating storm coastal erosion using both the large-scale experimental data (1D) and observed field data with high resolution covering from subaerial area to offshore model boundary (2D). In addition, model sensitivity to event-specific calibration data was examined using four storm datasets with four key parameters. The calibrated XBeachX successfully predicted wave transformation, dune erosion phenomena, various erosion volume changes above mean sea level, and observed formation of the longshore uniform sandbar after storm conditions. However, predicting exact crest position of bar and bed level changes across inner surfzone is the remaining problem. Analysis of model sensitivity to different incident wave conditions emphasize the need for storm datasets to include severe erosion in calibrating the erosion model. These results contribute numerical modelling for future dune erosion prediction and soft engineering such as artificial dune and provide practical guidance for the selection of event-specific data in model calibration.