Empirical modeling of beach evolution combining cross-shore and longshore processes

Overview & Synopsis

Sandy coasts are highly dynamic environments subject to erosion and marine flooding hazards, which are a threat to populations and economical activities near the coastline: methods and tools are needed to address coastal challenges, including models to predict future shoreline evolution at seasonal to decadal time scales and at beach to regional spatial scales. This PhD work is focused on improving the performance of empirical approaches based on a simplified representation of the dominant physical process, in particular by coupling a longshore one-line model with an equilibrium cross-shore model.The goal is to improve the cross-shore model prediction skill for long (decadal to centennial) time scales and to include the coast potentially more dominated by longshore processes. The empirical equilibrium shoreline change model of Yates et al. (2009), modified by Lemos et al. (2018) is used to model cross-shore processes, and a simple one-line approach is used to model longshore processes.The coupled model is implemented, validated and analyzed at the well-studied Narrabeen Beach (Australia). A sensitivity analysis is performed on the longshore model to errors in the forcing wave conditions, and an important sensitivity to errors in the wave direction is highlighted. Then, a method is proposed to correct a previously observed behavior of the one-line model generating a change in the coastline planform orientation that is not observed in the survey data, by assuming that it is due to biases in the forcing wave direction. Using a Monte Carlo approach, a set of relatively small wave angle biases is found to correct the model reorientation. Then, 7 combined models, with different ways of coupling the cross-shore model, the longshore model and a linear trend term, are implemented and tested at Narrabeen Beach to evaluate the model performance at reproducing the shoreline position at different temporal scales. Three criteria are used for this inter-comparison, based on the time scale at which the model skill in reproducing shoreline variability is assessed: short (~monthly), medium (seasonal) and long (pluri-annual) temporal scales.With the overall objective to extend the timescale of shoreline change predictions using empirical models such as the coupled model presented herein, 3 existing methods to predict long-term (from 10 to 100 years) shoreline change were tested at Vougot Beach (France) to examine the differences between existing methods, as well as the uncertainties of such long-term predictions. This work pointed out the uncertainties and the complexity of generating long term forcing conditions, that include potential climate change impact, when performing long-term predictions. It also highlights the necessity to improve the methods used to take into account changes in water levels in the equilibrium shoreline change model.