Prediction of beach morphological changes using a data-based approach

A data-based approach using linear transfer functions (TF) was adopted to predict the evolution of the nearshore beach profile volume at Duck, North Carolina, using different wave forcing variables. The best TF model relation was found with the squared monthly average direction resolved significant wave heights. This TF model explained 76% of the variance of the data and produced a very good fit of the long-term trend in beach volume. This suggests that the long-term behavior of the bulk morphology of the beach profile is strongly influenced by the monthly average wave conditions. Complimentary long-term patterns in behavior were also observed on comparing the beach morphology and wave data. The fit of this TF model was improved by including the inputs of past alongshore sediment exchanges between adjacent profiles. Here, the TF model reproduced 92% of the variance in the volume data and fitted the long-term trend as well as some short-term behavior. This model gave very good forecasts of beach volume over a 5 year period. Thus, the linear TF modeling approach shows strong potential for predicting beach morphological changes