Seasonal to Interannual Shoreline Variability in a Changing Wave Climate

Abstract

Understanding the drivers of shoreline evolution is critical to manage the world’s coastlines now and in the future. In view of anticipated future variability of the wave forcing drivers of shoreline change, this thesis aims to improve our understanding and ability to predict seasonal to interannual shoreline variability in a changing wave climate. First, a high-resolution dataset of shoreline evolution in SE Australia is used to examine the interannual variability in drivers of embayed beach shoreline variability and beach rotation. A novel time-varying EOF analysis reveals interannual variability of the dominant cross-shore and alongshore mechanisms of shoreline response, which in turn are controlled by different influences of wave intensity and wave direction parameters. Extending the analysis to regional scales, a dataset of satellite-derived shoreline time series (~200 embayed beaches) is used to classify regional variability in beach rotation across SE Australia. Time series representing beach rotation are subjected to correlation and cluster analyses to identify regional variability in beach rotation. Different patterns are classified into five statistically distinct clusters, each physically linked to the regional variability in embayment’s morphometry. Noting the interannual variability in shoreline behaviour presented in the first two components, next, a novel approach to improve seasonal to interannual shoreline change predictions is presented, whereby model parameters are allowed to vary in time. An Ensemble Kalman Filter (EnKF) technique is embedded within the shoreline change model ShoreFor. Evaluation of this technique over synthetic scenarios shows that the EnKF technique suitably detects time-varying parameters. Application to an 8-year real‐world shoreline dataset on the Gold Coast, Australia reveals that time‐varying parameters are linked through physical processes to changing characteristics of the wave forcing. Finally, the EnKF technique is applied to a 28-year satellite-derived shoreline dataset in the Gold Coast. The estimated time-varying parameters are modelled as a function of the multi-year variability in wave forcing and used to predict shoreline changes out of the training period. Predictions from this enhanced model outperform a previous approach based on time-invariant parameters and emphasizes the need for shoreline models that can adjust to multi-year changes in wave forcing in view of climate variability.