MACHINE LEARNING TECHNIQUES FOR CROSS SHORE BEACH CHANGE FORECASTING

The ability to predict beach morphodynamic changes at short to medium timescales is crucial for sustainable coastal management in a changing climate. The use of data driven approaches such as Machine Learning (ML) techniques to predict coastal change have gained interest in recent years (Kim, 2022, Montano, 2020). These techniques are proving to be a more computationally efficient alternative to the traditional process-based models. In this study tests two high-performance ML algorithms to predict cross-shore beach change. A deep learning neural network, Long-Short Term Memory (LSTM) and an ensemble ML model, extreme gradient boosting (XGBoost) are explored, and performances analysed, compared and contrasted.