Major Research Developing future projection model of coastal change in the context of climate change

There is a growing need to precisely predict beach morphological changes caused by future sea level rise and wave climate changes, and to devise appropriate countermeasures for addressing the coastal disaster risks induced by such coastal changes. In this research project, we aim to develop a numerical model predicting the shoreline changes caused by climate change applicable to various beaches based on long-term beach monitoring data. We have established a prediction model using LSTM network, which is one of the deep neural networks. Against this model, it was trained using the daily changes of coastal topographic section over the past 18 years observed at Hasaki coast and the daily changes of coastal topographic section for the following six years were predicted, and verified. Although the prediction accuracy of longshore bar morphology was not necessarily good enough, for foreshore topographic is stable for long period, it could be predicted high precisely. For climate change and related topographic change of prediction, further research will be needed.