Narrow Results

A new data-driven method for morphodynamic seabed modelling and prediction that is able to account for morphological relationships across space is presented. The new method uses a two-dimensional spatial statistical model that first derives surfaces that are representative of the spatial morphological properties in both the cross-shore and alongshore directions simultaneously. It evaluates the spatial changes that occur between successive surfaces across the time series to derive a spatial-temporal function of the behaviour evolution. The resulting function is then extrapolated to obtain a prediction. The model is initially applied to idealised morphological scenarios with known morphological and evolution properties to assess its validity for morphological applications. The outcome from the idealised cases indicates that the model is relevant to such applications as it was able to identify morphological properties and evolution behaviour. It was also able to predict future states of the idealised morphology within a margin that is less than the natural variability of the feature.