Prediction of Longshore Sediment Transport Using Soft Computing Techniques

An important task for coastal engineers is to predict the sediment transport rates in coastal regions with correct estimation of this transport rate, it is possible to predict both natural morphological or beach morphology changes and the influence of coastal structures on the coast line. A large number of empirical formulas have been proposed for predicting the longshore sediment transport rate as a function of breaking wave characteristics and beach slope. The main shortcoming of these empirical formulas is that these formulas are not able to predict the field transport rate accurately. In this paper, an Adaptive-Network-Based Fuzzy Inference System which can serve as a basis for consulting a set of fuzzy IF-THEN rules with appropriate membership functions to generate the stipulated input-output pairs, is used to predict and model longshore sediment transport. For statistical comparison of predicted and observed sediment transport, bias, Root Mean Square Error, and scatter index are used. The results suggest that the ANFIS method is superior to empirical formulas in the modeling and forecasting of sediment transport. We conclude that the constructed models, through subtractive fuzzy clustering, can efficiently deal with complex input-output patterns. They can learn and build up a neuro-fuzzy inference system for prediction, while the forecasting results provide a useful guidance or reference for predicting longshore sediment transport.