Modelling the cross-shore beach profiles of sandy beaches with Posidonia oceanica using artificial neural networks: Murcia (Spain) as study case

Abstract

This paper presents a model of the cross-shore beach profile taking into account the presence of the seagrass Posidonia oceanica whose ultimate objective is to reduce the volume of sand used in beach nourishment. The methodology describes the training, validation, testing and application of models of artificial neural networks (ANN) for computing the cross-shore beach profile of sandy beaches in the province of Murcia (Spain). Eighty ANN models were generated by modifying both the input variables and the number of neurons in the hidden layer. The input variables consist of wave and sediment data and data concerning the Posidonia oceanica. To select and evaluate the performance of the optimal model, the following parameters were used: R2, absolute error, mean absolute percentage error and percentage relative error. The results show a mean absolute error of 0.22 m (0.21 m in training and 0.28 m in test), representing an improvement of 85.1% compared to models that do not use the Posidonia oceanica and 69.8% against those that consider it. Although the ANN was developed for beaches with P.oceanica, it could be used in areas with other seagrass able to reduce wave energy and consolidate the sand such as Syringodium filiforme, Thaalassia testudinum, Laminaria hyperborea, Halodule wrightii and Zostera marina.