Numerical modelling of the equilibrium profile in Valencia (Spain)

Artificial neural network modeling of cross-shore profile on sand beaches: The coast of the province of Valencia (Spain)

“…Although, a study of variables should be performed again, since possibly the influence of the different variables in the profile would vary, and it would even have to study other possible variables that could affect the formation and movement of the bars, as for example the maximum wave height, period, etc. Nevertheless, the results obtained by the ANN in this area are lower (41.5% absolute error, 34% MAPE and 48.9% percentage relative error) than those committed by the model proposed by Aragonés et al (2016b). All this would result in a lower volume error when calculating the amount of sand needed in beach nourishment, and therefore a lower economic cost.…”

“…On the other hand, for evaluating the performance of the model, various statistical parameters were studied (Equation 7, 8 and 9). By analysing these parameters, it was observed that the neural network improves the results compared to numerical model obtained by Aragonés et al (2016b) for the equilibrium profile in the study area. Thus, it is observed that the mean absolute error of the ANN is 36.3% lower than the numerical model (0.21 m vs. 0.33 m).…”

“…The results were compared with the numerical model presented by Aragonés et al (2016b) for the equilibrium profile in the study area. It was observed that the mean absolute error committed by the ANN was 0.21 m while for the model of Aragonés et al (2016b) amounted to 0.33 m. In addition, the network offered better results in the bar area located near the port of Valencia. However, to improve the results obtained in the area a larger number of profiles affected by bars would be needed, since the number of profiles affected by these elements represents only 9.3% of the total.…”

“…Finally, the results were compared with the numerical model developed by Aragonés et al (2016b) to obtain the equilibrium profile in the study area. The results were also compared with the formulations of Dean (1977) and Vellinga (1983Vellinga ( , 1984) (supplementary material), obtaining errors much larger than the model of Aragonés and the ANN.…”

“…To select and evaluate the performance of the optimal model, the following parameters were used: R 2 , absolute error, mean absolute percentage error and percentage relative error. Finally, the results are compared with the numerical model proposed by Aragonés et al (2016b) for the equilibrium profile in the study area. The results show a mean absolute error of 0.21 m compared to 0.33 m Aragones' model, significantly improving the results of the numerical model in the bar area around de Valencia Port.…”