Suspended Sediment Load Prediction using Artificial NeuralNetwork Integrated with the Whale Optimization Algorithm

Estimating suspended sediment load (SSL) is an essential task in water resources management. This article proposes theutilization of a hybrid artificial neural network (ANN) model for predicting SSL using historical SSL data. Various inputscenarios involving streamflow (Q) and precipitation (P) were utilized to assess the performance of the ANN and ANNWhaleOptimization Algorithm (WOA) in SSL prediction at Sarab Seyedali within the Alashtar basin. Optimizationalgorithms were employed to adjust and optimize the parameters of the ANN model. Two statistical indices, the correlationcoefficient (R2) and the root-mean-square error (RMSE), were employed to assess the accuracy of the models. Acomparison of models indicated that the integration of ANN-WOA improved the accuracy compared to the standaloneANN mode. Results Obtained from Pearson’s correlation coefficient techniques showed that the most effectiveparameters in SSL prediction are Q (t), Q (t-1), and P (t-1). ANN-WOA exhibited superior performance compared to ANN,achieving an R2 value of 0.690 and an RMSE of 0.0666.

Suspended Sediment Load Prediction using Artificial NeuralNetwork Integrated with the Whale Optimization Algorithm Keywords:

Alashtar Basin , Neural Network , Suspended Sediment Prediction , Whale Optimization Algorithm