Evaluation of Selected Formulas and Neural Network Model for Predicting the Longitudinal Dispersion Coefficient in River

Longitudinal dispersion coefficient (LDC) is one of the most important parameters in the river water quality management. Several ways as empirical formulas and artificial intelligent techniques are proposed for predicting the LDC and it is necessary to evaluate the performance of them. In this study, a Multilayer perceptron neural network (MLP) model has been developed and 12 formulas empirical formulas were collected. To assess the performance of these formulas and MLP model in a case study problem, calculating the LDC by dispersion routing method for Severn River in UK was considered. Results shows that the best accuracy is related to the Tavakollizadeh and Kashefipour formula (R2?0.45) based on data set and for Severn River, its accuracy is R2?0.4. the (MLP) model has acceptable accuracy (R2?0.83) to predict the LDC in Severn River.