Comparison of Bayesian Neural Networks and Artificial Neural Network to Estimate Suspended Sediments in the RiverS (Case Study: Simineh Rood)

Background and Simulation and evaluation of sediment are important issues in water resources management. Common methods for measuring sediment concentration are generally time consuming and costly and sometimes does not have enough accuracy.
In this research, we have tried to evaluate sediment amounts, using bayesian neural network for Simineh-Rood, West Azerbaijan, Iran, and compare it with common artificial neural networks. Monthly river discharge, temperature and total dissolved solids for time period (1354-1383) was used as input and sediment discharge for output. Criteria of correlation coefficient, root mean square error and Nash Sutcliff bias coefficient were used to evaluate and compare the performance of models.
The results showed that three models smart estimate sediment discharge with acceptable accuracy, but in terms of accuracy, the bayesian neural network model had the highest correlation coefficient (0.832), minimum root mean square error (0.071ton/day) and the Nash Sutcliff (0.692) and the bias (0.0001) and hence was chosen the prior in the verification stage.
Discussion and Finally, the results showed that the bayesian neural network has great capability in estimating minimum and maximum sediment discharge values.