Comparison of River Suspended Sediment Load Estimation, using Regression and GA Methods

The rivers sediment load is determined using hydrologic methods. In the statistical methods, by measuring the rivers discharge and suspended sediment load in a long-term period, the relationships between the suspended sediment load and discharge is obtained. The aim of this study is to compare different estimation methods of suspended load and select the most appropriate relationship for the prediction of suspended sediment load which results in more accurate prediction of sediment load discharge in the studied rivers. In this research, by using the daily river flow rate and corresponding suspended sediment discharge in four hydrometric stations in West Azerbaijan, the rating curve power equation for the middle limit of the data was used. The optimum coefficients of the mentioned equation ware obtained by using genetic algorithm (GA) and the ordinary regression, and the necessary programs ware written in Matlab® software. Since GA does not require any restrictive assumptions on input data, so it showed better performance in real data validation phase. This improvement in the rating curve coefficients causes to decrease up to 25 percent Root Mean Squared Error (RMSE). Therefore using GA as an appropriate tool was proposed to estimate the sediment rating curve for the studied stations.