Suspended sediment prediction using time Series and sediment rating curve (Case study: Ghazaghly station in Gorganroud River)

Accurate estimation of suspended sediment in rivers is very important from different aspects including agriculture, soil conservation, shipping, dam construction and aquatic research. There are different methods for suspended sediment estimation. In the present study to evaluate the ability of time-series models including Markov and ARIMA in predicting suspended sediment and to compare their results to sediment rating curve (SRC) it was tried to use daily suspended data from Ghazaghly station of Gorganroud River, as average monthly values in Minitab 16 software and finally suspended sediment was predicted for 111 months. In the next step, different combinations of all types of SRC and bias correction factors were used in Excel software to evaluate ability of SRC in suspended sediment estimation. Based on the results of this study, monthly SRC without bias correction factor was the most appropriate SRC models. To compare efficiency of different models in estimating suspended sediment, Root Mean Square Errors (RMSE) and Normalized Mean Square Errors (NMSE) were used. Time series model performance measured by RMSE and NMSE was about 71.34 and 2.48 respectively, compared to SRC model with RMSE 220.75 and NMSE 28.62. Results showed a good ability of time series models in estimating average monthly suspended sediment