Estimation of the sediment amount in the rivers with support vector machine model

Accurate estimation of sediments content in river structures has a significant importance in engineering studies of water resources and depends on the precise design and projecting of these structures. If the amount of sediments is not determined by the correct analysis, water intake structures and dams can complete their economic life in a short time. This has negative consequences in terms of productivity and cost. The amount of sediments is mostly determined by measurements from observation stations, sediment key curve, artificial intelligence modeling methods and solids equations used. In this study, the estimation of the sediments content was performed by using hydro-meteorological parameters such as river flow, air temperature and precipitation measured between 2011-2017 years at Omaha Station in Nebraska State, Missouri. For the estimation of sediment amount, sediment rating curve (SRC) and support vector machines- radial basis function (SVM-RBF), support vector machines-polynomial kernel (SVM-PK) and generalized regression neural network (GRNN) which are among the artificial intelligence methods in literature are used. These models were compared by using correlation coefficient (R), root of mean square errors (RMSE) and absolute mean error (MAE). When the measurement and model results were compared, support vector machines- radial basis function (SVM-RBF), support vector machines-polynomial kernel (SVM-PK) and generalized regression neural network (GRNN) models gave consistent results in the estimation of sediments content in rivers. SVM-RBF, SVM-PK and GRNN models can be offered as an alternative to classical methods.