Prediction of suspended sediment load by artifical neural networks on rivers

Correct estimation of sedimet volume carried by a river is very important with water resources, dams, reservoirs planning and managements, water supply problems, channel navigability,hydroelectiric equipment longevity and river aesthetics and scientific interests.Because of these problems which were explained above, the nonlineer dynamic relationship between hydrological events such as rainfall, runoff and sediment yield, have to be determined truly and certainly.Computation on hydology and hydraulic engineering has concentrated primerly on Artificial Neuroal Networks(ANN) in the past few years.In this study,ANN works for the estimation of sediment volume carried by a river.For application area Sakarya River was selected. Sediment yield forecasting models having various input structures were developed using Artificial Neuroal Networks. The results of the neural networks,Multiple Linear Regression (MLR) and observed values were compared and performances were assessed by fitness criterias. The results of ANN models have shown that ANN can be applied successfully and provides high accuracy and reliability for sediment yield forecasting.