Estimation and forecasting of daily suspended sediment data by multi-layer perceptrons

The determination of the suspended sediment amount on the rivers is of crucial importance since it directly affects the design and operation of many water resources structures. In this study the performance of multi-layer perceptrons, MLPs, the most frequently used artificial neural network algorithm in the water resources literature, in daily suspended sediment estimation and forecasting was investigated. The forecasting part of the study was focused on sediment predictions using the past sediment records belonging either to downstream or upstream stations. The estimation of sediment values with the help of daily mean flows was the concern of the second part of the study. From the graphs and statistics it is apparent that MLPs capture the complex non-linear behaviour of the sediment series relatively better than the conventional models.