Development of Integrated Discharge and Sediment Rating Relation Using a Compound Neural Network

Abstract

The assessment of sediment transport in rivers is of vital importance in design and management of hydraulic structures such as dams, diversions, hydro-power projects, river training works, bridges, etc. Previously reported studies have shown that data driven techniques such as the artificial neural network (ANN) can give better results in modeling stage-discharge relations than the conventional rating curves. In view of the complexities of rating relationships, compound rating curves are frequently used in place of a single rating curve. Accordingly, this paper investigates the abilities of compound neural networks (CNNs) to model integrated stage-discharge-suspended sediment rating relationship. Using the data of two stations on the Mississippi River and one station on Conococheague Creek, CNNs were trained. A comparison of the results of applying a single ANN and a CNN shows that the estimates of CNN are closer to the observed values than those of single ANN.