Parameter Estimation Of Linear AndNonlinear Muskingum Models ForRiver Flood Routing

The importance of flood routing in rivers has been vastly recognized in hydraulic engineering practice. Field data scarcity often prevents the use of the Saint Venant equations to route floods in natural streams. As a consequence, approximate techniques such as the Muskingum models are commonly employed. The linear form of the Muskingum model has been widely applied to river flood routing. However, a nonlinear relationship between storage and discharge exists in most rivers, making the use of the linear form of the Muskingum method inappropriate. In this paper, three nonlinear forms of the Muskingum model are used. Different optimization techniques are applied to river flood routing for the estimation of the nonlinear Muskingum models parameters for the calibration period. To evaluate the perfomance of different parameter estimation techniques in the calibration period, the main inflow hydrograph and the partial inflows from tributaries and local ungaged inflow for the calibration period are routed to produce a computed outflow hydrograph for a given parameter set for each model. Then, the computed and observed outflow hydrographs for the calibration period are compared.