Variable-Parameter Muskingum Model

Abstract

The Muskingum model is not only a simple conceptual flood routing method but is also widely used by water resources engineers. Since Muskingum model parameters cannot be determined physically, the calibration of these parameters is the only challenge of this method. Although various techniques have been recommended for this purpose, a more accurate method with faster convergence rate is still required to improve the computational precision for parameter estimation. In this study, a new model was presented based on the linear Muskingum model for more accurate flood routing. In this model, the Muskingum parameters are allowed to change in the flood period, whereas in the classical model they are forced to be fixed over the entire flood interval. The Modified Honey Bee Mating Optimization (MHBMO) algorithm was utilized for the estimation of linear Muskingum parameters. The proposed model using the MHBMO algorithm was compared with constant-parameter model calibrated with other techniques in the literature for a case study. The results demonstrate that not only can the new model reduce the sum of the square of the deviations more than 50 %, but the MHBMO algorithm can also be confidently used to estimate the parameters of the Muskingum model.