Evaluation of some probability distribution functions for derivation of unit hydrograph in the Bar Watershed, Iran

Abstract

Considering the ever increasing financial damages of floods and the need to manage the surface water, the use of new and more sufficient methods seems to be necessary. Therefore, the present study aims to investigate capability of the Gamma, Beta, Chi-square, and Weibull probability distribution functions (PDFs) for flood hydrograph derivation. The present study was conducted in the Bar watershed, Iran using 14 suitable rainfall-runoff events. After base-flow separation and direct runoff determination, the three hour unit hydrograph of the watershed was computed to determine the coefficients and parameters of the mentioned PDFs. Finally, Nash-Sutcliffe (NS), relative error (RE), and root mean square error (RMSE) indices were used for evaluation. According to the results, Gamma, Chi-square, and Weibull PDFs had satisfying results for predicting the peak flow with RE values of 11.87%, 25.34% and 37.67% respectively. In the case of time to peak, Gamma, Chi-square and beta PDFs showed acceptable results with RE values of 26.66%, 39.17%, and 34.17% respectively. Totally, considering the high importance of peak discharge and time to peak in flood control projects, Gamma PDF is recommended to be used in flood control projects.