Storm resampling for uncertainty analysis of a multiple-storm unit hydrograph

Abstract

Due to various types of uncertainties involved in the estimation of a unit hydrograph (UH), the UH derived by any method is subject to uncertainties. Based on the concept of the bootstrap resampling technique, a practical methodology called storm resampling is proposed to quantify the uncertainties of multiple-storm UH ordinates and any parameters involved in the estimation of the multiple-storm UH. The important UH ordinates and parameters may include UH peak discharge, UH time-to-peak, UH volume, condition number related to the effective rainfall matrix, mean square error of the UH, and ridge parameter. The proposed bootstrap-based storm resampling technique, along with the least squares and ridge least squares solution techniques, is applied to typhoon storm events over a watershed in Taiwan. The methodology can be applied to other UH solution techniques and other hydrological and hydraulic simulation/optimization models. © 1997 Elsevier Science B.V.