Probabilistic nature of storage delay parameter of the hydrologic model RORB: a case study for the Cooper's Creek catchment in Australia

Abstract

In rainfall–runoff modeling, Design Event Approach is widely adopted in practice, which assumes that the rainfall depth of a given annual exceedance probability (AEP), can be converted to a flood peak of the same AEP by assuming a representative fixed value for the other model inputs/parameters such as temporal pattern, losses and storage-delay parameter of the runoff routing model. This paper presents a case study which applies Monte Carlo simulation technique (MCST) to assess the probabilistic nature of the storage delay parameter (k_c) of the RORB model for the Cooper's Creek catchment in New South Wales, Australia. It has been found that the values of k_c exhibit a high degree of variability, and different sets of plausible values of k_c result in quite different flood peak estimates. It has been shown that a stochastic k_c in the MCST provides more accurate design flood estimates than a fixed representative value of k_c. The method presented in this study can be adapted to other catchments/countries to derive more accurate design flood estimates, in particular for important flood study projects, which require a sensitivity analysis to investigate the impacts of parameter uncertainty on design flood estimates.