Probabilistic losses for design flood estimation : a case study in New South Wales

![CDATA[Australian Rainfall & Runoff currently recommends the design event approach for rainfall-based flood estimation. This method ignores the probabilistic nature of model inputs, such as losses, which can lead to significant inconsistences and bias in design flood estimates. Recently, the joint probability approach has been proposed to allow for the probabilistic nature of key inputs and their correlation; however, before this can be applied in practice, a probability distribution needs to be derived for each input. This paper investigates the probability distributions of losses and examines the impact of these inputs on flood estimates, for the Orara River Catchment in North-East NSW. For each loss parameter, 500 random values were generated, so that the HORB model could be run in a semi-stochastic manner. Results show that the use of distributed inputs can lead to more realistic flood estimates as compared to the design event approach.]]