Application of Monte Carlo simulation technique for flood estimation for two catchments in New South Wales, Australia

The currently adopted rainfall-based design flood estimation method in Australia, known as design event approach (DEA), has a flaw that is widely criticized by the hydrologists. The DEA is based on the assumption that a rainfall depth of a certain frequency can be transformed to a flood peak of the same frequency by adopting the ‘representative values’ of other model input variables, such as temporal patterns and losses. To overcome the limitation associated with the DEA, this paper develops stochastic model inputs to apply Monte Carlo simulation technique (MCST) for design flood estimation. This uses data from 86 pluviograph stations and six catchments from eastern New South Wales (NSW), Australia, to regionalize the distributions of various input variables (e.g., rainfall duration, inter-event duration, intensity and temporal patterns and loss and routing characteristics) to simulate thousands of flood hydrographs using a nonlinear runoff routing model. The regionalized stochastic inputs are then applied with the MCST to two catchments in eastern NSW. The results indicate that the developed MCST provide more accurate flood quantile estimates than the DEA for the two test catchments. The particular advantage of the new MCST and stochastic design input variables is that it reduces the subjectivity in the selection of model input values in flood modeling. The developed MCST can be adapted to other parts of Australia and other countries. Copyright Springer Science+Business Media Dordrecht 2014