Towards the application of the Joint Probability Approach to ungauged catchments : regional distribution for storm duration in Eastern Victoria

In recent years, there have been significant researches in Australia on the development and application of the Joint Probability Approach/Monte Carlo Simulation technique to design flood estimation. The superiority of the Joint Probability Approach is based on the fact that this accounts for the probabilistic nature of the major input variables to the runoff routing/unit hydrograph model in a more explicit manner. The application of the Joint Probability Approach so far has been limited to gauged catchments with reasonably long rainfall and streamflow records. However, in practical situations, many catchments are ungauged where there is little or no data available to identify the probability distributions of various input variables. To apply the Joint Probability Approach to ungauged catchments, it is necessary to regionalise the distributions, which involves the analysis of rainfall and streamflow data available in the region. This paper presents the regionalisation of distribution of rainfall duration in eastern Victoria. This uses pluviograph data from 49 stations across the region. Three goodness-of-fit tests were adopted. Chi-squared test, Kolmogonov-Smirnov test and Anderson-Darling test. It has been found that majority of the pluviograph stations in eastern Victoria satisfy regional exponential distribution.