Investigation of sub-tropical rainfall characteristics for use in the joint probability approach to design flood estimation

![CDATA[In Australia, for design flood estimation there has been a move to the use of Joint Probability Approach/ Monte Carlo Simulation (MCS) or Continuous Simulation modelling rather than the Design Event Approach as outlined in the Australian Institution of Engineers’ guideline known as Australian Rainfall & Runoff. This use of the MCS approach, on which this paper focuses, requires definition of independent rainfall events and their subsequent statistical analysis. However, the MCS methodologies to date have been mostly for areas with temperate climates. This paper considers the application of the MCS to sub-tropical areas of Australia. In Australia there are two prevailing definitions of independent events – the first named a complete storm is specified by a minimum gap between events and removal of pre and post event rainfall not exceeding some rainfall rate threshold criterion. The second definition is a subset of a complete storm and is the most rare rainfall burst within the complete storm. The rainfall depth and pattern associated with the most rare storm duration is known as the storm core. The paper shows that rainfall analysis adopting typical independent storm definition parameters for sub-tropical areas results in differing statistical storm durations than those reported for more temperate areas. In particular while complete storm durations are exponentially distributed, core storm durations’ distributions are non-exponential and a gamma distribution can be adopted. The application of multiplicative-cascade model for temporal pattern distribution is discussed in the paper. The relationship between rainfall intensities of complete storm and storm core is also investigated. The results of this study form an important basis of the application of the runoff routing model URBS with MCS technique in sub-tropical part of Australia.]]