Study of fixed-duration design rainfalls in Australian rainfall runoff and joint probability based design rainfalls

Rainfall, in terms of its intensity, duration and temporal patterns, known as design rainfall, is the basic input to many hydrological models. Estimation of design rainfall is based on frequency analysis of historical rainfall data. The design rainfalls in Australia are provided in its national guideline known as Australian Rainfall and Runoff (ARR)" a guide to flood estimation, published by the Institution of Engineers Australia. An alternative approach to ARR method is a joint probability based method, which considers rainfall event duration as a random variable. This allows accounting for the joint probabilistic nature of the rainfall duration and depth in a more rigorous manner. This joint probability based design rainfalls can be used to estimate design floods using a Joint Probability Approach, which considers the probabilistic behavior of the input variables and model parameters in an explicit manner in the flood estimation. This paper investigates the design rainfall estimates using the two above approaches using rainfall data from Victoria state in Australia. This in particular examines the sampling properties of the rainfall events in the two approaches to identify any systematic differences between them. It has been found that ARR design rainfall estimates are generally higher than the joint probability based estimates, however these differences vary with locations, durations and average recurrence intervals. The results of this study offer insights into the two methods of rainfall analysis, which will assist in developing a link between the two types of design rainfalls.