Statistical Modeling of Extreme Rainfall in Southwest Western Australia

Abstract

Rainfall over southwest Western Australia (SWWA; 32°S southward and 118°E westward) has been decreasing over the past decades, putting further constraints on water resources in an already dry area. In this study, daily rainfall over five geographically dispersed and homogenized weather stations within SWWA are analyzed. A peak over threshold method from the extreme value theory is used to model daily rainfall above a given threshold. The Mann?Whitney?Pittitt (change point) test was applied to detect changes in annual, winter (May?October), and summer (November?April) maximum daily rainfall. Change points for winter extreme daily rainfall were found around 1965, based on different individual stations, with the extreme daily rainfall reduced since then. To demonstrate the degree of change in the winter extreme daily rainfall, at 1965 the data were stratified, and generalized Pareto distributions were fitted to the tails of the distributions for daily rainfall in the prechange period of 1930?65 (including 1965) and the postchange period of 1966?2001. The fitted tail distributions also allow the estimation of probabilities and return periods of the daily rainfall extreme. Results show that return periods for the winter extreme daily rainfall have increased after 1965, implying that winter daily rainfall extremes in SWWA are lower after 1965 than they were before. There has been vigorous debate as to what forces the drying trend, that is, whether it is part of multidecadal variability or whether it is driven by secular forcings, such as increasing atmospheric CO2 concentration. In this paper, statistical modeling is also used to identify possible associated changes in atmospheric circulation. It is found that there is a change point near 1965 in a dominant atmospheric circulation mode of the Antarctic Oscillation (AAO). The result offers qualified support for the argument that the AAO may contribute to the drying trend.