An investigation into extreme rainfall variability in Australia

Abstract

The main objective of this project was to investigate variability in Australian extreme rainfall. This was achieved through the use of observational, reanalysis and climate model datasets. Firstly a gridded rainfall dataset was assessed against station observations and found to be suitable for the study of extreme rainfall variability across much of Australia. An empirical orthogonal teleconnection method was applied to the gridded dataset to examine the climate modes and atmospheric conditions associated with extreme rainfall. Variability in extreme rainfall was observed to be more closely related to mean rainfall variability in summer than winter. Strong associations exist between some climate modes, such as the El Niño-Southern Oscillation (ENSO), and extreme rainfall. Lagged regressions of sea surface temperatures onto summer extreme rainfall show potential predictability of extreme rainfall arising from the Pacific and Indian Oceans. A non-linear ENSO-extreme rainfall relationship was observed over eastern Australia using the gridded rainfall dataset. A reanalysis product was found to also capture this non-linearity. An attribution study was performed on the extreme rainfall over southeast Australia in 2011-12. Data from state-of-the-art climate models were used to show that the role of anthropogenic climate change in this event was limited in comparison to the La Niña event. The ENSO-rainfall relationship over eastern Australia was investigated in more detail in reanalysis and state-of-the-art climate models. The non-linearity in the relationship is associated with moisture availability over the south-west Pacific. Variability in thermodynamic processes also explains the differing ENSO-rainfall relationship in climate models. Climate projections in the best-performing climate models, based on their ENSO-rainfall non-linearities, point to lower increases or even decreases in east Australian extreme rainfall compared with the entire multi-model ensemble. Sources of uncertainty in projections dependent on different methods were also examined. This thesis documents a number of important findings that improve our knowledge and understanding of Australian extreme rainfall variability and relationships with climate modes, particularly ENSO.