One of the key aspects to better managing water resources in Australia is to understand the causes of medium- to long-term rainfall variability, which results in both droughts and periods of above average rainfall and flooding. Much of the research on this variability has focused on the El Ni¿o--Southern Oscillation (ENSO) phenomenon, using methods that assume the relationships between ENSO and Australian rainfall are both linear and stationary. In this paper we present an alternative approach based on wavelets to analyze the dominant modes of variability in three rainfall characteristics: (1) the total annual rainfall, (2) the annual number of wet days, and (3) the maximum annual daily rainfall. We then use a wavelet regression approach to examine the extent of the variability that can be associated with ENSO. The results show that time series of total annual rainfall and annual number of wet days exhibit significant variability at periods of 2.6, 4.6, 7 and 13 years in various locations throughout the country and that these periodicities are not caused directly by the ENSO phenomenon. While maintaining that ENSO still plays a significant role in influencing rainfall variability in Australia, these results highlight the importance of looking beyond ENSO to identify dominant sources of variability in the characteristics of annual Australian rainfall that were studied. In contrast, no coherent modes of variability could be found for the maximum annual daily rainfall time series, highlighting the greater level of random behavior in the intensity of larger rainfall events compared with the long-term averages.