Extreme rainfall represents a major hazard to life, property, and infrastructure, yet remains a significant challenge for operational forecasting, seasonal prediction, and long-term climate projections, due to the complex, multiscale processes involved in precipitation formation. Improving our understanding of rainfall extremes and the underlying physical processes, particularly for Australia, is one of the key objectives of the ARC Centre of Excellence for Climate Extremes. As part of this effort, the present study investigates the synoptic patterns associated with daily rainfall extremes for the densely populated cities of Brisbane, Sydney, and Melbourne. In particular, we look to answer the following question: “What ingredients, at the synoptic scale, distinguish extreme rainfall events (EREs) from lesser heavy rainfall events (HREs)?”

For the 40 year period from 1979 to 2018, HREs are identified in each domain based on daily rain gauge records, using thresholds that correspond to the 95th percentile of wet-day (≥ 1 mm) rainfall for each gauge. These events are then subdivided into EREs and non-EREs depending on whether rainfall exceeds the 99th percentile for wet days. Cluster analysis, applied to reanalysis data, is used to classify the dominant synoptic patterns associated with HREs in each domain. Then, for every cluster, ERE and non-ERE composites are produced and compared to identify the key synoptic ingredients associated with extreme rainfall in the three regions. As well as providing insight into historical events, this analysis suggests a framework for assessing the representation of rainfall extremes in climate models based on the associated synoptic-scale circulation.