Characteristic thunderstorm distribution in the Sydney area

The spatial distribution of deep convection in the greater Sydney area is examined. The primary source of data is the incoherent, broadbeam 10 cm radar at Sydney’s Mascot Airport. For 25 years significant storm cells on the radar screen were manually logged. A composite storm cell density map is produced. Under certain assumptions, this map can be converted to a storm probability distribution. Such distribution can be produced for any subset of the entire database (e.g. a given time of the day), and the anomalous storm probability under the conditions of the subset can be examined. We focus on anomaly maps for a set of six synoptic situations in which thunderstorms occur most commonly. These synoptic classes are defined solely on the basis of mean sea-level pressure (MSLP) patterns, because MSLP charts are available at the highest time and space resolution, and because MSLP, via its relation to the low-level flow, is an important synoptic-scale discriminant for local storm distribution. We find that in some areas the low-level flow has a significant impact on storm distribution. These anomaly maps directly link the local storm predictability to the regional weather predictability. The results are compared to data from much more recent, more objective but climatologically inadequate alternative data sources: automatically recorded significant radar echoes, a lightning network, and a storm spotters network. While the comparison is difficult, some basic patterns are verified.