With increased interest in the potential effect of anthropogenic climate change on the frequency and intensity of tropical cyclones, the Australian Bureau of Meteorology has undertaken a project to improve the available track and intensity data for the Australian region (0-40°S, 90-160°E).

The previously existing Australian region data set extends back to 1906. There are, however, a number of known issues with this data set, including:

• General underestimation of cyclone intensity prior to the introduction of satellite-based (Dvorak) techniques into routine analysis in 1984, and especially prior to the availability of the first regular satellite data in 1969.

• Missing systems, especially over the open ocean away from the coast, in the period prior to 1969, and especially prior to 1955.

• Inclusion of systems which are extratropical, or too weak to meet current definitions of a tropical cyclone, particularly between 1955 and 1970.

• Duplicate systems and inconsistent definitions of what constitutes a single cyclone, especially where a system regenerates after a period over land.

• Inconsistent definitions of the Australian region boundary.

• Inconsistent methods of intensity estimation (including the use of multiple pressure-wind relationships).

• A lack of information, prior to 1985, on intensity parameters other than central pressure.

The major focus of the first stage of this project has been to develop an accurate record of the number of tropical cyclones, and of key aspects of their tracks (such as coastal crossings and near-coastal approaches). The most important part of this has involved consolidating various existing data sets and resolving inconsistencies between them, in addition to the correction of other gross errors. (Major existing data sets, in addition to the Bureau's own set, include that of the Joint Typhoon Warning Center (JTWC), limited reanalyses carried out by the Bureau's Brisbane and Darwin offices and by private consultants on behalf of local resource interests, and written seasonal summaries published by the Bureau or in journals).

The second stage of the project, which has not yet commenced (as of July 2007) is a full-scale reanalysis from original source data, particularly satellite images. The major aspects to be addressed here will be consistency of intensity estimates, and of the start and end points of tracks (including the point of extratropical transition where applicable). It is expected that this will allow the development of a reasonably homogeneous record of tropical cyclone intensity in the Australian region back to about 1970, whilst it is also expected to be possible to develop a good record of coastal crossings back to about 1950 for the Australian continent as a whole, and to near 1900 for the more densely settled parts of the eastern Australian coastline.