This search will pick up (or should pick up) anything in the Bureau's electronic archives. For most sites these extend back to about 1957 (if the site was open then), although capital city sites extend further, and the extremes (particularly in NSW) extend further at a limited range of sites. This means that many extremes which appear in other sources do not appear in these lists. In general, these lists should only be considered as a reflection of extremes over the post-1957 period. Pre-1957 data is likely to be included progressively as manuscript data is digitised. The lists will be revised periodically to include this data.
I am aware that some sites, notably recently-opened automatic stations (AWS), have not been included in the monthly data tables from which these records were generally taken. I've checked daily data for some of the more obvious potential record-breaking sites (mainly alpine ones), but it is possible that some have slipped through the net.
In general, the daily maxima and minima are taken for the 24 hours to 0900 local time. There are some exceptions to this, notably recent AWS data which is read at 0000 and 1200 UTC, and major stations between 1932 and 1963 which reported on an 0000-0000 basis. It should also be noted that the years when extremes have been recorded may be a reflection of when stations were open; for example, the only periods when there have been stations open above 1200 metres in Tasmania have been between 1961 and 1972, and since 1990.
All values in these tables have been checked for internal consistency with observations at fixed hours, and against neighbouring stations, where there are any. Several notable observations have been excluded, most significantly the -20.6 at Charlottes Pass in August 1968 and the -19.5 at Perisher Valley in July 1979, the fifth and sixth lowest readings on record in Australia. Both were found to be suspect on undergoing the aforementioned checks.
There are also numerous extreme high temperatures which have been recorded prior to about 1910 using non-standard instrumentation, most notably a reading of 53.1 at Cloncurry in January 1889. It is likely that this will be struck from the official record in the near future. A discussion of the evidence behind this may be found in:
Trewin, B.C. (1997). Another look at Australia's highest temperature. Aust.Met.Mag. 46. 251-256.
Observations known to have been taken using non-standard instrumentation have been excluded from these lists.
The extremes tables have been compiled in two forms; by state and by month.
The tables are current as of March 4, 1998, although some of the most recent records are taken from real-time observations and should be regarded as provisional until the field books are submitted. As real-time observations are only, in general, communicated in whole degrees, the final observation may differ slightly from that shown here at present.
All readings are in degrees C.