Several observational datasets have been developed to identify, monitor and attribute variations and changes in the Australian climate. These datasets have been produced using a variety of quality control and homogenisation techniques to ensure they are comparable through time and free of spurious non-climate influences which might otherwise mask real trends.
Discontinuities (i.e. erroneous shifts) in climate data through time may be caused by changes in the location or exposure of the observation site, instrumentation type or observation practices. For example, prior to 1910 temperature measurements were often taken using non-standard temperature screens which gave different values to those taken in modern Stevenson Screens. For this reason temperature observations before this period are not currently included in Bureau datasets used for monitoring climate change.
Homogenisation of data allows discontinuities stemming from non-climate influences to be removed from the data before long-term changes are investigated. Procedures to identify non-climatic changes in historical climate data generally involve a combination of:
- investigating historical information (metadata) about the observations,
- using statistical tests to compare records from nearby locations, and
- using comparison data recorded simultaneously at old and new locations, or with old and new instrument types.
The Australian Climate Observations Reference Network – Surface Air Temperature (ACORN-SAT), monthly pan evaporation and monthly cloud amount datasets have been temporally homogenised. Rainfall data has not been homogenised in either the monthly or daily datasets; stations suspected of being affected by an inhomogeneity or data quality problem during their record were excluded from the dataset leaving a network composed of only the best and most reliable stations.
Scientific papers describing the homogenisation process have been published for each applicable dataset. This information and the datasets are available below.
- Daily temperature: ACORN-SAT
- Daily rainfall: Data | Information (PDF)
- Monthly rainfall: Data | Information (PDF)
- Monthly pan evaporation: Data | Information (PDF)
- Monthly cloud amount: Data | Information (PDF)
These datasets underlie operational monitoring of Australia's changing climate.
Content of the site networks application
The site networks application provides access to homogeneous historical climate data for Australian observing sites. Climate variables available include temperature, rainfall, pan evaporation and cloud amount, at timescales ranging from daily data through to annual averages.
Climate data are provided in both graphical and numerical formats for individual station sites. Graphs can be viewed as either mean/absolute data or as anomalies from the standard 1961–1990 base period (1971–2000 is used for pan evaporation due to shortness of record of the available data). Daily data are only plotted from 1995 onwards to avoid over-crowding the graphs, but the full daily record is available in the provided text files.
Basic supporting information (metadata) about each site is given to help you choose the most appropriate location.
The data format for daily data is date (yyyymmdd) in the first column, with data values in the second column. For all other timescales (e.g. monthly or annual averages/sums) the data format is start date (yyyymmdd) in the first column, end date (yyyymmdd) in the second column, with data values in the third column. Throughout all data series 99999.9 is used to denote missing data for that particular time. Data that is identified as missing is only applicable to these datasets. It is possible that there may be raw (i.e. unhomogenised) data for that period, but as a result of data quality control processes, these may now be denoted as missing in the current datasets. Further, new data is constantly being added to the raw databases as the large body of historical paper records are entered and these data take some time to flow into the homogenised datasets.
The daily rainfall dataset contains some multi-day totals of rainfall. These are identifiable as totals which follow one or more days of missing data (99999.9). Any rainfall recorded on the first day following a period of missing data may have fallen on any of the previous days with missing data, or may be an accumulated total (i.e. the sum total of rain which fell on more than one day).
Site data temporal averaging methods
For maximum and minimum temperatures, monthly values are calculated from the average of the daily temperature data from the ACORN-SAT dataset. There should be no more than 15 days missing in any month for a monthly value to be reported. Annual (and seasonal) values are calculated from the simple average over the applicable months. If there are insufficient daily data that results in a missing month, then the annual value (and applicable seasonal value) for that year (and season) will also be marked as missing.
For mean temperature, the average of the maximum and minimum temperatures is used for the appropriate time period; e.g. the mean monthly temperature for a given month is (monthly maximum + monthly minimum)/2. Therefore the annual mean temperature for a given year is the average of the annual maximum and annual minimum temperatures. A similar calculation is used for seasonal, monthly and daily mean temperatures.
It is possible that a station with more than 15 days with missing daily mean temperatures in a month could still report a monthly mean temperature value. This could occur if each of maximum and minimum temperature are missing on 15 days or fewer, but there are more than 15 days on which at least one of the two is missing.
The anomaly data are calculated with respect to a 30-year climate normal (1961–1990) for the appropriate month, season or annual period. The daily anomaly data are calculated with respect to the corresponding monthly normal (e.g. each day in January is referenced to the January monthly average). The climate normal for a given month is the average of all individual monthly values for that month over the 1961–1990 reference period. The annual climate normal is the average of the 12 monthly climate normals. The seasonal climate normal is the average of the three applicable monthly climate normals.
Annual and seasonal totals are calculated using monthly rainfall from the monthly rainfall network. The daily data are from the daily rainfall network. Annual totals are calculated from the simple average over the 12 calendar months.
The anomaly data are calculated with respect to a climate normal (1961–1990) for the appropriate month, season or annual period. The daily anomaly data are calculated with respect to the corresponding monthly normal.
Annual and seasonal values are calculated using monthly evaporation from the monthly pan evaporation network. Annual values are calculated from the simple average over the 12 calendar months.
The anomaly data are calculated with respect to a climate normal (1971–2000) for the appropriate month, season or annual period.
Annual and seasonal amounts are calculated using monthly cloud amount from the monthly cloud amount network. Annual amounts are calculated from the simple average over the 12 calendar months. For the daytime mean cloud amount, the average of the 9 am and 3 pm cloud amounts are used for the appropriate time period. Therefore the annual daytime cloud amount is the average of the annual 9 am and annual 3 pm cloud amounts, likewise a similar calculation is used for the seasonal and monthly mean cloud amounts.
The anomaly data are calculated with respect to a climate normal (1961–1990) for the appropriate month, season or annual period.
Features of the site networks application
The map displays the location of observing sites with available data for the variables, timescales and regions you select. Clicking on a site dot will display corresponding data. You may then select alternative timescales or switch to a neighbouring site. You may also switch to other variables available for the same site using links in the right-hand panel.
Mean/absolute and anomaly data can be downloaded using links in the right-hand panel. The entire datasets that these individual series come from are available for download from the links near the top of this page.