Below is an original web page built in early January 2012 to demonstrate the rounding errors common to the BoM's Fahrenheit and Celsius temperature recordings. See audit background.

Western Australia temperature trend analysis

Bureau of Meteorology climate records in Western Australia are influenced by the metrication of temperature scales in 1972 and erroneous thermometer readings in earlier years.

Rounded decimal trends and Fahrenheit errors creating an unknown but probable warming bias can be found in BoM temperature data that the Australian government accepted as accurate when legislating its 2012 carbon tax.

These influences are discussed in Australia's seminal study into the creation of an accurate historic temperature record:

"Some doubts remain regarding possible discontinuities in the early 1970s, when Australian temperature recordings changed from the Fahrenheit scale to the Celsius scale. New thermometers were issued at that time. Thermometers before and after the change to metric units underwent calibration tests and a search through the documentation regarding the change does not suggest possible instrumental discontinuities. As well, measurements with thermometers manufactured before and after the changeover date have been compared. This check revealed no instrumental reason for expecting discontinuities at the time of the change over to metric units.

One possible source of a non-climatic discontinuity is observer practice in recording temperatures. Some observers, prior to the change to metric units, recorded temperatures in whole degrees Fahrenheit, instead of recording to the nearest tenth of a degree, as specified in directions to observers. If many observers 'truncated' their measurements to the nearest whole degree below the actual measurement, prior to metrication, and after metrication recorded to tenths of a degree, this would result in an artificial warming in the early 1970s.

Discontinuities caused by such a practice, if it was common, would not be detected by the statistical programs used here. Examination of field books does not suggest that this practice was sufficiently common to produce a major discontinuity."

A historical annual temperature dataset for Australia (Torok and Nicholls, BoM Research Centre, Melbourne, 1996)

Weather stations around Australia recorded an upward shift in their mean temperature plateau in the early 1970s and this page addresses the issues by examining decimal fractions for trends at 37 BoM weather stations across Western Australia, examining all years of daily temperatures with data as early as 1897.

The available data demonstrates that:

• significant periods of raw temperature records at most BoM stations have months, years or decades with all daily temperatures exactly .5 or .0C, and in some instances these rounded degrees are in both raw and High Quality BoM datasets recognised as the official gauge of Australia's climate history
• most .5 and .0C daily temperatures in the weather station raw records have happened since September 1972, the month Australia's weather services were metricated and the Fahrenheit temperature scale was replaced by the Celsius scale
• most temperature readings before September 1972 were recorded as a full rather than a tenth of a Fahrenheit degree with consequently inaccurate readings that "would result in an artificial warming in the early 1970s"

Method

This analysis categorises ascending lists of BoM raw and HQ daily temperatures databases where available for 37 temperatures stations across Western Australia, counting only the number of .5 and .0C decimal fractions.

It then compares the total count of .5 and .0C degrees over the full duration of the station record with readings since September 1972, comparing the pre and post totals as a proportion of the station's timescale.

The total number of all decimals fractions from .0 to .9 is provided for each station, and a pre and post 1972 HQ comparison of average temperatures is provided where available.

Discussion

Frequently long and unbroken periods with raw data rounded at .5 and/or .0C only occur after 1972 and are of sufficient magnitude to affect the temperature trend accuracy of various weather stations.

Temperature records at most WA stations have unbroken months or years of .5 and .0C and/or periods when these rounded degrees are in high proportion.

The BoM has explained in the case of Perth Metro 9225 that the precise temperatures for the affected time period are available but have not yet been supplied to the BoM's national database.

The BoM argues that although in the case of Perth Metro the rounding can produce differences between 0.0C and 0.3C, the effect is negligible and the precise data rounds to exactly the same as the rounded data.

Most weather stations including those within the BoM HQ dataset consistently have two, three and sometimes four times the proportion of .5 and .0C days occurring in their record after September 1972 when Australia's temperature scale switched from Fahrenheit to Celsius.

The decimal fraction of .0C is by far the most frequent degree recorded at most stations and the least frequent is .5C. This analysis counts both as they are the fractions that can occur together over months or years to the exclusion of all other temperatures.

Torok et al (see above) find no evidence of discontinuities between Fahrenheit and Celsius thermometers but fail to consider a systemic or non-instrument bias.

The data in this analysis (click menu at left for stations) shows that abnormally low .5C and abnormally high .0C relate to the date of September 1972, suggesting that the conversion from Fahrenheit to Celsius has influenced WA's historic temperature trend.

This analysis notes that a large number of weather station records have temperatures before 1972 that convert from Celsius to Fahrenheit only as rounded degrees. For example, some BoM stations have unbroken periods or are peppered with temperatures such as 20C, 20.6C, 21.1C, 21,7C, 22.2C, 22.8C, etc (= 68F, 69F, 70F, 71F, 72F, 73F, etc). This excludes or is biased against temperatures that may have been 68.4F (20.2C), 70.8F (21.6C), 72.5F (22.5C), 69.9F (21.1C), etc.

The most likely explanation is that, as noted above to Torok et al, large numbers of observers recorded temperatures in whole degrees Fahrenheit instead of recording to the nearest tenth of a degree.

If observers were rounding to the nearest Fahrenheit degree, it is most likely they would round down to the closest whole degree visible on the thermometer, some possibly rounding down from more than .5F or even as much as .9F.

Torok et al acknowledge that If many observers 'truncated' their measurements to the nearest whole degree below the actual measurement prior to metrication, this would result in an artificial warming in the early 1970s.

Their research indicated that examination of field books did not suggest the practice was sufficiently common to produce a major discontinuity in the temperature record, but this analysis suggests field books were an inaccurate source for research.

With Australia's temperature trend rising sharply as of 1972 and overwhelming raw data evidence that metrication combined with the cessation rounded temperature observations, it seems a likely explanation for "an artificial warming in the early 1970s".

The BoM will argue that this analysis mostly studies raw instead of adjusted HQ temperatures and is thus invalid. This analysis finds commonalities between the raw and HQ data and nevertheless reasons that if the HQ data is based on the raw data, the latter is where errors are likely to be found.

This analysis is an amateur audit of BoM temperature data in Western Australia over the past 100 years. The BoM has almost 3,000 observation posts around WA and this analysis studies only 36, but they have the longest timescales and most BoM stations don't have electronic database records.

With the advent of a carbon tax that aims to alter the economic and lifestyle fundamentals of Australia, why is the BoM temperature record not audited to identify data patterns indicating possible bias?

This website produced by Scribeworks 2009/12