Monday, 15 September 2014

Is the Australian Temperature record Accurate?

Opinion: Anthony Cox

Dr Marohasy
The Australian Bureau of Meteorology [BOM] prepares the Australian temperature record. This record is called Australian Climate Observations Reference Network–Surface Air Temperature dataset (known as ACORN-SAT, or ACORN).

ACORN is prepared by the BOM from readings of temperature from sites around Australia. These readings are the raw data.

Recently a group of researchers including Dr Jennifer Marohasy, have questioned the accuracy and reliability of ACORN because the final temperatures which form ACORN differ from the raw data. Dr Marohasy has found that the raw data has been adjusted or homogenised by the BOM so that warming appears where no warming or less warming was shown by the raw data.

Sometimes there is a valid reason for adjusting raw temperature data. In his 1996 thesis Simon Torok analysed the temperature sites around Australia and found a number of sites which required adjustments. These sites needed an adjustment because there was a discontinuity in the data. A discontinuity could be either a gap in the data where for some reason the data was stopped, or there was a fluctuation in the data which was inconsistent with what could be expected from the climate.

Torok provided some amusing and typically Australian examples of discontinuities such as Cockatoos stealing or destroying the thermometers and their screen and a suspicion that unusual hot records at a site were due to the site recorder increasing the temperature so his council worker friends could have the day off with pay.

A common reason for an unusual temperature is a move in the position of the thermometer.

These potential reasons for adjusting the temperature are called metadata. It is crucial that adjustments have some metadata reason to justify them and most importantly that the adjustment not increase or alter the trend in the raw data.

However it seems that adjustments at several sites have occurred without genuine evidence in the metadata about a discontinuity in the raw data.

In the analysis of the ACORN records compared with the raw data Dr Marohasy and the research group have found adjustments which increase or even create a warming temperature trend without any support from the metadata.

For instance at Bourke a long temperature record of over 100 years going back to 1880 has effectively been truncated to 2000. The researchers discovered the raw data showed an Australian maximum temperature record of 51.7ºC on 3rd January 1909. This record is no longer used by BOM. 

Overall Bourke raw data show a cooling maximum temperature trend from 1880 of 1.7ºC per century. After adjustment ACORN at Bourke shows a slight warming temperature trend. This change of maximum temperature trend has a great effect on the whole of Australia’s temperature record. The BOM has offered no particular reason for these adjustments.

In the Newcastle region the temperature site at Williamtown also shows a marked difference between raw minimum temperature data and the ACORN temperature after homogenisation by BOM.

Researcher Ken Stewart has shown that adjusting the minimum temperature data can also affect
the overall temperature trend. Stewart’s research shows a consistent warming bias for the minimum temperature over all ACORN sites in Australia. This is uncontroversial with BOM head scientist Dr Blair Trewin conceding:
“negative adjustments are somewhat more numerous for minimum temperatures, which is likely to result in ACORN-SAT minimum temperatures showing a stronger warming trend than the raw data do"

In the case of Rutherglen because the BOM has decreased the minimum temperature backwards from 1973 to 1913 the difference between the raw data and the adjusted data in 1913 is 1.8C. BOM has said this was justified because the thermometer site at Rutherglen was moved. 
However, retired
natural resources scientist Dr Bill Johnston, who worked at Rutherglen says that no evidence of a site move at Rutherglen has been provided by BOM.

Government policy is based on climate data which needs to be reliable and this includes temperature records. If well-credentialed amateurs have revealed some possible faults with the temperature record then there is an obligation on government departments such as the BOM, which advise and determine government policy, to consider those possible faults in an open and transparent manner