Tuesday, April 6, 2010

Where is the Quality Assurance?

When the BoM+CSIRO released its "State of the Cliimate" prolefeed brochure in mid-March, the BoM website listed 103 climate stations that comprised the RCS network. At that time, this listing included the following three stations:
  1. 010035 - Cunderdin (1914 - )  [data] [graph]
  2. 029004 - Burketown PO (1886 - )  [data] [graph]
  3. 078031 - Nhill (1897 - 2009)  [data] [graph]
The proof of this is that - as at today - these three stations still have their photographs and station details included in the RCS "Site Details" database, and clicking on the station numbers above will take you to these records.

Following the coverage of the "Dodgy Post Offices" item (which included Cunderdin) on the Andrew Bolt blog, Cunderdin was summarily dropped from the RCS network. Dropped at the same time was Nhill - a station which had actually been closed last year. Substitute stations were added to the RCS network for both of these dropped stations. Just like that!

After completing the "Burketown - Which One?" item, it was pointed out that the 029004 station had also been summarily dropped from the RCS network listing.

As at today, the following three stations have been substituted for the ones noted above:
  1. 010286 - Cunderdin Airfield (1942 - )  [data] [graph]
  2. 029077 - Burketown Airport (2001 - )  [data] [graph]
  3. 078015 - Nhill Aerodrome (2003 - )  [data] [graph]
Apart from the glaringly obvious quality control issue whereby a station which has been closed a year ago (078031 - Nhill) continues to be a "high quality, long-term climate monitoring station", it also raises the question of just how valid it is to substitute newer stations for ones with a data history stretching back over a century.

Taking these three in sequence, we notice that:
  • Cunderdin, a station which began in 1914 (but only has data since 1950) has been replaced by Cunderdin Airfield, which began in 1942 (but only has data since 1996). A loss of 46 years' worth of data for this location.
  • Burketown PO, a station which began in 1886 (and has data since 1890) has been replaced by Burketown Airport, which began in 2001. A loss of 111 years' worth of data for this location.
  • Nhill, a station which began in 1897 and was closed in 2009, has been replaced by Nhill Aerodrome, which began in 2003. A loss of at least 106 years' worth of data for this location.
In simple terms, for Cunderdin, this:

...has been replaced by this:

For Burketown, this:

...has been replaced by this:

And for Nhill, this:

...has been replaced by this:

Apart from the obvious issue of older stations being substituted with newer stations, where are any of these changes documented? How can a closed station be simply left in the network for a year? And how can one station simply be substituted for another on the basis of bad publicity?

When the CSIRO+BoM released their Climate brochure in mid-March, Greg Ayers, Director of the BoM, was interviewed by Sarah Clarke on the ABC, and he offered the following self-assessment:

GREG AYERS: For the Australian region, we have around 100 climate reference stations, as we call them, where we pay a great deal of attention to doing the best possible measurements, making sure that there's quality assurance on the way the data is brought in and then used
Where is this quality assurance process and where is it documented? How do the above three station pairs fit within this quality assurance process? How much of this statement is a work of fiction?


In their Climate brochure, the CSIRO+BoM make the assertion that "since 1960 the mean temperature in Australia has increased by about 0.7°C". Over a 50 year period, this represents about 0.014°C per year, and for this conclusion to be drawn, some exceedingly accurate record keeping would clearly be required. Of the stations which have been examined so far, it is hard to see how either of these assertions (namely, a quality assurance process or the temperature increase) can be supported.

No comments:

Post a Comment