Ken Stewart, April 2010
Cairns, Queensland’s most northerly tropical city, is one of Australia’s premier tourist destinations. Once a quiet port city for the Atherton Tablelands and Cape York Peninsula, it is now the gateway to the Great Barrier Reef, Port Douglas, the Daintree, Kuranda and the Skyrail. Cairns is the 4th most popular destination for International Tourists after Sydney. Cairns’ population has grown rapidly since the 1980s to 122,731 (2006).
The Bureau of Meteorology (BOM) has two temperature records from Cairns, the Post Office (until 1952) and Cairns Aero, at the Airport. Let’s have a look at Cairns’ climate history.
First, some peculiarities:
GISS (Goddard Institute for Space Studies) regard Cairns Aero as being urban.
BOM regard it as being rural (although in 1996 it was Urban; and Innisfail, a town of some 8,000, is listed by BOM as Urban!)
My analysis runs from 1910 to 2009. I downloaded maximum and minimum annual means from the BOM Climate Data Online site, calculated the average of this, and plotted the results. I also downloaded the GISS pre- and post-homogenisation data and plotted it as well. Finally I downloaded and plotted the Cairns data from the Australian High Quality Climate Site page.
Ready for some fun?
Cairns- PO and Aero
I spliced these for ease of comparison.
Trend of 0.65 degree Celsius per 100 years. In the 1980s it became an International Airport. Bulldog44, in a comment on this site, writes:
“I was transferred to Cairns by Ansett in early 1984 to take over as manager of the Cargo operation. At that time our shed was just next to the tower where the temperatures were taken (and I presume still are).
At that time we handled 4 domestic flights a day, one cargo aircraft and one international flight a week (to a mine in Irianjaya).
After the new air port opened the number of aircraft movements exploded in just a few years to the point where my staff were handling 12 domestic flights a day and 60 international flights a week. And an entirely new International Passenger Terminal had to be added.
The relevance of this being that the engineering base(s) were directly in front of the tower and all the jet engines were tested, blasting their exhaust directly back over the area where the temperature equipment was stationed. That, combined with the extraordinary tarmac extensions adding vast areas of extra concrete to the air port could hardly have failed to affect any readings (in my opinion, of course).”
So it has a recognisable UHI signal!
GISS, in their 2 part homogenisation process, combine all temperatures at a location, then adjust this with the mean of the trend of “nearby” rural stations (up to 250km or 1200km away.)
To their credit, the first part of the process appears to correct for UHI:
A trend of 0.2 C/100 years.
But then, homogenisation increases the warming again to 0.6C.
How does that work? Well, I’m a little tired of trying to work out the GISS shenanigans, so for the sake of argument let’s assume the adjustment is perfectly correct in every way.
Thus, there are now three temperature records for Cairns. But wait, there’s more!
Cairns Aero is part of the Australian High Quality Climate Site network, a set of stations whose data is used to develop the Australian climate record. Only rural stations, not urban, are used in the climate record, theoretically to exclude UHI contamination. Note Cairns Aero is rural!
However, the “High Quality” refers to the fact that the data from all of these sites was adjusted in a homogenisation process by Torok and Nicholls (1996) and another unspecified process in 2004.
This is the result! Cairns Aero, High Quality:
The trend is now 0.9 degrees per 100 years. And this is used to develop our climate record.
So- Cairns has:
BOM raw data (trend: 0.65C/100 years)
GISS preadjusted ( 0.2C/100 years)
GISS homogenised (0.6C/100 years)
BOM High Quality (0.9C/100 years)
Confused? So am I. Will the real Cairns climate please raise its hand?
References:
Torok, SJ, and Nicholls, N (1996) : A historical temperature dataset for Australia, Australian Meteorological Magazine, 45 (1996), pp 251-260.
http://www.bom.gov.au/other/disclaimer.shtml
http://www.bom.gov.au/other/copyright.shtml#acknowledgements
kenskingdom 
April 19, 2010 at 4:07 pm
Gday folks, new post. Please remember, keep it civil, give references, don’t waste time, and stick to the topic. JM, you’re welcome back!
April 20, 2010 at 10:44 am
It would be interesting to know the complete reference(s) for: “Torok and Nicholls (2006) and another unspecified process in 2004.”
And perhaps a summary of the methodology/ philosophy used to make the adjustments.
The question arises – is or are these processes used for all of the so called “high quality” data? Depending on the answers one may need to put the “high quality” data on notice.
April 20, 2010 at 2:32 pm
Torok, SJ, and Nicholls, N (1996) : A historical temperature dataset for Australia, Australian Meteorological Magazine, 45 (1996), pp 251-260.
Did I really write 2006? Will update.
I have written to BOM requesting details of the 2004 process.
The methodology is pretty involved- briefly, 224 stations selected, minima and maxima discontinuities checked by calculating the median of year to year temperature changes of stations up to 6 degrees long. and lat. away but not over mountain ranges, then adding this to the first year of record, checking for discontinuities visually then subjectively deciding which to adjust and by how much. This was then checked with the Detect software. Annual means and diurnal ranges were calculated. The results were checked for reliability by comparing with Jones (1994) (!!!) who used many fewer stations including capital cities, and by correlating with annual rainfall , which not surprisingly correlate well.
To answer your question- yes. I will be posting about the HQ data for Qld within a week or so, and believe me, the Cairns adjustment is quite mild when compared with some others. I am asking BOM to respond about these others before posting.
Stay tuned.
KS
April 20, 2010 at 12:35 pm
“Only rural stations, not rural, are used in the climate record, theoretically to exclude UHI contamination.”
Should this read “Only rural stations, not urban, are used….”
Lies, damn lies and statistics is forever at work with these temperature records, hence the difficulty a lot of us have in believing the ‘science’.
April 20, 2010 at 2:38 pm
John: Thanks! Looks like I need readers to do my proof reading for me- and I’m a school teacher! The shame.
I’ll have a post maybe next week re BOM’s HQ data set.
KS
September 14, 2010 at 9:17 pm
Hi Ken. I made a couple of posts on JoNova. Salient point I made:
“I get Urban adjusted anomaly : 0.84 C per 100 years. And Raw anomaly of 0.51 C per 100 years. Or 60% increase”
Our numbers are in close agreement.
One thing reading your work alerted me to was that BOM gridding only includes rural stations not urban stations. My process was in error because I did rural + urban. Interesting thing though : the trend results for rural vs rural+urban at almost exactly the same, for both raw and homogenized data sets. So the rural signal completely overpowers the urban signal after gridding.
Which begs the question what exactly are the BOM doing? They publish data that on close scrutiny makes them look nefarious (60% to 70% signal is not something to be swept under the carpet), yet it has no impact at all on the final trend anomaly. Analytical rigour must be in short supply at the BOM.
September 15, 2010 at 1:55 am
fantastic work. great blog. thank you. i am subscribing to your site.