The BoM was asked to calculate the climate of the Canberra district in the process of deciding on a location for the Australian Federal Capital Territory – Chris Gillham alerted me to this report and he has his own comments on the data in a 1Mb pdf report.
Looking at the temperature data in the Table on page 7 & 8 of the pdf report I decided to check on Goulburn, Cooma and Yass getting data from BoM CDO.
In the case of Goulburn the annual mean temperature was 56.1°F for the 46 years prior to 1910 – this equates to 13.4°C and the station was at an altitude of 702m. Now Goulburn Airport is at an altitude of 640m so the 13.4 in town would equate to 14 at the Airport assuming a standard lapse rate. The Airport in the ten years 2004-2013 averaged 13.1 – so even if we take of 0.2 from the Goulburn Town temperature for possible non-Stevenson screen exposure – we still have Goulburn 105 years ago 0.7 degrees warmer than the last ten years.
In the case of Cooma the annual mean temperature was 54°F for the 44 years prior to 1910 – this equates to 12.2°C and the station at Lambie Street was at an altitude of 812m. Now Cooma Airport is at an altitude of 930m so the 12.2 in town would equate to 11.04 at the Airport. The Airport in the ten years 2004-2013 averaged 11.15 – If we take of 0.2 from the historic Cooma temperature for possible non-Stevenson screen exposure – Cooma 105 years ago might have averaged 10.85 or 0.3 degrees cooler than the last ten years.
Yass is complicated by stations closing in the last ten years with only one month overlap. I have done the calculations and allowing for the site change and altitude differences there is near zero difference between the Yass district pre 1910 and in the last ten years.
Upshot is there seems to be no global warming or significant climate change around Canberra since the 19th Century.
And that is an official BoM publication using data which at that time the BoM considered acceptable and accurate. If, as is likely, the Canberra data now may have some UHI effect added, so this part of the world has been no hotter than that pre-1909 period. I had better keep my opinion to myself!
Yesterday I updated the Hunt PDF to include a Year Book extract with the unadjusted raw temps to 1944 at the single location of Canberra. This provides another glimpse at national capital temps prior to Canberra Airport Comparison 70014 which starts 1939 in BoM Climate Data Online.
Year Book <1944: Min – 7.1C, Max 19.7C, Mean 13.4C
It will be argued that Canberra nowadays should be compared with the closest station to Canberra Airport in the <1909 Hunt analysis, which I believe is Queanbeyan.
Hunt Queanbeyan <1909: Min 5.9C, Max 19.4C, Mean 12.7C
That compares with Canberra AP Comparison 70014 1939-2010: Min 6.5C, Max 19.7C, Mean 13.1C. (or 2001-2010 Min 7.3C, Max 21.0, Mean 14.1C)
Alternatively, Canberrra AP 70351 2009-2014: Min 6.6C, Max 20.8C, Mean 13.7C
If it was to be assumed that Canberra AP 70014 and 70351 are comparable stations to determine temperature trends, we might rejoice that our national capital was 0.4C cooler over the past four years than in the preceding decade.
However, the Queanbeyan comparison tells a different story and it could be argued that Canberra AP 70351 2009-2014 is 1.0C warmer than <1909. However, if 70351 is compared with the Year Book Canberra <1944, it's just 0.3C warmer.
There is a possible Glaisher difference (av warming bias 0.2C over a year but probably less in the relatively cool ACT climate) pre 1909 but not in the Year Book temps pre 1944, which are all Stevenson. Of course, this ignores the UHI impact of building a national capital and an international airport around what was nothing but paddock and bush.
The Year Book doesn't nominate a start year for its averages but it's worth noting the hottest Canberra day recorded was 41.9C on 11 Jan 1939, and this compares with a hottest 41.4C on 31 January 1968 at 70014 and 42.0C on 18 January 2013 at 70351. Coldest day in Year Book was -7.7C on 20 June 1935 vs -8.0C on 29 July 2011 at 70351 and -10.0C on 29 July 1975 at 70014. All Stevensons.
It's the regional averages that give the best comparison of the ACT now vs <1909 as they're not so heavily biased by UHI and the international airport (Stevenson 100m from runway – pics at www.flickr.com/photos/87591849@N00/6301188055/in/photostream/). The historic records suggest that if Canberra itself has warmed over the past century, most of it happened before the 1940s when it's difficult to blame CO2.
My updated PDF is linked from www.waclimate.net/year-book-csir.html, which compares a CSIR collation of all temps including 1800s at 225 weather stations across Australia before 1931, 44 stations collated by the Year Book between 1911 and 1940, and their current stations in RAW and ACORN. RAW averages 0.2C warming at 225 locations while ACORN squeezes 0.5C warming out of the 79 stations that can be compared.
The original CSIR document, which you can download from the page, gives another historic glimpse at ACT temps separate from Hunt and the Year Books, this time with all recordings up to 1931 …
Canberra: Max 19.7C, Min 6.7C, Mean 13.2C
That produces 0.5C mean warming compared to Canberra AP 70351 2009-2014 at 13.7C.
The ACT/Yass district (as defined in the Hunt stations) from CSIR estimates up to 1931 …
Bombala: Max 18.3C, Min 4.8C, Mean 11.5C
Canberra: Max 19.7C, Min 6.7C, Mean 13.2C
Cooma: Max 19.8C, Min 4.7C, Mean 12.3
Goulburn: Max 20.0C, Min 7.0C, Mean 13.5C
Queanbeyan: Max 20.7C, Min 6.4C, Mean 13.5C
Cootamundra: Max 22.2C, Min 8.4, Mean 15.3C
Murrumburrah: Max 23.3C, Min 7.9, Mean 15.6
Those seven stations can be averaged for mean temps from the three alternative datasets (substituting 70351 for Queanbeyan with no alternative close station):
Hunt <1909 = 13.7C
CSIR <1931 = 13.6C
BoM CDO 1912-2014 = 13.5C
BoM CDO 2004-2013 = 13.7C
So unadjusted RAW from two separate historic documents suggest not much warming around our national capital, despite a lot of hot air in Canberra about climate change. The results are similar to points made on various climate research blogs that inner NSW hasn't warmed much over the past century.
A good find Chris. I bet there are a few more hidden in back cupboards that will shake the cobwebs at BoM.
Have you checked the raw BoM data for steps down in temperature? This is quite easy to do with a spreadsheet, I’m finding a lot of stations have 1-2C (seasonal average) steps down at various dates between 1890 and 1909, but not yet looked at this region. I’m guessing that these steps are the introduction of Stevenson screens. Sorry, but just quoting pre-1910 raw data temps has no credibility until the switch to Stevenson screens is accounted for, IMHO.
You need a good reference station, Adelaide or Melbourne would probably work OK. The reference station does not need to be nearby, it just needs to have a slowly varying deviation from the regional climate.
I work out and plot 4 seasonal averages per year, then manually offset the reference data (independently for each season) so that it has close to the level of the station of interest. The key plot is then the time series of seasonal differences (4 samples per year), any step in temperature greater than 0.5C should be clearly visible. Having 4 samples per year is crucial, annual data is not sampled at a high enough rate to make steps visually detectable, except in a few cases like Cape Otway.
I have looked at BoM data in spreadsheets for 25 years now – and am aware of what you are talking about. These issues are addressed in my article on 7th of this month – What reasonable conclusions could be drawn from results of the 1887-1947 Adelaide experiment comparing thermometer readings in a Glaisher stand and Stevenson screen ?
www.warwickhughes.com/blog/?p=3398
You will see the BoM published a paper on this in 1996 and said that pre Stevenson screen readings could be 0.2°C warm in mean annual data. So I applied that adjustment in case there was an exposure issue.
BTW – I think that to use Adelaide or Melbourne to check Canberra region data could be a stretch too far. The BoM criticized me for comparing Charleville and Cunnamulla.
The 1-2C steps I’m finding are in average maximum temperatures, and often there are quite big seasonal differences in the size of the steps, highest step in summer. Maybe the stations with steps did not have Glaisher stands. Many stations have no detectable (at least by me) steps at all, so maybe those ones went from Glaisher to Stevenson with little change.
I was surprised at the similarity of Adelaide to far flung stations, a similarity that extends all the way to Wagga Wagga (2C step detected there in 1908) and maybe beyond.
I only use Adelaide for detection of steps, then try to use nearby stations to get the size, a simplified version of the 2-stage process used by the BoM in ACORN-SAT.
If you want to email me a few graphics – I can decide if there is a basis for a guest post by you on the subject you have raised (assuming you would want to). I agree these are interesting issues. Odd that nobody has published on the nuts & bolts of temperature inhomogeneities around screen changes. No hurry – good to hear of people probing these issues.
OK, will send you a few plots to illustrate the steps I’m finding, seeing is believing when it comes to steps, everyone should be sceptical when it comes to claims of large changes. It would be good to get a complete list of major steps prior to 1910, dates and sizes. Ashcroft et al 2012 (Temperature variations of southeastern Australia, 1860 – 2011) probably have this info, but I like to do independent checks.
Since this post is about <1909 temps and if anyone's interested, Hunt also gives an insight to some northern Australia weather stations in a 1909 letter to his minister (courtesy of detective work by Sliggy, I believe).
In short, the climate table in his letter suggests the mean summer temp of Darwin, Daly Waters, Alice Springs, Wyndham, Townsville and Brisbane <1909 was 28.7C, which compares to the subsequent long-term summer mean at those locations of 28.5C and a 2004-2013 summer mean of 28.9C.
The <1909 average winter mean at the six locations was 20.6C, compared to long-term 20.1C and a 2004-2013 winter mean of 20.0C.
The averaged summer/winter mean <1909 was 24.7C, compared to long-term 24.3C and 2004-13 summer/winter mean of 24.5C.
The Hunt letter table also looks at the hottest/coldest days and months before 1909 which can be compared with current station records.
There is a Glaisher influence among most but not all stations, particularly affecting summer max, but Glaishers shouldn't artificially warm winter temps and <1909 winters were 0.6C warmer than 2004-2013 at the six northern locations. Similar to the southern ACT/Yass district … no warming since <1909 … with the same intractable argument about whether the thermometer screens artificially warmed the past or if UHI is artificially warming the present.
The Hunt letter with a temp spreadsheet at the end can be viewed at www.waclimate.net/northern-australia-pre1909.pdf