BoM made a mistake in announcing hottest January day for Mackay
February 8th, 2013 by Warwick HughesYou could not invent this. See my post from yesterday – now as Chris Gillham has pointed out – the BoM mistakenly entered 36.9° as 39.6° – and nobody had the nous to question this for a week. The Weatherzone announcement – Mackay sets January heat record -
was dated 5 Feb as was the ABC’s. So the media announcing this triumph of warming was a week after the instruments at Mackay Aero recorded 36.9 which somehow ended up as 39.6 in BoM data for a week. The 36.9 is still a record for Mackay Aero which has a short record, previous high was 35.8 on 7 Jan 2006 – but Mackay M.O. recorded 37.2 on 1st Jan 1985. To be correct the media stories should have said – “Mackay Airport sets January heat record” – and then the article could have stated that Mackay M.O. recorded 37.2 on 1st Jan 1985 when apparently there was no Airport data that we know of.
I wonder what will happen now – will the BoM or either media group be interested in publishing a correction ?
Posted in BoM Australia, Surface Record | 4 Comments »
February 8th, 2013 at 12:41 pm
To clarify the 39.6C vs 36.9C … I don’t think they misread the numbers and that’s apparent in their second sentence claiming the previous record was 38.5C in 2006.
Mackay Aero 28 Jan 2013 raw = 36.9C
Mackay Aero 07 Jan 2006 raw = 35.8C
36.9 + 2.7 = 39.6
35.8 + 2.7 = 38.5
No way have they misread both temps. The only explanation for 39.6C and 38.5C is a 2.7C upward adjustment of temps at Mackay Aero.
The question is what causes a 2.7C upward adjustment of temps, although it’s also debatable why they reference the secondary station (airport) over the official one with consistent records good enough for ACORN (MO) without saying so. The statement that “the mercury hit 39.6 degrees on January 28″ is best described as deceptive.
I believe a partial explanation for this mystery can be found at www.bom.gov.au/cgi-bin/climate/change/timeseries.cgi?graph=tmax&area=qld&season=01&ave_yr=0
If you download the annual dataset for January in Queensland (Jan 1961-90 average 34.8C), you’ll find that January 2013 had a statewide anomaly of 2.62C (second highest ever anomaly after Jan 1947 at 2.84C).
Now that’s not 2.7C because it’s a monthly anomaly and not a daily anomaly for 28 Jan, but I think it gives a clue as to what’s going on.
This has all the hallmarks of area averaging (see www.waclimate.net/hottest-day.html) wherein actual averages have several degrees added to compensate for areas with few thermometers, except area averaging is applicable to numerous stations within state land areas and should have nothing to do with an individual station such as Mackay Aero (or MO).
There are similar questions about how Australia’s hottest ever day on 7 January had an averaged absolute max of 35.1C at 721 stations but an area averaged 40.3C at 718 stations.
In the words of the BoM … “due to the different length scales of anomalies versus absolute temperatures, the ACORN-SAT data is best analysed as anomalies (departures from a climatological mean). The national average daily anomaly from ACORN-SAT for the 7th January 2013 was 5.36°C. This can be converted to an absolute temperature by adding the national climatological average for the period 1961–1990 of 34.64°C, giving a national temperature of 40.00°C. However we believe that the AWAP derived average of 40.33°C is likely to be more accurate, since the coverage of the network is maximised.”
There is nothing in the public record from which these daily area average adjustments can be sourced, although the BoM has stated that an analysis of the new AWAP network calibration will be published soon with the historic analysis thereafter released and hopefully including a daily dataset back to 1911.
I’m not sure how area averages relate to a max at just Mackay Aero on 28 Jan but it should be noted that these new mystery temps and records started in January and in the BoM’s December 2012 Monthly Statement (www.bom.gov.au/climate/current/month/aus/archive/201212.summary.shtml), they say:
“The new ACORN-SAT temperature dataset has been used for calculation of State and national temperature area averages in summaries from December 2012 onwards. The major change from earlier datasets is that the ACORN-SAT dataset commences in 1910, rather than 1950, and hence rankings are calculated using a larger set of years.”
I’ve got no idea how it all ties together but for the past month we’ve been stuck with an unknown daily anomaly adjustment procedure that seems pretty good at producing or exaggerating record temps, and a media that blithely quotes every media statement without question.
February 8th, 2013 at 2:25 pm
Chris
I have looked at all the data re the 38.5C in 2006, both BOM and ACORN, and none state that – they say 35.8C or
35.7C for that day.
All the data for Jan 2013 says 36.9C. I have sited the relevant data in the blog re Mackay earlier.
The highest temp for Mackay in Jan was in 1913 – 37.7C at the PO.
It will be interesting if some temps will be adjusted.
February 8th, 2013 at 4:56 pm
I am still opting for mistakes as the explanation – I have kept screen-shots of the two media articles. From my experience roaming thru ACORN daily data – it is normal that there are no or v few adjustments in current data and recent decades. Then at a chosen date maybe in the 80′s or 70′s the adjustments start and almost every day might be adjusted by varying amounts back to the start in early 20th Century. You have to see it graphed to believe it. So I would be staggered if they adjust Mackay ACORN for the 28 Jan 2013. We will see.
February 10th, 2013 at 8:26 pm
A little bit more information re Mackay’s ‘highest temperature’.
The Weatherzone site has the number ’24′ shown beside the 36.9C for Mackay Aero (Mackay’s so called highest Jan temp).
Explanation reads:-
‘A yellow cell indicates a probable monthly record for this site (sites with ≥ 10 years of records only). The number of years of records available for the relevant field for this month is indicated in black. See the station’s climate page for full details.’ See this at:-
www.weatherzone.com.au/station.jsp?lt=site&lc=33045&list=ds&of=of_a&ot=ot_a&mm=01&yyyy=2013&sub=go
Mackay Aero starts its daily data set in 1950 but has missing or no data from 1960 to the mid 1990s. Mackay’s MO has a Jan top temp of 37.2 on New Years Day, 1985 (ACORN agrees).
So can it be a record for Mackay Aero only based on its incomplete record but not for Mackay itself?