The Adelaide temperature history has been discussed at Jo Nova’s – Forgotten: Historic hot temperatures recorded with detail and care in Adelaide –
First – Adelaide weather data was collected at West Terrace for around a century then in Feb 1977 a new site commenced 3.2km away at Kent Town. The new BoM adjusted ACORN SAT data is named Adelaide – there is a downloadable ACORN SAT station catalogue if you go to the TAB ‘Data and network’ at the above link.
I wanted to make a start showing how the daily ACORN SAT (Australian Climate Observations Reference Network – Surface Air Temperature) adjustments vary from day to day – seemingly inexplicably. Interesting patterns appear – interesting results emerge – the reasons for which I can only speculate on.
ACORN used West Terrace data till Kent Town started Feb 1977. Then ACORN used Kent Town minimum untouched from Feb 1977. However ACORN adjusted the Kent Town max Feb 1977 until end 2001. West Terrace closed end Feb 1979. The overlap period 1977-1979 for West Terrace and Kent Town showed the raw minimums differed over a greater range than the max diffs.
Here are the daily ACORN minus West Terrace adjustments 1910-Feb 1977 – Three interesting step changes in the minimum adjustments are seen circa 1943-1951 and 1961 – were there site changes?
This Table shows daily details for Jan 1939 – month of record max at time of the 1939 bushfires – now the Adelaide all-time highest max of 46.1 on 12 Jan magically becomes 46.4 according to the god-like BoM.
Here is an example of ACORN producing seven outlier cold nights in April 1960.
Kent Town 23090 max was adjusted to end of 2001 in ACORN – Looking at ACORN max minus Ken Town max we see that mostly the adjustments make ACORN warmer except for a period ~1983-1990 where the adjustments are more balanced either side of zero. Were there site changes ~1983 and 1990?
If the steps are due to Kent Town site changes – then this should have affected the minimums – but ACORN does not touch the Kent Town minimums.
For those interested to pore through spreadsheets –
[1] ACORN has minimum on 16 April 1961 at 5.8 yet no other station in District 23A went that low.
[2] April 1970 example of mystery missing minimums in ACORN on 5th and 27th – what was wrong with the West Terrace data ? Or is it related to those days being slightly warm peaks. An example of many ACORN mysteries.
It is not uncommon for ACORN to reject a days data for no apparent reason. There are many examples of ACORN missing out an extra day if West Terrace misses a day eg Jan and Jul 1913.
[3] In May 1972 from 4th to 10th ACORN makes no adjustments to a string of warm nights where West Terrace is usually the warmest night site in district 23A. This puzzles me – I can understand that the minimums at West Terrace are often too warm and I thought ACORN was tweaking these down so I ask – why this string of zero adjustments to a run of standout warm nights. Is this an indication that the ACORN algorithm was fully up to quota in May 1972? A contra result like this leaves me puzzled.
I am sure readers can come up with many ACORN puzzles.
Yes, this is an on-going problem with ACORN. We discussed Adelaide and Hobart having a higher record temp on ACORN than the raw temp data on Apr 5th on your site. George Bailey will probably be on soon to explain why this could be so. It’s hard to believe that those earlier temps were adjusted because of ‘this-and-that’, yet more recent ones not so.
Warwick
I received a reply from the BoM to my query as to why there was a change to the 1998 and 2009 annual means. The BoM says in part;
‘In transitioning to this update in 2012, the annual temperature values for all
years from 1910 to 2011 changed slightly as a result of using new and different
historical data and from applying more sophisticated analysis techniques.’
By this the BoM means ACORN.
But, question – if the raw and ACORN data have remained the same since the 1980s, how could those years be reranked?
quote
“The Bureau maintains a layered approach to correcting data errors. Automated and semi-automated quality control systems are used to identify observational errors.”
www.bom.gov.au/climate/change/acorn-sat/#tabs=Methods
I can give an example of this
While studying daily forecast temps’ in Horsham a few years ago , l had taken some screen dump snaps of forecast max temps’ for Horsham and the final outcome from the bureaus daily observations
I had noted that the maximum temperature for the day was a few degrees higher than forecast
Not unusual..But what did surprise me was when l snapped the next day obs and forecast data
I went back to the obs’ and found that one deg C had been sheared off the weather stations obs at max temp , making it closer to the predicted forecast..I thought that someone had tampered
However the statement above suggests that the adjustments are automated? This process also enhances the rating /performance of BOM forecasts .
Here is a weather map for 2nd May 1977.
trove.nla.gov.au/ndp/del/article/110843919?
The minimum for Adelaide is given as “06”. So how did the raw get up to 9.2?
From the BoM raw 5.6 is the Kent Town figure. 9.2 is the West Terrace figure. 8.9 is the Waite Institute figure. 2.7 is the airport figure. 8 (in italics) from Belair. 8.7 from Northfield Research Centre. 2.2 at Stirling post office. 5 at Parafield Airport.
“Note that from 1 March 1979 to present time raw data was not adjusted”
Is it a coincidence that this is about the time that satellite records begin?
I tried to figure out what the word “acorn” means apart from nuts. Related words seem to be adulterate, debase, doctor and load (something to do with a coffee scam).
It also comes up related to corrupt officials and fraud.
www.urbandictionary.com/define.php?term=acorn
Oh and as for the patterns you see after using it, there are organisations that exist to help you rehabilitate.
www.stockportdrugsandalcohol.org/treatment-services/acorn
Weathercycles
Last year I took a capture of Sydney Ob’s temps for 18th Jan (Sydney’s hottest day).
The BoM claims the temp reached 45.8C at 1453.
However, the AWS displays 44.9C at 1449 and 44.7C at 1459. So the temp rose 0.9C in 4 mins and then dived 1.1C in six minutes.
The previous record was in Jan 1939 at 45.4C. I doubt if the thermometers in those days would have picked up a sudden spike in temp that the electronic Ts can pick up these days.
hi warwick good work but when i checked the data that you compared against it was the old weather bureau site. when you look at the kent town data it is almost exactly the same. so i think that you should redo the table with the kent town data. ben.
Thanks Ben – you are right – ACORN only used West Terrace data till Kent Town started Feb 1977. Then ACORN used Kent Town minimum untouched from Feb 1977. However ACORN adjusted the Kent Town max Feb 1977 until end 2001. West Terrace closed end Feb 1979. The overlap period 1977-1979 for West Terrace and Kent Town showed the raw minimums
differed over a greater range than the max diffs. Later in weekend I should have new graphics. If my XP system starts behaving.
Ben K
While you are correct as I showed above rather than Warwick being blamed for confusion that is caused by the BoM, would it not be better for them to re-establish a site at West Terrace so that the data from the good quality long record does not get compared to a different place. That day in 1977 had a wide range of temperatures around Adelaide and shows that mashing the records together is not an accurate way see anything much. ACORN just fails to show what the day was like. It is the ACORN method that is wrong not Warwick.
weathercycles – I have never noticed what you mention seeing at Horsham – daily obs being altered to fit forecasts closer – If you see it again please let me know. I have noticed often that obs are often slow by a day or two to be entered at these pages –
www.bom.gov.au/climate/dwo/IDCJDW3050.latest.shtml
which link off daily obs pages.
I assume these data are some distance from ACORN – note how the Climate Data Online monthly T numbers often lag up to a year – I consider CDO data to be raw.
www.bom.gov.au/climate/data/
I have assumed that only after CDO numbers were updated would ACORN be fed.
Not sure exactly how raw the daily obs’are.
According to the notes . They say
IMPORTANT NOTES..ACCOMPANYING DAILY OBSERVATIONS
“Data availability
These observations have been taken from the Bureau of Meteorology’s “real time” system. MOST.. of the data are generated and handled automatically…. Some quality checking has been performed…, but it is still possible for erroneous values to appear.
From time to time, observations will not be available, for a variety of reasons. Sometimes when the daily maximum and minimum temperatures, rainfall or evaporation are missing, the next value given has been accumulated over several days rather than the normal one day. It is very difficult for an automatic system to detect this reliably, so caution is advised. For more detail about this or any other data quality issue, contact us.
———————
Note the ..
“MOST DATA are handled automatically
and .”SOME checking has been performed…”
www.bom.gov.au/climate/dwo/IDCJDW0000.shtml
no explanation and links to exactly how this may be done and in what circumstances?