CRAIG KELLY: BUREAU DELETED HEAT RECORDS. BUREAU RESPONDS
Liberal MP Craig Kelly has gone through the National Archives and found four records – all before 1940 – of heat waves in NSW that were hotter than the Bureau of Meteorology’s list of the hottest ever days in the state. Why were these records deleted?
Australia’s highest recorded temperature (using standardised equipment) at an official government weather station was 51.7°C (125°F) recorded at Bourke on the 3rd January 1909, followed by 51.1°C (124°F) at White Cliffs on the 11th Jan 1939.
However these records have been purged and hidden by Australia’s Bureau of Meteorology.was 49.7°C.
I discussed this with Kelly:
The Bureau’s response:
The Bureau of Meteorology rejects any suggestion it has deleted temperature records or that a handful of individual extreme temperatures would alter the conclusions about Australia’s warming climate, which is based upon millions of observations from hundreds of sites.
While temperature records for a number of locations stretch back into the mid-nineteenth century, the Bureau’s national daily records for temperature begin in 1910. The standardisation of instruments in many parts of the country did not occur until 1910, and there is a lack of documentation available that outlines how temperature observations were recorded at many sites before this period.
The Bureau does not discount that certain sites may have recorded temperatures in excess of 50 degrees before 1910, but there is a lack of appropriate documentation and information at many sites, that would verify that the methods and equipment being used were consistent with post-1910 practices. This is essential if we are to appropriately compare things like daily temperature records.
The Bureau has previously carried out thorough analysis of the 3 January 1909 Bourke observation and is confident the recorded temperature was highly likely to be a reporting error. A comparison to observations of all inland sites with standardised equipment throughout NSW and Queensland on the same day, showed temperatures were more than six degrees cooler at all other sites. A temperature difference of this scale is highly unusual and likely implausible, as sites in inland Australia have a high degree of consistency on warm days.
The Bureau understands that an undigitised paper record from a site in Brewarrina was recently found in an archive. It’s important we understand the instrumentation and methods that were used at this site before making comparisons. Based on the data and evidence currently available to the Bureau, we remain confident the Bourke measurement from 3 January 1909 was either a reporting error or an observation from a non-standard piece of equipment.
The 1939 observations from White Cliffs were recorded on non-standard pieces of equipment that were outdated even for the time. This site was only upgraded with standardised Stevenson Screens in the mid-1940s. To appropriately compare daily weather observations, it’s critical that the instrumentation being used is the same. This way we are comparing apples with apples.
The Bureau is aware of the 1939
observation of 50.0 °C, and this is currently the highest daily record that has been discovered for NSW using a standardised Stevenson Screen. The record does not appear on official Bureau publications since the daily records from that period for Wilcannia are undigitised. Nonetheless, it has been referenced in a recent Special Climate Statement on extreme heat in NSW. Work to digitise historical daily temperature paper records from Wilcannia, along with other historic paper records, is ongoing. Once this has occurred, the Wilcannia observation will be made available on the public web, including on the table of historic extremes for NSW, which is based on observations that have been digitised. The Bureau notes the current NSW maximum temperature record based on data that has been digitised is also from 1939. It’s entirely possible that further extreme historic temperatures both maximum and minimum will be uncovered through the process of digitising paper records. The discovery of such historic extremes will have no influence on the well-established conclusions that Australia’s climate has warmed by more than one degree since 1910.
The Bureau has digitised a sufficient amount of monthly data from paper records, to provide a good understanding of how climate has varied and changed in Australia since 1910.
Craig Kelly responds:
Firstly the BoM’s response is an acknowledgement that their glossy brochure of “Extreme Records” is false and misleading. And it’s a vindication of the issues I have been raising.
It’s scandalous that the BoM continues to publish information showing that NSW never had a day 50C or over, when they confirm they are fully aware that is untrue and that Wilcannia hit 50C in 1939.
“The Bureau is aware of the 1939 Wilcannia observation of 50.0 °C, and this is currently the highest daily record that has been discovered for NSW using a standardized Stevenson Screen”.
But go to the BoM’s current published list of hottest days for NSW and this record is not there. The excuse given by the BoM for not including this in their published list that’
“The record does not appear on official Bureau publications since the daily records from that period for Wilcannia are undigitised”.
This is farcical. ‘’Undigitised’’ means the record exists, it’s detailed in the original Observation records, but simply it has not yet been put into a computer format.
And it’s of concern that 9 years ago, a report by the Independent Peer Review Panel into the BOM’s practices in 2011, in recommendation A4 stated;
“The Review Panel noted that most but not all of the possible historical monthly temperature records have been digitised, and that some daily ACORN-SAT data remains undigitised. The Panel welcomes the statements they received from Bureau staff that undigitised ACORN-SAT records will be targeted for future digitisation. The residual set of non-digitised data and metadata paper records for ACORN-SAT observation stations should be located as a matter of priority, and imaged. It is important that the digitisation of these remaining paper only daily temperature records then be completed.”
http://www.bom.gov.au/climate/data/acorn-sat/documents/ACORN-SAT_IPR_Panel_Report_WEB.pdf
And yet here we are almost 9 years later, and the BoM have still failed to get around digitising what they acknowledge as NSW’s hottest ever day, and then they use their failure to digitising the record as their excuse for not including it list of extreme temperatures.
The BoM should immediately, as in to today, take down the page of ‘Extreme Records’ from their website because it is false and misleading, and they should also immediately re-instate it with 50°C Wilcannia record.
THE BOURKE RECORD
The 125F (51.7C) recorded at Bourke in 1909 is recorded in the official Meteorology Observations for Bourke, and the Bureau admit that this was measured with standardised equipment.
The reason given for its exclusion is that the BoM’s ‘previous analysis’ concluded that the Observer that meticulously recorded all meteorology observations onto the official government form, at the Bourke weather station back in 1909 ‘highly likely’ made a ‘reporting error’. In other words, he was either too stupid or too incompetent to accurately record the temperature properly. And because we are so much smarter now, we know it couldn’t have been that hot back them (never mind all the newspapers reports of the day with the body counts of people dying for the heat).
In making reaching their conclusion of this “observer error” the BOM refer to;
“a comparison to observations of all inland sites with standardised equipment throughout NSW and Queensland on the same day, showed temperatures were more than six degrees cooler at all other sites. A temperature difference of this scale is highly unusual and likely implausible, as sites in inland Australia have a high degree of consistency on warm days.”
However, the this comparison overlooked the very closest weather station at Brewarrina, which recorded 50.6 C on the same day that Bourke recorded 51.7C. Such difference between Bourke and Brewarrina are par for the course. Further, the comparisons with the overnight minimum at Brewarrina on the 3rd Jan 1909 and the temperatures on the preceding and follow days all move in a similar manner to those at Bourke.
It remains a mystery to why the BoM would overlook a comparison with the very nearest official government weather station, especially when this assists in confirming the accuracy of the Bourke record.
The BoM’s responses states;
The Bureau understands that an undigitised paper record from a site in Brewarrina was recently found in an archive. It’s important we understand the instrumentation and methods that were used at this site before making comparisons.
Firstly this response demonstrates that the BoM doesn’t know what the ‘instrumentation and methods that were’ at Brewarrina in January 1909 – they are searching for an excuse for exclude the nearest official weather station.
Secondly, in considering the possibility of a ‘reporting error’ at Bourke it is irrelevant as to what the exact instrument was being used in 1909 at Brewarrina. This was an official government weather station that had been in operation since the 1871. What is relevant is if the the closest official government weather also recorded unprecedented heat very same day, 3rd January – not just the maximum but also the minimum, and if the movements of recorded temperatures over the preceding and following days on 3rd Jan 1909.
The BoM also argue;
A comparison to observations of all inland sites with standardised equipment throughout NSW and Queensland on the same day, showed temperatures were more than six degrees cooler at all other sites. A temperature difference of this scale is highly unusual and likely implausible, as sites in inland Australia have a high degree of consistency on warm days.
This is also false.
Firstly, the comparison that BoM refers to compares temperatures recorded at Bourke, with those of Walgett, Coonamble and Thargomindah (conveniently overlooking the nearby Brewarrina). However this comparison was not taken in the years around 1909, it was undertaken during the period 1959-95 – that’s half a century later.
Not commencing this comparison until 50 years after the record in question is significant, as during that period, the average maximums and the extremes at Bourke declined significantly. The reason that Bourke’s maximum temperatures (recorded at the rear of the Bourke Post Office) declined throughout the last century are possibly because the Stevenson Screen in the backyard of the Bourke Post office became shadowed at times from growing trees and/or new constructed buildings to the north and west of the Post Office.
However whatever reason that Bourke’s maximum temperatures cooled , they render the BoM’s comparison of Bourke’s maximum temperatures with other nearby stations during the period 1959-95 irrelevant when considering the 1909 record. It also leaves the question, why would the BoM jump half a century into the future and ignore decades of comparisons from around 1909 ? Where they ignorant of the fact of Bourke’s cooling ?
Further, if instead of ignoring these 50 years of comparisons, the BoM had of made comparisons in the years closest to 1909 they would have found that a 6 C temperature difference of between Bourke and other nearby stations as not highly unusual nor likely implausible on days of extreme heat.
For example;
Bourke v Walgett : On the 28th Jan 1913, Bourke recorded a maximum temperature of 47.5°C. In comparison Walgett recorded a temperature of 40.0°C – that’s 7.5°C cooler – substantially cooler than the 6 degrees cooler that the BoM claim is “likely implausible”.
Bourke v Coonamble : The daily Jan 1913 maximum temperatures for Coonamble are not digitised and it would require a visit to the national archives to determine the maximum temperature on the 28th Jan 1919. However, the BoM lists the highest daily maximum for the month of January 1909 at Coonamble as 40.6 C on the 22nd Jan. Therefore the highest possible reading at Coonamble on the 28th Jan 1909 would have been 40.5 C – which is 7 degrees cooler than what Bourke recorded in the same day.
The BoM’s excuses for ignoring the original observations at Bourke on 3rd Jan 1909 are a nonsense.
Even though the official government weather station at Bourke recorded 127 F on 17th Jan 1877, this was not with the standardised Stevenson Screen and therefore Bourke’s 125F (51.7 C) as recorded on the 3rd January 1909, in the official government meteorological observations of the day, should be officially recognised by the BoM as Australia’s hottest ever recorded temperature.