When the Bureau called a new record hot day for 23rd September 2017, based on the instantaneous spot value from a platinum resistance probe in an automatic weather station at Mildura, I wanted to know the maximum temperature recorded by the mercury thermometer on that day as a check.
I wanted to compare the new record temperature with the temperature as recorded by the mercury at the same weather station, for verification.
I understood that Mildura was one of the few official recording stations with parallel data.
I had asked The Australian Bureau of Meteorology for this type of data before and been ignored. On this occasion I contacted 2GB radio host Alan Jones. He intervened on my behalf contacting the relevant government minister.
The Bureau claimed the temperature on 23rd September 2017 to be the hottest day ever for Victoria since records began back in 1889.
But as I explained to Alan Jones, there weren’t probes measuring temperatures back then; rather mercury thermometers. So, I wanted to know how much the mercury had measured on 23rd September 2017, the claimed record hottest day.
In short, I really wanted to see the A8 Form for 23rd September 2017, to know what had been written down for the maximum temperature from the mercury and compare this against the spot reading for the probe. To check, to verify. This should be standard practice when calling a new record hot day, but it’s not.
Instead of being provided with that one day of data, I acquired 10,761 scanned A8 Forms for Mildura. But not the form for 23rd September 2017.
Apparently, the mercury had been removed sometime in 2015. So, so disappointing, I thought.
But nevertheless, I thought, with all this data I can know how the newest probe records relative to the mercury including during September 2012, 2013 and 2014.
Except, there are no recordings from the mercury for September 2013 and 2014, and September 2012 has so far been withheld.
“Thirty days hath September, April, June and November …” is a traditional mnemonic verse for remembering the number of days in each month.
I really just wanted the temperatures as written on the A8 Form for 23rd September 2017. But denied that, then for other Septembers for Mildura from the same probe, please, so I can compare values for the same location for the same month. I could see that there was seasonal variation in the difference between the first two probes used at Mildura and the temperature as recorded from the mercury thermometer.
I wrote to Anthony at the Australian Bureau of Meteorology about this in January and then February 2018.
From: Jennifer Marohasy
Date: 24 January 2018 at 3:14:19 pm AEDT
To: Anthony R
Subject: Missing A8Forms, September 2012
My best wishes to you for 2018.
I took some time-off over Christmas and New Year, and I am now back transcribing the scanned A8 forms that you so kindly provided to me early December.
I was intending to transcribe September 2012 today, but I have found that this month is missing. It is not on the USB that you emailed me, nor was it one of the months that you originally uploaded at the FTP drive.
Perhaps this month in that year was never actually scanned?
Could the September 2012 A8 forms please be scanned, and provided to me… so the record is complete?
Kind regards Jennifer
From: Anthony R
Date: Tue, Jan 30, 2018 at 9:38 PM
Subject: RE: Missing A8Forms, September 2012 [SEC=UNCLASSIFIED] To: Jennifer Marohasy
Thanks for bringing this to my attention. I have just returned from leave but will follow up on the missing month of data now.
Kind regards Anthony
From: Jennifer Marohasy
Date: Mon, Feb 12, 2018 at 11:09 AM
To: Anthony R
I have tried to phone you a couple of times chasing this data: the scanned A8 forms for September 2012. I left a message on your phone again this morning.
When can you make the scanned A8 forms for September 2012 available to me? I had not initially realized how critical that month of data is.
When you first wrote to me you indicate that there were parallel readings through until January 2015. So, I had assumed that there would be at least three Septembers (2012, 2013, 2014) with parallel data from the current Rosemount probe installed in June 2012.
When I wrote to you on 24 January 2018 explaining that the month of September 2012 was missing, I had not yet realized that the months of 2013 and 2014 (while scanned) did not actually have any recordings from a mercury thermometer i.e. no parallel readings. I had been working slowly through the forms chronologically.
You had initially indicated, and I quote from your email of 28 October 2017:
“A8 records also exist from 1 January 2001 to 28 January 2015 and these are in the process of being extracted from archive and scanned. The scanning is a labour-intensive activity and will take some time to complete… As the station was automated in 2015 there are no liquid-in-glass measurements for 23 September 2017.”
I don’t much believe in conspiracy theories, but it does seem somewhat suspicious that you have provided me with over 10,000 scanned pages representing a complete record for all the A8 forms from Mildura for the period 1 January 1989 to 31st January 2015, with the only exception being September 2012 – with every day for that month missing.
Of course, given the available manual recordings from the mercury thermometer this would be the only month of data that would provide a direct comparison with the recording on 23 September 2017; apparently the hottest day every recorded in Victoria since 1889.
Kind regards Jennifer
The feature image is a scanned A8 Form from 13th September 1996. This is the last September that the mercury was the official recording instrument at Mildura, from 1 November 1996 there was a change to the probe. An unusually cold 14.3C was recorded from the mercury as the maximum at 9am that morning, this being the maximum for the day before. Those interested in chasing down such things will see that this is also the maximum temperature recorded in the ADAM data archive for meteorology for 12th September 1996. Interesting the ACORN-SAT data base, used by the Bureau to understand climate variability and change, shows 15.1C for that day. To be clear, through remodelling alone the Bureau has changed this maximum daily value by 0.8C. Such dramatic changes via remodelling, also known as homogenisation, are not unusual.