Bureau Still Limiting Cooling to Minus 10 Degrees

TWO weeks ago, in response to my queries the Australian Bureau of Meteorology acknowledged that it had put in place limits on the lowest temperature that an individual weather station could record.  So, when the automatic weather station which now operates in the town of Goulburn (not far from Canberra), recorded a measurement of minus 10.4 degree Celsius, the ‘quality control’ system in place ‘corrected’ this to minus 10.0 degree Celsius. None of this is denied.  When we protested, the Bureau emailed:

“The correct minimum temperature for Goulburn on 2 July, 2017 is -10.4 recorded at 6.30am at Goulburn Airport AWS… The Bureau’s quality control system, designed to filter out spurious low or high values was set at -10 minimum for Goulburn which is why the record automatically adjusted.

The error was picked up yesterday internally and quality control processes are being reviewed for those stations where temperatures below -10 are possible.

Further attempts at clarification have been met with stonewalling. Though after three days the Bureau did insert the correct measurement of -10.4 degree Celsius into the CDO dataset, where -10.0 had previously been showing.

I reported some of this in a note for The Spectator.

I had assumed that after this unfortunate incident the Bureau would remove the new absurd limits on how cold a temperature could actually be measured.   Indeed, the implications for the integrity of invaluable historical meteorological records are too horrible to contemplate.

I suffer, however, from a propensity to always check things.  So, when I thought temperatures might yet again dip below -10.0 degrees Celsius somewhere, I set my alarm for 5am, to get up early that morning to check (yes check) the values actually being recorded by the new automatic weather stations — particularly those scattered across the Australian Alps, given the weather-setup.  Then I would be in a position to subsequently check (yes check), to be sure any values measured below -10.0 were also taken forward into the official raw temperature dataset.

So, last Sunday morning (16 July) I got up at 5am and started checking the latest weather observations for the Snowy Mountains, including the locations of Perisher, Cooma and Thredbo.  At 5.30am the automatic weather station at the Thredbo Top Station (Number 071032) showed -10.4 degrees Celsius. I took a screenshot, and saved it on my computer.

The screenshot taken soon after 5.30am on the morning of Sunday 16th July 2017. This clearly shows the Automatic Weather Station at Thredbo recorded a temperature of -10.4 degree Celsius.

This ‘latest weather observation’ pages automatically refreshes every 10 minutes. So, if this value was going to be erroneously reset to -10.0 — as the Bureau wrongly reset the -10.4 recorded at Goulburn two weeks earlier — I would want the screenshot as evidence.

I thought this would probably be the end of it.  Temperatures have actually got as low as -14.7 at Thredbo Top Station – with this value recorded on 28th July 1980.  Surely the Bureau would not reset the cold, but not exceptionally cold, value of -10.4 at Thredbo that I saw measured at 5.30 by that AWS on Sunday 16th July 2017.

Nevertheless, the next morning (Monday morning) I checked the ‘Daily Weather Observations’ sheet for the Thredbo Top Station at the Bureau website to be sure that the value of -10.4 had indeed been brought forward as the lowest value recorded so far this month at that weather station.

But alas! There was no minimum value recorded for Sunday 16th July.  And the sheet stated that -9.6 was the lowest/the coldest value so far recorded at Thredbo this July 2017.

The ‘Daily Weather Observations’ page does not show the lowest minimum recorded on Sunday of -10.4. Rather it suggests no minimum temperature value was recorded on that day.

This morning (Tuesday), I held out some hope that the correct value of -10.4 would nevertheless be recorded in the CDO database as the minimum for Thredbo on Sunday 16th 2017. But alas, they are again leaving a blank.

This is a screenshot of the ‘Daily Minimum Temperatures’ for the Thredbo Top Station (number 71032). These are the raw values that comprise the historical record for this location.  The value of -10.4 has not been recorded for Sunday 16th July.

This is unforgiveable, surely someone is going to be sacked?  This either reflects an extraordinary incompetence, or a determination to prevent evidence of low temperatures.

While the Bureau just two weeks ago was adjusting values below -10.0 to -10.0 creating something of a bodge at Goulburn, they now appear to be leaving a blank — at least at Thredbo.

We know from the Climategate emails that the Manager of Climate Monitoring and Prediction at the Bureau, Dr David Jones, has little regard for numbers.  He wrote in an email on 7th September 2007 (ten years ago) that: “Truth be know, climate change here is now running so rampant that we don’t need meteorological data to see it.”



Additional information

A simple frequency plot of all daily minimum temperatures record at Thredbo Top Station and entered into the CDO dataset indicates that it is likely temperatures will sometime drop below 10.0 degree Celsius.


While there is no minimum value recorded in the CDO dataset for Thredbo, a maximum value was recorded on Sunday 16th July 2017.


Thredbo is skewed normal as many data samples for the warmer months (January to April) are missing for the period from 1981-1997. Then for the period 1997-2003 nearly ALL daily values for Thredbo are rounded to the nearest degree! Thanks to Jaco for this information, and the charts.  [added 20th July]

The Higgins’s Storm Chasing crew also noticed the very cold -10.4 at Thredbo. Like me they must have been up early checking – because this temperature record has since been erased. More information about how cold it was that morning can be found here: https://www.sunshinecoastdaily.com.au/news/coldest-temperature-years-australia-shivers/3200942/ .  Thanks to Bob F-J for this link, and image. [added 20th July]

26 Responses to Bureau Still Limiting Cooling to Minus 10 Degrees

  1. Glen Michel July 18, 2017 at 9:44 pm #

    Dr. David Jones has form – poor form. His involvement with the CRU at UEA leaves no doubt that he is THE problem at the Bureau. Blair too! Too much political shenanigans;we’ve always known that.

  2. Siliggy July 18, 2017 at 10:15 pm #

    Hi Jennifer. As you know I was up early that morning also taking screen shots. I saw the Thredbo top station temperature hit -10.4 so many times I suspect it hit a limit and the real temperature got much colder as it did at Perisher (-12.1). It hit -10.4 that I know of at 4:09 AM, 4:29, 5:30, 6:02, 6:22 and 7:04. From you screen shot above we see something missing at 2:00 AM. When they delete it the grid position would be filled with system generated infill. It would be good to know what they have put into the daily grid at that location but my software cannot do it. Perhaps one of your readers can download the daily grid and check? Click on “Download the daily grid” here http://www.bom.gov.au/jsp/awap/temp/archive.jsp?colour=colour&map=minave&year=2017&month=7&day=16&period=daily&area=nat
    Siliggy = Lance Pidgeon

  3. Frederick Colbourne July 19, 2017 at 12:45 am #

    You have used the normal distribution, whereas the distribution is obviously skewed to the left. Possibly the distribution is lognormal.

    I am surprised by this as lognormal suggests the generation process is multiplicative instead of additive.

    Leads me to wonder if the mean and standard deviation have shifted over time, say decades. These lognormal parameters might define temperature change in a more robust manner than the normal parameters.

  4. Mark M July 19, 2017 at 5:44 am #

    Feb, 2009: David Jones, the head of the bureau’s National Climate Centre, said there was some risk of a worsening El Nino event this year, but it was more likely to arrive in 2010 or 2011.

    “In the past, La Nina usually meant wet in Victoria and El Nino meant dry,” Dr Jones said. “For the last 13 years, the El Ninos have been associated with what could be best called devastating droughts and the La Ninas have been associated with close to average rainfalls, at best.”


    2013: Australian floods of 2010 and 2011 caused global sea level to drop

    “But surface waters in the eastern Pacific were affected by a cyclic cooling phenomenon called La Niña:”

    2006: This year, Dr David Jones received the National Australia Day Council Achievement Medallion from the Bureau of Meteorology for leadership of the bureau’s Climate Analysis section.


    Incompetence has it’s rewards at the BoM.

  5. Robert Swan July 19, 2017 at 9:04 am #

    It is outrageous that they poison the record in this way, but you are wrong to expect a normal distribution or any other formal distribution.

    Local topography will have a big effect on wind and sunlight. Probably the biggest factor is hours of daylight and, over the year, it clusters around the solstices. IOW, hours of daylight is bimodal. In case I messed up the html for that link: http://home.comcen.com.au/rswanws/sh/daylight.png is a histogram of daylight hours at 35 degrees.latitude on a spherical Earth.

    I posted something similar at JoNova’s the other day but should have said something here too.

    Not a major point, but seems worth making all the same.

  6. Mike Borgelt July 19, 2017 at 9:53 am #

    “Truth be know, climate change here is now running so rampant that we don’t need meteorological data to see it.”
    Which planet do these people live on? Dr Jones would seem to be in need of medical help.

  7. jennifer July 19, 2017 at 10:17 am #

    Hi Mike Borgelt,

    Thanks for caring – and for your concern about David Jones’ health.

    But the reality is that successive Coalition governments (under first Abbott and now Turnbull) have done nothing about this. The Climategate emails have been publicly available online since November 2009.

    I rather think it is a case of the gutless leaving the cunning (not mentally ill) in-charge. When I’ve challenged the odd politician on this issue they have generally been quick to tell me how they would be crucified by our national broadcaster (the ABC) should they take-up this issue.

    Almost every politician in the current Australian parliament is complicit in this fraud. (Craig Kelly, George Christiansen and Malcolm Roberts being obvious exceptions.)

    Best regards

  8. Forrest Gardener July 19, 2017 at 10:45 am #

    It may be informative to search the records for Goulburn and Thredbo for minimums of exactly -10.0 and for blank minimums.

    The best case scenario is that the BOM has an automatic process to identify spurious temperatures and is slow in fixing them. The worst case scenario is that they don’t bother fixing them until somebody makes some noise.

    In either case they should not be displaying fake data and their filtering process is deficient. It would be helpful for the BOM to have a “we were wrong” section on their web site, but I expect to see pigs fly long before that happens.

  9. Alfred (Melbourne) July 19, 2017 at 11:32 am #

    Hello Jeniffer,

    Last week, in order to help out the website http://euanmearns.com/ keep track of the energy situation in Australia, I got a server of mine to poll a website with the live data so that I could record it for posterity.

    If you provide me with the necessary URL’s, I am sure I can do that for any weather stations that you suggest.




  10. Mike Borgelt July 19, 2017 at 12:01 pm #

    We think it strange that there are starving Hindus while all that beef on the hoof is wandering around but Australia will end up the same way. Starving for energy while we have it in abundance. This will be known as “bad luck”.

  11. William Taylor July 19, 2017 at 3:23 pm #

    You write that the Australian Bureau of Meteorology limits the lowest temperature that an individual weather station can record. Do they also limit the highest temperature that an individual weather station can record?

    Seems relevant. Thanks for all your work. It matters!

    William Taylor
    Brooklyn, NY

  12. Siliggy July 19, 2017 at 3:36 pm #

    Frederick Colbourne
    “lognormal suggests the generation process is multiplicative instead of additive.”
    Have looked at a few reasons for nonlinear temperatures like the Platinum resistance thermometer probe. They unlike thermistors produce a positive feedback (self heating increases as temperature goes up). The electronics or software compensation for this may enhance spikes that the system takes in as real. While a thermistors negative feedback would cut spikes (self heating decreases as temperature goes up). However it is always good to look at the obvious real world reason for nonlinearity. In this case it would be snow. Snow will change the shape of the terrain as well as the thermal momentum and albedo around the site until it melts. As it melts it needs to cross the enthalpy of fusion which will cause a step change. Another reason is hidden in the site record history. It was for many years only recording from mid June to mid October.

  13. Siliggy July 19, 2017 at 3:50 pm #

    Alfred how about this page every ten minutes for 50 hours from 530 PM this Thursday evening? http://www.bom.gov.au/nsw/observations/nswall.shtml

  14. Alfred (Melbourne) July 19, 2017 at 4:46 pm #


    I checked your link. It seems to have the temperature for Goulburn only for 09:00. In some other places there is also a reading at 16:30

    Jennifer must have some other URL.

  15. Jennifer Marohasy July 19, 2017 at 5:35 pm #

    William Taylor,

    Thank you for your insightful comment.

    The Australian Bureau of Meteorology has acknowledged that it has put limits on the lowest temperature that an individual weather station can record – specifically that a lower limit of 10.0 degree Celsius was implemented for Goulburn.

    Two weeks ago the following questions were asked of the Bureau, so far they have refused to answer them:

    Q1. When was the lowest possible daily minimum temperature for Goulburn set at 10.0 degree Celsius?
    Q2. How was it determined that temperatures at Goulburn should not exceed this minimum of 10.0?
    Q3. At what upper limit does the Goulburn temperature reset to a lower value?
    Q4. What is the maximum temperature that can be recorded at Goulburn?
    Q5. How was the daily maximum temperature limit determined?
    Q6. When was the daily maximum temperature limit determined?

    Kind regards

  16. Siliggy July 19, 2017 at 7:05 pm #

    Alfred the Goulburn site of interest is “Goulburn Airport” all those numbers update every 10 minutes. we are interested in the current reading the low and the high. They update and change every ten minutes. Watch the page. If it does not change refresh the page. Are you able to log whole pages that update?

  17. jennifer July 19, 2017 at 8:08 pm #

    Hi Alfred,

    Your offer is a very kind one. I’m thinking the issue through, and will get back to you.

    Kind regards

  18. Alfred (Melbourne) July 20, 2017 at 9:22 am #


    I just checked http://www.bom.gov.au/products/IDN60801/IDN60801.95716.shtml with my “sniffer”.

    If you keep on that page, it refreshes the whole page every 10 minutes – and the actual table within that page seems to get a new line of data every 15 or 30 minutes.

    Technically speaking, this is the way things were done 10+ years ago. Nowadays, you would only need to send the new line and use Javascript to reconfigure the table.

    “Are you able to log whole pages that update?”

    Absolutely. However, I would have to filter out the data that has already been recorded. I could also do something similar for (almost) any number of stations. The problem is that I know very little about Australia’s geography and places like Goulburn Airport have no real meaning to me. I just checked it on the map and it seems to be halfway between Sydney and Canberra – that is the totality of my knowledge about it. 🙂


    The offer is open, but I will not come back to this page too often. I think you have my email

  19. MikeR July 20, 2017 at 3:39 pm #

    Hi Jennifer,

    Just a minor correction.

    For July 22nd, 2016 the minimum temperature for Thredbo AWS was +1.8C. The lowest minimum of -14.7C was recorded on July 28th 1980.

    There hasn’t been a temperature below-10.4C this century.

  20. Jennifer Marohasy July 20, 2017 at 4:02 pm #

    Hi Mike R,

    Thank you for finding and reporting this error. I have just corrected the blog post proper.

  21. MikeR July 20, 2017 at 11:03 pm #

    Hi Jennifer,

    Glad to be of service again. Yet another demonstration of the importance of quality control.

  22. Geoffrey Williams July 24, 2017 at 5:47 pm #

    Jennifer, I think that the BOM shuld make available a website where the Australian public can ask genuine questions regarding climate science. The answers could then be published on the same site for all to see. I wonder if such a mechanism presently exists.
    Regards GeoffW

  23. Geoffrey Williams July 26, 2017 at 12:40 pm #

    Hi Jennifer, listend to your interview this morning on 2GB wirh Alan Jones.
    Well done and well spoken on this matter of the BOM temperature recording data.
    I hope some of our politicians will sit up and take note of this matter.
    When the World temperatures are up 0.1 degree (hottest ever) they scream about it.
    How much more has the Australian BOM been ‘adjusting ‘ that we just don’t know of.
    Regards GeoffW

  24. hunter July 29, 2017 at 1:57 am #

    More people are noticing the systematic misadjustments of temperature data.

    The author of the essay at that link is very well informed and has the tools to analyze the situation quite well.


  1. Four-Steps Needed to Restore Confidence in Bureau’s Handling of Temperature Data - Jennifer Marohasy - August 8, 2017

    […] 1. Lift any limits currently placed on the recording of minimum temperatures; 2. Make publicly available the dates on which limits were first set (e.g. minus 10.0 for Goulburn), and the specific weather stations for which limits were set; 3. Advise whether or not the actual measured temperatures have been stored for the weather stations where limits were set (e.g. Goulburn and Thredbo Top); 4. Make publicly available the stored values, which were not entered into the Australia Data Archive for Meteorology (ADAM) – known more generally as the CDO dataset; 5. Clarify, and document, the specific standard applied in the recording of measurements from the automatic weather station (AWS) equipment including period of the measurement (i.e. 1-second or 10-minute average), checks in place to ensure compliance with the standard, checks in place to monitor and correct any drift, and temperature range over which the equipment gives valid measurements. […]

Website by 46digital