By Jennifer Marohasy
A few years ago, I reported that the Australian Bureau of Meteorology changed a minimum temperature at the official Goulburn weather station from a very chilly minus 10.4C to a chilly minus 10.0C on Sunday 2nd July 2017. After we protested, the Bureau removed the value altogether, and no value was recorded in the official archive as the minimum for that day for Goulburn. After we protested some more, what I incorrectly assumed was the real value of minus 10.4 was inserted into the online database. I now regret that value having been recorded in the official archive, because I now know it to be a fraud.
Then two weeks later, on Sunday 16th July 2017, I caught the Bureau doing the same thing at the Thredbo weather station in Australia’s snow fields – changing, then deleting, a very cold temperature.
I now know the temperature archive for Thredbo for July 2017 is also a lie – in fact it is a lie right back to May 2007 when an MSI1 card was installed limiting the temperature that could be recorded to minus 10.
For more than a decade, many weather stations across the Australian mainland and in Tasmania had limits set on how cold temperatures could be recorded. The archived data is what some of my friends now call Seinfeld data: when you see -10.0 or -10.4 in the Bureau archive or a blank during winter at any of the cold weather station, well the correct value could well have been -13.2 which is the temperature recorded at Thredbo on 28th August 2018, the winter after Lance and I had the cold day limit of minus 10.0 lifted.
Graham Lloyd first reported the story in The Australian newspaper, with a photograph of Lance Pidgeon and I making the front page, on 1st August 2017. That same evening I was on Sky TV with Alan Jones and eventually an internal ‘Review of the Bureau of Meteorology’s Automatic Weather Stations’ was called. A report by the same name was issued in September 2017, carefully worded, and so implying, everything was fine: that the Bureau was faithfully recording temperatures from its 695 Automatic Weather Stations spread across the land mass of Australia.
Nothing could be further from the truth.
The Bureau operates an error prone network of automatic weather stations (AWS) that have never met International Organisation for Standardization 17025 or ISO 9001 requirements, nor does the Bureau operate its network of automatic weather stations in accordance with World Meteorological Organisation (WMO) recommendations.
I’ve been discussing these issues with my friend and electronics’ specialist Lance Pidgeon over the last couple of weeks, and he has written a list of ten myths about the Bureau’s automatic weather system. I plan to detail these in this new series about legacy electronics at the Bureau.
Lance will quote extensively from email exchanges published as part of Freedom of Information document 30/6150 that includes email exchanges between members of the committee appointed to review the operations of the Automatic Weather Stations following the botched recording from Goulburn and Thredbo back in July 2017. The emails between senior Bureau staff and technicians from the New Zealand MetService are public because of a Freedom of Information request initiated by former Institute of Public Affairs colleague, Evan Mulholland. Despite the odds, Evan persisted with this FOI request, with the relevant documentation now public.
In a key email, Bruce Hartley the Systems Engineering Manager at the New Zealand Met Service explains:
The equipment wasn’t faulty. The purchase specification required operation down to -10 ˚C, so the words need to be carefully written.
Bruce Hartley is referring to the need to carefully word the final report ‘Review of the Bureau of Meteorology’s Automatic Weather Stations’ because most of what the Bureau had communicated to the Australian public about temperatures as recorded at Goulburn and Thredbo to that point in time was a big lie.
I use the word ‘lie’ deliberately, and perhaps as George from Seinfeld used it when he infamously advised Jerry: Just remember, it’s not a lie, if you believe it.
I have come to understand that Bureau staff, especially the director Andrew Johnson, spend a lot of time trying to convince themselves that even if they can’t forecast the weather, they have a system that can reliably record temperatures and that the subsequence need for homogenisation, is not proof of further fudging – even if it is.
This will become more obvious as Lance explains the electronics, the focus of future posts in this series. There are at least 9 myths to follow.
Myth 1. The Bureau does not set limits on how low a cold temperature can be measured.
The Bureau had set a limit of minus 10 Celsius across its Australian automatic weather station (AWS) network including at locations likely to record temperatures below this value. We know from page 53 of the Bureau’s carefully worded internal review, Review of the Bureau of Meteorology’s Automatic Weather Stations, that this limit was in place at Thredbo for some 10 years from May 2007 until July 2017 and at Goulburn Airport from November 2002 until July 2017.
On page 4 of the same document the Bureau acknowledges lifting these limits for the locations of Tuggeranong, Mount Baw Baw, Butlers Gorge and Fingal in October 2017 by removing the MSI1 card which had a particular equipment configuration that created the artificial limit of variously minus 10.0C and minus 10.4C.
The Bureau is yet to explain how long the minus 10.0C/10.4C limits was in place at those locations.
It is also yet to assure the Australian public that it has removed the MSI1 card that created the limit from the rest of the network.
To be clear, the original ‘full range’ design specification of the Almos MSI1, as per Technical Report A2671 was -10 to +55.
As Bruce Hartley of the New Zealand Metservice explained to the Bureau in 2017:
The equipment wasn’t faulty. The purchase specification required operation down to -10 ˚C, so the words need to be carefully written.
To be clear, temperatures may have reached much lower than minus 10.4C at Goulburn on 2nd July 2017 and at Thredbo on 16th July 2017. We will never know because the MSI1 cards limited recording of temperatures to 10.0C/10.4C.
It is the case that after the limits were lifted ‘temperatures plunged’ at Thredbo, to quote from an article by Graham Lloyd published in The Australian on August 4, 2017:
Recorded temperatures at the Bureau of Meteorology’s Thredbo Top automatic weather station have dropped below -10C in the past week, after action was taken to make the facility ‘fit for purpose’.
A record of the Thredbo Top station for 3am on Wednesday shows a temperature reading of -10.6C. This compares with the BoM’s monthly highlights for June and July, both showing a low of -9.6C.
The BoM said it had taken immediate action to replace the Thredbo station after concerns were raised that very low temperatures were not making it onto the official record. Controversy has dogged the bureau’s automatic weather station network since Goulburn man Lance Pigeon saw a -10.4C reading on the BoM’s website on July 2 automatically adjust to -10C, then disappear.
Later independent monitoring of the Thredbo Top station by scientist Jennifer Marohasy showed a recording of -10.6C vanish from the record.
BoM initially claimed the adjustments were part of its quality control procedures. But bureau chief executive Andrew Johnson later told Environment Minister Josh Frydenberg that investigations had found a number of cold weather stations were not “fit for purpose” and would be replaced.
The BoM has admitted that, in addition to Goulburn and Thredbo Top, stations at Tuggeranong in the ACT, Butlers Gorge and Fingal in Tasmania and Mount Baw Baw in Victoria would be replaced.
It angers me that we will never know how cold it really got at Goulburn on the morning of Sunday 2nd July 2017 or at Thredbo on morning of Sunday 16th July 2017 because the Bureau did indeed have limits set on how cold temperatures could be recorded. That information is lost forever.
In fact, the Bureau continues to operate a legacy system of outdated electronics and computing software developed in the 1990s with inherent biases, that mostly hype maximum temperatures. This combined with setting a cold limit of minus 10.0/10.4C means that university researchers relying on Bureau data have been able to claim that ‘record hot days are now 12 times more likely in Australia than days of record-breaking cold’ – Peter Hannam from the Sydney Morning Herald quoting Sophie Lewis and Andrew King from the ARC Centre of Excellence for Climate System Science. This fits the human-caused global warming narrative that is a reliable source of funding for academics, catastrophe stories for mainstream media, and government subsidies that prop-up renewable energy industries.
If the temperature recording system at the Bureau was overhauled, all of this would be put at risk.
Reliable temperature records would likely show only a modest increase in temperatures from the 1960s and that most of the record hot days occurred during the first half of the twentieth century. Furthermore, there would be no decline in cold days.
And until the Bureau reopens the Charlotte Pass weather station which holds the record of minus 23.0C for the lowest daily minimum temperature ever recorded in Australia, set on 29 June 1994, the chances of a new record minimum cold day are reduced. The Charlotte Pass weather station was closed on 31st March 2015. The Bureau would never close any of its hot weather stations, like Oodnadatta or Onslow. But it does have an aversion to new cold day records. The Australian Bureau of Meteorology is a disgrace.
It is all incredibly sad for those of us who care about the integrity of Australia’s historical temperate data and for accurately assessing climate variability and change.
*****
The feature image includes Lance Pidgeon and I at the Goulburn Airport on the morning of the 31st July 2017.
Eventually I get to discuss Goulburn and Thredbo in this interview with Alan Jones:
Fire every single employee of BoM and start over. All management is permanently barred from government employment or employment related to government.
Yep. And in the US, the FBI, DOJ and EPA.
Clean house, and prosecute, with new staff, all wrongdoing and biased prosecution and rulings not based on science, and ban all sudo-scientists from ever receiving any grants or being employed at any institution that collects federal funds.
My question, is this being done elsewhere
Dear Jennifer and Mike,
Like all her hand-waving and shouting about Mildura, Cape Otway and Wilsons Promontory, she is also probably not aware that the low-T card issue was analysed and discussed by Greg Ayers in 2018 (https://www.publish.csiro.au/es/es18002).
To save Jen the difficulty that she seem to have with data, I point out that in relation to Goulburn Ayers said: “A nine-minute gap in the Tmin record at Goulburn on just one day, 2 July 2017, is assessed to have no effect on that data record, as the likely actual minimum temperature was estimated to be essentially unchanged at -10.4°C +0.0/-0.1°C”. He did a pretty simple graphical analyses on page 17.
As for Thredbo, Ayers identified “six days within the last decade in which Tmin was inadvertently limited to -10.4°C”, and used three different tests which “pointed to the conclusion that the small number of limitations has had no material impact on the Thredbo record”.
I guess your viewpoint depends on whether you think the problem was inadvertent, or a deliberate oversight on the Bureau’s part.
Yours sincerely,
Dr Bill Johnston
http://www.bomwatch.com.au
Oh wait,
When I went and photographed the weather station at Goulburn airport on 17 April 2016, the Vaisala Ceilometer was there but there was no Stevenson screen! Furthermore I chatted to a few people and no body knew where it was!
Cheers,
Bill
Bill, Not sure how you missed the Stevenson screen, as per my many photographs from Goulburn it is just the other side of the airstrip and very visible from the hangers.
In other places on the internet from 2017 Bill has given the date he was there as 7 June 2016 not 17 April 2016, which was a Sunday. Unlikely to have been anyone there let alone people who don’t know what a Stevenson screen is.
While the claim could throw Ayers conclusions out, it is an irrelevant distraction from the point that the BoM do not comply with the WMO standards that specify function of the AWS down to minus 80 degrees.
You keep saying this stuff Lance as though it matters. For the 30 years since 1991 (to 2020, which is last time I analysed Goulburn T-data), there have been just four occurrences of Tmin less than -10 degC. 1994, -10.9, 1999, -10.1, 2000, -10.2, and 2017 -10.4.
So why would they need an electronic probe that measures down to -80 degC? You know, or you should know, that PRT thermometers do not perform equally well across a wide T-range, which is why they specify the range they think they require, based on the data they have.
You and Jennifer are also loose with the truth as to NATA certification of the Bureau’s laboratory facilities. Instead of blowing-off you and Marohasy could have easily found out that the Bureau IS accredited (Bureau of Meteorology – Accredited Organisation (Site No. 24147) – NATA)
All the best,
Bill
I did not miss the screen Jennifer, had it been there I would have photographed it. No one else I spoke to knew where it was either. Up to March 2016 the site was at Lat – 34,8054, Long 149.7312; they may have moved the AWS infrastructure first and installed the Celio and moved the screen later. But who knows? You don’t.
You are right, thanks Lance, my typo, the photograph was taken on Wednesday 27 April 2016, at 11:46. I later straightened it. The airport was reasonably busy at the time I was there, with flight training and aircraft moving about and I was unable to walk beyond the hanger – there was no screen to see anyway.
All the best,
Bill
The metadata for Goulburn is here:
http://www.bom.gov.au/clim_data/cdio/metadata/pdf/siteinfo/IDCJMD0040.070330.SiteInfo.pdf
Here is the map as of 1 March 2016:
Thanks Nick, and yes I have been through the metadata (the earliest matadata I have is dated 26 November 2016, the most recent site diagram therein is dated 17 November 2010).
There was still no screen there on that day.
Cheers,
Bill
“There was still no screen there on that day.”
WOW.. so yet another sign of gross incompetence from BoM and its leaders. !
All that money they get, and you can’t even keep track of screens.
Again this is all a distraction from the basic reality that the WMO standards specify the range -80 to +60. If the BoM had met the standard the problem would not have happened.
However compare the Bill nonsense to a sensible report on what platinum resistance thermometers do and are. Bear in mind also that “electronic thermometers” are semi- conductor thermometers, something entirely different to a PRTD. He is really mixed up. This quote from link below.
“The platinum resistance thermometer—in which the principle of measurement is the variation in the resistance of a platinum wire as a function of temperature—is generally accepted as the most accurate temperature measuring instrument available. Its sensitivity and reliability are evident from the fact that it was first used in 1928 to define the International Temperature Scale from −190° to 660°C and has thus been the primary international standard for over thirty years.”
Link.
“Again this is all a distraction from the basic reality that the WMO standards specify the range -80 to +60.”
They do not.
I’d give you an upvote for this comment, but I don’t vote on comments. Science is not a popularity contest.
But anyway …
Lets say dear Jennifer, that instead of digitizing the 10,000 A8 forms for Mildura that they may have thought occupied you for the time being, that at another weather station far, far away say at Halls Creek in Western Australia, there was an overlap dataset of 4.5 years that you could compare with thermometer data measured in the same 60-litre screen. I’ve been researching Halls Creek and I know it well.
Let’s say that you compared daily overlap Tmax thermometer data for the MO site (ID 2071) with AWS data (ID 2012) observed in the same screen and that you used PAST from the University of Oslo to do the comparison (https://www.nhm.uio.no/english/research/resources/past/), which, because you don’t know much about statistics, I know you won’t.
Further, that as you thought that because data was observed on the same day you would ignore all the caveats and use paired t-tests. Just a scenario OK, and you can download the data and analyse it for yourself.
The paired t-test (Sign and Wilcoxon tests and Monte Carlo likelihood) show that without any shadow of doubt the difference between instruments of 0.04oC is highly significant (Psame <0.001).
But it is not true. Probability statistics are grossly inflated by autocorrelation – the propensity for data-pairs to predict values for the following day.
Examined as unpaired populations, there is no significant difference in means, medians, distributions or anything else including extremes.
Your credibility is on the line and so is that of the IPA, whom you represent.
You don’t study metadata for the sites you use for hand-waving, and you are making fake claims about differences that are immaterial in the context of temperature trends
In reality, as you cannot defend your claims, you are making a mess of the whole issue of climate trend and changes.
Yours sincerely,
Dr Bill Johnston
http://www.bomwatch.com.au
.
” you don’t know much about statistics”
Unbecoming insult, whether true or not
It is not an insult. It is an argument that has been going on for at least 5 years, which in anybody’s book is too long. Furthermore misuse of statistical tests is a critical issue in messaging. Claiming differences are significant when they are not is unethical and wrong and if the Bureau were doing it she would be up them for the rent.
The whole basis of this series of posts stems from Jennifer’s obsession with not being given overlap data for Wilsons Promontory in 2015. It is now 2023, and the obsession goes on. However, that data never existed in the first place – she simply ignored the relevant metadata and with the IPA behind her, raised it with politicians and ran around finger-pointing.
I had raised a similar issue in respect of Canberra, which is a bureau site, in 2017, but I don’t have financial backing like Jennifer has, so I let the matter rest. https://joannenova.com.au/2017/08/another-bom-scandal-australian-climate-data-is-being-destroyed-as-routine-practice/
Like Thredbo and Charlotte Pass, Wilsons Prom was NOT a Bureau-run site; it was run originally by the Lighthouse Service, then Parks Victoria, so why is she up the Bureau for a site they did not operate? Thredbo is/was run by NSW Parks & Wildlife Service, Charlotte Pass by people at the Chalet.
The same for Cape Otway – the Lighthouse Service, then Parks Victoria. She said that a dry-bulb thermometer they (the BoM) installed measured maximum temperature, which it does not. She is writing these posts as a scientist and “expert witness” and it is reasonable to expect that she knows this stuff. People reading this and other posts should be able to trust what she says. Cooperative sites that contribute data are NOT Bureau sites and they cannot be compelled to abide by Bureau policies, period, as they say.
But how do readers measure or evaluate trust? How many fundamental errors are too many? Whatever side you are on, misleading people is unacceptable.
At Mildura, which is a Bureau-run site, she said a new AWS-probe was installed when metadata shows it was not. She did no site research, not even basic stuff like satellite images. She has claimed in multiple forums that a difference of 0.22 degC is statistically significant, including in her recent WUWT post. But it is not – could not possibly be. I’ve argued the Mildura thing with her before, and now, as I don’t have her data, I’ve provided an example of the wayward use of paired t-tests on time-correlated differences using overlap data for Halls Creek.
A search of “paired t test assumptions” brings-up 27.6 million hits in 0.04 seconds, so it it is no secret what those assumptions are. It is also reasonable in the context of a scientific debate, that statistics be used appropriately. Crikey, imagine if this was Pfizer claiming some tiny number was significant when it was not!
I am also not asking anyone to take sides – present arguments by all means, but there are no sides, Since the 1990s, Australia’s Bureau of Meteorology has lost its way, it has become weponised and narrative-driven, basically from the top-down. This needs to be exposed, but whatever is said must be accurate and defensible.
All the best,
Dr Bill Johnston
Nice, only looking at Tmax not Tmin, which this post is about.
Why that choice?
Tmin data is available too. I thought I’d leave it you to examine.
Cheers,
Bill
Unbecoming to challenge an obviously honest claim for which Mr. Johnston has nothing to gain by lying.
One demerit.
So who are you going to believe, Bill Johnston or your lying eyes? Are you saying the picture above in the post is a fake?
-20 down votes for a reliable source who photographed the weather station and did not see a screen is outrageous.
Do you people trust no one?
I consider Bill Johnston to a reliable source of Australian climate information, based on my experience of reading up to 20 climate science and energy articles every day of the year.
It matters not if the problem is inadvertant or not. What matters is that science depends on accurate data. If anyone is hand waving, it is you. There is no excuse for provisioning measurement systems incapable of measuring the temps they are supposed to measure. You cannot say that it wouldn’t have been significant because you don’t have the data to prove that. Even if that were correct, there is nothing to say that future temps will be the same.
Every time I hear someone trumpeting a new high as a sign of impending doom, I wonder if there are new lows that went unreported. Again, you cannot say because you do not have the actual data.
So stop with the excuses, fix the equipment, and make it clear to the public what you’ve done and how you are correcting it. That’s how you build both accurate data and public trust.
Dear David,
Of course it matters. Specifications did not envisage that temperatures would decline below -10.4, which for 99% of stations is the case. Furthermore, the BoM immediately rectified the problem (see (https://www.publish.csiro.au/es/es18002)).
Don’t get me wrong, I have little regard for the Bureau as a propaganda tool of the warming-brigade, but not because of trifling decimal-point failures, which are nothing more than convenient distractions. They could regard that despite the cost to them, that keeping Marohasy busy entering data from 10,000 A8 forms for Mildura as a major victory for example. I would if I was them; but I’m not them. I am strictly unpaid and independent.
The Bureau has simply gone to the other side where shouting and alarmism means more than running a weather station network. They also fail on both counts.
Your pinkie would be challenged to tell the difference between -10.4 degC and -11.0 degC. Your eyeballs would be challenged to even measure that difference using a meteorological thermometer at 9am at Goulburn or Thredbo, and even if your pencil worked, your hand would be challenged writing the value legibly in an A8 Field Book, especially if it was raining or snowing and you had to write inside a plastic bag.
Because they have never observed the weather, Lance and Jennifer (and good old Nick Stokes) would know nothing of this. They think data just jumps-out to be accurately embedded in the Field Book … the Register … the internet …
Whatever your perspective, is is not such first and second decimal-point differences that meaningfully impact a time-series. The problem lies elsewhere which is what I am showing in the various analyses and reports I have presented thus far at http://www.bomwatch.com.au.
All the best,
Bill Johnston
(scientist@bomwatch.com.au)
Hi Bill,
I’m not sure how you can be so sure about the statistics when the Bureau has admitted that it had equipment that would only record to – 10.0. They lift the limit and the temperature goes down. There would be no recording of -13.2 from August 2018, except I got the limit removed in July 2017.
That equipment was in place at Thredbo from March 2003 until 27 July 2017, and at Goulburn from November 2002 until 10 July 2017. That is a long time, given those weather stations have not been around for very long. Many of the alpine stations were only installed in the 1980s.
To be clear, the MSI1 card was only replaced at Goulburn and Thredbo because I called them out, got some media.
And Goulburn was only discovered because it is the closest weather station to Lance and he was watching it on the morning of 2nd July 2017, worrying his pipes were going to burst. He had moved into a new house.
Thredbo was only discovered because I decided to watch in on 16th July 2017 because I thought it likely to be cold that morning at Thredbo, and I was wondering if it would ‘flip’ as Goulburn had for Lance.
I reckon there is more to be discovered, but not unless you take an interest in the issue.
Smokescreen
“Whatever your perspective, is is not such first and second decimal-point differences that meaningfully impact a time-series.”
Again, the issue isn’t the time series, it’s stating how many record hot and cold days there has been. You are avoiding the issue.
Dear Tim,
I am not avoiding the problem, Tim. I routinely analyse tails of data distributions and it is clear that due to infrastructure changes, warm extremes are increasing at the expense of low extremes. It is also clear this is unrelated to the true climate.
You may not be aware (Jennifer in particular) that by changing data distributions, counts of warm extremes can increase without impacting the mean or median. I’ve been routinely observing this, including applying a range of valid statistical tests since I first developed BomWatch protocols almost a decade ago.
If you think about it, switching the argument from ‘trends’ to ‘extremes’, while keeping Jennifer busy with 10,000 A8 forms is not a bad strategy. Maybe they could make another 10,000 A8 forms for mythical overlap at Wilsons Promontory; or for that dry-bulb thermometer at cape Otway, she thought was a Max; or for the second probe at Mildura that metadata showed was never installed.
In my mind, while there are lots of big issues, getting sweaty about tiny little numbers that have a frequency at Goulburn of just 4 occurrences in 30 years (4 days in 10,556 actual Tmin observations to 2020), is small cheddar. (I’ll have a look at Thredbo later in the day.)
To be clear, while Greg Ayers has done a credible job looking at the low-Tmin problem, I strongly disagree with his comment that “that the frequency of occurrence of cool days at Thredbo, including days below -10.4°C, has fallen compared with the late 1960s due to the general warming of Australia since then.”.
His own paper that he referred to in that regard, is as dodgy as some of Jennifer’s statistical analyses.
Yours sincerely,
Bill Johnston
Dear Tim and all,
I use a standard routine for analysing daily T data.
I have just completed a preliminary look at Tmin at Thredbo Top-station and this is what I found:
Summary of daily data downloaded 14 February 2023; summarized to the end of 2022.
Although the site (ID 71032) opened on 26 Nov 1966 observations during summer months declined markedly after 1976 to as few as 100 observations in total per year from 1986 to 1996. July August and September had the the most complete data (I’m surprised that Jennifer or Lance did not at least inspect the data … but oh-well.)
After the AWS was installed on 20 April 1997, it reported whole degrees only, until March 2003 and sporadically until 5 December 2003 (i.e., for about 6 years). Whether data were rounded-up or down or just truncated is unclear. This is understood to be a communication issue not a temperature probe issue. There are also data-gaps of up to 5 to 7 days duration, which for unmanned electronic gear operating in harsh environments is not uncommon.
Leaving out the data between 1997 and 2003 and previous manual data, total counts of Tmin less than or equal to -10 degC = 19 (i.e., 19 individual days). Of those only 2 days recorded less than -10.4 degC.
The frequency of low Tmin (<-10DegC) was higher before the AWS was installed, but this had nothing to do with the climate. Restricting analysis to just the three winter months may provide useful insights.
While I’ll look at the data more thoroughly another day, it is really not much use for climate studies.
All the best,
Bill
“Your pinkie would be challenged to tell the difference between -10.4 degC and -11.0 degC.”
Then why the heck are all the AGW clowns yelling and screaming about a degree or so of maybe warming in 200 or so years.
You have just destroyed the whole AGW meme with that one sentence..
Well done. !
Mr. nice: I see you beat me to it, well done.
Dr. Johnston: Has your pinkie detected global warming?
I’ve said repeatedly that warming is not detectable in any of the 300 or so Australian weather station datasets that I’ve examined. The Bureau’s favorite trick is to adjust for changes that made no difference to the data, or not adjust for changes that did. By doing that they can create the trend they want.
Learmonth is an example of a site they did not adjust, but they should have! (see: https://www.bomwatch.com.au/bureau-of-meterology/part-5-potshot-and-acorn-sat/).
The way to argue the case, is to use straightforward independent statistical methods to detect change-points, and make adjustments that are proportionate to the effect of impacts on the data. I build other checks and balances into my methods which are fully explained here: https://www.bomwatch.com.au/data-quality/part-1-methods-case-study-parafield-south-australia-2/ (Scroll to the end where full reports and data are available for download.)
I also use a range of supporting information including archived maps, plans and aerial photographs from libraries and the national archives, and I undertake diligent research into historical information about weather station sites.
Fogging the debate by worrying about four instances when Tmin dropped below -10 degC out of a total of 10,556 observations over 20 years at Goulburn AP is hardly rational or sensible especially when the worrywarts don’t pay close attention to detail – who ran the sites for example, were the WMO limits –80 °C to 60 °C or was that the likely range, with limits being set locally.
Yelling “fire” “fire” when there is no fire comes to mind.
All the best,
Bill Johnston
4 points where they just happened to be following the temperature readings in real time and found limitations and adjustments or outright removal.
How many times, where they were not in real time following the temperatures, did Tmin readings, adjustments or deletions happen?
A PROVEN 4 days of manipulation, how many manipulations happened at other times? And the 4 times Tmin were all shown to be the warmer than reality.
Just run that out to all low temperature readings, since the manipulations of “extreme” low temperatures reported were 100% of the time, then no Tmin temperature of -10c can be assumed to be valid, and the whole of the temperature records are invalid.
BTW, somewhere above someone stated the BoM is “accredited”, but it has been shown that the accreditation is faulty since the BoM cannot record the full range of temperatures at these, and other, locations. So the Accreditation is in itself faulty, and of no use.
Anthony’s study of station siting, what first brought me to WUWT, shows how badly government temperature data is collected, and this and many other posts show how really badly the records are “maintained”. When the equipment is moved from the middle of a field to between a paved parking lot, a building and beside an AC condensing unit, THAT is the epitome of UNSCIENTIFIC data. Without a LONG period of measurements from both locations and old and new equipment, the data is worthless to science, but great for showing a warming trend. Limiting Tmin, is also great for that, although if you keep the truncated data, using T sensors that read to below -10c will probably end up showing a more recent DOWNWARD temperature trend.
BTW: Is the BoM still using instantaneous temperature readings for Tmin and Tmax?
Yes David. And if only the weather station at Charlotte Pass was operating through July 2017 we might have a new coldest temperature for Australia. It is currently -23.0. There were blizzards through the high country that July 2017. It may have been as cold as it was back in 1994 at Charlotte Pass when it got down to -23.0. We don’t know, we can’t know, because the Bureau decommissioned that weather station. They decommissioned the weather station that holds the record for the minimum temperature for all of Australia.
If only, if only…
They closed the Kiandra Chalet when it ceased being a Chalet too and all along they have depended on volunteers or other organisations to take and contribute observations.
The data for Charlotte Pass was very sporadic and broken. No data from about 1975 to 2005 and lots of missing data after that. It was also never a ‘Bureau site’ that they closed; it was observed by people at the Chalet who probably got sick and tired or disinterested in doing the job. Most of the sites in the Mountains were run by SMHEA, not the Bureau, but you would not know that because you are too busy shouting and waving your arms to do the research.
Although I have not analysed the data, Thredbo-top was also very scatty up to April 1997 when the AWS was installed. Before that it would have been observed by resort or NSW NPWS people, not “The Bureau”.
Your obsession gets in the way of you undertaking investigations that you could do do yourself. Remember, you did not check metadata for overlap data at Wilsons Promontory before you raced around banging-on about data that did not exist.
There are also unanswered questions about Cape Otway and Mildura and your dodgy statistical methods. Now you are getting hot and sweaty about NATA certification, that does exist.
You have not looked at Goulburn data sufficiently to make a mature judgement-call about the importance of data colder than -10 degC. Perhaps you don’t have a protocol for examining data, like I use at http://www.bomwatch,com.au, or you are not disciplined in your approach to data and metadata.
All the best,
Dr Bill Johnston
http://www.bomwatch.com.au
“Perhaps you don’t have a protocol for examining data”
BoM has a protocol for weather stations
Pity that they totally ignore it !
The weather stations could be perfectly sited, well spaced for good coverage, with perfect measurements, but that would not matter if you can not trust BOM to compile an honest Australian average.
“that would not matter if you can not trust BOM to compile an honest Australian average”
You can do it yourself, with the data they post.
Really Nick
What would the first step be if wanting to confirm these three month means?
Here My gut feeling is they are too low.
The problem is not only the equipment, it’s mainly a problem with the people who own the equipment.
No, we don’t think it was any sort of oversight. It was designed in to produce exactly the false warming it did. The entire BOM has shown itself to be untrustworthy and should be abolished.
I would go with rebuilt from the ground up.
No “scientists” who ever were signatory to any “global warming” “peer reviewed” literature or previous BoM employees, less any whistleblowers, need apply.
All station siting to be reviewed to meet actual standards.
All previous stations to be maintained and records kept, side by side, with the new equipment and locations for 5 years minimum.
Just a couple of suggestions.
“A nine-minute gap in the Tmin record at Goulburn on just one day, 2 July 2017, is assessed to have no effect on that data record”
Very misleading. While it makes little difference in monthly averages it *does* make a difference when talking about the number of very cold days.
“university researchers relying on Bureau data have been able to claim that ‘record hot days are now 12 times more likely in Australia than days of record-breaking cold’ “
If you don’t know how many actual days of record-breaking cold days there were then this statement is meaningless.
you really went with the…. We investigated ourselves and found nothing wrong… defense. Come on man. Its not just one day and it does make a difference the way these clowns homogenize and infill. The final sentence in your provided link is pure political BS and not science. ” the frequency of occurrence of cool days at Thredbo, including days below -10.4°C, has fallen compared with the late 1960s due to the general warming of Australia since then.” really, so you manipulate the data and then claim that the data supports your conclusion because you manipulated it.
I don’t and have never worked for the Bureau. So “we have not investigated ourselves”. Neither do I agree that “the frequency of occurrence of cool days at Thredbo, including days below -10.4°C, has fallen compared with the late 1960s due to the general warming of Australia since then”.
You have a beef then contact Greg Ayers:
greg.ayers@bom.gov.au.
I don’t believe Thredbo data are sufficiently reliable to draw such a conclusion, I have not manipulated the data and I’m not satisfied that the site has never re-located. I am also unconvinced that those writing this post have undertaken a deep-dive into Thredbo data.
All the best,
Bill Johnston
“I guess your viewpoint depends on whether you think the problem was inadvertent, or a deliberate oversight on the Bureau’s part.”
Are you saying that the DIDN’T KNOW that the thermometers were limited to -10C (ie gross incompetence on behalf of BoM and it’s leaders)
or that they DID know.. ie basic fraud.
Can’t be the politicos’ fault since they just depended on the INCOMPETANCE of their “climate scientists” to set the range for the temperature sensors.
OR the sensors they chose were chosen due to their costs meeting politically set costs, as in smaller range, lower price. And when that plays into your scenario, all the better.
Reading through the ‘likes’ and ‘dislikes’ ticks on comments that Bill Johnston has made, it seems obvious to me that many readers seem to assume that just because Bill is critical of some of What Jennifer has to say, that he is automatically on ‘the other side’ to Jennifer.
It also seems to be that many of those giving negative ticks to Bill, are also automatically assuming that he is actually an advocate or defender of the BOM. In some cases, it also appears that some of the negative ticks/comments come from readers who even seem to assume that Bill is actually a staff member within the BOM.
As a colleague of Bill’s let me assure everyone that nothing could be further from the truth. If you really get down to his analyses on bomwatch you will come to realise that Bill is actually one of the BOM’s most severe critics.
Possible the feature of Bill’s protocol is that he does not start out with ‘the end in mind’ to prove that the BOM is wrong on all counts. In the case of each weather station analysed, he starts out with an open mind and begins his analysis of each site with an impartial, strictly rigorous, and statistically sound adherence to the scientific method.
As with some of his comments on this post, he is prepared to give the BOM credit where credit is due and this is the act of a fair minded scientist.
He has repeated this process – not starting out with the end in mind – over and over again in his work on the BOM weather stations in Australia. And, there results are utterly damning of the BOM.
And to state my position about bill once again, he is an absolutely rigorous scientist and probably one of the most hard hitting critics the BOM has.
I fully agree with this excellent comment!\ !
Good comment.
If you still have a bottom dollar you could bet that it is.
Jennifer Marohasy has been working assiduously to try to keep the Bureau and others honest. Any kind of acknowledgement from those organisations comes only after prolonged effort on her part, and even then there is no action to correct anything, they just work even harder to block her efforts. What a nightmare. Many thanks, Jennifer, for your efforts, and may you prevail eventually (or better, right now).
Really Mike.
Why doesn’t she post her analysis and data on https://jennifermarohasy.com/jenns-blog/, like I publish mine at http://www.bomwatch.com.au?
She is clearly in it for the selfies and the shouting, not the science.
Kind regards,
Dr Bill Johnston
Hi Bill,
I have never thought of my blog as a place for too much data.
If you are looking for my scientific publications you can always do a google search:
https://scholar.google.com.au/citations?hl=en&user=ZWlxl7wAAAAJ&view_op=list_works
If you scroll down the list there, you will find something by me and Jaco Vlok about Great Barrier Reef temperatures from a few years back that I think a bit more substantial than your recent waffle.
There is a list of my climate science publications, not up to date, at my climate lab website: https://climatelab.com.au/publications/
I keep my blog for more popular and newsy items.
Be honest: for the selfies and the shouting, not the science.
Anything fora noise.
Cheers,
Bill
Bill, I don’t think you are being honest here. Why don’t you just say you don’t like her methodology and process. But guys, you should be getting together to fight the scourge that is the BoM’s processes and clear AGW bias, not ticking off each other!
Keep up the good work, both of you.
In politics, the NOISE is incredibly important.
Is she getting RESULTS?
Are you?
“Is she getting RESULTS?”
Is she?
I have looked. For instance, your “something by me and Jaco Vlok about Great Barrier Reef temperatures from a few years back”, was actually Authored by Jaco Volk (and you as second Author) and it is basically a data catalogue, with some graphs.
My report on Great Barrier Reef temperatures (https://www.bomwatch.com.au/bureau-of-meteorology/trends-in-sea-surface-temperature-at-townsville-great-barrier-reef/) presented a deep-dive into most of that data.
While I tried to make it simple, I’m sorry you did not understand it.
All the best,
Dr Bill Johnston
I live in Walcha in northern NSW,1065m asl. Our AWS was removed in 1972 by the BOM. There is a limited data base from a few years prior to then. We reach very low winter minima due to topography & elevation. For example, in July 1970, between 13th -18th, the highest minimum was -10.6 and 6 morning average was -12.1. The nearest current AWS to us is at Woolbrook about 15km away. They too record very low minima. The record low is from 19/6/1994 at -14.5. -10 has been recorded there in five separate months between May & September.I have personal records which show similar lows. Jennifers expose makes me wonder whether similar BOM skullduggery has/is occuringin my neck of the woods!
Of course Walcha did not have an AWS removed There were no such things in 1975, when the temperature record ceased. Nor is there an AWS at Woolbrook. It has a mercury thermometer, with an alcohol one for minima. None of these is subject to any kind of cut-off.
Woolbrook
And at what temperature does the alcohol freeze?
You might want to check your comment of no cut-off? Science is clearly not your friend.
“And at what temperature does the alcohol freeze?”
-114°C
Which is why it is used in the minimum temp thermometer.
So it does have a limit, a cut off. And I think your answer is wrong. The alcohol normally used is ethanol, (see wiki..) and it is used down to -70C.
https://en.wikipedia.org/wiki/Alcohol_thermometer
Do you always type without checking your facts?
Care to be more specific with your answer?
Have you ever stuck your finger in some fluid held at -70 degC? Could you tell if is was colder; or say 10 degC warmer and would your finger really care?
Cheers,
Bill
I think it very important to know how cold it gets at Antarctica, and those are the type of temperatures that need to be able to be recorded to know if it is getting colder. :-).
Important for who, the penguins?
1970s cold
1980s cold
1990s cold
2000s cold
2010s cold
2020 to 2023 still cold.
More CO2 in the atmosphere obviously does not warm Antarctica.
case closed
You seem very familiar with testing temperatures using fingers.
Is that the method BOM defaults to?
Don’t forget the purposeful distraction and evasion by his throwing in the irrelevant;
When the discussion is about honest accurate temperature recordings, throwing around claims that a person could not tell a difference of -10°C is intentional sophistry.
That’s the best BOM can accomplish?
a) “Look, a squirrel” speciousness.
b) A blatant refusal to work on accurate temperature stations and temperature measurements. Until forced by public opinion?
Cheesy at best.
The Bureau operated very few of Australia’s weather stations in fact. How would define an honest, accurate observation and how would you test that it was?
Like it not, data are coarse at best and are not up to the task of measuring small trends and changes. The use of spread-sheet regression methods is entirely misleading.
If you want to know how temperature data should be analysed, check out any of my reports at http://www.bomwatch.com.au.
I recommend Charleville, for instance (https://www.bomwatch.com.au/bureau-of-meteorology/charleville-queensland/). You can access the full report near the end of the frontstory. (Or Amberley,Rutherglen, Townsvile, or the latest series focussing on data homogenisation).
If you want to check anything the data is embedded in the reports or available as a data package.
All the best ATheoK
Dr Bill Johnston
No I’ve always relied on thermometers. As I observed the weather and used A8 forms in rotation with colleagues for about a decade, I have a rough idea of the routine, the pitfalls, issues relating to infrastructure and problems relating to uncertainty. Its not the fun-job you think it is.
Bill
You are correct Bill, The world cannot tell a difference of 1C, so why all the fudging of once-was-data to support the climate agenda. ?
There is obviously no “climate emergency”.
Just a credibility emergency at BoM.
I don’t disagree at all.
Bill
“Do you always type without checking your facts?”
You asked for the freezing point of alcohol.
Ron
Hence the need in earlier days for the Walcha defrost –
Stand on the bonnet (hood) and pee on the windscreen
“It angers me that we will never know how cold it really got at Goulburn on the morning of Sunday 2nd July 2017 or at Thredbo on morning of Sunday 16th July 2017 because the Bureau did indeed have limits set on how cold temperatures could be recorded. That information is lost forever.”
This stuff just gets pettier and pettier. And more obsessive.
Australia has over 700 AWS stations. Almost all never get down near -10°C. A very few have done so, on very rare occasions. Pretty much just high places like Goulburn and Thredbo, on one or two very cold mornings. These are not major stations. They are not in ACORN, or GHCN. The low temperature would be of momentary interest to locals.
Whether Goulburn was -10°C or -10.4°C on the frosty morning of 16 July 2017 will have not the slightest effect on any climate statistic.
“I now regret that value having been recorded in the official archive, because I now know it to be a fraud.”
You have nothing to back that.
Read the story , the evidence shows the ‘electronics card’ had a preset minimum it would register of -10C or -10.4C
Wheres your evidence -10C or lower is ‘rare’, and ‘doesnt matter’ never been to Tasmania have you
Yes, this was thrashed out years ago, as Jennifer indicates. The Murdoch outfit liked the story. Anything for a bit of BoM-baching.
Tasmania? Well lowest ever minima:
Hobart -2.8°C
Launceston -5.2°C
New Norfolk -5.7°C
Swansea -5.0°C
Lying by intentional omission is lying. They intentionally omitted, they lied.
And then had an internal review, even my husband saw through it: https://www.spectator.com.au/2017/09/not-really-fit-for-purpose-the-bureau-of-meteorology/
“even my husband saw through it”
In the US, husbands would take that as an insult !
OFF Topic: Whatever happened to the great old family photo of you as a child with your parents, and their station wagon, that used to be somewhere on your blog?
Very difficult to get anyone to report on this. When community radio stations have me on, they suddenly find that the BoM withholds future weather reports etc.. When regional ABC have had me on, journalists tell me they are in trouble, and have been warned not to make the mistake twice. News Ltd is almost big enough, but not quite.
“When regional ABC have had me on, journalists tell me they are in trouble, and have been warned not to make the mistake twice.”
Gangster tactics, practiced by the Weather Guys. What kind of world are we living in?
I notice that you list a bunch of places that are heavily influenced by the UHI effect. Cherry picking at its best.
OK, do you know of any place in Tassie that has got down to -10? I think the only one is Butler’s Gorge, which is in a similar situation to Goulburn (high plateau).
Reply to Nick Stokes – You should check the BoM Daily Extremes pages
http://www.bom.gov.au/cgi-bin/climate/extremes/annual_extremes.cgi?climtab=tmin_low&area=tas&year=2023
1 -14.2 07 August 2020 Liawenee 96033 1057 -41.90 146.67
2 -13.0 30 June 1983 Tarraleah Village 95018 585 -42.30 146.45
2 -13.0 30 June 1983 Butlers Gorge 96003 666 -42.28 146.27
2 -13.0 30 June 1983 Shannon HEC 96021 939 -42.05 146.75
5 -12.5 01 July 1983 Butlers Gorge 96003 666 -42.28 146.27
5 -12.5 24 June 1972 Bothwell (Franklin Street) 95001 352 -42.39 147.01
7 -12.2 09 July 2013 Liawenee 96033 1057 -41.90 146.67
7 -12.2 24 June 1972 Campbell Town 93036 200 -41.93 147.48
9 -11.9 24 June 1972 Palmerston 93027 180 -41.79 146.99
10 -11.7 10 August 2020 Liawenee 96033 1057 -41.90 146.67
10 -11.7 24 June 1972 Oatlands Post Office 93014 406 -42.30 147.37
And the BoM table is not yet down to minus 10
Yes. So what is going on here? Jennifer tells us that BoM is covering up temperatures lower than -10. Yet here you are rattling off BoM data.
Looks like Liawenee got a proper thermometer..
Unlike Charlotte’s Pass and other places in the Snowy… etc.
Seems BoM KNEW that proper thermometers existed, and decide to use the ones limited to -10C in the Snowy region.
Why would that be ?
“Why would that be ?”
Well, OK, why?
That’s the thing about these conspiracy theories. They just don’t make sense. Why would the BoM try to lie about low temperatures in one region when it is happily reporting them elsewhere?
Anyway Goulburn is not in the Snowy region.
Only one is this century. Maybe still doesnt have the new hardware
Well, three. Maybe it is getting warmer.
Mostly earlier data with the old fashioned max and min type thermometers based on the years
You goose… All those places are near sea level. How about Liawenee in the Central highlands, August 6th 2020 -14.2 degrees Celsius.
Nick will try any deceitful/disingenuous little trick he can think of..
get’s caught and exposed most times. !
Lowest ever ‘minima’ ?
Just days after recording the first heavy snowfall in 40 years in Launceston, Tasmania has now recorded its lowest ever temperature.
Liawenee, in Tasmania’s Central Plateau, dropped to a bone-chilling -14.2C on Friday about 6am.
August 2020
‘The previous lowest temperatures were recorded on June 30, 1983. Tarraleah Village, Butlers Gorge and Shannon in Tasmania all recorded -13 that day.
Bureau of Meteorology supervising meteorologist Simon Louis said Liawanee actually got colder than Casey Station in Antarctica.’
Whats it called when you deliberately omit facts ?
Likewise whether Penrith (NSW) recorded 47.3C or 46.5C on Jan 7 2018 is neither here nor there statistic-wise but of course it’s not about statistics it’s about propaganda.
The BoM hypes maximum temperatures in the media recorded on super-sensitive instruments in comparison with older mercury or alcohol thermometer records to feed the ridiculous ‘climate crisis’ narrative.
Consequently the country that was one of the world’s cheapest energy producers thirty years ago is heading to be one of the most expensive for no rational reason.
The UAH temperature record for continental Australia shows a net 0.7C – 0.8C rise since 1979 when the world was supposedly entering another ice age.
What seems to be missed is that for at least Goulburn, Thredbo, Tuggeranong, Mount Baw Baw, Butlers Gorge and Fingal and probably a heap more … we don’t know how cold it was getting. When it went below -10.0 it would stop at -10.4 and then flip to a blank or to -10.0.
-10.4 WAS NEVER THE CORRECT TEMPERATURE, IT WAS THE CUT OFF. THEN IT WOULD FLIP, PROBABLY OFTEN TO BLANK.
Hi Nick
I think that you are missing the point about the -10.0 and -10.4.
That after it went to -10.4, it flipped to -10.0 suggested to Lance and me that something was up.
My guess is that the correct temperature for Goulburn on the morning of 2nd July is more like -13.8; but I can’t know because the BoM had a limit on how cold temperatures could be recorded.
You don’t have a problem with a thermometer unable to register temperatures below -10C being used to record daily lows? Really?
“university researchers relying on Bureau data have been able to claim that ‘record hot days are now 12 times more likely in Australia than days of record-breaking cold”
If you don’t know the number of record cold days then how can this statement be meaningful? Most stations never get near record breaking hot days either.
Nick is now saying that temperature below-10C DON’T MATTER..
A massive face-plant into his own BS.
Not that rare at all. Happens every winter here on the Northern Tablelands.
Try and find the ‘lowest temperature recorded in Australia’ in all of 2022 and you are out of luck
They cleverly tell you the ‘lowest daily maximum‘ instead. As a general search
Thredbo -4.8C on 7th june
Presumably if they have daily max, they have the daily min for that day and others.
They know you are onto them
They are closing the cold temperature stations, while opening stations in warm places, I’ve written about this here: https://www.spectator.com.au/2017/09/not-really-fit-for-purpose-the-bureau-of-meteorology/
And I am surprised there is no outcry from the closing of Charlotte Pass.
Jennifer must be squarely over the target
Nicks in here whining his ass off again
The bylines I look for when reading about Australia’s climate and energy are Jennifer Marohasy, Geoff Sherrington, Rafe Champion, Bill Johnston and Jo Nova. Their climate articles almost always get recommended on my climate science and energy blog’s daily list of up to 24 of the best articles I’ve read, which will include this one tomorrow. Honest Climate Science and Energy
BOM, by the way, stands for the Australian Bureau Of Mediocracy.
After BOM are done homogenizing the numbers, they will start pasteurizing the numbers.
In fact, to save money, the BOM in 2024 will begin pulling their average monthly Australian temperature number out of a hat, or from two feet below the back of a hat. They are currently debating what kind of hat to use.
Probably an ass-hat.
As the sun comes up this morning, I am feeding the barnyard. To pass time as I wait for the horses to finish, I pull up weeds, as is my wont. As is its wont, my brain starts feeding me a story, this time about a guy in Australia that checks a temperature probe, and find it reads 5 degrees too low. He proceeds to write a very learnéd paper on his finding, after which the whole world adjustificate their temperature records upwards by five degrees.
I told my mind not to be silly, and carried on with my work.
…and this article must be the first thing I open? Prophesy or just a mind soaked in bulldust?
One day there will be a low temperature that willl be the same accross the entire region, exposing the BOM for what it really is.
Edenhope vic bom units a disgrace
not only does it NOT record rainfalls corectly frequently IF it ever shows the Barometrics then the temps vanish!
its prob only shown bar about 5 times in the 26yrs Ive been here
and temps that WERE registered manage to go blank a day or two later
theyve done a reno on the airport(rare use) to extend runway for air ambos and firies and thrown new shed up
no idea where the setup is as they keep the gates locked
funnier is the approved very costly new temp setup in our main street on the library
its permanently 3c ABOVE what even the crapola airports reads are
It isn’t so much the electronics, which seem to have worked as designed. It’s the specifications. Replacing temperatures less than -10C with -10C is bad design, bad science and shows how corrupt the BOM was even then.
Yes even the new replacement printed circuit board is only rated down to -25 degrees C. So our record low of -23 Degrees C is only allowed to be broken by 2 degrees C.
“bad science”
It was a programming decision by whatever programmer set up the card. It dealt badly with a contingency which in Australia hardly ever arises.
BoM must have known the capability of the card.
And used it anyway.. or maybe BECAUSE OF?
No weaseling out of it this time, Nick
Incompetence of your fellow AGW stall-warts.
Odd how Bill is flipping off a half a degree or two, saying it’s nothing to see here or there. The whole climate change meme hinges on the arse pluck 1.5C of which half is already achieved without sky falling on our heads.
Laughable.
“Laughable.”
Quite HILARIOUS, actually.
Bill has just DESTROYED the whole AGW meme with one statement.
I’ve been busily destroying it for years, I’m glad the message id finally getting through!
Cheers,
Bill
I thought that one of the nefarious tricks of the climate alarmists was to lower the historical temperatures, not increase them.
This is about a design that prevents modern low readings after it is installed. Rather than cooling the past it is warming the future.
From the article: “Reliable temperature records would likely show only a modest increase in temperatures from the 1960s and that most of the record hot days occurred during the first half of the twentieth century.”
There’s that pesky “Early Twentieth Century Warming” again.
The same Early Twentieth Century warming, climate temperature data mannipulators tried to erase from the temperature record.
The Early Twentieth Century warming was worldwide. The historic, written regional temperature records tell the story. It’s a lie to claim today is the hottest time in human history. Not even close.
The Australian Bureau of Meteorology is a member of the World Meteorological Organisation.
For them to meet the WMO standards they need to meet this:
Guide to Instruments and Methods of Observation Volume I –Measurement of Meteorological Variables 2021 edition. ANNEX 1.A. OPERATIONAL MEASUREMENT UNCERTAINTY REQUIREMENTS AND INSTRUMENT PERFORMANCE REQUIREMENTS 1.1 Air temperature. Range Minus 80 to plus 60 degrees.
Annex 1.A. is a cut and paste in of an older document from earlier standards with a different page orientation. It was Annex 1.D in the 2008 version. If they really did meet these standards this problem with low measurements not being able to be recorded could not have happened.
https://library.wmo.int/doc_num.php?explnum_id=11386
Lance Pidgeon.
Dear Lance,
In the notes it also says:
Column 2 gives the common range for most variables; limits depend on local climatological conditions
So no, Lance, the BoM does not have to meet -80 to +60 degC. The limits they use “depend on local climatological conditions”. Clearly they use gear in Tasmania that records Tmin < -10 degC.
As I have explained above, Goulburn AP experienced 4 instances out of a total of 10,566 observations over 30 years where Tmin was less than 10 degC. Even if it was really five, six or 10, it is still trifling.
I make the point that the common range is different to limits and also that ignoring the fine-print, or caveats, does not help your argument in this case.
All the best,
Dr Bill
At the date in November 1996 when the Platinum thermometer became the primary instrument only the MSI 1 Almos card existed. This means that no AWS anywhere in Australia let alone Tasmania could measure below -10.4 degrees C until the MSI 2 card was tested and began to be installed many years later. This is not a limit it is a lack of function. This means that a design that did not function over that “common” range prevented readings from being taken let alone getting being flagged by a Q.A. limit for being outside a calibration range.
You may try to prove that there were no temperatures that low but that is because there could not be one by design.
May be true Lance and there is no denying that it was useful to call it out.
However, you are still only telling half the story. You have said repeatedly (in other posts and forums) that as a member of the WMO, the BoM needs to meet the standard of minus 80 to plus 60 degrees, which is simply BS. They are not obliged to do that at all.
You could also think about the fact that Goulburn AP experienced just four instances out of a total of 10,566 observations over 30 years where Tmin was less than minus 10 degC.
I just managed to download Thredbo data and I’ll have a look at it shortly (The BoM site is really playing-up, a conspiracy perhaps?).
All the best,
Bill
Far less than half the story. This is part 1. The electronics should be able to measure and report over the common range. Limits are for Q.A. flags. As i have shown the platinum will work from -200 to + 1000 with extreme accuracy from -190 to +660 degrees. There is no need for a design limit on function. Only a limit on what might pas Q.A.
Looking forward to seeing your data Lance. (Although I have no idea why they would us a scale that was hundreds of degrees higher/lower than the range they need to measure, which is around say -15 to 55 degC.)
Cheers,
Bill
Bill
“the range they need to measure, which is around say -15 to 55 degC.)”
Perisher Valley ski center has been down to -18 but you would not let the young Perisher AWS only opened in 2010 go down 0.8 of a degree more than it already has, knowing that nearby Charlotte pass has been down to -23.
Why…?
Lance,
In fact the BoM has two cards in common use.
“The MSI2 can record temperatures over a broader range (nominally –25°C to +55°C) than the MSI1 (nominally –10°C to +55°C).”
The MSI2 card is used where -10°C might be a problem. Apparently the issue at Goulburn and Thredbo arose when the MSI2 card originally installed was replaced by an MSI1 card, by people who weren’t aware of this difference. The simple remedy was to restore the MSI2 card.
I acknowledge that I have to do counts within years to come up with a more accurate figure than 4 out of 10,566 observations.
(At this point of time, for another day.)
Bill,
Here is a list of minima for Goulburn, in order:
Year Month Day Minimum.temperature..Degree.C.
1690 1994 8 17 -10.9
1683 1994 8 10 -10.6
10045 2017 7 2 -10.4
3819 2000 6 15 -10.2
3502 1999 8 3 -10.1
10044 2017 7 1 -9.7
Nick
You forgot the deleted days.
And which would they be?
They look like this without the quotation marks
” ”
This winter just gone at Thredbo gets interesting. At Thredbo village the temperature got down to -9.7 On July 20. This is very cold for that site. Not abnormal for the AWS site to be a full degree colder than this one. Perhaps -10.7?
http://www.bom.gov.au/climate/dwo/202207/html/IDCJDW2133.202207.shtml
On July 20 the temperature at Perisher went down to -11.7.
http://www.bom.gov.au/climate/dwo/202207/html/IDCJDW2112.202207.shtml
On July 20 the temperature from Thredbo AWS is gone, again.
http://www.bom.gov.au/climate/dwo/202207/html/IDCJDW2132.202207.shtml
Another one of those innocent coincidences, right?
Siliggy = Lance Pidgeon.
Again, none of this conspiracy stuff makes any sense. Why would they suppress a reading of -10.7 (your guess) at Thredbo when they are posting a minimum of -11.7 at the nearby ski resort of Perisher?
More to the point of the missing Thredbo cold day is why is it missing? Is it running on solar power and thus more likely to fail after days of cloud and snow during the shortest daylight season?
Is it poorly sited so that it gets covered in snow. Was a Q.A. limit triggered that was left to default to delete a low temp?
This time arounf all we know is it should not have been the data logger not functioning below -10.4. Did it go below the new Almos card’s abilities?
Here is the full data for Thredbo in 2022. I’ve marked missing data in red. There is quite a lot. Only one is a very cold morning (OK, I suppose you’ll claim that is suspicious too). Are you going to quiz all of them?
To set the record straight, I’ve been referring to absolute Tmin/year not counts of data <10 degC within years, which will be higher of course.
b.
Dear Dr Bill,
Bull. (in a word).
When installing a sensor at the north end of the Cape York peninsula, it may make sense to install a sensor which cannot read temps below 10C. If temps below 15 C are not uncommon, it would be prudent to install a 0C capable sensor (and probably more cost effective, there are a lot of 0-?? range applications) Given the trivial cost, it would be even more reasonable to just install -80 to +60 sensors in every AWS. Making all the units interchangeable solves all kinds of logistic and data problems.
However, it appears that the real issue was not sensors that could not measure low temperatures, but software that rejected the low temperatures and made some other entry instead. A blank is bad if you actually have a measurement, but inserting a default measurement is far, far worse. Now that we know some measurements are changed from what the sensor reports without annotation, how do we know that any given entry is correct? You can’t argue no one would do that, because we have just seen that somebody did.
So yes, it may make sense to install a narrow range thermometer on the Nicaraguan Plateau, or in Manila, or on Waikiki Beach, but it makes better sense to use the same thermometer everywhere. It makes no sense to replace a “out of range” reading with a default number in the instrument record. Either record the number as received, or flag it for validation, but never discard or overwrite. You argue as though just meeting the letter of the recommendation is the goal. The goal is a trustworthy record, not check marks on a list.
It is the card that does the work of converting the signal from the sensor/PRT-probe thermometer (whatever you want to call it) into T; not the sensor itself. Lance has made this clear. I have no idea of the actual range of the sensor itself (i.e,. the one they use) – whether it is capable of reading to 400 degC or minus 100 degC; but it is card that does the work. I just assess the data, and I like to understand how the system works. especially how it handles outliers.
I also make the point that I collaborated with a supplier in the 1980s to assess a particular brand of AWS/probe combination by comparing it in its own screen, with daily observations measured by thermometers in a Stevenson screen about 3m away. I did not publish anything, neither do I still have any data. I also deployed several brands of commercial AWS at remote field sites that ran for up to 5 years.
Rapid-sampling AWS-probes are prone to spiking – temperatures that can be out-of-range, for a number of reasons, including the shelter frosting or de-frosting, electrical connections etc. as well as briefly exceeding their calibration range.
Calibrations (the conversion of the pulse into a value) are not linear, but are quadratic. The problem identified in the 1980s was that ‘centering’ the calibration across the range being observed, could cause problems in the tails – i.e., out-of-range spikes, The tails of course is where Tmin and Tmax occur. While most days were unaffected, a really hot (cold) day could produce a spike, because of a calibration miss-fit (which I call an over-ranged value). However, lengthening the calibrated range reduced day-to-day precision (increased variance). I understood that to be the tradeoff, but Lance may have data that shows otherwise.
That problem may also not be the case now, but is was then, and the manufacturer at the time is now part of another company. It was also 40 years ago. However they made good gear and provided issues were understood, they performed well, unattended under field conditions.
Given all this, error-trapping is both valid and important. The problem of sitting-up all night observing data coming-in at 3am, is that some real-time data could be erroneous. The Bureau does use error-trapping routines that are outlined in various publications. If 10 1-minute values in a row exceed the limits of the error-trap, they are immediately flagged as NA – not available. Make of that what you will. However, at some point in the process erroneous data has to be flagged as such. Why not at source? In order to check anything, who is able to handle masses of 1-second or even 1-minute data anyway?
I accept that the problem of the cards is real – knocking-off low values in particular is of concern, but how much concern is warranted, when manually observed thermometers are not that ‘accurate’ either? After digitising 10,000 A8 pages, perhaps Jennifer will be able to tell us. While there is some variation, there is no difference in Halls Creek comparison data for example, but that is only one site that I happen to be working on.
I believe that given the overall quality of manually observed data, making hay out of first and second decimal-place differences is not warranted. A bubble in an alcohol thermometer above the index, would be sufficient to cause an error in the lowest value for Charlotte Pass for instance – it was NOT a Bureau site, but did anyone check the resets or undertake any other QA on the data?
(No-one pays me by the way, not the IPA, not the Bureau, no-one.)
Kind regards,
Dr Bill Johnston
“I have no idea of the actual range of the sensor itself (i.e,. the one they use)”
It is a WIKA TR-40. Maker gives a range from -196°C to 600°C. So the lower bound of -10, if it ever existed, came from somewhere else. Obviously it existed in the mind of whoever programmed the card. But it may simply have been that that is the range down to which the BoM did tests.
It would be interesting to know the percentages of hotter inland measurement sites that have been added in Australia, compared to the percentage of cooler sites, over the last ten years.
i know of one example west of Derby WA, where a cooler site on the islands was replaced with a much warmer site further inland at a military airfield. Then they touted the very hot temp measured one summer at the new site as an “example” of global warming!
When you’re a ‘sceptic’ it is OK to query data from so-called alarmists, but other sceptics usually really dislike it if you query anything from them or another member of their group. I guess at heart most people desperately want to be a part of a team and want to protect other members of that team. And they want everyone within the group to admire them and their contribution, uncritically. And they want everyone within the team to get on.
I tend to be trusting of other people, and I tend to share a lot of information. I also have a tendency to check information that I think is important, whether it is from my mother or Al Gore. I like to be a part of a team, but I am also happy to be what Irving L. Janis in his seminal works ‘Group Think’ describes as the devil’s advocate.
There is a big difference between being a devil’s advocate and a troll. A devil’s advocate takes the issue at hand, and considers it from different perspectives. A troll twists and distorts what was originally said first and/or throws up something unrelated and tries to derail the conversation that way.
Some years ago, as part of a private email group, I queried some of the information that was being provided to the group by Bill Johnson. I did that within the group, and in good faith. Bill Johnson really didn’t like it. He continues to dislike me.
The first time, from memory, it related to his statistical analysis of data from Rutherglen. He went on about his statistical method that showed a discontinuity in, from memory, the early 1960s for Rutherglen. I suggested the discontinuity he was finding was likely caused by the absence of data for a period of time, rather than anything else. I was very familiar with the same data. I did not publish this anywhere, or write a blog post about it. But rather I simply drew the issue to the attention of the group for discussion.
I was subsequently tossed out of that group. Not immediately. But I had so upset Bill he never forgave me. Not to his day.
Back those few years ago I was told by the moderator that every time I posted something, or comment on something within the group, that it so upset Bill, that it now upset everyone else so the easiest thing to do was to ask me to leave.
Not content to have me gone from that group, Bill Johnson now follows me about the internet often posting comments in threads that distract from the key point I am trying to make.
He is more troll than devil’s advocate.
Take one of Bill’s first and early posts at the above thread, complete with photograph, claiming there is no Stevenson Screen at the Goulburn Airport.
Bill actually sent this photograph (attached again) to that same private email group back some years ago. I remember he wrote about how there was no regular Stevenson Screen at Goulburn Airport. He wrote:
“Here is picture of the Goulburn AP AWS I took a few years ago. It seems to me that they use a new type of screen. Yes it is a fair distance from tarmac; I used a strong telephoto to get close from the door of a hanger (nice man would not let me get any closer!). It is possible that the non-Stevenson screen box is aspirated (has a fan that draws in air). I could not see any evidence of another screen …”
Bill could not see any evidence of ‘another screen’ because he didn’t get past the back door of the hanger. He never got permission to go onto the airfield or to venture infront of the hangers. It is indeed impossible to see the Stevenson Screen that existed at Goulburn Airport through 2016, and that still exists today at Goulburn Airport, from the backdoor of the hanger. I’ve had that confirmed just this morning in a phone conversation with a Goulburn local.
Unable to get permission to get onto the Goulburn airfield and unable to see the Stevenson Screen from the door of the hanger, Bill invents a story about a non-Stevenson screen with a fan.
In doing this he draws attention away from very real issues, while also seeding nonsense information into the minds of some of the best and brightest within our community of sceptics.
FYI.
Thanks Jennifer,
But there you go again .. I did not claim “there is no Stevenson Screen at the Goulburn Airport” – you just made that up, on the spot, out of thin air. Like the parallel data for Wilsons Promontory, … the second probe at Mildura … the nonsense about the site at Rutherglen, which that even the Bureau later confirmed DID MOVE in a special report done just for you. It is not possible to have a sensible conversation along these lines.
You have also put a lot of people to a lot of trouble over some of this stuff, including photocopying 10,000 A8 forms. You are a very good science communicator but your messages have to be solid, not flimsy and easily knocked-over. Why you published the Ruthergaln thing three tines or more in IPA publications, after you knew the site had moved, is beyond comprehension. I DON’T DISLIKE YOU, I don’t have to like you either, nor you me, but I think some of your claims are pretty fragile. Some are just untrue.
Demanding a greater level of scientific rigor in what you say is not trolling. I am also a scientist, not a skeptic, and I take care that my discussions are supported by data or metadata. Did you or Lance examine the data, metadata or satellite images for Goulburn or Thredbo, is a pretty good question at this juncture. My rapid assessment found Charlotte Pass data is better than what I thought. (Lots missing, but better data.)
Lack of a screen at Goulburn mystified me. I drew a line-of-sight diagram using Google Earth Pro, based on the image for October 2013, from where I was standing at the hanger to the anemometer mast (which is atop the AWS base), using coordinates provided by the camera (see attached). I later identified the white box as a Ceilometer, which metadata notes was installed on 13 October 2013. The previous GEP image (Sept 2012) shows lots of gear in the vicinity of the mast but none of it was there when I took my photograph. There was nothing outside the field of view either.
While you say “It is indeed impossible to see the Stevenson Screen that existed at Goulburn Airport through 2016” the photo you use on your blog shows you walking towards the hanger (or terminal) and the screen being to the right of the Ceilometer, would have been in clear view from where I stood at the hanger. Your picture at least solves that problem, but it creates another, which is that you say some mate at Goulburn said it would not have been in view of the hanger; which your own photo shows is BS – you just make it up, which is infuriating. Irrespective of anything else, the screen was NOT in that location the year before you photographed it.
I’m not intimating you are telling stories; I am just stating what I found a year earlier. I also noted that metadata provided different coordinates for the site in 1988 and that those were changed on the 1 March 2016 (I was there on 27 April).
Kind regards,
Bill
Well this is a great distraction from the BoM AWS’s not functioning over the -80 to +60 degree range specified by the WMO isn’t it.
However.
Bill seems to have forgotten that Stevenson screens slope toward North with sides narrower than the front. He also did not notice the pine trees behind you Jennifer in your photo that cannot be seen in his photo.
BS Lance. There is no -80 to +60 degC requirement. The door of the screen faces south, therefore the back faces north. The hangers et al. were to the east and Marohasy’s blog photograph shows that the screen would have been visible had it been there.
Before firing-off neither of you sat down and calmly looked through and considered the evidence – your photographs, the data, satellite images and metadata and arrived at a defensible position.
You have no idea how many daily observations could have been affected by the card issue at any of the sites. Furthermore you accused the Bureau of something they had no control over at Charlotte Pass – it was never “their” station in the first place. While it was desirable that they established an AWS at Charlotte Pass, it had nothing to do with the BoM, that the people at the Chalet decided to quit.
You have both consequently pushed the wrong angle onto the story and I can really understand if they (the BoM) treat JM as a serial pest. She and you need to present evidence not hand-waving.
This is not science and as an IPA member and supporter, she needs to lift her game – considerably.
I’m looking forward to whatever data that you have about the probe issue.
All the best,
Bill Johnston
The dictated minimum common functional range you and Nick Stokes both claimed is not there in the WMO standards, is there. You as usual attempt to insert extra confusion for your old CSIRO/BoM buddies by deliberately mixing up local climate software Q.A limits with the common WMO wide minimum functional range specification and starting unrelated bickering. .
Guide to Instruments and Methods of Observation Volume I –Measurement of Meteorological Variables 2021 edition. ANNEX 1.A. OPERATIONAL MEASUREMENT UNCERTAINTY REQUIREMENTS AND INSTRUMENT PERFORMANCE REQUIREMENTS 1.1 Air temperature. Range Minus 80 to plus 60 degrees.
The Stevenson screen slopes down to the North and as you say the door is to the South. The box is not a rectangle so the lack of pine trees in your photo shows Jennifer walking toward me and the photographer. Not the hanger door which faces at a near right angle to the screen location from memory. I think the door is under the dark sky light line near the plane you can see on the track between many hanger doors. I have no idea which hanger you gained access to but it would not be the door you can see clearly open behind us. There were planes in that area that they would not allow anyone near. I was warned at for taking a step in that direction. There are no pine trees in your background. .
Thanks Lance,
The Table you referenced (Anex 1.1.) clearly stated:
(See explanatory notes at the end of the table; numbers in the top row indicate column numbers.)
The notes say:
Notes:
1. Column 1 gives the basic variable.
2. Column 2 gives the common range for most variables; limits depend on local climatological conditions.
Which words don’t you understand and I’ll try to help.
Nick Stokes said the PRT they use is a “WIKA TR-40. Maker gives a range from -196°C to 600°C. So the lower bound of -10, if it ever existed, came from somewhere else”.
Importantly, as -196 degC is cooler than your -80 degC, and 600 degC is warmer than your 60 degC, the PRT unit itself (the WIKA TR040) does have the capability of measuring temperature across the range you specified, which is -80 to +60, which is not actually mandatory (see above under Notes).
Having chosen the WIKA, the two remaining problems were calibrating the device for “local climatological conditions”, and achieving a time-constant that was close to the response time of LIG thermometers (presumably Tmin, Tmax and dry-bulb thermometers, whose response times maybe slightly different – but only slightly; seconds not minutes).
As I am nowhere near a lab, I can’t help with that, but if I was and I had access to meteorological and lab-grade thermometers, using time-lapse photography I could probably rig-up a small experiment to test that.
WIKA have a technical datasheet that discusses calibration (WIKA data sheet IN 00.29), which is a separate, specialist exercise altogether. The Bureau does this in association with their supplier and in their NATA-certified laboratory in Melbourne.
As for the Stevenson screen that wasn’t there, it was simply somewhere else. Data are continuous, but the screen moved between the time I was there on 27 April 2013, and the time-stamp of the satellite image, which was October 2013 (no day stamp), which shows the screen in its current position on the far right.
Nevertheless, I remained curious why the screen was not in view in my photograph, but was clearly in view in yours, which is why I did the field-of-view pic, and why I raised it within the “informal email-group”.
The screen was not hiding, it simply was not there.
At the time, I thought the “white box” which was actually the Ceilometer, was a new type of screen, which I also thought was worth raising with the “group” – which your quote of my words makes clear. I also said “I could not see any evidence of another screen …”. Although I don’t have the exchange anymore, you jumped up and down in your usual way and eventually the “group” dissolved.
Metadata notes a change in coordinates on 1 March 2013 from Lat -34.8985, Long 149.73114 to -34.8085, 149.7311, which is current site. Looking at Google Earth Pro sat images, the first of which was for October 2005, I believe the previous site was about 360m to the NE (drive past the skydive centre about 170m and it would have been on your left). However, as nobody I talked to knew where it was, I did not go looking. (The October 2005 image shows the shadow of the AWS-anemometer.)
Between my photographs, your photographs and Google earth Pro, unless you have further to add, the mystery of the Goulburn Stevenson screen is now solved. I’ve attached a pic showing the two sites.
You both need to stop finger-pointing, do some research and lift your game.
All the best,
Bill Johnston
(That is it from me, you have wasted enough of my time.)
In all the discussion about temperature, it is rarely mentioned about how temparture is affected directly by a number of factors, none of which have ever been used when various organizations compare temperature around the world.
The area pressure, the wind direction, it’s speed and the surface the wind has traveled over, relative humidity and the height above sea level of the measuring station. All of these factors should be considered when temperature comparisons are made but they are not and that is why such data as average world wide temperatures or country average tempartures are nonsense.
In criticising the BOM in Australia about temperature, what is overlooked is that they are far worse in their forecasts and their knowledge and records of severe thunderstorms, cyclones and even fog. There is not one person in the BOM that now has first hand experience over a wide range of locations and in fact they are told not to look outside at the weather. They are bureaucrats in the true sense of the word and can’t even define what a severe thunderstorm is
Thanks R.K.. You make important points. I’m less concerned about climate variability and change from the perspective of the politics than to ensure reliable data for weather forecasting. I wrote to the relevant minister about all of this in 2014, https://jennifermarohasy.com/wp-content/uploads/2014/03/Birmingham_2014_08_12.pdf
In July 2011 I visited with Oscar Alves who heads seasonal rainfall forecasting at the Bureau. I wanted output from their POAMA simulation model to compare with output from my statistical model built using artificial neural network technologies. I eventually got the data I needed, and published the first of a series of papers with John Abbot, Application of artificial neural networks to rainfall forecasting in Queensland, Australia. Advances in Atmospheric Sciences, Volume 29, Number 4, Pages 717-730. doi: 10.1007/s00376-012-1259-9 .
We sent that papers to meteorological journals in Australia, the UK, and US. They all queried the relevance of a skilful seasonal rainfall forecasting based on AI technology. Eventually the papers was published in Advances in Atmospheric Sciences, by the Chinese Academy of Science.
The Bureaucrats you mention including Oscar Alves tell me that it is actually impossible to forecast seasonal rainfall, and that the method I developed with John Abbot using historical data/pattern recognition can’t work because the climate is on a new trajectory.
Jennifer,
Oscar Alves is correct in that you can’t forecast seasonal rainfall anymore than you can accurately forecast weather more than a couple of days ahead. The reason being that the pressure patterns and frontal systems control the weather and rainfall and too many influences affect them, not just on the surface but higher up.
Even the rainfall data they record is wrong and distorted. Back about three years ago in February a friend who owns a cattle station just south east of Blackall in Queensland told me that in speaking with a BOM person in Brisbane he was told that his property had received 300 mm of rain the day before as recorded on the BOM AWS on his property. When he told the BOM guy he was wrong, that his rain guage had only recorded 180 mm of rain and that no other property in the surrounding region had recorded more than that, the BOM guy said their data must have been correct and that’s what went into the records.
Their flood measuring of the Barcoo River was also out by 1.2 metres again measured on his property which borders the river and again the wrong data has been recorded. Quite arrogant they were. So, their records of everything are not to be believed – like cyclones Yasi, Ita and Marcia were never Cat 5 but only Cat 3 as all the recorded data of central pressures and wind gusts never supported a Cat 5 being proclaimed.