Mao of Australia's Acorn-sat temperature network.

Craig Kelly and the “warmist climate cult” conspiracy to rort Australia’s temperature records


Craig Kelly has claimed that “fraudsters” at the Bureau of Meteorology have deleted and hidden temperature records for “Australia’s hottest ever days”, and that these records are “inconvenient facts for the warmist climate cult”.


These claims are incorrect. No records have been deleted or hidden. Some data has been corrected and “homogenised” but it makes little difference to the clear overall warming trend.

Craig Kelly has focused on a few isolated readings which have minimal impact on the overall trend and then made baseless allegations against a large number of national and international experts. He has ignored the clear trend in the data that shows significant heating up.

This case is a good example of a conspiracy claim based on misunderstanding, but explaining it requires a little detail.

The claims in detail

On 10 July, Craig Kelly posted this comment:

Although the Bureau of Meteorology might delete and hide record hot days of the past (that are inconvenient facts for the warmist climate cult) the original records are not destroyed, they are stored by the National Archives and that’s where I spent a several hours today.”

Then on 12 September he posted this:

“And a memo to the fraudsters that have re-written Australia’s climate history by deleting Australia’s hottest ever days from the records – you haven’t been forgotten about. As soon as this Covid mess is sorted you’re next.”

On 15 July on Sky News, Craig Kelly claimed he had found 4 temperature readings higher than BoM’s stated maximum. He named a 50°C day at Wilcannia in 1939 as one of these. In a recent series of emails to Ian McLeod, he enlarges on this, saying:

“the BoM have deleted this record hot day of the past. The deletion is not part of any ‘homogenisation’ it’s simply been deleted from the BOM’s list of records.”


To understand why these posts are misrepresenting the facts, we need to understand the various Bureau of Meteorology (BoM) temperature records.

BoM temperature records

BoM measures temperature at hundreds of sites around Australia. The historical record also includes many sites that have been discontinued.

The data from stations can be inconsistent over time because:

  • the exact location of some stations may have changed, or
  • the environment around the station has changed – e.g. trees may have grown, buildings constructed, etc, or
  • the equipment used may have changed.

All of these can affect the records by introducing systematic errors. With changing land use and technological improvements in measuring equipment, these inconsistencies are inevitable.

For the records to be useful, especially for calculating trends, the data has to be “homogenised” – i.e converted to what it is estimated would be the case if the present weather station had done the measurement. This short video explains it. It isn’t “fraud” to do this – it would be misleading and inaccurate not to do it.


The Bureau has chosen 112 measuring stations for inclusion in its Australian Climate Observations Reference Network – Surface Air Temperature (ACORN-SAT). The maximum and minimum daily temperatures from these stations are used for the long term trend analysis.

Often the records from several nearby stations are combined and homogenised to give a consistent record for that location. The homogenisation procedures are well documented.

More than half the stations in this network have a full 110 years of homogenised records and almost all have more than 50 years. This means there are something like 3-4 million daily maximum and daily minimum records.

The homogenisation of ACORN-SAT records has been reviewed by a Technical Advisory Forum which includes 4 Professors, 2 senior public servants and two others, most of them experts in statistics. The processes used are common throughout the world, and the Australian procedures have been published and peer reviewed favourably.

Discussion of Craig Kelly’s claims

“the Bureau of Meteorology might delete and hide record hot days”

The records are not deleted, but are readily available. They are not hidden away, and can be viewed, as Mr Kelly has done . Some records have been adjusted and homogenised as discussed above, but the original reading is always retained.

Mr Kelly seems to think that changing a record (without destroying the original reading) is a sign of fraud, when in reality, if done for good reasons, is making the data more useful.

“inconvenient facts for the warmist climate cult”

There is no inconvenience and no warmist cult. Just good science recognised by almost all climate scientists.

With 3-4 million daily maxima in the ACORN-SAT record, a change in a few high temperature readings will make very little difference. BoM has calculated the unadjusted and homogenised data and shown there is little difference in the average and no practical difference in the trend (see graph below).

This example shows that it is the climate change sceptics who find the data contradicting their beliefs. Australia, like the world, is clearly getting hotter. Mr Kelly is ignoring the full set of data to focus on a few changes that make no tangible difference.

50°C day at Wilcannia in 1939

The record hasn’t been deleted as Mr Kelly claims. If it was, he could never have found it. The original record is in fact still available in the archive and was included in the 2009-10 Year Book Australia (Table 1.15) published by the government.

This year isn’t included in the ACORN-SAT data, which begins only in 1957. There has been a station at Wilcannia since 1879, but the earlier data has apparently not been considered reliable. The record page for Wilcannia notes: “when an event has produced suspect data [it] may result in an extreme event not being recorded.”

[Correction: 7 December]
This year isn’t included in the ACORN-SAT data, which begins only in 1957. In an email, BoM said:

“The Bureau is aware of the 1939 Wilcannia observation of 50.0 °C, and this is currently the highest daily record that has been discovered for NSW using a standardised Stevenson Screen.

The record does not appear on official Bureau publications since the daily records from that period for Wilcannia are undigitised. Nonetheless, it has been referenced in a recent report on extreme heat in NSW.

Work to digitise historical daily temperature paper records from Wilcannia, along with other historic paper records, is ongoing. Once this has occurred, the Wilcannia observation will be made available on the public web, including on the table of historic extremes for NSW, which is based on observations that have been digitised.”

But as noted above, the addition of one very hot day in 1939 would not make any discernible difference to the rising trend.


Some temperature records have been changed by BoM. The processes for doing this have been reviewed by experts and considered to be reliable.

Has this been part of a “warmist” conspiracy as claimed by Mr Kelly?

Neither Mr Kelly or I have expertise in meteorology and statistics (aIthough I have some knowledge and experience in both), nor do we have access to the reasons why certain records have been adjusted. We have insufficient information to make an informed judgment.

The advisory forum endorsed BoM’s approach. It included university, government and privately employed experts. One member, Dr. Phillip Gould, is currently the Assistant Secretary, Office of the National Data Commissioner at the Department of the Prime Minister and Cabinet. The BoM methodology has been internationally endorsed.

So I have a choice about who I believe.

Either, I can believe that the BoM scientists and data managers, these eminent people on the advisory forum, and the international reviewers, are all part of a conspiracy to mislead the government, the Australian people, and the world.

Or, I can believe that Craig Kelly doesn’t understand meteorology and is criticising for ideological reasons what he doesn’t understand.

I don’t find it difficult to make a choice. Mr Kelly has offered no analysis of the data checking process that led to the changes, no evidence of a conspiracy, no credible motive for a conspiracy, and no reason to believe the experts are incompetent or dishonest.

A personal note

I spent many years as a hydrologist working on stream and rainfall data, some of it going back to the 19th century. Part of my job was checking old data for inconsistencies, correcting it, and developing better methods to do this.

It was well understood among those of us doing this work that very few long term data records were without significant errors, many of which could be easily identified and corrected. (This was done as accurately as possible, but sometimes it could only be approximate.) Problems included unrecorded changes in location of equipment, misreadings by manual readers (before automatic recorders were in use), faulty transcription into computer records, use of wrong rating curves (conversion charts), etc.

It would therefore be very surprising to me if the same wasn’t true for BoM records. Modern records (say since about 1970) would be more reliable, but older records would be prone to occasional and systematic errors which require correction.


Accusations without evidence are worthless. And conspiracy theories that traduce a wide range of experts without offering any evidence that they have conspired and could have conspired, are both scurrilous and unbelievable.

Mr Kelly’s accusations don’t stand up. They don’t make sense. They show little understanding of data collection and statistics.

We can safely reject them and the ideology that appears to motivate them, and trust that the experts have done their best.

It is notable that Mr Kelly supports his anti-climate change stance by focusing on a few hot temperature readings that have been adjusted, yet he ignores the vast majority of the data that clearly shows that Australia is heating up.

It does seem like there is a conspiracy, but it isn’t where he is looking for it.

Further reading

Main graphic: Bureau of Meteorology ACORN-SAT temperature network of 112 stations

Leave a Reply

Your email address will not be published. Required fields are marked *