Stevenson screen for air temperature measurement.

Fact check: have US and Australian hottest day records been deleted?

Craig Kelly has accused Australian “fraudsters” of deleting temperature records. Is there any basis to this claim?


On 12 September, Craig Kelly claimed that Australian hottest day records had been deleted by “fraudsters”. He posted a video claiming a similar threat to maximum temperature records in the US, set in 1913.


There is no reasonable basis for either claim. But even if true they would be irrelevant.

Read on to find out why.

The facts

A temperature of 56.7°C was recorded in Death Valley, California in July 1913. That record has not been deleted or changed. For example, the World Meteorological Organization lists it as the hottest ever recorded. And news reports of recent hot days also reference the 1913 day as the hottest.

The Tony Heller video Mr Kelly links to is titled Erasing The Record Heat Of 1913, but the strongest claim I could find in the video is that: “climate alarmists are working very hard to try to erase the world’s record temperature of 1913”.

Mr Kelly’s claim Australian high temperature records had been deleted is not specific, but presumably refers to concerns that the Bureau of Meteorology (BoM) “homogenises” records to correct apparent errors (see this report by a climate sceptic and this report supporting BoM).

BoM has clarified that the original records haven’t been lost or destroyed. However datasets used in climate models and elsewhere are adjusted, using “world’s best practice and a peer reviewed process” where there is good reason to think they are inaccurate.

Interpreting the facts

It is well known that many errors and inconsistencies can occur in climate data, especially years ago when recording was done manually. Possible errors include:

  • simply misreading the equipment or wrongly writing the reading down;
  • reading at a different time of day – either by mistake, or because the standard time changed at some date;
  • in early days non standard measuring equipment was sometimes used, giving an inconsistent result;
  • changing the measuring equipment;
  • moving the measuring site; or
  • conditions around the measurement site changing.

To undertake accurate analysis, homogeneous records are required. So data analysts are forced to try to remove inconsistencies and errors, using statistical and practical methods. There is nothing subversive about this – if this wasn’t done, the analysis would be compromised.

And analysis shows that, overall, the modifications made to temperature data have made global warming appears slightly less, although some data adjustment has had the opposite effect.

Understanding the US data

The Tony Heller video makes the unsubstantiated accusations that any proposed amendment of the 1913 record would be unjustified and that it would be done to satisfy “climate alarmists”.

He doesn’t actually know if the record will be changed. If it IS changed, he doesn’t know in what way. Nor could he know what the technical justification might be until a report is published.

Nor will it have any impact on our understanding of global warming.

Why Tony Heller’s claim is irrelevant

Global warming calculations and presentations are based on global temperatures. The total area of USA is less than 2% of the earth’s surface. The records in question are for a week in the whole of 1913 (2% of the year).

Even if every US temperature record for 1913 was altered, it would make an insignificant difference to the overall calculation of global annual temperature.

It is worth noting that despite these record high temperatures for a week in July, 1913 wasn’t a very hot year overall, as the NASA map shows.

Understanding the Australian records

BoM’s processes for checking data for consistency are quite clear (see Temperature data methodologies at the Bureau of Meteorology). Their methods have been peer reviewed by international experts and published in widely available journals.

Those critical of BoM haven’t (to our knowledge) published their analyses and subjected them to the same level of peer review. Until they do, they cannot be taken seriously.

The BoM data is the best available.


It is easy to make a claim of a conspiracy to change data to suit an agenda. But the experts are well qualified and have documented their methods.

Conspiracy theorists say we cannot trust the experts. But the sceptics expect us to trust them, even though they offer no evidence.

Soon this site will have a page on how to assess which “experts” to trust and how to assess the climate claims.


No temperature data has actually been deleted, in US or Australia.

There is no evidence of any conspiracy to delete climate records.

Analysis show that, overall, adjustments have slightly reduced the degree of global warming.

The conspiracy claims are based on misunderstandings of data errors and consistency checking and unsubstantiated suspicions of what might happen.

Photo: Stevenson screen, used to measure maximum and minimum air temperature (Wikipedia).