Sutton Bonington Part 3 – Mysterious relocations, more mashed up data and trying to interpret errors.

This third part of my investigation into the spurious data from the Met Office “Historic” station covers the odd mystery of its original location, supplies further evidence of huge discrepancies in temperature readings and will lead into subsequent reports of other sites demonstrating how temperature history can be rewritten very easily indeed. Analysis of the data has been kindly supplied by a reader who is Senior Science Research Director (professor equivalent) at a leading European university.

Firstly regarding the original location, this is a cautionary tale against simply trusting something that was written before. The headline image is of the very first manual observations record for Sutton Bonington from 1924. Helpfully it gives the site coordinates of Latitude 53°45` N, 1°14` W. Unfortunately these are wrong and would put the site some 61 miles to the north- northeast of the present day location. Clearly these coordinates were wrong but unfortunately from 1924 to 1957 every subsequent report showed these incorrect figures. Only in 1958 were the correct coordinates of 52°50`, 1°15 ` eventually used.

These new coordinates were still not sufficient to pin down the exact location so I attempted the system I used at Bala from the “Objects used for estimates of visibility” – unfortunately again these were not the best of indicators. The “4th tel pole on Farm Road” is not that useful 100 years later!

Eventually after a lot of drawing imaginary circles on google maps I finally spotted the original location about 650 metres from its modern site in what is now a highly developed part of the college which no doubt enforced the relocation. It appears this rather compromised site was the one covered by the conversion period from manual to automatic reporting. The location and site renumbering to a recognised different climatology from DCNN4201 to 4202.

Looking at the comparative figures for February and March comparing the two maximums and minimums and the hybrid monthly means

February.

There are 12 instances of manual to automatic differences of more than the possible instrument deviations of 0.4°C ranging up to as high as 3.6°C. None of the potential derived means equate particularly close to the published “historic data” of 9.7°C maximum and 2.7°C minimum which disproved my suggestion of a hybrid mean being used.

March

Again another 12 readings showing very significant variations ranging up to 4.6°C difference. The monthly means of any combination do not correspond to those claimed of 11.5°C for maximums and 3.2°C for minimums. However any of these authority figures below are derived, they are clearly not from the prime readings.

Questions worth answering are how have these discrepancies occurred in the first place? A highly qualified reader in this subject area offered his interpretations based on experience and went on to tabulate them for me.

“Hi Ray,

I’ve finally had a little time to look at the Sutton Bonnington data.
Brief description of my deductions.

December: One missing set of TX and TN from the AWS. Three questionable readings for the DLY TNs – 6th, 17th, 18th.

January: AWS complete. 17 missing DLY readings. Inspection suggests observer turned up on weekdays. By correlation with the AWS, 4 TXs can be placed correctly (red boxes recorded date, red values probable date: 4th->2nd; 9th->7th; 20th->12th; 23rd->21st). 2 TNs can be placed (blue boxes recorded date, blue values probable date: 21st->16th; 31st->29th(Q)). One questionable TN (29th), one TX not reset (black box – 6th – not reset on 7th).

February: The observer at least turned up Monday to Friday of each week. 11 missing days. By correlation with the AWS, 3 TXs can be placed correctly (red boxes recorded date, red values probable date). 3 TNs can be placed (blue boxes recorded date, blue values probable date). 3 questionable TNs (2nd, 3rd, 18th), one TX not reset (black box – 2nd – not reset on 3rd). Possible scenario: The observer forgot not only to reset the TX but also didn’t write down the TNs, only to realise two days later and then swapped the values (I’m guessing here from having done it myself once or twice).

March: The observer turned up Monday to Friday of each week. 9 missing days. By correlation with the AWS, 1 TX can be placed correctly (red box recorded date, red value probable date). 4 TNs can be placed (blue boxes recorded date, blue values probable date). Three TX and one TN not reset (black box – 5th, 14th, 15th and 16th).

Conclusion: From the beginning of January the observer only turns up Monday to Friday, and sometimes not even that. So the TN and TX read on Monday mornings could be any day from Friday pm to Monday am. 8 TX and 9 TN can be placed correctly (a reasonable best guess given the average difference). Questionable readings – did the observer forget to write them down or was the observation made later than 09h? There were 6 mornings when the TX (or TN) was not reset.
If the questionable readings and the ‘not reset’ values are removed, the AWS-DLY difference for the TN is -0.15°C and +0.05°C for the TX.

Bottom line: The observer did not do their job properly and zero quality control by the Met Office.“

Further details followed:

Hi Ray, 

Two pairs of eyes are better than one! If in the future you want me to have a look at this sort of data before you post, I would be happy to help.

Had a second look this afternoon. I have deliberately marked the AWS values for the occurrences of 21h TNs and 09h TXs. In light blue and light orange.

I found a couple of errors on my part (a too narrow excel column rounds the 10.7 to 11!).

So yes, you can quote my analysis but please use this second excel file rather than the first one.

Additional remarks:

The questionable DLY 09h TN readings for Dec 6, Dec 17, Feb 18, Feb 23 coincide with the AWS TN from the 21h reading. I have no explanation for why they are higher!

The apparent non reset TX of Jan 6th can be explained if the TX occurred at or very close to 09h. The reading would then count for both days.

The same logic applies to the non reset TN at 09h on March 16th. The TN probably occurred close to 09h. So this would be an example of a doubly recorded minimum.

It is a shame that the Met Office no longer carries out the quality checks as they used to before 1959. Correspondence with the observer might have enabled some of the values to be corrected. See for example Rothamsted June 1952, the MO received a reply two days later enabling the the TN on June 18th to be corrected from 45°F to 55°F.” {n.b. the Rothamsted data is the subject of separate reports}

I quote all these details in full to emphasize that the Talkshop and the Surface Stations Project are not some online journalistic grifts simply looking for sensationalism to make a quick buck. These are unfunded projects, not carrying advertisements, that actually cost the authors to produce for no gain other than genuine scientific research and to produce accurate data. Our reader went on to supply this analysis by date/reading.

I am very indebted for this level of assistance that has demonstrated the complete absence of any quality control in modern Met Office practices. Readings which were completely incorrect are there in the archives seemingly forever. As I pointed out in my 500th post numerous readings, even incorrect National record highs, are sitting in the “records books.” and are never amended even if shown to be wrong. It seems somewhat preposterous that outside observers can identify errors and their causes whilst the official authority cannot.

Many of the highlighted problems with the manual observations would have gone unseen were it not for the overlap comparison period that I was able to study. Without this prior comparison there is no way of knowing how many readings were wrong in the absence of any ongoing quality control scrutiny.

This problem of non existent quality control is now seen to have become much worse. Automatic readings seem to be taken as always fully accurate when they too should be equally subject to individual scrutiny. Had it not been for the diligence of a separate reader, completely bogus readings (broadcast all over the Met Office website and X social media) would have gone unnoticed. A concerned member of the general public is now apparently the only check there is! If the ridiculous and obviously wrong readings at Giants Causeway seem a worry, consider this. A few days later Giant’s causeway came back online…..the readings were AGAIN wrong and similarly broadcast on the internet! The same reader again raised a query.

And the Met Office response to someone helpfully doing their own quality control for them and safeguarding against yet more embarrassingly wrong readings being so highly profiled was hardly reassuring – the Met Office had again not noticed yet more incorrect readings.

“Good afternoon, 

Our support teams have advised that we are aware of ongoing issues with this site and our field engineers are looking into fixing the issues. 

There is no time frame at this stage for a fix. 

Kind regards,

Richard

Service Desk“

Ultimately this three part review has been to demonstrate that in order to create a climate history for the UK, every single reading has to be carefully authenticated. Whilst there are poor sites, deteriorated sites, relocations, extraneous heat sources, poor observing practices (manual and automatic), instrument changes and archiving protocol changes, ……there is also the lack of adequate quality control allowing hugely incorrect readings to corrupt data files.

Sutton Bonington’s data is portrayed by the Met Office as from one continuous site when it is proven to be from at least two different locations of recognised different climatology. Multiple dubious and plain wrong readings are proven to be on file with no assurances that the other readings may be any more accurate particularly in the more modern era. Numbers proudly shown as “Historic Station Data” for 4 months appear to be contrived from nothing more than thin air. Despite all this the general public is expected to trust the Met Office, believe all their extravagant prognostications for the future and make major alterations and adjustments to their lifestyles.

I guess the best codicil is to consider this from a Met Office Media Advisor. Seems nothing ever changes!

Met Office Caught Fabricating Daily Temperature Record Claims

Source: https://talkshop.blog/2026/09/29/sutton-bonington-part-3-mysterious-relocations-more-mashed-up-data-and-trying-to-interpret-errors/