Marnhull DCNN8758 – Is Met Office data trustworthy?

The above image is from the Met Office’s X (formerly Twitter) page showing the “Daily Extremes” for the 17th August 2025. This data is produced just after midnight each day and also appears on their main website – https://weather.metoffice.gov.uk/observations/weather-extremes.

From this a very simple question to answer would be – What was the highest recorded temperature in the UK on the above date? Answer…….Yeovilton (a CIMO Class 1 site) at 27.7 °C Yes? Well no it was not. Confused? You probably should be because this is a review of the Marnhull weather station. From data received under Freedom of Information Request* I have the hottest recorded temperature of each day of this year to 10th September. Marnhull recorded 28.3°C on the above date and it is that temperature that what will go into the archives . The public have to take all of this on “trust” because the public actually have no reasonable way of checking any of it.

50.97978 -2.31783 Met Office CIMO Assessed Class 4 (disputed) Temperature readings from 31/8/2019

Marnhull is a manual weather station that reports once daily at 09:00 GMT thus maximum readings are attributed to the day prior. In common with all other manual sites (just over one third of the total of all sites) its readings are not available when the Met office excitedly rushes out its Daily Extremes to such online fanfare rather than wait “until all the figures are in”. Why the Met Office is in such haste to produce such partial figures from automated stations only is probably explained by the likes of Marnhull.

The site itself lies in rural Dorset just north of the village it bears the name of and about 6 miles WSW of Shaftesbury. The Met Office claims this site is CIMO Class 4, whilst I consider this site to be so unreliable that no data should be accepted from it. I will demonstrate why it is so poor and why the Met Office itself has become an unreliable source of recorded weather data.

Firstly to consider the site location, the Met Office has specific requirements for enclosures around the instrumentation to avoid issues such as vandalism, unwanted tampering with readings, distortion by animals and livestock and so on. Below are their general conditions.

On every stated requirement this site fails principally because there is no enclosure whatsoever. This complete lack of an enclosure is very common indeed for Met Office weather stations – everyone a failure. To demonstrate just how compromised this site can be, below is the most recent Google aerial image which I chose not to use above in order to ensure readers would not be confused by the many sheep in the image. It is difficult to tell the livestock from the screen without an indicator in this view. In a chill easterly wind what is there to stop the animals from sheltering by the hedge and clustering around the screen?

Whilst my remark may come across as possibly flippant, take another look at the first site locator image and consider those two metallic trough-like objects to the north of the red kite marked screen. They are almost certainly feed troughs. They are also clearly portable as can be seen with them just a few feet from the screen to the east and south in this differently dated image.

It gets worse, these objects are so reflective that in some dated blurry images they are really all that stands out. Reminiscent of Bainbridge and those ultra reflective former milk tanks alongside the screen.

There is visual evidence of highly reflective material so close to the screen at times as to render any readings irrelevant. Not only will there definitely be livestock compromising readings by clustering around the troughs and the screen to eat, but also those troughs will require filling by humans probably using motorised vehicles as evidenced by the worn trackways emanating from the building to the south. The northern perimeter hedging is now several metres high and completely obscures the site for roadside view AND any wind from that direction. The field location is in fact bordered by dense hedging to all compass points bar a warm southerly.

In the absence of any enclosure in such a working environment there are literally unlimited likely causes to distort readings.

How does the Met office describe this site on its official Weather Observations Website.

The “Reasons for running the site” are given as “Personal”. and there is no official Met Office logo despite this definitely being the one appearing on this list. The last noted readings are for 02:20 on the 29th September 2024 (almost a year ago). Attempting to obtain up to date information results in the following.

That is correct, for whatever reason (though probably lack of input by the site owner) there is “No Data Found“. Obviously though (or possibly not so) there must be data as the Met office has declared daily extremes for this site notably as above on 17th August.

Marnhull is a manual reporting station that has only reported temperature data from 31/8/2019. That the Met Office has chosen to accept data from such a poor site long after adoption of CIMO siting regulations is a complete indictment of its unacceptably poor operating standards. This poor Met Office practise is then made worse by the highly dubious, observations standard.

This site is equipped with a PRT and Data Logger with a LIGT for minimum readings as explained in this report showing the Floors Castle screen. The archived readings for 2024 demonstrate that the screen was only visited 347 days out of the 366 possible. However, of those 347 visits only 298 actually managed to obtain both maximum and minimum temperature readings. {Deriving a daily mean requires both readings.} Below is a typical extract showing some days where only one of either the maximum or minimum were taken and on some days both omitted. Columns I and J are max/min respectively with NA denoting no reading taken.

If 298 out of 366 is rather poor, consider that in 2023 only 269 screen visits were made with a further 31 omitting readings – just 238 (65%) days meteorological means were actually calculable. Referring back to the Facebook video clip https://www.facebook.com/watch/?v=1893457610801591 where the observer demonstrates the reading process, how difficult is it to actually take both readings? If the Marnhull observer for example on 15th December 2024 can manage to record the LIGT minimum of 5.2°C why could that person not read the PRT Data Logger for a maximum? Similarly if the Data Logger could be read on the 5th December why not the LIGT as well? Why were no temperature readings at all taken so many times?

Most importantly if there are such disparities in simplistic observations, how can the public be assured this “Personal Reasons” observer is actually competent enough to take accurate readings all the time?

Despite all the numerous shortcomings of a modern installed and equipped but atrociously sited and with no site control over multiple and blatantly obvious corrupting influences the Met office still accepted the readings of 17th August. Behind that “X” and website portrayed day’s “Extreme” of CIMO Class 1 Yeovilton the reality is that the national historic climate record will be constructed from almost comedic data taken at a complete JUNK site. And that is probably the reason why the Met Office so prematurely produces its daily extremes – it would not want to overly publicise that 96 of the 253 days highs highest temperature readings so far this year were recorded at the likes of manually reporting sites such as Marnhull.

Codicil

*The Met office rather foolishly refuses to respond to any Freedom of Information Requests that “I” make on the grounds that “I” am declared “vexatious“. However under EIR regulations there is no requirement for personal identification. “Others” may ask of course……….

Source: https://tallbloke.wordpress.com/2025/09/18/marnhull-dcnn8758-is-met-office-data-trustworthy/