Hemyock DCNN8895 – A further study into area sample bias. This surely cannot be “accidental”.

“50.908940 -3.227068 Met office CIMO Assessed Class 4 Temperature Records from 1/8/1978

Hemyock is a small village approximately 12km/7.5 miles north of Honiton in east Devon. This site may now be recently closed as there is some uncertainty as to the well being of the owner of this amateur site that has been a long term adopted weather station by the Met Office. Despite being in the owners rear garden it is unusually situated in its own enclosure and has clearly been well maintained and observed over a long period by them. This report highlights the unfortunate shortcomings of such amateur sites and also, more importsntly highlights yet again highly questionable location practises by the Met Office.

At first glance this looks a reasonable quality site from a “flat” aerial image, however, that is just a facet of the imagery as this site is anything but level, it is after all in the Blackdown Hills. The streetview image below is from 2009 taken from the hill adjacent to the site. The rather “Gallic” style of Stevenson Scrren has a pitched roof and is clearly long time served. The anemometer will be for the personal use of the owner as it does not meet Met Office requirements. The original LIGT was replaced by a PRT in 2017 but the site remains a manually observed site.

I have no wish to criticise the owner of this site who is clearly doing a very good job here in restricting outside interference to his equipment by fencing it in. My main concern is the proximity to the road and what that causes. The twig-like vegetation to the bottom of this image is actually the top of a recently flailed roadside hedge which is potentially outside the owners control. In a similar vein to the problems at Pershore , Bishopton and many others, the screen can be compromised by wind shading should the surrounding area vegetation become uncontrolled.

The same point image of 2021 shows again a recently flailed hedge that is now much taller and the screen is no longer visible from any angle. This wind break is to the south west of the screen – the exact direction of the prevailing wind and inevitably will cause low wind speed in the screen that would not have occurred in its earlier days. Aitkent effect will lead to increased events of still night stagnant air being insulated within the screen thus internal temperatures not falling as low as ambient. Overheating in strong insolation will also potentially exaggerate occasional summer highs. All of this is a great shame as the owner otherwise puts a lot of effort into his hobby that has been recognised locally..

Where this Hemyock station becomes more interesting is that in common with all other Met Office sites, its reading will most definitely contribute to the historic temperature record. This is a climate reporting site rather than a major synoptic one but its readings will be used alongside other stations to “fill in” blanks in observations at other stations and be added to the “homogenisation pot” for the area. Hemyock is just 5.4 km/ 3.3 miles from RAF Dunkeswell which the Met Office Location Specific Long Term averages site quotes as its nearest “Climate Averages Station.”

I recently posted details of the intriguing distribution of weather stations in London with its remarkably high concentration of sites particularly in the known hotspot area west of the capital. The whole subject of weather station distribution is one rarely discussed so I feel I should further put this into context – in the same way that there are 14 weather stations within a 40 km radius of central London there are quite inexplicably 14 within 40 km of Hemyock and that really is very telling when studied in greater depth. Bear in mind there are large swathes of the UK that have very few or even none at all in many areas of that size. More surprising is that while numbers of sites are reducing in nearly all areas, a “selected” few areas are increasing in site numbers.

Context is provided here – in 1994 there were around 635 weather stations in the UK recording temperatures. Even at 2000 there were still 600 but over the following few years the numbers massively reduced to under 400 by 2012.. From peak to trough a reduction of almost 40% – there are fewer weather stations now than in 1953, 73 years ago. How many people would thought that likely?

The fall in numbers was partly due to automation of sites requiring power supplies and data-comms – or so the official excuse goes – and this was often not possible in rural areas. This resulted in mostly these rural sites meeting their demise with more urban (warmer?) sites remaining. This wholly unnatural selection process was well demonstrated in my Dumfries and Galloway report demonstrating this drift from rural to coastal and urban sites.

So how does this stack up with Hemyock? Devon is yet another of those areas shown to have a much higher current site concentration than elsewhere.

Within a 40km radius circle of Hemyock there are

Teignmouth Manual installed 1903 (Class 5)

RNAS Yeovilton Automatic installed 1964 (Class 1)

Hemyock Manual installed 1978 (Class 4)

RAF Dunkeswell Automatic installed 1979 (Class 3)

Nettlecombe Manual Installed 1968 (Class 5)

Liscombe Automatic installed 1993 Class 4)

Westonzoyland Manual installed 2007 (Class 5)

Huntsham Manual Installed 2008 (Class 5)

Seavington Manual installed 2008 (Class 5)

Lapford Manual Installed 2009 – yet to be reviewed. (Class 4)

Exeter Airport no 2 Automatic installed 2009 (Class 4)

Broadwindsor Manual installed 2011 (Class 4)

RNAS Merryfield Automatic Installed 2021 (Class 4)

Stoodleigh:Dryhill Manual Installed 2024 – yet to be reviewed. (Class 5)

14 weather stations with no fewer than 8 of them installed in the 21st century. Of those installed in the 21st century half are Class 4 {additional estimated uncertainty added by siting up to 2 °C)} and half are Class 5 Class 5 (additional estimated uncertainty added by siting up to 5 °C). Also note that 6 of them are manually observed – so what was the real imperative to close down so many rural sites if so many new manuals are being installed?

As shown above, however, since 1994 there has been a 40% decrease in total sites so how many have been closed down in this same 40km radius? Just 5 at Withleigh 2005, Sidmouth 2017, Bovey Tracy 2008, Exmouth 2005 and Cannington 2005. Thus this area has shown a net increase of 27% in numbers of stations simultaneous to a national 40% reduction – a remarkable bucking of the trend.

Consider the average number of stations used by the Met Office to “formulate” data for their km square grids and how dramatically these have changed as shown below up to 2010 data from the same RMetS report. From 1990 to 2020 the number will have reduced further still, the 2000 to 2030 averages (when produced) will be from under 400 “carefully selected” sites.

Table 2. Number of stations used to generate long-term average grids

Climate variable	Average number of station values per monthly long-term average grid (1981–2010)	Average number of station values per monthly grid over period 1981–2010
Air temperature	1,203	545
Rainfall	9,547	3,857
Sunshine	611	226

When Grok AI was asked to put the number of weather stations into a historic table it supplied this.

Thus taking all these reductions into account, why is it that some areas (i.e. west London, east Devon and others) are showing an increasing trend when others such as Fife are demonstrating huge declines? My view is that this is not accidental and furthermore this is a deliberate modification to sample points to increase warming trends. This may be a controversial comment but how else can anyone explain opening new sites in modern history within walled gardens and known micro-climates such as Kingston Maurward in 2019 or Bisterne in 2025?

Is this all by accident or design?

Source: https://tallbloke.wordpress.com/2026/07/24/hemyock-dcnn8895-a-further-study-into-area-sample-bias-this-surely-cannot-be-accidental/