Devon – The historic changes in weather station sites over the last 50 years – the reverse of the national trend.

My recent reviews of historic changes in weather station sites over time have variously referenced geographic moves in Dumfries and Galloway, London, South West Wales and Derbyshire areas. I further went on to highlight how this type of geographic change alters the temperature readings recorded and ultimately derived averages specifically for the county of Kent.

Historic weather observations ALWAYS come from weather stations – they NEVER come from computer models or any form of simulations whatever anyone may claim. However, the change in sample points used (i.e. weather stations) over time and geography can, and most certainly does, alter any subsequent arithmetical conjuring tricks made from them. This post reviews Devon in similar style to that I produced of Kent but reveals a dramatically different history.

Devon is now home to the Met Office following its move from Bracknell in 2003/4. I have commented before that there seems to be a disproportionately large number of weather stations in the county but never fully analysed the situation. When examining my own home county of Kent I noted a dramatic drop in the numbers of sites over the last 50 years from 15 in 1976 to just 7 today. 10 stations had closed notably in cooler locations with one a replacement in a warmer site and the one genuinely new site (Frittenden) is now a record breaking warmer recording one. Simply modifying the sample sites used for an “apples with apples” comparison allowed me to “change” the climate of the past by simply gaming the system.

Similarly in London and Dumfries and Galloway it becomes increasingly apparent that sample point changes can have dramatic effects in derived “means”. Disregard “weightings”, “adjustments”, “homogenisation” or any other black box devices, simply obtaining the primary data from different locations will achieve almost any desired end result.

So to Devon in 1976 and that particularly pleasant summer. Again as in my Kent report I logged those significantly warm date’s temperatures from 23rd June to 8th July. What stations were operational in 1976?………..just 12. Rather than print out all the figures I will simply give the derived averages – these can be checked from the CEDA archives.

The sites operational in 1976 that are still operational today are marked in Red on the headline map

Those operational in 1976 that have since closed are marked in Blue.

Currently Still Operational Sites.

Site………………………………………..Mean Period Maximum

Exeter Airport …………………………29.3°C {Site now superceded for climate reporting by the Number 2 site} (Automatic from 2009 )

Teignmouth.……………………………25.6°C (Manual)

Plymouth Mountbatten.…………….25.3°C (Automatic from 2000)

Slapton.………………………………….24.6°C (Manual)

{n.b. Both Plymouth Hoe and Sidmouth were operational before and after 1976 but were not recording temperatures throughout the period and both subsequently closed.}

Operational 1976 Subsequently Closed

Tavistock………………………………28.5°C. (Manual – replacement site closed 2002)

Cheldon Barton……………………..28.4°C (Manual – closed 1978)

Exmouth……………………………….26.9°C (Manual – closed 1990)

Torquay…………………………………26.3°C (Manual – closed 2005)

Starcross……………………………….25.2°C (Manual – closed 1990)

Princetown……………………………24.5°C (Manual – closed 2001)

North Hessary Tor………………….24.5°C (Manual – closed 2005)

Hartland Point……………………….21.9°C (Manual closed – 1988)

{Prawle Point was also shown to be recording temperatures but maximum recordings went not recorded for the whole year uo tp closure that year}

In similar style to Kent with only 5 stations still open 50 years later, only 4 of the 12 Devon sites remained operational in 2026. However, there the similarity ends. Whilst nationally weather station numbers reduced from over 630 to well under 400 (an almost 40% reduction), Devon hugely bucked the trend with all the following additions marked Yellow on the headline map.

Dunkeswell Although this site predates 1976 there are no archived records either digital or manual available and it was in a markedly different location anyway. Now prone to “sileage heating” and artificial windbreaks. A seriously compromised aviation location all round. (automantic reporting)

Chivenor Again predating 1976 but only visible manual records run from 1943 to 1955. Digital records only start from 1984. Again a highly compromised aviation site close to taxiways and an artificial surface sports pitch. (Automatic)

Hemyock. Long term amateur site in Volunteer Climate Network (VCN) that unfortunately deteriorated over time through no fault of the owner. Opened 1978. (Manual)

North Wyke Excellent site installed and operated by Rothamsted Research from 1981 – if only all sites were as good as theirs. A site I propose to use for recent climate reconstruction for the west country. (Automatic from 2000)

Okehampton VCN CIMO Class 4 poor quality farm site by drystone walls – opened in 1992. (Automatic)

Holne VCN Class 4 poor quality farm site with temperature records from 2006 relocated 2019. (Automatic)

Huntsham VCN Class 5 very poor quality farm site from 2006 (Manual)

Lapford VCN Class 4 farm site that really should be Class 2. Good quality with reliable data from 2009. (Manual)

Virginstow VCN Class 5 very low quality back garden site. From 2015 (Manual)

Bratton Fleming VCN Class 4 low quality farm site from 2021. (Manual)

Hollacombe VCN Class 5 very poor farm site from 2024. (Manual)

Stoodleigh VCN Class 4 low quality farm site from 2024. (Manual)

From a 1976 total of 12 fully operational temperature recording weather stations in Devon, 50 years later this area has shown a third increment to 16 weather stations recording temperature in 2026. Unfortunately,There are few indications of attempting to improve the quality of station siting with North Wyke and Lapford (despite the Met Office’s peculiar assessment) being the only Class 2 standard new additions. Many of these new sites are Class 5 meteorological Junk.

I should point out though, that there never really was any halcyon days of predominently good climate reporting weather stations – a large number in the past were equally not very good. This site below ran from July 1977 to December 2008. It seems reading the temperature surrounded by woodland is a time honoured tradition, such as at High Beach, that never changes. Surely it is incumbent upon a largely tax payer funded government agency to improve on its estate rather than claim it was always not that good so, seemingly, adopt a “why bother” attitude. This problem was addressed in the US and should be in the UK if the Met office is to claim such “world leading” standards.

{N.B. the above Bovey Tracey site supplies “60 year climate averages” for its area – more later}

This bucking of the national site reducting numbers trend requires an explanation. The Met Office has claimed that automation resulted in the closure of many sites but the figures above do NOT substantiate that position. There are almost as many manual sites operating in 2026 in Devon as there were in 1976 with the most recent 6 all manually observed. Thus the automation argument does not stand up to scrutiny. It is also incredibly strange to have any new manual sites as modern day automation is very easy indeed with the advent of solar panel/battery power packs, PRT Data loggers (already installed at all sites since 2017 at the latest) and internet access now universally available. The “volunteers” can effectively still take observations across the full range without interfereing with their own interests.

I have gone to such lengths in supplying all this data to make sense of the headline map. Looking at the colour coding there is demonstrated (as in London, Dumfries and Galloway and Kent) a readily visible change in locations. The temperature sample points have all subtly moved: –

Hartland Point (at the time the only site on the North Devon coast) simply had to go for recording so low despite it being a significant forecasting station.

Princetown (a.k.a. Dartmoor Prison) and North Hessary are both typical high elevation moorland sites – can’t be recording those even though such upland accounts for about 20% of the county by area. The Met Office acknowledges the huge micro climate differences between Teignmouth and Princetown in one of its own “Factsheets” so they kept Teignmouth and ditched the moorland sites.

Prawle Point regularly previously recording fog and misty cold never stood a chance of retention.

Now notice how all the new (“yellow sites”) have migrated into sheltered low grade spots in back gardens or farm locations away from coastal moderating winds in summer but kept cosy and warm in winter by their niche locations such as this absurdity behind the hedges, alongside the vegetable plot, in front of the solar panels, polytunnel and greenouse and to the side wall of the house at Virginstow. Perfectly representative of its own garden and nowhere else at all.

Or perhaps in front of south facing walling, the vegetable plot and sheltered from any slightest breeze from every direction. Is the internationally recognised high standard of Class 5 what the Met office is aiming to achieve in its sophisticated selection process like Huntsham below?

Surely it is obvious that the two sites above are much more likely to produce the desired uplifts in temperature tthan this one below with the nearest tree over 150 metres away. “Micro not Macro” seems to be Met Office Mantra.

In conclusion, I am confident that whatever may, or may not, be changing with the UK climate over the last 150 years, it is certainly not going to be accurately recorded by such contrived changes over time to the sample points used by the Met Office. A completely independent historic national temperature record from only selected good quality and unchange sites is required – not gamed computer models.

Talking of computer models – remember that image of the Bovey Tracey station surrounded by the woodlands from 1978 to 2008? That is taken as representative of its region for a 60 year period when it barely existed for half that time……in a wood.

And also consider that Class 5 terrible weather station by the amusment arcade in front of the pier at Teignmouth is similarly deemed wide area representatrive climate averages site. Somehow I find this rather difficult to believe given the Met Office themselve publish evidence to the contrary.

My next post will move into the almost surreal manipulations that pass for the climate record of East Anglia.

Source: https://talkshop.blog/2026/10/20/devon-the-historic-changes-in-weather-station-sites-over-the-last-50-years-the-reverse-of-the-national-trend/