Stevenson Screens – ” The aesthetic and moral codes of the suburban domestic garden.” An explanatory report.

Above image: What a Met Office Stevennson Screen used to look like…….but doesn’t anymore.

This report is intended to clarify the many points I make regarding the siting of temperature sensors by the Met Office. Whilst many of the public will naturally assume the Met Office operates in a high quality and accurate manner, on detailed inspection it becomes quite apparent that their observational standard is often particularly poor and indeed, in most respects, quite antiquated. The headline image of the interior of a Stevenson Screen is an example of what is NO LONGER used and it is the modern variant that is the cause of much discrepancy and debate.

First of all, a game of “spot the difference” with this modern “manually” observed weather station screen below would take a long time, so many are the fundamental differences.

Leaving aside alterations to the level of ventilation via louvred walls, holes in the base, construction material (plastic not wood) and even different internal colour, the most striking difference is the completely different temperature measuring devices and the computerised “Data Logger”. 21st century electronics incongruously housed within a fundamentally 19th Century box. So how did this all begin?

A good description of the history of weather station screens can be found in this 2018 research paper from which the somewhat pithy headline quote is taken.

https://royalsocietypublishing.org/rsnr/article/73/2/203/48694/Thermometer-screens-and-the-geographies-of

{n.b. the above spelling of “Strathfield Turgiss” is incorrect and should be “Stratfield Turgiss“}

Prior to the universal adoption of the Stevenson Screen around 1884, there were a range of different means to actually measure daily temperatures. Some may seem almost bizarre by modern standards such as actually measuring outside air temperature by placing thermometers indoors in unheated outbuildings or in metal boxes fixed externally to north facing walls as at Armagh. As the above research paper details, a standardised type of screen was needed and thus a “competition“was organised by the Rev. Charles Higman Griffith, Rector of Stratfield Turgis, which studied 42 different designs of rain gauges and 10 different thermometer screens in his large back garden.

The Stevenson design (with modifications by Edward Mawley) was adopted purely on the basis it was adjudged the best operating system for the Rector’s back garden – neither a natural location nor one representative of the larger surrounding area. The least worst option available. The instruments inside for temperature readings were Liquid in Glass Thermometers ( then usually filled with Mercury) which were originally invented in the latter 16th Century and only minimally modified over the subsequent years.

The Stevenson Screen, even on adoption, was identified to have a major design flaw identified by Scottish meteorologist, John Aitken after whom the “Aitken Effect” was named. The problem was twofold, in low wind speed the entrapped air could stagnate and potentially overheat in strong sunshine leading to occasional over recording of highs. This aspect is perfectly exemplified by the variations between back garden Astwood Bank and nearby open field Wellesbourne. Similarly in low overnight wind speed (a more common phenomenon due to cooler conditions) the stagnant air could be insulated by the structure and not drop as low as the outside air. This latter effect is clearly noted by the observer at the Floors Castle Met Office station who notes in this video clip from Facebook how the apparent temperature immediately drops on opening the screen door and letting a bit of “outside air” in.

From Met Office Factsheet 17

The above problems with the Screen were thus well known to meteorology but apparently adjudged to not be too serious to warrant change and would, in theory, “even out” on universal adoption. Regarding highs, the LIGT were slow to react to sudden changes in temperature (e.g. intense sunshine appearing from behind clouds) so the over-recording was deemed much rarer. Under recording of night time lows was recently identified to be in the region of 12% (almost 1 in 8) of minimums but even this frequency was hand waived away as not significant and “spun around” into claiming the virtues of Victorian technology despite such notable problems . https://research.reading.ac.uk/research-blog/2024/09/02/victorian-weather-tech-remarkably-accurate/

It has been argued that providing this flawed system continued in use, historical readings would not be materially affected and any variations in climate would be accurately represented. However, several other issues most certainly have changed and it is these significant changes that are of major concern and the driver of my research for the Surface Stations Project.

So, for perspective, consider this image:

This is the world’s first internal combustion engine motor car which hit the roads in the same year that Stevenson screens were universally adopted. Nobody would voluntarily use this type of vehicle as their daily city runaround in 2026 would they? Consider any aspect of modern life, technology, medicine, indeed anything and virtually nobody would accept the standards of the late 19th century – antibiotics were first developed in 1928 and not many would prefer life without those. So why do meteorologists and climatologists seem to feel their area of “expertise” should be constrained by such antiquated and grossly inadequate systems? As ever, they do and they don’t!

So what has changed? – Perhaps the most significant change is how temperature readings are actually taken. Starting around the late 1970s automation of readings came into being. In order to take readings in a form that could be electronically transmissible, the thermometers were changed from LIGT to Electronic Resistance Thermometers (ERT). These latter thermometers measured the resistance in a circuit though a wire probe that varied with temperature rather than the older LIGT system dependent on the expansion and contraction of a liquid medium. These eventually were standardised to a particular length and thickness of a platinum wire probe and became known as PRTs.

The performance characteristics of PRTs and LIGTs are markedly different. Whilst LIGTs are slow to respond, PRTs respond much more quickly to the extent that instant readouts are taken electronically every 15 seconds with 4 consecutive readouts per minute averaged to take 1440 individual readings per day stored in the Data Logger.

Response times (i.e. how quickly thermometers take to adjust to temperature changes) are calibrated by time taken to both 63% (T63) or 90% (T90) of the final temperature. This assumes (in air) a wind velocity of 3 metres per second/10.8 kmh/6.7mph which equates to Beaufort Scale 2.  The T63 response is considered the most important figure.

The time taken for Met Office PRTs at 3 m/s air speed for T63 is in the region of 20 seconds. Conversely at the same wind speed the LIGT T63 is up to 80 seconds. However, this considerable difference widens dramatically as the wind speed ventilation WITHIN the screen drops. In still air the PRT T63 response time is typically 70 seconds (just over the 1 minute sampling period) whilst the LIGT drops down to the region of 280 seconds (just 20 seconds short of 5 minutes. )

It becomes very clear that ventilation within the Stevenson Screen is a critical factor and that changing technologies from the slow responding LIGT to the much more responsive PRT will result in differing readings. For example, the brief burst of exhaust from an internal combustion engine leaf blower alongside a screen in a walled garden (such as Lysdinam) forcing warm air into the screen will be detected by a PRT but will probably have largely dissipated before a former LIGT would have fully responded. It must be borne in mind that meteorological averaging simply operates on maximum + minimum temperatures ÷ 2 and each reading can be as brief as just 1 minute duration with the new PRT.

On top of this change to instrumentation comes the multiple and ongoing changes to the screen locations both reducing wind speeds (leading to overheating and insulation effects) as well as multiple novel extraneous heat sources. Very many of my reports refer to wind shading effects with classic examples being Pershore and Astwood Bank. Dr Eric Huxter has been very extensively researching the PRT issue and noted the extreme “spikiness” exhibited at such sites for these very reasons. His excellent blog notably recently featured Astwood Bank and Pershore and his research is a must view in this respect.

The issue of Urbanisation is well publicised regarding weather stations and it is easy to prove how surrounding former rural sites with building development, roads and modern infrastructure will both elevate nighttime minimums (the main contributor to increasing daily means) and impose “spikiness” into daytime highs. Lesser known, but equally important and a key feature of my reports are rural wind shielding effectively lowering internal screen wind speeds and are not only extremely common but often seem to be almost deliberately manufactured such as at Liscombe. Screen relocations regularly appear to significantly worsen exposure to almost surreal extents such as Bingley no 2.

Unquestionably problematic aviation issues are ignored on the absurd metrics of “internationally agreed distance” from the runway regardless of all the changes from propeller driven aircraft when originally installed to massively more powerful modern jet blasts or complete conversion of site use to anything from mass car parking to solar farms. Again, as an example, should this site below be considered a perfectly natural situation alongside which to continue operating a weather station? This use would never have occurred even until the very latter 20th century. The UK Met Office considers this as one of their best (Class 2) locations. Best for exactly what purpose is surely a valid question – “Best” to prove anthropogenic Global Warming? Well it certainly is anthropogenic but no CO2 is required!

In the United States the government established a network of “Climate Reference” stations of the very highest standards to overcome such problems and ensure an accurate climate record could be constructed using only best practices and quality instrumentation. Given the extremely lavish budget enjoyed by the Met Office surely it would be appropriate for a similar system to be established in the UK?

The ultimate condemnation of the Met office comes down to the simple fact that many of the well known problems with their seemingly unimpeachable low grade Victorian technology can be ameliorated by adding artificial fanned ventilation to their Stevenson Screens ensuring adequate and uniform wind speed within. The World Meteorological Organisation (WMO) recommends all “climate reporting” temperature sensors should be in artificially ventilated screens – the Met Office climate network has, precisely, NONE. The WMO considers 60 second PRT averaging from artificially ventilated sensors as acceptable, but recommends up to 5 minute averaging from naturally ventilated screens to overcome the comparative differences in response times I highlighted above. The Met Office seems to consider themselves above such guidelines. Is this an example of arrogant British exceptionalism over scientific credibility or a requirement to provide “evidence” (not real data) to “prove” a theory.

This official Met office climate reporting station below has temperature readings starting 31/8/2019 – less than 7 years ago!

Clearly this inappropriate and inaccurate corrupted data collection must stop – now.

Source: https://tallbloke.wordpress.com/2026/04/26/stevenson-screens-the-aesthetic-and-moral-codes-of-the-suburban-domestic-garden-an-explanatory-report/