Arran:Brodick Castle DCNN 6133 – A case study of multiple Scottish issues.

55.59212 -5.15423 Met Office CIMO Assessed Class 5 Installed 1/1/1965 (but more likely 1859)

Arran is the largest island in the Firth of Clyde and overall the 7th largest Scottish island. It enjoys a cool but mild climate with the east coast at Brodick having the benefit of shading from the prevailing westerly winds. Typical of western isles it is a largely frost free but high rainfall area. Brodick Castle itself is now administered by the National Trust for Scotland. The weather station is typical of very many Scottish sites and highlights problems they represent in both meteorology and climatology terms. This review covers issues that, whilst not exclusive to Scotland, heavily dominate the recording system there.

Walled Gardens These are common throughout the UK but have a greater prevalence in Scotland – a point that is noted in the Wikipedia article. The deliberate purpose of a walled garden is to create a modified micro climate with increased temperature and the moderated detrimental effects to crop growing of both wind and excess water (latent heat). The temperature modifying effects can be very notable with up to a 5°C differential above the natural ambient environment. Frosts are almost entirely eliminated in western Scottish sites.

Walled gardens were a feature of wealthy land owners and it became a norm for many such people to be both patrons of arts and sciences. Weather stations were frequently installed in walled gardens in the latter 19th century as indeed this Brodick Castle site originally was. This practise has, however, continued to modern times even though the initial motivations have since long gone. Whilst this Arran site is claimed to have been installed in the 1960s the Met Office has continued these walled garden installations well into the 21st Century with examples such as at Kingston Maurward added in 2019. I have so far identified 24 of the 105 weather stations in Scotland in some form of walled or botanic gardens and I have not yet completed the identification process of all sites.

I do not feel I can over-emphasise the simple fact that measuring the temperatures in a deliberately artificial environment is totally irrelevant to any form of recording producing a national historic temperature record. To suggest these sites can possibly be used to monitor “Climate Change” is no more logical than suggesting readings taken indoors are acceptable. There are many other issues that lead on from this.

Weather forecasting. Though this is not a subject within the Surface Stations Project remit, many commenters have in the past mentioned the poor quality of Met Office forecasting in Scotland. Conversely, I have not had any particularly adverse comments regarding England and Wales forecasts. It does not seem surprising though that if readings are taken from artificial environments they will not reflect the likelihood of future events in the natural world. Given that these walled garden sites are already in more benign climates (they were rarely installed on high ground sites preferring lower altitude and often coastal locations in Scotland) it is hardly surprising they are rather useless for predictions. This leads on to the next issue:

Manual/Automatic Recording. Whilst approximately a quarter of weather stations in England and Wales report their readings on a manual basis (i.e. once every 24 hours at 09:00/10:00 a.m in arrears) almost half of all Scottish stations, such as Arran, report manually. Although there are some benefits to forecasting in knowing the previous day’s data, it is not remotely as beneficial as the almost instant reporting from automated sites that dominate the England and Wales . The number of mainland (non coastal) surface stations the Met Office takes real time readings from in Scotland is so much lower than those available to the rest of the UK that it seriously compromises forecast accuracy. From the Surface Stations project viewpoint this gets even worse.

Lack of Readings A manual site that fails to report readings is effectively non- existent. I frequently make this type of strident remark that people question. So consider the recent readings history from Arran. The CEDA archived readings are freely available to view online here. A potential of 365/6 readings should be taken annually.

https://utils.ceda.ac.uk/cgi-bin/midas_stations/station_details.cgi.py?id=937&db=midas_stations

2023 ……………………………….103 days readings

2022…………………………………62 days readings.

2021………………………………..17 days readings.

2020…………………………………ZERO days readings……. blame it on Covid 19?

2019…………………………………31 days readings.

2018………………………………….21 days readings.

2017…………………………………ZERO days readings

2016…………………………………ZERO days readings

2015…………………………………Zero days readings.

2014…………………………………64 days reading.

2013………………………………..364 days readings.

For the last decade it appears the Met Office has been unable to get anyone adequately trained to take readings. As I pointed out regarding Margam no 2, any station only being intermittently read is worthless for all practical purposes. Automation of weather stations is now relatively cheap and simple in no longer requiring a electricity supply (solar panel and battery in lieu) nor landline communication (internet/ satellite comms) as demonstrated at Great Cumbrae. Any expenses currently incurred for this Brodick Castle site are a waste of money – the expense of automation would restore value to the site.

Observation and archive recording Some general points regarding observations and historic recording, though not exclusive to Arran, are worth clarifying here. The skill of taking readings has significantly changed over the years. As discussed under Metrology, temperature readings were formerly taken using Liquid in Glass (LiG) thermometers whilst now the most common practise is using Platinum resistances thermometers (PRT). The review of Floors Castle included a Facebook clip showing the observer recording the readings by simply writing down the digital display from a PRT. Accurately reading a LiG, however, actually requires specific training to avoid making parallax errors . A former observer advised me that even changing from light soled sandals in summer to heavy boots in winter could affect his sight line if he did not make personal body adjustments to accommodate the change.

The possibility of reading error may seem trivial but context is important. Below is the defined accuracy of recordings over time.

• 0.1C for all temperatures that are reported today
• 1C for hourly temperatures received in SYNOP code prior to 1982
• 1F for many maximum and minimum temperatures prior to 1961
• 0.1F for most air temperatures prior to 1961 and for some maximum and minimum temperatures prior to the 1920s
• 0.5F for maximum and minimum temperatures at a few stations prior to 1961

Readings prior to 1961 were usually to the nearest 1°F (which equates to 0.55°C). Given that modern day readings are recorded to the nearest 0.1°C, this represents a significant change in accuracy level and the potential of a 1°F parallax reading error would be hugely significant. It is also worth noting how this conversion in archives creates a “step point” change with only certain Celsius points being possible on conversion.

Arran Brodick weather station was, however, originally installed in 1859 and not the archived 1964 date and back then readings were recorded to 0.5°F as shown here.

All the above demonstrates the difficulties of historical comparisons and how reporting changes over time to the second decimal place of a degree celsius is well beyond the validity of the original data. This leads onto a further issue of “climate averaging“.

Geographic spread of Weather Stations. Island sites to the west and north of mainland Scotland alone represent over 15% of all Scottish weather stations. A further 30% of sites are now located on the western coasts. Whilst there has been an historic UK wide trend of reduction in the number of weather stations initially brought about by automation requirements of the latter 20th century, this has much less applied to these isles. Surprisingly new manual additional sites have been installed on Raasay:Osgaig – 7/3/2023 and Skye:Harlosh 6/4/2023. This contrasts starkly to the situation over on the much colder eastern coast as highlighted with Fife now only retaining 3 low grade sites.

This uneven mix is likely led by forecasting requirements monitoring the predominant direction of weather fronts but slews the averaging of such site’s data to falsely indicate a progressively “warming” trend. This situation is also demonstrated in the south west of England where Devon has a very high density of weather stations that is similarly increasing with new sites at Bratton Fleming from 2021, Stoodleigh Dryhill from 2023 and Hollacombe installed just 3 months ago on 4/11/2024. The Met Office explainer for this south west England concentration equally applies to the islands and western coast of Scotland.

The ever changing selection of the physical real weather station site mix (rather than artificially concocted numbers for non-existent sites ) is one of the causes of “warming” appearing in the averages being influenced by sample selections rather than actual change.

In conclusion Arran Brodick Castle weather station is unacceptably poorly sited in artificial surroundings. Readings are not being consistently nor reliably taken rendering those that are taken worthless. Even if reading frequency were improved the data (if used for climate reporting purposes), would only further continue to degrade the dataset. The site demonstrates many of the Met Office’s failings both in its primary forecasting role and even more so, in its function of the collation of useful historic data.

Source: https://tallbloke.wordpress.com/2025/03/03/arranbrodick-castle-dcnn-6133-a-case-study-of-multiple-scottish-issues/