Marlborough DCNN5877 – A very long term site for possible inclusion in climate history reconstruction.

51.416110, -1.7353645 No CIMO Assessment Archived Temperature data from 1874 to 30/11/1996

Although there are still around 18 UK official operational Met Office weather stations for me to review, I will also now continue reviewing those sites suitable for inclusion in the construction of an alternative historic temperature record. So far the Rothamsted site is an obvious contender and a number of other current sites are good candidates. In this report I will reiterate the conditions required for inclusion, some of the means to overcome inevitably missing data and relate to investigations of closed sites in similar manner to the investigation of Raunds. Apologies for a slightly “drier” post but it is important to rationalise exactly where I am going with the Surface Stations Project. This is intended as much as a discussion and suggestions piece so any readers please feel free to comment or contribute.

When Dave Woolcock kindly produced a historic data representation for Wye, I also stipulated the requirements for station inclusion into any temperature reconstruction I am attempting and these conditions are worth repeating.

“I would like to re-emphasize this is not a cherry pick. Wye simply represents the first long term operational site that I could establish meeting the following criteria:

1. CIMO Class 1 or 2 Site.
2. Not relocated over data reporting period.
3. Identified good maintenance with no significant change of surrounding conditions.
4. Not subject to significant Urban Heat Island effects.
5. No instrumentation changes (i.e. from LIGT to PRT)
6. Consistent good quality observations record. (WYE was excellent)

As I discovered with my research on Raunds, even very long term sites have had relocations and evaluating site conditions over 100 years ago is difficult, but not impossible. Thus there will inevitably be an element of uncertainty regarding the distant past though given the much lower level of urbanisation in the 19th century perhaps less problematical than modern site moves.

The aspect of site maintenance is surprisingly not really an issue at all. Older standards of maintenance and observation were much better than modern day practise. As I noted at Lake Vyrnwy, both sites were excellently well sited and impeccably observed regardless of the harsh conditions at times. {n.b. – Lake Vyrnwy is a definite contributor to my reconstruction attempt} Conversely, as I noted at Hayling Island (and others) Local Authorities declare “Climate Emergencies” but cannot manage to get observations taken reliably even in the most benign of locations. Modern day Cromer has recently looked disgustingly badly neglected whilst historic images display a traditional high quality level of “Civic pride” now long gone.

The issue of instrument changes is very tricky indeed. Even manually observed stations were switched from LIGT to PRT by 2017 (at the latest) and automatic conversions required switching to PRTs occurring mostly in the 2009/12 period completing a process that had started in the latter 1970s and accelerated in the early 1990s. Dave Woolcock has devised a means of “correcting” minimums and maximums to ensure they are comparable in date times to manual observations. Unfortunately the winter bias cannot be overcome (more posts on this shortly with some interesting comparative French real world data supplied by an expert reader) so as far as possible only manually observed stations will be used up to the early 21st century. Beyond the mass adoption of PRT recording in 2017 is a problem to still be worked on.

Geographic spread is a major issue for me in that the Met Office seems remarkably adept at manipulating sample points in a way I feel is deliberately designed to modify averages that I detailed here . To simplify the problem I have opted to operate on a solely “Great Britain” basis of only MAINLAND England, Scotland and Wales. The island of Ireland is a separate issue for others.

The Met office delineates 9 “climate districts” as shown below.

I find these demarcations to be risible and bear no logical relationship to anything other than a rather odd Civil Service administration boundary system. For example, on what basis are Bristol and Kirkby Stephen (Cumbria) both in the “Midlands”? How is RAF Brize Norton considered part of the “South East” whilst parts of West London drift into East Anglia? I feel the nationalist concept is seriously overrated for climatology and even the map is an area distorting “Tranverse Marcator Projection”. Geography can look very different from an “equal area” projection and changed viewpoint!

Period cover for my reconstruction is from the universal adoption of Stevenson Screens in 1884 to present day being just over 140 years.

For my part I feel that the GB climate historical record could be accurately monitored by around 20 good quality sites for which the exact locations I have yet to determine. It is worth remembering that the current Central England Temperature Series only uses 3 sites and frequently used even less in the past. Unfortunately, there are unlikely to be 20 continuously operating sites meeting the criteria going back so far into the Victorian era thus this will have to be a composite using a few stations as “Benchmarks” and building around those over time as indeed the CET does.

The final issue of infilling missing readings is very much open to discussion. My personal view was to use nearby sites shown to genuinely bear a correlation to infill but that this should be kept to a bare minimum. The Met Office seems to operate on the basis that up to 2 missing days readings in any calendar month was still acceptable so I propose to continue with that.

Any views, advice, comments or criticisms most welcome. There are other issues to consider (i.e. scale conversions) but I will cover those in future posts.

So Back to Marlborough

Obviously I am considering this site for possible inclusion so what is my justification given the strict parameters previously established?

Location: The site is believed not to have moved over its long 122 year history at Blackett Observatory. and is located within the grounds of Marlborough College.

Traditionally weather stations were associated with astronomical observatories (Oxford Radcliffe, Jeremiah Horrocks, Herstmonceux and Weston Park are a few of many examples) and this ensured a reasonable lack of interference from outside sources in their earlier days. Sadly this changed over time with astronomical instruments having to be moved from Greenwich to Herstmonceux for instance.

In the case of Marlborough the college itself is very much on the town’s outskirts and Marlborough has not been subjected to large scale development over the years to the extent other locations have.

I certainly cannot pretend that this site is not affected by an element of Urban Heat Island effect but it will not be anywhere near as adversely affected as most long term urban sites. Another example of using the best available and not making 100% perfection the enemy of the good. At this point I would like to emphasise I have not studied the data over time in any way other than its relative completeness, in order to avoid my own sampling bias. Simply put, in location terms this is not the best but seems reasonably good for a long term late 19th to 20th century comparison.

Observations. Digital records start from 1875 but with lots of gaps up to 1921 from when the archive looks complete. The manual archives do show copies for the missing years so these may be either in the process of entering or could be digitally transcribed myself.

Post 1996 data. As is often the case, although the Met Office may “de-list” sites, they often continue reporting privately – Greycrook is a perfect example. I am in the process of clarifying this point, however, anecdotally, readings have continued by the school itself and indeed may even be by traditional instrumentation. It is very common for educational establishments to host weather stations (especially “Independent” ones notably Stowe) and it would be remarkably helpful if this, or any other sites, continued to use Liquid in Glass Thermometers for continuity.

Potential Infilling sites. Larkhill is nearby with a history from 1920 but is frankly a ridiculous site alongside a sand school in its modern iteration. Manuscript records, however, indicate the original site was differently located much further from the built up area and may be of use for earlier 20th century readings. A likely better option may be the Porton Down weather station operative from 1918 to 1980. This highly secretive site was much less developed in the weather station’s operating period and the adjacent trackway was only lightly used.

Above: Porton Down fully equipped synoptic weather station in 2005 – note 25 years subsequent to cessation of archived records.

It is worth noting that Marlbough College does appear as a “Climate Station” on the Met Office Location Specific Long Term Averages web page BUT the figures noted there will obviously not be solely from Marlborough as it closed in 1996 thus the period 1990 to 2020 simply cannot be from the closed site. The Met Office quoted figures will represent adjusted, modified, homogenised and even potentially guessed data from unknown/unspecified surrounding locations that are continually changing. Any reconstruction by the Surface Stations Project will fully disclose methodology and, where other site’s readings are used, this will be fully disclosed.

I hope this post is a guide to where the project is aiming to get to and how it is intended to get there. I would very much welcome any comments guidance or views anyone may feel they can offer. Anyone may contact me confidentially at [email protected] if they prefer. Anonymity, if required, is perfectly acceptable.

Source: https://tallbloke.wordpress.com/2026/07/20/marlborough-dcnn5877-a-very-long-term-site-for-possible-inclusion-in-climate-history-reconstruction/