Coleshill WMO 03535 – An interesting comparison made with Private Weather Stations.

52.48022 -1.69063 Met Office CIMO Assessed Class 5S Installed 1/9/1997

Tim Channon reviewed this site in 2012 and back then felt it was probably good enough for a fully accurate and representative Class 2. What has happened to degrade the site to the Met Office’s own assessment as the worst possible is a typical tale of lack of oversight and diligence. Astonishingly this site was also used as a comparison site for testing private weather station equipment with Met Office equipment – only it wasn’t really!

Tim was a very astute and methodical judge of weather station sites, he would not have called this so wrong. The intervening 13 years of unfettered tree growth that he noted has been one factor creating the obvious heavy shade (those are very long shadows) but there are other problems now evident. The blurry headline image indicates a vehicle parking alongside the enclosure which is a recurring theme in google Earth historic imagery. Caravans, also mentioned by Tim, come and go and the extensive ground rutting and discolouration of the grass in the adjoining field indicates a high level of activity and occupancy period. Street View looks as below and is now getting worse. Even the Met Office inspectors could not fail to mark down such a publicly visible site but why nothing has been done to correct matters is a mystery. Not only temperatures but wind speed and direction readings will now be suspect.

As I have often emphasized in the past, the Met Office makes a big play about how its sites meet much higher standards than private sites. However, it far too frequently seems Met Office sites are inferior to a large proportion of private sites. In researching this issue of Met Office versus private sites I came across equipment comparative testing research carried out by Aston University.

There are a number of numeric errors in this piece i.e. “the Met Office runs 250 or so land surface stations in its professional Meteorological Monitoring System,”. “250 or so” seems rather vague considering it is actually 384. I am also concerned with the definitions given to private (“citizen”) weather stations which states:

“A citizen weather station (CWS) is defined as a weather station set up by a member of the public for whom the terms weather enthusiast, volunteer, hobbyist and amateur observer are fitting
descriptions. Crucially, these stations are set up out of personal interest (or, in schools, for
educational purposes) rather than because it is the owner’s job.”

The authors seem unaware that very many official sites are in schools, playing fields, domestic back gardens, walled gardens, botanic gardens, sewage plants, water treatment works, surrounded by solar farms, alongside jet engines and wind turbines, heliports and on and on.

The amateur suggestion is also quite incorrect as many private sites are genuinely for professional purposes. The nearest three PWS to me, as examples, are variously at a vineyard for horticultural purposes, a private airfield and a golf course/spa leisure centre. Certainly not the almost hapless, but well meaning, amateurs being portrayed above. This globally searchable map is a better indicator of PWS sites https://www.weatherlink.com/map

However, the main point of the above research that concerns the Surface Stations Project is the applicability of PWS readings versus Met Office stations for wider area representation. This section is very apt:

This research is starting from the unfounded assumption that Met Office sites are better locations. “Can CWS really compete with accurate equipment installed by professional organisations at well calibrated
and exposed sites?” Calibration is one issue but “well…..exposed sites” is clearly a false assertion for Met Office sites like Coleshill in question here in more ways than one. The key noted item

“5 Representativity error – It’s difficult to assess whether CWS observations represent a
scale suitable for the application.”

Is an issue far more relevant to the Met Office who openly have to admit to the vast majority of sites being Class 4 and 5 having such wide error margins of +/- 2°C or even +/-5°C making minute calibration variations rather irrelevant. I would argue that all three of the previously mentioned PWS near to my home are better sited than the majority of Met Office sites.

In the case of this research, the chosen site to monitor the private instrumentation against Met Office equipment was the Met Office/Birmingham University Winterbourne No 2 – a Class 5 (junk status) site subject to severe UHI where the Met Office parodied the term “extraneous heat sources” by concerning themselves withthe siting of low power electricity junction boxes. { talkshop comment Bingley No 2 with its attendant massive electricity sub station is not a concern though.} This site was certainly chosen for local convenience over site quality, but the astonishing admission of Met Office inadequacy came in the tiny footnotes to Table 2 in the “results”:

• Winterbourne No. 2 site does not have MMS MSLP readings, instead observations from the Coleshill MMS site 16km away were used. CWS pressure readings were set to match the Coleshill
  reading at the start of the period, except for the WMR200 for which the MSLP correction is based upon the elevation the user enters into the electronic console.

To clarify “Winterbourne No.2 site does not have Meteorological Monitoring System Mean Sea Level (air) Pressure (MMS MSLP) readings that the comparison technologies automatically include. So instead the “data” from the Coleshill Class 5S site subject of this review (even worse than Winterbourne) that is 16km (10 miles) away was inserted. This rather exposes the myth of Met Office equipment superiority and brings this research into question when casually dropping in data from a distant site.

In conclusion the issues with Coleshill reflect poor siting and ongoing maintenance that is the responsibility of the Met Office. Despite this clear lack of adequacy the Met Office is still seen by its “peer” academic institutions as credible in its claims of high quality even in the face of researchers having to “amend” comparison terms. The research subsequently found “possible” variations of calibration quality of readings measured in one or two tenths of a degree (assuming the Met Office equipment was correct) whilst completely ignoring the known plus/minus 5°C wider area representation which was a key factor noted in the testing.

Given all the above, I feel the Surface Station Projects objective of identifying good quality sites from which to reconstruct an accurate and reliable long term national temperature record is all the more important. The Met Office and its peers do not appear to be interested in genuine credibility.