Swanage DCNN8781 – Analysis of the issue of scales and the implications of conversion.

50.61361 -1.95870 Met Office CIMO Assessed Class 5 Temperature records from 1/1/1931

Swanage weather station is one of those coastal sites that are kept in first class pristine order by the local authority almost as a matter of “civic pride”. It is beyond any question one of the best maintained sites in the UK and, in its manual reading days up to 31/1/2011, the observations were impeccable with hardly ever a missed day or reading. Whatever the merits/demerits of the location, it cannot be denied this site was exceptionally well run. The Met Office assessment is Class 5 which seems unduly harsh in this case when compared to their ratings of many other sites.

Assessing the site, firstly it is coastal and on a very gently sloping greensward above the promenade with (from personal memory) a delightful view out to sea. I do not think the slope is a particularly troublesome feature, so what of any surrounding problems. The 10 metre radius circle area does not indicate any large area problems.

The roadways are respectively 16 metres to the south and 20 metres to the west. The potential shading tree to the north west is similarly too far distant to be an issue. The Met Office low grading seems to come down to similar issues they chose to identify at Bradford Lister Park (another excellently maintained site) of compromising features within the enclosure itself. Here is Swanage in close up from a southerly viewpoint.

Secondly (to allay my concern that the opening door to the screen was on the wrong elevation- it isn’t) here it from a northerly viewpoint n.b. this is a 2009 image for better clarity with the nearby small tree only just planted – a concern for the future.

This site being rated Class 5 whilst Heathrow is rated Class 3 tells you everything you need to know about the Met Office and motivated reasoning. They will no doubt claim the structures close to the screen are a “major” problem (and they are obviously undesirable) but they almost certainly date back to the original site construction so at least are not altering the site conditions. I would personally derate Heathrow to Class 5 and slightly elevate Swanage to Class 4 more on the basis of the slope not being desirable together with the brick structures making Class 3 difficult.

Swanage does though have a very long term and very reliable observing record and that in itself is valuable – until, that is, the Met office starts its manipulation.This introduces the issue of differing scales.

Modern western culture in Europe and later the UK chose to adopt a “decimal” (Metric) system for most scales largely due to making numeric calculations quicker and simpler. Most major world currencies (including the US Dollar) run on such decimal scales with even the UK finally adopting one in 1971. It is important to note though, that while Europe had already adopted the Celsius system and the Met Office changed over all weather stations to Celsius recording from 1961, there is nothing decimal at all about the Celsius scale nor can there be. The fact that the scale starts at Zero degrees for the “freezing point of water” and 100 degrees for the boiling point of water (at a set atmospheric pressure level) in no way whatsoever makes it preferable to the Fahrenheit scale which was based on other preselected parameters. A scale is a scale is a scale however you play with it. That the Kelvin scale starting from Absolute Zero adopted the same graduation points as Celcius in no way makes celsius any better a scale for meteorology than Fahrenheit which is the predominantly used US choice.

On this point, many years ago a “critical thinking” teacher read out to his students (I was one of them) the following multiple choice question for a show of hands answering.

“You are in a room with a temperature of 20 °C. The heating is turned on and the temperature rises 10% – what does the temperature become?

a.) 321 degrees (“that’s 123 backwards” he added)

b.) 49 degrees (“my age” he said, but was in his 60s)

c.) 22 degrees (he raised his voice to emphasize this one)

d.) 18 degrees. ( he rolled his eyes to suggest only an idiot would think this)

e. ) 2 degrees. (he pointed to the door if you were to answer that one)

Unsurprisingly everyone put their hand up for option C and 22 degrees. He paused and then asked “Why did you all get the answer ………wrong!

The whole lesson was actually about the reasons why people ask questions and was a demonstration of how we could all fool ourselves into thinking the wrong things courtesy of the clever salesman or the politician. The “false base” premise was duly explained as nobody could understand why they were wrong. He first started by converting the numbers into Fahrenheit thus 20°C = 68 °F. Adding 10% = 74.4°F. Converting 74.4°F back to celsius gave 23.6°C! The “obvious” and potentially “no brainer” mathematics simply did not apply and somewhat perversely the nearest “possibly” correct answer was 49°C. {293k x 1.1 = 322k = 49°C (though even that is not really true!)

All the above is to demonstrate there is nothing intrinsically better for climate recording purposes of whatever scale is used as long as it is consistent over time , which of course the changeover from Fahrenheit to Celsius altered. The Met office quote Location Specific Long Term rolling climate averages only from 1961 onwards on their general public website for probably two principle reasons. Firstly there was no change over in scale used, and (cynically) the period from 1945 to 1975 showed a significant cooling trend allowing the subsequent rebound figures to be shown from a suitable “false base”.

Swanage had 30 years (conveniently a climate averaging period) from 1931 to 1960 of completely Fahrenheit recorded figures that would have made an easy comparison and only the end resultant figures would need conversion not the tens of thousands in between. {Editors note: Yes I am on the case of doing just that}

However, by way of demonstrating just how distorting the Met Office conversions actually were, modern Celsius readings are to the nearest 0.1°C whilst most pre 1960 stations reported to the nearest degree Fahrenheit which represents 0.55°C. The Met Office has a policy of rounding up the 0.05 to 0.1 thus 33°F was transcribed into the archives as 0.6°C. So far so good, but they perversely do the same for minus readings thus 31°F became recorded as -0.6°C. Just 2°F variation suddenly could become 1.2°C. Of course 34°F (2°F above freezing point) does not require rounding and it was recorded as 1.1°C thus a tenth of a celsius degree variation for no reason.

The end result was a very “blocky” set of figures with only certain conversion figures actually possible and fahrenheit degrees having different values depending on where they appeared on the scale. To demonstrate this is an archive from Swanage below. Columns “I” and “J” are maximum and minimum respectively.

This set demonstrates that +0.6 to -0.6 distortion and the “blocky feel to the data with many “impossible” data points such as no 4, 4.1, 4.2 and 4.3 °C readings ever appearing. Despite this the Met Office still absurdly feels it can produce results to the second decimal place accuracy of a celsius degree.

This difficulty in accurately reconstructing temperatures over time was highlighted (amongst other issues) by Stephen Connolly in his work on comparing temperatures over time, and is demonstrated by sites like Swanage.

In summary Swanage is a well maintained site and whilst not perfect is probably considerably better than the Met Office seems to judge it. It does have a reliably taken long term record and the site conditions are unlikely to have significantly changed over time. I will try to produce an early period climate average to see quite how the older Fahrenheit readings can be compared (if genuinely possible) with the modern celsius variants. A work in progress.

Source: https://tallbloke.wordpress.com/2025/06/20/swanage-dcnn8781-analysis-of-the-issue-of-scales-and-the-implications-of-conversion/