I have always been fascinated by the Central England Temperatures series first assembled by Gordon Manley in 1953 and which I have written extensively about it my blogs over the course of the last 8 years. Back in 1983 one of the first programs I ever wrote in BBC basic was a program to visualise the monthly CET data series from 1659, and back then I even had to create my own data file to do it! The daily mean series introduced by the Met Office later, and which starts in 1772, is even more useful, but there is one thing that has always puzzled me about it, and that is the unusual month of December 1786, because on every day that month the temperature is exactly the same at 2.8°C. Don’t take my word for it, after all I may have parsed the values incorrectly, so here’s a snippet from the raw text file, which you can download from the Met Office website, so you can see for yourself.
So what’s going on?
The only thing I can think is that during December 1786 there was a great deal of missing data from the individual sites used to calculate the daily mean composite value. The Met Office may be reasonably confident about the mean for the whole month is correct, but can’t be confident about the individual daily mean temperatures for each of the 31 days that make up the month, so they replaced all the daily means with the monthly mean. I have never read anything on the Met Office website to indicate that this is indeed the case, but I am quite confident this is why.