Problems with Met Office Data? - the Central England Temperature record is broken
Central England Temperatures welcome page
These daily and monthly temperatures are representative of a roughly triangular area of the United Kingdom enclosed by Bristol, Lancashire and London (view map). The monthly series, which begins in 1659, is the longest available instrumental record of temperature in the world. The daily series begins in 1772.
The Met Office have also been compiling Maximum, Minimum and Mean Daily Central England Temperatures data files since January 1878. The following stations are used by the Met Office to compile the CET data: Rothamsted, Malvern, Squires Gate and Stonyhurst and revised urban warming and bias adjustments have now been applied ....
Please note that the Met Office data sets are available for bona fide academic research only..
Luckily the data are available at Met Office Hadley Centre HadCET observations datasets
So let's look at Maximum Central England Temperature, 2009
The pale dotted line in the middle is this year (click to enlarge) - Phew what a scorcher! Look at that late June peak, way above average!
Where was it? We know this record is a blend of records from apparently three stations, and their records are also graphed on the same page:
The graphs for minimum and mean temperatures also show the same inconsistency.
I think I better ask Tim Legg what is going on, unless you have an idea.
UPDATE - And Tim kindly replies:
I can assure you that the CET software is running correctly. The graphs
you include on your blog page as "max Pershore - Gt Malvern" etc. are
there to explore any differences between the stations which comprise the
CET series. As time goes on, sometimes a station closes, for any of a
variety of reasons. One thing we wish to ensure is the continued
homogeneity of the long-standing CET series, so we are careful when
choosing a replacement station to ensure that it is representative of
the 'Central England' area, and to allow for any systematic biases
between that station and the one it replaces. Only then can we be sure
that any long-term trends in the data are real.