The curious tale of the HadCRUT land station data

The global land station climate data used in the CRUTEM climate series is no longer available to download from the Met Office site as a plain text file. For many years this data was always made available in plain text format, but with the release of version 5.0,1.0 this is no longer the case. For some bizarre reason the Met Office can now only supply it in NetCDF format because they have “retired” the old ASCII text file. This is a pain because there are 10,640 station files in NetCDF format in the massive 121 MB zip file, and as far as I can see each file would require individually parsing from NetCDF into text, although there may well be a utility program that can automatically do this for you. To all intents and purposes the UKMO have deliberately obfuscated this data to the public.

The Good news

Luckily, after complaining to the Met Office, they have relented. They did this by cleverly foisting the hosting of a new fixed space delimited text file of land station data onto the Climate Research Unit. I have reworked my parser and can now view any of the monthly means from the 10,640 stations from around the world in the new format file. You can access the new file from the CRU site, hopefully they will regularly update it from now on.

Now the bad news

If you, like me, love climate data, then you’ll be disappointed to find that the new plain text file from the CRU only contains monthly climate data from January 1850. This is a great pity because the old version contained quite a number of valuable long temperature series which have now been shortened (see below). The oldest series was the one for Berlin which started in 1701, that’s not so long after the start of the composite Central England series of Gordon Manley in 1659.

The old CRUTEM4.6.0.0 data

Incredibly shrinking data

The other curious thing is how the number of land stations has drastically decreased in recent years. If this monthly land stations data is really used to construct their temperature anomalies with, you would logically think that the more of them there were, the more accurate the estimate of temperature over land would be.

  • Why has the number of global land stations used in recent years fallen by 66% from over 7,500 per year in the 1970’s to less than 2,500 in 2020?
  • And why has the number of land stations in WMO block #03 in recent years fallen by over 75% from over 120 in the 1960’s to around 30 in 2020?

Surely the UKMO could supply a few more land stations than 30 for their own global temperature series? On any day I can access temperature extremes from SYNOP reports for around 160 UK stations alone. So why has this been reduced in recent years to a rather odd selection of just 30 stations? I have posed these questions to a Climate Science Communicator at the Met Office and hopefully they will reply in due course.

Scroll to Top
%d bloggers like this: