Tuesday
Feb162010
by
Bishop Hill

JG-C in the Times



The Times covers John Graham-Cumming's discoveries of a series of errors in the Met Office's code for its HADCRUT temperature index. The article also quotes none other than Bob Ward speaking up in favour of data availability.
Reader Comments (14)
They asked me to write a short article to go with the piece that Hannah Devlin put together. It's significantly cut down in the printed paper, but the full text of it is online: http://www.timesonline.co.uk/tol/news/uk/article7028418.ece
Not strictly relevant to your subject here, but a funny comment on green mania nonetheless. At the Vancouver Winter Olympics, greenies insist that an eco-friendly ice making machine be used. It's a disaster, creating huge delays, nearly causing competitions to be cancelled and injuries to competitors. Organisers eventually relent and go back to the non-eco friendly model which has worked for decades.
http://www.ctvolympics.ca/speed-skating/news/newsid=43070.html
Thank you John. You are doing a great service. If only the Met Office and other government organisations were a bit more professional and a bit less arrogant.
Not exactly a gracious response from the Met Office ...
“We are grateful to Dr Graham- Cumming, but they are quite minor changes,” said Peter Stott, head of climate monitoring and attribution at the Met Office. “It shows how open we are. We have put an exhaustive amount of information out there to show people exactly what we do.”
...
During the checking procedure Met Office officials discovered further problems with US temperature calculations.
... one hopes that they are reviewing their methodology.
"Scandinavia-gate"
http://hidethedecline.eu/pages/posts/scandinavian-temperatures-ipccacutes--scandinavia-gate--123.php
How reliable are land and ocean records?
The Met Office is ISO-9001 and TickIT accredited. Good job they have procedures in place that enable outsiders to do their work for them. I hope Dr Graham-Cunning has sent the Met Office an invoice. It's no use the Met Office having a £20M computer and crap code.
Thank you for posting the results of your investigations. There are others out there doing the same thing, identifying the location, life and death of these temperature stations. This poster, for example, has spotted what he describes as Zombie thermometers, apparently dead but which become undead later:
http://chiefio.wordpress.com/2010/02/15/thermometer-zombie-walk/
It is unclear just what impact this has on the data or how you construct an audit trail to discover what has been happening if it does turn out to be significant. As a lay observer of all the questions raised here and elsewhere about the sources and statistical manipulation of data I do wonder whether the foundations of what we are asked to believe, and do in response, are as sound as they should be.
The stuff about the location and quality of the thermometer sitings is interesting. As you are probably aware there's a project called Surface Stations in the US taking pictures of them all. I've asked them to release their underlying data to me and they are currently discussing amongst themselves whether they want to do that.
For an interesting read on the impact of poor siting of thermometers you might like to read: ftp://ftp.ncdc.noaa.gov/pub/data/ushcn/v2/monthly/menne-etal2010.pdf
@PhillipBratby
Do not be fooled by ISO-9001. All it shows is that the institution has written down its procedures. Its says nothing at all about the quality of the procedures.
It would be perfectly acceptable in ISO-9001 land to have a procedure called FOI request. With the process 'Ignore, but abuse requester'. Sounds fanciful I know, but they would have achieved ISO-9001 standard just by having written it down.
That Times story has now moved around the world today. It's on the front page of US Fox News and there's even a Nature blog post about it: http://blogs.nature.com/news/thegreatbeyond/2010/02/how_a_computer_expert_correcte.html
@ Latimer Alder.
I said it tongue in cheek. I am fully conversant with ISO-9001 and know there is a difference between having good procedures and bad procedures, both of which can be allowed under ISO-9001.
May we have some clarification, please? Is Hadcrut in the Dodgy Dossier business or the Sexed-Up dossier business?
@ dearieme:
The answer to that question must be: "Yes". 25 years in global climate terms are equivalent to 45 minutes in international relations.
In the times article they quote the Met office as saying, "When all of the errors identified were corrected, the temperature trend remained well within the 95 per cent confidence range of the original plot, meaning that the difference would not be considered scientifically significant."
My understanding of a confidence range is that it quantifies the amount of sampling error which is caused by random sampling of the data. It does not and cannot say anything about mistakes in the calculation process. As such their comment is meaningless.
They need to calculate new confidence intervals based on the correct (big assumption here) data. The new confidence interval can then be used to test hypotheses. Since the edges of the new confidence interval will have changed, previous comparisons that were (or were not )statistically significant may have changed as well.
With those few identified errors, did the trend go up, down or stay the same?