Is the GISS temperature index fraudulent?

David Stockwell is an Australian statistical expert who has written a book covering, among other things, statistical tests for detecting datasets which have been manipulated in some way. He also has a blog called Niche Modelling which is well worth a visit.

His latest post outlines the results of running one of these fraud-busting tests on NASA GISS's global temperature index, and the results were rather interesting....

RESULT: Significant management detected.

David is quick to point out that he's in exploratory mode and hasn't actually drawn any conclusions yet, but this is definitely going to be one to watch.

Update on Jan 15, 2009 by Bishop Hill

Lubos Motls has picked up on Stockwell's story and taken it a bit further by running the test over the satellite data from UAH. No sign of any manipulation there.

What fun!

Comments

[Phillip Bratby]

I wonder how easy it will be, if people know their manipulations can be detected, to manage the manipulations in such a way so that they cannot be detected .

[jnicklin]

Without knowing exactly how the manipulation detection works, it may be quite difficult to avoid detection short of not releasing the data set for examination. With GISS that may become the preferred route, afterall, it is highly classified religeous material that should only be viewed by the high priests. We lowly parishoners shoudl just take their word for it.

[James P]

Hmm. I found the following quotation from Prof.Stephen Salter (of Duck fame) while revisiting his shameful treatment at the hands of the Atomic Energy Authority, whom Mrs T had unaccountably put in charge of renewable energy (!) in the early 80's.

"We must find a way of reporting accurate results to decision makers and have decision makers with enough technical knowledge to spot data massage if it occurs."

Plus ca change...

[Frank O'Dwyer]

So this would be the same index that is supposed to be

1) Flat
2) Showing cooling
3) Showing warming due to the urban heat island effect

To these mutually repugnant claims, we must now add the notion that the authors fudged the numbers to support the denialist case.

Well Amen and preach it. Can I get a 'good grief'?