The only thing we can be sure of about the future is that it will be absolutely fantastic. So if what I say now seems to you to be very reasonable, then I have failed completely. Only if what I tell you appears absolutely unbelievable, have we any chance of visualizing the future as it really will happen.

- Arthur C. Clarke (h/t Brin)

Thursday, March 10, 2011

Ben Santer on Christy Testimony

Via a mailing list, reproduced here with permission:
I have had a quick look at John Christy's recent Congressional testimony. Many aspects of it are deeply troubling. From my own personal perspective, one of the most troubling aspects is that Christy cites a paper by David Douglass, John Christy, Benjamin Pearson, and S. Fred Singer. The Douglass et al. paper appeared in the online edition of the International Journal of Climatology (a publication of the Royal Meteorological Society) in December 2007.

Shortly after its publication, it became apparent that the authors of the Douglass et al. paper had applied a flawed statistical significance test. Application of this flawed test led them to reach incorrect scientific conclusions.

Together with a number of colleagues (including Gavin), I prepared a response to the Douglass et al. paper. Our response was published by the International Journal of Climatology in October 2008. (DOI: 10.1002/joc.1756) I am also appending a "fact sheet" providing some of the scientific context for both the Douglass et al. and Santer et al. International Journal of Climatology papers.)

To my knowledge, the Douglass et al. International Journal of Climatology paper has never been retracted. Nor have the authors acknowledged the existence of any statistical errors in their work. The fact that John Christy has now cited a demonstrably-flawed scientific paper in his Congressional testimony - without any mention of errors in the Douglass et al. paper - is deeply disturbing.

It is my opinion - and the opinion of many of my scientific colleagues - that the Douglass et al. International Journal of Climatology paper represents an egregious misuse of statistics. It is of great concern that this statistically-flawed paper has been used (and is still being used) as crucial "evidence of absence" of human effects on climate.

----------------------------------------------
Benjamin D. Santer
Program for Climate Model Diagnosis and Intercomparison
Lawrence Livermore National Laboratory

8 comments:

Dana said...

Good point about the use of Douglass et al. I just kicked myself for not thinking to make that point in my rebuttal to Christy's testimony on Skeptical Science. I got caught up in the (worse, IMO) error of claiming the 'hot spot' is an anthropogenic fingerprint.

Just goes to show what a horrific Gish Gallop Christy's testimony was. Really hard to catch all the errors therein.

Tony Lee said...

Bad use of statistics in an academic journal? Isn't that Steve McIntyre's desmesne?

Tony Lee said...

Looks like a job for Climate Audit!

Steve Bloom said...

And what will the scientific community do to deal with this long-standing Christy problem? The usual solution of waiting for him to retire and die seems to fall short. Note that Andy Revkin relies on Christy for support of his view that things won't be so bad all that soon. The scientific community's seeming inability to deal with Christy and the handful of similar problem scientists is a major part of the climate science communications problem. But of course we can't bear grudges, can we?

rustneversleeps said...

Backstory video on Ben's personal and professional frustrations in rebutting that paper and the appalling aftermath.

Christy - wtf???

David B. Benson said...

deeply troubling

What an understatement.

Christy has previously made it quite clear that his ministry does not allow him to view global warming as problematic.

Marco said...

Tony Lee: McIntyre DID take it on. Well, actually, he didn't, he repeatedly declined to comment on Douglass et al, and sent in a paper attacking Santer et al ("you left data out, boohoo!").

Statistical trickery in that paper by McIntyre included comparing to the average of ALL models. There we go, statistically significant differences between the models and the observations!

Martin Vermeer said...

Yes, this was the paper that Tom Wigley called a 'scientific fraud', of course in an email intended to be private.

The error is so elementary that it takes uncommon credulity to accept that a trained scientist like Christy would not only allow it to be made in a paper co-authored by him, but continue to fail to grasp its wrongness, all these years after it has been pointed out and rubbed in.

Shameful.