Norwich, we still have a problem -

Norwich, we still have a problem


Laymen who have understandably decided to accept what much of the media now treats as axiomatic–that humans are causing potentially catastrophic global warming–must now be suffering some anxiety over the leaked e-mails from the Climatic Research Unit. Is an opinion leader like George Monbiot right to view this as a serious matter, or should they believe the reassurances of somebody like, say, Toronto Star environment columnist Peter Gorrie?

I ask solely as a matter of media-consumer interest, because, realistically, what Gorrie writes doesn’t matter to a climate-change skeptic, or to anyone with the time and the quantitative training to follow a scientific debate on his own. It matters to the guy on the subway who avoided Stats 101 as if it had horns and fangs, and that guy is now getting conflicting signals. I presume Gorrie would agree that his job is not just to confirm that reader’s prejudices–though people do like having their prejudices confirmed, and any argument a columnist can make will confirm somebody’s.

Like other columnists covering the CRU leak, Gorrie zooms in on just one “example” from the e-mails; although the etymology and sound of that word “example” would seem to imply some element of randomness in the selection, many of these columnists are choosing the same e-mail, because it contains an apparent faux pas that is relatively easy to explain away:

In one email, the research unit’s director, Phil Jones, refers to work by another scientist, Michael Mann, published in the journal Nature: “I’ve just completed Mike’s Nature trick of adding in the real temps to each series … from 1961 … to hide the decline.”

“Trick” doesn’t refer to sleight of hand; it’s jargon for a good, useful solution to a research problem. The problem in question relates to the fact that one method used to estimate temperatures over centuries – measuring tree rings – doesn’t give good recent results. But actual observations, the “real temps,” were available.

It’s much easier to understand “scandal” than even that simplified explanation.

He’s right about the word “trick.” Scientists do use the word to describe simple solutions to sincere research problems. It does not, on its own, imply deception. The real problem with the Jones e-mail is the part about “hiding the decline.” The issue, really, is right there in Gorrie’s paragraph: tree rings appear to have serious problems as a means of inferring global surface temperatures from before human records were kept. As an abstract of the Briffa study Jones was discussing notes:

…tree-ring density records become de-coupled from temperature after 1950, possibly due to some large-scale human influence that caused wood densities to decline. Thus, the reconstructed temperature record after 1960 is considered unreliable.

Jones’ “trick” was to graft observed temperature data from after 1960 onto a line showing temperatures merely inferred from tree rings. If you just reported the tree-ring data straight-up, they would suggest that the earth has cooled since 1960, which conflicts with what we know was happening (assuming there are no biases in the temperature observations, but that’s another battleground several miles away).

In one sense you could argue that this is a “trick” in the innocent meaning of the term, a real answer to a real problem: Jones only meant to “hide” a presumptively nonexistent “decline”. But an ordinary person looking at a graph doesn’t expect the underlying data to be spliced together from two different sources if the point of the graph is to highlight what one source (the tree rings) tells us. Moreover, the divergence between the predictions of the tree-ring model and real post-1960 temperatures is a legitimate problem in paleoclimate reconstruction. (“Some large-scale human influence” on “wood densities”? Oh, hell, what about the fairy hypothesis? Couldn’t woodland sprites have sprinkled magic dust on those trees?)

In “hiding the decline”, Jones was thus proposing to “hide” a weakness in the research itself. IPCC peer reviewers squawked about this “hiding” when it was done in another way, by simply cutting off the data at 1960. As a matter of scientific ethics, Jones’s “trick” sucks. Though it’s still probably not one of the four or five most ethically troubling statements in the leaked CRU e-mails, even considering just the ones made by Jones.

Gorrie could have minimized the offence in dealing with this cherry-picked example of malfeaseance; instead, he handwaved it away completely. But then there’s a lot of handwaving in this column, like the obnoxious complaint that environmental reporters are being asked to “parse e-mails” (which, as described above, he goes on to do in a tendentious, half-hearted way) instead of “focusing on the evidence of human-made climate change”. As if the debate over the CRU e-mails was anything other than an argument about the provenance and quality of the most important body of a posteriori evidence for human-made climate change.

Gorrie also says, sympathetically, that climate scientists “resent having to respond to skeptics.” Well, who the hell doesn’t? That’s like saying that prosecutors resent the threat of having unfairly acquired evidence excluded from the courtroom, or that ballplayers resent the danger of getting picked off first base. They can resent it all they like, but it’s there in the rules of the game, for good reasons. Q: What do you call a scientist who can’t accept criticism from “skeptics”? A: Anything you like, as long as it’s not “scientist”.