Admissions against interest

The polemicist I.F. Stone used to say that the government doesn't have any secrets. It publishes everything it knows, somewhere, you just have to find it. Although that isn't all true, it's true enough that a thousand reporters and lobbyists couldn't mine all the possible sources.

It is less true of private sources but, still, often true. Thus, in all the commotion over the CRUtape letters, a shocking and damning admission by the climate alarmists was published – even promoted – by one of the top Chicken Littles, and nobody seems to have noticed.

I mean Stephen Schneider, who by his own admission is the father of climate modeling. If he isn't doing it right, then suspect maybe nobody is.

I can understand how this got overlooked and underreported. Schneider's autobiography, “Science as a Contact Sport,” is a wretched, silly book; and I've already had my fun with it in a review (See Book Review 109, Dec. 9). Probably Schneider's friends, if he has any, didn't want to draw attention to his goof, if they noticed it; and his foes, of which he has many, probably didn't expect any revelations helpful to them to be delivered as a gift. The book is hard to read and to stomach, both. Still, you never know till you look.

While RtO's jeering review was partly in fun, this next is serious. Really, really serious.

First, from pages 147-148:

“Ben Santer, still at the Lawrence Livermore National Lab, did not want to spend his time making his line-by-line computer program accessible to public perusal. He knew – and obviously so did his attackers – that the programming codes would be virtually useless to any one trying to replicate his results. The beauty of systems science is that we come to conclusions through independent efforts that confirm one another – it's not merely a matter of rerunning someone else's computer code or models. We like independent groups using independent models coded by each separate group to try the same experiments or look at the same data set, and if reasonably conforming, we increase our confidence in the conclusions. Personal codes are so idiosyncratic to the programmers that it could take months to explain them to others who could, in much shorter time, do an independent audit by building their own code using the same equations or data sets. The National Science Foundation has asserted that scientists are not required to present their personal computer codes to peer reviewers and critics, recognizing how much that would inhibit scientific practice.

“A serial abuser of legalistic attacks was Stephen McIntyre, a statistician who had worked in Canada for a mining company. I had had a similar experience with McIntyre when he demanded that Michael Mann and colleagues publish all their computer codes for peer-reviewed papers previously published in Climatic Change. The journal's editorial board supported the view that the replication efforts do not extend to personal computer codes, with all their undocumented subroutines. It's an intellectual property issue as well as a major drain on scientists' productivity, an opinion with which the National Science Foundation concurred, as mentioned.

“Ben Santer was on the verge of quitting the Livermore Lab over this pernicious attack – it would have been a great loss to climate science. He worked for the Department of Energy, and it would have been admirable if the DOE had stepped in and informed McIntyre that his request was unreasonable. DOE, still under the Bush Administration at that time, did not. The Obama Administration, however, appears poised to fight these battles side by side with concerned scientists. Time will tell how it turns out.”

OK, I wanted you to read unadulterated Schneider first. Now, let's fisk.

“Ben Santer, still at the Lawrence Livermore National Lab, did not want to spend his time making his line-by-line computer program accessible to public perusal.

He works for us. Making his work product available to us is part of his job whether he wants to do it or not.

He knew – and obviously so did his attackers – that the programming codes would be virtually useless to any one trying to replicate his results.

When an investigator says his results cannot be replicated, believe him!

The beauty of systems science is that we come to conclusions through independent efforts that confirm one another – it's not merely a matter of rerunning someone else's computer code or models. We like independent groups using independent models coded by each separate group to try the same experiments or look at the same data set, and if reasonably conforming, we increase our confidence in the conclusions.

It's how lemmings get from Point A to Point B. But it isn't the issue. The interest in the codes is not, primarily, to rerun them but to look into them to see whether, in fact, they are capable of doing what they are claimed to do. Also, to detect possible mathematical or programming mistakes that could influence output. (We could here go into the issue of verifying code, and programming managers will know why that is a problem, but let's not, for now.)

Personal codes are so idiosyncratic to the programmers that it could take months to explain them to others who could, in much shorter time, do an independent audit by building their own code using the same equations or data sets. The National Science Foundation has asserted that scientists are not required to present their personal computer codes to peer reviewers and critics, recognizing how much that would inhibit scientific practice.

Auditors audit what someone else has done. Doing it yourself from scratch and getting a different result would suggest a problem, but wouldn't tell what it is. Peer review, if reviewers cannot look at the code, is only skin deep, then. (In climate science, there are also problems with knowing exactly which data sets are being used, but let's not go there now.)

It would not take qualified programmers months to figure out a competently written code. Steve McIntyre, to take one example, has managed to reverse engineer some badly written code within a few days.

“A serial abuser of legalistic attacks was Stephen McIntyre, a statistician who had worked in Canada for a mining company. I had had a similar experience with McIntyre when he demanded that Michael Mann and colleagues publish all their computer codes for peer-reviewed papers previously published in Climatic Change. The journal's editorial board supported the view that the replication efforts do not extend to personal computer codes, with all their undocumented subroutines. It's an intellectual property issue as well as a major drain on scientists' productivity, an opinion with which the National Science Foundation concurred, as mentioned.

“Undocumented subroutines” is the real revelation here. Why aren't they documented? Real scientists keep laboratory notebooks that meticulously record every input. There are hundreds of examples of experiments that gave startling results which, once the lab notebooks were reviewed, turned out to be spurious. But if you don't know what went in, you cannot learn where the project went awry.

Climatic Change is Schneider's own journal. Other journals have a different policy, including Philosophical Transactions of the Royal Society, and it was that journal's insistence on full reporting that pried open the door to the room of climate science's dubious practices.

In any event, replication is replication. In practice, investigators usually do not attempt to reproduce exactly the first report but to repeat it with variations. These test (prove, in the old meaning) the result; but you cannot do that if you don't know what the original investigator did.

“Ben Santer was on the verge of quitting the Livermore Lab over this pernicious attack – it would have been a great loss to climate science. He worked for the Department of Energy, and it would have been admirable if the DOE had stepped in and informed McIntyre that his request was unreasonable. DOE, still under the Bush Administration at that time, did not. The Obama Administration, however, appears poised to fight these battles side by side with concerned scientists. Time will tell how it turns out.”

Politics ain't science. That knife cuts both ways.

The request for Santer's work was made by McIntyre in 2008. In January, McIntyre was told the request was in a “queue” for consideration. Nothing has been heard in the year since.

In late November and early December, shortly after the CRUtape letters were revealed, New Scientist and Nature editorialized that climate scientists were being hampered in their work by repeated requests for their work. Gee, I wonder where they got that idea?

But as we see from Schneider's relation – unlikely to be underwrought – one request to Santer was nearly enough to make him abandon climate science. These people are desperate to prevent the world from seeing what they have done. Even the least suspicious minds should be getting curious about what's really in there by now.

I am willing to hazard a bet: If and when Santer's complete work product is revealed, he will leave climate science whether he wants to or not. As Schneider says, time will tell.