tdaxp

"Shrinking polar icecaps (and credibility)," by Gaijinbiker, Riding Sun, http://ridingsun.blogspot.com/2005/01/shrinking-polar-icecaps-and.html, 31 January 2005.

"You're exactly right," by Dan, Riding Sun, http://ridingsun.blogspot.com/2005/01/shrinking-polar-icecaps-and.html#110718078243455046, 31 January 2005.

A new report on global warming is out. I'm not a climate scientist, and people who I respect are on boths sides of the debate -- often in surprising way. But the article is a joke.

Gaijinbiker explains:

When you measure the same thing twice, you don't expect the second result to be double the first. If it is, that's a clue that your measurements are worthless. If the second try is 100% higher, perhaps a third try would yield results 100% lower -- that is, zero.

My response?

You're exactly right.

For my graduate degree (Computer Science) I had to build a model simulation large-scale systems. Results had to be consistent with itself and the real world. 100% variation is a failure -- or more academically, a field for future research -- not a conclusion.

But it doesn't end there. In the original post, Gaijinbiker points out another problem

Also notable is that the article mentions only the report's "worst-case" scenario. How likely is that scenario to occur? Ten percent? One percent? .00001 percent? And what are the other scenarios like? How likely are they? Are there any where the earth actually gets cooler?

It would be nice to know.

However, Mr. Connor apparently sees his purpose as terrifying Britons into immediate and unwarranted action, rather than skeptically assessing the most drastic outcome of a single new study.

Again, exactly right.

I'm not an expert on simulation design or criticism, but the guys on my committee where. If I would have presented, as my results, the most extreme outcome I'd be laughed out of the room, if not asked to leave the university.

The results of a simulation are taken. They are explained, and areas that seem particularly weak or interesting become "future research." I was lucky to complete my studies under very experienced and knowldgeable professors. They taught me the pitfalls of simulation building, and how a simulation can be perverted for personal or political gain (one had been contracted to simulate how to "win" a thermonuclear exchange using equipment from only one military contractor --- single-source the apocalypse!).

At best, the press coverage of this is biased and inaccurate. Alternatively, the scientists involved are shockingly unprofessional. At worst, they are academically dishonest.