People have been repeating over and over again that “scientists believe that anthropogenic global warming is real.”

I know nobody asked this scientist, and that if I did get a cold call asking for a yes/no on the issue, I would probably just hang up. It’s a complicated question, and I don’t think it’s going to be easy to get over 95% certainty in anthropogenic (CO2-driven) global warming. I don’t think it’s that easy to speak for all scientists, and I’m not sure when it would be a good idea. Almost all scientists are specialists, so some could be very mediocre in their appreciation of climate modeling.

It’s not that lower degrees of confidence would stop me from doing what’s prudent. When the issue becomes politicized, and policies directed towards reducing CO2 levels become a politically accessible outcome, then it’s easier for me to act on my approximate grasp of things. But I’m not a policy writer just yet, so I don’t need to discuss California’s Proposition 23, for example.

I’ve been doing some reading on climate change models, and climate change skepticism. I’ve read some of the documents released by IPCC and I found them pretty readable and credible.

I’ve read some of scientific criticisms too, probably more of those by volume. Here’s one of them by Lord Monckton of Brenchley, published (without peer review) by the American Physical Society (a group of real scientists). I have to admit, some of the arguments confuse me, some of them snow me, some of them could be right.

I think I’ll dissect Monckton’s paper more completely in other posts (the paper’s doppleganger, in the same issue, seems worth reading too). Pretty much it seems weak right from the start; the words ‘temperature’ and ‘phase transition’ are used way too loosely in Figures 1 and 2, for example. But at one point Monckton launches this turkey, which simply begs for a carving:

Here as elsewhere the IPCC assigns a 90% confidence interval to “very likely”, rather than the customary 95% (two standard deviations). There is no good statistical basis for any such quantification, for the object to which it is applied is, in the formal sense, chaotic. The climate is “a complex, non-linear, chaotic object” that defies long-run prediction of its future states (IPCC, 2001), unless the initial state of its millions of variables is known to a precision that is in practice unattainable, as Lorenz (1963; and see Giorgi, 2005) concluded in the celebrated paper that founded chaos theory – [technically in-context quote by Lorenz follows].

First Monckton complains about their confidence intervals, saying that they use too-low standards for significance.’Standard deviations’ in this sense refers to confidence tests based on the normal distribution. I don’t think this “customary 95%” is relevant for the IPCC model. There is, after all, only one future climate, not some randomized sampling distribution of future climates. The probabilities should come from information theory, not a theory of random distributions. Squashing a climate model into an orthodox statistical confidence test is wrong.

Perhaps Monckton is would agree with that last point, because he says that “any such quantification” is inappropriate because the climate is “chaotic in a formal sense.” (Formally, chaos is not the same thing as randomness.) He doesn’t justify his sudden invocation of formality, but instead switches to a discussion of the impossibility of accurately forecasting weather more than a few days into the future (a celebrated result due to Lorenz). He’s right; Lorenz did say that. (I’ve tried to read that Lorentz paper before but I definitely didn’t understand it.)

But weather is not the same thing as climate. Take for example a pot of boiling water. Can scientists predict specifically where the little boiling bubbles appear and how they move around? No, that’s a (formally) chaotic process, and it’s going to require unreasonably accurate measurements of the water immediately before it boils to model it exactly. But if we get some data (heat capacities mostly), it’s easy to accurately predict (guess with a correct estimate of the error in the guess) how long the water will heat before the whole pot boils. The whole-pot outlook is climate modeling; predicting bubbles is weather modeling.

As I do think Lord Monckton understands the difference between climate and weather, I assume he’s deliberately obfuscating things here. His rhetorical strategy seems to hinge on providing an enormous quantity of bad arguments, like scatter shot, even if I allow the possibility of some good criticisms along the way. Suspicion of bad faith (dishonest intentions) is a hefty scientific criticism. And that’s not even the worst part.

The worst part of Monckton’s invocation of chaos, and conflation of climate and weather, is that he spends the remainder of the very same paper offering a correction to the IPCC climate model! Apparently he thinks the whole endeavor is a fool’s errand, but he’s going to still going to correct the predicted values…pretty discordant behavior, actually.

I lack the means to think, in a disciplined way, about multiscale phenomena, which means I can’t really address Monckton’s substantial points, or say more interesting things about (formal) chaos. It’s obvious that the IPCC models do not predict a large temperature increase directly from another doubling in atmospheric CO2, but invoke several feedback mechanisms. By focusing on uncertainties in the IPCC’s feedback mechanisms, Mockton may have a valid criticism. Too bad it’s mixed-in with so much rhetorical flailing.

Ryan MB Hoffman has a B.Sc. in Biochemistry from Queen’s University in Kingston, Ontario, and a Ph.D. in Biochemistry from the University of Alberta in Edmonton, Alberta. He is mostly interested in how protein molecules fluctuate throughout their functional processes. During his doctoral work he studied troponin, which is a switch that regulates striated muscle contraction. His thesis is called Process and Troponin, and can be obtained here. He works at the University of California, San Diego, at the Center for Theoretical Biological Physics. He is active with the Intrinsically Disordered Proteins subgroup of the Biophysical Society. Ryan likes to remind people that his contributions to TRN are performed entirely using his personal resources.