Nice piece, Mike. And, it deals a ‘death blow’ only to those who choose to listen or to those who have the time to digest your words. To the ideologues and rubes who believe them, you are [insert pejorative here] and you’ll never change those minds.
I deal with politicians often in my work. What would be helpful in analyses like these is a short wrap-up for decision-makers; they do not have the time to wade thru this analysis – nor do their aides – and I suspect the effectiveness of your message may be lost on some.
For decision-makers to have a chance to make good policy when faced with the emotional agitprop of the indy-fundeds, they need to be able to access the material. Simple, cogent bullets somewhere (best at the beginning) will allow decision-makers to scan your work and get the rational message contained within.
All of this technical, statistical jargon is over my head, but I get the impression that the data on which the climate reconstruction is based is so sparse and uncertain that you can’t draw any firm conclusions supporting either MM’s or Mann’s side of the debate.
[Response: Even without technical training or a statistical background, you should have an adequate basis for discerning which of the two parties is likely wrong here. Only one of the parties involved has (1) had their claims fail scientific peer-review, (2) produced a reconstruction that is completely at odds with all other existing estimates (note that there is no sign of the anomalous 15th century warmth claimed by MM in any of the roughly dozen other model and proxy-based estimates shown here), and (3) been established to have made egregious elementary errors in other published work that render the work thoroughly invalid. These observations would seem quite telling. -mike]
The question that I wish someone would address is “What difference does it make?”. CO2 is reaching levels that haven’t existed for hundreds of millennia, and it seems to me that on that time scale the climate a few hundred years ago provides no better basis for extrapolating future climate than does yesterday’s weather. Can any conclusion regarding future climate change be drawn from the “hockey stick” reconstruction, and can any level of statistical confidence be placed on the conclusion?
[Response: Refer to this post (and references therein) for a detailed discussion of how comparisons of proxy-based climate reconstructions with theoretical climate model simulations can inform our assessment of the role of both natural and anthropogenic factors in recent climate change.
In considering the issue of earth warming, I believe that the question of “Is it or is it not occurring?” is irrelevant. The more important question is, it seems to me, “How should we behave?”
In this, I take my thinking from Pascal: There are two “realities” and two “behaviors”, paired to give us four scenarios:
1) We behave as though earth-warming is not occuring and it isn’t.
2) We behave as though earth-warming is not occuring and it is.
3) We behave as though earth-warming is occuring and it isn’t.
4) We behave as though earth-warming is occuring and it is.
If one thinks through the implications of each of these four scenarios, our course of action seems clear. What’s the problem???
Comment by Bill Hemmer (not on CNN) — 9 Jan 2005 @ 5:00 PM
According to lay people that I talk to who have been influenced by the op-ed pieces that you mention, “global warming has been disproved”. The MM story is indicative of a pattern in which industry (and now our own government) PR machines latch on to minority scientific articles to claim that an environmental issue has no basis. The journal Science (in the late ’70s or early 80′s) once published an article in which the author claimed that the major components of acid rain were weak acids. The article should have failed peer review and never been published: the scientist conducting the work titrated the samples in open air, effectively measuring not only the weak acids in the samples but also the carbon dioxide from the room. The work was plain wrong. Nevertheless representatives of the power companies parroted the “findings” for several years to claim that acid rain was NOT related to industrial air pollution.
Great Piece Mike. If you are familiar with Tim Lambert, he has spent much time discussing/discrediting MM.
As an aside, are you familiar with Positive Matrix Factorization (PMF)? We are using it in lieu of PCA for samples we analyzed from a dust storm experiment in Asia. It has several major advantages over PCA including that it doesn’t produce negative (non-real) results and you can incorporate uncertainty into the analysis so you can limit significance of low-level or missing data.
Keep it up, guys. I have a brother-in-law who is a Fox News/Talk Radio junky and he insists global warming is a myth yet has never read or seen a scientific article. Truly unbelievable.
[Response: Thanks very much for your comment Scott. Indeed, there are other statistical approaches, as you note, to the problem of 'climate field reconstruction', many of which are somewhat more sophisticated than (and arguably preferable to) PCA-based approaches. In this previous post here, we discuss the results from a recent paper in press in Journal of Climate by Rutherford et al that uses the Regularized Expectation-Maximization ("RegEM") algorithm to reconstruct past temperature patterns from proxy data--the results are remarkably consistent with past proxy-based reconstructions using other (e.g. PCA-based) methods. Applications to reconstructions of patterns of past continental drought from tree-ring data can be found here. RegEM is an iterative approach for estimating the data covariance of an incomplete data set (and imputing missing values in the process) using what is sometimes referred to as "ridge regression". In this approach, the main diagonal of the estimated data covariance matrix is inflated using a "smoothing" parameter determined by generalized cross validation (GCV) to insure that the estimated data covariance matrix is not underdetermined in the presence of noise and incomplete information. Estimated missing values are computed based on the smoothed data covariance estimate, and the process is continued until a stable solution is obtained. I believe the method may have some features in common with the alternative approach you describe. A fuller discussion of the methodology in specific, including a discussion of related past work, and of the more general issue of "climate field reconstruction" is provided in the references mentioned above.]
Comment by Scott Robertson — 10 Jan 2005 @ 12:40 PM
Actually, Mike, I’d be more interested in your views on the criticisms presented by Hans von Storch. I believe he claims that the graph exhibits much less variability than the natural climate.
von Storch et al purport to test statistical methods used to reconstruct past climate patterns from "noisy" proxy data by constructing false proxy records ( "pseudoproxy" records) based on adding noise to model gridbox temperature series taken from a climate simulation forced with estimated past radiative forcing changes. Several other researchers have published similar studies in the past. One thing that makes the von Storch analysis different is that they use a simulation that exhibits larger forcing, and more variability than most simulations, the very same "GKSS" simulation (Gonzalez-Rouco et al, 2003) discussed here. In their commentary on the paper, Briffa and Osborn note that the results may not generalize to the actual world, where the forced variability may be much smaller than in the GKSS simulation. While von Storch et al focus on the Mann et al (1998) reconstruction method, they argue that their results generalize to other proxy reconstructions methods as well.
The von Storch paper has appeared too recently for responses to have made their way through the appropriate peer-review process. While we are already aware of some recent work arriving at very different conclusions from von Storch et al, it is premature at this time to comment on that work. We hope to be able to comment more fully on the matter when this other work has appeared in the peer-reviewed literature. Respecting the peer-review process, however, we prefer not to post any further comments on this matter until it has played out more fully in the peer-reviewed literature -mike]
I don’t agree with your decision chart for the same reason that I don’t agree with Pascal’s original wager: there are not only two realities and two behaviors. Climate can occur in many degrees, have many causes, and follow many paths. Behavior is ridiculously variable. In this case, I don’t think boiling all the alternatives down to a two-by-two grid will yield anything like a useful approximation of the real world.