Rearranging The Deckchairs

Frank O'Dwyer's blog

Mann 4, Shollenberger 0

March 11: 4 more here

March 9: A followup post dealing with 4 more of Shollenberger’s claims is here

March 7: Updated to reflect clarification from Michael Mann regarding claim 1 below. Additions are marked and deletions shown with strikethrough. I have also changed the title from Mann 5, Shollenberger 0, to Mann 4, Shollenberger 0, for reasons explained below.

Brandon Shollenberger has posted a 15 page review of Professor Michael Mann’s book The Hockey Stick and the Climate Wars: Dispatches from the front lines.

To his credit he has clearly read the book (I have too), and the whole thing comes over as an intellectually honest attempt to fact check Mann’s claims. Reading Shollenberger’s remarks over at Climate Etc, he himself comes over as intellectually honest also, e.g. unwilling to associate himself with other ‘sceptical’ points made against Mann’s book that he perceives as poor.

Unfortunately many of the claims made in his review are nevertheless flat wrong, with many mistakes that can be put down to simple logic errors or reading comprehension failure on his part.

Other problems in his review include pedantic nitpicking, and hyperbole (“Goebbels” and “big lies” indeed—paging Mr Godwin!), with much of the pedantry turning out to be wrong as well. Often minor differences or no significant differences at all are claimed to be ‘contradictions’. Most of the points which have any validity amount to torturing of issues that imply little more than the possibility that Mann was mistaken about some minor detail of events that make little difference either to Mann’s basic point, or to anything important.

In light of this one despairs to see Judith Curry, a distinguished scientist, write approvingly of his review

Brandon, nice job. I assume the auditors will audit your audit, but your analysis seems to me to be correct.

It’s really hard to see how she could have read the review and arrived at this conclusion.

Shollenberger makes about 19 claims in his review. I address 5 of them here, in no particular order, and will deal with the rest in later instalments as I get time. Meanwhile I will point out in fairness to Shollenberger that his review does have better points than the ones I mention below, but unfortunately that would not be difficult. Though not all of the points made have no validity, the vast majority have little or none, and/or suffer from the other problems I mentioned above.

1. Incorrect claim that Mann contradicts himself re CO2 doubling

Shollenberger claims that Mann contradicts himself re temperature increase from doubling CO2. This claim begins with Shollenberger noting the following:

For example, if Mann says a doubling of CO2 levels “would lead to an additional warming of anywhere between 1.5 and 4.5°C (roughly 3-8°F)”, it should be easy to notice something is wrong on page 18 when he says (“of” added where it was missing):

There was increasing recognition by the mid-1990s that another 2°C (3.5°F) [of] warming beyond current levels (for a total of 3°C or 5°F warming relative to preindustrial times) could represent a serious threat to our welfare. Precisely what limitations in global greenhouse gas emissions would be required to avoid that amount of warming remained uncertain, and still does, because of the spread of predictions among models. If we choose to take the midrange model estimates as a best guess, avoiding another 2°C of warming would require stabilizing atmospheric CO2 concentrations at no higher than about 450 parts per million (ppm). Preindustrial levels were about 280 ppm…

and Shollenberger then says:

The midrange of 1.5 and 4.5°C is 3°C. This means Mann said there would be a total increase of 3°C with CO2 levels having only gone up by 60%. Something is obviously wrong.

Shollenberger is right that something is obviously wrong, but unfortunately the something that is obviously wrong is Shollenberger’s understanding of the word ‘avoid’. Mann did not say there would be a total increase of 3C. Mann proposed 450ppm as a means to avoid a total of 3C of warming, relative to preindustrial times. If you avoid 3C of warming, that means there wouldn’t be a total increase of 3C, not that there would be. Since Mann is trying to work out how to avoid an increase of 3C, not how to achieve one, it should be blatantly obvious that what he is saying here is that 450ppm would (probably) yield a total increase of less than 3C.

Furthermore 450ppm is not a doubling from 280ppm, as Shollenberger correctly states, but the midrange estimate (3C) he uses is for a doubling of CO2, so Shollenberger’s second sentence simply doesn’t follow from his first. As Shollenberger clearly understands, an estimate for CO2 doubling of 3C, implies that the total increase from 450ppm would be less than 3C—and that (again) is what avoiding another 2C of warming logically entails. Far from being ‘something wrong’ with Mann’s claim, this just underscores that Shollenberger has misread it.

Given all that, it is a mystery as to how Shollenberger has managed to decide that Mann is saying there would be a total increase of 3C at 450ppm. The only time Mann mentions a total increase of 3C is to mention that this could represent a serious threat to our welfare. Hence he sets out what we’d need to do in order to avoid that, i.e. he asks the question what we’d need to do to see less warming than 3C, and his answer is not to exceed 450ppm. That in no way contradicts “The midrange of 1.5 and 4.5C”, i.e. 3C as a best guess for warming from doubling CO2 to 560ppm.

It turns out Mann added an increase of 2°C to the already observed warming when he should have added it to the temperature of 200 years ago. Instead of “another 2°C” and “a total of 3°C,” Mann should have said “another 1°C” and “a total of 2°C.”

No, it turns out that Shollenberger had misread the paragraph he was quoting and attributed a statement to Mann which Mann never made. Mann should have said what he did say.

Update March 7: Michael Mann has clarified this on twitter:

So, it turns out what Mann meant to describe was the EU target of 2C total warming, rather than a goal of avoiding 3C total warming. That is why it should read “for a total of 2C warming”. So while Shollenberger’s conclusion just above turns out to be correct, his reason turns out to be wrong. He also remains wrong in claiming that Mann said that 450ppm means 3C warming, because it’s still obviously true you don’t avoid a warming of 3C by warming 3C (though oddly enough, a similar construction would work in terms of how the EU target is expressed, because in that case 450ppm is offered as a 50% chance of avoiding 2C!). And it still remains clear in the context of the book as a whole and now from the twitter conversation that Mann was saying that 450ppm implies about 2C warming as a best guess/median projection.

Despite the error on the target, there is simply no reasonable reading of any of that or the original text that gives you ‘This means Mann said there would be a total increase of 3°C with CO2 levels having only gone up by 60%’, as Shollenberger claimed. When I look at it again in hindsight, I see that one thing you could have queried in the original context is why the phrase ‘no higher than’ was present–after all there are levels above 450ppm that would also (probably) avoid a total warming of 3C, so why stop 1C short rather than, say, at 2.5C? Caution? You could also have pointed to the EU target and said that the 3C goal seems too high, it should have been a target of 2C (which it turns out is what Mann actually meant). But none of that was actually said, and none of it would have been enough to justify the claims of contradiction that Shollenberger actually did make. The worst he could have claimed is that ‘no higher than’ was unsupported/unexplained–but he didn’t say that either. In any case you couldn’t have said (and still cannot say) that a warming of 2C doesn’t avoid a warming of 3C, nor that it means a warming of 3C. And that is what Shollenberger did say.

This is confirmed on page 250 where Mann says:

When we reach concentrations of 450 ppm (about 2030, extrapolating from current trends), we will likely have locked in at least 2°C (3.5°F) warming of the climate relative to preindustrial levels…

The only thing this confirms is that Shollenberger is mistaken. Mann shows here that he knows fine well that 450ppm means a best guess of about 2C of warming, which is why he mentioned it still wouldn’t be wrong to offer 450ppm as a way to avoid a total rise of 3C, not to achieve it, as Shollenberger claimed Mann said. There is no contradiction at all between this statement and the earlier one.

That the first set of numbers was nonsense should be obvious. It contradicts a later statement by Mann, and it contradicts basic arithmetic. It’s a mystery how such an obvious error could slip past an editor, but it’s certainly something any reader should be able to spot.

Unfortunately the only nonsense and obvious basic errors here are from Shollenberger. Anyone can make mistakes, but Judith Curry should know better than this, and should be embarrassed to have posted that she finds Shollenberger’s analysis correct here.

Michael Mann: 10, Brandon Shollenberger: 0

2. Incorrect claim that MM03 said PCA manufactured hockey sticks from noise

Shollenberger claims that Mann is wrong to say that MM05 included a new assertion regarding hockey sticks from pure noise.

Mann goes on to discuss a later work by McIntyre and McKitrick, saying (on page 130):

McIntyre and McKitrick had quietly dropped their erroneous original assertion (in their 2003 paper discussed in chapter 8) that the hockey stick was an artifact of bad data. Their new, albeit equally erroneous, assertion was that the hockey stick was an artifact of the conventions used in applying principal component analysis (PCA) to certain tree ring networks, which, they argued, “manufactured Hockey Sticks” even from pure noise.

Mann clearly portrays the two papers as showing a change in argument. In reality, the assertion in their later paper was not “new.” It was the same as in their 2003 paper, a point evidenced by Mann’s own note #45.

Once again Shollenberger’s claim here is flat wrong and Mann is right. Search MM03 for the word “noise” and you will find nothing, because there is no mention of the specific claim Mann mentions. Do the same search for “noise”on MM05 and you will find that it is discussed prominently, because the claim that MBH98 produces hockey sticks even from noise is a new claim introduced in that paper.

The fact that the first paper claimed that PCA calculation mistakes were made and the new claim also ostensibly relates to the same PCA mistakes doesn’t change this. While MM03 did claim that a hockey stick from at least one specific proxy series was an artefact of mistaken PCA calculations, nowhere in that paper do they say the method meant you would get one even from pure noise. That was a new argument introduced in MM05, just as Mann said it was.

Michael Mann: 1, Brandon Shollenberger: 0

3. Incorrect claim regarding Wegman testimony to congress

Shollenberger says this is the most serious misrepresentation from Mann and that it amounts to libel. Let’s see. Here Shollenberger starts to make his case (my emphasis):

The claim is basically that the Wegman Report repeats McIntyre’s work and conclusions without due consideration. Mann doesn’t spend much time on this criticism in his book, but what he says is very important:

Not only had there apparently been substantial undisclosed collaboration between the WR authors and Stephen McIntyre, as hinted at earlier–something Wegman had denied in his testimony under oath in Congress…

Mann claims Wegman denied something, under oath, that was true.

Actually he only said it looks like (‘apparently’) the something which Wegman denied (substantial collaboration) was true (and in the book Mann elaborates on why there appeared to be substantial collaboration). Mann then backs up the fact that this was undisclosed in Note 66, which Shollenberger quotes in full:

See http://ftp.resource.org/gpo.gov/hearings/109h/31362.txt, specifically the following exchange between Rep. Stupak and Wegman:

Mr. Stupak: Did you or your co-authors contact Mr. McIntyre and get his help in replicating his work?

Dr. Wegman. Actually, no…

Shollenberger’s main claim here is that Mann is guilty of misrepresenting that this is the whole exchange. However he clearly isn’t. Mann clearly invites the reader to go look at the link, and even tells you what to look for - there is even a clickable link in the Kindle version. Shollenberger goes on to present the indicated section of the transcript and suggests that simply because the whole exhange was ‘lengthy’ and not included inline in Note 66, somehow this means it contradicts what Mann said, and that Mann was lying by omission.

Not only does this not follow, it’s not true either. Nowhere in the exchange does Wegman disclose any substantial collaboration, and his remarks do not suggest substantial collaboration and replication help from McIntyre. That presumably is why Wegman’s remarks start with ‘Actually, no’ and not ‘Why, yes!’—and despite Shollenberger’s claims, that fact is significant. The remainder of Wegman’s remarks elaborate on this to create the impression he is giving a full disclosure and that any help with replication from McIntyre was minor. However Mann is saying there appears to have been substantial collaboration. If there was, then that was not disclosed.

Yes, if that’s true it would have serious consequences for Wegman, as Shollenberger notes, and he’s hardly the first to notice that if Wegman misled congress that’s pretty serious. But that still doesn’t mean Wegman disclosed substantial collaboration, and Mann did not misrepresent anything when he said that he didn’t. Nor did he misrepresent Wegman by omission, because nothing in the full transcript contradicts Mann’s point that if it happened (as he says it appears) Wegman didn’t disclose it. Last but not least, Mann invites readers to use the link and see the whole exchange, and tells you where it begins.

Michael Mann: 2 Brandon Shollenberger: 0

4. Incorrect claim that Mann misrepresented Singer’s letter to Science

Fortunately, the first simple contradiction is found on page three:

In February 1996, for example, S. Fred Singer, the founder of the Science and Environmental Projection Project and a recipient over the years of substantial fossil fuel funding, published a letter attacking Santer in the journal Science. Singer disputed the IPCC finding that model predictions matched the observed warming and claimed–wrongly–that the observations showed cooling.

Singer criticized Ben Santer’s article. He didn’t “attack” Santer.

Shollenberger makes this remark only in passing, but really? To claim that Santer did not mention purported information that utterly undermined his case, and to make this claim in a public letter to a science journal, is a serious slur on his scientific reputation, his competence and/or his integrity. If that is not an attack, it will do until one comes along.

That alone should raise eyebrows, but it hardly compares to the fact Mann misrepresented Singer’s letter. That letter said:

The summary (correctly) reports that climate has warmed by 0.3° to 0.6°C in the last 100 years, but does not mention that there has been little warming if any (depending on whose compilation is used) in the last 50 years, during which time some 80% of greenhouse gases were added to the atmosphere.

Singer clearly acknowledged warming had been observed. You would never have guessed this from Mann’s description of his letter. How does Mann explain this discrepancy? Mann ignores that part of Singer’s letter and acts as though another part is all that exists:

The summary does not mention that the satellite data–the only true global measurements, available since 1979–show no warming at all, but actually a slight cooling, although this is compatible with a zero trend.

Singer says one set of observations shows cooling (which he mentions is statistically insignificant). He mentions other observations show warming. Mann portrays this as him saying observations only show cooling. This is a tame example

Nope, it’s not an example at all. Nowhere does Mann say that Singer said observations only show cooling, any more than the fact that Shollenberger wrote a 15 page review of Mann’s book means that he misrepresented Mann’s book as being less than 15 pages long, or the fact that Shollenberger uses quotations from Mann’s book means that he is misrepresenting that this is all that Mann wrote in his book. Mann simply states that Singer said observations showed cooling, and he’s right: Singer did say that. It’s completely obvious which claim Mann is referring to, and the fact that Singer said other things too doesn’t mean he didn’t make it. This again is simply reading comprehension failure on Shollenberger’s part.

Michael Mann:3 Brandon Shollenberger: 0

5. Incorrect claim that Mann contradicts himself re MM03

Mann contradicts his sources. Mann contradicts himself. It is hardly surprising he would do both at the same time. On page 123, he says:

The central claim of the McIntyre and McKitrick paper, that the hockey stick was an artifact of bad data, was readily refuted.

To understand Mann’s misrepresentation here, there is no need to understand any technical details. All you need to do is compare a few simple sentences. First, compare the above sentence with a quote from the abstract of the paper he discusses (emphasis added):

The particular “hockey stick” shape derived in the MBH98 proxy construction – a temperature index that decreases slightly between the early 15th century and early 20th century and then increases dramatically up to 1980 — is primarily an artefact of poor data handling, obsolete data and incorrect calculation of principal components.

Even though McIntyre and McKitrick’s conclusions refer to “poor data handling, obsolete data and incorrect calculation of principal components,” Mann claims their argument dealt solely with “bad data.”

No he does not. He says it is the central claim, not the only claim.

And it sure looks like the central claim. No need to take my word for it, here it is in the introduction to the paper itself, right after a long list of the things they did (mostly concerning the data, and including a one liner about PCA):

We find that the particular “hockey stick” shape derived by MBH98 is primarily an artefact of poor data handling and use of obsolete proxy records.

I haven’t truncated that sentence. (Sure, they say different in the abstract. Perhaps then Shollenberger will write a review accusing McIntyre and McKitrick of contradicting themselves—after all, they said the hockey stick shape was primarily an artefact of two things and also primarily an artefact of three things, all in the same paper). You can also look at the entire paper, and you will see that the majority of the discussion concerns data.

In any case Shollenberger is wrong to claim there is a contradiction here regardless of whether he agrees that this was the central claim of MM03. There is simply no inherent contradiction in saying that X is the central claim of an paper that also makes Y and Z claims. (I’m not sure why Shollenberger has such trouble with this kind of reasoning—it’s very similar to his mistake regarding Singer above.)

Michael Mann:4 Brandon Shollenberger: 0