The review of referee #2 for Bürger and Cubasch’s article in Climates of the Past is posted up here . From the style, can anyone doubt that the anonymous reviewer was Mann himself? Take a read.
The reviewer makes reference to this being the "2nd attempt by the authors to publish this flawed manuscript" – it had been rejected by GRL. I presume that referee #2 must have participated in the GRL process.
I think that there are many ways that the Bürger and Cubasch article could have been strengthened and one would hope that a reviewer would do so. In terms of comments that would be useful to a writer, TCO’s comments posted up here would have been far more useful than those of referee #2.
Also, not that I have a great deal of experience with reviewer comments, but my own experience has been that reviewer comments are usually a paragraph or two. I don’t object to the length of "Mann’s" review, but merely point out that such detail is atypical. For example, here is the GRL review of the Wahl and Ammann Comment and our Reply:
This paper should not be published in GRL. The exchanges between the present authors and everyone who has criticized their earlier paper have gone on too long. The point has been made over and over. The fact seems to be that there are choices to be made in this kind of analysis but the results seem to not be too sensitive to them when they are carried out in good faith (exactly what it means to be robust is of course a matter of judgement, but I have some experience on this and I think it is). The current authors have had their say in numerous other replies to criticisms of their work. It is time to get on with it!
While there were many reasons to reject the Wahl and Ammann submission to GRL, the reviewer’s comments were simply irrelevant to them. Also the reviewer brought up an unrelated issue. In this case, Famiglietti’s cover letter agreed that the review was "brief", but Famiglietti claimed to have had "significant communication" with the reviewer. He also apologized for delays because "it was very difficult to find a qualified and willing reviewer".
Now think back also to a comment by von Storch about how hard it used to be to get articles published on millennial paleoclimate methods, a comment where he generously credited us with opening the door and being lucky enough not to have had Mann as the reviewer for our article.
So here we see that the reviewer for Climates of the Past is probably Mann. It also seems that he was a reviewer for GRL. Are "Mann’s" review comments the sort of thing that should decide whether to publish or not publish – or are they things that should be dealt with in a Reply?
Update: I have reason to believe that Mann may not have been a GRL reviewer, although the GRL reviewer seems to have been on the Hockey Team. So how did Mann know that the Bürger and Cubasch paper had been rejected by GRL? Did the GRL reviewer – perhaps Schmidt – break confidentiality and inform Mann? Will GRL admonish Schmidt or whoever for breaking confidentiality?
Climate Audit 
156 Comments
A review as long as the original paper?
The references cited and the obvious omissions probally hold clues as to the author.
Aside from that, there are stylistic flourishes that are distinctly Mannian, both in terms of layout and even vocabulary.
I’m just starting to read the review. I agree that it has too much of a reply type content and that it fails to address the fundamental problems of poor writing and organization in the paper.
I don’t think that length per se is inappropriate. I write lengthe reviews. Because that’s how I read papers. What’s innappropriate is some of the rhetorical games like the the we’ve moved on argument. The comment about ascribing a flavor (detrending) as a practice when the flavor is switched on and off. is very nasty and really at core dishonest (maybe the guy is so twisted that he thinks this way, but if so he’s not a scientist). Also, I get a little tired of the ooh someone made a mistake and that invalidates their larger argument (it’s some form of logical fallacy). This review is really juvenile. Much more adult to say what factors are wrong in the analysis and then how that changes the larger picture of inferences when those factors are changed.
That said, when you (one) writes a poorly organized paper, it is much harder to combat these types of silly rhetorical argument given that the paper itself is quite hard and unpleasant to read.
I don’t think that it’s inappropriate to have Mann as a reveiwer. He obviously knows a lot about the subject. They asked you to review a paper related to you remember? A good editor, though would have also had other reviewers and abstracted significant criticisms from reply type arguments or rhetorical arguments. Note that the editor here DID publish the paper, so no big hardm was done.
All that said, I would be a little leary of publishing in this journal, Steve. They obviously do not enforce clear writing standards and that is one thing that I want a journal to enforce on you! 🙂
I don’t have any problem with length of a review. Most are too short in my opinion. Journal peer review is not a very severe form of due diligence.
I don’t disagree with your beefs on style. (But jeez, I don’t know why you beef so much about my style ( which is far from perfect), when surely our articles are models of clarity relative to other authors in the field.)
The issue with Mann’s review is that he is advoating his own POV; he’s not reviewing the article.
I think the comments about the Amman and Wahl review are not well tied to the present topic (this review). I would not expect perfection in reviews, Steve. Fami rejected the paper for whatever reason. He’s got several reviews to look at. He’s got your reply. He’s got your letter. At the end of the day, getting hung for a murder you did not commit is not something to complain about if you committed another.
There are pro’s and con’s to having an adversary as a reviewer. I’m not suggesting that it’s fine if I’m the adversary and not fine if Mann’s the adversary.
An editor has to be careful as to where the adversary is arguing his POV and where his intimate knowledge of the area is appropriate for a reviewer. It’s a tradeoff.
Mann falls into POV. For example, look at his consideration of RE significance. He cites Huybers criticism of us, but not our Reply (which actually completely dealt with Huybers’ criticism.)
Steve, I’m sorry if I seem all over you. There are many good things that you do and I’m very proud of you. My critical comments are just critical on these topics. Are not meant to in any way say that you don’t have a right to do work here. I am VERY capable of disaggregating issues. I don’t agree with it when Mann tries to imply that a single flaw invalidates a whole and I don’t like it when you buy into that same type of thinking (defensively or offensively). I just call what I see.
I’m a poor writer also, as you can see by my comments. But if you want to win an argument and change a field’s practices, is it too much to ask that you explain things clearly? Given that the field had accepted a past practice that was wrong, it would seem obvious to me that combining an intricate argument in opposition with poorstructure is not the way to go to change things (or better yet, to drive understanding and insight.)
My bigger kvetch is with failures of logical analysis, Steve: conflating two faults and discussing each as a fault, but only doing one test of them both jointly. I think that if you write clear papers, it will be one thing that will help to reduce this tendancy. That’s why I liked the Huybers comment where he shows the equation and clarifies that covariance matrix is two degrees of Kevin Bacon from the Mann off-centered work and that correlation matrix is only one degree of Kevin Bacon away. I actually think that if you force yourself to write clearly, that it will benefit your investigations.
Steve,
Welcome to the world of scientific publishing! Reminds me of how much time I wasted trying to figure out who was that nasty referee who would try to systematically block a seemingly innocuous paper. Then sending a revised version only to get the same irrelevant comments again… then sending again to another journal etc. etc. Then you look at all your colleagues/competitors with suspicion when you meet them at conferences. Friends or foes?
Yes what a waste of valuable time, and that’s one of the things that led me to make a move from academia to industry.
But there are alternatives now. Take a look at the new journal soon to be launched by the Public Library of Science, PLoS ONE. Not only is it open access, but every paper is published without review, and the review takes place AFTER. So it’s up to the reader to make up his/her mind if the paper makes sense or not, and that’s the way it should be. The review is only useful to outsiders to the field, and helps them figure out what paper is more significant. That’s the real future of publishing.
Francois
#6. Reviews are what they are. All I try to communicate to academics who puff “peer review” is that, as practiced by journals, it is a very limited form of due diligence and not an audit. Auditors of financial statements would not make irrelevant and incorrect statements.
Why should Mann be a reviewer of an article on statistical procedures when he admitted in public before the NAS Panel that he is not a statistician?
I object to the idea that people whose work is being investigated/criticized should be anonymous reviewers. I think its unethical. At least in politics and in business, they know what “conflict of interest” means and why its a bad idea to allow it to happen.
The tiny world of climate science review is dominated by the same small group of people over and over again. This is the central failure of peer review in climate science.
#11. Francois, I often feel like an anthropologist since I view such procedures not as written in stone but as customs. My main beef is not so much with what peer reviewers do, but at the subsequent reliance on “peer review” as some sort of talisman of truth.
And all this wailing and gnashing and the paper was published over that review! Sheesh.
Steve, believe me, beleive me. If you are trying to write a paper that corrects someone famous and even takes him to task a bit, you need to write a flawless paper. It can’t have language mistakes. Can’t be awkward. Can’t be in any way hiding unresolved issues. You need to lay everything out brutally honestly. Even the points that argue against what you are trying to say (or which minimize an effect). Write that kind of paper and you can get it published. Not a doubt in my mind. Might have one issue at a certain journal. But you can get it into a quality journal. The structure is not that rotten. Oh…and that kind of paper does the reader and the feild a real service!
Wikipedia has already updated Michael Mann’s info page to reflect the NAS report and it links to the topic about the report on RC as “Mann’s response” So I am reading RC and it says (in the critique of the NAS report) :
“that an independent study not cited, but published well before the report was drafted, comes to very different conclusions. This reflects one of a number of inevitable minor holes in this quickly prepared report”
so I click on the link provided, and it takes me to Mann’s list of publications.
The one he refers to is called “Testing the Fidelity of Climate Reconstruction Methods” Published: © 2005,
Journal of Climate, vol. 18, pp. 4097-4107.
So I go to the website for the “Journel of Climate”
and surprise! find “Gavin Schmidt” as an associate editor.
I actually agree with the broad parts of Mann’ (if it is him) review in the Some Other Problems. I think it would be much better to more tightly scope the paper to what it is about. Sure that leaves you open to the countercharge that you are just writing about one small area. But so what? You ARE.
TCo, he’s not wrong about everything. If I were reviewing the paper constructively, I’d agree with most of your comments.
That’s not the point. Most of Mann’s reveiw is merely POV – he cites Mann et al, Rutherford et al, Ammann .
Can someone check whether http://holocene.meteo.psu.edu/ is down? Otherwise, they’ve now blocked me from his site as well.
Yeah…I agree with that. maybe the editor did too, since the paper got published! Too bad, that he didn’t fix the valid points on theme and conclusion though…
Looks like it’s down, Steve.
Other sites at Penn State are down including http://www.met.psu.edu (which is the server which holds Michael Mann’s bio and term papers.
re: #16
And just what are the “broad parts” of “Mann’s” review? As far as I could tell he was just shouting “it’s not so” and then referencing his fellow HT members.
The only thing he might have traction on would be claiming that using Heinz 57 kinds of choices is wrong since that wasn’t what they (the HT) actually had to do to perform RegEM (or whatever that abbreviation is). But I need to re-read the B&C article as I don’t think they did what our putative Mann says they did.
Dardie, points a and b under “some other problems”, towards the end. I make the remark broad, because “Mann” mixes in some gratitious asides, but regardless, I agree with (had previously made!) his points.
a. title (and scope) are misleadking, not well thought out.
b. paper does not support (does not even lay out a chain of argument!) the claim that it is impossible to resove the MWP quesiton.
I was trying to keep an open mind about the review and decide for myself whether or not it truly seemed as if Mann were the reviewer. The first paragraph alone sure led me in that direction. By the end of the first paragraph of S-141, it seems almost certain.
The intimidation attempts in the 1st paragraph are embarassing.
Re #23, TCO
B&C’s paper is titled “On the verification of climate reconstructions”. Which correctly describes its contents.
I can understand Mann having an attack of paranoia, but I don’t see what you are objecting to, TCO.
Wow, Herr Mann certainly hasn’t given up on his lousy arguments. Fightin’ for his bloody career, perhaps…
All that crap about criticizing a “decade old paper” actually suggests to me that Mann knows the paper is completely flawed. “Moving on” doesn’t excuse that fact. The whole review is a subterfuge.
Fred, it is not clear how many climate revisions have the issues that BC are going after. Is it just the Mann ones? Also, a little unclear without reading the entire paper, to what extent they are going after RegEM or earlier paper.
If people find flaws in Mann’s procedures, he should examine them and either admit or disagree with them. The impact of a particular flaw on the overall or on his ability or whatever, is a whole different issue. I find it annoying when he tries to bypass a discussion of particulars with the “it doesn’t matter” segue. Let’s figure out the particulars and THEN look at or argue the significance. On the occasional times when Steve or a cheerleader does similar things, that bugs me too.
If I ran that journal, I’d publish the paper plus Mann’s detailed rebuttal, plus the authors’ reply to Mann, plus Mann’s reply to that reply, plus the authors’ second reply to Mann, plus others who may want to weigh in. And I’d do it all in one edition.
For Mann and his friends, it’s the perfect opportunity to debunk his critics, if they can.
For the journal, it’s both a scientific service and a circulation boost.
Re #30, David Smith
Heh. In other words, you’d do it as a blog ?
Re #30
You must be under the misapprehension that Mann is as other men, and is aware of his own fallibilities.
Well we think it’s a version of a good old boy network growing in a field of science.
A cutting-edge-young field of science too. A perfect atmosphere for it. You could say like the early days of oil. Are they “evil chimpy mcbuckies”? No. But I don’t think oil men are either. But warmers do, and they even say the evil people run my country and are ruining it. They also inforce other countries to look at mine a “bad”.
High style-living (ie: travel-fame) busy important “work”, busy social life- everybody wants to be your friend, is an appealing situation to everyone in any micro-cosmic sect of society, group or institution, political party.
I don’t view the work of ClimateAudit or the comments of the “cheerleaders” falling into that spell. People speaking anything skeptical about the current state of this field of science barely gets heard. Certainly doesn’t make you “famous” as it stands now.
That’s the small difference I see missing in what TCO says in #29. TCO is stating an obvious way to be fair, and I agree with him. We all know how to do that. Yet, I don’t see RC et als ever being that fair, with or without the limelight shining on them. And they always sound irritated in the press or on the blog if they have to. LOL
(weird how all those links I referred to aren’t working)
ugh, of course typo:
They also inforce other countries to look at mine a bad–
should be something like “encourage other countries to view mine, as bad”
Sheesh! Happy 4th Americans!
Wow. The big picture is sort of a mini-Marxist conspiracy w/Mann as a puppeteer manipulating the IPCC, supporting big-media, & climate modelers on different strings, working frantically.
Captain Mann (over ominous rumblings):
Scotty, is the Hockey Stick gonna hold?
Scotty (over torturous screaming of the overloaded Hockey Stick):
She canna take it much longerrrrrrrr….
There are computer programs that compare texts, to see if the texts were authored by the same person. Thus, you could feed in some work authored by our dearly beloved Michael Mann, and then feed in the review, and you would get an estimate of the probability that the review was authored by Mann…. (This has the advantage of objectivity.)
Steve,
I’m surprised the reviewer didn’t accuse Bürger and Cubasch of being on the payroll of Exxon.
There are two problems with the way the editor of GRL handled this manuscript: (1) he submitted it to a reviewer who had an interest in not having the manuscript published; and, (2) when the differences of opinion betwen the reviewer and the authors are of such a great degree it becomes incumbent upon the editor to publish in the same issue of his magazine the submitted paper with revision, the reviewers’ comments written up as manuscripts with attribution, and a rebuttal by the the authors of the original manuscript. This is done quite frequently in other areas of the physical and social sciences, but doesn’t seem to be very common in climatology.
The arrogance of the reviewer is insufferable. Reviewers forget, or never learned, that constructive comments are important too.
Did you get the comments of the other reviewer?
JSP
The term “Conflict of Interest” comes to mind.
It’s interesting that Anonymous Referee #2 just can’t bring himself to cite McIntyre and McKitrick. This is depite several mentions in the text. Whoever it is, this points to a serious psychological problem of of denial! (Unless they are just REALLY sloppy in their bibliographical practises and forgot to include it. 🙂 ).
I’ve been too byusy to follow all this in depth, and I might have missed any discussion on this, but Nature has a quite farcical news story on the NAS Panel at:
http://www.nature.com/news/2006/060626/full/4411032a.html
‘Academy affirms hockey-stick graph
But it criticizes the way the controversial climate result was used.’
by Geoff Brumfiel
Interesting take! Not ‘NAS Panel reaffirms what we already know (hotter now than depths of LIA), says rewriting of previous 600 years only “plausible”, problems with methodology.’
Interesting graph there, too: splices together Mann with observational record plus projections obtained using SRES scenarios. Is there any observational science left in climatology?
Welike: Of course, RC is unfair. That doesn’t change my attitude of curiosity and wanting to understand what is going on. I would have no problem were I the editor, disaggregating the argumentative points from the substantive ones and would still want the benefit of hearing the other side.
JSP: I agree. You would think that the reviewer would at least maintain the fiction of constructive criticism. As it is, it reeks of desperation.
I wonder if it might not be best to leave the estimations of the identity of the Mann booster responsible for comment 2 for a later time. This business of trying to figure out who wrote what is work for brutes like johnA’s hateclub,not for gentlemen.
Reading the quote
“This paper should not be published in GRL. The exchanges between the present authors and everyone who has criticized their earlier paper have gone on too long. The point has been made over and over. The fact seems to be that there are choices to be made in this kind of analysis but the results seem to not be too sensitive to them when they are carried out in good faith (exactly what it means to be robust is of course a matter of judgement, but I have some experience on this and I think it is). The current authors have had their say in numerous other replies to criticisms of their work. It is time to get on with it! ”
I suspect this might get our host to roll his eyes, but one contributing factor to the prolongation of the discussion is probably that still no-one (or at least the anti-Hockey team) knows exactly what MBH did. Being able to get all their data, and run it through a code, to get EXACTLY their results (and I mean to +3 significant digits for everything) would naturally lead to quicker and simpler resolution of points of disagreement and agreement. Things then become more cut and dried. So, all I can say, is “Do not collect $200 and Return to Go”.
BTW, most journals have appeal procedures. Whether the article has any hope of getting through would depend on what the other reviewer wrote.
#42. In this case, the Comment that was rejected was Wahl and Ammann’s comment. It was a scurrilous article that deserved to be rejected. In fact, it was originally rejected last June without even requiring a Reply. The editor who rejected it was removed by Famiglietti and the article revived. After another process, the exchange was rejected once again. The only thing that pissed me off was the snippy comment by the reviewer in effect blaming us for the process.
There was an irony to this. In the re-submission, Wahl and Ammann included a claim about RE statistic significance levels, which they relied upon their Climatic Change submission without proving it in their Climatic Change submission. In my original review of their CC article, I said that they should not rely on results from a rejected article. This recommnedation was ignored. Thus the article continues to rely on (incorrect) results from a rejected article.
The first sign that our points about RE significance have penetrated into the climate science community is the recent Bürger and Cubasch article which deals with this topic, albeit not very concisely.
Re 18
The site is back up today.
The review looks like a reply, and I agree that all of its features could indicate that it was written by Mann itself.
He says, among other things:
The papers in 1998 and 1999 were likely wrong, but it does not matter because it is already a decade ago, they were confirmed by dozens of other papers following the same wrong methods, and this huge amount of consensus can’t be ignored. It is surprising that the authors (BC) ignore the overwhelming pile of nonsense that our team has written. M and M were not even written in a journal from ISI – social science journal instead. Blah Blah Blah.
Yes, I agree that it is the same antiscientific immoral stream of references to consensus and authorities that Mann promotes whenever he can, but on the other hand, my experience has told me that there are always several solutions and sometimes the guesses that a review had to be written by XY turn out to be incorrect.
#54. Luboà…⟬ my opinion is based not just on the points, but on things like layout style, vocabulary and sentence structure. Schmidt, Connolley, Ammann all have different quirks.
Did you see the link to the blog report of the Holivar conference where Bradley was spewing about ExxonMobil? In this case, causing even a climate scientist to be irritated.
FWIW, there were a few threads at RC weeks back where BP got a shocking amount of praise and admiration. If the BP CEO said it or if it was printed on the BP website, then it was “true.” I thought big bad oil was the villain?
But the real news (to me anyway, and I suspect TCO), is that a wonderful typo made it all the way through the Interactive Comment editing (do reviewer comments get edited?).
Anyway, it is at the top of S146, which is the middle of page 4 on my .pdf. One more comma and we would have had a classic moment in scientific correspondence.
The reviewer (whoever it was) came off a bit riled up, angry, and probably needs to do some relaxation exercises. Interestingly, I think the typo accurately reflects his (her?) underlying emotions.
I am so glad I am in a (currently) more cordial arm of climatology; this stuff can’t be good for one’s health.
#48 “Anyway, it is at the top of S146, which is the middle of page 4 on my .pdf. One more comma and we would have had a classic moment in scientific correspondence.”
lol! Thanks for the pointer. You’re right, one more comma and that would have gone into the Hall of Fame for inadvertant candor. Pride of place, too, in the entrance foyer.
Re: Update – The thought had crossed my mind that “no way would they allow Mann to do the review” but that he had it handed to him by another reviewer and contributed (heavily) to that reviewer’s “comments.” But I thought that was REALLY stretching things ethically, so I dismissed it entirely as impossible.
How is it that you know Mann knew it had been rejected ahead of time?
#50. In Mann’s CPD review which is online, he says that the Burger and Cubasch article had already been rejected elsewhere. How did Mann know that unless (say) Gavin (it looks Gavin or another team-mate was the GRL reviewer) told him?
Steve, I got confused and thought you were saying the review linked above was not done by Mann, not the GRL review.
Delete my post if you wish – I might confuse someone else into thinking I’ve read the GRL review!
1. Maybe it was in the cover letter for the submission?
2. Maybe Mann didn’t write this review (or did write the GRL review)?
3. Maybe someone told him?
I would move on, though. It’s not that interesting or damning.
Back on topic of an op ed.
I think Ross McKitrick’s “What is the “Hockey Stick” Debate About?” is an excellent starting point. I don’t know how to write the op ed to show the background and the non-support for Mann approach by the NAS.
However, if you want to do more in the op ed than just “preach to the choir”, I know what you must not do.
No sarcasm. No personal attacks on Mann or any else. It is ok to attack his work, but it is not ok to attack him. No attempts at humor related to the subject at hand because these are “inside jokes”. No great detail in the statistical methods because that will loose the bulk of the general readers.
I used to work with a plant manager who said he had the easiest job in the world because everyone told him how to do it. lol
I find that incredibly hard, personally.
We have now formulated a reply to the reviewer comment.
Will look forward to it. I liked your example of a nonsense regressor. You might consider making reference to some of the classic articles on nonsense regressors e.g. Yule; Granger and Newbold; Phillips 1986. Nearly all of the classic spurious regressions will seem significant under an RE test. That was definitely a point that we did not sufficiently attend to in our original discussion. But it would be worthwhile doing RE calculations on the classic nonsense regressions – drunkenness versus number of C of E ministers; Honduran burths against South African wine sales.
Phillips 1986 approach to spurious regression was,I thought, highly original and that’s the model of what happens with “spurious significance” (used technically not as a term of Mannian disapproval) that I had in mind in our own more modest efforts.
We surely had Yule and others on our list, but GRL article length prohibited us from digging a little deeper. You know you have to prove somehow that the series in question are actually non-stationary, and that is not easy, probably impossible given the relatively small sample size.
Publish a real article then, instead of a letter. Sheesh. And write well.
#59: Can’t you do that yourself instead of complaining all around?
Re:#54
Sorry, Bard, but Michael Mann himself has gone way beyond politeness in his own criticism of his critics. I would, therefore, suggest that some very “robust” commentary on his character would be more than fair.
Of course, it might not be wise, nor good form, but it certainly wouldn’t be unwarranted.
re #58: Gerd, I don’t know exactly what you are looking for, but this recent article might be useful.
#62: One could use the Dickey-Fuller test for unit roots.
Folks, you really need to pay attention to the reviewer comments on Burger and Cubasch here . Eduardo Zorita posted a comment. Now Mann has posted up 2 more comments in his sock-puppet (to borrow Lambert’s term) "Anonymous Referee #2". Mann sure needs to take a valium.
Don’t know about the valium, but he has more typos than Lee.
Re#64 – Laughable. The 2nd reviewer response (S163) resorts in it’s 2nd paragraph to, “I’m pleased that BC raise the subject of the recent NRC support. The reported(sic) supported the key findings of Mann et al (1998, 1999)…”
On S-164, the reviewer references a Mann publication “in press” and goes on posting RC links on S-165.
I also like the reference on S-168 under #10 to Rutherford et al (2005), after which the reviewer says, “Give us a break!” (the “et al” in that case would be Mann, Bradley, Hughes, Briffa, Jones, and Osborn).
You were spot-on from the start – it’s Mann, either as the designated reviewer or doing the work by proxy.
I enjoyed this from Reviewer 2’s last post:
“This the sort of cherry-picking one is used to seeing on contrarian websites claiming to “debunk” global warming by finding one or two locations on the globe that have cooled during the 20th century. One doesn’t expect this sort of behavior, however, from serious scientists in the field.”
One or two locations?? Seriously? It would be interesting to ask Reviewer 2, who clearly regards himself as “a serious scientist”, to substantiate this interesting comment.
While I am at it, I would also like to ask Reviewer 2 to comment on the assumption of a linear relationship between tree ring thickness and temperature. As a gardener, I have noticed that in hotter years plants are stressed, and there is less growth. And that is also true in colder years. Maximum growth occurs under optimal conditions of moisture, temperature etc.
67: Yes, it is very well established that the optimal temperature is 20-25 deg. C for most plants. This could possibly explain the “divergence” problem, if indeed temperatures are getting hotter.
Regarding the one or two locations showing cooling: every week the CO2Science.com folks (Idsos) report a location somewhere in the USA that shows a significant cooling trend. Of course, this is cherry-picking, but so many sites across the country are showing this trend that it really makes me wonder about whether the climate is actually getting warmer.
BTW, the “divergence problem” could also be explained by it becoming cooler! Wouldn’t it be weird if there is actually a cooling trend going on and we are getting fooled by “UHI” effects?
re 65:
jae, that is not possible.
Lee: Why?
Just read AR2’s remarks and one point he made at the end is true:
Of course, what it deserves better of is a reviewer, whether Mann or not, who doesn’t spew venom but has some degree of decorum. Such an open forum isn’t RC or CA and should not allow such antics.
And, BTW, if AR2 isn’t Mann, why doesn’t he show some Co-Jones and come over here and deny it? It’s not like someone hasn’t told him of the speculations here. It’s his reputation that’s at stake, IMO, not B&C, V&Z or M&M.
re: #71
Hey, cut Lee some slack! That was self-deprecating humor. (Though he rather spoiled the effect by not having a typo in that post.)
Is it just me or does “Anonymous Referee #2” sound an awful lot like a typical Yahoo Message Board poster?
Yeah, give Lee a break. Anyone who can have a joke at his own expense is ok. Although when he first came here he was breathing fire and brimstone I really like his attitude now, he says that he is here because the discussion challenges what he thinks he knows. That’s precisely what we all need to be doing all the time. There are two sides to every argument and this site needs more Lee’s who challenge what is stated here, usually by constructive counter-argument. (That’s what RC needs even more desperately because they don’t allow counter-argument). I find the shouting match between John A and Lee very unfortunate but I don’t think Lee is all (or even mostly) to blame there.
Having said all that I don’t usually agree with his point of view but his arguments are a lot more constructive than AR2.
What about me? How do I rate?
Fang.
Mark
I “Marked” you.
Re #73, and #75,
Dave and IL,
You might want to exercise caution in making assumptions regarding a possible, if not probable, con artist.
There are people who are more than willing to impose on the complaisance of others. If you review Lee’s posts since April 20, while considering techniques of argumentation,
you might be somewhat less complaisant.
So I’m a con artist now. How extraordinarily revealing.
Waaaaaaah! 🙂
Next instalment from Gerd is here.
Steve/John, it might be worth putting the link to the discussion page at the top of the main post.
£10 says “Anonymous Reviewer #2” stomps off in a huff and we see a snarky article on RC about how “peer review has fallen down in recent years”
#83 I, for one, would welcome a “snarky article on RC about how ‘peer review has fallen down'”, but I would not hold my breath waiting for it. Here is why: 1) The peer review process is dominated by “consensus view” researchers (this is to be expected and is not necessarily bad), so to criticize the process is to criticize one’s colleagues — never a winning strategy; 2) It is my sense that a lot of excellent, interesting, and critical, manuscripts have been rejected from the best journals because of mischievous anonymous reviews from members of the “hockey team” (this is very bad) — a topic some people would prefer to discuss only in private.
In any case, I see cause for optimism. Following the NAS report, I would imagine that editors of serious scientific journals would realize what has been going on. My hope is that they would begin to seek out manuscripts critical of the supposed “consensus” — this is how science progresses. Perhaps they could go further, and reconsider manuscripts rejected in the past based on anonymous review comments from “hockey team” members.
Just to be clear, I am entirely in favor of public, “onymous”, debate on technical issues, the sharper the better.
In conclusion, if RC wants to start this conversation, I’m all in favor of it. It is entirely possible that my suspicions are completely unjustified, and, if so, I’d be delighted to hear it.
Hmmm… I never seen such a lenghty, nasty ad-hominem attacks with so little criticism about the actual manuscript in any review report before. Someone should remind him/her about the General Obligations for Referees of the journal. At least the following points came to my mind:
We have an anonymous referee #3.
TCO, is that you ?
re #86: Coincidentally the two statistics books in the reference list are soon to be reviewed in Bulletin of the American Meteorological Society 😉
Missed the link somehow: http://www.gps.caltech.edu/~tapio/papers/bams_miss_data.pdf
No, but it reflects points that I’ve made and points that I would make, had I the ability. For people like Willie, who seem to think that they got the whole paper and that the title is good and such and that I must be a dumbie or a troll not to get it, maybe they will change their minds, given this type of review.
On the actual content: I still have to go through the paper. It is just so damn laborious to read poorly written and ambigious papers in a field that is already difficult. (Note that I can and do read papers in fields outside mine all the time, so it’s not impossible to do so, just a pain when they are poorly written.) Bottom line, authors need to make simple clear assertions that they then show analysis to support. Then we can engage. But when they do things like say there are bad degrees of freedom and not even say how many degrees of freedom there were or what would be a right amount? How do we engage with that?
How do we fix the errors of Mann if we can’t even express our own thoughts clearly? It is no good to say that Mann should have to justify himself. That horse is out of the barn…and the field obviously lacks mathematical rigor. To move discussion forward, issues need to be layed out clearly and obfuscation evaporated, not increased. Loved the full factorial BC paper. Loved the Huybers comment. This BC paper is not up to speed. I think if they just made it much more simple and plain and vanilla and avoided some of the bridge too far implications, that this would benmefit the field. When they have built a few more bricks, they will know how to build the whole house. But it’s evident that they don’t know now.
Jean, do you suspect Schneider wrote the review?
Oh, I see, it now turns out it wasn’t Michael Mann…well, well there’s a thing. So, the guns turn onto GS…
Yet the highly unpleasant comment #1 here (amongst several other dotted about in various threads this was the worst) is still there. Again, well, well…
Dude, we are talking about a different reviewer now.
re #89: TCO, I almost completely agree. I actually read yesterday the BC paper, all related comments. I even read the RegEM paper by Schneider (2001) and also the Rutherford (2005) paper. I agree that BC paper needs some improvement, but on the basis of AR3 comments, e.g., the Rutherford (2005) would not have been accepted either. I am afraid though that now the BC paper gets rejected because of two negative reports (AR3 fair, AR2 unjustified): I would have recommended “major rewrite” if the journal allows it.
re #90: Well, let’s put it this way: the editors tend to pick the reviewers from the reference list. The editor wants definitely persons who are technically (statistically) qualified for the review. The AR3 definitely knows (unlike AR2) what he/she is talking about (read Schneider (2001)).
Re #92, and I’m talking about the update.
re #94: Peter, update is talking about Mann being a reviewer for GRL. That was issue just for the fact that the Anon2 for this new submission knew about the rejection. Most comments here are talking about the fact the Anon2 is, very likely, Mann himself. You should actually read the Anon2 comments, is that the way you think a scientist writes review reports?
Jean, nope, most people here find MM to be anon 2 – period. Then then, in the post I mention, blow that supposition up into the kind of personnal attack I find quite chilling.
But, humble pie time, you’re (and TCO) right, I have misread the update – apologies.
It’s cool. I love you, Pete.
And Jean: don’t worry about the paper getting rejected. It’s a service to the readers. BC need to write more clearly. They can do it.
I just read Referee #3’s comments and I agree that it’s probably Tapio Schneider. Yes, B-C could be written better, but it’s more clearly than 90% of the articles in this field. Other than a few points – which have not been brought up by this referee – I can understand what they are saying. I don’t think referee #3 understood the spurious RE argument at all and spun off into his own hobbyhorse. B and C use the number of gridcells as an example of an irrelevant and spurious predictor; I would have used dot.com stock prices, but they both are the same point. The referee interpreted this as being a missing data bias – which is the hobbyhorse of the referee. There was nothing unclear about how B-C stated this point; the referee just grabbed the wrong end of the stick. Look at his closing:
I don’t think that the referee has any business suggesting other interesting topics. Let me give an analogy. Let’s suppose that I ddid a prospectus with a business plan for a certain type of business. The job of someone considering the prospectus is whether the business is properly disclosed. Maybe the regulator thinks that some other business would be better – but so what? The analogy isn’t exact.
Here the job of the reviewer is to deal with the issue at hand, not to suggest other lines of research. How pompous is that?
Is the question of RE significance important enough in itself to merit an article? Well it was important enought to be considered by the NAS panel. Maybe someone else is interested in "To what extent does the fact that temperature values may not be missing at random lead to biases of reconstructed temperatures?" but B-C aren’t discussing that question and they shouldn’t be obliged to. This is just pompousness on the part of the reviewer.
Also one wonders about bias. There are lots of points by reviewer #3 that I agree with. The reviewer here is about 5 times more diligent than most reviewers in the field and I think that the diligence is related to the criticism of the Team. How could Wahl and Ammann be passed by Climatic Change if these standards were applied to them? They must have had puffball reviews by their pals.
Oh get over it, Steve. I don’t care about what Amman and Wahl get away with. I care about an objective standard. And it’s just not that good on that standard. And I’ve read papers in psych, soche, econ, chem, clima, physics, etc. etc. Great that you who are immersed in the subject and often post non-layed out arguments yourself can tell what is going through BC’s head. But they need to do better then an audience of one. It’s not unreasonable to ask for clear arguments. For not confusing the reader with things like scope and title and degrees of freedom and ability to solve the millenial controversy that are all not layed out in the paper and that cause confusion as you read the paper to try to figure out what points they really can validly make.
They can get it rewritten and submit elsewhere. Heck, maybe it will even tighten their thought process to figure out what points they are making.
Oh…and when you see what is happening with this paper, maybe you better grasp my point that style and writing brutally clear and simple papers DOES MATTER when you are trying to change opinion in a field.
Write a frigging flawless paper, in scope, assertions, footnotes, following the guidelines, obeying Strunk and White and the like and you will be blown away what you can do!!!
I read the Tapio review. What do you think of his review, of the issues with missing values. Have not heard this addressed much before.
Actually it might not be Tapio. But it’s a stupid review. It’s crazy to think about proxy reconstructions as a problem in filling in missing values in temperature gridcells. I agree with David Stockwell.
What? I liked his style comments. Just blew past the technical things that he raised, since I still haven’t had the will to bull through the BC paper to figure out what issues that they raise, they support.
wait, we or I am getting all confused. In 101, I am asking about his published book review. So that IS Tapio. What do you think of that?
With respect, I just don’t see the strength of this example. Couldn’t you be accused of ‘cheery picking’ spurious correlation. You could find others that didn’t correlate.
But if there is no correlation of series with local temps, and teleconnections only extend to 500km at most, then the series themselves have been ‘cherry picked’ for local temps, end of argument.
I agree and have made this comment before, Dave. I really think that Mann has had a bad effect on Steve. He tends to go for debate points at times (and sometimes unfairly) rather then just trying to understand things and to clarify them.
#105 Should be ‘cherry picked’ for correlation with GLOBAL temps. #106 Yeah well I have got bitter and twisted by something like this. Its easy to sit in the peanut gallery.
Oh…agreed on the peanut gallery. If Steve stops putting up content, what would I have to criticize. That said, I still have good points…
Why are you bitter and twisted? Something I did? Steve did? Another person? Dish!
http://www.co2science.org/scripts/CO2ScienceB2C/articles/V7/N19/EDIT.jsp
The underlying problem with most of thes arguments is that statisticians are looking for “unsupervised” methods to kick out ridiculous relationships and it’s not easy. In the Philllips 1998 article (which I quoted last summer and is quoted here by Eduardo as well), he mentions an example from Hendry 1980 of a statistical relationship between price levels in England and cumulative rainfall, that was resistant to broad ranges of statistical tests.
Or go ack to the original examples of spurious regression. (It’s always a good idea to work with statistical sets that have nothing to do with any topics in dispute to see what’s agred on.)
Yule 1926 pointed to a spurious correlation between the mortality per 1000 in England from 1866 to 1912 and the proportion of Church of England marriages to total marriages. This had an r2 of 0.95. I’ve done the calculations and it has an RE of 0.97.
David, did Yule cherry pick the marriages data to correlate with mortality? Maybe so, maybe not. It doesn’t matter for the argument. Yes, stock prices in some periods yield spurious RE; in others they don’t. But the problem is that they sometimes do.
It’s important to keep in mind Hampel’s dictums that statisticians can identify outliers, but the issues ultimately have to be decided scientifically. For the bristlecones, Mann wants to argue that there is some deus ex machina statistical rule that mandates their inclusion; e.g. you get higher RE scores with bristlecones, therefore they are a valid proxy and it is obligatory to include them. That’s why I’ve emphasized the nits of bristlecone botany so much. There are obviously problems with this proxy; there’s a high probabilty that it’s growth trend is not a unique world thermomenter. End of story.
Yeah, that extinction thing seems like fancy doodling as is the general impression I get from all ENM studies. 😉
To your points, I think that species saves because of expansion of range (for example American crocodiles) may be small. But I also think that the definition of species is a bit fluid. I mean red wolves or things like that, that are really more races then species the way I would think. We would get new races, I assume.
110, the gridcell thing should be sufficient here. I think that the burden of proof for physicality should be on the includer for any teleconnection or climate field crap. Otherwise, you’re just feeding stuff into the hopper until you get something to use.
But isn’t the issue the proportion of spurious corelation and their effect. It does no good to point to one or two series that aren’t used anyway and see look! If Mann’s argument depends on ALL the series being bonafide then, sure, but if it is resistant to a small proportion of spurious correlations then he can always argue that they ‘don’t matter’.
Now you have argued that Bristle Cones drive the reconstruction and that is a different case, because the proportional contribution from them is high, and they probably are spurious.
Now if Mann is saying BECAUSE they have high RE they have no chance of being spurious, that is the flawed argument. You have shown they contribute greatly to the reconstruction, that they don’t correlate with local temperature, and the biology is bad for temps, that should be enough. I think the strong argument is with Bristlecones being spurious, but dot.com’s being spurious is not a strong argument for a flawed reconstruction.
Steve needs to disaggregate issues better. It’s like he tries to argue that both algorithm and inputs are wrong. BOTH. But when we start getting into the nitty gritty of algorithm, he tries to drag in issues with the inputs.
I’m mostly saying this to bump the link down so John A’s screwed up side bar gets fixed.
#111. Yeah well I don’t know what the proportion increasing range would be in a specific case, because it depends on the configuration of land mass, but in a theoretical sense, for range shifts over a flat sphere, the increases should equal the decreases. So to just consider decreases builds in a bias towards reductions in area.
I have found more extreme contributions to bias in this study since. The main one is the use of the Hadley model low, medium and high scenarios for temperature increase, to provide the prediction of temperatures for 2050. The actual rate of increase in temperatures has been lower that the lowest scenario (e.g. +0.75C expected by 2050 not +1 or 1.5C). But the authors take the mid scenarios to represent the most likely scenario.
I agree with your fancy doodling characterization too. Though sometimes its amazing, finding new species and everything. Other times its dissapointing. Thing is, we don’t measure things from the critters perspective.
I want alligators in VA. If Steve can just confuse the Kyoto advocates long enough, we might be able to get a good tipping point and make me happy with some palm trees in VA.
I love alligators. you know they are almost up to the VA border in parts of NC? They were in the midatlantic at certain times in history or paleohistory or something. can’t find good info actually by googling.
#117 I come from a place where the crocs eat people regularly, thanks to conservation laws. I saw a guy in Thailand put his head inside ones mouth. Anyway +1C might be enough to do it for you, thats about 500km.
Let’s try another example – the Gaspe extrapolation. Is this a problem of data or method or something else?
Mann extrapolated one out of 415 series at the start. Only 4 years, but still unique. The extrapolation affected 15th century results which “wanted” to go higher. There are problems with the series – # of trees, correlation with temperature etc.
Is the method wrong? At best, it’s opprtunistic.
Is the problem that the data is bad? Well, yes, but that’s not really the issue.
What would bother an accountant by this is the combination of things: – the unique treatment of the series AND the especially the failure to disclose (and even misrepresentation of the start date). If you have a failure to disclose combined with a unique treatment of an account and with it mattering, the auditors would be really angry. Also if you have unique treatment like this, you have to justify it to the auditors, by showing it doesn’t matter, so they’d probably ask why you did it. I’ve never found an academic yet that understood this line of reasoning.
In the Gaspe case, I don’t think that it makes sense to separate the method from the data. Same with other issues.
#114 I think all algorithms can be ‘right’ in some situations — the case where their assumptions are met — unless they are miscoded which is common. So all criticisms come down to a mismatch between inputs and assumptions. Its not one or the other. If the method downweighted bristlecones due to, say, lack of correlation with local temperatures, then you might say that method was better even though the data was poor. I don’t understand the method enough. Does it calibrate on a global temperature and then predict a local temperature? That would incorporate some strange assumptions.
118: I swam in a river for hours in Queensland that had crocs. THought the Aussies were taking the piss out of me with croc stories. Heard afterwards that this was very unsafe. Did not see any though. Including at the boat landings.
I don’t think that 1c is 500km at the NC/VA line. I think 500km gives you a lot more then 1C.
Steve, to the extent that both methods and data are wrong and both errors materially affect the answer, this helps your case. To the extent that both need to be wrong and changed to affect the answer, then this makes your case more precarious.
WRT Gaspe: The extrapolation is an error of method. The issue with the trees is an error of data.
Steve: I think you need to be very careful (not in a CYA mode, but in an ethical Feynman scientist mode) to say exactly what your case hinges on.
#121 Think you’ll see the one that gets you? Steve our messages crossed but I think we agree.
#122 I think you are right. 1C annav. is about 1degLat is about 110km.
You usually don’t see them until the last fraction of a second
How would you know that unless you got bitten? All the movies I’ve ever seen, they are like logs in the river or else they are moving around on the banks or such.
A young girl got taken by a croc in the Northern Territory yesterday. She was fishing by a river with her parents after dark. She went to the river’s edge to fetch some water. Her parents heard a “splash” and she was gone. Her body hasn’t been recovered.
Two or three people get taken by crocs every year in the Territory. Just because you don’t see them doesn’t mean they aren’t there. They are very efficient predators, and it’s true that generally the victim doesn’t see them, if at all, until it’s too late.
A few years ago I went on a boat tour of the Mary River mouth at Point Stuart, east of Darwin. It’s claimed to have the highest concentration of salt water crocs in the world, a statistic that I wouldn’t care to dispute. There were hundreds on the river banks, and god knows how many (unseen) in the water.
Crocodile numbers have exploded in Northern Australia since a ban on culling ten or fifteen years ago. (Or maybe it’s due to global warming!)
I think several of you have been too hard on Steve regarding the dot.com stock example. With this example, Steve is giving a very effective reductio argument against one aspect of Mann’s method. He says in #110, “For the bristlecones, Mann wants to argue that there is some deus ex machina statistical rule that mandates their inclusion; e.g. you get higher RE scores with bristlecones, therefore they are a valid proxy and it is obligatory to include them. ” If this line of argument were valid, the dot.com stock prices also should be included as a valid temperature proxy. This is clearly an absurd conclusion. Therefore, there has to be something worng with the argument. Steve’s example is more effective for making this point than the spurious correlation example BC give precisely because it is so obvious in the dot.com case that the correlation has to be spurious — just like in the classical examples from Yule etc that Steve quotes.
I agree that dot.com may be a better example than BCs example of number of data points because it is so irrelevant. But the dot.coms would not be included because they are not trees. It makes you look like you are seriously proposing adding dot.com’s which BTW probably don’t correlate with temperature any more, given the up and down shape. What are you going to do if dot.com’s don’t correlate — use a different example of course, which shows the example is cherry picked.
Its not that the argument is incorrect. Its just that I think there are stronger arguments. Why say that bristlecones MIGHT be spurious when you can show that they ARE spurious.
Yes, I agree that one also needs to argue about what is in fact going on with the bristlecones. However, the dot.com example shows that the argument that “something must be included because it yields a high RE score” must be wrong. The bristlecones or other proxies related to trees do not reveal the weakness in this methodological claim nearly so well because there is a hint of plausability that they could be related to temperature. It is precisely because “the dot.coms would not be included because they are not trees” or, I would add, plausably related to temperature, that they are effective for revealing the weakness in the method.
I think you are better off making the point with shapes. Bristlecones: it is debatable how bad of a proxy they are. Dotcom stocks: it is a cherrypicked example to work well with the math, but not be physical.
So the point is that if you look at enough variables with no physical or logical relationship to the variable under consideration, you are likely to find some in there that have correlations significant at .95 or .99 levels?
Well, duh!
The relvevant question is, what is the probability that this particular correlate, chosen for its plausible biological link to that variable under consideration, is a ‘nonsense’ predictor.
IOW, not, what is the probability that there are nonsense predictors out there (high, of course) but what is the probability that THIS is a nonsense predictor.
TCO: One comeback with the shapes is that you have chosen shapes that do not correspond to any series in the real world. The dotcom series is an actual series that yields a high RE but is not plausibly related to temperature. It shows how one cannot rely on a mechanical rule like high RE. One needs to ask whether there is any sensible theory or physical mechanism expalining the correlation. To borrow a phrase from economics “measurement without theory” is not acceptable. I think the shapes idea is best used for showing things like how the short period centering biases the selection of proxies. That is effectively what Steve and Ross did with their “red noise” examples. The dotcom example is making a different point. It is not just the shape of the dotcom series that is important. It is that it shows how a real world series can produce a spurious relationship. Just using made up series does not quite do that for you.
Oh…ok. It’s still a crappy example for that, since they stretched it, no?
#134. Lee,
In the case of the bristlecones, it seems to me that there is a substantial probability that it is a nonsense correlation. If Ammann or Mann went back to basics and proved that the bristlecones were valid temperature proxies, then that’s fair enough. But the overwhelming specialist evidence is that their 20th century growth is not a temperature proxy. You can’t just ignor that because their inclusion improves an RE score.
It’s Mann’s side that attempts to include them as “mathematically” required or to argue that their exclusion is a “mathematical error”.
Let’s go back to Mann and “nonsense regressors”. “Nonsense regressors” have a specific somewhat technical use in statistics to mean something completely unrelated and usually to denote a trending unrelated series. Mann says that noise is a nonsense regressor. It’s a nonsense regressor, but not a “nonsense regressor”, if you get my drift. Referee #3 had no clue either about the nonsense regressor point.
Climate scientists need to take a couse in time series from a business statistics viewpoint to get a feel for the pitfalls. It’s hard to believe they are so naive.
Having said that, I think that maybe the way to make the point is through the concept of the “power” of a statistic. If you look at Mann’s argument in favor of RE significance, he argues that the RE statistic has excellent power against AR1 noise (probably with low AR1-coefficient. ) It probably does.
I think that the way to progress this discussion is by discussing the power of the statistic against other multivariate situations that are plausible in a paleoclimate situation.
What we did in MM05a is the right way to go, but the results can be generalized and the point probably emerges more clearly from the generalized point.
What paleoclimate people need is a statistic or statistics that have “power” (in a statistical sense) against say, (a) a composite formed from biased selection from red noise. (b) series contaminated by end-period fertilization. The RE statistic is completely unable to deal with such situations. I think that’s the moral of the discussion and latent in our earlier articles.
Verification r2 is one suggestion, but that was merely using a standard statistic at hand. Perhaps a good applied statistician could design a better test with power against nonsense regressors. HOwever the task may not be easy. In the Phillips 1998 reference, theres a mention of an example of nonsense regression from Hendry (1980) entitled interestingly “Econometrics: Alchemy or Science”, which proved resistant to testing. It sure seems like climate science is re-living issues dealt with by econometrics. I’m going to track down HEndry’s article , which led to a recent collection of essays with the same title.
137:
So Mann can’t mathematically argue that they are valid predictors, and you can’t mathematically argue that they are nonsense predictors. That means this issue reduces to a nindependent questin fo the phenomenological validity, revolving around biology and the specific analyses of these samples, not methodology in general.
It seems to me that Mann has a point (which he isnt making well) that if there is a predetermined criteria for inclusion, then kicking out data that meets that criteria without good reason is methodologically suspect. You may argue that his criteria are not valid, or that there is good reason to exclude these data, but those are different issues than this one.
138: Usage of the term “nonsense” is unfortunate and usage of obscure examples like dotcoms is a distraction. Look at how Huybers clarified the situation wrt correlation, covariance and off-centering. Use that as a model in your own thinking. Same with the flavors. MECE, MECE. Disaggregate issues, even interactions.
I’m rapidly getting confused, but have some sympathy towards the point in 139. If things devolve to a single issue, then you need to be upfront with that. If they require two things to be wrong, then be upfront with that. It there are two independant torpedos either of which can sink the ship, then show this clearly. But don’t confuse the situations.
I’m growing increasingly concerned about our need to rely on statistics, and in some cases dubious statistics, in climate reconstructions. I think the fact that we are having this debate about statistical methods, the phenomenological relationship between proxies and temperature, and selection of data sets is telling us a very simple fact. The current list of proxies are not reliable for use as temperature estimators.
There is a clear need for development of a suitable palaeothermometer that can be applied to reliable archives using sound phenomenological principles and with a well understood constitutive relationship between temperature and the parameter that is measured. The closest we possibly come to this is the oxygen isotope composition of high latitude ice cores. Some other proxies, such as the noble gas mixing ratios in groundwaters may also be good proxies but these probably lack the temporal resolution to answer questions relating to the MWP for example.
I think we might be in danger of arguing around an ever decreasign circle with respect to statistics and tree ring records. My own hunch is these are not reliable proxies for temperature.
I don’t disagree with any of this, not do I believe that there is anything in our articles that would suggest this.
Look at what we’ve actually said. We said that Mann’s results fail verification tests, including ones that he said they passed. This has been confirmed. Mann is now arguing that the some of the verification tests – which were his selection not ours – are not appropriate. OK, he can argue that and it’s being argued now. But he should have argued this at the time.
We’ve said that the results are not "robust" to presence/absence of bristlecones, known ahead of time to be a flawed proxy. CO2 fertilization was said in IPCC SAR to be problematic. MAnnignored that and included them. No one knows what Mann’s inclusion criteria. Mann et al 2000 should not have said that their results were dobust to presence/absence of all dendro indicators, when they weren’t.
The lack of robustness shows that the bristlecones are an outliers.
If you read Hampel on Robust Statistics, he agrees entirely -sometimes being an outlier is a sign of something interesting; all it means is the data has to be separately considered – but it’s a scientific issue and not a statistical issue. I have no trouble with that. It’s Mann that wants to avoid scrutiny of the bristlecones by invoking concepts of mathematical and statistical "correctness" that exist only in his own head. Preisendorfer’s Rule N is not a test of whether bristlecones are a temperature proxy. I’m using math to show his arguments are wrong, not to put something else in its place.
But in simple terms, how can one reasonable base a view of world climate history on bristlecones? It’s not just me saying it. Tree people had said it – Graybill and Idso, Biondi, now the NAS Panel.
#142. Paul, I agree entirely. If you’d like to a couple of interesting articles showing alternative approaches with tree rings, look at Millar et al 2006, and Naurzbaev et al 2004, both mentioned previously on this blog with citation. The NAS Panel even favorably mentioned Naurzbaev et al. They use the excellent dating properties of trees, but then apply information about changing tree lines.
The Hockey Team HATES changing tree lines as a class of evidence, because there is rather strong evidence of higher medieval tree lines around the world. A survey of this evidence is urgently needed.
I think that their strong advocacy position has interfered with a scientific appraisal of lines of evidence. For example, there’s a discrepancy between the temperature history from the Polar Urals Update and Yamal. There’s tree line evidence in Siberia to consider.
Briffa garnered a lot of attention with his original Polar Urals study which showed a cold 11th century (based on incorrectly dated early series IMHO). But with updated data in 1998, the Polar Urals data showed a warm MWP. That should have been presented and discussed. At around the same time, Briffa found a nearby data set (Yamal) a site with significant north-south movement and created a chronlogy with a hugely strong 20th century increase, which seamlessly entered into multiproxy compilations without ANY discussion or reconciliation with the Polar Urals Update. I can think of no acceptable justification of this sleight of hand.
Steve, this is exactly the kind of data we should be looking at. Dendrochronolgy is a wonderful tool and can give us very precise dates. We understand temperature lapse rates with altitude and the limits imposed on tree growth by cold temperatures. Thus we have an excellent indicator of temperature combined with a wonderful dating tool.
I need to go back and look more closely at the tree line data. Incidentally has anyone looked further back than the MWP. What about the Holocene climatic optimum. My understanding is that tree lines in Scotland were significantly higher about 5000 years bp?
Lack of selection criteria seems to me the biggest issue. The bristlecone feritilization thing seems like more of a stretch: given how crappy so many of the other tree-ring proxies are and prone to confounding, are bristlecones really any worse?
This discussion is not unlike others I see here and sometimes the cogent point is made (as Peter Hartley does in a comment in his post #132 that is listed below and with my emphasis added) without an acknowledgment from those to whom it was directed and then the discussion goes on and on and on.. without me and I would guess others here ever really knowing whether it was understood.
#145. Paul, there is much evidence that treelines were higher and more northerly all over the Arctic in the Holocene Optimum. I’ve got many references and will try to post some up some time. Not just there, but also in the California Sierra Nevadas and recent work (Schlucter) suggests no glaciers in the Alps. Personally I am very skeptical that the Kilimanjaro glacier existed than and think that Thompson’s dating is very, very shaky.
Just for fun I know no one would buy it but I could argue that the dot.coms are related to temperature. For instance dot coms need servers which use a lot of energy. The size of the industry would be roughly correlated with the value of the stocks. A greater dot com industry could imply a greater bank of servers. Servers both give off energy and CO2 emissions are required to produce that energy. Consequently the temperature is dependent on the dot com stocks.
How could that question be resolved (Kilamanjor, Thompson dating)? What would be the appropriate analysis or test?
Now that I’m apparently back out from under the filtering, let me try this again.
Tree line data is probably a good qualitative comparison of PAST trends, not of current conditions, **IF** one takes into account certain major assymetries.
Rates of change in treeline downward are probably far faster than rates of change upwards. A tree can die in one year, while growing even a small grove takes far longer. Also, recruitement of new young trees is generally far more sensitive to bad conditions than maintainaince of established trees. A year that wipes out a stand of seedlings might not bother an exttablished grove at all. Thus, failure of a treeline to extend upwards is NOT necessarily evidence that conditins are worse now than when a grove existed there.
Taken together, this means that one can not compare TODAY’s treeline to past treelines, to interpret CURRENT relative temperatures.
Given that lots of the high altitude US locations don’t show any instrumental warming, it makes sense that the tree lines don’t move either. Unlike rings, they can’t respond to global climate fields.
150. the start would be for Thompson to properly publish his material the way an archaeologist would instead of writing Reader’s Digest articles for popular magazines like Science. He’s not provided an archive for Dunde drilled in 1987 so why would he provide adequate documentation of Kilimanjaro.
I agree that his articles are not logical, clear expositions, but seem more textualist. But, I asked you seriously. What would be a good test? Do we need to recore? Look at it like a Martian for a second, like a disinterested scientist or a CEO who just wants an answer and doesn’t care about the squable or who looks bad: what kind of test is needed?
Certainly, when Thompson can rely on the likes of a Tom Brokaw to do that job, he has even less of an excuse for his writing style and publishing choices.
The current main stream thinking on AGW becomes apparent from a preview given in a USA Today interview with Tom Brokaw here:
1. The tipping point argument will encourage the “average” Americans to think seriously about GW.
2. The evidence is overwhelming and precludes even listening to what the “minimizers” (not denialists or even skeptics, but mininmizers) might have to say.
3. The medicine will not be all that bad (at least at the economic level of a former network anchor)
In summary, I’d like to think of this program’s potential as “The Hockey Stick on Steroids with No Side Effects”.
#148 The Holocene Optimum is in particular optimum for high-latitude summer time insolation which is why the Artic is comparably warm (Milankovitch). Low latitudes (Kili,Andes etc) are however less affected by that.
The Alps passed in the last 10-20 years the glacier line of the Medieval. In the Tauern there was active gold mining upto 2500meter for over 200 years emptying nearly all possible seams. Now with the retreating glaciers never seen soils appear and people are already making fun of a new gold rush.
Actual retreat of glaciers is so rapid and strong that the glacier tongues are no longer retreating (since this needs a dynamical reaction of the glaciers), they are disappearing where they are melting from the former accumulation zone to the tongue.
Would be very interested to see the papers showing an icefree Holocene Optimum.
Re#155, I wonder why his enviro wife doesn’t drive a hybrid? Wikipedia says there’s a 3rd daughter – why no hybrid for her?
I like this comment: “It’s the same science that we are drawing upon and it’s irrefutable,” he said.
Brokaw dropped out of the U of Iowa after one year, earned (according to wiki) the nickname “two point Tom” as a slap at his GPA while getting a poly sci degree at U of South Dakota, and now he can “draw upon” the science of climate change?
I’m more qualified to be a network anchor than he is to talk climate change.