Reverse Engineering Hegerl et al.

Lee, a new poster, complained that my attempting to guess the series in the CH-blend was insulting to Hegerl et al. [SM note - This sentence is added on Friday evening. I had a sentence somewhat like this in this post. Lee said that the sentence misrepresented his viewpoints - see one of his posts below - and so I removed it. He then complained that I had removed the sentence - see another of his posts. I didn't save the sentence, so I've restored the point from memory but it's undoubtedly somewhat different from the original reading.]

Here’s my first attempt at replicating the CH blend without knowing what proxies were used:

Black- CH "long"; blue- emulation of CH "long".

I emulated the CH-long blend using the predictions in my earlier post as follows. All of the 12 predictions are in the 14-series Osborn and Briffa [2006] data set. I removed 2 series from the smoothed Osborn and Briffa data set (the Foxtail series and the Chesapeake Mg-Ca series) , took the average of the 10 series available in 1251 (that’s one more than CH so there’s an adjustment to come) and then scaled the average to the CH-long blend. I’ve obviously been able to replicate the CH-blend pretty accurately without them even having to say what proxies they used. Their weighting methodology is not an unweighted average of the proxies. So it’s hard to tell whether the remaining differences relate to weighting systems or different proxies. There’s at least one proxy that I’ve not matched. Also, I’d be surprised if Hegerl used the Alberta version from Luckman and Wilson [2005] – they probably used an older version. I’ll try this.

Does this "prove" anything? Right now I’m just amusing myself by trying to guess what’s in the CH-blend. I think that this should prove to Lee that I’m probably going to be substantially right in my predictions. I’d be amazed if I don’t have at least 5 guesses right.

For the purposes of their Nature study, if Hegerl et al had written the same study and, wherever they used CH-blend, had substituted OB-blend, would the Nature referees have cared? Off the top of my head, I can’t see why they would. Varying this slightly, suppose they had listed the items in the CH-blend and an alert Nature referee noticed that they overlapped the OB-blend, would he have required some discussion of this? Who knows. If the CH-blend is pretty much the same as the OB-blend, should Journal of Climate referees care? Is it something that should be discussed in the submitted paper?

For our purposes, if the overlap is substantial, we can safely say that the two studies are not "independent" in their proxy selection. Does this show that their selection protocols are biased? Maybe they independently arrived at the same proxy choices because these are what result from objective protocols? If so, I’d like to see an objective statement of the selection protocols and how they were applied. Otherwise, there is certainly the appearance of cherrypicking. It’s hard to go much beyond this until the JClim paper comes out.

105 Comments

  1. IL
    Posted Apr 21, 2006 at 10:43 AM | Permalink

    The point Lee should be getting worked up about is that it should not be necessary to guess how the series is arrived at or what its components are! The study should be transparent and replicable by anyone else competent to do the analysis with enough information to do that. This just isn’t the case and nor has it been in almost all other cases.

  2. John A
    Posted Apr 21, 2006 at 10:50 AM | Permalink

    Well Hegerl did say that cherry-picking was essential (to get the best quality Hockey Sticks?) so why should we be surprised when she does precisely that?

  3. Posted Apr 21, 2006 at 10:56 AM | Permalink

    Wow.. I’m surprised how close that matches.

    Good job!

    Is it just me or is the climate reconstruction field really not coming up with anything new lately. Seems like it’s mostly the same warmed-over series recombined in trivially different ways.

  4. IL
    Posted Apr 21, 2006 at 11:15 AM | Permalink

    Yeah, and each one makes it into Nature!

  5. TCO
    Posted Apr 21, 2006 at 11:27 AM | Permalink

    Nice, Steve. You do know the field.

  6. Jean S
    Posted Apr 21, 2006 at 11:28 AM | Permalink

    Steve, be careful here. Since you do not know the series to be used, you have just created, with realclimate terminology – “an independent proxy-based reconstruction that also suggests that the late 20th century warmth is anomalous in multi-century context”, i.e., confirming the hockey stick. Moreover, the close match to CH-long just shows how “robust” the CH-construction is. Unfortunately, this is a non-peer reviewed website, so I guess the construction won’t be good enough for RealClimate (I only wish their web site is peer reviewed). Maybe you should submit your construction to Nature :)

    Another thing, where can I find the MBH98 construction time intervals (are they known?) and the (approximate) number of proxies used in each interval, I’m trying to see if I can figure out something about those confidence intervals…

  7. John Hekman
    Posted Apr 21, 2006 at 11:41 AM | Permalink

    following up on the suggestion of Jean S that you submit this to Nature, you could pick different cherries that showed a radically different “reconstruction.” Putting your close replication of Hegerl alongside your very different one, the point would be that the reader knows absolutely nothing from the Hegerl article unless the cherries are revealed and some justification is given for which cherries were selected.

    Of course, Nature would at this point consider you an avowed enemy of climate science, so is there some other journal that would appreciate zinging Nature by publishing your criticism?

  8. TCO
    Posted Apr 21, 2006 at 11:47 AM | Permalink

    There are plenty of other journals that will publish Steve. He needs to finish his arguments, produce well-written documents, and consider which publication goes to which journal more effectively.

  9. John Lish
    Posted Apr 21, 2006 at 11:48 AM | Permalink

    #2 wasn’t that D’Arrigo John A? I thought that Hegerl spoke of confidence intervals running from the floor to the ceiling?

  10. Steve McIntyre
    Posted Apr 21, 2006 at 12:08 PM | Permalink

    #2,9. John Lish is right. Hegerl’s remark is pretty funny in the context of these confidence intervals though.

    #6. The reported 11 MBH98 steps are 1400, 1450, 1500, 1600, 1700, 1730, 1750, 1760, 1780, 1800, 1820. Why these steps- who knows? Mann also has published a reconstructed PC starting in 1650 so he might have used a 1650 step at times. MBH99 takes one more step back to 1000. The confidence intervals in MBH98 are 2-sigma of the standard error in the calibraiton period. The calibration period is 79 years long. If you look at my notes on the linear algebra of MBH98, in the AD1400 step and AD1000 step, it boils down to a type of partial least squares procedure (which I need to write up), but which can be approximated by a multiple regression of the temperature PC1 against the proxies in the AD1400 network (22 predictors in the AD1400 network are surprisingly orthogonal).

    The confidence intervals were re-stated in MBH99 (I posted this last year in May or so – see the MBH98 Category). If you can decode the MBH99 confidence interval methodology , which comes at it from a frequency perspective, I’d be much obliged. I’ve asked help from a top-level time series expert who couldn’t make head nor tail of it.

  11. jae
    Posted Apr 21, 2006 at 12:23 PM | Permalink

    Whoa! The HS guys and gals are calibrating their proxies using part of the instrumental temperature record and then trying to make their reconstructions fit the rest of the temperature record, by careful selection of the proxies (aka, cherry picking). What if the instrumental temperature record, itself, is bogus (which is very possible, because of heat island effects, bogus data massaging)? How embarassing would that be?

  12. fFreddy
    Posted Apr 21, 2006 at 12:41 PM | Permalink

    Re #11, jae

    How embarassing would that be?

    Not even vaguely, unless the scientific establishment and the journalists start doing their job.

  13. TCO
    Posted Apr 21, 2006 at 12:48 PM | Permalink

    Jae, why don’t you do a sensativity analysis. How much would it change the validity if we find that balloons were right and ground stations wrong?

    a. What does it do to the error bars?
    b. Can you perform the same analysis (e.g. Mannian one) using a different training period and what does the end result look like. How much does it change.

    This is what Burger and Cubasch (my heros*)would do.

    *Ok…Steve is my hero too! :)

  14. Dave Dardinger
    Posted Apr 21, 2006 at 1:07 PM | Permalink

    re:#13

    I think you’re on to something there, TCO. If the Hockeyteam is at all correct in their climate reconstructions, then why can’t we assume that their reconstruction of say 1400-1500 is correct? If so then use it as the training period and go back to the original pool of proxies and do a new PC analysis. Then reconstruct to the present using this analysis and see what the “instrumental era” would be predicted to be from that standpoint. If it doesn’t give the actual instrumental record then that implies the method isn’t reliable. Of course if it does that’s a feather in the HockeyTeam’s cap. But somehow I don’t expect that will be the case or we’d have seen it already from them. They’re not dummies even if they are rather biased.

  15. Ross McKitrick
    Posted Apr 21, 2006 at 1:15 PM | Permalink

    #6: you can get the tabulations in our E&E paper, and the spreadsheet at http://www.uoguelph.ca/~rmckitri/research/MBHproxydata.xls has various breakdowns including the proxy availability in each step.

  16. TCO
    Posted Apr 21, 2006 at 1:16 PM | Permalink

    I think that is a differnt (but interesting) idea from what I’m suggesting. Would need to think for a while if failure (or passing) of such a test is relevant.

  17. Lee
    Posted Apr 21, 2006 at 1:18 PM | Permalink

    Well, no Steve. That is a misrepresentation of what I said.

    What I said was that I found it distressing to read, as a stated reason for your making those guesses, implications of dishonesty. It wasn’t the precise guesses I was objecting to, or even the fact that you were guessing at all. It was your assumption that the selection would be based on fraudulent grounds that I objected to.

    You said:
    Just for fun, I’m going to try to guess what the Hegerl et al sites are, from the limited information available and from the other guiding Hockey Team principle -maximum non-independence of data.
    -and-
    “…there’s a good chance that they used the Yang composite – it has the added attraction to the Hockey Team of including the Dunde and Guliya ice cores, which impart a HS to it.”

    These constitute statements that you are guessing based on an assumption that they are choosing so as to get the data that gives them their desired ends, for the reason that it gives them their desired ends. This is a direct claim that they are selecting their data for particular results, and it was that claim, which IMO constitutes a veiled accusation of academic dishinesty, made in the ABSENSE OF DATA SUPPORTING YOUR CURRENT CLAIM, to which I objected.

    My initial statement regarding this, in my first post in that thread, was as follows, embarassing typos and all:
    “The fact that it is obviosu that this is critical information, it is referenced as not yet published, is referenced as being submitted, and is clearly a sufficient body of work to deserve independent publication, argues strongly agaisnt the imputation that anything is shady here, as Steve strongly implies and as some of the responses are pretty overtly saying.”

    And further down the thread:
    “I was referring to a first impression of style and substance. And the entire middle of Steve’s post contains guesswork based on an hinted on (at least) assumptions that the authors were going to be cherrypickng data based on desired conclusions. How on earth is that NOT an implication of dishonesty.”

    Note that I am not complaining about the “guessing” of the series. I am talking directly about the imputations insinuated or made toward the authors. Not the complaints that we’d like to see the data, nor the complaints about about withholding of data, which may or many not be valid and are valid criticisms, nor about their choices or their analysis, which may or may not be valid, and not about attempts per se to guess the series. But rather, about what to me looked like a pretty overt imputation of dishonesty made in THAT article, with no supporting evidence for a claim of dishonesty there to back it.
    -
    And now, you choose to mention me, by name, in the first sentence of a new article, and in doing so misrepresent what I said and what my point was. And this, as you mention, after I’ve made only a handful of posts as a new poster here. I could care less about your guessing at which proxies they used, I did not say or insinuate that yor guessing was itself insulting to anyone. I did not complain in any way about your *guessing* of the proxies. My argument was based on the REASONS you were using for guessing the proxies, and the imputations behind it. And that holds whether or not you guessed the series correctly; if you are in fact correct in your guesses, the fact that you picked the set for one set of reasons is NOT in itself any kind of evidence that they used the same reasons, and imputations of dishonesty require and deserve a lot more evidence than that. And yes, I’m belaboring the point; I want to make sure it is perfectly clear.

    Now you did, IMO, accuse them of dishonesty. When I pointed that out, and that the evidence in your article does not support such an accusation, you respond in the first sentence of a new article by misrepresenting what I said, in quite a fundamental (and if I may say so, insulting) way.

    For someone who is evaluating whether this blog (as opposed to the literature itself, which is always worth understanding and evaluating) is honest, accurate, and rigorous enough to be worth the effort of digging through it and understanding the arguments, this is not a promising start. At least. And an apology would be nice.

  18. Lee
    Posted Apr 21, 2006 at 1:28 PM | Permalink

    My response to your misrepresentation of what I said has just been blocked by your spam filter. I can not easily find an email address or admin response link to use in order to get my response posted. Please provide, or point to if I am missing it, such a method.

  19. Dave Dardinger
    Posted Apr 21, 2006 at 1:45 PM | Permalink

    re:#16 The question is whether the proxies used to produce the original “training period” would be allowable in the backward analysis. I’d think they’d have to be, but it would probably be worthwhile doing both ways and by dropping a couple of them just to see what it looked like. The trouble is that if one or two proxies so imprint the initial hockeystick they’d likely imprint all parts of it to some extent and therefore would given high rankings in training against another part of the run. But I expect they’d be given rather a lower ranking; after all, that’s what the PC4 thingee was all about, wasn’t it? Choosing another “proxy” training period would just be pushing them further down the slope as they wouldn’t be able to use their “fitness” to the instrumental record at all.

  20. Lee
    Posted Apr 21, 2006 at 2:15 PM | Permalink

    Am I still blocked?

  21. Jean S
    Posted Apr 21, 2006 at 2:33 PM | Permalink

    re #10: Thanks! Well, I also may or may not ;) have some qualifications for the time series analysis, but I think that may be only an obstacle… these guys have demonstrated several times (in my opinion) rather amatourish statistical understanding. So what to look: something simple and “intuitively appealing” rather than any sound and time tested thing, something that you might not want to let other people know later…

    re #15: Thanks Ross! This is what I was looking for!

  22. Jaime
    Posted Apr 21, 2006 at 3:53 PM | Permalink

    Re #7

    Of course, Nature would at this point consider you an avowed enemy of climate science, so is there some other journal that would appreciate zinging Nature by publishing your criticism?

    Hmm… Yes. Britannica comes to mind. Oops! Sorry, not a journal!

  23. Lee
    Posted Apr 21, 2006 at 5:09 PM | Permalink

    Just checking to see if I am still blocked from Climate Audit.

  24. Steve McIntyre
    Posted Apr 21, 2006 at 5:38 PM | Permalink

    Lee, we have a Spam filter on the site. These sorts of sites often get dozens of spams per day and it’s impractical to manually check. One of the things that Spam Karma checks is a sudden increase in posting frequency and new posters tend to get blocked. It recognizes you after a couple of manual recoveries. I’ve recovered several posts. Let me know if I’ve missed any. Email me if you get blocked – contact info in right frame.

    I’ve edited out the reference to your post on the basis that I’ve misconstrued your remark.

  25. TCO
    Posted Apr 21, 2006 at 6:22 PM | Permalink

    Lee (you ********, I don’t have proof that the Iranians are building a bomb. But I would put money on it. Do you understand the difference, you ****? Oh…and if you’ve read anything about human tendancies in science, you would know that it is a common (and useful to watch for) failing for people to skew data to serve their ends, consciously, unconsciously and shades in between. Of course, you would ****** to understand that. ***

    Steve: c’mon TCO. I can’t be bothered editing like this.

  26. Steve McIntyre
    Posted Apr 21, 2006 at 6:29 PM | Permalink

    Here’s what I said:

    Just for fun, I’m going to try to guess what the Hegerl et al sites are, from the limited information available and from the other guiding Hockey Team principle -maximum non-independence of data.

    Lee, you said about this (and a comment about the Yang selection):

    These constitute statements that you are guessing based on an assumption that they are choosing so as to get the data that gives them their desired ends, for the reason that it gives them their desired ends. This is a direct claim that they are selecting their data for particular results, and it was that claim, which IMO constitutes a veiled accusation of academic dishinesty, made in the ABSENSE OF DATA SUPPORTING YOUR CURRENT CLAIM, to which I objected.

    Lee, don’t put words in my mouth. I was trying to guess the 12 proxies used in Hegerl et al out of the hundreds of proxy series at the World Data Center for Paleoclimatology with minimal information.

    For the purposes of my guessing, I had to have a method for picking 12 proxies. I simply stated my method – that the choices would be maximally non-independent; using this hypothesis, I examined recent selections made by Osborn and Briffa and, from the 14 series selected by Osborn and Briffa, chose the 12 series that seemed most likely to have to have been chosen by Hegerl et al. I tested this hypothesis by then applying 12 series from Osborn and Briffa and comparing the result to the CH-blend.

    I made no comments about what their motives were for selecting these series. I merely described an operational protocol by which I was guessing at what was in the CH-blend. If someone like yourself wishes to evaluate what motives would lead to near-identical selections, you are free to do so. I myself deliberately refrained from any such speculation.

  27. JerryB
    Posted Apr 21, 2006 at 6:56 PM | Permalink

    “Well, no Steve. That is a misrepresentation of what I said.

    What I said was that I found it distressing to read, as a stated reason
    for your making those guesses, implications of dishonesty. It wasn’t
    the precise guesses I was objecting to, or even the fact that you were
    guessing at all. It was your assumption that the selection would be
    based on fraudulent grounds that I objected to.”

    Your few posts suggest that you suscribe to some notion of the efficacy,
    and accuracy, of the art of mindreading. I do not, and I suspect that
    most, if not all others, of the regular visitors to this blog also do not
    so subscribe, which you may regard as quaint, but so be it.

    In any case, your statment “That is a misrepresentation of what I said.”
    exhibits such a notion in that it suggests that other visitors to this
    blog can magically percieve to which “that” your refer.

    It is rather late in my timezone for me to go in to much detail, but let
    me mention that in case you might prefer to be perceived as other than
    pathological, you might try to be somewhat less preposterous.

  28. Pat Frank
    Posted Apr 21, 2006 at 7:12 PM | Permalink

    #17 “These constitute statements that you are guessing based on an assumption that they are choosing so as to get the data that gives them their desired ends, for the reason that it gives them their desired ends.”

    If you’d been around here for a while, and had seen the analyses of one “independent” proxy study after another, revealing their lack of independence, you’d not have thought Steve’s comments to be at all untoward or factually unjustified. You were, in short, jumping to a conclusion.

  29. Lee
    Posted Apr 21, 2006 at 7:25 PM | Permalink

    OK, enough. I’m out of here after this.

    Steve, are you seriously arguing that "the other guiding Hockey Team principle" and "it has the added attraction to the Hockey Team of including the Dunde and Guliya ice cores, which impart a HS to it" are not "comments about what their motives were" and (at the very least) overt insinuations that they are selecting their data in order to get the desired result?

    And JerryB, before Steve edited it out (thanks for the additional display of basic dishonesty, Steve), the very first phrase of the very first sentence of the article by Steve that leads this thread refers to me by name and attributes to me a particular statement. IF Steve has nay honesty a tall he will put that first sentecne back into the article. It seemed that my first post responding to an article that mentioned me by name, where I respond to Steve by name, might have been interpreted by any reasonable reader as referring to the article to which it was posted as a response. No mindreading required. But Steve (that paragon of honesty) edited the article before finally the spam filters, which he claims accidentally trapped me as a spammer after a total of three posts to the site.

    I beleive I have learned everythign I need to know about the honesty of this site, and I’m gone.

    Steve: As to spam filters, that’s how they work. I don’t know how to whitelist someone; I can manually recover posts which is what I’ve done. You complained that the sentence had misrepresented you, so I removed it. I thought that’s what you wanted. If you want it back in, I’m happy to oblige. I don’t exactly remmber what the sentence was, so I’ll do the best recovery that I can off hand – feel free to correct it.

  30. Doug L
    Posted Apr 21, 2006 at 7:26 PM | Permalink

    Re #27

    re: “it suggests that other visitors to this blog can magically percieve to which “that” your refer.

    Maybe it’s so late in JerryB’s time zone he/she didn’t notice “that” has been removed.

  31. TCO
    Posted Apr 21, 2006 at 7:32 PM | Permalink

    You know it’s amazing how much time we waste on this bs versus discussion of the science. I’m really happy that Steve brought this paper out. The f***ed up confidence levels are histarical. I like the science.

  32. JerryB
    Posted Apr 21, 2006 at 7:37 PM | Permalink

    Re #29,

    Doug.L,

    FWIW, I’m a he,but let me congratulate your expression of awareness of possible differences of perception. Otherwise, let me ask which “that” has been removed.

  33. Doug L
    Posted Apr 21, 2006 at 8:17 PM | Permalink

    Re # 31

    Jerry, I can’t give a quote, it’s gone. Adding to the confusion, Lee’s post was added out of order because of spam filter. He/she said:

    “And now, you choose to mention me, by name, in the first sentence of a new article [this one], and in doing so misrepresent what I said and what my point was.”

    Later Steve indicated in #24 he removed it, thus proving that Lee is not preposterous :-).

  34. mark
    Posted Apr 21, 2006 at 8:26 PM | Permalink

    Part of the problem with referencing #s in a spam controlled blog… confusing indeed. For a sec, I thought Lee = JerryB… ??? :)

    Mark

  35. John M
    Posted Apr 21, 2006 at 8:55 PM | Permalink

    For what it’s worth, I just started to post about a week ago and got a couple of my posts gobbled up by the spam monster too.

    I wasn’t being that critical of anyone who posts here, it’s just the way the thing works.

  36. bkc
    Posted Apr 21, 2006 at 10:30 PM | Permalink

    I think rather than impugning their honesty, Steve M. is subjecting the authors to ridicule, which perhaps offends Lee even more so.

  37. Willis Eschenbach
    Posted Apr 21, 2006 at 10:51 PM | Permalink

    Lee, I understand your confusion about the tone of the site.

    Unfortunately, a large number of us have been badly jerked around by various members of the “Hockey Team”, as well as the editors of several “scientific” journals, and by those members of the Hockey Team who run the (Un)RealClimate site. This has resulted in a somewhat cynical approach to whatever the newest, latest, and greatest dendroclimate study from the Hockey Team might happen to be. In particular, based on bitter past experience,we expect them to use questionable proxy series, and to hide, obscure, or be coy about what those series actually are. In this expectation, we are rarely disappointed.

    Since Hegerl have not disappointed us, and since they have once again hidden the nature of their data, we come Steve’s light-hearted attempt to guess what the Hegerl secret blend might contain. Now stop and think about what Steve having to guess their data source means, Lee, and consider this question:

    WHY SHOULD ANYONE HAVE TO GUESS?

    I mean, this is (supposed to be) science, where transparency is the watchword. But they’re still trying to hide their calculations, methods, and data, even after falling on their faces the previous times they’ve tried that.

    Now, you have said that you don’t like what you think Steve is saying about the motives of Hegerl et. al.

    In that case, perhaps, Lee, you might offer up to us your take on what you think might be the motives of “scientists” like Hegerl et. al. who hide their data? Are they doing it because they are so honest? Because they are so ethical? Because they are so devoted to science?

    I invite you, Lee, in all seriousness, to consider this question. It will help understand the tenor and the tone of this site.

    I await your answer, and I thank you for your comments and your honesty about what you feel is happening here.

    All the best,

    w.

  38. Dave Dardinger
    Posted Apr 21, 2006 at 10:53 PM | Permalink

    bkc, you’re on to something there. Being made fun of by people you dislike is always worse than being called a fool or dishonest. OTOH, these others should have thought about that before they started insulting Steve first. Lee wasn’t aware of how ridiculous a lot of the excuses made by the hockey team on why their data isn’t available is. I think someone did a list one time here. As far as I know none of the team have claimed their dog ate it, though if if was stored on a floppy disk or CD….

    But the point Steve was making initially was that as he’s pried more and more of the usual data suspects into the open, it’s getting harder and harder to hide what’s going on. Now he can make good guesses on what’s going on without any additional help and was tweaking a few noses to let them know it!

  39. McCall
    Posted Apr 22, 2006 at 12:00 AM | Permalink

    Enough thread space has been wasted on Lee and his views — let Mr. McIntyre’s earlier prediction of 6 out of 12 stand on its own. I just want to know what Hegerl used — the pattern of non-disclosure and non-independent HS verification is likely to most here except Lee, but would likely be so to even him at 6 of 12 hits. If less than 6 of 12, than more analysis is warranted to make the non-disclosure/non-independence/BCP-proxy case; but I suspect that will be made even with lower hits (why hide the proxy list).

  40. HANS KELP
    Posted Apr 22, 2006 at 6:17 AM | Permalink

    Honestly, I think that posters like mr. Lee should very much consider what kind of blog they´re going to submit posts to before they involve themselves and jumping to conclusion as mr. Lee did because of a few remarks Steve M. has done. But maybe mr. Lee was just another Hockey Team zealot, who in an odd way just wanted to tell Steve M. that he doesn´t like him because of what Steve M. is doing to the HS-teams. I don´t know. What I do know is that this site is all about auditioning of “scientific” proxy research of which results, and use of , has been proven by Steve Mcintyre to be dubious and sloppy at best. Some are good and some are bad, but too many of the researcher s in the proxy field have by their stonewalling and obfuscation of Steve M. proven to be dishonest in a way that is outright disgraceful, not alone to Steve M. but certainly to the field of climate science as such, but that fact is obviously not a consideration of the ethical mr. Lee. So mr. Lee, before jumping to conclusion I would advise you to put your ethical ego aside and instead “dive” into this blog and see what actually is being thrown at it and watch the responses and I´m shure you will come to change your mind a bit about Steve M. and CLIMATE AUDIT. And just one more thing for you mr. Lee. I point you to that what you accuse Steve M. of is what you actually are doing very very much yourself . You write

    “I beleive I have learned everythign I need to know about the honesty of this site,
    and I’m gone.”

    As for me, a site includes also the users of the site , posters and readers as well. Authors, posters and readers make the “site ” along with the issues contained herein and without them all,- no site. If your remarks about honesty of this site is to be taken for what it says, then you are actually accusing me along with all of the other users of this site to be a bunch of dishonest people, because thats exactly what you imply. As you so accuse people you are not familiar with of dishonesty , is to me much worse than what you accuse Steve M. of doing about people he is familiar with , and it implies to me that either you must be god, which I consider rather unlikely, or you are just a poor misguided human soul which I have to forgive for it sins. So dear mr. Lee, I think you should reconsider your attitude to this “site” and I would personally expect and accept an apology from you.

    All the best.

    Hans Kelp

  41. TCO
    Posted Apr 22, 2006 at 6:45 AM | Permalink

    If Steve gets more than half right, I will buy him an expensive steak dinner (wine included). If the opposite, he has to do me. Ok…he doesn’t have to do me–I’ll get a gf for that. He has to pay for what I can drink and eat. And there will be some gin in the mix. If it’s at 50%, we go dutch. Let’s see what happens, competetive squash guy….

  42. Posted Apr 22, 2006 at 7:32 AM | Permalink

    It is possible that Spam Karma blocked Lee, but it is also possible that it was done manually by Steve or John. I can’t see any way to know for sure.

    MCall: how about a bit of transparency from you? Have you conspired with Lott on his lawsuit against Levitt?

  43. TCO
    Posted Apr 22, 2006 at 7:42 AM | Permalink

    Is it possible that Tim Lambert kills babies and sucks Al Queda shlong? Hard to tell…

  44. Posted Apr 22, 2006 at 7:48 AM | Permalink

    TCO: I see that it went right over your head.

  45. TCO
    Posted Apr 22, 2006 at 8:04 AM | Permalink

    Don’t get all intellectual on me, bitch.

  46. Dave Dardinger
    Posted Apr 22, 2006 at 8:10 AM | Permalink

    Well it sure is fun watching Tim be Tim and TCO be TCO. I think I’ll go for a long walk with my wife.

  47. TCO
    Posted Apr 22, 2006 at 8:13 AM | Permalink

    I want a wife also.

  48. John M
    Posted Apr 22, 2006 at 8:24 AM | Permalink

    What are the odds?

  49. Steve McIntyre
    Posted Apr 22, 2006 at 8:30 AM | Permalink

    #41. Tim, I’m happy to work up a bet more or less along the lines that you propose. This is more fun than wrangling. I was actually going to offer out a bet, suspecting that you’d be likely to respond: the terms that I had in mind were 6 or more – my win; 4-5 a draw; 3 or less, I lose. I realize that I didn’t post up those exact terms, but it’s what I had in mind.

    I’d like to play with my list a little before finalizing – I may want to vary the Quebec series, but I’ll present a final list tomorrow. One other issue – there are a couple of versions of the Alberta series and the Polar Urals/Yamal series. It’s possible that Hegerl used the older Luckman version of Alberta rather than the newer Luckman-Wilson verion; likewise, she might have used the older Briffa version of Polar Urals, rather than the newer Briffa version of Yamal; Esper has a Polar Urals version that differs from both. My terms would be: if I get the site right, that’s a plus for me; I don’t have to match the exact version.

    So let’s say the following: US$150 covering the cost of a good dinner for two; money to be wire transferred in advance to an agreed referee; 6 or more- I win, 5 – a draw, 4 or less you win; site identification is the criterion not the exact version; I can slightly modify my list up to 8 pm tomorrow my time, not to exceed 3 variations.

  50. TCO
    Posted Apr 22, 2006 at 8:34 AM | Permalink

    dude. I’ll do it.

    BUT:

    1. 50% is a wash.
    2. you have to freaking break bread with me. Mano y farking mano.

  51. TCO
    Posted Apr 22, 2006 at 8:36 AM | Permalink

    And don’t worry about the wire transfer. I’m a f****** USNA grad. You can trust me. And I’ll take me chances with you.

  52. jae
    Posted Apr 22, 2006 at 9:10 AM | Permalink

    Methinks Lee is an avid HST fan and he is having a very hard time facing the truth.

  53. Lee
    Posted Apr 22, 2006 at 11:25 AM | Permalink

    Response on a few points, before I dive into Saturday housecleaning with my family.

    First, I have never before experienced being cited by name, in what is really a publication available to the public, and in the first phrase of the article no less, and being simultaneously rendered unable to respond. BTW, the Contact Steve link takes one to another blog thread, with no indication that it is treated any differently by the spam filters, and NOT to an email link. Steve, when you point people to the email link on the blog, and there IS no email link, it isnt all that helpful. It appears that you monitor posts to the contact thread, but you need some better way to indicate that this actually does serve as a way to get through to you even if the spam filter says it is filtering that thread, which it does.

    Regardless of how it was that I ended up being blocked from the site after being misrepresented by name in a top article, it is incumbent upon the adminstrator to ensure that you do not in future allow such a thing to happen; at best, it reflects very badly upon you.

    Second, as I indicated at least once in my posts, it IS a reasonable (at least) expectation and requirement that the source of their analysis is public. I **said** that. I also pointed out that the relevant source is in an article under submission, which they say has the full details and rationale for their selection and use of that set. A separate publication is reasonable, IMO. It is also incumbent upon them to work to get that info published in a reasonable time scale, and to ensure that it becomes available if it does not get published. I never said otherwise.

    None of which is applicable to the initial issue I raised, which is that Steve was, at least indirectly, making a very serious accusation of academic dishonesty, which was not backed by the evidence under consideration but was being backed by insinuation. Such accusations are VERY serious; if Steve wants to make them, he needs, IMO, to create a detailed analysis in a single avaiable article or essay of the basis of those charges, and make them available for analysis, and cite them whenever he makes that charge. And if he is not willing to do so, then he needs to stop making those insinuations. This is a more clearly stated version of my point (with some added advice for how to properly handle accusations of academic dishonesty, which is what they amount to), but it WAS my point, and was pretty clearly stated from the first.

    Third, I have not yet come to a conclusion about the “Hockey Stick” analysis and results. That is part of why I’m looking at what peope have to say. I am convinced by the evidence that we are at this moment rapidly warming and have been for a few decades, that absent unknown mechanisms the current statement about the effects of carbon, that the most probable climate sensitivity to CO2 doubling is about 3 +/- 1.5 C (with that lower limit pretty solid and the upper limit possibly subject to upward revision) is probably reasonably solid, and that therefore the known CO2 modification of the atmopshere is probably going to be taking us into new climate territory, if it hasn’t already.

    From that perspective, the “Hockey Stick” analysis is clarification of detail, not any kind of key evidence in the AGW story. If true, that HS detail *is* illuminating, which is why I’m interested. The bald guesses about my state of mind here are simply wrong. I am gathering primary papers, but I am not a dendrologist, nor am I qualified in the statistical analysis of time series (although I know somethig aobu tthe area in general), so I am not qualified to judge the accuracy of those reports. Thus my interest in various science-oriented sites.

    When my first introduction to the most visible ‘contrarion’ site (in absense of a better word), involves snide unsupported accusations of fraud, mispreresentations of what I said **by name** in a new top article, misrepresentations by Steve that he wasnt saying anything about their motivations at all (I cited this in a post further up), an inability for whatever reason to respond to being cited by name, an editing (for whatever reason) of the original article such that when my response was let through it was rendered partly incomprehensible, and various attacks by readers here on MY motivations, I have to say that, at least, I am not impressed.

    And one last point, on the “unReal Climate’ bit. I will note that the current top article at RC is a post DEFLATING overstated claims in the press, and pointing out that it is incumbent upon the scientists to work with the press to try to keep such things from happening. I’ve seen their posts disputing slowdown of teh atlantic conveyor, and siputing links between warming and atlantic hurricane frequency as well. Are they some kind of perfect oracle? Of course not. They are a useful resource, with links to the things they cite, and with some care not to overstate and to take on overstatements. My opinion.

  54. ET SidViscous
    Posted Apr 22, 2006 at 11:31 AM | Permalink

    You expect anyone to read through that whinging.

    Can you say S_U_C_C_I_N_C_T

    No, How about C_O_N_C_I_S_E

  55. Steve McIntyre
    Posted Apr 22, 2006 at 11:49 AM | Permalink

    #41, 49. Here are my final picks. There are a couple of changes after looking at Darrigo et al 2006 as well as Osborn and Briffa, since a couple of the choices were not maximal overlap:

    Yang composite
    Taymir
    Polar Urals – on balance, I expect the Briffa MXD version, but it could be the Yamal version. Either qualifies.
    Mongolia
    Tornetrask
    van Engeln
    Greenland dO18
    Jasper – on balance I expect the Luckman version, rather than the Luckman-Wilson version which is too recent, but either qualifies.
    MBH PC1

    There are 3 series in my previous guess, which were simply taken over from Osborn and Briffa without fully weighing maximum overlap. I am making three substitutions to be more consistent with usual Hockey Team choices.

    Yakutia instead of Mangazeja. Mangazeja is a bit unusual; Yakutia is more consistent with maximal overlap and is used in DWJ06.

    Quebec – I’m going to go with the Jacoby version rather than
    Schweingruber version (which isn’t used outside of Esper/OB). either. One point for the Jacoby version; half point for SChweingruber version.

    Tirol isn’t used outside of Esper/OB and isn’t maximum overlap. So something from the Jacoby treeline series, it could be the composite or it could be something like TTHH. One point for TTHH, half-point for the composite or half-point for another Jacoby treeline site. I’ll be mad if it’s Tirol.

    As amended, I’ll go with a saw-off at 6 points, more than 6 points I win, less than 6 points I lose. Are you on, Tim??

  56. JerryB
    Posted Apr 22, 2006 at 12:10 PM | Permalink

    Re #33 Doug L,

    Steve had to wade through most of Lee’s fantasies in that comment, including Lee’s misrepresentation of what Lee said in the previous thread, to get to the sentence that you quoted, so I regard the “that” as having been ambiguous, especially in view of Lee’s apparent faith in mind reading.

  57. ET SidViscous
    Posted Apr 22, 2006 at 12:16 PM | Permalink

    I don’t think it’s faith in mind reading. I think it’s an inablility to get a point across.

    First person I’ve ever seen who’s writing style is best described as blancmange

  58. fFreddy
    Posted Apr 22, 2006 at 12:21 PM | Permalink

    Lee, glad you returned. Taking you at your word from your earlier posts, I’m assuming that the initial source of your irritation is that you are experienced in an area of science where the system works as it supposed to, and you are assuming that the same is true of climatology.
    If you take the time to read through this site, you will come to the same conclusion as most of us here hold: the system has been grossly polluted by political activists, and you need to check everything you think you know on the subject.
    In particular, I am entirely unconvinced about 1.5C for CO2 doubling. You might like to ask yourself why you regard it as solid – is it just because of what you have seen in the non-specialist media ? If so, you are, of course, doing the right thing by looking for primary sources. I look forward to hearing your view once you have had chance to read up on this blog.

    Regarding RealClimate – you are right that the current top post is about deflating excessive claims by supposed scientists, but you should note that it is in response to a BBC radio program exposing these excessive claims. If you have not already done so, you can listen to that program on the BBC’s web site for the next couple of days. I recommend that you do.

    Regarding the use of your name – I hardly think one syllable is a particularly strong identifier on the wide world of the interweb. But if you don’t like it being used, you could always use an alias. After all, I am not posting under the name my mMother called me.

  59. Lee
    Posted Apr 22, 2006 at 12:53 PM | Permalink

    fFreddy,

    It isn’t even so much that my name identifies me; Lee is pretty common, even if restricted to biology. And I don’t mind it being used; the only reason I don’t use my full name is that I have to wade through enough irrrelevant stuff evry day as it is,a nd as Steve says, spam is pervasive. In areas where I do feel I’m staking my reputation, let key people at least know who I am.

    But it was a body of ideas identified with “Lee” as an identifier that was being called out. I’m not worried about my reputation when I was doing science, I did good work that had an impact, and that stands on its own. I do get perturbed when what I say is misrepresented in a way that diminishes the point I’m trying to make.

    I’m willing to entertain that there are reasons to doubt this science, or even to believe that it is dishonest. But saying so my unsupported insinuation, EVER, rather than marshalling the support and making the argument available in a piece (which Steve could theen cite whenever he wishes) seems to me to be avoiding a very serious responsibility that comes with seriously making such a charge. Insinuating it rather than charging it and documenting it is, IMO, the improper way to deal with it if he feels there is a serious issue.

  60. TCO
    Posted Apr 22, 2006 at 12:59 PM | Permalink

    Well…go read the board, then, Lee. Get cracking, ********John says: This is your last warning TCO. Stay away from colorful metaphors and I’ll stay away from the spam button.

  61. fFreddy
    Posted Apr 22, 2006 at 1:11 PM | Permalink

    TCO, please go and sober up.

    Lee, I am not a professional academic, and I don’t know what is the correct approach to dealing with this problem.
    If, having reviewed the site and come to your own conclusion, you agree that something is violently wrong, then I would be very interested to know what you think is the appropriate way to deal with the issue.

  62. Dave Dardinger
    Posted Apr 22, 2006 at 1:33 PM | Permalink

    Regarding the use of your name – I hardly think one syllable is a particularly strong identifier on the wide world of the interweb. But if you don’t like it being used, you could always use an alias. After all, I am not posting under the name my mMother called me.

    More importantly, the responses to these blogs are just as public as what Steve posts at the top of a thread. I just did a search on Google on dardinger +”Climate Audit” and got 529 hits. Hardly an anonomous set of postings.

  63. John A
    Posted Apr 22, 2006 at 2:41 PM | Permalink

    Re: #62

    And there could be thousands of “Dave Dardingers” out there who have an interest in climate science

  64. Doug L
    Posted Apr 22, 2006 at 2:47 PM | Permalink

    Re #56

    JerryB,

    I appreciate your difficulty at reading minds. Otherwise you’d know, I knew you had other some other point. I was more interested in alerting readers that info was being subtracted and that it could make your comment seem unfair.

    It may be of interest that jurors are routinely asked to read the minds of defendants. My knowledge of law is rudimentary and a bit rusty on these points, but a pattern of behavior could be construed as malice which could be an element needed to be proven in a libel suit in some jurisdictions. Also in the UK libel laws are much stricter than the US, they might not require that malice be demonstrated in the UK.

    Consequently, readers of the blog may have quite different expectations about what kinds of insinuations are legally actionable. What may seem preposterous in the US may be serious business in the UK.

    Having not been quite as succinct as I could have been, let me just add that while Steve’s defense sounds quite adequate to me, a sharp lawyer might well call it “cute” in some jurisdictions, in others I suspect they’d never get a chance to.

    Apologies if I’m magnifying the seriousness of the alleged insinuations with all this legal talk. I have no clue whether these proxies MUST be independent in this case. It certainly doesn’t sound unusual to me, a regular reader of this blog, and non scientist.

  65. Dave Dardinger
    Posted Apr 22, 2006 at 3:07 PM | Permalink

    re: #63 There might be thousands of “lee”s with an interest in climate science, but since there’s ever only been one David Dardinger in the entire world, AFAIK, if you find someone of that name, it’s me. Though with the simple surname Dardinger there’s also a cousin’s daughter who teachs environmental stuff at Cornell. I suspect she’s a warmer. I’m not sure I’ve ever met her, though it’s possible she was at some Dardinger Reunions years ago.

  66. John A
    Posted Apr 22, 2006 at 3:10 PM | Permalink

    Dave,

    I was being ironic.

  67. Spence_UK
    Posted Apr 22, 2006 at 3:57 PM | Permalink

    OK sorry guys I’m fanning the flames tonight.

    Second, as I indicated at least once in my posts, it IS a reasonable (at least) expectation and requirement that the source of their analysis is public. I **said** that. I also pointed out that the relevant source is in an article under submission, which they say has the full details and rationale for their selection and use of that set. A separate publication is reasonable, IMO. It is also incumbent upon them to work to get that info published in a reasonable time scale

    OK, this is a good thing. You recognise the importance of this. Just out of curiousity, what is a reasonable time scale? It took something like six years to get the real data and methods behind MBH98 – and MBH98 is one of the studies we know the most about.

    First, I have never before experienced being cited by name, in what is really a publication available to the public, and in the first phrase of the article no less, and being simultaneously rendered unable to respond.

    I don’t see you’ve been rendered unable to respond. How many posts have you made that have not appeared? Many blogs operate moderator queues and spam prevention mechanisms, so to assume this just because your post did not instantly appear strikes me as being rather naive. Contrast this with Steve, who has also been named on RealClimate in posts, and who has been refused the ability to respond – not just waiting hours for posts to appear, but months and years. I’d say you’ve been treated pretty fairly. Also I have e-mailed Steve on a number of occasions, using the e-mail quoted on this site. You have to look for it admittedly, but I managed without great difficulty. You can even get a correspondence address for snail mail if you look carefully enough. I hope this isn’t indicative of your typical level of research on a topic.

    But on to your substantive point:

    None of which is applicable to the initial issue I raised, which is that Steve was, at least indirectly, making a very serious accusation of academic dishonesty, which was not backed by the evidence under consideration but was being backed by insinuation.

    This is your opinion Lee, and not one I particularly share. Steve stated his methods, not those of Hegerl/Crowley. If anything, this little thought experiment is far more likely to damage Steve’s credibility (if he gets it wrong) compared to Crowley/Hegerl.

    Even if Steve is right, it is far from an accusation of scientific dishonesty. It doesn’t prove anything – merely throws up the fact that aiming to get a hockeystick produces a coincidental data set. I’ve pointed out on a few occasions the potential for subconcious bias in proxy picking evidenced from the thinking processes of some of the scientists on RealClimate. Subconcious bias is not a “serious accusation of academic dishonesty”, but an observation about studies that rely heavily on statistics with poorly defined sampling criteria. Most multi-proxy studies fall into exactly this category, something that could be trivially resolved with a decent study protocol. This kind of thing is common place in the field of medical statistics (google study protocol and you’ll see what I mean). Why is it lacking in the temperature reconstructions?

    As you can see, you have to conveniently ignore a lot of reasonable possibilities to draw your conclusions that Steve is insinuating scientific dishonesty. When that situation arises, you have to recognise your original point is more than a bit of a stretch.

  68. Hans Erren
    Posted Apr 22, 2006 at 4:28 PM | Permalink

    I think journals should adopt the strict policy that only published papers can count as reference.
    Saves a lot of searching, I came across “In press” references that actually were never published, quite embarassing.

    So in the case of Hegerl et al the publication in Nature should at least have been delayed.

  69. McCall
    Posted Apr 22, 2006 at 4:42 PM | Permalink

    re: 49 (to 41) TCO is also Tim?

    re: 43 Huh?

    re: 42 No.

  70. Lee
    Posted Apr 22, 2006 at 4:51 PM | Permalink

    Spence,

    You disagree; fair enough. I will simply repeat that it isn’t the fact that Steve attempted to guess what they used based on some criteria taht was teh reason bfor my belief, but rather his statements about what criteria he used (and therefore, clearly IMO, suspected they used) that form the basis of my argument. I’ve already made the point, indlucing citing the specific statemsnt and why I thought they wer problematic, several times. So I’ll leave it at that.

    On the delayed response: My first posts at this site WERE added instantly, with no perceptible filtering or delay.

    But only after this particular article was added, calling out what I had previously said, only from that point on was every post I made for several hours blocked from the site. BTW, the time stamp on my posts in this thread was the time I posted them (and they got filtered); it was several hours later that they were all added back en masse.

    Meanwhile, several other people WERE responding. Some of them were responding specifically to the attribution of my previous posts, to the misrepresentation of what it was claimed I had said, and their posts WERE getting through, but mine were not. It wasnt just the fact that my posts were delayed per se; it was the fact that other’s posts were getting through while mine were not.

    I am willing to be charitable and credit that it was a comedy of errors, but (especially after responding to someone else’s statement with attribution, regardless of the semi-anon stuff) it was a serious set of errors. My point stands that Steve has an obligation to make sure it doesn’t happen again. Perhaps by addressing the spam filter over-sensitivity? By using the required email address to let someone know they are being cited before the article is posted, with an email response address to guarantee a route to respond? Some such thing. If Steve is trying to use this site as a serious and rigorous response to what he believes are serious issues, then he must treat it seriously and rigorously.

  71. Lee
    Posted Apr 22, 2006 at 5:30 PM | Permalink

    Oh, and BTW, I finally did just find Steve’s email address.

    Funny thing; when you click the “Contact Steve link and go to that thread, there is no email link. There is a single line of text email discription at the top, which is obscured as soon as one scrolls down to the post area in the thread.

    I suspect I’m not the only person who wouldn’t immediately figure out where that is.

    Oh, and I’m being filtered again.

  72. John A
    Posted Apr 22, 2006 at 5:35 PM | Permalink

    I’ve not had reason to mark any of Lee’s posts as spam. However Spam Karma does have a policy of newish posters having only five posts per day before it gets used to them.

  73. Lee
    Posted Apr 22, 2006 at 5:56 PM | Permalink

    JohnA,

    Unless I’m misremembering (possible; I had been surfing several sites and not paying all that much attention to which was which for a couple days) I believe it was my 4th post ever here that got blocked – and it was posted the next day from my first three.

    And I’m being blocked again right now; I emailed Steve about it half an hour ago.

  74. Terry
    Posted Apr 22, 2006 at 6:35 PM | Permalink

    Lee:

    Your contributions are appreciated. Please continue to grace us with your presence.

    You have to get used to the constant level of snarkiness in weblog comments. Usually about half the posts should simply be ignored. (Here as well as on RealClimate.) That’s just the nature of the beast. By focusing on the snark, you miss the valuable stuff.

  75. Steve McIntyre
    Posted Apr 22, 2006 at 7:05 PM | Permalink

    #70. Lee, the spam software has scoring system to try to detect spam. Sometimes we get over 50 spams a day and sorting them out manually would be impractical. At realclimate, you can’t post in real time; they censor everything. Here you can post in real time. As I mentioned to you, if a new poster gets suddenly active, the spam software thinks that it’s a spammer since that’s a symptom of how messages from poker sites and porno sites and mortgate sites operate. Maybe there’s better spam software, but that’s what we have.

    The software also will rach triggers when it "retro-spanks" all posts. So what’s happened is some posts are up; people reply and then when there’s a seeming overload, the spam software goes back and "spanks" all the posts from the IP address. I’ve gone back and recovered them.

    So you’re not being censored. I apologize for any teething issues with Spam software, but again that’s what we have. On balance, I think that it’s a much fairer and more reasonable system than realclimate’s censorship of all posts, where they are not simply filtering spam but censoring adverse comment.

  76. Steve McIntyre
    Posted Apr 22, 2006 at 7:36 PM | Permalink

    Lee, you say:

    None of which is applicable to the initial issue I raised, which is that Steve was, at least indirectly, making a very serious accusation of academic dishonesty, which was not backed by the evidence under consideration but was being backed by insinuation. Such accusations are VERY serious; if Steve wants to make them, he needs, IMO, to create a detailed analysis in a single avaiable article or essay of the basis of those charges, and make them available for analysis, and cite them whenever he makes that charge. And if he is not willing to do so, then he needs to stop making those insinuations. This is a more clearly stated version of my point (with some added advice for how to properly handle accusations of academic dishonesty, which is what they amount to), but it WAS my point, and was pretty clearly stated from the first.

    I don’t think that I’ve been shy about calling a spade a spade in formal presentations. Look at my comments to the NAS panel about Mann or my Reply to Ammann and Wahl submitted to GRL or my review of Ammann and Wahl for Climatic Change. In my opinion, both parties withheld adverse results. I presented detailed analyses in both cases. In my opinion, UCAR’s press release about Ammann and Wahl contained serious misrepresentations and I’ve complained about ito UCAR in detail to no avail. I don’t get the impression that everyone in climate science takes withholding of adverse results as a serious matter.

    Over the years, I’ve learned to avoid adjectives and to avoid trying to go a bridge too far. There’s a legal maxim that, if there’s an honest interpretation that you can put on something, you are entitled to presume that. Thus it was our presumption that Mann had probably made a programming error or inadvertent assumption in his principal components algorithm rather than that he had known what he was doing and intentionally misrepresented his methodology. Somewhat surprisingly, Mann has denied that there was any error or inadvertence in what he did.

    My projection for Hegerl et al. is that there will be very little "independence" in the proxy selection and that most, if not all of them, will be the same as similar studies, up to and including Mann’s notorious PC1. One of the issues that I hear a lot about is that there are supposedly a "dozen independent studies" which confirm Mann. My take on Hegerl et al, which was more sarcastic than anything, as one poster observed, is that it won’t be "independent" in its proxy selection. Does a lack of independence in proxy selection "imply" academic dishonesty? Not necessarily. One can envisage other explanations – I’m sure that the proponents would argue that these were the "best" proxies.

    My beef is really with the claims of "independence". If Hegerl et al were to claim that their study was "independent", then they’d be taking matters another step. But to my knowledge they haven’t made such claims yet. However, you might evaluate actual claims of independence made at realclimate or by Mann in his evidence to the Barton Committee and get back to me on whether you think that those claims can be made honestly.

    You might not be aware that Mann has written to a journal accusing me in writing of "dishonesty" and that Crowley of Hegerl et al has published a scurrilous and untrue account of our correspondence in EOS. So don’t shed too many tears for these folks.

  77. Louis Hissink
    Posted May 6, 2006 at 5:42 AM | Permalink

    # 76

    Just came back from the AIG 25th Year Conference. Bill Shaw, of Golder Associates, spoke about “auditing”. In terms of that my support for Steve here is even more focussed and strengthened.

    When it comes to funding I have discovered academics are as truthful as shonky mining entrepreneurs.

    I have asked Bill Shaw to summarise the requirements for auditing Oz companies might be required to present to future audits. Knowing part of what those audit requirements are, Mann et al fail though by how much would need to be determined in committee.

  78. bender
    Posted Aug 3, 2006 at 5:03 PM | Permalink

    On proxy interdependence & Hegerl et al. (2006):

    All four of the recons used in Hegerl et al. (2006) [refs 11,12,14,15] use BCPs, correct?

    It would be great to see a continually updated version of Wegman’s Figure 5.8, p. 46, “Relationship of major temperature reconstruction papers and the proxies that they use”. This sort of question of proxy inter-dependence comes up very frequently. Maybe once a day.

  79. Steve McIntyre
    Posted Aug 3, 2006 at 8:30 PM | Permalink

    11. Mann, M. E. & Jones, 2003: uses Mann’s PC1; uses Yamal as part of 3-series average; uses Thompson’s Dunde and Guliya in the Yang composite;

    12. Briffa, K. R. et al. (2001). Only goes back to 1400 so it doesn’t affect MWP. I don’t know what sites are used. Has Divergence Problem i.e. declines after 1960. Briffa deals with this by deleting post-1960 values – you’ll be amused by the snip in IPCC TAR – browse the Jones et al 1998 category for May or June 2005 for a post on this. This network is recycled in Rutherford et al 2005, where post-1960 values are again deleted.

    13. Esper, J., Cook, E. R. & Schweingruber, F. (2002); 14. Cook, E., Esper, J. & D’Arrigo, R. E (2004). These are the same network. Two of 14 sites (and two of 7 or so MWP sites) are foxtails. This is the only study that uses the Polar Urals Update instead of Yamal and has a quite elevated MWP – that’s why they need two foxtail sites and not just one.

    15. Hegerl, G. C. et al. 2006. I don’t know what sites are used. But my guess (based on the reverse engineering described above) is that their network is very similar to Osborn and Brifa 2006 and will include two BCP/foxtail sites: one as Mann’s PC1 and one as the average of two foxtail sites used in Esper; (b) they use Yamal; (c) they use Dunde and Guliya as part of the Yang composite.

    16. Moberg, A., et al 2005. This uses 3 BSP sites, although, in this case, both because of the sites selected and their system, the BCPs are not what cause any HS-ness. Curiously, they inadvertently used 2 different versions of the same site from the White Mts – one from Graybill and one from Lamarche or Schulman from the 1960s (the data from which was incorporated into the Graybill version – how sloppy is that?). They also use Yamal and Dunde and Guliya as part of the Yang composite. However the coldwater diatom percentage is a very active ingredient here.

  80. TCO
    Posted Aug 3, 2006 at 8:47 PM | Permalink

    If you find this sort of sloppiness (reusing series and the like) why don’t you publish little notes in the dendro literature. I mean forget the AGW fight for a second. Just cite issues with basic dendro. (But it will help the AGW fight in the end. And if it doesn’t, it will at least help science…which is what I care about…know you don’t…but still…have a heart.

  81. Jean S
    Posted Oct 24, 2006 at 10:23 AM | Permalink

    Hegerl et al is available from Hegerl’s web page:

    http://www.nicholas.duke.edu/people/faculty/hegerl2.html

    I don’t know if it is accepted yet. The manuscript only says “resubmitted June 12, 2006″. Anyhow, it contains an appendix describing all the used proxies in some detail! Seeing who is the last author, I wonder if discussions over CA have had any effect on that ;)

    Oh, and the key technique used is … drums please … total least squares!

  82. Steve McIntyre
    Posted Oct 24, 2006 at 10:37 AM | Permalink

    #81. Well, I guess that we’re having some impact. IT sort of says what the sites are. Here’s something amusing:

    western U.S.: this time series uses an RCS processed treering composite used in
    Mann et al. (1999), and kindly provided by Malcolm Hughes,
    and two sites
    generated by Lloyd and Graumlich (1997), analyzed by Esper et al. (Boreal and
    Upper Wright), and provided by E. Cook. The Esper analyses were first averaged.
    Although there are a number of broad similarities between the Esper and Hughes
    reconstructions, the correlation is only 0.66. The two composites were averaged.

    An “RCS-processed treering composite used in Mann et al 1999″ ????? Surely this is Mann’s PC1 – Eduardo, what on earth were you thnking??

    It will be fun to check on my predictions.

  83. bender
    Posted Oct 24, 2006 at 10:53 AM | Permalink

    Re #81
    1. When they smooth data series like Fig 4 do they make use of Emanuel’s “pinning” effect? (i.e. Failure to exclude endpoints, thereby exagerrating HS when series end during a rising trend.)
    2. The modern instrumental data (ending in 1998) in Fig 3 do not exceed the reconstructed 95% confidence interval at 1100AD.
    3. 1998 was warmer than any years since, so 1999-2006 are well within the MWP confidence interval at 1100AD. Why do the authors not show the later data, which flatten out post-1998?

  84. Steve McIntyre
    Posted Oct 24, 2006 at 11:03 AM | Permalink

    #83. bender, speaking of bin-and-pin, have you read Mann’s discussion of endpoint smoothing – it was a Comment by Willie Soon and Reply by Mann in GRL a couple of years ago. Should be up on Mann’s website. I didn’t try to sort it out and, since it’s Mann-speak, it always takes time.

  85. bender
    Posted Oct 24, 2006 at 11:06 AM | Permalink

    Check out Fig 2. Focus on 1010AD and 1120AD. Why the divergence between Moberg (M-shaped double-hump, purple) and CH (W-shaped double-dip, red/blue)? Is this an analog for the present-day divergence problem? And where are the CI’s on Moberg?

  86. bender
    Posted Oct 24, 2006 at 11:08 AM | Permalink

    Re #84 You’re kidding me. They know about this problem, and still use it? Now THAT’s moving on.

  87. Jean S
    Posted Oct 24, 2006 at 11:13 AM | Permalink

    The “Dark Ages”-version is based on 5 series! I kind of like the name… ;)

  88. bender
    Posted Oct 24, 2006 at 11:15 AM | Permalink

    Re #83
    Last line in Fig 4 caption suggests smoothing may have been done correctly:

    Esper and CH-blend are shortened due to the application of the filter.

  89. Steve Sadlov
    Posted Oct 24, 2006 at 11:33 AM | Permalink

    RE: #83 – They needed to make the MWP go away. Got to give credit where credit is due. Now, because of that ground work, we are now seeing Hockey Sticks with the handle going all the way back to the 3rd Century AD!

  90. Steve McIntyre
    Posted Oct 24, 2006 at 11:38 AM | Permalink

    #88. bender, I wouldn’t get too excited about end-point smoothing in most of these studies- that seems to be an Emanuel one-off.

    BTW my predictions as to what’s in this study are nearly all bang-on. I’ll post up a detailed comparison.

  91. bender
    Posted Oct 24, 2006 at 11:44 AM | Permalink

    Canadians will be unimpressed with Appendix A. The data from Mackenzie don’t fit the global picture and the data from Alberta and Quebec yield conflicting results for no apparent reason. Good thing “the consensus” is strong.

  92. Jean S
    Posted Oct 24, 2006 at 11:45 AM | Permalink

    I think they have fixed their confidence intervals although I can not find any written confirmation for that…

  93. bender
    Posted Oct 24, 2006 at 11:50 AM | Permalink

    These “confidence intervals” are crap. I want to see uncertainty envelopes, where they trace the error through the entire calibration/reconstruction. That’s where the real error is.
    TCB

  94. Jean S
    Posted Oct 24, 2006 at 11:54 AM | Permalink

    re #91: I think it is also unfair that Swedes have almost a consensus climate (series #8) whereas only about 200 km away we have a nontypical climate… (see here and here).

  95. Steve McIntyre
    Posted Oct 24, 2006 at 12:00 PM | Permalink

    bender, you’ll like their sense of Canadian geography:

    Examination of the NGDC data base indicates that the original Esper et al. reconstruction appears to be from the Boniface site. A record from nearby St. Anne also shows many similarities to Boniface (r=0.66), extends closer in time to the present, but is also slightly shorter (the Boniface/St. Anne correlation is 0.70).

    I’ve determined from analyzing Esper that “Boniface” is Schweingruber’s Bonif historisch site (CAN; 55N,77W; Schweingruber,F.H.) cana169w_crns.crn which is located on the east shore of James Bay. “Nearby” St. Anne is the Gaspé series(CAN; 48N,65W; Cook,E.R) cana036.

  96. bender
    Posted Oct 24, 2006 at 12:13 PM | Permalink

    Their p. 19:

    The detection of the anthropogenic fingerprint is robust for all results except Moberg, where significance levels of that signal are dependent on details of the analysis. The Moberg et al. (2005) reconstruction is also the only one that shows a detectable solar signal, although this is somewhat sensitive to details of the analysis.

    As pointed out in #85, the Moberg series does some other interesting things as well.

    Re #95
    James Bay is only about a thousand miles from Gaspe. Not bad.

    So you mean that big pink box covering E.N.America in Fig 1 is just two sites?

  97. Steve McIntyre
    Posted Oct 24, 2006 at 12:24 PM | Permalink

    w. Siberia: in order to avoid any heavy biases of the mean composite by a number of sites from one region, the west Siberia time series is a composite of three/four time series from this region: two “polar Urals” records east of the Urals — Yamal (Briffa et al. 1995) and Mangazeja (Hantemirov and Shiyatov 2002 – both by way of Esper et al.) and two records from west of the Urals (Hantemirov and Shiyatov 2002). The records from each side of the Urals were first averaged and then combined for the w.Siberia.short composite; the w.Siberia.long composite involved Yamal and the west.Urals composite. The sites from Esper have been RCS processed.

    I guess that they feel that identifying the sites is somewhat of a concession and they don’t need to be encumbered by any considerations of making the identifications easy. For example, Mangazeja is not mentioned in Hantemitov and Shiyatov 2002, but is used in Esper et al 2002. I presume that this is the version carried forward.

    We’ve had lots of discussion of Yamal on this site, but they don’t seem troubled by it. Here they’ve attributed the “Yamal” site to Briffa et al 1995, which discusses the different Polar Urals site; the Yamal substitution took place in Briffa 2000. However, Esper et al 2002 is the ONLY study that used the Polar Urals update – so it’s still a guess what series they used. The start date shown is in the 10th century – so it’s possible that they might have used the Briffa et al 1995 series that started then without the update.

    They mention two sites east of the Urals attributed to Hantemirov and Shiyatov 2002, which doesn’t mention any such sites. I wonder what they are.

    I’ve posted up Hantemirov and Shiyatov here (which is an interesting read in its own right.)

  98. Steve McIntyre
    Posted Oct 24, 2006 at 12:30 PM | Permalink

    #96. yes.

    All the blue triangles in North America look like Mann’s North American tree ring network – all the sites available after 1750. But the MBH99 PC1 only uses 28 sites, and is, of course, almost entirely weighted towards bristlecones, especially Sheep Mountain – (I guess Hughes didn’t send Hegerl the updated Sheep Mountain data.) I would bet dollars to doughnuts that most of the sites shown do not contribute to the PC1.

  99. Willis Eschenbach
    Posted Oct 24, 2006 at 2:09 PM | Permalink

    Re #83, Steve M., you say:

    #83. bender, speaking of bin-and-pin, have you read Mann’s discussion of endpoint smoothing – it was a Comment by Willie Soon and Reply by Mann in GRL a couple of years ago. Should be up on Mann’s website. I didn’t try to sort it out and, since it’s Mann-speak, it always takes time.

    The reply by Mann in GRL compared three methods. Here’s his description of the methods:

    To approximate the “minimum norm’ constraint, one pads the series with the long-term mean beyond the boundaries (up to at least one filter width) prior to smoothing.

    To approximate the “minimum slope’ constraint, one pads the series with the values within one filter width of the boundary reflected about the time boundary. This leads the smooth towards zero slope as it approaches the boundary.

    Finally, to approximate the “minimum roughness’ constraint, one pads the series with the values within one filter width of the boundary reflected about the time boundary, and reflected vertically (i.e., about the “”y” axis) relative to the final value. This tends to impose a point of inflection at the boundary, and leads the smooth towards the boundary with constant slope.” (Mann, 2004)

    He recommends using the “minimum roughness” constraint … apparently without noticing that it pins the endpoints.

    I wrote a reply to GRL pointing this out, and advocating another method than one of those three, but they declined to publish it. I’m resubmitting it.

    w.

  100. bender
    Posted Oct 24, 2006 at 3:02 PM | Permalink

    Good show, Willis.

  101. Steve McIntyre
    Posted Oct 24, 2006 at 4:02 PM | Permalink

    In terms of methodology, Hegerl et al is amusingly similar to MBH98. They have a Mannian PC1 for the North American tree ring network but their regression method is amusingly similar.

    Before they get to their Total Least Squares step, they weight the proxies according to their correlation with NH temperature – I’ve written about Partial Least Squares regression previously and , of course, this is Partial Least Squares. I’ve shown that the Mannian regression methodology reduces to PArtial Least Squares in the early steps – of course they didn’t realize it. So here we have two climate reconstrucions both using Partial Least Squares, both supposedly as novel inventions.

    Mann made a PArtial Least Squares estimator and re-scaled it to match the variance of the temperature PC1. Hegerl et al re-scale their PLS estimator a little differently – by a Total Least Squares regression on NH temperature.

    The Hegerl proxies are, as I predicted, virtually identical to the Osborn and Briffa “independent” reconstruction: PC1, foxtails, Yamal, Tornetrask, Yang composite, etc. It’s pretty funny.

  102. bender
    Posted Oct 24, 2006 at 4:05 PM | Permalink

    Can we please “move on” from this issue of “independence”?! :)

  103. Steve Sadlov
    Posted Oct 24, 2006 at 4:19 PM | Permalink

    It’s like a cult. The inside loop of inside loops.

  104. bender
    Posted Oct 24, 2006 at 4:28 PM | Permalink

    Please, no social network analysis, either, Sadlov. Let’s just keep moving.

  105. Douglas Hoyt
    Posted Oct 25, 2006 at 6:02 AM | Permalink

    they weight the proxies according to their correlation with NH temperature

    This doesn’t make any sense physically since trees respond to local climate and not global climate. Perhaps they should have weighted them by the r^2 of local temperature. I wonder if this would cause the LIA and MWP to reappear?

3 Trackbacks

  1. [...] as published in Nature simply using the proxies in Osborn and Briffa 2006, describing the process as follows: I emulated the CH-long blend using the predictions in my earlier post as follows. All of the 12 [...]

  2. [...] I commented about my lack of success on the web. I described my findings over at Climate Audit, saying: [...]

  3. [...] I commented about my lack of success on the web. I described my findings over at Climate Audit, saying: [...]

Follow

Get every new post delivered to your Inbox.

Join 3,381 other followers

%d bloggers like this: