How science does and does not work (and how skeptics mostly fall in the latter category)

by

Another kick-ass article on The Conversation, providing some insights into the scientific process and the place that pseudo-skepticism takes in (or rather outside of) that process.

In science, to actually contribute at the forefront of a field one has to earn credibility, not demand it. Being taken seriously is a privilege, not a right.

In science, this privilege is earned by not only following conventional norms of honesty and transparency but by supporting one’s opinions with evidence and reasoned argument in the peer-reviewed literature.

This is what makes science self-correcting. If arguments turn out to be wrong, in time they are caught and corrected by other scientists. It is virtually impossible to publish long-refuted nonsense in good peer-reviewed journals.

(…)

an overwhelming scientific consensus does not imply the absence of contrarian voices even within the scientific community.

Over time, those contrarian voices simply fade away because no one takes them seriously, despite their shouts of “censorship” and accusations of bias.

This is not to say that a scientific consensus is never overturned.

There are well-known examples such as the Helicobacter pylori discovery in medicine, and continental drift in geology. But in both cases the arguments were won and lost in the peer-reviewed literature, not by contrarians sitting on the side-lines writing opinion pieces about how they were being oppressed.

A ‘change in paradigm’ occurs when the evidence for the prevailing theory is shown to be weak, and the evidence for a competing theory is getting stronger. That is the opposite of what has been happening in climate science over the past 150 years: The evidence of human influence on climate has been steadily accumulating from the time that it was first postulated as a prediction. Arguments against it have been shown to be either wrong or irrelevant for the big picture.

Even more so, the prevailing paradigm (that us tiny humans can’t possibly compete with the great forces of nature in affecting the earth’ climate) has been gradually overturned by the evidence which pointed out that yes, we can.

Oreskes gives a good overview of how climate science stacks up against the scientific methods (Highly recommended: “How do we know we’re not wrong?” slides and book chapter).

Back to The Conversation. Read and shudder:

One self-proclaimed “rocket scientist” who has published junk science in the opinion pages of The Australian has been quoted on aNew Zealandwebsite as saying:

“To win the political aspect of the climate debate, we have to lower the western climate establishment’s credibility with the lay person. And this paper [an accompanying picture book of thermometers] shows how you do it. It simply assembles the most easily understood points that show they are not to be entirely trusted, with lots of pictures and a minimum of text and details. It omits lots of relevant facts and is excruciatingly economical with words simply because the lay person has a very short attention span for climate arguments. The strategy of the paper is to undermine the credibility of the establishment climate scientists. That’s all. There is nothing special science-wise.”

Undermine credibility.

That’s all.

Nothing science-wise.

Are these the people one should entrust with the welfare of future generations?

The tried and tested strategy of sowing doubt in the minds of the public seems to be supplemented by a strategy of lowering the scientists’ credibility. The latter strategy seems to be at least as successful as the former, if recent events are any guide.

PS: I’m well aware that the bulk of public skepticism is not based on a well orchestrated campaign, but stems more from individual reasons (which I’ve discussed in a previous post). That does however not negate the fact that certain lines of reasoning are repeatedly used in the public debate.

Another necessary element of denial is conspiratorial thinking. Any denier sooner or later, whether an academic or not, must resort to a global conspiracy theory to negate the overwhelming evidence arrayed against them.

(…)

Just imagine the devastating rebuttal of climate change that Bob Carter could submit for peer-review if he wasn’t being oppressed by the Archbishop of Canterbury and Prince Charles.

(…)

Time to close the phony debate on climate science

At a time when the oceans are accumulating heat at the rate of five Hiroshima bombs per second, are conspiracy theorists the people whom a nation should entrust with the future of our children?

The so-called “debate” on climate change has been over for decades in the peer-reviewed literature. It is time to accept the scientific consensus and move on, and to stop giving air-time to the cranks.

It is time for accountability.

Other articles on The Conversation that I enjoyed reading:

Part Two: The greenhouse effect is real: here’s why.

Part Three: Speaking science to climate policy.

Part Seven: When scientists take to the streets it’s time to listen up.

PS: Even though I chose not to use the word “denier” to describe those who don’t accept mainstream science, I did not change it from the original text as quoted. However, I do not intend to host a discussion here on the arguments pro and con of using this or the other label.

About these ads

Tags: , , , ,

451 Responses to “How science does and does not work (and how skeptics mostly fall in the latter category)”

  1. Jeff Id Says:

    Well, if ‘science’ would start applying reasonable standards to paleoclimate, perhaps it would be less easy to criticize.

    BTW, perhaps paleoclimatology should be called paleomathmashology.

    Sorry, not happy with a few of the recent turds being published.

  2. Jeff Id Says:

    Bart,

    Please feel free to snip this and my previous comment. It really comes from a bit of grumpiness with the poor quality of recent paleo papers and continued lack of reasoning from the climate community. Note the plural b/c I’m not happy with those whom I am expected to be either.

  3. Chris Colose Says:

    The problem with this article is assuming that the “deniers” are interested in advancing the science (regardless of whether it is a traditional or roundabout way).

    While some of the so-called skeptics may be well-intentioned but misinformed, most are not well-intentioned and seek only to cause confusion and take issues of very little significance, feed them steroids, and give the impression that they are a supporting scaffold of modern climatology. In truth, there is not a single peer reviewed article by the skeptic community that has convincingly challenged some of the “broad” topics that underlie global warming, such as the radiative properties of CO2, the water vapor feedback, etc. Of the very few articles they generate, most are in odd journals like E&E with a very poor reputation. In the case of articles that actually do contribute to the science, such as Jeff Id’s Antarctic paper, they have very little influence outside a very specialized field and regardless of their conclusions, are very unlikely to affect most people. These tend to become overhyped to make them much more significant than they actually are. Steve McIntyre and the discovery that 1934 was really the hottest U.S. year was another case in point of where an error was pointed out, but then the signal-to-noise ratio hit the basement. There’s no excuse for this.

    Many academics who know better (e.g., Richard Lindzen) say a lot of bulls**t in non-refereed forums where he is not under the scrutiny of peer review. He does it because he can get away with it. The ones with some credentials (Lindzen, Spencer) argue for low climate sensitivity rather than some of the nuttier “greenhouse effect doesn’t exist!” stuff, but everyone in the scientific community knows they are wrong at this point.

    You will notice that there has been a progressive evolution of failed scientific arguments to more and more conspiracies. “CO2 lags temperature, “its the sun”, “Water Vapor rules!!” and all those memes never worked out, so the obvious fall-back mechanism is climategate, the hockey stick, unfair PNAS reviews; the next meme will be this new “Hansengate” thing, which appears to be another attack from a similar source (just like the old “Soros” stuff); there are also attacks on Gavin, and naturally the original attacks on Mann were to attack a powerful icon, also known as the hockey stick (even though very few people actually care about the hockey stick, and virtually nothing surrounding AGW rests on it).

    The attack on the historical paleoclimate community is interesting. I have tried to figure out why the focus has turned on them so much, as opposed to the aerosol community, or the feedback community, or the carbon cycle community, or the planetary atmosphere community. It is particularly interesting because the implications of differing conclusions (for example, whether the MWP was 0.1 C hotter or cooler than today) have no impact on the attribution of modern global warming or the potential threat of future climate change. It is also interesting because despite a decade of relentless attacks and threats against Mann and others, no one has come up with a more credible paleo-reconstruction, and at least a dozen other papers are in agreement with Mann’s (within uncertainties). If people REALLY cared that much, they would go out and try to do better, but they don’t, which shows their is some ulterior motive.

    I think that part of the issue is that the arguments involving Principal Components, statistical methods, etc are not likely to be familiar to many bloggers, so naturally many people will just side with whoever supports their preconceived notions. It is also much easier to argue a point than it is to argue whether CO2 absorbs photons. This is because one topic is well constrained by the laws of physics, while the other inevitably requires choices, subjective judgment, the use of imperfect data, etc. As long as no one is constrained to show that these things matter, they are free to poke jabs all they want. This, of course, is not constructive criticism, it is just criticizing for the sake of it.

  4. Bart Says:

    Jeff, I see nothing worth sniping.

    Chris, good comment. Indeed, if the valid criticisms wouldn’t be packaged in such conspiratorial/accusative/exaggerated (c/a/e) ways, they would be welcomed much more than they currently are. The art that mainstream scientists and their defenders must learn is to take the valid parts of the criticisms and deal with/respond to them, and leave the c/a/e packaging for what it is. That is increasingly difficult because the critics and their supporters will try to keep the c/a/e in (presumably because this packaging is what is most effective at decreasing the scientists’ credibility and sowing doubt). That dynamic needs to be broken. It needs effort from both sides, as difficult and unfair as it sounds.

  5. Deech56 Says:

    *snipping – although it is tempting to find a use for the word “sniping.” :-)

  6. Dan Olner Says:

    I was reading Mark Lynas’ blog and noting who was turning up in the comments. One chap (David Bailey) has posted an old WUWT story, with the argument: look, here’s a pic of a submarine at the North Pole in 1958, so a melty North Pole is nothing new.

    I couldn’t help but try and point him towards improving his google-skills a little, but it underscores the point that ‘rocket scientist’ wants to exploit: a single picture on WUWT does have a lot of persuasive power, and is much easier to integrate into a narrative than spending time checking up on how ice actually changes / the history of nuke subs trying to get through the ice / looking at the actual data. It’s just “no smoke without fire” writ large, and once it gets stuck in people’s minds, it’s incredibly hard to dislodge.

    The problem is, of course, we cannot counter with the same tactics – and indeed, of course, you are attacked if you post (say) pictures of polar bears swimming across open water. Cherry picking nonsense! Unscientific!

    It’s a communication problem we still haven’t solved, and the only way I can think it can be solved is long-term education in critical thinking. Except we don’t have that sort of time.

  7. The Climate Change Debate Thread - Page 768 Says:

    [...] [...]

  8. Bart Says:

    Pun unintended.

    Dan, well said. That’s indeed the problem and there’s no easy solution.

  9. MIkeN Says:

    Chris, the hockey stick was very significant to AGW. That’s why it is used in so many press releases and such. The visual is needed to get political results.

  10. Chris Colose Says:

    MikeN,

    If the goal is public appeal rather than advancing science then you might have a point. I’m not a psychologist or a sociologist, so I don’t have special insight into what catches people’s attention, nor am I particularly interested in sacrificing truth for rhetoric just to get political results. I’m quite sure that the hockey stick is a nicer visual for a press releases than a tutorial on how the quantum physics of photon absorption works, but the latter is the more important knowledge for understanding climate in a generalized sense…the former is a detail, and is most useful to a relatively specialized group of people. It is certainly nice to put modern global warming in a historical context, but we already have a good working knowledge of Earth’s history to say useful things about that, as well as make attributions to specific causes with good confidence, and can do so completely independent of the hockey stick.

    Whatever the “true” shape of the hemispheric or globally averaged temperature is over the last millennia, aside from having no bearing on the attribution of modern global warming, it also does not have valuable insight into the impacts associated with a 2x or a 3x CO2 world. For this, you need to turn more to deep-time paleoclimate.

    But you’ve only highlighted the main point of my post which was that the skeptic industry is interested in everything else– public relations, conspiracies, gotcha games, complaining, etc but have absolutely no focus on the end goal, which is to advance our understanding of climate.

  11. intrepid_wanders Says:

    The sad part of this post is the very tactics that are described in this post go both ways. There are peer reviewed documents going both ways, but feel free to enlighten me to ANY of the scientific academies that voted on any of the elements of the statement in support of the consensus. Everyone on ALL sides have the privilege of credibility and the crux of proving a theory’s validity. The “peer review theory” was a decent one, but it can fall prey to the small population of “peers” in a small field.

    Anyhow, the conspiracy of stupidity stands. All carbon solutions ALWAYS exclude the nuclear power option and mostly exclude hydro-electric. Obviously, Greenpeace and World Wildlife Federation views are plain as day on this, that is what I suspect ticked off Mark Lynas.

    As I have stated and argued on other blogs, I may not even care as to the latest scare tactic that comes from “peer reviewed literature”, but I am interested in reducing the “dirty fuels”. So far, I find renewables are weak in energy density and environmentally unsound.

    Both sides have the same problems with the other. Now what?

  12. Marco Says:

    Bart, the post by Sacha Davilak is spam. It’s a direct copy of a comment from a Bernard Walsh here:

    http://www.climatespectator.com.au/commentary/weird-and-wacky-world-climate-change-denial

  13. Bart Says:

    Chris,
    Very good summary of putting the hockeystick in the context of both the scientific and public debate. Indeed, in the former its role is much more limited than in the latter context.

    Intrepid,
    Your “ALWAYS” is factually incorrect. Many energy scenarios and visions for the decarbonizing the future energy system include nuclear. E.g. World Energy Outlook, European Climate Foundation, but there are plenty more.

    Marco,
    Thanks for the heads up.

  14. Eli Rabett Says:

    Rejectionist.

  15. MIkeN Says:

    Chris, you may not be interested in it, and I mostly agree with your conclusion regarding the significance of the hockey stick. I think Jeff has posted a counterargument at his site.

    However, you are ignoring a key detail that is being missed so much. The hockey stick has moved to a primary position along with other scare stories from WG2. These have done so because people thought it was important to move public opinion. Now when these are found to be false, it brings the whole picture into question. Why give credibility to the atmospheric physics people when the most prominent stories are found to be wrong? If the IPCC procedures themselves lead to weak science, then why give credibiliity to any part of the report? If scientists can’t call out obvious errors like upside-down usage of data, then can they be considered credible?

  16. Chris Colose Says:

    MikeN,

    I think that the WG2 and WG3 have a good amount of evolving to do, and are where the WG1 one was a decade ago. It’s unfortunate, but that’s how it is.

    That said, for a set of reports ~1000 pages each, the track record for an extremely small amount of errors is astounding. No one has raised substantive issues with the physical science, and the only errors found in the WG2 are minimal in nature (such as getting a date wrong in a sentence embedded deep inside the report; the significance was small enough that the claim did not even land a spot in the SPM). I don’t follow WG3 material at all. From the ‘denial’ point of view, the goal however is not to place such errors in the broader context or rationally point out that they need to be fixed, but to use them as “gotcha!” tools to promote their own end. The people like Steve M, regardless of their motives, that continue to play to the “it’s a hoax” crowd will not ever be taken seriously because of this.

    I also disagree with your version of what “was shown to be wrong” (as I said, its general shape has been reproduced by over a dozen studies without any alternative reconstruction that has stood up to the test of time) or why even if the hockey stick is wrong, why that should extend to the atmospheric physics community. The people who study radiation, fluid dynamics, synoptic scale frontal systems, aerosols, air pollution, atmospheric chemistry, solar variations, El Ninos, monsoons, physical oceanography, Quaternary paleoclimate, deep-time paleoclimate, etc are completely different groups with generally different methods of research. Of course, saying something that appears to be a “public icon” is wrong plays well to the public ear, which is precisely my point in what these people are actually interested in.

    The response to the recent Kemp et al. paper is another case in point of how eager the “rejection Team” (to combine the phrases of Steve M himself and Eli above) is to pounce on anything they don’t like. There was absolutely NO RESEARCH done into Kemp et al and the inside references, no attempt to email the authors for clarifications, no attempts at reproducing the results and playing with different parameters. There were only accusations starting with *day 1* of its public release, including “issues” with the paper that spread like wildfire on the blogs that are completely untrue (see the second update of the RC post on the topic). One blogger (I believe Steve Mosher) said he read the supplementary material for 45 seconds and found “issues.” I’m sorry, but that’s just not how it works. The chief “auditor” himself could not even get the affiliation of the lead author right. I’m sure he wasn’t too busy to google search his name (or read the author affiliations in the very paper itself), s the issue must have been that he wanted to jump on a topic with “fresh meat” right away. Even without raising any real issues, they are already succeeding in generating doubt, because that is what they are experts at.

    There is a large amount of desperation to generate doubt now, from all the “gates” to the PNAS reviews to the WG3 energy public release…and the ratio of signal to noise in doing so has never hit such a low.

  17. Paul Kelly Says:

    If the goal is to win the arguments about climate science, this post is very relevant. If the goal is to somehow prevent or slow climate change, this post is an instrument of delay.

    There is nothing pro-active in argument over climate or any other reason for energy transformation. The pro-active approach focuses entirely on the transformation itself. It focuses on immediate action, now. The pro-active approach does not require political majority or government mandate to succeed.

  18. Steve Bloom Says:

    MikeN, don’t forget that what the hockey stick actually shows is relative insensitivity to the natural forcings applied in the late Holocene. More bumpiness would be bad news. Alternatively, if one really, really disliked (to the point of abandoning any pretense to the scientific method) the implications of ~3C+ CO2 Charney sensitivity, one could then turn to attacking the modern T record. But while you were busy with that, a vast body of work was developed (thanks to the ODP, the gift that keeps on giving) showing that CO2 really is the “big control knob” and that even current CO2 levels are likely very big trouble. Got anything on that?

    Chris, thanks for all your efforts here and elsewhere. Your patience is admirable.

  19. steven mosher Says:

    “This is what makes science self-correcting. If arguments turn out to be wrong, in time they are caught and corrected by other scientists. It is virtually impossible to publish long-refuted nonsense in good peer-reviewed journals.”

    Not so. Even AFTER it was debunked Piltdown Man was referred to in an article in Nature.

    generally, science is self correcting. However, there are certain cases, usually those tinged by politics or religion, where good science is kept out and bad science (even hoaxes) are kept in.

    Generally true, but the exceptions are fascinating sociology

  20. Bart Says:

    Mosher,

    Climate science, like any other science, is indeed self correcting. Your insinuations as to a pervasive bias in the whole field are entirely off base and would basically mean there’s a massive conspiracy going on. I thought you were above that kind of stuff.

  21. MIkeN Says:

    Steve Mosher, I assume you are referring to Piltdown Man, and not a mocking reference to Piltdown Mann reusing Tiljander in Science?

  22. MIkeN Says:

    >why even if the hockey stick is wrong, why that should extend to the atmospheric physics community. The people who study …are completely different groups with generally different methods of research.

    Indeed they are. All the more reason to distance yourself from the errors of others. To this point I have not seen members of these groups call out Mann for using data upside-down. Martin Vermeer acknowledged it, but then says Mann was correct in his PNAS reply to McIntyre. Instead Mann is more likely to have his 2008 paper with the upside-down data referred to by the members of those other groups as reinforcing the hockey stick.

  23. Dan Olner Says:

    “To this point I have not seen members of these groups call out Mann for using data upside-down.”

    That’s because, AFAIK, this particular meme has been dealt with, repeatedly, some time ago – what, some years ago now? It’s interesting that it won’t die. I’ve stopped providing links because if I’m having to say this sort of thing, I know it’ll take me about 90 seconds to find the information. C’mon! Google, MikeN, and come back and tell us why that “Mann got it upside down” stuff is wrong. You don’t even need to believe it at this point – just treat it as an exercise in collaboration. Save us having to post links for you.

  24. MIkeN Says:

    Well the most prominent link I can find is to Grumbine science, and that doesn’t settle anything. Indeed in the comment links on that one, I find that I have already posted there as to why he is likely wrong.

    Indeed, I have posted at WUWT that this line is wrong, but on a more technical detail. The data was used upside-down.

  25. MIkeN Says:

    >The Korttajarvi record was oriented in the reconstruction in the way that McIntyre said.

    From the ClimateGate e-mails.

  26. Chris Colose Says:

    Well, these “other groups” are doing their own thing, and might not even be familiar with what exactly is “upside down” or even care. From your parroting, it sounds like you don’t really understand either. Certainly, if the error and its implications were so obvious, there would be corrections and other things to hear about by now.

    Maybe they used the data wrong, maybe they didn’t. I’m not an expert here, and I’ve seen people who actually study this disagree with your interpretation. Mann didn’t buy into it, and as far as I can tell no one has yet taken it seriously in the literature, so I don’t judge this to be of significance. I really don’t have time to do my own research into every single claim made about every single academic field, nor does anyone else, and historical multiproxy temperature reconstructions is one of those I have not familiarized myself with in any detail. I realize every single blogger has made this field their area of expertise, but nothing has been convincingly demonstrated here. It has already been shown that the main conclusions still hold even if the Tiljander proxies in the EIV surface temperature reconstructions are removed, so once again, I see little point in debating this. You’re just trying to add to the noise level.

  27. Eli Rabett Says:

    Let us understand what Mike N is doing which will give us a good idea of where he is coming from, and maybe, some people, not Eli might say, who is providing the scripts and maybe who is paying the bills

    The attacks on Michael Mann (and Al Gore before him, and Joe Romm) use any brickbat that can be manufactured to try and get folk to disown them. This removes strong advocates from the discussion.

    Too often this tactic has been successful. Eli really doesn’t give a crap if Al Gore is fat, Michael Mann bald, and Joe Romm shrill. What they are is mostly right, not perfect, but mostly right. See what CC said about the IPCC above

  28. MIkeN Says:

    Chris, I think it is you who are playing into skeptic hands with your response. As it stands, it is very easy to point out that climate scientists use data upside-down, it can’t be trusted. Even further I can say, it’s not just that they messed up and used something upside-down, which is merely embarrassing, but when pointed out, they didn’t issue a correction, denied the obvious, and used it upside down again.

    I told Kaufmann at the time that his misuse of TIljander made the problem very obvious as it was in a spreadsheet and very easy to compare to the original paper. His correction(which you say don’t exist) made things more difficult to show the errors of climate science.

    >It has already been shown that the main conclusions still hold even if the Tiljander proxies in the EIV surface temperature reconstructions are removed,

    Actually Gavin posted that without TIljander, the no-dendro case doesn’t pass validation prior to 1500. Plus the medieval warm period in this invalid reconstruction becomes substantially higher.

  29. MIkeN Says:

    By the way, I agree with you Chris from a scientific standpoint. But you are missing the political impact of letting bad science stand. The circle-the-wagons approach that Eli advocates doesn’t work if you are defending bad science.

  30. Marco Says:

    Yes, MikeN, without Tiljander the *no-dendro* case doesn’t pass validation prior to 1500. WITH dendro’s…no problem.

    And perhaps in the no-dendro/no Tiljander case the MWP gets warmer. WITH dendro’s…not at all.

    Why do you move the goalposts?

  31. luminous beauty Says:

    As it stands, it is very easy to point out that climate scientists use data upside-down, it can’t be trusted. Even further I can say, it’s not just that they messed up and used something upside-down, which is merely embarrassing, but when pointed out, they didn’t issue a correction, denied the obvious, and used it upside down again.

    It is easy to point out the ignorance of this comment. ‘Inverse correlation’ is not ‘no correlation’. Just as a photographic negative holds the same information as a print, the fact that the correlation methodology is not sensitive to sign, means no correction is necessary.

  32. Deech56 Says:

    Since when were scientists required to toss out the dendro data?

  33. Eli Rabett Says:

    Mike N confuses circular firing squads for circling the wagons. Why is Eli not surprised?

    Mike N recommends that those of us interested in reality and meeting the serious threat of climate change should attack each other to make his job easier.

  34. Deech56 Says:

    …as opposed to ensuring scientific coherence among the critics of mainstream science.

  35. PeteB Says:

    I thought the best discussion on Tiljander (particularly Martin Vermeer’s comments) here

    http://arthur.shumwaysmith.com/life/content/michael_manns_errors

    I also came across this rather wonderful post from arthur smith

    http://www.realclimate.org/?comments_popup=4431#comment-183104

    …..Let’s look at the issue of belief in three points straight from the title of Montford’s book
    * is the “Hockey stick” an illusion?
    * is “climategate” significant?
    * is [any part of] science corrupt?

    On the method of “tenacity”, Peirce explained his meaning:

    why should we not [...take] as answer to a question any we may fancy, and constantly reiterating it to ourselves, dwelling on all which may conduce to that belief, and learning to turn with contempt and hatred from anything that might disturb it?

    Once you have formed any opinion on a subject, it’s a perfectly natural human tendency to cling to that opinion, and “both sides” do it here. From the start most of us with some scientific background and trust in the methods of peer review would have tended to the “No” answer on each of the three questions, though many of us may have had doubts on each point. Clearly on the McIntyre/Montford side, the tendency was to answer “Yes”.

    The question regarding Peirce’s classification is to what degree those beliefs were *fixed* by this method of simply clinging to them, despite contrary evidence. Tamino’s review here and past discussions (Mann 2008 in particular, but even all the other reconstructions like Loehle’s) clearly show a certain “robustness” in the “hockey stick” form and the conclusion regarding recent warmth being highly anomalous. What is the evidence on the other side? There were clearly some mistakes made in the analyses (in Mann’s papers, and far worse in Loehle’s) but there doesn’t seem to be any modern reconstruction published in the peer reviewed literature that shows anything other than essentially the picture in MBH98.

    So the weight of the evidence on question 1 is very heavily in favor of those who fixed belief in the “No” side of the question, and heavily against those on the “Yes”. Have any of those on the “Yes” side changed their minds in regard to this evidence? If not, then they are clearly stuck in Peirce’s “tenacity” mindset.

    Similarly on climategate and corruption, we’ve now had 5 reviews that found at worst some poor communication and data sharing practices. The weight of the evidence is very strongly on the “No” side of these questions. Have any of the “Yes” folks changed their minds as a result? There are some examples: The Guardian’s George Monbiot, for instance, at first thought climategate was significant and called for Phil Jones to resign, but has now stepped back from that (his concern seemed to be chiefly from a deep faith in the importance of FOI-type laws)……

  36. Hank Roberts Says:

    Scientists challenge industry to cooperate with research rather than criticize results:

    http://www.nicholas.duke.edu/hydrofracking/responses-about-gas-shale

    —excerpt follows—-

    Proposed Joint Experiment with Industry

    More sampling is needed in the Marcellus and in other regions of the United States. With the goals of having more pre- and post-drilling data and obtaining as close to a random sample as possible, we propose to work with industry, the DEP, and homeowners to perform such an experiment. Here is one possible design.

    Industry has hundreds, even thousands, of archived pre-drilling estimates of water quality from individual homeowners in Pennsylvania alone. Working together, we would randomly select 100-200 homeowner wells from the pre-drilling database in Pennsylvania. After obtaining homeowner permission, we would resample those homes that have had gas wells drilled in their area. The goal would be to see to what extent water quality changed after drilling and hydraulic fracturing. An independent entity, Isotech Laboratories Inc., could conduct the methane measurements, without knowledge of any pre-drilling data or identifiers for specific wells. If desirable, independent laboratories could also perform the other measurements, including tracers of hydraulic fracturing fluids.

    At the end of the experiment, we would have at least 100 pre- and post-drilling comparisons for individual homes, enough to obtain a clearer picture of what is occurring in the Marcellus in Pennsylvania. The experiment could be completed in a year, free of charge to homeowners. If needed, Duke University would even seek to raise the money to perform it. We look forward to collaborating with industry, the DEP, and homeowners in the endeavor.

  37. MIkeN Says:

    PeteB, in either that discussion or the companion one, Arthur Smith gave a definition of fraud. I posted that I didn’t think Mann was guilty of committing fraud, then I found that Mann’s behavior did meet Smith’s definition. I still don’t think the former, but I found that interesting. Smith eventually reached no conclusion, and just ended with he is looking into it.

    Luminous beauty, you are missing the part where in this proxy, the correlation flips at a later point in time, and the proxy is now inversely correlated. That’s why it shows ‘cold’ in the modern period. So if you correlate this, then you are now showing cold in the past when it was warm.

  38. PeteB Says:

    MikeN, What I liked about that thread was that all sides seemed to agree the technicalities of what has happened. I certainly don’t think there was any case of fraud or of any wrongdoing at all.

    The point where we disagree is neatly summarised by Martin

    “AMac, I disagree strongly for reasons that should be clear to any reasonable reader by now. Mann’s reply was fully responsive to the statement as made by McIntyre in the Comment. He could not be expected to engage in mind reading, or to follow, or even be aware of, discussions on fringe blogs. I am disappointed that you continue to fail to see this, AMac, and have nothing further to add”.

  39. luminous beauty Says:

    MikeN

    Luminous beauty, you are missing the part where in this proxy, the correlation flips at a later point in time, and the proxy is now inversely correlated.

    This is incorrect. The Lake Korttajarvi X-ray density proxy is negatively correlated (the y-axis is upside down) and the ‘light sum’ and ‘dark sum’ proxies are positively correlated. None of them change the sign of correlation over time. What is problematic is that upstream land uses create step changes, altering the calibration metrics post about 1720, making their use ‘difficult’, in Tiljander’s words, for paleo-climate reconstruction. Difficult, but not impossible, though as Mann et al. point out, problematic.

    McIntyre likes to play ‘hide the pea’. He creates some rhetorical ambiguity about what the problem really is, which so many of his acolytes (such as yourself) rush to misconstrue.

  40. MIkeN Says:

    You’re being a bit semantic there. In the modern portion, the value is going in the direction of ‘cold’, while temperatures of course went in the direction of warm. In older times, the proxy did not behave this way. Depending on which way you orient your chart, the proxy was either negatively correlated with temperature in older times and positively correlated in modern times(what is in Tiljander’s paper), or positively correlated in older times and negatively correlated now.
    Either way, it could not be used, and in the CPS screened version it was used upside down to pass screening. Roger Pielke argued the Kaufmann correction showed that McIntyre was right and Mann was wrong, yet still people defend this.

    http://rogerpielkejr.blogspot.com/2009/10/point-resolved-in-hockey-stick-wars.html

    On minor details, I think Pielke gets it wrong and the critics get it right.

  41. mikep Says:

    It’s the complete unwillingness of apparently well-respected scientists to acknowledge obvious errors like de-centred principal components and the trivially obvious misuse of the Tiljander lake sediments that has made me deeply sceptical. Combine this with other more sophisticated errors – like thinking that autocorrelation in the dependent variables in a regression is of necessity a problem, instead of examining autocorrelation in the residuals – and I conclude that these are people I cannot take seriously as authorities.

  42. Dan Olner Says:

    MikeN/MikeP… well, starting by quoting mikeP: “It’s the complete unwillingness of apparently well-respected scientists to acknowledge obvious errors like de-centred principal components and the trivially obvious misuse of the Tiljander lake sediments that has made me deeply sceptical // I conclude that these are people I cannot take seriously as authorities.”

    Could you be more specific: *which* people are you saying cannot be taken seriously? Could we look at the other lines of evidence regarding climate change? Are you saying *none* of them can be taken seriously? What’s so important about this tiny example?

    OK, so you’re deeply sceptical about this particular analysis. I’d like to know what you think about the rest of the science – the massive, multiple lines of evidence pointing in the same direction.

    What I *think* I pick up from people arguing about this: the very fact that others in the climate community don’t call out the errors shows its all about circling the wagons, and no-one can be trusted. For myself, I’m satisfied I’ve seen those errors acknowledged, and that the people involved have moved on and done better analyses that supersede the ones you’re talking about, so it just comes across to me as keeping a focus on some irrelevant minutiae.

    But there must be some way to get past this level of argument. In the end, seems to me all we can do is talk about all lines the evidence. The vast majority of these are well within the reach of basic science knowledge and simple logic / common sense – which is why I’m also sceptical about anyone who returns the focus to this upside-down stuff. It’s a much bigger investment to find the truth for those things built on complex technique, and it’s also then much harder to communicate in a clear way to others, unless they have also done the digging. Whereas the overall case for the soundness of the physical science is just much clearer and stronger. Why can’t we talk about those? What would it take to put this upside down stuff to bed? If it’s an acknowledgement that the entire scientific community is covering for some small error, then it’s going to be difficult to carry on having a conversation.

  43. J Bowers Says:

    MIkeN Says:

    “The Korttajarvi record was oriented in the reconstruction in the way that McIntyre said.”

    From the ClimateGate e-mails.

    Let’s put that into context and see what more of the email says. (my bold for emphasis)

    “The Korttajarvi record was oriented in the reconstruction in the way that McIntyre said. I took a look at the original reference – the temperature proxy we looked at is x-ray density, which the author interprets to be inversely related to temperature. We had higher values as warmer in the reconstruction, so it looks to me like we got it wrong, unless we decided to reinterpret the record which I don’t remember. Darrell, does this sound right to you? This dataset is truncated at 1800, so it doesn’t enter the calibration, nor does it affect the recent warming trend. The attached plot (same as before) shows the effect of re-orienting the record on the reconstruction. It doesn’t change any of our major or minor interpretations of course.
    Nick”

    So, what did they do? The clue is in the email:

    “Hi Darrell
    Sorry to hear that you are getting trouble for doing such a nice paper….I by the way agree completely with Peck that we should not be rushed and that a correction probably should go into Science.

    And it was indeed published in Science.

    Climate scientists are self-correcting.

  44. J Bowers Says:

    steven mosher Says:
    “Not so. Even AFTER it was debunked Piltdown Man was referred to in an article in Nature.”

    Once? What year? I’ve looked but can’t find it. Any help would be appreciated.

  45. Bart Says:

    Dan, very well said. That’s indeed the dynamic as I see it as well.

  46. AMac Says:

    First, there are three (not four) Tiljander data series (they aren’t clear proxies for temperature).

    Second, referrals to justifications of the use of the Tiljander data series in Mann08, Mann09, Kemp11, and other papers are often talked about in comments, as here. Advocates should follow up by providing citations for arguments in the literature, and by linking blog arguments. (I’ll then add them to this page for publications and the corresponding page for blog posts, as time permits.)

    At this point, I am unaware of the existence of justifications for Mann08’s and following papers’ use of the four [sic] Tiljander data series in paleotemperature reconstructions. To be specific, I am referring to the existence of explanations that meet all of these criteria:

    * They address the actual concerns of critics. For example, to me, the key issue is the answer to this question: “Are the Tiljander data series calibratable to the post-1850 instrumental temperature record?”

    * They employ logic, deduction, inference, etc. in the manner that is common to the scientific literature. For instance, in contentious discussions of whether arsenate could substitute for phosphate in the DNA backbone (e.g. here), nobody proposed something like, “I don’t know, and it doesn’t matter.” That’s a statement of taste, not a technical argument.

    * They are informed as to the physical meaning of lakebed varves, and how these could be related to temperature by physical processes. One Philip Mulholland discussed Tiljander this way in, of all places, a recent comment thread at Watts’ WUWT.

    For an instance involving the use of a Tiljander-dependent reconstruction in Kemp11, you can compare the current RealClimate.org post (Update 2) with what I show here and in the immediately preceding post. Which spaghetti graph traces speak to the point Prof. Mann raises at RC: the Blue-Red pair he shows in the Update, or the Blue-Green pair I discuss? (The two Blues are different from each other, by the way.)

    I won’t have the time to comment further on this interesting thread. For what it’s worth, essentially everything MikeN has said here strikes me as correct.

  47. Dan Olner Says:

    Amac: “For example, to me, the key issue is the answer to this question: “Are the Tiljander data series calibratable to the post-1850 instrumental temperature record?”

    I’m confused why this should be your most crucial question. At a certain point, it’s well-known that dendro proxies are *not* going to be useful – they diverge from the temperature record around the 1950s (and wikipedia has lots of leads in the literature on that). Your point is what?

    Shame you’re not coming back, I would have been much more interested to find out what you think the overall meaning of all this Tijlander stuff is for climate science as a whole.

  48. PeteB Says:

    to be fair I think MikeN is mostly correct on the technicalities of what actually happened (some of this is copied from the MikeN, Martin Vermeer + AMac posts elsewhere)

    1) the original Tiljander03 (which was all pretty tentative in the first place and Mann08 reflects fully this caution from the original authors) assigned these temperature correlations to the pre-1720 period:

    Darksum – higher signals correlate to warmer temperatures;
    Lightsum – higher signals correlate to cooler temperatures;
    X-Ray Density – higher signals correlate to cooler temperatures.

    2) From then on Agricultural activity, roadbuilding activity, and lake eutrophication led to these non-climate-related trends:
    Darksum – local activities led to generally increasing signals over that time;
    Lightsum – local activities led to generally increasing signals over that time;
    X-Ray Density – local activities led to generally increasing signals over that time.

    3) The result from the method used in Mann08 paper is

    Darksum – higher signals correlate to warmer temperatures;
    Lightsum – higher signals correlate to warmer temperatures;
    X-Ray Density – higher signals correlate to cooler temperatures.

    So 2 out of the 3 correlations are different to the original authors interpretation. I’m pretty sure that Mann was unaware of this or he would have raised it in the paper.

    There was a published comment from McIntyre in PNAS (which was very poorly worded IMO) There was a reply from Mann, which IMO fully addressed the points from the comment.

    I don’t think Mann should be expected to read discussions on Climate Audit and issue a correction (and considering the amount of abuse he and his colleagues get, I would be surprised if he reads it)

    If anybody thinks it is worth it somebody should publish this as a comment, clearly stating what the issue is and clearly state the impact on the Mann conclusions (which I think is pretty minimal)

  49. PeteB Says:

    correction

    3) The result from the method used in Mann08 paper is

    Darksum – higher signals correlate to warmer temperatures;
    Lightsum – higher signals correlate to warmer temperatures;
    X-Ray Density – higher signals correlate to
    warmer temperatures.

  50. PeteB Says:

    I sort of would be interested on the effect on the Mann08 paper if the method was tweaked so that the ‘Tiljander’ proxies were all correlated the same way as the original Tijander03 paper – (I know what effect removing Tiljander completely has) – Does anybody know if anybody has attempted this ?

  51. AMac Says:

    Here is the correct link to articles touching on Tiljander data series questions, from my comment of 13:17 earlier today.

    Dan Olner @ 13:48 — The question of calibratibility of the Tiljander series is central to understanding Mann08 and what has come afterwards.

    Here’s my view of the middle part of this history.

    * Mann and others presented their “hockey stick” paleotemperature reconstructions based on tree-rings.

    * McIntyre and others criticized those recons, with one major cirticism being that their shape depends heavily on the outsized influence accorded to a few non-representative proxies. E.g. Sheep Mtn. bristlecone pines, Yamal larches.

    * Mann (et al) responded by saying that the series aren’t non-representative. But that even if they were, they weren’t overrepresented. McIntyre (et al) riposted that their emulations showed that this was not so. Back and forth.

    * To answer this, Mann undertook an analysis of non-tree-ring proxies. This was Mann08. He showed that similar hockey stick recons could be constructed from (1) tree-rings alone, (2) tree-rings plus non-dendro proxies, and (3) non-dendro proxies alone. Because he recognized that the Tiljander series might be “problematic”, he did recons that excluded Tiljander. Same result claimed.

    * A McIntyre reader discovered that Mann08 used two of the Tiljander data series “upside down”. More than that, none of the Tiljander series could be calibrated to the instrumental record — that’s how darksum and XRD got flipped upside-down (with respect to the sole authority’s interpretation).

    * Mann08’s Methods rely absolutely on direct calibration to the CRUTEM3v record 1850-1995. Therefore: Series that cannot be meaningfully calibrated cannot be meaningfully used by that approach.

    * Mann denied using any proxies upside-down, calling the charge bizarre.

    * The Mann08 SI figure presenting the various alternative no-Tilj recons had some problems. Mann re-visited the issue multiple times, most recently in Mann09’s SI (see the link in my comment of 13:17). It turns out that no-tree-ring recons require Tiljander to pass Mann’s validation test prior to 1500. And the shape of the no-Tilj/no-tree-ring recon diverges significantly from that of the various yes-Tilj recons prior to 1200 AD or so.

    If the Tiljander proxies indeed cannot be calibrated to the instrumental temp record, it follows that all the yes-Tilj recons are unreliable — they satisfy the “Garbage In” part of the famous aphorism.

    Then we are left with only the no-Tilj recons to discuss.

    I should point out my own view — If the Tiljander data series cannot be calibrated, they should be discarded from analyses like Mann08, Mann09, M&W, Smith10, and Kemp11 before going any further.

    Full stop.

    Then we can decide whether the no-Tilj/no-dendro recons tell us very much.

    Life is complicated enough as it is, without adding postmodern “I don’t know, and it doesn’t matter” conundrums. And all the rest of the convoluted argumentation that this matter has spawned.

  52. J Bowers Says:

    Amac, is the cryosphere melting?

  53. AMac Says:

    What a question.

    Provide a link, explain how, in your view, your topic relates to anything I’ve written here, and I will be glad to offer my opinion.

  54. Marco Says:

    AMac, could you tells us what the confidence setting is for the claim “It turns out that no-tree-ring recons require Tiljander to pass Mann’s validation test prior to 1500″ ?

    I happen to know what it is for going back to 1300 and 1100 AD.

  55. AMac Says:

    Over at The Blackboard, Lucia discourages commenters from argument by rhetorical question. I agree with her that discussions are generally better when people state their positions directly.

    In my comment of 13:17 supra, I link to a post in which I quote Gavin Schmidt: “…it’s worth pointing out that validation for the no-dendro/no-Tilj is quite sensitive to the required significance, for EIV NH Land+Ocean it goes back to 1500 for 95%, but 1300 for 94% and 1100 AD for 90%.”

    There is additional detail in one of the tables near the end of Mann09’s SI. PDF here.

  56. Dan Olner Says:

    Amac, thanks for stating your position. To pick up on J Bower’s question (is the cryosphere melting): it doesn’t relate to your topic, and that’s exactly the point. Bart’s post is *not* about Tiljander: it’s about the difference between the full sweep of climate science versus the methods used by its detractors – one of which I and others are arguing is: keep the focus on some minor part of the overall picture, preferably one with an easily communicated meme like “the scientists got it upside down!” (Which you’ve repeated, incorrectly I think: “Mann denied using any proxies upside-down, calling the charge bizarre.” No, he said claiming that it was upside-down mattered for the correlation was bizarre. Notice how your version still makes Mann out to be a liar.)

    I’m not saying we shouldn’t talk about hockey sticks, but if someone asks you a non-Mann-related question directly – as J Bowers did, as I did, and as I’ve asked the Mikes – it’s not unreasonable to expect a straightforward answer, given what the topic is here. I want to talk about climate science as a whole please, and how you and others think this Tiljander stuff affects it, if at all. (MikeN has given a good summary of how he thinks it affects it, I think, but I’d like to hear more on that.) J Bowers’ approach is great: a specific, simple, straightforward question. Your answering it in no way implies anything specific about Tiljander, and we can come back to talking about it – but equally, answering it merely requires a yes or no from you. So, is the cryosphere melting? Oh and, is the earth warming up due in large part to the co2 people are putting in the atmosphere? Both simple yes or no questions…?

  57. Marco Says:

    Amac, the reason I asked the question in the way I did is exactly because the devil is in the details. By not pointing out what “Mann’s validation test” was, i.e., the 95% confidence interval, the suggestion is made that the reconstruction has NO confidence.

    At the same time the “significant” differences you claim require removal of BOTH tree rings and Tiljander (+others). If you only remove Tiljander, the difference from the yes-Tiljander is minimal (before 900 the differences are largest, but excluding a small section around 800 well within 2-sigma).

    The devil is in the details, and if you discuss details, you better make sure you make every detail very clear.

  58. AMac Says:

    Dan @ 10:08 —

    > Bart’s post is *not* about Tiljander…

    Thanks for linking to MikeN’s earlier comment. He explains (as I have done elsewhere) why Tiljander *is* important to Bart’s argument. It isn’t because Mann08 contains mistakes, even ones that are potentially fatal to its main point. And it isn’t because scientifically-literate laypeople defend the flawed work on teh intrawebs. Scientists make mistakes all the time, and advocates craft arguments out of whatever material is at hand. We should expect that.

    The problem lies in the failure of the paleoclimate community and the broader climate-science community to apply the corrective measures that are standard for non-postmodern and non-postnormal physical science (per Judy Curry’s definitions).

    J. Bowers tried to rebut this point earlier (June 30, 2011 at 10:39) by quoting a Climategate email. In my opinion, in context, that email strengthens my immediately-prior assertion. I can go into more detail if you wish, time permitting.

    Part of the full sweep of a scientific field is precisely in undertaking this function. That is why I linked to a comment thread on the arsenical DNA debate that erupted last December. That controversy is an instance of science done right.

    > you’ve repeated, incorrectly I think: “Mann denied using any proxies upside-down, calling the charge bizarre.” No, he said claiming that it was upside-down mattered for the correlation was bizarre. Notice how your version still makes Mann out to be a liar.

    As to whether “Mann denied using any proxies upside-down, calling the charge bizarre,”here is the full copy-paste from Mann et al.’s Reply in PNAS

    The claim that “upside down” data were used is bizarre. Multivariate regression methods are insensitive to the sign of predictors. Screening, when used, employed one-sided tests only when a definite sign could be a priori reasoned on physical grounds.

    Notice how your version makes me out to be calling Mann a liar. I am not a mind reader and have made no claims to that effect. I have repeatedly written that the entire affair is more parsimoniously explained by a series of honest mistakes. I could dig up quotes to this effect, but you first. Look on my blog or use my easily-searchable handle at The Blackboard, Climate Audit, C-a-s, or RealClimate and back up your case. Or retract the assertion.

    If you want answers from me on unrelated issues, please do the following. (1) Provide a link to the data you wish me to comment on. (2) Engage with me in a more civil fashion.

  59. AMac Says:

    Marco,

    > the suggestion is made [by AMac] …

    I don’t think you read the blog post that I linked before submitting your rhetorical question as a comment. The fact that I clicked there to cut-and-paste a quote of myself quoting Gavin Schmidt rebuts your idea.

    > The devil is in the details, and if you discuss details, you better make sure you make every detail very clear.

    Thank you for this advice. You seem to be suggesting that I must hold to a standard whereby every blog comment I write must discuss every relevant detail, very clearly.

    Compared to the use of links, this might satisfy you, but it would be unlikely to please other readers. I started that Tiljander-themed blog for three reasons. (1) To organize the material for myself. (2) To provide a resource for other interested parties. And (3) to address the criticism I’ve repeatedly encountered when leaving comprehensive comments at other blogs… too much detail!

  60. Dan Olner Says:

    Amac, apologies if I came across as uncivil: I agree civility is very important, and more-so in online communication. I don’t read Mann’s comment as a denial of using anything upside-down, only that it doesn’t matter / the idea of ‘upside down’ in that context doesn’t make any statistical sense. One might then want to go on and argue that Mann is wrong or right – but as far as I read it, it’s not a “denial that he used any proxies upside down”, it’s a denial that the idea of “upside-down” has any relevance for his analysis.

    As for providing you with data: I confess to being confused. Obviously I can do that – but have you not already come to a solid conclusion on whether the earth is warming due in large part to human-introduced co2, or whether the cryosphere is melting?

  61. AMac Says:

    Dan Olner —

    This is tiresome.

    > if I came across as uncivil

    Your statement “Notice how your version still makes Mann out to be a liar” remains, unsupported and unwithdrawn.

    > I don’t read Mann’s comment as a denial of using anything upside-down

    versus Prof. Mann’s formal response to the assertion that he used a Tiljander proxy upside-down —

    The claim that “upside down” data were used [by us in Mann08] is bizarre.

    > As for providing you with data: I confess to being confused.

    Immediately above are two instances where conversation about seemingly-straightforward issues have gone quickly off the rails. Now you ask me to provide a yes-no answer to a complex question, without making the effort to provide alink to the data you think we should be talking about.

    Under these conditions, I decline.

  62. Dan Olner Says:

    I agree, the fact that we can’t seem to have a straightforward online interaction is unpromising. You say you presume Mann made “an honest mistake” – OK then, I withdraw my line about “your version makes Mann out to be a liar.” Happy to, if only because it’s just the sort of minute detail I’d like to avoid. Your version makes Mann out to have made an honest mistake. Sorry.

    It isn’t any effort to provide links to data: I just fail to see why you need links to data to be able to make simple statements about the main findings of climate science. Neither of the two questions asked are actually complex. The *methods* used to answer them may be complex, but the answers are not. I mean: can you tell what season it is where you are? The reasons behind that are massively complex, but presumably you’d be perfectly capable of giving me a straightforward answer. As far as I’m concerned, answers to those two climate questions are equally straightforward. Aren’t they?

    I’m sorry you feel this is tiresome. I’m genuinely keen to try and find out how to talk about this stuff with people I disagree with, without it going off the rails, as you say.

  63. AMac Says:

    > Neither of the two questions asked are actually complex.

    I have never used the word “cryosphere” in conversation, so you and I must travel in different circles.

    This page presumably includes the data that J. Bowers and you had in mind. Why it is somehow better for you guys to not link to its “Global sea ice area” graph or something else, I still don’t understand.

    Short “cryosphere” answer — Yes.

    Long “cryosphere” answer — That graph clearly shows a significant decline in the area covered by sea ice. The record starts in 1979, and my Mark I Eyeball puts the start of the drop at 2002. There are other ways of slicing and dicing the data, but the overall conclusion of warming since 1979 is clear.

    It is also consistent with the clear evidence that the Earth has warmed about 1 C in the past 100-plus years.

    Things get less obvious when one considers the past 2,000 years instead of the past 200 years ago (I’m not even sure that the Little Ice Age was a worldwide phenomenon — I think so). As far as Arctic ice, the historic record includes evidence of an icier Iceland (colder than at present) and a greener Greenland (warmer than at present). A trustworthy multi-century temperature reconstruction would sure be helpful in placing current warming in context — too bad we don’t have one, and may never get one.

    Short “whether the earth is warming due in large part to human-introduced CO2″ answer — Yes.

    Longer answer — One of my few comments to WUWT concludes as follows: “By the way, regular readers of Lucia’s Blackboard know that I am convinced by the evidence of the role of CO2 and other greenhouse gasses in AGW. And also that the temperature record (despite its problems) clearly shows that the Earth has warmed by about one degree C in the past century or so. But to me, that’s the start of the conversation, not its conclusion.”

    That conversation would center around strengths and weaknesses of the current crop of GCMs, and how to evaluate them. As models of complex dynamical processes, and as prediction engines. In simple terms, it’s not the direct effect of CO2 (~1 C per doubling), but the impact of indirect forcings that are uncertain. “Lucia’s Blackboard” regular SteveF has offered a perspective that is close to my own sense of the matter, albeit one that is better informed and more oriented to the underlying mathematics. Which is the key. As far as “I’m genuinely keen to try and find out how to talk about this stuff with people I disagree with” — you might profit more from engaging with him.

    By the way, to my surprise, my treatment by the bloggers and commentariat at WUWT has been much more civil and productive than has been the case at pro-AGW-Consensus darling RealClimate.org. Don’t take that as a compliment to WUWT. Also, the earlier part of that linked comment would have been part of my rebuttal to the false notion that I characterize Prof. Mann as a liar.

  64. Deech56 Says:

    I think Marco’s comment bears repeating:

    “At the same time the ‘significant’ differences you claim require removal of BOTH tree rings and Tiljander (+others). If you only remove Tiljander, the difference from the yes-Tiljander is minimal (before 900 the differences are largest, but excluding a small section around 800 well within 2-sigma).”

    I kind of raised a similar point earlier.

  65. AMac Says:

    Deech56 and Marco —

    I agree with your observation about the similarity of yes-Tilj/yes-dendro and no-Tilj/yes-dendro.

    To me, the broader point is that looking at how the recons compare isn’t a very good way of getting a handle on the robustness of the approaches being used. I’d object to Tiljander being added to Mann08’s proxyhopper no matter how the comparisons turned out.

    That said, if one accepts that Tiljander from the early 19th century on is corrupted by non-climate-related influences, it’s a curious picture. Adding Bad Data to a reconstruction makes it turn out… better?!?!

    Since you’re engaging in a technical discussion, it seems fair to ask each of you two Yes/No/Don’t-Know technical questions. Bart, you and everyone else are welcome to play, too.

    1. Do you think the Tiljander data series can be meaningfully calibrated to the post-1850 instrumental temperature record?

    2. Do you think that the methods outlined in Mann08 require meaningful direct calibration of each proxy used, to the post-1850 instrumental record?

    I’m off to work, maybe see you tonight.

  66. luminous beauty Says:

    1. Do you think the Tiljander data series can be meaningfully calibrated to the post-1850 instrumental temperature record?

    Yes. Difficult but not impossible. It requires corrections for discrete step changes related to physically meaningful land use metadata (what you choose to characterize as corruptions). There is nothing to indicate Mann, et al. bother to do this, however, instead simply choosing to truncate the series after 1800.

    2. Do you think that the methods outlined in Mann08 require meaningful direct calibration of each proxy used, to the post-1850 instrumental record?

    In the particular case of the Lake Korttajarvi series, since they were cut off post 1800, obviously not. Perhaps this is why they were considered ‘problematic’, no? Can you imagine an indirect calibration method?

    (hint: If A is to B as B is to C; then A is to C.)

    Can you say this is not meaningful?

  67. PeteB Says:

    Amac,

    II’m a bit wary of ‘throwing out’ proxies when there is a problem. You can find some sort of problem with most proxies and you will quickly end up with none at all. I suspect Tiljander was correct with his relationship between the proxies and temperature and the relationship that Mann’s method came up with was wrong for 2 out of the 3 proxies. Rather than throw out the proxies completely, is there no way of making some ‘best guess’ calibration for the remaining proxies, either for some part of the instrumental record or by estimating the effect of the contamination or calibrating against some other period where we have a reasonably good handle on temperatures?. It seems there is some useful signal in the earlier part of the record, it seems a shame to throw it out completely.

  68. MIkeN Says:

    JBowers, there is the issue. You say science is self-correcting. That was not the Mann paper which was corrected, but rather Kaufmann’s Arctic warming paper. After that correction was issued, another paper came out with upside-down TIljander by Mann who shared at least one coauthor with Kaufmann.
    The Kaufmann correction makes it hard to attack that paper.
    Now, without the correction, if other climate scientists had been issuing a defense, saying that data was not used upside-down, t would make the other scientists look not credible.

    It is Mann who is not issuing a correction.

  69. MIkeN Says:

    >What I *think* I pick up from people arguing about this: the very fact that others in the climate community don’t call out the errors shows its all about circling the wagons, and no-one can be trusted. For myself, I’m satisfied I’ve seen those errors acknowledged, and

    That does about sum up the differences. Let me say, that as a skeptic, I love sending people to RealClimate. People who are aware of some skeptic arguments will end up asking a few questions, and the attitude of the responses, or perhaps having their question deleted, ends up creating more skeptics than they destroy.

    You speak of the multiple threads of other evidence. It’s true, and I may still change my mind on the issue, but I always wonder if the same cutting of corners is happening as I see in other areas when I took a deep look.

  70. Quiet Waters Says:

    ” I don’t read Mann’s comment as a denial of using anything upside-down

    versus Prof. Mann’s formal response to the assertion that he used a Tiljander proxy upside-down –

    The claim that “upside down” data were used [by us in Mann08] is bizarre.”

    If either way is the right way up then the claim it is upside down IS bizarre. I don’t see why this seems to be so hard for some people to grasp. Can you kick an upside down soccer ball?

    Of course the defining word in that first sentence is “if”, Mann believed this to be the case, there is conversation to be had as to whether he is right. There is no benefit to conversation in claiming he denied* using the proxy upside down.

    *note this use of the word denied has no relation to any events from German history or subsequent revisionist claims.

  71. Deech56 Says:

    AMac, I have not followed this particular proxy enough or gone through enough analysis to answer your questions. Now if you had asked me about vaccines…

    You state your ideas about the broader point, but it seems to me that the broader point is that there are proxies for past temperatures, and that as more proxies are analyzed and better analysis is performed, we get a better sense of the temperature record over the last couple of millennia.

    That some discount this mass of physical evidence in favor of anecdotes (“Greenland was green.”) is baffling. And I am not surprised by your civil treatment on “skeptic” blogs – they don’t care about coherence, they erect a big tent and welcome anyone outside the scientific process who goes after the scientists.

  72. AMac Says:

    Deech56 —

    > they erect a big tent and welcome anyone outside the scientific process who goes after the scientists.

    Heh. Not the point of what I wrote, at all, but that’s a clever turnaround. Touche.

  73. AMac Says:

    Reading through the comments, I belatedly recognize that most of the contributing pro-AGW-Consensus advocates have a vague or incorrect idea of what a “proxy” is. So perhaps I could say a few things about that.

    As a thought experiment, suppose that we discover the records of the recently-vanished Qui-qui culture on an equatorial island off the coast of South America that had kept certain climate records. Back in 1400, they invented their version of a thermometer. They recorded daily average temperature, but using the 26 letters of the Qui-qui alphabet. Transliterated to the Roman alphabet, “A” was the coldest, with B, C, D, etc., being progressively warmer. All the way through to Z, the warmest.

    Unfortunately, the last Qui-qui died in 1995, and there’s no way of using the records to figure out just where A, B, C, etc. lie on the Celsius scale.

    However, a whaling station was established on the island in 1850, and we have daily average temperature records in degrees Celsius for 150 years, through 2000.

    Since the two records overlap for 145 years, we can try to calibrate the Qui-qui record to the whalers’ record. For instance, suppose that the first two weeks of 1910 ran as follows:

    Qui-qui — A, B, E, G, G, I, K, M, O, N
    Whalers (deg C) — 1, 2, 5, 7, 7, 9, 11, 13, 15, 14

    This would enable us to establish that A = 1 C, B = 2 C, and so forth.

    We would then use this relationship to decode the pre-1850 Qui-qui records.

    If the first ten days of 1410’s record were

    B, D, F, F, F, E, E, E, D, C

    we would impute

    2, 4, 6, 6, 6, 5, 5, 5, 4, 3

    Of course, no proxies are this clear-cut or this highly correlated to temperature. All valid temperature proxies have noise, and all have complex correlations to other factors (e.g. precipitation, wind, humidity) as well.

    Methods of analysis may be quite sophisticated — Principal Component Analysis comes to mind — but the relationship I’ve just described is at the heart of the idea of a “temperature proxy”.

    For the Qui-qui record, if P > N > L for the calibration period, then P > N > L for the reconstruction period, as well. The point of the calibration is to establish the relationship (as I did, above) and its strength.

    This is the end of the thought experiment.

    — Continues —

  74. AMac Says:

    — Continuing… —-

    Now consider Tiljander’s lightsum. You can visualize the proxy via the graph here, and download the data as an Excel spreadsheet via a link from that post. (The numbers that follow are not real values; I chose them for illustrative purposes.)

    The annual varve thickness 1941 – 1950 might look like this (millimeters) —

    0.9, 1.0, 1.1, 1.0, 1.1, 1.0, 1.2, 1.2, 1.1, 1.2

    While the CRUTEM3v gridcell temperature anomaly looked like this over these years (degrees C) —

    0.0, 0.0, 0.1, 0.1, 0.2, 0.1, 0.2, 0.2, 0.3, 0.3

    Clearly, the valve thickness is increasing over this period. And clearly, the temperature anomaly is also increasing. The two are positively correlated.

    While this illustration used pseudo data, you can inspect the actual lightsum and CRUTEM3v data for the 1850-1995 calibration period used in Mann08, at the link, above. Overall rising lightsum, overall rising temperature anomaly.

    By a more-sophisticated variant of this method, Mann08 established the positive relationship (positive correlation) of lightsum to temperature.

    This is the relationship that was used for the reconstruction period, 500 AD to 1849 AD.

    In the calibration period, warmer temperatures lead to higher lightsum values.
    In the reconstruction period, higher lightsum values signify warmer temperatures.

    However, Mia Tiljander had a different interpretation in her 2003 paper. She suggested that a cold, snowy winter led to a deep snowpack and a vigorous spring runoff, bringing more mineral silt into the lake. A warm, less-snowy winter would cause a thin snowpack and a modest spring runoff, with less mineral material settling to the lakebed

    colder and snowier winters lead to higher lightsum values.

    Post-1720, Tiljander cautioned that local human activities progressively overwhelmed this relationship.

    More road-building, bridge reconstruction, and farming causes mud to wash into the lake, leading to higher lightsum values.

    Prof. Mann and coauthors discussed this issue, but decided that Tiljander’s cautions could be overridden. He used lightsum as I have described.

    In Mann08’s 500 AD to 1849 AD reconstruction period, higher lightsum values signify warmer temperatures.

    This relationship is inverted with respect to the relationship that Tiljander proposed between lightsum and temperature. More than that, the correlation that Mann08 established for the 1850 to 1995 calibration period is invalid, irrespective of its orientation. During those years, the climate signal in lightsum was overwhelmed by the non-climate signals of road-building, bridge reconstruction, and farming.

    Look at the graphs of lightsum and ask yourself what Tiljander asked herself: “How could I extract a climate-related signal from lightsum for the 19th and 20th centuries, given the massive contamination by local human activities?”

    Her answer: It can’t be done.

    What’s your answer?

  75. PeteB Says:

    Amac,

    Do you think there is a useful signal for the Tiljander proxies in the earlier part of the record that can be used, combined with other proxies in some way to improve a temperature reconstruction ?

  76. PeteB Says:

    Amac,

    another question sorry, – it is asked in good faith, I’m genuinely interested.

    Do you disagree with the AR4 claim (particularly the bit in bold)

    It is very likely that average Northern Hemisphere temperatures during the second half of the 20th century were higher than for any other 50-year period in the last 500 years. It is also likely that this 50-year period was the warmest Northern Hemisphere period in the last 1.3 kyr, and that this warmth was more widespread than during any other 50-year period in the last 1.3 kyr.

    In other words, do you think there is a greater than 33% chance that there were other 50 year periods in the last 1300 years that were warmer ?

    This seems a very modest claim, given that all the studies from all the groups show this (and I know there are some areas of debate in most of the studies)

    http://www.ipcc.ch/publications_and_data/ar4/wg1/en/figure-6-10.html

  77. PeteB Says:

    oh damm formatting didn’t work as I expected – the bit in bold was

    It is also likely that this 50-year period was the warmest Northern Hemisphere period in the last 1.3 kyr, and that this warmth was more widespread than during any other 50-year period in the last 1.3 kyr.

  78. AMac Says:

    PeteB,

    > Do you think there is a useful signal for the Tiljander proxies in the earlier part of the record that can be used, combined with other proxies in some way to improve a temperature reconstruction?

    Good question. This was alluded to earlier in the thread by you, and by J. Bowers when s/he quoted one of the Climategate emails on last-minute revisions to the Kaufman09 paper.

    The final draft of Kaufman09 used one of the Tiljander data series as did Mann08 (I don’t recall whether it was lightsum or XRD). When that draft appeared on Science’s website as a preprint, McIntyre wrote a blog post criticizing Kaufman et al for their mistake. Unlike the authors of Mann08, the authors of Kaufman09 recognized that McIntyre was correct, and that their use of Tiljander was in error (with respect to the interpretations offered in Tiljander03).

    As a result, Kaufman et al pulled their not-quite-formally-published preprint and revised it by (1) changing from the upside-down to the rightside-up orientation w.r.t. Tiljander03, and (2) not directly calibrating it to the instrumental record. Rather, they used an indirect method, correlating this proxy to other proxies in the later part of the reconstruction period.

    So a rise in this proxy is taken to mean “warmer” by Mann08, Mann09, Kemp11, and other papers, but “colder” by Kaufman09. Mann08 and Kaufman09 share an author. Curious state of affairs.

    This was covered at CimateAudit; you will have to search there for details (and my memory could be wrong on a few of the minor points). The account at RealClimate and other pro-Consensus sites is so bowdlerized as to be uninterpretabie.

    My own view is that Mia Tiljander was likely mostly wrong in claiming that lightsum, darksum, and XRD are proxies for temperature prior to 1720. I discussed that with TCO/Scientist and others here. However, a warning. It’s heavy going, and of the contributors to this thread, only MikeN has exhibited a grasp of the fundamentals of proxy-based reconstructions that is adequate for understanding the arguments, pro and con, that are made in that and the following post.

  79. AMac Says:

    PeteB —

    > do you think there is a greater than 33% chance that there were other 50 year periods in the last 1300 years that were warmer?

    Short answer — Not enough is known about the climate of the past 1300 years to be able to answer that question.

    Long answer — The multi-century temperature reconstructions published by top researchers in this field in high-impact journals have multiple fundamental shortcomings. In data (e.g. wrong use of Tiljander) and in methodology (which we haven’t even touched on here). The publication process gets a Fail from a science-centric point of view. From some of the Climategate emails, one can see that peer-review was very superficial, at least at times. Worse, there appears to be a culture of silence, whereby scientist/advocates mute any criticism of the work of like-minded scientist/advocates. Likely for a combination of cultural, political, and career-advancement reasons.

    Given this dreary set of conditions, I have no faith that any of the multi-century reconstructions say much of anything about the actual global temperatures of the past 2,000 years.

    Happily, the picture is better when one considers longer-range recons going back thousands of years to the Younger Dryas, Wisconsin Ice Age, and earlier. These are based on proxies with a more clear-cut physical association with temperature. Notably, oxygen isotope variations in crystallizing snow.

    Unhappily, these proxies generally lack the precision in time to be useful for recons that can resolve temperature on the scale of years or decades. And, obviously, paleo-ice is very restricted in its geography.

  80. MIkeN Says:

    AMac thanks for the link. I wasn’t aware that you consider the proxies to be not correlated to temperature ever.

    Looking at your comparison of the little ice age, I think it is there, even if it’s not as obvious as for the other chart in Finland.

    I also think this thread has gotten hijacked by TIljander, and I am the primary culprit.

  81. MIkeN Says:

    A separate question, in the post it refers to 5 hiroshima bombs per second of heat accumlation, yet Roger Pielke Sr posts that ocean heat has flatlined for about 8 years now. What is the issue here?

  82. AMac Says:

    MikeN —

    I strongly disagree that this thread has been hijacked from the theme of Bart’s post.

    Bart, Chris Colose, and others have made strong claims about “how [climate] science works.” In my opinion, it is futile to tackle such a broad theoretical subject without grounding remarks in specifics. Since the theme is “science”, those details are necessarily technical in nature. The story of the use of the Lake Korttajarvi data series illustrates how one branch of climatology has diverged from “best practices” as they are known in the other physical sciences.

    I’d also note that Naomi Oreskes took a similar approach in “Merchants of Doubt”. She stated her general theme, and then filled in her picture with specifics, incident by incident.

  83. PeteB Says:

    The multi-century temperature reconstructions published by top researchers in this field in high-impact journals have multiple fundamental shortcomings….

    …. I have no faith that any of the multi-century reconstructions say much of anything about the actual global temperatures of the past 2,000 years.

    I find this very hard to accept (and yes, I have spent a lot of time a CA reading about the supposed issues, and I can just about keep up with the maths and statistics). I can see that there are some
    legitimate issues raised there, but the effect of them (IMO) tends to be very much exaggerated.

    I can see that there are lots of problems trying to tease out the signal from a very noisy background and people have tried different techniques. The IPCC ‘spaghetti graph’ of temperature reconstructions covers a fairly wide band, but all of them that go back that far would agree with the IPCC claim quoted above (which seems a pretty modest claim, allowing a 33% chance of a 50 year period of NH temperatures being higher in the the last 1300 years. If the IPCC claim was false, wouldn’t you expect at least one reconstruction to get somewhere near that ?

    Are you sure you are not falling for the ‘If we don’t know everything, we know nothing’ fallacy ?

    Sorry, another question. Do you think there is sufficient signal available from the proxies to be able to create a reconstruction and the scientists are messing it up, or there is insufficient signal available to be able to develop a reconstruction ?

  84. PeteB Says:

    I strongly disagree that this thread has been hijacked from the theme of Bart’s post.

    I agree – I think we should discuss whatever the contrarians think is their strongest point

  85. AMac Says:

    < If the IPCC claim was false, wouldn’t you expect at least one reconstruction to get somewhere near that?

    How many proxy sets are there that are reliable prior to, say,1500? 1200? 800?

    Do the different reconstructions use different proxies, or are they mostly based on the same data? If that's the case, should they be thought of as independent from one another?

    Are the different reconstructions built with different methods, or do most of them share the same approach?

    Is there diversity in the outlook of the groups doing this work, or do most of them prretty much see things the same way?

    On the last point, it seems to me that McShane & Wyner and Smith are two recent entrants into the multiproxy recon area that have brought fresh methodological approaches to this area.

    > Are you sure you are not falling for the ‘If we don’t know everything, we know nothing’ fallacy?

    Yes, I am sure. Lots of poorly-conceived work based on questionable data isn’t validated by further poorly-conceived work based on further questionable data. Would you prefer that the FDA start approving drugs and devices as if that was a good idea? Why do you describe that position as a “fallacy”?

    > Do you think there is sufficient signal available from the proxies to be able to create a reconstruction?

    Good question. I don’t know. It’s largely a matter of how big the error bars can be, and still have the recon be considered to be “informative” or “useful.”

    > whatever the contrarians think

    I’ll be interested in what you credulostrians think of my explanation of how proxy-based reconstructions are constructed, and whether you grasp the magnitude of the problem that Tiljander poses for the “move along now, nothing to see here” position.

  86. PeteB Says:

    Amac,

    I apologise if you take ‘contrarian’ as an insult – it wasn’t meant as such, I was actually trying to use a non loaded term for people that have views contrary to the ‘consensus’, I am happy to use any term apart from ‘sceptic’, – hopefully we are all sceptics.

    How many proxy sets are there that are reliable prior to, say,1500? 1200? 800?

    quite a few

    http://www.ipcc.ch/publications_and_data/ar4/wg1/en/figure-6-10.html

    Do the different reconstructions use different proxies, or are they mostly based on the same data? If that’s the case, should they be thought of as independent from one another?

    Well, they are certainly not all based on the same set of data, but I take the point that there is a limited (although gradually increasing) set of proxies. There are some that do have varying methods (e.g. Moberg)

    Surely if somebody could publish a surprising result this would happen.

    McShane & Wyne

    OK, I read the CA report + the deep climate + RC links (I’m a Mechanical Engineer by training, so I tend to get most of the maths and stats but probably miss some of the subtleties)

    From RC

    http://www.realclimate.org/index.php/archives/2010/12/responses-to-mcshane-and-wyner/

    On that specific issue, presumably just an oversight, MW apparently used the “Start Year” column in the M08 spreadsheet instead of the “Start Year (for recon)” column. The difference between the two is related to the fact that many tree ring reconstructions only have a small number of trees in their earliest periods and that greatly inflates their uncertainty (and therefore reduces their utility). To reduce the impact of this problem, M08 only used tree ring records when they had at least 8 individual trees, which left 59 series in the 1000 AD frozen network………
    ……..With just this correction, applying MW’s own procedures yields strong conclusions regarding how anomalous recent warmth is the longer-term context

    Would you prefer that the FDA start approving drugs and devices as if that was a good idea?

    Absolutlely not ! But that is not what we are asking scientists to do. We want their best assessment,and they seem to be very cautious about what they are claiming

  87. AMac Says:

    peteB —

    > I apologise if you take ‘contrarian’ as an insult – it wasn’t meant as such, I was actually trying to use a non loaded term for people that have views contrary to the ‘consensus’…

    I appreciate the effort and the thought behind it, and try to use non-loaded terms as well. They are hard to find. I was just testy…

    I’m not sure AAR4 WG1 Figure 6-10 or the table it cites provide an answer to the question, How many proxies? Probably more than “a few” and less than “a lot” in ~1500, going back to “very few indeed” prior to 1200.

    My point in mentioning McShane & Wyner and Smith was that these two papers *are* using procedures that are rather different from the rest of the crowd. All of which seem “about the same”, one to another. So if there are problems common to the PCA methods that are commonly used, those problems will be reflected in the consensus spaghetti graph.

    RC is a good source for learning about problems with work that disagrees with the Consensus. Balanced perspective and exchange of ideas — not so much.

    Thanks for writing such a thoughtful comment, and including relevant links.

    Question for you: Suppose the authors of Mann08, Mann09, or Kemp11 read this thread and realize, “By golly, what MikeN and AMac are claiming is correct. Lightsum and XRD are upside-down! We can’t use thicknessmm! All three (not four) data series are uncalibratable to the instrumental record!”

    What, if anything, should they do?

  88. Dan Olner Says:

    Amac: thanks for your reply above.

    “But to me, that’s the start of the conversation, not its conclusion… That conversation would center around strengths and weaknesses of the current crop of GCMs, and how to evaluate them. As models of complex dynamical processes, and as prediction engines. In simple terms, it’s not the direct effect of CO2 (~1 C per doubling), but the impact of indirect forcings that are uncertain.”

    Nothing there I’d disagree with, nor indeed would the last IPCC report, I don’t think: huge uncertainty exists. So, some of things I still don’t get –

    What exactly motivates your interest in the Tiljander stuff, apart from academic interest? What do the supposed errors mean for you? Perhaps to give something more concrete, how much of this realclimate post do you agree or disagree with? Specifically, the drive of the article being that millenium-long reconstructions may help us understand climate change, but that the key scientific work was done before they’d even been published, and still stands.

    Given that you agree human-induced warming is occurring, and you say that’s only the start of the conversation: OK. Why start with discussing GCMs? What about: is human-induced climate change likely to cause damage to us? How likely? What kind? With what probability? Should we be comforted by the amount we don’t know, or worried? Do you think it’s possible that climate change could potentially be harmless, or not damaging enough to bother about? Do you think we should be taking any action?

  89. dhogaza Says:

    Amac:

    > I don’t read Mann’s comment as a denial of using anything upside-down

    versus Prof. Mann’s formal response to the assertion that he used a Tiljander proxy upside-down –

    The claim that “upside down” data were used [by us in Mann08] is bizarre.

    He’s saying it’s bizarre because the signedness of the data won’t affect the multivariate analysis, therefore the claim shows a certain lack of understanding of such analysis. If they mistakenly thought the correlation was positive rather than inverted, it doesn’t change the result and it is bizarre to think it would (if you understand the basic mathematical argument, which boils down to -i*-i = i*i for all values of i).

    His statement is being (intentionally, I believe) misrepresented as is clear when it’s printed in full, rather than quote-mined as you and many others have chosen to do.

    The important claim is whether or not there’s a climate signal in the data, be it a positive or inverse correlation. Luminous Beauty quite clearly explained the actual problems related to use of those datasets for climate proxies. Mann was clearly aware of the possibility that the proxy data was possibly unusable as a climate proxy, since after all they brought up the issue in the paper itself.

    And, Amac, your rather paternalistic claim that we don’t understand what a proxy is, is rather … off the mark.

  90. Deech56 Says:

    AMac, I’d like to bring the conversation back to the point of this article. I’ve read some of the back-and-forth and some thoughtful blog discussions in which you have participated, but have you pursued normal scientific channels (letter to journal, note with alternative analysis, etc.) to offer your point of view? ( I apologize for not knowing this – maybe you covered it elsewhere.) Tamino, Eli, Ryan O’Donnell and many others have done so, and their contributions are matters of record, while contributions on various blogs aren’t. It’s like an argument in the stands at a football game (choose your definition ;-)) while the game’s on the field.

  91. J Bowers Says:

    Did Spencer and Christy ever publish a correction to their “novel statistical technique” in the literature?

  92. Steve McIntyre Says:

    Here is a translation of a Finnish interview with Tiljander coauthor Matti Saarnisto (translation by Jean S originally posted at CA in Feb 2010 http://climateaudit.org/2010/02/06/say-my-name-%E2%80%93-february-rerun/). Our PNAS Comment reporting the problem was submitted in Dec 2008 and published in Feb 2009, well before the repetition of the error in Mann et al 2009 and Kaufman et al 2009. (As an aside, for those wondering whether Mann was monitoring CA discussions of Mann et al 2008, he was. Several changes were made to his SI immediately after they were pointed out at CA, though no citation or acknowledgement was provided.)

    Saarnisto had some additional comments as well, available at the CA thread linked above:

    Sari Huovinen (journalist): Matti, your own research result has been distorted in public. Tell us shortly, what was done.
    Matti Saarnisto: Well, indeed, here … one of the persons who have been [lately] in public, professor Mann from The Pennsylvania State University. He has published several articles about the climate history of past thousand years. The last time it was last the history of last two thousand years [published] with many colleagues. In that [article], research material, from Korttajärvi near Jyväskylä, of my group was used such that the Medieval Warm Period was show as a mirror image.
    SH: That is, the graph was flipped?
    MS: The graph was flipped upside-down. And, and, … it was in Science in last August, and, …
    SH: Why was that done, how do you interpret that?
    MS: That is something I’ve tried to sort out … in this e-mail I received yesterday from one of the authors of the article, from my good friend prof. Ray Bradley in Chile, where he was traveling. There was a large group of researchers who had been handling an extremely large research material, and at some point it had happened such that this graph had been turned upside-down.
    SH: So it was not done in purpose, it was a mistake?
    MS: Well, when Bradley says so to me, I don’t doubt even a slightest moment. I hold him in high regard. He is one of the best paleoclimate researchers, and … a frequent visitor in Finland. But then that this happened yet another time in Science … in Apr… in November last year, a little before Christmas … again this Korttajärvi material, which was a part of Mia Tiljander’s PhD Thesis, Mia Tiljander is a known person worldwide, and … the article where the material appeared was published in 2003. Mia Tiljander was the first author, I was the second, and good, younger collogues of mine, Timo Saarinen and Antti Ojala, were then after…
    SH: … yes …
    MS: It has been turned twice upside-down in Science, and now I doubt if it can be a mistake anymore.

  93. Neurathian_Ship Says:

    I too would like to know whether Amac has actually contributed to the scientific literature? If not, why not?

    This is important because there are very good reasons to be suspicious of anyone who makes seemingly important claims but fails to test them anywhere other than blogs. Despite the suggestions from some in the blogosphere, for all practical purposes, science is not and never will be conducted on blogs or random websites.

    This isn’t really a comment about Amac, but the very fact that much of the criticism of climate science comes from blogs is fairly strong — albeit indirect — evidence that the people who are criticizing are, at the very least, not particularly interested in furthering the science, and in many cases are exactly as they are often described — merchants of doubt.

    I’m certainly not competent enough to know whether what Amac is suggesting is accurate or not, but I would have thought that a basic commitment to intellectrual integrity alone would be enough to persuade someone to test their ideas in the scientific literature if they believe that important mistakes have been made.

  94. AMac Says:

    Dan Olner —

    > What exactly motivates your interest in the Tiljander stuff?

    Much of it is academic, in that it’s a puzzle. What happens when you pull on a string? Another piece is that when I first read about the issue at Pielke Jr.’s blog, his the problem with the use of Tiljander in Mann08 was obviously correct — “obviously” “correct”, if you wish — after I’d done under an hour of background reading. And yet Pielke was being mocked and scorned for his stance. The “Through the looking glass” quality of the discussion continued at William Connolly’s blog Stoat (see here), and for that matter continues to the present.

    Why I think this kerfluffle is relevant to the Big Question that Bart addressed in the body of this post — we’ve already discussed that in these comments.

    “Tiljander” is a topic that I know pretty well, and I can contribute novel material to conversations that concern it. Sometimes my comments are welcome, other times, not so much. As to “what do I think about the likelihood of human-induced climate change damaging us,” etc. etc. — by now, you’ve read a range of eloquent and well-informed opinion on that and other subjects. I don’t have any particular insight, and I’ve already written more than usual, in response to your questions, upthread. It’s not clear to me why you and others should be particularly thrilled to read more.

  95. AMac Says:

    dhogaza —

    Thanks for your comment.

    > He’s saying it’s bizarre because the signedness of the data won’t affect the multivariate analysis…

    Yes, that’s what Prof. Mann is saying, and at other venues you’ve concurred with his explanation.

    > [Prof. Mann's] statement is being (intentionally, I believe) misrepresented as is clear when it’s printed in full, rather than quote-mined as you and many others have chosen to do.

    Please refer upthread to my comment of July 1, 2011 at 10:47, where I link to Mann et al’s Reply in PNAS. The relevant text in its entirety is blockquoted just below the link.

    > your rather paternalistic claim that we don’t understand what a proxy is…

    Perhaps my explanation of the basic idea of a proxy-based reconstruction (upthread, July 2, 2011 at 05:10) is helpful to you. Perhaps not. Either way is fine.

    – – – – – – – – – –

    Deech56 and Neurathian_Ship —

    No, I haven’t written any of this up for submission to a peer reviewed journal. It hasn’t been clear to me what I would write — none of this is very novel, and it likely wouldn’t reach the threshold of originality that most journals demand. Also, I don’t know R, which makes it hard to produce new work as opposed to criticism. And this is a hobby — not my field, not my day job.

    I did have an idea a year back involving pseudoproxies, but none of the people I contacted (from “both sides of the aisle”) had the time, interest, and skills to collaborate. Funnily enough, I came up with a different and much better idea this afternoon, involving only Excel. If that works out, maybe I’ll go forward as you suggest. :-)

    BTW, Neurathian, your comment is quite meta, as you no doubt intended. Arguing in blog comments about whether arguing in blog comments is a fit pasttime for Gentlemen and Ladies.

    > This isn’t really a comment about Amac… [ISTM that] the people who are criticizing are… not particularly interested in furthering the science, and in many cases are… merchants of doubt… a basic commitment to intellectrual integrity alone would be enough to persuade someone to test their ideas in the scientific literature if they believe that important mistakes have been made.

    Well, I’m glad that wasn’t a comment about me. :-)

  96. PeteB Says:

    Question for you: Suppose the authors of Mann08, Mann09, or Kemp11 read this thread and realize, “By golly, what MikeN and AMac are claiming is correct. Lightsum and XRD are upside-down! We can’t use thicknessmm! All three (not four) data series are uncalibratable to the instrumental record!”

    Well, I don’t really mind who does it (whether it was the original authors or someone else), but it would be interesting to re-do the analysis with the ‘indirect calibration’ method and see what the effect was.

    Slightly off-topic, and there’s a good chance you’ve seen it already but via this analysis of McShane and Wyner

    http://klimazwiebel.blogspot.com/2010/08/mcshane-and-wyner-on-climate.html

    …..Two schools of thoughts depart here. One, represented for instance by Mann, attempts to use all proxy records available and design a statistical method that some how can extract the signal from the noise. The paper by McShane and Wyner also fits within this school, trying to apply the Lasso method for proxy selection. It fails, according to the authors, but this may be due to characteristics of the Lasso method that render it inadequate or perhaps due to the impossibility of designing a statistical method that can successfully screen the proxy data. It is not clear what the real reason is.
    Other idea, portrayed in our recent paper argues that another way forward is to select good proxies, for which there is a priori a very good mechanistic understanding of the relationship between proxy and temperature. Here, the observed correlation between proxy and temperature is secondary and proxies with good correlations would be rejected if the mechanistic understanding is absent or dubious. Once a good set of proxies is selected, with minimal amounts of noise, any method should be able to provide good reconstructions….

    The paper zorita refers to is here

    http://coast.gkss.de/staff/zorita/Frank_etal_WIRESCllmChange_2010.pdf

  97. PeteB Says:

    From the Zorita paper

    [i]….Expert assessment to evaluate the signal of a particular record from a particular proxy archive (e.g., the lowfrequency skill of a new speleothem record) will be invaluable in trying to minimize ‘wrong figures’ being put into a large-scale reconstruction. It seems advisable at this point to use fewer, but expertassessed proxy records, rather than hundreds of proxy series, and hope that reconstruction algorithms will overcome the often huge noise components typical for many of the available time series….[/i]

  98. Dan Olner Says:

    Amac, thanks. Generally a question: could anyone please help me get a handle on the basic issue with the “upside-down” question? My understanding is the same as dhogaza’s:

    “He’s saying it’s bizarre because the signedness of the data won’t affect the multivariate analysis, therefore the claim shows a certain lack of understanding of such analysis. If they mistakenly thought the correlation was positive rather than inverted, it doesn’t change the result and it is bizarre to think it would (if you understand the basic mathematical argument, which boils down to -i*-i = i*i for all values of i).”

    Reading Amac’s questions has helped me not at all to understand why the “upside-down” down keeps on returning. This could be my own stupidity, but I would greatly appreciate any attempts to bring some clarity to the issue.

  99. Dan Olner Says:

    Amac, your thought experiment example wasn’t a proxy, it was two temperature records. A proxy would be if they, oo… er, happened to be a culture that spent more time outdoors (on average) when it was warm, and happened to record how much time they spent outdoors. Or perhaps a record of crop output, which would of course have plenty of confounding factors, but if we had a direct temperature record to compare to, we’d be able to assess it’s worth.

    ‘Proxy’ implies having to deduce the values you’re after from something else: your example was just a direct temperature record that hadn’t been calibrated.

  100. AMac Says:

    Dan Olner —

    Your question at 11:33 on “upside down” and your observation at 11:52 about my as-simple-as-possible thought experiment are related.

    In that thought experiment, the key point is that the relationship of (A, B, C… X, Y, Z) to temperature is established for the calibration period, and then applied to the reconstruction period.

    In this instance, in the calibration period, A < B < C and so on. The magnitude of the difference between A and B is 1 degree, and so on. Validation statistics would show r^2=1, and so on (per Zorita, this would be a perfect "proxy").

    These are the relationships that are then applied by investigators to the record during the reconstruction period.

    As you suggest, we could make the thought experiment more “realistic” — but more complicated — by having the Qui-qui alphabet record represent time spent outdoors or crop output, with an r^2<<1. But the assertion of the immediately-prior paragraph would stand.

    A simple way to conceptualize the "upside down" issue is to extend the thought experiment to consider the effect of the actions of two data-entry clerks, Goofus and Gallant. Goofus worked on calibration-period data, while Gallant handled the reconstruction-era record. Somehow, Goofus messed up the key file. In it, he systematically transposed the letters of the Qui-qui record. The result is that a comparison of the Qui-qui record with the whaling station archives yields the following relationship:

    * A > B > C and so on.
    * The magnitude of the difference between A and B is 3 degrees, and so on.

    Meanwhile, Gallant input the information for the reconstruction period with perfect fidelity.

    When these two files are used as input for the engine that establishes the temperatures for the reconstruction period, what will be the result?

    (This is the end of the Goofus/Gallant thought experiment.)

    Of course, in the case of Mann08, there are no careless clerks at fault. However, there are post-1720 non-climate-related local activities to consider. These activities had two effects on varve thickness, with progressively greater impact through the 19th and 20th centuries:

    * Increasing deposition of mineral matter, due to farming, road-building, and bridge reconstruction, etc. (higher lightsum).

    * Increasing deposition of organic matter, due to peat cutting, eutrophication, etc. (higher darksum).

    I hope this helps.

    You might also find the thought experiment contained in the post The Newly-Discovered Jarvykortta Proxy — II to be useful.

  101. dhogaza Says:

    AMac:

    He’s saying it’s bizarre because the signedness of the data won’t affect the multivariate analysis…

    Yes, that’s what Prof. Mann is saying

    Earlier, you said (repeatedly) than Mann wasn’t saying this, but rather was being dishonest about whether or not the proxy data was used “upside down”.

    Are you now admitting that Mann’s statement wasn’t a lie regarding the proxy usage, and that your previous statement was a bit short of the truth?

    However, there are post-1720 non-climate-related local activities to consider…

    And, of course, Mann’s paper included caveats regarding the proxies, provided an alternative analysis without them, etc.

    I suppose your point is that if Mann hadn’t so kindly laid the potential issues out on a silver platter McI and his acolytes such as yourself would’ve never realized the existence of said issues?

    I see that you’re not simply moving the goalposts, but sort of dragging them around randomly as you’re pinned on specifics.

    Hmmm …

  102. AMac Says:

    dhogaza —

    Thanks for your comment.

    I think you have made assertions of fact that are not so. Your opinions are your own, of course, but you offer no support for them. Your approach to discussion is less than civil.

    Under the circumstances, the chance that we will have a productive dialog is on the order of a rounding error.

  103. luminous beauty Says:

    Amac.

    Perhaps you could clarify what seems to me a difference between your and my understanding of the calibration process. You seem to imply that low frequency variability has some profound effect on calibration. My understanding is, the first step in calibration is to normalize the series, i.e., detrend it. Therefore, one is looking, not at decadal or longer scale variance, but annual, or in the case of tree-rings, semi-annual resolved variance. E.g., boreal tree-ring series that show decadal scale divergence, still show annual variability correlated to temperature, strongly suggesting that divergence is not a non-linear response to temperature, but has some other cause. Thus, they are useful for calibration purposes even though they reduce statistical confidence in the modern period reconstruction.

    Much the same can be inferred from land use changes surrounding Lake Korttajarvi. It is not unreasonable to assume most, if not all, of these changes represent relatively linear changes in sedimentation rates, and have little effect on the annually resolved climatic variation.

  104. AMac Says:

    luminous beauty —

    That seems to me to be a really complicated way of starting to think about proxy-based reconstruction.

    I’m beginning from a simpler place.

    Premise 1 (per Zorita, quoted upthread) is that the candidate has some meaningful relationship to temperature. Be it tree-ring width, varve thickness, or Qui-qui letter.

    Premise 2 is that this relationship holds throughout the calibration and the reconstruction period. E.g. if “wider tree-ring” meant “warmer temperature” in the 1920s, “wider tree-ring” also meant “warmer temperature” in the 1630s.

    Re: the Lake Korttajarvi data series: have you looked at the actual data (millimeters of varve thickness for lightsum, darksum, and thickness; Arbitrary Units of X-ray absorbance for XRD)? See here. The first two graphs show all four data series for 200 AD through 1995 (the final 10 years are “infilled,” i.e. not real data).

    For instance, check out the figure titled “Lightsum 1700-1995″. Lightsum is in red. Note the huge spikes in 1930 (4.4 mm), 1931 (1.7 mm), 1962 (1.5 mm), 1965 (1.3 mm), 1966 (1.6 mm), 1967 (4.6 mm), and 1968-1985 (all years between 1.0mm and 2.8 mm).

    Recall Tiljander03’s interpretation: Prior to local-activity contamination, a higher lightsum means a colder, wetter winter.

    As far as using the 20th-century lghtsum record to meaningfully calibrate lightsum during the reconstruction period: Tiljander03 implicitly advised against it. My conclusion is that it can’t be done.

    Can you suggest a way to do this? Put another way, do you think that Mann08 successfully accomplished this task?

  105. dhogaza Says:

    AMac:

    I think you have made assertions of fact that are not so. Your opinions are your own, of course, but you offer no support for them. Your approach to discussion is less than civil.

    Hand-waving dismissal and a claim of victory …

    Why am I not surprised?

  106. dhogaza Says:

    So AMac used the quote-mining technique to misrepresent Mann over at Arthur Shumway’s blog, also …

    (4) The authors have not acknowledged and corrected the error, or committed to not repeat the mistake. Yes. In their Reply published in PNAS in 2009, Mann et al. called the claims of improper use of the Tiljander proxies “bizarre.”

    Sigh …

  107. MIkeN Says:

    Dhogaza, I think you are wrong on the facts as well. Notice a link above, to a thread at Arthur Smith’s with Martin Vermeer’s comments. If I remember that thread correctly, upside-down usage was agreed to by now at least one Mann co-author.

    Again this is where I think Mann is incorrect in his reply:

    The claim that “upside down” data were used is bizarre. Multivariate regression methods are insensitive to the sign of predictors. Screening, when used, employed one-sided tests only when a definite sign could be a priori reasoned on physical grounds.

    I interpret this to mean as follows, and from the comments, I think you and Dan Olner reached the same interpretation. The second sentence implies that the algorithm is blind to the sign of the proxies, therefore the accusation of upside-down usage is bizarre. The third sentence means that while there is an exception to when the algorithm is blind to the sign of the predictor, Tiljander does not meet this exception.

    Note, these are not facts, but my interpretation of Mann’s reply,
    Looking at the facts, I think all 3 of these sentences are in error.

    Looking at Mann’s code, and it appears Martin Vermeer reached the same conclusion, sentence 3 does not apply to Tiljander. That is, the CPS screened version does assign an a priori definite sign to Tiljander. Therefore, the algorithm(CPS screened) was not blind to the sign of the predictor. And with the way the proxy was used, it was in fact used upside-down.

    I will note that for the EIV version, the algorithm is blind to the sign of the predictor. Proxy was still used upside-down, but it at least is a different issue of flipped correlations.

  108. PeteB Says:

    If you look on the Arthur Smith thread, Martin Vermeer basically agreed on the technicalities, i.e. contamination of proxies during the calibration period(the instrumental record) caused the calibration to orient 2 out of the 3 proxies in the opposite direction to Mia Tiljander and this was probably wrong.

    However he thought Mann’s reply was correct

    http://arthur.shumwaysmith.com/life/content/michael_manns_errors?page=1

    “He was specifically asked about Tiljander, and responds that the algorithm is blind”.

    And he was clearly correct, for both cases, EIV and CPS. The regression algorithm is blind. The pre-screening is a different story… that’s preprocessing, logically separate from the regression algorithm.

    “Several commenters went off of this and thought that a proxy would be flipped back if it was entered upside-down”.

    Those commenters were a bit quick perhaps… the flipping back only happens in the case of CPS: this is what the sign() operator does. For EIV, there is no “flipping”, and yet that algorithm too is “blind” to proxy orientation. So, flipping is a bit of a red herring.

    This “blindness” is a well known property of regression methods of this kind. I cannot think of a valid algorithm where this is not so. Shouldn’t even be controversial… I can see where Mann’s “bizarre” comes from.

    and finally

    AMac, I disagree strongly for reasons that should be clear to any reasonable reader by now. Mann’s reply was fully responsive to the statement as made by McIntyre in the Comment. He could not be expected to engage in mind reading, or to follow, or even be aware of, discussions on fringe blogs. I am disappointed that you continue to fail to see this, AMac, and have nothing further to add.

  109. AMac Says:

    MikeN and PeteB —

    There are two related but distinct issues here.

    1. In the last two comments, you address how Mann08’s code for EIV and for CPS each handle the Tiljander data series, specifically lightsum. Does the EIV code a priori assign a sign to lightsum? Could it flip the orientation of lightsum? What about the code for the CPS reconstructions? In the discussion at Arthur Smith’s site, Martin Vermeer took this approach as well. While I don’t read MatLab, my sense is that MikeN’s analysis is correct for both EIV and CPS.

    2. In my comment of July 3, 2011 at 14:14, I ask a simpler question.

    What is the relationship of lightsum to CRUTEM3v gridcell temperature during the calibration period (1850-1995), and how would any reconstruction engine (EIV and CPS included) apply this relationship to the reconstruction period (500 – 1849)?

    You can inspect the relationship for yourself at this blog post. Scroll to the graph labelled “Ligthsum 1700-1995″.

    * CRUTEM3v gridcell temperature rises over time.

    * Lightsum rises over time.

    Therefore, rising temperature is correlated with rising lightsum.

    Since Mann08 log-transformed the Tiljander data series, I graphed ln lightsum verus temperature. Scroll further, to the graph titled “ln Lightsum (infilled) vs CRUTEM3v.”

    Again, the positive correlation of temperature and lightsum is clear. I calculated an R^2 value of 0.0311. This is similar but not identical to the correlation coefficient presented for lightsum in Mann08’s SI.

    PeteB, have you clicked on the link and inspected these graphs?

    The point I made upthread (July 3, 2011 at 14:14) is this: The relationship between lightsum and temperature that is established for the calibration period is the one that will be used during the reconstruction period. By EIV, by CPS, by M&W10’s lasso, or by any other direct-calibration method.

    MikeN — Do you agree?
    PeteB — Do you agree?

  110. PeteB Says:

    Hi Amac,

    I’ve had a look – this all seems to agree with what Martin said at Arthur’s thread. (I get the feeling I am missing something)

    Yes – Lightsum has a positive correlation during the calibration period, Yes, of course, this positive correlation will get used during the reconstruction period under any method (EIV, CPS)

    The difference, as I understand it, between EIV and CPS is that CPS has a screening step, which if the correlation during the calibration period was negative would have thrown it out because all the Tiljander proxies were categorised to have a positive relationship.

    There is no screening step for EIV, so if there had been a negative correlation during the calibration period this negative correlation would have been used for the reconstruction period

  111. AMac Says:

    PeteB —

    Thanks for the response.

    > Lightsum has a positive correlation during the calibration period, Yes, of course, this positive correlation will get used during the reconstruction period under any method…

    So, Mann08 uses lightsum as follows:

    * A higher lightsum value means a warmer year.

    * A lower lightsum value means a colder year.

    It seems that you, MikeN, and I are in agreement on that point.

    Tiljander defined lightsum as the measure of the mineral content of a varve. From her 2003 paper —

    The accumulation of the mineral lamina is most likely a short-term event and the thickness of the mineral layer is directly related to the duration and strength of the spring flood… In mild winters the snow cover remains thin, but the ground is often frozen to prevent erosion. Mild winters, with frequent snow thaws, lead to weak spring floods that cause only minor mineral matter influx into the lake system… [page 570]

    …the amounts of inorganic and organic matter, form the basis of the climate interpretations. Periods rich in organic matter indicate favourable climate conditions, when less snow accumulates in winter by diminished precipitation and/or increased thawing, causing weaker spring flow and formation of a thin mineral layer… More severe climate conditions occur with higher winter precipitation, a longer cold period and rapid melting at spring, shown as thicker mineral matter within a varve. [page 571]

    So, Tiljander03 interprets lightsum as follows:

    * A higher lightsum value means a colder and snowier winter.

    * A lower lightsum value means a warmer and drier winter.

    Do you think Mann08’s calibration of lightsum is consistent with Tiljander03’s interpretation?

    As I have stated, my opinion is that Mann08’s calibration is inverted (“upside-down”) with respect to Tiljander03’s interpretation. This came about because during the 19th and 20th centuries, climate effects on lightsum were overwhelmed by local-activity effects (e.g from farming, road building, and bridge reconstruction). This means that lightsum cannot be meaningfully calibrated to the instrumental record.

    Do you agree or disagree?

    I hope that Chris Colose, Deech56, Eli Rabbett, Dan Olner, luminous beauty, J. Bowers, Marco, and Quiet Waters will weigh in with their views on this, as well.

  112. AMac Says:

    By the way, there is a common pattern to discussions of Mann08’s use of the Tiljander data series. Earlly on, scientifically-literate advocates of the pro-AGW Consensus position offer their evaluation of the merits of my comments on Tiljander (basically, “Mann08 is right and AMac is wrong”). But as we go through the technical pieces (as on this thread), interest in the subject dwindles, and active commenters fall silent. Arguments aren’t followed to their conclusions, and their implications thus go unexplored.

    Arthur Smith’s burst of coverage of Tiljander (previously cited on this thread) is an example of this phenomenon.

    Upthread (June 28 at 23:05 and again on July 4 at 08:41), PeteB quoted Kemp11 author Martin Vermeer’s parting remarks to me on the second of Smith’s posts, “Michael Mann’s errors.” In light of this pattern, they bear repeating.

    AMac, I disagree strongly for reasons that should be clear to any reasonable reader by now. Mann’s reply was fully responsive to the statement as made by McIntyre in the Comment. He could not be expected to engage in mind reading, or to follow, or even be aware of, discussions on fringe blogs. I am disappointed that you continue to fail to see this, AMac, and have nothing further to add.

    Vermeer states three grievances. The first (“Mann’s reply was fully responsive…” appears incorrect, given Vermeer’s earlier agreement with my interpretation of the orientation of lightsum. The second (“mind reading”) and third (“following… fringe blogs”) are not relevant to a discussion of Mann08’s methods in employing lakebed varved sediments as climate proxies.

    PeteB — As you quoted Vermeer’s comment twice in this thread, this seems like a fair question. What do you see as the substance of his objection to my analysis of Mann08’s use of the Tiljander data series (“AMac, I am disappointed that you continue to fail to see this…”)? Why did it follow that Vermeer would “have nothing further to add”?

  113. Dan Olner Says:

    Amac:

    “The problem lies in the failure of the paleoclimate community and the broader climate-science community to apply the corrective measures that are standard for non-postmodern and non-postnormal physical science (per Judy Curry’s definitions).”

    It’s going to be some months before I have time to dig into the technicalities enough to satisfy myself either way; in the meantime, it’s the thrust of this quote that interests me. You’re saying, effectively, all of climate science has shown itself to be unscientific. The first thing there I don’t understand: why do you expect “the broader climate science community” to be responding to papers about proxy temperature reconstructions? If there was an error in a paper on stomatal growth in arabidopsis, would you expect the broader plant biology community to automatically take an interest? Would you conclude that, if they didn’t, plant biology wasn’t being done properly?

    If you’re interested in understanding temperature proxy analyses, OK. But you’re arguing for much more than that. If I can put it as a statement: “Mann08 and related work reveals that the broader climate science community are not doing proper science.” That’s what your line I quoted above appears to say. Is that correct?

    The realclimate article I linked to, in effect, says the opposite: if Mann’s stuff turned out to be completely wrong, climate science as a whole carries on regardless, and is right to do so. To come back to my previous example, it would need to be something as ground-breaking as showing the Calvin cycle was fundamentally wrong (which is pretty much a physical impossibility at this point.)

    For myself, this is what I’m interested in: finding out ways to understand how some people conclude – as MikeN does, and as you appear to, though you haven’t quite said it outright – that ‘the broader climate science community’ is tainted and untrustworthy, and that IPCC errors mean the whole thing is tainted and untrustworthy. (And equally, if you’re able to convince me that they are…)

    So another question then: do you think we should be taking any action to reduce carbon output?

  114. PeteB Says:

    AMac,

    Yes, yes – I already agreed all that ages ago (as did Martin on the Arthur Smith thread)

    http://ourchangingclimate.wordpress.com/2011/06/22/climate-science-scientific-method-skeptics-not/#comment-13455

    to quote myself further upthread (I think before you joined in the
    discussion)

    to be fair I think MikeN is mostly correct on the technicalities of what actually happened (some of this is copied from the MikeN, Martin Vermeer + AMac posts elsewhere)…..
    ..So 2 out of the 3 correlations are different to the original authors interpretation. I’m pretty sure that Mann was unaware of this or he would have raised it in the paper…..

    Hopefully this confounds your expectation that

    scientifically-literate advocates of the pro-AGW Consensus position offer their evaluation of the merits of my comments on Tiljander (basically, “Mann08 is right and AMac is wrong”).

    Where we have different views is Mcintyre’s comment and Mann’s reply – let me compose a detailed reply

  115. PeteB Says:

    Amac

    McIntyre and McKitrick comment

    http://www.pnas.org/content/106/6/E10.full

    ….Their non-dendro network uses some data with the axes upside down, e.g., Korttajarvi sediments, which are also compromised by agricultural impact (M. Tiljander, personal communication), and uses data not qualified as temperature proxies (e.g., speleothem δ13C)…..

    Mann’s reply

    http://www.meteo.psu.edu/~mann/shared/articles/MMReplyPNAS09.pdf

    …The claim that ‘‘upside down’’ data were used is bizarre.

    Multivariate regression methods are insensitive to the sign of
    predictors. Screening, when used, employed one-sided tests
    only when a definite sign could be a priori reasoned on physical grounds. Potential nonclimatic inf luences on the Tiljander
    and other proxies were discussed in the SI, which showed that
    none of our central conclusions relied on their use…..

    To me, it just looks like they are talking past each other.

    Their non-dendro network uses some data with the axes upside down is not a clear statement of the problem. As Mann says, if you try and use data upside down the correlation would just get flipped from positive to negative. That’s why he called it a ‘bizzare’ claim.

    If McIntyre had said something like ‘contamination of proxies during the calibration period(the instrumental record) caused the calibration to orientate 2 out of the 3 proxies in the opposite direction to Mia Tiljander’ then Mann might have had a chance of understanding what he was on about

    It’s obvious this is the one point of disagreement that Martin Vermeer had, he explicitly agreed with what you on what had happened, but just disagreed that Mann’s reply was misleading.

    Also, just out of interest I tried posting a question on Tiljander on the latest Real Climate open thread to see if it got censored – it got approved.

  116. AMac Says:

    Dan Olner —

    It’s helpful when you accompany paraphrases with direct quotes. I wish everyone would consider doing that.

    > You’re saying, effectively, all of climate science has shown itself to be unscientific.

    No. I’m saying that the paleoclimate community has failed to apply the corrective measures that are standard for non-postmodern and non-postnormal physical science. Since well-known members of the broader climate-science community have weighed in forcefully in Mann08’s defense — and none have taken a different position — this observation also has some broader applicability.

    > If there was an error in a paper on stomatal growth in arabidopsis, would you expect the broader plant biology community to automatically take an interest?

    The most revealing part of your analogy is that it’s fictional! It would be much better to cite an actual parallel case, and then discuss its similarities and differences.

    I assume that you cannot do so.

    Returning to your hypothetical: was this faulty paper published in the Annals of the Simple Flowering Plant Society of Western Pennsylvania, or in PNAS? Were its findings trumpeted as being key insights to the field? Was it cited often, or did it lapse into well-deserved obscurity? Did other high-profile articles build on its methods and conclusions?

    > The realclimate article I linked to [says that] if Mann’s stuff turned out to be completely wrong, climate science as a whole carries on regardless, and is right to do so.

    The article’s observation is trivially true, e.g. Mann08 says nothing about GCMs. But this is a classic “have your cake and eat it, too” moment. Mann08 was a landmark because, RealClimate said at the time, its findings are important and novel and add much to the confidence we can place in the Consensus view of the last 2,000 years of the Earth’s climate. Yet, you quote RealClimate opining that it wouldn’t mean anything if Mann08’s findings were wholly invalid. Important or inconsequential: which is it?

    > do you think we should be taking any action to reduce carbon output?

    Simple answer: yes.

  117. Dan Olner Says:

    Amac: “It’s helpful when you accompany paraphrases with direct quotes.”

    Eh? I started by quoting you directly.

    “The most revealing part of your analogy is that it’s fictional!”

    Yup, I used a thought experiment. You’ll have to let me know what your criteria are for when that’s OK (Qui-qui) and when it isn’t. Or perhaps I’m reading snark into your comment that isn’t there. If so, I apologise for being passive-aggressive back!

    “Mann08 was a landmark because, RealClimate said at the time, its findings are important and novel and add much to the confidence we can place in the Consensus view of the last 2,000 years of the Earth’s climate.”

    I didn’t find anything saying that in the article. It did say, “in comparison with previous reconstructions, the current analysis does not provide many surprises.”

    And the basic point still remains: you don’t need millenial temp reconstructions to make strong conclusions about climate change, only physics.

    Next question then: if you think action needs to be taken to reduce co2 output, is that because you think climate change stands a chance of being dangerous? (I nearly said, “I presume that’s because you think there’s a likelihood climate change will be dangerous,” but didn’t want to put words in your mouth.)

    It is, of course, possible to argue that a) climate change is dangerous and b) climate science has severe problems, which is what you appear to be saying. But, another question: do you appreciate the effect that focusing on minor errors has on the public’s view of climate science? Do you think there are any problems with how those minor errors get turned into an illusion of major flaws with the whole of climate science?

    I’m presuming you don’t think the flaws are major – otherwise, why would you say you thought co2 should be reduced?

  118. AMac Says:

    PeteB —

    I don’t think you mean it to be, but this is tiresome.

    > Hopefully this confounds your [AMac's] expectation that “
    scientifically-literate advocates of the pro-AGW Consensus position offer their evaluation of the merits of my comments on Tiljander (basically, “Mann08 is right and AMac is wrong”).

    Why would it?

    * Recall that Martin Vermeer ended our exchange by saying, “I am disappointed that you continue to fail to see this, AMac, and have nothing further to add.” He has since co-authored a paper that uses upside-down Tiljander.

    * Upthread, you make two narrow statements. “MikeN is mostly correct on the technicalities of what actually happened…” OK, specify where MikeN is in error re: Mann08. And, “So 2 out of the 3 correlations are different to the original authors’ interpretation…” No, all four — actually four out of three! — correlations are different from Tiljander03’s interpretations. And what broader conclusions have you drawn after reaching those two narrow insights?

    Regarding McIntyre and McKitrick’s Letter to PNAS, you argue that their text was so curt as to be too difficult for Mann08’s authors to understand. M&M claim that the editor at PNAS required such extreme brevity; it wasn’t their choice. Do you have reason to think otherwise? More tellingly, when the matter of upside-down Tiljander was brought to their attention, at least two of Kaufman09’s authors (McKay and Kaufman) immediately understood the issue, and revised their manuscript. See J. Bowers’ quote of McKay’s email upthread (June 30, 2011 at 10:39).

    > I tried posting a question on Tiljander on the latest Real Climate open thread to see if it got censored – it got approved.

    Yes, and interested parties should click on the link I added to read that comment for content and for tone. Do you think that you offered a forthright portrayal of the issues we are discussing to RealClimate’s readership?

  119. AMac Says:

    Dan Olner —

    > Amac: “It’s helpful when you accompany paraphrases with direct quotes.”
    > Eh? I started by quoting you directly.

    Dan, that wasn’t meant to be a snarky or passive-aggressive statement. It was a sincere expression of thanks, precisely because you took the time to copy-paste my words into your comment, before adding your own remarks.

    > Or perhaps I’m reading snark into your comment that isn’t there.

    Sorry, again, for the misunderstanding. No snark intended. The serious point is that an actual example of an equivalent faulty process in another physical science would be very helpful. With your arabidopsis one, we would have to spend a lot of time fleshing out the hypothetical particulars, before beginning a meaningful exchange of views.

    Actually, there ought to be numerous partially-analogous instances in the medical literature, concerning drug testing and clinical trials. I’ve come across one website where the scientist-blogger comments acidly on cases of misconduct, but I don’t have a link. And it’s a big problem that many such cases involve accusations of fraud, which is not at issue here. There’s also last week’s story of Medtronic’s behavior in bone-regeneration trials, which is still unfolding.

    Anyway — yes, such episodes do bring some discredit to the medical specialty in question, and even to medicine as a whole.

    > if you think action needs to be taken to reduce co2 output, is that because you think climate change stands a chance of being dangerous?

    Short answer: I think that, all things being equal, it would be prudent to reduce CO2 output, keeping in mind the potential benefits of the proposed actions (i.e. their effects on future CO2 levels) and their costs. I think climate change does stand a chance of being dangerous.

    > do you appreciate the effect that focusing on minor errors has on the public’s view of climate science?

    I have opinions on that subject. First, define “minor”.

    > Do you think there are any problems with how those minor errors get turned into an illusion of major flaws with the whole of climate science?

    I don’t know of anyone who self-describes as “pro-illusion.”

    The answers to your final two questions are painfully simple. And also painful. Scientists should “return to their roots” and behave like scientists. Or perhaps that’s an absurd notion, and the only alternative is to muzzle technically-minded critics. We had to destroy the village to save it.

  120. AMac Says:

    And now, I have to get some work done… Thanks to Bart for hosting this discussion.

  121. PeteB Says:

    Amac,

    OK, this has probably run it’s course, but I felt we made progress

    Hopefully this confounds your [AMac's] expectation that scientifically-literate advocates of the pro-AGW Consensus position offer their evaluation of the merits of my comments on Tiljander (basically, “Mann08 is right and AMac is wrong”).

    Why would it?

    Because all through this thread (even before you joined it) I have agreed with all the technical points that you and MikeN have made, apart from the fact that you thought Mann’s reply to McIntyre was misleading (and MikeN thought it met the someone else’s definition of fraud!) and I didn’t.

    I even said

    I suspect Tiljander was correct with his relationship between the proxies and temperature and the relationship that Mann’s method came up with was wrong for 2 out of the 3 proxies.

    So I didn’t say “Mann08 is right and AMac is wrong” I said the opposite.

    And what broader conclusions have you drawn after reaching those two narrow insights?

    That automated calibration / reconstruction algorithms will get it wrong for some time series

    you argue that their text was so curt as to be too difficult for Mann08′s authors to understand. M&M claim that the editor at PNAS required such extreme brevity

    Well I think it is not primarily the brevity that is the problem, it’s the clarity. Can you not see what is wrong with this : ?

    Their non-dendro network uses some data with the axes upside down, e.g., Korttajarvi sediments, which are also compromised by agricultural impact (M. Tiljander, personal communication),

    Yes, and interested parties should click on the link I added to read that comment for content and for tone. Do you think that you offered a forthright portrayal of the issues we are discussing to RealClimate’s readership?

    It directly raised the the use of the Tiljander proxies and possible problems trying to calibrate to the instrumental record because of contamination during the period of the instrumental record.I thought this was all censored at real climate ?

  122. willard Says:

    AMac,

    These two suppositions deserve due diligence:

    > **Since** [1] well-known members of the broader climate-science community have weighed in forcefully in Mann08′s defense — and [2] none have taken a different position [...].

    The first supposition looks like a paraphrase. So this applies:

    > It’s helpful when you accompany paraphrases with direct quotes.

    The second supposition is not a paraphrase. It would be tough to paraphrase an absence of testimony. Setting aside for now the appeal to ignorance, I’m not sure how to verify or falsify that claim.

    First, we might need enough examples of “members of the broader climate-science community” to induce who’s in and who’s out this set. Second, we might need a criteria to determine how many counter-examples suffice to prove the claim wrong, i.e. how crisp is “none”?

    ***

    And then, we’d need to pay due diligence to the implicit inference regarding the broader applicability of the argument:

    > I’m saying that the paleoclimate community has failed to apply the corrective measures that are standard for non-postmodern and non-postnormal physical science. Since well-known members of the broader climate-science community have weighed in forcefully in Mann08′s defense — and none have taken a different position — this observation also has some broader applicability.

    We can see that Judith Curry’s definitions of “non-postmodern and non-postnormal physical science” have been paraphrased. Until we have the always helpful direct quote to these definitions, we’ll “presume” that this reverts to what was being referred almost one year ago as:

    > Inerrancy. Cargo cult science.

    http://noconsensus.wordpress.com/2010/08/02/amac-on-tijander-and-hockeysticks/#comment-33714

  123. Jim Bouldin Says:

    Amac is addressing, using the specific example of sedimentation, the more general issues raised by Loehle (2009) using tree rings–the implications for climate reconstructions, of biological or physical processes that change in rate or (even worse) in direction.

    This is a very important issue, it needs to be addressed, and he is raising it without ad hominem attacks or larger accusations. A number of people are working on it, including myself. And without going to details, since it’s unpublished work in progress, with some good success.

    Loehle, 2009. A mathematical analysis of the divergence problem in dendroclimatology.

  124. Tom Says:

    Clicking through the link that willard provides leads to a series of statements that seem pretty obviously true and certain statements that have not been successfully contested at all.

    Amac has since moderated his tone, and I salute him for that. But I don’t see anything in what he wrote there that requires even rephrasing, let alone an apology.

  125. Jim Bouldin Says:

    And I just posted a delayed answer to PeteB’s question at RC, for those interested.

  126. Quiet Waters Says:

    “The serious point is that an actual example of an equivalent faulty process in another physical science would be very helpful.”

    Would this qualify? http://www.nature.com/nature/journal/v390/n6655/full/390100b0.html

    (The high level of genetic variability was because they were unknowingly sampling more than one species of vole.)

  127. J Bowers Says:

    Amac — “I’m saying that the paleoclimate community has failed to apply the corrective measures that are standard for non-postmodern and non-postnormal physical science.”

    But Kaufman et al did correct themselves, in the peer reviewed literature for all to see. It’s in the email I quoted. Why do you say “the paleoclimate community” when a clear example has been given of the complete opposite?

    And why do you lump “the paleoclimate community” in with post normal science, when PNS isn’t even science (Ravetz’s own words).

  128. AMac Says:

    PeteB @ 17:43 —

    Thanks for the review of your current view of the use of Tiljander’s data series in Mann08. You are right, on the important points you discuss, we are in basic agreement. I recall one other person changing their mind on this subject after a blog-comment discussion. So I appreciate your open-minded approach.

    > I thought this was all censored at real climate?

    “Censorship” is what a government might do, IMO. RealClimate’s owners pay to host their blog, and have the right to run it as they see fit, aggressive moderation included. Of course, others have the right to complain that RC doesn’t live up to its professed standard (“Science for Climate Scientists” or something). And to not trust it, etc. At any rate, their moderators do sometimes allow skeptical comments in, as you can see. Makes threads more interesting, I guess.

    Willard @ 19:16 —

    > These two suppositions deserve due diligence
    > [quote of AMac]
    >The first supposition looks like a paraphrase. So this applies:
    >> It’s helpful when you accompany paraphrases with direct quotes.

    I usually follow you, even if it takes a re-read. This time, not so much. Me offering an opinion doesn’t look like a paraphrase of another participant in a blog-thread discussion. Because it isn’t.

    Anyway. Well-known members of the broader climate-science community who have weighed in forcefully in Mann08′s defense… I’d include Gavin Schmidt and Martin Vermeer. OK, you got me, that’s thin gruel for that forceful statement. I have to modify it.

    I’d say implicit support has come from Mann08’s authors, Mann09’s authors, Kemp11’s authors, Kaufman09’s authors, and the PNAS and Science editors and peer-reviewers through whose hands these manuscripts passed. I would add the co-bloggers at RealClimate to this list (but see Jim Bouldin in this thread).

    As far as my opining that no well-known members of the broader climate-science community have challenged Mann08’s use of the Tiljander data series, you say —

    > …Setting aside for now the appeal to ignorance, I’m not sure how to verify or falsify that claim…

    You’re being modest. It’s easy, and there’s no appeal to ignorance. Just name the candidate well-known members of the broader climate-science community who you think have questioned Mann08.

    At her blog, Curry has written about postmodern and post-normal approaches to science. If you want to supply quotes, we can discuss them. In the meantime, the link you supply to the Air Vent will do. I don’t like the tone I used in that comment — too strident — but it lays out my view of the broader problems with Mann08’s use of Tiljander.

  129. AMac Says:

    Jim Bouldin 21:24 —

    Thanks for the kind words and for writing that response to PeteB’s comment at RC. Helpful.

    J. Bowers @ 1:17 —

    > But Kaufman et al did correct themselves…

    Yes, greatly to their credit, they did. You may not have read the correction, since it’s behind the Science paywall (here). I think it’s fair to say that few readers would get the sense that there was an interesting story behind Kaufman and co-authors’ decisions to truncate the Lake Korttajarvi data set and flip the relevant series back to rightside-up.

    So the tally would be:

    Mann08 — no error acknowledged, no correction.
    Kaufman09 — error acknowledged and cryptically described, corrected.
    Mann09 — no error acknowledged, no correction.
    Kemp11 — no error acknowledged, no correction.

  130. Dan Olner Says:

    Amac: I’ve just been trying to read up on “post-normal science”, now my head hurts. (e.g. Here’s Judith.) As far as I understand it, some people are arguing: climate science is so embedded in politics and so run-through with uncertainty that it’s no longer amenable to normal scientific method. But as I read it – e.g. over at WUWT, who hosted the originators of the idea – it appears to do two opposite things: 1) berate climate science for not being a ‘proper science’ 2) enthuse for a future where science is ‘democratised’ by input from the web.

    Just reading Ravetz at WUWT: there are so many clearly false ideas in there, I don’t know where to start. e.g.

    “We can begin to see what went seriously wrong when we examine what the leading practitioners of this ‘evangelical science’ of global warming (thanks to Angela Wilkinson) took to be the plain and urgent truth in their case. This was not merely that there are signs of exceptional disturbance in the ecosphere due to human influence, nor even that the climate might well be changing more rapidly now than for a very long time. Rather, they propounded, as a proven fact, Anthropogenic Carbon-based Global Warming.”

    Let’s ignore the pointless ‘evangelical’ stuff, and look at the last line: propounded as proven fact? When did that happen? All of the essential radiative physics is about as close to proven as physics ever gets, and is a solid foundation for arguing that AGW is highly likely – and that’s exactly what’s happened. An argument has been built saying it’s highly likely. Ravetz is perpepuating this idea of scientists as Gods of Certainty that’s been growing (See e.g. this stuff, my blog, going back to Hayek and now doing the rounds) – it’s utter nonsense.

    I’m not implying you agree with any of this, but you’ve been accusing some scientists of this postnormal/postmodern thing – what exactly do you mean? Does your take on it match Ravetz?

    This seems all upside-down to me, ironically enough. Do you not see any irony in Watts hosting an argument that climate science is `postnormal’?

    On your last point to Jim Bouldin, the ‘tally': what do you think should happen with those? As far as I know, if more work gets done that supersedes the older stuff or improves it, that’s how science progresses. No scientist is required to respond to blog posts. It doesn’t actually need scientists to declare mea culpas either. It *does* need them to stop using ideas that have been superseded – so I guess you’re argument is, this hasn’t happened? How do you think it’s *going* to happen unless someone gets a paper in or does a better analysis?

    On my “minor” point above: I think the Mann stuff is minor, compared to the massive damage being done to science as a whole by attacks on climate science. The ‘postnormal’ stuff is part of that, and Ravetz’ article nicely captures the sort of narrative I’m talking about.

    I mean – here’s the Heartland Institute arguing for ‘restoring the scientific method’. It’s so Orwellian, it makes me feel a little queezy. A couple of questions to end on, then: do you agree that the Heartland Institute has not exhibited good scientific standards up to now (to say the least)? Do you agree that is damaging to science as a whole? Does its behaviour bother you?

    I guess my point: I think – could be wrong – that you’re interested in tree rings etc because, you’ve argued, science isn’t being done properly. If good science is your motivator, are you not worried by… I’m repeating myself…

  131. J Bowers Says:

    Kaufman et al’s correction changed nothing in their concusions. I once read a newspaper article on a murder
    which had a grammatical error in the text. It didn’t overturn the coroner’s verdict of homicide into accidental death.

    If you’re so hung up on Mann et al not publishing a correction, how come you’re not so hung up on Spencer and Christy not publishing a correction? The same goes to Steve McIntyre.

  132. AMac Says:

    Dan Olner —

    Seeing how much background reading you did, and how much thought you put into the 10:46 post, I wish I’d never mentioned the word “postnormal.” Before going farther, then, let me apologize.

    > All of the essential radiative physics is about as close to proven as physics ever gets, and is a solid foundation for arguing that AGW is highly likely…

    I agree. RTE counts as “settled science,” as in “suitable for inclusion in textbooks for high school and university students.” The contentious issues (and the uncertain issues) have to do with other forcings than the direct effect of CO2 concentration.

    > you’ve been accusing some scientists of this postnormal/postmodern thing – what exactly do you mean? Does your take on it match Ravetz?

    > I didn’t know Ravetz was posting at WUWT, and haven’t read his material there. I will follow your link at some point. I think the conversation (the one I’m interested in, anyway) would have been better served if I had stuck with simpler ideas.

    > Do you not see any irony in Watts hosting an argument that climate science is `postnormal’?

    Yes.

    > On your last point to Jim Bouldin, the ‘tally’: what do you think should happen with those?

    That was a response to J. Bower’s comment of 1:17. When mistakes enter the literature in obscure, low-impact journals, the formal result when they are noticed is, typically… nothing. To the extent that those in the sub-field care, the reputation of the authors takes a hit, which can show up in things like how much respect their other work is accorded, invitations to speak at conferences or serve on panels, career advancement prospects, and so on.

    For high-profile work published in high-impact journals, people are going to pay more attention. The outcome varies. Upthread at 00:45, Quiet Waters offered one example. When brought to the attention of the authors, they acknowledged it. Its implications for the published work turned out to be so severe that the authors didn’t just issue a Correction, but retracted the paper. Embarrassing I’m sure, but schadenfreude on the part of the authors’ peers was probably tempered by “there but for the grace of God go I.” At least to an extent.

    > No scientist is required to respond to blog posts… How do you think it [correction?] is *going* to happen unless someone gets a paper in or does a better analysis?

    Authors are required to notice Comments to the journal that published their work. They can expect to be held accountable for their response — the whole “bizarre” thing, for Mann08’s authors. “Well, that response was written that way because the authors didn’t understand what McIntyre and McKitrick were trying to say.” That seems pretty lame to me.

    If authors respond to another’s blog post, it seems to me they lose the cover of, “Well we didn’t read that post, and were unaware of the issues raised there.” Do you agree?

    > do you agree that the Heartland Institute has not exhibited good scientific standards up to now?

    Yes, that seems right to me, but I haven’t paid a lot of attention. And that doesn’t say that somebody is in error, because they are at the Heartland Institute. If someone there says 2+2=4 while others are claiming 2+2=5 … 2+2 still equals 4.

    > Do you agree that is damaging to science as a whole?

    There’s a whole lot of ferment about creationism and intelligent design. Does that “damage science”? Well, it depends what you mean. Does Science need everybody to agree with reasonable ideas? With sophisticated ideas (e.g. evolution) that are correct? Science itself is a messy business, as we have been discussing. So it’s unsurprising that Science’s interaction with the broader society is even messier. This isn’t a replay of the dark days of the 1930s and 1940s, with the Nazis’ “Jewish science” and Lysenko’s triumph over genetics, where dissenters were sent to the gulag.

    > If good science is your motivator, are you not worried by… I’m repeating myself…

    I have had some original things to say about Mann08 and the Tiljander data series. The resistance of the Consensus establishment has interested me. Holy Crusade? More like a fascinating puzzle combined with a pebble in the shoe. I have lots of other opinions, you probably do as well. Why mine should be consistently featured on the front page of the Huffington Post, I’m not sure… I’m repeating myself…

    If I had a time machine, I would excise my use of “post-normal” from this thread. Perhaps “postmodern” as well. The parable of “The Emperor’s New Clothes” is a better concept for illuminating certain aspects of Tiljander/Mann08. The caveat is that the story begins with the readers knowing that the emperor is indeed nekkid. In real life, we don’t know the factual circumstances, until we investigate.

    So In simple terms,

    1. What clothes is the emperor wearing, if any?

    2. What are the implications, if any, of the emperor’s wardrobe? For the royal court, for society as a whole, etc.

    If the best we can do with #1 is akin to “I don’t know, and it doesn’t matter,” it becomes very hard to think clearly about #2.

    Off to work. I’ll check in again tonight.

  133. AMac Says:

    J. Bowers @ 14:19 —

    See my immediately-prior post. I don’t think you like my explanation, but it’s all I’ve got.

  134. Jim Bouldin Says:

    Amac,

    We are not in lockstep in what we do at RC, even if we are in general philosophical agreement about what we want to accomplish and the limits of what we will tolerate in the comments. We have definite differences of opinion on a whole number of things which we are not afraid to state to each other, and which are respected and listened to. In fact, RC is quite amazing in this respect.

  135. willard Says:

    AMac,

    Sorry about the floridity. You claim you did not understand me, and yet had an answer for about everything I was asking:

    (1) We have names like Gavin Schmidt and Martin Vermeer as candidates for the existential proof of:

    > Well-known members of the broader climate-science community have weighed in forcefully in Mann08′s defense.

    If we’re these collaborated with the usual suspect, we need to adjust our identifier “well-known members of the broader climate-science community”.

    (2) If I read you well, only one name would suffice to disprove that (rephrasing your supposition):

    > No member of the broader climate-science community have taken a different position.

    I believe Jim Bouldin might be someone belonging in the broader climate-science community. (Have you ever forcefully weighted in Mann08’s defense, Jim?)

    Judith Curry might belong in that set too, I believe. We can presume that she has taken a very different position from “weighting in forcefully in Mann08’s defense.”

    From the top of my hat, I also recall Andy Russell:

    > Ok, that sounds pretty convincing. Thanks for the info.

    http://andyrussell.wordpress.com/2010/06/15/the-hockey-stick-evolution/#comment-597

    Some qualification is needed here too, I believe.

    (3) You’re willing to let go of the “post-normal” epithet, or at least Judith Curry’s interpretations. My own gut feeling is that she has misrepresented Jeremy Ravetz’ position. But I stopped trying to share my understanding of Ravetz’ concept:

    http://rabett.blogspot.com/2011/02/that-ol-devil-rabett.html

    http://shewonk.wordpress.com/2011/02/05/pns-pretty-nonsensical-stuff/

    “Post-normal” is too ugly to be worth the effort.

    (This is my lame excuse for not having answered back Rattus. Sorry, Rattus. Email-me if you want it.)

    (4) Incidentally, while reading back the threads mentioning Tiljander, I came across this mention of PostModernScience, in response to TCO’s opinion that Mann or Gavin do not seem to concede the contamination of the Tiljander sediments:

    > No, they clearly don’t. In my opinion, they aren’t playing by the rules of Science, but by the rules of PostModernScience. The First Rule of PMS is, “there’s no such thing as PMS, we are all upstanding traditional Scientists, and I’m offended that you suggest otherwise!” The second rule is, all the methods of scientific inquiry can be used, as long as they lead to the desired result. If they don’t, all the methods of rhetoric and forensics can also be employed, by our side only.

    http://climateaudit.org/2010/08/01/the-no-dendro-illusion/#comment-237955

    I would be interested in a more serious development of that idea. The only thing I can offer for now is “post-modern contamination”.

    (5) In any case, here are my own conclusions about l’affaire Tiljander:

    Even a seeming scientific case like this one can get really complicated enough to understand without delving into it. If we trace back all the conversations you had about Tiljander this past year, we can see that lots of people have prejudices that do not stand up scrutinity. I have observed this “niche effect” before.

    A seemingly technical point always steps out to reach “larger” conclusions, if I may borrow Jim Bouldin’s adjective. These conclusions do not even have to be spelled out: sometimes, dog-whistling suffices. How these conclusions are constructed deserves due diligence, more so when they become accusations.

    You know where I stand on this last point, AMac. We already had this conversation almost one year ago:

    http://deepclimate.org/2010/07/19/open-thread-4/#comment-4945

    (6) Interestingly, a way to discuss these conclusions might be by analyzing at what Gavin said about what he referred to “the pathological dynamic”:

    http://www.collide-a-scape.com/2010/06/16/the-main-hindrance-to-dialogue-and-detente/#comment-7585

    (Casual readers should note that the whole thread is worth the read.)

    I’m not sure if the dynamic is pathological. But I bet there’s some magic to it. There sure are many tricks.

    So one can say that I’m interested in magic tricks.

    (7) Let me finish, AMac, by applauding your attitude, your tone, and most importantly to me, your style. You seem to take pride and pleasure in writing and it shows. If we had a fantasy draft, with some chance, I’d pick you to play on my team.

    That Jim Bouldin recognized something like this proves that this matters in affairs between humans.

  136. willard Says:

    This :

    > If we’re these collaborated with the usual suspect, we need to adjust our identifier “well-known members of the broader climate-science community”.

    should be read:

    > If we’re to name only those who collaborated with the usual suspect, we need to adjust our identifier “well-known members of the broader climate-science community”.

  137. J Bowers Says:

    Willard — ““Post-normal” is too ugly to be worth the effort.”

    When I started to think of PNS as “that which is done after (latin – post) the normal science has been done”, it was much easier to come to terms with.

    It’s really for engineers and insurance agents IMHO (think Three Mile Island). I suspect Ravetz likes to milk it, though.

  138. AMac Says:

    Jim Bouldin —

    > We are not in lockstep in what we do at RC, even if we are in general philosophical agreement about what we want to accomplish and the limits of what we will tolerate in the comments.

    It sounds like blogging there is a good experience for you. My own views of the site have been colored by the failure of perhaps two-thirds of my +/- dozen comment submissions to pass moderation. Obviously, such a policy drives potential commenters like me away. RC owns the RC printing press, so that’s the way it should be. I don’t really have more to say on the subject.

    willard —

    I’m glad I covered most of your questions (I wasn’t sure!)

    >> No member of the broader climate-science community have taken a different position.

    I apologize for the lack of clarity. “A different position” was supposed to mean something like “offering criticism of Mann08″, not “the absence of support for Mann08″. Thanks for the link to Andy Russell’s comment thread; I had lost that somehow. Deep Climate (not a climate scientist but an A-list blogger) had made a similar remark at one point.

    Gotta run, response on your (4), (5), and (6) later. Thanks for the kind words in (7)!

  139. MIkeN Says:

    PeteB, I didn’t say that Mann’s response to McIntylre represented fraud, but the totality of behavior with regards to Tiljander, meets the definition of fraud presented by Arthur Smith. A definition that I consider to be incorrect as well.

    The paragraph you quote from Martin Vermeer is wrong. I think it’s a typo, or perhaps he is using different terminology, because we were in agreement on the behavior of the software. Namely, the CPS algorithm will not flip the proxy and the EIV will flip the proxy.

    So in the case of CPS, Mann simply used the data upside-down, so the McIntyre criticism of upside-down axes is valid without going into change of correlations or contamination, etc.

  140. AMac Says:

    At 01:04 supra, I wrote, “RC owns the RC printing press, so that’s the way it should be.” That is unclear. I meant that RealClimate.org‘s owners set their site’s polices as they see fit; bloggers should be free to do so.

    – – – – – – – – – –

    willard @ 18:30 —

    (4) Postmodern Science is a factor I’ve overestimated

    One of the things I’ve realized from this thread is that many scientifically-literate blog commenters don’t understand the basic concepts behind proxy-based reconstructions at a level that permits informed discussion of technical issues. In a better world, individuals would know their limits in regard to such matters. But in that better world, I would have realized this point a long time ago.

    In retrospect, this has likely been a significant contributor to the Alice-in-Wonderland aspect of many of the blog-conversations that I have engaged in concerning Tiljander (at Lucia’s Blackboard, Collide-a-scape, Climate Audit, other places). If one looks at the analytical steps in reconstruction-building as a very complex black box, it’s reasonable to orient oneself in the debate on the basis of who one trusts. For skeptics, that might be Steve McIntyre and allies. For pro-Consensus advocates, it could be Gavin Schmidt.

    Of course, if one has a clearer view of what goes into a proxy-based reconstruction, it’s more interesting to look at the substance being discussed, rather than at who is claiming what. 2+2=4, irrespective of whether it’s McIntyre or Schmidt making the assertion.

    So, I think that some parties to this dispute have been pounding the table, drawing conversations away from “the facts” and “the law” (as it were). That’s not behavior aligned with scientific inquiry, though it’s commonplace in politics. Such conduct probably fall somewhere on a continuum between Postmodern and Cynical.

    However, most of the time, I think folks just don’t know any better. That’s not postmodernism. Rather, it’s passion mixed with inadequate education.

    (5) I have my doubts about the utility of the dog-whistling narrative. Very often, such concerns can be flipped around to cut the other way. And a goodly amount of accurate mind-reading is required. I find that talent to be in shorter supply than most people suppose.

    As a companion to the Deep Climate comment where you express your concerns about the morally appalling nature of “insisting” upon “publicly condemn[ing] Mann”, I’ll excerpt a comment from the Collide-a-scape thread you also reference. A year later, it’s still topical.

    – – – – – begin quote – – – – –

    40. amac78 Says:
    June 17th, 2010 at 11:02 am

    As the subject of Gavin’s comment that was the impetus for this post, I will respond to his comment upthread that begins

    “As expected, the pathological dynamic worked it’s magic again…”

    1. In my opinion, Gavin’s comment makes many claims that merit a technical rebuttal…

    2. One approach to a complex and contentious problem is to look for component issues that might be resolvable. Ideally, such a component can be expressed as a plainly-worded and easy-to-understand question that (if answerable) has the promise of simplifying the overall problem. Sometimes this is not possible. Instead, a series of If/Then premises have to be considered in sequence, or simultaneously.

    While this [latter] approach is [sometimes required by] individuals tackling advanced suduko, it isn’t preferred. A better way to solve a hard suduko is to begin by identifying the easier parts of the puzzle, and filling them in. Keeping things holistic and complex is a strategy that is particularly ill-suited to occasions where participants bring different skill sets to the table. And where people have strong preferences as to the outcome. And where trust is low and emotions run high. All these caveats apply to Mann08.

    3. I’ve distilled my concerns about the use of the Tiljander proxies in Mann08 to one question.

    “Are the Tiljander proxies calibratable to the instrumental temperature record, 1850-1995?”

    To date, to my knowledge, this question has never been directly answered by an author of Mann08, or by a working climate scientist, or by a prominent AGW Consensus advocate.

    [snip]

    – – – – – end quote – – – – –

    > (6) Interestingly, a way to discuss these conclusions might be by analyzing at what Gavin said about what he referred to “the pathological dynamic”

    You link the Gavin Schmidt comment from earlier in that C-a-s thread that begins,

    “As expected, the pathological dynamic worked it’s magic again. I say pathological advisedly because in both of the responses Amac and Willis both simply regurgitated points they’ve made before without bothering to read what I said, and without clicking on the link (where, if they had cared to look, a reconstruction without both tree rings *and* the Tiljander proxies is shown – and yes, it looks similar to the others).”

    As best I can tell, you find his remark to be insightful. That isn’t a view I can share. For instance, ad hominems aside, I note three to five misstatements of fact (depending on how one counts).

    That style of discourse is, I think, a hindrance to dialogue.

  141. Dan Olner Says:

    Amac:

    “1. What clothes is the emperor wearing, if any?

    2. What are the implications, if any, of the emperor’s wardrobe? For the royal court, for society as a whole, etc.

    If the best we can do with #1 is akin to “I don’t know, and it doesn’t matter,” it becomes very hard to think clearly about #2.”

    Ah – OK. So this is where we differ. I see climate science as a whole as a large redundant system: I don’t think Tiljander/Mann08 can possibly have any important impact on the main findings of climate science. From the answers you’ve given on co2 cuts, I would normally conclude that you agreed: otherwise, you appear to be saying co2 should be cut, even though you have no trust in the science that tells us about the effects of co2. Which wouldn’t make any sense.

    So I’m still confused about exactly what you think Tiljander/Mann08 implies for climate science as a whole.

    In your post to Willard, you make a very good point about trust. It’s true. You also say:

    “many scientifically-literate blog commenters don’t understand the basic concepts behind proxy-based reconstructions at a level that permits informed discussion of technical issues. In a better world, individuals would know their limits in regard to such matters.”

    For myself, I’m saying I don’t know enough about it to have a conclusion – but that it doesn’t matter. That is, it doesn’t matter, because – as you seem to agree – our job at this point is to stop putting co2 into the atmosphere, we have quite enough independent scientific evidence telling us that’s what needs to happen, and spending our lives arguing over some tree ring data is going to make doing that harder. Well: not arguing over tree-ring data, but doing what you’re doing here – claiming that if we *don’* fix whatever problems exist with Tiljander/Mann, we can’t go on to defend climate science (your points 1 and 2 above, as I read them.) That just isn’t true.

    Perhaps you could answer your own question 2: “2. What are the implications, if any, of the emperor’s wardrobe? For the royal court, for society as a whole, etc.”

  142. PeteB Says:

    MikeN,

    OK, I’m not sure what comment we are talking about, but my understanding is (and I think this is the same as Martin’s understanding and I thought it was the same as Amac’s and your’s)

    1) the original Tiljander03 she (see I have learnt something :-) ) assigned these temperature correlations to the instrumental period:

    Darksum – higher signals correlate to warmer temperatures;
    Lightsum – higher signals correlate to cooler temperatures;
    X-Ray Density – higher signals correlate to cooler temperatures.

    2) There has been recent (over the instrumental period) contamination by non climatic signals

    Darksum – contamination increased signals over that time;
    Lightsum – contamination increased signals over that time;
    X-Ray Density – contamination increased signals over that time

    3) The Mann08 calibration was done to local temperatures over the instrumental period. This will have affected all the signals. The calibration found a positive correlation with all 3 signals

    Darksum – higher signals correlate to warmer temperatures
    Lightsum – higher signals correlate to warmer temperatures;
    X-Ray Density – higher signals correlate to warmer temperatures.

    So obviously all these correlations are used for the reconstruction period for both CPS or EIV methods (and two correlations are likely to be the wrong sign and the other one the calibration factor is likely to be invalid because of contamination during the calibration period)

    OK the we come onto the two methods

    The difference, as I understand it, between EIV and CPS is that CPS has a screening step, which if the correlation during the calibration period was negative would have thrown it out because all the Tiljander proxies were categorised to have a positive relationship (I guess if this had happened, it would have caused them to look at it again.)

    There is no screening step for EIV, so if there had been a negative correlation during the calibration period this negative correlation would have been used for the reconstruction period

    This is all a bit hypothetical anyway because the correlation during the calibration period was positive.

    I think SteveM’s comment would have been much more informative if he had said that contamination during the calibration period invalidated all the time series (and actually reversed the correlation on some of them)

  143. PeteB Says:

    Oh bother,

    point 1) should have said

    the original Tiljander03 she assigned these temperature correlations prior to the instrumental period:

  144. AMac Says:

    Dan Olner —

    In my view, there isn’t this thing called “science” that is right, and that tells us what the climate history was, what its present state is, and what its future likely holds.

    If there was, the relevant task for us literate bystanders would be to pay attention to professional climatologists and absorb their lessons. We could then move straight to implementing policy.

    This whole Tiljander kerfluffle would be a terrible distraction, and nothing more.

    “Science” is a process that was invented during the Renaissance. It’s a way of looking at the world and learning about the world. As progress has accelerated from the mid-20th century to the present, its procedures have gotten more complex and more black-box-like — but they aren’t really black boxes.

    An analogy (and I try to be cautious with them). Perhaps you cannot trace the steps that are involved in your bank’s computer producing your monthly account statement. I can’t — not being a banker or a computer science grad, the process seems black-box-like to me.

    But if my starting balance was $1,000.00 and the sole transactions are a deposit of $100.00 and a service fee of $1.00, I can tell you what my ending balance should be. $1,099.00. If that balance is anything else, the bank manager should be able to explain why. The rules of arithmetic weren’t modified by the movement of banks from adding machines to cloud computing.

    Arithmetic isn’t a mystical or arcane procedure. It can be understood by a numerate layperson.

    A longer treatment of this analogy here.

    Consider PeteB’s comment, just above this one.

    His points 1), 2), and 3) reflect a scientifically-literate layperson’s correct understanding of the major problem with Mann08’s use of the Tiljander data series. Contamination during the calibration period.

    (There are additional twists to the tale, but that’s fine. And, unable to read MatLab or R code, I can’t vouch for MikeN’s dissection of EIV or CPS procedures. That’s fine, too.)

    So a key question has come up. This question was first raised by a Climate Audit commenter within a couple of days of Mann08’s online publication. From that time to this, the existence of contamination and the effects of the inclusion of contaminated proxies on reconstructions have been extensively discussed.

    The issue was formally raised in the peer-reviewed literature (M&M Comment) and formally responded to by Mann08’s authors (the “bizarre” Reply).

    The authors, their colleagues, and scientifically-literate advocate-defenders have extensively addressed the issue, as well. E.g., see willard’s link to the C-a-s post The Main Hindrance to Dialog (and Detente).

    In my opinion, the authors’ and the authors’ defenders’ rebuttals are almost entirely without merit.

    This simple question stands, unaddressed:

    Can the three Tiljander data series be calibrated to the instrumental temperature record, 1850-on?

    That brings us full circle.

    Scientists have careers and personalities and biases and blind spots and so forth. Scientists make mistakes. Because they are people.

    The scientific process isn’t designed to work perfectly, once scientists are perfected. Rather, one of its virtues is that its people and institutions interact to correct the significant errors that develop. Notwithstanding the flawed natures of those people and institutions.

    Nearly three years on, this hasn’t worked.

    The dilution of my faith in “climate science” is proportional to the fierceness and speciousness of the defenses of Mann08.

    I’ve made this point a number of times in a number of ways, and I understand that you (Dan Olner) and most other readers see things quite differently. As far as understanding my position, I think this explanation will have to suffice.

  145. Dan Olner Says:

    Amac: Sounds like you don’t like analogies much, but I’m going to carry on with yours…

    You find a dollar discrepancy in your bank balance. The bank refuses to acknowledge the mistake. You then work tirelessly to demonstrate not only that the bank is mistaken but that the entire banking industry cannot add up, and that their basic mathematical understanding means none of our money is safe. Bank runs follow, entire economy collapses.

    In reality, what you appear to be arguing – and it’s this I can’t get: you work tirelessly to demonstrate not only that the bank is mistaken but that the entire banking industry cannot add up, but you then happily carry on banking at the same place, and seem surprised when others won’t.

  146. AMac Says:

    Dan Olner —

    Thanks for your response. Each of our respective positions remains something of a puzzle to the other. But I think the chasm is a little narrower than it was at the beginning of the thread.

  147. PeteB Says:

    Amac,

    The scientific process isn’t designed to work perfectly, once scientists are perfected. Rather, one of its virtues is that its people and institutions interact to correct the significant errors that develop. Notwithstanding the flawed natures of those people and institutions.

    But that seems exactly what is happening to me, from the Frank et al paper referenced above surely :

    ….Expert assessment to evaluate the signal of a particular record from a particular proxy archive (e.g., the low frequency skill of a new speleothem record) will be invaluable in trying to minimize ‘wrong figures’ being put into a large-scale reconstruction. It seems advisable at this point to use fewer, but expert assessed proxy records, rather than hundreds of proxy series, and hope that reconstruction algorithms will overcome the often huge noise components typical for many of the available time series….

    addresses this very issue – if this approach is adopted, then it would avoid this problem

    and Jim

    Amac is addressing, using the specific example of sedimentation, the more general issues raised by Loehle (2009) using tree rings–the implications for climate reconstructions, of biological or physical processes that change in rate or (even worse) in direction.

    This is a very important issue, it needs to be addressed, and he is raising it without ad hominem attacks or larger accusations. A number of people are working on it, including myself. And without going to details, since it’s unpublished work in progress, with some good success.

    seems to be researching the general issue, which Tiljander is a specific example ( I’ll be very interested to see the results of this when they are ready )

    I guess people will continue to use the Mann08 reconstruction until someone comes along with something better

  148. MIkeN Says:

    PeteB, that is roughly correct, except for maybe 2 minor things. CPS, the screening will accept negative correlations for some proxies, but TIljander is not qualified as such. This is what is meant by the third sentence of Mann’s reply, which I find confusing. Screening, when used, employed one-sided tests only when a definite sign could be a priori reasoned on physical grounds.

    Tiljander does meet this criteria, and his software does the same, so this sentence should have been left out in a response to was Tiljander used upside-down.

    The other possible error is I think in Mann 08 the EIV does not use local temperatures, while CPS uses a gridcell.

    So the error here is that some of the Tiljander proxies were fed into screening with the axes upside-down, as the software assumes higher numbers means warmer. Mann should have done an invert on the appropriate proxy, and used it as such.

    Martin Vermeer gives Mann a pass because in general regression algorithms are blind to the sign of the predictor, and the screening should count separately. I don’t think that is good enough. The accusation was that axes are used upside-down specifically in the case of Tiljander.

    So Mann could have just said
    The argument is bizarre. Tiljander was not oriented upside-down.
    Implication McIntyre is an idiot or lying. Except this is not the case.

    The argument is bizarre. Regression algorithms are blind to the sign of the predictor.
    except this is not the case for CPS.

    The argument is bizarre. Regression algorithms are blind to the sign of the predictor. The only exception is when we know which way is up.
    The implication is that Tiljander is not one of the exceptions, only Tiljander is one of the exceptions.

  149. MIkeN Says:

    Dan Olner, I think you have to extend the analogy further to get to AMac’s point. If a prominent banking industry website is asked whether bank X operates correctly, and the reply is unresponsive, If other banks say it looks OK, and it doesn’t matter, etc.

    Now you can say, well the banks are still going to operate properly regardless of what this bank is doing.

  150. J Bowers Says:

    MikeN — “The other possible error is I think in Mann 08 the EIV does not use local temperatures, while CPS uses a gridcell.”

    Of any use? (read responses)

  151. Jim Bouldin Says:

    Amac,

    Simply stated, if you’re concerned about the importance of the misuse of the Tiljander proxies themselves, your concerns are far out of proportion to the overall significance of same. and you are badly missing the forest for the trees.

    If your concern is the more general issue of how well calibration relationships hold up throughout reconstruction periods, generally, then you are asking an important question that needs to be thoroughly examined. This is the essence of the Loehle paper I cited.

  152. Dan Olner Says:

    Amac, I’d be interested to know what you think of this review of Ray Bradley’s new book.

  153. MIkeN Says:

    The review strikes me as a proper review of the book. I haven’t read it, but I would probably write it the same way.

  154. kuhnkat Says:

    PeteB says,

    “Are you sure you are not falling for the ‘If we don’t know everything, we know nothing’ fallacy ?”

    Are you sure you are not falling for the fallacy that if you know something that it is automatically applicable to the issue at hand or even useful at this point??

  155. dhogaza Says:

    Are you sure you are not falling for the fallacy that if you know something that it is automatically applicable to the issue at hand or even useful at this point

    Not relevant, doesn’t apply to the discussion in which PeteB’s point was made.

    And, regardless, there are many known facts in physics (for instance) that *are* automatically applicable to every situation, so your noise-post is meaningless unless you make a specific claim.

    As you state your supposed fallacy, you simply appear ignorant …

  156. willard Says:

    Back to the grid, and having had the time to read back the thread, I will take the opportubuty to preface forthcoming comments with this quote:

    > The art that mainstream scientists and their defenders must learn is to take the valid parts of the criticisms and deal with/respond to them, and leave the c/a/e [conspiratorial/accusative/exaggerated] packaging for what it is.

    http://ourchangingclimate.wordpress.com/2011/06/22/climate-science-scientific-method-skeptics-not/#comment-13399

    In my humble, this wins the Internet of the thread.

    Until later,

    w

  157. willard Says:

    Reading back the many resources related to L’Affaire Tiljander (LAT), I found this tidbit that add one name to the set of people belonging to what AMac refers to the “broader AGW Consensus community”:

    > That’s not just Nick Stokes [...] and Gavin Schmidt. It’s many others as well (including Bob Grumbine, who commented earlier on this thread).

    http://initforthegold.blogspot.com/2010/02/grim-amusement.html?showComment=1266086825237#c6443642510592774951

    So now we have Robert Grumbine, viz. here:

    http://moregrumbinescience.blogspot.com/2009/10/sound-and-fury-at-wuwt.html

    I’ll return to the question of naming names in another comment. Naming names bears relevance to the identity of the bankers, in AMac’s analogy (endorsed by MikeN) above. The impatient reader can start there:

    http://amac1.blogspot.com/2009/11/blog-links-mann-08-and-korttajarvi.html

    ***

    For now, I’ll concentrate on the two sentences that follow in that comment by AMac. The next sentence of the comment I underlined is interesting:

    > _Nobody_ is dissenting from the “party line.”

    The emphasis on “nobody” is not mine. We do not yet know from what line from which nobody is dissenting. We do not know if that implicates all those (like me) that would prefer to refrain from assenting anything related to old Finnish lake mud.

    The next sentence is also interesting:

    > I’ve been surprised at the number of Lukewarmer commenters whose sentiments largely mirror my own.

    The two collective identifiers are “Lukewarmer commenters” and “broader AGW Consensus community”. A bunch of unorganized individuals versus a community as a whole, cohesive to the point of sometimes being portrayed as indulging in a Cargo cult. For sake of symmetry, we might think of replacing “Lukewarmers commenters” with “broader-anti-AGW-community” or “anti-broader-AGW-community”.

    This might not be too farfetched, when taking into account this sentence of AMac’s comment, which preceded the first one I quoted:

    > It’s the reactions of the broader AGW Consensus community that really matter. Those who defend the indefensible–because of whatever mix of ignorance, team loyalty, hubris, and confirmation bias–forfeit the presumption of trust on related issues.

    Speaking of party lines, indefendable defenses, and other sociological traits might very well forfeit the presumption of neutrality on related issues.

  158. danolner Says:

    “It’s the reactions of the broader AGW Consensus community that really matter. Those who defend the indefensible–because of whatever mix of ignorance, team loyalty, hubris, and confirmation bias–forfeit the presumption of trust on related issues.”

    Question: in what ways has the reaction of the ‘broader AGW consensus community’ differed for Lindzen, Christy, Spencer? Reaction of the blogosphere may have been different, but as far as I can tell, it would be the same for them as for all these Mann/Tiljander goings-on – which is, mostly silent in public, and playing it out in the peer-reviewed literature.

    Related: it seems mighty peculiar to me that anyone claiming they think we risk destabilising the climate with our co2 exhaustinz should spend their time exclusively on proxy reconstructions. The most common line of attack for the ‘peer-reviewed skeptics’ is equally as theory-laden, but several orders of magnitude more important. Impacts come down to sensitivity. Amac, why don’t you spend any time digging into this issue? If you want a technical challenge, there’s very little to beat it. And, as I understand it, you certainly won’t be short of interesting errors to pick apart in some of the common work claiming sensitivity is low / low enough for no-one to get phased.

  159. MIkeN Says:

    Dan Olner, again that speaks to credibility. If errors are pointed out and the relevant people are not responding to the errors, that says something. I have recently pointed out on one site that I find Fred Singer to be non-credible for reasons like this. What you are arguing actually reinforces the idea that climate science should not be trusted.

  160. danolner Says:

    “I have recently pointed out on one site that I find Fred Singer to be non-credible for reasons like this. What you are arguing actually reinforces the idea that climate science should not be trusted.”

    One of my points was, it doesn’t matter what you or me are doing. We’re arguing in the blogosphere – here, there really is a large difference to how proponents of different theories get treated. I was wondering whether – to take a specific example – Lindzen has been treated any differently in the peer-reviewed literature to Mann/Tiljander. I don’t have any thorough answer to that question. My only point was: if you’re arguing that Mann etc have received specially lenient treatment of some sort, you could do with demonstrating that e.g. Lindzen has been treated differently. There are plenty of blog posts about about Lindzen’s errors, of course, but I’m not aware that ‘the broader climate science community’ has done anything but the usual – i.e. work through the peer-review process.

    And I’m still confused as to why dodgy calculations about climate sensitivity – which appear in the literature – do not seem to attract the attention of people who are otherwise fascinated by the nitty-gritty detail of other corners of climate science. What’s the difference? In terms of impact on whether climate change actually matters, getting climate sensitivity as accurate as possible is waaaay ahead of tussling over a subsection of temp proxies.

  161. Bart Says:

    Excellent points Dan.

    Thanks Willard for bringing my own words to my attention; I though I recognized them and when I checked it suddenly all made sense. Funny that when re-reading your own stuff it suddenly falls in place.

    Apologies for my silence here as of late, and thanks for the good discussion.

    During summer my posting will be light (but hopefully not entirely absent). The weather is just too nice.

  162. MIkeN Says:

    >And I’m still confused as to why dodgy calculations about climate sensitivity – which appear in the literature – do not seem to attract the attention of people who are otherwise fascinated by the nitty-gritty detail of other corners of climate science. What’s the difference?

    Too much of an assumption there. I do follow some of this.
    Easier answer is that if the dodgy calculations involved upside-down data, it’s easier to follow. There’s a reason Tiljander errors are talked about more than the other Mann errors, including in that same paper. As a whole, the paper convinces me of the existence of a Medieval Warm Period, far more than any thing else.

  163. willard Says:

    Unless one is an impatient reader, one might not has started to read

    http://amac1.blogspot.com/2009/11/blog-links-mann-08-and-korttajarvi.html

    to find out who belongs to the “broader AGW Consensus community” or the “broader pro-AGW-consensus community”. Perhaps a way to distinguish the former from the latter is that the latter’s darling has been identified above as realclimate.org. But for the sake of our discussion, we won’t use this distinction.

    Our known candidates know that Gavin Schmidt, Nick Stokes and Bob Grumbine. Three names are not enough. So let’s take a look at the compedium of blog links related to LAT:

    First, we see that the posts, instead of being listed chronologically, are organized by order of some kind of importance:

    – CA
    – RC (the darling of the pro-AGW consensus community)
    – Stoat
    – the Air Vent
    – Roger Pielke Jr
    – Cruel Mistress
    – More Grumbine Science
    – Watts Up With That?
    – Climate Progress
    – Delayed Oscillator
    – AGW Observer
    – Climate Observations
    – Climate Skeptic
    – Collide-a-scape
    – Not Spaghetti
    – AGW Observer (named twice)

    Second, if we replace the name of the blogs with the name of the blog curators, we get the cast of characters:

    – Steve McIntyre
    – Gavin Schmidt and the Kyoto Flames
    – William Connoley
    – Jeff Id
    – Roger Pielke Jr
    – Ben Hale
    – Bob Grumbine
    – Anthony Watts
    – Joe Romm
    – Ari Jokimäki
    – Bob Tisdale
    – Warren Meyer
    – Keith Kloor
    – Arthur Smith

    This impressive cast of characters would be incomplete if we omit Bart Verheggen, curator of this very blog. It might prove interesting to add the names of the main commenters (MikeN, Martin Vermeer, etc), but we’ll do that later. Many more people comment, so many that we might need a more formal Social-Network analysis. This will have to wait.

    Third, we notice that each blog and each curator are being assigned a type:

    – skeptical
    – pro-AGW consensus
    – skeptical/pro-AGW consensus

    Fourth, readers will notice that there are two curators that have the two types: Roger Pielke Jr and Keith Kloor. That means one can be pro-AGW consensus and skeptical, at the same time or in alternation. We’ll assume that the former is both at the same time and the former is both in alternation.

    Fifth, we note we could build symmetrical types. To do that, we’ll have to find an inverse of “pro”. I suggest contra, and not “anti” (the usual prefix), as I suggested previously. It fits perfectly with “contrarian” (Steve’s choice) and “dissenter” (Carrick’s favorite). So we would have:

    – pro-AGW consensus
    – contra-AGW consensus
    – neutral

    Sixth, let’s note that Bart is using the epithet “mainstream scientists and their defenders”. We could use it as a synonym to AGW consensus.

    With this symmetrical typology at hand, it might prove easier to identify who’s who. But that is insufficient. We have not associated any claims to these types.

    This needs to be done, if we want to analyze further.

  164. danolner Says:

    MikeN: “As a whole, the paper convinces me of the existence of a Medieval Warm Period, far more than any thing else.”

    Could you give some details on your views about the MWP?

  165. AMac Says:

    willard (July 10, 2011 at 19:19)

    You quoted our host upthread: “The art that mainstream scientists and their defenders must learn is to take the valid parts of the criticisms and deal with/respond to them, and leave the c/a/e [conspiratorial/accusative/exaggerated] packaging for what it is.”

    I agree with your praise of that comment.

    .

    willard (July 11, 2011 at 02:13)

    > Speaking of party lines, indefendable defenses, and other sociological traits might very well forfeit the presumption of neutrality on related issues.

    A few general remarks on this and your following comment (July 12, 2011 at 03:39).

    You seem to have little interest in the technical question of whether the paleotemperature reconstructions in the highly-cited Mann08 PNAS article represent strong science or weak science. That’s fine — it’s a wide world.

    The only reason I have engaged in this thread is because I believe it is relevant to the topic of the post. Earlier:

    Bart, Chris Colose, and others have made strong claims about “how [climate] science works.” In my opinion, it is futile to tackle such a broad theoretical subject without grounding remarks in specifics. Since the theme is “science”, those details are necessarily technical in nature. The story of the use of the Lake Korttajarvi data series illustrates how one branch of climatology has diverged from “best practices” as they are known in the other physical sciences.

    Also see your quote of Bart, above.

    Your seem to have alighted on a combined textual analysis and Wegman (Said?) style social network analysis. On finding that you’ve put my writing at the center of this venture, I first felt uneasy. The subtext of whether I am a Good Enough Person to qualify as a blog commenter is somewhat creepy.

    As well as irrelevant to the calibratability of the Tiljander data series to the instrumental temperature record.

    As well as a commonplace in this brave new world of blog discussions of hot-button issues.

    On the other hand, I started my Tiljander-focused blog to provide an easily accessed archive of links, references, data, and arguments. The point was to try and pull the discussion up from the he-said-she-said was-too-was-not template that is all too common in teh intrawebs.

    So, have at it.

    A few reminders and caveats.

    * This is a hobby, or a pebble in my shoe. Believe it or not as you will, but nobody subsidizes, supports, or coaches. I have a life. A comprehensive blog-trawl or SNA might mistake oversight for evil intent. To pick one example.

    * Blog commentary is informal writing. I try to “play the man, not the ball” during cocktail-party conversation, but sometimes my efforts fall short. So, too, with comments. Look around, and you might find others who fail your standards, as well.

    * In the course of this conversation, I have belatedly come to believe that many scientifically-literate people who use technical vocabulary to describe proxy-based reconstructions nevertheless view the process as a black box. This explains much of the banal and illogical quality of the discourse on Tiljander in Mann08. The insight makes the process of engagement less frustrating to me. If I’d arrived at this place sooner, I’d have strayed less into play-the-man territory, I think. I particularly wish that I hadn’t used the term “full-throated” to describe Bob Grumbine’s initial commentary on Tiljander in Mann08. It offended him. Dr Grumbine is aware of this regret.

    * I’m unlikely to comment further on your SNA (or on personal critiques, if your interest trends that way).

    * My goal with the two References pages on my blog was to provide a reasonably-comprehensive list of resources on Tiljander in Mann08. But curation is unending, and I am usually behind. As time (I’m out of it for today) and motivation permits, I will update the those pages. And I’ll thank you for highlighting errors, so that I can correct them. When I get a Round Tuit, that is.

  166. AMac Says:

    Dan Olner (July 11 at 14:04) —

    > Amac, why don’t you spend any time digging into [shortcomings in the works of Lindzen, Christy, and Spencer]? If you want a technical challenge, there’s very little to beat it.

    Thank you for flagging subjects that you think will interest me.

    Upthread, I’ve offered reasons for why I have written on the Tiljander in Mann08 story, and have little more to add.

  167. AMac Says:

    In my comment of 14:36 supra, I wrote,

    “I try to ‘play the man, not the ball'”.

    A slip. Heh. Should have been, “I try to “play the ball, not the man‘”.

  168. Bart Says:

    No worries, Freudian slips happen all the time ;-)

  169. MIkeN Says:

    >Could you give some details on your views about the MWP?

    I would say that there was a time period around 1000 which was warmer than now. I’d never heard of it until after I read the critiques of the hockey stick. It was in an early IPCC report that is now considered to be a poorly sourced graph. I had no idea whether this time period was warmer or cooler than present and was ambivalent about it. However the Mann 08 paper suggests that this time period was warmer. The algorithm mines for hockey sticks, filtering out proxies that do not show modern warming, while the past wiggles would balance out against each other, almost guaranteeing a hockey stick if you feed in random data. Jeff Id has a number of posts on the subject at The Air Vent. Despite this algorithm that enhances and produces hockey sticks to order, that you need to use a proxy upside-down, turning a medieval warm period into a medieval cold period, to get a hockey stick shape, tells me the actual non-random climate proxies are signalling against a hockey stick.

  170. J Bowers Says:

    MikeN — “I would say that there was a time period around 1000 which was warmer than now.”

    What’s your error margin on that?

    MikeN — “The algorithm mines for hockey sticks, filtering out proxies that do not show modern warming”

    Can you point to the specific piece of work you are referencing in order to make that statement?

  171. neverendingaudit Says:

    AMac,

    Thanks for your general remarks.

    In my opinion, Bart’s quote wins the Internet because it does two things.

    First, it provides a fundamental tasks:

    – learning to respond and deal with the valid parts of the criticisms;
    – accepting the conspiratorial/accusative/exaggerated packaging for what it is.

    Second, it underline that these two tasks belong to the fine arts, be they liberal, conservative, or libertarian.

    These two tasks presume we can recognize what is this “conspirational/accusative/exagerated packaging.” My own interests lie specifically in the accusative packaging, i.e. the judiciary mode. For me, this is a technical question.

    ***

    I agree with you when you say that

    > [I]t is futile to tackle such a broad theoretical subject [“how [climate] science works”] without grounding remarks in specifics. Since the theme is “science”, those details are necessarily technical in nature.

    You would never guess how many thinkers agree with this, including post-modern ones. In fact, this is what distinguishes traditional epistemology from current philosophy of science. Nothing beats a good stock of examples.

    That’s why indefinite terms like “pro-AGW consensus community” do not suffice to build any conclusion, however technical the analysis upon which it is built. That’s why appeals to Cargo cults, Feynmanian moralism, falsifiability, post-normal science, Newspeak or worse do not suffice either. These two speech patterns are being problematic. They do deserve due diligence and further clarification.

    I agree with you that the calibratability of the Tiljander data series to the instrumental temperature record is a scientific question. Many members of the pro-AGW consensus have declared that Finnish lake mud is no good proxy. For what it’s worth, my personal impression is that there seems valid criticism being adjudicated in favor of the contra-AGW consensus community.

    I can also concede that many members of the pro-AGW consensus community did respond “suboptimally,” to use an auditing term. This might indirectly acknowledged by Bart, if I read correctly the first part of his quote. The importance of the second part of this quote might be illustrated by this example:

    On the 2010-07-22, at 12:37 (PST?), in one of Arthur Smith’s thread here are three answers by Martin Vermeer , :

    AMac, the only question that has some scientific relevance is (3). The rest is lawyerly stuff for which I have no patience. But to answer (3), I must answer part of (1), so here goes.

    (1) I believe that the Tiljander orientations are likely correct, and their caution on the contamination in later years almost certainly correct.

    (3) I do not believe this is realistically possible. Any effect from the mishandling of Tiljander drowns in the other, substantial uncertainties both statistical and structural, especially going back in time before 1000AD. These uncertainties are well described in the paper. I would have no hesitation to use these reconstructions in my own work.

    Ad (2):

    Should Mann and coauthors have figured out the problems with calibration of the Tiljander proxies to the instrumental record from the discussion in the blogosphere,

    Thanks for making me laugh ;-)

    Source: http://arthur.shumwaysmith.com/life/content/michael_manns_errors

    I believe that my own perspective is quite relevant to arbitrate any conclusion regarding (2). This perspective in no way necessitates an opinion on the calibratability of Finnish lake mud. While, I understand your feelings of uneasiness, creepiness, and subtextual worthiness, that you need to budget your time, that these are informal conversations based on imperfect and unsubsidized hobby, and that you might not respond to further analysis from my part, my own perspective can only work with have I read. And there are enough to read as it is. So when you’re done with your Round Tuit, please share it with me.

  172. AMac Says:

    willard (03:10)

    “Pro-AGW-Consensus” was my awkward neologism, an attempt to find a neutral description. I guess you can’t please everyone.

    Reading you, it strikes me that “Mainstream” could be a better term. Paired with “Minority view,” perhaps. What do you think?

    As far as Martin Vermeer’s comment, here’s the context. My point #2 was:

    (2) Should Mann and coauthors have figured out the problems with calibration of the Tiljander proxies to the instrumental record from the discussion in the blogosphere, then from the M&M Comment in PNAS, and then from the correction to the Kaufman Science 2009 manuscript? Is the Mann et al. Reply to M&M in PNAS responsive to the criticisms of the calibration issues that were raised? In particular, does the Reply’s discussion of “upside down” constitute a failure to deal with the calibration issues?

    Re: #2, Vermeer’s response in its entirety was:

    Ad (2):

    Should Mann and coauthors have figured out the problems with calibration of the Tiljander proxies to the instrumental record from the discussion in the blogosphere,

    Thanks for making me laugh ;-)

    So we can break down #2 as follows:

    2a) Should Mann and coauthors have figured out the problems with calibration of the Tiljander proxies to the instrumental record from the discussion in (2a.1) the blogosphere, then (2a.2) from the M&M Comment in PNAS, and then (2a.3) from the correction to the Kaufman Science 2009 manuscript? (This is relevant because paleoreconstruction expert RS Bradley is an author on both papers.)

    2b) Is the Mann et al. Reply to M&M in PNAS responsive to the criticisms of the calibration issues that were raised [in the M&M Comment]?

    2c) In particular, does the Reply’s discussion of “upside down” constitute a failure to deal with the calibration issues [raised in the M&M Comment]?

    As you can see, Martin Vermeer restricted his response to the (2a.1) portion of #2. It’s always gratifying to make people laugh, but I confess that my humor wasn’t clear to me at the time. Still isn’t, for that matter.

  173. MIkeN Says:

    >>MikeN — “The algorithm mines for hockey sticks, filtering out proxies that do not show modern warming”

    >Can you point to the specific piece of work you are referencing in order to make that statement?

    I refer you to The Air Vent, where Jeff Id has many posts on the subject. Also, the code for Mann 08, like gridproxy.m.

  174. PeteB Says:

    Amac,

    It seems clear to me that Martin was laughing at the idea that authors should be expected to keep with discussion in ‘the blogosphere’

    I think the blogosphere is useful for bouncing around ideas and sharing information, and hopefully this discussion with the link from rc and Jim’s comment and responses at rc and here help..

    But ultimately it is what is in the peer reviewed literature which really counts.

    This sounds like what Jim is doing, because it sounds like a specific example of the problem where proxies that change in rate or (even worse) in direction in response to temperature. This sounds very exciting to me, because if Jim comes up with some technique to compensate for this, we will have better quality reconstructions with more confidence.

  175. Bart Says:

    J Bowers, all,

    On open thread nr 2 we had a constructive hockeystick discussion (my first; I had avoided the topic until then). Jeff Id helpfully explained his position with as few words as possible:

    Temperature proxies are millenia long series of suspected but unknown temperature sensitivity among other things (noise). Other things include moisture, soil condition, CO2, weather pattern changes, disease and unexpected local unpredictable events. Proxies are things like tree growth rates, sediment rates/types, boreholes, isotope measures etc.
    In an attempt to detect a temperature signal in noisy proxy data, today’s climatologists use math to choose data which most closely match the recently thermometer measured temperature (calibration range) 1850-present.
    The series are scaled and/or eliminated according to their best match to measured temperature which has an upslope. The result of this sorting is a preferential selection of noise in the calibration range that matches the upslope, whereas the pre-calibration time has both temperature and unsorted noise. Unsorted noise naturally cancels randomly (the flat handle), sorted noise (the blade) is additive and will average to the signal sorted for.
    The blade of the hockey stick is the sorted signal and noise in post 1850 times, the handle has a flattened shape due to the canceling of the noise. The difference between the blade and handle is referred to as variance loss.

    (left c/a/e that follwed out ;-)

    I replied, trying to paraphrase his position (that’s my way of trying to understand someone else’s position):

    To me it sounds entirely reasonable to weigh the proxies based on how well they reproduce the instrumental temperature record. You seem to assume a good correlation over this period is based on noise, i.e. coincidence? Or at least, that noise could have contributed to the good correlation, which is fair enough. (With sorting, you mean weighing, right? (giving it more weight, i.e. importance, in the final reconstruction) )
    You argue that inasfar as by coincidence the noise correlated with the measured temperature increase since 1850, that takes care of the upswing in the proxie reconstruction (the blade), whereas the pre-1850 proxies have random noise which causes the flat shaft.
    Is that good paraphrasing of your position?
    In that case the critical point is really, to what extent is a good correlation between measured temperature and proxies coincidence (ie not due to a causal relation between the proxie and temperature), and to what extent is it due to a real causal relationship? Inasfar as the latter dominated, there shouldn’t be a problem.
    This could -and I think has- been investigated by people studying the actual dynamics of the proxies involved, eg plant physiologists for tree proxies (Jim Bouldin frequenly mentions this at RC).
    It also shows that in the end, finding statistical relation still has to rest on physics (or chem or biology) in order to be properly interpreted. I take that as some sort of vindication.
    And quite interesting that you alude to the fact that small temperature variability in the past would be consistent with a low climate sensitivity and large past temperature variations would be consistent with high climate sensitivity. This is of course only the case if all else is equal, which it is not, but still. It’s a point I’ve also brought up before, as it is a funny mixing up of tribalist thinking in a way.

    Jeff replied with a long comment in which he gave some links to his hockeystick posts. The discussion went all over the place thereafter, but these few exchanges were very helpful to me in understanding an archetypical skeptical position on these issues. FWIW.

  176. AMac Says:

    PeteB (10:37) —

    > It seems clear to me that Martin was laughing at the idea that authors should be expected to keep with discussion in ‘the blogosphere’… But ultimately it is what is in the peer reviewed literature which really counts.

    Oh. That’s a funny (‘odd’ rather than ‘humorous’) position. Science’s most durable and authoritative record is the peer-reviewed literature, as you say. In both theory and practice, it’s an intellectual and social enterprise that includes periodic lab meetings, impromptu discussions over coffee, scheduled happy hours, departmental seminars, email exchanges, national and international meetings, grand rounds, graduate-level teaching, dissertation defences… and so on.

    I wouldn’t have expected that RC blogger Prof. Mann or blog commenter Prof. Vermeer would disqualify science-themed blogs from that list. In that regard, RC blogger Gavin Schmidt explained in the comments at Collide-a-scape that co-blogger Prof. Mann had revised Mann08 SI Fig. S8 in order to respond to blog criticisms of the shortcomings of that figure, as it appears at PNAS.

    A re-read of point #2 (AMac at 04:36), should show that the puzzle in question was the inadequacy of Mann08’s authors Reply to M&M’s Comment. Mann et al might have cleared up Tiljander. Instead, they entered those cryptic and incorrect paragraphs into the peer-reviewed literature. Launching these thousand ships.

  177. danolner Says:

    “it’s an intellectual and social enterprise that includes periodic lab meetings, impromptu discussions over coffee, scheduled happy hours, departmental seminars…”

    True – but I think you’d probably laugh at the idea that authors should be expected to keep up with what’s happening in every Friday cake n coffee session around the world, even if they were streamed live to the internet.

    It’s an interesting point though. I’d say the point of peer-reviewed literature is, in part, precisely to overcome this informational problem. It needs to provide a centralised repository, an auditorium clearly visible to all interested parties. All of the informal science Amac mentions, in the end, has to pass through there.

    So the point isn’t that `informal science’ goes on – of course it does. But that peer-review is a way of making sure everyone’s on the same page, informationally.

    Someone responding to a blog-post doesn’t change that, any more than if they’d had an argument at coffee morning. If both end up in the literature, job done (and, of course, you wouldn’t expect anyone to cite the coffee morning.)

  178. AMac Says:

    Dan Olner (12:48)

    I agree with what you wrote.

    As I see it, the puzzle that I invited Prof. Vermeer to respond to in #2 was why Mann08’s authors wrote their Reply to M&M such that it was cryptic and incorrect. And then repeated the Tiljander-themed mistakes of Mann08 in subsequent peer-reviewed papers.

    “They didn’t understand the issues that M&M’s Comment had raised concerning Tiljander” is one possibility.

    In writing their Reply, they might have Googled “Tiljander Mann” etc. If lack of comprehension of the Calibration and Upside-Down issues was the problem, that would have provided a solution.

    At some point, “We were familiar only with what had been entered into the peer reviewed literature” stops being a very compelling explanation. And if one chooses to revise a SI figure to respond to blog criticism, it seems fair for others to assume that one is familiar with that blog criticism.

  179. PeteB Says:

    Bart,

    thanks for highlighting Jeff Id’s concise explanation and your reply

    I found this very interesting :

    http://www.ncdc.noaa.gov/paleo/reports/trieste2008/tree-rings.pdf.

  180. J Bowers Says:

    Thanks Bart.

  181. Jim Bouldin Says:

    Bart at 11:41:

    Oh boy. BIG problem with Jeff ID’s point that you quoted above.

    To summarize: He is arguing that a hockey stick emerges as an artifact of the method used for screening proxies to include in a reconstruction (with specific reference of course, to *Mike Mann’s* reconstructions). The cause of this artifact production is supposed by him to be due to the fact that: “…The series are scaled and/or eliminated according to their best match to measured temperature which has an upslope. The result of this sorting is a preferential selection of noise in the calibration range that matches the upslope, whereas the pre-calibration time has both temperature and unsorted noise.”

    This statement is entirely *false*, and it is so on several levels (including use of poor terminology such as “sorted” to mean screened). Not only is it false, it shows a phenomenal lack of attention to the most basic of facts, as presented by Mann et al in their 2008 paper in PNAS, both in the main paper and in the supplemental material. To wit:

    There were 1209 proxies (from some larger candidate set) that met three initial screening criteria, (such as minimum length of record and stated minimum correlation among the individual members at a given site). From these 1209 records, a nominal screening cutoff of p < .10 with either of the two closest instrumental temperature grid points, was established. (After accounting for temporal autocorrelation, this p value rises slightly to p < .128). If one assumes a positive relationship between ring measure and temperature (i.e. one tailed test), the expected number of sites meeting this criterion is: 1209 * .128 = 155. (If one assumes that either a positive or negative relationship might occur, which they do not, the number is half that, about 78.)

    The actual number of sites that passed this screening: 484, or over 3 times the number expected based on chance alone, (i.e. assuming no relationship between rings and temperature, and using a one tailed test).

    Furthermore, the mean correlation for sites with records that went back to 1000 AD was 0.33. The probability of getting an r value that high by chance, over a 150 year calibration period, for 59 sites, is very small indeed. Note that Mann et al pointed all but the last of these things out in either the paper, the supplement, or both.

    In short, the probability of getting 484 sites that pass the p < .128 screening by chance, is very small, and his argument is utterly wrong. The only way it could be true is if somehow the temp-ring relationship magically arose in 1850 but didn't exist beforehand, which of course is ludicrous.

    I imagine you had no intention that this post would devolve into another paleoclimate debate, which in the minds of some, is of course perfectly synonymous with the hockey stick "debate".

    I note also that Jeff ID states in the thread you mentioned: "Even though I am certain this is one hundred percent correct, this changes little about the climate story. What it does do is make one wonder how math skilled individuals still refuse to acknowledge it."

    Well Jeff, maybe because it's perhaps, patently wrong?

  182. Jim Bouldin Says:

    Not to mention the fact that all 1209 of the set passing the first screening, had mean inter-member correlations (e.g. the pairwise correlations between all cores at a tree ring site), of at least 0.5.

    Now what is the chance of having 484 sites with p = 0.5. To what factor of the environment–which they must all be sensing in order to achieve such high inter-member correlations–are these trees collectively responding, pray tell?

    Biased choice of spurious noise vs actual signal in the calibration period is what creates the blade of a hockey stick eh?

  183. Jim Bouldin Says:

    PeteB:

    “But ultimately it is what is in the peer reviewed literature which really counts”

    Yes, and it would be funny if it weren’t so sad to see the way the full range of issues that scientists are *actually* dealing with there, gets truncated and twisted by those with a limited viewpoint.

    A number of these folks don’t even know what the real problem areas are. Hence, back to Bart’s original point here.

  184. AMac Says:

    Jim Bouldin —

    You offered an informed summary of Mann08’s procedures, so the odds are good that we could have a worthwhile conversation about a technical issue.

    Do the methods in the text and SI of Mann08 require that each proxy be directly calibrated to the instrumental temperature record, 1850-1995?

    Can any of the four Tiljander data series be meaningfully calibrated to the instrumental temperature record, 1850-1995?

    To be clear, these four data series are among the 1209 that you discuss. Three of the four passed Mann08’s screen (for the EIV recons, IIRC).

  185. Jim Bouldin Says:

    Amac:

    Yes, they do require it (calibr. to instrumental)–that is the final screening procedure they used.

    I would not make any statement regarding the calibrate-ability of the four Lake K sediments until I had read Tiljander et al’s paper (which I have not), given that I know utterly zip about the geophysical determinants of boreal lake sediments.

    But if you want to discuss the general issue of tree ring calibrate-ability, that I can do.

  186. AMac Says:

    Jim Bouldin —

    Thanks for the straight answer on whether calibration is required. I agree with you that it is a mandatory step in Mann08.

    The question of whether the Tiljander data series is calibratable is important to the central conclusions of Mann08.

    Without Tiljander, the EIV non-dendro reconstruction doesn’t resemble the EIV dendro reconstructions. And the No-dendro/no-Tilj recon only passes Mann08’s validation test for the 350 years after 1500. It fails from 500 to 1499 (wrt the pre-defined cutoff).

    It might help to eyeball the Tiljander data; check my blog for the graphs.

  187. Jim Bouldin Says:

    MikeN at 22:19:

    “I would say that there was a time period around 1000 which was warmer than now….the Mann 08 paper suggests that this time period was warmer. ”

    The *EIV, NH, land only* reconstruction suggests that it might have been ~ equal to the present. Is it not strange that the fellow who is, according to some, supposedly doing every thing in his mortal power to kill the MWP, has published, in a very prominent journal, a reconstruction that gives among the best credence to the existence of an MWP?

    By the way, I would add that it is quite possible that somewhere between about 900 and 1400 it was warmer than the present (especially around 950 and 1350). But the reasons for this are not apparent by simple inspection of Mann et al 2008, figure 3.

    “The algorithm mines for hockey sticks, filtering out proxies that do not show modern warming, while the past wiggles would balance out against each other, almost guaranteeing a hockey stick if you feed in random data.”

    NO, it does not. See my comment at 01:11 above and also this response by MIke Mann some time ago at RC:

    http://www.realclimate.org/index.php/archives/2009/06/a-warning-from-copenhagen/comment-page-3/#comment-127626

  188. MIkeN Says:

    >If one assumes that either a positive or negative relationship might occur, which they do not, the number is half that, about 78.

    Shouldn’t you expect more matches if you loosen the criteria.
    Looking at the code, it appears these numbers are used as minimum correlations in the screening step. You are using them as probabilities.

  189. Bart Says:

    Jim,

    Thanks for your reaction, that’s helpful in understanding these issues. I haven’t looked into the hockeystick issues in detail at all (for reasons I have explained on this blog before), but since it’s such a recurring theme I am interested in at least understanding the basic reasoning of different actors in this (public rather than scientific) debate. See also my reaction to Jeff, where I wrote:

    To me it sounds entirely reasonable to weigh the proxies based on how well they reproduce the instrumental temperature record.

    That seems to be one of the central points of your reply as well if I’m not mistaken.

  190. AMac Says:

    Bart 11:10 —

    > I am interested in at least understanding the basic reasoning of different actors in this (public rather than scientific) debate.

    That strikes me as a reasonable starting place for looking at this issue.

    An atypical feature of this discussion is that one side (myself, MikeN) is consistently saying, “let’s examine the technical issues, because our position is supported by facts, logic, and the literature.”

    Proponents of “Mann08 is strong science” are silent (or lawyerly) on these matters.

    I can declare, “the cube root of eight plus two-squared is six,” because it is. Others can counter that that sum is an irrational number or unknowable, but they’ll likely end up with “Racehorse Haynes” arguments. There’s no there there.

    Jim Bouldin offered a good synopsis of Mann08’s proxy-selection methods at 01:11. At 02:21, he and I agreed on this statement:

    Mann08’s methods require that each proxy be directly calibrated to the instrumental temperature record, 1850-1995.

    This sets the stage for the central issue in this debate:

    Can any of the four Tiljander data series be meaningfully calibrated to the instrumental temperature record, 1850-1995?

    As you know, I offer a simple and direct answer:

    No.

    Jim Bouldin won’t answer off-the-cuff, as he hasn’t looked into it (02:21) — a very reasonable position.

    OK. As you try and gain understanding into the basic reasoning of the different actors in this (public rather than scientific) debate, here are two questions to mull:

    Among the “Mann08 is strong science” advocates, who has advanced the plain statement that “Yes, the Tiljander data series are calibratable”?

    To my knowledge: Nobody.

    What might explain knowledgable advocates’ reluctance to advance this simple and necessary defense of Mann08’s methods and conclusions?

  191. J Bowers Says:

    AMac, use of Tiljander by Mann et al may be a blip, but, if so, it’s exactly that; a blip. Not everyone shares your fascination with it, especially when Tiljander is removed from the reconstructions; upside down or right way up, it makes no difference. When you say, “Proponents of “Mann08 is strong science” are silent (or lawyerly) on these matters.”, I’m afraid all I see from those using the Tiljander question to cast doubt on the wider field, and the Hockey Team in particular, are lawyerly arguments and “gotchas”. It’s a sideshow.

    Meanwhile, Arctic sea ice is playing to the script, just as predicted using physics. That’s a main event.

  192. willard Says:

    AMac,

    I only have a few minutes until at least tomorrow, so I’ll close two loops, ask one question and provide an answer to one of your latest question.

    First, I do like the sound of Mainstream and Minority. With Mud, we have an alliterative device. Speaking of Mud device, here is one:

    http://tomgerhardt.com/mudtub/

    Second, I would like to quote how Martin Vermeer and you closed your conversation around point (1).

    On the 2010-07-22, at 15:95, Martin continued:

    AMac, I see now that I was ambiguous. What I meant was that the understanding, or interpretation, by Tiljander et al, on what was the proper orientation of these proxies, was likely correct.

    Mann et al. of course assumed all lake sediment proxies to be oriented ‘upward’. The paper says so, and — if what MikeN asserts, that 4000 is the code for this proxy type, is true — so does the code.

    On the same evening, at 19:24, you replied:

    Martin Vermeer, thanks for the clarification.

    Here is what I understand you to be saying in the just-prior comment.

    * Tiljander03 was likely correct in their interpretation of Lightsum and XRD: that higher values correspond to lower temperatures.

    * Mann08 assumed all the Tiljander proxies were oriented ‘upward.’ Thus, Mann08 interpreted Lightsum and XRD in this fashion: that higher values correspond to higher temperatures.

    Is this paraphrasing correct?

    Later in the night, on the 2010-07-23, at 00:58, Martin confirmed the paraphrase:

    Yep.

    Next evening, on the 2010-07-23, at 18:00, you thank Dr. Vermeer:

    Martin, your clarity is much appreciated. Ari Jokimäki appears to have arrived that this conclusion, as well.

    Source: http://arthur.shumwaysmith.com/life/content/michael_manns_errors

    ***

    So here’s my question: am I right in believing that Jim Bouldin, Martin Vermeer and Ari Jokimäki are not advocating “Mann08 is strong science”, whatever that means, while still being Mainstream? In fact, considering that *nobody* said that the Tiljander data series were calibratable, I fail to see who’s in Minority Mud. Everyone seem to agree to be in Mainstream Mud, pending further analysis.

    ***

    So a plausible answer to your question:

    > What might explain knowledgable advocates’ reluctance to advance this simple and necessary defense of Mann08′s methods and conclusions?

    would be that either these knowledgeable advocates do not exist, or, if they do existe, they reject that this defense is necessary, for instance because of some disagreement over what is implicit in the expression “strong science”.

    And so we yet again we might have to fall back on another interesting statement

    > It doesn’t matter.

    This statement deserves due dligence.

  193. AMac Says:

    J. Bowers and willard,

    Interesting points; response tonight. Meanwhile, a housekeeping question for willard — when I update the 2 “links” posts at my blog to reflect recent activity, would you prefer that I keep a “frozen” copy of each, that a Social Network Analysis could refer back to?

  194. willard Says:

    AMac,

    No need for frozen copies. Updates are fine. Interested readers will follow the discussion anyway.

  195. AMac Says:

    J Bowers (16:04) —

    I agree with much of your comment. I don’t expect folks to agree with my interest in Tiljander. It’s a wide world. The matter came up because MikeN and others thought it relevant to the issue Bart raised in his post.. As I noted upthread (July 2, 2011 at 18:40), I also think so.

    > Tiljander is removed from the reconstructions; upside down or right way up, it makes no difference.

    Well, it depends what “makes no difference” means. If “teaches us nothing about Arctic sea ice,” then I agree. If “the main conclusions of Mann08 hold, Tiljander notwithstanding”… then not so much.

    > lawyerly arguments and “gotchas”. It’s a sideshow.

    My question is the opposite of a lawyerly argument, gotcha, or sideshow.

    Can any of the four Tiljander data series be meaningfully calibrated to the instrumental temperature record, 1850-1995?

    > Meanwhile, Arctic sea ice is playing to the script

    Arctic sea ice is a different topic, important in its own right.

  196. AMac Says:

    willard (17:30)

    Re: your link, I anxiously await Varve tub!

    On Martin Vermeer and me at Arthur Smith’s —

    The cordial end of our conversation did not prevent Prof Vermeer from alluding derisively to my stance in that exchange, at a later date. But tame stuff, and I don’t have a link to hand.

    Prof Vermeer went on to use Tiljander-including recons in Kemp11 (he’s a coauthor), so presumably, yes, he holds to the belief that “Mann08 is strong science.”

    > am I right in believing that Jim Bouldin, Martin Vermeer and Ari Jokimäki are not advocating [Mann08 is strong science]… while still being Mainstream?

    Prof Vermeer, we’ve considered. Jim Bouldin won’t opine before reading Tiljander, he said upthread. Ari Jokimäki seems to have abruptly lost interest; Arthur Smith as well. So I don’t know their current thinking.

    > either [AMac's] knowledgeable advocates do not exist, or, if they do exist, they reject that this defense [that the Tiljander data series are calibratable] is necessary, for instance because of some disagreement over what is implicit in the expression “strong science”.

    Yes, anybody can hold any position that they like. At least some Mainstreamers do appear to think that Mann08 is strong science, notwithstanding the non-calibratability of the Tiljander data series. Not to worry–that leaves as many as five additional impossible things to believe before breakfast.

  197. danolner Says:

    Amac:

    “Well, it depends what `makes no difference’ means. If ‘teaches us nothing about Arctic sea ice,’ then I agree. If ‘the main conclusions of Mann08 hold, Tiljander notwithstanding’… then not so much.”

    Assuming I’ve got the right paper –

    Michael E. Mann et al., “Proxy-based reconstructions of hemispheric and global surface temperature variations over the past two millennia,” Proceedings of the National Academy of Sciences (2008), http://www.pnas.org/content/early/2008/09/02/0805721105.abstract.

    This is my first attempt at reading the paper itself, so quite likely to be missing a lot – but on a first read, the conclusions absolutely still hold without any of the tree-ring proxies. More than that, one of the paper’s main goals is clearly precisely to ask “exactly what difference does it make to use or leave out the tree ring proxies?” So clearly, their conclusions do hold. Amac, what did you think their conclusions were?

    Before getting onto some quotes, a question jumps out at me from reading the paper itself: on Amac’s question, “Can any of the four Tiljander data series be meaningfully calibrated to the instrumental temperature record, 1850-1995?” Would it matter if they couldn’t? Isn’t there a bigger job of cross-validating the proxies? Each doesn’t have to independently calibrate to the instrumental record, does it? Also, there’s the larger issue of spatial reach (mentioned below), but perhaps we can come back to that.

    Secondly, would it matter if they couldn’t, because plenty of their conclusions are based on non-tree-ring analyses? As I say, they go to some lengths, repeatedly, to be clear on exactly what difference all the tree-ring proxies make. Some quotes, first from the start:

    “For both methods, we perform reconstructions both with and without dendroclimatic proxies to address any potential sensitivity of our conclusions to issues that have been raised with regard to the reliability of tree-ring data on multicentury time scales.” [13253]

    And the conclusion –

    “We find that the hemispheric-scale warmth of the past decade for the NH is likely anomalous in the context of not just the past 1,000 years, as suggested in previous work, but longer. This conclusion appears to hold for at least the past 1,300 years (consistent with the recent assessment by [Jansen et al 2007]) from reconstructions that do not use tree-ring proxies, and are therefore not subject to the associated additional caveats [see below]. This conclusion can be extended back to at least the past 1,700 years if tree-ring data are used, but with the additional strong caveats noted.”

    Some larger quotes that leapt out at me. This one drawing attention to the uncertainty analysis, and tree-ring data again:

    “The recent observed decadal warmth recorded in the instrumental observations exceeds the uncertainty range of the reconstructions over at least the past 1,600 years for NH land temperatures as reconstructed by CPS (Fig. S5) and the past 1,700 years forNHland plus ocean temperatures as reconstructed by EIV. Because this conclusion extends to the past 1,300 years for EIV reconstructions withholding all tree-ring data, and because non-tree-ring proxy records are generally treated in the literature as being free of limitations in recording millennial scale variability the conclusion that recent NH warmth likely exceeds that of at least the past 1,300 years thus appears reasonably robust. It is intriguing to note that the removal of tree-ring data from the proxy dataset yields less, rather than greater, peak cooling during the 16th–19th centuries for both CPS and EIV methods (see Figs. S5a and S6b, respectively), contradicting the claim (33) that tree-ring data are prone to yielding a warm-biased ‘Little Ice Age’ relative to reconstructions using other high-resolution climate proxy indicators.” 13255

    And that holds with or without tree-ring data. As far as I understand it, the reason for using tree-ring data at all is – well, quote: “When tree-ring data are eliminated from the proxy data
    network, a skillful reconstruction is possible only back to A.D. 1500 by using the CPS approach but is possible considerably further back, to A.D. 1000, by using the EIV approach.” So you can’t go back to 500 A.D. – although actually it sounds like that’s because of the spatial reach of the tree-ring data, compared to other proxies.

    “we place greatest confidence in the EIV reconstructions, particularly back to A.D. 700, when a skillful reconstruction as noted earlier is possible without using tree-ring data at all.”

    For the Southern Hemisphere: “skillful CPS reconstructions are not possible without tree-ring data before A.D. 1700, implying additional caveats as discussed above. Recent warmth exceeds that reconstructed for at least the past 1,800 years in the EIV reconstructions, and this conclusion extends back at least 1,500 years without using tree-ring data. However, the estimated uncertainties are compatible with the possibility that recent SH warmth might have been breached during brief periods in the past. Similarly, for global mean temperature, the CPS reconstruction suggests that recent warmth is anomalous for at least the past 1,500 years, but with the caveat that tree-ring data are required for a skillful long-term reconstruction. The EIV reconstruction indicates recent warmth that exceeds the reconstructed warmth (past 1,500 years with caveats related to the use of tree-ring data, and the past 1,300 years if tree-ring data are excluded), but like the SH, the uncertainties are compatible with the possibility of brief periods of similar warmth over the past 1,500 years. More confident statements about long-term temperature variations in the SH and globe on the whole must await additional proxy data collection.” [13256]

  198. Jim Bouldin Says:

    Amac at 5:03

    “> am I right in believing that Jim Bouldin, Martin Vermeer and Ari Jokimäki are not advocating [Mann08 is strong science]… while still being Mainstream?

    Prof Vermeer, we’ve considered. Jim Bouldin won’t opine before reading Tiljander, he said upthread. Ari Jokimäki seems to have abruptly lost interest; Arthur Smith as well. So I don’t know their current thinking.”

    Hold on there.

    What I won’t opine on is the issue of how to calibrate the Finnish lake sediment record. I don’t know where this line of argument started, or where it’s attempting to lead, nor do I understand what whoever introduced the term “strong science” here (you?) means by it, but I most certainly *do* believe that Mann et al (2008) is a strong piece of science, for several reasons. It sure as hell doesn’t pass or fail based on the Tiljander data when you’ve got 480 other carefully chosen proxies spread out over the globe. It’s your obsession with this minute component of the data that leads to your opinion that the paper is bad science.

    Please don’t put words in my mouth based on your understanding of the science.

  199. Jim Bouldin Says:

    Amac, the primary reason that you think Mann et al 2008 is not “strong science” is because you allow the search for errors to dominate over the search for validity. The McIntyre approach in other words.

  200. Jim Bouldin Says:

    Bart at 11:10:

    Yes I thought your previous reply to Jeff Id was good. The issue of the weighting of proxies based on their correlation to instrumental is a little bit different from the issue that I expounded on at 01:11. There, I was referring to the issue of inclusion/exclusion of candidate proxies into one’s data set, based on p values of a correlation with instrumental data, and Jeff Id’s false claim that the random variation component that is included in the set of screened proxies, by this procedure, is what causes a hockey stick to appear (it doesn’t). Hope that helps.

  201. Jim Bouldin Says:

    MIkeN at 06:54:

    “Shouldn’t you expect more matches if you loosen the criteria.
    Looking at the code, it appears these numbers are used as minimum correlations in the screening step. You are using them as probabilities.”

    You’re confusing “loosening the criteria”, (by which I assume you are referring to either the p or r values of the correlation of proxy with instrumental over the calibration period) with a one vs two tailed test. Yes, if you increase the permitted p value (or lower the permitted r value), you will include more candidates into the set. My point was more of a side point: if you didn’t necessarily assume a positive proxy variable to temp correlation, you would use a 2-tailed test, and thus expect about 78 of the proxy series to show a positive relationship with temp, instead of 155. Shouldn’t have even mentioned it.

    I can’t read Matlab code but the paper is quite clear that the proxies’ final screening criterion was the p value over the full calibration interval, not the r value.

    And by the way, the fact that a high percentage of the 484 proxies showed a positive correlation with T is, by itself, highly meaningful, with respect to the issue raised by Loehle (2009, A mathematical analysis of the divergence problem). But this is too involved to go into here. Suffice it to say that it limits the likelihood that certain, defined time points in the past had proxy to T relationship direction changes. It also relates to why I said that it is possible that around 1350 or 950, temperatures were warmer than the present. Still unlikely when you consider all the reconstructions, but possible.

  202. AMac Says:

    Jim Bouldin (18:57) —

    > Please don’t put words in my mouth based on your understanding of the science.

    Believe it or not, I try hard to characterize other people’s positions correctly.

    You wrote (July 14, 2011 at 02:21) —

    I would not make any statement regarding the calibrate-ability of the four Lake K sediments until I had read Tiljander et al’s paper (which I have not)…M

    I noted to Bart (July 14, 2011 at 14:00) —

    …This sets the stage for the central issue in this debate:

    Can any of the four Tiljander data series be meaningfully calibrated to the instrumental temperature record, 1850-1995?

    As you know, I offer a simple and direct answer:

    No.

    Jim Bouldin won’t answer off-the-cuff, as he hasn’t looked into it (02:21) — a very reasonable position.

    willard then asked me (July 14, 2011 at 17:30) —

    So here’s my question: am I right in believing that Jim Bouldin, Martin Vermeer and Ari Jokimäki are not advocating “Mann08 is strong science”, whatever that means, while still being Mainstream?

    I responded (July 15, 2011 at 05:03) —

    Prof Vermeer, we’ve considered. Jim Bouldin won’t opine before reading Tiljander, he said upthread. Ari Jokimäki seems to have abruptly lost interest; Arthur Smith as well. So I don’t know their current thinking.

    You hadn’t offered an opinion on “Mann08 is strong science (whatever that means)”. I mistakenly equated your not-commenting with your being in fact-gathering mode (reading Tiljander). I thus mistakenly answered willard’s question with “Jim Bouldin won’t opine before reading Tiljander,” rather than with the cautious and correct “I don’t know what Jim Bouldin thinks.”

    You have now cleared that up by writing (18:57), “I most certainly *do* believe that Mann et al (2008) is a strong piece of science…”

    Perhaps you can see my 05:03 response to willard as an error made in good faith. In any case, I apologize for putting words in your mouth.

  203. Jim Bouldin Says:

    Above should be:

    Suffice it to say that it limits the likelihood that certain, defined time points in the past had proxy to T relationship directional changes, (due to unimodal responses of proxy to temp) and hence, badly under-estimated temperatures.

  204. Jim Bouldin Says:

    OK, thanks much for clarifying Amac. I see that it was a misunderstanding introduced by Willard’s earlier comment (and Willard, that is not my position, nor do I think it likely that it is Martin’s). I may be in fact finding mode regarding how to calibrate the sediment data, but I’m not in terms of whether Mann et al is a good piece of work. Those are very separate questions, to a number of us here.

  205. AMac Says:

    Jim Bouldin (19:04) —

    > the primary reason that you think Mann et al 2008 is not “strong science” is because you allow the search for errors to dominate over the search for validity.

    Sorry. I don’t know what that means.

    Here is a set of ideas.

    H1. One of Mann08’s central findings was that multi-century temperature reconstructions could be successfully constructed, without using tree-rings.

    H2. The paper showed that Tiljander-including non-dendro recons could be validated back as far as 700.

    H3. The paper showed that Tiljander-including non-dendro recons’ portrayals of paleotemperatures are quite good matches to tree-ring-based recons’ portrayals.

    H4. Non-dendro recons were built with dozens of data series, including the Tiljander data series.

    H5. Mann08’s methods require direct calibration of proxies to the instrumental temperature record, 1850-1995.

    H6. The Tiljander data series cannot be directly calibrated to the instrumental temperature record, 1850-1995.

    H7. Together, H5 and H6 compel the exclusion of the Tiljander data series from recons such as those presented in Mann08.

    H8. When the Tiljander data series are not used in EIV non-dendro recons, those recons fail validation earlier than ~1500. (Cf. H2).

    H9. When the Tiljander data series are not used in EIV non-dendro recons, those recons’ portrayals of paleotemperatures do not resemble tree-ring based portrayals. (Cf. H3).

    H8 and H9 argue against one of Mann08’s central findings (Cf. H1).

  206. MIkeN Says:

    Jim Bouldin, I got that you were referring to two sided tests.
    Shouldn’t I expect more proxies to be positively or negatively correlated with temperatures, than proxies that are positively correlated with temperature?

    What is the p probability?

  207. MIkeN Says:

    Dan Olner, you are correct that there is not much difference when you leave out tree-rings or include them, except Tiljander is not a tree-ring proxy. if you take out Tiljander, then there is a difference between the main chart and the no-treering chart that is substantial, and arguably overturns the conclusions of the paper.

  208. MIkeN Says:

    Jim Bouldin, the code for CPS refers to the variable corra and comments say rvalues.
    corra=0.106; %% 9000,8000,7500,4000,3000,2000

    %———– r-table description ————–
    % row 1: longitude
    % row 2: latitude
    % row 3: proxy ID (e.g. 8000 annually-resolved ice)
    % row 4: r -value for raw data over 1850-1995(entire),

    if … z(ia,i)>=corra

    Your description of the algorithm is at odds with Jeff Id’s description.
    He responded to Mann’s RealClimate comment here.

    http://noconsensus.wordpress.com/2009/06/20/hockey-stick-cps-revisited-part-1/

  209. Jim Bouldin Says:

    MikeN:

    @22:03
    No, because you are including both tails of the distribution–therefore you have to divide whatever number you would expect to get if a one tailed test were the appropriate model, by 2. So, if tree rings were hypothesized to respond either positively or negatively to T, you would expect to get about 155 sites that exceeded p values of either + or – 0.128 (assuming you started with 1209 series). Then you would compare whatever number you did observe, with that number. Again, I shouldn’t have mentioned it, because it’s not important. The appropriate model is that the relationship between rings and T is positive,therefore, the one tailed test is the right one. There were a very small number–about 10 or 12–of the ring series that had a negative relationship. Thus my earlier point about the importance of this finding in relationship to the points raised by Loehle in the paper I’ve cited, and which I hope people are following, but which are tangential to this conversation.

  210. Jim Bouldin Says:

    and at 22:18:

    As I said, I don’t read Matlab, so the code snippet you supply doesn’t help me, and even if I did, same response, because it’s a snippet.

    I remember looking at the post you cite some time ago. Jeff Id has *no idea* what he is talking about, and he has formulated an entire theory of bad science based upon it. This is evidenced very clearly by the following statement he makes there:

    ” When I read about step 1 I actually didn’t sleep that night. Throw away the data that doesn’t fit your pre-determined conclusion of a match to temperature. Really, I was so mad I didn’t sleep, remember we’re paying big $$ for this stuff and we must not pretend it’s somehow ok. NOT good.”

    I explained this very point very clearly to Amac on his blog post titled Voldemaart’s question.

    Now we are finally really getting to what Bart is raising in this post.

  211. Jim Bouldin Says:

    And lastly, looking at the data table, you may well be right that the screening involved an r value instead of a p value criterion (since the minimum r value listed for the proxies that passed screening was just over 0.10). However, this distinction is completely meaningless, because p and r are going to correlate strongly–you can use either one. The important point is the number of proxies passing screening that exceed those from expectation by chance.

  212. AMac Says:

    Jim Bouldin (01:22) —

    > I explained this very point very clearly to Amac on his blog post titled Voldemaart’s question.

    I think (but am not sure) that you are referring to the comment that you made here.

  213. AMac Says:

    Dan Olner (July 15, 2011 at 15:15) —

    Yes, you have the right paper.

    > on a first read, the conclusions absolutely still hold without any of the tree-ring proxies.

    This is the authors’ claim. I disagree. I have tried to outline why, most recently in the comment of July 15, 2011 at 20:02.

    > Amac, what did you think their conclusions were?

    From your narrative and your quotes of the paper, I can say that we agree on what Mann08’s conclusions were. E.g. “one of the paper’s main goals is clearly precisely to ask ‘exactly what difference does it make to use or leave out the tree ring proxies?’”. More generally, the authors claim that the recons built from non-tree-ring proxies are in general agreement with those built from tree-ring proxies. They believe that the convergence of these two lines of independent evidence strengthens the case that the overall “hockey-stick” appearance of their paleotemp recons are a good representation of the earth’s temperature history.

    > Would it matter if [the Tiljander data series] couldn’t [be meaningfully calibrated to the instrumental temperature record, 1850-1995]?

    Yes. It matters. That’s the assertion that I’ve gone to some length to substantiate in this thread, with links back to my blog for fuller treatment of the reasoning, evidence, and background.

    > Isn’t there a bigger job of cross-validating the proxies? Each doesn’t have to independently calibrate to the instrumental record, does it?

    I’m not sure what you mean. I think the answers are No, and Yes.

    > would it matter if [the Tiljander proxies] couldn’t [be calibrated], because plenty of their conclusions are based on non-tree-ring analyses?

    Well, the failure to properly calibrate non-dendro proxies wouldn’t affect any of the tree-ring-only reconstructions.

    As far as I can tell, this failure calls all of the non-dendro recons into question, except the No-Tilj/non-dendro ones. And the No-Tilj/non-dendro recons in the SI (SI Fig S8) seem to have other errors. The latest version of the No-Tilj/non-dendro recons appear to be in the Mann09 SI. See my post Voldemort’s Question for those traces.

  214. MIkeN Says:

    Jim Bouldin, I think I see where I got confused. You listed 78 sites as having a POSITIVE correlation when two-sided test is used.

    I think you are missing Jeff Id’s point that these are noisy proxies. Now if you have noisy proxies, some of those will pass screening, while the historical noise will balance out to 0. You already have over half the proxies not passing screening. How many passed only because of random noise? White noise correlates to the temperature record 13% of the time.

  215. MIkeN Says:

    Make that 11.6% of the time, for the .106 level used in the code.

  216. willard Says:

    AMac,

    Let’s review you questions:

    (1) Was the original Tiljander03 paper correct in assigning these temperature correlations to the pre-1720 period?

    (2a.1) Should Mann and coauthors have figured out the problems with calibration of the Tiljander proxies to the instrumental record from the discussion in the blogosphere?

    (2a.2) Should Mann and coauthors have figured out the problems with calibration of the Tiljander proxies to the instrumental record from the discussion from the M&M comment in PNAS?

    (2a.3) Should Mann and coauthors have figured out the problems with calibration of the Tiljander proxies to the instrumental record from the discussion from the M&M from the correction to the Kaufman Science 2009 manuscript?

    (3.a) Is it possible, despite the third version of Fig. S8a showing no (or “modest”) substantive difference to the temperature anomaly trace in the reconstruction period, that use of the Tiljander proxies made a material difference to some of the conclusions of Mann08?

    (3.b) Would that criticism extend to some conclusions of later papers using the same data in similar fashion?

    We already saw Vermeer’s first direct answer to this set of questions at Arthur’s, on the 2010-07-22, at 12:37?. From this answer, we can presume that his answer to (1) is “yes”. This is good news, as it is the precondition to discuss the other ones.

    Let’s put aside for the moment (2a.1) and repeat the parts we have not discussed yet:

    AMac, the only question that has some scientific relevance is (3). [...]

    (3) I do not believe this is realistically possible. Any effect from the mishandling of Tiljander drowns in the other, substantial uncertainties both statistical and structural, especially going back in time before 1000AD. These uncertainties are well described in the paper. I would have no hesitation to use these reconstructions in my own work. [...]

    [http://arthur.shumwaysmith.com/life/content/michael_manns_errors#comment-9982]

    In fact, I do believe that Vermeer’s answer to (3) can be paraphrased as “(1) and (2) do not matter”. Vermeer’s formulation of (3) makes it immune to any accusation of postmodern wording. Do you agree that the type-3 questions are scientific and that the type-2 questions are not?

    ***

    I suppose you do not consider type-2 questions lawyery. Considering his answer to (3), and in the light of your proviso to (3.a), it might be understandable why he has no patience for type-2 questions. Perhaps this lack of patience can explain his sardonicism, but this is not enough.

    So let’s follow with this other comment by Martin Vermeer, on the 2010-07-25, at 4:40?:

    AMac, I disagree strongly for reasons that should be clear to any reasonable reader by now. Mann’s reply was fully responsive to the statement as made by McIntyre in the Comment. He could not be expected to engage in mind reading, or to follow, or even be aware of, discussions on fringe blogs. I am disappointed that you continue to fail to see this, AMac, and have nothing further to add.

    [http://arthur.shumwaysmith.com/life/content/michael_manns_errors#comment-10041]

    This seems to answer (2a.1) and (2a.2). I doubt you find it satisfying. They are still of note.

    This other comment by Martin Vermeer, on the 2010-07-22, at 12:04 (i.e. earlier than the two above), might provide more evidence:

    TCO, admitting being wrong is not easy for anyone…. still Mike Mann has published his share of corrigenda in journals, and less formal corrections on his own web site. He will acknowledge real mistakes.

    About being unrepentant about errors, I don’t see the problem with that. Sin belongs in church, and the only way not to make errors is to not do anything, and certainly not anything for the first time ever.

    About “it doesn’t matter”: that’s very, very often the right question to ask. “All models are wrong”, remember? Ask if the wrongness matters. And get the order of magnitude right. One problem I see with McIntyre is a lack of intuition on the influence of various “wrongnesses” in data and statistical procedures that he goes on about. That broke him up in the Yamal spat too. Strange for someone with the reputation of being a skilled statistician.

    [http://arthur.shumwaysmith.com/life/content/michael_manns_errors#comment-9981]

    This answer, posted earlier than the first one we analyzed, contains material that deserves due diligence.

    For instance, “unrepentance” belongs to the same semantic field as “inerrancy”. There is also this interesting opposition between unrepentance, inerrancy, and “real mistakes.”

    More importantly, we see more reasons why the “it does not matter” should not be simply dismissed as postmodern science.

    By reviewing what has been written so far on Tiljander, it might be possible to dig what matters, what mattering means, to what avail and according to which criteria.

  217. AMac Says:

    willard (03:55) —

    > In fact, I do believe that Vermeer’s answer to (3) can be paraphrased as “(1) and (2) do not matter”. Vermeer’s formulation of (3) makes it immune to any accusation of postmodern wording.

    I agree with both of your assertions in the two sentences I quoted. In addition, Vermeer’s claims may be correct, or they may be incorrect. In this thread and elsewhere, I have offered substantial evidence that they are incorrect. If Vermeer or others have offered reasoned counter-arguments consistent with the facts of the matter — I haven’t seen them.

    > Do you agree that the type-3 questions are scientific and that the type-2 questions are not?

    ?? They are all reasonable questions.

    We could also discuss the use of color in abstract-expressionism. You might think that the best artists make extensive use of Blue paint. I might counter that the major elements should be Green.

    Upthread (July 14 at 14:00), I wrote that I can assert that the cube root of eight plus two-squared is six. Others can counter that that sum is an irrational number.

    In the end, I’m discussing small set of technical issues with people who are asserting that they can believe whatever they wish. That is, of course, true. Blue! Green! Six! Irrational number! Matters! Doesn’t matter!

    Can any of the four Tiljander data series be meaningfully calibrated to the instrumental temperature record, 1850-1995?

    If the answer is “No” — and it is — that leads to the set of ideas I outlined on July 15, 2011 at 20:02.

    If “Tiljander doesn’t matter,” then the differences between the Green trace and the Blue trace in Figure S8 from Mann09 “don’t matter.”

    In my opinion, the meta-discussion won’t be very productive, given that Vermeer, you, MikeN, and I — among others — can’t agree on a set of fairly simple facts and fairly simple conclusions to be drawn from them.

  218. Jim Bouldin Says:

    MikeN at 00:34

    Right, 78 would have a positive corr. and the same number would have a negative corr., under a model in which the relationship is uncertain between the variables. Under a model in which you can exclude one correlation direction, a test of randomness with p = 0.128 gives twice that number. That’s what they did.

    Your 2nd paragraph really makes no sense. You state: “You already have over half the proxies not passing screening. How many passed only because of random noise?” What is the significance of this “over half” that you bring up. As already stated, 484 of them *did* pass the screening, and the expected number, due to chance, given a p =.128 criterion and 1209 candidates, is about 155. Now, is there likely a temperature signal in those 484-155 = 349 proxies, or not?

    I’m not missing his point at all. Of course they’re noisy proxies–have you ever seen any that aren’t? Basically, he has claimed to nullify the entire concept of using correlation over a common interval, to predict values over the non-common interval (i.e., in which the values of only variable is known). Not just paleoclimate estimation, but any science or engineering field at all that uses such an approach. Now this is a rather astounding discovery that will change the way things are done in a wide range of fields, is it not? So where’s the publication? In a blog post? How have all the statisticians of the world missed this important concept so far do you think?

  219. willard Says:

    AMac,

    You agree that Martin Vermeer has provided an answer to:

    (3.a) Is it possible, despite the third version of Fig. S8a showing no (or “modest”) substantive difference to the temperature anomaly trace in the reconstruction period, that use of the Tiljander proxies made a material difference to some of the conclusions of Mann08?

    (3.b) Would that criticism extend to some conclusions of later papers using the same data in similar fashion?

    You also agree that Martin Vermeer provided an answer that does not qualify as postmodern.

    This might suffice to confirm your impression in your comment above:

    > If I had a time machine, I would excise my use of “post-normal” from this thread. Perhaps “postmodern” as well. The parable of “The Emperor’s New Clothes” is a better concept for illuminating certain aspects of Tiljander/Mann08.

    http://ourchangingclimate.wordpress.com/2011/06/22/climate-science-scientific-method-skeptics-not/#comment-13542

    If not, I’m sure we can find other instances where “it does not matter” get some ice time.

    I like the Emperor’s New Clothes parable. The character the child who reveals the Emperor’s nudity is interesting. In Andersen’s version, he was not old enough to understand the rules of sociable behavior. But nothing should prevent us to imagine other kinds of characters. The parable makes me think of Nelson from the Simpsons. Not because of you, I hasten to add. On the contrary.

    It would interesting to collect all these mataphors, analogies, and parables.

    ***

    There is need to talk about abstract-expressionnism, cube roots, or any other metaphor if we’re interested in technical issues. Talking about metaphors, analogies, or parables is substantial evidence that we’re not talking about technical issues.

    As far as I can see, the choice of metaphor, analogy or parable tends to illuminate the fact that we’re talking about the behavior of people (e.g. the “resistance of the establishment”), not only technical issues.

    Describing people with whom we are discussing as “asserting that they can believe whatever they wish” is another way of saying that people are not rational. This might prove to be an even stronger hindrance to a rational discussion that talking about post normal science, postmodern science, Cargo cult, Newspeak, inerrancy, or any other metaphor, analogy, or parable.

    Or not. There is always hope. You can testify that from your experiences with TCO and Nick Stokes.

    The shift between technical issues and other kinds of issues creates a dynamic that deserves due diligence.

    Discussing this dynamic will perhaps be the most productive thing we can down here. Or not. It’s a big world.

    ***

    And so I will repeat my question:

    Do you agree that the type-3 questions are scientific and that the type-2 questions are not?

    This is a simple question. It can be answered with a simple answer.

    Here is mine: yes.

    Both types of questions can be reasonable questions, depending on the dynamic that leads the discussion. But my concern for now is that we agree that type-2 questions are not scientific. This is not a matter of fact. So I have to ask.

  220. Jim Bouldin Says:

    484 – 155 = 329 of course

  221. Jim Bouldin Says:

    Amac, is there likely a temperature signal in somewhere around 330 of the proxies in Mann et al 2008 or not?

    If the answer is yes, then why are you so concerned about possible errors from one lake in Finland? Is it because the correlation to T is so much higher than the average r over all sites, that the values would change the global reconstruction substantially?

  222. MIkeN Says:

    Jim Bouldin, the different proxies are not weighted by correlation.

    They are recentered and rescaled. This was a second factor for Jeff Id, saying that the noise in the calibration period was filtered thru, but the noise in the older period was not just canceled out, but the overall signal is reduced, shrinking any medieval warm period.

  223. AMac Says:

    willard (16:39) —

    > Do you agree that the type-3 questions are scientific and that the type-2 questions are not?

    Perhaps the problem arises because I wrote the type-2 questions as if the uncalibratability of the Tiljander data series is fact (I believe that it is a fact). There is thus a dependent clause in each of those sentences, e.g.

    (2a.2) Should Mann and coauthors have figured out the problems with calibration of the Tiljander proxies to the instrumental record from the discussion from the M&M comment in PNAS?

    If one believes that the uncalibratability issue is still undetermined, this would have a structure akin to “Do you still beat your wife?”

    As far as “scientific”, I don’t know what that word is supposed to mean in this context. Here is dictionary.com

    scientific (adjective)

    1. of or pertaining to science or the sciences: scientific studies.

    2. occupied or concerned with science: scientific experts.

    3. regulated by or conforming to the principles of exact science: scientific procedures.

    By definitions 1 and 3, I would say that the Type-2 questions pertain to scientific issues.

  224. AMac Says:

    Jim Bouldin (17:02)

    > is there likely a temperature signal in somewhere around 330 of the proxies in Mann et al 2008 or not?

    I think there likely are temperature signals in many of the data series in Mann08.

    Did Mann08 select the data series that have more temp and less other-stuff? Did Mann08’s procedures extract temp signals from other signals? Did these procedures combine the data series to give meaningful portraits of global temperature anomalies? Do the error ranges adequately reflect the uncertainties of the process? If Mann08 included errors, would this be clear to climate scientists from the results, e.g. from internal inconsistencies?

    My opinion is that the answer to many of those questions is likely to be “No.”

    > why are you so concerned about possible errors from one lake in Finland? Is it because the correlation to T is so much higher than the average r over all sites [?]

    No.

    > [Is it] that the values would change the global reconstruction substantially?

    The addition of the four mis-calibrated Tiljander data series do substantially change the global reconstruction, unless bristlecone pines or other hockey-stick tree-ring series are included.

    This is what the comparison of No-Tilj/no-dendro (Green trace) to Yes-Tilj/no-dendro (Blue trace) teaches in Mann09 SI Fig. S8. Voldemort’s Question.

    Do you think that the Green trace is substantially the same as the Blue trace, with respect to its shape? With respect to its validation according to Mann08/09 criteria?

  225. willard Says:

    AMac,

    Thank you for your response.

    > If one believes that the uncalibratability issue is still undetermined, this would have a structure akin to “Do you still beat your wife?”

    I agree. There is more to say about these matters. But I need to be AFK for a while

    > As far as “scientific”, I don’t know what that word is supposed to mean in this context.

    “Scientific” as in “pertaining to the science theories”, or “scientific question”. But I like your answer, as it makes me feel as a scientist.

    Your reply has a structure akin to a rhetorical device one can witness at the Blackboard. This is a bad omen.

  226. Deech56 Says:

    >The addition of the four mis-calibrated Tiljander data series do substantially change the global reconstruction, unless bristlecone pines or other hockey-stick tree-ring series are included.

    Again, to repeat a point I made earlier, “Since when were scientists required to toss out the dendro data?”

  227. MIkeN Says:

    Deech, they of course are not required to do so, but Mann chose to do so to demonstrate a conclusion that even without dendro certain results can be achieved. On top of that you have a NAS Panel report that said bristlecone pines were suspect as temperature proxies and should not be used.

  228. AMac Says:

    Deech56 (17:34) —

    > “Since when were scientists required to toss out the dendro data?”

    Mann08’s authors chose the topic of their paper, not MikeN or me. Like them, I think it’s a worthwhile subject.

    Upthread, these two comments touched on this issue: July 1, 2011 at 00:08 and href=”http://ourchangingclimate.wordpress.com/2011/06/22/climate-science-scientific-method-skeptics-not/#comment-13472″>July 1, 2011 at 14:18.

  229. AMac Says:

    Ugh.

    July 1, 2011 at 00:08 and July 1, 2011 at 14:18.

  230. Deech56 Says:

    But weren’t the NAS criticisms regarding bristlecone pines superseded by Salazar, et al.? From what I read, the topic of Mann08 was temperature reconstruction, and included in this paper were various alternative reconstructions. I didn’t see them tossing out the dendro data.

  231. Deech56 Says:

    Ugh. ;-) At least the hyperlink works, I think. If not, it’s

    http://www.pnas.org/content/early/2009/11/13/0903029106.abstract?sid=1c81cc57-d8a5-47ac-9652-9664d86f01cf

  232. AMac Says:

    Deech56 —

    Yeah, that’s an interesting paper. IIRC, it made me uneasy because of that whole post-hoc data selection question. Upthread, Jim Bouldin didn’t think those concerns had merit (AFAICT), so presumably you won’t either.

    In any event, Mann08’s authors set out to show that non-dendro proxies told the same story as dendro proxies, within reason.

    I think they did not succeed, for the reasons given. MikeN agrees (again, AFAICT).

  233. MIkeN Says:

    Deech, I read it as the scientists find bristlecones very convenient for creating hockey sticks, so they are trying to resurrect them in the face of critiques. There is still sufficient problem that Mann chose to leave them out to try and make a point. Of course Salzer had not yet been published at that time. Even with the paper, I don’t think the controversy goes away. Here is McIntyre’s first look at the paper.

    http://climateaudit.org/2009/11/17/salzer-et-al-2009-a-first-look/

  234. danolner Says:

    MikeN: “scientists find bristlecones very convenient for creating hockey sticks”

    Just so we’re clear – you’re saying they’re cherrypicking data to produce the result they want?

  235. grypo Says:

    “Even with the paper, I don’t think the controversy goes away. Here is McIntyre’s first look at the paper.”

    Only if blog controversy is important. Scientific controversy occurs in journals. We can leave it to the reader to decide which is more to their liking. One is easier to understand than the other.

  236. Deech56 Says:

    I would have to echo grypo. If the conclusions of Salazar, et al. really bolster the use of tree rings, McIntyre would really have to write a comment to the journal for it to get noticed by the scientific community, unless that’s not his audience. A commentary that offers a valuable critique shouldn’t be hidden under a bushel basket.

  237. Deech56 Says:

    May I quote from the original post (quoting from The Conversation)?

    “In science, to actually contribute at the forefront of a field one has to earn credibility, not demand it. Being taken seriously is a privilege, not a right.

    “In science, this privilege is earned by not only following conventional norms of honesty and transparency but by supporting one’s opinions with evidence and reasoned argument in the peer-reviewed literature.”

  238. AMac Says:

    Dan Olner, grypo, Deech56 —

    It would be interesting to talk about Salazar09, and their ideas about the Divergence Problem. With respect to Bart’s thoughts in the original post, Salazar09 would likely lead to another conversation. Not necessarily better or worse — different.

    Salazar won’t teach us anything about Mann08’s claims on the similarities between tree-ring-based and non-dendro reconstructions. It’s unlikely to offer insight on whether Mann08’s approach to proxy calibration has merits.

    If such things matter. If we care.

    Deech56 (01:54) —

    Taking your quote at face value, it seems that, for discussing peer-reviewed literature on climate-science-themed blogs, the choice is between agreeing with mainstream papers’ authors.

  239. willard Says:

    At 2011-07-02, at 18:40, AMac strongly disagreed that this thread has been hijacked from the theme of Bart’s post by **L’Affaire Tiljander** [LAT]:

    > Bart, Chris Colose, and others have made strong claims about “how [climate] science works.” In my opinion, it is futile to tackle such a broad theoretical subject without grounding remarks in specifics. Since the theme is “science”, those details are necessarily technical in nature. The story of the use of the Lake Korttajarvi data series illustrates how one branch of climatology has diverged from “best practices” as they are known in the other physical sciences.

    http://ourchangingclimate.wordpress.com/2011/06/22/climate-science-scientific-method-skeptics-not/#comment-13490

    Either LAT is a unique example or not. If LAT is a unique example, we won’t be able to infer anything regarding paleoclimatology, climate sciences, or science in general, which was the theme of the post. If LAT is not a unique example, we need to know which common characteristics it shares with other cases, in related fields of science, or not.

    Enough examples so we can see some common characteristics. Not a favorite stock of a few pet examples: this amounts to stacking up the deck. If what we want to to ground remarks about a broad theorical subject in specifics, any detailed example should be welcome, be they from paleoclimatology, climate sciences, or other fields of science.

    Focusing exclusively on technical details of LAT will not suffice to ground remarks that could be generalized to the topic of the thread. Using the topicality of LAT to push technical points can be seen as a way to shift the focus to a very specific instance. This is even more obvious when we suppose there is only one rational combination of answers to the technical points.

    If it’s ok to talk about LAT when discussing science, it should be reminded that is taken as an object of sociology of science. The focus should always be on the topic, which is how to do science, or not, and not technical details of one example.

    If it’s ok to talk about LAT when discussing science, it’s ok to talk about any relevant example.

    So LAT should be one example among many.

    If we also care about science and the topic of the thread, that is.

  240. grypo Says:

    “Taking your quote at face value, it seems that, for discussing peer-reviewed literature on climate-science-themed blogs, the choice is between agreeing with mainstream papers’ authors.”

    The quote never mentions “mainstream”. It mentions peer-review, but more importantly, it mentions credibility of participating in the process of science. So when discussing science, in the context of Mike N’s “controversy” one has a choice between discussing differences in opinions in the literature, or discussing differences in refereed literature and unrefereed blogs. There’s a reason why the “pal review” meme is so important to the skeptic narrative. Without it, the choice is too simple. With it, not so much.

  241. AMac Says:

    willard —

    Here’s Bart, quoting from The Conversation.

    The strategy of the paper is to undermine the credibility of the establishment climate scientists. That’s all. There is nothing special science-wise.

    Undermine credibility.

    That’s all.

    Nothing science-wise.

    [snip]

    Time to close the phony debate on climate science… The so-called “debate” on climate change has been over for decades in the peer-reviewed literature. It is time to accept the scientific consensus and move on, and to stop giving air-time to the cranks.

    It is time for accountability.

    Words like these are bandied around so much, we might sometimes neglect to appreciate their sharp, hard edges.

    There are softer words: “It seems,” “Usually I find that,” “Often,” “On the other hand.”

    For these examples, a contrary case study might not mean much. “Yes, it’s an unsettled situation, so not every instance fits into the overall mental model I’ve proposed.”

    Undermine credibility. That’s all. Nothing science-wise.

    If these are the claims, a contrary example might help us in our thinking. Perhaps the situation is more nuanced than we first thought.

    Maybe Tiljander-in-Mann08 is a unique example. Maybe it’s not.

    And most of the participants in this conversation seem to hold to the idea that Tiljander-in-Mann08 isn’t an example of anything of the sort. Recall, upthread, Jim Bouldin labelled Mann08 strong science.

    > Using the topicality of LAT to push technical points can be seen as a way to shift the focus to a very specific instance.

    You quoted me explaining why I think Tiljander-in-Mann08 is a worthwhile focus. And if we use the Cocktail Party Model and engage in a dozen or so free-flowing conversations that range from shoes to ships to sealing wax — well, we might never arrive at an understanding of the key points of any one topic. To say nothing of agreement.

    At any rate. The discussion has been sufficiently focused and careful to help me explore whether I have overlooked factors that remedy the seemingly-severe shortcomings of Mann08.

    I recognize, too, that cocktail parties are typically more fun than journal clubs. And climate-blog-commenting is a hobby, for all of us. It shouldn’t be a chore. So I thank you, Deech, Dan Olner, Jim Bouldin, J Bowers, and the folks upthread for engaging with me (and with MikeN). I hope we each gained something from these interactions.

  242. MIkeN Says:

    I agree discussion of Salazar probably goes off topic. I brought up Tiljander because it seems like when an error is pointed out, climate science defends the error rather than correct it, leading to a loss of credibility.

  243. MIkeN Says:

    Dan Olner, there are alternatives that Steve McIntyre has uncovered for some of the most hockey-stick proxies. Polar Urals update that was never published by Briffa but has been used, shows a correlation to modern temperatures, but a larger medieval warm period. Instead, people use Yamal that shows a smaller medieval warm period and a very large modern warming, helped along by a single tree that is unusually high. Then Sheep Mountain has an updated version produced by a PhD student, which does not show the same results of high bristlecone warming.

  244. danolner Says:

    MikeN: “there are alternatives that Steve McIntyre has uncovered…”

    Which is potentially good, of course: finding new data, adding to the analysis, great. But you said: “scientists find bristlecones very convenient for creating hockey sticks”. So I’m still not clear if you’re saying they’re cherrypicking data to produce the result they want?

    To re-iterate, nothing wrong with using alternatives or adding new analyses. But claiming that they’re better than other analyses because “scientists find x very convenient for creating hockey sticks?” You’re going to have to justify that rather more. Also, if that’s true, what makes the following false (if anything): “McIntyre and others find y convenient for *avoiding* hockey-sticks?”

  245. PeteB Says:

    Thanks all for the discussion,

    My position (FWIW) is that I don’t think the Tiljander proxies can be directly calibrated to the instrumental record so when looking at Mann08 it is better to look at the sets of data which don’t contain them.

    When comparing all proxies versus all proxies “- 7 bad series” as they are called in the supplemental information spreadsheet it makes no difference.

    When comparing non-dendro versus non dendro – “7 bad series” it does make a small difference as they acknowledge in the supplemental information

    http://www.meteo.psu.edu/~mann/supplements/MultiproxySpatial09/

    Update 22 Aug 2010: Additional significance tests that we have performed indicate that the NH land+ocean Had reconstruction with all tree-ring data and 7 potential “problem” proxies removed (see original Supp Info where this reconstruction is shown) yields a reconstruction that passes RE at just below the 95% level (approximately 94% level) back to AD 1300 and the 90% level back to AD 1100 (they pass CE at similar respective levels). So if one were to set the significant threshold just a bit lower than our rather stringent 95% significant requirement, the reconstruction stands back to AD 1100 with these data withheld. Recent work by Saltzer et al [ Salzer et al, Recent unprecedented tree-ring growth in bristlecone pine at the highest elevations and possible causes, Proc. Nat. Acad. Sci., 2009] suggests there is little reason to withhold tree-ring data however.

    I certainly think it is ‘strong science’.

  246. PeteB Says:

    MikeN,

    Why doesn’t someone submit an analysis using the alternatives to a journal ? Are you sure that the analyses published on climate audit are subject to rigorous and informed scrutiny (I’m not)

  247. Deech56 Says:

    AMac, I too thank you for the discussion. As you can probably guess, I did not intend to veer into a discussion of Salazar09, but was using that paper as an example of how science advances within the peer review literature. I would say that Manno8 is an improvement over MBH99/99 and the next Mann opus will be an improvement over Mann08. Science does progress and each study is a piece of the overall picture. My point was well illustrated by PeteB’s quote from the SI of Mann08.

    I also believe that the criticisms from the NAS had a greater influence in subsequent publications than any blog posts did.

  248. Marco Says:

    Guys, it’s Salzer, not Salazer. And some more attentive people will have noted that Salzer and Hughes already had a paper out in 2007 showing good correlation between bristlecone pines and climate events:

    http://www.sciencedirect.com/science/article/pii/S0033589406000998

  249. Deech56 Says:

    Thanks, Marco.

  250. MIkeN Says:

    PeteB, I don’t think these alternatives are any more valid than the originals. The question is whether the original hockey stick studies are convincing.

  251. MIkeN Says:

    >“scientists find bristlecones very convenient for creating hockey sticks”. So I’m still not clear if you’re saying they’re cherrypicking data to produce the result they want?

    If I wanted to produce a hockey stick, I would like to use Yamal, bristlecones, and the like. I could then throw in as many other proxies as I want, with proper processing, I can make the hockey stick show up. After the NAS Panel suggested bristlecones, or at least strip-bark bristlecone proxies, should not be used, they still were used in several papers.

    It is climate scientist themselves who have claimed to be cherry picking.

    “However as we mentioned earlier on the subject of biological growth populations, this does not mean that one could not improve a chronology by reducing the number of series used if the purpose of removing samples is to enhance a desired signal. The ability to pick and choose which samples to use is an advantage unique to dendroclimatology. “

  252. MIkeN Says:

    >“McIntyre and others find y convenient for *avoiding* hockey-sticks?”

    Well yes, if I wanted to avoid a hockey stick, I would use proxies that do not show a hockey stick, and leave out ones that do, or use an algorithm that manages to elevate non-hockey sticks in stature. Tricky to do with warmer temperatures, but I suppose it could be done.
    The question is, there is a Polar Urals that was done later than Yamal(the same area), yet the updated Polar Urals doesn’t get used except for one paper. Then the latest bristlecone series with newer data gets skipped over in place of one with more of a hockey stick shape. I suspect that if the updates had been more in line with the desired result, these would be published and used more. At a certain point, the lack of result is being used to say the proxies are flawed.

  253. danolner Says:

    “It is climate scientist themselves who have claimed to be cherry picking.”

    It took me a while to track down the quote you’ve given. Got it eventually, though along the way I learned –

    a) McIntyre first quoted it, I think, didn’t give a full reference, and did the usual “ooo you’d never get away with this in the mining business” stuff.

    b) The quote gets over a thousand hits, all people saying what you’ve said, none actually looking at the piece of writing as far as I can tell. Cherry-pick? My irony-o-meter has gone off the scale, I’m afraid to say. Basically, they’re stress-testing a particular method, and part of that is a process of working out what’s noise, what’s signal.

    As I say, I’ve had a look at the bit of writing. I’ll leave that for you to find, perhaps when you have we can discuss ways it, perhaps, didn’t mean “we pick whatever little bit we like to support our previously held position.”

    Actually, the more I think about it, the crosser I get: they’ve clearly been headbanging against some hard problems. Same as with Mann, it’s a) covered in caveats and b) actually designed to, as I say, stress-test a given method to try and work out what it can and can’t do. That one sentence should get cherry-picked like that, and then used to accuse them of cherry-picking…? Think I need to go and have a litte sit-down…

    p.s. to save you the bother.

  254. PeteB Says:

    I suspect that if the updates had been more in line with the desired result, these would be published and used more

    That doesn’t ring true to me. Even if you thought (and I don’t !) that there were some scientists that were wedded to the idea of a hockey stick and were making choices of methodology or proxies to reinforce those results. Surely there are some young scientists coming up, who would really make a name for themselves by doing an alternative reconstruction, using alternative proxies that got a very different result. In over a decade there seems to be none forthcoming. Journals have a bias to publish new and interesting results.

  255. AMac Says:

    At numerous points in this thread, I’ve brought up the severe problems that result from the decision to use post-hoc analytical methods. In that regard, the text quoted by MikeN at 19:44 may resemble a Kinsley gaffe (“an occurrence of someone telling the truth by accident”).

    Here is the citation and the full context for that excerpt.

    Esper, J., Cook, E.R., Krusic, P.J., Peters, K., Schweingruber, F.H. 2003. “Tests of the RCS method for preserving low-frequency variability in long tree-ring chronologies.” Tree-Ring Research 59(2):81-98. Scanned PDF available here; most of the quoted text is on PDF pg 42 / printed pg 92.

    Before venturing into the subject of sample depth and chronology quality, we state from the beginning, “more is always better”. However, as we mentioned earlier on the subject of biological-growth populations, this does not mean that one could not improve a chronology by reducing the number of series used if the purpose of removing samples is to enhance a desired signal. The ability to pick and choose which samples to use is an advantage unique to dendroclimatology. That said, it begs the question, how low can we go?

  256. AMac Says:

    Dan Olner —

    I missed your comment at 20:23 prior to posting at 20:36.

    You are correct in that Esper et al. are discussing the validity of their “RCS” method.

    My concerns about the acceptance of post-hoc methods remains. See the Discussion (PDF pg 46-47 / printed pg 96-97). For instance,

    When using RCS one has to decide whether differences between subsamples reflect (i) climatic signals and should be preserved or (ii) non-climatic signals and should be eliminated.

    Imagine making an argument along these lines to a group of biostatisticians who were designing a pivotal clinical trial. You would be met with shocked silence, I’d wager.

  257. danolner Says:

    That’d be why dendro isn’t used isn’t clinical trials then, huh? Seriously: dendro is clearly a messy, complex process of attempting to tease climate signals out from a mass of overlapping statistical problems. The kind of strictures required in clinical stats is not a useful comparison.

  258. willard Says:

    AMac,

    It seems to me that commenting on threads should not be a chore to anyone, including you and scientists. Usually I find that comment threads end up because one of the conditions to bring a conversation to a closure has been met. These conditions deserve due diligence.

    The part from the Conversation article you selected contains words that are neither soft nor dull. They contain words that often bring conversation to a closure. On the other hand, this seemed to be the point of the article: it makes no sense to keep arguing with what appear to be cranks.

    In my humble opinion, the harder word “cranks” convey about the same idea than the softer expression “people that assert that they can believe whatever they wish”. I would even be tempted to say that softer expressions are better suited to push the limits of one can say than harder ones, in cocktail parties, more serious conversations or in front of the camera. Notwithstanding the importance of cheerfulness, perhaps:

    http://scientistscitizens.wordpress.com/2010/03/07/case-marc-morano/

    http://scientistscitizens.wordpress.com/2010/03/07/morano-analysis-1/

    http://scientistscitizens.wordpress.com/2010/03/08/morano-analysis-2/

    http://scientistscitizens.wordpress.com/2010/03/10/morano-3-whats-the-issue/

    http://scientistscitizens.wordpress.com/2010/03/11/morano-analysis-4-bringing-the-arguments-home/

    ***

    Gavin Schmidt talked about a certain dynamic that is important for my Neverending Audit, a dynamic that might not interest everyone, which is alright, after all it’s a big world. You appear to have focused on a part in which you find incorrect factual statements. This focus is not unlike the focus about limnological points, which import to sociology of science.

    These two aspects of the dynamic might be of some importance for the technique I am tempted to call **scientific criticism**. They are so commonplace that I might stay for a while describing them “as they are”, to take an expression that echoes Bart’s wording. I believe you already said that you took no interest in commenting on that subject, so I don’t expect you to follow that discussion.

    ***

    In any case, I did find many of your comments to be insightful. Among them, I took the liberty to post this one and to name it **Between Postmodern and Cynical**:

    http://neverendingaudit.tumblr.com/post/7709607897

    I sincerely believe that this provides a simpler, stronger, and more fecund explanation than any analogy, metaphor or parable we could invent. (Not that I find no interest in these images, on the contrary.) This explanation reinforces the idea that L’Affaire Tiljander does not rest on any “simple” question:

    http://deepclimate.org/2010/07/19/open-thread-4/#comment-4879

    I apologize for not having acknowledged in due time your offer to turn that comment into a blog post on your site.

    Thank you for the discussion,

    Farewell,

    w

  259. dhogaza Says:

    Imagine making an argument along these lines to a group of biostatisticians who were designing a pivotal clinical trial. You would be met with shocked silence, I’d wager.

    On the other hand, those trying to understand the early 20th-century spanish flu epidemiology might also be met with shocked silence.

    I mean, how many people designing a clinical trial would plan to disinter Inuit who died decades after the experiment was run?

    You do understand what the word “paleo” means, no?

  260. dhogaza Says:

    It’s interesting to see how people like AMac (and McI and the rest) really rest their case on impossible standards, yet insist that analysis proves that the MWP was much warmer than various hockey stick papers suggest.

    On the one hand … the data and statistical treatment isn’t robust enough to tell us what happened 1000 years ago.

    On the other hand, the same non-robust data is sufficient to tell us that the MWP was as warm or warmer than today.

    Hmmmm.

  261. AMac Says:

    dhogaza (06:02) –

    > people like AMac (and McI and the rest)… insist that analysis proves that the MWP was much warmer than various hockey stick papers suggest.

    You’ve made a claim about what I think about the MWP. My claim is that many people aren’t very skilled at paraphrasing.

    Since you think this is worth discussing, please supply a direct quote or a link.

  262. danolner Says:

    Amac: well, I think the interesting point from dhogaza was the bit you left out of your quote: again – “It’s interesting to see how people like AMac (and McI and the rest) really rest their case on impossible standards…”

    The impossible standard, in this case, is you saying “Imagine making an argument along these lines to a group of biostatisticians who were designing a pivotal clinical trial. You would be met with shocked silence, I’d wager.” Which is what McInytre does when he says “ooo you wouldn’t get away with this in mining.”

    Why would you be using clinical trial methods for dendro, or even comparing to them? How is it in any way similiar? Clinical trials are designed, controlled experiments generally with a p<0.01 standard, because people lives depend on them. The dendro people are trying to tease signal out of a very complex, multilayered problem, where just working out roughly what the error level is equals a big achievement. We probably wouldn't want anyone's life to depend on it, but that doesn't make the task any less worthwhile, does it?

    So for me, anyway: I'd like to know more on why (or indeed *how*) you think dendros should be doing something able to meet the approval of clinical trial people? Why make that comparison?

    The quote you gave is, again, a central problem for dendro: figuring out ways to separate climatic from non-climatic signals. What exactly is the problem with that? It's in the context of a much larger effort to detect a climate signal. It can, of course, be cherry-picked and interpreted as "look, dendros pick the records to suit their pre-determined idea of what the climate record should be" – but you don't think that, do you?

  263. Deech56 Says:

    AMac, there’s a big difference between how one approaches prospective and retrospective studies. In my real life, I work with biostatisticians on prospective non-clinical efficacy studies (the equivalent of clinical trials, with interim analysis and all that). The first thing one learns is that even statistics is open to interpretation – when to use one-sided tests, adding arms and something about Bonferroni.

    The study of paleoclimate is much like paleontology, where one has to go with the available data, and epidemiology, where there are multiple factors to consider. I’ve spent my scientific life working with living systems, and have come to appreciate the nuances and uncertainties.

    I believe that what those who study paleoclimate are doing is developing the statistical tools that can give them the most accurate signal from proxies. They have to try out different techniques to find out what works – that’s why Mann08 and the SI have the various different types of analysis. If someone believes that the goal is hockey sticks, then there’s not much point in arguing, but if one wants to weigh in on which tools work or don’t work and why, that’s another story. Unfortunately, much of what is written in blogs does not achieve the same visibility as a published paper or comment.

  264. Adder Says:

    Amac:

    “For instance,

    When using RCS one has to decide whether differences between subsamples reflect (i) climatic signals and should be preserved or (ii) non-climatic signals and should be eliminated.

    Imagine making an argument along these lines to a group of biostatisticians who were designing a pivotal clinical trial. You would be met with shocked silence, I’d wager.”

    As far as I understand, you have been saying that Tiljander subset should not be used because it cannot be correlated with climate. Now you are saying that leaving out Tiljander would be met with shocked silence by biostatisticians?

  265. AMac Says:

    Adder (11:04) —

    > you have been saying that Tiljander subset should not be used because it cannot be correlated with climate.

    Correct.

    > Now you are saying that leaving out Tiljander would be met with shocked silence by biostatisticians?

    No. I should have been clearer. The inadvertant misuse of the Tiljander data series in Mann08 is a discrete problem that becomes clear upon understanding of Mann08’s methods (requirement for direct calibration), and reading the authority’s cautions (Tiljander03), and examining the evident patterns in the data series themselves. Becomes clear in my opinion I should say, considering the length of this thread.

    Statistical concepts aren’t part of this reasoning.

    In contrast, the issues concerning post hoc analysis are general ones.

  266. AMac Says:

    Dan Olner (9:59) —

    You seem to think that I am under an obligation to tease the least-flawed arguments out of every submission to this thread, irrespective of how flawed its premises are, or how offensive its writer has chosen to be.

  267. danolner Says:

    Amac: well, of course, you’re not. You’re right, you’ve made no claim (that I’ve seen) about the MWP – but as I say, you have just made some claims about dendro vs clinical trials, which I asked some questions about.

  268. AMac Says:

    Dan Olner (9:59) and Deech56 (11:02) —

    “What McIntyre says” is of no great interest to me, except insofar as his points are perceptive. Likewise for Schmidt, Verheggen — and you. I suggest you take that up at his blog.

    You seem to be suggesting that there is one set of statistical principles that govern experiments where people’s lives are at stake, and another where tree-ring time series are the subjects of analysis. I do not understand the basis of this apparent claim. If people’s lives did ride on the results of paleo recons, would the statistical approach to the problem have to change?

    As to “what dendros could do differently,” that has come up in the literature. E.g. in a fairly recent paper by von Storch, I believe — but I don’t have a citation to hand. I offered my own suggestion in remarks to Ron Broberg, repeated here to Jim Bouldin at my blog. This line of attack would be tedious. Supplanting post-hoc analysis in certain clinical trials was also an unappetizing prospect, which was one reason it took decades. See link to remarks by Norman Wolmark in the linked comment.

    The problem with post-hoc analysis can be thought of as succumbing to a subtle temptation to calculate error bars that are unrealistically tight. In other words, there are fundamental assumptions about hypothesis-testing that are built into the idea of most of the calculations that go into determining significance or P-value. For the results to be valid, these assumptions have to be made explicit, and accounted for. To me, the Esper quote shows that this is not the case in the dendro field as a whole. Jim Bouldin’s response to the linked comment reinforced that impression.

    The introductory section of the Wikipedia entry on post-hoc testing has an edge of frustration to it, but I think it gives a good outline of the issue. If one insists on post-hoc approaches — as Esper and others apparently do — the next question should be about how to calculate uncertainties. The Bonferroni correction? Another approach? As I’m not a statistician, I won’t be of any help there.

    Anyway — post-hoc analysis opens up a whole new issue. At this point in the thread, I’m not up for a second protracted discussion. You considered my remarks and responded sensibly, for which I thank you.

  269. Deech56 Says:

    AMac, I am not saying that there are different statistical principles, but that there are different statistical tools for the different studies. I would think that the proper comparison for paleoclimate stats isn’t a prospective clinical trial, but a retrospective study such as a case control epidemiological study. Of course, I am not a statician either, so I can’t really get into this into any meaningful detail (and it is veering OT).

  270. willard Says:

    Taking stock, let’s see how L’Affaire Tiljander came to be introduced into the thread. All times should be in EDT, the officiel auditing timezone. All miscalculations are mine.

    On the 2011-06-23, at 07:10 EDT, Jeff Id suggests the neologism “paleomathmashology”. At 9:39, Jeff Id blames his grumpiness on some amount of paleo papers. Note the plural.

    On the 2010-06-23, at 14:04, Chris Colose tries to explain in a long and somewhat general comment why contrarians focus on paleoclimate community. At 15:50, Bart refuses to snip Id and introduces the abreviation “c/a/e” for conspirational/accusatory/exagerrated.

    At 17:30, Deech56 is being a smartass. At 18:18, Dan Olner recalls an old story at Tony’s. At 19:47, Bart +1s Olner.

    On the 2010-06-24, at 2:37, MikeN responds to Colose in a comment worth repeating:

    > Chris, the hockey stick was very significant to AGW. That’s why it is used in so many press releases and such. The visual is needed to get political results.

    http://ourchangingclimate.wordpress.com/2011/06/22/climate-science-scientific-method-skeptics-not/#comment-13405

    At 3:28, Chris Colose responds to MikeN.

    On the 2010-06-24, at 12:29, intrepid_wanders interjects that scare tactics go both ways. At 13:58, Marco reports spam. At 16:15, Bart +1s Colose, corrects intrepid_wanders and thanks Marco. At 21:38, Eli Rabett parks links to his “rejectionist” post.

    On the night of the 2010-06-25, MikeN and Chris exchange ideas, Paul Kelly returns to his usual comment with different words, and in the morning Steve Bloom piles on MikeN.

    Then it gets even more interesting.

    On the 2010-06-25, at 7:31, Moshpit thinks of a counterexample of self-correction in science while admitting it’s generally true:

    > Even AFTER it was debunked Piltdown Man was referred to in an article in Nature. [...] Generally true, but the exceptions are fascinating sociology

    http://ourchangingclimate.wordpress.com/2011/06/22/climate-science-scientific-method-skeptics-not/#comment-13417

    This was to counter this quote from The Conversation:

    > [T]his is what makes science self-correcting. If arguments turn out to be wrong, in time they are caught and corrected by other scientists. It is virtually impossible to publish long-refuted nonsense in good peer-reviewed journals.

    Notice the wiggling room that provides the expression “long-refuted nonsense”.

    The problem now is to spring from Pitdown Man to Tiljander mud-wrestling. A neat way to solve this problem is to ask a Jeopardy question, as MikeN did on the 2011-06-26, at 5:18:

    > Steve Mosher, I assume you are referring to Piltdown Man, and not a mocking reference to Piltdown Mann reusing Tiljander in Science?

    http://ourchangingclimate.wordpress.com/2011/06/22/climate-science-scientific-method-skeptics-not/#comment-13420

    On the 2011-06-26, at 5:22, MikeN follows on his idea without mentioning explicitely “Tiljander”:

    > Indeed they are. All the more reason to distance yourself from the errors of others. To this point I have not seen members of these groups call out Mann for using data upside-down. Martin Vermeer acknowledged it, but then says Mann was correct in his PNAS reply to McIntyre. Instead Mann is more likely to have his 2008 paper with the upside-down data referred to by the members of those other groups as reinforcing the hockey stick.

    http://ourchangingclimate.wordpress.com/2011/06/22/climate-science-scientific-method-skeptics-not/#comment-13421

    It did not take long to shift from Man to Mann.

    It took four minutes.

    All we now need is someone to call MikeN’s trick. This time, that is on the 2011-06-26, at 7:09, it was Dan Olner:

    > C’mon! Google, MikeN, and come back and tell us why that “Mann got it upside down” stuff is wrong.

    http://ourchangingclimate.wordpress.com/2011/06/22/climate-science-scientific-method-skeptics-not/#comment-13422

    The rest is anotherstory.

    ***

    AMac’s first comment in the thread was on the 2011-06-30, at 19:17.

    Notice this comment on AMac’s blog, posted the 2011-06-30, at 2:19:

    > Bart Verheggen has a post on climate science, and it got into a discussion of Tiljander.

    http://amac1.blogspot.com/2011/06/voldemorts-question.html?showComment=1309414792548#c2235922335859586387

    The reader will never guess who’s the author of this comment.

  271. Chris Colose Says:

    willard,

    The problem I have is that this is no longer a scientific debate or an “audit.” It’s also not an issue of Mann/Steve M being right or wrong.

    I also think that ~99% of the people who choose to comment on this issue lack the statistical background or have not went through the fine details of the many proxy studies to really comment confidently about it (I am, admittedly, in this crowd as well). The arguments have become very nuanced and technical, and generally boil down to most people agreeing with whoever they decided has the better position about something completely irrelevant, like whether CO2 causes warming or what climate sensitivity is. The whole thing is used primarily a “gotcha” tool and as some straw man argument against the foundation of anthropogenic climate change; it justifies countless people living in an odd Groundhog’s day repetitive version of how conspiracy science operates, instead of learning something new and interesting and placing it within the context of the big picture . It is not an interesting argument in the scientific literature and quite honestly, the attention it gets it makes no sense.

    So to be blunt, aside from maybe a few specialists, no one actually cares whether Mann used Tiljander properly or not.

  272. mikep Says:

    “So to be blunt, aside from maybe a few specialists, no one actually cares whether Mann used Tiljander properly or not.” The attitude displayed by this quote (and the whole comment) is exactly what makes me sceptical. There is nothing difficult about the misuse of the Tiljander proxies but apparently serious people will still defend the misuse or say its too difficult to understand. The “unprecedented in 1000 years” argument about global temperature have, whether wisely or not, been extensively used in public debate. The evidence for the claim looks, at best, weak and depends a limited set of dubious proxies. When attempts to examine the evidence are just brushed as because “no-one is interested” I get annoyed (and more sceptical).

  273. Deech56 Says:

    > At 17:30, Deech56 is being a smartass.

    It’s a tough job. but someone’s got to do it. ;-)

  274. Deech56 Says:

    Mikep, a major controversy is something that generates a number of papers containing claims and counterclaims. The mud wrestling, as willard so aptly put it, is basically a blog discussion, important not for the science itself, but as a proxy for the larger battle.

  275. AMac Says:

    Deech56 —

    > The mud wrestling [over Tiljander-in-Mann08 is] important not for the science itself, but as a proxy for the larger battle.

    A question. Is it appropriate to discuss the findings of peer-reviewed articles in blog comments, with the focus on the methods, reasoning, and conclusions of the papers?

    Well, the world being a big place and all, I guess we’d all agree on “yes, sure, go for it, why not?” For some people, it’s as interesting as Justin Bieber’s new haircut (try +”justin bieber” +”new haircut” on Google, heh).

    So, a more focused question. Is the Comments section of this post at Bart’s climate-science-themed blog an appropriate venue for a discussion of Mann08’s strengths and weaknesses?

    That gets to what’s “important”.

    At Zap2It, dishing on stylin’ Justin is what matters.

    As stated upthread, I think scientifically-literate laypeople can sift through a paper like Mann08 and can reach conclusions as to whether it fulfilled the claims in its Abstract. And I think that’s relevant to the argument from ‘The Conversation’ in the body of Bart’s post.

    But is it “important?”

    Where one stands on “the larger battle” seems to map tightly to where one stands on Tiljander-in-Mann08. Unfortunately. I’ll sketch three stances.

    1. The debate on climate change has been over for decades. It is time to accept the scientific consensus and move on to implementing the policies that will help society ameliorate the problems that climate change is causing, and will cause.

    2. The fundamentals of human-caused climate change are well-established. But mainstream scientists are overconfident in their understanding of magnitude of these changes, and the historical context of these changes. Strong science is a prerequisite for wise policy, and climate science is immature.

    3. Human-caused global warming is an unproven theory that is being advanced by scientists and politicians as a means to various ends. These range from career advancement to atoning for Western Guilt to deep-seated belief in the moral superiority of communitarian policies.

    Thus, another question. How do you define “the larger battle”?

  276. AMac Says:

    In my opinion, there’s also an ongoing “smaller battle” among the commentariat of climate-science-themed blogs.

    That’s between the Journal Club model and the Cocktail Party model of conversation in the comments following blog posts.

    Examples on request :-)

    Cocktail Party is easier and, I think, more fun for most people. That puts me in the minority.

  277. Deech56 Says:

    AMac, I think you are very much in journal club mode, and that’s to your credit, but having participated in various journal clubs, I can attest that all our discussions didn’t change the scientific literature.

    Not every datum in every paper is perfect, but science still advances. Every paper has strengths and weaknesses, especially for cutting edge research. Regarding Mann08, if the mud is taken out, the results still hold (if one accepts tree-ring data), so I guess our differences revolve around whether we accept the tree ring data. I feel that the issues brought up by the NAS were addressed by Salzer and Hughes. To me, this is how science works.

    But my concerns about the effects of continued increases in CO2 have little to do with climate over the last 2000 years and more to do with the sensitivity of the climate to CO2 increases, and the effects of these temperature increases (based on studies that go back even further).

    By “larger battle” I mean the climate wars, and I have concerns about the attacks that climate science is and scientists are enduring. I would probably be in category 1 in your list.

  278. grypo Says:

    Category one isn’t really valid unless you use the word “risk”. It is the risk that is undeniable and this is the best that predictive science can tell us. The rest is up to our collective understanding of the time scales and long term delayed effects of CO2 emissions and how that interfaces with our collective value system.

    The problem with Tiljander is that it is used as way to prove category 3 is reality, which it doesn’t and isn’t. Whether or not people intend this is difficult to ascertain.

  279. AMac Says:

    grypo —

    > Category one isn’t really valid unless you use the word “risk”.

    Well, you think so. But here’s the text I used as the basis for category 1 (see the main post).

    The so-called “debate” on climate change has been over for decades in the peer-reviewed literature. It is time to accept the scientific consensus and move on, and to stop giving air-time to the cranks.

    It is time for accountability.

    > The problem with Tiljander is that it is used as way to prove category 3 is reality

    No anecdote is going to prove an argument of that breadth, whatever the cheer squad at WUWT might wish to believe. Reality often comes in shades of gray. “Merchants of doubt” have a stake in the outcome. Climate scientists aren’t all members of a detached and objective priestly caste. They can be expected to have their own interests, too.

    I’d turn what you say around: the problem is that some advocates feel they must defend Mann08, because they fear that acknowledgment of its weak science will be perceived as strengthening the Category 3 argument.

    If there were strong technical rebuttals to the specific criticisms leveled at Mann08, they would have been made, by now.

  280. PeteB Says:

    Amac,

    Where one stands on “the larger battle” seems to map tightly to where one stands on Tiljander-in-Mann08. Unfortunately. I’ll sketch three stances

    But everyone that I have found from the ‘consensus’ side that has looked into it seems to have the same viewpoint as you on the technical issue..

    e.g.

    Martin said

    AMac, I see now that I was ambiguous. What I meant was that the understanding, or interpretation, by Tiljander et al, on what was the proper orientation of these proxies, was likely correct.

    Mann et al. of course assumed all lake sediment proxies to be oriented ‘upward’. The paper says so, and — if what MikeN asserts, that 4000 is the code for this proxy type, is true — so does the code.

    you replied:

    Martin Vermeer, thanks for the clarification.

    Here is what I understand you to be saying in the just-prior comment.

    * Tiljander03 was likely correct in their interpretation of Lightsum and XRD: that higher values correspond to lower temperatures.

    * Mann08 assumed all the Tiljander proxies were oriented ‘upward.’ Thus, Mann08 interpreted Lightsum and XRD in this fashion: that higher values correspond to higher temperatures.

    Is this paraphrasing correct?

    Martin confirmed

    Yep.

    you say:

    Martin, your clarity is much appreciated. Ari Jokimäki appears to have arrived that this conclusion, as well.

    Andy Russell

    Ok, that sounds pretty convincing. Thanks for the info.

    Why did McIntyre not focus on these issues instead of the “upside down” thing, which seems to be a criticism of the method of incorporating the data rather than not using it at all?

    Deep Climate

    Do I think the proxies can be usefully calibrated to the instrumental record? No.</quote?

    http://deepclimate.org/2010/07/19/open-thread-4/#comment-4879

    Me

    My position (FWIW) is that I don’t think the Tiljander proxies can be directly calibrated to the instrumental record so when looking at Mann08 it is better to look at the sets of data which don’t contain them.

    Even Michael Mann on his last update to the SI isn’t attempting to defend the ‘problematic’ proxies, just emphasising the results not using the problematic proxies and that
    …there is little reason to withhold tree-ring data…

  281. PeteB Says:

    All 3 “stances” seem pretty hopeless to me

    Let me give my stance :

    I actually think the more the uncertainty (particularly if very high values of climate sensitivity are plausible) the stronger the policy case. It would be more harmful to delay action if there is a real chance of 6 Deg C climate sensitivity (which personally I don’t think is plausible) than to take action and find it is less than 2 Deg C.

    I actually don’t think there is much chance of very high climate sensitivities and would work on the 2 – 4.5 Deg C range. That is fine from a policy perspective, you can work out the economically optimum policy and fine tune it if and when we can narrow the range and better judge the impacts.

  282. AMac Says:

    PeteB (17:11) —

    You’re right, concerning Andy Russell, Deep Climate, and Martin Vermeer conceding the uncalibratibility of Tiljander’s data series to temperature.

    The problem is this: disqualify the Tiljander proxies, and the no-treering reconstruction isn’t represented by the Blue line in Mann09 SI Fig S8, here. Instead, it is represented by the very-different Green line (which fails Mann08/09’s validation test prior to 1500).

    In other words, if you think the dendro-based recons are valid, this shows that the no-dendro recons are invalid.

    There might, possibly, be a reasoned argument that goes, “Tiljander is uncalibratable, and the main conclusions of Mann08 still hold”.

    To my knowledge, nobody has advanced such an argument. In this thread, in the peer-reviewed literature, or anywhere else.

  283. grypo Says:

    Yes, I think the Conversation article misses that crucial point. By repeating it, they only replay an often heard skeptic meme, that all scientists believe the science is settled, and we can no longer discuss it. It is important to mention what is “settled” and why having a good idea about the risk is the time to collectively discuss how to best avoid it.

  284. PeteB Says:

    Amac – but isn’t that the exact argument that Mann is making in his Aug 2010 update to the supplementary Information – he concedes than non dendro minus problem proxies is not significant at 95% level before 1500, but he still argues

    1) It is significant at a slightly lower level before that date.
    2) That more recent work suggest there is little reason to withhold tree ring data.

    http://www.meteo.psu.edu/~mann/supplements/MultiproxySpatial09/

    Update 22 Aug 2010: Additional significance tests that we have performed indicate that the NH land+ocean Had reconstruction with all tree-ring data and 7 potential “problem” proxies removed (see original Supp Info where this reconstruction is shown) yields a reconstruction that passes RE at just below the 95% level (approximately 94% level) back to AD 1300 and the 90% level back to AD 1100 (they pass CE at similar respective levels). So if one were to set the significant threshold just a bit lower than our rather stringent 95% significant requirement, the reconstruction stands back to AD 1100 with these data withheld. Recent work by Saltzer et al [ Salzer et al, Recent unprecedented tree-ring growth in bristlecone pine at the highest elevations and possible causes, Proc. Nat. Acad. Sci., 2009] suggests there is little reason to withhold tree-ring data however.

  285. AMac Says:

    PeteB (17:50) —

    > but isn’t that the exact argument that Mann is making in his Aug 2010 update to the supplementary Information

    We have to pay attention to indefinite articles. What, exactly, is “that” referring to?

    My argument is: The Tiljander data series are uncalibratable by direct methods to temperature, and thus their use in Mann08 is invalid. With Tiljander removed from the no-dendro recons, those no longer resemble the dendro recons. One of Mann08’s central points is that recons built from non-dendro proxies support the validity of recons built from treering proxies. Without Tiljander, this key assertion fails.

    Is that equivalent to what the text quoted from the non-peer-reviewed August 2010 update to the Mann09 Science paper’s SI’s update says, as far as Mann08?

    I would venture, “No.”

  286. PeteB Says:

    Amac,

    OK – looking at this

    could you clarify what the red line is please ?

  287. AMac Says:

    PeteB —

    The image you link is Mann09 SI Fig S8, where I have overwritten the faint Green dashed line prior to 1500 with a continuous Green line, for improved visibility. The legend, reproduced here, is

    Figure S8: Sensitivity of NH mean reconstruction to exclusion of selected proxy record. Reconstructions are shown based on “all proxy” network (red, with two standard error region shown in yellow) proxy network with all tree-ring records removed (blue), proxy network with a group of 7 long-term proxy with greater uncertainties and/or potential biases as discussed in ref. S1 (brown) and both tree-ring data and the group of 7 records removed (green; dashed before AD 1500 indicates reconstruction no longer passes validation).

    So the Red line is the NH recon based on the “all proxy” network. In shorthand, Red is “yes-dendro/yes-Tilj.”

    If you think yes-dendro/yes-Tilj to yes-dendro/no-Tilj is a useful comparison, that’s Red to Brown.

    (That’s the same as Blue to Red, in the color scheme that Prof. Mann used in the June 23rd Update 2 of the Kemp11 post at RealClimate, reproduced at the linked blog post.)

  288. PeteB Says:

    OK thanks,

    Although it confuses me a bit !

    So taking out Tiji from the non-dendro reconstruction warms the little ice age !

    I thought the point was that the calibration reversed Tiji so the effect of including Tiji would warm the Little Ice Age and cool the MWP

    OK, comparing the effect on Tiji on non-dendro reconstruction it agrees pretty well to 1300 ish, but the non-tiji is a bit warmer 1100-1300

    1300-1500 is very close

    1500 onwards taking Tiji out ends up with a slightly warmer reconstruction

    Interesting with Jim’s comments that he thought the warmest points were about 950 and 1350

    I’d actually never realised how close the dendro/non-dendro ‘average’ reconstructions were (with the problematic reconstructions included). I suspect this is a bit misleading given the structural uncertainties involved

    Just out of interest – what points do you take away from this graph

  289. Jim Bouldin Says:

    Well, a phenomenal amount of electrons spilled since I last checked in. Not an ice cube’s chance in the melting arctic of responding to it…

    but PeteB I do want to respond, to clarify what I meant with the reference to 950 and 1350 being the warmest points–possibly warmer than the present. This statement was based on the NH, land only, EIV reconstruction–the one that gave the warmest medieval temperatures. Only if that reconstruction is accurate does my statement hold, and it is based on something that requires a little explanation and reference to the concepts described in Loehle, 2009. This has to do with the possibility of an asymmetric, unimodal response to T overall, but a positive, roughly linear response during the calibration period. If so, then if you track T back in time from the 20th century, the first point where you get T values that might equal the present, is around 1400. If the response is steeply negative on the right side of the mode–which is biologically quite possible–then the steep dip that you see around 1350, holds the possibility of actually being an extreme *warm* period (the opposite of the sharp cooling that it appears as). However, it could also be exactly as it appears–a cooler period. This is because, once you reach the point where the mode occurs, there is no way to discern between which of these two possibilities is the most likely one. The same thing is true for around year 950, and there is also the possibility of extended warmth of uncertain magnitude, over that entire interval, but again, impossible to say.

    However, as I said, this is only if the reconstruction mentioned is the most accurate one. If any of the others presented are more likely, then this won’t hold. You have to find a point in time in the past where the *apparent* temperature–as indicated by the smoothed reconstruction–is similar to the present (more precisely, to the late 1980s early 90s, when the ring series used, ended). This may occur at some other time points for other reconstructions, or not at all in others.

    Make sense?

    The above is the explanation of why it is actually Loehle who has, by far, the most meaningful things to say on the whole tree ring reconstruction topic.

  290. AMac Says:

    PeteB (20:13) —

    > what points do you take away from this graph

    The Blue line and the Red lines are bogus. They are built with proxies that we know are completely wrong (the uncalibratable Tiljander data series).

    The possibly-valid “all-proxy” recon would be yes-dendro/no-Tilj, which is Brown.

    So if you want to compare treering-only with non-dendro only to see how the two classes of proxies look compared to each other, you can’t. There’s no treering-only recon shown.

    However, treering-only “validates” back many centuries (I can’t recall its stopping point). For most of that time, it is going to look a lot like the Brown line, because most of the proxies that Brown is built on are treerings.

    Brown doesn’t look all that similar to Green, to my eyes.

    The strange thing about all this is what happens when you add the bogus Tiljander proxies to other “good” proxies. (I’m not claiming that they are good proxies for temperature — but that seems to be what everyone else believes.)

    Add bogus Tiljander to the Green recon, and you get the Blue line. Blue looks pretty close to Brown. And it “validates” back to 700.

    Add bogus Tiljander to the Brown recon, and you get the Red recon. Except for 800 to 900, Red looks almost identical to Brown.

    These are weird goings-on. In the first case (Green -> Blue), adding in bad data seems to improve the recon?! In the second case (Brown -> Red), adding in bad data seems to have little effect.

  291. Jim Bouldin Says:

    And I should add that it is exactly this problem that I am trying to solve–by looking for a richer (i.e. multivariate) signal in the response variables (the rings, which traditionally have been analyzed as a univariate response), which might have the ability to discriminate between the two states mentioned above. If so, then it will be possible to say something, statistically, about the likelihood of actually being on one side or the other of the mode, and hence, whether it was actually warming, vs cooling, for any given slope direction of the smoothed reconstruction. As it stands, you can’t. Note that this a completely different application of multivariate analysis than has been used up until now, which involves the extraction of principle components from multiple sites over multiple years. This is the extraction components of variation from multiple rings over mulitple years, at single sites.

    It is also possible that the method will not have the ability to discriminate. It is unknown how effective it will be–that depends on intricacies of nature of the tree ring response to climate, on a site by site basis.

  292. PeteB Says:

    Jim,

    Yes,thanks, makes sense – fascinating stuff. OK, memo to myself, I will read the Loehle paper.

  293. PeteB Says:

    Amac,

    OK – seems fair points to me (but what do I know!)

    The ‘average’ dendro / non dendro (including problematical proxies) seem very close given the structural uncertainties involved. I actually find brown – green more convincing !

  294. Jim Bouldin Says:

    I should clarify further my comment at 21:05

    I said: “This is because, once you reach the point where the mode occurs, there is no way to discern between which of these two possibilities is the most likely one.” The assumption there is that you *may* have reached a mode (the apex of a unimodal response) at point where the apparent reconstructed T level is equal to the present (or recent past)–and at that point you lose predictability of what the curve before that point represents. (Also, and as a side, if you have, this has implications for the current state, with specific reference to the divergence effect). Hope that helps.

  295. Jim Bouldin Says:

    “If there were strong technical rebuttals to the specific criticisms leveled at Mann08, they would have been made, by now.”

    They have been made. The problem is that you are either not listening to them, or you don’t understand them. How many times do we have to go around on this ferris wheel before you do? Whether you realize it or not, you are providing a classic example of what Bart was trying to get at with this post. You want to practice science by harping on supposed fatal errors, when they are in fact not fatal WHEN YOU LOOK AT THE TOTALITY OF THE EVIDENCE PRESENTED, which you are failing to do.

    I have already replied to your basic argument. It rests on the fact that you have–after accounting for the possibility that some proxies were significant strictly by chance alone–somewhere around 330 proxies, all of which passed either a p value, or correlation, criterion for acceptance into the analysis data set. And many of these had correlations *far above* whatever level was needed for inclusion. This means that they VERY VERY LIKELY are a reflection of temperature in the past. And they are spread throughout the planet, and they represent tree rings, speleotherms, corals, ice cores, and sediments.

    This argument is then met by the completely bogus claim, stemming from Jeff Id, that spurious correlations in the calibration period are what lead to hockey sticks, thus magically sweeping away with one wave of the hand the validity of these 330 proxies. 100% BOGUS!!!!

    Or by the similarly bogus red herring that bristlecone pines don’t respond to temperature mentioned I think by MikeN at some point, generalizing wildly from the issues with strip barked specimens, to the species as a whole. Well, I can tell you that THERE IS NOT A TREE SPECIES ON THE PLANET THAT DOES NOT RESPOND TO TEMPERATURE! That would be a biological, evolutionary impossibility.

    Yeah, if you remove enough proxies from the mix, castrating the power of the full data set, you can find some proxies that differ wildly in their signal from some other small set thereof. And my response to that is, to quote Miles Davis….wait for it..

    SO WHAT?

  296. Jim Bouldin Says:

    mikep at 9:30:

    It is unfortunate that these kinds of “warmer than any x year period since year y” statements get made, IMO. They don’t really help anything. I for one could not care less whether it was warmer in the medieval period than now–maybe it really was (and as stated above, I think the one reconstruction line lends some possibility to it). We shouldn’t be doing paleoclimate in order to make such statements–I mean go back to the Mesozoic and it was a lot warmer. So what? We need to do these reconstructions with a goal of improving our mechanistic understanding of the forcings, and the natural variability, oscillations etc., operating in the past, with the ultimate goal of improving our physical models.

  297. PeteB Says:

    Jim,

    just one question, I can think of a mechanism for a sudden cooling period (volcanic), but not a sudden warming -can you speculate on why we could get a sudden warming ?

  298. Jim Bouldin Says:

    Pete,

    No, I can’t, and I agree that a very sudden cooling is more plausible than a very sudden warming. Couple of points though. First, the steepness of the curve slope does not necessarily indicate that the T rose rapidly. It might have risen only marginally–but did so at that range of temperatures at which the rings respond most drastically, and negatively. That is, the response function is very steep right there. But since a linear model between T and ring width over the calib. period is used, it makes it look like temps changed greatly. They may not have–but the rings did in response. There is good evidence for such responses in biology.

    The second point is that we’re apparently looking at more decadal-scale changes, roughly (although these are smoothed curves, so hard to say). It wouldn’t have to have been an instantaneous warming. A volcano might provide a fast change, but unlikely a multi-decadal one.

    And of course, as stated, it could actually have been a cooling. We don’t know, given the stated issue.

  299. AMac Says:

    Jim Bouldin (22:02) —

    > They [strong technical rebuttals to the specific criticisms leveled at Mann08] have been made. The problem is that you are either not listening to them, or you don’t understand them. How many times do we have to go around on this ferris wheel before you do?

    You are making arguments. But they are not responsive to any of the points I have raised. Writing in capital letters does not make it so.

    If we move to “Journal Club mode,” I can demonstrate this. A strong technical rebuttal would address one or more of the claims I have made, as far as facts or reasoning. I outlined those claims as nine ideas, H1 through H9, in my comment of July 15 at 20:02.

    > LOOK AT THE TOTALITY OF THE EVIDENCE PRESENTED

    Yes, we should look at the whole picture as well as at particulars, with Mann08 and every other paper we discuss. For what it’s worth, I agree.

    > I have already replied to your basic argument. It rests on the fact that you have–after accounting for the possibility that some proxies were significant strictly by chance alone–somewhere around 330 proxies, all of which passed either a p value, or correlation, criterion for acceptance into the analysis data set. And many of these had correlations *far above* whatever level was needed for inclusion. This means that they VERY VERY LIKELY are a reflection of temperature in the past. And they are spread throughout the planet, and they represent tree rings, speleotherms, corals, ice cores, and sediments.

    If this paragraph is indeed a reply to my basic argument, it must be responsive to one or more of the points labelled H1 through H9 in the above-linked comment.

    Which ones?

    My assessment, for what it’s worth, is none. Instead, you are bringing up other issues. This is welcome. But it’s not a rebuttal.

    I’ll offer some thoughts on the points you raise in a follow-up comment.

    > This argument is then met by the completely bogus claim, stemming from Jeff Id, that spurious correlations in the calibration period are what lead to hockey sticks, thus magically sweeping away with one wave of the hand the validity of these 330 proxies. 100% BOGUS!!!!

    Here is what I wrote upthread (July 23, 2011 at 18:31) —

    My argument is: The Tiljander data series are uncalibratable by direct methods to temperature, and thus their use in Mann08 is invalid. With Tiljander removed from the no-dendro recons, those no longer resemble the dendro recons. One of Mann08′s central points is that recons built from non-dendro proxies support the validity of recons built from treering proxies. Without Tiljander, this key assertion fails.

    A simple visual inspection suffices to show this. Look at the graphics in the post at my blog titled Voldemort’s Question. This was also the topic of an exchange between PeteB and me on this thread, a few hours ago. See my remarks of July 23 at 19:24 and July 23 at 21:09.

    > Or by the similarly bogus red herring that bristlecone pines don’t respond to temperature mentioned I think by MikeN at some point, generalizing wildly from the issues with strip barked specimens, to the species as a whole.

    One of the main reasons Mann08’s authors wrote that paper was to address that criticism. Quoting from page 13253,

    For both methods, we perform reconstructions both with and without dendroclimatic proxies to address any potential sensitivity of our conclusions to issues that have been raised with regard to the reliability of tree-ring data on multicentury time scales (4, 11, 16, 19, 33, 34).

    > Well, I can tell you that THERE IS NOT A TREE SPECIES ON THE PLANET THAT DOES NOT RESPOND TO TEMPERATURE! That would be a biological, evolutionary impossibility.

    Taking that at face value, you would be suggesting that all tree-ring records could serve as temperature proxies. But all parties agree that, at best, a tiny percentage of such records are suited to that purpose.

    > Yeah, if you remove enough proxies from the mix, castrating the power of the full data set, you can find some proxies that differ wildly in their signal from some other small set thereof.

    That may well be case, but it is not an claim I have made. For a summary of my actual argument, refer back to the blockquoted text in this comment that begins, “My argument is”.

  300. AMac Says:

    Jim Bouldin (22:02) —

    I am returning to what you describe as your reply to my basic argument.

    [My reply] rests on the fact that you have–after accounting for the possibility that some proxies were significant strictly by chance alone–somewhere around 330 proxies, all of which passed either a p value, or correlation, criterion for acceptance into the analysis data set. And many of these had correlations *far above* whatever level was needed for inclusion. This means that they VERY VERY LIKELY are a reflection of temperature in the past. And they are spread throughout the planet, and they represent tree rings, speleotherms, corals, ice cores, and sediments.

    This is a recapitulation of what Mann et al wrote in the Methods section of their paper (pg. 13253 col. 2), and in the SI (pg, 2 col. 1). From the SI:

    Although 484 [proxies] (c. 40%) pass the temperature screening process over the full (1850–1995) calibration interval, one would expect that no more than c. 150 (c. 13%) of the proxy series would pass the screening procedure described above by chance alone. This observation indicates that selection bias, although potentially problematic when employing screened predictors… does not appear a significant problem in our case.

    There are potential problems with this approach.

    1. The screening/validation/calibration exercise is limited to the years for which temperature data are available, 1850-1995. For pre-1850 reconstructions, the underlying premise is obviously that the relationships established for the screening period hold throughout the reconstruction period. That only a certain fraction of candidate proxies pass Mann08’s test does not address this concern. This weakness in all proxy-based reconstructions has been highlighted by the much-discussed “Divergence Problem.”

    2. Three of the four Tiljander data series passed screening and validation. (NB, there are actually only two Tiljander data series, not four. And the ten years of Tiljander data from 1986 to 1995 are not real, but synthetic. Let me know if you want to discuss those issues.) This success captures a major shortcoming of Mann08’s procedure: it cannot distinguish meaningful correlations from spurious ones. Statistician J. Udny Yule flagged the challenges of nonsense-correlations in 1926 (PDF). Are all Mann08’s proxies save the Tiljander data series free from nonsense-correlations? I don’t know, as it would take a careful examination to come to an answer. But I wouldn’t take that on a bet.

    3. Even stipulating that proxy selection was well-conceived and correctly executed: that is hardly a validation of the paper’s results. “Necessary but not sufficient” is the applicable concept. Indeed, concerning key questions about the effects of the Tiljander proxies on recons, Mann08 SI Fig. S8 had to be re-done as a result of calculation errors (the version in the SI at pnas.org is incorrect).

  301. mikep Says:

    You might also look at the classic paper by Granger and Newbold

    http://www.climateaudit.info/pdf/others/granger.1974.pdf

    which shows how two random walks will show high correlation far more than is indicated by the simple application of standard t tests. This was the paper was one of the sources of the unit root literature. Basically you can avoid these problems if you check the behaviour of your time series(plural) before attempting to regress them against each other.

  302. J Bowers Says:

    Tamino has a couple of interesting posts on random walks. Perhaps they are relevant.

    * Not a Random Walk
    * Still Not

  303. AMac Says:

    J Bowers —

    Here’s another link to Tamino’s 3/16/10 post Still Not.

    Bart hosted a long and contentious discussion last year on VS’ application of Random Walk dynamics to long-term temperature trends. My impression is that VS was unable to show that his models made sense with respect to the physics of climate. Or that they provide a more useful way of thinking about trends, than do other approaches.

    My other impression is that Demetris Koutsoyiannis and coauthors have gained some traction with their work on applying Hurst-Kolmogorov dynamics to time series in climate. It’s the same general theme: the extent to which ideas such as autocorrelation, red noise, ARMA, and ARIMA aid in interpretation of records such as river-flow or temperature histories. Recent Blackboard post.

    But I don’t have sufficient confidence (heh) in my statistical knowledge to offer a judgement.

  304. J Bowers Says:

    AMac, thanks for the corrected link.

  305. Jim Bouldin Says:

    Amac at 7:44:

    “If this paragraph is indeed a reply to my basic argument, it must be responsive to one or more of the points labelled H1 through H9 in the above-linked comment. Which ones?”

    Really? By “basic” I mean something larger than what you conceive of. I mean your underlying message. I’m interested in the sum of the evidence, not your legalistic approach (obsession actually) with one tiny component thereof. And yes, you have a very legalistic approach to the whole topic, exactly in the mode of McIntyre. This was mentioned upthread, and you denied it; you appear to be completely blind to it. You think that whatever rule set you design (1) is the right one and (2) therefore has to be followed by others. This is one way you demonstrate the difference that Bart’s post is getting at.

    Later in that comment you confuse the issue I raise (over-generalization of the problems with strip barked vs all bristlecones that MikeN introduced, and which is patently false), with the issue that Mann et al was addressing in the paper (i.e. how do the non tree proxies compare with the tree proxies), and you confuse statements I made about what others have said, with saying that you said them.

  306. Jim Bouldin Says:

    at 7:44 again:

    “Taking that at face value, you would be suggesting that all tree-ring records could serve as temperature proxies. But all parties agree that, at best, a tiny percentage of such records are suited to that purpose.”

    Uh no, that’s not what I was getting at, but nice job of flipping the meaning to your ends. My point was that the claim that bristlecone pines as a whole don’t respond to temperature is ludicrous. Now however, since you’ve raised your version, let’s discuss it. All parties agree to what you state? Are you sure about that? What do you mean by “all tree ring records”? By “tiny percentage”? Care to quantify? Do you disagree that all species are capable of recording a T signal? Your evidence for this claim?

  307. Jim Bouldin Says:

    and finally, at 8:25:

    “There are potential problems with this approach.

    1. The screening/validation/calibration exercise is limited to the years for which temperature data are available, 1850-1995. For pre-1850 reconstructions, the underlying premise is obviously that the relationships established for the screening period hold throughout the reconstruction period. That only a certain fraction of candidate proxies pass Mann08′s test does not address this concern. This weakness in all proxy-based reconstructions has been highlighted by the much-discussed “Divergence Problem.””

    Straight up moving of the goalposts. This is *exactly* why I have been bringing up the issues raised by Loehle. It’s nothing you yourself have raised; I raised it *for* you. Now you’re citing it as an argument against my position!!? I’m *well* aware of it–I was addressing your bigger –and much more important—failure to see the totality of the evidence.

    “2. …This success captures a major shortcoming of Mann08′s procedure: it cannot distinguish meaningful correlations from spurious ones. Statistician J. Udny Yule flagged the challenges of nonsense-correlations in 1926 (PDF). Are all Mann08′s proxies save the Tiljander data series free from nonsense-correlations? I don’t know, as it would take a careful examination to come to an answer. But I wouldn’t take that on a bet.”

    Well by “nonsense” I guess you mean spurious. This is exactly what I’ve been addressing, and which was addressed in the paper with the very quote you cite. Are you paying attention? And as mentioned by me above, you do realize that you are now addressing a much wider issue that affects anyone who uses a known correlation to predict an unknown value, don’t you?

    “3. Even stipulating that proxy selection was well-conceived and correctly executed: that is hardly a validation of the paper’s results. “Necessary but not sufficient” is the applicable concept. Indeed, concerning key questions about the effects of the Tiljander proxies on recons, Mann08 SI Fig. S8 had to be re-done as a result of calculation errors (the version in the SI at pnas.org is incorrect).”

    Moving of the goalposts, version 2.

    It’s unlikely that there’s much to be gained from discussing this topic any more with you. Your approach is clearly to cast doubt rather than to try to assess the overall strength and meaning of the paper. If you were really interested in methodological issues, you wouldn’t be harping exclusively on Mann et al 2008, but rather discussing the literature more generally, because I’m sure you can find some issue or weakness in almost any paper you read. The reason you harp on it is because others have harped on it before you.

  308. sidd Says:

    The comment thread illustrates the first sentence quoted by Dr. Verheggen:

    “In science, to actually contribute at the forefront of a field one has to earn credibility, not demand it. Being taken seriously is a privilege, not a right.”

    As a sometime scientist, it is clear to me that Dr. Bouldin has earned the privilege of being taken seriously, and I thank him for his patient explanations. Not so, his debaters, who seem to be numerologists at their best. At worst they resemble hieromancers, gazing upon the entrails of papers they do not understand, and divining profound irrelevancies therefrom.

    sidd

  309. AMac Says:

    Jim Bouldin (17:12) —

    > By “basic” I mean something larger than what you conceive of. I mean your underlying message.

    My underlying message is: The Tiljander data series are uncalibratable by direct methods to temperature, and thus their use in Mann08 is invalid. With Tiljander removed from the no-dendro recons, those no longer resemble the dendro recons. One of Mann08′s central points is that recons built from non-dendro proxies support the validity of recons built from treering proxies. Without Tiljander, this key assertion fails.

    > I’m interested in the sum of the evidence, not your legalistic approach (obsession actually) with one tiny component thereof.

    My approach isn’t legalistic, or obsessive, or — in Gavin Schmidt’s turn of phrase — pathological. My prior paragraph ends with a strong claim: “Without Tiljander, this key assertion fails.” Thus, I am contesting your belief that Tiljander is a “tiny component.” I’m saying that it is a major component.

    You should consider a strong claim to be a gift. If you are correct, you should be able to refute such a claim. You can do this by pointing to a non-dendro reconstruction that excludes Tiljander, and retains characteristics that are meaningful. “Meaningful” in this case translates to “supportive of the claims in Mann08’s abstract, and/or in the press releases announcing Mann08’s publication.”

    > You think that whatever rule set you design (1) is the right one and (2) therefore has to be followed by others.

    That’s (1) trivial and (2) silly. (1) Of course I think my reasoning has merit. As new facts and insights emerge, I change my mind. What do you do, sir? (2) People can and do follow whatever rule sets they wish. I’m not the emperor’s-clothes police.

    > Later in that comment you confuse the issue I raise (over-generalization of the problems with strip barked vs all bristlecones that MikeN introduced, and which is patently false), with the issue that Mann et al was addressing in the paper (i.e. how do the non tree proxies compare with the tree proxies)

    I treat others’ comments as carefully as I can. Reading back, I see the source of your annoyance. Still, I quoted Mann08 in context on their reasoning for seeking and analyzing non-dendro proxies. That is hardly evidence of confusion. I would suggest that quoting others directly or providing links (in this case to whatever MikeN said) would take care of most of this problem.

    > and you confuse statements I made about what others have said, with saying that you said them.

    Without a quote or a link, I’d just be guessing as to what you are referring to.

    Jim Bouldin (17:30) —

    > [Quoting AMac at 7:44] “Taking that ["THERE IS NOT A TREE SPECIES ON THE PLANET THAT DOES NOT RESPOND TO TEMPERATURE! That would be a biological, evolutionary impossibility."] at face value, you would be suggesting that all tree-ring records could serve as temperature proxies. But all parties agree that, at best, a tiny percentage of such records are suited to that purpose.”

    > Uh no, that’s not what I was getting at, but nice job of flipping the meaning to your ends.

    I suggest less mind-reading. Write more carefully.

    > My point was that the claim that bristlecone pines as a whole don’t respond to temperature is ludicrous.

    I agree.

    > All parties agree to what you state? Are you sure about that? What do you mean by “all tree ring records”?

    I thought it was common knowledge that tree-rings primarily respond to temperature only in certain restricted settings. I thought that much effort goes into locating stands of trees that qualify in this way, and that much thought goes into figuring out if present-day treering: temperature correlations extend into the past. If these impressions are not true, I stand corrected.

    The first listing that Google brings up is Wikipedia’s entry on Dendrochonology. Quoting

    In areas where the climate is reasonably predictable, trees develop annual rings of different properties depending on weather, rain, temperature, soil pH, plant nutrition, CO2 concentration, etc. in different years. These variations may be used to infer past climate variations.

    That is consistent with what I think and with what I wrote.

    > By “tiny percentage”? Care to quantify?

    As a SWAG, how about “fewer than 2% of all tree stands in a given broad geography (e.g. the Rocky Mountain states, France) Europe) can be expected to include trees that display variations in tree-ring widths over time with a detectable temperature signal. Other changes (e.g. precipitation, humidity, insect infestation) often contribute “noise” that can overwhelm the temperature signal.”

    > Do you disagree that all species are capable of recording a T signal?

    As above, No.

  310. AMac Says:

    Jim Bouldin (17:53) —

    > Straight up moving of the goalposts. This is *exactly* why I have been bringing up the issues raised by Loehle. It’s nothing you yourself have raised; I raised it *for* you. Now you’re citing it as an argument against my position!!?

    As far as I can tell, this is a case of Cocktail Party versus Journal Club. I generally restrict my comments to a narrow, technical question because I have found that it is extremely difficult to get most people to apply evidence-based approaches when more than two or three steps are required — if the conclusion might clash with their overall views. This is true of “skeptics” — see WUWT. From what I see, it is equally true of those in the Mainstream of climate science.

    That said, as willard observed upthread and PeteB noted yesterday (July 23 at 17:11), those in the Mainstream who have looked at the evidence and then offered a clear-cut judgement have agreed with my claim that “The Tiljander data series cannot be directly calibrated to the instrumental temperature record, 1850-1995.” (H6).

    So there’s something to be said for maintaining a narrow focus.

    In your responses to me, you have declined to address the points I raise, preferring to broaden the focus to discuss “the big picture”. With some reluctance, I joined you in Cocktail Party mode, and offered my opinions on the subjects you raised (my comment of July 24 at 08:25).

    In one sense, that’s fine. There’s no single rule set that’s the right one, and therefore has to be followed by everybody. See my immediately prior comment of 18:19.

    In another sense, it’s a problem. You scorn what I wrote on problems with Mann08’s screening/validation/calibration (#1 at 8:25, supra). As best I can tell, this is because Craig Loehle has raised similar points, and you had already mentioned his work several times in this thread. You suppose that by not switching the discussion of Mann08 to Loehle’s concerns, I am moving goalposts. Again.

    Or something.

    > Well by “nonsense” I guess you mean spurious.

    Yes. That’s why I said “spurious.” Yule introduced the concept with the term “nonsense-correlation” in the PDF I linked.

    > Are you paying attention?

    Yes.

    > And as mentioned by me above, you do realize that you are now addressing a much wider issue that affects anyone who uses a known correlation to predict an unknown value, don’t you?

    Yes. The inadvertent misuse of the Tiljander data series is a single instance of a much wider issue.

    > It’s unlikely that there’s much to be gained from discussing this topic any more with you.

    That’s a sentiment we share. You should be gratified by sidd’s comment at 18:03, supra. I expect that most of Bart’s readers share his perspective.

    > The reason you harp on it is because others have harped on it before you.

    You would be a better sparring partner if you would restrain the impulse to mind-read.

    Still, it’s been nice discussing this with you — thanks for returning to the thread. Enjoy the rest of the weekend.

    PS – If I can find an email for Dr Loehle, I’ll point him to this discussion.

  311. willard Says:

    Chris Colose,

    Thank you for your comment.

    You say that we do no longer have a “scientific debate.” This use of “scientific” is worth diligence. Let’s recall what Martin Vermeer said:

    > AMac, the only question that has some scientific relevance is (3).

    You seem to be using the mean something not unlike that when you say:

    > It is not an interesting argument in the scientific literature and quite honestly, the attention it gets it makes no sense.

    I know you’re not exactly talking about the same thing as Vermeer. What I want to underline is that you’re using the word “scientific” is a way that agrees with Vermeer’s usage.

    It also seems that this meaning agrees with one of MIkeN’s comments, on the 2011-06-26, at 11:03 EDT:

    > By the way, I agree with you Chris from a scientific standpoint. But you are missing the political impact of letting bad science stand.

    http://ourchangingclimate.wordpress.com/2011/06/22/climate-science-scientific-method-skeptics-not/#comment-13430

    The expression “bad science” is not unlike the “strong science” we also witness in the thread. Both deserve due diligence, but we have to deal with “scientific” for now.

    ***

    Let’s recall what I called type-2 questions:

    (2a.1) Should Mann and coauthors have figured out the problems with calibration of the Tiljander proxies to the instrumental record from the discussion in the blogosphere?

    (2a.2) Should Mann and coauthors have figured out the problems with calibration of the Tiljander proxies to the instrumental record from the discussion from the M&M comment in PNAS?

    (2a.3) Should Mann and coauthors have figured out the problems with calibration of the Tiljander proxies to the instrumental record from the discussion from the M&M from the correction to the Kaufman Science 2009 manuscript?

    I wonder if you’d say these are scientific questions. From what I can gather from reading you, Martin Vermeer, and MIkeN, I believe the answer should be **no**. For instance, you say:

    > The whole thing is used primarily a “gotcha” tool and as some straw man argument against the foundation of anthropogenic climate change; it justifies countless people living in an odd Groundhog’s day repetitive version of how conspiracy science operates, instead of learning something new and interesting and placing it within the context of the big picture.

    Am I reading you right?

    I would also be interested by what you mean by “conspiracy science”.

    What you call “gotcha” tool might have an interesting mix of questions that are scientific or not, that are beaten to death or hinted by way of images, that are matter of interpretation or so obvious to the querent as to to imply one and only one answer. This mix deserves due diligence.

    ***

    Here are expressions with the adjective “scientific” that one can find on this very page, removing duplicates among the 80 or so occurences and ordering alphabetically:

    – scientific academies
    – scientific arguments
    – scientific background
    – scientific case
    – scientific channels
    – scientific coherence
    – scientific community
    – scientific consensus
    – scientific controversy
    – scientific (and public) debate
    – scientific evidence
    – scientific field
    – scientific inquiry
    – scientific literature
    – scientific method
    – scientific process
    – scientific question
    – scientific relevance
    – scientific standards
    – scientific standpoint
    – scientific work

    Looking at the sentences, paragraphs, and whole comments in which these expressions are being used would be one way to determine what “scientific” means among many. This would be one way how the journal clubs I know determine these things. Quoting from dictionary (as did for instance AMac on the 20111-07-19 at 07:34) would in fact go against the best practices of these clubs.

    I know at least one journal club where we can witness lots of “appeals to dictionary.” Incidentally, AMac is a frequent contributor to this journal club. Not uninterestingly, if one searches for “site:http://rankexploits.com/musings tiljander”, it might return more than 100 results.

    Looking back into the most active threads, we can see 25 hits on this one:

    http://rankexploits.com/musings/2011/blah-blah-blah-mt-and-communicating-climate/

    I believe that the thread that has the more hits (45!) is this one:

    http://rankexploits.com/musings/2010/litigation-copyright-fair-use-us-speech-and-debate-clause/

    We can witness interesting “arguments from dictionary” there.

    The “Journal Club” exchanges from this journal club surrounding Tiljander deserve due diligence.

    ***

    I’ve introduced the expression **scientific criticism** in the discussion. It seems to be a better fit for journal club activities than the “audit” metaphor. Some clubs concentrate on litterary criticism, others on scientific criticism: whatever criticism partakes of these discussions, criticism it is.

    It would be interesting to know the background of the many members of scientific criticism clubs. While this question deserves due diligence, we can acknowledge this agreement between you and AMac. In your last comment, you say:

    > I also think that ~99% of the people who choose to comment on this issue lack the statistical background or have not went through the fine details of the many proxy studies to really comment confidently about it (I am, admittedly, in this crowd as well). The arguments have become very nuanced and technical, and generally boil down to most people agreeing with whoever they decided has the better position about something completely irrelevant, like whether CO2 causes warming or what climate sensitivity is.

    Here’s AMac’s take on this, on the 2011-07-05, at 11:46:

    > One of the things I’ve realized from this thread is that many scientifically-literate blog commenters don’t understand the basic concepts behind proxy-based reconstructions at a level that permits informed discussion of technical issues. [...] If one looks at the analytical steps in reconstruction-building as a very complex black box, it’s reasonable to orient oneself in the debate on the basis of who one trusts.

    http://ourchangingclimate.wordpress.com/2011/06/22/climate-science-scientific-method-skeptics-not/#comment-13550

    (Also reproduced here: http://neverendingaudit.tumblr.com/post/7709607897)

    Your impression seems to agree with AMac realization. Let’s hope we do not have to rely on the dictionary for any word contained in both quotes to mare sure of this agreement. Appeals to dictionary usually do not end well.

    ***

    If you’re both right, most people do not know enough to answer these technical questions.

    And yet, these technical conversations happen all the time, in the modes of journal clubs, cocktail-parties, or else.

    Here’s where an auditor asks her, him, or itself:

    Why?

    This why-question, and many more, deserves due diligence.

  312. Craig Loehle Says:

    I hesitate to post here since I only have 20 out of my 131 papers on climate…not sure if that makes me a “real” scientist or denier. Anyway, a few points:
    1) A series of papers such as http://geography.cz/sbornik/wp-content/uploads/2011/06/g11-2-1ljungqvist.pdf and others by the same author show the MWP to have been consistently warmer than today, and the mid-Holocene even warmer than the MWP (and about 2 deg C warmer than today). The importance of this is that species did not go extinct then nor did ice sheets slide into the ocean. I think this a serious issue that needs to factored in.
    2) As lucia has shown, a plot of GMT from GCMs without GHG additions shows little agreement between the models, with some as much as 4 deg C warmer than reality. Given that black-body radiation is proportional to the 4th power of surface temperature, this seems like a pretty serious problem with their radiative physics or heat transfer across the globe.
    3) Which version of global warming do supporters of the idea here agree with? The version which says only the arctic will be habitable by 2100? The one with 20m sea level rise? The one which says 2 deg C warming means the end of civilization? Don’t sneer–these make headlines all the time and are what people like Gore and Hansen believe. And let’s say you believe the median 3 deg warming predicted by the IPCC and think this will cause…discomfort, some droughts, some floods, etc (ie, not the end of the world), then you should perhaps forgive people for not getting excited about draconian price hikes on energy and people telling us what light bulbs to buy to prevent possible discomfort and a few droughts.

  313. J Bowers Says:

    For reference:

    * New temperature reconstruction vindicates …

    A new temperature reconstruction has been published in the Swedish journal Geografiska Annaler: Series A, Physical Geography. The reconstruction (hereafter Ljungqvist 2010) covers the past 2000 years for the middle and high latitudes of the Northern Hemisphere.
    [...]
    Readers may wonder how this new reconstruction compares to previous hemispherical and global temperature reconstructions. In his conclusion, Ljungqvist (2010) reports that:

    “Although partly different data and methods have been used in our reconstruction than in Moberg et al. (2005) and Mann et al. (2008), the result is surprisingly similar. The inclusion of additional records would probably not substantially change the overall picture of the temperature variability.”

    * How does Ljungqvist’s reconstruction compare to others?
    * Shooting the Messenger with Blanks
    * Tai Chi Temperature Reconstructions
    * Hockey Stick Own Goal

  314. Craig Loehle Says:

    In response to the Tiljander series debate above, 3 pts:
    1) Since the Tiljander sediments are contaminated for the ENTIRE period for which there is climate data (post-1850) there is no way to calibrate this data to make it a proxy. It is as meaningless as sales of slinkies or barbie dolls.
    2) The no-dendro series of Mann was a test to prove his method and results were the same without tree rings (since the tree rings are heavily influence by strip-bark bristlecones)–but included Tiljander sediments. Steve M showed that this data without Tiljander does not match the dendro data (different results…bad) and has lousy fit stats. Thus Mann’s claim is false.
    3) If the result depends on a nonsense (known to be nonsense) correlation, ask yourself why you are spending electrons defending such a thing.

  315. AMac Says:

    willard (00:26) —

    Interesting observations, thanks for compiling and presenting them.

    .

    Craig Loehle (03:32)–

    I am glad that you found the thread. However… the points you made at 03:32 have all been discussed at some length, over the time this thread’s been in existence.

    I suggest that the odds that many readers will change their minds at this point is… low.

    I fear that Discussion 2.0 might try the patience of our host. Though he would know best.

  316. grypo Says:

    “And let’s say you believe the median 3 deg warming predicted by the IPCC and think this will cause…discomfort, some droughts, some floods, etc (ie, not the end of the world), then you should perhaps forgive people for not getting excited about draconian price hikes on energy and people telling us what light bulbs to buy to prevent possible discomfort and a few droughts.”

    Could you please point out the literature that makes you so confident in your predictions? And do you understand clearly what the IPCC predicts and for what scenario on what time scale? No one expects people to get excited about “draconian” price hikes, thanks.

    “The one which says 2 deg C warming means the end of civilization? Don’t sneer–these make headlines all the time and are what people like Gore and Hansen believe.”

    Could you point out where they say that? I know Hansen calls 2C a “prescription for disaster” due to the time scales at which the heat energy moves from the ocean to the ice and shows what rises have been like in the past at similar temperatures to the possible predictions. These changes will have quite a profound effect on future civilization, to say the least. You seem to think that as long as we can’t predict that the effects of a changing climate aren’t the ‘end of the world’ that there is no need to act. All I can say is that I could not possibly agree with that, and I believe most people would be with me. Risk seems to be in the eye of the beholder. I’d be respectful of that.

  317. Craig Loehle Says:

    For grypo at 03:56
    Re: 2 deg C rise and disaster, in my post above I pointed out that it was 2 deg C warmer from 6000 to 8000 yrs ago, and the world did not end. If Hansen is your only source, good luck to you.

    Yes, perception of risk is personal–I was asking people to consider that others may not agree with their perception of risk.

    Ok, here is some literature: 2009 IUFRO (International Union of Forest Research Organizations) report on climate change and forests agrees with the IPCC that there is no detectable impact of climate change on the world’s forests so far, and that none is likely over coming decades. My paper Loehle, C. 2003. Competitive Displacement of Trees in Response to Climate Change or Introduction of Exotics. Environmental Management 32:106-115.
    argues (and I have several others) that climate shifts should cause a slow change in vegetation zonation. In my latest paper Loehle, C. 2011. Criteria for Assessing Climate Change Impacts on Ecosystems. Ecology and Evolution doi: 10.1002/ece3.7
    I show that published simulation model studies of various ecosystems (other people’s published work) tend to predict increased vegetation growth and more drought tolerance in response to IPCC median (or even A2) scenarios over the next 100 yrs (not destruction), partially due to model increased precip and also CO2 effects on enhancing plant growth. The “negative” impacts in the IPCC for ecosystems are either taken from greenpeace reports or are vague and not even within the realm of what the GCMs can do–for example, GCMs do not simulate tornados or match historic drought extent in test regions.

  318. PeteB Says:

    Craig Loehle.

    1) I agree Tiljander shouldn’t be used
    2) This effects the shape and fit characteristics

    OK, I agree to some extent, but, according to the update to supplemental information, the non dendro reconstruction with problem proxies removed passes RE at just below the 95% level (approximately 94% level) back to AD 1300 and the 90% level back to AD 1100 (they pass CE at similar respective levels)

    1100 maybe is pushing it a bit, but 1300 is pretty close.

    Comparing the dendro reconstruction with problem proxies removed with the non-dendro reconstruction with problem proxies removed gives a fair degree of overlap given the structural uncertainties involved.

    the yellow area is the two sigma error region for the dendro including problem proxies, so I guess it will be even larger for the other reconstructions

  319. PeteB Says:

    Craig Loehle.

    There seems 3 steps to me

    1) Forecasting the climatic changes for different scenarios
    2) Working out what are the impacts and the societal costs of these impacts
    3) working out what we should do (if anything) about these

    Of course, there is uncertainty in all these, but that doesn’t really help us policy wise

    For example, very high climate sensitivities will have a disproportionally high impact, and even a relatively modest chance of these would push for a more severe policy response. I actually am convinced by James Annan that these very high climate sensitivities are unlikely, but the principle seems to me that uncertainties don’t help the argument to delay mitigation policies, in fact the opposite

  320. PeteB Says:

    Incidentally has anyone attempted to use the Mann method including all the dendro except the strip-bark bristlecones and including all the non-dendro except those already identified as problematical

  321. Marco Says:

    Craigh Loehle, that paper of Ljunqvist you refer to does NOT show what you claim. He compares to pre-industrial temperatures, which we know are significantly colder than today. Did you not notice this at all?

    You also previously tried to claim your results were vindicated by Ljunqvist, only to be shown decidedly wrong:

    http://tamino.wordpress.com/2010/09/28/vindication/

    http://rankexploits.com/musings/2010/comparing-proxy-reconstructions/

    Just wrong, or an attempt at obfuscation?

  322. PeteB Says:

    Marco,

    That post at Lucias seemed to show remarkable agreement between Mann 08 eiv, Ljunqvist and Moberg 05

  323. PeteB Says:

    Oh and Ljunqvist doesn’t use strip-bark bristlecones

    http://noconsensus.wordpress.com/2010/10/07/another-mathematically-honest-reconstruction/#comment-37980

  324. Craig Loehle Says:

    PeteB & Marco: I would appreciate you not claiming to know my motives which are strictly about good science. I don’t do food fights. If you add the warming of recent decades to Ljunqvist MWP recon, the MWP is still warmer. If you look at his Holocene recon, it shows a mid-Holocene period of thousands of years (not just a few years) 2 deg warmer than pre-industrial. The Moberg 05 recon uses non-treering proxies for the low frequency signal detection and my data partially overlaps with Moberg. I don’t take Tamino posts to show anything “decidedly”. How temporal patterns match up is pretty subjective (part of the problem). In the IPCC report, they claim their spagetti graph shows “remarkable agreement” when to my eye and to a statistical coherence test (which someone did) they are pretty incoherent. An incoherent set of reconstructions does not indicate that the proxies work very well.
    As far as PetB 3 steps–yes, this is correct procedure. I show in my latest paper Loehle, C. 2011. Criteria for Assessing Climate Change Impacts on Ecosystems. Ecology and Evolution doi: 10.1002/ece3.7
    that much of the literature on ecosystem impacts leaves out critical factors which tend to produce alarmist outcomes–but the simulation models which do the best job of including everything project positive to only slightly negative effects on ecosystems (forests, grasslands mainly in the lit I reviewed) by 2100 even with the A2 scenario. I would say there is a consensus on this. Much of this literature is too recent to have made it into AR4, which used very qualitative and sloppy assessments of impacts from climate change on crops and ecosystems and focused mostly on putative droughts and extreme events which mostly bother humans more than ecosystems. Ecosystems and forests in particular are my specialty.

  325. grypo Says:

    “For grypo at 03:56
    Re: 2 deg C rise and disaster, in my post above I pointed out that it was 2 deg C warmer from 6000 to 8000 yrs ago, and the world did not end. If Hansen is your only source, good luck to you.”

    Your basically strawmanning again. The ‘end of the world’ argument just a logical fallacy. I know it’s easier to argue with people if you make up the argument that you want argue against, but it’s usually not appreciated. And I’m not the one who brought up Hansen, you did. Are you aware of any of the work from the Eemian period? The Mid-Holocene while it may help to predict to N Hemisphere drought (which was severe by all accounts) we don’t have good global numbers to know how it compared to today or how it compares to what will happen decades from now.

    “2009 IUFRO (International Union of Forest Research Organizations) report on climate change and forests agrees with the IPCC that there is no detectable impact of climate change on the world’s forests so far, and that none is likely over coming decades ”

    I invite people to read the report:

    http://www.iufro.org/download/file/4373/387/op23_pdf/

    I don’t see where there is your kind attitude toward is expressed toward climate and forests. While severe impacts may be decades away, it certainly does not mean they won’t happen, and in fact the document you cite and the IPCC predicts it will, as well as many other man-caused impacts to forests. That’s the whole point — preparing and mitigating for the future. I’m sure Jim Bouldin could give us a good run-down of the forests’ many vulnerabilities to the changing climate moving forward.

    But since you point out that report, I deduce that you would agree to many of the adaption strategies suggested, as well as using that as a guide to look at all effects of minimal and maximum climate change going forward, such as sea-level rise, extinction, drought, floods, etc etc. Right? And can we at least agree that by not mitigating, we are forcing generations beyond ours to dealing with the effects of climate change, what ever they may be, and dictating that they use their resources for adapting to changes that they did not cause?

    “The “negative” impacts in the IPCC for ecosystems are either taken from greenpeace reports or are vague”

    This just isn’t the case at all.

    “Yes, perception of risk is personal–I was asking people to consider that others may not agree with their perception of risk.”

    I try. But that’s not what you did. You strawmanned and mocked.

  326. Marco Says:

    Craigh Loehle, add modern temperatures to Ljungqvist’s reconstruction, and the MWP is clearly cooler than the current temperature. Heck, Ljungqvist even noted his MWP is cooler than Mann08 and 09!!

    At least you stepped away from your initial claim that the HCO was several degrees warmer than modern temperatures.

    This is not about a food fight, it is about you making false claims. If you’re concerned about good science, you’d not have made such false claims.

  327. PeteB Says:

    PeteB & Marco: I would appreciate you not claiming to know my motives which are strictly about good science

    I didn’t !

  328. Eli Rabett Says:

    Ah, the wounded warrior gambit.

  329. J Bowers Says:

    Craig Loehle — “2 deg C rise and disaster, in my post above I pointed out that it was 2 deg C warmer from 6000 to 8000 yrs ago, and the world did not end.”

    Are you an archaeologist now? How do you know this?

  330. Chris Colose Says:

    Craig Loehle is simply wrong about it being 2-3 C warmer than today during the early or mid Holocene, at least if one is to interpret his statement as an annual and global average. The higher summer latitudes were probably warmer as a function of orbital forcing, but with differential winter and tropical responses. You have to go back at least to the Pliocene to find global temperatures several degrees warmer than today, and these generally correspond with much higher sea level and reduced ice cover.

    But in fact, there were many changes associated with the ITCZ, monsoon systems, etc in the mid-holocene that today affect billions of people (presumably there weren´t this many people around to be impacted in the Pliocene or even mid-Holocene); today we also have much more coastline infrastructure, and a significantly changed landscape (today, a certain species might have to cross a wal-mart parking lot in order to migrate from the effects of climate change; in the mid-holocene, not so much), so the comparison is not really useful.

    I understand Loehle is the resident expert in dendroclimatology, ocean heat content analyses, satellite data, archeology, holocene variability, ecosystems and forests, CO2 trajectories, etc…with generally one or two publications in spurious journals like E&E on each topic, but it´s these simple errors and phantom arguments like ¨the world is going to end¨ that are the reason no one actually takes him seriously.

  331. MIkeN Says:

    Jim Bouldin, I see your point about Jeff Id’s criticisms, and am reworking his code to do a more exact replication to my satisfaction, as he is using different correlation numbers.

    The issue I have along the lines of Jeff Id,
    there is the suppression of historical signal that comes from calibrating to the modern temperature level. This is separate from the spurious correlation issue.

    You wrote above that they only selected treerings where there was .5 correlation between individual cores. What do you think they are correlating to? You are implying that they are correlating to temperature, yet from the results of Mann 08, we see that the answer for about 2/3 of the proxies is ‘not temperature’, at least at a level of .106. So you have .5 correlating to each other, but not .1 correlating to temperature?

    You point out 400 proxies that correlate to temperature. If we assume a 10% random correlation level, then this is maybe 320 valid proxies and 90 correlated by chance. Again 2/3(this number is sensitive to the types of proxies you include in the count) did not correlate. I’m working on how sensitive this would be.

    Technical question, the .5 correlation between cores. Is this a .5 correlation between every pair of cores, or is there a group correlation calculation? At the time the paper came out, the Yamal cores were not even released yet, so how did this get included. I realize you just got the details from the paper, but anything you can surmise would be helpful.

  332. willard Says:

    On the 2011-07-25 at 7:28 EDT (Auditor’s Official Timezone), Craig Loehle introduced the concept of “good science” and “food fights”:

    > I would appreciate you not claiming to know my motives which are strictly about good science. I don’t do food fights.

    If I read the timestamps properly at 7:34, so six minutes later, Craig Loehle had suspicions:

    > I have my suspicions about the aerosol input data to IPCC which look totally made up (long straight lines, no data as input) so I will read the Hoyt papers (great, another specialty to come up to speed on, thanks a bunch!! )

    http://judithcurry.com/2011/07/23/theories-vs-theories/#comment-90087

    One day earlier, on the 24-07-24 at 20:26, Craig Loehle was using the scare quotes:

    > There is a difference between the scientific debate and those who claim the only habitable land by 2100 will be in the Arctic, or who proclaim that we must destroy the cities and half the world’s population to save nature. There is a complete spectrum of belief, which makes it hard to even have a discussion. There are also “experts” like Hansen (an astrophysicist) making pronouncements about topics he knows nothing about, like species extinctions or economics. Murky thinking and hyperbole are everywhere (I am not accusing you of this, just an observation).

    http://judithcurry.com/2011/07/23/theories-vs-theories/#comment-89969

    Four minutes before that comment, on the 2011-07-24 at 20:22, Craig Loehle recalled ancient Greeks and corrected some misunderstandings about Columbus:

    > One of the ancient greeks, by going to egypt and noting the date when sunlight penetrated to the bottom of a well and comparing the angle of the sun on the same date in Greece and the distance to greece from Egypt got the circumf of the earth pretty close to correct (which is not to say that everyone knew this centuries later). And the story about Columbus is backwards–those opposed to his voyage thought he would starve in the huge ocean between Europe and Asia –and if the Americas had not been in the way, they would have been right.

    http://judithcurry.com/2011/07/23/theories-vs-theories/#comment-89966

    One day before that comment, on the 2011-07-23 at 20:37, Craig Loehle was using the “/sarc” operator:

    > I guess they [Climate Scientists who combine parameters to arrive at a final answer that has smaller uncertainties than many of their inputs] must just be smarter than engineers /sarc

    http://judithcurry.com/2011/07/23/theories-vs-theories/#comment-89684

    One day before that comment, on the 2011-07-22 at 16:20, Craig Loehle was spotting a trend:

    > The trend of making “scary scenarios” predates climate change. Rather than saying “it is shameful that there are homeless people in the US” advocates claim millions of people are homeless. People make up wild numbers about child abductions or claim that essentially all women have been sexually assaulted (raped) by including pinching or tickling in the definition. This is perpetuated by journalists who don’t bother checking facts. It pollutes discourse and is a bullying tactic.

    http://judithcurry.com/2011/07/21/stephen-schneider-and-the-%E2%80%9Cdouble-ethical-bind%E2%80%9D-of-climate-change-communication/#comment-89250

    Five minutes before that comment, on the 2011-07-22 at 16:15, Craig Loehle was using more scare quotes:

    > In Europe, many of the “grass-roots” organizations lobbying for carbon reduction are in fact funded by the government itself. Some “popular movement”.

    Twelve minutes before that comment, on the 2011-07-22 at 16:03, Craig Loehle was explaining how AGW alarmists deny others the right to not be scared of the things they are scared of:

    > They do it by avoiding cost/benefit analysis, avoiding risk assessment, trying to shut down debate, and jumping right from effect (2 deg warming) to policy prescription (and laws).

    http://judithcurry.com/2011/07/21/stephen-schneider-and-the-%E2%80%9Cdouble-ethical-bind%E2%80%9D-of-climate-change-communication/#comment-89244

    Less than seventy minutes before that comment, on the 2011-07-22 at 14:51, Craig Loehle was just guessing:

    > They probably spent a ton of money to create a wonky tool worse than the R code for accessing climate data that Mosher put together in his spare time (just guessing, based on their track record).

    http://climateaudit.org/2011/07/20/east-anglias-toxic-reputation-manager/#comment-298919

    About 3 hours before that comment, on the 2011-07-22 at 11:44, Craig Loehle was talking about virtue ethics:

    > The seduction of virtue is very tempting. I like the cop shows, but the violation of basic rights in these shows during a week would be hard to even count, in the name of catching the bad guy. In the name of catching drug smugglers, carrying cash in your car is de facto illegal if you are caught with any drugs (how do you prove you got it legally?). A fundamental problem with AGW alarmists is that they deny others the right to not be scared of the things they are scared of. (like 2 deg warming). In their pursuit of virtue (saving the world) they assume too much. Maybe others fear unemployment more than 1 ft rise in sea levels.

    http://judithcurry.com/2011/07/21/stephen-schneider-and-the-%E2%80%9Cdouble-ethical-bind%E2%80%9D-of-climate-change-communication/#comment-89181

    ***

    Readers can witness “good science” in action.

  333. Eli Rabett Says:

    Curry has become the go to place for posting any flotsam that passes by with the “this is interesting” flag. She is the Sgt. Schultz of the rejectionists.

  334. Chris Colose Says:

    Eli,

    Indeed. Such is Loehle and Scafetta’s new “60-year cycle means climate sensitivity is low and attribution is wrong, and by extrapolation, future global warming will be small” nonsense, that Curry couldn’t resist hosting.

    I really wish I knew how these people do it; even the authors managed to write an abstract with glaring fallacies in it. You’d think whatever drives their B.S. motivations could up with something better…

  335. Quiet Waters Says:

    It may or may not be relevant that, in this age of emails, Craig Loehle still sends reprint request postcards.

  336. J Bowers Says:

    I’d still like to know more about this apparent steady state hypothesis of early civilisation.

  337. AMac Says:

    Eli Rabbett & Chris Colose —

    A brief aside. Pursuant to the earlier conversation I had with willard, could I invite you each to weigh in on two questions? Of course, you’re under no obligation to answer — just that I’m curious.

    * In your opinion, are the Tiljander data series directly calibratable to the instrumental temperature record?

    * In your opinion, is Mann08 “strong science”?

    The first question is one of willard’s “Type 3,” while the second would be a “Type 2.” For context, you can search this thread for these phrases.

  338. willard Says:

    Craig Loehle introduced the expression “good science”. We also saw how Loehle himself was advocating good science. While reading the many threads dealing with the Tiljander affair, we encountered a related expression.

    In pages 27-28, Andrew Montford introduces Deming and his research using this description:

    > Deming had recently created a temperature reconstruction for the last 150 years, based on boreholes in North America. In his study, he concluded that North America had warmed somewhat in the period since 1850, but had little to say bebond that. This was good, solid science but not the stuff of newspaper headlines.

    Here are some expressions in the following sentences:

    – “considered highly important in climatic science circle”
    – “with the expectation that temperatures were being driven upward”
    – “storyline of rising temperatures”
    – “global warming industry”
    – “who [the global warming industrialists] thought they saw”
    – “they thought I was one of them, someone who would pervert science”
    – “flash of light [...] murky shadow”
    – “the aim was to erase it in the climatological record in its entirety”

    My interest here is not in the Deming affair as such [1], but in the function of portraying Deming as doing “good, solid science”, immediately before the intriguing portray of consideration, expectation, storyline, industry, and other institutionalized thought processes and interests. Andrew Montford is not known to be knowledgeable in borehology. The statement of his opinion regarding Deming’s work deserves due diligence. As the Auditor might ask his readership to ponder: _why_?

    Here’s my hypothesis. Montford is framing the Deming affair as the story between a noble scientist versus global warming industrialists. One (or two, if we count Richard Lindzen) against a powerful multitude. A sudden revelation of tainted intentions. Pure light among the murky shadows of climatology.

    We can see that defining “good, solid science” is quite secondary. We’re not into the realms of scientific criticism, but more something like **scientific opposition research**. The way scientific opposition research operates deserves due diligence.

    [1] The Montford Dossier certainly deserves due diligence. For instance, it is claimed that “Lindzen of MIT has confirmed that the email was written by Jonathan Overpeck.” But note 12, which follows this claim, points to an Arxiv document authored by Lindzen. There is one mention to Overpeck in that document: a signature to an international conference invitation. The only mention of “getting rid” of MWP cites (Deming, 2005) as authority. Here is when the Auditor might revive yet another introduction to check-kiting.

  339. willard Says:

    In pages 27-28 of the **Hockey-Stick Illusion**, of course.

  340. Jim Bouldin Says:

    Craig:

    I’ve been fairly strongly supporting your 09 paper on the divergence issue here–because I think you’ve really got something (potentially quite) meaningful to say there–and you did a good job laying it out in that paper. Now, I do think that there are potential solutions to the conundrum you describe there, but that is a technical discussion for another time and place.

    But I think you’re far over-simplifying the current and future effects of climate change on forests, and ignoring some strong evidence of ongoing effects of various kinds, which include things such as effects on mortality rates, fire regime characteristics, and herbivorous insect dynamics. And the claim that continued climate changes in the future decades will have little effect, is even more problematic.

    MikeN: Saw your post and will get back to you asap.

  341. J Bowers Says:

    willard — “Montford is framing the Deming affair as the story between a noble scientist versus global warming industrialists.”

    “The king died and then the queen died” is a story.

    “The king died and then the queen died of grief” is a plot.

    – E. M. Forster, Aspects of the Novel

  342. MIkeN Says:

    OK, Jim. Here or at AMac’s is fine with me, as I think it is off topic here.

  343. Jim Bouldin Says:

    MIkeN on July 25 at 22:55:

    ” The issue I have along the lines of Jeff Id, there is the suppression of historical signal that comes from calibrating to the modern temperature level. This is separate from the spurious correlation issue.”

    Yes, very different topics. The former will always be a potential issue.

    “You wrote above that they only selected treerings where there was .5 correlation between individual cores. What do you think they are correlating to? You are implying that they are correlating to temperature…”

    No, I wasn’t meaning to imply that. I was trying to make the point that the individual trees at the sites selected are responding +/- in concert to *some* environmental signal(s)–the .5 criterion is the mean correlation of all the cores with each other–evidence of a common forcing/signal. This means that the trees were in fact responding to a signal(s): another piece of evidence negating the argument that spurious correlations in calibration period are what produce hockey sticks, especially when you consider that the inter-tree correlations with each other are computed over the entire chronology, not just the calibration period.

    “…yet from the results of Mann 08, we see that the answer for about 2/3 of the proxies is ‘not temperature’, at least at a level of .106. So you have .5 correlating to each other, but not .1 correlating to temperature?”

    Right–good observation. And that is important information, because it does mean that a lot of trees at the sites that didn’t pass the T screeniing, were either (1) responding to something else, or (2) that they were so far from the closest met station that their r values with station temps were seriously degraded. A very interesting question itself, and addressable using different techniques.

    “You point out 400 proxies that correlate to temperature. If we assume a 10% random correlation level, then this is maybe 320 valid proxies and 90 correlated by chance. Again 2/3 (this number is sensitive to the types of proxies you include in the count) did not correlate. I’m working on how sensitive this would be.”

    484, from memory, was the number that passed the screening. But, as you noted above, it was likely an r value (r>= .128) screening criterion, not p value. This complicates the question of how many should be expected by chance. It lowers it, because an r criterion of 0.128 (the original criterion of .10 goes to .128 when accounting for AC), with n = 146 years, is almost certainly << a p criterion of the same value. Thus, I think my original statement that about 154 should pass the screening by chance, is high, and possibly way high.

    "Technical question, the .5 correlation between cores. Is this a .5 correlation between every pair of cores, or is there a group correlation calculation? At the time the paper came out, the Yamal cores were not even released yet, so how did this get included. I realize you just got the details from the paper, but anything you can surmise would be helpful."

    Good question on the correlation –I had it myself. The answer is buried in the guts of the ARSTAN program or documentation somewhere. I looked briefly at the ITRDB site and couldn't find it. Will look again.
    The Yamal data are completely inconsequential to this discussion. And they've been in the database with all the others from the beginning, AFAIK.

  344. MIkeN Says:

    Steve McIntyre got the Yamal core counts and individual core data a little bit before ClimateGate.
    It’s possible that Yamal was included as part of a larger chronology, and thus didn’t go thru the .5 screening.

    Also, .106 is the correlation amount used. .128 is used for the hundred year early period and late period calculations.

  345. Jim Bouldin Says:

    Don’t get sidetracked by Yamal. It’s completely inconsequential to the overall validity of the paper–I don’t know that it was even used.

    And the difference in r criteria also not important

  346. Eli Rabett Says:

    AMac, as Willard would say, IEHO, Tiljander has become the Schleswig Holstein of the climate wars and Eli is Bismarkian in his reaction.

  347. AMac Says:

    Funny bunny

  348. Eli Rabett Says:

    Think about it if you can

  349. AMac Says:

    silly rabbit

  350. willard Says:

    There are many ways to analyze the expression “it does not matter”. I’ll try to separate the issues the best I can. There are so many notes that it would be tough to find an optimal ordering. So I’ll simply start with some loose ends.

    Before delving into my subject, I will note one trackback (an underestimated tool) I received this morning. On the July 27, 2011, at 23:14, Jeff Id repied to Jim Bouldin with his usual colorfoulness. It starts like this:

    > I’ve been gone for a while working on other things. MikeN called my attention to a criticism by Jim Bouldin, of my ‘probing’ of the hockey stick CPS methods. Since the Air Vent wouldn’t even be a climate blog, if it weren’t for Mann 08, it does seem important to address the criticisms by Jim. As it is my blog, the cool thing is that I can shove comments right in the middle of his criticisms to point out the issues of disagreement.

    http://noconsensus.wordpress.com/2011/07/27/123-2/

    ***

    Onto our subject matter: what “matters”.

    The opposite of what does not matter is of course what matters most, better yet only what matters. What only matters is what is on topic, or what is the question at hand. Among the many possible questions that an auditor can him, her, or itself (not forgetting robots, hi bender) try to identify *the* question.

    For instance, on the 2011-07-21, at 13:39 EDT, MIkeN uses this formulation:

    > The question is whether the original hockey stick studies are convincing.

    http://ourchangingclimate.wordpress.com/2011/06/22/climate-science-scientific-method-skeptics-not/#comment-13677

    Interestingly, on the 2011-07-21, at 13:50, MIkeN identifies another question:

    > The question is, there is a Polar Urals that was done later than Yamal(the same area), yet the updated Polar Urals doesn’t get used except for one paper.

    http://ourchangingclimate.wordpress.com/2011/06/22/climate-science-scientific-method-skeptics-not/#comment-13679

    MIkeN seems to have spotted two different questions. Both are identified as *the* question. In these two quotes, the *the* device does seem to add some focus.

    ***

    Indicating what the topic is can be seen as the same kind of focus device. On the 2011-07-25, at 12:55, Steve McIntyre orients the discussion by way of this device and a related one, *the* problem:

    The failure of Jean Goodwin and others to recognize the corrosiveness of these refusals is part of the problem.

    This post is about a specific FOI refusal that makes me and many readers think that the 2006 results are “unfavorable” and that’s why they weren’t reported. of course, people’s trust in the data refusers is going to be reduced. Again, I don’t know why Jean and other concerned parties don’t simply write to Briffa and tell them their data obstruction is hurting their cause.

    http://climateaudit.org/2011/07/23/building-trust-and-foi-refusals/#comment-299200

    Notice how the first sentence assumes that Jean Goodwin failed. Notice how Jean Goodwin’s failure is part of the problem. Notice: *the* problem.

    Interested auditors should look how many “the problem” we can read on that CA page. On this very page, I see more than 40 occurences, all of them providing cues as to what *the* problem is. For the life of me, I would never say that *the* problem is the problem, but I sure would be tempted to say that it is not *the* solution. The multiplicity of *the* problems certainly deserve due diligence, if we want to understand what “matters” to most commenters and bloggers.

    ***

    Recalling the beginning of the exchange between Jean and Steve would be too long. An interested auditor would probably see that how the focus is being redirected. A quote from the article is more expedient and helps us identify another way to redirect the auditors’ attention:

    > The moral of today’s post are the points of my introduction.

    Notice: *the moral*. Here is the introduction:

    In yesterday’s post, I discussed the inconsistency between the climate community’s desire to rebuild trust and CRU/East Anglia’s continuing refusal of FOI requests, most recently for the 2006 version of the Yamal regional chronology. The moral of that post was that providing such information – even if they didn’t “have” to – was the sort of small concession that the community should willingly make as a means of “rebuilding trust” as opposed to the polarization caused by refusals that merely lead to further FOI appeals.

    Given their refusal to make even the smallest concession voluntarily, today’s post is going to be more pointed and will directly address issues of hypocrisy and mendacity that are directly raised by the most recent CRU/East Anglia refusal.

    The relationship between what a post “is about” and “the moral” of a post deserves due diligence. The first sentence clearly shows that two consecutive posts are related. Somehow, it seems that what has been discussed in past posts are not to be discussed anymore, and in fact assumed to be true, that is not to be discussed.

    In our case, yesterday’s post was about trust, while today’s post is about hypocrisy and mendacity. Trust seems to be related to hypocrisy and mendacity. Yet, Jean Goodwin or any other auditor can only discuss *the* failure, i.e. *the* problem, at the risk of being off topic (OT), topic which has been declared to be *about* “a specific FOI refusal”.

    The problems (note the plural) that surround the question (note that I’m not saying that this *is* the question) of determining what is on and off topic deserves due diligence. Delving into this questions might very well explain why moderation usually sucks. I believe it is an open (note the *an*) problem. Many other aspects of publishing (choice, frequency, timing, etc) would deserve due diligence. So much to do for auditors, so little time.

    ***

    We now have identified three devices: *the* problem, *the* question, and *the* moral. Perhaps all we need now would be *the* plot (with an hat tip to E. M. Foster) to have *the* story.

    Being able to decide *the* problem, *the* question, *the* plot and *the* moral brings great power. With great power comes great responsiblity. This responsibity rests on “what matters”.

  351. J Bowers Says:

    So, the aforementioned don’t even seem to know themselves what *the* problem is? Thanks, Willard.

  352. Deech56 Says:

    “Colorfoulness.” I like that. *The* purpose? “Doubt is our product.”

  353. PS Says:

    willard,
    When you say “So much to do for auditors, so little time” I see what you mean. For instance there is this phrase from one of your Steve quotes – “climate community’s desire to rebuild trust”

    “climate community”, “desire”, “rebuild” and “trust” all need auditing before we can evaluate the sentence in which the phrase appears much less the post in which the sentence appears.

  354. Jim Bouldin Says:

    MikeN:

    “You point out 400 proxies that correlate to temperature. If we assume a 10% random correlation level, then this is maybe 320 valid proxies and 90 correlated by chance. Again 2/3 (this number is sensitive to the types of proxies you include in the count) did not correlate. I’m working on how sensitive this would be.”

    How sensitive what would be? The absolute number doesn’t matter (other than w.r.t. sample size related uncertainty issues), and if you mean the percentage passing, that’s trivial: it doesn’t change with type of proxy, unless the AC changes drastically.

    It was 484, not 400, and (1209-484)/1209 = 0.6 that didn’t pass (and there’s a difference between “didn’t correlate” and “didn’t pass the screening criterion”). There were probably a whole bunch of them that correlated, but not at the specified criterion.

  355. Jim Bouldin Says:

    There is a great editorial in Nature this week, titled “Heart of the matter” that many will be interested in. It discusses the topic of this post. Among the many quotable phrases is this one:

    “Many climate sceptics seem to review scientific data and studies not as scientists but as attorneys, magnifying doubts and treating incomplete explanations as falsehoods rather than signs of progress towards the truth.”

    And they appear not to have any awareness of this at all I might add.

  356. Bart Says:

    Chris (on July 23rd) echoes what Amac wrote earlier: That most people commenting on Tiljander have no real grasp of the issuies at hand, and that in the end it is not such a significant issue to the larger question of AGW anyway.

    Amac, did I get that right? In the spirit of reconcilliation and such it would be nice to voice agreement with the other side at each opportunity that arises.

    Craig Loehle,
    Not sure what your dig at your first comment here is supposed to refer to, but everyone is welcome to comment here (provided they keep themselves to some basic groundrules of respectful communication as laid out in the very lenient comment policy).

  357. Jeff Id Says:

    Jim,

    I made a post to demonstrate the weakness in your commentary here. You really should read the comments in the thread, there are quite a few sharp people there.

    If you would like to discuss the issue, you are welcome to guest post on it at my blog. – I don’t moderate much but you will find qualified discussion.

  358. Novick Says:

    >How sensitive what would be? The absolute number doesn’t matter (other than w.r.t. sample size related uncertainty issues), and if you mean the percentage passing, that’s trivial: it doesn’t change with type of proxy, unless the AC changes drastically.

    The AC does change with type of proxy drastically.

    >It was 484, not 400, and (1209-484)/1209 = 0.6 that didn’t pass

    Luterbachers raise that number a bit more, as they are built from temperature and guaranteed to correlate.

    >(and there’s a difference between “didn’t correlate” and “didn’t pass the screening criterion”). There were probably a whole bunch of them that correlated, but not at the specified criterion.

    My other technical question that I forgot. Is a correlation level of .106 high or low? If a proxy gets to .099 correlation, is that valuable? Also, the program actually doesn’t just pick ones that correlate to temperature, but picks proxies that correlate to either the local temperature, or the temperature of the nearest gridcell. Not sure how much this would boost the probabilities for random proxies.

  359. AMac Says:

    Bart (23:08) —

    When I was in grad school, attendance at student seminars was mandatory, and some of those hours really dragged. One lab had a particularly complicated and boring set of projects, concerning some virus or other.

    I finally managed to tune in for one of these talks. It turns out that certain strains of Coxsackie viruses target the heart, with the result that masses of muscle cells are killed. In mouse as in man, this leads to heart failure, sometimes followed by death.

    The lab’s mission was to investigate the mechanism of this myocarditis.

    Using a fairly arcane set of reagents, they were able to interfere with different parts of the mouse’s immune response to Coxsackie virus infection. When they got around to disabling a certain obscure subpopulation of T cells, this is what they found: the mice survived.

    Their heart cells were teeming with virus.

    But there was almost no damage to their hearts.

    At first, this seemed like a puzzling set of results: more virus, but less viral injury?

    With hindsight, the explanation was clear. Most of the myocarditis wasn’t caused by the viral enemy. Instead, the damage was due to over-reaction of the body’s own defensive mechanisms,

    Sometimes, unimportant details aren’t. And I learned that virology can be pretty darn interesting.

    I’m fond of that story.

  360. danolner Says:

    Nice story Amac. Though that’s generally the case for viruses, isn’t it? Flu kills people the same way, by triggering an over-keen immune response, hence why it can get young, strong folk.

    Find myself still spinning in circles trying to understand where you’re coming from. “Sometimes unimportant details, aren’t”…? I’m reminded of your line:

    “1. What clothes is the emperor wearing, if any?

    2. What are the implications, if any, of the emperor’s wardrobe? For the royal court, for society as a whole, etc.

    If the best we can do with #1 is akin to “I don’t know, and it doesn’t matter,” it becomes very hard to think clearly about #2.”

    Some examples, see what you think. “Unimportant details” is the point in Kuhn’s paradigm shifts where whole theories pivot. Kuhn: “proponents of competing paradigms are always at least slightly at cross purposes. Neither side will grant the other all the non-empirical assumptions that the other needs to make its case.” So what do you think might/should happen if these unimportant details become important? Are you expecting some fundamental scientific shifts, or just a correction in one or two statistical tools?

    We’re not talking about a competing paradigm, though, are we? Not in the scientific sense. You have have a question about some proxies and calibration, and then some suggestive points about immune responses and emperors clothes that, to me, read as insinuations about people’s motives (‘the royal court’). That is, it doesn’t have any actual scientific content. Your calibration question, of course, does: it should have a yes or no answer (though the supplementary question, `does it matter’, is important, as is the next one, `matters for what?’)

    Perhaps there is a scientific point to what you’re getting at. Let’s assume there is. Then there are two possible kinds of paradigm shift if it turns out the emperor’s nekkid/it’s all a massive immune response against an assault from a paradigm-shifting blog-science virus (er, though I may have misunderstood the moral of your story…!)

    Let’s call the two types of paradigm shift Galilean and Einstein-like. In the former, the shift turns the science around by 180 degrees and the usurped theory is cast in the darkness. In reality, of course, it doesn’t quite, since there wasn’t any science to oppose, only religious dogma. A lot of climate skeptics like the metaphor for precisely that reason.

    Second, Einstein overturns Newton – but unlike Galileo vs the Bible, both frameworks are spectactularly good at physical prediction. For me, that’s the central point: *any* successful scientific theory at this point needs to do pretty damn well at not being falsified.

    The key point for me, again: a better theory may come along, even one that completely overturns something as fundamental as assumptions about space itself. I like Kuhn’s “non-empirical assumptions”; the idea of “emperor’s new clothes” not so much, though we’re talking about the same thing there, aren’t we? (If we’re just talking about errors in one statistical method, it sounds like hyperbole.)

    But crucially, the two theories had to be pretty close to each other, predictively, to have survived that long. The arbiter is reality. Ultimately, it’s reality that provides the stability that scientific theory has, not people’s motives or level of clique-i-ness. (Is that a word?) Conclusion: if anyone comes up with some incredible paradigm-shifting scientific findings at this point, it’s going to change the outcome for anything very marginally, if at all.

    I’m talking into the wind a bit, since I’m still stumbling about trying to understand where you’re coming from.

  361. Deech56 Says:

    AMac, not to be to picky, but your sentence should read “…immunology can be pretty darn interesting.” That was my field of study. Very interesting thoughts, Dan Olner.

  362. close Says:

    I know it is tedious that Tiljander keeps being brought up, but it equally tedious that people don’t acknowledge the problems with the series. Jim Bouldin is under a fundamental misunderstanding about the problem with the sediments – it is not just a spurious correlation, it was a clear mistake made by Mann et al in the screening procedures (by inverting the series in a one-sided test).

    This can be explained very simply:
    1) Mann et al used a screening process for proxies
    2) The screening test was one-sided in cases where a definite sign could be a priori reasoned on physical grounds
    3) the sediment series were included in the cases where a definite sign could be reasoned
    4) a one-sided test was used on the sediment series
    5) the series was corrupted in the calibration period by land use
    5) the sediment series should have failed the one-sided screening test if the test was correctly designed (testing for the physical meaning the right way up)
    6) The series actually passed the test because (a) the series was corrupted in the calibration period by land use and (b) the series was tested upside down

    The alternative is that Mann et al actually believed the series orientation used was correct on physical grounds. If that was the case, they should have explained why they had a different interpretation to Tiljander herself, This explanation seems quite unlikely.

    It would be nice if Jim, Eli, Willard etc would acknowledge this clear mistake and not just treat it as a statistical artifact.

  363. MIkeN Says:

    Close, that is easy for others to acknowledge if they review it clearly. Not so easy for Mann to acknowledge because, it meant he used data upside-down, real embarrassing. Also, he denied using data upside-down when asked about it. Even worse, and it appears that many others were fooled by this denial.

  364. MIkeN Says:

    Close, 6b is a point that doesn’t get emphasized enough. People like to excuse it as Mann didn’t understand Steve McIntyre’s statement about upside-down axes, when in fact the issue was calibration. However, even if the calibration was not an issue, for the CPS version, the proxy was used upside-down.

  365. Jeff Id Says:

    “Jim Bouldin Says:
    July 16, 2011 at 01:22

    I remember looking at the post you cite some time ago. Jeff Id has *no idea* what he is talking about, and he has formulated an entire theory of bad science based upon it. This is evidenced very clearly by the following statement he makes there:”

    Jim,

    You are good at throwing around conclusions, yet it is obvious to me and a few others, that you have little understanding of the paper. Therefore, since you are so certaint that I’m without a clue, I flatly challenge you to find any single error in my analysis of M08.

    My guess from your commentary is that you don’t have half of the skill set required to pull it off. If you did, you wouldn’t be making these arguments.

  366. willard Says:

    On the 2011-07-29 at 7:49, in a thread with more than 350 comments, an onlooker that seemingly never commented (at least under the name “close”) at Jeff’s, at Steve’s, at Lucia’s, at Keith’s, nor at Judith’s, observes that the Tiljander affair keeps resurfacing from time to time.

    That our onlooker prefaces his observation by saying how tedious is l’Affaire Tiljander deserves to be acknowledged. Contrary to what assumes our onlooker, we should also remind ourselves that perhaps not everyone believes l’Affaire Tiljander is *that* tedious. We do live in a big, colourful world, with a broad range of tastes, in matters of scientific criticism as in everything else.

    Our onlooker also observes that *people* do not acknowledge *the problems* of the series. This observation deserves due diligence. There are at least two readings of that statement:

    (Strong) No one from Mainstream ackowledge the problems of the series.

    (Super Strong) Not everyone from Mainstream acknowledge the problems of the series.

    The Super Strong version rests on the assumption that everyone from Mainstream *must* acknowledge the problems of the series. This assumption deserves due diligence, but not now. So much to do, so little time.

    The strong version is perhaps easier to analyze. It should suffice to recall what PeteB, on the 2011-07-23 at 11:11, already observed:

    > [E]veryone that I have found from the ‘consensus’ side that has looked into it seems to have the same viewpoint as you on the technical issue.

    http://ourchangingclimate.wordpress.com/2011/06/22/climate-science-scientific-method-skeptics-not/#comment-13708

    This echoes what has been said: *everyone* seems to be Mainstream Mud. This also echoes what is presupposed by a question AMac already asked the auditors: why *noone* (not even Mike, Gavin, nor any other members of the Kyoto Flames) seems to stand up for Minority Mud? (If nobody stood up for Mainstream Mud, the Strong version of our onlooker’s claim would obtain.)

    ***

    In any case, I believe that AMac agrees with PeteB, since on the 2011-07-23 at 11:34 he said:

    > You’re right, concerning Andy Russell, Deep Climate, and Martin Vermeer conceding the uncalibratibility of Tiljander’s data series to temperature.

    The next paragraph starts with an interesting expression:

    > The problem is this: [...]

    http://ourchangingclimate.wordpress.com/2011/06/22/climate-science-scientific-method-skeptics-not/#comment-13710

    In this specific instance, the device looks a lot like a “Yes, but.” Notice the swift shift from is being conceded to what is *the problem*. Notice the restriction of the agreement: three names. Saying

    > Everyone PeteB found believed that *p*.

    does not seem to have the same effect as

    > A, B, and C believed that *p*.

    Shifting back and forth from definite (like Vermeer and DC) to indefinite epithets (like *people*, *everyone* and *pro-climate consensus advocates*) partakes of two modes: accusing and conceding. This speech behavior can be witnessed in this very thread: accusing brought indefinite epithets, conceding definite ones. This seems quite natural: the accusing party tries to maximize the scope of the scientific criticism, while the conceding party tries to minimize it.

    ***

    Since PeteB seems confident enough to talk about “everyone from the consensus side”, auditors should wonder if any other name could be brought in the discussion. Here is for instance Lazar’s testimony (on the 2010-02-13 at 5:30) at Michael’s:

    > To my judgment Mann and Jones have both behaved dishonestly — Mann in his ultra-defensive refusal to admit error, Jones whilst digging himself deeper in the FOI kerfuffle. I would say the same for Lucia and McIntyre, e.g. Lucia’s interminable dodging of Nick Stoke’s simple request here [http://rankexploits.com/musings/2010/multi-model-mean-projection-rejects-gisstemp-start-dates-50-60-70-80/#comments], and McIntyre’s green carding of Douglass et al. But can I judge ‘what the science says’ from any of these issues? Hell no. Can I extrapolate to the whole field from Mann and Jones? Hell no.

    http://initforthegold.blogspot.com/2010/02/grim-amusement.html?showComment=1266067850166#c738395479135694029

    Auditors will observe that this comment was directed at AMac. To be sure if he’s Mainstream Mud, reading the discussion that ensued, however tedious, might be necessary. Perhaps this is a proper time for auditors to recall the Physicist [1]:

    > [T]ry to give all of the information to help others to judge the value of your contribution; not just the information that leads to judgment in one particular direction or another.

    because its reading is never tedious, however idealized it may portray scientific criticism. At the very least, we could meditate on this quandary: who’s burden is it to properly narrow down, disprove, and bookkeep the claims of our own scientific criticism? Auditors that enjoy puzzle will no doubt like to chew a bit on the tension between our Physicist Golden Rule, falsification, and scientific criticism.

    ***

    When interpreted in its Strong sense, and considering PeteB’s testimony, our onlooker’s observation can only be false. This implies, of course, that we take our onlooker’s observation at face value. But notice how the claim is built around *people*, which is quite indefinite, even more so than *Pro-AGW consensus advocates*. It is so muddy as to make us wonder if its truthyness really “matters” at all.

    And yet this is so frequent as to deserve due diligence. Auditors should recognize that it is a very neat way to take for granted that this question begs an answer. For a first comment ever made hereunder, our onlooker sure looks like a natural rookie.

    [1]: http://www.lhup.edu/~DSIMANEK/cargocul.htm

  367. AMac Says:

    danolner (July 29, 2011 at 10:07) —

    > Find myself still spinning in circles trying to understand where you’re coming from. “Sometimes unimportant details, aren’t”…?

    That’s it.

    At this point in the thread, another goal is to not repeat myself. That’s boring for both of us.

    As far as Kuhn, paradigm shifts, Newton, and Einstein — I wouldn’t put Mann’s work in that category. Or Tiljander’s. Much less the mistakes that Mann08 made in building upon Tiljander03.

    By including Tiljander03’s data series, Mann08 made an implicit but very strong claim: that the Tiljander03 data series are suitable for use as paleotemperature proxies (for rebuttals to “but Mann08’s described potential limitations in their text,” see upthread).

    I’ve made an explicit and very strong counterclaim: that the Tiljander03 data series are unsuitable for use as paleotemperature proxies by Mann08’s methods (I think they are likely unsuited for use by any methods, but better to keep things simple).

    In their Reply to McIntyre & McKitrick in PNAS, Mann08’s authors strongly — if cryptically — rejected a very similar claim (see upthread for discussion).

    Gavin Schmidt at Collide-a-scape, Jim Bouldin upthread, and others elsewhere have indicated that the counterclaim is trivial, or false.

    As willard (07:04) notes, other Mainstream scientist/advocates have agreed with part of the counterclaim — specifically, with my assertion that the Tiljander data series cannot be directly calibrated to the instrumental temperature record.

    (That leaves open the fallback position that Mann08 doesn’t require direct calibratability, but now we’re deep into Racehorse Haynes territory.)

    That brings us to the question of “does it matter? (matter for what?)”

    For the latter, the obvious and useful answer is “matters with respect to the objectives of the paper.” As the authors explained that purpose in the Abstract and, in Mann08’s case, in press releases.

    A mistake might be as trivial as a typo, or a split infinitive. Or, a mistake might concern a detail that’s key to an argument that underpins a central thrust of a paper. It depends.

    Mann08 brought non-dendro proxies into their paleotemperature reconstructions to rebut criticisms of tree-ring-based recons. In particular, McIntyre and others have claimed that tree-ring data series are suspect, and that the analytical methods used by Mann and other Mainstream climatologists are deficient.

    Mann08 showed (1) non-dendro proxies yield recons that are generally similar to the recons yielded by tree-ring proxies; and (2) recons constructed by different methods (EIV and CPS) yield recons that are generally consistent with each other.

    With Tiljander, (1) can be argued to succeed. Without Tiljander, (1) fails. (See upthread.)

    (This failure has adverse implications as to the robustness of Mann08’s methods, but that’s a subtler argument — and we haven’t done well with subtle, on this thread.)

    Thus, the use of the Tiljander proxies is not a trivial detail. It’s crucial to the conclusions of the paper, as the authors laid them out.

    … But now I’ve largely repeated myself.

    Of course, there’s a reason that I keep coming back to this simple and somewhat boring narrative.

    That’s what this episode is about.

    If this was a story about proxies for precipitation, or for economic activity, or for a kangaroo population — I don’t think we would be having such a drawn-out conversation. The evidence and the logic are fairly straightforward. However, it seems to be very hard to use the tools of ordinary science and ordinary life to approach controversies related to AGW.

  368. Jeff Id Says:

    Jim,

    “It lowers it, because an r criterion of 0.128 (the original criterion of .10 goes to .128 when accounting for AC), with n = 146 years, is almost certainly << a p criterion of the same value. Thus, I think my original statement that about 154 should pass the screening by chance, is high, and possibly way high."

    I think you should spend some more time on this topic. If you actually correct for true autocorrelation, you will find that this number is actually low — WAY low. I have done it and found a potential SNR with a single test pass of under 5%. If you consider the fact that M08 used a pick two scenario, it is easily demonstrable that these 484 series would pass by accident alone.

    I also used the true autocorrelation values on M07 and found that paper dramatically invalid as well.

  369. J Bowers Says:

    Are Jeff Id, MikeN and AMac actually able to summarise *the* probelmS into a concise list, and actually stick to such a list? All it seems to the disinterested, or possibly even interested, observer right now is a lot of moving of goalposts. Small wonder many groan.

    If they can’t, why can’t they?

  370. AMac Says:

    J Bowers —

    Moving of goalposts?

    That seems like an extraordinary claim, at least for the case I know best (my own).

    I alluded to a quite different issue at 12:47, supra: that I have to try not to bore with repetition of the same points. You can use that comment as the concise list of the main problems I see with Tiljander-in-Mann08.

    In your opinion, how does what I wrote there represent “a lot of moving of goalposts,” compared to what I’ve written earlier (e.g. upthread, or at my blog)?

  371. J Bowers Says:

    AMac, it’s neither a single issue nor you specifically. I didn’t ask for Jeff Id or MikeN or AMac to summarise their problem/problems, I asked if you guys were capable of summarising all of the problems?

    Why don’t the three of you get together and write a paper? Share the work. Or something. Steve McIntyre’s another with problems. There are four of you who could do it.

  372. AMac Says:

    J Bowers —

    OK. Can I take it that moving-of-goalposts isn’t a problem as far as my blog commentary on Tiljander-in-Mann08?

  373. J Bowers Says:

    Not as an individual. But the sum total amounts to disjointed noise with no signal…. or signalS.

  374. AMac Says:

    > But the sum total amounts to disjointed noise with no signal…. or signalS.

    For signal, can I refer you back to the content of my recent comment of July 30 at 12:47.

    By including Tiljander03′s data series, Mann08 made an implicit but very strong claim: that the Tiljander03 data series are suitable for use as paleotemperature proxies…

    I’ve made an explicit and very strong counterclaim: that the Tiljander03 data series are unsuitable for use as paleotemperature proxies by Mann08′s methods…

    [This counterclaim] matters with respect to the objectives of the paper. As the authors explained that purpose in the Abstract and… in press releases.

    Mann08 brought non-dendro proxies into their paleotemperature reconstructions to rebut criticisms… that tree-ring data series are suspect, and that the analytical methods used by Mann and other Mainstream climatologists are deficient.

    Mann08 showed (1) non-dendro proxies yield recons that are generally similar to the recons yielded by tree-ring proxies; and (2) recons constructed by different methods (EIV and CPS) yield recons that are generally consistent with each other.

    With Tiljander, (1) can be argued to succeed. Without Tiljander, (1) fails…

    Thus, the use of the Tiljander proxies is… crucial to the conclusions of the paper, as the authors laid them out.

    … But now I’ve largely repeated myself.

  375. jeff Id Says:

    J Bowers,

    I speak for myself, if you don’t mind. I have summarized the primary issues in many different ways. Any of the primary issues prove the paper worthless so I wonder, why do you want more than one? and then, why don’t you just suss one of them out yourself and tell me where I’m wrong?

  376. J Bowers Says:

    It’s obviously just wishful thinking on my part.

  377. Close Says:

    Willard,

    Your long meandering post is faulty in many respects. To start with, this is certainly not my first comment on the topic, notwithstanding your repeated assertion. On the contrary, I have posted a number of comments on Tiljander at CA under my name (Close). You need to brush up on your googling skills. (You can look at this thread for example: http://climateaudit.org/2011/07/06/dirty-laundry-ii-contaminated-sediments/) So you start off your thesis on a mistaken premise and it unfortunately goes downhill from there.

    I never said “all people” and you should not assume that. My frustration comes from comments in this very thread above that assert that the Tiljander problem is a normal spurious correlation that you see in statistics (it is not, it is a calibration error). It comes from the failure of Mann et al to plainly admit the error. It comes from the disingenuous PNAS response. It come from Real Climate not conceding the slightest point on this clear error. It comes from Gavin on a long thread over at Keith Kloor’s refusing to admit any weakness on this point.

    As Steve M has said may times, a corrigendum needs to published for M08. Until this happens, my post remains valid. I hope you will agree.

  378. J Bowers Says:

    Make that a paper by Id,AMac,MikeN,McIntyre,Close. Citing it could be fun.

  379. willard Says:

    Close,

    I apologize for having missed your conversation with Nick Stokes. It made little sense you were a rookie. I hope we can agree that this impression was justified, that it did not affect the main part of my comment, and that your conversation with Nick brought you some experience.

    Sorry to hear about the source your growing frustration, which agreeably does not affect what I said so far in the thread.

    Without further ado, here is the wish you expressed:

    > It would be nice if Jim, Eli, Willard etc would acknowledge this clear mistake and not just treat it as a statistical artifact.

    First, since you “never said all people”, I’m not sure what to do with your

    > etc [et al.]

    We need an interpretation for this *et al.* that does not extend to “all people”. Do you have a specific list of “people” in mind?

    Second, we see that “it would be nice” to grant your wish. This is not unlike AMac’s invitation to Eli and Chris Colose above:

    > A brief aside. Pursuant to the earlier conversation I had with willard, could I invite you each to weigh in on two questions? Of course, you’re under no obligation to answer — just that I’m curious.

    http://ourchangingclimate.wordpress.com/2011/06/22/climate-science-scientific-method-skeptics-not/#comment-13767

    Declining this invitation may frustrate you furthermore. This would not be nice at all. By chance, here is something that could perhaps satisfy you:

    Many members of the pro-AGW consensus have declared that Finnish lake mud is no good proxy. For what it’s worth, my personal impression is that there seems valid criticism [the calibratability of the Tiljander data series to the instrumental temperature record] being adjudicated in favor of the contra-AGW consensus community.

    I can also concede that many members of the pro-AGW consensus community did respond “suboptimally,” to use an auditing term.

    http://ourchangingclimate.wordpress.com/2011/06/22/climate-science-scientific-method-skeptics-not/#comment-13587

    WordPress autofilled “neverendingaudit” by taking the first part of my tumblog’s name. I hereby declare to be the author of the quote above. I still believe that l’Affaire Tiljander does not look pretty.

    I could add that the validity of your post is independent from the causes of your frustration. And also offer to talk about why this mistake *matters* so much as to frustrate you. But I won’t meander.

    ***

    The fortrightness of your comments makes me wonder if you mean more than an invitation. Here’s one sent to Nick on the 2011-07-08 at 13:52, with this a spirited ending:

    > Will you salvage a shred of credibility and accept this [the correlation was inverted because Mann et al made an a priori decision to invert it]?

    http://climateaudit.org/2011/07/06/dirty-laundry-ii-contaminated-sediments/#comment-297763

    If I read the conversation correctly, Nick did salvage a shred of his credibility, and you did concede that the topic of your request was a sideshow. Am I correct?

    It would matter to know the difference between your request to Nick and the one in this very thread. Credibility followed from Nick’s responsiveness. By contrast, only some kind of nicety is expected to follow from the answers from the *people* here. I hope you do not mind if that channels my inner John McEnroe.

    ***

    In a recent post, Jean Goodwin presented the concept of *commitment* in a dialogue:

    > Key to any successful debate is managing the basic responsibilities: who is obligated to defend what. If responsibilities aren’t limited, a debate e.g. over some immediate political issue can easily devolve back into a debate on how we know anything at all–philosophically interesting, perhaps, but far from the original topic. And if the responsibilities aren’t clearly defined, participants can find the debate slipping from one issue to another in what may be an unproductive fashion.

    http://scientistscitizens.wordpress.com/2011/07/26/debate-in-the-blogosphere-a-small-case-study/

    To illustrate this idea, she described her experience on CA, and explained how commenters tried to burden her with commitments that were not hers. Any commenter with some experience (including you, after your exchange with Nick) should recognize this common trick. In this very thread, AMac had a variety of questions to answer. CA readers see how Nick Stokes is being asked to defend about any choice made by the Kyoto Flames [KF]. This oftentimes includes the blog curator:

    > More Omerta, Nick?

    http://neverendingaudit.tumblr.com/post/8195679480

    The commitments presumed by this question deserve due diligence.

    ***

    It would then be more than nice to know which commitments l’Affaire Tiljander force (not all) “people” to opine on on your list of claims. Let’s suppose, for example, that every commenter were required to to opine the calibration process of some Finnish lake mud to salvage some shred of credibility. Imagine one rabbit (let’s call it R*) demonstrating reluctance to opine on it. Then the Super Strong statement from my previous comment obtains:

    (Super Strong) Not everyone from Mainstream acknowledge the problems of the series.

    The credibility of R* would then be put into shreds. Would the Mainstream community be affected by this? Probably not: R* is only a silly rabbit, after all. (Let’s not contemplate the possibility that R* carries a virus affecting the human heart.)

    By chance, you did not mean *all* people. Only me and some other more. Yet I do not know if this is only for curiosity’s sake, or to shred credibility. Nor do I know what commitments warrants you to target not “all people”, only some. Nor do I know why these social niceties matter to you.

    I thank you for your kind words and hope this sounds less meandering to your scientist ear.

    [KF]: http://web.archive.org/web/20050222224714/www.climate2003.com/blog/hockey_team.htm

  380. Jim Bouldin Says:

    close at 13:49:

    “I know it is tedious that Tiljander keeps being brought up, but it equally tedious that people don’t acknowledge the problems with the series. Jim Bouldin is under a fundamental misunderstanding about the problem with the sediments – it is not just a spurious correlation, it was a clear mistake made by Mann et al in the screening procedures (by inverting the series in a one-sided test).”

    No, that’s not the case. I know what the contention(s) is/are about the Tiljander issue–that they either cannot be calibrated at all, or that an existing relationship with T was inverted. My point regarding spurious correlation was not with reference to that issue. It was with reference to the claim that spurious correlations cause hockey sticks, which was introduced into this discussion at some point.

    And yes, it is tedious.

  381. AMac Says:

    Jim Bouldin (15:57) —

    > My point regarding spurious correlation was not with reference to that issue [that the Tiljander data series either cannot be calibrated at all, or that an existing relationship with T was inverted].

    Well, it depends which of your comments you are referencing. The first time, the topic was the spurious correlation the Mann08 established between the Tiljander data series and the instrumental temperature record.

    In my comment of July 24 at 08:25, I wrote,

    2. Three of the four Tiljander data series passed screening and validation… This success captures a major shortcoming of Mann08′s procedure: it cannot distinguish meaningful correlations from spurious ones. Statistician J. Udny Yule flagged the challenges of nonsense-correlations in 1926 (PDF)…

    Jim Bouldin countered with sarcasm at 17:53 with:

    Well by “nonsense” I guess you mean spurious. This is exactly what I’ve been addressing, and which was addressed in the paper with the very quote you cite. Are you paying attention? And as mentioned by me above, you do realize that you are now addressing a much wider issue that affects anyone who uses a known correlation to predict an unknown value, don’t you?

    I responded at 19:01

    …The inadvertent misuse of the Tiljander data series is a single instance of a much wider issue.

    Later in the thread, MikeN brought up “spurious correlations” a second time, writing on July 25 at 22:55:

    Jim Bouldin, I see your point about Jeff Id’s criticisms, and am reworking his code to do a more exact replication to my satisfaction, as he is using different correlation numbers.

    The issue I have along the lines of Jeff Id, there is the suppression of historical signal that comes from calibrating to the modern temperature level. This is separate from the spurious correlation issue.

    Jim Bouldin replied (July 27 at 18:07) —

    Yes, very different topics. The former will always be a potential issue.

    [snip]

    I was trying to make the point that the individual trees at the sites selected are responding +/- in concert to *some* environmental signal(s)–the .5 criterion is the mean correlation of all the cores with each other–evidence of a common forcing/signal. This means that the trees were in fact responding to a signal(s): another piece of evidence negating the argument that spurious correlations in calibration period are what produce hockey sticks, especially when you consider that the inter-tree correlations with each other are computed over the entire chronology, not just the calibration period.

    It is encouraging to see Mainstream scientists/advocates grasping the first part of the Tiljander-in-Mann08 puzzle: that the correlations of the Tiljander data series to the instrumental temperature record are spurious.

    As far as I know, no Mainstream climate scientist or advocate has taken the next step, and figured out the implications of this insight as far as the robustness of Mann08’s non-dendro-proxy-based paleotemperature reconstruction results.

  382. luminous beauty Says:

    Einstein was known to caution his students to make their arguments as simple as possible, but to take care not to make them too simple.

    In my opinion Amac’s arguments suffer from over-simplification. Most seriously, in his unwillingness or inability to probe the depths of ‘physical reasoning’ and rely, instead on the most superficial interpretations of some interested party’s (McIntyre, Tiljander, Mann, et alia) ‘testimony’ concerning ‘physical reasoning’.

    This causes me to suspect Amac’s expertise might lie in the legal profession rather than the scientific. I may well be wrong. I have very little evidence on which to base my conclusion, but is it not a reasonable inference? It is just so with Amac’s arguments. They are superficially reasonable, but lacking in much evidential corroboration, and are ultimately prejudicial based on who he finds more credible. I.e, they are mere opinions that don’t amount to a hill of beans.

    JeffId, makes, again, the error in thinking the auto-correlation coefficient lies additively on the noise side of the S/N equation, rather than being a physically deterministic distortion of the annualized signal, correction for which strengthens the signal side of the equation.

  383. AMac Says:

    luminous beauty (20:14) —

    > This causes me to suspect Amac’s expertise might lie in the legal profession rather than the scientific.

    Well FWIW, I trained as a biologist and have had only occasional brushes with the legal profession. And no great sympathy for legalistic reasoning. But I’m pseudonymous, so make of that what you will.

    > [Amac’s arguments] are superficially reasonable, but lacking in much evidential corroboration

    In that regard, what I have done is download Tiljander’s data series, format them as a spreadsheet, graph them, perform simple statistical analyses on them, and discuss them. I’ve provided links to relevant data archives, peer-reviewed literature, blog posts, spreadsheets, and figures. I’ve made evidence- and logic- based arguments as to why:

    * we can know that the Tiljander data series cannot be meaningfully calibrated to the instrumental temperature record, and

    * there are good reasons to think that Mann08’s key results concerning non-dendro paleotemperature recons depend on the use of the Tiljander data series. These findings are probably in error.

    I’ve also discovered other subtler but likely important mistakes in Mann08’s uses of the Tiljander series, but as this thread shows, it’s hard enough to have a productive discussion on the more clear-cut issues.

    I suspect it might not be possible to provide enough evidential corroboration to satisfy many of the readers of Bart’s blog. That’s okay, it’s a wide world. At any rate, thanks for considering what I’ve had to say.

  384. Jim Bouldin Says:

    Such a phenomenal boatload of confusion.

    Amac:
    There have been several charges levied here. One is that it’s not possible to calibrate the Tiljander proxies at all, due to human effects over-riding the T signal. The second is that, yes, it is possible to calibrate them, but that Mann et al mistakenly inverted the direction of the correlation. The third is that a spurious correlation can arise simply by chance.

    These are all different issues and they’ve all been thrown together into a stew here (along with several other diversions, such as your incredulity at Esper and Frank’s comment and “post-hoc” analysis) such that it’s no longer possible to follow the discussion, because it’s an incoherent jumble of accusations and various statements of uncertain relation either to each other or to some central coherent argument. In other words, a kitchen sink approach.

    It was this third option that I was referring to when I stated with my original comment here that you were addressing the issue of the “implications for climate reconstructions, of biological or physical processes that change in rate or (even worse) in direction.” That’s what I thought at the time–for whatever reason, I thought you were onto that issue.

    However, based on statements at your blog and here, I began to doubt that that was really the generalization you were onto, which instead appeared to be the typical denialist innuendo devoted to the idea that the Tiljander issue is just the iceberg tip for a lot of either sloppy or fraudulent science. This is evidenced by the fact that you did not respond to my statements that what you were raising was essentially Loehle’s points about changing physical relationships between temp and proxy measure. Some time later you did mention it, but I’m not sure you realized it on your own, apart from me bringing it up first. And what I read at your blog gave me no indication that you were really onto that idea either, especially after you used an example from clinical trials to illustrate, not some statistical concept, but rather the fact that some group had done some bad research.

    And then of course there is the whole issue that I alluded to earlier. Simply stated, I really don’t trust that *anyone* who harps constantly on the works of Mike Mann and Mike Mann alone, has any real interest in advancing the science. And why? Because if their concern was really with the various shortcomings in the field of paleoclimate reconstructions, they would be bringing up issues from papers by different authors on different topics, and the discussion would be MUCH more generalized than it is.

    This is why Loehle’s paper in 09 is by far the most significant work discussed here–because he addresses a general, potential issue. Not saying his contentions are necessarily critical to the science or hold true in practice–that depends on some further analyses, including via development of new analytical methods. But at least he addresses the most widely used proxy (tree rings), and has a definite and reasonable biological basis for his contentions, which give them some potential significance. It is ONLY in that context that Tiljander proxy issues have significance–that is, if the physical relationship changes direction depending on physical conditions, then you could have a problem in interpretation

    And *that* was my point with my very first comment here.

  385. Eli Rabett Says:

    Well since AMac asked.

    Bismark was once asked how many people understood the dynastic succession issues concerned with Schleswig Holstein.

    He responded three

    The first is dead
    The second went mad
    And I myself have forgotten all about it.

    Demands that Eli take a position for or against are somewhat of that ilk. Eli’s position is that he trusts the weasel. That and the update over there.

  386. AMac Says:

    Eli Rabbett —

    > Demands that Eli take a position for or against are somewhat of that ilk.

    I made no demands. I asked a question. Which you answered as you saw fit. Thank you for this subsequent clarification.

    .

    Jim Bouldin at 12:49 wrote:

    There have been several charges levied here. One is that it’s not possible to calibrate the Tiljander proxies at all, due to human effects over-riding the T signal. The second is that, yes, it is possible to calibrate them, but that Mann et al mistakenly inverted the direction of the correlation. The third is that a spurious correlation can arise simply by chance.

    Yes, much has been said in this thread. Earlier, I even alluded to shoes, ships, and sealing wax, which are quite far afield.

    However. What I have said about the main issue of Tiljander-in-Mann08 has been consistent and simple. Really simple.

    The nth iteration was on July 31, 2011 at 20:40. Here is n+1 —

    * we can know that the Tiljander data series cannot be meaningfully calibrated to the instrumental temperature record, and

    * there are good reasons to think that Mann08′s key results concerning non-dendro paleotemperature recons depend on the use of the Tiljander data series. These findings are probably in error.

    Any correlations of any of the Tiljander data series to temperature for the 1850-1995 period were bound to be spurious. Why? Due to human effects over-riding any T signal.

    That said, to get tediously basic: of course it is possible to correlate any string of numbers against any other string of numbers. Excel (etc.) will not gasp in dismay — instead, input the lists, and it will deliver a value for slope, intercept, r^2, and the like. The investigator (not the program) is responsible for guarding against spurious correlations.

    The specific correlations that the authors of Mann08 established for the Tiljander data series were as follows:

    XRD – correlation between XRD and T established in Mann08 was spurious; slope and r^2 thus meaningless. In direction Mann08 correlated higher XRD to higher temperature, the opposite of Tiljander03’s interpretation. Mann et al’s procedure thus mistakenly inverted the direction of the correlation.

    Lightsum – correlation between Lightsum and T established in Mann08 was spurious; slope and r^2 thus meaningless. In direction Mann08 correlated higher Lightsum to higher temperature, the opposite of Tiljander03’s interpretation. Mann et al’s procedure thus mistakenly inverted the direction of the correlation.

    Darksum – correlation between Darksum and T established in Mann08 was spurious; slope and r^2 thus meaningless. In direction Mann08 correlated higher Darksum to higher temperature, qualitatively the same as Tiljander03’s interpretation. Mann et al’s procedure thus did not invert the direction of the correlation.

    Thickness – correlation between Thickness and T established in Mann08 was spurious; slope and r^2 thus meaningless. Tiljander03 offered no interpretation of Thickness. Thus, the spurious correlation cannot be called inverted or not-inverted (upside-down or rightside-up) with respect to Tiljander03’s (nonexistent) interpretation.

    Jim Bouldin at 12:49 wrote:

    These are all different issues and they’ve all been thrown together into a stew here… such that it’s no longer possible to follow the discussion, because it’s an incoherent jumble of accusations and various statements of uncertain relation either to each other or to some central coherent argument. In other words, a kitchen sink approach.

    Perhaps you will find the immediately prior re-iteration helpful.

    Jim Bouldin at 12:49 wrote:

    It was this third option [that a spurious correlation can arise simply by chance] that I was referring to when I stated with my original comment here that you were addressing the issue of the “implications for climate reconstructions, of biological or physical processes that change in rate or (even worse) in direction.” That’s what I thought at the time–for whatever reason, I thought you were onto that issue.

    Yes, spurious correlations can arise by chance; the Yule (1926) paper that I cited earlier gives several examples.

    I do not recall anybody in this thread suggesting that the correlations that Mann08 established for the Tiljander data series “arose by chance” — what comment are you referring to? That strikes me as awkward and imprecise phrasing. It invites misinterpretation.

    Jim Bouldin at 12:49 wrote:

    However, based on statements at your blog and here, I began to doubt that that was really the generalization you were onto, which instead appeared to be the typical denialist innuendo devoted to… This is evidenced by the fact that you did not respond to my statements that… And what I read at your blog gave me no indication that you were really onto that…

    It would be fun to launch into a discussion of subtle and important concepts concerning multiproxy temperature reconstructions.

    Though the venture would be more fruitful and more pleasant if you could learn to omit inflammatory and pejorative insults from your writing — perhaps willard could help you with “denialist” and “innuendo”, for instance.

    It is my experience that this sort of productive exchange is not feasible, so long as participants cannot agree on what should be simple and important concepts concerning multiproxy temperature recons.

    If you would like an example of silly commentary on simple and important concepts concerning Tiljander-in-Mann08, follow the link that Eli Rabett provided above (August 1, 2011 at 02:53) to Stoat‘s threads on the subject.

    First, the idea of what sudoku is.
    Then, easy puzzles.
    Finally, difficult, complex puzzles.

    We don’t even have agreement about this idea-about-discussing-ideas!

    Simply stated, I really don’t trust that *anyone* who harps constantly…

    You say that my commentary hasn’t earned your trust. Simply stated, your commentary hasn’t earned my trust. So be it.

    Jim Bouldin at 12:49 wrote:

    It is ONLY in that context [Loehle's definite and reasonable biological basis for his contentions] that Tiljander proxy issues have significance–that is, if the physical relationship changes direction depending on physical conditions, then you could have a problem in interpretation.

    This is a difficult sentence to parse. I agree with you — If you are meaning to concur with the two starred sentences I repeated at the start of this comment (beginning with ” * we can know that the Tiljander data series cannot be meaningfully calibrated”).

    But I have the feeling that you are actually trying to convey something quite different.

  387. willard Says:

    Since Eli has voiced his opinion by way of metaphors at the beginning of the thread, they are worth repeating.

    Second, on the 2011-06-27 at 1:37, there are two metaphors:

    Mike N confuses circular firing squads for circling the wagons. Why is Eli not surprised?

    Mike N recommends that those of us interested in reality and meeting the serious threat of climate change should attack each other to make his job easier.

    First, on the 2011-06-25 at 23:03, the figures of speech were cruder and clearer:

    The attacks on Michael Mann (and Al Gore before him, and Joe Romm) use any brickbat that can be manufactured to try and get folk to disown them. This removes strong advocates from the discussion.

    Too often this tactic has been successful. **Eli really doesn’t give a crap if Al Gore is fat, Michael Mann bald, and Joe Romm shrill. What they are is mostly right, not perfect, but mostly right.** See what CC [Chris Colose] said about the IPCC above[.]

    I suppose the comment by Chris that was referred is this one:

    http://ourchangingclimate.wordpress.com/2011/06/22/climate-science-scientific-method-skeptics-not/#comment-13398

    Eli seems willing to concede that Mike Mann could have been wrong somewhere, and why not about Tiljander. Eli does not seem to think that this matters much in the grand scheme of things. Eli does seem cautious of the trust he puts into discussions made to achieve some kind “firing squad effect.” In a carotshell, Eli does seem to distrust what looks to it (i.e. Eli) like a mud or brickbat.

    When a discussion looks like a trap, no wonder why our rabbit hesitates before jumping in.

    ***

    Let’s compare Eli’s metaphors with AMac’s very neat viral story. Here is the conclusion:

    > Sometimes, unimportant details aren’t.

    http://ourchangingclimate.wordpress.com/2011/06/22/climate-science-scientific-method-skeptics-not/#comment-13793

    In my opinion, Eli’s and AMac’s images show different reasons why l’Affaire Tiljander matters. The images also show some level of distrust. Perhaps what matters and what in which we put our trust goes hand in hand.

  388. Eli Rabett Says:

    Weeelll to add (or to correct) what Willard wrote, Eli is fundamentally not very interested (and that is an exaggeration) in l’affaire Tiljander. Like everything else Steve McIntyre touches it is a giant stew of gish gallop, and frankly there is the day job to think about. As pointed out above it is a lot easier to trust the ones you trust who have some clue on this than go take a course.

    And yes, AMac you repeated requests that everyone admit you are right are demands. The usual blog comment defensive crouch

  389. Eli Rabett Says:

    Oh yes, elsewhere someone has pointed out that the problem with Michael Mann is not that he is bald, but the goatee.

  390. AMac Says:

    > AMac you repeated requests that everyone admit you are right

    It’s still conceivable that there are plausible defenses of the approaches taken by the authors of Mann08 with respect to the Tiljander data series. But so far, none have been offered by the Mainstream scientist/advocates who have engaged (however reluctantly) on this topic.

    And here we are, back with more Racehorse Haynes fodder.

    BTW I’ll be offline for a while.

  391. luminous beauty Says:

    Amac,

    Your simple statistics are not so simple. Being able to plug & play into Excel doesn’t necessarily mean you know what you’re about. Log normalization, if implemented properly, is a calibration step, which you keep claiming is impossible. What it does is make the two data sets to be compared orthonormal, or in simple language, proportionately equal, i.e., calibrated to the same scalar mean. Just glancing at your scatter plots indicates ur doin’ it rong.

  392. AMac Says:

    Thanks, Luminous. That’s a helpful observation. I’ll look into it when I get back, and probably update the post.

    I don’t see how the re-plotting will change any of the substantial pieces of the Tiljander-in-Mann08 story, but you never know. If it does, I will report it.

    > is a calibration step, which you keep claiming is impossible.

    I’m not sure what you are claiming that I claim is impossible. Could you quote or cite instead of paraphrasing?

  393. J Bowers Says:

    “Oh yes, elsewhere someone has pointed out that the problem with Michael Mann is not that he is bald, but the goatee.”

    Not just him. Shave it or grow a proper one. ;)

  394. Jim Bouldin Says:

    “You say that my commentary hasn’t earned your trust. Simply stated, your commentary hasn’t earned my trust. So be it.”

    You have a blog devoted entirely to the use/misuse of proxies from one lake in Finland by Mike Mann and coworkers. I’m working on more fundamental and general issues in dendroclimatology methods which I think have some significance and which I expect to publish, in more than one paper, in the peer reviewed literature.

    You and the other readers here can make of that whatever you like.

  395. MikeN Says:

    That’s fine Jim, This thread was about scientific method, and I brought up Tiljander in that context, that the failure to correct an error casts doubt on the credibility of the field. The specific analysis of Mann08 methods aren’t that relevant to this thread, but I am interested to hear about them.

  396. oarobin Says:

    fascinating thread but i am getting lost by the various arguments if i may summarize my understanding and have it corrected by the participants.

    Amac main point seems to be that by calibrating the tiljander proxies to the modern temperature period (1850-1995) Mann08 *necessarily* introduced an interpretation of the correlation of proxy values to temp that is the exact opposite of Tiljander in the pre 1720 period. This is because the post 1720 contamination of the proxy is so severe that CPS, EIV when calibrated and scaled against modern temp data results in an interpretation opposite to that of tiljander pre 1720. so when mann08 CPS, EIV methods projected temperatures back pre 1720 based on this 1850-1995 calibration they have an interpretation oppsite to that of tiljander.

    this i think amac argues should mean tiljander should be dropped from all the reconstructions in mann08. this further implies that the no dendro(now without tiljander and 3 other problematic proxies) does not validate back 1300 years and thus *the* central conclusion of mann08 should modified to reflect this change.

    now while this is a straightforward argument the issues are clouded by question of whether or not tiljander original interpretation of the proxies is correct, whether or not contamination reversed the interpretation of the proxies post 1720 , what is the quantitative effect of tiljander on CPS, EIV methods that are deployed in mann08 and whether or not the no dendro reconstruction was *the * central conclusion of mann08.

    Jim Bouldin point seems to take the specific instance of the tiljander reverse interpretation of proxy values to temp and genaralise it to various physical processes that operate in these proxies that vary in interpretation ,as the tiljander does, or whose rate of change vary , a topic i would love to hear more about.

    there are also unresolved nontechnical questions of what are the implications to mainstream science of the failure of Mann or PNAS to acknowledge and issue a correction based on the arguments put forward by Amac et al in the blogosphere?

    what is to be accomplished by the polling of scientist to agree/disagree with the arguments against mann08? is to get mann to issue a correction? or to have the behaviour of Mann denounced by his colleagues? others reasons?

  397. Jeff Id Says:

    Jim,

    If you ‘expect to publish’, I wonder if you would mind addressing my more poignent and relevent critiques. After all, you claim that I have no understanding. I do speak the language of paleo, so where am I going wrong?

    Seriously, I’ve admitted error more than any blogger on the climate internet. I won’t give easily but if you truly have a point, stop with the unsubstantiated comments and let me have it!

    To be fair, I don’t think you have the depth of Mann. He knows how to play the political version of science better than anyone. It is a team sport for him. You have given not one point which is difficult to refute and I think if you try and dispute any of my points directly, you will be exposed as a lawyer or shallow. Mann will either find the loophole or hide from it. Prove me wrong or even demonstrate some form of shallow misunderstanding on my own part and I will take it back.

  398. Jim Bouldin Says:

    “Jim, If you ‘expect to publish’, I wonder if you would mind addressing my more poignent and relevent critiques.”

    Sure Jeff, point me to the journal article(s) in which you have shown that the use of the correlation between two variables over the common interval, to predict the values of one of the two over the non-common interval, is now null and void, and I’ll see what I can do. As I said above, this is quite a profound finding that affects a very wide range of fields in science, so it would seemingly be suitable for at least one article in Nature or Science, wouldn’t you say?

    And good to see that you have a high opinion of your work, that should help.

  399. Jim Bouldin Says:

    MikeN:
    Yes I understand. Comment appreciated.

  400. MikeN Says:

    Jim Bouldin, in the comments of Jeff’s hockey stick posts, he actually wrote against what you attribute to him. I took that position when arguing in favor of a medieval warm period, which is what originally brought you into this discussion. I agree that you want to use items which are reliable temperature proxies, and not those that aren’t. That’s part of the problem with Tiljander after all.

  401. MikeN Says:

    oarobin, the CPS version is upside down not because of the calibration, but because of the way the data were put in the list of proxies, upside-down. If they were put right-side up, it would not have been used.

  402. J Bowers Says:

    Is it worth a reminder of the title, ‘How science does and does not work’?

  403. luminous beauty Says:

    Amac,

    I’m not sure what you are claiming that I claim is impossible. Could you quote or cite instead of paraphrasing?

    * we can know that the Tiljander data series cannot be meaningfully calibrated to the instrumental temperature record.

    In order to determine the correlation of two data sets of differing qualitative measurements, they must first be ‘meaningfully’ calibrated, i.e., normalized to the target measurement.

    If you can’t do this properly, you cannot say anything ‘meaningful’ about whether they are correlated or not, weak nor strong nor in between.

    Your conclusions about the series aren’t based on deep consideration of the physical reasoning about what would constitute mechanical, as differing from statistical, correlations, but predicated on Tiljander’s reductive ‘interpretations’, which are little more than simplistic speculative suggestions by a doctoral candidate, outside the bounds of her remit, being necessarily correct. I give you credit for making an amateurish stab at it, but it is really much more complex than you imagine.

  404. Jeff Id Says:

    Jim,

    Math is math. I have demonstrated a number of effects which I’m pleased to note have been verified by scientists in a number of journals. If you are truly interested see Christiansen and Zorita’s work. Some of which I disagree with. I do have a high opinion of my work, because I understand regression, have a VERY detailed understanding of M08, am ready to be wrong and have posted my own rationale endlessly. Since you have thoroughly lashed me for not understanding, I would have thought you could point out just one thing – one single little tiny thing- that I’ve not understood.

    I have pointed out in my recent article, many points you have missed, you’ve addressed zero. So let me know Doc, enlighten me.

  405. willard Says:

    Luminous Beauty, a nickname imposing glowing responsibilities, tells AMac:

    > Log normalization, if implemented properly, is a calibration step, which you keep claiming is impossible.

    AMac responds:

    > I’m not sure what you are claiming that I claim is impossible. Could you quote or cite instead of paraphrasing?

    I’m not sure what Luminous Beauty has in mind, but here is AMac Voldemort question:

    > Are the Tiljander proxies calibratable to the instrumental temperature record, 1850-1995?

    On the June 17th, 2010 at 2:23 pm, AMac answers his own question, saying he laid out a case laid out a case for a one word answer:

    > No.

    http://www.collide-a-scape.com/2010/06/16/the-main-hindrance-to-dialogue-and-detente/#comment-7705

    Here are the evidences provided in that thread at Keith’s. We can see Tiljander quotes, from her thesis and from her 2003 paper, showing that she cautioned against contamination of some part of the proxy she studied. We can read a summary of Tiljander’s interpretation of three of her proxies. We can read that AMac is at the time of the opinion that it was “an honest mistake.” We can read the non-responsiveness of the reply to M&M, at least regarding Tiljander’s interpretation. We finally read that Kaufmann (2009) corrected his manuscript to reconcile his use with Tiljander’s interpretation.

    ***

    I duly submit that AMac’s answer only looks categorical. Here’s what a categorical answer to Voldemort’s question would look like:

    > No, no one will ever be able to calibrate Tiljander’s proxies.

    This conclusion holds for an infinite amount of time and for virtually every rational beings, from perfectly omniscient and eternal Dutch mathematicians (imagine Bart with spiked white hair) to small furry creatures from Aplha Centauri.

    This conclusion does not obtain from the premises I’ve seen so far. It is, strictly speaking, what we call a _non sequitur_. With such a strong statement, no wonder the premises of the argument are too weak. That’s why evidence-based reasoning is seldom categorical.

    The formulation asks something about the possibility to calibrate. Unless we have an existential proof, it asks too much. Dispositional concepts are problematic.

    ***

    Here how an answer should look like, if we’re to make an evidence-based claim:

    > Considering Tiljander’s interpretation of the proxies and her warning that it might be “a demanding task” to calibrate the physical varve data, no, we have no reason to believe that Tiljander’s proxies can be calibrated.

    Another example of the same kind:

    > If Tiljander and nor Kaufman can calibrate the physical varve data, chances are we won’t be able to calibrate it.

    A last example:

    > Until somebody offers a reasonable alternative interpretation of the proxies, **we** should stick with Tiljander’s.

    To put it another way, AMac can answer “no” because all the details to understand this “no” has been shoveled under his argumentation.

    ***

    Perhaps we would like to answer another question.

    Something like:

    > Would you calibrate Tiljander proxies?

    Personally, I find this version interesting. Nick Stokes seems to answer this question when he says something like:

    > Considering the evidence, I would not use lake mud in my reconstructions.

    We also had something like this:

    > Should Mann have calibrated Tiljander proxies?

    Personally, I find that second-guessing Mann is not that interesting.

    Some may disagree. It’s a big world after all.

  406. willard Says:

    We should read:

    > If Tiljander and nor Kaufman won’t calibrate the physical varve data, chances are we won’t be able to calibrate it.

    I’m not saying that these are valid answers, by the way. I’m simply saying that they would look more natural than a single *no* in a critical discussion. Appealing to simplicity can be overdone.

  407. Jeff Id Says:

    I’ve taken the time to write yet another post on this:

    http://noconsensus.wordpress.com/2011/08/02/3456-2/

    There is one issue though that is best addressed here is the ‘ poor terminology sorted’

    “(including use of poor terminology such as “sorted” to mean screened)”

    A duck is a duck dock and until you point out a statistical text which supports ‘screening’ by arbitrary correlation threshold on a small subset of the timeseries, I think ‘sorted’ is more accurate. Sorry if it irritates you.

  408. MikeN Says:

    >but predicated on Tiljander’s reductive ‘interpretations’, which are little more than simplistic speculative suggestions by a doctoral candidate, outside the bounds of her remit, being necessarily correct.

    I don’t think that’s particularly relevant. The authors of Mann 08 decided to use the proxy, and made no mention of a different interpretation, just that it was potentially flawed. Then McIntyre informs them they’ve used it upside-down, and still no mention of we’re interpreting it differently. Just that the accusation is bizarre, meaning they either don’t know their own programming or some other incompetence, or they were lying to make McIntyre look bad and confuse others who would support them. The talk of Tiljander’s lousy interpretation is an attempt at a save, kind of like the bristlecone pines attempt referenced above by Deech. At least that has a paper behind it.

  409. AMac Says:

    luminous beauty (August 2, 2011 at 20:25) –

    > Your [AMac's] conclusions about the [four Tiljander data] series aren’t based on deep consideration of the physical reasoning about what would constitute mechanical, as differing from statistical, correlations, but predicated on Tiljander’s reductive ‘interpretations’, which are little more than simplistic speculative suggestions by a doctoral candidate, outside the bounds of her remit, being necessarily correct.

    You seem to be making two claims. The first is that I haven’t offered “deep consideration” of the Tiljander data series as representatives of physical entities. I have. Expositions of data and reasoning along those lines are the subject of most posts at my blog. See here for data and here for one such discussion.

    The second claim is that the interpretations offered in Tiljander03 were simplistic speculative suggestions by an out-of-her-depth grad student. Tiljander03’s authors included Professor Matti Saarnisto of the Geologic Survey of Finland (Tiljander’s advisor), junior limnologist Antti E.K. Ojala, and senior limnologist Matti Saarnisto. Their (cf. ‘her’) discussion wasn’t correct on all points (in my opinion), but it was certainly informed.

    If you review Arthur Smith’s two threads on Tiljander (starting here), you will see Mainstream scientist/advocates advancing the argument that Mann08’s interpretations of the data series were as good as Tiljander03’s, or better. On further consideration, they abandoned this defense of Mann08 — it’s a very weak case. If it were correct (it’s not), it would cause more problems for Mann08’s authors than it solves.

  410. AMac Says:

    willard (August 3, 2011 at 00:38) —

    That’s a lengthy discourse, almost 600 words. But I don’t think it sheds much light on its subject. That subject is this question I asked:

    Are the Tiljander proxies calibratable to the instrumental temperature record, 1850-1995?

    And my one-word answer:

    No.

    “No” remains the best answer.

    Willard’s essay melds two related topics. I would like to separate them.

    The first is about establishing a relationship between the time-series strings of numbers that Antti E.K. Ojalla submitted to archives such as NOAA (here), and the instrumental temperature record, 1850-1995. This is the task that Mann08’s authors undertook. They ended up with four correlations that were spurious — because establishing meaningful correlations is impossible. That is thanks to the progressive and ultimately overwhelming contamination of post-1720 records with non-climate-related signals, as noted in Tiljander03.

    The second topic concerns whether some future investigators might be able to establish a meaningful correlation between information buried in the Lake Korttajarvi sediments and the instrumental temperature record. For instance, time series on magnetic orientation, hydrogen isotopes, or carbon isotopes could be collected from the varves. (These have already been collected and published, in fact.) We can’t rule out that such new time series will enable direct calibration of the varves to the instrumental temperature record — in fact, we hope that will be the case.

    This second topic is not relevant to Tiljander-in-Mann08.

  411. willard Says:

    If that is true:

    > We can’t rule out that such new time series will enable direct calibration of the varves to the instrumental temperature record — in fact, we hope that will be the case.

    then categorically answering

    > No

    to

    > Are the Tiljander proxies calibratable to the instrumental temperature record, 1850-1995?

    does not follow from the premises.

    ***

    If what we can say is:

    > The best answer we have so far is that we can’t.

    that means that “no” is not enough and that we need to unpack it with its evidence base and its qualifications.

    This matters to Voldemort question and AMac’s answer.

    ***

    Asking for a simple answer (when it can’t be) is a trick. Asking it over and over and over and over and over again is another trick.

    Asking it over and over and over again without noting the answers does not bode well with Feynmanian ideals.

  412. AMac Says:

    Let me offer this longhand version of the Calibratability question.

    Can the Tiljander data series be calibrated in a meaningful (non-spurious) way to the instrumental temperature record, 1850-1995? The “Tiljander data series” are the annual time series for Lightsum, Darksum, and XRD, submitted by Antti E.K. Ojalla to NOAA, plus Thickness, which is the sum of Lightsum and Darksum. These four series comprise the “Tiljander proxies” as used in Mann08.

    No.

    .

    willard, I do not see a trick.

    If you think the question isn’t answerable, or if the answer somehow misleads the audience: can you explain why?

    If you think the question should be answered, “Yes,” could you sketch out this reasoning?

    If you think the question should be answered, “Maybe,” or “It isn’t knowable,” what are the sources of uncertainty?

  413. Carrick Says:

    Jim Boudin:

    Sure Jeff, point me to the journal article(s) in which you have shown that the use of the correlation between two variables over the common interval, to predict the values of one of the two over the non-common interval, is now null and void, and I’ll see what I can do.

    This is a heavily oversimplified version what what Jeff is trying to say. I’d recommend Jeff’s blog posts for what he says over this artifice resembling his arguments that you’ve erected.

    The CPS method, if that’s what you’re trying to refer to, isn’t that great and, among many other people, Mann admits to that in his more recent publications.

    (Also see this comparison of of Mann’s CPS method compared to other Ljundqvist. I think this meets willard’s requirement for not “over simplifying.”)

    Jeff has attempted to write up his comments in a comprehensible, mathematically rigorous series of blog posts. It’s there for anybody to look at, no need to go through a peer review before we’re allowed to make our own conclusions.

  414. luminous beauty Says:

    Amac,

    Note, I didn’t say you haven’t attempted plumbing the depths of physical reasoning, just that those attempts are shallow. The key series being lightsum. Colder winters = greater snowpack = higher spring runoff = greater lightsum is just chock full of non sequitur and over simplification. First, cold season lacustrine sedimentation (lightsum) is likely to be reduced in colder winters because of lower overall stream flow from surface melting, deeper inflow stream freezing, and later ice break-up. Second, a warmer melt season is likely to increase sedimentation disregarding snow pack size. However, the greater problems have little to do with the particulars of limnology, but, rather, meteorology and climate. Put simply, greater precipitation in polar regions is dependent on the transport of warmer wetter air from lower latitude circulation and, of necessity, somewhat reduced incursion of colder drier air from the Arctic Gyre. This is still a bit of over simplification. I’ll leave it to you to figure out the greater complexities, if you can.

    Whether the calibration against the temperature record is ‘meaningful’ for the proxy prior to the temperature record is indeed ‘problematic’. It isn’t a simple qualitative yes or no question, though, nor is it properly the real problem. It is a question of how much the climate signal is distorted by local engineering projects and farming practices. These may have somewhat masked the signal, in effect, reversing or exaggerating the trend in the signal, either temporarily or long term, but, on an annualized basis, excluding obvious outliers due to periodic construction disturbances, some portion of the climate signal is most likely still quantifiable. For this, it isn’t a question of meaningful/not meaningful, but what level of significance we can attach to this quantification. Is it a close enough fit to include these proxies in a reconstruction where these problems prior to the 18th Century really, really don’t matter? Does the uncertainty of the 20th Century calibration make the prior period’s variability so utterly inaccurate that it changes the reconstruction significantly? By themselves, obviously not. That is the principle concern for including them in the reconstruction. Not at all some supposed primary nor secondary nor tertiary justification for using dendro series, as you would suggest. This is an on going ClimateAudit zombie fixation which has really been dead almost from conception. These exclusion results are standard significance tests into which you are reading more than is there based on your pre-conceived bias.

    In the end it is a judgement call. I think Mann, et al., are at least somewhat justified in making this call with the obvious caveats. You obviously don’t. That is your amateur and ill-founded opinion and you are welcome to it. This, I think, is what willard is trying to convey, in his subtly inductive fashion.

    For the purpose of comparing the proxy to the temperature record, correct calibration is not only trivially possible, but absolutely necessary. On this, IMHO, you have totally screwed the pooch. You’ve missed a vital step, and whatever conclusions you might draw from faulty and improper analysis are pretty much worthless.

  415. Jeff Id Says:

    luminous beauty,

    I continue to marvel at the ability of people who don’t know about what they are writing, yet continue to do so. You write of uncertainty which in the context used, I can’t figure out any possible scenario of correctness. You write of judgement calls, this has nothing to do with any of the paleo discussion. You discuss calibration as ‘trivially possible’ calibration yet the entire point is that it is nearly impossible.

    You are not alone, as Jim Bouldin has jumped both feet into the same mire so don’t feel bad but the willing claims of knowledge/understanding where little exists is disturbing.

  416. Carrick Says:

    Luminous Beauty:

    For the purpose of comparing the proxy to the temperature record, correct calibration is not only trivially possible, but absolutely necessary.

    Please tell me you didn’t mean to say this.

  417. grypo! Says:

    …but I play one on the internet!

  418. willard Says:

    AMac,

    Thank you for these many questions. You are now asking:

    > If you think the question should be answered, “Yes,” could you sketch out this reasoning?

    On the 2011-08-02, at 18:38, which is not too far from now, I provided many examples of what you just ask. A bit earlier, on the 2011-07-12, at 21:10 EDT, I wrote my own answer, an answer I quoted on the 2011-07-31, at 6:18, so not long ago:

    http://ourchangingclimate.wordpress.com/2011/06/22/climate-science-scientific-method-skeptics-not/#comment-13813

    The first time I wrote this answer, it was meant for you.

    This is a bad omen.

    ***

    Before these questions, you claimed that I was talking about two topics, and that

    > This second topic is not relevant to Tiljander-in-Mann08.

    In my last response, I believe to have provided enough evidence to show that this claim is false. The Voldemort question is asking about the possibility or the impossibility of calibration, which is too categorical for an evidence-based reasoning setting. This argument needs to be acknowledged. If that could help you, one of your authority (Jeff Id) seems to be conceding in passing this argument, with his “nearly impossible” phrasing today at 14:10.

    This needs to be acknowledged.

    ***

    You also have said about that same comment:

    > That’s a lengthy discourse, almost 600 words. But I don’t think it sheds much light on its subject.

    My lengthy has 540 words and four parts and mainly contains quotes, paraphrases, and examples. For instance, the first part has 205 words, most of which (151) are words used to quote Luminous Beauty (16) and you (140). Asking for complete quotations and short comments looks like a double bind to me.

    This is not the Feynman way.

  419. AMac Says:

    luminous beauty (August 3, 2011 at 19:01) —

    > I didn’t say you haven’t attempted plumbing the depths of physical reasoning, just that those attempts are shallow. The key series being lightsum…

    Some folks seem have a hard time grasping some simple concepts, one of them being “what Tiljander 03 says.” So I’ve repeatedly quoted and then paraphrased the paper on this subject. Yes, Tiljander03’s portrayal of pre-1720 Lightsum is pretty simplistic.

    If you wanted to know what I think and why — a different concept — you would have to ask. Or read the posts I’ve linked.

    > Whether the calibration against the temperature record is ‘meaningful’ for the proxy prior to the temperature record is indeed ‘problematic’…

    I’ve highlighted the phrase “prior to”, because it makes your paragraph difficult or impossible to interpret. As far as I know. the sole method described in Mann08 is to screen, validate, and calibrate proxy series from 1850 to 1995 against temperature series from 1850 to 1995 (this 145-year period is split into two Early and Late divisions).

    What is your evidence that Mann08 used procedures involving a prior period?

    > [Local engineering projects and farming practices] may have somewhat masked the signal, in effect, reversing or exaggerating the trend in the signal, either temporarily or long term, but, on an annualized basis, excluding obvious outliers due to periodic construction disturbances, some portion of the climate signal is most likely still quantifiable.

    As far as I know, there are no procedures in Mann08 to unmask a somewhat masked signal that has reversed or exaggerated a trend in the [climate] signal. As far as I know, Mann08 did not exclude any obvious outliers due to construction or other disturbances.

    I don’t know what such an unmasking routine would be.

    Could you quote the text you are thinking of from Mann08 or its SI? Or withdraw the claim?

  420. AMac Says:

    Sorry, that’s me.

    Here is the sort of information that the Tiljander data series contain. It may be helpful for Luminous Beauty, willard, and other readers to work through the numbers. Perhaps we can say something about the plausibility of our competing ideas.

    Let’s consider the mean and standard deviation of lightsum, century by century. Such an analysis ignores autocorrelation, but it’s good enough for a first pass. Excel file and graphs here.

    Here’s Lightsum in micrometers (mean +/- std dev) for the calibration period, by half-century interval (retaining Mann08’s synthetic data for 1986-1995):

    1850 – 1899 480 um +/- 140 um
    1900 – 1949 640 um +/- 600 um
    1950 – 1995 1,325 um +/- 770 um

    Those are very thick varves!

    Here are the average temperature anomalies for the CRUTEM3v 5-degree by 5-degree gridcell that includes Lake Korttajarvi:

    1850 – 1899 -0.5 C +/- 1.0 C
    1900 – 1949 -0.1 C +/- 1.1 C
    1950 – 1995 +0.1 C +/- 1.0 C

    Tiljander03’s Figure 2 provides data from the nearest weather station. For the period 1881-1993, temperatures were more-or-less steady:

    Summer average: 14.3 C
    Winter average: 8.1 C

    We are now equipped to look at the pre-instrumental reconstruction period. From 500 to 1720, Lightsum was lowest in the 11th century, and highest in the 8th century:

    1000 – 1049 150 um +/- 40 um
    1050 – 1099 140 um +/- 30 um

    700 – 749 270 um +/- 100 um
    750 – 799 280 um +/- 90 um

    If there is a meaningful (non-spurious) correlation of Lightsum to average temperature for the calibration period (1850-1995), this correlation will tell us something about average temperatures in the 11th century. It will tell us something about average temperatures in the 8th century.

    (At this semi-qualitative level, a log-transformation of data isn’t necessary. If you think it is, it’s easy enough to perform; see the Excel spreadsheet.)

    What can we learn from Lightsum about the temperature of the Lake Korttajarvi region in the 11th Century,and in the 8th Century? What might the summer average have been? What might the winter average have been?

    Recall my answer: Since the correlation of Lightsum to Temperature 1850-1995 is spurious, there is no information to be gained from this exercise.

    What is your appreciation?

  421. AMac Says:

    willard (August 4, 2011 at 00:21) —

    Your comment is puzzling.

    You make no argument that I can follow, that the answer to the calibratability question (here) is anything other than “No”.

    Jeff Id seeming to concede the argument in passing with his “nearly impossible” phrasing — that is not an argument. It’s a quip.

    “No” as an argument is strong, for reasons given throughout this thread. It is strengthened by the illustration of Lightsum thicknesses in the immediately-prior comment (though I don’t realistically expect most readers to work through it).

    And yet, because I hold to “No” in the absence of counter-arguments, you remark “This is a bad omen” and “This is not the Feynman way.”

    We may have reached a dead end in this exchange.

  422. willard Says:

    AMac,

    Earlier today you said:

    > We can’t rule out that such new time series will enable direct calibration of the varves to the instrumental temperature record — in fact, we hope that will be the case.

    http://ourchangingclimate.wordpress.com/2011/06/22/climate-science-scientific-method-skeptics-not/#comment-13849

    If that’s true, then categorically answering

    > No

    to

    > Are the Tiljander proxies calibratable to the instrumental temperature record, 1850-1995?

    does not follow from the premises.

    Do you agree with me?

  423. AMac Says:

    willard (01:24) –

    > Do you agree with me?

    We have to define “the Tlljander proxies” in a consistent and meaningful way.

    Because I do not think they are well-suited as temperature proxies, I prefer to speak of “the Tiljander data series”. But either phrasing should suffice for this purpose.

    “The Tiljander data series/proxies” are four time-series of annual values that were described in a 2003 paper with Mia Tiljander as first author. These four time series were later used in Mann08. They are Thickness, Lightsum, and Darksum (in millimeters or microns) and XRD (in arbitrary greyscale values).

    There are other descriptors of the varved sediments that underlie Lake Korttajarvi. Examples would include magnetic susceptibility and H and C isotope compositions. Mia Tiljander was involved in collecting or analyzing some of them, but not others. There are also new analyses of old cores that may be undertaken in the future, and new cores that may be drilled.

    None of these additional time series from Lake Korttajarvi (real or hypothetical) are included in the set being labeled “the Tiljander data series/proxies.” To my knowledge, that label has been restricted to Thickness, Lightsum, Darksum, and XRD. That is the only way that I have employed this terminology.

    Given that definition, here is the calibratiblity question, copied from August 3, 2011 at 14:15, supra.

    “Can the Tiljander data series be calibrated in a meaningful (non-spurious) way to the instrumental temperature record, 1850-1995? The “Tiljander data series” are the annual time series for Lightsum, Darksum, and XRD, submitted by Antti E.K. Ojalla to NOAA, plus Thickness, which is the sum of Lightsum and Darksum. These four series comprise the “Tiljander proxies” as used in Mann08.”

    And the answer:

    No.

    Therefore, given that definition you and I are not in agreement.

    Earlier today, I stated, “future investigators might be able to establish a meaningful correlation between information buried in the Lake Korttajarvi sediments and the instrumental temperature record… We can’t rule out that such new time series will enable direct calibration of the varves to the instrumental temperature record — in fact, we hope that will be the case.”

    If these hopes are fulfilled, one or more of these new data series — data series that were not used in Mann08 will turn out to be calibratable to the instrumental temperature record. These new data series are not members of the set of “the Tiljander data series/proxies.” That set is restricted to Thickness, Lightsum, Darksum, and XRD — the four data series used in Mann08.

    I have not stated that no time series from Lake Korttajarvi will ever be meaningfully calibrated to the instrumental temperature record. I would not make such a prediction — it may well be false; I hope it turns out to be false.

    Thus, if you and I broaden the discussion to include the prospect of calibrating any data series from Lake Korttajarvi, then we are in accord.

  424. luminous beauty Says:

    Amac,

    I speak of annually resolved variance and you throw out half century averages. It’s like you are making an effort not to understand.

  425. mikep Says:

    Jim says

    “Sure Jeff, point me to the journal article(s) in which you have shown that the use of the correlation between two variables over the common interval, to predict the values of one of the two over the non-common interval, is now null and void, and I’ll see what I can do.”

    No statistician I know would try to use correlation between variables over one period to predict the values in another period without a lot of careful qualifications. Udney Yule wrote a famous paper in the 1920s examining what he called “nonsense” correlations. His example was the correlation between the proportion of marriages that were Church of England and the standardised mortality rate 1866-1911, which was over 0.95, but was clearly non-causal and not robust outside the time interval. His paper discusses how these nonsense correlations can arise. The short answer is that the standard random sampling assumptions are violated. Granger and Newbold in 1974 showed that two independent random walks (so unrelated by construction) would show very high correlations much more of the time than you would expect using t statistics interpreted using standard tables. Using this correlation outside the sample period would give most misleading answers. There is a huge associated literature. Jeff doesn’t need to write any journal articles making points which are well established in the literature.

  426. AMac Says:

    luminous beauty 07:51 —

    In response to statements that you made on Aug. 3 at 19:01, I asked (Aug. 4 at 00:44) –

    > What is your evidence that Mann08 used procedures involving a prior period?

    > I don’t know what such an unmasking routine would be… Could you quote the text you are thinking of from Mann08 or its SI? Or withdraw the claim?

    I later presented half-century averages of Lightsum that strongly suggest that no analysis will be able to derive information on temperature from the post-1850 record (the annual data are available at the link).

    AT 07:31, you ignored the two questions and dismissed the data.

    This is fine, in that it’s a wide world, and each of us may hold and advance opinions of our choosing.

    But it suggests to me that our conversation is at an impasse. Neither one of us — nor any other readers — are likely to learn much from continuing the dialog.

    Best wishes to you.

  427. willard Says:

    AMac,

    Thank you for your clarification, which explains what you meant and makes me understand your reaction:

    > The second topic concerns whether some future investigators might be able to establish a meaningful correlation between information buried in the Lake Korttajarvi sediments and the instrumental temperature record. [...] We can’t rule out that such new time series will enable direct calibration of the varves to the instrumental temperature record — in fact, we hope that will be the case.

    Somehow I thought the last sentence (“We can’t rule out…”) that you were entertaining the possibility produce a directly calibrated series from the Tiljander proxy, e.g. by filtering out the noise from its contaminated part.

    My apologies for having misread you.

    ***

    I’m still unsure what in my comment gave you the idea that I was talking about “future investigations” of something else than Voldemort’s question. Here are again my examples:

    (E1) Considering Tiljander’s interpretation of the proxies and her warning that it might be “a demanding task” to calibrate the physical varve data, no, we have no reason to believe that Tiljander’s proxies can be calibrated.

    (E2) If Tiljander nor Kaufman won’t calibrate the physical varve data, chances are we won’t be able to calibrate it.

    (E3) Until somebody offers a reasonable alternative interpretation of the proxies, **we** should stick with Tiljander’s.

    (E4) Considering the evidence, I [Nick Stokes] would not use lake mud in my reconstructions.

    The only example I can see is my paraphrasing of Nick. He said many times he would discard Tiljander’s proxies. This answers this question:

    (N) Would you calibrate Tiljander proxies?

    I’d rather consider this question than Voldemort’s. We should not mention his name that often anyway.

    ***

    To be sure I get your opinion correctly, here is how I interpret your “no” answer:

    > No, no one will ever be able to calibrate Tiljander’s proxies.

    This excludes the possibility of a new way to calibrate these proxies. This excludes the possibility of another limnological interpretation in contradistinction to Tiljander’s.

    To me, this is what “impossible to callibrate” means. To me, this means you have a formal proof of your claim. To me, this is stronger than all the examples I’ve shown above.

    Is this what you mean by your “no” answer?

    ***

    Evidence-based reasoning requires conclusions that are not expressed in categorical statements, i.e. an unqualified yes or no. Answering with a categorical “no” to Voldemort question either does not follow from the evidence base put forward as premises, or it only looks categorical, for instance by shoveling the qualifiers somewhere else.

    In that case, the “no” answer is there for a rhetorical effect, i.e. an appeal to simplicity. This is what I mean by “trick”.

  428. dhogaza Says:

    No statistician I know would try to use correlation between variables over one period to predict the values in another period without a lot of careful qualifications.

    Which, of course, is why possible proxies are chosen carefully. We know (and have endlessly pointed out) that trees growing at the extremes of their latitudinal and altitudinal ranges in areas with sufficient water and nutrients are sensitive to summer temperatures and are useful as proxies for past temperatures.

    So selection based on knowledge of plant physiology paired with correlation with the instrumental record over time (and, yes, I know that *some* dendro series diverge over about *1/3* of the instrumental record) combine to provide a strong argument for their use.

    Regardless of Jeff’s opinion. Or yours. Or the rest of the denial sphere’s.

  429. AMac Says:

    willard –

    Thank you for your remarks at 14:21.

    > here is how I interpret your “no” answer

    For precision, let me reword what you wrote next.

    No one will ever be able to directly establish a non-spurious calibration of Tiljander’s proxies to the instrumental temperature record.

    “Directly” is important because this is a discussion of Tiljander-in-Mann08, and what that incident does (or doesn’t) say about the position from “The Conversation” that Bart quoted in the body of his post.

    Mann08 (and Mann09) used direct calibration only.

    (When the authors of Kaufman09 learned from McIntyre et al of the problems with direct calibration, they re-wrote their paper to substitute an indirect calibration of XRD to other proxies in the pre-1720 period.)

    I am not a logician or a philosopher, but can sketch the general direction of a formal proof based on information theory as it applies to signal processing. The four Tiljander data series contain 405 discrete pieces of information, as three time series of 135 numbers, with each number describing one year’s varve (it’s actually 270 numbers in two time series, a story for another day).

    The climate of southern Finland in, say, the 8th and 11th centuries was similar to the climate of the calibration period. Back then, southern Finland didn’t support, say, tundra underlain with permafrost. Nor did palm trees sway in a tropical breeze.

    Prior to post-1720 local disturbances, we can see what the climate-influenced Tiljander data series looked like. For, say, Lightsum (Aug. 4 at 00:51), the (mean +/- standard deviation) was in the range of (140 +/- 30 microns) to (280 +/- 100 microns).

    Therefore, this is what a hypothetical climate-influenced-but-not-local-activity-influenced Lightsum data series would look like for the 1850-1995 calibration period, with its essentially-similar climate. No permafrost, no palm trees.

    However, we know that local effects overwhelmed climate effects. Tiljander03’s authors checked peat-cutting and bridge-reconstruction records, and told us so. Lightsum for 1950 through 1995 has a mean of 1,325 microns and a standard deviation of 770 microns. 1,325 +/- 770 microns!

    The information required to extract a signal of, say 80 microns from a {signal plus contamination} sum of, say 1,325 +/- 770 microns is not contained in the 405 discrete numbers that comprise the four Tiljander data series.

    .

    > Evidence-based reasoning requires conclusions that are not expressed in categorical statements.

    That is a categorical statement of your opinion. I do not believe that you have shown that it has practical application to the narrow case that we are discussing here.

    > the “no” answer is there for a rhetorical effect

    Because I made that statement, I have access to the reason it was made. You are incorrect. The “no” answer is there because, in my opinion, it succinctly and accurately describes a central issue of Tiljander-in-Mann08.

  430. AMac Says:

    Clarification. At 15:46 immediately prior, I wrote,

    > The four Tiljander data series contain 405 discrete pieces of information…

    and

    > The information required to extract a signal… is not contained in the 405 discrete numbers that comprise the four Tiljander data series.

    These two statements refer specifically to the Tiljander data series over the calibration interval of 1850 to 1995.

  431. luminous beauty Says:

    Amac,

    > What is your evidence that Mann08 used procedures involving a prior period?

    Duh. Figure S8: Sensitivity of NH mean reconstruction to exclusion of selected proxy
    record.

    The brown line, not the green.

    Verification statistics for cases where the tree-ring proxies and/or the 7 problem proxies are withheld [XLS]. Also.

    > I don’t know what such an unmasking routine would be… Could you quote the text you are thinking of from Mann08 or its SI? Or withdraw the claim?

    There is an obvious spurious trend in sedimentation rates over the late 18th and 20th Century as the result of human intervention.

    Log normalization, (when done correctly by calibrating the compared series to an equivalent scale, i.e., normalized mean of both series = 0, and scalar adjustment of distribution amplitude and width so that: SDA/SDB = 1), automatically detrends the data by calculating the distribution of annualized variance around a zero mean.

    I don’t think Mann, et al. adjusted for the obvious outliers because including them as is didn’t ‘meaningfully’ affect statistical significance.

  432. mikep Says:

    dhogaza,

    I’m glad you agree with me about correlation. I take it that you therefore agree that Jim’s argument about Jeff’s incompetence is wrong. We can return to the question of whether the selection procedures for proxies are adequate to solve the problems later on (though the Tiljander proxies in the calibration period are surely classic cases of nonsense correlations).

  433. MikeN Says:

    dhogaza. you entered this thread with an incorrect argument:
    He’s saying it’s bizarre because the signedness of the data won’t affect the multivariate analysis, therefore the claim shows a certain lack of understanding of such analysis. If they mistakenly thought the correlation was positive rather than inverted, it doesn’t change the result and it is bizarre to think it would (if you understand the basic mathematical argument, which boils down to -i*-i = i*i for all values of i).

    This is false. By any chance do you retract this statement, or a portion?

  434. dhogaza Says:

    I’m glad you agree with me about correlation. I take it that you therefore agree that Jim’s argument about Jeff’s incompetence is wrong. We can return to the question of whether the selection procedures for proxies are adequate

    That’s not JeffId’s argument at all, and no, I don’t agree with you.

    Your lack of reading comprehension is just more evidence that I’d be wasting my time posting here more often.

    Not sure why Jim bothers, no other scientists, do (well, we have Bart, our host).

    And the whole spew you folks put forth have absolutely no impact on science at all …

    Live with it … embrace it … and watch it have no impact whatsoever.

  435. dhogaza Says:

    This is false. By any chance do you retract this statement, or a portion?

    No, of course not, because you’re leaving out the part where part of the paper was devoted to the possibility that the Tijlander proxies were worthless.

    My guess is that you and your heroes would’ve never noticed problems with the Tijlander proxies if Mann hadn’t laid them out at your feet in the paper.

    You certainly don’t credit him with being the first person to say, “hey! these proxies might not be valid!”

    Of course, your true goal – and the goal of JeffID’s pseudo-mathematical “proofs” – is to suggest that paleo dendro records must be entirely ignored.

    Which is crap, but keep shitting on your crap, because clearly you’re interested in building an ever increasing pile of it.

  436. dhogaza Says:

    Amac,

    > What is your evidence that Mann08 used procedures involving a prior period?

    Duh.

    Amac doesn’t understand the paper … duh.

    And all these insults against LB who is, herself, a practicing scientist.

    It’s like all these wannabes want to redefine science so only those without a PhD should be listened to …

  437. willard Says:

    AMac,

    Thank you for your precisions. V’s question is indeed about the possibility to calibrate to something specific:

    > Are the Tiljander proxies calibratable to the instrumental temperature record, 1850-1995?

    http://amac1.blogspot.com/2011/06/voldemorts-question.html

    There is also this variant, on the 2011-07-14 10:00 EDT at Keith’s:

    > Can any of the four Tiljander data series be meaningfully calibrated to the instrumental temperature record, 1850-1995?

    http://ourchangingclimate.wordpress.com/2011/06/22/climate-science-scientific-method-skeptics-not/#comment-13606

    which asks for the possibility to meaningfully calibrate at least one data serie.

    ***

    It is important to distinguish V’s question from this more general one:

    > Are the Tiljander proxies calibratable?

    because we can respond to this other one by

    > Yes, if Kaufman and his team succeeded in doing so, albeit indirectly.

    The difference between direct and indirect calibration is important to limit the scope of your inquiry. While I appreciate the knowledge that Kaufman and his team (indirectly) calibrated the proxies, I do not think it matters much for my conceptual analysis. What matters is the way evidence-based reasoning is made explicit.

    ***

    Before returning to this analysis, I’ll note a remark you made on the 2010-06-16 (17:25 EDT):

    > That question–originally asked by Steve McIntyre–has yet to be answered. By Prof. Mann, by his co-authors, by RealClimate co-bloggers, by William Connolley… by any climate scientist.

    http://www.collide-a-scape.com/2010/06/16/the-main-hindrance-to-dialogue-and-detente/#comment-7537

    The origin of this question is not why you named it that way:

    > One can imagine a Question that would evoke similar dread, and must not be named.

    I get the impression that the three dots include any member of what you referred to at the beginning of the thread by “pro-AGW consensus advocates” or “pro-AGW consensus community”.

    I can understand the commitment of Prof. Mann and his co-authors to answer this question. But I’m not sure why any climate scientist would have this responsibility. Or any other person except Mann and his co-authors, for that matter.

  438. luminous beauty Says:

    For Amac,

    http://www.zi.ku.dk/popecol/ape/PDF/030821/IJC1.pdf

    http://finland.fi/public/default.aspx?contentid=160033

    Yes, Finnish winter temperatures, precipitation and snowpack are positively correlated.

  439. MikeN Says:

    >He’s saying it’s bizarre because the signedness of the data won’t affect the multivariate analysis, therefore the claim shows a certain lack of understanding of such analysis. If they mistakenly thought the correlation was positive rather than inverted, it doesn’t change the result

    Regardless of Mann’s additional statements on Tiljander, he used the data upside-down, and you have shown yourself to be misinformed as to the workings of his paper.

  440. willard Says:

    Before we pursue our analysis, it might be interesting to recall the activity of Steve’s Journal Club during the first month after the press release of Mann08.

    ## Back from Rome

    On the 2008-09-02, at 9:16, Steve McIntyre announces he’s back from
    Italy, observes and then wonders:

    > I see that Mann et al have a welcome-home present for us. I see that there are press releases announcing the article before it’s even available online. And while mongabay.com was given a preprint, it seems that climateaudit.org wasn’t. I wonder why.

    http://climateaudit.org/2008/09/02/back-from-italy/

    ## Notices

    On the 2008-09-02, at 10:06, in a post named **Mann et al 2008**, Steve notices:

    > Notice of a new reconstruction by Mann and the Team is in many press clippings today, citing a PNAS article that is not (as I write) online. Rather than clutter other threads, here’s a placeholder thread pending my own response which may take a few days.

    http://climateaudit.org/2008/09/02/mann-et-al-2008/

    This placeholder post contains more than 130 comments. The category
    “Mann et al 2008″ is being introduced.

    ## UC

    On the 2008-09-02, at 10:58, Steve publishes a comment by UC

    > UC writes in that there’s another Mannian problem: [...]

    Tags: plagiarism, smooth, smoothing, uc, uc00. More than 60 comments.

    ## Sep 3, 2008 – 10:13 AM – Mann 2008 MWP Proxies: Punta Laguna

    > I’m spending a little extra time examining the “new” Mann 2008 MWP non-dendro “proxies”.

    http://climateaudit.org/2008/09/03/mann-2008-mwp-proxies-punta-laguna/

    ## Sep 3, 2008 – 1:14 PM – Mann et al 2008: Korttajärvi

    > Larry Huldén writes: The Finnish lake sediments can not be used for temperature interpretations in the 18th to 20th century unless you know exactly the history of the regional lake environment conditions.

    http://climateaudit.org/2008/09/03/mann-et-al-2008-korttajarvi/

    ## Sep 3, 2008 – 8:19 PM – Mann et al 2008 Proxies

    > Here are 5 graphics which show all the Mann 2008 proxies in a consistent format, highlighting two periods 950-1100 and 1850-1980.

    http://climateaudit.org/2008/09/03/mann-et-al-2008-proxies/

    ## Filling in Punta Laguna (Sep 5, 2008 – 8:39 AM)

    > Yesterday Mann archived a new version of his data, a version supposedly without any infilling. Gavin Schmidt is pretending that this was there all along and demanding an apology from one of his readers, who had the temerity to question Schmidt.

    http://climateaudit.org/2008/09/05/filling-in-punta-laguna/

    ## MWP Non-Dendro Proxies #2 (Sep 5, 2008 – 1:14 PM)

    > OK, I’m starting to get the feel of the new proxy network and have some ideas of what the new Mannomatic is doing.

    http://climateaudit.org/2008/09/05/mwp-non-dendro-proxies-2/

    ## Phil B. Borehole Inversion Calculations (Sep 5, 2008 – 3:21 PM)

    > Phil B writes: My day job does include parameter and state estimation using Least Squares and Kalman filtering.

    http://climateaudit.org/2008/09/05/borehole-inversion-calculations/

    ## Proxy Screening by Correlation (Sep 5, 2008 – 9:00 PM)

    > I’ve made histograms of reported proxy correlations for 1850-1995, as reported in r1209.xls (which contains results for all proxies, unlike SI SD1.xls which withholds results below a benchmark.)

    http://climateaudit.org/2008/09/05/proxy-screening-by-correlation/

    [Erice hiatus filled with Joliffe.]

    ## Will the Real Slim Shady Please Stand Up? Re-Mix. (Sep 12, 2008 – 7:01 PM)

    > There’s an amusing little incident with the deleted “original” data set that was posted up for a few minutes at Mann’s website – you know, the data set that was first demonstrably referenced by a CA reader in the early morning of Sep 5.

    By Steve McIntyre| Posted in Briffa, Mann et al 2008 | Tagged Mann et
    al 2008, plagiarism, Proxies, substitution, Yamal | Comments (74)

    ## Mann 2008: Impact of the Missing Data (Sep 15, 2008 – 8:45 AM)

    > Jeff Id has an interesting post in which he examines the 148 “missing” series. Check it out.

    http://climateaudit.org/2008/09/15/mann-2008-impact-of-the-missing-data/

    ## Mann 2008: the Bristlecone Addiction (Sep 15, 2008 – 12:33 PM)

    > I see that the BBC is taking the position that the “evidence” of an MWP-modern differential now is confirmed in so many “independent” studies that the matter is now “incontrovertible”.

    http://climateaudit.org/2008/09/15/mann-2008-the-bristlecone-addiction/

    ## Mann 2008: the Luterbacher Mystery (Sep 19, 2008 – 10:21 PM)

    > Jeff Id has identified another intriguing mystery in the arduous problem of determining what Mann’s realdata was.

    http://climateaudit.org/2008/09/19/mann-2008-the-luterbacher-mystery/

    ## Jeff Id (Sep 20, 2008 – 10:24 AM)

    > Another interesting post from Jeff Id: http://noconsensus.wordpress.com/2008/09/20/online-experiment-with-the-latest-hockey-stick/

    http://climateaudit.org/2008/09/20/jeff-id/

    ## The Mann Correlation Mystery (Sep 20, 2008 – 11:42 AM)

    > Here’s another interesting mystery in Mann et al 2008.

    http://climateaudit.org/2008/09/20/the-mann-correlation-mystery/

    ## Mann 2008: the Briffa MXD Network (Sep 22, 2008 – 1:32 PM)

    > Those who forget the past are condemned to repeat it.

    http://climateaudit.org/2008/09/22/the-new-rain-in-maine/

    ## The Mystery in Kenya (Sep 22, 2008 – 8:01 PM)

    > Today’s Mannian mystery takes us to Kenya Tanzania, not to Kilimanjaro itself, but to the great plains, still home to prides of lions, herds of wildebeest and giraffes.

    http://climateaudit.org/2008/09/22/the-mystery-in-kenya/

    ## Mann 2008 Correlation Benchmarks (Sep 23, 2008 – 6:09 PM)

    > The purpose of working through frustrating details of Mannian lat-longs and so on was to start testing the assertion that the network contained 484 “significant” proxies and that this meant something.

    http://climateaudit.org/2008/09/23/mann-2008-correlation-benchmarks/

    ## Mann 2008 Correlations – A New Graphic (Sep 24, 2008 – 10:05 AM)

    > I’ve adopted a graphic from the R-gallery to better illustrate the issues that we’re working on with the Mann proxy correlations.

    http://climateaudit.org/2008/09/24/mann-2008-correlations-a-new-graphic/

    ## Mann 2008 – Replication (Sep 24, 2008 – 1:01 PM)

    > Do not post anything other than programming comments. Absolutely no piling on please.

    http://climateaudit.org/2008/09/24/mann-2008-replication/

    ## Another Interesting Correlation Graphic (Sep 25, 2008 – 12:08 PM)

    > In my last post, I observed an interesting bimodality which almost certainly appears to originate in Mann’s pick two procedure on low-correlation tree ring networks.

    http://climateaudit.org/2008/09/25/another-interesting-graphic/

    ## Mann’s PC1 in Esper and Frank 2008 (Sep 26, 2008 – 3:16 PM)

    On previous occasions, we’ve noticed some strange appearances of the
    Mann hockey stick under different disguises.

    http://climateaudit.org/2008/09/26/manns-pc1-in-esper-and-frank-2008/

    ## Mann Sediments and Noise Simulation (Sep 27, 2008 – 3:07 PM)

    > Both in climate blog world and the financial world, there has been much talk recently about the interaction of models and data distributions.

    http://climateaudit.org/2008/09/27/models-and-weird-distributions/

    ## When Good Proxies Go Bad by Willis Eschenbach (Sep 28, 2008 – 3:51 PM)

    > Many of the good folks who write the papers and keep the databases seem not to use their naked eyeballs.

    http://climateaudit.org/2008/09/28/when-good-proxies-go-bad/

    ## Wouldn’t it be nice… (Sep 29, 2008 – 8:49 AM)

    > if Team methodological descriptions were correct?

    http://climateaudit.org/2008/09/29/wouldnt-it-be-nice/

    ## The MBH98 Corrigendum (Sep 29, 2008 – 3:25 PM)

    > The SI for MBH98 listed 34 tree ring series that were not actually used.

    http://climateaudit.org/2008/09/29/the-mbh98-corrigendum/

    ***

    These are the CA posts related to M08 the month it was out. Please consider that during this month, Steve went to Erice for a week that month.

    Considering the latest activity of Steve’s Journal Club, and to complement the *scientific criticism* we already noticed, we might be tempted to introduce the concept of *scientific investigation*.

  441. AMac Says:

    luminous beauty (August 6, 2011 at 20:28) —

    Thanks for the two links. The first is to a PDF of a 2003 paper by C.B. Uvo (Int. J. Climatol. 23: 1185–1194). It is on the relationship of winter precipitation in Northern Europe to the North Atlantic Oscillation.

    The second is to a website maintained by the Finnish Tourist Board, giving a general overview of the country’s weather.

    .

    willard (August 6, 2011 at 14:11) —

    > While I appreciate the knowledge that Kaufman and his team (indirectly) calibrated the proxies…

    Note that I do not claim that Kaufman et al were successful in their effort. For starters, we would have to establish what the word means, in this context. I think there are reasons to think that Kaufman’s correlation of XRD to temperature is not meaningful. However, that’s a problem with nuances and complications. This thread has convinced me that it is overly ambitious to attempt to discuss subtleties, at this forum, on this general subject.

    > What matters is the way evidence-based reasoning is made explicit.

    To the best of my abilities, I have been explicit in the positions I have taken on Tiljander-in-Mann08. I have presented the evidence and the logic behind these positions. As far as I know, there’s no substance to the challenges that have been offered to what I’ve presented (or to the points that MikeN has made, either).

    The only exception is luminous beauty’s contention that certain scatterplots should have shown the logarithm of thickness, lightsum, and darksum versus temperature, rather than the thickness values themselves (XRD is already on a log scale). On reflection, that’s a debatable point, but it’s thought-provoking (and I’ll do that re-plotting). But it’s peripheral — those plots are mainly aids for visualizing data.

    The bigger issue is that I’ve responded to many misconceived comments by gamely repeating correct information. This gets boring to write, and is likely boring to read. Since the nth version changed no minds, it was (is) a poor use of my time to proceed with n+1.

  442. Rattus Norvegicus Says:

    AMac,

    luminous beauty (in the first link) provides evidence that Tiljander’s interpretation of lightsum (IIRC this is the winter temperature) was wrong.

    When I read the original Boreas paper I was, well to put it lightly, bemused, given my experience in snowy country (warm winters == greater snowfalls and bigger snowpack) but I was willing to accept the conditions in Finland might be different. The paper LB linked to seems to say otherwise. This would mean that 3 of 4 Tiljander proxies were correctly oriented (although they may have been polluted by human activity).

    Mann ran a basic sensitivity analysis and decided that the polluted proxies did not make much difference. Including them was a judgement call. There are open source matlab interpreters use his matlab code and try the various permutations. Let us know what you find. Not all of the options of inclusion or non inclusion of the Tiljander proxies are valid.

  443. AMac Says:

    Rattus Norvegicus (06:32) —

    > luminous beauty (in the first link) provides evidence that Tiljander’s interpretation of lightsum was wrong.

    As far as I can tell, this is incorrect.

    However, separately, some time ago, I became convinced that Lightsum is a proxy for temperature that is poor, at best, in the orientation that Tiljander03 describes. See my discussion with ‘Scientist’ (‘TCO’) at my blog, which I’ve linked n times on this thread.

    Lightsum is as a poor proxy for temperature in the upside-down (with respect to Tiljander03) orientation. If not worse.

    I suspect that both Lightsum and Darksum may be positively correlated with precipitation.

    These are subtle issues. Your comment suggests that you are not grasping the major points of this story, e.g. that Mann08’s interpretation of Lightsum was entirely spurious. Mistaken. Without merit.

    You’ve weighed in on Tiljander-in-Mann08 ever since those early threads on Stoat. So I don’t see much value in you and me pursuing iteration n+1. You don’t have relevant evidence to support your position (in my opinion). I have no expectation of influencing your views.

  444. J Bowers Says:

    AMac — “Your comment suggests that you are not grasping the major points of this story, e.g. that Mann08′s interpretation of Lightsum was entirely spurious. Mistaken. Without merit.”

    Shades of EM Forster.

    “Mann08′s interpretation of Lightsum was flawed.”

    A single aspect in the paper is flawed. The science can be discussed.

    “Mann08′s interpretation of Lightsum was entirely spurious. Mistaken. Without merit.”

    Mann uses spurious methods with no merit. Mann can be discussed.

  445. willard Says:

    AMac,

    I have no reason to believe your opinion of Kauffman is very different from what has being claimed so many times at Steve’s. By chance this opinion is irrelevant to the distinction about direct and indirect calibration. So we should thank your sense of moderation, which prevents you from discussing subtleties, nuances and complications of a red herring. If you don’t mind, I’d rather return to your evidence-based approach [1].

    ***

    First, V’s question is not a question, but a claim:

    (V) It is impossible to calibrate the Tiljander proxies to the instrumental temperature record, 1850-1995.

    It is not an open question anymore. You already have your answer: the fact that you answer “no” makes it a claim. It is a claim you believe must be challenged:

    > As far as I know, there’s no substance to the challenges that have been offered to what I’ve presented (or to the points that MikeN has made, either).

    You offered two main arguments for the V claim: Tiljander’s limnological interpretation and the failure of the calibration test when corrected from Mann’s mistakes. I believe that this summarizes well the evidence you offered so far, which readers can read on your site along with many interesting analogies, mataphors, or parables. This summary serves my purpose if we accept that the evidence is both limnological (i.e. someone physical) and statistical (i.e. formal).

    ***

    Second, the modal of V is of a stronger type (i.e. I see it as logical) than the limnological and statistical evidence on which it rests. This is the only point I made so far about your evidence-based reasoning. You conceded that your evidence does not contain any proof of impossibility. This is progress, although we did have to work for it, not unlike in journal club discussions I’ve seen elsewhere.

    In your last reply, you sketched a proof based on information theory. So far so good: I’m not excluding that possibility, in fact I would not be that surprised that we find one. But for now, all we have is a test showing an estimate of the unreliability of the data. (As an aside, if you can show that statistical testing equates to formal verification, tell me so in private and we could become very rich.)

    ***

    Third, I believe all this leads us astray. And this would be my substantial point. I don’t see the need to ask any calibratability question to get to the claim that Tilj’s two times series should not have been used.

    So here’s a five-minutes suggestion to adjust your argument to prevent that:

    (P1) Until we find a way to calibrate T’s proxies, we can’t use them.
    (P2) We have not found a way to directly calibrate T’s proxies yet.
    (C) We can’t use them for direct calibration yet.

    I believe everyone could agree with (P1). Here is Gavin [2]:

    > If there is a contamination in the modern period by non-climatic influences (which the originating authors suggested there might be), then they just can’t be used.

    Here is the Ferret [3]:

    > If the proxy is unreltaed [sic.] to temperature over the last 200 years then it is unusable for Mea’s method.

    I believe we have enough evidence for P2, e.g. the discussion between Close and Nick Stokes. In any case, we can fully appreciate that Luminous Beauty needs to find a way to directly calibrate T’s proxies to beat that argument. Trying to find a direct way to calibrate T’s proxies seems a bit tougher than simply raising the possibility that T’s interpretation could be wrong.

    ***

    Fourth, since almost every interested commenter (this is not a quip) seem to agree with your most basic points, so much that there seems to be only one rational answer, auditors could wonder where if the issue is made explicit at all. Besides providing many testimonies of other “members of the pro-AGW community”, I already conceded that you had a strong case. More than once, actually. In fact, to the reader that might wonder what is the opinion of Rattus Norvegicus on the subject, here is a sample:

    > I just read the Tiljander paper and came away with this takehome message: this is a lousy climate proxy. [...] So upside down, inside out or right side up, this is a lousy proxy and should probably not have been used. Of course Mann anticipated these attacks and showed that it didn’t matter whether he used it or not, leaving McI fuming.

    http://scienceblogs.com/stoat/2009/10/tiljander.php#comment-2033940

    Since 2009-10-29, this rat is of the opinion that we should *probably* not use the proxies. Here is how portrayed it:

    > [Rattus Norvegicus has] weighed in on Tiljander-in-Mann08 ever since those early threads on Stoat. So I don’t see much value in you and me pursuing iteration n+1.

    Auditors need to compare their impressions of our Norvegian rat after reading the first quote, and after reading the second one.

    In any case, the value of your Tiljander website should exactly be to bypass these iterations. Alright, you’re not neutral. So what? Nothing prevents anyone from specializing in the Tiljander gambit. It provides a portable repertoire, and it would alleviate a problem you see, a bigger problem than l’Affaire Tiljander, which is responding to

    > Many misconceived comments by gamely repeating correct information.

    We have evidence that MIkeN asked you to intervene in this very thread. Readers can observe that your very first comment starts with:

    > First, there are three (not four) Tiljander data series (they aren’t clear proxies for temperature).

    http://ourchangingclimate.wordpress.com/2011/06/22/climate-science-scientific-method-skeptics-not/#comment-13453

    If correcting misconceived comments gets boring, nobody’s making you do it. Not even MIkeN. A link to your site should suffice, with the issue of a challenge, as you often did at Lucia’s.

    ***

    In my first comment here, on the 2011-07-04, I’ve alluded to the fact that we’d need to pay due diligence to conclusions regarding the broader applicability of the argument. In my second comment, on the 2011-07-05, I’ve mentioned that we need to pay due diligence to the way these conclusions are constructed. And this is without mentioning the question how and why any of this matters for Mann08 and for the grand scheme of things.

    To me, these are more important areas of disagreement, and the most we have so far are analogies, metaphors, and parables. Notwithstanding closure marker after closure marker, the older one (directed at me) already being one month old. There are so many things to discuss that this dynamic can only remind me when VS kinda “forgot” we could simply assume **for argument’s sake** anything he needed to make his case.

    Auditors might ask why we don’t talk about about this dynamic, or about these analogies, metaphors, and parables. To quote Paul Graham [2]:

    > It’s not just the mob you need to learn to watch from a distance. You need to be able to watch your own thoughts from a distance.

    Perhaps a reason why this debate keeps resurfacing is that people keep entertaining incorrect beliefs. Perhaps this is not the only reason. Perhaps this is not even the main reason. Whatever the reasons, they can’t simply be analogies, metaphors, and parables.

    [1] http://ourchangingclimate.wordpress.com/2011/06/22/climate-science-scientific-method-skeptics-not/#comment-13740
    [2] http://www.collide-a-scape.com/2010/06/16/the-main-hindrance-to-dialogue-and-detente/#comment-7517
    [3] http://scienceblogs.com/stoat/2009/10/tiljander.php#comment-2046320
    [4] http://www.paulgraham.com/say.html

  446. AMac Says:

    On August 1, 2011 at 17:14, luminous beauty noted:

    Amac,
    Your simple statistics are not so simple. Being able to plug & play into Excel doesn’t necessarily mean you know what you’re about. Log normalization, if implemented properly, is a calibration step, which you keep claiming is impossible. What it does is make the two data sets to be compared orthonormal, or in simple language, proportionately equal, i.e., calibrated to the same scalar mean. Just glancing at your scatter plots indicates ur doin’ it rong.

    I’ve belatedly reviewed the post in question. It turns out that its scatterplots already use log-transformed data. However, since Absorbance is itself a logarithmic scale, I was in error to subject that data series to log transformation. I’ve redone that scatterplot; see Update 2. Thanks to luminous beauty for drawing my attention to the issue.

    Separately, I explain that only two independent series can be represented by the four Tiljander time series — either Lightsum and Darksum or Thickness and XRD. But not both pairs.

  447. Cameron Says:

    As an ecologist I have taken great interest in this scientific debate. I avoid taking sides on these matters as I try and look at it from an objective standpoint. However, I do have a tendency to side with the minority on issues because I like to question things that a majority of people seem to just accept despite any lack of understanding.
    So I have recently been asked to review a proposed framework from a major U.S. university and some leading authors in the world of ecology and land management sciences. I will not share the paper nor will I use any reference to who or where because this is a dradft document.
    This framework is attempting to focus land management around anthropogenic climate change. However, nowhere in the paper does it mention “CO2.” I found this to be quite puzzling. After all if you are attempting to address a problem, is it not important to be upfront and identify the problem? Out of dozens of climate articles cited in the paper, not a single one had co2 in the title.
    Then I began to notice something odd about the wording; it has temporal implications. For example, they state that “science has shown that species are likely to shift as a result of changes in climate.” This may not seem odd to someone well equipped with the latest in co2 science. However, as someone who has studied ecology for a reasonable amount of time, this is a clear error on their part; we have known species shift in response to climate change long before co2 was ever an issue. This is not new knowledge, yet they word it as such. Now one could assume, based on the fact that the intended audience is “highly informed” and “highly technical,” that one must already have an understanding that co2 is driving climate change. However, the reason we know species shift in response to climate change, which seems like common sense to me, is because it has been observed. This error occurs not once, but multiple times throughout the document. On top of this, the wording gives me the impression that they are intentionally trying to talk over people and make it hard to understand; simple ecological principles are described using long drawn out paragraphs using words that are rarely if ever used in scientific papers.
    So I get the impression that the authors have no knowledge of previous changes in climate and previous shifts in species, which seems a bit odd because previous studies noting previous changes in climate and shifts in species are cited in the research supporting anthropogenic climate change. It also became very clear reading this draft framework that they lack an experience in actually working with natural resources and are in fact just academics sitting in an office with no real connection to any ecosystem.
    So it leads me to place more weight on the claim that anthropogenic climate change is bordering on religious doctrine or political propaganda. What reason is there for supposedly credible scientists to discount, or completely ignore previous peer reviewed research that is essentially common knowledge? And why would they propose a framework to address a problem, but not clearly state what the problem is?
    This paper has left me confused and questioning a lot of things. And the sad thing is someone thinks this is going to be a valuable tool, when in reality it will accomplish virtually nothing, nor will change how we manage lands. In fact, it is doomed to be ineffective because they failed to clearly present the problem. And again I ask why?

  448. dhogaza Says:

    This framework is attempting to focus land management around anthropogenic climate change. However, nowhere in the paper does it mention “CO2.”

    This is one of the silliest things I’ve ever read, sorry.

    Anthropogenic climate change is known to be *largely* but not *entirely* due to increased CO2 concentrations in the atmosphere. Land-use change contribute, as do other GHGs.

    So strictly speaking “CO2 warming” or whatever wouldn’t be correct.

    And, of course, everyone knows what “anthropogenic climate change” refers to. You’re being extremely disingenuous is suggesting the while you do, they don’t.

    This is even sillier:

    Then I began to notice something odd about the wording; it has temporal implications. For example, they state that “science has shown that species are likely to shift as a result of changes in climate.” This may not seem odd to someone well equipped with the latest in co2 science. However, as someone who has studied ecology for a reasonable amount of time, this is a clear error on their part; we have known species shift in response to climate change long before co2 was ever an issue.

    They state that species are known to shift as a result of changes in climate – regardless of what has caused such a change, if you are quoting accurately.

    Then you accuse them of not being aware that climate change that’s not caused by anthropogenic changes can cause species ranges to shift”.

    That’s an extremely poor grasp of reading comprehension you’re showing, dude.

    And you’re being paid to evaluate and review this, even though you don’t appear to understand plain english?

    And, the topper, your lack of reading comprehension leads you to conclude …

    So it leads me to place more weight on the claim that anthropogenic climate change is bordering on religious doctrine or political propaganda.

    Bollocks.

    This paper has left me confused and questioning a lot of things.

    I’m left questioning your competency.

    And the sad thing is someone thinks this is going to be a valuable tool, when in reality it will accomplish virtually nothing, nor will change how we manage lands. In fact, it is doomed to be ineffective because they failed to clearly present the problem. And again I ask why?

    Facts not in evidence.

  449. Deech56 Says:

    Cameron, your citing “religious doctrine or political propaganda” sounds more like the response of one who is biased on the side of contrarianism (the “I do have a tendency to side with the minority on issues…” is a bit of a giveaway) rather than accepting that climate scientists know what they are doing.

    There’s nothing wrong with healthy skepticism, but when approaching a science as an outsider, it is best if one at least respects the general competency of those who practice that science. I would have to agree with dhogaza’s criticisms.

  450. Bart Says:

    Cameron,

    “have a tendency to side with the minority on issues because I like to question things that a majority of people seem to just accept” sounds like a logical fallacy to me. The more logical null hypothesis would be to assume a competent majority to be more likely right than wrong. Upon close examination of course ones judgment can change, but to start with the opposite null hypothesis strikes me as odd.

    From what you quote I see no evidence whatsoever “that the authors have no knowledge of previous changes in climate and previous shifts in species”, which you then use to proclaim a whole scientific field to be no more than “religious doctrine or political propaganda”. Sounds like a very rushed judgment to me, and one that invariably involves conspiracy thinking. Plus, the fact that the papers didn’t mention CO2 at all actually goes against your argument if you ask me.

    I don’t have an opinion on whether or not the authors “clearly presented the problem” (too little information for me to base an opinion on), but your characterization of the science of climate change that you seem to deduce from it clearly doesn’t follow.

    Dhogaza,

    Please tone down the aggresive tone in your response; in many cases it may work counterproductively.

  451. Tom Fuller Says:

    “Has history ever recorded an instance where the majority was right?”

    He was a bit tough on ol’ homo sap, but I live in a country that elected George Bush… twice. And Richard Nixon… twice.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s


Follow

Get every new post delivered to your Inbox.

Join 128 other followers

%d bloggers like this: