Climate Audit

Jeff Id

The evidence that Mann used the Tiljander proxies upside down is easily understood by non climate scientists;

Ok, now this is an interesting point which was a huge discussion on tAV. Ryan and myself got into a big battle at tAV with others over using ACTUAL thermometers upside down. This situation happened in Steig’s paper on Antarctic trends and it was defended by a number of people as reasonable through various odd conceptualizations of inverse covariance and how the upside down thermometer might explain them – completely randomly through fancy math of course. Mann 08 is the exact same situation. If we can’t determine that a thermometer may NEVER be read upside down!!, how can we discuss if a thermometer PROXY is reasonable UPSIDE DOWN.

It hurts my engineer brain.

Its like different PCs in a PCA analysis being attached to individual physical processes i.e. rain, CO2, temperature. Now my just be engineer but my think math don’t care about physical processes. AT ALL, ever!

Mann didn’t answer anything. The reply was out of context just enough that it didn’t make any sense if you’re familiar with the subject but it sounded scientific, aloof and obviously dismissing if you’re not. Since I’m not unfamiliar, with the subject the reply did nothing but confirm the criticism and my own belief that they understand exactly what the paper was comprised of. Just imagine pretending to not understand the criticism of ‘upside down’ thermometers and responding with ‘multivariate methods don’t care’. There isn’t a science PHD on the planet who couldn’t understand the criticism better than that. NOT ONE!

Think about it.

[snip away if need be]

• TAC
  Posted Oct 15, 2009 at 8:17 AM | Permalink

  Re: Jeff Id (#2), I completely agree. It is hard to believe Mann’s ‘response,’ a response consisting of nothing more than annoying misdirection. In contrast, McIntyre’s analysis (e.g. here) is lucid, correct, and compelling. McIntyre’s analogy to securities trading makes the point. To quote:

  In Mann et al 2008, there is a truly remarkable example of opportunistic after-the-fact sign selection, which, in addition, beautifully illustrates the concept of spurious regression, a concept that seems to baffle signal mining paleoclimatologists.

  Anyone can understand this.

  It surprises me that Revkin, among others, continues to tolerate (and, worse, rely upon) Mann’s Alice-in-Wonderland responses to rigorous criticism.

Jason

As interesting as Steve’s discoveries about Briffa are, THIS is the issue that I would hope outsiders look at and examine.

Nobody can no what Mann’s intent is, but he said something false and indefensible in a journal, and now repeats it citing the journal comment as proof.

Numerous climate scientists have confirmed the problem. No evidence has been offered to refute it.

If the climate science community is not able to correct an obvious and blatant falsehood in its peer reviewed journals, why should anyone take climate science journals seriously.

John A

Unlike some (in my view) overly patriotic CA readers, I do not regard blogs as a substitute for journals. While (in my opinion) my recent journal submissions have not been fairly reviewed (and there are many points that I take issue with in how academic journals handle reviews), this is not a point that I’ve placed at issue at this blog because, in my opinion, I haven’t fully put the system to a full test (but this issue is definitely on my radar.)

I don’t think of this blog (or any blog) as a substitute for journals. On the other hand, the way that journals deal with combative and censorious people like Mann is totally inadequate, allowing him to walk over criticism as if it didn’t exist.

Realclimate was setup for the express purpose of refuting McIntyre and McKitrick and to defend Mann, so we know and expect that Mann will exercise total control of postings in order to avoid any hits on his massive ego. And if that means censorship of Steve while allowing defamatory comments then so be it as far as Mann is concerned.

The journals are setup for quick debate and quick commentary because every page of paper that they print on costs money. They have not adjusted to the reality of the Internet. On the other hand, there has been a continuing problem with journals publishing wrong or damaging statements and then refusing time and space to set the record straight which suggest that the problem is institutional rather than monetary.

• C. Baxter
  Posted Oct 15, 2009 at 12:42 AM | Permalink

  Re: John A (#5), This is a good point. I’m sure in all sciences (it’s certainly true in physics) there are often one or two people who regard their own research as having said everything there is to say about a particular area. These people are frequently highly combative and overly assertive (a polite way of saying arrogant) and take great offence at anyone presently anything new. If you get one of these people as a reviewer then God help you. Appeals to the editorial board of the journal are usually not successful as they invariably side with the reviewer. In physics journal (APS and the UK IoP journals), the solution of allowing the reviewer to publish his comments along side that of the submitted (and dammed paper) is not an option.

• MrPete
  Posted Oct 15, 2009 at 1:51 AM | Permalink

  Re: John A (#5),
  I don’t really think blogs substitute for journals either. They are different beasts.

  In some ideal world, I think the intense dialogue of a properly run blog would provide a great level of multidisciplinary feedback not found anywhere else… and would be a great precursor to “archival” publishing in a journal.

  Journals are great for posterity. They seem pretty awful for getting the science right in the first place.

don

I suppose in the ideal world journals are about doing science and blogs are about doing policy–that is politics. Is that a valid assumption? In the real world the institutional problem seems to be the current overlap between the two. Is this because of rogue behavior violating defined methods to achieve desired policy goals or are even the methods and what constitutes the science of climate the issue? If one cannot agree on what a fact is–the thermometer is up side down– much less its meaning–is an up side down thermometer’s readings significant–then intelligent conversation is impossible. Needless to say, the lack of personal accountability for bogus meter readings provides the incentive for abuse, especially for egos on a mission to save the planet for a profit, playing the prophet.

michel

The most extraordinary think about Revkin’s comments is that he thinks there is some kind of public policy issue to be debated. There is not. There is no decision to be made. It does not matter what anyone thinks about blogs, and discussions on all kinds of topics in them, they are here to say. It is equally pointless to ask whether newspaper reports on new developments are ‘useful’ or whether such reports should be confined to the science press.

Revkin should have asked another, reasonable, and quite legitimate question. This would be something like this: when people without academic affiliations are researching topics with public policy implications, and arriving at conclusions different from the consensus, how is it most effective for them to communicate their results? Should they confine themselves to the peer reviewed journals? Or is it more effective to communicate through the cut and thrust of debate on blogs?

Another legitimate question is whether, for the informed and curious voter, should he confine himself to reading the peer reviewed journals, science mags, and national newspapers, or are the various science blogs now cost effective for him, in terms of time effort and rewards, to add to this reading? And if so which blogs on which topics?

The underlying idea that there is any sort of public policy issue in whether things of various sorts are or are not discussed in blogs is positively sinister, and Revkin should know better.

Stuart Huggett

I find an interesting parallel between the founding of the Royal Society by a bunch of ‘natural philosophers’ who take the intellectual high ground away from religion and the astrologers and the way in which this medium is severely taxing the authority of the journals. The print media – mainly newspapers – seem to be having a particularly hard time of it and new and more effective ways of dealing with information flow are inevitably emerging. I would not expect Revkin to be able to look at this situation very creatively – or can he? As an aside, I think the fact that an architect from Fiji is able to make a comment here on this matter is an extraordinary thing in itself. (Irrelevant note: Christopher Wren was a founding member of the RS).

Jean S

It is true that Mann et al used consistently wrong sign for Korttajärvi series in the CPS method. However, the situation in EIV method is even more “bizarre”. In EIV method, the sign is determined independently for each step (century), and signs may change from step to step. This can been seen here (compare for example “tiljander_2003_xraydenseave” series in AD500 and AD600 steps; blue indicates positive and red negative weight relative to the orientation in “raw” data files):

Let me try to elobrate this a bit. Suppose CRU was flipping signs of their station data anomalies, say, every decade in their NH temperature average calculations. How many people would defend that? Some people (not me, I hasten to add) might say the analysis would be complete rubbish.

• Jason
  Posted Oct 15, 2009 at 3:14 AM | Permalink

  Re: Jean S (#23),

  I’d suggest that CRU flipping a single station (i.e. treating high temperature readings as indicating a lower temperature) would be sufficiently scandalous; whether or not they did this inconsistently.

  Changing the sign between centuries is merely another indication that Mann is using methods that have not and could not survive peer review by an actual statistician.

  • Steve McIntyre
    Posted Oct 15, 2009 at 11:44 PM | Permalink

    Re: Jason (#24),

    I’d suggest that CRU flipping a single station (i.e. treating high temperature readings as indicating a lower temperature) would be sufficiently scandalous; whether or not they did this inconsistently.

    Actually the MBH98 Mannomatic did this. Some of their “proxies” were actually temperature series. When you worked through the weights from the Mannomatic in the reconstruction step, something that I did a few years ago, some of the weights for instrumental temperature series were negative.

    It didn’t “matter” in the sense that the only thing that “mattered” were the bristlecones, but it was amusing in terms of the absurdity of the Mannomatic methodology.

• Nick Stokes
  Posted Oct 15, 2009 at 5:31 AM | Permalink

  Re: Jean S (#23), I read a lot that says there must be something wrong with changing signs. But I’ve asked for a pointer to where the effects are demonstrated. My hope is that someone will have done calculations with alternate signs to show the different results.

  • Jason
    Posted Oct 15, 2009 at 6:49 AM | Permalink

    Re: Nick Stokes (#35),

    The problem with changing signs is NOT that it changes the results. The problem is that it demonstrates that the fundamental assumptions underlying the statistical methodology are flawed.

    The basic assumption behind any multiproxy temperature reconstruction is that the individual proxies contain a signal which has a constant relationship with temperature over time. Thus proxies which are known to be related to temperature during the last 100-200 years (when humans recorded temperature) are assumed to have the same relationship during the period preceding human recorded temperature. It is only on the basis of this assumption that Mann claims that his calculated graph has any relationship to temperature.

    To the extent that a proxy is treated as being positively related to temperature at some times, and negatively related to temperature at other times, this assumption is being disregarded.

    A scientific argument is logically invalid if it assumes X, and then proceeds based on calculations that clearly violate X.

    Mann attempts to resuscitate his effort by calculating how likely it is that random data would have produced similar results. Even if his calculations during this step were correct, it says nothing about temperature. All this test shows (again, if we act as if Mann’s calculations were correct) is that the input data was not randomly selected. The only basis for attributing this non-randomness to temperature is Mann’s fundamental assumption (which again is violated when he calculated the shape of his reconstruction).

    There is an alternative explanation for the lack of randomness which is well documented here, and which is not contradicted by Mann’s intermediate calculations. Specifically, the decisions and methods that produced the initial pool of proxies may have artificially selected for proxies containing a hockey stick. The hockey stick signal would then be amplified by Mann’s statistical process which checks for similarity between the input proxies and the recorded temperature record. Since the recorded temperature record contains a pronounced hockey stick blade, proxies possessing a blade were amplified, while other were minimized (or, in this case, deselected entirely). Consequently, a relatively small number of proxies displaying bias would have a disproportionate impact on the end result, quite possibly enough to pass a test for statistical significance.

    • Dave Dardinger
      Posted Oct 15, 2009 at 7:53 AM | Permalink

      Re: Jason (#39),

      the initial pool of proxies may have artificially selected for proxies containing a hockey stick.

      This is indeed the essence of what was done, but of course it isn’t described that way. It’s described as correlating with instrumental temperature records. Since the recent instrumental temperature record is, in fact, a HS, that is what is selected. If, in fact, there is indeed a true signal for temperature in the proxies, all is well. Selecting such records will enhance the signal and allow determining its value. But there are three other possibilities which don’t give such good results:

      1. The proxies don’t respond to temperature at all, in which case the whole project is worthless, but the very process of assuming it’s true will result in a homogenization of the early temperature reconstruction; the temperature in the early 1800s will be pinned low since that’s what the instrumental record says and modern temperatures will be pinned high.

      2. If the proxies respond somewhat to temperature but somewhat to other things, there will still be a structure to earlier temperatures but it will be reduced in amplitude and thus the modern temperatures will be higher WRT earlier temperatures than the real temperatures would be if they could be independently determined.

      3. If some proxies are good temperature responders but most aren’t then there will be an admixture of correct earlier temperatures with incorrect ones resulting in a situation similar to #2. This is because some of the proxies selected will randomly have modern values which look like they’re responding to modern temperatures but aren’t.

      I suspect 2 or 3 are more likely than 1, but in all cases comparing say, MWP to modern temperatures will be bogus. And, IMO, anyone who can read the above and understands it needs to have some sort of star next to their name in posts to indicate they get it.

    • Jason
      Posted Oct 15, 2009 at 8:12 AM | Permalink

      Re: Dave Dardinger (#46),

      I think I’m saying something different.

      As I understand his methods, Mann claims to test for spurious signals introduced by his methodology by calculating the probability that random data (examined via the same method) would have caused a signal of equal or greater magnitude.

      He claims to show that it is unlikely (to the point of statistical significance) that such a large spurious signal would be detected. Were this calculation correct, then any signals introduced by Mann’s methods alone could be be ruled out.

      When I say that “the initial pool of proxies may have artificially selected for proxies containing a hockey stick”, I an specifically saying that the set of proxies that Mann chose to input into the first step of his calculation (BEFORE checking for correlation with local temperature) were created and/or selected in such a manner as to create a bias.

      The latter calculations then only have to enhance that bias (which, assuming the presence of a bias, they undeniably would) to escape detection via his test for statistical significance.

  • RomanM
    Posted Oct 15, 2009 at 8:05 AM | Permalink

    Re: Nick Stokes (#34),

    I read a lot that says there must be something wrong with changing signs.

    … but you don’t seem to understand the problem.

    It is basically the same flaw as with the earlier Alice-in-Wonderland use of teleconnection in climate reconstructions.

    Here, we have a proxy the form of whose response to climate has been identified by the original researchers. Dr. Mann and crew decide to use the proxy upside down with the nebulous justification that it doesn’t matter (“Multivariate regression methods are insensitive to the sign of predictors”).

    The missing portion is a proper explanation of the physical reasons for why, in this particular case, the relationship between climate and proxy response has been not only reversed, but with equal magnitude of effect. Unsupported by such physically tenable reasons, the whole exercise is nothing more than opportunistic selection of numbers to produce a previously ordained outcome. The use of teleconnection for including proxies which did not match their local environment, but had the proper hockey stick “global” shape were also never justified by a specific accounting of an actual physical process which could make that a genuine explainable effect.

    This still would not explain the validity of a statistical process which “swings both ways” using the proxy in each orientation (while the other proxies remain stable) at different times as demonstrated by Jean S in comment 23. I think the term “bizarre” might apply to that even more strongly.

    Your attempt to divert the thread to a “It doesn’t matter so it’s OK” line is simply disingenuous, but that just seems to be your preferred operating procedure as past experience has demonstrated.

    • bender
      Posted Oct 15, 2009 at 8:17 AM | Permalink

      Re: RomanM (#50),
      Yes, the sign going in doesn’t matter. But the coefficient coming out sure does. If it’s directly opposite to what’s a;ready been published you kinda sorta have a problem.
      .
      And one that is easily diagnosed and rectified, I might add! The lack of attention to such obvious, and easily rectified, contradictions is quite disturbing evidence of confirmation bias.

steven mosher

what is a paper in a journal? A paper in a journal is a blog post. The post button is pressed by an editor after the post is reviewed by a group of people who deem it worthy of posting. The comments on this ‘blog’ are heavily moderated and word limited. The thread closes fast.

It would be interesting if CA did an experiment.

1. select a topic to be “blogged” as we know it.
2. use peer review before its posted. get rid of reviewer anonymity.
3. Comments are posted without moderation or word limits.
4. readers can digg comments. ( basically vote)
5. reviewers review the top comments and select those that are most germane and should be addressed by the author

or something like that.

John A

Why can Michael Mann not, with Steve consent and clear credit for source, use some of Steve’s idea’s approaches or data for inclusion in his own work? Not of the same rank as a proper citation, or dual authorship but instead some form of creditation.

Erm. Where do I begin?

Craig Loehle

Considering that Mann et al believe in “teleconnections” of local sites with hemispheric climate, or that it is valid to view a site as a global indicator rather than reflecting strictly local temperature history, any correlation is a good correlation and there is no such thing as “upside down” thus SM is speaking nonsense. There is no sense that things like varves reflect a physical process and have a physical meaning. It is simply a proxy. It only appears that Mann and SM are speaking the same language, but they are not.

• Jean S
  Posted Oct 16, 2009 at 1:14 AM | Permalink

  Re: Craig Loehle (#37), Re: RomanM (#50),
  IMO, Mann et al took this “teleconnection” idea up to a new, borrowing some Team terminology, unprecedented level in their EIV reconstructions.

  First, it is required (screening) that proxies correlate with local temperature (or actually, some temperature nearby as this is done with the “pick best out of two correlations”-method). Then, in the actual reconstruction exercise, proxies are calibrated (through RegEM) against hemispheric temperature. Now the determined proxy weights may (and do) have different sign than indicated by the screening with respect to local temperature. And not only that, the reconstructions are done in a stepwise manner leading to a situation, where the weight signs (and of course magnitudes also) may change (and are changing) over the time.

  In other words, proxies are required to respond local temperature in a certain way, but their “respond” to hemispheric temperature (“teleconnection”) may be of opposing sign, and the sign and magnitude of the respond are changing from century to century! The story is so amazing, that I doubt even the best science fiction writers could come up with that!

  • bender
    Posted Oct 16, 2009 at 5:41 AM | Permalink

    Re: Jean S (#173),
    Oh, but Nick Stokes spotted that instantly. I’m sure there is a defense in the pipeline.

Steve Geiger

not quite sure added space for comments (PNAS comments < 250 words) would make much of a difference, would it? Seems it would still be just as easy to label an argument ‘bizarre’, without further discussion.

bender

The orientation would not “matter” if all they were doing was a one-stage multivariate regression. In that case they would simply have a proxy whose interpretation is (bizarrely) opposite to that of the scientist who originally collected the data. But they are not only doing a multivariate regression, they are doing (i) a modern calibration and then (ii) a historical reconstruction. They calibrate on the modern uptick portion – the piece that should have been removed (as Kaufman recognized) – and reconstruct on the remainder. Thus the “inversion” doesn’t happen in the calibration stage (which is probably what Mann thought, and why he might have used the term “bizarre”). It happens when the bogus calibration is misapplied in the reconstruction stage. This forces the historical portion of the series to take on the opposite interpretation of what Tiljander herself proposed.
.
Along the lines that Nick Stokes suggests, it would be interesting to see the difference with the bogus blade included versus removed. The former should force an “inversion” (i.e. proxy interpretation bizarrely opposite to Tiljander’s). The latter not.
.
IMO it would be tough to explain this effect in 250 words. Also, you would need at least one figure to demonstrate, and that is often not allowed in replies.

Morgan

Peer review is simply a way to make sure that what gets published in the journal is of interest to the target readers of the journal, which usually means creating an elevated sense of the credibility of the articles. Any “quality control” function is incidental – if the editors believe the readers want quality, they will arrange the review process so that it successfully signals quality to the readers, and as cheaply as possible. That usually means through proxies (quality itself requiring a great deal of effort to ascertain) – things like “well-respected reviewers” who are “published within the field”. It also means that a comment system that makes it difficult to demonstrate the low quality of previously published research is highly desirable.

Peer review will attempt to ensure actual quality (as opposed to proxy quality) only when readers demand it. I hope that blogs like Steve’s will eventually cause a groundswell from journal readers wishing to improve the actual quality of published research.

• Steve McIntyre
  Posted Oct 15, 2009 at 9:13 AM | Permalink

  Re: Morgan (#63),

  Let’s not discuss peer review in general terms. The issues are well understood by readers here.

  I’m trying to focus here on a very specific case. It seems unsatisfactory from a quality control perspective that such a simple issue as whether Mann used the Tiljander series upside down should remain in dispute after being specifically raised in an exchange in the PeerReviewedLitchurchur.

  This is a very elementary point and yet it remains in dispute nearly a year after being raised.

  It seems to me that the editor of the journal should have required them to directly deal with the comment, rather than permit them to introduce a lot of irrelevant fog.

  But even more, it seems to me that the authors of Mann et al 2008 had a responsibility to concede that they had used the series upside down from the Tiljander orientation once it was brought to the attention of the journal. A question for readers – is there a point at which the failure to concede a point like this creates a situation where the research is not accurately represented in the research record?

  Despite my request that readers avoid generalities, too many readers are trying to litigate every issue all at once. There are some important distinctions in this case. As I observed in my post, in my opinion, there is an important difference in the state of mind at the time of the original article and at the time of the Reply. Does this create different obligations?

  • bender
    Posted Oct 15, 2009 at 9:23 AM | Permalink

    Re: Steve McIntyre (#65),
    Steve, the issue is how diversity of expert opinion is to be summarized when the goal is consensus. The IPCC model is flawed. But they will argue that the flawed process can be defended under the “precautionary principle”: “We don’t have time to cope with uncertainties and caveats, never mind the right wing skeptics. The science is complex enough as it is.”

  • bender
    Posted Oct 15, 2009 at 9:46 AM | Permalink

    Re: Steve McIntyre (#65),

    it seems to me that the authors of Mann et al 2008 had a responsibility to concede that they had used the series upside down from the Tiljander orientation once it was brought to the attention of the journal. A question for readers – is there a point at which the failure to concede a point like this creates a situation where the research is not accurately represented in the research record?

    They had the right to complain about your vernacular. If they understood your point, they had a moral obligation to concede it. If they did not understand the point, they might have obtained clarification before commenting further, but exercised their right to not do so. I do not believe the prevailing culture would oblige them to do so. Pleading ignorance is always available as an out. (Only the Editor can prevent this endpoint outcome. He is not obliged to do so. But one does expect it of a good Editor.)
    .
    This points to a serious flaw in the peer review process as a mechanism for ensuring the level of due diligence that policymakers need so that they can sell their policies as “science-based consensus”. More eyeballs. More inclusivity. More back-and-forth. When you are talking global meltdown, you want maximum inter-neural connectivity. If the alarmists are right, then we can’t afford to get this wrong.

    • bender
      Posted Oct 15, 2009 at 9:56 AM | Permalink

      Re: bender (#73),
      … and if they did concede the point – that their interpretation of the proxy is opposite to that of the originator of the data – they would be obliged to explain that fact. They would then be forced to either defend their interpretation, putting them in argument Tiljander, or accept that it is wrong. If they accept it is wrong, or could be wrong, then the only obvious solution is dropping Tiljander from the network and recalculating the results. If nothing changes then it greatly diminishes the weight of the original complaint.
      .
      But there’s that pea under the thimble. Removal of Tiljander “doesn’t matter” … as long as something else is in there, acting as a backup crutch. Take out both crutches and the patient falls over.

    • Nick Stokes
      Posted Oct 15, 2009 at 4:13 PM | Permalink

      Re: bender (#76),

      If they accept it is wrong, or could be wrong, then the only obvious solution is dropping Tiljander from the network and recalculating the results.

      That seems to be what is done in the thread to which Jean S (#36),directed me. But I can’t see why the more obvious solution isn’t to correct the sign error and recalculate (with Tiljander). And since I believe the calcs have been replicated here, I’ve been asking whether anyone has done that, and can report the difference. It seems that according to Mann, there would be none, and that can be checked.

    • Nicolas Nierenberg
      Posted Oct 15, 2009 at 4:39 PM | Permalink

      Re: Nick Stokes (#119),

      I believe this is missing the point. Changing the sign of the series wouldn’t make any difference to the result, the sign of the coefficient would just invert, and the result would be the same.

      The problem is that this would still be the opposite sign to the physical meaning of the series.

      Take tree rings as a simple example. Imagine that during a period of measured increasing temperature a tree ring series got smaller. This type of algorithm would assign a negative coefficient to the tree rings in that series, that would then propagate to the non-instrumental period. If you try to fix it by inverting the sign of the tree ring series the algorithm would then sign a positive coefficient and the result would be unchanged, but this would then be inverse to the already known physical property of increasing tree rings with increasing temperature. (And please don’t let this set everyone off on whether that relationship is really always true.)

    • Nick Stokes
      Posted Oct 15, 2009 at 5:32 PM | Permalink

      Re: Nicolas Nierenberg (#121),

      Take tree rings as a simple example. Imagine that during a period of measured increasing temperature a tree ring series got smaller. This type of algorithm would assign a negative coefficient to the tree rings in that series, that would then propagate to the non-instrumental period.

      This is the issue of consistency. If you are using treerings “upsidedown”, then they will behave similarly in the calibration period and the proxy period, and nothing is lost. But if the tree really responds inversely in the two periods, then there is a real problem whatever convention you use.

      Re: Steve McIntyre (#125),

      Huh??
      You’re completely misunderstanding the issue.
      In a CPS multiproxy study (one of the two legs of M08), the sign absolutely does “matter”.

      Well, I’m looking for numbers to help me understand it. It seems reasonable for Mann to look for an indication that someone else had response to a paleoclimatic signal as a proxy guide, and I don’t see why that binds him use their sign convention (although I would have). And I’m trying to understand the sense in which the portion is “contaminated”. I’m trying to distinguish between using a different convention and actual wrong results, and looking for some quantitative measure. A recalc with the “right” sign would be a good guide.

    • RomanM
      Posted Oct 15, 2009 at 5:57 PM | Permalink

      Re: Nick Stokes (#129),

      Well, I’m looking for numbers to help me understand it. It seems reasonable for Mann to look for an indication that someone else had response to a paleoclimatic signal as a proxy guide, and I don’t see why that binds him use their sign convention (although I would have). And I’m trying to understand the sense in which the portion is “contaminated”.

      You don’t need numbers. You need to understand what the CPS procedure does. From an earlier post by Steve:

      As a technique, CPS is little more than averaging and shouldn’t take more than a few lines of code. You standardize the network – one line; take an average – one more line; and then re-scale to match the mean and standard deviation of your target – one more line.

      In CPS the proxies are averaged. Do you need to see numbers for an explanation of what effect flipping a proxy has? While other proxies are going up, this one is going down and vice versa. The whole thing is “contaminated”!

      You don’t think it is incumbent that someone who decides(?) to use it the opposite way should be aware of how it was originally used AND give a reason for flipping it. Saying the methodology will flip it back doen’t cut it.

      Get a reality check!

    • bender
      Posted Oct 15, 2009 at 6:46 PM | Permalink

      Re: RomanM (#132),
      And if all they are doing is CPS, then it is even simpler and this makes sense.

      Steve: they do both in M08.

    • Nick Stokes
      Posted Oct 15, 2009 at 6:53 PM | Permalink

      Re: RomanM (#132),

      In CPS the proxies are averaged…

      Thanks, that may help. It seems the key step in CPS as described in Mann’s SI is
      “All proxies available within a given instrumental surface temperature grid box were then averaged and scaled to the same mean and decadal standard deviation …”. So if at that stage the Tiljander proxy was added in with inverted sign, it would make a difference. In the paper a “potential” weighting factor (which could be negative) is mentioned, but not in the SI. Do we know that the Tiljander proxy had the wrong sign at this averaging stage?

      The SI also says “As a cross-check we considered two alternative CPS procedures. The first procedure attempts to preserve signal amplitudes by matching trends rather than variances …”. Unlike variance, the trend is signed, so that could compensate for any inversion. They say this (and another procedure) “generally produced similar results, with isolated exceptions”.

    • Nicolas Nierenberg
      Posted Oct 15, 2009 at 6:32 PM | Permalink

      Re: Nick Stokes (#129),

      In the specific case being discussed the proxy is known by the author’s of the original paper to behave differently in the calibration period. But since the upward slope in that period correlates with temperature the algorithm assumes that is the correct direction for the proxy and propagates that back to the earlier periods. This is why Steve describes it as being “upside down.”

    • bender
      Posted Oct 15, 2009 at 6:47 PM | Permalink

      Re: Nicolas Nierenberg (#136),
      And this makes sense.
      .
      So what doesn’t Nick Stokes get?

    • Nick Stokes
      Posted Oct 15, 2009 at 6:59 PM | Permalink

      Re: Nicolas Nierenberg (#136), I’m sorry I’m catching up here – do you have a reference for that known kinky behaviour of the Tiljander proxy? At the moment, Romanm’s explanation looks more straightforward to me.

    • bender
      Posted Oct 15, 2009 at 7:07 PM | Permalink

      Re: Nick Stokes (#143),
      Read Tiljander’s own paper. It’s available by searching CA.

    • Nicolas Nierenberg
      Posted Oct 15, 2009 at 7:26 PM | Permalink

      Re: Nick Stokes (#143),

      Two different topics. Read the original post by Steve on this thread. I believe it has all the references you need on this topic.

    • Nick Stokes
      Posted Oct 15, 2009 at 9:16 PM | Permalink

      Re: Nicolas Nierenberg (#149),
      OK, thanks, the “Saturday Night Live” thread seems to have the explanation, and does indeed describe the kinked behaviour. In recent times agriculture etc makes the proxy appear to get colder (more clay), and if you calibrate with instrumental temperature it comes out upside down in the earlier period. On that basis, presumably Tiljander data can’t be reliably calibrated at all with instrumental temperature, and should indeed be omitted.

      This seems to relate to the note in the Mann 2008 SI which says

      These records include the four Tijander et al. (12) series used (see Fig. S9) for which the original authors note that human effects over the past few centuries unrelated to climate might impact records (the original paper states ‘‘Natural variability in the sediment record was disrupted by increased human impact in the catchment area at A.D. 1720.’ and later, ‘‘In the case of Lake Korttajarvi it is a demanding task to calibrate the physical varve data we have collected against meteorological data, because human impacts have distorted the natural signal to varying extents’)

      Incidentally, although most of these high-latitude sites seem very remote, I have indeed seen Lake Korttajarvi from the airport bus.

    • Steve McIntyre
      Posted Oct 15, 2009 at 9:22 PM | Permalink

      Re: Nick Stokes (#158),

      Nick, again, you seem to be wilfully obtuse to what is going on here. Mann’s notation in the SI does not quarantine the problem, as I stated in the post. Having noted the problem with the Tiljander sediments, why did he use them at all? What’s so pressing about having a contaminated series? Mann’s sensitivity study without the Tiljander sediments was not a no=dendro sensitivity study without the Tiljander sediments. For that study, he included Graybill bristlecones.

      The Tiljander sediments are used in the recon that supposedly shows that they don’t need tree ring proxies – the recon that Gavin Schmidt cites so proudly.

      You’re not watching the pea.

    • bender
      Posted Oct 15, 2009 at 9:25 PM | Permalink

      Re: Steve McIntyre (#159),
      #160 Crosspost.
      There are 2 peas, 22 thimbles. Makes the game a bit harder.

    • bender
      Posted Oct 15, 2009 at 9:23 PM | Permalink

      Re: Nick Stokes (#158),

      presumably Tiljander data can’t be reliably calibrated at all with instrumental temperature

      Did you read Tiljander yet? These problems were known BEFORE Mann et al. 2008. There’s a reason they didn’t do a calibration. There’s a reason why they go into tremendous detail about the paleohistory of the area. So why – despite all the caveats – does the thing get fed into another Mannomatic?
      .
      Note further that Tiljander questions Lamoureux & Bradley’s rather bold and questionable assertion that lake sediment varves ought to be considered a temperature proxy until proven otherwise. Why’s everyone so dismissive of Tiljander’s hard work?

    • Pat Frank
      Posted Oct 16, 2009 at 7:03 PM | Permalink

      Re: Nick Stokes (#158), “…can’t be reliably calibrated…”(bolding added)

      An abuse of the term that is endemic in tree ring thermometry. ‘Correlated’ is the correct term. And, of course, correlation = causation. True science, alright.

    • Micky C
      Posted Oct 15, 2009 at 4:47 PM | Permalink

      Re: Nick Stokes (#119), I don’t believe it would make a difference and that Mann is indeed correct because any PCA-like reduction of the data would pick up the largest variance not the sign. I believe the sign would be transfered to the weight assigned to the series when the PCs are created (this is a bit vague, I know PCA is a linear transformation). I’m trying to think if it would have any impact on the magnitude of output PCs that would be used in the calibration regression? Maybe not.
      Someone needs to check. However although this is a mathematically consistent argument it is not a physically consistent argument (the series is upside down) and since the outcome of the reconstruction is to produce a physical reconstruction, the Mann answer is in its full context, incorrect.

    • Mesa
      Posted Oct 15, 2009 at 5:27 PM | Permalink

      Re: Micky C (#123), The coefficients have to be constrained to make physical sense. The math would “work” either way, but you wouldn’t want to use the series with a net sign in the wrong sense to the presumed physical temperature/sediment relationship. I think this example is a low variance signal with a high variance noise (bridges, peat ditching, forest clearing) added at the end with a spurious correlation to global temperature anomalies in the instrument period. The regression (or PCA) will probably minimize the error of the high variance noise – ie pick the sign the wrong way to what is the accurate physical temperature/sediment relationship. To the extent there is a physical relationship to temperature that is significant, this should reduce the accuracy of the overall proxy in the pre-calibration period. However, it’s quite possible that the spurious non-physical correlation may increased the R2 during the calibration period, without any physical meaning. If you understood that the wide rings in the 20th century were outliers or at least non-temperature related data, you would want to address this issue in some other way, to put it mildly.

    • bender
      Posted Oct 15, 2009 at 6:44 PM | Permalink

      Re: Mesa (#128),
      That’s my understanding of it. Steve?

    • Nick Stokes
      Posted Oct 15, 2009 at 7:25 PM | Permalink

      Re: bender (#76),

      If they accept it is wrong, or could be wrong, then the only obvious solution is dropping Tiljander from the network and recalculating the results. If nothing changes then it greatly diminishes the weight of the original complaint.

      Oddly enough, I see they have actually done that in the SI. Not because of the sign issue, but because of other quality doubts about the 4 Tiljander series, and three others. So they dropped them, and showed the alternative results in Fig S8. Unfortunately, the colors in my version are unclear (to my eyes), but the difference seems noticeable but not huge.

    • bender
      Posted Oct 15, 2009 at 7:30 PM | Permalink

      Re: Nick Stokes (#148),
      You are not reading everything that is relevant. Read the entire thread here. I’m aware of what’s in the SI. And what isn’t.

    • bender
      Posted Oct 15, 2009 at 7:36 PM | Permalink

      Re: Nick Stokes (#148),

      the difference seems noticeable but not huge

      1. We’re not looking for HUGE. We’re looking for “small” differences that nevertheless tip the balance toward OMGIWTWT.

      2. See my comment above:

      But there’s that pea under the thimble. Removal of Tiljander “doesn’t matter” … as long as something else is in there, acting as a backup crutch. Take out both crutches and the patient falls over.

      You say they take out the offensive Tiljander series. What do they leave in?

    • Mesa
      Posted Oct 15, 2009 at 7:52 PM | Permalink

      Re: Nick Stokes (#148), From original post:

      The SI to Mann et al 2008 conceded that there were problems with the recent portion of the Tiljander proxies (without mentioning that they were using them upside down from the interpretation of Tiljander and Finnish paleolimnologists), but argued that they could still “get ” a Stick without the Tiljander sediments. However, as I observed at the time, this case required the Graybill bristlecone chronology (where they failed to mention or cite Ababneh’s inability to replicate Graybill’s Sheep Mt results, even though Malcolm Hughes, a member of Ababneh’s thesis panel was a coauthor of Mann et al 2008). Thus their “robustness” analysis used either upside down Tiljander sediments or Graybill bristlecones.

    • bender
      Posted Oct 15, 2009 at 8:00 PM | Permalink

      Re: Mesa (#152), bingo!

    • bender
      Posted Oct 15, 2009 at 8:01 PM | Permalink

      Re: Mesa (#152),
      Someone’s reading. Hallelujah.

    • steven mosher
      Posted Oct 15, 2009 at 9:27 PM | Permalink

      Re: Nick Stokes (#148), ah +-noticeable rather than +-huge.

      You see nick here is where I have problems with your defense. When mann mislocated the rain in seine, when he got the coordinates demonstrably WRONG, he wouldn’t correct them. he even repeated the mistake ” it didnt matter”

      Well if it didn’t matter then why not just THANK STEVE and fix it. just for neatness sake. there is something similar happening here. If it doesnt matter, or if it matters just a little, why not admit.

      1. the original study had the sries oriented one way, a way that made physical sense.
      2. Run the analysis with the data oriented this way.. since it doesnt matter! or matters just a little.

      What is it with not being able to admit errors, even when they don’t matter or matter just a little

      ( ps I’m not granting this problem doesnt matter)

      can you SIMPLY and without buts say.

      1. mann inverted the series.
      2. Any one with a math background can see this.
      3. I nick can see this.
      4. I nick would orient the series the way the original author intended.

      let’s grant it doesnt matter. that actually works IN favor of orienting it the “right” way.

      Do you invert it just because you can? just for the hell of it? because you don’t know how to keep it in the
      right orientation?

    • bender
      Posted Oct 15, 2009 at 9:54 PM | Permalink

      Re: steven mosher (#162),
      I can’t even grant that much grace. These “small” errors are always in the direction required to elevate the paper from ho-hum Journal of Quaternary Research or Holocene to Nature or Science. These guys are racing each other to sound an ever more alarming alarm.

    • Nick Stokes
      Posted Oct 15, 2009 at 10:51 PM | Permalink

      Re: steven mosher (#162),

      Run the analysis with the data oriented this way.. since it doesnt matter! or matters just a little.

      No, that was my first thought too, from the talk of “upside down”. Just turn it right way up. That’s why I asked initially if anyone could point where that had been done.

      But it won’t work. As NN said, it’s wrong either way. Inverted again, it’s wrong for the instrumental period. As Steve said here, and I said in the previous comment, the Tiljander data should not be used here. Mann’s Fig S8 was more appropriate.

    • jeez
      Posted Oct 15, 2009 at 11:05 PM | Permalink

      Re: Nick Stokes (#166),

      Addressing

      Re: steven mosher (#162),

      While bypassing

      Re: Steve McIntyre (#159),

      Classic logical sophistry.

    • Nick Stokes
      Posted Oct 15, 2009 at 11:20 PM | Permalink

      Re: jeez (#168), I didn’t bypass it at all
      “As Steve said here, and I said in the previous comment, the Tiljander data should not be used here“.

    • Steve McIntyre
      Posted Oct 15, 2009 at 11:24 PM | Permalink

      Re: Nick Stokes (#166),

      once again, you’re being wilfully obtuse about bristlecones and Mann 2008.

      I’m glad that you agree that the hockey stick that Gavin SChmidt displayed so proudly at the Yamal thread at realclimate – the supposed hockey stick without bristlecones – is not appropriate due to its use of upside down Tiljander (and this isn;t the only problem).

      But, as I noted before, Fig S8 used bristlecones – a proxy that the NAS panel said should be “avoided”. I presume that you agree that bristlecones should not be used in temperature reconstructions, as the NAS panel recommended (and Mann said that he complied with NAS recommendations but actually flouted them).

      Unless you are willing to justify the continued use of Graybill bristlecones (also repudiated by Ababneh), you must also concede that Fig S8 also is inappropriate.

    • Nick Stokes
      Posted Oct 16, 2009 at 12:31 AM | Permalink

      Re: Steve McIntyre (#170),

      once again, you’re being wilfully obtuse about bristlecones and Mann 2008.

      Gosh, there’s always something 😦 I don’t think I’ve mentioned bristlecones, and I wasn’t planning to. I’m glad you folk have given me some pointers to understand the Tiljander issue. Bristlecones might have to wait a while.

    • bender
      Posted Oct 16, 2009 at 5:38 AM | Permalink

      Re: Nick Stokes (#172), Yes, there IS always something, isn’t there? And whose fault is that? The auditors, of course. Willfully obtuse, you can say that again.

    • Patrick M.
      Posted Oct 16, 2009 at 7:08 AM | Permalink

      Re: Nick Stokes (#172),

      No, it is the perfect time to look at bristlecones and PC’s too. If you want to understand how CA views Mann, you need to follow the trail. It’s the pattern that is disconcerting. That’s what you are missing; you are looking at each tree as completely separate. Look at the forest.

      Then once you understand how Mann operates, look at the other members of the “Team”.

    • Steve McIntyre
      Posted Oct 16, 2009 at 7:56 AM | Permalink

      Re: Nick Stokes (#172),

      pending your consideration of bristlecones, a topic that has been endlessly discussed here, you might post over at realclimate that the non-dendro HS of Mann 2008 shown in their Yamal thread cannot be used because of upside down Tiljander.

      I remind you of Atte Korhola’s (an eminent Finnish paleolimnologist) comment (as translated) on the upside down use of Tiljander and other proxies( in this case, Kaufman, but Mann’s use is far worse because he used the contaminated portion):

      It is concluded in the article that the average temperatures in the Arctic region are much higher now than at any time in the past two thousand years. The result may well be true, but the way the researchers ended up with this conclusion raises questions. Proxies have been included selectively, they have been digested, manipulated, filtered, and combined, for example, data collected from Finland in the past by my own colleagues has even been turned upside down such that the warm periods become cold and vice versa. Normally, this would be considered as a scientific forgery [ falsification??], which has serious consequences.

    • Richard Sycamore
      Posted Oct 16, 2009 at 9:12 AM | Permalink

      Re: Steve McIntyre (#179),

      I’m not sure why you wouldn’t post it yourself, but I have taken the liberty of doing it for you:

      Richard Sycamore says:
      Your comment is awaiting moderation.

      16 October 2009 at 10:04 AM
      There are two graphics in the opening post that contain problems and I ask whether they should be removed or amended.

      The first is the one derived from Kaufman et al. (2009). It makes inappropriate use of the Tiljander lake sediment proxy.
      On this topic, eminent Finnish paleoclimatologist, Atte Korhola, has stated [translation from Finnish]:

      “It is concluded in the article that the average temperatures in the Arctic region are much higher now than at any time in the past two thousand years. The result may well be true, but the way the researchers ended up with this conclusion raises questions. Proxies have been included selectively, they have been digested, manipulated, filtered, and combined, for example, data collected from Finland in the past by my own colleagues has even been turned upside down such that the warm periods become cold and vice versa. Normally, this would be considered as a scientific forgery [falsification?], which has serious consequences.”

      The second problematic graph is the one beneath it, derived from Mann et al. (2008) which makes the same innapropriate use of this proxy, attributing to it an interpretation that is directly opposite to that proposed by the orginal authors. It does not make sense to have proxies being interepted one way by one group and the other way by another. Given the contradiction, perhaps fix the graph or remove it?

      [Feel free to edit this text if any part of it is considered offensive.]

      Thank you.

      Please let me know if I have mis-represented the case. I rather dislike posting the comments of others, but in this case there is a need for resolution on the matter.
      Good day.

    • Steve McIntyre
      Posted Oct 16, 2009 at 9:40 AM | Permalink

      Re: Richard Sycamore (#188),

      I already submitted a related (but less pointed) post and, needless to say, it was censored.

    • bernie
      Posted Oct 16, 2009 at 12:07 PM | Permalink

      Re: Richard Sycamore (#188), I have looked at Real Climate for this post but no comments seem to have been added since yesterday.

    • Nick Stokes
      Posted Oct 16, 2009 at 10:30 PM | Permalink

      Re: Steve McIntyre (#179),

      you might post over at realclimate that the non-dendro HS of Mann 2008 shown in their Yamal thread cannot be used because of upside down Tiljander.

      I don’t think I have any special influence at RC. But I see that the question has previously been raised:

      Gavin, what is your opinion on the Finnish Lake sediment proxies? They certainly jump out when looking at the series used. Would you consider them quality temperature proxies?

      [Response: I have no particular opinion, but leaving them out (and a few other potentially problematic ones) was tested as described in the SI (p2 and fig S8). It makes a little difference, but nothing particularly striking. It’s worth pointing out that in these kinds of projects there are no guarantees – all you can do is attempt to find out how robust answers are to various issues (this is one, but so is the dependence on tree ring data, reconstruction methodology, etc, etc.) and seeing what decisions make a significant difference to the final answer. If they do, then obviously you are constrained about what can be concluded, while if they don’t matter, it’s not worth delving deeper. – gavin]

    • Ron Cram
      Posted Oct 16, 2009 at 10:45 PM | Permalink

      Re: Nick Stokes (#210),

      So Gavin is saying if a reconstruction is filled with nonsense from a variety of sources and you remove one bit of nonsense and the difference is not “particularly striking,” then you can put it back before you test the next bit of nonsense? That hardly seems scientific to me. I would much prefer “delving deeper” to make certain you are looking at data warm side up so to speak. I would also like to see a little more honesty when errors are found. All of this posturing and circling of wagons is a disservice to science.

    • Nick Stokes
      Posted Oct 16, 2009 at 10:52 PM | Permalink

      Re: Ron Cram (#211), There was a follow-up question:

      54
      Timo Hämeranta says:
      5 September 2008 at 12:32 PM

      Re 50

      The authors of the study explicitly state

      “. Natural variability in the sediment record was disrupted by increased human impact in the catchment area at AD 1720.”

      Ref: Tiljander, Mia, Matti Saarnisto, Antti E. K. Ojala and Timo Saarinen, 2003. A 3000-year palaeoenvironmental record from annually laminated sediment of Lake Korttajärvi, central Finland. Boreas, Vol. 26, pp. 566–577. Oslo. ISSN 0300-9483, December 2003

      When the study was published I discussed with Mia Tiljander and she stated that human impact, especially land cover change, agriculture, accelerates erosion, more material flows to lake sediments, and thereafter the lakes are of no use in proxy reconstructions anymore, Korttajärvi from 1720 on, most of the other 180.000 Finnish lakes from 1850s on.

      [Response: That is, I presume, why Mann et al did a test that didn’t include them. – gavin]

    • Michael Jankowski
      Posted Oct 16, 2009 at 11:29 PM | Permalink

      Re: Nick Stokes (#212),

      [Response: That is, I presume, why Mann et al did a test that didn’t include them. – gavin]

      Why include them in the first place if (a) there are serious issues that preclude them from use and (b) if it “doesn’t matter” whether or not they are included? Talk about incredible!

    • Peter
      Posted Oct 17, 2009 at 3:42 AM | Permalink

      Re: Nick Stokes (#210),

      Nick this has been spelled out to the point that fifth graders get it. You are now obsfucating, and nothing more.

    • Nick Stokes
      Posted Oct 17, 2009 at 5:04 AM | Permalink

      Re: Peter (#217),
      Obfuscating? What? Steve suggested I post something at RC, and I responded by pointing out related queries that had been posted with responses.

    • bender
      Posted Oct 17, 2009 at 5:26 AM | Permalink

      Re: Nick Stokes (#218),
      You’re “obfuscating” only in the sense that the issue is not whether the question was raised but whether it was resolved. That it wasn’t is evidenced by the two incorrect graphs.

    • bender
      Posted Oct 17, 2009 at 10:05 AM | Permalink

      Re: Nick Stokes (#218),
      Dodge noted.

    • Peter
      Posted Oct 17, 2009 at 12:04 PM | Permalink

      Re: Nick Stokes (#218),

      Nick, that is pathetic. You cite a RC question answered by Gavin citing figure S8!! The fact that figure S8 is Mannomatic bristlecone sausage and nothing more has been demonstrated here more than once, and is not in dispute. Please don’t cite S8 as proof of anything other than Mann’s intransgience.

      One more time. Tiljander upside down or otherwise is not a valid proxy by Mann’s own admission. The only thing which keeps the stick alive in M08 if you take out Tiljander is to add stripbark bristlecones, and the NAS said no stripbark bristlecones, full stop.

      The hockey stick breaks without one of these peas in your shell game.

    • Nick Stokes
      Posted Oct 17, 2009 at 3:30 PM | Permalink

      Re: Peter (#227), Again you and bender are missing the point. Steve suggested that I post something at RC, presumably to get their response. I pointed to where someone had done that. I’m not saying RC gave a good response; I’m just saying that the result of posting that query is known.

    • Steve McIntyre
      Posted Oct 17, 2009 at 4:12 PM | Permalink

      Re: Nick Stokes (#229),

      Nick, your reference is from last year. Maybe if they heard the comment from you (rather than last years’ commenter), it would help them understand. Wouldn’t do you any harm to try. Let us know how you fare.

      As to your point that they’ve already responded, I’ve noticed that you don’t immediately refrain from commenting here after I’ve answered a question. You’ve been known to voice your disagreement.

      I see no legitimate reason for you to express disagreement here (or at Lucia’s) and not doing so at realclimate, merely because Gavin said something a year ago.

    • Nick Stokes
      Posted Oct 17, 2009 at 5:41 PM | Permalink

      Re: Steve McIntyre (#230), I’m not averse to pressing an argument. But only if I have something new to say. In this case, we have a comment offerred in challenging style at RC by Tilo, who seems to be a scientist in the field who has discussed the matter with Tiljander. I could not muster that authority, and I can’t see what content I could add.

      Gavin responded by pointing to the Mann SI which flagged the Tiljander issues and did a T-free calculation. He seems to say that is sufficient; others (me too, I hasten to add) think the data should have been omitted entirely. I don’t see how the passage of time would change that difference of views.

    • EdBhoy
      Posted Oct 17, 2009 at 4:21 PM | Permalink

      Re: Nick Stokes (#229), This is repeatedly the case with any dissenting or difficult questions at Realclimate. They give a poor response which does not answer the question and then their acolytes point to the fact that it has been “discussed” as being evidence that they were right all along. The fact that Gavin moderates (ahem!) dissenting voices out of the blog renders it a useless forum for academic discussion.

    • Niels A Nielsen
      Posted Oct 17, 2009 at 4:38 PM | Permalink

      Re: EdBhoy (#231), Yes, Nick Stokes has the chance of improving the reputation of both himself and Realclimate in the eyes of many by voicing disagreement there. Do you think you would be censored, Nick?

    • Peter
      Posted Oct 17, 2009 at 5:56 PM | Permalink

      Re: Nick Stokes (#229),

      Nick, This point is more direct. Do you agree that the M08 hockey stick fails, breaks etc. without Tiljander and stripbark bristlecones. Yes or no?

      Do you agree that the NAS said “Stripbark bristlecones should not be used” or words to that direct effect, yes or no?

      Do you agree that the Tiljander series has demonstrated flaws? Yes or no.

      Please don’t lawyer up and parse meanings on this simple, direct post, it serves no purpose.

    • Nick Stokes
      Posted Oct 17, 2009 at 6:19 PM | Permalink

      Re: Peter (#234), Peter, everyone here seems to be an expert on the statistical vagaries of bristlecone pines. But as I said above, I’m not.

      I have now said several times that I think the Tiljander series is unsuitable for use as a temperature proxy. Incidentally, being careful with words has some worth. I’m sure Mia agrees with that statement, and disagrees that the series has demonstrated flaws. I believe it is good data for the right uses.

      One other annoying query, it isn’t actually clear to me which M08 hockey stick (with dependence on Tiljander etc) your first question refers to.

      Steve: Again, Nick, that’s not what she said in her article. She said it was suitable for use as a temperature proxy prior to modern disturbance – however, in the opposite orientation to Upside Down Mann.

    • Peter
      Posted Oct 17, 2009 at 6:31 PM | Permalink

      Re: Nick Stokes (#235),

      Nick, I repeat:

      Do you agree that the NAS said “Stripbark bristlecones should not be used” or words to that direct effect, yes or no?

      One word answer please.

    • Nick Stokes
      Posted Oct 17, 2009 at 7:02 PM | Permalink

      Re: Peter (#236), Peter, I know there is a history of discussion of this at CA, but I haven’t followed it. Could you point me to that NAS statement?

    • bender
      Posted Oct 17, 2009 at 5:07 AM | Permalink

      Re: Nick Stokes (#210),
      1. That the question “was raised” doesn’t mean it was answered. (It wasn’t.)
      2. Not all the relevant questions were raised. (The bcps.)
      Your behavior is now what I would call “willfully obtuse”.

    • Peter
      Posted Oct 16, 2009 at 8:28 AM | Permalink

      Re: Nick Stokes (#172),

      Nick, you think wrong. You offered figure S8 as demonstrating that the reconstruction was not hugely different without the inverted Tiljander series, but since S8’s “hockey stickedness” is flat stuff + bristlecones, you certainly did mention them.

      This lawyerly parsing of words is so antithetic to the scientific method and search for truth that it becomes maddening.

      Nick: The hockey stick lives! Look at S8
      NAS panel: Don’t use stripbark bristlecones.
      Nick: The hockey stick lives! Look at M08, no trees.
      Steve, many others and even Mann himself: Tiljander’s got problems.
      Nick: The hockey stick lives! Look at S8, no Tiljander.
      NAS panel: Don’t….I think even you can follow this.

  • Morgan
    Posted Oct 15, 2009 at 10:45 AM | Permalink

    Re: Steve McIntyre (#65),

    Let’s not discuss peer review in general terms. The issues are well understood by readers here.

    I know. But sometimes it’s just so tempting!

    • Morgan
      Posted Oct 15, 2009 at 11:12 AM | Permalink

      Re: Morgan (#83),

      No, I take that back. It wasn’t just temptation, I really do see the quality-signaling function of peer review as relevant to the question at hand. Steve asked “Why did the exchange have so little impact?” I say it’s because the mechanism for the exchange has not evolved in response to incentives for creating that impact; rather, the incentives (from the journal’s perspective) are to minimize the impact of after-the-fact findings of error. That makes it not just a question of “more ink costs more money”. It’s not good for science as a whole, but it is good for maintaining the prestige of the journal, and it’s certainly related to peer review being the primary signal of quality.

      So while Mann had an obligation (to his peers in the scientific community as a whole) to admit the problem, and a matching obligation not to smear the person who correctly identified the problem and brought it to light, he didn’t. And in my experience, denying that there is a problem, or that the problem matters, is completely typical of responses to comments, regardless of the journal or the branch of science. And I think it’s typical because it is (or has been) an adaptive attribute for journals.

    • bender
      Posted Oct 15, 2009 at 11:17 AM | Permalink

      Re: Morgan (#89),
      The Edior just let it slide. There is every incentive for him to do that. The only disincentive would be the shame that we can now pile upon him.

    • Morgan
      Posted Oct 15, 2009 at 11:28 AM | Permalink

      Re: bender (#90),

      There is every incentive for him to do that.

      It would be fun to enumerate them. Maybe another day, maybe in an unthreaded topic. I quite like this blog, and don’t want to get myself banned…

    • steven mosher
      Posted Oct 15, 2009 at 12:19 PM | Permalink

      Re: bender (#90), If you want to know the reason why the editor let it slide I would offer this. As I contemplated my blog experiment ( where a paper is publsihed as a blog without space restrictions on comments ) it became evident ( duh) that the principle difference between a blog and a journal article is closure. That is, a blog for the most part can theoretically be left open, left open to comment ( can you say like the talmud). How does one reach closure in such an open ended media. the conversation can go on and on an on. In fact a comment could overtake the whole article and disrupt the whole affair. How would one cite that in a bibiliography? Journals like to view what they publish as established science. Citeable. there is a certain finality to the publication process. Comments are like expanded footnotes in a journal. in a blog, as I noted, the comment could overtake the article. Further, allowing back and forth, extended back and forth would undermine the whole peer review edifice.

    • Steve Geiger
      Posted Oct 15, 2009 at 12:28 PM | Permalink

      Re: steven mosher (#101),

      Further, allowing back and forth, extended back and forth would undermine the whole peer review edifice.

      sounds like wikipedia….

    • steven mosher
      Posted Oct 15, 2009 at 2:16 PM | Permalink

      Re: Steve Geiger (#103), ya citizen science.

    • Dave Dardinger
      Posted Oct 15, 2009 at 1:04 PM | Permalink

      Re: steven mosher (#101),

      How does one reach closure in such an open ended media.

      I was thinking about that yesterday in the context of Steve’s assumed lack of peer reviewed output and I think something like the following would work.

      An interesting result appears in a post by Steve M and either he or someone else decides it should be “published.” The interested party produces a mock-up article and posts it as an open thread. After a fixed period the “author” makes a post saying comments for the first round are closed and then posts a new version containing changes, new citations, etc and it’s open as well. After 2 or three iterations, a “closed” version is posted where only specially invited people can post (perhaps after vetting) and when this is done the semi-final version is posted and outside peer reviewers send reviews and if necessary more changes are made. Finally the paper is “published” in a special place. Perhaps it could even be outside Climate Audit. This would serve as a repository for the paper and would signal “closure” I.e. complaints and further discussion would go back to the blog level.

    • Steve McIntyre
      Posted Oct 15, 2009 at 1:10 PM | Permalink

      Re: Dave Dardinger (#103),

      I’ve posted our Santer comment and its revision on arxiv.org and plan to do this in the future as much as possible.

      But again, this thread should not be about me.

      It’s about why climate scientists cannot seem to resolve such a simple issue as whether Mann used the Tiljander series upside down to the temperature interpretation of the author.

    • PhilH
      Posted Oct 15, 2009 at 1:45 PM | Permalink

      Re: Dave Dardinger (#103), While peer review, properly and fairly done by qualified and relatively unbiased reviewers is something to be desired and aimed at, given the present atmosphere in the climate science journals, the widely touted stricture that Steve must to be ignored until he “publishes” is nothing but a delaying tactic, a smokescreen which seeks to take advantage of the status quo. These scientists are reading this site and are well aware of what he is doing and the quality of what is going on here.

      The simple answer to Steve’s question in #104…why “they cannot resolve such a simple issue”… is that they don’t want to.

    • steven mosher
      Posted Oct 15, 2009 at 2:29 PM | Permalink

      Re: Dave Dardinger (#104), Ya something like that

      I think it would be an interesting experiment. I don’t know if wordpress can handle all the things
      I would like to see handled in the review process.. Steve hasnt indicated if he thinks such an experiment
      is feasible or to his liking.

      As for the specific topic here> why won’t people address this problem with Mann. For christ sakes on time over at tamino I trued to get ONE PERSON to admit that mann got coordinates wrong and nobody would dare say that he
      did. i suppose they look at how to this day they triumphantly roll out UAHs math error or an error ( degrees versus radians er something like that) that Ross made and they realize that if they cop to an error ( like the Y2K error)
      people run through the streets with somebodies head.

      Think of this way.. climate science has been highly PERSONALIZED.. there are key persona.. gore.. gavin.. mann.. hansen. That personalization of science ends with people not being able to admit errors.

      Put it this way if Mann didnt make the error if the anonymous “science process” made the error correcting it would be trivial

  • Kenneth Fritsch
    Posted Oct 15, 2009 at 4:43 PM | Permalink

    Re: Steve McIntyre (#65),

    I’m trying to focus here on a very specific case. It seems unsatisfactory from a quality control perspective that such a simple issue as whether Mann used the Tiljander series upside down should remain in dispute after being specifically raised in an exchange in the PeerReviewedLitchurchur.

    This is a very elementary point and yet it remains in dispute nearly a year after being raised.

    I am of the frame of mind to generally favor lack of understanding and knowledge as the root of these types of recognition problems over any motivational or conspiratorial influences.

    The peer review process could handle these types of problems, given a general understanding and knowledge of those people administrating and reviewing the process. Without those attributes, we can have a rather frustrating situation for those that do have the required knowledge to understand. Is not the blog a great outlet for those frustrations and pointing them out to the thinking person?

    Take the case at hand with the comment by a mathematician here:
    Re: Nick Stokes (#34),
    who is apparently a sincere and intelligent mathematician who is answered here: Re: RomanM (#50),
    by a statistician at the university level and reputation.

    Could the problem be that these issues are processed in peer review more by the likes of Nick Stokeses than RomanMs? The less innocent alternative would require that the peer review owners in this area of publishing are somehow and at some levels blinded by their advocacies on policy. Without the capability of getting into these people’s heads, I personally can more readily handle the reasoning given in first sentence in this paragraph.

Fred2

If Tiljander trees, for example, get less dense as it got warmer, doesn’t it mean that trees are bad thermometers?

• bender
  Posted Oct 15, 2009 at 9:48 AM | Permalink

  Re: Fred2 (#72),
  Read the background material. It’s sedimentary varve thickness and/or density that is at issue.

AJStrata

Steve,

I appreciate all your work exposing the sloppy process of testing conclusions and challenging their validity as it pertains to the global climate debate. I work in the space community and we could never allow this kind of peer analysis and debate in our processes – lives are at stake. Many lives. This is a rolling disaster from where I sit.

When the scientific community decided to go from debating theory to imposing world wide policy effecting the lives of literally billions of people, their cozy little debate society needed to grow up and perform quality analysis and debate as is required in any endeavor that can adversely impact human beings through shoddy work.

The journal based peer review process, with its quirks, biases and good ‘ol PhD network, is a complete failure in this matter. What bloggers have done is bring these conclusions and theories into the crucible of serious analysis and debate by a variety of minds and experience. You and others have simply shown the process is too lax, allows too many errors through, and the conclusions based on this process are basically invalid.

If the climate science community is prepared to step out of the theoretical debate society and get into policy and human safety, they need to completely overhaul their peer review process. What I have seen from reviewing this debate is that the ‘science’ here is basically unfounded. You could never design a system like this in the real world. There have been too many instances the data you being used was corrupted, tainted or error filled. The history you went through above would cause people to be fired for poor work and hiding their poor work in my field.

I don’t have the answer to fix this, but the climate debate has moved into an arena where folks like me should have a voice in challenging how solid the conclusions and theories are. Our lives and livelihood are at stake. We who do similar activities using similar methodologies are well versed in doing this process in a manner the requires solid, high quality, bullet proof analysis and results.

I know you and others would prefer to salvage the journal review process. But this is not a debate of theory, but a debate of human safety and trillions of dollars. We need to make the peer review process as broad as possible, so that the results that survive are as solid as possible – before we ask the world to invest at this level.

Cheers, AJStrata

• bender
  Posted Oct 15, 2009 at 9:58 AM | Permalink

  Re: AJStrata (#74),

  I don’t have the answer to fix this

  I do.

AJStrata

Re: Bender 77:

Why am I not surprised! 😉

I see you and I agree on this, more eyes to drive out quality results and conclusions:

This points to a serious flaw in the peer review process as a mechanism for ensuring the level of due diligence that policymakers need so that they can sell their policies as “science-based consensus”. More eyeballs. More inclusivity. More back-and-forth. When you are talking global meltdown, you want maximum inter-neural connectivity. If the alarmists are right, then we can’t afford to get this wrong.

So please enlighten me – what is the answer? How do we get this process into a more rigorous methodology we can all have high confidence in?

• bender
  Posted Oct 15, 2009 at 10:05 AM | Permalink

  Re: AJStrata (#78),
  It’s OT for this thread because it’s generic, not specifically pertaining to this case. Maybe later.

AJStrata

OK, but I will not know when or where the answer will show up. Still got my day jobs!

PaulD

snip – policy

MikeN

Bender, the sign going in doesn’t just change the coefficient coming out. Steve has said the program assumes a higher value is warmer orientation for the proxies. It won’t flip a proxy back if it is inversely correlated with temperature.

• bender
  Posted Oct 15, 2009 at 10:53 AM | Permalink

  Re: MikeN (#85),
  I was referring to a standard multivariate regression, not any Mannomathomagical procedure.

craig Loehle

In a multivariate regression (e.g., stock prices from various economic indicators) it indeed does not matter what the sign is. In a Mannomatic, this would be ok also except now it means that Tiljander site is assumed to correlate inversely to the rest of the world: a hypothesis that can be tested.

jae

Another terrific post, Steve!

Re: bender (#79),

Re: AJStrata (#78),
It’s OT for this thread because it’s generic, not specifically pertaining to this case. Maybe later.

How about taking this to Unthreaded?

AJStrata

Oh no – not the dreaded unthreaded!

😉

• kim
  Posted Oct 15, 2009 at 4:54 PM | Permalink

  Re: AJStrata (#98),

  Only the most unworthy get banned, but even the most worthy get snipped.
  ==============================================

Mesa

In doing a regression analysis like this, the coefficients would typically be constrained to all go in one direction (assuming the data was all in one direction) to force the model to be physical. If the unconstrained coefficients came out with different signs (or zeros if constrained), it would be a good bet that the assumptions of the model were wrong in some cases, or at least not supported by some of the data.

steven

For those of us mathamatically challenged this sounds like the ice increasing on lake A and the ice decreasing on lake B can both mean it is warming when they are seperate data sets and a regression analysis is conducted without worrying about the physical meaning of the change in ice.

• bernie
  Posted Oct 15, 2009 at 2:38 PM | Permalink

  Re: steven (#101), I think it is worse than that, because they are saying that 100 years it may have been vice versa.

hengav

The models expect a hockey stick. Those reconstructions that have a high correlatability to the model “stick” out like a sore thumb. Tiljander, Graybill, Yamal … a list that is growing. It might be time to submit the list (and the detailed critique of each) to several publications for simultaneous review, then get them to agree to publish.

Tom C

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 influences on the Tiljander and other proxies were discussed in the SI, which showed that none of our central conclusions relied on their use.

I’ve been assuming that the paleos out there remain silent out of fear. But maybe they actually buy the Mann statement above. Is it possible that their analytical skills are that bad? It would be really great to get a comment from luminous beauty and Jim Bouldin to clarify.

Steve Geiger

Ah, the Riemann hypothesis paper…a lovely light read ;-). I think the point was that a certain reviewer found a fatal flaw and the paper was withdrawn….very much unlike a few cases in the climate-related literature where neither the originating authors or his/her peers cared to uphold similar (fatal) criticisms.

• Dave Dardinger
  Posted Oct 15, 2009 at 4:12 PM | Permalink

  Re: Steve Geiger (#115),

  That’s right. I always get those two confused.

charlesH

Steve,

Would it be possible and useful (in your view) to have a brief bio attached to each CA posters name? Something about the education and experience each brings to the discussion? For example BS physics, MBA, semi conductor marketing (in my case).

I would really like to know which comments come from those with science/engineering educations and work experience. Also, if a poster has a particular expertise in a topic under discussion(e.g. statistics, dendro etc).

• charlesH
  Posted Oct 15, 2009 at 5:45 PM | Permalink

  Re: charlesH (#116), Let me expland just a bit.

  The journal peer review process doesn’t reliably result in full disclose and reproducibility, the foundation of the scientific process. However, while the reviewers are not public, the authors of the paper are.

  The blog process seems to be working for full disclosure and reproducibility but the credibility of the commentators cannot be easily judged. I think the blog process could be improved if commentators are allowed and encouraged to state their competency’s.

  I think a blog’s credibility will improve if observers (particularly those without scientific training) could have some idea of the competency of the commentators.

  • Kenneth Fritsch
    Posted Oct 15, 2009 at 6:20 PM | Permalink

    Re: charlesH (#131),

    The blog process seems to be working for full disclosure and reproducibility but the credibility of the commentators cannot be easily judged. I think the blog process could be improved if commentators are allowed and encouraged to state their competency’s.

    I think a blog’s credibility will improve if observers (particularly those without scientific training) could have some idea of the competency of the commentators.

    I would think that at some point the observer must judge for themselves and without undue reliance on the credentials of the observed. If one does not have much confidence in one’s capability to comprehend the subject at hand, then one should not partake in the discussions and rather let the “experts” speak for them.

    We must be ever vigilant to judge the expert as well as the blogger and use all the information available – of which credentials can be a part, but certainly not the major part of the judgment. And, of course, we are all aware of two people of equal credentials arguing nearly diametrically opposed views. What then? A coin toss?

    • charlesH
      Posted Oct 15, 2009 at 6:56 PM | Permalink

      Re: Kenneth Fritsch (#135), I agree with you for those, like myself, who have technical training. However, most of the press, politicians and general public are not capable of discerning scientific truth.

      I think it would be very beneficial if the press, for example, could see the excellent qualifications of many of the frequent posters on CA. Thus far, they rest almost totally on the “peer reviewed good housekeeping seal of approval” and do not understand the serious flaws especially as it relates to climate science.

    • bender
      Posted Oct 15, 2009 at 7:07 PM | Permalink

      Re: charlesH (#142),
      Follow the blog and judge by what people write. “Credentials” mean nothing in this brave new world where “retired mining consultants” can obtain PhD-level understanding just by reading, thinking, doing, writing, publishing.

    • Gavin (not Schimdt)
      Posted Oct 15, 2009 at 7:14 PM | Permalink

      Re: bender (#145),

      Bender,

      If “PhD-level understanding” you are referring to is like crying out “the emperor has no clothes” , then yes, “retired mining consultants” can definitely do it.

    • charlesH
      Posted Oct 15, 2009 at 7:53 PM | Permalink

      Re: bender (#145), Bender,
      Steve has a very strong technical background.

      I suspect you do to. I have a basic technical background. You and I can form our own independent judgement based on fundamentals.

      If Steve wants to move CA’s influence beyond the technical crowd then CA needs to publicize the credentials of CA posters. Otherwise “peer review” in mainstream journals with all it’s shortcomings will remain the “good house keeping seal of approval” to the press, political and general public.

Kenneth Fritsch

He added: “Those such as McIntyre who operate almost entirely outside of this system are not to be trusted.”

This comment by Mann at the Revkin blog tells it all. Revkin writes for the main stream media and I doubt that the main stream media – with reporters that are usually in a technical position to do no more than report the “official” line – will take a bloggers analysis over that of the more peer reviewed “experts”.

So be it and for the person who thinks for herself who cares.

This is the same for such fields as economics and even within the framework of peer review. Currently held conventional thinking is reported and accepted and whether it be right or wrong is for the thinking person to judge.

Once understood, this crap or jewels of wisdom are much easier to take.

Mesa

So, peer review questions aside, how should this data be handled? If there is knowledge of fat tailed outliers in the 20th century, should they be eliminated? Clearly using the data in the “right” sense and constraining the coefficients isn’t necessarily the best solution, if there is specific knowledge about the behavior of certain outlier years, etc.

jae

An aside: I’ve been observing for a long time and can’t recall one instance where Nick has agreed or even supported ANY comment that just might take some wind out of the sails of the DOOMSDAY SYNDROME. How about it, Nick, can you refute this?

Steve McIntyre

#162. Actually it’s even worse. Even after the upside down was pointed out in a PNAS comment, he denied it. The first mistake may have been unintentional, but the second time, in the Reply, was with knowledge. Why wouldn’t that feel into being a distortion of the research record?

40 Shades of Green

Steve,

Why not start your own journal. Or more specifically, why not turn this excellent website into a journal? Perhaps even do a joint venture with some other like minded folks.

40 Shades

Severian

OK, I have to make one comment I think is relevant to the subject of the thread, regarding “peerreviewedliterature.” I’ve been on peer review panels for a couple of niche publications related to tech industries I’ve worked in. I know just how political they can be, and just how thorough they usually are, and for me peer review is hardly a UL stamp of approval. I hesitate to say it means nothing at all, but that’s close to my feeling after my experiences, which I freely admit are my experiences only and may not be (but probably are) transferable to other pubs and subjects. Certainly the experiences presented here and in other climate related sites regarding the kinds of things that get thru and get published, the refusal to abide by archiving and openness standards, the shutting out of dissenting views, certainly leads me to believe that my experiences are not unusual.

I’ve seen far more thorough, and intensive, scrutiny of papers and research on this and other blogs than I have ever seen in real peer review panels. While blogs have their good and bad points, ones that are seriously run as this one is allow a large number of eyes and minds to review and comment on an issue, and have proven far more useful at rooting out problems than any publications peer review panel (or thesis committee I’ve ever seen). The comments against sites like this are remarkably similar to what the news media make regarding bloggers, that they don’t have the professionalism of “real” journalists (who have shown striking lack of honesty and competence at times), and smacks more of a jealous attempt to hold off progress and change more than a real interest in accuracy or fairness. Same here. The paradigm is changing, old methods and channels of information are dying off to be replaced by new ones, and I find it a fascinating thing to watch. I see the same thing everywhere, either in journalism, scientific publication, and going further back in the music and video industry and their attacks on new distribution channels.

Change is never embraced easily, particularly by the moribund institutions who feel threatened by it. It would seem that a more open, logical, and useful approach would be to endeavor to determine a good way to embrace and incorporate the power of blogs and such into the review process, but being human it’s doubtful change will be happily embraced, it will have to be forced. And the egg on faces that has arisen from work like our host’s will undoubtedly be one of the things that causes a change, so naturally it’s feared and attacked. Sadly, but not unexpectedly.

• bender
  Posted Oct 16, 2009 at 8:42 AM | Permalink

  Re: Severian (#184),
  Do you ever notice that freshmen ask the greatest questions? The reason is because they have watchful eyes and thoughtful minds and they don’t trust authority. Multiply that watchfulness and thoughtfulness and skepticism a million fold and that is what open blog-based review offers. I’ll take that over two tired, established guys with “credentials” any day. Put the two together and you have a vetting mechanism of unprecedented parallel processing power.

Espen

Can anyone point me to a 1-2-3 of how Mann et al. use PCA? I’m a mathematician by profession (speciality logic, but with statistics as a side subject), but currently working as a software developer. I used to do some biometrics statistics for a biology professor (as a student many years ago – I applied stepwise multiple regression which is a somewhat related technique, and I reinvented cluster analysis 🙂 which is also a loosely related technique). From what Steve and Jeff Id writes, I get the impression that “mannomatics” is misuse of statistics methods at its worst, but I may be missing something (e.g. it seems to me they commit the sin of reusing data used for model building for the actual analysis – is that really true?). (btw., and trying to be on topic wrt. peer review: I didn’t get a very convincing impression of the statistical competence of biology journal reviewers at that time)

Steve: Look at our two 2005 articles and the left-frame categories on MBH and Principal Components. Also Wegman. There are layers of issues with MBH, not just, PCA, but it was an interesting one. 🙂

• AndyL
  Posted Oct 16, 2009 at 11:39 AM | Permalink

  Re: Espen (#186),
  Espen
  Open Mind had a series of five posts on PCA a while back. Here is the most recent
  http://tamino.wordpress.com/2008/03/19/pca-part-5-non-centered-pca-and-multiple-regressions/
  In these posts, Tamino used material from Joliffe (an expert on PCA) to justify use of ‘non-centred PCA’. A while later Ian Joliffe joined the discussion, pointed out that Mann had not used the techniques he said he had used, and basically called the analysis meaningless. Joliffe begins by asking for an apology from Tamino for mis-representing his work

  Discussion is on this entry
  http://tamino.wordpress.com/2008/08/10/open-thread-5-2/
  starting here:
  Ian Jolliffe // September 8, 2008 at 9:36 am

  Needless to say, none of this is in the published literature.

  Steve: Jolliffe was one of the reviewers of our Nature 2004 submission, which he gave favorable reviews, while also saying that he was unable to determine who was right and wrong in the dispute.

MikeN

Nick Stokes, Kaufman use Tiljander upside-down, and chop off the period after 1800. You can download their data and see that they have it warmer in the 1600s and coldest around 1000 and 1200.
If a proxy were found to go down when the temperature goes up, I would expect the algorithm to invert the proxy and use that. It appears Mann’s CPS algorithm would just drop it, or perhaps it would keep it in inverted orientation, since the correlation is still good.

theduke

190.

Blogs are not a substitute for journal publication, but they are a unique and potentially excellent way of establishing the sort of high-level discussion that the public is starving for.

Exactly. And they are a corrective for attempts to impose a monopoly on information about climate. Being the focus of a column in the NYT is a stunning victory in the effort to hold world-renowned scientists accountable.

• Richard Sycamore
  Posted Oct 16, 2009 at 10:20 AM | Permalink

  Re: theduke (#192),
  Blogs are a shady place where misinformation passes for information and skeptics can spend their days endlessly reaping what they sow in the coprophagic echo chambers of the denialosphere.

  Re: Steve McIntyre (#191),
  Respectfully, I question your sense of what it means to write “less pointedly”. Some of your communications are unnecessarily provocative. Yet you complain when those provoked are “unresponsive”. Your game is to obtain “unresponsive” responses. Admit that nothing pleases you more.

  Re: MikeN (#189),
  I’ve not been censored there before, and yet I have challenged some of their assumptions. When the comment is printed you will be asked to eat your cyncism the way you eat up half-truths.

  Steve: Sometimes I write pointedly. All I meant in this instance is that my comment at RC did not include the portion of Korhola’s quote where he describes the continued use of upside-down Finnish series as a “scientific forgery”. I merely quoted the sentence where he says that they were used upside-down. I posted it simultaneously at DotEarth.

  • TerryS
    Posted Oct 17, 2009 at 5:59 AM | Permalink

    Re: Richard Sycamore (#193),

    I’ve not been censored there before, and yet I have challenged some of their assumptions. When the comment is printed you will be asked to eat your cyncism the way you eat up half-truths.

    Re: Richard Sycamore (#201),

    Something might have gone wrong in the submission process. Perhaps my typographical error on the word “interpreted” Or perhaps it will materialize later. I decided to submit the following:

    Congratulations! You are no longer a Real Climate Virgin. You will never again be able to use the words “I’ve not been censored there before”

    • Richard Sycamore
      Posted Oct 17, 2009 at 9:20 AM | Permalink

      Re: TerryS (#221),
      Whether the comment appears or not is somewhat immaterial. What matters is whether the graphics are corrected – by removing the Tiljander lake sediment series and the defective pine tree series. I will wait the week-end before passing judgement.

  • Jeff Alberts
    Posted Oct 17, 2009 at 2:19 PM | Permalink

    Re: Richard Sycamore (#193),

    Blogs are a shady place where misinformation passes for information and skeptics can spend their days endlessly reaping what they sow in the coprophagic echo chambers of the denialosphere.

    Sounds exactly like RC…

MikeN

Richard, looking forward to it. Hope you ask some followups about the lake varve proxies, I think it’s #1,#2 and #4 in the list. There are a number of posts on the subject last month.
Just throwing out Tiljander and Yamal changes things, taking away the claim of it being ‘warmer now than at any point in the last 2000 years.’ The Medieval Warm period is still cooler when you take both of those out, so they have something to cling to.

• Steve McIntyre
  Posted Oct 16, 2009 at 12:43 PM | Permalink

  Re: MikeN (#199),

  I looked at Kaufman sensitivity to several accounting issues here showing that a few changes in accounting policy (using Finnish series upside-up and using three reputable non-Briffa tree ring series (Grudd’s Tornetrask, Polar Urals and Indigirka River) instead of the three Briffa series; and a couple of other things mentioned in the post. I haven’t parsed the precise impact of each accounting decision – my guess is that the Yamal-Polar Urals sensitivity is the largest.

  Please be careful about presuming that any of these squiggles means very much.

  • bender
    Posted Oct 16, 2009 at 1:24 PM | Permalink

    Re: Steve McIntyre (#200),
    You’d get a 2-minute penalty if you were ever caught using the top graph.

• Richard Sycamore
  Posted Oct 16, 2009 at 12:43 PM | Permalink

  Re: MikeN (#199)
  Something might have gone wrong in the submission process. Perhaps my typographical error on the word “interpreted” Or perhaps it will materialize later. I decided to submit the following:

  Richard Sycamore says:
  Your comment is awaiting moderation.

  16 October 2009 at 1:38 PM
  Are you quite sure the graphs posted above are all correct? There are two that make use of a lake sediment proxy that I’d like to discuss.
  Thank you.

  • bender
    Posted Oct 16, 2009 at 1:28 PM | Permalink

    Re: Richard Sycamore (#201),
    Once that’s submitted maybe you can try getting “Deep Climate” and “Delayed Oscillator” over here for a dendro lecture from “Luminous Beauty”. There’s gonna be some learnin done out behind the woodshed tonight.

  • MikeN
    Posted Oct 16, 2009 at 1:57 PM | Permalink

    Re: Richard Sycamore (#201), I don’t think any typos would hold up a post. Someone got in with a post 40 min after yours.
    I’d say the second one might get through, since there’s no direct attack.

PhilipM

Steve

Your link to Atte Korhola’s Blog posting “Climate Science doldrums” via Google Translate is broken. Here is the direct link to the text in Finnish:-

http://www.co2-raportti.fi/index.php?page=blogi&news_id=1370#

The poor quality Google Translate version of this item follows:-

Atte Korhola:
Climate Science doldrums
27.09.2009 11:15

Thesis as soon as I put the tip of the’d put to public criticism: When you later learn about climate science, knees, they classify the early-2000s history of science an embarrassment to the figures. Kummastelemaan they come in and use our time as a warning of how the core values of science and criteria, and were given little by little forgotten in the research areas – climate change – changes in political and social playground.

The truth, universality, objectivity, reproducibility, publicity, systematic, rationality, criticality and self-korjaavuus are signs that enable science has traditionally been separated from other types of world hahmottamistavoista and pseudo-science. The revolutionary new science of research results has always taken criticism. The University’s own operational strategy of “engaging in research in the basic attitude of a university is critical.”

Therefore calls for the development of scientific skepticism. The American sociologist Robert K. Merton speaks of systematic doubt – conclusions, and estimates of the postponement, until sufficient empirical data to give them a solid foundation. If this is not a systematic science is no doubt, threaten to change research dogmeiksi, which imprisons the progress of science.

These dogmeista was caught in time, including the church, but now the researchers themselves can sometimes become dogmatic. Researchers identify with the feeling of ease at the level of its outputs, and long-puurtamisen through the findings will be easily generated as the ultimate truth. This is a lot of examples on climate science. Ilmastomallintajat For example, some may start to keep their own designs as reliable descriptions of the world, when in fact they are crude simplifications and simulations of the complex and chaotic reality as statistical virherajoineen. Climate change is an area where researchers are still a lot to do before the “final truth” achievement.

Climate research critical to eat ilmastoskeptikoiden sharp dichotomy and the human contribution to highlighting the researchers. Scientists do not wish to be identified with skepticism, when many all-natural, materials, methods and conclusions of the criticism fails to mention fear of stigma. At the same fear of new alternative models of explanation does not dare to look, with the result that the oxygen is being lost throughout the scientific field.

It is highly symptomatic that criticism or alternative models must be presented as soon as the stress that he is far from being a skeptic. This is what the UN Climate Conference in Geneva recently talked Mojib Latif rushed to point out to the telling of the near-term climate change projections could go into new and that the world’s climate may cool down in the next 10-20 years. Such prudence and living on tiptoe is not in the long run in honor of science.

Korjaavuuden themselves, along with the criticality and the objectivity of science is a key criterion for autonomy. Scientific progress can only be the scientific community through the activities of their own and do not be pressured by external parties or the hope of certain types of delivery. Scientific knowledge is sought and set out on his own personal interests and prestige notwithstanding.

Climate research is the subject of political heat because of the enormous external pressure. Community of researchers in this field has not previously been used to. In particular, the Copenhagen Climate Change Conference is below the pressure outside the research community seems to energisoineen voltage level at which the normal critical study of the basic principles of easily forgotten.

Public putkahtelee ever and anon a variety of reports and studies are maalaillaan still hurjempia horrors of climate change. In this report the genuine scientific criticism, publicity and repeatability is often forgotten. For example, the UN Environment Program has just released a report on climate (McMullen, CP and Jabbour, J. (2009). Climate Change Science Compendium 2009. United Nations Environment Program, Nairobi, Earth Print) presents impressively Wikipedia receipt for “hockey racket” to describe the last hundred years the temperature rises and inform lähteekseen Hanno (2009), which is one of the occurrence of Wikipedia checkmate meikäläisen pseudonym. Master’s thesis of such sources in the mocking.

Another example is the prestigious journal Science recently published a study in arctic regions, average temperatures are found to be higher now than at any time in the past two thousand years. Result may well be true, but the way the researchers conclude that raises questions. Proxy-material has been included selectively, they have been digested, manipulated, glazed, and the combined – for example, own and my colleagues collected data from Finland in the past has even turned upside down when the warm periods become cold and vice versa. Normally, this would be considered as a scientific falsification, which has serious consequences.

The ultimate goal of science is to seek truth and to broaden tietopiirimme. Scientific knowledge is never absolutely certain: to err is always possible. It does not mean that all information would be as shaky, or that should be adopted in a general skepticism in relation to all information. However, it is what we perceive truth may now change as you rise to new ideas, theories, methods and better research tools. This is science always happened and is happening still.

This humility and a sense of proportion is now also needed on climate science.

The author is a professor of environmental change, the University of Helsinki

jae

Whatever one thinks about peer-review, this “hockey stick saga” demonstrates clearly and conclusively that it is a totally unreliable and potentially corrupt construct. I don’t want to say it’s worthless, but it is damn close to that. It’s very similar to the American political system: good-ol-boy networks, pay to play, scratch my back and I scratch yours, $$$$, etc.* We need something better. I don’t have a solution and I don’t know what’s “better,” but peer-review means absolutely nothing, relative to the veracity of some printed material. I have just as much faith in “Car and Driver” as I do in “Nature.” My point is that peer-review just doesn’t add any value to a publication. I’ll take a heavily debated blog post, anytime. Peer review may already be an anachronism!

*I know, because I’ve been there, many years ago. Sometimes the damn reviewers don’t even read the publication!

PhilipM

Atte Korhola’s concluding remarks from “Climate Science doldrums”

Korhola commented 02/10/2009
12:37
Atte Korhola debate has evolved into a very interesting, so thanks to everyone’s participation.

Garfield asked why I write such a science forum, when things is not here, can not influence. Could just as well ask why here, 2 in secondary target; mekö that we decide? Hypothesis was that the researcher will also visit saitilla bunch, and the liveliness of the debate shows that the interest is sufficient.

The most recent debate spoke of “hockey racket.” I know it very well. Hockey Mailan ilmastotieteesssä significance in itself is a limited (albeit important), but the public image of climate change unprecedented. It has become a sort of climate change, logo, symbol ikonografinen.

I am deeply entangled in matter, as the International Geosphere-Biosphere Program’s (IGBP) PAGES program has recently launched a new sub-program called Arctic2k. The program focuses on the last 2000-year climate history of northern polaarialueilla. I am taking this program as coordinator.

But I know what I inject my head. Hockey racket, bat or rather the plant, with very many open questions related to proksisarjojen choice of data sets for statistical analysis and data archiving and accessibility. I will do anything to make data access and increase the transparency of any kind. I will also make use of experts in mathematics of statistical time series analysis.

McIntyre and ClimateAuditin The criticism is taken seriously. Mann and Partners Real Climate essentially makes it a mockery of a new mass blog. It may, however, in the long run give rise to self-destructive.

MikeN

>I’ve not been censored there before, and yet I have challenged some of their assumptions. When the comment is printed you will be asked to eat your cyncism the way you eat up half-truths.

That cynicism is burning.

kuhnkat

Nick Stokes,

I am flabbergasted that you could not find this at RealClimate. You DID lead us to believe that the Bristlecone issue was well discussed over there.

• Nick Stokes
  Posted Oct 17, 2009 at 8:22 PM | Permalink

  Re: kuhnkat (#239),

  I am flabbergasted that you could not find this at RealClimate. You DID lead us to believe that the Bristlecone issue was well discussed over there.

  Did I? Where? I’ve only mentioned BC pines to protest my lack of expertise.

  I actually haven’t followed this closely for a reason. I don’t think the paleo evidence is central to the AGW argument. AGW is about what happens if you burn hundreds of Gtons of fossil carbon over a few decades. That has never been done before, so there will only be very indirect guidance at best from whatever we can glean from the older history.

  But thanks for the pointer to the report, and Chap 4. It does seem very informative. And to respond to Peter, yes, I do see a sentence that starts
  “While “strip-bark” samples should be avoided for temperature reconstructions…”.
  I’ll need to read more to get a feel for the context.

  Steve: “Big picture” questions are important but are not relevant to the questions at hand. Editorially, I don’t permit two sentence exchanges debating the big picture to avoid every thread being exactly the same after 12 comments. So please do not respond on this thread to Nick’s musings here on the big picture.

  • bender
    Posted Oct 17, 2009 at 9:05 PM | Permalink

    Re: Nick Stokes (#241),
    As luminous beauty says: strip-bark is normal for this species. Pretty tough to avoid it. Graybill’s were particularly well-stripped. Try Ababneh’s sample instead.

Nick Stokes

snip – OT on this thread as is one responded to

Craig Loehle

To return to the topic of this post, I would suggest that the problem with peerreviewedliterature has always been there. In the past, flawed papers that were not that important were simply never corrected. Those more important would be revisited as people tried to include the new results in their own work or to make practical use of them. This would start a cycle of refinement of the results, often over decades. The idea that a result is a flawless gem simply because it appeared in Science is preposterous. Still, a simple matter such as using Tiljander upside down should have been addressed by the editor. In my experience, however, editors take a completely hands-off approach to replies.

• Raven
  Posted Oct 18, 2009 at 1:36 PM | Permalink

  Re: Craig Loehle (#248)
  The problem are not the scientists but the legions of activists and media who treat the label “peer review” as an stamp of truth. This means that nonsense papers will get repeated and their conclusions accepted as general knowledge even if the scientific community collectively ignores the work.

  For this reason, editors should be allowing much more detailed comments and ensuring that the author’s replies actually address the criticisms (a comment submitter should automatically be a reviewer on the author’s reply). This would make it much easier to identify papers that were junk in the first place.

Mrjthomas

It’s a long time since I had anything published, but I don’t recall stringent examinations of results, data, methods or code. That just didn’t seem to be the point of peer review. Which is fine if your work is of interest to a limited number of specialists and did not look like it was going to have practical consequences to the public (mine fell into this category). Looking back, I don’t think I would have welcomed a whole heap of blog feedback a la CA on my research and methods, but it sure would have forced me to sharpen my act (which incidentally is what happened when I went into industry …) It is not healthy to deify “peerreviewedliterature” in this way – it just doesn’t operate in the way that is being assumed.

Wendy

My present for the team this Xmas:

Wendy

snip – prohibited language and venting.

Steve McIntyre

Where did he say that he submitted a correction?

I notified him of the upside-down problem on Sept 3, pointing to our PNAS Comment. He replied at the time.

I wasn’t aware that Mann et al. had flipped the Korttajarvi series, and I’m afraid that I didn’t see your correction in PNAS. I’ve plotted our original series along with the re-orientation of the Korttajarvi data to get a feel for the effect on the overall result (attached).

I offered him the opportunity to post a thread here without any editorial restrictions – he was indignant at the very idea, he didn’t want to catch cooties/kudies. “Zach has kudies/koodies/coodies. He’s gross, I don’t want to sit next to him! EWWWWWW! Yucky”

I also asked him for some of the “publicly available” data that wasn’t actually publicly available. He refused. Then had a temper tantrum and said not to contact him in the future. I notified Science of the data refusal and asked them to intervene. Nothing from them yet; I should refresh the request.

MikeN

“I offered him the opportunity to post a thread here without any editorial restrictions – he was indignant at the very idea, he didn’t want to catch cooties/kudies. “Zach has kudies/koodies/coodies. He’s gross, I don’t want to sit next to him! EWWWWWW! Yucky””

“I also asked him for some of the “publicly available” data that wasn’t actually publicly available. He refused. Then had a temper tantrum and said not to contact him in the future. ”

I was here at the time, and that’s not how I would characterize his reactions. Joe Romm has temper tantrums.

• Steve McIntyre
  Posted Oct 20, 2009 at 9:47 PM | Permalink

  Re: MikeN (#262),

  SM:

  …
  As you may know, I’ve been discussing Kaufman et al 2009 at http://www.climateaudit,org and would welcome your participation. In addition, if you wish to post a thread setting out the paper in whatever terms you wish, you are more than welcome. LEt me know and I’ll provide you a password.

  One of the questions that I’ve been discussing at Climate Audit is the apparent disconnect between the sampling and archiving program originally outlined for the 30 NSF sites (the original program description stated that data would be collected from the sites in a consistent format), the limited archive to date and what was used in Kaufman et al 2009, the topic of a current thread. Any light that you can shed on this would be greatly welcome.

  In one of your meeting notes, you mention that you intended to “publicly defend” the findings of these studies. Climate Audit represents a relevant forum for such public discussion and I hope that you can participate.

  Kaufman

  I did log onto the Climateaudit website about a week ago. I have no desire to engage in vicious commentary…

  and later…
  SM:

  As I told you previously, I expect politeness from myself and my readers and have no plans to alter these policies. If there are any comments that, in your opinion, offend against these policies, I would appreciate it if you would draw them to my attention. If you are unable or unwilling to identify such comments, I would appreciate it if you would withdraw your offensive comment that I had engaged in “vicious commentary” at Climate Audit.

  In Kaufman et al 2009, you stated: “We compiled available proxy climate records that (i) were located north of 60°N latitude, (ii) extended back at least 1000 years, (iii) were resolved at an annual to decadal level, and (iv) were published with publicly available data”. Given your failure to provide URLs for proxies 6- Lake C2; 19- Lake Nautajarvi; 21- Lake Lehmilampi; 20- Lake Korttajarvi or the annual versions of proxies 7,12,13 and 16, I take it that the claim to have used “publicly available data” is, at least in part, incorrect.

  Your email attaching pdfs of various articles is unresponsive to my request, as is your suggestion that I contact the various original authors for the supposedly “publicly available data”.

  Kaufman:

  I really, really tried.
  Please do not write to me again.
  Thank you.

  There’s a bit more – I think that I posted it before. I think that my characterizations of his responses is fair.

MikeN

I saw what you posted last month.
I don’t think his responses are fairly characterized as
“he was indignant at the very idea, he didn’t want to catch cooties/kudies. “Zach has kudies/koodies/coodies. He’s gross, I don’t want to sit next to him! EWWWWWW! Yucky”

I also asked him for some of the “publicly available” data that wasn’t actually publicly available. He refused. Then had a temper tantrum and said not to contact him in the future. ”

Anyways, I don’t really want to have that argument again.

• bender
  Posted Oct 21, 2009 at 8:07 AM | Permalink

  Re: MikeN (#264),

  I don’t really want to have that argument again.

  Then why did you start it? Kaufman basically had the grown man’s equivalent of a toddler’s “temper tantrum”. It’s a figure of speech. Sure, he withdrew quietly. *After* flinging a couple of spitballs.

Lucy Skywalker

as Plato noted, to those accustomed to living in the dark, light is uncomfortable.

MikeN

And 6,7,12,13.

MikeN

Kaufman’s correction, not yet reviewed. Calibration changes are still being reviewed.

http://www.arcus.org/synthesis2k/synthesis/index.php

• bender
  Posted Oct 25, 2009 at 10:45 AM | Permalink

  Re: MikeN (#272),

  We thank those who have pointed out errors and have offered suggestions.

  That is not explicit enough.

• bender
  Posted Oct 25, 2009 at 10:47 AM | Permalink

  Re: MikeN (#272),
  They did not correct the time-frame of the GCM run. They did not even recognize it as an error.

• bender
  Posted Oct 25, 2009 at 10:49 AM | Permalink

  Re: MikeN (#272),

  The original conclusions of the paper have been strengthened as a result.

  That is not true. The original conclusions of the paper have been weakened as a result of the inclusion of the error bars.

MikeN

Bender, what is the timeframe of the GCM run you are talking about? I missed that part.

• bender
  Posted Oct 25, 2009 at 7:44 PM | Permalink

  Re: MikeN (#278),
  We discussed it on the original thread. I don;t have time to hunt it down right now, but it’s there. Steve M noticed this strange discrepancy too.

DrDweeb

Thanks for the link, it is appreciated.

As for the minor correction, my gut feeling is that “some” is probably “most”, but i think it is a given that the the very vocal Climate Science community don’t care much for lots of discpilines.

Of course a lot of geologists are working for energy companies, and that is guilt by association, so they have an uphill battle to start with.

J. King

Help me out, do I have this right? If I regress height of son on height of father I find height of father is a very powerful predictor. If I regress height of son on negative height of father, I find the same but the sign of my regression coefficient changes. No problem making future predictions if I remember what my variables mean.

Now suppose I have a father/son height dataset, but it’s corrupted so that some tall fathers have their sons height grossly understated, say they’re all 1 foot tall. Suppose too I’d really be happy if people were getting taller.

I do a regression on my corrupted dataset. I look at what it predicts about future average height – uh oh, we’re going to get shorter. But what if I accidentally flip the data before regressing, say I use 10 minus height instead of height? Now my tall fathers have sons who are 9 foot giants.

Now I take this model and project on a dataset where the corruption is even worse (late 20th century sediment data?)- people are going to be really huge. But if I get a clean dataset (distant past) I have a problem which is the height relationship is upside down – the really tall people look short because of the flipping, and vice versa. Hence the reversal of the MWP and little ice age historically, and the emergence of a hockey stick if the x-ray density is even more inverse to temperature outside the calibration period.

Further clarification would be welcome.

• AMac
  Posted Nov 28, 2009 at 4:45 PM | Permalink

  Re: J. King (#285),

  “Corrupt” may not be the right word. Tiljander (2003) warned that the climate signal in the varves was accompanied by progressively stronger human-activity signals, from circa 1720 on. Mann et al. (2008) acknowledged this problem… and then seemingly proceeded to ignore it.

  Very few people understand what the temperature signal is, that the Tiljander proxies were wrongly calibrated to, 1850-1995. I believe (but am not certain) that it was correlated to a composite, regional temperature grid that was computed by a worldwide model. Jean S, Steve Mosher, Steve McI, bender, and some others know this aspect. It is significant, because Tiljander has a figure in her paper that clearly shows that the local average temperature hasn’t displayed much if any trend since about 1880. This “at-a-distance” influence of temperature upon data series might be a bug or a feature, depending on one’s desires, and one’s sense of rigor.

  The Upside-Down episode and the ongoing efforts to justify Mann et al’s procedures often appear to be more in the spirit of Inspector Clouseau than Professor Moriarty.

  • J. King
    Posted Nov 28, 2009 at 6:39 PM | Permalink

    Re: AMac (#286),

    Thanks a lot this helps. Substitute “signal distorted by non-climate influences” for corrupt, that’s what I really meant.

    So from 1720 onward the the temperature signal is obscured by human activities, and the imapct of these activities was mostly to increase x-ray density, so a literal reading of the proxy would make it appear colder than it was. This proxy was calibrated to 1850 to 1995 temp data, but flipped upside down before doing so. If the 1850 to 1995 temp data showed an increase, which I’m guessing it did, this would incorrectly show increasing x-ray density to be correlated with rising temperatures. If x-ray density continued to increase post 1995, for non temperature reasons, the regression would predict increasing temperatures (I’m not 100% clear on if they used the sediment series to forecast post 1995 temps and produce an artificial hockey stick). Using the regression on pre 1720 data would cause an inversion – low x-ray densities (suggestive of warmth) would predict cold, and vice versa, causing the reversal of the MWP and the LIA (if I’m following correctly).

    Is this the issue in a nutshell?

J. King

Actually this suggests that the whole enterprise of reconstructing 1500 years (!) of temperature history from 150 years of real data by regressing it against various proxies is a hopeless exercise. At least Tiljander had an explanation of the divergence of the proxy from the real temperatures post 1930 (forest clearing, bridge building). No agreed upon explanations seem to exist for the divergence of the northern forest tree ring proxies post 1960. The 150 years we have available for calibration are the worst possible ones because the amount of human influence on the proxies is higher than at any previous time.

If high quality proxies without these kinds of problems existed we wouldn’t be talking about tree rings and sediment data. It seems to me the whole exercise is an elephant balanced on a rubber ball.

• dougie
  Posted Nov 29, 2009 at 9:29 AM | Permalink

  Re: J. King (#291),
  yep, you got it.

• AMac
  Posted Nov 29, 2009 at 11:27 AM | Permalink

  Re: J. King (#291),
  As a latecomer to this story, I have realized that pieces of the story of how Mann et al (2008) employed the Lake Korttajarvi varve proxies are spread out at Climate Audit and beyond. Here is a compilation of some links to primary source material.