Trip Report

Back from Washington. I think that I’ve had enough Washington for a while. It takes me a lot of time to prepare; I don’t begrudge it and it’s useful to try to put what you’re thinking about into short statements, but it still takes time and it’s tiring. It’s a nuisance that Mann couldn’t get a babysitter last week and this had to be done twice.

For this trip, the Committee paid my airfare (I had to pay my own airfare last time). However, they would not pay for travel from Canada as it is an “international” flight. So even though you clear U.S. customs in Toronto and there are planes to Washington every few hours, I had to drive to Buffalo on Wednesday, stay overnight and then fly to Washington. I thought about staying down to dinner in Washington, but headed home instead. Flight was delayed and I didn’t get in until after 11 pm and then had to drive to Toronto. I got home about 2 am.

After hearing stories about divisiveness in the House, I liked the teasing between the Dems and Repubs about the Energy and Commerce Committee baseball game and that at least Barton, Whitfield and Innslee seemed to be on pretty good personal terms. I hope that Barton and Innslee follow through and go to see Al Gore’s movie plus one another of Barton’s choice. BTW I was astonished when Rep Blackburn mentioned being in high school as long ago as 1960. She must have been a prom queen and then been lucky enough to keep her looks. Despite all the criticism that Barton’s received for holding these hearings, I thought he had a genuine interest in both sides of the quantitative significance of AGW, plus he’s smart. For someone like Cicerone, Barton is exactly the type of person that you want to be engaged in dialogue with. He’s smart enough to be persuaded. Innslee came over and thanked the panel individually. I wish that he were less strident, but I thought that his closing note of optimism was appropriate.

I introduced myself to Mann at the start of the panel; he seemed startled when I came up to shake hands. Although I’m pretty good about gritting my teeth and socializing in such circumstances, I didn’t end up chatting with Mann. I talked to Gulledge and Cicerone quite a bit. If I was a bit less tired, I probably would have made more of an effort to chat with Mann.

UPDATE: Here’s picture of the panelists.

As I mentioned last week, I thought that the energy had gone out of the committee’s interest in temperature reconstructions, once they understood that there was some kind of problem with Mann’s hockey stick, but that this didn’t mean the end of global warming. Once again, I think that their foray into climate reconstructions was extremely constructive, as neither of the NAS or Wegman reports would have been written otherwise and both, in different ways, have helped clear the air in this rancorous debate.

If I were on the committee, I wouldn’t be very interested in anything that either Mann or I said about the hockey stick, however convincing either of us seemed. All either Mann or I could do was convey an impression of whether we seemed believable or not, or whether we knew what we were about. My guess is that both of us seemed competent enough that, if I were on the committee, I wouldn’t even attempt to draw a conclusion about who was right or wrong. That’s why you have Wegman and North – they were the only opinions that a committee member was interested in.

Last week, Mann got tossed overboard. I think that the only thing that would have got Mann completely back on the island would have been if Wegman had been discredited. I don’t think that they got close o shaking Wegman, who, if anything, got stronger as it went on. IMO, the question that scored the most was when Cicerone was asked to comment on Wegman’s qualifications. Cicerone agreed that Wegman was highly qualified – what else could he say about someone who was chairman of a NAS committee. Cicerone was there on the request of the Democrats, but that one answer made it impossible for Mann to get back on the island.
Since nothing in Wegman’s testimony was dislodged, the only conclusion that the Committee could reasonably leave with was the one that they came in with: there was some kind of schmozzle with Mann’s work, but global warming was still an issue whether or not Mann had “ever been born”.

I don’t think that the Republicans had the faintest interest in taking Mann to the wall on verification statistics or bristlecones or anything like that. They seemed to intentionally stay away from confrontational questions and stayed on a pretty high road (as compared especially to Waxman going after Christy). I think that the Republicans had lost interest in Mann and, while a litigation lawyer approach would have been interesting theatre, I can see good reasons why experienced people wouldn’t bother. There was no need to.

Wegman himself could scarcely believe his ears when he heard Mann citing Rutherford et al 2005 (where Mann was a coauthor) and Wahl and Ammann as independent validation of his work. I’m used to it and you forget how inane this sort of thing is, but it was new to Wegman. I get tired of Mann and others citing von Storch and Zorita’s tests on “tame” networks as supposedly having any relevance to MBH, but no one cared. ( I wish that von Storch and Zorita would actually check the impact on MBH network rather than a pseudoproxy network.)

I thought that Waxman’s attack on Christy misfired. I couldn’t see the point. If he was trying to support Mann on not sharing source code, what was the point of arguing about Christy sharing code? In Christy’s shoes, I’d sure be mad at Wentz’ role in this, although it’s probably not worth his time pursuing it. In his first email response to Wentz’ request for code, Christy did indeed send the unresponsive email that was read into the record, but he then changed his mind and sent Wentz the relevant code – obviously. Wentz even sent him an acknowledgement thanking him for his cooperation. It’s hard to see what legitimate purpose Wentz had in sending Waxman Christy’s first email especially since just the one letter was entered into the record.

Gulledge was cross that Wegman did not deal with Wahl and Ammann, which impressed him as a study. He thinks that Wahl and Ammann have shown flaws in our work and that I’m just being stubborn in not admitting it. He criticized me for not studying the impact of the PC error on the NH reconstruction, but didn’t even cite our 2005 E&E article where we did so. Nor did the NAS panel. The carelessness is frustrating. I find it hard to take Wahl and Ammann seriously after it took an academic misconduct complaint to get them to put in the adverse verification statistics. I suppose that Wahl and Ammann needs a reply, but it’s a turgid piece of work and there’s other stuff that interests me more. But I guess I’ll need to wade through it. It would have been a lot more sensible if Ammann had accepted my offer at AGU rather than continuing with old controversises.

Speaking of both the NAS panel and Wegman report, it amazes me how little due diligence is actually done in these sorts of reports. The only due diligence that the NAS panel did was to check the tendency of the PC method to make hockey sticks. Everything else is just a literature review – by slightly less biased people than usual, but still just one more literature review. For example, as I point out, they reject bristlecones as a proxy, but do not assess the impact of this.

Wegman did more due diligence – they checked the biased PC method and also checked its impact on the North American network. It’s too bad that they didn’t also express opinions on the impact on reconstructions – I agree with Gulledge on that. However, Gulledge should then have been equally mad at the NAS panel who didn’t independently check this either. Wahl and Ammann and Rutherford et al cannot be used as evidence. It’s too bad that Wegman didn’t assess the von Storch and Zorita argument as well. I think that they would have endorsed our reply and it would have been helpful to have some validation.

At this point, it’s actually amazing how much due diligence has been done our work. Compare the energy spent on trying to prove us wrong compared the non-existence energy spent verifying MBH in the first place.

Another thought: it was crazy for Mann not to concede the biased PC argument in the first place, such as here.

Here, however, we choose to focus on some curious additional related assertions made by MM holding that (1) use of non-centered PCA (as by MBH98) is somehow not statistically valid, and (2) that “Hockey Stick” patterns arise naturally from application of non-centered PCA to purely random “red noise”. Both claims, which are of course false, were made in a comment on MBH98 by MM that was rejected by Nature , and subsequently parroted by astronomer Richard Muller in a non peer-reviewed setting–see e.g. this nice discussion by science journalist David Appell of Muller’s uncritical repetition of these false claims. These claims were discredited in the response provided by Mann and coworkers to the Nature editor and reviewers, which presumably formed the primary basis for the rejection of the MM comment. Contrary to MM’s assertions, the use of non-centered PCA is well-established in the statistical literature, and in some cases is shown to give superior results to standard, centered PCA

In effect, he tried to argue two alternatives: “there isn’t anything wrong in the PC method, but, if there is (which I deny), then it doesn’t matter.” Although his argumentation is notoriously confused – he’d sometimes use supposed evidence that the bias supposedly “didn’t matter” im his reconstruction (with his ad hoc Preisendorfer rationalizing the bristlecones) as an argument against the first leg of the alternative – to which the argument is irrelevant.

Mann obviously fought on both counts. realclimate is now implying that they conceded the bias already, but there’s no evidence of such a concession on any previous occasion; had they conceded the bias, they should have issued another corrigendum, in which they could have argued the alternative that it doesn’t matter. But Mann himself has never publicly conceded any bias in the PC method.

In litigation, people often argue alternatives, but there’s a risk in it. For example, it’s a risky strategy to argue: I didn’t kill him, but, in the alternative, if I did (which I deny), it was self-defence. If you want to argue self-defence, you have concede the act and argue the justification. Mann took a chance in arguing both legs of the alternative. Only one leg has been specifically verified by the panels and here Mann lost totally.

Another thing that I don’t get about the Team is their decision to go to battle on data and code. Any lawyer would advise them to not fight these issues. They are going to lose. And losses on irrelevant issues make it harder for them to persuade people on major issues. For people that profess to be concerned about global warming, it is irresponsible of Mann and Jones and Briffa to put their cause at risk by making a spectacle of themselves refusing data. It’s unwinnable.

After all is said in done in this, we get back to: does the Hockey Stick matter? Not any more to the Energy and Commerce Committee.

Two follow-on thoughts – I’m beginning to wonder: if the Hockey Stick doesn’t matter, then does the IPCC matter any more either? Think about it. Originally the IPCC had a function, but maybe its time is past. Right now the Fourth Assessment Report is a massive literature review and it’s important for individual climate scientists to get mentioned in it. So it ended up being a vast, unwieldy and self-indulgent document, that’s more concerned about citations than about policy.

For example, paleoclimate authors WANT to be mentioned in IPCC. But if the hockey stick doesn’t matter, why have a paleoclimate section? Why distract policy-makers with this irrelevancy? Shouldn’t the section be deleted in its entirety if the stick doesn’t matter? Even worse – IPCC 4AR has a history of climate science section – how self-indulgent is that?

I think that the most interesting outcome of the hearings was Barton’s interest in having the climate models looked at by a NAS panel from an engineering/applied statistics panel – a fresh and independent look, rather than climate modelers taking in one another’s laundry. If this happens – and I’m betting it will – and it’s done right, i.e. properly staffed with people commissioned to do independent due diligence with an actual and serious budget, and not just another literature review – I’m not so confident of this – then it might be more productive for climate policy than the IPCC Fourth Assessment Report. Scientists worried about climate should welcome Barton’s initiative and encourage him in this enterprise. What’s the over/under on the number of climate scientists who will publicly welcome such a look? Probably in single digits.


  1. Posted Jul 28, 2006 at 3:23 PM | Permalink


    You have our permission (except maybe TCO) to take a few days off. Thanks for being the squeaky wheel, though the grease is still hard to come by.

  2. Dave Dardinger
    Posted Jul 28, 2006 at 3:24 PM | Permalink

    So where does this put you now, Steve? Are you going to concentrate on the other paleoclimate reconstructions or widen your scope to some other aspect of Climate change?…. Or become a lobbyist in Washington for Exxon (Just kidding!)

  3. Dave B
    Posted Jul 28, 2006 at 3:27 PM | Permalink

    hey, i wish i knew you were in buffalo last night. i live about a mile from the airport, and would have loved to have a beer with you. thanks for your efforts.

    question: if it’s now “mann overboard”, and by extension, rutherford, wahl, and amman, then what next? may i respectfully request an audit of a key model?

  4. Jeff Weffer
    Posted Jul 28, 2006 at 3:27 PM | Permalink

    Great job Steve.

    The hockey stick is dead.

    And climate scientists are going to think twice before distorting data again (because they now know there are consequences such as being called to testify before Congressional committees.) The latter achievement is actually the greater.

  5. Dave B
    Posted Jul 28, 2006 at 3:28 PM | Permalink

    oh, by the way, when you were on c-span2 last night, the screen said, “Stephen McIntyre, Mining Industry Consultant”…just so you know.

  6. fFreddy
    Posted Jul 28, 2006 at 3:42 PM | Permalink

    After all is said in done in this, we get back to: does the Hockey Stick matter? Not any more to the Energy and Commerce Committee.

    That sounds depressingly final. Don’t you think they will be interested in a presentation on what all the other reconstructions look like once the bristlecones and the MBH PC1 have been removed ?
    I thought Blackburn was very sharp. She was the only one on the republican side who held focus, and noticed that she was being given the run-around by Mann. The rest were just going through the motions.

  7. Steve McIntyre
    Posted Jul 28, 2006 at 3:49 PM | Permalink

    BTW we’ve had a lot of people online when I checked in. We’ve been over 30 on a couple of occasions when I checked in which is a lot for us.

  8. John Hekman
    Posted Jul 28, 2006 at 3:59 PM | Permalink

    Steve, your submitted testimony was really clear and made a lot of terrific points. You just get better and better at defining the issues.

    I wonder what will happen with all of the studies that are already published that used Bristlecones. Will they continue to be cited as authority? And the hockey team has to come up with a new way to get “results” without Bristlecones, unless they think they can rationalize keeping them in.

    I hope that the funding agencies can be convinced to fund some auditing of the surface temp records. Christy has helped to shine a light on this. That is the most important black box that needs to be opened up.

  9. fFreddy
    Posted Jul 28, 2006 at 4:05 PM | Permalink

    Steve, is anything actually happening about the NAS panel’s recommendation that the bristlecones be removed ? Are all the other studies that used them being rewritten, on pain of being cancelled from whatever journals they were published in ? Or are the Team just carrying on with business as usual ?

  10. KevinUK
    Posted Jul 28, 2006 at 4:17 PM | Permalink

    Steve (and not forgetting Ross as well of course)

    One day (some time from now probably in the midst of another record winter) some eminent historian will chronical the history of the AGW myth and you will rightly (with Ross and others like John Daly, Steve milloy etc) receive due credit for revealing the poor science (that’s a euphemism for deliberate perversion of the truth for political ends) that has under-pinned the AGW myth. If there is also any justice in this world the HT will disappear into oblivion along with their fellow eco-theologian mis-guided, mis-informed supporters and political backers. Do they have the equivalent of a Congressional Medal of Honour in Canada? If so then yourself and Ross deserve one.

    Bravo Steve


  11. John A
    Posted Jul 28, 2006 at 4:20 PM | Permalink

    If the Committee members think that archving and sharing code is important, then its a win.

  12. Dane
    Posted Jul 28, 2006 at 4:37 PM | Permalink

    #10, no offense but the CMH is for bravery in combat against the enemy, not exactly the same thing here. I would recommend maybe the Presidential Medal of Freedom? Nobel Prize in Math? That sounds more like it!

  13. Cameron
    Posted Jul 28, 2006 at 4:51 PM | Permalink

    Steve – personally I thought the attack on you by Ms Illinois when she was asking for a show of hands was outrageous and condescending, and you conducted yourself well.

    But getting back to the show of hands questions since she so rudely interrupted the flow – what level of confidence would you ascribe to the statement “The warming of the earth in the past 25 years is in significant part due to anthropogenic CO2 emmissions”? Note for KevinUK if this statement is true then AGW is not a myth, and the only question remaining is what if anything to do about it?

    In general I though the republicans came off in a better light than the democrats – particularly Barton and Whitfield. This was kind of a surprise although, I suspect they were acutely aware of the political need not to come across as persecutors. Whatever the reason they made the dems look kind of wild…

  14. KevinUK
    Posted Jul 28, 2006 at 4:55 PM | Permalink


    Sorry for showing my ignorance of American honours. This debate (is GW caused by man or largely by natural causes?) often does seem like a battle and Steve and Ross have certainly shown ‘bravery’ in ‘the face of the enemy’. In the UK it would be an OBE at least and perhaps even a knighthood. Sir Steve or Sir Ross has a certain ring to it don’t you think?


  15. Reid
    Posted Jul 28, 2006 at 4:59 PM | Permalink

    Re #2: “Or become a lobbyist in Washington for Exxon”

    Exxon should hire Steve to produce a movie called “Climate Audit”. Starring Patrick Stewart as Steve McIntyre.

    The repercussions of Steve’s work may be huge.

  16. per
    Posted Jul 28, 2006 at 5:14 PM | Permalink

    Fantastic job.

    as you rightly infer, Barton must be delighted. The evidence from the NAS panel and Wegman was clear and consistent, and it was that MBH98/99 was flawed on several counts. It was singularly notable that Mann’s evidence did not address the criticisms of the statistics raised by Wegman; instead he focused on “other people get similar results” and “it doesn’t matter”. Mann left wegman’s evidence unchallenged. Mann’s testimony did not even mention the bristlecones; maybe he thinks that if he keeps quiet, no-one will remember that was a criticism.

    One hearing and the iconic MBH’98/99, the centrepiece of the IPCC TAR, has been blown out of the water. i imagine that Barton is probably looking forward to his next audit 🙂


  17. Jonthan Schafer
    Posted Jul 28, 2006 at 5:22 PM | Permalink

    GW, to what extent it exists, could still be caused by man. IMO though, for whatever it’s worth, is that CO2 is not the likely culprit. We may eventually learn it’s due to land-usage changes over time. Or we may find out that our temperature gauges are recording surface temperature changes due to land-use modifications, but that atmospheric temperatures record no overall change, this leading to no real GW at all. We simply don’t know at this point. We do know some things for sure, most importantly, that natural climate variability has occurred in the past, is occurring in the present, and will occur in the future. We need to focus on understanding that better before we ascribe the big “A” to any GW, or at least to any modifications to natural climate variability, assuming there is even a way to distinguish what is a modification and what is a natural event. Anything else is speculation. And I’m not willing to make severe modifications/restrictions on human activity to combat something we aren’t sure of, unless there are other benefits. For example, I support alternative fuel research, decreased pollution, reduction of dependency on foreign energy sources, etc. What I find most interesting is that these are some of the same goals as those on the warming side, but rather than focusing on the positives that both sides can agree on, they appear to want to be correct more than to work together where we have common ground. Maybe that’s due to an ideology vs taking a pragmatic approach.

  18. KevinUK
    Posted Jul 28, 2006 at 5:22 PM | Permalink


    I missed a lot of the broadcast as I only tuned in at the point when Steve was giving his testimony. Is “The warming of the earth in the past 25 years is in significant part due to anthropogenic CO2 emmissions” a quote from someone at the hearing?

    If not then are you wanting to conduct a poll of what visitors to this blog think is their level of confidence in this statement or are you asking for some real science based quantitative numbers? If the latter then its clear that the numbers don’t exist and are difficult if not impossible to derive as the statement is largely claimed based on the predictions of IMO (and many others) flawed GCMs. Much of the evidence for and against this statement has been and continues to be discussed on this blog. If its down to a poll of visitors to this web site then I’d guess it would be about 10%, if you asked the same question at RC I suspect it would be at least 90% (for what it would be worth). What is your assessment?


  19. welikerocks
    Posted Jul 28, 2006 at 5:27 PM | Permalink


    The search for truth is more precious than its possession
    –Albert Einstein (03/14/1879-1955)

    I never give them hell. I just tell the truth and they think it’s hell.
    –Harry S. Truman

    Truth has no special time of its own. Its hour is now — always
    –Albert Schweitzer (01/14/1875-1965)

    Thank you so very much!

  20. Follow the Money
    Posted Jul 28, 2006 at 5:36 PM | Permalink

    “”I thought that Waxman’s attack on Christy misfired. I couldn’t see the point.””

    Steve, don’t you know? Waxman just sponsored the “Safe Climate Act.” AGW is going to be an election year issue.

    Here’s a blurb:

    “”The bill achieves these targets through a flexible economy-wide cap-and-trade program for greenhouse gas emissions, along with measures to advance technology and reduce emissions through renewable energy, energy efficiency, and cleaner cars.””

    Although I’m generally for the latter without having to resort to AGW fictions, the first part is first because it’s the more important – Waxman and other democrats are fronting for a Kyoto-like Carbon Credit trading racket. (I often term it casino, I mean that in the money laundering not gaming perspective.) No greenie wrote a “economy-wide cap-and-trade program” and a “flexible” one at that! That is expensive business lobbying work. True, Democrats like Waxman may not like to be associated with the lobbies behind these schemes and perhaps the bill was presented to them laundered through an environmental front-group funded by others. Flexible? Sounds like a lot of money can be made around that squishy term.

    Did you know Waxman has his own hockey stick? A chart on his own site!

    Click to access scavsbau.pdf

    Steve, you were in a hornets nest. I don’t know what Waxman’s angle is, he’s from California. Could be some scheme to subsidize gas plants the lobbies should be paying for themselves. Maybe some old coal/oil plants need decommissioning and they would like to gain some extra carbon credit moolah for shutting down the plants to be shut anyway. Lots of games can be played here, the key is to find out who wrote Waxman’s bill – trust me, his staff did not. Expensive lobbyists did.

  21. Cameron
    Posted Jul 28, 2006 at 5:47 PM | Permalink

    Re #19 KevinUK…No I don’t want a poll, I was just asking for steves confidence level…this was related to the Dem from Illinois asking the whole panel for their agreement on similar such statements and the associated confidence level that I think she was getting from the NAS report. Anyway when Steve said simply that he didn’t know for one about that the warming was unprecedented = confident she shot down his qualifications and told him off for answering her question – really bitchy. Anyway, I was just curious on Steves level of confidence there..

    Me personally on that statement: I would say CONFIDENT (with significant being >30%), GCMS are far from perfect but they are also far from a waste of time, and there fact remains that there is no way to explain recent warming without using GHG forcing given our current understanding of climate…but like I said I don’t want a poll just Steve a chance to answer…

  22. Posted Jul 28, 2006 at 6:01 PM | Permalink

    Hey Steve, Thanks for the informative and frank report. I am going to write a letter to the committee commending them on their course of action. It had a lot of risks, least of which opportunities for grandstanding, but I think an honest intent by the committee to perform due dilligence came through. I have observed an warm current in US politics, where it is appreciated that being Democrat or Republican is better than being nothing at all.

    I can pretty much relate to everything you say and think it sums up what is wrong in environmental sciences today. Sorry you had to pay, but I would have thought if you take a US carrier you would have been reimbursed.

    This could be the death of IPCC. By the last meeting it was clear that it was just a massive literature review, and relying on this (along with peer review) was inadequate due dilligence for policy decisions. So why fund it. Barton came across very well in this regard.

    I agree that the most interesting outcome of the hearings was Barton’s interest in having a rewiew of the climate models. Like waiting for the other shoe to drop. Lets hope they get someone with some guts like Wegman and not a group of sheep.

  23. Cameron
    Posted Jul 28, 2006 at 6:07 PM | Permalink

    Re #13 and #19…I went to the recording.

    The line of questioning begins around 3:53:30..The questions they get to before she starts impugning Steve’s credentials (aroung 3:55:30) were The level of CO2 is has gone up = unequivocal (full agreement), The earth has warmed = unequivocal (full agreement), the warming is unprecedented…this is where it gets interrupted

    So I wanted to know Steves take on the statement in 13…

  24. per
    Posted Jul 28, 2006 at 6:08 PM | Permalink

    i will just note that Barton said he would ask ciccerone to review the wegman and M&M recommendations on public archiving of methods and data. I think that Mann agreed that was a good step in principle. That’s a big set of issues.

    Christy also made the point in evidence that it is important that scientific data and code are openly available. Ciccerone sat on the fence a bit, but agreed.

    I also note that wegman took a dent when he admitted that his group had been in contact with Steve to get the M&M code to work, but not with MBH’s group. That looks like unequal treatment to me.


  25. Reid
    Posted Jul 28, 2006 at 6:15 PM | Permalink

    The Barton hearings are 100% due to Steve’s work. Steve hasn’t produced any seminal work that has changed science. He has only justly proclaimed “The Emperor has no clothes”.

    From a longtime (multi-decadal) skeptic, thank you Steve.

  26. Kevin
    Posted Jul 28, 2006 at 6:30 PM | Permalink

    Welcome back, Steve! I too am encouraged by Barton’s interest in having the climate models looked at by a NAS panel. As I’ve said in an earlier post the bottom line is the GCM provide a poor fit to bad data.

    FYI, the other day I ran some univariate ARMA models on the most recent MSU (satellite) data on the internet. (This was mainly out of curiousity and I have no intention of publicizing my results.)

    Without troubling you with details, I modeled the northern and southern hemispheres separately. In the northern hemisphere results suggest a slight warming trend or level shift beginning in early 1998. There also is some evidence that the series process changed around 10/94. The southern hemisphere exhibited no trend whatsoever, which is clear enough from simply examining the data. Differencing was not necessary in either case, meaning that both series were essentially stationary. There were quite a lot of outliers in both series. Not being a climate scientist I have no idea why NH would show some recent warming while SH is flat and we obviously don’t have a sufficiently long satellite record to tell if this might be cyclical phenomenon.

    I ran 10-year forecasts for both hemispheres based on alternative models. All models forecast no warming or cooling in SH; the results for NH were ambiguous and dependent upon whether I included a linear trend variable in the model (which can be unwise).

    While I recognize there has been controversy about the satellite data, the surface temperature data are clearly unreliable. All the temperature data need to be looked at very closely by an independent panel which includes statisticians. Once data issues have been settled (to the extent that will ever be possible) the next step in my view is to conduct univariate forecasts, using modern TS modeling techiques, to ascertain whether in fact there is a trend, or trends, that we need to worry about. If there are, then we need to start examining what might be causing them. Not writing computer code based on preconceptions.

    From all I’ve read about this issue, these first basic steps in the research process have yet to be conducted properly, and I find this shocking.

    Again, hats off to you Steve. You’ve won an important battle for science and for integrity.


  27. Douglas Hoyt
    Posted Jul 28, 2006 at 6:31 PM | Permalink

    There is a new paper by Runnalls and Oke that describes problems withe the surface temperature record. It is summarized by Pielke Sr who starts his decription as follows:

    “A March 2006 paper in the Journal of Climate by K.E. Runnalls and T.R. Oke entitled “A Technique to Detect Microclimatic Inhomogeneities in Historical Records of Screen-Level Air Temperature” further demonstrates significant problems with the accurate quantification of multi-decadal land surface temperature trends.”

    The rest of the paper description is at

    Personally I think this paper is one of the most important papers on climate change this year. It would not surprise me at all that systematically applying this method to the entire surface network would reduce the measured global warming for the twentieth century to about 0.3 C versus the now reported figure of 0.57 C.

  28. Steve McIntyre
    Posted Jul 28, 2006 at 8:04 PM | Permalink

    #25. per – here’s one answer to your point.

    when I originally archived the code, it was to show the workings.
    and some of the directories in it were local. I wasn’t thinking in terms of a totally turnkey operation, although I’m not sure why I wasn’t. It was easy to do by changing a couple of file definitions – and why not?

    A user had to make some modifications. Huybers did that without any effort with our GRL work.

    I hadn’t put my code up in response to a request from a congressional Committee. The Committee had been in contact with Mann and he failed to deliver either turnkey code or even code that could be worked. He was contacted formally and failed to deliver turnkey code. By contrast, when someone contacted me, I ensured that the code was turnkey. Mann had a chance to ensure workability in the first instance and didn’t do it.

  29. Steve McIntyre
    Posted Jul 28, 2006 at 8:21 PM | Permalink

    #27. Kevin, last summer I played around with ARMA models of gridcell data sets. AR1 seems to be the usual take of climate guys when they venture into this, but ARMA(1,1) is a much better fit and typically yields ARMA (.6, -4 to -.6) series. I did calculations for all tghe gridcells in the CRU data set and then plotted up a map of the coefficients as a color code. It turned out some interesting anomalies. Some oddball coefficients showed coding errors in the CRU data set. Take a look at the category on the right – satellite: I didn’t do much work on it.

    Because of the type of ARMA series, one sure can’t have much confidence in trend significance as opposed to a noise process.

  30. Kevin
    Posted Jul 28, 2006 at 9:04 PM | Permalink

    # 30: Steve, I too have seen many examples – and not just in climate science – of arbitrarily assuming an AR(1) process. I supect this is a legacy of the old Cochrane-Orcutt method, the DW statistic and that generation of thinking about TS modeling. Worse yet, of course, is the use of linear regression on TS data. This is often used to indicate trends and, even as a graphical device, is bad practice.

    Sorry, I didn’t quite catch what you meant by “Take a look at the category on the right – satellite: I didn’t do much work on it.”

  31. Deanster
    Posted Jul 28, 2006 at 9:27 PM | Permalink

    Steve …

    I just want to thank you for taking the time to actually get involved in all this. You would think that science, with all its prestige would have taken more time to verify the results, but apparently not.

    On another note (can’t remember the post #) regarding AGW, CO2 and temperature, I’m still wondering about the answer to the question of which came first, the CO2 or the Warming. Given the inverse relationship between dissolved gases and water temperature, and a few reports regarding the increased ocean temps, it would seem that the increases in CO2 could be a result of increased ocean temps just as easily as increased ocean temps could be the result of increased CO2. Makes me wonder about the whole GW debate. Then throw in the manipulation of models to make data conform to preconceived notions, (as has been pointed out by MM), then one really begins to wonder.

    Anyway … glad you came along and questioned the statistics.

  32. John Baltutis
    Posted Jul 28, 2006 at 10:01 PM | Permalink

    Re: #31

    Go to the top of the page, scroll down the right-hand side, and clikc on the Satellite and gridcell link under Categories. That’ll take you to a page with topics discussing his work.

  33. John Baltutis
    Posted Jul 28, 2006 at 10:01 PM | Permalink

    Steve and Ross:

    Congratulations! Well done!

  34. Steve McIntyre
    Posted Jul 28, 2006 at 10:15 PM | Permalink

    #31 – Kevin, did you see the discussions in late May and early June trying to relicate Mann’s Figure 7 – his attribution figure?

  35. McCall
    Posted Jul 28, 2006 at 10:22 PM | Permalink

    Re: “The Big Q” in #13, 19, 24
    Apologies for my lack of diplomacy on what appears to be a thoughtful Q, but IM(Informed)O “the big Q” is worthless in climatology. This period is slightly over two 11 year Schwabe solar cycles, or one 22 year Hale cycle — virtually (if not absolutely) meaningless, climatically. What distinct or high impact earth event would warrant such a question, a super-volcano eruption we haven’t heard about?

    The forecast of a weak Solar Cycle 25 (peaking ~2022, and may be the weakest in centuries) and its direct and latent impact on climate are still received with great skepticism in AGW community. Similar AGW skepticism greets the Landscheidt minimum (peaking ~2030).

    No, most if not all of this latest round of bets on AGW will be settled by 2034-5 (though some will be settled much sooner). And all of the “we will reach the AGW tipping point” wagers will be paid off as losses, up to the year 2036. But for the AGW-lurkers out there, go ahead and make our retirements wealthier by stepping up the window, and well, you know…

  36. Steve McIntyre
    Posted Jul 28, 2006 at 10:26 PM | Permalink

    #13 – the main issue with CO2 impact is the sensitivity. The GCMs think that 2xCO2 will be about 3 deg C, Doug Hoyt thinks that the impact is small and I don’t KNOW why or if Doug is wrong.I’m not saying he’s right; I’m just saying I don’t know why he’s wrong, if he is. Different policy approaches would be mandated depending on what you think that the quantum is.

    Having said that, I think that nuclear programs are mandated on many accounts including climate concern. The sandbagging of nuclear plants a generation ago has resulted in a nuclear power deficit. Just imagine how different the power supply situation would be if we’d kept at nuclear plants on a steady basis.

  37. Kevin
    Posted Jul 28, 2006 at 11:03 PM | Permalink

    Steve, John. Thanks. I’ll have a look.

  38. Kevin
    Posted Jul 29, 2006 at 1:52 AM | Permalink

    Steve: Regarding the modeling you did last year on the satellite and CRU data, nice job. I am a relative newcomer to this blog and had missed it. I only would have small observations and most of them would pertain to comments and questions in that thread, now nearly a year old. (I’ve already said be wary of performing linear regression on TS data!) I will say that your work provides more evidence that much basic research remains to be done. We have serious problems with the surface temperature data. RC and others can claim this is a non-issue but this seems to be true only to the extent that it hasn’t been given the attention it warrants.

    As for Mann’s Figure 7 and his obsession with the RE statistic, I never was able to follow his line of reasoning. BTW, correct me if I’m wrong, but solar forcings are left out of all or most GCMs, right?

  39. julian braggins
    Posted Jul 29, 2006 at 2:43 AM | Permalink

    hello McCall,
    if you havn’t come across Cameron before, he is a pro AGW who crops up all over the climate debate and researches his facts well in the CO2 field,so I limit myself to conclusions, not being a scientist. ‘Unprecedented warming’ is debatable, but not really necessary to be debated if the fact that Solar activity in its entirety is greater than at any time in the last 2000 years. I agree with you about the coming cooler spell, and Cameron might care to look at

  40. julian braggins
    Posted Jul 29, 2006 at 2:50 AM | Permalink

    Sorry about the link, it must have been swollowed by the spam eater, you’ll have to search ‘solar influences data centre’ for some very interesting graphs and the official sunspot record back to the middle 1700’s

  41. Demesure
    Posted Jul 29, 2006 at 6:16 AM | Permalink

    I would like to thank Steve for his fantastic job.
    The concrete repercussions of his results will be found in the IPPC’s 4AR : hockey stick out, bathtub in, excellent to combat hot air.

  42. muirgeo
    Posted Jul 29, 2006 at 6:29 AM | Permalink

    Re: #18


    Sorry but your logic is tortured beyond compare. So some how we know that past climate is variable. I assume you are referring to proxy data which is the whole goal of this site to call into question. But the directly measured data of the current era is apparently to imprecise to say anything about current climate changes????? Wow. Do you really think you are being objective here?

  43. muirgeo
    Posted Jul 29, 2006 at 6:52 AM | Permalink

    And then there isTHIS.

    For each of these cores, the team — including research partner Ellen Mosley-Thompson, professor of geography at Ohio State — extracted chronological measurements of the ratio of two oxygen isotopes — O18 and O16 — whose ratio serves as an indicator of air temperature at the time the ice was formed. All seven cores provided clear annual records of the isotope ratios for the last 400 years and decadally averaged records dating back 2000 years.

    “We have a record going back 2,000 years and when you plot it out, you can see the Medieval Warm Period (MWP) and the Little Ice Age (LIA),” Thompson said. During the MWP, 70O to 1000 years ago, the climate warmed in some parts of the world. The MWP was followed by the LIA, a sudden onset of colder temperatures marked by advancing glaciers in Europe and North America.

    “And in that same record, you can clearly see the 20th Century and the thing that stands out — whether you look at individual cores or the composite of all seven — is how unusually warm the last 50 years have been.

    “There hasn’t been anything in the record like it — not even the MWP,” Thompson said.

    “The fact that the isotope values in the last 50 years have been so unusual means that things are dramatically changing. That’s the real story here.”

  44. Steve McIntyre
    Posted Jul 29, 2006 at 7:54 AM | Permalink

    #41. George, I posted up a comment on Thompson’s PNAS article a few weeks ago here . Perhaps you’d care to comment on that or on a subsequent related post.

  45. Posted Jul 29, 2006 at 7:58 AM | Permalink

    #28 This is interesting, particularly due to the methodology that acknowledges LTP.

    After transforming the time series using Hurst rescaling, which identifies long-term persistence in geophysical phenomena, cooling ratio records show distinct discontinuities, which, when compared against detailed station metadata records, are found to correspond to even minor changes in the station environment. Effects detected by this method are shown to escape detection by current generally accepted techniques.

  46. Jonthan Schafer
    Posted Jul 29, 2006 at 8:28 AM | Permalink


    No one disputes that the climate of the earth has changed dramatically over the historical record over millions of years (unless you are, in which case I think that says all it needs to about where you stand). Otherwise, I think we all agree that is not in question.

    The current surface temperature record IS in question. See the link referenced in #28 above. All that refers to is questioning the current amount of warming that has been claimed. IOW, whether we are on a spike upwards in temperature, and whether that change is “unprecendented”. That speaks directly to whether we can hone in on CO2 as being the MAJOR factor in any increase or not, not whether proxy data indicates natural climate variability or not. They are not the same thing. I’m sorry you can’t understand that. But when one is driven by an ideology instead of a search for knowledge, that’s to be expected.

    If anyone has twisted logic beyond compare, it is the warmers who want to ignore the natural climate variability and blame all current warming, as such exists, solely on increased CO2 in the atmosphere. That’s not science, that’s religion.

  47. Steve McIntyre
    Posted Jul 29, 2006 at 8:46 AM | Permalink

    #28, 46. T.R. Oke of this article is either the first person or one of the first people to study urban heat island effect. He proposed that the UHI varied as log(population) – which certainly doesn’t imply that it would not occur in certain types of rural settings. One sees lots of population growth in tropical villages and towns.

  48. Allan M.R. MacRae
    Posted Jul 29, 2006 at 9:08 AM | Permalink

    Re above posts #28 (Hoyt) and #37 (McIntyre).
    For background, see post 321 on

    1-D closed-form solution, assuming logarithmic relationship and no amplifiers:

    Case A
    If 20th Century warming is 0.6C, future warming (due to doubling CO2 from pre-industrial levels) is 0.76C
    k CO2a CO2b deltaT Case
    1.965 280 380 0.6 Assumes 100% delta
    1.965 380 560 0.76 due to >CO2.

    Case B
    If 20th Century warming is 0.3C, future warming (due to doubling CO2 from pre-industrial levels) is 0.38C
    k CO2a CO2b deltaT Case
    0.982 280 380 0.3 Assumes 100% deltaT
    0.982 380 560 0.38 due to >CO2.

    Note that the above calculations are unconservative in that it is assumed that all the current warming is due to CO2.

    To produce the typically reported (IPCC, etc.) GCM results of 3 degrees C or more requires huge amplifiers. Richard Lindzen says there is no evidence for the existence of such amplifiers.

    As I posted earlier based on Douglas Hoyt’s previous comments, rerunning the history-match of GCM’s without all the alarmist assumptions would be of interest:
    1. Choose several series of 20th Century temperature profiles, ranging from (say) from 0.3 to 0.6 C total warming.
    2. Eliminate the fabricated aerosol data and assume no trend in aerosols.
    3. Ascribe the 1940-1975 cooling to natural (solar) causes rather than to aerosols.

    Then rerun the GCM forward projections, without introducing the unproven amplifiers.
    I suggest that these runs will project future humanmade warming (due to a doubling of CO2 from pre-industrial levels) of less than 0.1 to 0.2 degrees C (respectively, for the 0.3 and 0.6 cases mentioned in my point #1 herein).

    Regards, Allan

  49. Paul Dennis
    Posted Jul 29, 2006 at 9:13 AM | Permalink

    #44 muirgeo, as Steve says check out the thread that deals with the Thompson et al. PNAS article. I’ve had a good look at the data and note that the authors talk about 20th century effects but omit to say the effects are also clear for the 19th and latter part of the 18th century as well.

    It doesn’t need a detailed disection of the data to see there are virtually no detectable trends.

  50. Terry
    Posted Jul 29, 2006 at 9:51 AM | Permalink

    Re: #41, #44, #45 about Thompson’s recent work on isotope measurements.

    The graphs in comment #46 of the Thompson thread (see are extremely helpful.

    The graphs in #46 tell us something very interesting about the RATE of temperature change, which is a very hot item lately as pro-AGW folks are currently claiming that the rate of temperature change is unprecedented and therefore strong evidence of AGW. The graphs allow us to compare the rate of change before the claimed AGW and after claimed AGW began.

    The Huascaran plot shows a higher absolute rate of change pre AGW than post AGW (ignoring for a minute the problem that the “AGW” temperature change begins in the 1700s). Hence, the Huascaran rate of change is not anomalous post-AGW.

    The Quelccaya plot shows a higher absolute rate of change post-AGW than pre-AGW (again, ignoring the proble that the “AGW” change begins in the 1800s). While it is larger post-AGW, it is still smaller than the pre-AGW Huascaran rate of change, so it cannot be said that the post-AGW change at Quelccaya is anomalous.

    The Sajama plot, of course, shows no trends, so theres is no anomalous post-AGW rate of change either.

    Bottom line: Thompson’s article provides evidence that the rate of temperature change in the twentieth century is not anomalous.

    I look forward to 1) Thompson enthusiastically emphasisizing this valuable contribution to the science, 2) Muirgeo and Steve Bloom citing to this paper whenever the topic of anomalous twentieth century temperature changes comes up and chiding any posters who refuse to acknowledge the science on the issue.

  51. G Haven
    Posted Jul 29, 2006 at 10:49 AM | Permalink

    Whither now?
    As a newbie I have been watching congressional hearings, and reading this blog, RealClimate, Pielke Sr.’s site, scattered papers, and various summations, reports, and compendia on the web for the past several weeks. Necessarily, I am at best a dilettante and amateur in this area after only a month of reading; albeit a moderately capable one, having done both statistics-based and physics-based computer modeling for image scene analysis, language production, protein structure prediction, and elastic fluid flow (I’ve had a nicely varied career *g*). For insight into the machinations of politicized science at its worst, my recent month’s reading has been without parallel and I’d like to thank Steve and everyone here for the fantastic job you’ve been doing.
    There have been scattered posts on this blog in recent days to the effect of “whither now?’ — the implication being that, having won this battle, perhaps it is time to move on to other debates in climatology. First, it is worth emphasis that the battle over the hockey stick is not won (see the post of 19 July 2006) — it appears that the hockey stick has beat a strategic retreat by conceding only part of the issue and is simply planning to bypass the skirmish. The line is that the criticisms of MM do not matter to the results because, “although the centering conventions make a difference to the first PC” the centering changes make “no practical difference at all” to the final reconstruction (quotes and context from the RC post of 19 July 2006). Questions about the effect of Bristlecones on the reconstruction are being shunted to confront Wahl and Ammann 2006 (see comment 8 and response to the RC post of 19 July 2006). Thus, although it might well seem like drudgery after all Steve et al. have been through, the battle over the hockey stick is not finished and the controversy ought to be rejoined.
    It would be great to keep the pressure on wrt validity of proxy choices, error bars, statistical methods, scientific reproducibility and interpretation of paleoclimate reconstruction results. Steve is obviously the independent expert here but it is only reasonable to wonder whether Steve’s energies at some point will flag. Forgive my ignorance, but how is Steve able to devote so much time to climatic pursuits? Independent wealth? An academic position? A patron? Sheer orneriness?
    Paleoclimatic reconstruction is, however, only one topic worth a serious (and continuing?) climate audit. As Steve has pointed out, it is heartening to imagine that the Barton/Whitfield investigation could actually lead to an independent review of GCMs and the conclusions being drawn from their performance. Trusting that this will occur without prompting, argument, and lots of homework is likely a comfortable self-delusion however. It is quite possible that control over the relevant Congressional committees will change hands this November and, judging from the comments of Waxman, Stupak, et al. during the recent hearings, I do not anticipate that they would be interested in pursuing such a review.
    As a modeler myself I have lots of questions about how the favored climate models with their favored feedback processes get parameterized, tested, verified, run, and reported on. A faintly distasteful odor hangs in the air over some of the science here. Unfortunately, although methodological critiques can be generated from an armchair, methodological critiques will only go so far without actually running the models at issue and doing simulations to test their most egregious physical assumptions. As with the paleoclimate kerfuffle, I think there is a very real question as to whether the academic scientific community has the necessary instincts to assess the models dispassionately. Unfortunately the task for the contrarian community in auditing the climate models on an amateur basis is at least an order of magnitude larger than the task in auditing the paleoclimate reconstructions.
    Another area worth investigating is the subject of multiple posts on Pielke’s blog: what is the confidence that can rationally be placed in the surface temperature data? Pielke has highlighted biases that likely exist in the surface temperature data — along with plausible physical mechanisms that could cause them. An authoritative paper estimating the degree to which the surface temperature data is measuring the same thing from year to year has yet to be written (or am I wrong?). The results could have significant ramifications.
    The investigation of surface temperature data seems like the lowest hanging fruit to me. A couple of people with some good stats and good judgment ought to be able to bang out a defensible estimate of uncertainty in the data within a few months by taking into account Pielke’s, Runnall’s and Oke’s concerns.

  52. Posted Jul 29, 2006 at 12:12 PM | Permalink

    From the same article muirgeo quoted:

    The researchers say a major climate shift around 5,000 years ago in the tropics had to have cooled the region since the ice cap quickly expanded and covered the plants. The fact that they are now being exposed indicates that the opposite has occurred — the region has warmed dramatically, causing the ice cap to quickly melt.

    What’s this? Quick climate change 5000 years ago. Hmmmm. Not many humans driving round pumping gobs of CO2 into the atmosphere. Does this not show how quickly the worlds climate can change without our influence? Proves the climate can and HAS changed on a dime due to non-anthropogenic force. Interesting the article didn’t mention what the CO2 levels were 5000 years ago. THIS theory suggests we prevented a new ice ago from occuring. And it’s increased methane that was the lead forcer, not CO2. So we may be chasing after the wrong gas. And we currently have the highest concentration of CO2 in 650,000 years, yet it was hotter 7 – 5000 years ago than it is today. Once again, not saying extra CO2 has no effect on current warming trend, but saying that it is not THE primary cause.

  53. TCO
    Posted Jul 29, 2006 at 12:34 PM | Permalink

    1. Good work, Steve.

    2. fFred: NAS was a lit review. Their remark about bristlecones has no depth or proof to it. Appealing to it is an opportunistic appeal to an authority that has not shown deep thinking on the subject under question. I agree that it would be an interesting EXERCISE to look at with/without bristlecone reconstructions. But I have not yet seen sufficient work to convince me that bristlecones are bad in some way different from the rest of the proxies. That may be, but the argument needs to be fleshed out. In a paper. A full-length paper. A fair, objective assessment of a paper (not slanted).

    3. Steve, you brought a lot of the PC1 versus reconstruction thing down on your own head and “the E&E paper” is not adequate to cover confusion which you generated in other venues of discussion of the issue. I have seen some tendancy towards these type of logic/ethics errors on other factors (and always in the direction of overstating your case). This is (in a small manner) “Mannian” and as what I care about is Feynmannian honesty and clearness, I will continue to call you on it or on other lapses of similar nature.

  54. Posted Jul 29, 2006 at 12:46 PM | Permalink

    Great job, Steve, and congrats. All the best, Lubos

  55. Dave Dardinger
    Posted Jul 29, 2006 at 12:46 PM | Permalink

    TCO, since you brought up PC1, bristlecones and the final reconstruction, here’s my quick summary for agreement or criticism.

    1. Bristlecones get concentrated in a PC because they all show the 20th century spike, which matches the rising trend in the instrumental record.

    2. Which PC they get concentrated in depends on the PC method used. For normally centered they’re PC4, for off-centered they’re PC1.

    3. As long as the PC with the bristlecones are included, we get a HS. If they’re not included we don’t.

    4. That’s because what’s being used for training is the rising instrumental record. Therefore if the trainer is to train what’s picked it will show the bristlecone rise.

    5. Therefore the claim that “it doesn’t matter” leaves out “if you don’t touch our bristlecones.

  56. Ken Fritsch
    Posted Jul 29, 2006 at 2:07 PM | Permalink

    I guess my feelings about Steve M’s participation in the Washington hearings could be summarized by an adaptation of the old bar statement from a viewer at the hearings stating: “What’s a nice man like you doing in a place like this?”

    I do think it appropriate to ask again as I did on another Washington hearing thread: When all is said and done what are the likely actions that the world’s governments will take and how will the market place react to the current energy trends? The currently constructed legal Kyoto GHG limits appear to be totally inadequate in significantly reducing GHG emissions vis-a-vis temperature/climate by anyone’s climate model calculations. Of course the kicker is that the accords allow for actions beyond these current limits by way of rather vague wording that man’s influence on the climate must be somehow negated by future actions/limits if necessary.

    Last time I asked I received no responses to this query. I would think that those that have made up their minds about AGW being more of less proven by circumstantial evidence and to be of such a magnitude that it is a threat to future generations would have formulated in their own minds what either should be done or at least what they would expect might actually occur given the current political climate and limitations of governments to act and/or accomplish.

    In my view, if AGW were real and a tipping point were upon us, people would react provided that they were experiencing real negative effects in their everyday lives and not simply being feed media worst case scenarios of current and future climate events. In the meanwhile I suspect that nations of the world will continue going through the motions and do a lot of talking about reducing GHG while doing little that would adversely affect their economies or voter contentment. With continuing high oil prices, the market will push forward the search for alternative energy sources and with, in my mind, nuclear leading the way as a most practical single source alternative to fossil fuels.

    I would think those convinced of AGW and its adverse effects would not be so concerned with arguing with us skeptical holdouts and spend more time and energy revealing how they are going to fix the problem that they are so very certain will be occurring.

  57. Ken Fritsch
    Posted Jul 29, 2006 at 2:48 PM | Permalink

    3. Steve, you brought a lot of the PC1 versus reconstruction thing down on your own head and “the E&E paper” is not adequate to cover confusion which you generated in other venues of discussion of the issue. I have seen some tendancy towards these type of logic/ethics errors on other factors (and always in the direction of overstating your case). This is (in a small manner) “Mannian” and as what I care about is Feynmannian honesty and clearness, I will continue to call you on it or on other lapses of similar nature.

    TCO, I have heard Steve M’s and Ross M’s explanation of MM’s criticism of the PCA approach of Mann. I have heard Wegman’s, the NAS committee’s and several poster’s at this blog and I feel I understand them. On the other hand, I have a very difficult time understanding what it is that you keep objecting to here. Perhaps a little of that “Feymannian” clearness could help your case. It’s almost like you are saying, “Steve M, I do not understand or appreciate the finer points of your criticism but unless you can deliver them to me, a layman in this area, in a completely black and white and yes or no approach, they must be flawed.” Besides I doubt that Steve needs to listen to you anymore than you do to me.

  58. muirgeo
    Posted Jul 29, 2006 at 3:32 PM | Permalink

    Re 47

    No one disputes that the climate of the earth has changed dramatically over the historical record over millions of years (unless you are, in which case I think that says all it needs to about where you stand).

    Tell me what is the best evidence for dramatic climate change over the “historical” record over millions of years.

  59. muirgeo
    Posted Jul 29, 2006 at 3:36 PM | Permalink

    Re 48

    Well I’m guessing it (UHI effect) is not happening over the melting glaciers, melting permafrost or the warming oceans.

  60. G Haven
    Posted Jul 29, 2006 at 3:42 PM | Permalink

    Re #57: Ken,
    If it is established to my satisfaction that A) anthropogenic global warming (and/or climate change) is occurring and will have significant negative social costs which outweigh the benefits and B) we can identify the causes and their effects and C) it is probable that the costs of obviating the causes will be outweighed by net benefits then I would personally support a political solution. Suppose a cause of AGW is identified as sulfur dioxide emissions, then

    1. Create a transferable property right in the emission of sulfur dioxide based on a Lockean notion of first possession: to wit, those of us who currently emit SO2 at some rate have a right to continue emitting SO2 at that rate;
    2. A market in the emission rights will develop;
    3. The government, through taxing the general population, and interested private individuals, through voluntary action, may buy the emission rights and retire them — thus reducing SO2 emissions without violating any property rights or requiring any great additional bureaucracy beyond that necessary to enforce the emissions restrictions.

    A libertarian solution along those lines is probably a pipe dream, alas — especially given the bungled experience with Kyoto-type CO2 emissions trading.

    More probably, what will actually happen will take the form of a compromise between factions like this: “You let us build some new nuclear power plants without squealing too much (to placate the business/utilities crowd) and maybe spend a bunch of money on “clean fuels research’ (to placate the academic/research crowd) and we will let you establish a new climate change regulatory bureaucracy within the EPA (to placate the rent-seekers and political power junkies).” In that way everyone could declare victory while being seen to have “done something’, the politicos could distribute more power and money, the green advocates could still scream to high heaven that “not enough is being done’ and continue milking their donor lists to fight the evil-doers, and the real fight would transfer to the crepuscular innards of the EPA where fortunes and careers would be created and destroyed on the basis of power, pull, and cocktail parties. Business as usual, in other words.

  61. muirgeo
    Posted Jul 29, 2006 at 3:47 PM | Permalink

    Re 49

    And that’s interesting because even the most conservative data for the lower troposphere MSU satellite data shows a trend of 1.33 C/century and increasing.

  62. Phil B.
    Posted Jul 29, 2006 at 4:51 PM | Permalink

    Re 62 muirgeo, your link only has 28 years of data. How are you drawing century trend conclusions??

  63. Martin Ringo
    Posted Jul 29, 2006 at 5:40 PM | Permalink

    Re #62

    The ARMA(3,3) with trend and trend square fit for the global MSU is as follows:

    Dependent variable MSU monthly global anomalies
    Sample(adjusted): 1980:03 2005:07

    Variable Coefficient Std. Err t-Stat Prob.

    C -0.123653 0.076711 -1.611930 0.1080
    @TREND 0.001047 0.000403 2.598118 0.0098
    AR(1) -0.344179 0.373784 -0.920797 0.3579
    AR(2) 0.754875 0.075041 10.059470 0.0000
    AR(3) 0.246683 0.326293 0.756017 0.4502
    MA(1) 0.920876 0.374923 2.456170 0.0146
    MA(2) 0.055647 0.255350 0.217926 0.8276
    MA(3) 0.071334 0.142180 0.501718 0.6162

    R-squared 0.716414
    Mean dependent var 0.034810
    Adjusted R-squared 0.708750
    S.D. dependent var 0.201771
    S.E. of regression 0.108891
    Akaike info criterion -1.567879
    Sum squared resid 3.509732
    Log likelihood 248.1015
    F-statistic 93.47206
    Prob(F-statistic) 0.000000
    Durbin-Watson stat 2.009112

    If you stick a trend^2 into the fit, the coefficients of both trend and trend^2 are insignificant. The trend coefficient by itself estimates a 1.25 C warming for a century (a ARMA(1,1) with trend model also gives 1.25 C per century trend). But, and this is one of those big “BUT”s, a forecast accounting for the ARMA structure predicts a insignificant cooling (approximately -0.04 C per year). A forecast based on the time series takes into account the estimated error terms for the last several periods and applies them to ARMA structure for the forecast. Hence, the forecast can actually be in the opposite direction that the trend coefficient would imply.

    I offer this ARMA-with-trend fit not as a good model or to push or deny any warming trend. It is only offered to show how problematic forecasting trends are when there is significant autocorrelation and probable moving average.

  64. Kevin
    Posted Jul 29, 2006 at 6:33 PM | Permalink

    Re 62: I believe the trend figure at the bottom of the MSU data comes come a linear regression and should be interpreted (if at all) as a very rough estimate. I’d check but the computer I’m using at the moment does not even have Excel.

  65. TCO
    Posted Jul 29, 2006 at 6:59 PM | Permalink

    Dardie: The issue is with quantitative expression of your point (3). For instance, is the impact (say using Steve’s metric of “HS Index” of the mining tendancy for the overall reconstruction overstated if we look at the impact on PC1 versus the impact on the overall reconstruction.

    Ken F.: Sorry that you don’t understand it, but Steve does, Wegman does, etc.

  66. TCO
    Posted Jul 29, 2006 at 7:18 PM | Permalink

    Ken F., take a look at the Burger (full factorial) paper. It is useful to understanding different levers and their effect on the end result.

    Both: The issue of Preisendorfer’s n come in as an interesting confounding factor. An argument can be made either way as to use of that for number of retained PCs. So fine. Steve’s paper would have been more thoughtful to have showed both possibilities and (if he really thinks this to be the case) made the argument for why it is fairer to look at the non-n case wrt Mann. The after the fact sputtering on Preisendorfer is a bit strained (my general impression, but I do promise to reread it) and misleading (in the sense of arguing a side versus being explanatory and objective and fair). Remember Feynman said that you always show all the reasons why your case might be wrong (so others understand and can assess that). For the same reasons that Mann should show MBH with and without bcp (even if he thinks they belong in there), Steve should do the same wrt Preisendorfer n or correlation/covariance* matrix.

    *Actually here he really is guilty of a logic or communication flaw as he changed two variables at the same time when discussing the effect of changing only one. Huybers nails this by showing the actual formula for the transform.

    Ken: I agree that these are fine points. But if we are in the middle of discussing the fine points, we discuss them. No reason to elide like a Michael Mann.

  67. TCO
    Posted Jul 29, 2006 at 7:22 PM | Permalink

    Ken: Of course Steve doesn’t have to listen to me, nor I to you. But Steve runs an uncensored blog that allows criticism of him and general debate of subjects. So my comments are in line with discussion by the community of the issues. I don’t think Steve minds my points so much and he’s amenable to being shown wrong as am I. He doesn’t like my loqaciousness or my misbehavior, but who does?

  68. Ken Fritsch
    Posted Jul 29, 2006 at 7:50 PM | Permalink

    Ken F.: Sorry that you don’t understand it, but Steve does, Wegman does, etc.

    Don’t be sorry, TCO, be Feymannian.

    What you are saying in effect is that the centering when done correctly pushes the Bristlecones from PC1 to PC4 and explains about 8% of the variance but since it still shows an HS that the off centering issue and its effects on the final interpretation of the reconstruction are somehow immaterial and therefore its influence should not be exaggerated. If this be the case than please explain what PC1, PC2 and PC3 mean and why a PC4 can have so much influence on the final result when it often might be eliminated from the final analysis. If the PCA shows that Bristlecones dominant the results and that without them we see what appears to be white noise why would or should not the results be based completely on the Bristlecones — as I think this where your argument is leading.

    Why would not the above analysis lead to looking at the final reconstruction result without the Bristlecones and through up a big red flag when the HS goes away?

  69. TCO
    Posted Jul 29, 2006 at 8:10 PM | Permalink

    Ken, do you understand the following:

    A. Quantification of impact of hockey stick index on PC1 does not equal what the impact is on the overall reconstruction because
    1. more then one PC is retained.
    2. other components come into the reconstruction (not PCs)
    Thus the impact on the PC1 is likely to overstate impact on the whole since “dilution” is not accounted for.
    B. Causing/allowing confusion amongst readers as to PC1 versus “the hockey stick” is mildly dishonest.

    If there are parts of this that you don’t understand, let me know exactly what parts troubles you. Refer to the words or phrases, that you don’t get.

  70. Dave Dardinger
    Posted Jul 29, 2006 at 8:48 PM | Permalink

    re #70 TCO,

    2. other components come into the reconstruction (not PCs)

    What components are you talking about? I thought the whole point of PCA was to only use a certain number of them to reduce the degrees of freedom or whatever.

    PS, you can skip cutesy appelations like “dardie”. (I’ve been called worse, mind you, but what’s the purpose? I don’t know you well enough for pet nicknames. And what could I do when you’re using an alias anyway, call you TKO?)

  71. TCO
    Posted Jul 29, 2006 at 9:01 PM | Permalink

    It was meant to be cute in a familiar manner. Will use Dave in future.

    My understanding from reading this blog is that the reconstruction itself actually contains a rough average of different components, some of which are PCs and some of which are not. I’m on dialup now so it is complex finding the place where Steve/Ross talked about this.

  72. TCO
    Posted Jul 29, 2006 at 9:50 PM | Permalink


    See this post by Steve. Lot of information in the comment replies. Read past the fun and games to the insights about the Mannomatic that are revealed.

  73. TCO
    Posted Jul 29, 2006 at 9:50 PM | Permalink


  74. Jonathan Schafer
    Posted Jul 29, 2006 at 10:56 PM | Permalink


    How about the sedimentary deposits left by the advance and retreat of glaciers. That evidence exists in many places. Or do you deny that glaciers advanced when the earth was colder and retreated when the earth warmed? This happened more than once in the past, regardless of knowing the exact temperatures, we know there were enough fluctations to cause these things to occur, even without man spewing that nasty old CO2 into the atmosphere.

  75. Allan M.R. MacRae
    Posted Jul 29, 2006 at 11:15 PM | Permalink

    Muirgeo I cannnot agree with your post #62. You may have used my post, #170 to source your LT data, but you failed to acknowledge my analysis of this data. Instead, you quoted a simple linear regression through all the data, which is quite misleading, imo.

    To summarize my previous comment, the LT data does not track surface temperature at all well – the LT temperature trend is essentially flat from ~1978 to ~1998, then there is the large 1998 El Nino spike and reversal, and then a ~0.2C step-up which has stayed ~flat for the most recent ~5 years. In comparison, the surface temperatures increased from ~1975 to 1998, and have been fairly flat since 1998.

  76. Kevin
    Posted Jul 30, 2006 at 12:29 AM | Permalink

    Re 62, 76: The USCCSP report – particularly the summary – was a very crude effort to cook the satellite data and remove any inconvenient discrepancies with the surface record. They failed in my opinion. Roger Pielke ( resigned from the committee for what seem like perfectly valid reasons to me. I think this is what you were referring to, Allan?

  77. James Lane
    Posted Jul 30, 2006 at 3:21 AM | Permalink

    TCO, the relationship between centering, the bristlecones and the MBH reconstruction is fully discussed in MM2005 (Energy & Environment). This has been explained to you before.

    Ken, do you understand the following:

    A. Quantification of impact of hockey stick index on PC1 does not equal what the impact is on the overall reconstruction because
    1. more then one PC is retained.
    2. other components come into the reconstruction (not PCs)
    Thus the impact on the PC1 is likely to overstate impact on the whole since “dilution” is not accounted for.

    Actually, not, because the other components are essentially “noise” and cancel themselves out (talking here about the 15th century step). In fact if you look at MBH PC1 and the MBH reconstruction, they have a remarkably similar shape

    Maybe it is confusing for Steve to talk on the blog about PC1 instead of the reconstruction, but only if you haven’t read the literature.

    As for Preisendorfer’s Rule N, the same is true. The literature shows the influence on the reconstruction (hell, realclimate does). The appeal by Mann to Preisendorfer in including PC4 is weakened by the fact that he didn’t use it as a selection criteria in MBH (Steve has a post or paper on this)*

    Even more important, if you are going to attribute a temperature signal to PC4, what is your interpretation of PC1, PC2 & PC3?. This issue is directional – if your temperature signal is in PC1 (as it is in non-centred MBH) I’ll concede you could interpret lower order PCs as “noise”, but how can you claim PC4 as your signal and disregard the higher order PCs? In my experience with PCA, this seems to be nonsense, and I can imagine Wegman rolling his eyes.

    * STEVE: Can you remind of this reference, I lost it in one of the long threads? I wanted to test the eigenvalues against other “standard” PC retention criteria.

  78. Jeff Weffer
    Posted Jul 30, 2006 at 4:57 AM | Permalink

    One thing the climate models can not explain or even attempt to explain in the regular cycle of glaciations.

    How can you build a model of the climate when a component that shows changes of 10 deg. C in a regular pattern is not included?

    Anything that can affect the historical climate by 5 to 10 degrees is certainly having an effect on today’s climate. When you can not explain that kind of variation, your model is incomplete. Even if the factors are more muted today than they are at times of glaciation shifts, they clearly would be having an impact on trends of today.

    For that reason alone, the climate models should not be used.

  79. Allan M.R. MacRae
    Posted Jul 30, 2006 at 7:09 AM | Permalink

    RE #77 (& 76 & 62): Thank you Kevin for your question:

    Excerpt from my post at, #170, in which I referred to the recent US NAS (Dr. North et al) report for the Whitfield Committee:

    The June 22, 2006 US NAS Report on Climate Reconstruction is located at

    On the bottom of Page 30 of this report, it states:
    “Since 1978, instruments on satellites have monitored the temperature of the deep atmospheric layer above the surface, and though regional differences occur, global average trends agree with the surface warming of +0.16 degrees C per decade within +/-0.04 degrees C per decade (CCSP and SGCR 2006).”

    I cannot agree with this statement. It might be a worthwhile exercise for some of the statisticians on this site to examine the data and comment.

    Kevin, I am not familiar with the politics of this “CCSP and SGCR 2006” report – I simply plotted the data (Lower Troposphere temperatures and surface temperatures) and there is no correlation, unless the analysis is skewed by running linear regression lines through the two datasets and then comparing the slopes of the two regression lines. I am a poor statistician, but I think this is not an acceptable process for this data.

    The Lower Troposphere (LT) temperatures are measured by satellites uniformly over most/all the planet and are quite precise. The so-called “error” in this data (due I recall to orbital decay of satellites) has been corrected and was quite minor – I believe far too much has been made of this correction, which in my opinion was inconsequential.

    The surface temperature (ST) data is much poorer quality than the satellite data. It is gathered at irregular locations all over the planet, typically by different government agencies in countries ranging from first- to third-world. Oceans and polar regions are very poorly covered. It would be interesting to compare the USA NOAA ST data (which is relatively high-quality) to the global ST data to analyze where the differences exist, and why. It is interesting to note that the USA ST data does not show the same catastrophic warming as the global ST data, particularly if you plot from 1930 to Present – the period of during which humanmade greenhouse gases increased significantly. The USA ST data actually shows very slight summer and fall cooling from 1930 to present, within a slight annual warming trend. It also shows annual cooling since 1998. I have posted all this somewhere on climateaudit.

    Notwithstanding the frailties of the ST data, both the USA and global ST data shows the cooling trend from ~1940 to ~1975, the warming to 1998, and moderating/flat temps since 1998. This is nothing like the LT data, which shows no warming trend from (data inception in) ~1978 to 1998, then the 1998 El Nino spike and reversal, and then a ~0.2 degree warming step-up after 2000.

    All this post is from memory, but I think you will find it accurate enough. It would take too much time for me to check each reference – as I said here, I have posted most/all of this info elsewhere on climateaudit and checked it at that time.

    Best, Allan

  80. TCO
    Posted Jul 30, 2006 at 9:35 AM | Permalink


    1. If discussion in EE article was appropriate, that does not excuse misleading arguments elsewhere. Steve, on this blog, once referred to the PC1 as “the reconstruction”: there was no footnote to the EE article. Steve has a pattern of confounding the two issues and allowing/causing readers to be confused as well.

    2. “Remarkably similar” is not a mathematical description. What is the ratio of hockey stick indices? (Are you SURE that the HS index of the PC1 is not higher than that of the MBH? Would you bet your ass on it?) What’s the rsq? Etc, etc.

    3. Just because the PC1 and MBH have a similar shape does not guarrantee that an effect which causes the shape of the PC1 also causes it in the MBH hockey stick.* If you think that a knob (off-centering) causes the shape of the reconstruction, you need to do a test on the end result. Measuring the result on an intermediate (especially given the complexity of the overall Mannomatic reconstruction) is not a guarantee of the overall effect.

    4. When testing the effct of that, you should be clear that you turn that knob and no others (or do a full factorial). For instance shifting from off-centered Mannian to covariance matrix is changing TWO effects at the same time. If you just have to do this (for God knows what reason), you need to make it clear in discussion that you can’t say that the test tells you the effect of your knob (off-centering) since you changed two effects at the same time.

    5. The Preisendorfer’s n discussion is arguable either way. Which is keeping things the same (staying with Preisendorfer or using same number of PCs)? In any case, Steve should have made his argument for why “keeping things the same” within HIS PAPER. He needs to be fair to the reader. The after the fact arguments are Mannian.

    *If I have 10% profit in the fourth quarter and 10% for the year overall, it does not NECSESARILY mean that the cause of my year’s result comes from an effect that I can show impacted the fourth quarter. To prove that it does, I need to prove that either the fourth quarter dominates the result (as with say a seasonal retailer) or better yet, just show the impact of the effect on the overall year itself.

  81. Ken Fritsch
    Posted Jul 30, 2006 at 10:44 AM | Permalink

    If there are parts of this that you don’t understand, let me know exactly what parts troubles you. Refer to the words or phrases, that you don’t get.

    Actually, TCO, I was attempting to assess the extent of your detailed knowledge in this area and, of course, what I could learn from it. Your replies have pretty much cleared up that issue for me, although I must admit that I do have a difficult time following your arguments.

    I sometimes feel that your comments stem from your need to bolster your basic argument that Steve M must publish more to be heard above the crowd and to put his analyses in more precise and articulate forms. I think that you set yourself up as some kind of sounding board who, while a layman, was capable of passing judgment on the MM publications (as well as Mann’s). The NAS report came out and substantiated the correctness of the MM analyses and the weaknesses of Mann’s approach to an extent beyond what I think most here would have anticipated given the makeup and task of the committee. Then Wegman and company came along, as well recognized and published statisticians in their field, and substantiated the statistical approach of MM over that of Mann and his colleagues and went further in suggesting the need for more statistical analyses and less group think in the area of climatology.

    I do not personally see any major immediate repercussions from these findings, and particularly so with regards to public policy, but in the longer term they would have to be a positive for revealing the truths about climate, past, present and future.

    In my view Wegman was correct about blogs not being the place to conduct science, but that doesn’t mean a blog cannot be a place for learning, while having a little fun at the same time and publicizing a point that is more easily made there than in other venues. Where else could I tell The Commanding Officer to back down from giving exaggerated and critical personal assessments and overly paternalistic advice.

  82. TCO
    Posted Jul 30, 2006 at 10:56 AM | Permalink

    I understand your umbrage at my attitude. My point on PC1 versu reconstruction stands.

  83. Ken Fritsch
    Posted Jul 30, 2006 at 11:35 AM | Permalink

    This is nothing like the LT data, which shows no warming trend from (data inception in) ~1978 to 1998, then the 1998 El Nino spike and reversal, and then a ~0.2 degree warming step-up after 2000.

    By my Excel accounting, from 1979 to 2005 the LT data yields y = 0.0129x – 25.7 with r2 = 0.317 and from 1979 to 1997 it yields y = 0.0026x – 5.26 with r2 = 0.012. Which shows the effect of the 1998 to 2005 LT data on the temperature slope and trend for the overall time period of 26 years — when applying simple linear regression. Obviously the analysis must make a more complicated accounting than that. What are the current thoughts from climate science on that accounting, particularly in light of the measured increase in cabon dioxide (and other changes in GHG) over that time period.

  84. Dan Hughes
    Posted Jul 30, 2006 at 12:33 PM | Permalink

    A request for assistance.

    I have read several times that the climate is chaotic. It seems to me that proxies that are functions of the state of the climate and are exposed to chaotic varations in climate should reflect the chaos. Do proxy records that should be a function of the state of the climate show chaotic responses. If so, how are deterministic signals recovered from the data. If they do not, why not?

    Maybe someone can point me to online info.


  85. Allan M.R. MacRae
    Posted Jul 30, 2006 at 2:20 PM | Permalink

    Thank you Martin for your post #64 and Ken for your post #84.

    Ken from my post, #170, there is a source for atmospheric CO2 data:

    No warming trend in LT (Lower Troposphere temperature) from 12/1978 to 04/1997, just oscillation around zero – then the huge 1997-98 El Nino spike peaking in 04/1998 which quickly reversed itself; possibly 0.2 degree C warming from 2000 to 2005 but note the complete lack of correlation with atmospheric CO2 levels, which have been rising consistently, at least since measurements began in 1958 – (see

    I have found some published climate science reports (such as CCSP and SGCR 2006) to be less than helpful, so I have been seeking out the raw data, plotting it and analyzing it – with my limited statistical skills. I have posted many of the data sources here on climateaudit, in the hope that those with better statistical skills would look at it and examine some of the AGW claims, as Steve has so ably done with MBH.

    I prefer using 1/80 for my LT temp initial date – you will understand why when you plot the data. Here is a simple analysis I did about one year ago (simply slopes of linregs – crude, but some interesting results):

    Analysis of Lower Troposphere (LT) temperatures for different time intervals.
    Using 10*SLOPE function (10* for decade) for varying dates (positive slope is net warming, negative cooling).
    Note Month 1/98 is close to the huge 1998 El Nino peak, and 04/97 is prior to that peak.


    1/80-1/98 0.0342 0.1173 -0.0488
    1/98-7/05 -0.0243 -0.0769 0.0281

    1/80-4/97 0.0104 0.0964 -0.0756
    4/97-7/05 0.0687 0.0475 0.0896

    Best regards, Allan

    P.S. I also hope someone looks more closely at the climate computer models and particularly the questionable aerosol data for ~1940-1975 – the recent postings of Douglas Hoyt here on climateaudit are of particular interest.

  86. Kevin
    Posted Jul 30, 2006 at 2:26 PM | Permalink

    Re #80: Allan, you might want to look at the technical appendix of the report. Their methodology is not very state of the art to say the least.

  87. Cameron
    Posted Jul 30, 2006 at 3:03 PM | Permalink

    Dan – weather is chaotic, climate not so much…

  88. David Smith
    Posted Jul 30, 2006 at 3:24 PM | Permalink

    With regards to a plot of the lower-tropospheric satellite data:

    What I see are two periods (1985,1993) of lower temperatures in the early part of the period, due to volcanic eruptions, then one period (1998) of elevated temperature, due to the famous 1998 El Nino. After all three, the temperature returned to a “near-normal” state until about 2001. Since 2001, the temperature appears to have made a step change to an elevated value.

    It seems to me that one could eliminate those special-cause periods ( two volcanoes and the 1998 El Nino) and see what the temperature trend then looks like. I realize that, over a long period, special causes tend to cancel each other, but since the two cool special causes were early in the period, and the El Nino was late in the period, that timing makes any regression slope upwards.

    To me, that upward slope due to special causes seems a bit misleading if that data is included in a search for the underlying trend.

    What I’m wondering is, what do you statisticians think about eliminating special-cause periods from a data body? I know it’s risky and needs to be clearly and loudly noted, but it seems like it’s a technique that has some value.

    I imagine this is Statistics 101, so pardon my fundamental ignorance.



  89. bruce
    Posted Jul 30, 2006 at 3:27 PM | Permalink

    Re: #80: The difficulties with the surface temperature record that you describe seem to me to be crucially important. If it is demonstrated that not only the “A” part of AGW is sus, but also the “GW” part, then not only the Hockey Stick falls apart, but the whole alarmist scaremongering of the AGW crowd falls apart.

    Clearly this is why Jones and co refuse to allow examination of their work, and particularly the adjustments they make to the record. “Trust us” they say. “WE are Climate Scientists!”

    Well after M&M have done their job, backed up (largely) by the NAS Panel, and now by Wegman et al, the term “Climate Scientist” and particularly “Real Climate Scientist” are becoming phrases that carry connotations that the proponents might not be happy with.

    The problems with the temperature record are discussed over at Warwick Hughes blog – see – but at this stage haven’t attracted a lot of traffic or discussion. It would be good to have a thread here where this important topic can be addressed.

  90. jae
    Posted Jul 30, 2006 at 3:52 PM | Permalink

    LOL. It looks like Muirego, Steve Bloom, and Peter H. are taking turns on this blog, providing the same old crap. The Hockey Stick is dead, although it will take some time for some die-hards to accept it. It is just laughable that until 1998, it was generally accepted that there was a strong MWP and a LIA. All of a sudden, with the publication of the now destroyed MBHxx studies, a bunch of envirofreaks concluded there was no MWP or LIA. What amazes me most is that a HANDFUL of scientists have convinced the politicians and the news media that there is a concensus on AGW. One brief perusal of the site shows that this is not the case. There are HUNDREDS of articles summarized at that site that show that the science is NOT settled.

  91. Posted Jul 30, 2006 at 4:02 PM | Permalink

    Re 90, Thanks bruce. Can I suggest that if anyone wants to understand what the Jones et al group has done with global T data, then they have to understand the 1986 Jones et al papers and the 1988 Fred Wood critique, which I have scanned along with the Wigley & Jones reply.
    All discussed at various pages under;
    These early papers are the only sources to glimpse the Jones et al methodologies.

  92. Dan Hughes
    Posted Jul 30, 2006 at 5:58 PM | Permalink

    Cameron, RC says climate is chaotic. Or that’s the way I understand it. Maybe I’m wrong. I also understand that the ensemble-of-a-bunch-of-AOLGCM-runs approach is absolutely necessary because of chaos. BTW, I think this ensemble approach is unique in all of science and engineering. Can anyone point me to another application area where this approach is used? I suspect it can only be found in climate change AOLGCM applications.

  93. gbalella
    Posted Jul 30, 2006 at 5:59 PM | Permalink

    RE 75

    Or do you deny that glaciers advanced when the earth was colder and retreated when the earth warmed?

    Comment by Jonathan Schafer “¢’‚¬? 29 July 2006 @ 10:56 pm

    BINGO! Thank you very much Jonathan. Yes, indeed glaciers are great thermometers. Now back to the modern times again. Why do you think that the current surface record is in question when glaciers are melting at accelerating paces all around the world? To quote you,

    “Or do you deny that glaciers advanced when the earth was colder and retreated when the earth warmed?”

  94. Ken Fritsch
    Posted Jul 30, 2006 at 6:57 PM | Permalink

    TCO, I think that you have a problem keeping your observations in context. Do you understand what Wegman is saying here?

    Principal component analysis is a method often used for reducing multidimensional datasets to lower dimensions for analysis. In this context, dimensions refer to the number
    of distinct variables. The time series proxy data involved are transformed into their principal components, where the first principal component is intended to explain most of the variation present in the variables. Each subsequent principal component explains less and less of the variation. In the methodology of MBH98/99, the first principal component is used in the temperature reconstruction, and also has the highest explained variance.

    This method is intended for dimension reduction. In most datasets, the first principal component should be the least smooth (because of the higher variance). However, in
    MBH98, MBH99, the proxy data are incorrectly centered, which inflates the variance of certain proxies and selectively chooses those decentered proxies as the temperature reconstruction.

    TCO, do you recall that Steve M found the censored file from Mann’s work that provided evidence that Mann looked at the temperature reconstructions without the bristlecone pines, i.e. he saw the red flag and evidently chose to ignore it? How many times have we heard this from Steve M before having it stated in the Wegman report as presented below. Steve M had also pointed to the fact that Mann had to have been aware of these warnings against the use of bristle cones as a temperature proxy.

    Although we have not addressed the Bristlecone Pines issue extensively in this report except as one element of the proxy data, there is one point worth mentioning. Graybill and Idso (1993) specifically sought to show that Bristlecone Pines were CO2 fertilized. Bondi et al. (1999) suggest [Bristlecones] “are not a reliable temperature proxy for the last 150 years as it shows an increasing trend in about 1850 that has been attributed to atmospheric CO2 fertilization.” It is not
    surprising therefore that this important proxy in MBH98/99 yields a temperature curve that is highly correlated with atmospheric CO2. We also note that IPCC 1996 stated that “the possible confounding effects of carbon dioxide fertilization need to be taken into account when calibrating tree ring data against climate
    variations.” In addition, as use of fossil fuels has risen, so does the release of oxides of nitrogen into the atmosphere, some of which are deposited as nitrates, that are fertilizer for biota. Thus tree ring growth would be correlated with the
    deposition of nitrates, which, in turn, would be correlated with carbon dioxide release. There are clearly confounding factors for using tree rings as temperature signals.

    TCO, do you understand the impact of this statement from the Wegman committee on the ability of Mann and company to do temperature reconstructions?

    There are several time series models that exist for the purpose of modeling series with dependence, including autoregressive, moving averages, autoregressive moving average models, and long memory processes. MBH98 and MBH99 focus on simple signal plus superimposed noise models for paleoclimate temperature reconstruction. Because of complex feedback mechanisms involved in climate dynamics, it is unlikely that the temperature records and the data derived from the proxies can be adequately modeled with a simple temperature signal with superimposed noise. We believe that there has not been a serious investigation to model the underlying process structures nor to model the present instrumented temperature record with sophisticated process models.

    Then the Wegman report, in effect, throws done the gauntlet to all modelers of climate to try harder and to be more cautious before making “conclusive statements without specific findings with regard to atmospheric forcings” as it suggests “a lack of scientific rigor and possibly an agenda”.

    In their works, Mann et al. describe the possible causes of global climate change in terms of atmospheric forcings, such as anthropogenic, volcanic, or solar forcings. Another
    questionable aspect of these works is that linear relationships are assumed in all forcing-climate relationships. This is a significantly simplified model for something as complex
    as the earth’s climate, which most likely has complicated nonlinear cyclical processes on a multi-centennial scale that we do not yet understand. Mann et al. also infer that since there is a partial positive correlation between global mean temperatures in the 20th century and CO2 concentration, greenhouse-gas forcing is the dominant external forcing of the climate system. Osborn and Briffa make a similar statement, where they casually note that evidence for warming also occurs at a period where CO2 concentrations are high. A common phrase among statisticians is correlation does not imply causation. The variables affecting earth’s climate and atmosphere are most likely to be numerous and
    confounding. Making conclusive statements without specific findings with regard to atmospheric forcings suggests a lack of scientific rigor and possibly an agenda..

    ..The work initiated by Mann and his colleagues is still in its infancy, and as such further study, the use of wider proxy networks and the development of more sophisticated climate models will all be necessary future steps in propagating this research. It is not expected or likely that after preliminary research, definitive conclusions can be made about the earth’s climate over the past millennium.

    Steve M’s audit and detective work also shows rather conclusively that Mann had calculated the r2 statistic for temperature reconstruction and particularly for the 15th century and on finding it to be nil or near nil chose to ignore it.

    Putting all these findings in context, one would reasonably have to find, in my mind anyway, that Steve M made a strong case against the Mann approach to and execution of temperature reconstructions from proxies. He did not do it by way of an epiphany at some point that put all the pieces together at one time in a perfect published article, but by clever auditing and tenacious digging. That you seem to want to detract from that effort with your constant nit picking makes me want to say that, you da Mann, you da Mann.

  95. Ken Fritsch
    Posted Jul 30, 2006 at 7:30 PM | Permalink

    re #89

    It seems to me that one could eliminate those special-cause periods ( two volcanoes and the 1998 El Nino) and see what the temperature trend then looks like.

    You and I, as non-statisticians, can do things like that, but most statisticians that I have been exposed to would be very, very cautious in doing it and warn that you must have a very good reason/explanation before throwing out data points. I think it is sufficient to simply look at the plots and conjecture that relationships are probably more complex than we can currently explain. Throwing out the three years in question gives essentially the same r2 as with them and reduces the trend line gradient by approximately by 20%.

  96. John Baltutis
    Posted Jul 30, 2006 at 7:31 PM | Permalink

    Re: #93

    I consider Meta-analyses conducted by epidemiologists without regard for confounders, limited data, and lack of significance in any individual study would qualify.

  97. N. Joseph Potts
    Posted Jul 30, 2006 at 7:53 PM | Permalink

    The hockey stick is an indestructible hydra, and it may soon occupy the Oval Office in the person of Al “Medusa” Gore. Being a masochist, I have both seen “An Inconvenient Truth” AND (in hopes of securing pointers to sources for its many statistics and graphs) purchased the “book”. The book is the first I’ve acquired since childhood that lacked not only an index, but also a TABLE OF CONTENTS (but the pages ARE numbered).

    The hockey stick appears TWICE – once vertically (on Page 63) with no identification of source and again on the next page horizontally, there ascribed to “IPCC.” The narrative (sound track, actually) around the first presentation APPEARS to ascribe it to one Dr. Lonnie Thompson, while the narrative around the second includes this sentence: “Those global-warming skeptics – a group diminishing almost as rapidly as the mountain glaciers – launched a fierce attack against another measurement of the 1,000-year correlation between CO2 and temperature known as ‘the hockey stick,’ a graphic image representing the research of climate scientist Michael Mann and his colleagues.”

    I fear the effect of this group’s work, along with NAS and the Wegman group, on the political strength of the global-warming hysteria may be even less than the effect (whatever it is) of man’s activity on global climate change.

  98. Steve McIntyre
    Posted Jul 30, 2006 at 8:19 PM | Permalink

    As you can no doubt tell, I’ve been offline for the weekend and will check back in tomorrow.

  99. JP
    Posted Jul 30, 2006 at 8:39 PM | Permalink

    “It seems to me that one could eliminate those special-cause periods ( two volcanoes and the 1998 El Nino) and see what the temperature trend then looks like. I realize that, over a long period, special causes tend to cancel each other, but since the two cool special causes were early in the period, and the El Nino was late in the period, that timing makes any regression slope upwards.”


    Statistically you are highlighting a rather short time length (23 years) -within that period you have pointed out 3 atmospheric anomalies -2 volcanic events which led to brief cooling, and the 1998 ENSO event. You cannot ignore or censor thses years just because they were abnormal. If you go back far enough you can dig up hundreds if not thousands of these events worldwide. As we speak, most of NAmerica and Western Europe is in a heatwave (2-4 deg C above the global mean), while all of Siberia, Russia, Mongolia, and West China are in a very sharp cold snap. All of these anaomalies will over time present a trend in global temps despite thier variabilities.

  100. TCO
    Posted Jul 30, 2006 at 9:01 PM | Permalink

    Ken (#95):

    Answers to your questions:
    1. Yes.
    2. Yes. Several
    3. I understand Wegman’s point about wanting to see more thoughtful examination of autocorrelation in the research. I don’t think he is making a point about Mann’s “ability”, but about what kind of work should be done.

    To your points: I agree with your recital of Stevian acheivements. I disagree with the idea that we should not examine each claimed fault on it’s own merits to see who is right/wrong and to learn something about the intricacies. I also disagree with your idea that I am trying to put Steve down by discussing individual issues to detail.

  101. Allan M.R. MacRae
    Posted Jul 30, 2006 at 11:44 PM | Permalink

    Thank you David for your post #89.

    Can you please provide the details (names, locations and dates) of the volcanoes that allegedly caused the cooling periods that you have labelled as “1985” and “1993”.

    Please also provide any reference that supports your claim as to the volcanic cause of these cooling periods.

    I am not doubting your comments – I am just trying to save some time searching for this information.

    Please note the starting times for these cooling periods were ~2 years prior to 1985 and 1993:
    Your “1985” cooling period started in 1Q1983 and bottomed out in 3Q1984.
    Your “1993” cooling period started in 2Q1991 and bottomed out in 3Q1992.

    Do you have any explanation for the shorter cooling intervals that “bottomed out” in 1982, 1989, 1997 and 2000?

    Regards, Allan

  102. Greg F
    Posted Jul 31, 2006 at 4:47 AM | Permalink

    Can you please provide the details (names, locations and dates) of the volcanoes that allegedly caused the cooling periods that you have labelled as “1985”‚Ⱡand “1993”‚Ɱ

    The two volcanos that are considered to have had a world wide effect on temperature were El ChichàƒÆ’à‚Ⲯ in Mexico in March 28 – April 4, 1982 and Mount Pinatubo – June 15, 1991.

  103. David Smith
    Posted Jul 31, 2006 at 5:04 AM | Permalink

    RE #100, #102

    As #103 mentions, they were Pinatubo and El Chicon. Of course, I can’t prove anything, but I believe that people of all persuasions agree that these events lowered temperatures worldwide.

    Allan, I find it interesting that the satellite temperatures go through short-interval (6 to 18 month) dips and rises which look, to me, like something other than random variation. I’ve never seen an attempt at explaining them, other than those that might align with El Nino / La Nina.

    JP, I agree completely about the short satellite data interval and the risk of reading anything into it about the rate of global warming. Over time, things even out, but over a short interval (1979-present) the inclusion and timing of special events is quite important. If the data analysis began in 1985 and ended in 1998, the regression would show a huge rise. Likewise, if the El Nino was in 1985 and the volcano was in 1998, the data “analysis” would indicate sharp cooling.

  104. David Smith
    Posted Jul 31, 2006 at 5:32 AM | Permalink

    Here’s a link to the satellite-derived temperature chart (second chart): link

  105. Dan Hughes
    Posted Jul 31, 2006 at 7:18 AM | Permalink

    John Baltutis, I understand meta-analysis to use experimental data. The AOLBGCM ensembles are constructed from calculations and are said to be the only way to apply the results of the calculations. A single calculation by a single code is of no use. At least that is my understanding. In contrast, each of the data sets used in a meta-analysis is of value standing on its own. Again, my understanding of the situation.

  106. Paul Penrose
    Posted Jul 31, 2006 at 8:18 AM | Permalink

    I don’t know why there is such a confusion about chaos and climate. Weather and climate are describing the same thing, really, just on different time-scales. Thus weather and climate are both equally chaotic, within their useful time-scales. For example: weather models tend diverge rapidly after about three days, which is why most forcasts past three days have low accuracy. However, climate models take longer to diverge (a few years) since they are working on a longer time-scale.

    One more thing: Chaos is not the same as Random. So many people confuse them.

  107. Peter Hearnden
    Posted Jul 31, 2006 at 8:35 AM | Permalink

    The experience of weather watchers in the UK would be that weather models give quite a good idea of the broad patterns further out than you say (indeed I remember a record breakingly N. Atlantic low being picked up day’s in advance what about decade ago – the bigger the feature/change the better they seem to be) and this frontier is continually being pushed at. I suspect the same applies to climate models.

    If I’d lived a hundred years ago and I’d montitored the English weather, formulated an average for July (they did do this btw) and then said “I don’t see any reason to think in 100 years the July average will not be the same” how far out would I be out and why? And would I be wrong because of changes to forcings (atmosphere physics) or chaos? Well, I think I’d be about 1C out and I think it would be because of forcings changes not chaos. What’s your view?

  108. Paul
    Posted Jul 31, 2006 at 8:44 AM | Permalink

    Here in the central upper-midwest of the US, we’re having a “heat wave.” The radio station I listen to gives out the record high and low temperatures for the past on the same day. Here’s what I’ve learned. About 70 years ago, they had a stretch of heat that was worse that what we’re having now. About 30 years ago, they had a stretch of cold. We’re now having a stretch of heat, again. To me, that would say we have a 30 year pattern of very warm summers and very cold summers.

    But, they only have records for this area for just over 100 years. It’s probably not enough to infer anything about climate patterns, is it?

  109. Dan Hughes
    Posted Jul 31, 2006 at 8:45 AM | Permalink

    I had asked for assistance in #85 above but the discussions seems to have drifted off that. Then in #93 I threw in a comment about the AOLBGCM codes, so maybe I added to the drift.

    I’m especially interested in the issue in #85. I have found lillte info on the Web.

    Thanks for all pointers to info.

  110. Peter Hearnden
    Posted Jul 31, 2006 at 9:32 AM | Permalink

    Re# 109

    But, they only have records for this area for just over 100 years. It’s probably not enough to infer anything about climate patterns, is it?

    Well, inquiring minds like to know. So, I guess those minds look for evidence of why your regional climate varies as it has over the last century? Haven’t ideas been put forward for why the thirties were warm and dry where you are? Do the make sense, or should they be dismissed? Or should we just not bother to try and find out why our world is as it is, was and might be because we like to dismiss the results?

  111. Dave Dardinger
    Posted Jul 31, 2006 at 9:34 AM | Permalink

    re:#107 (& others)

    Chaos is not the same as Random. So many people confuse them.

    I know what you mean, but let me throw out a theoretical question. How, exactly, do you know they’re not the same? We have a concept of what “random” means, but exactly how do you actually produce truly random numbers? The only two methods I know of either use pseudo-random numbers generated mathematically or produce the numbers physically. In neither case are they truly random numbers? Both rely on a lack of knowledge by the observer. Chaos is much the same, starting with deterministic equations and relying on lack of knowledge of one or more initial conditions. Quantum Mechanics Postulates that actions result from truly random instantuations of statistically based physical laws, but I have considerable doubt that that’s actually the case. I’d like to know if there’s even any theoretical possibility of producing true random numbers. If there isn’t, then just how does randomness differ from chaos?

  112. Mark T.
    Posted Jul 31, 2006 at 9:58 AM | Permalink

    I’d like to know if there’s even any theoretical possibility of producing true random numbers.

    There’s not. At least, it has been proven that there is a number large enough to account for all randomness. The number is Graham’s number (G), and has been shown in graph theory to produce any given association if enough samples are taken. I.e., you can know for certainty a certain occurence if there are G permutations. I cannot explain it well enough to give it justice, but suffice it to say, the number is so large that if all the matter in the known universe were converted to a pen and ink, there would not be enough to write it out (paraphrased from the article I read – I cannot recall where, however, but try wiki).

    Theory aside, there is enough randomness in many natural processes to produce reasonably random numbers. At least, random enough that we could never discern any real pattern in the outcomes given current (or any foreseeable future) technology.


  113. Allan M.R. MacRae
    Posted Jul 31, 2006 at 10:44 AM | Permalink

    Thank you for your posts #103 & #104 Greg and David,

    There is an interesting timing issue:
    Pinatubo shows immediate cooling in the Lower Troposphere (LT) whereas El Chichon does not – data included below. There is also stratospheric temperature data, and reportedly El Chichon caused considerable warming in the stratosphere (I have not verified this from data). All quite interesting and complex. If you are aware of any good papers that help to explain this data, I would be most interested in reading them.

    As David has mentioned, the LT data also shows global lows every few years – 1979.6, 1982.6, 1984.8, 1989.1, 1992.6, 1996.0, 1997.3, 2000.0, 2004.5. It appears that these lows would have happened with or without the volcanoes, but the volcanoes made the lows deeper and longer in duration. Presume the primary driving force is natural – seems to correlate pretty well with La Nina, based on my fifteen-minutes of in-depth research.

    Best, Allan

    Data source for LT temps:

    El ChichàƒⲮ in Mexico in March 28 – April 4 , 1982.
    1982 1 -0.090 -0.216 0.037
    1982 2 -0.088 -0.193 0.017
    1982 3 -0.234 -0.366 -0.101
    1982 4 -0.131 -0.045 -0.218
    1982 5 -0.154 -0.294 -0.014
    1982 6 -0.113 -0.258 0.031
    1982 7 -0.256 -0.249 -0.263
    1982 8 -0.189 -0.287 -0.090
    1982 9 -0.150 -0.254 -0.045
    1982 10 -0.228 -0.351 -0.105
    1982 11 -0.120 -0.419 0.179
    1982 12 -0.001 -0.050 0.047

    Mount Pinatubo – June 15, 1991.
    1991 1 0.145 0.191 0.099
    1991 2 0.177 0.214 0.140
    1991 3 0.300 0.451 0.148
    1991 4 0.138 0.240 0.036
    1991 5 0.168 0.328 0.009
    1991 6 0.330 0.309 0.352
    1991 7 0.190 0.206 0.174
    1991 8 0.220 0.212 0.228
    1991 9 0.067 0.151 -0.017
    1991 10 -0.055 0.001 -0.111
    1991 11 -0.105 0.004 -0.213
    1991 12 -0.130 -0.139 -0.122

  114. Michael Jankowski
    Posted Jul 31, 2006 at 11:40 AM | Permalink


    I found the explanation for the dust bowl to be very interesting.

    Unfortunately, I imagine that should those identical circumstances produce a dust bowl in the future, the explanation will be “global warming due to the atmospheric increase in greenhouse gases due to man’s activities.”

  115. KevinUK
    Posted Jul 31, 2006 at 11:57 AM | Permalink

    #106 and #110

    Shouldn’t that be ALOBGCM and not AOLBGCM where ALOBGCM stands for ‘A Load of BO****KS’ GCM?


  116. Paul Penrose
    Posted Jul 31, 2006 at 1:28 PM | Permalink

    Re 112:
    You are correct in that some chaotic processes do indeed produce what appear to be random numbers, especially if you sample them for short periods. However all the processes I’ve seen eventually settle into longer-term trends (as they oscillate between and around attractors) and so are not truely random in the larger scale. For exampe, if I were to plot daily high temperatures here for the last month, it would appear to be random, however if I plotted for an entire year there is definately a pattern. If I were to plot the temperature every hour for that same month, however other patterns would appear.

    So it really depends on what you mean by random. If you mean that in a series we can’t possibly (or even theoretically) predict what the next value is, then clearly weather and climate are not random. Real world data seldom, if ever, meets this strict definition of random, so then we get into talking about how random something is. It all gets pretty messy, but I will say this: only a subset of chaotic processes produce series for which we currently can’t predict future values, making them for all practical purposes random. Since this is only a subset, however, this means that Chaotic is not the same as Random.

  117. Dave Dardinger
    Posted Jul 31, 2006 at 3:07 PM | Permalink

    re:#“7 Paul,

    I don’t think anyone applies “random” to something as gross as weather. We first say that certain cycles are obvious or we learn of them by careful study. The big ones are concerned with orbital parameters, solar cyles, maybe even the earth’s movemnt through the galactic plane. Some of these cycles are pretty much predictable while others consist of a base cycle and then some adjustments to it. For instance the phase of the cycle or the wavelength may change. If these changes are predictable by some higher level equation then they may be either deterministic or they may seem random. But whether Chaos is involved depends on whether the parameters present in the equations are critically dependent on the initial conditions or not.

    And Chaos, if it is present, can be present in more than one cycle or on more than one level in a given cycle. Thus we might find that the sunspot cycle either in it’s length or intensity might be chaotic (or there might just be other cycles which explain most of the differences.) But we may also find that cloud cover, which regulates just how much solar energy gets to the surface, is sensitive to an interaction between Ice and ocean currents and that there’s a chaotic regime during parts of the cycle connecting them and that this is logically disconnected from the solar sunspot cycle. Then we have the old “butterfly effect” sorts of chaos where short range interactions produce large-scale chaotic conditions which could affect ocean currents, storm cycles or any number of other cycles.

    You know, we have these Carbon Cycle Charts and the Radiation Charts which should be familiar to most people here. Perhaps we need an Interaction Chart which shows ways in which various things interact with other things overall in climate. Something like CO2 increase might produce specific feedbacks, positive or negative, but the CO2 or it’s feedbacks might also shift cycles and thus produce chaotic effects which couldn’t be very easily predicted, especially since we haven’t studied the climate long enough to have much of a feel on what affects what.

    Of course, this ultimately leads to climate models, but I think what gets forgotten in the process of constructing them is that the reason they’re necessary in the first place is that the connections are too complicated for humans to understand directly. I think the modelers get so involved in the construction process that they forget their task is essentially impossible in the first place. They start trusting their outputs, at least on a statistical basis, and don’t point out as forcefully as they should that the actual result will certainly be much different than the model average and in unpredictable ways.

  118. Steve Geiger
    Posted Jul 31, 2006 at 6:41 PM | Permalink

    RE 52.

    I have the same question, mainly has anyone addressed this yet:

    “The line is that the criticisms of MM do not matter to the results because, “although the centering conventions make a difference to the first PC” the centering changes make “no practical difference at all” to the final reconstruction (quotes and context from the RC post of 19 July 2006).”

    I caught the same post on RC, and am looking forward to a reply by someone with (much) more expertise than myself on the topic. Can it be true that the errors identified in the early Mann papers had ‘no practical’ effect on his final reconstruction?

  119. Gerald Machnee
    Posted Jul 31, 2006 at 7:38 PM | Permalink

    Re #119 – Which results are they talking about? If it is wrong already(hockey stick), and you just move it around, it is still wrong. In that case it does not matter, but RC will not clarify that. Hopefully Steve M can clarify that in a few sentences. It is possible that nobody replied because it was another of their meaningless statements.

  120. Steve McIntyre
    Posted Jul 31, 2006 at 8:24 PM | Permalink

    #120. because of the erroneous PC method, Mann thought that the HS shape of the PC1 was the “dominant component of variance”, when in reality, it was a distinctive pattern particular to the bristlecones.

    The questions are 1) whether the PC4 should be excluded either because it is a lower-order series or 2) because it is made up of bristlecones, which should be excluded; 3) does the reconstruction mean anything anyway?

    PC simply identifies patterns in the data; it does not show that the pattern is a temperature proxy. If you replaced bristlecones in the NOAMER network with dotcom stock prices, they also make a distictive pattern, but that does not make them a temperature proxy.

    Bristlecone growth in the 20th century is not believed to be a temperature proxy, but due to non-climatic fertilization. Merely because it is a distinctive pattern in the network does not make it a temperature proxy. NAS said that bristlecones should be “avoided” in temperature reconstructions – a view also expressed in IPCC SAR and other specialist literature. Aside from PC arguments,if the bristlecones are excluded, the MBH reconstruction does not have HS shape.

    In our EE article long ago, we pointed out various permutations and combinations of PC retention that either led to high 15th century or low 15th century values, contradicting MBH claims that their reconstruction was “robust” to presence/absence of all dendroclimatic indicators.

  121. TCO
    Posted Jul 31, 2006 at 9:11 PM | Permalink

    There is plenty wrong with the Mann work, but some imprecision and mild overstatement in Steve’s criticism of it. For instance, saying that x1, x2, x3, x4 are each causes of a wrong y value, when actually it all hangs on x4. Or when they don’t each affect y, but might affect z. And a little bit of a misdirection when the glaring spotlight is put on effect of x2 on y (to bring in the x4 or to appeal to the z implication).

    I also think a more straightforward and less argumentative (and stoical) answer is in order. “IOW, if you use Preisendorfer’s n, then the centering has little effect.” And I’m still waiting for a NUMERICAL assessment of “little” from either Mike or Steve. I think for both of them, backing up from the personal pride and public policy issues and just considering the thing as one big logic puzzle to look at objectively, would be more useful to understanding and to explication. This is what was so useful about the BC framework.

  122. TCO
    Posted Jul 31, 2006 at 9:14 PM | Permalink

    Steve G: also Steve has a lot more criticisms of the Mann work then just the off-centering. So even if that criticism is overstated or found to be minimal, that does not get Mike out of the woods.

  123. bender
    Posted Jul 31, 2006 at 9:44 PM | Permalink

    This is not about the man so much as the Mannomatic “overdetermined” pattern matching algorithm, which is geared to finding patterns in the clouds. i.e. This is about the statistical methods by which paleoclimatologists reach their conclusions about what the important processes are affecting climate variability. (1) A drop from PC1 to PC4 is hugely significant. It is the sole reason MBH98 was published in Nature in the first place. (2) Wegman’s indictment of the whole pseudoscience of time-series pattern matching is equally damning. It is the Mannomatic method that allowed MBH98 to interpret PC1 as the A in AGW. Other reconstructions can hint at A, but MBH98 claimed to actually have isolated A. Big difference.

  124. TCO
    Posted Jul 31, 2006 at 10:02 PM | Permalink

    The drop from PC1 to 4 is a seperate issue from impact on the reconstruction. One question is does the off-centering “mine” and thus produce a reconstruction that his HS shaped. The other has to do with the sentence about PC1 showing the “dominant mode of variation”. We should be careful to distinguish between these, not to confound them (from either side of the debate).

  125. bender
    Posted Jul 31, 2006 at 10:53 PM | Permalink

    Quite right, the two effects are distinct. The drop of the HS eigenvector from PC1 to 4 is caused by the off-centering; the untenable interpretation of the HS as A is the result of the Mannomatic. They are separate effects. I didn’t mean to confuse by mentioning both in one posting. (Anyone actually bothering to read MM05 will instantly understand this, as the paper is clear and 4p succinct.)

    My point is that this audit does NOT focus on one person, or one paper, or one group, or even on paleoclimatology. It reflects more broadly on anyone who uses similarly untenable pattern-matching methods.

    Although the Mannomatic is the crux of my concern, and the off-centering is a separate issue, it is important to note that there would have been no pattern to match to in the first place had the first PCA not been done on the instrumental record. [And this, to me, is just as curious a procedure. If the HS PC1/4 pattern is a “global” pattern, then why does it load much more heavily on some areas of the globe than others? It should load weakly everywhere. This decomposition of the instrumental record seems unorthodox – but I don’t think anyone has really questioned this aspect. (Could be wrong, there’s alot written on the subject here that is hidden from view.)]

    Notably it was this first step (PCA on the instrumental record) that led MBH98 to start hunting for proxies that would match the HS pattern in the first place. The Mannomatic was just the last, worst piece of the puzzle.

    The search for HS patterns will continue. What Wegman has proven to us, and what MM suspected all along, is that overdetermined pattern matching algorithms such as the Mannomatic are untenable.

    The reason for writing this, TCO, is that some are arguing that CA is all about one person, paper, one team. It is important to clarify that it is about a single untenable method – which without CA would risk gaining wider usage in the literature. This was a three-step method where all three steps needed to be wrong to get the sought after outcome. And what do you know: all three steps had errors in them that accomplished just that.

    I contend that the people who claim that CA is about one person, one paper, one team, do not understand the Mannomatic/RegEM method. Thus they do not understand the significance of the outcome of all this: if you want to estimate the A in AGW, you can forget about paleoclimatic pattern-matching methods. They’ve been soundly discredited. All of them.

    Some will say I am merely repeating what Ross and others have said before only more succinctly. Not really. I am stating my findings in my own words after having reached the same conclusions they have. Independent validation is worth something.

  126. James Lane
    Posted Jul 31, 2006 at 10:58 PM | Permalink

    Actually, that’s a nicely nuanced point, TCO. The drop from PC1 to PC4 (effect of centering) is not so much aboout the effect on the reconstruction, as the credibility of the reconstruction.

  127. don
    Posted Aug 1, 2006 at 4:50 AM | Permalink

    Video of some of the opening statements is available here:

    McIntyre opening statement – Whitfield subCommittee II

    Mann opening statement – Whitfield subCommittee II

    Christy opening statement – Whitfield subCommittee II

    Wegman opening statement – Whitfield subCommittee II

    – Thanks to for free hosting.

  128. Steve McIntyre
    Posted Aug 1, 2006 at 4:58 AM | Permalink

    #128. Don, thanks very much for this. ANy possibility of doing the same thing for hearing #1?

  129. don
    Posted Aug 1, 2006 at 5:21 AM | Permalink

    Re #129

    Unfortunately I do not have the recording from C-Span for hearing #1. I will have to use the lower quality RP-feed from the archive here:

    Expect that some of the opening statements can be uploaded to putfile. If you want something specific, just let me know.

    – Thanks for asking

  130. Steve Geiger
    Posted Aug 1, 2006 at 7:40 AM | Permalink

    RE 123:

    Thanks for clarifications. I guess my next question as a newbie would be: has anyone taken the similar datasets (sans the bad ones) as Mann (and/or others) and conducted a reconstruction using the tenable statistical approaches? Is this sort of study published? Seems that a finding of no hockey stick would be a hugely important paper in the reviewed literature(?) (or does this study exist?)

  131. Dave Dardinger
    Posted Aug 1, 2006 at 8:05 AM | Permalink

    re:# 131 Well Steve and Ross showed what happens if you remove the bristlecones from MBH , but don’t like to have it called a reconstruction. The point being that if you do so the results don’t pass statistical review for significance. Of course the Hockey team then had the chutzpah to claim that this shows M&M were doing something wrong not themselves; apparently thinking that it’s legitimate to cheat on a test as long as this allows you to get a passing grade. It wouldn’t have be so bad that the team did this except that Mann had claimed that it didn’t matter if you removed all tree-ring proxies; you’d still get a significant result. (BTW, Steve M, do you actually get a significant result if all tree-rings are removed? I realize that it’d be a bit paradoxical for removing all the ring proxies to give you “better” results than just removing the tainted ones, but as you always say, hey, it’s the hockey team.)

  132. TCO
    Posted Aug 1, 2006 at 8:08 AM | Permalink

    James: Thanks. We’ve covered it in the past on the site. It’s what I was alluding to in my y versus z output variable remark.

    Steve G: Look at Burger and Cubasch, GRL, 2005:

    Click to access 2005.burger.pdf

  133. Mark T.
    Posted Aug 1, 2006 at 8:41 AM | Permalink

    re: 131

    Thanks for clarifications. I guess my next question as a newbie would be: has anyone taken the similar datasets (sans the bad ones) as Mann (and/or others) and conducted a reconstruction using the tenable statistical approaches?

    Actually, Mann did this himself and it is known as the CENSORED data set in his archive.

    Is this sort of study published?

    What do you think? 🙂 It has been discussed heavily here, however.


  134. bender
    Posted Aug 1, 2006 at 3:45 PM | Permalink

    Re #69 On the need to interpret PC1, PC2, PC3 if PC4 is deemed interpretible, and
    Re #124 “The drop from PC1 to PC4 is hugely significant. It is the sole reason MBH98 was published in Nature in the first place.”

    For an honest example where PC1 was not interpretible in terms of a dendroclimatological mechanism, see Buckley et al.’s (2004) Can. J. For. Res. 34: 2541 paper on cedars in Canada. Notice this good paper did not make it to Nature.

  135. Don
    Posted Aug 7, 2006 at 11:07 AM | Permalink

    C-SPAN2 is running this show again soon, at 02:22 PM EDT 1:09 (est.)

    “The Committee on Energy and Commerce Subcommittee on Oversight and Investigations held a hearing titled “Questions Surrounding the “Hockey Stick’ Temperature Studies:”

    Link to C-SPAN2:

  136. fFreddy
    Posted Aug 7, 2006 at 11:36 AM | Permalink

    Re #136, Don
    Is it common for them to rerun Committee hearings ? Or does this imply there is an unusual amount of interest in this program ?

  137. Don
    Posted Aug 7, 2006 at 11:51 AM | Permalink

    Reruns is quite common for some of the hearings, usually edited. 3rd time this Mann-episode is running that I have seen.

  138. Steve Sadlov
    Posted Aug 14, 2006 at 12:19 PM | Permalink

    Here is an anecdote which attests to the level of ethics that some (but thankfully not all) amonst the warmist crowd sink to. Last Friday I quickly looked at:

    Someone had quite oviously hacked the server and manipulated the images. They had put in a massive crack in the Arctic ice, going from ~ Barrow reaching nearly to the North Pole. They had gone in and removed all the ice from the West Antarctic Peninisula, and had gone in and done the same S of Australia also creating a large “cracked off” piece that had been made to drift westward reattaching to the mass. If it was not done in such an obvious and exagerated manner, it might have resulted in a global panic. Nice work alarmists. I only wish that I’d done a snap shot of it.

  139. Steve Sadlov
    Posted Aug 14, 2006 at 12:24 PM | Permalink

    That image (from the 10th) is now missing from the 30 day. Indeed, it had to have been damaged, that would explain its removal. The mad hacker Gaia extremists … amazing.

  140. KevinUK
    Posted Aug 14, 2006 at 2:21 PM | Permalink

    Steve S,

    How can you be sure that it was a warmer who did this? To be fair to the warmers, from what I can see a skeptic could have had an equal motivation to conduct such a graffitic exercise. I somehow doubt that this act of obvious graffiti could have resulted in a ‘global panic’. Make such a statement makes you sound like an little alarmist yourself.


2 Trackbacks

  1. […] McIntyre has put up on Climate Audit several posts on a Congressional hearing, held over a couple of days, in which he was a witness. Mann was slated to give evidence too but, oddly enough, as Steve relates, he chickened out, claiming a sudden yet fogotten appointment as his excuse, as Steve relates. […]

  2. By The Climate Change Debate Thread - Page 3837 on Mar 30, 2014 at 11:24 AM

    […] […]

%d bloggers like this: