A Few Things Ill Considered

A layman's take on the science of Global Warming featuring a guide on How to Talk to a Climate Sceptic.

Wednesday, March 29, 2006

send this to... Digg it! | Technorati | Del.icio.us | Reddit | Furl | Spurl

The Hockey Stick is Broken

(Part of the How to Talk to a Climate Sceptic guide)

This article has moved to ScienceBlogs

It has also been updated and this page is still here only to preserve the original comment thread. Please visit A Few Things Ill Considered there. You may also like to view Painting With Water, Coby Beck's original fine art photography.



  • At March 29, 2006 1:37 PM, Anonymous Anonymous said…

    "Most striking is the fact that each record reveals that the 20th century is the warmest of the entire record, and that warming was most dramatic after 1920."

    How many times in history, as shown by the various graphs, is it possible to say. "Most striking is the fact that each record reveals that the "xxx period" is the "yyy" of the entire record, and that "zzz" was most dramatic after "ddd"."?

  • At March 29, 2006 3:00 PM, Blogger coby said…

    Since these various graphs are about temperature, we can constrain yyy to be either 'warmest' or 'coolest'. And since the only concern about a temperature change would be its rapidity, then zzz has only the value 'rate of change'. So, two times in history we could make a statement about warmest or coolest, and once we could say the change was most dramatic. The 20th century is the warmest and the most dramatic.

    But all baiting aside, do you have a real point to make?

  • At March 29, 2006 7:08 PM, Blogger Dano said…

    Coby cross-checks the contraHockey Teamers!

    Until someone goes out and actually collects their own data and shows that someone hid something or is not doing the statistics properly (which hasn't been done, so this is all finger pointing), the conclusion must be that one must go with the folks who do this for a living.

    And as you say, coby, there is plenty of empiricism there. It's a no-brainer until that happens.

    And as an aside/bonus: the tone at CA is such that no decision-maker's aide will go there and linger to see merits of the doubt-casting. It just smells too much like teen spirit there to be a serious inquiry site.



  • At March 29, 2006 7:25 PM, Blogger coby said…

    Hey! That was a clean body check, no stick at all! I thought there has been enough bickering and sniping about Bristlecones and R^2 statistics, ad nauseum so I took the high (as in step over the doodoo) road. Hope no one feels let down ;-)

    "Can't we all just get along?" lol

  • At March 31, 2006 5:20 PM, Blogger Dano said…

    Well, Canucks and Red Wings have been disagreeing for quite some time now, Coby, and I saw that stick go up.

    Not that I mind you checking that Ontarian anyways...



  • At April 04, 2006 4:26 AM, Blogger Wag the Dog said…

    The climate audit URL is incorrect. Should be www.climateaudit.org.

    It is fun to see them take the Bible Codes strategy -- that given enough mathematical munging and statistical tuning, you can get any conclusion out of noise, or indeed any body of data. It's a bit like watching a fish flapping around trying everything it can to get back to water. A comedy of incompetent decorators trying to hide a bumb in an overlarge carpet.

    Someone should catalog all their techniques and apply them to prove that Enron and Worldcom are still solvent, and their employees still have a pension.

    Having worked with principal components analysis and statistical modelling in my own work, I can appreciate the funny side. But it is sad that journalists lack the mathematical skills to report on this honestly. Tell people that one third of deaths occur during weekends, or that 40% of sickdays are taken on a Monday or Friday, and they immediately conclude some nefarious intention at work.

  • At April 04, 2006 6:25 AM, Anonymous Anonymous said…

    You've consistently misspelt "McIntyre" in the blog entry.

    Of interest is that McIntyre himself says that the "hocky stick" is not important to current global warming research. I agree that it has become something of a team flag/mascott that opposing teams want to tear down or do the strawman thing.

    On the topic of what laypersons can do? Lots. If you have a little computing skills there are plenty of holes you can poke into McKitrick's work. You can look at the impact factors of the journals. Nature and Science have the highest impact factors in science. McIntyre & McKitrick had to turn to a social science journal, Energy and Environment, to get published. The editor of this journal freely admits to an anti-global warming agenda:

    The journal's editor, Sonja Boehmer-Christiansen, a reader in
    geography at the University of Hull, in England, says she
    sometimes publishes scientific papers challenging the view
    that global warming is a problem, because that position is
    often stifled in other outlets. "I'm following my political
    agenda -- a bit, anyway," she says. "But isn't that the right
    of the editor?"

    The publishers of "Energy and Environment" think the M&M paper so important that they give it prominent positioning on the front page of their website, waiving the access fee they'd normally charge for downloading articles.

    So to believe M&M one must assume a failure of the peer review system. Both Nature ans Science have been fooled before in the case of Jan Hendrick Schon however this was quickly found out -- by peers reviewing the published work. Climate research is some of the most scrutinised research out there, yet the hockey stick papers have withstood scrutiny for quite some time. The same cannot be said of M&M who keep adapting their attacks , making fundamental flaws, and publish in non-peer reviewed journals on social aspects of science.

    So do we assume that everyone who reads Science and Nature are all in on one big conspiracy? And the review and editorial process of all the competing science journals are equally tainted? If so how can we bring ourselves to trust technology based on this conspiracy science? We should all just go back to living in a cave.

  • At April 04, 2006 12:26 PM, Anonymous Anonymous said…

    A simple-minded but fair approach might be to require visualizations of the data with and without the Bristlecone pine series. If the hockey-stick becomes a straight line by the removal of just one rather suspect methodology the study itself is suspect and should only be referenced with strong caveats.

    Imputing seasonal temperatures from visual inspection of tree rings has to be an error prone undertaking.

    It seems to me that Mann, et al. could easily provide the graph with and without the Bristlecone data. Why have they not provided this?

    I'm personally inclined to believe anthropogenic activity affects the climate. However, the last thing I want to see is data massaged to make a point. That's shameful science.

  • At April 04, 2006 2:40 PM, Blogger coby said…

    "A simple-minded but fair approach might be to require visualizations of the data with and without the Bristlecone pine series."

    You may be interested to read about a very recent analysis that seems to address this kind of point, ie how sensitive is the conclusion to any single data series. Real Climate presented a discussion here, I have not reread it but I think it is apropos your query.

    As for "why havn't they..." whatever, I think you need to recall the context. They published a paper, it was peer reviewed, they are experts in this field. Now without the benefit of hindsight and the knowledge of how public this single study has become, how much time should any of these researchers spend defending this study versus continuing their research? The way science works is that you publish and if someone has an issue they do their own analysis and publish that and science debates and progresses. What you are asking for is the scientists to take a direct interest in public relations and public education. I think that would be good but is not part of the job description and a scientist who choses not to do this should not be therefore suspect in some way.

    All that aside, Mann is a contributer to RealClimate which is a public outreach and they have devoted a lot of cyber real estate to the hockey stick issue.

  • At April 04, 2006 3:52 PM, Anonymous Anonymous said…

    No, the Bristlecone pines do matter, because the claims about the level of the Medieval Warm Period depend heavily on the Bristlecone pines (in nearly all of the studies--so far as I can tell) It's not unscientific to say so. Fortunately this question may be answered fairly soon, and it may be answered in favor of Anthropic Global Warming (or not), but it's not a trivial issue. (And this is not a troll comment!)

  • At April 05, 2006 1:16 AM, Blogger Wag the Dog said…

    If the hockey-stick becomes a straight line by the removal of just one rather suspect methodology the study itself is suspect and should only be referenced with strong caveats.

    It would prove nothing of the sort. All peer-reviewed scientific aready state what data the conclusions are based on and hence are already implicitly caveated. You are projecting your own biases by claiming bristlecones are suspect methodology and should be removed. What is the scientific basis for this? Removal of a dataset without justification would change the well documented method of the Mann98 paper as to make it a completely different paper. But I suppose that is what McIntyre&McKitrick are hoping for.

    the last thing I want to see is data massaged to make a point.

    Then disregard anything McIntyre or McKitrick say. The types of data manipulation contortions they do would make a high school maths wiz cringe in embarassment. Apparently, one only needs a working knowledge of spreadsheets or high school physics to be able to spot the flaws in McKitrick's methodology.

    It is a serious charge to imply Mann et al are massaging their data to reach a preordained conclusion. They have made their data and software freely available for download so that others can doublecheck their method and results, which others have done. They have applied the PCA method according to recommended statistical guidelines accepted by experts in statistical theory. In layman's terms, they have done everything by the book.

    The way principal component analysis works, if the bristlecone data is found to be a bad temperature proxy, it will be ignored relative to better proxy data -- automatically doing what you claim should now be done. Forcibly removing the series because one has a hunch it is unreliable violates the statistical rules. If it is unreliable, the method is designed to discount it.

    Only M&M seem to have objections. Judging from some of the stuff on ClimateAudit, they are clearly targetting a less mathematically and scientifically inclined audience, by aiming to confuse people with pseudo-methodology and spinning the science to make it look like global warming conclusions are snapped out of thin air. What they do is PR, not science.

  • At April 05, 2006 7:55 AM, Blogger Wag the Dog said…

    Until someone goes out and actually collects their own data and shows that someone hid something or is not doing the statistics properly

    Actually, the dispute is not that someone hid some data, but with the statistical processing of said data. But with regard to showing someone not doing statistics properly, then you'd be interested in the work of Ammann and Wahl 2006 which compares two of McIntryre&McKitrick's proxy reconstructions with their own and also a reproduction of the Mann et al. 98, subjecting them all to rigorous statistical testing.

    And they also perform a proxy reconstruction omitting the contentious bristlecone/foxtail data -- exactly what the anonymous commenter requested above. Read the paper online for the details.

    Or if you prefer a summary: Omitting the bristlecone did not flatten the hockeystick but that wasn't the point -- M&M want that data omitted so that the 15th century temperature will come out higher than today to refute the claim temperatures are hotter than in the last millennium. They obtain a double hockeystick instead but at the expense of higher variance and failing calibration and verification tests for the 15th century period.

    However, there is still some leeway left to adjust the Mann98 prediction upward for the the 1400-1450 period by 0.05 deg C, resulting in the slightly diluted claim that today's temperatures being hotter than in the last 600 years.

    So what have M&M achieved? The hockeystick is still there. The current rise in temperatures is still anomalous and more rapid than anything seen in any other time period of any of the reconstructions (even the statistically bad ones of M&M). And there are a whole load of other temperature reconstructions that M&M haven't even addressed. More and more it seems like a personal vendetta against one paper.

  • At April 05, 2006 8:03 PM, Anonymous Anonymous said…

    I'm not necessarily trying to defend M&M here, but I will say that McIntyre's criticisms seem to be made in good faith. Both authors even admitted that their reconstructions aren't statistically valid (and that was kind of their point...) and McIntyre, at least, has stated that he regards many studies since then the same way he does the original 1998 paper because they basically use the same datasets and analysis (And I should add that he seems less opposed to the more recent studies, especially those that don't use data he finds suspicious...) They've stated their a priori reasons why they don't like the data they don't like. They might be wrong, but that doesn't mean it's wrong for them to raise questions. Note that even the so-called "hockeystick team" recognized these potential problems, and even tried to correct for them. It does at times verge on becoming a personal vendetta, but I can kind of understand...

  • At April 13, 2006 12:20 PM, Anonymous Anonymous said…

    Wasn't all of this data based on a single tree? Can you really extrapolate global tempratures from one tree? Why would anyone considered good practice?

  • At April 13, 2006 12:30 PM, Blogger coby said…

    No, it is not based on one tree. You most certainly can not extrapolate global temperatures from one tree, nor even 1000 trees all in the same region. No one would consider that good practice.

    Now one for you: why would anyone make such an assumption?

  • At April 25, 2006 2:38 AM, Anonymous Anonymous said…

    It is of dubious value to even attempt to consider 'tree growth' in terms solely of 'temperature' and the quantity of refutation makes worthless the concept of doing such, regardless of tree/species numbers.

    When a 'modern' situation IS known, ALL the determinates of such would ALSO need to be validly defined for ANY period one attempted to produce such inference, regardless of 'when' such a period was.

    This has included soil structure, composition, humidity, global surface position, inter-flora 'species competition/symbiosis' (within a forest or other community, not all 'plants' are 'tress' i.e. have woody trunks) along with nutrient and water availability.

    In terms of any validatable measures (from suitable apparatus capable of 'reproducible measures'), this is limited to only the past 200 years.

    Too many unknowns produces too many assumptions..

    Your's, Peter K. Anderson a.k.a. Hartlod(tm)
    From the PC of Peter K Anderson
    E-Mail: Hartlod@bigpond.com

  • At October 26, 2006 11:00 PM, Anonymous Anonymous said…

    dano writes:
    "Until someone goes out and actually collects their own data and shows that someone hid something or is not doing the statistics properly (which hasn't been done, so this is all finger pointing), the conclusion must be that one must go with the folks who do this for a living. "

    Like Hegerl et al? Their newest "independent" reconstruction uses old data, most prominent MBH's PC1.

  • At December 04, 2006 7:00 AM, Anonymous Anonymous said…

    Until someone goes out and actually collects their own data and shows that someone hid something or is not doing the statistics properly (which hasn't been done, so this is all finger pointing)

    No. Actually it has been done.

    Like the independent group of statisticians (see report: http://energycommerce.house.gov/108/home/07142006_Wegman_Report.pdf) that concluded with respect to Mann's incorrect statistics:

    "It is not clear that Dr. Mann and his associates even realized that their methodology was faulty at the time of writing the MBH paper. The net effect of the decentering is to preferentially choose the so-called hockey stick shapes. The graphics’ prominence together with the fact that it is based on incorrect use of PCA puts Dr. Mann and his co-authors in a difficult face-saving position."

    And finally, "Overall, our committee believes that Mann’s assessments that the decade of the 1990s was the hottest decade of the millennium and that 1998 was the hottest year of the millennium cannot be supported by his analysis."


    Do scientific facts matter anymore?

  • At February 20, 2007 6:04 PM, Anonymous Anonymous said…

    Is the Hockey Stick discredited? Says who? And what if it were?


    Every year the hockey-stick projections (of years past) are discredited.

    And every year new hockey-stick projection is made, Doomesday is pushed back one year, a convenient ploy for the neo-Communists to NEVER HAVE TO PROVE ANYTHING!

  • At March 11, 2007 8:08 AM, Blogger Diogenes said…

    I note that there have not been any recent discussions of the (Hockey Stick) HS.

    Let me bring the reader up to date by referring to two reports on the HS that came out of the US Congress in 2006: the NRC report and the Wegman Report.

    The NRC Report stated: (1) that evidence exists that the 20th century was the warmest period in the last four centuries; and,(2) we could place less confidence in climate reconstructions for the period from 900 AD to 1600 AD. I think the long sought consensus exists on point (1). No one would argue against the proposition that we are warmer now than we were in the Little Ice Age. Point (2) calls into question 60% of the temporal length of the HS. So at best, the NRC report states the HS is 40% correct.

    The Wegman Report was chaired by Edward Wegman, Chairman of the National Academy of Sciences Committee on Applied and Theoretical Statistics. This report found fatal flaws in the statistics and data which are used to generate the HS.

    The rebuttals I read on Real Climate were less than convincing. They seemed to claim the NRC report vindicated the HS--60% wrong--and that the Wegman Report was not written by Climatologists--100% irrelevant since the Wegamn report addresses statistical problems that transcend disciplines. I also believe the Wegman Report validated the work of MacIntyre and McKintrick in that the algorith used to generate the HS would generates similar patterns if fed random data.

    Why is the HS important? Because without the HS, the Mideval Warm period sticks out as a hughe problem. If it was warmer then, than it is now, the argument that increased Carbon Dioxide concentrations in the Troposphere lead to global warming in on its death bed.

  • At March 11, 2007 1:48 PM, Blogger coby said…

    Hi Diogenes,

    I have no problem with your summation of the two reports. I do however disagree with translating "less confidence" as "wrong".

    This study is now very close to ten years old. Can't we move on? There have been many other proxy reconstructions, aren't we all just interested in learning as much as possible about past climates? There are not only two possible states for the MWP: barely there (HS) or large and globally pronounced (um...exactly what scientific research has shown this?). As near as I can tell, the most likely reality is that the MWP was indeed a generally warmer period but it was not globally sychronized and it was not as warm as now and it was not as sudden a change as now.

    BTW, you are hugely wrong about what a pronounced and global MWP would imply.

    It's almost like you did not read this post before replying...?

    Thanks for the comment nvertheless!

  • At November 26, 2007 7:31 PM, Anonymous Anonymous said…

    "Until someone goes out and actually collects their own data and shows that someone hid something or is not doing the statistics properly (which hasn't been done, so this is all finger pointing), the conclusion must be that one must go with the folks who do this for a living."

    Actually, I hate to break it to you, it's been done. And statistics is EXACTLY what proved Mann's methodology wrong.

    And what's with "collect their own data?" People were trying to VALIDATE the original conclusion! You can't do that with "your own data." You have to do it with the original data FIRST to ensure that the original conclusion is valid! THEN you go out and pick more data and test it with the hypothesis. The fact that Mann and company REFUSED to divulge their data speaks VOLUMES about their credibility and objectivity.

    Coby, you may sit there and claim that all of "the Canadians'" claims have been refuted, but the fact is, NONE OF THEM HAVE BEEN REFUTED. If you dispute this fact, then take a look at this link:
    Granted, it's by one of "the Canadians," but it shows that NO ONE has refuted one of their claims.

    I'd also challenge you to do a point-by-point rebuttal of "the Canadians'" testimony BEFORE CONGRESS, where they explicitly state WHY proxy studies are wrong and WHY there have been no truly "independent studies" since MBH98.

    The fact is, the stick IS broken. The original claims of MBH98 have all been crushed. Thoroughly. NO ONE has been able to replicate MBH98/99 with methods that pass scientific and statistical scrutiny while also replicating the stark statistical anamolies of MBH98. And until SOMEONE can replicate MBH98 scientifically, then MBH98 is hooey.

    The stick IS broken. And it DOES matter. Because without the stick, we have NO BASIS to claim that the current warming is unprecedented or even unusual.

  • At December 22, 2007 4:09 AM, Blogger Will Nitschke said…

    "Everything from this point on is hotly disputed and highly technical."

    To clarify: it is only "hotly disputed" by the authors of the paper and their associates who were criticised and all their supporters. It's not hotly disputed by the two independent review panels who found the paper to have serious errors.

    "All the claims made by M&M have been rebutted in detail by many other climatologists and they insist that these folks are completely in error."

    It's not a shouting contest. But I'm sure either side can drum up a few dozen colleges and friends to say 'I'm right, you're wrong' but is that really what this comes down to?

    "...the worst indictment from the climate science community came from a study led by Hans Von Storch that concluded M&M was right about a particular criticism of methodology but correcting it did not change the study results."

    So you are unaware that removing one proxy from their study (Bristlecone pines) completely eliminates the hockey stick shape? The tenuousness of the evidence presented by Mann doesn't concern you?

    "If you want to try to evaluate this issue fairly you must read the copious material at the sites mentioned above. You must also be prepared to get into dendrochonolgy and statistical analysis.
    Where does that leave the rest of us?"

    Untrue, you can simply read the summary evaluations of the independent review scientists and their opinions. The above statement seems to want to confuse an issue that is fairly straightforward.

    "The fact is there are dozens of other reconstructions. These other reconstructions do tend to show some more variability than MBH98, ie the handle of the hockey stick is not as straight, but they *all* support the general conclusions that the IPCC TAR came to in 2001: the late 20th century warming is anamolous in the last one or two thousand years and the 1990's are very likely warmer than any other time in the last one or two thousand years."

    I've heard this line of argument before, such at places like this:


    The problem is, this is propaganda, not science. The graph claims to show that 10 independent studies support the hockey stick.

    It turns out that the proxy reconstructions were done by the SAME group of scientists who did the first hockey stick. These are reconstructions: 1, 2, 3, 4, 5, 6 and 7. So the methods are not independent at all. Only 8-10 may be independent reconstructions.

    So the evidence for support of the hockey stick graph are more hockey stick graphs done by the same group of scientists who all know each and who all work together. No bias, eh?

    And this is apparently the best evidence we have supporting the merits of the hockey stick...

  • At March 09, 2008 12:22 PM, Anonymous Anonymous said…

    i wanna see both versions of the graph of climate over the years. then i can decide who is more accurate

  • At March 09, 2008 1:38 PM, Blogger coby said…

    A couple of points...firstly, how does seeing them help you decide which you believe unless you already "know" the answer? Secondly, there is no "sceptics" reconstruction, no opposing temperature history in the peer reviewed literature. That is one of the indications that they are just trying to cloud and confuse the issue rather than trying to uncover any truth.

    Thanks for dropping by.

  • At July 11, 2008 11:10 PM, Anonymous Anonymous said…

Post a Comment

<< Home