My Hegerl Predictions – Results

The reconstruction in Hegerl et al (J Clim) was previously used in Hegerl et al (Nature) which provided some information on the number of series and some other particulars, but did not identify the series. In order to show how farcical the Hockey Team claims of “independence” were, I made guesses last spring about the proxies that would be used in the Hegerl et al reconstruction using a principle of least independence. I made an extremely accurate emulation of the Hegerl et al reconstruction as published in Nature simply using the proxies in Osborn and Briffa 2006, describing the process as follows:

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

A few days later, I made the following prediction about what proxies would be in Hegerl et al:

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

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

10. Yakutia instead of Mangazeja. Mangazeja is a bit unusual; Yakutia is more consistent with maximal overlap and is used in DWJ06.
11. Quebec – I’m going to go with the Jacoby version rather than the Schweingruber version (which isn’t used outside of Esper/OB). either. One point for the Jacoby version; half point for Schweingruber version.
12. Tirol isn’t used outside of Esper/OB and isn’t maximum overlap. So something from the Jacoby treeline series, it could be the composite or it could be something like TTHH. One point for TTHH, half-point for the composite or half-point for another Jacoby treeline site. I’ll be mad if it’s Tirol.

I made a bet with TCO that I’d get at least 6 predictions right.

How Did I Do?

I think that even the most ungenerous reader would acknowledge that I nailed my predictions, far exceeding the terms of my bet – I pretty much got every location correct, I got most of the series versions exactly correct and, even where I did not get the version exactly right, I was nearly always substantially right. Here’s the report card.

Yang e.Asia:

this is the high resolution record (10-yr average) from Yang et al. (2002).

Correct

Taymir

Taimyr Peninsula: this is from Naurzbayev et al. (2002) by way of Esper.

Correct

Polar Urals

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

I haven’t figured out exactly what they’ve done, but I got the Polar Urals correct and I’m awarding myself a full point.

Mongolia

Mongolia: this is from the D’Arrigo et al. (2001) study. However, the full composite illustrated in this paper is not available. We reconstructed the composite from nine records from tree ring sites sent to the NGDC sites. The early growth part of the treering series from overlapping records was removed without further removal of low-frequency variability.

Again I haven’t figured out exactly what they’ve done, but I got the location bang on and I’m awarding myself a full point. Note the lack of availability of data to Hegerl. There’s something weird here as even Science was unable to get the Mongolia data as used by Esper, although they got the other 13 sites. I think that Esper must have lost the Mongolia version that he used.

Tornetrask

- n. Sweden: this is from Grudd et al. (2002) by way of Esper.

No surprises here. Full point.

Van Engeln

European historical: this composite was kindly provided by J. Luterbacher et al. (2004). I got slightly wrongfooted here.

I was right about a European documentary series, but they used Luterbacher starting in 1500 instead of van Engeln. I was sure that van Engeln was in because the CH-version in Nature started in 1251 which was the date that Luterbacher started. The plot in Figure A1 starts in 1500. Do I deserve part of a point here – it doesn’t matter.

Greenland dO18.

west Greenland: this composite is from Fisher et al. (1996).

Again no surprise. Full point.

Jasper
-

Alberta, Canada: this time series is also a composite of two different analyses of the 1997 reconstruction of Luckman (1997) – one is unchanged from Luckman’s paper, and the other (Athabasca) has been RCS processed by Esper et al. (2002). The correlation between these analyses is unimpressive (0.11); the records were simply averaged. Note that although the correlation with the decadally smoothed 30-90N instrument (land) temperatures varies greatly between the two records (0.14 for Athabasca, 0.82 for Jasper), the composite correlation is 0.84.

Full point, although I didn’t anticipate the average of the Luckman and Esper versions.

MBHPC1.

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

I didn’t anticipate the averaging of the PC1 and the foxtails (which were used in Osborn and Briffa and certainly within the Maximal Overlap concept. I think that I deserve a full point.

Yakutia

e.Siberia: the Esper et al. (2002) composite used the Zhaschiviresk time series from Schweingruber. However, this composite only went to 1708. We combined it with a ring width (by Schweingruber, available NGDC) series from the nearby Ayandina River site after removing the obvious growth overprint in the early part of the younger record. Ayandina River is at 6825-14310 from 1553 1991 russ063w. Zhaschieviersk RUS; 67N,142E; Schweingruber,F.H russ053w .

Got the right idea, but not the exact series. If I was scratching for points, I’d argue for part of a point. In terms of my equivocation about whether to go with Mangazeja or Yakutia (leaving Mangazeja off), it looks like they rolled Mangazeja in with Yamal somehow and it’s in after all.

Quebec -

Quebec: The situation with Esper’s Quebec reconstruction is somewhat similar to what was experienced for the Mackenzie Delta time series; the correlation between their Quebec record and the 30-90N average is only 0.25, partly because the time series ends in ~1930. Examination of the NGDC data base indicates that the original Esper et al. reconstruction appears to be from the Boniface site. A record from nearby St. Anne also shows many similarities to Boniface (r=0.66), extends closer in time to the present, but is also slightly shorter (the Boniface/St. Anne correlation is 0.70). Although the Boniface/St Anne composite has a very high correlation with the 30-90N (land) record (0.88), inspection of shorter records from Fort Chimo and No Name Lake showed a different 20th century response – earlier warming and late cooling. In order to preclude a Quebec composite from indicating a potentially unrealistic magnitude of late 20th century warmth for the whole region, we created a shorter composite of the four sites that averages records from a Fort Chino and No Name Lake composite after 1806. The new composite still shows significant warming in the 20th century, but not as extreme as the Boniface/St. Anne sites alone. It is not claimed that this solution represents the best possible way to deal with the conflicting evidence from Quebec; the problem can best be resolved if more long records from other regions of Quebec can be uniformly stacked together without any late-century adjustments.

The Ste Anne series is an alter ego for Gaspé. I think that I deserve a full point.

Northern Treeline

- Mackenzie Delta: The original time series (Szeicz and MacDonald 1995) provided by Esper et al. only had a 0.04 correlation with the 1880-1960 decadal average of NH temperature, which yields a very small weight if used for the hemispheric composite. We experimented with various other data from the National Geophysical Data Center (NGDC, http://www.ncdc.noaa.gov/paleo/ftp-search.html) to determine if other reconstructions for that area would yield more information for a hemispheric reconstruction. We found generally that proxy data for that region show little correlation with hemispheric mean temperature. We nevertheless included this site for the sake of completeness and in order to include as many long sites as possible.

I didn’t expect the Szeicz-MacDonald series, but it’s only used in Esper et al 2002. I was obviously uncertain about this series, but directionally got the right idea.

Anyhow, I think that my predictions of Hegerl proxy selection based on the principle of “least independence” have been massively vindicated and show how farcical the claims of proxy “independence” are.


40 Comments

  1. Posted Oct 25, 2006 at 1:09 PM | Permalink

    Is it possible at this stage to work out the various weights given to each of these proxy series (and I use that term in its loosest possible sense about MBH’s PC1)?

  2. bender
    Posted Oct 25, 2006 at 1:34 PM | Permalink

    Steve M, why is it commenters here in the past have insisted that YOU go out and get new independent data, when they don’t even do that THEMSELVES? They are recycling the same stuff over and over and over. Not sure if it’s ironic, or by design, but as they continue to do this, the multiproxy reconstructions are going to become less and less independent, and more and more in agreement, thus serving to shore up the “consensus”. That’s a mudhole that’s going to make it pretty hard to “move on”. At some point an independent test of the groupthink is going to come, and when it happens, their response will provide a clear indication as to who the real “stooges” in this debate are.

  3. Steve McIntyre
    Posted Oct 25, 2006 at 3:09 PM | Permalink

    Something that’s even funnier about this – the methodology of Hegerl et al is essentially the same as MBH99. MAnnian principal components on the North American tree ring network; then Partial Least Squares regression of NH temperature against the PC1 and other proxies; then re-scaling done a smidge differently. Now having cherry-picked the proxies, they didn’t need to do Mannian Partial Least Squares regression, but they did – perhaps because neither Mann et al nor Hegerl et al seemed to know what the hell multivariate methodlogy they were using.

    Then they have the audacity to claim that this is an independent methodology. The Team seems to have trouble keeping track of who’s on the ice.

  4. MarkR
    Posted Oct 25, 2006 at 3:14 PM | Permalink

    I’d like to know who’s paying for all these rehashes.

  5. Mark T
    Posted Oct 25, 2006 at 3:15 PM | Permalink

    And they’re supposed to be playing basketball anyway.

    Mark

  6. Mark T
    Posted Oct 25, 2006 at 3:17 PM | Permalink

    I’m guessing we are, MarkR… somewhere along the line, the money is probably coming from the taxpayers.

    Mark

  7. bender
    Posted Oct 25, 2006 at 3:17 PM | Permalink

    Excellent point. Independence of source data and independence in the peer review process are distinct issues, both addressed by Wegman & NAS. Independence of analytical methods is a different one that has not received nearly as much attention as the other two.

  8. Steve McIntyre
    Posted Oct 25, 2006 at 3:22 PM | Permalink

    #4. NOAA, NOAA’s office of global programmes and the DOE Office of Science, Office of Biological and Environmental Research, and by Duke University. GH was additionally supported by NSF grant ATM-0002206 and ATM-0296007. EZ was partially funded by the European research project SO&P. HP and JS were supported by NASA Grant GEWEC-0000-0132, and JS was additionally supported by a Lamont-Doherty Postdoctoral Fellowship

  9. Steve McIntyre
    Posted Oct 25, 2006 at 3:24 PM | Permalink

    Nanne Weber of KNMI and the MITRIE project have another one of these “independent” studies coming out. Nanne said that they used the series that were “available”. I offered to bet her 500 Euros that they used the Yamal series and not the Polar Urals Update – she had a free swing, she presumably knew the answer and I didn’t, but she didn’t take the bet. It’s going to be another re-hash of the same stuff.

  10. Mark T
    Posted Oct 25, 2006 at 3:25 PM | Permalink

    NOAA and NASA, of course, being taxpayer (US) funded organizations.

    Mark

  11. Steve McIntyre
    Posted Oct 25, 2006 at 4:13 PM | Permalink

    Hegerl et al stated:

    Examination of the NGDC data base indicates that the original Esper et al. reconstruction appears to be from the Boniface site. A record from nearby St. Anne also shows many similarities to Boniface (r=0.66), extends closer in time to the present, but is also slightly shorter (the Boniface/St. Anne correlation is 0.70). Although the Boniface/St Anne composite has a very high correlation with the 30-90N (land) record (0.88), inspection of shorter records from Fort Chimo and No Name Lake showed a different 20th century response — earlier warming and late cooling. In order to preclude a Quebec composite from indicating a potentially unrealistic magnitude of late 20th century warmth for the whole region, we created a shorter composite of the four sites that averages records from a Fort Chino and No Name Lake composite after 1806. The new composite still shows significant warming in the 20th century, but not as extreme as the Boniface/St. Anne sites alone

    I calculated the correlations between the Esper Bonif series as provided to me and the Ste Anne/Gaspe (cana036) series. I gor a correlation of 0.08, not 0.70 as reported. Here is a graphic of archived chronologies for what I presume to be the series in question.

  12. Steve McIntyre
    Posted Oct 25, 2006 at 4:32 PM | Permalink

    I’ve not waded through the Luterbacher reconstruction; it looks very Mannian in methodology, principal components; Mann cites Luterbacher as authority for the idea of only using RE statistics. If you look at Lutrebacher’s SI, on page 11, he uses 10 series for his 1500-1659 reconstruction, including, somewhat surprisingly, Yamal; also van Engeln, Swedish tree rings near Tornetrask attributed to Kirchhefer 2001; Greenland dO18 (Vinther version not Fisher version).

  13. bender
    Posted Oct 25, 2006 at 5:13 PM | Permalink

    But gaspe and james bay are so close together. LOL

  14. jae
    Posted Oct 25, 2006 at 5:48 PM | Permalink

    Congrats, Steve. I’m especially glad that you won the bet with TCO.

  15. Posted Oct 25, 2006 at 7:36 PM | Permalink

    Haha, I hope that TCO will have to eat a spider or something like that because he deserves it because of his occassionally overly limited confidence in Steve! ;-)

  16. Steve McIntyre
    Posted Oct 25, 2006 at 9:45 PM | Permalink

    The sloppiness of their data identifications is really quite awful. They say:

    Yamal (Briffa et al. 1995) and Mangazeja (Hantemirov and Shiyatov 2002 – both by way of Esper et al.) and two records from west of the Urals (Hantemirov and Shiyatov 2002).

    In a couple of sentences, this has the following errors:
    (1) Yamal is not mentioned in Briffa et al 1995, which is the pre-update Polar Urals data set.
    (2) Mangazeja is not mentioned in Hantemirov and Shiyatov 2002
    (3) The data that I got from Esper was the Polar Urals Update rather than Yamal – I’ve had trouble replicating the late portion of Esper. I wonder if he actually used the Yamal data set (which he sent to Hegerl et al) but accidentally sent me the Polar Urals update. Or perhaps Hegerl have mis-described what they got. If I ever get the exact data set, I’ll be able to tell.
    (4) Hantemirov and Shiyatov 2002 describes the Yamal data set and does not describe any data sets west of the Urals. In effect, they’ve attributed all the series to Hantemirov and Shiyatov that were not in it.

    What a schmozzle.

    The foxtail attribution is just as bad. Lloyd and Graumlich 1997 do not mention the Boreal and Upper Wright Lake sites. Does any of this “matter”? Who knows, but it sure is sloppy. In his testimony to Congress, Crowley said that Crowley and Lowery 2000 used the “bonehead” method, a term that I felt at the time should be reserved for the MBH method. However, I think that there is now a legitimate beauty contest.

  17. BradH
    Posted Oct 25, 2006 at 10:19 PM | Permalink

    Headline: New Study Confirms All 17 Year Olds Have Red Hair

    A new study confirms what the experts have been telling us for years – that all 17 year olds have red hair. “Amazing as it sounds, it’s true”, said the study’s lead researcher, “There have been many studies in the past showing this to be the case and our new, independent study confirms those results.”

    Over the past ten years, no less than 10 independent studies, using different methodologies have confirmed that all 17 year olds have red hair, but this study extended that research further and, astonishingly, reached the conclusion that all 17 year olds are also males.

    “There have been some people in the past who have refused to believe the results,” he said, “But I think now that there is no doubt about it. What surprised us was that even though we used a novel methodology, our results so closely mirror the results of other researchers over the past 10 years that they should satisfy even the most hardcore carrot-top skeptics. What we didn’t expect was our discovery that all 17 year olds are also male. Of course, there will need to be further studies done to confirm this latest finding.”

    While the weight of evidence seems to be overwhelming, still there are a small number of critics who maintain the methodology is flawed and the results are less certain than the majority of scientists maintain.

    One red hair denier contacted by this news agency was scathing of the new study. “It’s just like all of the other studies and suffers from the same flaws,” he said, “About 15 years ago, the first researcher knocked on the door at 1 Bradford Street, Greenborough and it was answered by Kevin Smith, who was 17 years old and a red head. From that limited sample, they concluded that all 17 year olds had red hair. Since then, all these so-called ‘scientists’ have done is re-work that sample. It’s a load of bollocks.”

    However, the author of the present study refutes this. “There will always be some hardcore elements out there who are impossible to convince,” he said, “But people have looked at the data in every possible way. In our own study, for example, we applied a partial least squares regression to the data, which has never been done before.”

    Previous studies have used other methods, such as principal components analysis, Monte-Carlo studies and so on. If there was a problem with the data, the author said, the results of these studies would not be so consistent.

    A spokesman from the Ministry of Youth said that, the study highlighted the appalling number of teenagers abusing hair dye and that a national education campaign in schools had already commenced. “We know that all 17 year olds have red hair,” he said, “So obviously we’re very concerned at the number of adolescents dying their hair blonde, brown and even black. They need to come to terms with their ginger heritage and that’s what our program aims to achieve.”

  18. TAC
    Posted Oct 25, 2006 at 10:51 PM | Permalink

    #17 LOL (or is it ROFLOL?). Whatever. Nice job!

  19. BradH
    Posted Oct 27, 2006 at 7:28 AM | Permalink

    Re:#2

    Steve M, why is it commenters here in the past have insisted that YOU go out and get new independent data, when they don’t even do that THEMSELVES? They are recycling the same stuff over and over and over.

    I’ve regularly thought the same thing myself, bender.

    Their data and methods are so similar, many of these papers border on plagiarism. Here’s a definition from dictionary.com:-

    1: a piece of writing that has been copied from someone else and is presented as being your own work

    2: the act of plagiarizing; taking someone’s words or ideas as if they were your own

    Not exactly within the definition, but my goodness, it’s close.

  20. Paul Penrose
    Posted Oct 27, 2006 at 7:42 AM | Permalink

    Actually, it’s more like inbreeding, and you know what that causes.

  21. Gerald Machnee
    Posted Oct 27, 2006 at 8:13 AM | Permalink

    Steve M’s predictions make the sportswriter’s predictions look very dismal even when they know all the players.

  22. bender
    Posted Oct 27, 2006 at 8:28 AM | Permalink

    Re #19
    BradH, Note the Wilson et al. abstract over in the AGU thread. Looks like they actually might have done a truly independent reconstruction. We’ll see what the paper says when it comes out.

  23. Dave Dardinger
    Posted Oct 27, 2006 at 9:00 AM | Permalink

    re: #20

    Actually inbreeding is good for the species, even if it’s not good for the individual. It brings recessive lethals to the surface and thus allows them to be weeded out. And in one sense, the same may be true of the multi-proxy climate crowd. The very fact that they’re being exposed may make science in general take a second look at the appropriateness of the methods and procedures in use and ultimately improve itself.

  24. BradH
    Posted Oct 27, 2006 at 9:02 AM | Permalink

    Re: # 22

    Yes, bender, I looked at that.

    However, is it an independent reconstruction of the same data? I’m not clear on that point, from what has been discussed.

    My point is that you can re-work any data set in different ways, if you wish. However, if there are inherent inaccuracies in data it doesn’t matter whether or not your method is independent or not. As I understand it, tree ring widths are not 100% correlated with temperature – we understand that, at the very least, they are affected by precipitation and fertilization. I have not seen any convincing corrections for these phenomena.

    If precipitation or fertilization variations can – either theoretically or observationally – alter the temperature signal and you have no effective means to adjust for this, you have a dud proxy, QED.

    In such a case, it wouldn’t matter how you studied the data.

    [BTW, for the record, I believe Rob Wilson to be a genuinely honest person, who believes in tree rings as accurate temperature proxies. I still have personal problems with precipitation and fertilization, but freely admit that I am not in a position to judge those issues scientifically.]

  25. Steve McIntyre
    Posted Dec 5, 2006 at 12:21 PM | Permalink

    I wrote to Gabi Hegerl enquiring about the Mongolia and Urals versions and for a version of the original data as used. She promptly replied sending me smoothed and truncated data and pointing me to the sources for Mongolia and Urals in the text (where it isn’t given). I replied:

    The original paper does not provide an exact data citation for the Urals and Mongolia series that I asked about. It says:

    “Mongolia: this is from the D’Arrigo et al. (2001) study. However, the full composite illustrated in this paper is not available. We reconstructed the composite from nine records from tree ring sites sent to the NGDC sites. The early growth part of the treering series from overlapping records was removed without further removal of low-frequency variability.

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

    I can’t tell what you used from this information or what you did. Can you provide details on these series – preferably in the form of exact digital citations (e.g. according to AGU data citation policies). The data that you sent has been smoothed and truncated. Do you have the unsmoothed untruncated versions that preceded these? Thanks, Steve McIntyre

    She replied that Crowley compiled these series, that he had a disk failure and it may be some time before he can reply.

  26. Posted Dec 5, 2006 at 12:40 PM | Permalink

    She replied that Crowley compiled these series, that he had a disk failure and it may be some time before he can reply.

    This is why we have backups.

  27. Hans Erren
    Posted Dec 5, 2006 at 4:30 PM | Permalink

    At my university we used to have a network where the system administrator would do the backups.
    Storing unique data that is essential for publication on a local disk drive without backup is [...]

  28. Hans Erren
    Posted Dec 5, 2006 at 4:31 PM | Permalink

    … unprofessional

  29. Brooks Hurd
    Posted Dec 5, 2006 at 4:43 PM | Permalink

    Hans,

    I could not agree more.

    My data is backed up on 2 or 3 physcially separate hard discs. I have a commercially available program that automatically backs up my notebook’s hard disc to the desktop’s hard disc.

    Hard drives are quite reliable these days with >200,000 hours MTBF, however they still do fail. It is a question of “when” not “if.” Anyone who does not plan for it, will suffer from hard disk crashes.

    I remember that Steve M was told of other hard drive failures as a reason why data was unavailable.

  30. Armand MacMurray
    Posted Dec 5, 2006 at 8:00 PM | Permalink

    It shouldn’t stress Crowley’s budget too much to hire a data recovery company, especially given the expense of trying to resurvey/resample a site. I can’t think of a good reason for not doing so. Perhaps Steve should suggest one or two at his next conference hobnob with the Team, just so they can’t claim ignorance.

  31. Steve McIntyre
    Posted Dec 5, 2006 at 8:32 PM | Permalink

    Since Crowley also lost the data for Crowley and Lowery 2000, you’d think that he’d be a little more careful. At the House Energy and Commerce Committee, he described his methodology as the Bonehead Method. I thought that this term was more properly reserved for MBH, but there seems to be an active competition within the Team.

  32. Steve McIntyre
    Posted Dec 5, 2006 at 9:22 PM | Permalink

    I didn’t formally predict the series in the Juckes composite, but I might as well assess the Juckes composite in terms of my Hegerl predictions. I would have done it a bit differently given notice of 18 series. Here were my 9 core predictions for HEgerl:

    1. Yang composite
    2. Taymir
    3. Polar Urals – on balance, I expect the Briffa MXD version, but it could be the Yamal version. Either qualifies.
    4. Mongolia
    5. Tornetrask
    6. van Engeln
    7. Greenland dO18
    8. Jasper – on balance I expect the Luckman version, rather than the Luckman-Wilson version which is too recent, but either qualifies.
    9. MBH PC1 (I’d add in the Graumlich Foxtails used in Osborn and Briffa).

    1. Yang Composite: Bingo.
    2. Taymir – bingo
    3. Yamal – double bingo. Both the Briffa 1995 and Yamal versions are used.
    4. Mongolia – surprisingly not in the Juckes set. There’s something quite weird going on here. It is HS and it fits the criteria; I think the problem relates to the unavailability of the version used by Esper, which seems to have gone AWOL. Maybe it’s on Crowley’s disk drive.
    5. Tornetrask – double bingo. It was used twice.
    6. van Engeln – all of the Juckes versions go to AD1000, so this would not have been on my list.
    7. Greenland dO18. Bingo
    8. Jasper – not in the Juckes set – presumably because the version in play in 2005 ends in 1072.
    9. Bristlecones/foxtails – both Graumlich series, plus 2 bristlecone series used in Moberg (they are Methuselah Walk and Indian Garden – odd choices since they are not strongly HS).

    This makes 11 of the 18 Juckes series. Of course, since Juckes’ criteria included prior use of the proxy by the Team, it’s not very surprising to have made these predictions.

    Juckes HS-ness comes from the 2 foxtail series, Yamal, Yang (Guliya and Dunde) and the Arabian Sea G Bulloides series. The results from the other 13 series have a higher MWP than modern.

    The Team doesn’t really need to keep publishing permutations and combinations of these 5 series over and over again. I am prepared to stipulate that manipulations of foxtails, Mann’s PC1, Yamal, G Bulloides and Dunde-Guliya yield a HS result under a variety of methods. OF course, permutations of Polar Urals update, Yakutia/Indigirka, Sargasso Sea SST, GRIP borehole, Khim’s Bransfield Strait don’t. I wonder if that’s why they aren’t in the Juckes Union.

  33. John Norris
    Posted Dec 5, 2006 at 10:28 PM | Permalink

    Re #32:

    Dr. Juckes appeared to set out to defend/support MBH by making more Hockey Sticks with new twists of previous used proxies. I find it rather ironic that the net result is he is providing more opportunity for you to make your point, and in turn further substantiating MM critique of MBH methods.

    I also half expect that you would like Hockey Sticks to keep coming from Dr. Juckes, or whomever. With respect to credibility, I certainly don’t see MBH stock going up with Dr. Juckes latest effort. And with him finding new ways for you to make your point I suspect you are getting a fair degree of entertainment value out of this as well.

  34. Steve McIntyre
    Posted Dec 6, 2006 at 7:35 AM | Permalink

    #30. Having had a computer crash myself recently, I can tell you exactly. I took my computer into the retailer. It cost me about US$65 to restore the system including recovery of the data in the hard drive. I took the computer in at 10 am and had it back at 7 pm. I can’t imagine why it would take Crowley any longer.

  35. James Erlandson
    Posted Dec 6, 2006 at 10:26 AM | Permalink

    Re 30: Not all hard drive failures are the same. They range from broken and fixable to scrambled sectors and unrecoverable. Although you may not have though so at the time, you were lucky. For about twice your $65 repair bill most on-line retailers well ship you 200+GB backup hard drive that installs in seconds (USB cable).
    No regulated financial institution or publically traded company can operate without implementing a data recovery plan and no university or research institution should expose themselves to the expense and ridicule that comes from losing valuable data. These organizations bask in the glow of their faculty’s research and notarity. Don’t they have an obligation monitor (audit?) their scientists’ work to make sure they don’t do anything stupid?

  36. TCO
    Posted Aug 3, 2007 at 3:19 PM | Permalink

    Steve, who owes who dinner?

  37. Steve McIntyre
    Posted Aug 3, 2007 at 3:27 PM | Permalink

    I won by a landslide. And then Juckes used almost exactly the same series. The lack of “randomness” is pathetic. Imagine someone being able to predict their selections in advance.

  38. TCO
    Posted Aug 3, 2007 at 3:31 PM | Permalink

    Thanks for the definitive response, Steve. I will go check on the specifics and then respond. There were some things that bugged me about what you wrote (‘could be argued’ and the like). I owe it to you to dig into it though.

  39. TCO
    Posted Sep 27, 2007 at 8:49 PM | Permalink

    This is a loose thread, that I need to wrap up. (My hands are still very washed.)

    When I made this bet with Steve, it was my hope/expectation that he would win. My desire to get him some cash, for the fun, I’ve had reading the blog. (This was pre tip jar.) His guesses got a bit legalistic, when we really set a bet up. Not exactly just 12 guesses, but some either-ors to give him more chances. This takes away from the message of cherrypicking and distinctiveness and max overlap (since more combinations still give a similar pattern). But, ok.

    What really bugged me was after the paper came out, Steve’s reaction to it. I was disturbed that there was no direct statement by Steve that he had “won the bet” or of exactly how many series he had guessed correctly. Several of his comments, seemed to anticipate an argument (wanting “fractional credit” or his ‘I could argue that’, type statements.) It’s the legalistic style of holding a concession in advance rather than the scientific style of testing a hypothesis and reporting results either way.

    In addition, a very big difference that I learned from the Hegerl paper was that these were not 12 series or even 12 composites as we normally identify them. But that several of the series were actually meta-composites covering several studies in a region. Like finding out that a reporters source was a “composite”. Steve had not predicted that. (There’s actually a hint of it in a parenthetical statement in Hegerl’s Science article, but Steve was so locked onto OB2006 and the guessing game, that he thought in terms of normal series.) And he did not note it when, we learned it in the paper. This is a major difference of the structure of the Hegerl reconstruction from what Steve predicted.

    So, I don’t accept Steve’s comments on face value and look at the actual Hegerl paper and Steve’s bet guesses (http://www.climateaudit.org/?p=642#comment-21075), myself to make a comparison.

    In several cases, Steve refers to series by colloquial names This can be ambigious. For instance, “Mongolia” is a country, not a temperature reconstruction! I understand that Steve, knows what he means when he says this, but it’s still a poor practice, like referring to VS04, when Von Storch wrote more than one paper in 2004 (and you don’t have a specific citation either). To understand these, I went back to the Osborne and Briffa study (table S1) to check for the “principle of max overlap”, Steve was addressing.

    _____________________________

    Yang composite:
    Good job, Steve. (1)

    Taymir:
    Same original series, different version (OK). Good job, Steve (1+1=2)

    Polar Urals – on balance, I expect the Briffa MXD version, but it could be the Yamal version. Either qualifies.:
    This is actually a “meta-composite”. It has 4 records mixed in. You predicted one of four (it was the Yamal). No granting yourself a “full point”. A meta-composite is not a specific series, no partial credit, this was a major thing you missed about what was being done and that you did not address squarely when we learned about it. Bad job, Steve. (2+0=2)

    Mongolia:
    Good job, Steve. (2+1=3).

    Tornetrask:
    Same series (not sure if Esper versus Briffa is a version difference, but those are kosher anyhow.) (3+1=4)

    van Engeln: Your comment about “do I deserve a point, it doesn’t matter” seems legalistic rather than observational. Luterbacher is not Van Engeln (and doesn’t even cite him, I read the article). Also, they differ in the 1500s: http://home.casema.nl/errenwijlens/co2/errenvsluterbacher.htm (4+0=4).

    Greenland dO18: Looks good. (4+1=5)

    Jasper – on balance I expect the Luckman version, rather than the Luckman-Wilson version which is too recent, but either qualifies.:
    I assume you mean the 1997 cited article. If so, good job. But you MISSED THE META-COMPOSITE and ESPER SERIES. It’s not the series you say it is, if it’s an average of two series. (5+0=5)

    MBH PC1:
    This is a very imprecise reference (I know you know what you mean by it, but others are not brainreaders and it’s ambiguous). Also not clear at all if “Hughes provided western US series” is same as Mannian PC1. Not sure why you give yourself credit without knowing this. In any case, THIS IS ACTUALLY A META-COMPOSITE. He mixes in others by straight up average combination with “Hughes”. BTW: “I think that I deserve a full point”. NO!!! First of all, there ain’t a bunch of damn fractional points running around here, anyhow, so what’s this “full” crap? And second, you missed something big with the meta-composites. That’s not principle of max overlap. NOt the uncanny accuracy that you seem to imply is so special. Bad job. (5+0=5)

    Yakutia instead of Mangazeja. Mangazeja is a bit unusual; Yakutia is more consistent with maximal overlap and is used in DWJ06.:
    You didn’t get it right. That there are multiple series in various river basins, giving similar results speaks to non-cherrypicking and usefulness of proxies in general and speaks to independance of the exercises. Your comments about “If I was scractching for points, I’d argue for a partial point” is odious. You got the WRONG SERIESes. You missed the meta-composite. And the basic idea of changing your analysis based on how it helps you is ethically wrong. Do you call the in/out bounds in tennis (a sport where one calls one’s opponents shots), based on how much you need the point? (5+0=5)

    Quebec – I’m going to go with the Jacoby version rather than
    Schweingruber version (which isn’t used outside of Esper/OB). either. One point for the Jacoby version; half point for SChweingruber version.:
    This is another one, where you gave yourself credit for more than one guess within a guess. But you still didn’t get it! And your comment “The Ste Anne series is an alter ego for Gaspé. I think that I deserve a full point.” is weasely. What is with all this partial credit crap? And besides, it’s a different series, different trees. No soup. The point of the exercise is max overlap. (5+0=5)

    Tirol isn’t used outside of Esper/OB and isn’t maximum overlap. So something from the Jacoby treeline series, it could be the composite or it could be something like TTHH. One point for TTHH, half-point for the composite or half-point for another Jacoby treeline site. I’ll be mad if it’s Tirol.:
    You gave yourself a multitude of choices here. One of them seems like it isn’t even a third series “another Jaboy treeline site”, but just a RESEARCHER!! But, in any case, you got it wrong. (5+0=5)

    ————-

    I also think that the point of “maximum overlap” became much less maximum, when we dug into it. Also that you were able to create a similar looking trend series for matching with some significant differences in series, may show that there’s more sense in “somewhat independant” combinations giving “somewhat similar” reconstructions. Of course, I think 5 out of 12 is still a lot of overlap. Definitely, not two independant exercises. And it just has to be that way, I guess based on how many good records are out there.

    ————-

    If there’s something I’ve missed in series identification, let me know. I still want to get you a couple more points, so you can get your money. I won’t listen to lawyerly arguing about, “what you choose to argue” or about partial credit or even full credit for “series” that were meta-composites.

  40. Steve McIntyre
    Posted Sep 27, 2007 at 11:17 PM | Permalink

    OK, TCO, fair enough, let’s have some fun with this and review the bidding.
    The bet was:

    6 or more- I win, 5 – a draw, 4 or less you win; site identification is the criterion not the exact version;

    To which, TCO replied: “1. 50% is a wash.”. I don’t see any final agreement on this, but I’m over 6 anyway.

    I’ve probably got a bit information on these sites than before, but I’ve been unable to totally pin down what they did. Nonetheless, I think that I’m way over 6.

    1. Yang composite -conceded
    2. Taymir – conceded (1+1=2)
    3. Polar Urals – on balance, I expect the Briffa MXD version, but it could be the Yamal version. Either qualifies.

    They’ve cocked up their description of the series, which is confusing you a little. They credit the “Yamal” series to Briffa et al 1995, but this is the Briffa MXD series that I refer to above. They mention two sites records from “west of the Urals (Hantemirov and Shiyatov 2002)”. Hantemirov and Shiyatov 2002 is a report on Yamal. Are there “two” sites and what are they? I suspect that there’s only one series. Mangazeja is in Esper (and was a site that was in my initial list and changed for Yakutia.) Given that I predicted either Briffa MXD or Yamal and the result (as best as I know right now) was the average of Briffa MXD, Yamal and Mangazeja, you’re being cheeky to say 0. The bet was on the sites, not the precise version. So I propose 2/3 of a point here, round down to .6. (2+ .6= 2.6)

    4. Mongolia – conceded (2.6+1 = 3.6)
    5. Tornetrask – conceded (3.6+1=4.6)
    6. van Engeln – (4.6+0 = 4.6)

    Your comment about “do I deserve a point, it doesn’t matter” seems legalistic rather than observational.

    Jeez, you’re getting chippy. I said that it didn’t matter in the sense that I was well over 6 with or without it. I was right in getting a European documentary version, but I didn’t get the exact documentary version. I don’t think that I was totally off-base and it’s not that I was totally wrong about this prediction. But I didn’t ask for the point.

    7. Greenland dO18 – conceded. (4.6+1= 5.6)

    8. Jasper – on balance I expect the Luckman version, rather than the Luckman-Wilson version which is too recent, but either qualifies.

    TCO, I totally disagree with you here. You say -0. I say a full point. “Athabasca” and “Jasper” are the same site. The Hegerl version seems to be an average of the Luckman version and Esper’s version using Schweingruber’s sample. The bet was for the site, not the exact version and I got this one bang on. Full point. (5.6+1=6.6)

    9. MBH PC1

    You say that this is “a very imprecise reference”. Puh-leeze. If you thought it was “ambiguous” , you should have asked for more particulars last spring. You also say that :”Also not clear at all if “Hughes provided western US series” is same as Mannian PC1.” You’re right that it isn’t clear. If I remember correctly, I think that I got some additional data on this – or that I matched a graphic. Anyway, I’m 100% sure that this is what they used. They say that they blended the Mannian PC1 with foxtails used in Osborn and Briffa, taken about 30 miles from the bristlecones. At a minimum, I get 0.5 points, but this is very conservative in that I pretty much nailed what they were going to do. (6.6+0.5=7.1)

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

    TCO, you’re arguing about a point a point that I didn’t claim. Jeez, you’re being ridiculous. But I was right to guess something from Siberia(7.1+0=7.1)

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

    They say that they used a blend of 4 sites: 1) Ste Anne/Gaspe (these are the same site despite your protests to the contrary, trust me on this), which is what I meant by the “Jacoby version” in my prediction, although this isn’t very clear; 2) Boniface – this is the “Schweingruber version” used in O and B; 3) Chimo 4) No Name Lake. They used a composite, of which 50% were my prediction. So 0.5 points (7.1+0.5).

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

    I guessed something like TTHH in the Yukon; it turned out to be the Campbell Dolomite site in the Yukon. I didn’t claim points as I didn’t need them.

    You can cavil all you want, but if you step back – if these sites are “randomly” selected, should I have been able to predict the selections as well as I did? Of course not.

    Reviewing the bet, TCO, you’re completely offbase on the Jasper site. I nailed it for a full point – no two ways about it and, together with the 5 conceded points, gives me the 6 points.

    Now let’s go back to Polar Urals. I predicted Polar Urals (either Briffa MXD or Yamal version.) These were used with Mangazeja blended in. The Mangazeja blend doesn’t affect things much. Is my prediction “wrong”? OF course it isn’t. However you cut it, I’m entitled to partial points to put me over 6 points and win the bet.

    Alternatively, I predicted the MBH PC1. It was blended with foxtails, but again my prediction was not wrong. Any part points put me over.

    Alternatively, I predicted two of the 4 Quebec sites: Ste Anne/Gaspe, Boniface.

    You cannot seriously be saying that having predicted the Polar Urals, the MBH PC1, Gaspe and Boniface, I’m entitled to 0 points on these sites. And you accuse me of being legalistic.

Follow

Get every new post delivered to your Inbox.

Join 3,329 other followers

%d bloggers like this: