Briffa on Yamal Impact

Keith Briffa has a couple of posts today on Yamal – one discussing its impact in other multiproxy studies and the other on the Yamal chronology itself. His post on the Yamal chronology includes a careful consideration of various issues involved in the development of the Yamal chronology and is accompanied by an extensive archive of original data. The concurrent inclusion of data is to be commended.

It will take a little while to consider this as there’s a lot of data to assimilate. First, let me discuss the post on the impact, which I can do quite quickly.

Briffa considers three aspects of Yamal use: (1) in Osborn and Briffa 2006; (2) in IPCC 2007; and (3) in other multiproxy studies. He concludes:

Thus, with the exception of Briffa (2000), the reconstructions shown by the IPCC (Figure 6) either do not use the Yamal record or they combine the Yamal record with many others and this reduces their sensitivity to the inclusion of any individual records.

I disagree with this summary for several reasons outlined below.

I considered the impact of Yamal on other multiproxy studies in a CA post here, which is, unfortunately, not linked in the corresponding Briffa comment. There are many points of empirical agreement, but considerable differences in emphasis. The differences in emphasis are significant: I considered the use of both Yamal and strip-bark bristlecones/foxtails, identifying several “families” of reconstructions: a family in which strip bark bristlecones/foxtails were highly influential (but not Yamal); a family in which Yamal was highly influential (but not bristlecones); a belt-and-braces family in which both were used and which typically asserted “robustness” to removal of individual proxies; and a few odds-and-ends which were affected by other questionable proxies e.g. by the “cold” 11th century Polar Urals of Briffa et al 1995.

Yamal arose as an issue in the wake of our criticism of MBH, because of Team assertions that the Mann reconstruction was supported by various “independent” reconstructions. We observed (and this point has been also observed by Briffa in the past) that the “independent” reconstructions are not, in fact, “independent” because of their re-use of the same proxies over and over, especially strip bark bristlecones/foxtails (both directly and as Mann’s PC1). Strip bark has been the more contentious issue; as I noted in my prior post ( a point also made in Wegman 2006), many of these supposedly “independent” reconstructions use strip bark bristlecones. However, not all of them do. Yamal became important as an issue because it is influential in reconstructions that do not use bristlecones (Briffa 200, D’Arrigo 2006, Kaufman 2009). To the considerable extent that Briffa’s response to Yamal impact is that they can “get” a stick using strip bark, the response is, shall we say in Mannian-speak, a little “disingenous”.

Briffa’s statement contradicts empirical observations at CA 7229 in only a couple of places and, in each case, I can demonstrate that Briffa is wrong. Both errors are relevant.

First, Briffa said that Hegerl et al (2006) did not use Yamal, while I said that they did. Hegerl et al (Nature 2006) url does not list the series that it uses; the companion Hegerl et al 2007 url states:

the west Siberia long composite involved Yamal and the west Urals composite.

Precisely how Yamal is “involved” in the west Siberia long composite is not reported. I’ve been trying to get the exact versions used by Hegerl since fall 2005 and despite many emails (mostly involving Crowley) have thus far been unsuccessful. (As I’ve mentioned elsewhere, my original attempts to get this data led to IPCC WG1 Chair Susan Solomon threatening to expel me as an IPCC reviewer.) On the empirical point of whether Hegerl et al used Yamal, given the explicit statement in Hegerl et al 2007, it is my opinion that Briffa is wrong.

Second, as regards d’Arrigo et al, Briffa said that Yamal was used, though “possibly labelled as Polar Urals”. To use a Mannian term, it is “disingenuous” to say that it was “possibly” labelled as Polar Urals. It is definitely labelled as Polar Urals. Briffa’s aware of the issue and has had an opportunity to check this out both by inspection of the article, where the labeling is evident and by asking the authors.

In my post on the impact, I identified the following reconstructions as ones which were “dependent” on Yamal, defining dependence as “equivalent calculation using plausible alternatives (e.g. Esper’s Polar Urals version instead of Briffa’s Yamal) yield different MWP-modern relationships”: Briffa 2000, the “closely related D’Arrigo et al 2006 and very recently, Kaufman et al 2009 (despite its first impression of a very different network)”.

Although the words “Polar Urals version” do not pass Briffa’s lips, he conceded the dependence in respect to Briffa 2000. He doesn’t consider Kaufman 2009, a recent reconstruction that we’ve discussed here and which can be seen to be dependent on Yamal in the above sense.

There is an empirical difference of opinion on the impact of Yamal on the D’Arrigo reconstruction. Here it is my opinion that use of Polar Urals rather than Yamal has a direct impact on the medieval-modern differential in the long D’Arrigo et al 2006 reconstruction. I’m not 100% sure of the size of the impact. I’ve been trying since 2005 to get the individual chronologies used in D’Arrigo et al 2006 in order to test precisely this sort of issue, but thus far have been unsuccessful. (I recently renewed this effort in the wake of present discussion.) The one that I am missing right now is their Coastal Alaska series. My surmise is based on the close relationship between the D’Arrigo series and the Briffa 2000 series. There are 6 series in the “long” D’Arrigo et al 2006 version (see their Figure 7) url are Yamal (labelled as “Polar Urals”), Tornetrask, Taimyr, Mongolia, Jasper (Icefields) and Jacoby-D’Arrigo’s Coastal Alaska. There are 7 series in the Briffa 2000 reconstruction: Yamal, Tornetrask, Taimyr, a shorter version of Mongolia (“Tarvagatory”), a slightly shorter version of Jasper (“Canadian Rockies”), the Jacoby-D’Arrigo NNorth America composite plus Yakutia(aka Indigirka River). My guess is that the it would alter the results materially, but I’m not 100% sure. Briffa guesses otherwise. We’ll see. (Obviously this shouldn’t be a matter of guessing. If D’Arrigo archived their data, then the guesswork would be eliminated.)

Briffa doesn’t mention strip bark bristlecones. Even the NAS Panel stated that strip bark should be “avoided” in temperature reconstructions. While these reconstructions are not affected by Yamal, neither can they be considered any longer to be relevant reconstructions without subtracting the strip bark series. In the case of MBH99, even Wahl and Ammann conceded that the MBH stick did not survive a sensitivity study without bristlecones. The IPCC did not squarely address the strip bark problem nor the sensitivity of multiple reconstructions to strip bark. Briffa should have addressed this point in his table. In my earlier post, I observed the following about this “family” of reconstructions (including, obviously MBH99).

One important “family” of spaghetti graph reconstructions are highly dependent on strip bark bristlecones/foxtails (a topic which has been much discussed here and elsewhere) but which do not use Yamal. These are “highly dependent” on strip bark bristlecones/foxtails in the sense that their methods do not yield a HS without them. Examples include MBH98-99, Crowley and Lowery 2000, Esper et al 2002 plus the re-statements of the MBH network in Rutherford et al 2005, Mann et al 2007 and Wahl and Ammann 2007.

In my earlier post, I pointed to recent reconstructions that used both strip bark (sometimes multiply) and Yamal as follows:

A third “family” of reconstructions wears both belt and braces – i.e. using both strip bark and Yamal. Key examples are Mann and Jones 2003, Mann et al (EOS 2003), Osborn and Briffa 2006, Hegerl et al 2007. The recent UNEP graphic uses the Mann and Jones 2003 version. A common stratagem in these studies is a leave one out sensitivity – where they show that they can “get” a similar result by leaving out any individual proxy. They can do so safely because they have both Yamal and bristlecones.

To this list, we can now add Tingley and Huybers (submitted Clim Chg). Briffa discusses the supposed “robustness” test of Osborn and Briffa 2006, the network used by Tingley and Huybers (submitted Clim Chg), upon which Clapton et al have already commented. In a network of only 14 proxies, Briffa used Yamal, Mann’s PC1 and strip bark foxtails as three of his 14 proxies. The trouble is that this sort of study has been data snooped: it doesn’t use Polar Urals, Indigirka River, Ababneh or Grudd’s Tornetrask.

In my earlier post, I pointed out that some spaghetti graph proxies didn’t go back to the MWP and were thus irrelevant to the MWP-modern comparison e.g. boreholes, Oerlemann, Briffa’s MXD reconstruction (the one where the “divergence problem” in the late 20th century was chopped off.) Briffa cites these as unaffected by Yamal, which is true but totally irrelevant to MWP-modern comparisons.

In my earlier post, I also noted that there were a few studies that were not materially affected by Yamal or strip bark bristlecones, but that these had their own problems, mentioning Jones et al 1998 and Moberg et al 2005 in this class.

There are a couple that are a bit sui generis, but these unfailingly have some serious problem. Jones et al 1998 uses neither Yamal nor bristlecones, but still has a slight modern-medieval differential. In its early portion, it uses only three series, two of which are early Briffa series (Tornetrask and Polar Urals pre-update). Both these series have serious problems – Briffa’s original Tornetrask series contains a gross manual adjustment to increase the 20th century relative to the MWP. See early CA posts on this.

Moberg uses both bristlecones and Yamal, but I view it as sui generis as well. Moberg used some unorthodox wavelet methods, that I’ve sort of emulated, but gave up trying to do so precisely. However, I can confirm that the bristlecone versions used in Moberg are not Graybill versions and don’t affect the result; they are merely fill. I’m not sure what impact Moberg’s filtering method will have on Yamal – I’ve not analyzed that in detail, but may do so some day. I’ve discussed Moberg problems in the past and, for present purposes, merely note that it is not a safe haven, but that it does not appear to stand or fall with Yamal and thus is not discussed further today.

I can guarantee 100% that the Jones et al 1998 reconstruction is materially affected by updates to Polar Urals (the Esper version) and updates to Tornetrask (Grudd). Briffa says that Moberg’s method removes all but the “high frequency” component of Yamal. No code is available for Moberg’s method; I can sort of emulate his results and will at some point examine the impact of Yamal. Moberg has some sui generis issues: e.g. its counterintuitive reliance on increased upwelling of cold water (subarctic G. Bulloides foraminifera in the Arabian Sea) as evidence of 20th century warming.

Revisiting Briffa’s table, I’ve added a column commenting on each reconstruction.

Study Briffa (2000) Yamal chronology  
Jones et al. (1998) Not used Briffa 1995 Polar Urals.
Mann et al. (1999) Not used Strip bark bristlecone/foxtail. Briffa 1995 Polar Urals.
Briffa et al. (2001) Not used Short reconstruction. Divergent portion truncated.
Esper et al. (2002) Not used Strip bark bristlecone/foxtail
Briffa (2000) Briffa (2000) Yamal was used Yamal
Mann and Jones (2003) Briffa (2000) Yamal was used in a composite of three ring-width chronologies from northern Eurasia Yamal and bristlecone PC1
Rutherford et al. (2005) Not used Strip bark bristlecone/foxtail
Moberg et al. (2005) Only high-frequency information from the Briffa (2000) Yamal chronology was used  
D’Arrigo et al. (2006) Briffa (2000) Yamal was used, though possibly labelled as Polar Urals Yamal.  It is definitely labelled as Polar Urals.
Hegerl et al. (2006) Not used "the west Siberia long composite involved Yamal and the west Urals composite."
Pollack and Smerdon (2004) Not used Short reconstruction
Oerlemans (2005) Not used Short reconstruction
Kaufman et al (2009)   Yamal
Mann et al (EOS 2003)   Yamal and bristlecone PC1
Osborn and Briffa 2006   Yamal and 2 strip bark series (including Mann’s PC1)
UNEP graphic (Mann and Jones 2003)   Yamal and bristlecone PC1

IPCC 2007
Briffa’s response provides a new bit of information about the IPCC spaghetti graph which even I was unaware of. He says:

In this analysis [IPCC 2007], the Yamal chronology was used cautiously because the series was truncated in 1985 for the purposes of constructing this Figure. Thus, the high recent values from Yamal were not shown in this Figure.

This truncation is nowhere mentioned in the graphic. Here is the figure showing the supposedly “cautious” use. The Yamal series is the one going off into the stratosphere at 4 sigma. To fully show Yamal in the version used by Kaufman, the height of the figure would have to be increased to 7 sigma! In my opinion, “cautious” use would mandate showing the actual data – all 6.97 sigma of it, so that readers could carefully consider the matter.

This is not the only instance of Briffa truncating data. As reported previously at CA, Briffa truncated the “divergent” portion of Briffa 2001 in IPCC TAR and, despite one IPCC reviewer insisting that this truncation not be repeated in IPCC AR4, did so once again.

 

Combining with Many Other Records

One of my major points of disagreement with Briffa and other Team authors is on whether “combining” Yamal and bristlecones with a lot of other records accomplishes the reduction in sensitivity that they assert using either CPS or Mannian methods. The graphic below, taken from a CA post here from a few years ago shows the impact of replacing all the “proxy” series in the MBH98 AD1400 network (other than Gaspe and the Mann PC1) by white noise. The reconstruction is virtually identical to the reconstruction with actual proxies. In actual reconstructions, most proxies are like the pills that Grace Slick’s mother gave her – they don’t do anything at all. In CPS and MBH-style methods is that the “white noise” “proxies” cancel out – the more “proxies”, the more effectively they cancel out under garden variety Central Limit Theorem considerations. In CPS and MBH methods, the average is not itself used. It is re-inflated to match the instrumental variance in the 20th century. If there are a couple of HS series lurking in the weeds (Yamal or the Mann PC1), they get re-inflated and you end up with the Mann PC1 plus a little static. That’s why the Mann reconstruction looks so much like the Mann PC1 (and the Graybill Sheep Mt series.) The same thing happens with CPS.

Thus, Briffa’s conclusion about the other studies: “they combine the Yamal record with many others and this reduces their sensitivity to the inclusion of any individual records” really doesn’t work as well he thinks. The “active ingredients” in study after study are strip bark and, secondarily, Yamal. Think of the screeching when sensitivities are done on MBH without strip bark – it’s as though the world had ended.


123 Comments

  1. Posted Oct 28, 2009 at 9:10 AM | Permalink

    You know, given that Briffa has had to respond to your silliness in the face of a serious illness. Given that there is very little left of your either your claim re Briffa’s cherry-picking or the claim that he hid data on you (which apparently you had sitting in a closet somewhere anyhow). And given that the various defamatory attacks on Briffa that were launched from the comments section of this blog, d’you not think that a simple “I’m sorry” is in order?

    Its never too late to do the right thing.

    • minimalist bender
      Posted Oct 28, 2009 at 9:13 AM | Permalink

      Re: bigcitylib (#1),
      Pay attention. There will next be some science that will be putting you in a position to apologize yourself.

    • jeff id
      Posted Oct 28, 2009 at 9:36 AM | Permalink

      Re: bigcitylib (#1),

      I fail to understand how accusations of data sorting are defamatory when it’s standard practice in the field.

    • frost
      Posted Oct 28, 2009 at 9:45 AM | Permalink

      Re: bigcitylib (#1),

      given that Briffa has had to respond to your silliness in the face of a serious illness

      And yet, Briffa did respond. Maybe it is not so silly after all.

    • Paul Penrose
      Posted Oct 28, 2009 at 10:59 AM | Permalink

      Re: bigcitylib (#1),
      Good way to lose a debate! I’m going to sit back and watch you shoot your other foot now.

  2. minimalist bender
    Posted Oct 28, 2009 at 9:11 AM | Permalink

    From Briffa’s response

    Briffa has also been attacked by McIntyre for not releasing the original ring-width measurement records from which the various chronologies discussed in Briffa (2000) and Briffa et al. (2008) were made.

    The primary criticism – and it was not an attack – is that the small sample size was not made known at the time of publication. Does CRU claim no freedom to release metadata? Then they are in no position to publish. Curious business model they seem to have. Shut it down.

  3. Clark
    Posted Oct 28, 2009 at 9:17 AM | Permalink

    I don’t understand how the last figure [mbh98_whitenoise.gif] doesn’t completely invalidate every study using CPS or MBH in the eyes of every climate scientist.

    If you can throw a single HS proxy into a ton of white noise samples and get the HS back out, that means your algorithm is completely discredited.

  4. minimalist bender
    Posted Oct 28, 2009 at 9:22 AM | Permalink

    Ah, the old “cautious use” criterion. Cautious for me, risky for you.
    .snip

  5. minimalist bender
    Posted Oct 28, 2009 at 9:31 AM | Permalink

    Briffa’s conclusion is exactly what I predicted could happen: he could pretend to salvage the hockey stick by dragging in other series to ressurrect the blade, effectively painting the Schweingruber sample as biased. (How could he have known this in 2000, with only twelve specimens?)
    .
    The cordial tone of the reponse aside, there is still a scientific pea-and-thimble game going on here, and McIntyre will get at the root of it. The clues to the game lie in the gentle admissions:

    The last 8 years of our chronology ARE based on data from a decreasing number of sites and trees and this smaller available sample does emphasise the faster growing trees, so this section of the chronology should be used cautiously. The reworked chronology, based on all of the currently available data is similar to our previously published versions of the Yamal chronology demonstrating that our earlier work presents a defensible and reasonable indication of tree growth changes during the 20th century, and in the context of long-term changes reconstructed over the last two millennia in the vicinity of the larch treeline in southern Yamal.
    .
    This does not mean that these chronologies will not change as additional data become available and as the RCS processing technique evolves, but the results we show here do suggest that McIntyre’s sensitivity analysis has little implication for those other proxy studies that make use of the published Yamal chronology data.
    .
    When using the RCS technique, it is important to examine the robustness of RCS chronologies, involving the type of sensitivity testing that McIntyre has undertaken and that we have shown in this example.

    IOW McIntyre, Jeff Id, et al are on the right track in scrutinizing RCS end-point issues when dealing with relatively small, heterogenous samples.
    .
    And Tom P? Out to lunch. Trying to sweep under the rug the problems that Briffa here admits are non-trivial.
    .
    And note the other pea-and-thimble game going on that Steve details above. Briffa asserts robustness of HIS work, but does not comment on Mann or Kaufman – which is why we went down the path to Yamal in the first place.

    • Tom P
      Posted Oct 28, 2009 at 11:25 AM | Permalink

      Re: minimalist bender (#6),

      Trying to sweep under the rug the problems that Briffa here admits are non-trivial.

      Not at all. Yamal ends up with five samples and Esper illustrated an example where below five samples the chronology is no longer robust. The divergence in the sensitivity test for Yamal I showed on the RCS thread is almost all in the final few years as was evident from the final offset in the chronologies. Here’s a close up of the unsmoothed RCS chronologies for the full Yamal series, without the three slowest growing trees and without the three fastest growing treees:

      The last decade is a region for which caution is warranted – in my sensitivity test the number of samples has fallen below five at the end. But in any event the chronology is higher here than at any time before the modern period, as my earlier plot showed.

      • Layman Lurker
        Posted Oct 28, 2009 at 11:57 AM | Permalink

        Re: Tom P (#34),

        Good graph Tom. It deals in part with the small sample size issue raised by Briffa. However, your plot makes an implicit assumption of population homogeniety. Something which minimalist bender reminds us of in his comment:

        IOW McIntyre, Jeff Id, et al are on the right track in scrutinizing RCS end-point issues when dealing with relatively small, heterogenous samples.

        • Tom P
          Posted Oct 28, 2009 at 12:14 PM | Permalink

          Re: Layman Lurker (#38),

          However, your plot makes an implicit assumption of population homogeniety.

          Actually it’s a sensitivity test of population homogeneity – I have removed the extreme growth samples to see how the chronology is affected. There’s not much difference apart from where the sample count falls very low anyway, indicating that sample heterogeneity is not a problem except for when the replication falls below five.

          If the CRU 12 population were heterogeneous with respect to the RCS reconstruction we would expect considerable differences to appear across the entire period by removing the top or bottom quartile of samples by growth rate.

        • ChrisZ
          Posted Oct 28, 2009 at 12:59 PM | Permalink

          Re: Tom P (#41),

          eyeballing your plot I’d say you have proved that this data set can be trusted up to 1950 when the problems with poor replication and ensuing divergence begin. I daresay Steve would be the first to agree that using the data this way (truncated in 1950 or thereabouts) is quite harmless. Would you be so kind as to suggest this to Briffa and your friends at RC?

        • Tom P
          Posted Oct 28, 2009 at 1:40 PM | Permalink

          Re: ChrisZ (#43),

          eyeballing your plot I’d say you have proved that this data set can be trusted up to 1950 when the problems with poor replication and ensuing divergence begin.

          No, the difference between the two sensitivity tests only becomes a significant fraction of the signal in the last decade or so:

        • Layman Lurker
          Posted Oct 28, 2009 at 1:10 PM | Permalink

          Re: Tom P (#41),

          To appreciate the full range of sensitivity wrt potential population heterogeneity, would you not need to add in the plots corrected by a separate growth function such as Roman’s loess fit: his comment #125..

        • Ron Cram
          Posted Oct 28, 2009 at 6:24 PM | Permalink

          Re: Layman Lurker (#44) (#46),

          I believe you were trying to link to roman’s headpost or to his comment #125.

          [RomanM: Thanks, Ron. I fixed the link.]

        • Layman Lurker
          Posted Oct 28, 2009 at 6:38 PM | Permalink

          Re: Ron Cram (#73),

          Thanks Roman.

  6. Posted Oct 28, 2009 at 9:34 AM | Permalink

    Thanks for this. I often wonder, what would we know without your tireless work.

    I noticed that Briffa apparently reads CA in that he noticed the Yamal/Ural substitution.

  7. Posted Oct 28, 2009 at 9:34 AM | Permalink

    snip – prohibited reference (religious)

  8. Posted Oct 28, 2009 at 9:51 AM | Permalink

    If by data sorting you mean cherry picking and are claiming that this is standard practice in the field–which is the claim made origonally made re Briffa and his team–then that is a pretty low remark as well and only justified if you swath yourself in the kind of conspiracy theories Steve generates with this blog.

    • jeff id
      Posted Oct 28, 2009 at 10:28 AM | Permalink

      Re: bigcitylib (#11),

      In my opinion any sorting I’ve seen so far is nothing more than cherry picking so there is no difference. However, there is a great deal of precedent in the field for this sorting of ‘temperature sensitive’ proxy.

      your claim re Briffa’s cherry-picking

      BTW: While I understand that you don’t like Steve or his work you do need to be specific when you make the statement that Steve made a claim. Since we all know he didn’t actually make this claim, it’s libelous to make untrue statements for the purpose of disparaging others. You are committing the injustice you accuse Steve of. I’m not going to write any more on this because it’s a waste of time.

  9. Andy
    Posted Oct 28, 2009 at 9:53 AM | Permalink

    snip – please don’t get distracted

  10. Posted Oct 28, 2009 at 10:00 AM | Permalink

    Isn’t Figure 2 from the first Briffa link fairly convincing, though? [1]

    They take out each individual proxy record one by one and it doesn’t affect the results. This strikes me as Briffa’s most important point, and I’d like someone to point me to a sufficient reply by McIntyre (if he’s written one). The reams of text above don’t seem to speak to it directly.

    Now, I suppose the argument is that 2 or 3 bad series, plus a statistical technique that inflates upward trends, would be impervious to a single proxy being removed. Is that the claim being made?

    • Steve McIntyre
      Posted Oct 28, 2009 at 10:08 AM | Permalink

      Re: Thras (#13),

      They contain two strip bark series and Yamal – thus belt and braces. You can take out any individual HS, because there is a fallback HS. If you make a biased MWP selection instead of the biased selection of Osborn and Briffa 2006 – Indigirka River, Polar Urals, Grudd Tornetrask,etc.. you can get different results. I’m not saying that one is “truer” than the other – only that things are much less robust than represented.

      • Posted Oct 28, 2009 at 10:25 AM | Permalink

        Re: Steve McIntyre (#14),

        “You can take out any individual HS, because there is a fallback HS.”

        And you are saying that any HS-shaped proxy will then be amplified by the statistical techniques used, I take it? Then it would seem that talking about cherry-picking data or the equivalent is more or less beside the point — a random selection of samples could conceivably always have one or two HS-shaped proxies.

        • Clark
          Posted Oct 28, 2009 at 10:37 AM | Permalink

          Re: Thras (#17),

          And you are saying that any HS-shaped proxy will then be amplified by the statistical techniques used, I take it?

          Look at the last figure in the post. 2 HS proxies + a bunch of white noise ‘proxies’ gives the same result as the two HS proxies plus many non-HS proxies.

          a random selection of samples could conceivably always have one or two HS-shaped proxies.

          Well, you’d have to exclude anti-HS proxies as well. You could do this either by:

          1. Requiring proxy correlation with the recent temperature record.
          2. Truncating the modern record.
          3. Inverting the series.
          4. Ignoring them entirely.

          I believe all four strategies have been documented here.

        • EddieO
          Posted Oct 28, 2009 at 10:44 AM | Permalink

          Re: Thras (#17), Yes but if the series only has a few samples there might not be a HS shaped sample but you can always choose another series that does. Hence the reason that the Polar Urals/ Yamal switch makes a difference.

        • Dave Dardinger
          Posted Oct 28, 2009 at 10:44 AM | Permalink

          Re: Thras (#17),

          a random selection of samples could conceivably always have one or two HS-shaped proxies.

          This was the problem with the original Mann off-center PCA method. Steve showed exactly what you say. Wegman and the NAS Panel (indirectly) agreed that the method wasn’t valid. But once it was known that certain proxies showed high correlation with the instrumental record, simply assuring that two or more of them are in any given reconstruction assures that weighing according to, or selecting via, correlation with the instrumental record will necessarily produce a HS. This is what makes BCLs complain so silly. As Steve shows in the head post, there is no reconstruction (with the exceptions noted)which does not use either strip bark or Yamal.

          What might be fun would be to assume a particular shape, say in the middle ages, for a “quasi-instrumental” period and then see what proxies would be selected or could be picked to prove this new shape. A square wave might be appropriate.

        • Al S.
          Posted Oct 28, 2009 at 11:11 AM | Permalink

          Re: Dave Dardinger (#26),
          JeffID’s Air Vent blog has several posts on hockey stick problems, showing that you can get almost arbitrary shapes from either noise/pseudodata or “real” tree-ring series (using Mann’s CPS technique, I believe).

        • Earle Williams
          Posted Oct 28, 2009 at 11:17 AM | Permalink

          Re: Al S. (#31),

          Thanks for mentioning that Al S. As I recall Jeff Id looked at red noise and red noise + “climate signal” and found that not only can you make any shape you want, you depress the non-random climate signal while accentuating the arbitrary pattern you are correlating against.

        • jae
          Posted Oct 28, 2009 at 12:56 PM | Permalink

          Re: Thras (#17),

          a random selection of samples could conceivably always have one or two HS-shaped proxies.

          Yes, but the researchers have an OBLIGATION to explain that these proxies represent the exception, not the rule. And then when there are so many studies that use the same few proxies, and when the authors claim “independence,” one has to wonder…

  11. Joe Black
    Posted Oct 28, 2009 at 10:11 AM | Permalink

    The way this is going, we soon may reach a state where the simplified, unified theory of many things expounded in California, Cassidy, Locke, Andes & Ferguson (1968) may come to include Treemometry.

    This work apparently was overlooked by the mainstream in the UK at the time of its original release.

  12. Charlie
    Posted Oct 28, 2009 at 10:24 AM | Permalink

    Steve McIntyre should put together a boilerplate template for articles like this. While the details have changed slightly, it seems like I’ve read this article before. In a more perfect world, SM would just have to point out the errors in one peer-reviewed article and future authors would avoid those mistakes.

    And where are the peer reviewers and magazine editors?

    Instead, time after time we see various combinations of

    1) data not available.
    1a) unavailability of data allows the scarcity of data to be hidden
    2) overall processing algorithm not disclosed
    3) unique modifications to particular datasets not disclosed (e.g. truncations, flipping the sign, filtering, etc.)
    4) apparently biased selection of which datasets to use. (Yamal vs Polar Urals, which version of Polar Urals, etc.)

  13. Charlie
    Posted Oct 28, 2009 at 10:25 AM | Permalink

    I forget an even more fundamental repetitive problem —-

    some sort of proof that the proxies actually have any sort of correlation with the metric they are supposedly a proxy for.

  14. Tomas E Rivas
    Posted Oct 28, 2009 at 10:36 AM | Permalink

    On the first comments of the Osborn and Briffa response on cautiously use of Yamal, they say:

    … the final years of the Yamal ring-width chronology (Briffa, 2000; Briffa et al., 2008) should be used cautiously on the basis that the values for the most recent part of this chronology are based on relatively few individual measurement series and this smaller available sample emphasises the faster growing trees. Despite this caveat, it should not be assumed that the Briffa (2000) and Briffa et al. (2008) chronologies are “wrong”: a reworked version of the Yamal chronology, using more data than were used in the published chronologies, shows essentially the same picture of comparatively high tree growth throughout most of the 20th century (i.e. 1901-1990).

    This means:
    1. The recent Yamal chronological data are from a small sample, with few individual measurements and faster growing trees… That’s not enough basis to eliminate Yamal data for any reconstrution?, since is not representative at all for temperature reconstructions, or just with a truncate Yamal series the data becomes representative?, I mean, the complete Yamal series are representative just in old data and not in recent data?… confusing for me.
    2. The chronologies from Yamal are not “wrong” by itself, but are wrong statistically, right?… how a small series can be reworked with more data proxies and becomes representative if you can not add data from Yamal itself?

    The faster growing signal from Yamal trees comes from a small sample, if you not add more data from Yamal (exclude proxies) how the signal becomes the right one?

    Yamal trees, are a faster growing trees or trees that have a high tree growth?… what’s the difference?… it’s just my poor english?

  15. Ian L. McQueen
    Posted Oct 28, 2009 at 10:41 AM | Permalink

    Small typo needs correction: “Briffa 200″”

    Yamal became important as an issue because it is influential in reconstructions that do not use bristlecones (Briffa 200, D’Arrigo 2006, Kaufman 2009).

  16. hmmm
    Posted Oct 28, 2009 at 10:43 AM | Permalink

    Also I would note there is still no reference to the reasoning and impact of using only Yamal and not Polar Urals. What is the impact of replacing Yamal with Urals? Or what is the impact of combining them if that is supposedly such a great way to reduce sensitivity? What would the impact me on the error bars/confidence levels? Why is the Urals data thrown in the garbage? There might be a great answer but it would be useful to have it available, no?

    bigcitylib- Pathos is not effective in a technical debate. I suggest you consider a little more logos.

  17. Posted Oct 28, 2009 at 10:44 AM | Permalink

    I was thinking of his remark on September 26th:

    Unfortunately, to date, people in the field have not honored this responsibility and, to an outside observer, seem to have done no more than pick the version (Yamal) that suits their bias.

    • hmmm
      Posted Oct 28, 2009 at 11:34 AM | Permalink

      Re: bigcitylib (#24),

      Bigcitylib,
      Your interpretation of that sentence is definitely off.

      All this says is that if the selection criteria isn’t provided and valid, outsiders who aren’t privy to the reasoning involved will talk about cherry picking and question the selection. Unfortunately, nobody who can and should has provided that reasoning.

      Saying that their silence on the subject will make people question their selection process is a good point and is not an accusation that they did anything wrong.

      Steve’s making a case for transparency and replication in science. He clearly didn’t accuse anyone, he just pointed out one of the obvious side-effects of continuing business as usual in the climate science world.

      This was obviously a good point Steve made, since here we are talking about it. A simple standardized scientific explanation on why Yamal was used and Polar Urals wasn’t would end most of this speculation. Don’t you agree?

      So where is that answer?

      • GTFrank
        Posted Oct 28, 2009 at 1:30 PM | Permalink

        Re: hmmm (#36),

        Yes, it seems to be a common misconception about Steve’s position. At Delayed Oscillator the other day, DO quipped that McIntyre would not being happy with one of DO’s examples because he was not “using all of the data”
        Steve has said many times that not using all of the data is ok. Making choices of data is ok, also. Just explain why some is used, and some not. Figure out the riddle of the tree rings. That’s where the real mystery lies. That’s where the intense scientific effort is needed. And then explain it to the number crunchers. (I’m probably offending a bunch of folks here)

        Not to diminish their obvious skills and attention to detail – as Steve and others have demonstrated, running the raw data through various statistical analysis and creating graphs is just not that “special”. Lots of folks have those capabilities.

  18. nanny_govt_sucks
    Posted Oct 28, 2009 at 10:51 AM | Permalink

    Thras,

    They take out each individual proxy record one by one and it doesn’t affect the results.

    What they didn’t do was then take out each combination of 2 proxies and see if the results were affected, then each combination of 3 proxies, then each combination of 4, etc…

    snip – over-editorializing

  19. bernie
    Posted Oct 28, 2009 at 10:51 AM | Permalink

    Steve:
    The table of studies with your comments is very helpful.

  20. Posted Oct 28, 2009 at 10:57 AM | Permalink

    Briffa’s 1 s.e. bands (in gray) for “Yamal All” in Figure E of http://www.cru.uea.ac.uk/cru/people/briffa/yamal2009/sensit.htm seem to indicate that no period was anomalous (relative to a 68% CI) except the vicinity of the “year with no summer” of 1816.

    I’m not sure how these were computed, but they seem to indicate that the fluctuations are almost all just noise.

    BTW, I’m not sure which sites are included in the Polar Urals composite. Is Briffa including them in his “Yamal All” series, or are they excluded? If the latter, why?

  21. AnonyMoose
    Posted Oct 28, 2009 at 11:31 AM | Permalink

    The phrase “url are Yamal” is not good. Better is something such as “(pdf url), these are Yamal”.

  22. Edward
    Posted Oct 28, 2009 at 11:35 AM | Permalink

    Charlie #16
    You ask “where are the peer reviewers and magazine editors” in relation to four problems. The crux of the debate is that the reseachers, reviewers and editors do not feel that: 1)data is being “hidden” or that hiding it is even an issue, 2) that disclosure of unique processing or modifications to datasets are necessary or required or 3) that the method of dataset selection is all that biased.

    All that matters is that the results conform to a concensus view of the temperature record even if their methods are indicted years later. By the time skeptics figure out what was done, new methods will have already been developed to allow the publication of a new study that confirms the old results and the debate begins anew and will go on and on. There is a continual moving target that skeptics will never be able to catch up to no matter how hard we watch the “pea in the thimble”.

    Shiny
    Edward

  23. BDAABAT
    Posted Oct 28, 2009 at 12:01 PM | Permalink

    BCL: Interesting response.

    1. BCL asked: Does Biffra have to respond to silliness when he’s sick? I guess it’s not “silliness” if he takes the time to read this blog and feel sufficiently motivated to comment. Note also that Biffra’s comments don’t address the real issues Steve raised.

    2. BCL asked: Is there little left to your (meaning Steve’s) claim of Biffra’s cherry picking? Steve never made such a claim. The fact that Biffra withheld data and methods allowed other people to insert possible reasons for withholding of data. But that’s not a claim that Steve made.

    3. BCL stated that Steve accused Biffra of hiding data and BCL said this wasn’t true because Steve had it in a closet. Biffra didn’t hide the data… he never provided it when requested. This is demonstrably true. Steve repeatedly asked for the data and was denied by Biffra. Steve asked for it as an individual researcher as well as in his capacity as IPCC reviewer. It was never provided by Biffra. It would likely never have been provided had not the Phil Trans B (Roy Soc) actually followed through on their data archiving policies and forced Biffra to do what was required as part of the publication process. The fact that Steve had acquired a copy of the data (without knowing that it was actually the data he had been trying to get) is irrelevant. Biffra did withhold data from Steve and from other researchers.

    An, “I’m sorry!”, is in order, but not from Steve. This entire issue could have been addressed had Biffra provided the information in his publications. It could have been addressed had Biffra archived his data. There have been numerous failures that have been identified in this entire exchange, none of which appear to have been committed by Steve.

    Bruce

  24. Stephen Parrish
    Posted Oct 28, 2009 at 12:26 PM | Permalink

    Big City Lib of Shagged Sheep Fame! Good to see you again!

  25. Sean Houlihane
    Posted Oct 28, 2009 at 1:14 PM | Permalink

    The focus still seems to be on the recent couple of decades. Doesn’t this recent divergence only serve to emphasise that the historical part of the reconstruction could quite plausibly demonstrate similar uncertainty, regardless of the correctness of selecting specific series based on correlation with modern records?

  26. Layman Lurker
    Posted Oct 28, 2009 at 1:16 PM | Permalink

    I must have done something wrong trying to link back to Roman’s growth function in the comments of the “RCS – one size fits all” thread.

    [RomanM: Fixed]

  27. henry
    Posted Oct 28, 2009 at 1:17 PM | Permalink

    Says they ran the test (removal of proxies) but only showed the chart with one at a time bieng taken out.

    Our use of smoothed (i.e., low-pass filtered) series could, however, spread the influence of extreme values that are short-lived or occur near the ends of the series over a broader range of years. We undertook additional sensitivity tests, therefore, to assess the impact on our results of completely excluding one, two or three out of the 14 proxy series (using all possible combinations of which proxies to exclude).

    Where can we see 1) list of all proxies retained, and 2) charts showing the removal of more than 1.

    Would be interesting to see which series did remain after the exclusion of one, two or three out of the 14 proxy series.

    With 14 available series, there had to have been one series run that removed all the HS series. THAT’S the one I’d like to see.

    Did they only stop at removal of three because there are 4 HS series?

  28. henry
    Posted Oct 28, 2009 at 1:38 PM | Permalink

    Maybe we should revisit the CA post of Feb 10, 06 to see how robust the removal of up to three proxies is:

    Readers of this blog will be familiar with certain proxies which I’ve criticized both individually and in their repeated use. So I can’t be accused of making an ad hoc response if I criticize them one more time. There are 8 proxies out off 14 with strongly elevated 20th century values that contribute the hockey stickness to O&B: the Mann PC1 (Sheep Mountain); the adjacent foxtails; Jacoby’s Mongolian site; Briffa’s Tornetrask and Briffa’s re-processed Yamal; Thompson’s Dunde and Guliya (expressed through the Yang composite); the van Engeln documentary series and Wilson’s Alberta series. There are 6 with not anomalous 20th century results: Chesapeake Bay Mg/Ca; Greenland dO18; Tirol; Mangazeja; Taimyr; and the Quebec series (although a longer version of the Quebec series does have higher values).

  29. Erasmus de Frigid
    Posted Oct 28, 2009 at 1:50 PM | Permalink

    I have read Grudds Tornetrask report and it looks ok, wonder why its
    not included in the proxy studies?

  30. BDAABAT
    Posted Oct 28, 2009 at 2:00 PM | Permalink

    LL #39: Thanks for the correction! Did I mention I’m a supporter of the “DAM” team: Mothers Against Dyslexia? :D

    Bruce

  31. Craig Loehle
    Posted Oct 28, 2009 at 2:06 PM | Permalink

    Nice summary table. Briffa tries to handwave that “it doesn’t matter” but should know better–Steve knows these studies. IF (and only if) these reconstructions were simple averages, then one polluted study would not matter. But they mostly use methods that weight certain series heavily, effectively data mining as steve showed in his last figure. Quite clear.

  32. David
    Posted Oct 28, 2009 at 2:08 PM | Permalink

    TomP, no matter how consistent the 10 tree cores are amongst themselves over the 20th C, the spike they result in is not consistent with temperatures from the area. Whatever was responsible for the spike, it was not temperature.

    So aside from the statistical issues with the low sample size (which was not reported at the time), they are simply not good proxies for temperature. Reconstructions depending on them are depending on statistically invalid samples which are ANYWAY not doing a good job of representing temperatures.

    If you ask me, a plausible reason for the spike is a side effect of the Tsar Bomb in 1961 http://en.wikipedia.org/wiki/Tsar_Bomba or severe pollution in the area (“industrial production pollute air, water, and land to an almost unimaginable extent”, http://arcticcircle.uconn.edu/NatResources/gasdev.html ) causing the death of less well established plantlife which competed for soil resources but which had a higher percentage dosage of pollutants taken from topsoil, rather than the deep roots of the established trees.

    Something does seem to have caused a spike for long life trees starting in the 60’s (unless its a statistical aberration) but since the local temperature has not risen noticeably it must have been something else.

    Steve: There’s little purpose in discussing this sort of hyopthesis on this thread. There are lots of issues to deal with in Briffa’s response, without getting sidertracked/

    • Tom P
      Posted Oct 28, 2009 at 2:50 PM | Permalink

      Re: David (#54),

      TomP, no matter how consistent the 10 tree cores are amongst themselves over the 20th C, the spike they result in is not consistent with temperatures from the area.

      No, the Yamal chronology is well correlated (r > 0.4) with temperature for the growth season over the period from 1950 to 1994: http://rstb.royalsocietypublishing.org/content/363/1501/2269.full.

      If you ask me, a plausible reason for the spike is a side effect of the Tsar Bomb in 1961.

      Interesting hypothesis. There is a common signal evident in Briffa’s plot of the chronology of the region’s subpopulations in the mid sixties, but nothing sustained:

      http://www.cru.uea.ac.uk/cru/people/briffa/yamal2009/fig/figC.pdf

      • Jon
        Posted Oct 28, 2009 at 3:40 PM | Permalink

        Re: Tom P (#58),

        In reference to the Tsar Bomba thing – I think it is much more likely that nitrogen in the form of acid rain or refinery emission of ammonia would cause spikes in growth, rather than radiation from nuclear tests or radioactive materials. Nitrogen actually makes plants grow, after all.

      • MikeN
        Posted Oct 28, 2009 at 8:51 PM | Permalink

        Re: Tom P (#58), Schweingruber’s Yamal chronology is also well correlated, a bit better than Yamal.

        • MikeN
          Posted Oct 28, 2009 at 9:11 PM | Permalink

          Re: MikeN (#82), the point is conceded in the note:

          Judged according to this criterion it is entirely appropriate to include the data from the KHAD site (used in McIntyre’s sensitivity test) when constructing a regional chronology for the area.

  33. Vg
    Posted Oct 28, 2009 at 2:28 PM | Permalink

    I hope I wont be snipped as I was a ice posting re BDAABAT who has editorialized at length about this whole issue here. I find the response by Briffa to be very salutory and a civilized way to deal with this issue rather than the usual RC way. ie at least they have taken this issue very seriously and departed from the usual “you haven’t published so we wont listen to you”

  34. Posted Oct 28, 2009 at 2:35 PM | Permalink

    “Oerlemann”
    Freudian slip Steve?
    Hans Oerlemans knows where to stop with his reconstruction. He hates the current climate hype, he is a scientist.

  35. David
    Posted Oct 28, 2009 at 2:37 PM | Permalink

    One wonders what the effect of over 1500–3500 Bq/kg is on topsoil plant life.

  36. David
    Posted Oct 28, 2009 at 2:57 PM | Permalink

    TomP
    “No, the Yamal chronology is well correlated (r > 0.4) with temperature for the growth season over the period from 1950 to 1994″

    Isn’t that the famous non-verified correlation where anyone else who tries it gets very low correlation, but mysteriously the Polar Urals gives around 0.4 leading to the assumption that somewhere in the paper the Polar Urals correlation was accidentally inserted instead of Yamal.

    You tell me, perhaps YOU can replicate the calculation and explain to the rest of us where we are going wrong ?

    The Team are mysteriously silent on it, and noone seems to be pressing them for an explanation (other than Steve).

    • Tom P
      Posted Oct 28, 2009 at 3:19 PM | Permalink

      Re: David (#59),

      Isn’t that the famous non-verified correlation where anyone else who tries it gets very low correlation, but mysteriously the Polar Urals gives around 0.4 leading to the assumption that somewhere in the paper the Polar Urals correlation was accidentally inserted instead of Yamal.

      You’re very confused. Steve McIntyre himself presented the following numbers last week at http://www.climateaudit.org/?p=7374:

      Both chronologies have statistically significant relationships to June-July temperature, but the t-statistic for Polar Urals is a bit higher (Polar Urals t-statistic – 5.90; Yamal 4.29; correlations are Polar Urals 0.50; Yamal 0.55).

      • Pat Frank
        Posted Oct 28, 2009 at 7:28 PM | Permalink

        Re: Tom P (#61), You must have missed this at the link you cited, Tom: “The Polar Urals chronology has a statistically significant relationship to annual temperature of the corresponding HadCRU/CRUTEM gridcell, while Yamal does not (Polar Urals t-statistic – 3.37; Yamal 0.92). For reference the correlation of the Polar Urals chronology to annual temperature is 0.31 (Yamal: 0.14).
        .
        It was immediately above the text you chose to quote. Guess you were distracted.
        .
        Steve went on: “Both chronologies have statistically significant relationships to June-July temperature, but the t-statistic for Polar Urals is a bit higher (Polar Urals t-statistic – 5.90; Yamal 4.29; correlations are Polar Urals 0.50; Yamal 0.55). Any practising statistician would take the position that the t-statistic, which takes into consideration the number of measurements, is the relevant measure of statistical significance, a point known since the early 20th century.
        .
        Interesting that the issue is annual average temperature, but when convenient we can focus opportunistically on June-July temperature.
        .
        Steve’s final point regarding the Yamal-Polar Urals dichotomy is the most revealing: “Thus, both chronologies have a “statistically significant” correlation to summer temperature while being inconsistent in their medieval-modern relationship.
        .
        That’s the ball game. Use Polar Urals, and the MWP trend is higher than the 20th century. Use Yamal, and the MWP is ‘correctly’ lower. And that’s today’s Q.E.D., as Lubos might write.

  37. MattN
    Posted Oct 28, 2009 at 3:02 PM | Permalink

    In summary: It doesn’t matter…..

  38. Hoi Polloi
    Posted Oct 28, 2009 at 3:21 PM | Permalink

    Sometimes I wonder if some AGW scientists are deliberately withholding data and/or comments to Steve’s articles. Inevitably Steve’s articles create a lot of unnecessary speculating, hype, straw men and other white noise (like McSteve would have accused Briffa of cherry picking) which conveniently can be adressed instead of adressing the real issue. Some self discipline would be welcome here methinks…

    • bender
      Posted Oct 28, 2009 at 8:28 PM | Permalink

      Re: Hoi Polloi (#62),
      That’s good advice. Unfortunately the blogosphere never sleeps and there is always somebody anxiously wanting the next card to be played.

  39. Carrick
    Posted Oct 28, 2009 at 3:48 PM | Permalink

    Tom P:

    Both chronologies have statistically significant relationships to June-July temperature, but the t-statistic for Polar Urals is a bit higher (Polar Urals t-statistic – 5.90; Yamal 4.29; correlations are Polar Urals 0.50; Yamal 0.55).

    Now I’m the one confused.

    Since when is correlating to June/July temperatures helpful, if what is needed is regional annual temperature?

    • EW
      Posted Oct 28, 2009 at 4:29 PM | Permalink

      Re: Carrick (#67),

      The Yamal larches grow only in June-July. And even during that period there could be abnormal tree-rings caused by low temperatures or “missing rings” due to cold summer. It was rather interesting reading about the extreme conditions they live in. For e.g., they can grow only in the river valleys, where they are a bit protected.

  40. David
    Posted Oct 28, 2009 at 3:53 PM | Permalink

    TomP
    Aha! You caught me by saying “growth season” while I automatically interpreted “plain yearly temperature”.

    So now you are holding a statistically insignificant number of samples which have a moderate correlation to the two months of June and July, and poor correlation to yearly temperatures.

    Does that really make you think you have a good proxy for annual temperature fluctuations ?

    Does a hotter than average June/July mean hotter than average yearly overall temperature ? Apparently for The Team supporters, it does.

  41. Posted Oct 28, 2009 at 4:01 PM | Permalink

    The raw data now takes YAD061 up to 1996 rather than stopping in 1995, so now there really are 5 samples in 1996 rather than 4.

    Latest version of Yamal AD ring raw data:
    YAD061 1990 1610 1580 1710 2870 1800 1720 2520 -9999

    Earlier version of the data:
    YAD061 1990 1610 1580 1710 2870 1800 1720 -9999

    Was is Romanm who was thinking that the previously released data couldn’t be what was used in the chronology?

    Is there a quick and painless way of matching the two data sets to see if there are other differences?

  42. EW
    Posted Oct 28, 2009 at 4:34 PM | Permalink

    On a second thought – Hantemirov seems to stress in his Thesis abstract, that all these data and reconstructions concern summer temperatures at Yamal. He doesn’t mention yearly temperatures or temperatures in general ;-)

  43. minimalist bender
    Posted Oct 28, 2009 at 5:44 PM | Permalink

    Did Tom P just dodge Layman Lurker’s #44? Because LL’s got it. If it was a dodge, can we ask why?

  44. Posted Oct 28, 2009 at 8:06 PM | Permalink

    Timothy J. Osborn is the main author of that note and has been mentioned 8 times, whereas Briffa has been mentioned 96 times to date.

  45. Posted Oct 28, 2009 at 8:10 PM | Permalink

    Timothy J. Osborn is the main author of that note; he has been mentioned 8 times to date in this thread, whereas Briffa has been mentioned 96 times.

    • bender
      Posted Oct 28, 2009 at 8:13 PM | Permalink

      Re: willard (#76),
      I was just getting to that. See my crosspost.

  46. bender
    Posted Oct 28, 2009 at 8:13 PM | Permalink

    In the first link above, Osborne writes:

    In a related comment, we noted that the final years of the Yamal ring-width chronology (Briffa, 2000; Briffa et al., 2008) should be used cautiously on the basis that the values for the most recent part of this chronology are based on relatively few individual measurement series and this smaller available sample emphasises the faster growing trees. Despite this caveat, it should not be assumed that the Briffa (2000) and Briffa et al. (2008) chronologies are “wrong”

    The issue was never whether the chronology was “wrong”. The issues were (1) whether it was based on a biased sample and (2) whether there were critical biases introduced by the use of the RCS method on a limited sample of heteregoneous sequences that had an impact in subsequent studies that used it.
    .
    The second point is a mouthful, but it is very important to understand this. Because the MWP is fairly close to the CWP (in both Yamal and Polar Urals) it does not take much of a bias to push the CWP the tiniest fraction warmer than the MWP, which then allows one to just barely conclude “modern temperatures are unprecedented”. So we are not searching for gross errors. We are searching for small biases (and significant cover-ups of the origins of these small biases) just large enough to make the difference between a run-of-the-mill reconstruction in a third-tier journal versus an OMGIWTWT in a top-tier journal such as Nature or Science. So first point: small biases in the warm direction are what matter.
    .
    Second point: whether these biases “matter” can only be judged by examining the multi-proxy reconstructions in which the biased data are embedded. To argue they’re not “wrong” is to erect a strawman. I repeat, the issue is not (so much) the quality of the CRU Yamal RCS chronology so much as its relative importance in multiproxy studies. It is very significant that neither Melvin nor Osborne in these two notes are willing to touch this issue with a ten-foot-pole. You don’t hear them defending Mann’s or Kaufman’s use of Yamal. You hear them saying “we told you it should be used cautiously”. Caveat emptor. That silence rings loud. The repeated cautionary note rings clear.

    • Tom P
      Posted Oct 29, 2009 at 7:58 AM | Permalink

      Re: bender (#76),

      So first point: small biases in the warm direction are what matter.

      Are you sure you really mean this? In general any conclusion that is sensitive to small changes in the result has to be rather tentative – there can’t be a sharp threshold for significance in what is a noisy signal. You’re overemphasising the exact shape of these chronologies here.

      Re: MikeN (#82),

      Schweingruber’s Yamal chronology is also well correlated, a bit better than Yamal.

      Thanks for the information. Could you point me to where these numbers are given?

      • Erasmus de Frigid
        Posted Oct 29, 2009 at 8:06 AM | Permalink

        Re: Tom P (#97),

        Just do a search here on the site for Schweingruber and Yamal, I believe Steve’s 9-26-09 post showed the
        difference between the larger Schweingruber data set and the Briffa, which selected only 12 trees.

        In the Gaspe vs AD1400 graphs above, looks like the Team has developed some Radar software for picking
        small aircraft signatures out of deep noise. Wonder if they can sell this to Raytheon?

        • Tom P
          Posted Oct 29, 2009 at 8:29 AM | Permalink

          Re: Erasmus de Frigid (#99),

          Just do a search here on the site for Schweingruber and Yamal

          I’ve already searched the site, without success. I’d prefer a location to a guess as to where the correlation between Khadyta and local temperature is to be found.

        • bender
          Posted Oct 29, 2009 at 8:44 AM | Permalink

          Re: Tom P (#100),
          My advice is to … READ THE BLOG.

        • steven mosher
          Posted Oct 30, 2009 at 2:34 AM | Permalink

          Re: Erasmus de Frigid (#99), its UWBBS radar. ultra wide band bull s**t

  47. bender
    Posted Oct 28, 2009 at 8:36 PM | Permalink

    The leave-one-out test done by Osborn is not at all convincing. If you have a heterogenous set of samples that interact poorly with the RCS method you need to apply a leave-some-out test.
    .
    And here is a noteworthy Team double-standard. When you’re hunting for correlations that boost the evidence in favor of your hypothesis it is ok to pick-two (or pick m, if you recall propick.m), but when you’re doing a robustness sensitivity test against the hypothesis, just leaving one out is adequate. That one is going in the database. (kim collects AGW ironies. I collect AGW double-standards.)

    • kim
      Posted Oct 29, 2009 at 7:16 AM | Permalink

      Re: bender (#80),

      I got BigCityLib’s comment #1 stuffed into the collection cabinet, but when I closed the door the hinges popped off.
      =================================

  48. Posted Oct 28, 2009 at 8:40 PM | Permalink

    How could a chronology not be wrong, if based on (however minutely) biased samples?

    • bender
      Posted Oct 28, 2009 at 9:03 PM | Permalink

      Re: willard (#81),
      Because all chronologies are approximations and right & wrong are just too categorical when you’re working with approximations.

  49. MikeN
    Posted Oct 28, 2009 at 8:54 PM | Permalink

    Did B&O really show 391 graphs eliminating any up to 3 proxies?

    • henry
      Posted Oct 28, 2009 at 9:46 PM | Permalink

      Re: MikeN (#83),

      Even if they had removed up to 3 proxies, the image probably wouldn’t have changed.

      Remember:

      There are 8 proxies out off 14 with strongly elevated 20th century values that contribute the hockey stickness to O&B

  50. MikeN
    Posted Oct 28, 2009 at 8:56 PM | Permalink

    Why refer to Clapton et al, especially with the young author taking comments?

  51. Posted Oct 28, 2009 at 9:44 PM | Permalink

    Mr. Bending Rodríguez,

    “Critical bias” is categorical enough to imply badness, from which wrongness is so close a strawman could not fit in between.

    • bender
      Posted Oct 28, 2009 at 9:50 PM | Permalink

      Re: willard (#87),
      That’s an opinion. I’m not here to debate opinions.

      • steven mosher
        Posted Oct 30, 2009 at 2:35 AM | Permalink

        Re: bender (#89), ya leave all the hard work to me, slacker.

  52. Raven
    Posted Oct 29, 2009 at 3:53 AM | Permalink

    Has Steve posted a response to Briffa’s claim that he had no obligation to provide anything more than the composite because the source data was not his? Briffa also claimed that Steve is wrong to claim that Science and Nature’s policies actually required Briffa to provide the source data.

    I remember reading a post on this but cannot find it now.

    • pinkr
      Posted Oct 29, 2009 at 6:43 AM | Permalink

      Re: Raven (#91),

      Nature policy

      http://www.nature.com/authors/editorial_policies/availability.html

      Science policy

      http://www.sciencemag.org/about/authors/prep/gen_info.dtl#dataavail

      • Raven
        Posted Oct 29, 2009 at 7:07 AM | Permalink

        Re: pinkr (#93)
        Thanks. It appears that Science provides a loophoop by allowing for “discipline-specific” exceptions to apply. The NAS report referenced by Science also provides a loop hole through the use of the words “other researchers” and “encourages”. Nature’s policy is more definitive but Briffa appears to claim that once the composite was published in Science then that is all that needed to be provided to support the Nature paper.
        .
        I can see how the spirit of the policies was violated by Briffa but it is not clear that the letter was violated. Am I missing something?

        • kim
          Posted Oct 29, 2009 at 7:24 AM | Permalink

          Re: Raven (#94),

          Raven, it’s the spirit that knocked the hinges off; you can always stuff another letter in.
          ======================================

  53. Posted Oct 29, 2009 at 6:06 AM | Permalink

    Bigcitylib #1 If scientific objectivity had been followed, we wouldn’t be discussing Yamal and Hockey Sticks during Briffa’s unfortunate illness. A lack of objectivity, a lack data sharing/archiving, and the promotion of one’s own work and perspctive as an IPCC lead author are some of the reasons have brought us here. Scientific sins not restricted to Briffa – the others know who they are. Not too late to do the right thing.

  54. theduke
    Posted Oct 29, 2009 at 8:03 AM | Permalink

    Bender: thanks for the clarity in 76 and 80. For those of us with small latin and less greek, it’s helpful.

  55. MikeN
    Posted Oct 29, 2009 at 9:40 AM | Permalink

    Tom, I posted a correlation comparison. I don’t remember what thread it was on. Also, see the quote from the note above. The authors concede Schweingruber was valid.

    • Tom P
      Posted Oct 29, 2009 at 4:49 PM | Permalink

      Re: MikeN (#102),

      You refer to the correlation on the Yamal Divergence thread, but as for an actual number or calculation, it’s nowhere to be seen.

      • steven mosher
        Posted Oct 30, 2009 at 2:36 AM | Permalink

        Re: Tom P (#104), u know R knock yourself out.

  56. Jimmy Haigh
    Posted Oct 29, 2009 at 3:40 PM | Permalink

    I was thinking about the divergence problem over a beer. Could it be that the increased growth in tree rings, which has been seen in the last century or so and which causes the divergence problem, be due to an increase in atmospheric CO2? Would that not be a beautiful irony?

  57. bender
    Posted Oct 29, 2009 at 10:14 PM | Permalink

    New techical note from Briffa’s lab:

    http://www.cru.uea.ac.uk/cru/people/briffa/yamal2009/sensit.htm

    • JS
      Posted Oct 30, 2009 at 2:13 AM | Permalink

      Re: bender (#105), very interesting. Compare his results with mine in the RCS thread (see comment #381 (for a picture) and #400). What I did is allow for a different age profile for YAD, POR and JAH estimating the climate effects using a random effects model which simultaneously calculates an age curve (with no functional form restrictions – it individually estimates a point for every age) for both sub-fossil and live trees as well as the climate profile. I did not add KHAD in. (You can read about this area of panel data analysis here and think of trees instead of people – or just use Google to search for “random effects model” or “panel data” if anyone is unfamiliar with my jargon.)

      Of note is that the standard errors on any of these estimates in the 20th century are pretty wide. I suppose that is something that isn’t presented in Biffra’s note but comes out of the random effects estimation I’ve done.

      NB The picture you see, while technically for a slightly different test, is pretty much the same as you get using the full procedure described in my comment #400 (the actual results are inaccessible to me until Monday because they are on a different computer). My results show the current peak (on a 21 year centred moving average basis) is about equal to the peaks in 200 and 1000 which seems different to the results presented by Biffra. This is not adding any new data at all, this is just allowing for a different age profile.

  58. MikeN
    Posted Oct 29, 2009 at 10:17 PM | Permalink

    OK, I was planning to do a bigger comparison with all the possible variations, so maybe I left out the number, or maybe I had already done it on another thread. Now I don’t see the point since the authors concede the point.

  59. bender
    Posted Oct 29, 2009 at 10:18 PM | Permalink

    The 3 HS series (YAD, POR, JAH) are clustered in one tight area spanning just 20km x 80km.

  60. bender
    Posted Oct 30, 2009 at 6:45 AM | Permalink

    Briffa’s most important statement, in his conclusions:

    McIntyre’s sensitivity analysis has little implication for those other proxy studies that make use of the published Yamal chronology data.

  61. steven mosher
    Posted Oct 30, 2009 at 3:31 PM | Permalink

    I has questions Dr. Bender. Here are the gross stats for the yamal 10.

    start end age RW st.dev corr. rbar
    CRN
    YAD011 1947 1996 50 1.94 0.79 0.61 0.52
    YAD021 1930 1996 67 1.73 0.64 0.76 0.65
    YAD031 1808 1996 189 0.67 0.47 0.63 0.59
    YAD041 1803 1996 194 0.66 0.45 0.59 0.58
    YAD061 1803 1996 194 0.70 0.63 0.50 0.58
    YAD071 1849 1996 148 0.51 0.35 0.63 0.61
    YAD081 1875 1996 122 1.22 0.72 0.66 0.62
    YAD091 1906 1996 91 1.21 0.54 0.72 0.61
    YAD101 1906 1996 91 1.49 0.60 0.80 0.66
    YAD121 1805 1996 192 0.54 0.44 0.61 0.61

    Just as a sample this kinda bothers me. All living, 4 of the specimin are within what one could arguably call
    the early growth period ( say less than 100 years) 50% of them starting in the industrial period. plus those starting
    in 1900 to 1940 “natural” rise in temperature..

    So, Im wondering how one decides which other samples need to be added to this one. Simply. the Yamal 10 are
    unique from a sampling perspective, if you want to use some “signal based” measure to make a homogenity argument don’t you have a problem there realted to the quality of the signal you can get from 10 samples ( thinking type 2 error)

    Or am being stupid again?

    heck its 2009. 13 more rings on those suckers sure would be fun to look at.

  62. Sean
    Posted Oct 30, 2009 at 5:19 PM | Permalink

    It’s hard to read the spaghetti graph, but there does not appear to be any line representing any sort of average of all the series; the chart just has individual lines for the individual proxies. If that’s the case, would there be any difference between Briffa’s statement that:

    “In this analysis [IPCC 2007], the Yamal chronology was used cautiously because the series was truncated in 1985 for the purposes of constructing this Figure. Thus, the high recent values from Yamal were not shown in this Figure.”

    And simply saying:

    The line reresenting the Yamal chronology goes off the chart in 1985.

    ?

    The latter point is obvious just by looking at the thing.

    Is there anything cautious about lines that literally go off the charts?

  63. Posted Oct 31, 2009 at 2:00 AM | Permalink

    Deep Climate comments on the Briffa response.

    • Michael Jennings
      Posted Oct 31, 2009 at 8:53 AM | Permalink

      Re: Bishop Hill (#119),

      Bishop, I am surprised you would even bother to link that tripe here (except for educational purposes of what NOT to do). It consists entirely of distortions, lies, and ad hom attacks and serves no useful purpose in furthering dialogue.

      • Brian B
        Posted Oct 31, 2009 at 10:06 AM | Permalink

        Re: Michael Jennings (#121),

        Bishop, I am surprised you would even bother to link that tripe here (except for educational purposes of what NOT to do). It consists entirely of distortions, lies, and ad hom attacks and serves no useful purpose in furthering dialogue.

        The usefulness of linking to tripe like that is not in the realm of furthering scientific dialogue.
        It does have a use however in reminding people of the second tier attacks on valid criticisms.
        Mann, Schmidt, Briffa etal respond fairly testily all too often but preserve some level of professionalism. Unfortunately they also give a pat on the head to places like Deep Climate (or a Tamino, who slides back and forth from official RC posts to his own site for ad homs) and rely on them for just the kind of cheerleading and ad homs that set up a meme to discredit SteveM and others’ arguments that Gavin, Mann, Briffa and others can’t discredit with the numbers.

        • Posted Nov 4, 2009 at 2:27 AM | Permalink

          Re: Brian B (#124), ah, and Deep Climate quotes Briffa himself

          McIntyre’s use of the data from a single, more spatially restricted site, to represent recent tree growth over the wider region, and his exclusion of the data from the other available sites, likely represents a biased reconstruction of tree growth.

          This is from the Abstract to the second Briffa piece Steve is talking about here, Examining the validity of the published RCS Yamal tree-ring chronology which addresses Steve throughout, and which DC quotes in full. Ad hom DC may be, but their quote of Briffa’s abstract did draw my attention to this – IMHO – significant (misleading) wording.

  64. MikeN
    Posted Oct 31, 2009 at 9:29 AM | Permalink

    Did you guys notice the implication that Steve was cherry-picking by selecting Schweingruber’s Khadtya site instead of the other sites in the area. Suspicious indeed. Avam was 400 km away, so clearly Steve should have included every site within that distance. The authors would have been perfectly OK with that.

    • bender
      Posted Oct 31, 2009 at 9:49 AM | Permalink

      Re: MikeN (#122),
      Yes, of course it was noticed. Steve cherry picked nothing. He noted a divergence between the two series and asked “why combine them this way, and not some other?” “Why substitute Yamal for Polar Urals?” They are still going with the strawman fabrication that Steve concocted an alternative chronology. He noted an anomaly and asked for the criteria used to decide (1) how to aggregate sites into a regional chronology, and (2) how to screen out chronologies, such as Polar Urals, as “unsuitable”. And for this – pointing out what really IS a disturbing pattern (arbitrary choices with a lack of disclosure on screening methodology) – he is demonized.

  65. bender
    Posted Oct 28, 2009 at 8:15 PM | Permalink

    Re: Hockeystickization Revisited « the Air Vent (#74),
    Acerbic? I’m not sure about that. I found them to be somewhat collegial. Just missing the point.

5 Trackbacks

  1. By Hockeystickization Revisited « the Air Vent on Oct 28, 2009 at 8:01 PM

    [...] by Jeff Id on October 28, 2009 I’ve got plenty of posts right now to work on but today Steve McIntyre called our attention to a couple of acerbic replies from Keith Briffa to Steve’s discovery that the Briffa Yamal [...]

  2. [...] Climate Audit: ‘Briffa on Yamal Impact’ [...]

  3. [...] Email 1072 (1247199598) – cruts tmp to 2008 (Dec 21, 2009) [...]

  4. [...] in October 2009 in two posts during the pre-Climategate Yamal controversy ( Sept 29 here and Oct 28 here, responding to Briffa’s post on the topic [...]

  5. By Response to Briffa #2 « Climate Audit on Mar 31, 2011 at 12:47 PM

    [...] noted at CA last week, Briffa published a partial response to Yamal issues at the CRU website, one post [...]

Follow

Get every new post delivered to your Inbox.

Join 3,254 other followers

%d bloggers like this: