It’s Saturday Night Live

When you’re trading in puts and calls (or derivatives), it’s important to know the sign of the relationship between the value of the derivative and the market. Short positions will go up in value as the market goes down. And, unfortunately, you don’t get to decide afterwards whether you wanted to be short or long. Proxies in climate can, in a sense, either be “short” or “long”, in the sense that the values of some proxies (e.g. coral dO18) are said to go down with higher temperatures, while the values of other proxies (e.g. ice core dO18) are said to go up with higher temperatures.

One feels that it is not asking too much of paleoclimatologists to know the expected sign of a proxy derivative. Traders would like to decide on the sign of a proxy derivative after the fact, by taking a correlation to market performance, but this luxury is denied to them, as it should be denied to climate scientists.

In Mann et al 2008, there is a truly remarkable example of opportunistic after-the-fact sign selection, which, in addition, beautifully illustrates the concept of spurious regression, a concept that seems to baffle signal mining paleoclimatologists. For this example, we turn to the highly HS-shaped Finnish sediment series of Tiljander et al 2003. (Update: We reported this in a PNAS Comment, to which Mann responded. See continued discussion at Upside Down Mann and the Peer Reviewed Literature.”)

Tiljander et al cored varved sediments from Lake Korttajarvi, Finland, going back through most of the Holocene. In Tiljander et al 2003, they distinguished the amount of mineral and organic matter in each varve. The basis for using mineral and organic matter as climate proxies is set out as follows:

The amounts of inorganic and organic matter, form the basis of the climate interpretations. Periods rich in organic matter indicate favourable climate conditions, when less snow accumulates in winter by diminished precipitation and/or increased thawing, causing weaker spring flow and formation of a thin mineral layer. In addition, a long growing season thickens the organic matter. More severe climate conditions occur with higher winter precipitation, a longer cold period and rapid melting at spring, shown as thicker mineral matter within a varve.

The caption to their Figure 5 reports the following link between X-ray density and their climate mechanism:

High X-ray density corresponds to high amount of mineral matter (light grey value tints in X-ray film) and low X-ray density corresponds to dark grey values caused by a higher proportion of organic matter.

Putting the two paragraphs together: warmer climate favors more organic material and thus a low X-ray density. In order to show warm values at the top of a graph, you need to invert the plot (i.e. you have to pay attention to the sign of your climate derivative.)

In the figure below, on the left, I show an excerpt from their Figure 5 which they show vertically (only the X-ray density is shown here – consult the original paper for the other plots.) The left portion of their Figure 5 shows an organic-rich period in the MWP, about which they say:

An organic rich period from AD 980 to 1250 in the Lake Korttajarvi record is chronologically comparable with the well-known ‘Medieval Warm Period’ (e.g. Lamb 1965; Grove & Switsur 1994; Broecker 2001). The sediment structure changes, less mineral material accumulates on the lake bottom than at any other time in the 3000 years sequence analysed and the sediment is quite organic rich (LOI ~20%). Thus, the winter snow cover must have been negligible, if it existed at all, and spring floods must have been of considerably lower magnitude than during the instrumental period (since AD 1881). According to the scenarios presented by Solantie & Drebs (2001), a 2°C increase in winter temperature would decrease the amount of snow in southern Finland significantly. Under such conditions, winter snow accumulation and intense spring floods would be rare events….

The Lake Korttajarvi record also indicates a climatically more severe period in the 17th century. Two periods, AD 1580–1630 and AD 1650–1710, are marked by an increase in both sedimentation (varve thickness) and mineral matter accumulation (relative Xray density). Also, magnetic susceptibility values are high between AD 1650 and 1710, indicating increasing mineral matter input into the lake.

They cite literature, including Hulden 2001, showing mild conditions in Finland in the MWP.

On the right, I’ve plotted the corresponding data so that “warm” grey values are on the top. I’ve also highlighted the (MWP) period identified as having elevated values of organic matter. If you squint, you can satisfy yourself that the left-hand and right-hand panels are showing the same data.

   

Fig. 1. Left from Tiljander et al 2003 Figure 5; right – plot of X-ray density (inverted).

Plotted according to the climatic interpretation offered by Tjilander et al, the modern warm period shows as colder than the Little Ice Age, something which makes no sense if this data is to be used as a climate proxy. Tiljander et al provide a plausible interpretation of the “divergence” of the proxy from its climatic interpretation as a result of agricultural and construction disturbance to sediment patterns, actually tying several especially thick varves to ditch and bridge construction:

This recent increase in thickness is due to the clay-rich varves caused by intensive cultivation in the late 20th century. …

There are two exceptionally thick clay-silt layers caused by man. The thick layer of AD 1930 resulted from peat ditching and forest clearance (information from a local farmer in 1999) and the thick layer of AD 1967 originated due to the rebuilding of the bridge in the vicinity of the lake’s southern corner (information from the Finnish Road Administration).

Now let’s see what Mann et al did with this data. All of the 20th century values of varve thickness, X-ray density etc go up like crazy with the agricultural and construction activities as shown below for 2 of the 4 series (the other two are similar). Instead of using the climatic interpretation of the data described by Tiljander et al, Mann correlates the increases in varve thickness and changes in density and color, originating from local construction and farming, to world climate.

   

Figure 2. Two of 4 versions used in Mann et al 2008

By flipping the data opposite to the interpretation of Tiljander et al, Mann shows the Little Ice Age in Finland as being warmer than the MWP, 100% opposite to the interpretation of the authors and the paleoclimate evidence. The flipping is done because the increase in varve thickness due to construction and agricultural activities is interpreted by Mann et al as a “nonlocal statistical relationship” or “teleconnection” to world climate. Mann:

the EIV approach, which makes use of nonlocal statistical relationships, allowing temperature changes over distant regions to be effectively represented through their covariance with climatic changes recorded by the network

A more convincing example of spurious regression in “peer reviewed” literature will be hard to find. After reading through this, I keep expecting someone to say:

Live from New York, it’s Saturday Night.

H/t to Howard Wiseman: http://www.youtube.com/watch?v=M1owcncKCHg

Reference:
TILJANDER, MIA, M. SAARNISTO, AEK OJALA, and T. SAARINEN. 2003. A 3000-year palaeoenvironmental record from annually laminated sediment of Lake Korttajarvi, central Finland. Boreas 32, no. 4: 566-577.


90 Comments

  1. Dave Brewer
    Posted Oct 2, 2008 at 8:58 AM | Permalink | Reply

    What a joke! Surely this and the other egregious errors you have found in Mann’s latest paper warrant a letter to the journal?

  2. PhilH
    Posted Oct 2, 2008 at 9:04 AM | Permalink | Reply

    Whatever it costs, Congress and the Executive branch need to immediately hire Mann to solve the financial crisis. He can just flip the charts over. Voila! Problem solved. Close the hood. The IPCC will be around to watch his back and Gavin can start a new blog: “Real Finance.”

  3. Raven
    Posted Oct 2, 2008 at 9:05 AM | Permalink | Reply

    Any feel for how much weight these proxies are given?

  4. Steve McIntyre
    Posted Oct 2, 2008 at 9:13 AM | Permalink | Reply

    #3. Not yet. I’ve been trying to get a handle on the proxies before wading into the mysteries of the newest weighting system.

    However, these “proxies” appear to be influential in the non-dendro reconstruction going back to the MWP which was the key tout of the present paper.

  5. TerryB
    Posted Oct 2, 2008 at 9:44 AM | Permalink | Reply

    Dear Climate Auditors,
    I’m sure you’ve all got thick skins these days, but this one might irritate you ever so slightly. According to George Monbiot (who will be familiar to UK readers as a staunch defender of Mann’s findings):

    “Mann’s paper, published in the Proceedings of the National Academy of Sciences, uses every uncluttered high-resolution proxy temperature record in the public domain(25).

    And any crtitique of Mann’s work are merely “…the claims of unqualified bloggers…”

    http://www.monbiot.com/archives/2008/09/23/the-patron-saint-of-charlatans/

    Over to you.

  6. GTFrank
    Posted Oct 2, 2008 at 9:46 AM | Permalink | Reply

    Just up:
    Press release from Penn State Journal of Climatology
    http://www.met.psu.edu/joc/issue=130a,p=24

    Headline: Tijlander all upside down and got it wrong!
    “In their latest paper, Mann et al have successfully invalidated the conclusions of Tijlander et al 2003 with their robust re-evaluation of the raw data from the Finnish study. This follows on the heels of his landmark papers (1998, 2003, 2005) which are most responsible for removing the Medieval Warming Period from the history books.”

    I wonder if Tijlander will respond.

    • IainM
      Posted Oct 2, 2008 at 11:04 AM | Permalink | Reply

      Re: GTFrank (#6),

      This link is dysfunctional, could you repost ?

      Steve:
      I think that it’s a joke.

      • GTFrank
        Posted Oct 2, 2008 at 11:15 AM | Permalink | Reply

        Re: IainM (#15),

        Yes, it is, somewhat like the paper, apparently.

        an attempt at satire, totally fake link.

        • GTFrank
          Posted Oct 2, 2008 at 11:24 AM | Permalink

          Re: GTFrank (#20), Re: IainM (#15),

          But the point is: Is Tijlander supportive of this use? Will she (Mia?) step up and criticize – or clarify (a la Joliffe) or stay quiet and bask in the glow of being referenced by a “leading” climatologist) surely to be included in the next IPPC report.

        • IainM
          Posted Oct 2, 2008 at 11:30 AM | Permalink

          Re: GTFrank (#23),

          I’d be surprised if the author was actually aware of it, so unless someone actively brings it to her attention I’d expect her to stay quiet.

    • Nylo
      Posted Oct 2, 2008 at 11:08 AM | Permalink | Reply

      Re: GTFrank (#6),
      The link doesn’t work, page not found.

  7. MrCPhysics
    Posted Oct 2, 2008 at 9:54 AM | Permalink | Reply

    This is among the clearest examples given on CA of bad science at work. It looks like they just didn’t bother to look at the evidence at all other than to flip it whatever direction suited the modern warming idea.

    An error of this silliness by itself, this would demand a retraction and rewrite of any paper in physics.

  8. Posted Oct 2, 2008 at 9:57 AM | Permalink | Reply

    My back calculations show this particular series was given a greater weight than the other 3 Tijlander sediment series by from 2 to 12 times.

    http://noconsensus.wordpress.com/2008/09/28/mann-08-series-weight-per-year/

    The gray line across the bottom of the multicolor graph is all 4000 type proxies from SD1 file in M08.

  9. Kusigrosz
    Posted Oct 2, 2008 at 10:14 AM | Permalink | Reply

    While researchers have been wandering all over the world since the end of the Medieval Warm Period in search of good temperature proxies, modern technology together with advanced statistical methods should open new possibilities. Proxies for temperatures in, say, southern Spain could be found on the Moon, on Mars, and beyond.

  10. Posted Oct 2, 2008 at 10:16 AM | Permalink | Reply

    You want climate science comedy????

    I’ll give you climate science comedy!!!!

    Gavin, the well known statistical expert, wrote a post criticizing Dennis Bray and Hans von Storch’s ongoing attempt to survey the climate science community.

    The series of questions Q15 through Q17, typify a key issue – precision. Q15 asks whether the “current state of scientific knowledge is developed well enough to allow for a reasonable assessment of the effects of turbulence, surface albedo, etc..”. But the subtext “well enough for what?” is not specified. Thus any respondent needs to form their own judgment about what the question is referring to..

    Well duh! It’s a question designed to have broad interpretation, so that experts in each field relating to climate can determine the answer without guidance from the inquisitors. Gavin completely misses the point and goes on to list all manner of specifics.

    The best rib tickler is in the comments:

    Wonderful project!

    Isn’t it amazing how climate science must reach out to other disciplines.

  11. Posted Oct 2, 2008 at 10:29 AM | Permalink | Reply

    The r value in SD1 for this series is positive. It must have been flipped before correlation. The darksum series also was weighted fairly heavily in my backcalc.

  12. Clark
    Posted Oct 2, 2008 at 10:32 AM | Permalink | Reply

    I wonder if the flipping is reversed back to the original Tijlander form, would Mann decide to repeat his dendro approach and simply chop off the last 50-100 years and “replace” the data with interpolations from other, more “accurate” proxies.

  13. Steve McIntyre
    Posted Oct 2, 2008 at 11:01 AM | Permalink | Reply

    #11. Be careful in how you use “flipping”. The climatic interpretation of Tijlander inverted the X-ray density measurements (and thicknesses) etc. When Mann used the raw data without checking to see whether it was a “long” or “short” position, he, in effect, inverted the Tijlander interpretation (by not inverting the original data.).

    All 4 series go up a lot in the 20th century and in amounts that are inconsistent with a climate interpretation – orders of magnitude. I see no purpose in trying to attribute a 20th century meaning to darksum rather than lightsum for recent climate interpretation. The original authors don’t. Neither should third parties.

  14. Posted Oct 2, 2008 at 11:01 AM | Permalink | Reply

    The thing that bothers me about the r value, is that the calibration methodology performs the flip automatically (if I understand it correctly). If he didn’t flip the data beforehand we would have seen a negative r and inquired ‘why the flip’. This way there is a +r and no red flag is introduced in a series which by its alleged physical relationship to temp could not be flipped.

    Steve doesn’t like us to go on about intent so I’m not going to.

    Steve: Again , be careful in what you’re saying is flipped. Tijlander inverted the series for interpretation; Mann didn’t. So it’s not that he had to manually do something to get a wrong interpretation. This could have happened by not paying attention. However, in the SI, he quotes from Tijlander et al – it’s the only study that he actually quotes in the entire SI, so he could hardly have een unaware of the issues here. Also he did a special calculation to show that they were “profitable” without this data, though, as far as I can tell, this profit calculation requires the inclusion of strip bark “derivatives”.

  15. Francois Ouellette
    Posted Oct 2, 2008 at 11:04 AM | Permalink | Reply

    #10 Sonic, in my recent exchange at RC on the “Simple questions…” post, I was heavily criticized for making comments on the inner workings of the climate science community (basically how being mostly academics they were shielded from the real world), and was told that I had a “profound ignorance” of how science, and in particular climate science, really works. I was even told that if I did not know a climate scientist personnally, I had no right to comment. I had to defend myself endlessly by telling them about my experience in academia, and in industry, and so on, but it never seemed to be enough. It seems only a climate scientist can comment on climate science.

    But then we have this post, where a climate scientist is telling a social scientist (not really what Dennis Bray’s background is…) how to do his job! Of course, it makes sense that climate scientists are experts at making polls…

  16. Posted Oct 2, 2008 at 11:10 AM | Permalink | Reply

    #13

    I see, so he or his grad students didn’t understand the data was supposed to be inverted. It makes more sense now.

    Very nice find again.

  17. Steve McIntyre
    Posted Oct 2, 2008 at 11:11 AM | Permalink | Reply

    #10, 16. OT. Maybe we’ll deal with this on another occasion.

  18. Steve McIntyre
    Posted Oct 2, 2008 at 11:15 AM | Permalink | Reply

    #18. IT’s a familiar story. If the series had gone “down”, they would have been all over the divergence. In this case, the increases are actually a sort of “divergence” on the up side, as there are exponential increases in sum series – orders of magnitude, something impossible in a normal distribution. But do they care?

  19. Steve McIntyre
    Posted Oct 2, 2008 at 11:19 AM | Permalink | Reply

    I’m thinking Tina Fey as Sarah Palin as Rosanne D’Arrigo, explaining in a nice homey way how you need to pick cherries to make cherry pie. The mind boggles at all the possibilities when you make three-stage comedy like this.

  20. Posted Oct 2, 2008 at 11:36 AM | Permalink | Reply

    It is amazing to me that he would even consider this series. It clearly isn’t temperature.

    On the other thread, I was being facetious about putting the curve in front of a class of 5th graders to vote on what is a proxy and what isn’t. It would still be an improvement.

  21. Woodpecker
    Posted Oct 2, 2008 at 11:38 AM | Permalink | Reply

    A very interesting story Steve. If this is all correct, perhaps it points the way towards explaining why the hockey stick shape is the way it is. If some proxies can result in downward-pointing sticks while others result in upward-pointing sticks, shouldn’t the real shape of the plot (without applying ex post facto analysis) be one where there is a huge increase in noise in the 20th century (due to the abrupt increase in all kinds of human activities), following 1000+ years of relative stability. This would, of course, mean that there could be no real correlation between temperature in the 20th C and any proxy, making temperature reconstructions before then totally meaningless.

    It goes without saying (but I’ll say it anyway just so we’re clear) that Mann et al appear to have chosen (or mis-chosen, in this case) only those sticks pointing in the ‘right’ direction.

  22. tty
    Posted Oct 2, 2008 at 11:51 AM | Permalink | Reply

    A bit of nitpicking, the name is “Tiljander”, and nothing else. I understand that it may be a bit confusing, since it is a Swedish name, rather than a Finnish.

    Steve
    : OK, thx.

  23. Posted Oct 2, 2008 at 11:57 AM | Permalink | Reply

    Let me check how gridboxcps.m handles this proxy. Here’s gridboxy(:,2202) when n=9 without (*) lowpassmin :

    and variance matched with local temperature (temp. in red)

    And then the same figures with Mann-smoothing, freq=1/10

    Smoothing + variance matching seems to nicely remove those inconvenient peaks ;)

    *) easy way: rewrite lowpassmin.m

    function[smoothed0,icb,ice,mse0] = lowpassmin(indata,frequency)
    %
    %

    smoothed0=indata;

    icb=0; ice=0; mse0=0;

    disp(‘!’)

  24. Steve McIntyre
    Posted Oct 2, 2008 at 12:01 PM | Permalink | Reply

    He only reports CPS versions back to AD1500. Have you figured out the EIV version?

  25. varve
    Posted Oct 2, 2008 at 12:02 PM | Permalink | Reply

    I’m a little confused by the somewhat ambiguous comment in the original Tiljander paper that suggests to me that perhaps one shouldn’t even be using these sedimentation series as temperature proxies (at least not without some pretty large and explicit warning labels):

    Even though the sedimentation in Lake Korttajarvi most likely reflects relatively long-term changes in
    local hydrology rather than temperature, several studies indicate a relative temperature rise during the Medieval period in Scandinavia.

    Tiljander at 574. Of course, Tiljander goes on in the paper to ascribe a temperature relationship to her sedimentary analysis (at least up until the last few centuries) and provides relatively little discussion about the hydrology relationship.

    I guess my question is: Even assuming one can identify the proper sign of the relationship here, how does one tease out a temperature signal (for the entire Northern Hemisphere no less !!) in these sedimentary series distinct from the local hydrology signal without being at least a little hesitant about how much confidence one should place in the analysis they’ve done?

  26. Steve McIntyre
    Posted Oct 2, 2008 at 12:08 PM | Permalink | Reply

    In Mann-world, such reasonable caveats are swept aside by the theory of teleconnection. All that matters is the correlation. Of course, elsewhere he says that the r2 is no use, so it’s a little hard to make sense of it all.

  27. Mark T.
    Posted Oct 2, 2008 at 12:15 PM | Permalink | Reply

    ^UC, I guess he could also just set r2 to 1.0 and be done with it, right? :)

    Oh, Steve, “Live from New York, it’s Saturday Night!” No “Live” on the end.

    Mark

    Steve:
    OK.

  28. John Lang
    Posted Oct 2, 2008 at 12:33 PM | Permalink | Reply

    This just gets stranger and stranger each time you make a new post.

  29. Steve McIntyre
    Posted Oct 2, 2008 at 12:43 PM | Permalink | Reply

    #32. There is a very strong Canadian (or more Toronto) influence on SNL. Lorne Michaels is from Toronto, a few years older than me. Other comedians from Toronto include Mike Myers, John Candy, Jim Carrey, Catherine O’Hara, Eugene Levy, Rick Moranis, Robin Duke, … A lot of the original SNL cast came from a Canadian show “Second City TV”.

    • Tolz
      Posted Oct 2, 2008 at 1:53 PM | Permalink | Reply

      Re: Steve McIntyre (#34),

      Dave Thomas (ya hoser!), and I think Martin Short, too. The old SCTV shows were just great. But I thought Chicago was the “Second City”. In any event, thank you Canada.

  30. Patrick M.
    Posted Oct 2, 2008 at 1:34 PM | Permalink | Reply

    Re Steve #34:

    Ah yes, who can forget Bob and Doug McKenzie, (Rick Moranis and Dave Thomas)?

    Perhaps the thread title could have been, “Welcome to the Great White North”?

  31. Hans Erren
    Posted Oct 2, 2008 at 2:16 PM | Permalink | Reply

    You could ask if Mia Tiljander still works with the Finnish geological survey
    http://en.gtk.fi/ContactInfo/index.html

  32. Steve McIntyre
    Posted Oct 2, 2008 at 2:17 PM | Permalink | Reply

    A reader has notified me that he has emailed Tiljander.

  33. DennisBoz
    Posted Oct 2, 2008 at 2:33 PM | Permalink | Reply

    One of the hallmarks of science is the idea of falseafiability. Science can be tested and if predictions do not match the observed results, the hypothesis is rejected.

    Since we don’t know what the world temperatures have been for the last 1000 years, does that make these reconstructions ‘not’ science?

    And if ‘not science’, then what is it?

    Steve: EDitorially, I prefer that the blog stick to smaller and more specific questions.

  34. Wolfgang Flamme
    Posted Oct 2, 2008 at 2:52 PM | Permalink | Reply

    Here’s another flipping-candidate: yu_1999_mgcaratio

    HF-use: no (decreasing in ref)
    LF-use: yes (increasing i.g.)

    Now things need to be turned upside down.

  35. Howard
    Posted Oct 2, 2008 at 2:55 PM | Permalink | Reply

    What is the problem, Mann deserves the Nobel prize for his new groundbreaking discovery. For millennia, scientists and engineers have consistently sought to maximize signal to noise ratios for their experiments and trials. Mann proves that we have all been wrong as he so clearly proved it is more better to maximize the noise to signal ratio.

  36. Posted Oct 2, 2008 at 3:16 PM | Permalink | Reply

    I would never know about puts and calls, in spite of my study in Economic University, but thanks to reading books like Rich dad poor dad I know everyting about traiding in the Stock Exchange!

  37. Wolfgang Flamme
    Posted Oct 2, 2008 at 3:46 PM | Permalink | Reply

    … and another one: dongge.

    Opening a can of … Mann?

  38. Dave Andrews
    Posted Oct 2, 2008 at 4:38 PM | Permalink | Reply

    Varve,

    Even assuming one can identify the proper sign of the relationship here, how does one tease out a temperature signal (for the entire Northern Hemisphere no less !!) in these sedimentary series distinct from the local hydrology signal without being at least a little hesitant about how much confidence one should place in the analysis they’ve done?

    Its easy once you have decided a few BCPs in a small part of the US can be representative of global temperature ( even though those specimens were not regarded as temperature proxies by the original collectors)

  39. hswiseman
    Posted Oct 2, 2008 at 10:17 PM | Permalink | Reply

    Careerbuilder superbowl ad with Chimps flipping the financial charts and celebrating.

  40. hswiseman
    Posted Oct 2, 2008 at 10:17 PM | Permalink | Reply
  41. hswiseman
    Posted Oct 2, 2008 at 10:18 PM | Permalink | Reply

    Sorry cannot seem to embed. Here is the link

  42. hswiseman
    Posted Oct 2, 2008 at 10:20 PM | Permalink | Reply

  43. tty
    Posted Oct 3, 2008 at 12:31 AM | Permalink | Reply

    Re 30

    “Even assuming one can identify the proper sign of the relationship here, how does one tease out a temperature signal (for the entire Northern Hemisphere no less !!) in these sedimentary series distinct from the local hydrology signal without being at least a little hesitant about how much confidence one should place in the analysis they’ve done?”

    I think you can be fairly confident in Tiljanders et al. MWP winter temperature interpretation. There is a sort of threshold effect here. At the winter temperatures now prevailing in southern Finland there is almost invariably a lasting snow cover and a strong spring flood. At slightly higher temperatures each snowfall tends to melt fairly quickly, there is no snowpack accumulation and little spring flooding. Living about 500 km southwest of Korttajärvi I have actually seen this change during my lifetime due to the recent warming. In the 1950′s through 1970′s we almost always had a lasting snow cover and strong spring flooding. In recent years this has become intermittent and only happens maybe every second or every third year. Incidentally the MWP must have been appreciably warmer in winter in northern Europe (1-2 degrees?) than the current climate for this zone to move to south-central Finland.

  44. Posted Oct 3, 2008 at 2:59 AM | Permalink | Reply

    Using both hockey sticks and anti-hockey sticks to improve the case for a hockey stick reconstruction is an advanced phase of cherry-picking. If there is no independently known sign of the correlation, it suggests that the correlation is a coincidence. Noise.

    Incidentally, Craig Loehle

    http://motls.blogspot.com/2008/09/craig-loehle-trouble-with-tree-ring.html

    recently argued in his paper that tree rings typically face the “intermediate” problem from the same universality class: the sign of the dependence actually flips above a certain realistic temperature. Nonlinearities are thus important and all the linear methods to obtain temperature from tree rings are incorrect. The paper is recommended (I suppose you have written about it, too, but I’ve forgotten).

  45. gens
    Posted Oct 3, 2008 at 7:55 AM | Permalink | Reply

    One minor nit – puts and calls are of course derivatives so no need to differentiate in your first sentence.

    To extend the metaphor, one of the interesting / difficult issues in financial risk management is that in times of crisis, relationships actually can (and do) change sign. A great example is Morgan Stanley losing in excess of $9 billion (!) last year where a major market move resulted in a hedge changing sign and becoming positively correlated with the underlying exposure. Well I think we can say that the hockey stock is in crisis and Mr. Mann is facing a large loss.

  46. Luis Dias
    Posted Oct 3, 2008 at 8:56 AM | Permalink | Reply

    This is shocking and makes me want to puke. Sorry mr. Steve. Snip my comment at will, but there are just some things that are completely over the top.

  47. Wolfgang Flamme
    Posted Oct 3, 2008 at 12:09 PM | Permalink | Reply

    @Luboš (#50)

    If that ‘simple’ optimum theory were correct and we propose even warmer temperatures for the past, then we should see all pasttime TRWs leveling-off around recent maximum growth rates.

    Or else one has to suggest very abrupt changes in climate = ‘switching between solutions’ of TRW response. However these changes shuld have been recorded by proxies that respond differently.

  48. Posted Oct 3, 2008 at 11:48 PM | Permalink | Reply

    This is totally devastating, Steve, especially considering that this series is the one that “proves” that the Mann et al 2008 reconstruction is “robust” to the elimination of BCP, and vice-versa!

    Please remember that PNAS policy is that any errors not corrected by a communication within 3 months become Pravda (uncontrovertable “truth”)!

    Perhaps Mann should sell this paper to the Treasury TARP program while he still can get something for it. That is, if the Treasury has not already followed PhilH’s advice (#2), and solved the subprime mortgage reckonning by means of a Mannomatic chart-flip!

  49. Steve McIntyre
    Posted Oct 4, 2008 at 8:22 AM | Permalink | Reply

    Here is the link referred to in #45, showing an example of using upside down graphs. The clip was made for purposes unrelated to climate science and resemblance to any real characters is explicitly renounced.

  50. jae
    Posted Oct 4, 2008 at 7:47 PM | Permalink | Reply

    I’m laughing so hard I’m almost crying!

  51. hswiseman
    Posted Oct 4, 2008 at 10:09 PM | Permalink | Reply

    Ad troglodytem attacks are expressly forbidden under CA policy

  52. Nylo
    Posted Oct 5, 2008 at 12:47 AM | Permalink | Reply

    Priceless video!

    But for me, whether characters in the video resemble or not some very specific real characters, I will tell you in a few months, depending on whether Mann apologises and uses his Korttajarvi graphics right or decides to keep them upside down because they look damn better :)

  53. Posted Oct 5, 2008 at 11:43 AM | Permalink | Reply

    #49

    While you may be right about an effect on the sediment rate. I’m sure you agree we can’t possibly call it linear.

    Love the video.

    #54 – Hu, really good point, again. This could force M08 to go back and redo a significant amount of statistical painting … M09 anyone?

    BTW:I cut the MXD out from my backcalc of the M08 NH graph I worked with and plotted the difference in reconstructed temp on my blog.

  54. John M
    Posted Oct 5, 2008 at 5:50 PM | Permalink | Reply

    It looks like our old friend TCO may be stirring things up over at “Open Mind”.

    link

    I didn’t see the inverted graph on his list.

    • Jean S
      Posted Oct 6, 2008 at 4:51 AM | Permalink | Reply

      Re: John M (#60),
      Me thinks TCO also forgot the use of bristlecone pines from the list. The most dominating proxy in CPS (pre AD1500) is this:

      Chronology file name : NV512.CRN
      Measurement file name : NV512.RWL
      Date checked : 08DEC94
      Technician’s name : MARIETTE SEKLECKI
      Supervisor’s name : HENRI D. GRISSINO-MAYER
      Beginning year : 320
      Ending year : 1985
      Principal investigators: DONALD A. GRAYBILL
      Site name : PEARL PEAK
      Site location : NEVADA, USA
      Species information : PILO BRISTLECONE PINE
      Latitude : 4014N
      Longitude : 11532W
      Elevation : 3170 M
      Series intercorrelation: 0.647
      Avg mean sensitivity : 0.310
      Avg standard deviation : 0.124
      Avg autocorrelation : 0.612
      Number dated series : 44
      Segment length tested : 50
      Number problem segments: 32
      Pct problem segments : 1.85

      Maybe nobody has told him that yet. Well, now you know, TCO.

      Anyone guessing what is the second most dominating (and the dominating non-dendro) (it’s NOT Korttajärvi)?

      • IainM
        Posted Oct 6, 2008 at 5:11 AM | Permalink | Reply

        Re: Jean S (#61),

        Would it happen to be a speleothem?

      • Posted Oct 6, 2008 at 6:42 AM | Permalink | Reply

        Re: Jean S (#61),

        Anyone guessing what is the second most dominating (and the dominating non-dendro) (it’s NOT Korttajärvi)?

        Ah, time for reverse-engineering ;)

        Let’s see,

        temp=reg_ts(1001:end,:,9);a=sum(isfinite(temp)); ii=find(a==996);ii’, A=gbcpsfull(:,9);
        Aa=A(isfinite(A)); P=[temp(:,ii) ones(996,1)]; g=pinv(P)*Aa,
        max(abs(Aa-P*g))

        ans =

        670
        714
        719
        978
        1102
        1486
        1487
        1603
        1714
        1733
        1748
        1764
        1786
        1790
        1808

        g =

        0.1784
        0.1784
        0.1784
        0.2256
        0.2392
        -0.0070
        0.0048
        -0.0000
        0.0000
        -0.0000
        -0.0000
        0.0000
        0.0000
        -0.0000
        -0.0000
        -0.1894

        ans =

        1.8874e-015

        Steve (Nov 2): I can tie this in to my emulation. The first 5 series are the 5 SH series more or less weighted by their cosine latitude (which equating the first coefficient yields 0.1784000 0.1784000 0.1784000 0.2235524 0.2362357)
        0.1784 – arge091
        0.1784 – Tasmania recon
        0.1784 – Oroko recon
        0.2256 – S AFr speleo
        0.2392 – Quelc O18
        The other coefficients in the reverse engineering appear to be noise coefficients from NH series not used in the calculation.

  55. Posted Oct 6, 2008 at 6:45 AM | Permalink | Reply

    (gbcpsfull in the above without hemispheric variance matching, that’s illegal so I had to modify the code )

  56. Posted Oct 6, 2008 at 6:59 AM | Permalink | Reply

    Ooops, that was SH. NH gains:

    g =

    0.0000
    -0.0000
    0.0000
    0.0000
    0.0000
    0.1100
    0.1649
    0.1353
    0.1301
    0.1154
    0.1154
    0.0885
    0.0557
    0.0557
    0.0190
    -0.0634

  57. Jean S
    Posted Oct 6, 2008 at 7:08 AM | Permalink | Reply

    UC! It’s spliced again! (Is gbcpsfull spliced or is it the the AD1000 step?) Notice also that ‘rebuiltproxymatrix’ does not have the proxies in the same order as they are in rtable1209 (or in “itrdbmatrix”)!

    Oh, I forgot. #61 and #63 are for the individual series. If you take series in one location combined, the dominating (pre AD1500) one comes from Socotra. It’s a kind of beautiful marriage with the Chinease spaleothem: apart form the modern period, the ups in Socotra are matched with downs in China. The result, well, you guessed it. I think one needs to be a skilled craftsman to find a better match! :)

    • Posted Oct 15, 2008 at 5:10 AM | Permalink | Reply

      Re: Jean S (#67),

      Is gbcpsfull spliced or is it the the AD1000 step?

      gbcpsfull(:,9) is AD1000 step. gridproxycps.m saves spliced ones as ‘gbcps’.

      #70, 71

      This means that P in #64 is rank deficient. pinv(P)*Aa gives different answer than P\Aa .

  58. Steve McIntyre
    Posted Oct 6, 2008 at 7:35 AM | Permalink | Reply

    #67. The modern portion of the Socotra speleothem has never been published. Mann’s reference to Socotra is to an article describing Socotra in the Pleistocene. The lead Socotra author is from Bradley’s department.

  59. Sam Urbinto
    Posted Oct 6, 2008 at 11:04 AM | Permalink | Reply

    Wasn’t the term something like “Serial contributor”?

  60. Posted Oct 6, 2008 at 11:52 AM | Permalink | Reply

    reg_ts 1486 and 1487 seem to be the same series, but scaled differently.

  61. Steve McIntyre
    Posted Oct 6, 2008 at 12:14 PM | Permalink | Reply

    The only proxies in Mann gridcells 1486-1487 are the two Burns Socotra speleothems in 1486.
    mannid lat long jones
    1486 1486 12.5 47.5 1126
    1487 1487 12.5 52.5 1127

    The longitude is exactly 50 :) So I guess the test is less than or equal to and that’s why it gets included twice – once in each adjacent cell. Good spotting.

    id lat long jones r1850_1995
    202 burns_2003_socotrad13c 12.5 50 1126 0.5066
    203 burns_2003_socotrad18o 12.5 50 1126 0.4759

    As noted elsewhere, this is still an unpublished data set. Also I haven’t seen ANY authorities for using C13 as a temperature proxy.

  62. Posted Oct 6, 2008 at 12:40 PM | Permalink | Reply

    This might help to visualize ( bypassed lowpassmin.m to show unsmoothed series )

    for i=6:15, subplot(5,2,i-5); plot(P(:,i)), ylabel(g(i));end

    result:

    http://signals.auditblogs.com/files/2008/10/p6_15.png

  63. Jud Partin
    Posted Oct 6, 2008 at 1:17 PM | Permalink | Reply

    #68 & 71. Steve – both you and the SI are wrong. The SI lists the wrong ref. However, it has been published: twice. In 2002 & in 2004. So there’s no need for people to start gossiping…

    Burns, S. J., et al. (2002). “A 780-year annually resolved record of Indian Ocean monsoon precipitation from a speleothem from south Oman.” Journal of Geophysical Research-Atmospheres 107(D20).

    Fleitmann, D., et al. (2004). “Palaeoclimatic interpretation of high-resolution oxygen isotope profiles derived from annually laminated speleothems from Southern Oman.” Quaternary Science Reviews 23(7-8): 935-945.

    Steve: Jud, thanks for this. Always nice to resolve this sort of issue. The two publications here describe a period of 780 years (i.e. do not cover the 11th century usually at issue), while the Mann data set goes back to 2754 BCE. I take it that the data set was subsequently extended. I notice that Burns et al 2002 says: “The d18O and d13C values of speleothem calcite are inversely related to precipitation” and Fleitmann et al 2004 says: “Nor is the isotopic composition of rainfall related to temperature in the tropics.” It’s odd candidate for an influential NH temperature proxy.

    • Steve McIntyre
      Posted Nov 1, 2008 at 5:21 PM | Permalink | Reply

      Re: Jud Partin (#73),

      Jud, I’ve had an opportunity to look at your two references and I believe that your above comment is incorrect. Socotra Island is said to be in Yemen, while the two references are to Oman (Kahf Defore), nearby but in a different spot.

  64. Steve McIntyre
    Posted Oct 9, 2008 at 8:09 AM | Permalink | Reply

    Mia Tiljander has just sent me a digital version of the data in Tiljander et al (Boreas 2003), which I’ve plotted below, comparing to the Mann version. In virtually all years, the data matches to 3 decimal places. The only exception are the spikes at 1326 shown in red, where the spike does not exist in the Mann version. Odd.

    TILJANDER, MIA, M. SAARNISTO, AEK OJALA, and T. SAARINEN. 2003. A 3000-year palaeoenvironmental record from annually laminated sediment of Lake Korttajarvi, central Finland. Boreas 32, no. 4: 566-577.

  65. Posted Nov 1, 2008 at 9:49 PM | Permalink | Reply

    Yes, Socotra – Suqutra – is an island that is part of Yemen, though closer to the African continent than the Arabian Peninsula, and populated by a different ethnic group than Arabs, speaking a different language. However, there is some intriguing overlap of certain fauna in Oman and Suqutra.

  66. Jud Partin
    Posted Nov 2, 2008 at 11:25 AM | Permalink | Reply

    Steve-

    Then it’s probably the record from Dimarshim Cave. Ref is:

    “Fleitmann, D., S. J. Burns, et al. (2007). “Holocene ITCZ and Indian monsoon dynamics recorded in stalagmites from Oman and Yemen (Socotra).” Quaternary Science Reviews 26(1-2): 170-188.”

  67. mccall
    Posted Nov 30, 2008 at 3:17 PM | Permalink | Reply

    So based on this thread analysis of the Mann’08 misuse of Tiljander’03
    http://www.climateaudit.org/?p=3967

    especially combined with Eschenbach & Company at
    http://www.climateaudit.org/?p=4428#more-4428

    it’s safe to say that the following is inaccurate:

    In a paper on 9 September 2008, Mann and colleagues published an updated reconstruction of Earth surface temperature for the past two millennia.[61] This reconstruction used a more diverse dataset that was significantly larger than the original tree-ring study. Similarly to the original study, this work found that recent increases in northern hemisphere surface temperature are anomalous relative to at least the past 1300 years, and that this result is robust to the inclusion or exclusion of the tree-ring dataset.

  68. Steve McIntyre
    Posted Nov 30, 2008 at 4:14 PM | Permalink | Reply

    The possibility of this result was noted some time ago and does not rely on the form of analysis used by Willis.

    I state it this way not to deprecate Willis’ analysis or methods which I find interesting. However, it’s important that this sort of applied claim not rely on novel methods – just because you happen to like the results.

    So far I haven’t posted up my own emulation of Mann’s results using MAnn’s methodology with a correct orientation for the Tiljander sediments, as I’ve been working my way through the nits of Mannian methods. However, it is definitely a reasonable hypothesis that reversing the orientation of the Tiljander sediments will have a noticeable impact on a Mann-style recon. Stay tuned.

  69. John M
    Posted Jan 26, 2009 at 7:59 AM | Permalink | Reply

    Steve,

    I’ve been having a discussion with “foinavon” over at WUWT, and along the course of the discussion (long thread), I mentioned this thread. Since he seems bashful about appearing over here himself, I’ll take the liberty of posting his comment myself.

    It took me about 5 minutes to download Mann et al’s paper [Proxy-based reconstructions of hemispheric and global surface temperature variations over the past two millennia Proc. Natl. Acad. Sci. 105, 13252-13257] , to browse the Supplementary Information posted on the website of Proc. Natl. Acad. Sci. and to discover the following:

    1. Mann et al. have a large section in their Supplementary info (right near the front on page 2) entitled:

    “Potential data quality problems”

    In this section they describe that of the 1200 or so proxy series used in their analyses, 7 are potentially problematic due to data quality issues. The authors state that these records include the 4 Tiljander lake varve series (that Mcintyre is making a fuss about) and three other series (Mono lake; Isdale fluorescence data; McCulloch Ba/Ca data).

    Since these data are potentially problematic an entire reanalysis of the data was performed leaving these data sets out.

    These reconstructions are shown in Figure S8 on page 14 of the Supplementary info, as a comparison with the reconstruction using the full data set. Leaving out the Tiljander (and Mono, Isdale, McCulloch data sets) makes a trivial difference to the long-term CPS Northern Hemisphere land reconstruction and a small difference to the EIV Northern Hemisphere land plus ocean reconstrucrion.

    I haven’t read the paper fully so won’t comment further for now. However this addresses your question about peer-review. Mann et al very clearly highlight those data sets they consider to have potential problems and perform analyses leaving these out. When they do so the interpretations of the paper don’t require any material change. Therefore a review should consider that particular matter adequately dealt with.

    The more interesting question is why McIntyre doesn’t indicate this glaringly obvious point. Perhaps he does somewhere else on his blog, but certainly not on the web page you linked to.

    That’s pretty much why I’m skeptical of blogs and prefer scientists and the scientific literature when it comes to assessing science.

    Comments?

  70. Steve McIntyre
    Posted Jan 26, 2009 at 8:31 AM | Permalink | Reply

    Leaving out the Tiljander (and Mono, Isdale, McCulloch data sets) makes a trivial difference to the long-term CPS Northern Hemisphere land reconstruction and a small difference to the EIV Northern Hemisphere land plus ocean reconstrucrion.

    John M, my take on this is that Mann’s two sensitivities are rather artfully designed. His “no-Tiljander” sensitivity includes Graybill bristlecones and the resulting recon is heavily weighted by these chronologies that even the NAS panel agreed shouldn’t be used. His “no-dendro” i.e. “no-bristlecone” recon uses upside-down Tiljander proxies. I mentioned this previously on one of the contemporary threads http://www.climateaudit.org/?p=3951.

    I agree that he noted problems with the Tiljander series, but, having noted them, his method of handling the problems was inadequate. IT wasn’t just that it was “possible” that there was non-climatic disturbance; there was known non-climatic disturbance. There was no obligation to use such problematic data; other less problematic data wasn’t used (the selection criteria being obscure.) There was no good reason to use the series upside-down other than this contributed to a big HS. The decision to retain this flawed data seems very odd. However, it’s not as easy as all that to assess the final impact of this sort of goof. You need to be able to replicate the calcs, which I’ve been working at on and off. I’m not convinced that S8 is a comprehensive test of the effect of the upside down Tiljander series.

    I agree that it would be better if climate scientists assessed Mann’s goofs in the literature, but they don’t. For example, does Mann’s no-dendro recon survive removal of the Tiljander proxies? Dunno, I’m working on it. I can tell you that there’s lots of hair on every step.

  71. John M
    Posted Jan 26, 2009 at 8:32 AM | Permalink | Reply

    Steve,

    I figured you were aware of Mann’s “doesn’t matter” argument.

    http://www.climateaudit.org/?p=3951

  72. Steve McIntyre
    Posted Jan 26, 2009 at 8:41 AM | Permalink | Reply

    #83. the link is written in a more lively style than many CA posts and (I think) deals with the issue.

    In business, if opportunistic accounting practices are identified, wise investors simply steer clear, rather than trying to figure out what the “real” accounts are.

  73. Steve McIntyre
    Posted Sep 3, 2009 at 9:23 AM | Permalink | Reply

    Bump.

    Keep your eyes peeled today for upside-down Tiljander. :)

  74. Posted Sep 3, 2009 at 9:41 AM | Permalink | Reply

    RE Steve #85 –
    I thought this had been resolved already, in Comment #55 above.

  75. AMac
    Posted Jun 28, 2010 at 7:08 AM | Permalink | Reply

    Re: the Trackback “Tiljander << AGW Observer (Jun 28, 2010 at 4:39 AM):

    Within the past fortnight, there was a useful discussion of the use of the Lake Korttajarvi proxies in Mann08, at Keith Kloor's blog Collide-a-scape, The Main Hindrance to Dialogue (and Detente). The post leads off with Gavin Schmidt’s complaint that my comments on Tiljander exemplify a pathology that obstructs communication between climate scientists and skeptics.

    On June 27, Arthur Smith posted on Mann08′s use of Tiljander, Where’s the fraud?. As the title indicates, his focus is on accusations of fraud. I point out in his (heavily moderated) comment section that this emphasis is misplaced and unhelpful.

    In March, I walked through the use of the Tiljander X-Ray Density (XRD) data series by Mann08 in the post The Newly-Discovered Jarvykortta Proxy — II.

    • Tom Gray
      Posted Jun 28, 2010 at 2:21 PM | Permalink | Reply

      The post leads off with Gavin Schmidt’s complaint that my comments on Tiljander exemplify a pathology that obstructs communication between climate scientists and skeptics.

      How would Schmidt say that the following practices are relevant to that kind of obstruction:

      a) extensive censoring of adverse comments

      b) editing and otherwise distorting comments

      c) replying to comments with personal and off topic comments

      d) making comments about someone’s postion and not providing the opportunity to reply

7 Trackbacks

  1. By Tiljander « AGW Observer on Jun 28, 2010 at 4:39 AM

    [...] has also presented some claims relating to this before in his website. Here’s to my knowledge the first of his posts on the issue, saying: By flipping the data opposite to the interpretation of Tiljander et al, Mann shows the [...]

  2. By Mannian CPS « Climate Audit on Jun 30, 2010 at 10:03 AM

    [...] 12.5N 47.5E and 12.5N 52.5E. This problem was noted a couple of weeks ago by Jean S and UC, who observed that that the regts series in two gridcells were identical in an early network (Socotra [...]

  3. [...] Atte Korhola. Mann’s upside down use of the Tiljander proxies was originally reported at CA here in fall 2008 and then reported to PNAS in a published comment by Ross and [...]

  4. By The Kaufman Corrigendum « Climate Audit on Aug 2, 2010 at 11:24 PM

    [...] use of the Tiljander series was originally raised at CA in the wake of Mann et al 2008 in Sept 2008 here, that it was further pointed out in a published comment on Mann et al 2008 (McIntyre and McKitrick [...]

  5. [...] use of the Tiljander sediments was originally discussed here on Oct 2, 2008, a post which discussed the multiple problems with Mann’s use of these [...]

  6. [...] Well … read on here and be [...]

  7. [...] elementary phenomenon that we reflected on in connection with Mann and upside-down Tiljander here, where a reader linked to an amusing video in which the protagonists didn’t care whether data [...]

Post a Comment

Required fields are marked *

*
*

Follow

Get every new post delivered to your Inbox.

Join 2,900 other followers

%d bloggers like this: