MBH98 and Correlation

One more bit of review before we get to Ammann’s answer. As an excuse for not answering the request of the House Energy and Commerce Committee about the R2 statistic, Mann told them that his "colleagues and [himself] did not rely on this statistic" in the following terms:

The Committee inquires about the calculation of the R2 statistic for temperature reconstruction, especially for the 15th Century proxy calculations. In order to answer this question it is important to clarify that I assume that what is meant by the “R2″ statistic is the squared Pearson dot-moment correlation, or r2 (i.e., the square of the simple linear correlation coefficient between two time series) over the 1856-1901 “verification” interval for our reconstruction. My colleagues and I did not rely on this statistic in our assessments of “skill” (i.e., the reliability of a statistical model, based on the ability of a statistical model to match data not used in constructing the model) because, in our view, and in the view of other reputable scientists in the field, it is not an adequate measure of “skill.” The statistic used by Mann et al. 1998, the reduction of error, or “RE” statistic, is generally favored by scientists in the field.

This sounds plausible, but it’s usually worthwhile checking what the Hockey Team says. MBH98 itself describes how they went about "not relying on" the r2 statistic as follows:

àƒÅ½à‚Ⱐ[RE] is a quite rigorous measure of the similarity between two variables, measuring their correspondence not only in terms of the relative departures from mean values (as does the correlation coefficient r) but also in terms of the means and absolute variance of the two series. For comparison, correlation (r) and squared-correlation (r2) statistics are also determined. Significance levels were determined for r2 from standard one-sided tables, accounting for decreased degrees of freedom owing to serial correlation

They made their lack of reliance on the verification r2 statistic very clear through the following illustration, showing the verification r2 statistics for each gridcell in the AD1820 step (which had 112 "proxies" including 12 actual temperature series as so-called "proxies" for temperature:

Original Caption: Figure 3 Spatial patterns of reconstruction statistics. … bottom, verification r2 (also based on 1854–1901 data). … For the r2 statistic, statistically insignificant values (or any gridpoints with unphysical values of correlation r < 0) are indicated in grey. The colour scale indicates values significant at the 90% (yellow), 99% (light red) and 99.9% (dark red) levels (these significance levels are slightly higher for the calibration statistics which are based on a longer period of time). A description of significance level estimation is provided in the Methods section.

Elsewhere in MBH98, they repeatedly demonstrated that they were "not relying" either on the correlation r or the correlation squared r2 through the following statements:

Time-dependent correlations of the reconstructions with time-series records representing changes in greenhouse-gas concentrations, solar irradiance, and volcanic aerosols suggest that each of these factors has contributed to the climate variability of the past 400 years, with greenhouse gases emerging as the dominant forcing during the twentieth century….

Figure 3 shows the spatial patterns of calibration àƒÅ½à‚Ⱜ and verification àƒÅ½à‚Ⱐand the squared correlation statistic r2, demonstrating highly significant reconstructive skill over widespread regions of the reconstructed spatial domain. Although a verification NINO3 index is not available from 1854 to 1901, correlation of the reconstructed NINO3 index with the available Southern Oscillation index (SOI) data from 1865 to 1901 of r = -0.38 (r2 = 0.14) compares reasonably with its target value given by the correlation between the actual instrumental NINO3 and SOI index from 1902 to 1980 (r =-0.72)


Original Caption: Figure 7 Relationships of Northern Hemisphere mean (NH) temperature with three candidate forcings between 1610 and 1995. … Bottom panel, evolving multivariate correlation of NH series with the three forcings NH, Solar, log CO2. The time axis denotes the centre of a 200-year moving window. One-sided (positive) 90%, 95%, 99% significance levels (see text) for correlations with CO2 and solar irradiance are shown by horizontal dashed lines, while the one-sided (negative) 90% significance threshold for correlations with the DVI series is shown by the horizontal dotted line. The grey bars indicate two difference 200-year windows of data, with the long-dashed vertical lines indicating the centre of the corresponding window….

We estimate the response of the climate to the three forcings based on an evolving multivariate regression method (Fig. 7). This time-dependent correlation approach generalizes on previous studies of (fixed) correlations between long-term Northern Hemisphere temperature records and possible forcing agents. Normalized regression (that is, correlation) coefficients r are simultaneously estimated between each of the three forcing series and the NH series from 1610 to 1995 in a 200-year moving window. The first calculated value centred at 1710 is based on data from 1610 to 1809, and the last value, centred at 1895, is based on data from 1796 to 1995″¢’‚¬?that is, the most recent 200 years. A window width of 200 yr was chosen to ensure that any given window contains enough samples to provide good signal-to-noise ratios in correlation estimates. Nonetheless, all of the important conclusions drawn below are robust to choosing other reasonable (for example, 100-year) window widths….

We test the significance of the correlation coefficients (r) relative to a null hypothesis of random correlation arising from natural climate variability, taking into account the reduced degrees of freedom in the correlations owing to substantial trends and low frequency variability in the NH series.

We use Monte Carlo simulations to estimate the likelihood of chance spurious correlations of such serially correlated noise with each of the three actual forcing series. For (positive) correlations with both CO2 and solar irradiance, the confidence levels are both approximately 0.24 (90%), 0.31 (95%), 0.41 (99%), while for the “‹Å“whiter’, relatively trendless, DVI index, the confidence levels for (negative) correlations are somewhat lower (-0.16, -0.20, -0.27 respectively). A one-sided significance test is used in each case because the physical nature of the forcing dictates a unique expected sign to the correlations (positive for CO2 and solar irradiance variations, negative for the DVI fluctuations).

The correlation statistics indicate highly significant detection of solar irradiance forcing in the NH series during the “‹Å“Maunder Minimum’ of solar activity from the mid-seventeenth to early eighteenth century which corresponds to an especially cold period. period. In turn, the steady increase in solar irradiance from the early nineteenth century through to the mid-twentieth century coincides with the general warming over the period, showing peak correlation during the mid-nineteenth century. Greenhouse forcing, on the other hand, shows no sign of significance until a large positive correlation sharply emerges as the moving window slides into the twentieth century. The partial correlation with CO2 indeed dominates over that of solar irradiance for the most recent 200-year interval, as increases in temperature and CO2 simultaneously accelerate through to the end of 1995, while solar irradiance levels off after the mid-twentieth century.

Their "not relying" on correlation or correlation squared is further demonstrated in Mann et al. [2000], which re-capitulates the two figures shown above with only slightly different legends.


Original Caption: Figure 4. Spatial patterns of (top) calibration beta, (middle) verification beta, and (bottom) r-squared statistics for annual-mean reconstructions. The calibration statistics are based on the 1902–80 data, while the verification statistics are based on the sparser 1854–1901 instrumental data (see Figure 2) withheld from calibration… For the r-squared statistic, statistically insignificant values (or any grid points with unphysical negative values of correlation) are indicated in gray. The color scale indicates values significant at the 90% (yellow), 99% (light red), and 99.9% (dark red) levels (these significance levels are slightly higher for the calibration statistics that are based on a longer period of time). More details regarding significance level estimation are provided in Mann et al. (Mann et al, 1998). [Reprinted with permission from Mann et al. (Mann et al., 1998).]….

Our winter Nino-3 reconstruction exhibits a highly significant correlation with largely independent reconstruction of the winter (Dec–Jan–Feb) SOI of Stahle et al. (Stahle et al., 1998). The two reconstructions are correlated at r = 0.63 over the full period of overlap (1705–1976) and r = 0.60 during the precalibration interval (1705–1901). This is nearly as high as the observed correlation (r = 0.7) between the instrumental SOI and Nino-3 series during the twentieth century….


Original Caption: Figure 17. Relationship of annual-mean NH mean temperature reconstruction to estimates of three candidate forcings (see Mann et al., 1998) between 1610 and 1995. ….. (e) Evolving multivariate correlation of NH series with the three forcings (a, b, and c). The time axis denotes the center of a 200-yr moving correlation window. Significance levels are based on the null hypothesis that the surface temperature series is a realization of natural variability represented as represented by a red noise process with the persistence structure of the observed NH series (see Mann et al. 1998 for details). One-sided significance levels for correlations with the different forcing agents are shown, under the assumption that only positive relationships with GHG and CO2, and negative relationships with DVI, are physically meaningful. These confidence levels are approximately constant over time and are thus represented by their average values over time for simplicity (although the number of degrees of freedom in the CO2 series is somewhat decreased prior to 1800 when the series is essentially flat, so that the confidence intervals are slightly too liberal in this case). Significance levels for correlations of temperature with CO2 and solar irradiance are nearly identical, and the 90%, 95%, and 99% (positive) significance levels are shown by the horizontal dashed lines. The 95% (negative) significance level for DVI is shown by a horizontal dotted line. The lower dotted line indicates the 99% significance level for correlation with GHG if a two-sided hypothesis test is invoked (this is only added to emphasize that the seemingly spurious negative correlation of NH with GHG apparent during the late eighteenth–early nineteenth century is in fact not statistically significant if the a priori physical requirement of a positive relationship between CO2 and temperature is not taken into account in hypothesis testing). The gray bars indicate two different 200-yr windows of data in the moving correlation, with the long-dashed vertical lines indicating the center of the corresponding windows…

For lags of 10–15 yr the relationship between greenhouse gas (GHG) increases in recent decades and increasing temperatures is considerably more significant, while the relationship with solar irradiance is considerably less significant. For the shorter (100 yr) window there are few enough degrees of freedom in the temperature and forcing series that the statistics are not as stable (i.e., the results are much “‹Å“”‹Å“noisier”). In particular, larger negative correlations with GHGs are achieved prior to 1800 in this case, although these are not significant taking into account the decreased degrees of freedom in the series. Nonetheless, even with the large sampling variations that arise in the 100-yr window case, the relationship between recent warming and increasing greenhouse gas concentrations is the dominant statistical feature. It is evident that the inclusion of a representation of the lagged response of temperatures to forcing heightens the evidence for a recent anthropogenic impact on twentieth century climate beyond that presented in Mann et al. (Mann et al., 1998)

Just to make it totally clear that they did "not rely" on correlation statistics, the SI to the MBH98 Corrigendum in July 2004 stated the following for "each" of the 11 steps in the stepwise reconstruction:

4. Statistical Verification

An essential step in the procedure of Mann et al (1998), as described therein, was the use of conventional verification procedures to establish the level of skill in the proxy-based surface temperature reconstructions. Verification estimates based on correlation and Reduction of Error (‘RE’ or, ‘beta’ in the language of Mann et al, 1998) were established for each of the 11 separate procedures contributing to the stepwise reconstruction procedure, based on comparison of the proxy reconstructions…It should be stressed that reconstructions that did not pass statistical cross-validation (i.e., yielded negative RE scores) were deemed unreliable.

I was about to editorialize a little on this, but words escape me. If this is "not relying" on correlation, I’d hate to think what would happen if they actually relied on it. You’d think that any of Mann’s friends who had read his draft evidence to Barton (or his lawyer for that matter) would have said: Uh, Mike, maybe you’d be better off just answering the question.

44 Comments

  1. John A
    Posted Jan 12, 2006 at 1:19 PM | Permalink

    Can I clarify something, Steve? In Mann’s reply to the House Committee, he notes that there are two statistics with the same name (differentiated only by the capitalisation):

    In order to answer this question it is important to clarify that I assume that what is meant by the “R2″ statistic is the squared Pearson dot-moment correlation, or r2 (i.e., the square of the simple linear correlation coefficient between two time series) over the 1856-1901 “verification” interval for our reconstruction.

    So which statistic are you actually referring to?

  2. Steve McIntyre
    Posted Jan 12, 2006 at 1:58 PM | Permalink

    They are both the same thing. The use of the capitalized form is the usual practice in econometrics and social sciences, while the use of the uncapitalized form is more common in tree ring research. Dano initially twitted me for this elsewhere, but then cordially agreed that both usages were acceptable. Mann is engaging here in a pointless bit of pedantry that a good lawyer would have told him to avoid.

  3. J. Sperry
    Posted Jan 12, 2006 at 2:04 PM | Permalink

    Re #1:
    I don’t think Mann was referring to two different statistics. Instead he was pointing out that the Barton letter used capital R while the convention is lower case r (not that it makes a difference). In other words, Mann said “I assume R2 is r2.”

  4. John Hekman
    Posted Jan 12, 2006 at 2:32 PM | Permalink

    Steve, this whole method of using tree ring width etc. as a proxy for temperature seems questionable because there are both cyclical and (possibly) long-term patterns in temperature. An example from economics would be business cycle effects on GDP growth versus long-term trend growth in GDP. A variable like the unemployment rate is significantly related to cyclical GDP growth (with a negative sign) but has no theoretical or empirical relationship to long-term GDP growth. So if you ran the GDP-unemployment correlation and then tried to use that to get a 200-year history of GDP, it would be garbage.

    Similarly with temperature, one could get a high correlation of tree ring width to temperature because of drought or other short-term cyles of climate, but there might be no detectable signal at all of a long-term “climate change” effect.

  5. John A
    Posted Jan 12, 2006 at 4:11 PM | Permalink

    Mann is engaging here in a pointless bit of pedantry that a good lawyer would have told him to avoid.

    So after all of this bluster, he doesn’t answer the question? Sheesh. That makes it worse.

    What this comes down to, is he did calculate the r2 statistic and didn’t report it because it was insignificant. I believe that you have already calculated a lot of other statistical tests for MBH98 and they lined up with the r2 test and not the RE test.

    What have we done? Some of our leaders have spent lots of time and lots of money based on a mirage.

  6. Steve McIntyre
    Posted Jan 12, 2006 at 5:10 PM | Permalink

    #5. John A. – people have quite reasonably argued that MBH was not the sole ground for their decision-making. So you can’t conclude that the time has necessarily been spent in vain. All you can conclude right now is that the disclosure and due diligence processes are flawed and that the MBH study is flawed.

  7. Steve McIntyre
    Posted Jan 12, 2006 at 5:11 PM | Permalink

    #4 – and when you add data mining to the methodology as in MBH, you get a witches’ brew.

  8. Pat Frank
    Posted Jan 12, 2006 at 5:32 PM | Permalink

    Re: #6 Given the TAR emphasis, we can further conclude that the IPCC is either incompetent or in cahoots.

    The other ground for decision-making, by the way, were the ‘realistic’ climate maps put out by GCM modelers. I think this is a case of being mesmerized by lovely computer graphics into seeing models and thinking ‘data.’ I’ve seen the same thing happening in chemistry (my field), and I find the trend alarming. In any case, the GCMs are incomplete and provide no grounds for prediction. The whole AGW thing (emphasis on “A”), as JohnA observes, really is a mirage.

  9. Posted Jan 12, 2006 at 5:43 PM | Permalink

    Another interesting belief from Mann abour linear correlation. “CLIMATE OVER PAST MILLENNIA” (Jones, Mann 2003) states:

    “Empirical analyses employing simple linear correlations or multivariate regressions between forcing series and climate reconstructions can provide insights into the relative roles of such forcings in past centuries [e.g., Lean et al., 1995; Mann et al., 1998a, 2000a; Beer et al., 2000; Waple et al., 2002]. Model simulations, however, driven with estimated forcings (see section 5.2) and comparisons of these model simulations with empirical paleoclimate reconstructions (see section 5.3) will likely yield more detailed physical insights.

  10. Steve McIntyre
    Posted Jan 12, 2006 at 6:07 PM | Permalink

    #9 – nice spotting, Justin. If we consider Kaufmann’s findings, it looks like the second sentence is untrue.

    I’ve really spent no time with the forcing correlations in MBH98 etc., but it looks increasingly like a good project.

  11. Douglas Hoyt
    Posted Jan 12, 2006 at 7:21 PM | Permalink

    These authors state: “For lags of 10–15 yr the relationship between greenhouse gas (GHG) increases in recent decades and increasing temperatures is considerably more significant”, which implies a short time constant for climate response which, in turn, implies a low climate sensitivity compared to IPCC claims.

    On another note, the study includes 3 types of forcings but omits forcings due to land cover changes. As a first approximation, land cover changes could be approximated as being proportional to the logarithm of population density. It would have regional variations, of course, and on the global scale would look almost identical to greenhouse gas forcing. They probably don’t reference any papers by Pielke or Kalnay. The fact that they fail to include this forcing means their study should be taken with a grain of salt.

  12. Posted Jan 12, 2006 at 8:34 PM | Permalink

    I haven’t tracked this site for all that long, though I have gone back and read many of the past postings. I apologize if this has been stated elsewhere, but it is unclear to me if Mann has released all of his data and/or source code. And with the data and source that has been released, is it possible to mimic his calculations for the missing correlation statistics in question?

    Re: my previous post, there is also an interesting paragraph in “Testing the Fidelity of Methods Used in Proxy-Based Reconstructions of Past Climate” (Mann et.al. 2005), that discusses the value in using R^2 for evaluating statistical reconstruction skill in certain situations where RE is inaccurately “enhanced”.

  13. Steve McIntyre
    Posted Jan 12, 2006 at 10:25 PM | Permalink

    #12: 1) Mann has not released all of the statistics that he calculated. He has withheld the verification R2 for the various temperature averages and the RE for the El Nino.
    2) He has released source code purusant to Barton in summer 2005, which I’ve posted about, and data in the Corrigendum SI. Unfortunately the two data sets do not match so that you can’t reproduce any of his output series from the source code that’s been produced so far. You can get somewhat close to his reconstruction but so far it’s impossible to exactly replicate it. There’s no reason for this – for example, I can exactly replicate the Wahl and Ammann replication, which was almost exactly the same as ours. They tried to make it seem like they were a lot different but that was just disinformation.

    The reason for insisting on the cross-validation statistics is not because I don’t know what they are; it’s that they haven’t admitted it. In the case of Ammann, this is going from the sublime to the ridiculous – he has got almost exactly the same verification statistics in his reconstruction as we reported in our GRL article and yet has the gall to say that our claims are “unfounded” all the while withholding his R2 calculations. I will refrain from adding an adjective as the facts yield many obvious adjectives without my editorializing.

  14. Peter Hearnden
    Posted Jan 13, 2006 at 3:50 AM | Permalink

    Re #6. You know what, Steve? I can actually live with this.

    Of course people like me don’t base our view just on MBH ’98, and thus you can’t conclude as ‘John A’ does. And, given the time you’ve spent on these matters, to come up with ‘that the MBH study is flawed.’ when we all KNOW it isn’t perfect (is any science, or science process?) sends us back to the start.

    Science moves on, other recons are produced (not by you though or any of your ilk). People look at trees and see methane (but it’s one study so we’ll see). CA comments section continues it’s merry role of hounding and ridiculing those who have the temerity to comment but off message. On message contributors continue to claim that your pointed out ‘flaws’ have overturned the entire science of climatology – yes, honestly. People here continue to look inwards not out of the window, to critice not observe.

    I rekon skirmishes will continue. It wont be clear who’s ‘right’, one way or another, for a few years, maybe a decade or two. And, for all you sceptics, the news is ‘good’. Nothing serious will be done about the problem (OK ‘problem’ if you like) during that time – so your money is safe :)

  15. Hans Erren
    Posted Jan 13, 2006 at 5:05 AM | Permalink

    Science moves on, other recons are produced (not by you though or any of your ilk).

    1) Did you read the comments placed here on this blog on these “other recons”?
    2) Who is Steve’s ilk?

  16. per
    Posted Jan 13, 2006 at 6:45 AM | Permalink

    Re: #14

    “to come up with “that the MBH study is flawed.’ when we all KNOW it isn’t perfect… “

    well, actually, we didn’t all KNOW that it wasn’t perfect. In fact, the IPCC TAR highlighted this study, and used it in a great deal of its publicity material, and I very much doubt whether they would have done that with a flawed study. As was, the MBH study was untouchable, had passed the most prestigious peer-review into Nature, and was widely lauded.

    Oh, how times change. We now KNOW that the original materials and methods was seriously defective; that there are undocumented procedures in the PC analysis that bias the results; that it relies upon a very few pine and cedar records to get the results that it does, and that these tree records are not related to temperature; and that MBH withheld adverse statistics.

    Isn’t KNOWLEDGE a great thing ?
    yours
    per

  17. John A
    Posted Jan 13, 2006 at 7:45 AM | Permalink

    Of course people like me don’t base our view just on MBH “98, and thus you can’t conclude as “John A’ does.

    Can’t help putting words in my mouth, can you Hearnden?

  18. Steve McIntyre
    Posted Jan 13, 2006 at 8:24 AM | Permalink

    When I studied math at university, “great” papers were ones that students could continue going back to. You don’t just “move on” from Gauss or Galois. If MBH was any good, then it should have led to insight. It doesn’t. As to the new papers, they really aren’t any better. I don’t think that any of these papers will stand the test of time. I get tired of the nomadic nature of these guys – they always “move on” when you look closely at something. There’s something ne; trust us, but you can’t see the data.

  19. muirgeo
    Posted Jan 13, 2006 at 8:55 AM | Permalink

    Re # 18

    Steve,

    Sorry if you’ve answered this else where. John Daly, I believe, was the first to bring up the hockey stick issue. My challenge to him was to show us a published reconstruction that shows something significantly different from the Mann or other existing studies. The challenge went unanswered until his untimely death. I truly wish the poor fellow would have lived right up to 2050. My challenge to him continues to you. If the Hockey stick is not valid show me what is the most valid estimate of recent Holocene climate. If you are apparently the only one capable of getting the statistics right why not forget all the others and their flawed studies and publish you own multiproxy study. This way we can see what results we get when its finally done right by some one who is actually capable of handling all the statistics, the rigorous methodologies, capable of spotting the “useful” proxies from the non-usable and is also unlike the others highly ethical and well intended. Please tell me you’re working on it.

  20. Hans Erren
    Posted Jan 13, 2006 at 9:12 AM | Permalink

    http://www.ncdc.noaa.gov/paleo/pubs/mangini2005/mangini2005.html

    ABSTRACT:
    The precisely dated isotopic composition of a stalagmite from Spannagel Cave
    in the Central Alps is translated into a highly resolved record of temperature
    at high elevation during the past 2000 yr. Temperature maxima during the
    Medieval Warm Period between 800 and 1300 AD are in average about 1.7°C higher
    than the minima in the Little Ice Age and similar to present-day values.
    The high correlation of this record to d14C suggests that solar variability
    was a major driver of climate in Central Europe during the past 2 millennia.

  21. Igor J.
    Posted Jan 13, 2006 at 9:28 AM | Permalink

    Re 18: “As to the new papers, they really aren’t any better.”

    Steve, I understand that you were able to replicate MBH’s work and even find some statistical problems in it, so you seem to be skilled to do the research. At the same moment you seem to be really interested in temperature reconstructions.

    ==> Isn’t it time to come up with some independent temp. reconstruction of your own, i.e. not to be only looking for errors in other studies but to write one “correctly” from scratch??? [[to find an error in other study only says that the study is wrong but does not really say what is the right answer to the original question etc...]]

    IJ.

  22. John A
    Posted Jan 13, 2006 at 9:44 AM | Permalink

    John Daly, I believe, was the first to bring up the hockey stick issue. My challenge to him was to show us a published reconstruction that shows something significantly different from the Mann or other existing studies. The challenge went unanswered until his untimely death.

    Crap. There are plenty of recontructions of past temperature that John Daly referred to, including the IPCC’s own reconstruction from 1995. The fact that you’ve closed your eyes and put fingers in your ears about them doesn’t mean that John Daly didn’t provided ample evidence, well before anyone else, on how the Mann Hockey Stick was radically different from all previous reconstructions.

    Which planet are you on to be making such ridiculous statements? Because it’s clearly not the Earth

  23. Brooks Hurd
    Posted Jan 13, 2006 at 10:20 AM | Permalink

    Muirgeo,

    The reconstructions by Mann, Jones, Crowley, et al are based on one underlying assumption: tree ring widths and other proxies are directly related to temperature and other factors have minor influences. I have not seen a calibration of these proxies to temperature or other factors.

    Can you provide a link to studies which attempt to calibrate the proxies used by Mann, Jones, Crowley et al? With all the countless billions of dollars going into climate research, surely someone has tried to calibrate these proxies.

    Mann, Jones, Crowley, et al assume that their proxies are predominantly indications of temperature. Is this based on anything beyond their assumptions?

  24. JerryB
    Posted Jan 13, 2006 at 10:57 AM | Permalink

    It seems that some people have difficulties coping with the idea of their heroes’ work being diagnosed/audited, and would rather that Steve stop doing what he seems to do well, and to enjoy doing. Instead, he should do something else, preferably something very time consuming, and make life a little easier for some so called “climate scientists” to pretend their concoctions should be regarded as “science”.

  25. Igor J.
    Posted Jan 13, 2006 at 11:41 AM | Permalink

    Re 24:
    “…heroes’ work being diagnosed/audited…”
    Most times the fastest audit method is actually to do/calculate the whole thing by your self than to try to go through all the formulas and calculations done by the “audited”. Then you simply compare the results and resolve the possible (statistically) significant differences…

    The discussion like “you should look on r^2″ vs “no we should not, because…” is not helping much to get the correct answer…

  26. per
    Posted Jan 13, 2006 at 12:35 PM | Permalink

    re: #25
    Igor
    the r2 issue is more central than your comment might suggest.

    MBH’s method is such that they set up a statistical test as a necessary validation of their method; they themselves are saying that without that validation, their method fails. There are multiple methods of testing significance, and MBH used r2 and RE. The RE “worked”, and those were the results they published. They didn’t publish that they had done the r2, and found that their method failed.

    Whether you like it or not, that is incredibly sharp practice, and some might say it is on the fine line separating bad scientific practice from misconduct.
    yours
    per

  27. muirgeo
    Posted Jan 13, 2006 at 1:09 PM | Permalink

    Re # 20

    Thanks Hans and thanks for defending me in your prior post. You are right I’m no sock-puppet.

    The graphs you reference are very interesting. It is my guess that they might reasonable represent the conditions as they might look if we had a very good global multi proxy study. Indeed it is significantly different from the hockey stick. In fact, I posted a comment the other day stating that even if you took the extremes of the error bars of the Mann study the difference in the peak warmth of the MWP and the trough coolness of the LIA would be about 1.3 C…similiar to the graphs you’ve cited. My point remains. On the graphs you’ve provided extend the x-axis 100 more years and the y-axis up 2.0 c. We are currently warming about 0.2 C per decade. So if you assume this for the next 100 years complete the graph and look at it as our children’s children may be seeing it. No hockey stick but now more of a sling blade…ummmm hhmm…

    Am I wrong to be concerned the graph will/might look so in 100 years? It’s quite dramatic.

  28. Michael Jankowski
    Posted Jan 13, 2006 at 1:19 PM | Permalink

    If you are apparently the only one capable of getting the statistics right why not forget all the others and their flawed studies and publish you own multiproxy study.

    For starters, it goes well-beyond the statistics. Take a look around at the articles on this site which refer to deficiences with the proxies themselves.

    Even if the proxies were valid, do you really think we have enough spatial coverage to accurately represent global temperatures of the past? Take a look at the proxy locations in this Mann and Jones paper http://www.ncdc.noaa.gov/paleo/pubs/mann2003b/mann2003b.html Do you really think that’s enough coverage to represent the “average global temperature” of our past with any degree of accuracy?

  29. muirgeo
    Posted Jan 13, 2006 at 2:11 PM | Permalink

    RE # 23

    Mann, Jones, Crowley, et al assume that their proxies are predominantly indications of temperature. Is this based on anything beyond their assumptions?

    Comment by Brooks Hurd

    Brooks,

    Let me give you the short answer….mmmmm….Yes.

  30. muirgeo
    Posted Jan 13, 2006 at 2:15 PM | Permalink

    RE # 24

    Jerry I have no problem with M&M or anyone trying to debunk Mann. But assuming they HAVE debunked them as so many of you are willing to believe, then aren’t you even the least bit curious what the curves really should look like?

  31. jae
    Posted Jan 13, 2006 at 2:34 PM | Permalink

    muirgeo:
    You keep insisting that Steve should conduct his own proxy study. Why? It looks to me like this is a flawed way to look at climate changes and Steve would be wasting his time: (1) it is probably not possible to cover enough of the planet to make the data represent “global temperatures” (whatever that really means); (2) It looks like it hasn’t been demonstrated that many (if not all) of the proxies, expecially tree rings, can be used to track temperature; (3) even if a proxy were shown to track temperature well, we are still left with the question of what caused the temperature changes; (4) there is so much evidence that any warming that is occurring is part of one of the many natural cycles that the Earth goes through. Why don’t YOU fund a correctly done proxy study?

  32. Tony Knudson
    Posted Jan 13, 2006 at 2:43 PM | Permalink

    RE # 27

    Extrapolation and prediction beyond existing data is a dangerous game. Especially with huge leaps in extrapolation of 50 or 100 years. One has to assume the increase in temperature over the last 100 or so years will continue at the same rate. From graphs of temperatures over the last few centuries, I don’t think that is a good assumption. The temperature could start to decrease after peaking in a decade or so. If you graph a stock for a few weeks and it is trending higher, can we predict what the stock is worth in 10 years? Climate data and financial data both suffer from persistence, so similar statistical methods can be used in their analysis. One can’t predict the future in the stock market or the climate, although many people will be willing to sell you software to do the former.

  33. muirgeo
    Posted Jan 13, 2006 at 2:44 PM | Permalink

    Re # 28

    Do you really think that’s enough coverage to represent the “average global temperature” of our past with any degree of accuracy?

    Comment by Michael Jankowski “¢’‚¬? 13 January 2006 @ 1:19 pm

    Mike,

    If you go to the station data site at NASA/GISS and randomly pick one rural station with about 100 years of data from each continent and then constructed a global trend from these 6 or 7 trends I think you’d be surprised how similar the results would look to the trends using the full data set (most of the times) . Again, imprecise does not mean useless.

    http://data.giss.nasa.gov/gistemp/station_data/

    http://data.giss.nasa.gov/gistemp/graphs/

  34. muirgeo
    Posted Jan 13, 2006 at 2:47 PM | Permalink

    RE # 31

    Why don’t YOU fund a correctly done proxy study?

    Comment by jae “¢’‚¬? 13 January 2006 @ 2:34 pm

    I already have. See Mann and the spaghetti graphs. I’m not the one who is casting doubt and claiming I have a better way to make widgets.

  35. Brooks Hurd
    Posted Jan 13, 2006 at 4:03 PM | Permalink

    Re: 33, Muirgeo

    I have done this for US stations. In addition, Hans Erren provided a link to Central European temperatures. I graphed the results to see what they looked like.

    The results do not validate the Mann, Crowley, and Jones reconstructions. Furthermore, they do not look like Jones’ instrument data.

    I do not make the claim that what I have graphed is a representation of “global temperatures”. However, I do not understand how Mann, Crowley, and Jones can claim that their work represents “global temperatures.” They have not, so far, released their data and their methodologies.

  36. Brooks Hurd
    Posted Jan 13, 2006 at 4:19 PM | Permalink

    Muirgeo

    My question in 23 had to do with calibrations: “Can you provide a link to studies which attempt to calibrate the proxies used by Mann, Jones, Crowley et al?”

    You indicated that calibrations exist in your partial answer above (29): “Let me give you the short answer….mmmmm….Yes.”

    OK, where are the calibration studies?

  37. Armand MacMurray
    Posted Jan 13, 2006 at 5:01 PM | Permalink

    Re: #34

    I’m not the one who is casting doubt and claiming I have a better way to make widgets.

    You’ve misunderstood again. No one here is claiming to have a “better way,” all we are claiming is that MBH doesn’t actually make proper widgets. Perhaps nobody knows how to make widgets properly; in any case, don’t you think that populations being asked to spend billions of dollars for what are claimed to be proper widgets should be told that the claimed widgets don’t live up to the manufacturer’s promises?

  38. muirgeo
    Posted Jan 13, 2006 at 5:16 PM | Permalink

    Re: 36

    See these;

    (Briffa et al., 1995; D’Arrigo et al., 1996; Jacoby et al., 1996; D’Arrigo et al., 1998; Wiles et al., 1998; Hughes et al., 1999; Cook et al., 2000, Briffa et al., 1996, Briffa et al., 1998b; Briffa, 2000).

    When you’re done I have more…lots more.

    Here’s a good textbook discussion;

    http://www.ace.mmu.ac.uk/Resources/gcc/3-3-3.html

  39. muirgeo
    Posted Jan 13, 2006 at 5:20 PM | Permalink

    Re #37

    Armand,

    When you say stuff like that above….are you able to say it and still feel intellectually honest? The whole of paleoclimatology swished a way on a whim because you don’t like the results?

  40. Armand MacMurray
    Posted Jan 13, 2006 at 5:31 PM | Permalink

    Re: #39
    I don’t remember equating MBH with “the whole of paleoclimatology;” if you really think that is true, you may want to widen your reading list. You might also consider toning down your imagination and putting a bit more effort into reading what is actually written. Just a suggestion. :)

  41. Brooks Hurd
    Posted Jan 14, 2006 at 4:29 PM | Permalink

    Re: 38

    All that I asked for was a link to a calibration study, preferably one which compares different proxies. I did not read a description of any calibration study in Mann’s work. Is their a weather station in close proximity to the bristlecones which have such a major impact on the MBH reconstruction?

    The link you provided was a very good explanation of how one would perform a calibration. Thanks, but I really do understand how to do this since I have performed numerous calibrations over the past 25 years. There is a difference between writing a good description and actually performing the work. It is the latter that I am interested in reading about.

  42. john emery
    Posted Jan 14, 2006 at 7:08 PM | Permalink

    muirego

    Sometimes, you almost make sense; other times you are very strange. I think you should not post when you are drinking or toking.

  43. Steve Bloom
    Posted Jan 15, 2006 at 9:03 PM | Permalink

    Re #42: Careful, if you apply that principle across the board you’ll put a stop to the bulk of the activity on this blog!

  44. Michael Jankowski
    Posted Jan 16, 2006 at 11:56 AM | Permalink

    If you go to the station data site at NASA/GISS and randomly pick one rural station with about 100 years of data from each continent and then constructed a global trend from these 6 or 7 trends I think you’d be surprised how similar the results would look to the trends using the full data set (most of the times).

    I’ve toyed with that before (randomly selecting stations in the area of Mann and Jones’ 2003 proxies) and not been impressed. If such a few samples could be used to represent the average global temperature to some degree, it should have been justified statistically long ago in a publication. Of course, as mentioned in my previous posts, that’s assuming the proxies themselves don’t have issues, which they most certainly do.

    Again, imprecise does not mean useless.

    What you are looking for here is accuracy, not precision.

Follow

Get every new post delivered to your Inbox.

Join 3,329 other followers

%d bloggers like this: