Bristlecone dC13

I’ve recently run across an article on changing water use efficiency in bristlecones, which pretty much put the nail in the coffin on any lingering ideas that 20th century bristlecone ring widths might be a temperature proxy. Tang et al. [1999], "The dC13 of tree rings in full-bark and strip-bark bristlecone pines in the White Mountains of California", shows a remarkable non-climatic increase of dC13 ratios of bristlecones at Sheep Mountain, the most important site in the MBH98 PC1. The dC13 values are stated to show an equally remarkable nonclimatic increase in water use efficiency at Sheep Mountain. Improved water use efficiency was the predicted mechanism for CO2 fertilization. See their Figure 3 below. Any attempt to argue that bristlecones are a temperature proxy on scientific grounds (something that has been conspicuously absent from any response by realclimate or their associates) would need to adjust for non-climatic changes in dC13 ratios and water use efficiency.

Bristlecone pines (and Gaspé cedar) have obviously featured prominently in our critique of MBH98. In our EE article, we included an extensive literature review of issues affecting the validity of these ring widths as temperature proxies, but weere then unaware of Tang et al [1999] (which I’ve posted up here). To my knowledge, neither Mann nor any realclimate associate has attempted to defend bristlecones as temperature "proxies" other than by arguing [Wahl and Ammann] that their use improves the reconstruction RE score ( a statistical argument leading directly into questions of spurious significance which they avoid.)

Where are the defences of bristlecones as valid proxies in tree ring terms? While we’ve pointed out a number of possible mechanisms for fertilization in the 20th century, the most prominent theory is CO2 fertilization [Graybill and Idso, 1993; Lamarche at al, 1984.] The usual mechanism by which CO2 fertilization is hypothesized to lead to increased ring widths is through increased water use efficiency. Water use efficiency changes are studied in a series of remarkable articles by Xiahung Feng of Dartmouth, of which Tang, Feng and Funkhouser [1999] is notable. "Plant water use efficiency" was defined by Farquhar et al 1989 (cited by Tang et al) as the ratio of net carbon fixed to the total water cost. The connection between CO2 fertilization and increased water use efficiency (W) is summarized by Tang et al. as follows:

Experimental work has strongly demonstrated the positive response of photosynthesis and plant water use efficiency to increasing CO2 concentration [ e.g. Strain and Cure 1985; Bazzaz 1990; Mooney et al 1991; Idso 1992; Korner and Arnone 1992; Norby et al 1992; Polley et a 1993; Wullscheger et al 1995)] and the negative response of stomatal conductance of plant leaves [Woodward 1987; Beerling and Woodward 1993; van de water et al 1994]. For example, by studying a number of C3 and C4 species, Polley et al 1993 showed that both plant water use efficiency and biomass increased with increasing ambient CO2 concentration. This led to the idea that CO2 fertilization may be evaluated by measuring plant water use efficiency.

Tang et al. discuss findings of Graybill and Idso and proceed to discuss whether water-use efficiency of bristlecones has increased at this site; and the relationship of stem growth to strip bark forms. They cored trees in June 1994, 5 of which were the SAME trees as studied by Graybill and Idso 1993 and analyzed 4 trees. Tang et al. :

To separate the atmospheric CO2 signal [direct fertilization] from the climatic noise, we assume the variation of dC13 with CO2 concentration is long-term and of low frequency, while climatic and other environmental factors contribute only to high frequency fluctuations in a dC13 tree ring series. This assumption has been shown to be essentially valid for this site [Feng and Epstein 1995; Feng 1999]. The superposition of high-frequency with low-frequency variations can be easily identified from the dC13 time series in Figure 3. The long-term trends mimic the trend of dC13(atm) and form a mirror image compared with the CO2 concentration of the atmosphere. We consider that the trends of carbon isotopes of tree rings contains the signal of atmospheric conditions [CO2 concentrations] for the past 200 years. …

Our results compare well with previously published results. Using carbon isotope data of tree rings from the White Mountains published by Epstein and Krishnamurthy 1990 and Leavitt and Long [1992], we have calculated water use efficiently for these trees and found that the water use efficiency shows the same trends as trees we present in their study. gh frequency variations of dC13 of tree rings for trees in an arid environment often correlate with he amount of annual precipitation..

Original Caption: Figure 3. Carbon isotopic compositions of tree rings for bristlecone pines from Sheep Mountain. The dC13 values were obtained from each tree for at most every 5th tree ring from 1796 to 1995. The smooth solid lines are modeled curves assuming exponential trends.

They removed the high-frequency variations and compared the low-frequency results in their Figure 5. Tang et al:

Figure 5 indicates that for all trees analyzed, water use efficiency increased with concentration of atmospheric CO2 increased. This may be related to the increase in the growth rate observed from tree ring widths by Graybill and Idso 1993 for bristlecone pine on Sheep Mountain and earlier by Lamarche et al 1983 in the White Mountains….

It is possible that the partitioning of biomass between roots and shoots was systematically different for the full-bark than for the strip-bark trees. For young full-bark trees to maintain foliage and reproductive growth, much assimilated carbon is allocated to the root system, Older strip bark trees may also be doing this to some degree but they use a substantial fraction of fixed carbon for cambial growth Graybill and Idso 1993. ..Using carbon isotope data of tree rings from the White Mountains area published by Epstein and Krishnamurthy 1990 and Leavitt and Long 1992, we have calculated water use efficiency for these trees and found the water use efficiency shows the same tends as trees we present in this study.

Original Caption: Figure 5. The relative rate of change in water-use efficiency for full-bark (solid) and strip-bark (dashed) bristlecone pine trees on Sheep Mountain. There is no significant difference in the relative rate of change in W between full-bark and strip-bark forms.

They also showed the climatic high-frequency variations after taking out the non-climatic increase in water use efficiency:

Original Caption: Figure 4. Correlation of high frequency variations among four bristlecone pines. The detrended value of dC13 represented the difference between the measured dC13 values of a given ring of a given tree and the value of the exponential trend at that year for that tree.

Funkhouser, a coauthor of Tang et al. [1998], is an associate and sometime coauthor of Hughes, the H of MBH. The potential for CO2 fertilization to distort tree ring proxies was specifically noted as a caveat in the 1995 IPCC Second Assessment Report concerning tree ring proxies. MBH99 touched on CO2 fertilization, referring to CO2 fertilization as a potential problem, but did not discuss the dC13 evidence. Astonishingly (and only for pre-1400 results), they adjusted 19th century values of the NOAMER PC1 and bizarrely claimed that the CO2 fertilization effect had somehow reached a "saturation" point in the 20th century – a claim not supported by any evidence and obviously at odds with the explicit statements of Lamarche et al [1984], Graybill and Idson [1993] and the dC13 evidence of Tang et al [1999] and predecessor studies.

Tang et al. [1999] confirms that there has been an important non-climatic effect on bristlecone pines, creating a trend. Because of non-existent statistical control in MBH98-99 for spurious relationships, this non-climatic trend had a spurious relationship with the temperature PC1, which imprinted the NH temperature reconstruction as we’ve pointed out in our articles. Bristlecones do not just affect MBH; they also are integral to Crowley and Lowery [2000] and Mann and Jones [2003].
Epstein S. and Krishnamurthy R. V. (1990) Environmental information in the isotopic record in trees. Philos. Trans. R. Soc. London Ser. A 330, 427–439.
Xiahong Feng and Samuel Epstein, 1995, Carbon isotopes of trees from arid environments and implications for reconstructing atmospheric CO, concentration, Geochimica et Cosmochimica Acta, 59( 12),. 2599-2608.
Xiahong Feng, 1998. Long-term ci/ca response of trees in western North America to atmospheric CO2 concentration derived from carbon isotope chronologies, Oecologia (1998) 117:19±25 8b.
L Tang K., Feng X., and Funkhouser G. S. (1999) The d13C of tree rings in full-bark and strip-bark bristlecone pine trees in the White Mountains of California. Global Change Biology 5(1), 33–40.


  1. Douglas Hoyt
    Posted Aug 28, 2005 at 12:57 PM | Permalink

    For future reference, would you list the references to other temperature reconstructions that use the bristlecone pines? It would seem to me that the studies that “confirm” the MBH claims may actually all be wrong in the same way via the common use of bristlecone pines.

  2. Steve McIntyre
    Posted Aug 28, 2005 at 1:23 PM | Permalink

    It’s not just bristlecones. There are other issues affecting other multiproxy studies. But bristlecones definitely have a major impact on Crowley and Lowery 2000 and Mann and Jones 2003.

    Bristlecones (White Mountains, Central Colorado) are 2 of 13 series actually used in Crowley and Lowery. (2 of 15 series, 2 with high MWP, are not used in his reconstruction. Crowley doesn’t know where he got the actual digital versions from. These plus the questionable Polar Urals and Dunde series impart whatever hockey stick-ness there is to CL 2000.

    Mann and Jones [2003] has a North American PC1 which is just bristlecones. I haven’t been able to relicate the methods of Mann and Jones 2003, but my surmise is that the bristlecones impart hockeystick-ness to Mann and Jones 2003. There are some other data mined series that have an impact – the Jacoby Sol Dav series, for example.

    Jones and Mann [2004] use a North American PC1, again a bristlecone series (also Dunde and Polar Urals.)

    Moberg uses several bristlecone series: Methesulah Walk, strangely an obsolete version of the White Mountains series ending in the 1960s and one other. They are not material to the "low frequency" shape, since he does not use tree ring results for his "low frequency". Moberg puts a negative coefficient on the bristlecones, since he calculates that they are inversely related to temperature.

    Jones et al 1998 does not use bristlecones. Its hockeystick-ness is dependent on Polar Urals and Tornetrask, both of which I’ve discussed at length. It also uses the data mined Jacoby northern treeline series, which I’ve posted about.

    Briffa et al 2001 has not released its sites. I don’t think that they use bristlecones. But this index goes DOWN after 1960 – so they delete this portion in the spaghetti diagrams (including in IPCC) without telling anyone – I’ve posted on this.

  3. Ray Soper
    Posted Aug 28, 2005 at 2:41 PM | Permalink

    The MBH98 Debate for Dummies

    I am doing the best that my limited capacities allow to try to keep up with the fascinating debate on these pages. Would you mind if I try to summarise the Bristlecone issue in simple terms for fellow lay readers? Maybe there is a simple summary somewhere on this site. If so, I have missed it.

    1. MBH proposition is that the thickness of the annual rings on living and fossil Bristlecone pine trunks over the past 1000 years is a proxy for temperature.
    2. The rings thickened during the 20th century thus “proving” that temperatures were warmer than the previous 9 centuries.
    3. The “proven” rise in temperature correlates with rising CO2 levels and is therefore caused by the rising CO2 levels.
    4. The rising CO2 levels are due entirely to man’s activities in burning fossil fuels.
    5. The conclusion from this chain of “logic” is that if the world population wants to stop temperatures rising by perhaps 5-6 deg C over this century, we have to work together to get CO2 levels back down, and temperatures will follow.

    The opponents argue:

    1. That on available data released by MBH there are clearly errors and anomalies in how MBH drew their conclusions. MBH are not exactly being helpful in tracking these errors and anomalies down, and correcting them.
    2. There may be factors other than temperature that correlate with, and perhaps drive, the thickness of the Bristlecone pine annual rings. These include differences in water-use efficiency that may have caused more growth in the 20th century. They could also include CO2 fertilisation due to the increased levels of CO2 in the atmosphere.
    3. It is not demonstrated that temperatures are in fact rising. There are significant difficulties in measuring the temperature of a large complex system such as earth, where there are many heterogeneous process going on. This is especially true when we try to gauge average temperatures going back centuries.
    4. It is not demonstrated that even if temperatures are rising, that they correlate with or in any way are caused by rising CO2 levels. It is much more likely that temperature fluctuations correlate most closely with solar activity.
    5. It is also not demonstrated that the rising CO2 levels (at least that much seems to be not controversial) are caused by man’s activity. We don’t even seem to know how much of the rising CO2 levels are accounted for by man’s activity. There are natural sources of CO2 that can make a significant impact – bush fires, volcanoes etc. There are also natural systems that regulate CO2 levels including absorption by the ocean and forests, and we understand little about these systems.
    6. It is not demonstrated that reducing anthropogenic CO2 emissions will reverse rising CO2 levels.
    7. It is not demonstrated that lowering CO2 levels will in any case lead to a global cooling.

    There are clearly many questions involved that require rigorous gathering of data, and rigorous application of scientific method and sound scientific practice. Given that the Kyoto proponents are proposing the expenditure of trillions of dollars and the disruption of whole communities, and even whole economies, it is incumbent on all to adhere to the highest standards of objective science to ensure a) that we understand what is going on and b) are choosing the right course. Joe Barton is merely being a responsible custodian of public moneys in asking the questions that he is.

    It is also incumbent on all to avoid the temptation to resort to sloganeering propaganda such as “the consensus of scientists” (never supported by numbers, and demonstrably wrong) agree (science was never decided by opinion polls – ask Copernicus!) that global warming is occurring, that it is caused by rising CO2 levels, that the rising CO2 levels are caused by man’s activities, and that if we want to stop temperatures rising, we must find a way to bring CO2 levels back down again.

    We need the leading journals in the field to demonstrate their understanding of the issues, and their commitment to truth and sound scientific practice. Unfortunately, many of those journals seem to have lost their way and forgotten that their main role is to educate the wider public on matters scientific.

  4. Ross McKitrick
    Posted Aug 28, 2005 at 4:41 PM | Permalink

    Re #3: Points 1-7 stake out more territory than is argued by Steve and I in our papers. The specific point in this thread, which relates to the discussion at large on this site, is that the hockey stick graph is an incompetent representation of world climate history. The hockey stick matters because of its dominant role in IPCC Reports, as well as many other continuing citations (see the ‘Spot the Hockey Stick’ thread on this site for more examples). The methodology of Mann et al was not properly disclosed in their papers: it contained an algebraic tweak that heavily favours hockey stick shapes. It loaded heavy weighting on the bristlecone pine series, overstated the bristlecones’ explanatory power for the temperature data and overstated the overall statistical significance of the results. The hockey stick data base looks large but most of it is just for show: the bristlecones determine the shape and the appearance of statistical power (the Gaspe cedar series enhances the effect). Remove them from the data base and the famous results collapse. Even if the bristlecones were temperature proxies this would indicate the fundamental weakness of the hockey stick: the fact that the bristlecones do not even constitute a temperature proxy renders the graph worthless.
    At that point, the fact that these problems were never discerned by the publishing journals during peer review, nor by the paleoclimate community itself in follow-up analysis over 7 years, nor by the IPCC, government ministries and other high-level authorities before they grabbed the hockey stick graph and began promoting a policy agenda with it; puts all sorts of other questions on the table. If they got this argument that wrong, how do we know they didn’t get their other arguments wrong too? What exactly is the level of due diligence being applied to climate studies and assessment reports that now loom so large over the world’s energy policy agenda?
    So questions 1-7, and many others, do follow on from this episode, beyond the specific issues arising from Steve’s and my papers.

  5. T J Olson
    Posted Aug 29, 2005 at 4:40 AM | Permalink

    I appreciate Ray Soper’s attempt (in comment 3) to distill much of the debate and ongoing inquiry in “MB98 Debate For Dummies” since I belong to that class of interested but informed observers to the high-level debates pro- and con- like those covered above and elsewhere (such as ray makes a valiant stab. We both, if I may speak for us, find the ebb and flo, the patience and surprise, and the puzzle-finding and probing stimulating – an interesting window on how scientific debate forms and evolves. We feel priviledged to lurk and honored that Ross would read and respond: thank you.

    Ross (in comment 4) quite rightly points out that M&M’s work is necessarily narrow, as science goes – it critiques paleoclimatology as well as deficient practices currently accepted by insitutional science itself, courtesy of big government funding. However, because this focused criticism hits priviledged “royalty,” its consequences are larger than than the former might imply.

    Ray’s “Opponents Points” 3 and 4 are perhaps the most debated, but point 5 – that humans have increased atmospheric CO2 by roughly one-third (although sometimes one reads one-fourth) since the industrial revolution – is not controversial. The precise measure is 19 percent increase from 1958 to 2003 as measured at Mauna Loa, Hawaii.
    Thus, anthropogenic increase in CO2 among the several greenhouse gases is not disputed.

    What the consequences of anthropogenic increases in CO2 are, how we can know it (the debate Ross sums up in post 4), and what we as humans can do about it (IPCC science summaries, Kyoto Treaty, or alternatives to it), if anything, are the whole enchilada. This is the AGW issue we read about in the news,

    In April, I attended the annual meeting of the American Geographical Association in Denver. I mentioned to a grad student doing her novice presentation in bio-geography that the public enjoys dendrochronology. It’s intuitively appealing. But as Steve’s nice summary above shows, it poses many technical mysteries to the nonspecialist.

    Thus, Ray puts the pro-AGW side at point 2 inadequetly as: “The [tree] rings thickened during the 20th century thus ‘proving’ that temperatures were warmer than the previous 9 centuries.” MBH uses some 30 or 40 proxies, including modern temperature methods during the last century. What makes Steve’s current post vital is an earlier post (Steve, February 20, 2005) on the inadequecy of norming proxy measures like tree rings when virtually all of them are so seriously dated.

    “if IPCC or others want to use “multiproxy” reconstructions of world temperature
    for policy purposes, stop using data ending in 1980 and bring the proxies up-to-
    date. […]
    Let’s see how they perform in the warm 1990s — which should be an ideal period
    to show the merit of the proxies. I do not believe that any responsible policy-
    maker can base policy, even in part, on the continued use of obsolete data ending
    in 1980, when the costs of bringing the data up-to-date is inconsequential
    compared to Kyoto costs.”

    In other words, cross checking and cross correlating dendochronological proxy temperature measures with modern satellite measures are called for but aren’t available without such updates. Worse, MBH uses Jones’ “more lurid CRU surface temperature history,” as Steve put it. This and related mysteries, I think, led him to Tang et. al. and the current post.

    Steve has wondered if increasing CO2 concentrations don’t confound contemporary dendrochronological temperature proxies, and the news above is that it does indeed. What can or cannot be done about this I hope subsequent commenters might help us know.

  6. Knut Knutsen
    Posted Aug 29, 2005 at 5:34 AM | Permalink

    So, since you are an expert. Please let us know the reasons carbon isotope ratios in wood cellulose should lead to a dismissal of relationships between ring with and ring density. Why should d13C have anything to do with ring widths?

  7. Steve McIntyre
    Posted Aug 29, 2005 at 6:54 AM | Permalink

    The theory (and it is not my theory) is that increased CO2 levels increase water use efficiency by reducing stomatal conductance. This is particularly significant for bristlecones because of their high altitude (thus lower CO2 concentrations) and arid climate (bristlecone growth is limited by soil moisture later in the season: Fritts, 1969]. Increased water use efficiency results in more biomass per unit of water availability. More biomass leads to greater ring widths.

    I haven’t ever seen density reported for bristlecones: it’s curious when you think about it. Note that density across the majority of conifer sites has been decreasing since ~1960. This decline has been attributed by Briffa to an “unknown” anthropogenic cause for about 10 years without resolving it. So they delete the adverse portion of the record from the spaghetti graphs, as I pointed out before.

    The dC13 trend is stated by Tang et al. to be a non-climatic trend – it’s not a conclusion that I am presenting on my own.

    BTW,I am knowledgeable in the statistical methods used in dendrochronology. I don’t claim to be an “expert” in tree botany. However, when supposed experts do present studies relying on tree botany, I think that they have an obligation to consider advrse salient material.

  8. Knut Knutsen
    Posted Aug 29, 2005 at 8:12 AM | Permalink

    Re. #7. So you are confident that it isn´t possible to have d13C trends that are unrelated to ring width, hence ring widths can do their thing and respond primarily to T while carbon isotopes reflect entirely different processes, or perhaps both processes happening at the same time? I cannot undestand why C-isotope variations disproves that ring width changes can be caused by T. And even if there is a CO2-fertilization effect, are you sure that all reported growth increases are attributed to this? Has it been scaled vs CO2 changes? If the ring width changes are caused by CO2-fertilization, then how do you explain changes occurring prior to the start of atmospheric CO2 increase. They must have been caused by something else, perhaps T?

  9. beng
    Posted Aug 29, 2005 at 8:56 AM | Permalink

    I wonder if there are other hockeystick series like the US Sheep Mnt series? If the CO2/water-efficiency effect is real & most evident at the driest sites, “hockeystickness” may show up in similar arid sites & other tree species where water is by far the limiting growth factor.

    The issue of growth enhancement in the Sheep Mnt series from mineral-dust blown from the dry Owens lakebed is another complicating issue.

  10. Steve McIntyre
    Posted Aug 29, 2005 at 9:24 AM | Permalink

    RE #8: Knut, ring width changes are caused by all sorts of things: precipitation, temperature, competition, cloudiness, insects, etc.

    If our view of climate history is to depend on bristlecones, the obligation on the original authors is to ensure that their data is free of non-climatic fertilization: potential CO2 fertilization was already an issue on the table for bristlecones and increased water use efficiency was said to be a probable mechanism which made sense in botanical terms.

    dC13 isotope variation was proposed by Feng and others as a method of testing for changes in water use efficiency. I am obviously not qualified to provide a warranty on whether it does or doesn’t. Hoever, this was exactly on point to the issue of CO2 fertilization – an issue already recognized by IPCC.

    Proper due diligence by the original authors and/or IPCC would have discussed the issue: it is their responsibility to show that bristlecone ring widths are a vlaid temperature proxy. There was lots of reason to be suspicious of these series as temperature proxies in 1998 and even more so now. In particular, it is unacceptable that the dependence of MBH98 results on bristlecones (vide the CENSORED directory) should not have been disclosed) and that they made claims that the MBH98 reconstruction was “robust” to the presence/absence of dendroclimatic indicators, when they knew that they were not even robust to the presence/absence of bristlecones.

    Even though their salvage defence depends entirely on back-door inclusion of bristlecones, they have not withdrawn their claims of robustness to all dendroclimatic indicators. It is bizarre beyond belief.

    Re #9: anomalous 20th century growth is especially characteristic of high-altitude bristlecone (and cousin foxtail)pines, at other sites besides Sheep Mountain.

  11. ET SidViscous
    Posted Aug 29, 2005 at 12:34 PM | Permalink

    I’ve been waiting for a post like this to roll out something I’ve been bouncing in my cavernous (empty) skull. Unfortunately I’m a bit busy on the road right now. Maybe this evening.

    But there has always been something about the bristlecone proxy data, tree rings and all that that has always bothered me. All of this data supposedly “proving” global warming, which they then infer is bad for everyone seems to overlook one basic fact.

    They say thicker rings, warmer temperatures. What it also means is more tree growth, which is a GOOD thing. Regardless of it’s cause (Aeolian fertilization, longer growing season, whatever). More tree growth is a good thing. What Mann et all have done is shown that more warming and more CO2 is great for tree growth.

    How they correlate that into a sky is falling theory I have no idea.

  12. John G. Bell
    Posted Aug 29, 2005 at 2:36 PM | Permalink

    I enclose part of an informal email exchange I had with Harrold Frits back in May. Knut you might email Tang and get a gracious reply like I did from Fritts if Steve didn’t answer you to your satisfaction.

    Re: Process Model of Cambial Activity
    From: hfritts@—————
    To: john.bell@————–

    Dear John,
    I am inclined to agree with you. LaMarche and I found at least
    increased ring width in the Pinus longeva forests at high elevation, so
    we proposed it reflected the CO2 increase.
    LaMarche, V. C., Jr., D. A. Graybill, H. C. Fritts and M. R. Rose.
    1984. Increasing atmospheric carbon dioxide: tree ring evidence for
    growth enhancement in natural vegetation. Science 225:1019-21

    A lot of people disagreed. Graumlich could not find it in the Foxtail
    Pine which was in the Sierra Mountains upwind of the BCP in the white
    mountains. We pointed out that our trees were strip-barked trees while
    the Foxtail Pines were whole bark. I don’t recall the exact reasoning,
    but it was in some of the rebuttals in later issues of Science. BCP has
    little latewood and I don’t believe LaMarche considered a change in the


    John G. Bell wrote:
    >I had wondered on a blog if CO2 fertilization could cause an increase in
    >ring widths relative to their maximum densities. This is based on a look
    >at Briffa’s Polar Urals data. Perhaps it was Tornetrask? Anyway what I
    >noticed was that up until the mid 1800s changes in RWs and MXDs agreed
    >in magnitude and direction very well. Note for all the data up to the mid
    >1800s, more than a thousand years! After that point RWs increased relative
    >to MXDs and this change tracked the known change in CO2 in the atmosphere.
    >Could CO2 fertilization change the RW / MXD ratio in a tree limited in
    >carbohydrates? Might a change in this ratio tell us something about the
    >amount of CO2 a tree was exposed. The real world doesn’t control all
    >the variables so perhaps this is at hint level.
    >What was fun about this was a few decade period back in the 1400s? that
    >would pan out to low CO2 levels. Quite unique. I can’t find my original
    >comment so I apologize for not being more exact.
    >John G. Bell

    I don’t know what Fritts thinks of Tangs mechanism, but it looks like Tang has put the earlier work of Fritts and others on even more solid ground.

  13. Pat Boyle
    Posted Aug 29, 2005 at 5:58 PM | Permalink

    Is there any hard science on any of this? For medical questions we often begin with epidemiological studies but there are so many confounding variables that understanding is difficult. In biology there are field observations by ethologists but again the complexity of the environment limits understanding. It sounds as if the state of the art currently in dentrochronology is at the ethological or epidemiological level of understanding and dispute. At this level you can’t form firm conclusions. You only get hints and suggestions. We have had, I suspect, more than enough natural observation. Now we need some reductionism. Now we need a laboratory experiment where we can control and manipulate the environment.

    Surely we needn’t speculate about the effect of CO2, temperature, and Owens Valley dust on tree ring formation. Hasn’t anyone done the obvious experiments? We can measure these effects – we don’t have to argue.

    This dispute can be put to bed by setting up a simple series of sealed tree growth chambers where light, air, water, and soil are controlled and manipulated. This is such an obvious experiment it must have been done. What were the results?

    If this exeriment hasn’t been done – for shame.

  14. Steve McIntyre
    Posted Aug 29, 2005 at 7:55 PM | Permalink

    There’s been lots of work on factors affecting tree growth, but there are still controversies about how ring widths in natural forests result. Ross and I spent time with specialists at the University of Guelph last year and their opinion was that many of the most prominent dendrochronologists ignore botanical factors. Some of the issues seem so banal that it defies belief, but I’ll present one important issue. There’s lots of botanical evidence that conifers grow best at an optimum temperature – too hot is bad for growth, too cold is bad for growth. Cedars, for example, prefer cool moist. However, the proxy theory is premised on a linear relationship between temperature and ring widths. (If you have an upside-down U relationship, then proxy reconstructinos are unable to distinguish between hot and cold years on ring width alone.) There have been a number of experiments showing enhanced growth with elevated CO2 – look at references in Graybill and Idso 1993 and Tang et al 1999, both of which I’ve posted up here. Opponents of a general CO2 fertilization hypothesis have conceded its existence for strip-bark bristlecones of Graaybill and Idso 1993, which makes their appearance in the alter ego of the MBH98 PC1 (driving the MBH reconstruction) particularly ironic. This data is known to be problematic: the problem for MBH is that they are dependent on proxies that they know to be bad.

  15. John G. Bell
    Posted Aug 29, 2005 at 9:15 PM | Permalink

    OT – Steve, John Ioannidis has an essay “Why Most Published Research Findings Are False” in PLOS Medicine. You might find it interesting.

  16. Paul Gosling
    Posted Aug 30, 2005 at 3:42 AM | Permalink

    Two points

    1) CO2 fertilisation will not be a factor prior to 1850? After that we have reasonable instrument records. So why is everyone so interested in the proxies after 1850? Is anyone at all saying they are a better record of temperature than the instrument record? If not, then ignore them because of all the complicating factors and concentrate on the instrument record.

    2) There have been lots of experiments done on the fertilisation effect of CO2. Sadly the results are not clear. I think, though I am not 100% sure, the latest FACE experiments are indicating that doubling CO2 (overnight) results in a large but short lived growth increase. We can only speculate what the effect of doubling CO2 over say 100 years will do.

  17. Ed Snack
    Posted Aug 30, 2005 at 4:12 AM | Permalink

    Paul, I think you miss the point slightly. MBH uses proxies to compare todays temperature to historical temperatures. The intention being to show that modern temperatures are “unprecendented”. What the data appears to show is that no such comparison can safely be made, the proxies are too uncertain, in fact I suggest that the preponderance of the evidence suggests that at times within the last 2,000 years temperatures have been at least as high or higher. The reason to have modern proxies is to attempt to calibrate the results from before 1850 by comparing the proxies with the local instrumental record.

    Thus the AGW viewpoint that modern temperatures are solely as a result of human actions is not supportable on the evidence provided to date. If temperatures over the past 2-4,000 years say have been higher than current temperatures, on what basis can we assert that only by invoking anthropogenic CO2 can we explain the instrumental record ?

    There is also the point to check the modern instrumental record, although it would seem that for tree rings at least, this is fraught with significant difficulties.

  18. Bruce Alan
    Posted Aug 30, 2005 at 4:46 AM | Permalink

    Re #15. Wow! I heartily agree with the gist of “Why Most Published Research Findings Are False”. Although ego, greed and politics are eternal, I think Enronitis stalks Big Science. I applaud the efforts of M&M to restore scientific integrity and honest criticism in climate research.

  19. Paul Gosling
    Posted Aug 30, 2005 at 5:24 AM | Permalink

    Ed RE 17

    On what basis do you say that temperatures have been higher than at present in the last 2000 years? Where is your fully audited data to show this? Statements such as, the Romans grew grapes in northern England, are not really good enough.

  20. Michael Jankowski
    Posted Aug 30, 2005 at 5:48 AM | Permalink

    After that we have reasonable instrument records. So why is everyone so interested in the proxies after 1850?

    (1) If you’re going to compare the instrumental records to pre-1850 proxies, you must be able to calibrate the pre-1850 proxies using the instrumental record. Therefore, you need a sufficient calibration period post-1850 where the instrumental record and proxies overlap. And for many/most locations, you must go much closer to the present than 1850 to find suitable calibration data. To some extent this is done in MBH and other reconstructions.
    (2) As others have noted, the correlation between proxies and temperature isn’t cut-and-dry. There are other factors that can affect things like ring width. In order for the pre-1850 proxy data to provide an accurate representation of temperature pre-1850, it’s important to study the validity of the assumed relationship of tree ring width vs temperature used with pre-1850 proxies. We have plenty of data in recent decades concerning rainfall, CO2 concentrations, and temperatures in the area of proxies, so updated proxy analysis could confirm, call into question, or completely throw-out the window the assumptions used in reconstructions about the relationship of ring width vs temperature.

  21. Ed Snack
    Posted Aug 30, 2005 at 1:56 PM | Permalink

    Sigh, Paul, stop distorting what I claimed. I said, “I suggest that the preponderance of the evidence suggests that at times within the last 2,000 years temperatures have been at least as high or higher”. The evidence is not all that solid, and I made no claim that it was. However evidence from a range of sources does suggest that the MWP had temperatures comparable to today. If you want to overturn that view, find some evidence. The IPCC has tried very hard, but so far has no evidence that stands up to scrutiny. Do you dispute that MBH98 & 99 are now totally discredited ?

  22. ET SidViscous
    Posted Aug 30, 2005 at 2:13 PM | Permalink


    On what basis do you say that temperatures have always been lower than at present in the last 2000 years? Where is your fully audited data to show this? Empirical statements are not good enough.

  23. Steve McIntyre
    Posted Aug 30, 2005 at 2:20 PM | Permalink

    Paul, there’s a differnce between saying that the MBH and similar proxy data do not support the conclusion that the 20th century is warmer than (say) the MWP and thus we don’t know; and between claiming that say the 12th century was warmer. I think that there is considerable evidence of a warmer MWP, especially higher treelines in many different parts of the world, but have not gone so far as to draw conclusions from that.

    I am convinced that the Pliocene and many ealier periods were much warmer than at present, so to say that the temperatures in the 1990s were “unprecedented” [Nature] makes sense only if you limit time-scales to those proposed by Creationists.

  24. Dave Dardinger
    Posted Aug 30, 2005 at 3:14 PM | Permalink

    Re: #23

    Foul, Steve! Using the C word in science is like using the N word in politics or the other N word in race relations.

  25. Mats Holmstrom
    Posted Aug 30, 2005 at 9:35 PM | Permalink

    Re: #13, #14. There is a paper on a controlled CO2 fertilization experiment in the latest issue of Science:
    Carbon Flux and Growth in Mature Deciduous Forest Trees Exposed to Elevated CO2, Kàƒ⵲ner et al., Science, Vol 309, Issue 5739, 1360-1362, 26 August 2005.

    Apparently this is the first study for large (30+ m) trees. The conclusion relevant to tree rings is that the growth seems to be highly specie dependent and they conclude that

    …one species showing a transitory growth stimulation and others not.

    It is a four year study where they release C02 in the canopy of the trees by tubes.

  26. Ed Snack
    Posted Aug 30, 2005 at 11:50 PM | Permalink

    I understand that the Idso’s have been conducting long term experiments with CO2 fertilisation using large covered enclosures, but since their site has become a pay site I haven’t been back to check. It may be worth doing so as they are one of the acknowledged experts in this area.

  27. Peter Hearnden
    Posted Aug 31, 2005 at 12:28 AM | Permalink

    Re #21,

    Ed, you’re trying to have it both ways. Either you think the proxies aren’t up to it or you don’t. You can’t dismiss some of them and then, in effect, accept the one you like (the warm ones, well well…) and say ‘oh, but there is good evidence it was warmer’. Why? becuase you look what you are, baised against a cold last 1000 years (becuase you’re biased aginst the possibilty we’re changing the climate seriously). That’s bad science. Open you’re mind!

    Re #23, Steve, I’m inclinded (though not ‘convinced’) to agree with you. In what context was ‘unprecidented’ used?

  28. Ed Snack
    Posted Aug 31, 2005 at 3:31 AM | Permalink

    Peter, here you go again. I’m not trying to make any particular point, I simply observe that “on the preponderance of the evidence” which is (as an explanation) the weaker form of “proof” used in the civil courts as opposed to the “beyond reasonable doubt” of the criminal courts (purely as an illustration, no specific legal reference intended) that there is such evidence. I don’t think the evidence we have to date can be used to “prove” higher or lower temperatures with any degree of accuracy. That is in contradistinction to MBH and the AGW proponenets who are, it seems, still arguing that they have shown that temperatures have always (over the last 1-2,000 years) been colder.

    Now Peter, care to stop the petty picking and contemplate the evidence ? I’d say it is now “beyond reasonable doubt” that MBH’s results are unsound and that it should be gracefully withdrawn. I’ve put that to another AGW fan in another thread, and they won’t answer the questions on the data either. Come on, it’s quite simple:

    1. The Bristlecones are not a temperature proxy
    2. The Gaspe Cedar records cannot be relied upon as the original records and location are now lost
    And without the BCPs and Gaspe, no “hockeystick” shape emerges
    3. The 15th Century step lacks statistical significance
    4. With respect to 3. MBH almost certainly know that, and they witheld that data on publication

    All in all, genuine junk. Now, try addressing the substance.

  29. Paul Gosling
    Posted Aug 31, 2005 at 5:34 AM | Permalink

    I agree MBH was rubbish. I have said so on several threads here. That does not mean there was a MWP any more than proving that there was no big bang proves there is a god. So we have bits and pieces of evidence that there was a MWP, and the Romans in northern England grew grapes and didn’t wear any socks, and we know that 18,000 bp there were huge ice sheets over the NH. Clearly this is evidence that there is no such thing as AGW and doubling atmospheric CO2 will have no effect on climate …NO. All we can say is that we don’t know to any degree of accuracy global temperatures in the way that MBH and others have claimed.

    Steve re 23. Did MBH claim unprecidented temperatures on a geological timescale or just in the last 1000 years. Its rubbish, but there is no need to distort what it said.

  30. Steve McIntyre
    Posted Aug 31, 2005 at 6:39 AM | Permalink

    Paul, I’m pretty careful not to distort what people say. I attributed the term “unprecedented” to Nature, not to MBH. It occurred in their editorial about the Barton letters as follows: “Subsequent studies have supported the observation that unprecedented warming occurred in the 20th century, while multiple lines of evidence support the notion that anthropogenic influences are contributing to it.”

    Click to access natureedit.pdf

    As I observed in my post, the warming is “unprecedented” only in a Creationist time scale. The temperatures are far from “unprecedented” in a geological time scale. In my opinion, this is a very non-trivial observation. Indeed, this is the context of my original interest: a geologist pointing out to me the discrepancy between IPCC theories and the geological record. Indeed, he used pretty similar terminology.

    Changes on a geological time scale do not appear well understood to me. Or at least, it seems to me that there is sufficient disagreement among authorities that a civilian cannot possibly conclude which explanation is right. If anyone wishes to provide references otherwise, I’d be interested.

    The following is only thinking out loud. For example, episodes like the Eemian or the Holocene are unusual within the Pleistocene period, but within a longer geological context, it looks to me like the Pleistocene itself is unusual (for its coldness). You can think of a lot of puzzles once you start thinking this way, especially if you think about it statistically.

    So back to the Nature editorial: maybe you can argue that the context implied that they were only talking about the millennium. But that’s not what they said. The editorial was written for a non-specialist audience and many readers might reasonably conclude that it was the view of Nature that the warming was actually “unprecedented” in Earth terms.

  31. JerryB
    Posted Aug 31, 2005 at 7:33 AM | Permalink


    The Idsos revised their site pay policy: parts are free, other parts require subscription:

    A link to free parts is:

  32. Paul Gosling
    Posted Aug 31, 2005 at 7:42 AM | Permalink


    I should read what you write more carefully!

    However, though you are correct in asserting that the changes in climate which have occurred over the last century, or even those predicted for this century, are small and even if the IPCC predictions are correct, the climate will still be well within the range in the geological record. That does not mean that the effects will be trivial. A return to full ice age conditions would also not be unprecedented but would be catastrophic in terms of the human and natural environment.

  33. Steve McIntyre
    Posted Aug 31, 2005 at 9:40 AM | Permalink

    Paul, my original issue was simply how one could say with any confidence that 1998 was the warmest year of the millennium, and, past that, how one calculated the response of climate to increased CO2 – why do people think that the response will be 2.5 deg C for 2xCO2 rather than 0.5 deg C? I realize that there are inter-related lines of argument for 2.5 deg C, but at least some of them seem to be tuned to proxy arguments that I regard as suspicious. I can’t investigate everything; I’m already looking at too many things and not finishing stuff. If I had a big policy responsibility, I would try to ensure that the various lines of argument were disentangled – so that I knew what was tuned to proxy arguments and what wasn’t. My guess is that much of the policy concern arises out of feedback arguments, which I regard as being far from proven in this sense: I would have thought that with the postulated positive feedbacks, the earth would have run out of control a long time ago and would either be an icehouse or a greenhouse already. It’s had an odd combination of what seem to be large climate fluctuations in geological time, but a curious boundedness to them.

    As to Ice Age return scenarios, if you were worried about this (and based on the Pleistocene, it’s far from an unrealistic concern) wouldn’t you want to do something to try to avoid an Ice Age return? I think that someone has recently argued (Ruddiman?) that human activities since the stone age have already had enough impact to stop this.

  34. Greg F
    Posted Aug 31, 2005 at 10:22 AM | Permalink

    …why do people think that the response will be 2.5 deg C for 2xCO2 rather than 0.5 deg C?

    An interesting blog of climate modeler James Annan may offer some insight.

    From “An overview of probabilistic climate prediction
    One crucial take-home message is that there is no way to generate a truly objective estimate of climate sensitivity, and never will be.

  35. Dave Dardinger
    Posted Aug 31, 2005 at 12:44 PM | Permalink

    Re: #34

    On problem, it seems to me, is that the feedbacks operate on different scales. This is because there are different sorts of mechanism, many non-linear. Warmers will readily accept some feedbacks if they have the potential to scare the general public into accepting their policy desires, but will ignore or downplay other such feedbacks, even though the obvious fact is that the Earth has never suffered a irreversible climate excursion in 5 billion years (though some have been pretty catastrophic to be sure) and that the reason must be the existence of negative feedbacks.

  36. Michael Mayson
    Posted Aug 31, 2005 at 1:16 PM | Permalink

    Re 34: Annan also says (my bold):

    Disclaimer: my recent (as yet unpublished) work points towards an upper limit of about 6C (and a most likely value rather lower than this). But I honestly did not go looking for such a result, and much higher values would certainly have been more readily publishable 🙂

  37. John G. Bell
    Posted Aug 31, 2005 at 2:11 PM | Permalink

    Re #34

    You might note that the emphasis was in the original and the only emboldened string in a long article. I am having a tough time reconciling it with this earlier quote from the same article.

    “The IPCC says that climate sensitivity is likely to be in the range of 1.5-4.5C, an estimate which has not changed for many years. Originally this was based on very limited evidence, but subsequent research appears to confirm that this early estimate was a pretty good (if lucky) one.”

    I think he is inconconsistent. Perhaps he trying to clarify his thinking and is sort of talking out loud in order to figure things out. A very useful thing to do.

  38. Peter Hearnden
    Posted Sep 1, 2005 at 12:48 AM | Permalink

    Peter, here you go again. I’m not trying to make any particular point, I simply observe that “on the preponderance of the evidence” which is (as an explanation) the weaker form of “proof” used in the civil courts as opposed to the “beyond reasonable doubt” of the criminal courts (purely as an illustration, no specific legal reference intended) that there is such evidence.

    You mean there is evidence there is evidence? OK…

    I don’t think the evidence we have to date can be used to “prove” higher or lower temperatures with any degree of accuracy. That is in contradistinction to MBH and the AGW proponenets who are, it seems, still arguing that they have shown that temperatures have always (over the last 1-2,000 years) been colder.

    But you know, don’t you, that temperatures were higher. Come on come clean, that’s what YOU know. You’ve convinced yourself. Me? I accept this as what we know – . It’s not the end of the story, but, since people like Steve, you and the rest of the amatuer climatologists/anti warmers wont give us an alternative view of what temperatures have done, it’s the best we have.

    Now Peter, care to stop the petty picking and contemplate the evidence ?

    ME, pickey, LOL…

    I’d say it is now “beyond reasonable doubt” that MBH’s results are unsound and that it should be gracefully withdrawn. I’ve put that to another AGW fan in another thread, and they won’t answer the questions on the data either. Come on, it’s quite simple:

    1. The Bristlecones are not a temperature proxy
    2. The Gaspe Cedar records cannot be relied upon as the original records and location are now lost
    And without the BCPs and Gaspe, no “hockeystick” shape emerges
    3. The 15th Century step lacks statistical significance
    4. With respect to 3. MBH almost certainly know that, and they witheld that data on publication

    All in all, genuine junk. Now, try addressing the substance.

    Calling something junk, especially from a POV that demands it be junk, isn’t going to convince anyone whose’s open minded. I don’t accept your points one and two as ‘shown’. I doubt three is of much relevance (where is the jump in the graph I quoted?) and I think four is just another of a vast series of similar insults from people of you ilk.

  39. Ed Snack
    Posted Sep 1, 2005 at 1:33 AM | Permalink

    Ah Peter, at last you decide to look at Data. Excellent, OK, the Bristlecone Pines, you think the Bristlecones are a temperature proxy ? Perhaps you should read the original paper by Idso & Graybill, available off this site, perhaps you should read the post heading this thread. Also, look at the increase in the ring widths, and then compare those to the instrumental record for the same period. The ring widths go up, but the temperature doesn’t. You can’t find anyone outside of Realclimate who is prepared to defend the use of the BCPs when they are so clearly reacting to something other than temperature. You see, if they were, you would expect the local temperature to have risen, but it hasn’t.

    Gaspe, OK, the problem there is a little different. A number of trees were sampled, and a subset that showed a clear increase in ring width were selected as proxies. Can you see the problem with this ? It means that the results are cherry picked to show the wanted result. This would be recoverable by going back to the original data to check the entire sample, and if necessary by going back to the site to check on locations and other factors. However the original data is not available, and the site location has been lost. So the data just might be good, but we can’t tell, the only reasonable way to deal with this is to exclude it. This should not be a problem as Michael Mann has claimed that his results are “robust” to the presence or absence of any particular proxies.

    And there have been emulations attempted with the BCP records, Mann did one and the results were (are ?) in the censored directory on the web site. Wahl and Amman did one also, and it is in their paper, not prominently displayed, but it is there, have a look. Interestingly, without the BCPs, the hockeystick shape tends to disappear.

    Point three is precisely relevant. You cannot support a statement that 20th century temperatures are higher than 15th century temperatures if your data does not show that. And Mann;s data does not support that statement. Steve has shown that by treating the MBH original data correctly the 15th century appears to have a higher temeprature than today, but that the result lacks significance. That is, we can’t claim that the 15th century is warmer than the 20th century even though the graph shows that, as the result is not statistically significant. We can only state that we cannot conclude anything about the relative temperatures using MBH data.

    As for the last being one of a series of insults, you continue to be as obtuse as ever. Let’s put it plainly in short statements:

    1. Mann has released code that he claims was used in the calculations associated with MBH98 (and presumably MBH99).
    2. That code calculates the r2 statistic for each step
    3. When using the most precise emulations of MBH98 (M&M and W&A), the r2 statistic for the 15th century step is effectively zero;
    4. That result was not released in MBH98 (or MBH99)
    5. So we have the almost certainty that r2 was calculated, that it was all but zero, and even you cannot dispute that it wasn’t released for the 15th century step. Now, just as a kicker, Mann apparently did release it for the 1820 step, where the statistic is OK.

    So what are you disputing ? Are you claiming that the code Mann freleased is the wrong code ? Are you disputing the evidence in that code that r2 is calculated (see Steve’s posts with the code) ? Are you disputing that the r2 value is very low ? Are you disputing that it was released ? Or, finally, are you accepting that one can calculate a result that undermines your conclusion, and still publish without revealing that failure ?

    If you are, as you claim to be, open minded, you will look at the data, and not simply believe what you are told to believe. Try it, look at the data and answer. Wiki just don’t cut it.

  40. Douglas Hoyt
    Posted Sep 1, 2005 at 4:43 AM | Permalink

    Another problem with the Mann and similiar approaches to climate reconstruction is the assumption of stationary or time invariant PCs. It is a ridiculous assumption and obviously not true in the real world. Even GCMs don’t produce stationary PCs. I suppose that this approach is adopted so they can make some sort of calculation, but the numbers have no validity. I am surprised this shortcoming is not commented on more often.

    They are aware of the problem since Jones, for example, has commented that the spatial pattern of warming for 1910-1940 is completely different from the spatial pattern of the most recent warming.

  41. Posted Sep 1, 2005 at 5:23 AM | Permalink

    Now here is an interesting reference to Bristle Cone Pines –

    E. Schulman, National Geographic Magazine, March 1958, pp. 356, 368

    paraphrasing “Bristlecone Pines during years of stress do not produce tree-rings to indicate that stresses occur”

    QF Ginenthal. P. 275, Extinction of Mammoths.

    Therefore Bristle Cone Pines are climatically problematical.

  42. Steve McIntyre
    Posted Sep 1, 2005 at 5:27 AM | Permalink

    Louis, in fairness to the bristlecone pine people, because of the great bristlecone longevity, I think that their crossdating is pretty good and that they can picku up missing rings – you don’t get missing rings in ALL trees and other trees will have narrow rings. I think that thre are crossdating problems in the MWP portion of Briffa’s important Polar Urals dataset, which does not have a big impact on MBH, but does have an important impact on Jones et al 1998 and Crowley and Lowery.

  43. Posted Sep 1, 2005 at 6:00 AM | Permalink


    Ok, issue accepted,

    Now what is their chronology based on ?

  44. àƒ'?anàƒËœ
    Posted Sep 2, 2005 at 3:03 PM | Permalink

    this isn’t quite right in preview, so I may have to repost

    Steve, I just ran across this post. Dave over on another post is hitting me up on this issue and I should post here, too.

    1. the correct ref is Global Change Biol 5:1.

    2. I hadn’t read this paper, but a number of folk here sporadically say the proxies should be updated and measured to date. This is another paper that indicates this idea isn’t a good one unless you are real careful, as the recent record is polluted with CO2 (I presume you are saying this as well).

    3. I don’thave the time to track it down, but a couple of the refs in the first paper I excerpted here discuss the long-term variability at these sites, which the authors hint at in the beginning of the Discussion. This speaks to the WUE portion of this paper.

    4. This paper is about the above- and below-ground biomass responses to elevated CO2 [We report the results of a study that intended to test, using carbon isotopes of tree rings, (i) whether the water-use efficiency of bristlecone pine trees at this site has increased, and (ii) whether the magnitude of any increase in W for these two tree forms corresponds to the rate of stem growth (and possibly rate of increase of stored biomass).]; the theory that plants (presumably all plants) respond to elevated CO2 in a particular way (specifically, by sequestering carbon) is a big deal and hence lots of studies are happening. Side studies look at WUE, stomatal conductance, root responses, etc.

    5. The purpose of the paper is to determine response to elevated CO2, not to validate a negative hypothesis about response to temp, so caution must be used in using this paper for another purpose.

    6. I see as I read the comments above (your 10 and the 12, plus your 14) there are some refs I used in the other one here that are germane to the point. The comment 12 is good and illustrative.

    Now, I’m a plant guy. It is very hard to get enough information to speak definitively about plants. We are constantly talking about needing more information. So,

    6a. What must be considered here is the Foxtail pine is a different spp and in a different biome, and in the #12 it is very important to have the context, and this guy should be engaged in a long discussion before using this to say he is backing the fact that…whatever.

    So, what does all this mean?

    The paper used for this post has found, along with other authors, that there is evidence for CO2 fertilization. We don’t know how it affects most spp. There are a few studies, but when the authors mention C3 and C4 plants, they don’t mention N apportionment differences (you have to be an ag guy and dredge that up) and other varying responses. The paper also mentions the below-ground biomass response [Although the results are controversial, many studies suggested that, at an elevated atmospheric CO2 concentration, an increase in biomass allocation to roots is generally expected ] without discussing many authors have found that water and N are limiting factors. Plus, the authors did not measure change over time, and imply limitations to their paper in the discussion immediately following what I just excerpted.


    What the authors conclude [The set of experiments reported in this paper indicates that neither water-use efficiency nor cambial growth rate can be a sufficient indicator for changes in the biomass of natural forests. However, studies of this sort may provide some information for long-term physiological responds of trees in natural forests to increasing atmospheric CO2 concentration.] is that they, too, find mixed results when considering the question of response of Pinus l. to increasing CO2, just like others do, and come up with a hypothesis to explain another reason forn the increase [Another possibility is that the relative increase in the photosynthetic rate of strip-bark trees is greater than that of full-bark trees, but such a difference is not related to the adjustment of water-use efficiency. There are no data to evaluate this hypothesis. ].

    The paper uses C isotopes to look at the tree-rings to test WUE. They don’t mention temperature. That is your assertion, Steve, and I don’t see in the literature that your assertion that Pinus l. or a. aren’t good temp. indicators.

    Perhaps I haven’t found that statement yet in the literature and you’ll correct me. But the dendro folk aren’t saying that and this paper does not say that. It is a leap, and you can make it if you wish.

    If you write the dendro guys about your hypothesis and you get some funding to core some trees, please let me know. I’ve done some site work in the Whites, but just to calculate tree distribution by aspect up there. I’d like to climb that road on my bike when we aren’t coring trees, if you’ll have me along to collect data to test your hypothesis.



  45. Steve McIntyre
    Posted Sep 2, 2005 at 3:16 PM | Permalink

    Dano, this is interesting stuff. It will take me some time to look at the refs, but I will do so.

    Some of these bristlecone sites look like terrific places to visit. Earlier this year, I posted up some geographical and quasi-tourist info on a couple of sites. I meant to do some more.

    I tried to get the location of the Gaspe cedar site from Jacoby to commission re-sampling there. He refused to provide it. I doubt that I’d get a whole lot of cooperation from the bristlecone guys, but you never know until you ask. What sort of budget do you think would be required?

    I’m planning to visit my sister in Colorado Springs in November and there are some bristlecone sites near there – just past Cripple Creek as in the song. They might be easier to re-sample than White Mountains.

  46. àƒ'?anàƒËœ
    Posted Sep 2, 2005 at 4:56 PM | Permalink

    1. Pinus a. is slightly different than Pinus l. so that might be something, and you have a place to stay and that is something too. The trees aren’t as old, though.

    2. The Whites have limited access to old trees – I was on the lower elev. aspects that had numerous young trees to sample soil (they wouldn’t let us sample the oldest tree sites or even tell us where exactly they were).

    Budget depends upon how well you sample and how many times you think you need to go back down and resample what you missed, or whether someone still has cores and you can go there instead to make up for what you missed (or add to it, depending) – travel expenses, IOW and not much else, but you never know what you get with cores and you always seem to want to go back if you aren’t sure (or get a result like this paper dC13 here b/c you don’t have time and money to do more (and who does?). I never know what the h*ll I’m doing with cores and end up giving them to someone else to look at anyway.

    3. You might be better off going north and using already sampled black spruce in sheltered, non-ice blasted locations; black spruce has some characteristics that may extend this investigation, and there certainly should be plenty of cores around.



  47. TCO
    Posted Sep 3, 2005 at 3:58 PM | Permalink

    1. I am all in favor of you doing (or at least commissioning) some new experimental work, Steve. At some point, there are diminishing returns to sticking to Mann like a bulldog (especially given no new data released by him). Also, it would have synergy with your sunk costs (you’ve spent the time learning the field) as well as possibly instructing your critiques of other meta-analyses. And it would make you a bit less of an enemy foreigner in the camp.

    2. I don’t think that finding a new discussion of confounding factors on the bristlecones is a “nail in the coffin”. (overstated title…you hanging around Lambert too much?) We know that there are all kinds of confounding factors. I’m wary of trusting a meta-analysis dependant on one study or set of closely linked studies given all the confounding factors. However, a real “nail in the coffin” would be definitive understanding of all the confounding factors, how they’ve played out in the species, and that temp is wrong. It seems obvious from this one study (and DanO’s comments hit same point) that you don’t have that level of understanding of the parameter space.

    3. I consider the response of the trees when local temp did not go up to be much more of a coffin nail.

  48. Steve McIntyre
    Posted Sep 3, 2005 at 4:10 PM | Permalink

    TCO – you’re right about the Nail in the Coffin in the title. It is too Lambertian. I’m deleting it.

  49. Steve McIntyre
    Posted Sep 3, 2005 at 4:15 PM | Permalink

    Dano, I talked to the Guelph cedar guys about going down and doing some bristlecone work. There is an amazing "forest" of cedars in southern Ontario along the Niagara escarpment, some of them are over 1000 years old. The "forest" is about 600 miles long and 100 years wide, along a long cliff. The cedars are strip bark, like bristlecones, slow-growing. Many interesting articles by Doug Larson and Pete Kelly.

  50. TCO
    Posted Sep 3, 2005 at 4:22 PM | Permalink

    No one has picked on me yet for using sunk costs and synergy in my posting language. Fine, I’m working in NPV somewhere…

  51. àƒ'?anàƒËœ
    Posted Sep 3, 2005 at 5:39 PM | Permalink


    Outstanding. Always better to get data closer to home, and you don’t have to cross our border. You should be able to see a long record of the AO in the record as well. The Pinus l./a. are marginally (if at all) affected by the AO, so you may have issues there making direct comparisons. In addition, you may want to consider these plants may experience some maritime influence, whereas the Pinus have solely a continental influence, further complicating matters. Lastly, I seem to recall an article about that area when I was studying adaptive management and it seemed quite interesting, so that’s reason enough to give it a go.

    IIRC, the strip-bark feature in slow-growing tree spp affects metabolism and apportionment, but shouldn’t be a reason for preferential selection when coring (e.g. it is something to account for, but won’t tell you something that a tree completely encircled by bark would).

    Anyway, it will be interesting to see whether these trees indicate, in this place, the MWP was warmer than present. And an acknowledgement of the utility (sure, TCO, in the economic sense) of TCO’s well-written comments WRT to points made at ends of paras as an important purpose.

    Gotta go,


  52. Steve McIntyre
    Posted Sep 3, 2005 at 7:31 PM | Permalink

    Larson and Kelly are really knowlegeable guys, who have had lots of interesting experience with cliff ecology. They think that nearly all dendrochronologists (especially Hughes) pay far too little attention to botanical and ecological factors. We spent a lot of time talking to them about bristlecones and cedars in 2004 while we were writing our 2005 articles.

    The cedars like cool moist climate and do not have a linear response to temperature. More like an upside down quadratic. I think that the upside down quadratic is abig problem for the tree ring jockeys. Briffa has complained like a stuck pig that tree rings haven’t been getting wider in the 2nd half of the 20th century. They chop off the graphs after 1960. Briffa blames an “unknown anthropogenic” factor and drives on. I think that it’s equally likely that there’s an upside down U relationship and the proxies are really in trouble in the 1990s – which is why we don’t hear about them.

    As I’ve said before, if the tree rings in the 1990s were off the charts, we’d be hearing about it from the tree ring jockeys. Their silence is deafening. That’s why they’re talking about glaciers now.

    The other trick that’s in the works is SO&P. They seem to have focused on tree rings after 1500 and will probably have hysterical articles timed for IPCC 4AR. But no ones’ arguing about the post-1500 proxies.

  53. TCO
    Posted Sep 24, 2005 at 2:05 PM | Permalink

    The one thing that I strongly urge for you Steve, if you fund/do/push new experimental work, is that you let the chips fall where they may regarding the results. Even if it hurts your “sell” recommendation. Even if after enough time goes by, your sell recc is proved right, you still should have allowed the initial study that went the other way, to be quickly/publicly disclosed. Remember the opposite behavior is what bothers you in Jacoby.

    Also (of course) the same issue applies to your auditing/examination of people’s papers. If you find errors the minimze(!) stickishness, you need to share those as well. I won’t go so far as to say you need to look for them with the same ardor. But definitely if you get wind of something or get a reasonable impression that such an error in the numbers exists, you ought to finish that little subanalysis and bring it forward. Forensic accounting finding errors either way….

  54. TCO
    Posted Apr 2, 2006 at 7:53 PM | Permalink

    Reading through this thing again…I have about 100 questions on this paper. Overall, I don’t see that it really says anything new. Sure, Mann and others have not included CO2 as an independant variable in their proxy training. Or come up with a botanical exlanation for CO2 or temp as the factor promoting growth. But this article really doesn’t do much new. This article doesn’t have temp in it. Doesn’t give a good explanation for why we should care about C13 versus RW itself. Doesn’t prove assertions of biomass in places other than the trunk. And only samples 4 trees.

    However, they did sample same trees as earlier study, which certainly should be the standard. I guess all these trees should be tagged really. Except the ones that are cut down for sectioning. Need to do that sometimes. I guess if someone wants to really understand these arguments about places where the biomass is added, we might need to cut whole trees down and pull out the roots and weigh it all.

  55. TCO
    Posted Sep 16, 2006 at 8:02 PM | Permalink

    Phil, this is the paper that I was thinking of in my head when I mentioned having read one Idso paper (which was ok, but not really controlling for confounding variables, so not the killer analysis.) Anyhoo, it is obviously not by the Idsos. I will read the one you referenced.

  56. bender
    Posted Nov 15, 2006 at 3:38 AM | Permalink

    Interesting that Dano in #44 is hitting all four cylinders growth = f(T,P,C,N) – but does not consider the model mis-specification error that I believe all dendroclimatologists make when they neglect the synergy from the 2,3, and 4-term interactions. Increased water use efficiency via reduced stomatal opening under elevated CO2 is, for example, part of the two-way P*C interaction.

    I’ve made this argument in other posts in previous weeks.

  57. bender
    Posted Nov 15, 2006 at 4:01 AM | Permalink

    Growth model mis-specification error discussed on two previous occasions:
    comment #33 in post Rob Wilson on bcps
    comment #7 in post New CPD paper on reconstruction

  58. Chris O'Neill
    Posted Nov 22, 2006 at 10:39 AM | Permalink

    In #16, Paul Gosling wrote

    “CO2 fertilisation will not be a factor prior to 1850? After that we have reasonable instrument records. So why is everyone so interested in the proxies after 1850?”

    To which Michael Jankowski in #20 wrote:

    “If you’re going to compare the instrumental records to pre-1850 proxies, you must be able to calibrate the pre-1850 proxies using the instrumental record. Therefore, you need a sufficient calibration period post-1850 where the instrumental record and proxies overlap.”

    So the issue is calibrating the proxies and MBH9x do it by calibrating during the 20th century when the problematic growth occurs. One way around this issue would be to make a reconstruction back to say, 1600, for which there are far more proxies than earlier and then use this reconstruction to calibrate the earlier proxies over say, 1600 to 1800, thus avoiding calibrating the proxies that are affected during the 20th century. This approach can easily be done using a REGEM reconstruction and illustrates an important benefit about REGEM that is not directly available in MBH9x, i.e. information from later-beginning proxies is used to calibrate earlier-beginning proxies. In MBH9x, the information from later-beginning proxies is completely ignored in calibrating earlier-beginning proxies.

    I don’t think they’ve applied the approach I described above in Rutherford et al 2005, (i.e. leave out the Bristlecone proxies after 1800), but the reconstruction before 1500 (for example) does depend on how the Bristlecone proxies correlate with all other proxies after 1500 and the instrument record, and not just on the instrument record in the 20th century.

One Trackback

  1. […] Water Use Efficiency (a topic previously discussed at CA last year in connection with bristlecones here) This increase in Water Use Efficiency meant, in her words, that trees were behaving as though it […]

%d bloggers like this: