Five Monsoon O18 Series

Jud Partin observed yesterday that a “fantastic new record” had been recently (early Nov 2008) published from Wanxiang, China. Zhang et al report that their new record is “broadly similar” to the reconstructions of Esper, Mann and Jones 2003 and Moberg as follows:

The Wanxiang record, with a d18O range of ~1.3 per mil (‰) (Fig. 1), exhibits a series of centennial to multicentennial fluctuations broadly similar to those documented in Northern Hemisphere (NH) temperature reconstructions, including the Current Warm Period (CWP), Little Ice Age (LIA), Medieval Warm Period (MWP), and Dark Age Cold Period (DACP) (5–8). [Esper, Mann and Jones; Moberg]

They illustrate this “broad similarity” with the following image:

[Dec 4 - the following paragraphs have been revised to incorporate comments from Jud Partin below].

The data for Wanxiang (33°19’N, 105°00’E, 1200 m) was promptly archived in the paper SI. Almost concurrent with publication, another somewhat nearby speleothem from Heshang (30°27’N, 110°25’E; 294 m) was archived at WDCP, previously published in Hu et al EPSL 2007. As Jud Partin had done in 2007 with his Borneo data, the speleothem data for all three caves has been made available with commendable promptness.

The authors of Zhang et al 2008 observe in their SI, a point that needs always to be kept in mind when placing interpretations on autocorrelated data.

In interpreting such records in terms of changing climate, we are pushing the limits of these archives in terms of signal to noise ratio.

I expressed frustration the other day with the handling of the monsoon affected proxies, but didn’t entirely explain the frustration. Jud Partin observed that the orientation for the Dongge O18 record was consistent with the Wanxiang cave record (more negative dO18 up). This observation appears to me to have the corollary (though this goes beyond Jud’s specific comments) that, even if Mann’s reasons for showing the Dongge proxy more negative up originated in through data mining, the actual orientation of the Dongge O18 proxy was not objectionable.

(In defence of my prior post, I didn’t actually take a position on how these proxies should be interpreted as I’m still finding my footing with these proxies. My point was the narrower one that the Mangini orientation was not “unique” and that Gavin Schmidt’s slagging of Loehle for use of the Mangini proxy therefore required a more substantive argument. It now appears that a number of Chinese speleothem proxies have the same orientation as Mangini used.)

The other issue that I had on my mind was the opposite orientation of Socotra speleothem O18 and Dasuopu ice core O18, both monsoon proxies as well and both oriented oppositely to the Chinese speleothems. Indeed, the Dasuopu O18 record (negative dO18 down) is by far the strongest contributor to the Thompson hockey stick.

For reference, I’ve plotted five O18 series below: Socotra (from Mann data), Wanxiang, Heshang HS4, Dongge D4 and Dasuopu. All of these series are O18 series and all are monsoon records. All are oriented with negative dO18 down (this is opposite to the orientation of the Zhang et al graphic). Jud Partin pointed out below that it is customary in paleoclimate literature that “warmer and/or wetter conditions” be plotted up. For now, given the different orientations that result from the interpretations of different authors, I want to show dO18 for all series in a consistent way (saving the warmer/wetter interpretation for a second step), since, for now, I’m interested in consistency between O18. If they were all shown with negative dO18 up in accordance with the interpretation of the Chinese speleothems, it wouldn’t accord with the usual orientation of Dasuopu where more negative dO18 is interpreted as colder rather than warmer/wetter.)

Based on a visual inspection of the 5 series, I find it hard to think up a reason why the Wanxiang, Heshang and Dongge records should be oriented with negative dO18 up, while the Socotra and Dasuopu records are oriented with negative dO18 down. (I note, in passing, that the Dasuopu record has a very odd appearance relative to the well-dated speleothems. The Dasuopu ice core is in a high-accumulation area; errors in dating would increase exponentially and I’m really wondering how certain the dates of the Dasuopu ice core are, but that’s a big topic.)

Perhaps Jud or someone else can explain why some monsoon O18 records should be oriented with negative dO18 up and some with negative dO18 down. For me, such an explanation needs more than saying – Dasuopu is an ice core and Socotra is in southwest Asia not southeast Asia. I know both those things, but don’t consider these particular points to be an “explanation”.

[Dec. 5: Jud has provided a citation for the Socotra record, which I will review. He is not familiar with the Dasuopu (ice core) record. This is the one that particularly interests me, since it is so widely used and an opposite behavior is attributed to it. It's collected under quite different circumstances: by raising the issue, I am not precluding a plausible reconciliation, merely observing that it seems like something that would be nice to see in the literature.]

References:
Hu, Chaoyong, Gideon M. Henderson, Junhua Huang, Shucheng Xie, Ying Sun, and Kathleen R. Johnson, 2008. Quantification of Holocene Asian monsoon rainfall from spatially separated cave records. Earth and Planetary Science Letters Vol. 266, No 3-4, pp. 221-232, February 2008

Zhang, Pingzhong, et al., 2008. A Test of Climate, Sun, and Culture Relationships from an 1810-Year Chinese Cave Record. Science Vol. 322, No 5903, pp. 940-942, November 7, 2008

Reblog this post [with Zemanta]

107 Comments

  1. jim edwards
    Posted Dec 4, 2008 at 12:11 PM | Permalink | Reply

    Crackpot theory #1

    Maybe Heshang and Dongge samples were actually stalactites, rather than stalagmites. The authors decided to flip the orientation of these samples out of respect to gravity, and to make the data feel more comfortable in the inverted position it’s been physically locked into for the past several centuries.

    • Craig Loehle
      Posted Dec 4, 2008 at 12:41 PM | Permalink | Reply

      Re: jim edwards (#1), Oh…a very good explanation…

    • Geoff Sherrington
      Posted Dec 7, 2008 at 7:39 AM | Permalink | Reply

      Re: jim edwards (#1),
      Crackpot theory number 2. Select a growth where a stagmite grows up to meet a stalactite coming down, to where they fuse. Then gravity cancels anti-gravity and there should be a wonderful, reproducible, paired data set.

  2. Timo Hämeranta
    Posted Dec 4, 2008 at 12:15 PM | Permalink | Reply

    Steve,

    “Hu et al EPSL 2007″ was published before print Oct 14, 2007, but the proper reference is

    Hu, Chaoyong, Gideon M. Henderson, Junhua Huang, Shucheng Xie, Ying Sun, and Kathleen R. Johnson, 2008. Quantification of Holocene Asian monsoon rainfall from spatially separated cave records. Earth and Planetary Science Letters Vol. 266, No 3-4, pp. 221-232, February 2008

    and

    “Zhang et al (early Nov 2008)” is

    Zhang, Pingzhong, et al., 2008. A Test of Climate, Sun, and Culture Relationships from an 1810-Year Chinese Cave Record. Science Vol. 322, No 5903, pp. 940-942, November 7, 2008

    I prefer proper references.

  3. Jean S
    Posted Dec 4, 2008 at 1:09 PM | Permalink | Reply

    I don’t fully accept that there are different scalings in the “broad similarity” figure, especially in the last two (Esper is in units of tree-index?)? Aren’t the last two both NH temperature anomalies in degrees celsius? How was the “pairing” done? Variance matching?

  4. Robinedwards
    Posted Dec 4, 2008 at 1:21 PM | Permalink | Reply

    As someone who knows nothing of the technology behind the use of of stalactite/stalagmite measurements as proxies for temperature (or more broadly /climate/) I cannot possibly comment on the rather fascinating inversions that seem to have been a part of the interpretations made by well-known names in the reconstruction industry.

    ***
    Steve wrote:- The other issue that I had on my mind was the opposite treatment of Socotra speleothem O18 and Dasuopu ice core O18, both monsoon proxies as well and both oriented oppositely. Indeed, the oppositely oriented Dasuopu O18 record is by far the strongest contributor to the Thompson hockey stick.

    **

    However, I again have to confess that I remain baffled by the apparent ability of a single record (or possibly a very few records) to dominate, or at least heavily influence, the graphical appearance of climate reconstructions. I have noted that this is a recurrent theme in CA in the discussions of assorted “hockey stick” attributes. Somehow plots of these data are apparently heavily weighted by a few crucial series relative to the overall collection and I can still not understand why this should be the case.

    Current assemblies from different – allegedly independent – sources (which is questioned in CA fairly regularly) of climate related information seem to run to several hundred individual time series from a variety of sciences. There is heavy emphasis on techniques for calibrating (or possibly re-scaling?) data from disparate sources so that some kind of acceptable (smooth?) transition from the relatively modern instrumental era back to the time when proxies are the only available source of pseudo-scientific numerics for climate can be made. There has been much discussion of the so-called “divergence problem” whose emergence seems to have generated plenty of weasel-words or other get-outs by climatologists. It is indeed a pity that completely acceptable explanations for the divergence have not yet emerged (to my knowledge). Craig Loehle’s papers showing that a simple linear mapping of dendrochronological data to temperatures over long time scales must not be assumed, and in fact might be very misleading.

    Such work makes me wonder why so much effort goes into the splicing of data from different provenances. The main objective is presumably to try to establish just where the “current warm period” lies with respect to other climate episodes. “Is it warmer now than in the MWP?” is a typical query to which a definitive answer would be most welcome.

    However, I don’t really regard this as being crucial. What is necessary is a very firm grip on what has been happening over the last hundred or so years, and more especially over the most recent times. Is there any reliable data set that covers the last 30 years, for example? Or, where can /unbiased/ candidate sets be found. The probable answer is that there are none, I fear.

    When I was investigating the data used by Mann et al in their first well-known reconstruction, the 1400 data set (kindly supplied to me by Steve McIntytre) I set out from the then reasonable-seeming premise that Mann et al used specialist knowledge to select data sets that were believed by them to be reasonably representative of Northern hemisphere climate from 1400 to about 1980. Whilst that pious hope turned out to be untenable, it was nevertheless possible to make an assessment of the data despite its vastly disparate scales and locations by quite simple means, working on the assumption that every series was worth equal weight in the overall assessment. Now this is known not to be exactly the case, but to distort the overall outcome by a noticeable amount individual series would have needed grossly enhanced weights.

    Is there not a case for bypassing all the amazing manipulations of Mann, which have been revealed by Steve’s phenomenal efforts to replicate Mann’s methods, and to revert to analyses that might be comprehended by “lay” people, amongst whom I include the media, politicians and any non-specialist scientists?

    I fear that in their current convoluted state the analyses are just going to be ignored by the sorts of people who might have a real influence on the general perception of what is happening or has happened to the Earth’s climate.

    Robin

  5. Jud Partin
    Posted Dec 4, 2008 at 1:36 PM | Permalink | Reply

    I think there may be a bit of confusion that needs to be cleared up.
    The Zhang et al Science 2008 article is from Wanxiang Cave (33°19’N, 105°00′).
    The Hu et al. EPSL 2008 is from Heshang Cave (30°27′N, 110°25′E).

    I think you may have confused the two records. They are from different caves. In fact, independent groups worked on these records. The Wenxiang cave record would make for a 5th monsoon series. (BTW, the data for this record is archived in the SOM – in case you want it)

    Second, in general (but not always), paleoclimate records are plotted with warmer and/or wetter conditions as “up” on the y-axis. I believe this convention comes from having modern (Holocene) conditions higher than colder conditions during the Last Glacial Maximum on the y-axis. Most paleo-papers you read plot their figures like this.

    Third, I said nothing about the Mann paper.

    Lastly, what I did say was to take a look at the Zhang et al. 2008 paper. Specifically Figure S4 in the SOM. They have modern calcite d18O and instrumental records of temperature and precipitation (as well as many other records plotted with their entire timeseries). The authors observe a positive correlation between d18O and temperature after the 1960’s. Someone was asking for a calibration study, so I posted this article.

    Steve: Jud, thanks for this. You’re right about my conflating the two series. I’ll edit this post accordingly. I understand the convention about warmer being “up”, but it seems like a useful precaution to plot the dO18 as they are measured before an orientation is interpreted.

  6. Steve McIntyre
    Posted Dec 4, 2008 at 2:08 PM | Permalink | Reply

    I’ve amended the post to incorporate the Wanxiang data from the SI. My apologies. Nice to have peer review from someone as able as Judd.

    Judd, I realize that your point was not specifically about the Mann paper, but, as someone interested in these series, I still don’t understand why the Chinese speleothems should have one orientation and the Socotra and Dasuopu series another.

    And reconciling Dasuopu to the speleothems goes far beyond the Mann paper as this is a very widely used proxy (in the Thompson hockey stick). I’m not saying that there isn’t an explanation – but, if there is, I don’t know what it is and I’d be interested in what you believe the explanation to be.

  7. Jud Partin
    Posted Dec 4, 2008 at 2:30 PM | Permalink | Reply

    No worries. I hope I’m representing the speleothem community in a good light.

    I have no clue about the Dasuopu ice core. I’m not that familiar with that record yet.

    As to Socotra, try this reference. I sent it to you a while back when you thought the record wasn’t published.

    Fleitmann, D., Burns, S. J., Mangini, A., Mudelsee, M., Kramers, J., Villa, I., Neff, U., Al-Subbary, A. A., Buettner, A., Hippler, D., and Matter, A., 2007. Holocene ITCZ and Indian monsoon dynamics recorded in stalagmites from Oman and Yemen (Socotra). Quaternary Science Reviews 26, 170-188.

    In it, the author does a really good job of examining a whole bunch of monsoon records. He finds that monsoon intensity and duration are both important. Just glancing at it, I saw the passage on p. 180.

    “In contrast to the gradual decrease in ISM
    precipitation recorded in stalagmite Q5 from Southern
    Oman, the D1 stalagmite d18O record from Socotra shows
    a long-term increase in inter-monsoon (spring and autumn)
    precipitation, as indicated by a long-term decrease in d18O
    since ~4.4 ka BP (Figs. 4 and 7). This anti-phase behavior
    between Southern Oman and Socotra seems to be apparent
    not only on millennial but also on multi-decadal timescales.”

    I suggest reading it. It may help you sort out the Socotra inversion you are mentioning.

    Also, double check your four/five monsoon plot. Did you want the y-axes the way you have them? Your text and plot don’t match up. Oh, and the Science plot has lighter d18O up, where you said it didn’t. And I said “warmer and/or wetter” goes up. Which set of words used depends on site location. …and it’s Hu et al 2008, not 2007.

    Ok, I really gotta get back to work.

    Jud …with one “d” ;)

  8. David L Hagen
    Posted Dec 4, 2008 at 2:39 PM | Permalink | Reply

    Perhaps a Mideastern speleothem might give you a refreshing break and perspective. Isn’t it equally part of the global temperature? (Not counting PDO, AMO, ENSO etc.)

    News: Climate history may explain empires’ fall
    Paper: Climate deterioration in the Eastern Mediterranean as revealed by ion microprobe analysis of a speleothem that grew from 2.2 to 0.9 ka in Soreq Cave, Israel

    Ian J. Orlanda, Miryam Bar-Matthewsb, Noriko T. Kitaa, Avner Ayalonb, Alan Matthewsc and John W. Valleya, Quaternary Research online 25 October 2008.
    Abstract
    “Analysis of oxygen isotope ratios (δ18O) by ion microprobe resolves a sub-annual climate record for the Eastern Mediterranean from a Soreq Cave stalagmite that grew between 2.2 and 0.9 ka. In contrast to conventional drill-sampling methods that yield a total variation of 1.0‰ in δ18Ocalcite values across our sample, the methods described here reveal up to 2.15‰ variation within single annual growth bands. Values of δ18O measured by ion microprobe vary in a regular saw-tooth pattern that correlates with annual, fluorescent growth banding where calcite grades from light to dark fluorescence. Modern records of precipitation and of cave dripwater indicate that variable δ18Ocalcite values record regular seasonal differences in δ18Orainfall modified by mixing in the vadose zone. Large differences in δ18O values measured across a single band (i.e., between the dark and light fluorescent calcite, or Δ18Odark-light) are interpreted to indicate wetter years, while smaller differences represent drier years. Oxygen isotopes record: 1) month-scale growth increments, 2) changes in Δ18Odark-light that represent seasonality, 3) a systematic, long-term decrease in maximum Δ18Odark-light values, and 4) an overall increase in average δ18Ocalcite values through time. These results suggest a drying of regional climate that coincides with the decline of the Roman and Byzantine Empires in the Levant region.”

  9. thefordprefect
    Posted Dec 4, 2008 at 3:03 PM | Permalink | Reply

    These may be of interest:
    http://www.pages.unibe.ch/products/newsletters/NL_2008-3lowres.pdf
    http://www.pages.unibe.ch/products/newsletters/2008-3/Special%20section/science%20hilight/Wang_2008-3(31-32).pdf

  10. thefordprefect
    Posted Dec 4, 2008 at 3:09 PM | Permalink | Reply

    Apologies again!
    The first links to:
    Advances in Speleothem Research Aug 2008
    The second links no where unless you copy and paste the link

    Millennial-scale climate variability recorded in Brazilian speleothems

  11. Posted Dec 4, 2008 at 5:22 PM | Permalink | Reply

    Just how strong is this statistical test called “broad simularity”? Does it involve squinting?

  12. Mark T
    Posted Dec 4, 2008 at 6:10 PM | Permalink | Reply

    Holding the pictures at arms length, I think.

    Mark

  13. jae
    Posted Dec 4, 2008 at 10:19 PM | Permalink | Reply

    Do I have it correct that this is even more convoluted than the tree-ring proxies? No more direct teleconections from Main to Spain. But here, we have even other steps? We go from stalagmite isotopes to rainfall amounts to wind direction to latitude to temperatures? I’m confused.

  14. tty
    Posted Dec 5, 2008 at 1:55 AM | Permalink | Reply

    There have been a fair amount of work done on Chinese speleothems in the past. The records have been used to study variations in intensity and extension of the monsoon, particularly between glacials and interglacials. This seems to be a “robust” way of using them. However, chinese geologists have up to now not tried to convert these d18O records to temperatures. I know because I have been looking for temperature estimates for previous interglacials from China. So this latest paper seems to be something new from a methodological point of view and should be used with caution.

  15. Paul Dennis
    Posted Dec 5, 2008 at 2:50 AM | Permalink | Reply

    I think a few words on speleothems as climate proxies, especially oxygen isotopes, might be in order. First a lot of attention was focussed on speleothems in the late 1960’s and early 1970’s with seminal work by Hendy on isotopic equilibrium in growing speleothems. The simple idea was that caves, especially in temperate climate zones tend to have constant average annual temperatures that is the same as the average annual air temperature above the cave. Thus by isotopically recording speleothem growth temperatures and dating the material using U-series, or by counting growth laminae for rapid growing, young specimens a climate record could be produced.

    When calcium carbonate (calcite or aragonite) precipitates from water there is a partitioning of the oxygen isotopes between the water and the carbonate that, under equilibrium conditions, is temperature dependent. This partitioning is known as the fractionation factor (alpha) and can be written as:

    alpha = R(calcite)/R(water)

    where R is the 18O to 16O ration in calcite, or water. Alpha varies by about 0.2 per mille (parts per thousand) for every degree change in precipitation temperature and has been subject to numerous theoretical (based on molecular vibrational energies of molecules), experimental and field calibrations. The measurement precision for 18O/16O ratios in both calcite and water is on the order of +/-0.05 per mille in many labs, possibly slightly better. Thus, roughly combining errors, we might expect to be able to determine precipitation temperatures to better than 1 degree C, and possibly better than 0.5 degrees C.

    There are two key problems: 1) To determine temperature we need to know both the isotope composition of the calcite and it’s parent fluid from which it grew. 2) We need to demonstrate that the speleothem grew in isotopic equilibrium.

    With respect to 1) the drip water composition will reflect the surface precipitation though not necessarily be the same. For example recharge may vary with season, stronger evapo-transpiration will reduce summer recharge and the groundwater might be biased towards a winter composition. Secondly precipitation isotope composition varies depending on a variety of factors: In temperate to high latitudes temperature seems to be the dominant controlling factor; For the tropics and monsoonal areas it is an effect known as ‘rain-out’ and isotope composition may reflect the strength of the monsoon. To add a further complication the areal temperature coefficient of precipitation composition determined by comparing precipitation isotope composition from different locations with different temperatures, but at the same time is not the same as the temperature coefficient determined at a single site by comparing, for example, summer and winter precipitation. Finally on longer time scales there is a different temperature coefficient that operates between ice age and interglacial precipitation!

    By now you will be gathering that decoding speleothem compositions as a function of temperature is difficult and not unlike using tree rings. I’ll throw another spanner in the works here just to make it that itty bitty bit more difficult. There is plenty of evidence that some speleothems do not grow in isotopic equilibrium with their host water.

    So to derive a true growth temperature record for a speleothem one needs to measure both it’s 18O composition and that of it’s host water. This is difficult but not impossible. I have published a 5000+ year record (now extended to 11,000 years but yet to be written up) of drip water compositions using fluid inclusions trapped in a growing stalagmite. We haven’t published temperatures because we have clear evidence that the stalagmites in the cave we studied are not in isotopic equilibrium.

    As someone who works with speleothems I’m interested in determining a growth temperature history using a well calibrated method that is based on equilibrium thermodynamics rather than wiggle matching my record to some other record that has itself been wiggle matched to yet another record. I can then match that to, for example, a modern temperature record and compare the two. I think then we can say we have an independent continental palaeothermometer. My motto is always caveat emptor!

  16. Posted Dec 5, 2008 at 6:11 AM | Permalink | Reply

    Steve,

    You may want to download the data excel sheet of the Global Network of Isotopes in Precipitation to see that seasonality is the main driver for isotope ratios. It also shows that the correlation between temperature versus isotope ratio is not that strong, especially in the tropics where ratios can be negative due to the monsoon effects.

    Also, it’s quite certain that the isotopes of the Greenland ice cores are a much better proxy for aridity than for temperature. The reason is simple. The most important factor determining the isotope ratio is the temperature at condensation. This temperature however is known as dew point and is merely dependent of the absolute humidity of the air parcel. Moist air has a high dewpoint, drier air has a lower dewpoint. During convection and advection the temperature of the rising air lowers adiabatically and condensation of the water vapor starts at that dewpoint, at lower and warmer levels with moist air and higher colder levels with dry air. Hence the isotopes in precipation are foremost a proxy of the dewpoint, not the ambient temperature.

    Nevertheless I have still to read the first ice core study that mentions “dew point”, yet they all show a pretty neat match between annual snow accumulation and isotope ratio.

  17. bernie
    Posted Dec 5, 2008 at 7:10 AM | Permalink | Reply

    Paul:
    Many thanks for the very clear explanation of the complexities associated with using speleothems as temperature proxies. Following Andre’s point is “the most important factor determining the isotope ratio” different for speleothems than for ice cores?

    • Paul Dennis
      Posted Dec 5, 2008 at 7:43 AM | Permalink | Reply

      Re: bernie (#19),
      Bernie, Andre is right to say that in polar regions there is a marked reduction in the dew point temperature because of the dryness of the air masses. The drier the air, the lower the dew point required to promote condensation. However, he is not quite right in saying that this is the dominant control on high latitude precipitation isotope composition. For temperate and high latitude precipitation there is a remarkably good correlation between mean annual surface air temperature and precipitation isotope composition. To find the reason for this we need to consider the life history of an air parcel and it’s moisture.

      One can view the global precipitation cycle as a simple distillation column. The bulk of moisture in the air is evaporated from the global oceans in the sub tropics where there is a marked excess of evaporation over precipitation. From here air masses migrating polewards will begin to condense water as the dew point is reached. The first water to condense out is enriched in 18O with respect to the vapour phase. The effect of this is to drive the remaining water vapour towards isotopic compositions more depleted in 18O. This process is very simple to describe mathematically and model and is called Rayleigh distillation. Thus as the air mass evolves, cooling and precipitating water its isotope composition is continuously changing to become more depleted in 18O.

      It is this continuous evolution of the isotopic composition of water vapour in an air mass that produces the beautiful relationship between surface air temperature and the isotopic composition of precipitation. The local effect of the dew point temperature on the isotopic fractionation between vapour and liquid is only a second order term in describing the local surface air temperature-isotope composition relationship.

      Now you ask about speleothems and ice core. In high latitude ice core one should expect that negative excursions in delta 18O (i.e. more depletion in 18O) are strongly correlated with local surface air temperature. This accords with very many observations. However, the exact constant of proportionality between air temperature and isotope composition is likely to change from location to location and needs to be evaluated.

      In speleothems the situation is more complicated because we are not measuring the water isotope composition, but the isotope composition of the calcite or aragonite. Here we need to take account of several effects: i) the local recharge water composition. For temperate and high latitude sites this will be controlled by very similar processes to that for polar ice core.; ii) The local cave temperature and it’s effect on the 18O fractionation between calcite and water, and iii) the degree of isotopic equilibrium. Not with standing these observations I would expect to find that more depleted 18O composition of speleothems would indicate cooler climatic conditions.

      The situation becomes very much more complex in low latitude areas. Here we lose the signal between temperature and precipitation and a so called amount or monsoon effect takes over. We might also find that the source region for precipitation changes on a seasonal, or longer time scale. I think that for many low latitude sites, especially those in monsoon areas the link between temperature and isotope composition is very difficult to constrain. This applies to both low latitude, high altitude ice core and speleothems.

      • Craig Loehle
        Posted Dec 5, 2008 at 8:22 AM | Permalink | Reply

        Re: Paul Dennis (#21), I think Paul Dennis makes it clear that input from experts is valued here and that if someone sticks to the science no one is going to poke them in the eye. Thanks, Paul.

        • Kenneth Fritsch
          Posted Dec 5, 2008 at 10:26 AM | Permalink

          Re: Craig Loehle (#23),

          Yes, thanks much, Paul Dennis for your posts. I need to print, reread them and save them for future references.

        • Paul Dennis
          Posted Dec 5, 2008 at 10:34 AM | Permalink

          Re: Kenneth Fritsch (#25),

          Thanks Kenneth

  18. Bob North
    Posted Dec 5, 2008 at 7:23 AM | Permalink | Reply

    I would like to thank both Jud and Paul for coming on here and helping enlighten me on the process of attempting to develop a temperature reconstruction from speleothems and the various ways that dO18 responds to environmental factors. Paul and Jud -this old geologist thanks you very much.

  19. bernie
    Posted Dec 5, 2008 at 8:01 AM | Permalink | Reply

    Paul:
    Thanks again for the clear and quick response. Two more questions – if they are not too simple-minded.
    First, does the movement of the airmass and the rate of precipitation from that air mass may affect the isotope composition? In Massachusetts, for example, is the isotope composition different for precipitation from an “Alberta Clipper” than it is for a regular Nor’easter?
    Second, and again simplistically, it sounds like you would have greater confidence in ice core temperature proxies then mid and high latitude speleothem temperature proxies and least confidence in tropical speleothem proxies based on isotope composition?

    • Paul Dennis
      Posted Dec 5, 2008 at 10:33 AM | Permalink | Reply

      Re: bernie (#22),
      Bernie, interesting and very pertinent questions.

      There is a fair bit of research spread across the meteorological and isotope literature looking at precipitation isotope compositions as a function of air mass and synoptic weather patterns. I fully expect your Alberta Clipper to be very different from a Nor’easter in terms of its isotope composition, and I suspect air temperature too!

      There’s a growing number of national and international programmes to monitor precipitation and vapour isotope composition on sub-event, event, daily, monthly and annual basis that are feeding into the GNIP (Global network for Isotopes in Precipitation) programme. This data is readily accessible if you follow the link that Andre gave in 18 above. Interestingly, I and a colleague have an application in for funding to support a Schools Network for Isotopes in Precipitation (SNIP). We’re looking for schools and others to collect rainfall/snowfall on a daily basis and record max and min temperature, wet and dry bulb temperature, barometric pressure, wind direction etc. We will analyse the rain for oxygen and hydrogen (deuterium and possibly tritium) composition and feed the results into the GNIP database. We will also organise and develop a web site with the results plus educational material suitable for a range of student ages and abilities.

      Finally, if I was asked to put in order my confidence in different archives it would be as you say: polar ice core, mid latitude speleothems then low latitude/tropical speleothems. This reflects our understanding of the chemistry and physics of processes that lead to isotope fractionation in these systems.

      We’re hearing a lot about some of the recent Chinese speleothem data and their widespread interpretation as a proxy for a weakening Asian monsoon over the last 9000 years. However, there are other proxy records, notably the loess palaeomagnetics record that shows that rainfall hasn’t weakened. An alternative interpretation is that the Chinese caves cover an area that receives two different monsoon signals: a declining isotopically depleted Indian monsoon and strengthening isotopically enriched east Asian monsoon (Maher, B., 2008, Holocene variability of the East Asian summer monsoon from Chinese cave records: a re-assessment, The Holocene, Vol. 18, No. 6, 861-866). This one example shows the complexities in trying to decipher a single speleothem record.

      Thank you Craig and Bob for the encouragement.

  20. Timo Hämeranta
    Posted Dec 5, 2008 at 8:52 AM | Permalink | Reply

    Re 8.
    Dear all, if interested you’ll find

    Fleitmann, Dominik, Stephen J. Burns, Augusto Mangini, Manfred Mudelsee, Jan Kramers, Igor Villa, Ulrich Neff, Abdulkarim A. Al-Subbary, Annett Buettner, Dorothea Hippler, and Albert Mater, 2007. Holocene ITCZ and Indian monsoon dynamics recorded in stalagmites from Oman and Yemen (Socotra). Quaternary Science Reviews Vol. 26, No 1-2, pp. 170-188, January 2007,

    online at
    http://www.manfredmudelsee.com/publ/pdf/monsoon-extremes-fleitmann-etal-2007-qsr.pdf

  21. Sam Urbinto
    Posted Dec 5, 2008 at 12:25 PM | Permalink | Reply

    NOAA has some good data source links at http://www.ncdc.noaa.gov/paleo/speleothem.html

    [Steve: most of my graphics are derived from NCDC/paleo data sets]

    An overview may be helpful.

    ———-

    A speleothem (from the Greek for “cave deposit”), commonly known as a cave formation, is a secondary mineral deposit formed in a cave. Speleothems are typically formed in limestone or dolostone solutional caves.

    Water seeping through cracks in a cave’s surrounding bedrock may dissolve certain compounds, usually calcite and aragonite (both calcium carbonate), or gypsum (calcium sulfate). The rate depends on the amount of carbon dioxide held in solution, on temperature, and on other factors. When the solution reaches an air-filled cave, a discharge of carbon dioxide may alter the water’s ability to hold these minerals in solution, causing its solutes to precipitate. Over time, which may span tens of thousands of years, the accumulation of these precipitates may form speleothems.

    Types; Dripstone, Flowstone, Cave Crystals, Speleogens, others.

    Many factors impact the shape and color of speleothem formations including the rate and direction of water seepage, the amount of acid in the water, the temperature and humidity content of a cave, air currents, the above ground climate, the amount of annual rainfall and the density of the plant cover. Most cave chemistry revolves around calcite; CaCO3, the primary mineral in limestone. It is a slightly soluble mineral whose solubility increases with the introduction of carbon dioxide, CO2. It is paradoxical in that its solubility decreases as the temperature increases, unlike the vast majority of dissolved solids. This decrease is due to interactions with the carbon dioxide, whose solubility is diminished by elevated temperatures; as the carbon dioxide is released, the calcium carbonate is precipitated.

    Most other solution caves that are not composed of limestone or dolostone are composed of gypsum (calcium sulfate), the solubility of which is positively correlated with temperature.

    Samples can be taken from speleothems to be used like ice cores as a proxy record of past climate changes.[2] A particular strength of speleothems in this regard is their unique ability to be accurately dated over much of the late Quaternary period using the uranium-thorium dating technique. Stalagmites are particularly useful for palaeoclimate applications because of their relatively simple geometry and because they contain several different climate records, such as oxygen and carbon isotopes and trace cations. These can provide clues to past precipitation, temperature, and vegetation changes over the last ~ 500,000 years.

    Source: Wikipedia

  22. Jud Partin
    Posted Dec 5, 2008 at 5:02 PM | Permalink | Reply

    Steve, did you see Fleitmann’s explanation for why Socotra is inverted on the d18O axis in relation to other monsoon proxies?

    Also, 1) Line one in your post still reads “Heshang” when it should read “Wanxiang”.
    2) I observed nothing about Mann’s orientation. I have only commented on speleothem literature. Please annotate your post above accordingly.
    3) I said “warmer and/or wetter” are generally plotted up on the y-axis in paleoclimate literature. Please change what you have me saying there as well.

    Thanks.

  23. Posted Dec 5, 2008 at 5:37 PM | Permalink | Reply

    I’m afraid I don’t see that there would be any correlation at all, one way or the other, between the green Wanxiang lines in the first figure and instrumental temperature 1850-present. The latter is not plotted, and varies with which index is used, but always has a general uptrend that is lacking in Wanxiang.

    As has been pointed out in this thread and earlier CA threads, d18O often indicates factors other than annual average temperature — seasonality of precipitation, altitude of precipitation, dew point, even wind direction, etc. It may sometimes correlate “correctly” with local or hemispheric or global temperature (eg Dasuopu, Quelccaya Summit), sometimes “backwards,” and sometimes not at all (eg Sajama, Bona Churchill, etc). Until someone runs a regression of Wanxiang on a temperature series and finds a significant t-stat (after correcting for any serial correlation), it appears to be in the “not at all” category, and therefore meaningless as a global or hemispheric temperature proxy.

    Simply correlating with a reconstruction from another set of data does not vindicate the earlier reconstruction as a temperature proxy. It may just mean that the new series is correlated with the non-temperature noise in the earlier reconstruction, even if the earlier reconstruction is valid. (A big if, I gather, in the case of these particular reconstructions.)

  24. Deep Climate
    Posted Dec 5, 2008 at 7:02 PM | Permalink | Reply

    Steve, where exactly did Jud Partin say this?

    Jud Partin observed that the inverted orientation that Mann used for the Dongge O18 record was the same orientation used in the “fantastic” Wanxiang cave record, with the corollary that, even if Mann’s reasons for inverting the Dongge proxy were not wonderful, the actual inversion of the proxy was not objectionable.

    Also, could you clarify what mean by asserting that “Mann’s reasons for inverting the Dongge proxy were not wonderful”? What are these supposed “not wonderful” reasons and how did you infer them?

    Anyway, I don’t see how “inversion” applies to a case like Dongge where Mann’s orientation follows that of the original researchers. Or am I missing something here?

  25. Deep Climate
    Posted Dec 5, 2008 at 8:17 PM | Permalink | Reply

    My point was the narrower one that the Mangini inversion was not “unique” and that Gavin Schmidt’s slagging off Loehle for use of the Mangini proxy therefore required a more substantive argument.

    Steve,
    Gavin Schmidt’s criticism of the Mangini “inversion” (as you call it) as “unique” was based on a comparison with other mid-latitude O18 speleothems. Discussion of low-latitude speleothems, or ice cores for that matter, do not address that particular point.

    Now, I do find Schmidt’s overall analysis of Loehle’s paper quite convincing (not to mention turning up a lot of problems that should have been caught in the E&E review).

    Having said that, the issue of the Mangini negative O18-temperature correlation is perhaps not as clearcut as Schmidt makes out. But he is correct in asserting that positive correlation between O18 and temperature is the norm among mid-latitude speleothems, a point echoed by Paul Dennis above. However, besides Mangini, I’m aware of at least one other exception (Lauritzen and Lundberg in Northern Norway), which was pointed out in the Fairchild speleothem reference I gave you. You are presumably already familiar with that proxy in any case. The gradient there (about -.32 per deg C) is not as steep as Mangini, which still has the steepest negative correlation among mid-latitude speleothems that I’m aware of.

    Perhaps Paul Dennis can comment as to whether he has the same “confidence” in these type of “negative” mid-latitude or high-latitude speleothems as he does in the more common “positve” ones. I presume that as long as there is good evidence of isotopic equilibrium that such proxies could pass muster. I also wonder what local conditions would lead to such an atypical relationship between O18 and temperature and if those would have any implications.

    • Paul Dennis
      Posted Dec 6, 2008 at 9:16 AM | Permalink | Reply

      Re: Deep Climate (#31),
      Deep Climate asked about my views concerning so called ‘inverse’ isotope temperature relationships for mid-latitude speleothems. I’m concerned about them and suspect that they may reflect either changes in groundwater recharge regimes, or isotopic disequilibria rather than specifically being a proxy for temperature.

      The two factors that control speleothem oxygen isotope composition are i) the thermodynamic oxygen isotope fractionation, and ii) the isotopic composition of the cave drip waters.

      The temperature dependence of the fractionation factor varies by -0.2 per mille per degree C.

      The relationship between precipitation isotope composition and temperature areally across the globe is about +0.7 per mille per degree C. i.e. it is a positive relationship. Locally the relationship may differ, but is always positive for mid to high latitude sites. Winter precipitation is isotopically depleted with respect to summer precipitation.

      Combining these two effects we find that the relationship between speleothem isotope composition and temperature is dominated by the precipitation component (typically +0.5 to +1 per mille per degree C) as opposed to the fractionation term (-0.2 per mille per degree C).

      To get strongly negative correlations in mid to high latitude speleothems requires unusual conditions. One way I can envisage that this might come about would be to have changing seasonality of recharge or significant melting of continental ice. One can imagine situations where a warming climate might encourage a larger component of winter recharge through melting of frozen ground. Equally a warming climate might result in melting of isotopically depleted continetal ice. In each case warming allows isotopically depleted water to recharge.

      What worries me about such calibrations is they may reflect a change of state of the hydrologic system in which speleothem isotope composition is not a smooth function of temperature. I think such systems need very careful study and calibration in order to pin down the processes.

      One other thing I think worth mentioning is isotope equilibrium. We often read that a speleothem has been subject to the Hendy test and passed. So what is the Hendy test? Basically it involves measuring oxygen and carbon isotopic composition along a single growth layer from the point where a drip falls onto a growing stalagmite. If the oxygen isotope composition remains constant whilst the carbon isotope composition moves towards 13C enrichment the speleothem is thought to have grown in oxygen isotope equilibrium. The reasoning for this is that as drip water flows as a thin film over a speleothem surface it loses CO2 by degassing. 12C enriched CO2 is preferentially lost. At fast degassing rates so is 16O enriched CO2 and there is a covariation in the C and O isotope compositions. However, if the degassing rate is slow enough then oxygen isotopic equilibrium is maintained between dissolved bicarbonate and water, thus preserving the oxygen isotope composition as the surface film of water evolves and precipitates calcite.

      The problem as I see it is that the Hendy test is not necessarily reliable. I have measured many speleothems that appear to conform to the Hendy test but are clearly out of isotopic equilibrium. I have great concerns that many published speleothems that are claimed to be in isotopic equilibrium may in fact be out of equilibrium. The degree of departure from equilibrium may vary over time, vary with rainfall amount and drip rate, vary with temperature etc.

      You now see why I’m a great believer in trying to recover both water and carbonate isotope compositions, or in using clumped isotopes to try and estimate speleothem temperatures.

  26. Posted Dec 5, 2008 at 9:49 PM | Permalink | Reply

    Re #29, I see now that Ken Fritsch (#50 on the 11/30 thread, “Gavin Schmidt and ‘Uniquely’ Oriented Speleothems”), does show a much more detailed graph of Wanxiang versus temperature at nearby Wudu station, and there is clearly a positive correlation, mostly driven by the behavior post-1980. Jud Partin (#53 of that thread) mentions that Zhang et al report a correlation of R = .8 with (Wudu?) temperature since the 1960s.

    However, Jud goes on to state that Zhang et al believe that the forcing at Wanxiang switched “from natural to anthropogenic around 1960,” and that in the longer-run (pre-1960?), stalagmite d18O

    is negatively correlated with Chinese air temperature taken from a fully coupled GCM with solar and volcanic forcing (r = –0.46 and n = 345) data points.

    The possible reasons for the switch after 1960 include, according to Zhang et al, as quoted by Jud,

    “differences in the nature of solar versus greenhouse forcing, the effect of anthropogenic black carbon on the AM, and/or the effect of anthropogenic sulfate aerosols.”

    In other words, the positive correlation suggested by Ken’s graph may just be due to modern air pollution. The more natural longer-run correlation is distinctly negative, if anything.

    I guess Gavin will be now writing to Science, pointing out that they have just published yet another impermissible “uniquely” negatively-oriented speleothem.

    But seriously, since what we are interested is global temperature, not the temperature at Wudu station per se, for me the natural correlation would not be between Wangxiang and Wudu, but between Wangxiang and a standard global (or even hemispheric) index. This should be done back to 1850, both with and without the possibly pollution-tainted post-1960s data.

    (Incidentally, how do Zhang via Jud get 345 data points out of less than 158 years of data? Are these monthly? Doesn’t instrumental temperature vary a lot throughout the year? If caves reflect the average annual surface temperature, are they using a 12-month average of surface temp? What is the serial correlation of this regression, as measured by DW?)

  27. Deep Climate
    Posted Dec 5, 2008 at 11:17 PM | Permalink | Reply

    #32 References:
    Mangini et al (2005)

    Lauritzen and Lundberg (1999)

    There’s a graphic of Lauritzen here on climateaudit.org, but I’m not sure of its copyright status.

    Fairchild, I.J., Frisia, S., Borsato, A. and Tooth, A.F. 2006. Speleothems. In: Geochemical Sediments and Landscapes

    (ed. Nash, D.J. and McLaren, S.J.), Blackwells, Oxford (in press)

  28. bernie
    Posted Dec 6, 2008 at 11:03 AM | Permalink | Reply

    Paul, what influences degassing rates?

    Your cautious interpretation of the data and acknowledgement of the complexities illustrates what I had assumed was the norm for scientific research. The Hendy test story is interesting – one where field observations conflicts with what I take to be an accepted practice.

    • Paul Dennis
      Posted Dec 6, 2008 at 11:28 AM | Permalink | Reply

      Re: bernie (#35),
      Bernie, I don’t fully undersatnd all the process that go on during speleothem growth. One of the key influences is the difference in pCO2 between the drip water and the cave atmosphere. Most groundwaters have equilibrated with a high partial pressure of carbon dioxide in the unsaturated soil zone. In this zone, because of plant respiration, CO2 pressures can be anything from several times to many 10’s to hundreds of times higher than in the open atmosphere. On entering a cave there is a large pressure gradient between the gas dissolved in the drip water and the cave atmosphere. I suspect the size of the pressure difference may be one controlling factors. The other is the role of the thin water film that drapes over a growing speleothem’s surface. If this film is very thin with a high surface area to volume relationship then degassing is likely to be rapid. On the other hand if the water pools then degassing may be slower.

      Several years ago whilst working on a speleothem from the UK (Mendip Hills) I measured both the water content (trapped in micro fluid inclusions) and the carbon isotope composition. There is a very high degree of inverse correlation between the two. I observed high delta 13C when there was very low water contents and vice versa. My interpretation was that at very slow drip rates ( periods of low precipitation) the water film covering the speleothem was reduced in thickness resulting in less trapped water, higher degassing rates and enriched carbon isotope signatures. I never published the data and now think it might be time to revisit the work.

      Incidentally this speleothem was not in isotopic equilibrium with the local drip water despite looking as though it had passed the Hendy test!

  29. Steve McIntyre
    Posted Dec 6, 2008 at 12:42 PM | Permalink | Reply

    Maher (Holocene 2008) attributes Chinese speleo O!8 variations to changes in source region (specifically between the Indian monsoon and the E Asian monsoon:

    Oxygen isotope records from stalagmites in caves in southern China, interpreted as proxy rainfall records reflecting the intensity of the East Asian summer monsoon, indicate gradual monsoon weakening for the last ~9,000 years, as also documented for the Indian monsoon. Coupled with high-precision dating, the speleothem proxy records have been used to test monsoon links with orbital forcing, solar changes, iceberg discharges in the North Atlantic, ocean currents, and atmospheric methane. However, these ‘benchmark’ cave records do not match other published, dated E. Asian proxy rainfall records (specifically here, independently calibrated rainfall records from loess/palaeosol magnetic properties, and cave oxygen isotope inter-comparisons), which show variable E. Asian monsoon intensity through the entire Holocene. The strong correlation of the cave records with the extra-regional Indian monsoon record yet their mismatch with these other dated Chinese rainfall records might be reconciled if the speleothem isotope variations reflect not changes in Holocene rainfall amount but in rainfall source. Declining Holocene influence of isotopically lighter, Indian monsoon–sourced moisture over China would have resulted in increasing proportions of isotopically heavier rainfall, sourced from the more oceanic E. Asian monsoon. Individual speleothems may thus regionally record Holocene changes in Indian monsoon intensity and isotopic influence. Conversely, the other Chinese proxy records described here reflect changes in rainfall amount, and thus in E. Asian summer monsoon intensity. For the Holocene, the E. Asian and the Indian monsoon responses to orbital forcing are likely to have differed, due to specifically E. Asian internal feedbacks and the seasonal contrasts between the two monsoon systems.

    Given that Jud Partin reported a similar conclusion for the Banda, Borneo speleothem, he could hardly object to a similar interpretation in China. That was also Fisher’s conclusion for the Mt Logan ice core (an ice core near Lonnie Thompson’s unreported Bona-Churchill core, where I successfully predicted lower O18 values in the 20th century based on publishing delay – this is a prediction technique well known to penny mining stock speculators).

    Maybe something like this is going on at Dasuopu as well. But wait, O18 goes up at Dasuopu. Up means global warming, down means changes in regional source. What was I thinking.

  30. Deep Climate
    Posted Dec 6, 2008 at 1:02 PM | Permalink | Reply

    #34
    Paul,

    Thanks for that very lucid explanation.

    To get strongly negative correlations in mid to high latitude speleothems requires unusual conditions…

    What worries me about such calibrations is they may reflect a change of state of the hydrologic system in which speleothem isotope composition is not a smooth function of temperature. I think such systems need very careful study and calibration in order to pin down the processes.

    One of the next obvious questions is this: Are there any high-resolution d180 mid or high-latitude speleothems that would be more suitable as a temperature proxy over the last 1000 to 2000 years? (I realize that speleothems are often used to analyze longer term changes than this). If so, could you point us all to the key studies?

    Again thanks for the benfit of your experience and knowledge – fascinating stuff.

    • Paul Dennis
      Posted Dec 6, 2008 at 2:03 PM | Permalink | Reply

      Re: Deep Climate (#38),
      Deep Climate, your question is highly relevant and I wish I could point you to an unambiguous record. I’m not in my lab/office right now but will post some references next week.

      One of the things that makes the tropical and sub-tropical speleothems attractive, apart from their relevance to the monsoon, ITCZ studies etc. is their very rapid growth rates. I an’t remember the details of the Wanxiang Cave speleothem but think it’s about 12cm long and 1800 years old. This gives it a growth rate of about 8cm per thousand years.

      Most temperate speleothems grow at about 10% of this rate so producing a high resolution record from one is not quite so easy. Having said this we only need 20 microgrammes of carbonate and can micro-mill 10-20 micron slices under computer control so it’s not impossible to produce a record with near annual resolution.

      • Paul Dennis
        Posted Dec 6, 2008 at 2:05 PM | Permalink | Reply

        Re: Paul Dennis (#41),
        Clearly my maths is not very good! A growth rate of over 6cm per thousand years is closer to the mark!

  31. Steve McIntyre
    Posted Dec 6, 2008 at 1:14 PM | Permalink | Reply

    Jud Partin stated that Chinese speleothem plots were conventionally displayed with heavier O18 up. An apparent exception to this convention occurs in Ma, Z., Li, H., Xia, M., Ku, T., Peng, Z., Chen, Y. and Zhang, Z. 2003. Paleotemperature changes over the past 3000 years in eastern Beijing, China: A reconstruction based on Mg/Sr records in a stalagmite. Chinese Science Bulletin 48: 395-400. which shows the following light less depleted O18-up graphic for Chinese speleothems:

    Nonetheless, since our eyes are creature of habit, I’ve replotted the graphic in the figure heavy more depleted O18 up, adding in the Shihua Cave O18 record. I’m having a little trouble discerning a recent uptrend in negatively-oriented dO18 records, other than, to some extent, in the “fantastic” Wanxiang record. LAtitudes: Socotra 12N, Dongge 25N, HEshang 32N, Wanxiang 33N, Shihua 40N.

    • Paul Dennis
      Posted Dec 6, 2008 at 1:57 PM | Permalink | Reply

      Re: Steve McIntyre (#39),
      Steve, when my students use your terminology with respect to isotopes it drives me crazy. You talk of both light and heavy O18! I think what you are saying is that the Shihua speleothem plot in the Ma et al paper is shown with an increasing enrichment in O18 upwards.

      In contrast most Chinese speleothem results are plotted the reverse way round with an increasing depletion in O18 upwards.

      Your point about Barbara Maher’s paper is well made. I’ve discussed these ideas at length with her and I think she has a very valid point. The independent rainfall proxies, such as loess magnetic properties, are not in agreement with the present interpretations of the isotope data as representing a waxing and waning of the east Asian Monsoon. Her model which allows for an interpaly between two different source regions is attractive but needs testing. One obvious test would be to look at the fluid inclusion isotope compositions at key periods in the record.

    • Jud Partin
      Posted Dec 6, 2008 at 2:45 PM | Permalink | Reply

      Re: Steve McIntyre (#39), check out Paul’s #40 comment. I think you have the terminology flipped. By plotting the more negative d18O values as “up” in #39, you have values which are depleted in the heavier 18O as up. You say “heavy O18 up”, but it’s really a depletion in the heavier 18O up. More positive values of d18O indicate an enrichment in 18O:16O (relative to a standard), and vice versa. You may want to change your posts accordingly.

      And thanks for changing the main post to reflect what I said regarding the Wanxiang record. But please refer to the summary in #43 about what I said regarding warmer and/or wetter conditions on the y-axis.

      Steve:
      I’ve substituted the terms more depleted and less depleted. Given the concurrent illustration, I think that the point was clear, but I’m happy to be more precise. On a larger issue, Jud, what do you make of the Dasuopu series?

  32. Deep Climate
    Posted Dec 6, 2008 at 2:16 PM | Permalink | Reply

    #39
    Steve,
    You said:

    Jud Partin stated that Chinese speleothem plots were conventionally displayed with heavier O18 up.

    What Jud actually said:
    #6

    Second, in general (but not always), paleoclimate records are plotted with warmer and/or wetter conditions as “up” on the y-axis. I believe this convention comes from having modern (Holocene) conditions higher than colder conditions during the Last Glacial Maximum on the y-axis. Most paleo-papers you read plot their figures like this.

    #8

    Oh, and the Science plot has lighter d18O up, where you said it didn’t. And I said “warmer and/or wetter” goes up.

    #28

    3) I said “warmer and/or wetter” are generally plotted up on the y-axis in paleoclimate literature.

  33. Posted Dec 7, 2008 at 8:28 AM | Permalink | Reply

    I’ve now gotten the Zhang et al Science paper to download, and obtained the SI (which is publicly available). I see why Jud calls this d18O series “fantastic”, in that it has much more detail and precision of dating than most other series.

    However, it’s just another octodecaoxymometer, like Lonnie Thompson’s ice core data, until it’s formally calibrated to temperature or whatever it is that it proxies. Zhang et al are primarily interested in its relevance to the intensity of the Asian Monsoon, but they do mention temperature, and compare it (inverted so that more d18O, ie less depletion, is down) to three previous temperature reconstructions in the diagram Steve presents above in the original post.

    A fourth panel, “A”, of this digram compares the same green series to an ECHO-G simulation of China temperatures from 1000 to 2000 or so. This is the source of the correlation Jud cited, of r = -0.46 with n = 345 points, so this correlation is not with actual instrumental data, but merely with model-simulated data. Unfortunately, Steve cropped this panel, but the correlation is as good or better than the others.

    But a mere correlation, either end up, between a proxy and a model’s output does not tell us whether either is an accurate measure of temperature. The correlation is interesting, but it may just mean that the model is actually simulating d18O levels, which in turn just tell us wind direction or some such, and not temperature.

    If it could be shown, using the instrumental record, that the proxy is a valid indicator of local or hemispheric or even global temperature, then its pre-industrial correlation to the model output would vindicate the model’s ability to handle natural forcings and therefore would suggest that it might be able to handle anthropogenic forcings as well. But this has not been done.

    In an earlier thread, Ken Fritsch showed Figure S4 from the SI, showing a positive correlation between temperatures at nearby Wudu station and d18O (this time right-side-up), for about 1955 to 2000. The paper reports that this correlation is +.8 after the 60’s, but then suggests that modern anthro factors may be disturbing the historical correlation, whever that is.

    So while it is indeed a fantastically detailed series, its value as a temperature proxy is decidedly less than fantastic at the present.

  34. Steve McIntyre
    Posted Dec 7, 2008 at 8:49 AM | Permalink | Reply

    #46. Hu, the connection to the Thompson Z-mometer is pretty important. The Wanxiang Z-mometer has an opposite orientation to the Thompson Z-mometer tho both are based on O18. Jud Partin d had no comment on this question due to lack of familiarity with the Dasuopu data. Both Z-mometers record O18 of monsoon precipitation in sites affected by the Asian monsoon. At both sites, summer precipitation is more negative (depleted in O18).

    One would hope that someone in the “Community” would be able to explain why both Z-mometers are “significant”, but one O18 Z-mometer has to be placed upside-up and the other upside-down. I’m sure that there’s a “good” reason, but Jud, who’s very knowledgeable, didn’t have the information at hand.

    • Jud Partin
      Posted Dec 7, 2008 at 12:14 PM | Permalink | Reply

      Re: Steve McIntyre (#47), so it’s like that?

      Jud Partin ducked this question due to lack of familiarity with the Dasuopu data.

      I didn’t “duck” anything. I have been upfront and honest. You asked about Socotra and Dasuopu. I pointed out the Socotra ref and their explanation – which is what you asked for. I try not to make the all-too-common human judgment error of commenting on things I haven’t read about. I have not read the Dasuopu paper. I leave on Friday for a 5 week trip to the West Pacific – which I previously mentioned. I don’t have the time to read and comment on Dasuopu. There’s no need for you to say I “ducked” the issue. You have mis-quoted me and made many errors (both scientifically and editorially) in this post. I have not made any negative comments about you in doing so. I just tried to help out and fix things.

      …so now I ask myself: self, what do I gain by coming here, getting mis-quoted, and having people talk down to me???

      BTW, you STILL have not fixed the “warmer” comment that you have me saying. See several of the posts above…

      Steve: Jud, in the same comment, I said “Jud, who’s very knowledgeable, didn’t have the information at hand” – so there was no slight intended. I’ve amended this to say that you had “no comment” on the question.

      • Jud Partin
        Posted Dec 7, 2008 at 3:49 PM | Permalink | Reply

        Re: Jud Partin (#52), thank-you for the clarification. Also, in the main body post there are still two things that need changing.

        For reference, I’ve plotted five O18 series below: Socotra (from Mann data), Wanxiang, Heshang HS4, Dongge D4 and Dasuopu. All of these series are O18 series and all are monsoon records. All are oriented with lighter O18 up (this is opposite to the inverted orientation of the Science graphic). (Jud Partin suggests below that “warmer” be plotted up; for now, I want to show O18 for all series in a consistent way, since, for now, I’m interested in consistency between O18 and, if I invert all of them to accord with the interpretation of the speleothems, it doesn’t accord with the usual interpretation of Dasuopu.)

        It should read “with heavier 18O, or enriched d18O values up”

        And “Jud Partin pointed out that it is customary for warmer and/or wetter conditions be plotted up in paleoclimate literature”.

        All you have to do is cut and paste.

  35. Posted Dec 7, 2008 at 9:56 AM | Permalink | Reply

    Re Steve #47.
    I agree that there are some very important issues here. At least the Thompson 6 (CC03) ice cores, when significant, are all correlated with global temperature in the same positive direction. Even the two utterly insignificant cores (Dunde and Sajama) have positive point estimates. (I used HadCRU3v GL and was constrained by the reported data to run the correlations on decadal averages, but I think these results would be fairly robust.)

    If speleothem d18O series are responding to something irrelevant, like wind direction or seasonality of precipitation, then one would expect them to be uncorrelated with temperature at the 5% test size, except for 5% of the time.

    Mangini et al find a negative correlation between their alpine speleothem and temperature. They do not report a t-stat, but the 95% confidence interval in their graph shows that the slope must be highly significant despite using only 5 points. I agree with Gavin that this correlation is suspicious. Craig correctly used it for his exercise on the grounds that it met his criteria of being a peer-reviewed, author-calibrated published temperature series, but that doesn’t mean we shouldn’t now reconsider it. (Thanks, DC, for providing a link to the paper in one of these threads.)

    Now that Mann et al 2008 and Zhang et al 2008 are all on record that negative correlations are OK for speleothems despite Gavin’s view to the contrary, the issue should be all means be investigated. It sounds as though pros like Jud and Paul, at least, are giving it some hard thought.

    Once it’s decided which way the correlation “should” go, that still leaves the general question of what one should do with the occasional mavericky series that significantly (or even insignificantly) have the “wrong” correlation? Should they just be included in multivariate calibrations and hope that they are dominated by the ones with the “right” correlation? Or should one impose a prior that the correlation must have a certain sign and so exclude them? Would excluding them cause the precision of the calibration to be overestimated? Presumably one would still test the validity of the prior with the full data set.

  36. Posted Dec 7, 2008 at 10:23 AM | Permalink | Reply

    I may be the last person to have figured this out, but I have just discovered that it is easy to extract digital data from PDFs like the Zhang et al. SI.

    The only trick is that you have to turn off the “hand” tool and replace it with the “select” tool using the “tools” menu or the appropriate icon on the toolbar. Then select the desired data, copy it to the clipboard, paste it into a Notepad (or whatever) file, and clean it up to where it can be read as a space-delimited text file into your program or spreadsheet.

    This would have saved me a lot of time (and transcription errors) hand-copying some of Thompson’s ice core data from a PDF he had archived!

  37. Steve McIntyre
    Posted Dec 7, 2008 at 10:41 AM | Permalink | Reply

    #49. Yup. I’ll try to remember to archive the cleanedup versions when I do it. (I also did this process with the Thompson PNAS data.) I’ve got some other odds and ends on Thompson ice core data – Hans Erren has a way of getting digital information from graphics in pdfs that’s very helpful and has been very generous in doing this for me from time to time.

    • Not sure
      Posted Dec 8, 2008 at 7:41 PM | Permalink | Reply

      Re: Steve McIntyre (#50),

      Hans Erren may have purchased one of the many commercial PDF extraction and conversion tools out there. Let me know if you’re ever stuck trying to extract data from a PDF. There are many open-source libraries for PDF manipulation.

  38. Paul Dennis
    Posted Dec 7, 2008 at 11:29 AM | Permalink | Reply

    There is a serious issue as to whether or not Wanxiang cave speleothem WX42B, and possibly other spelothems, are in oxygen isotopic equilibrium with their parent drip water. I’ve been thinking this through for several weeks since having had a paper for peer review on the equilibrium fractionation factor for oxygen isotope equilibrium between calcite and water. This new work suggested that published fractionation factors were incorrect and about 1.5 to 2 per mille too low. This is not a new suggestion. In 2007, Coplen of the USGS published a paper in Geochimica Cosmochimica Acta on Holocene calcite deposited in the Devils Hole system in Nevada. These waters are warm (33 degrees C) and precipitate calcite very slowly. Coplen found that using the standard published fractionation factors the calculated precipitation temperature was too low by 8 degrees C. Alternatively if he fixed the precipitation temperature and calculated the composition of the precipitating water this was now found to be 1.5 per mille depleted in 18O wrt to the measured composition.

    With this in mind I’ve gone back to the Wanxiang speleothem WX42B to see if I can check if it is growing in equilibrium. The cave is at 11 degrees C, and receives drip water with an average oxygen isotope composition close to -8.95 per mille (VSMOW) during the period 1999-2002 (Liu et al. 2008).
    The top most layer of the speleothem has an oxygen isotope composition of -8.15 per mille (VPDB) (Liu et al. 2008).

    If Coplen and the new work are correct then the calculated speleothem composition for Wanxiang cave should be between -6.4 and -6.8 per mille. i.e. the speleothem is not growing in isotopic equilibrium. I stress that this estimate is based on Coplen’s and a more recent studies new assessment of the equilibrium oxygen isotope fractionation factor. If we utilise existing estimates of the fractionation factor then the speleothem looks as though it is in equilibrium.

    I also checked the results of the Hendy test, and whilst both Zhang et al (2008) and Liu et al (2008) suggest the speleothem passes the test I would disagree. 4 layers were studied and are plotted in Figs 2A and B of Liu et al (2008). In 3 of the layers there is a distinct enrichment in 18O as one moves away from the growth axis. Moreover, also a positive correlation between 13C and 18O.

    What does this show. Well it’s possible that speleothem WX42B is in isotopic equilibrium, but also possible that it is far from equilibrium. This shows just how far we still have to go in developing and refining the oxygen isotope thermometer for speleothems and other carbonates. My verdict at the moment is that the jury is out. The recent work on fractionation factors is persuasive and if upheld, then clearly WX42B is far from equilibrium.

    Coplen, T., 2007, Calibration of the calcite–water oxygen-isotope geothermometer at Devils Hole, Nevada, a natural laboratory Geochim. Cosmochim. Acta, V71, 3948-3957

    Liu et al., 2008, Asian monsoon precipitation recorded by stalagmite oxygen isotopic composition in the western Loess Plateau during AD1875-2003 and it’s linkage with the ocean-atmosphere system, Chinese Science Bulletin,V53, 2041-2047

    Zhang et al., 2008, A test of climate, sun and culture relationships from an 1810 year Chinese cave record, Science, V322, 940-942

  39. Posted Dec 7, 2008 at 1:08 PM | Permalink | Reply

    Jeff #53 —
    Thanks for the info. However, the Zhang et al SI is not copy-protected, so this procedure works fine for it, at least. The only complication is how it’s arranged in 4 columns on each page, but that can easily be rearranged once it is in the computer.

  40. Deep Climate
    Posted Dec 7, 2008 at 1:22 PM | Permalink | Reply

    #54
    Steve,
    I don’t know if this is the only correction needed, but you still have this misstatement in #39.

    Jud Partin stated that Chinese speleothem plots were conventionally displayed with heavier O18 up.

    Please reread #43 as Jud has already suggested.

    generally, warmer and/or wetter goes/is plotted up [x3]

    Also it appears that at least some of the series in #39 are not displayed in the orientation suggested by Jud.

  41. Deep Climate
    Posted Dec 7, 2008 at 2:00 PM | Permalink | Reply

    #32

    I guess Gavin will be now writing to Science, pointing out that they have just published yet another impermissible “uniquely” negatively-oriented speleothem.

    #48

    Now that Mann et al 2008 and Zhang et al 2008 are all on record that negative correlations are OK for speleothems despite Gavin’s view to the contrary.

    Hu,
    Like Steve before you, you appear to have misconstrued Gavin Schmidt’s statement regarding the negative correlation in Mangini.

    In mid-latitudes, the relationship is positive – higher d18O in precipitation in warmer conditions. This is a function of the increase in fractionation as water vapour is continually removed from the air. Most d18O records – in caves stalagmites, lake sediment or ice cores are usually interpreted this way since most of their signal is from the rain water d18O. However, only one terrestrial d18O record is used by Loehle (#9 Spannagel), and this has been given a unique negative correlation to temperature. [emphasis added]

    So Gavin’s satement clearly indicates that the “unique negative correlation” is with respect to mid-latitude speleothems, not low-latitude ones like Dongge and so on. I’ve already pointed this out in #31 and numerous previous comments, yet the misunderstanding continues.

  42. Posted Dec 7, 2008 at 3:09 PM | Permalink | Reply

    RE DC, #56, thanks for patiently clarifying Gavin’s position.

    Although Dongge (25 deg 17 min N) is technically in the Temperate Zone, it is below 30 degrees, and therefore broadly in the Hadley Cell circulation rather than the Ferrell Cell. However, Wanxiang (33 deg 19 min N) is above 30 degrees. Would it not be “mid-latitude”, or would the “Horse Latitudes” (30-35 or so) still count as “low”?

    • Deep Climate
      Posted Dec 7, 2008 at 11:42 PM | Permalink | Reply

      Re: Hu McCulloch (#57),
      Hu,
      As far as I can tell, “middle latitude” is demarcated at somewhere between 30 and 35 degrees, depending on the context and particular academic tradition. The Wikipedia definition seems to be quite common and states:

      The middle latitudes are between 33 degrees 33′ 33″ North and 66 degrees 33′ 33″ [North,] and 64 degrees 33′ 33″ South and 33 degrees 33′ 33″ South latitude.

      Of the speleothems listed above, only Shihua Cave falls squarely in the mid-latitude range by this definition.

  43. Deep Climate
    Posted Dec 7, 2008 at 3:19 PM | Permalink | Reply

    #52: Jud, in addressing Steve, you said:

    You have mis-quoted me and made many errors (both scientifically and editorially) in this post. I have not made any negative comments about you in doing so. I just tried to help out and fix things.

    …so now I ask myself: self, what do I gain by coming here, getting mis-quoted, and having people talk down to me???

    Jud, I understand your frustration and can accept that you personally have little to “gain by coming here”. On the other hand, Climateaudit.org is currently registering almost twice as many readers as Realclimate.org according to the latest daily statistics from Alexa, so you are providing a greatly appreciated public service in correcting the “many errors” you and others have identified.

  44. Steve McIntyre
    Posted Dec 7, 2008 at 3:53 PM | Permalink | Reply

    I have edited this post describing the orientaiton of the graphics as “negative dO18 up ” or “negative dO18 down”, in order to have an interpretation-neutral description of the orientation, which is all that I was trying to do here.

    If I’ve stumbled on speleothem terminology (and I expressly noted that I’m in the process of learning this data), I apologize.

    However, the point of the post remains despite the sniping: the orientation of the Dasuopu ice core series as used by Thompson and by Mann and the orientation of the Socotra speleothem as used by Mann (though not in the original article) are opposite to the orientation of the three Chinese speleothems and that the orientation of the Chinese speleothems (Dongge, Heshang and Wanxiang) and that the orientation of the Chinese speleothems is the same as the orientation of the Mangini temperature reconstruction.

    • Jud Partin
      Posted Dec 7, 2008 at 6:02 PM | Permalink | Reply

      Re: Steve McIntyre (#61), It’s not that you

      stumbled on speleothem terminology (and I expressly noted that I’m in the process of learning this data), I apologize.

      You are confused about the concept of d18O. That is basic stable isotope geochemistry, not speleothem terminology. Which happens to be the question that you are asking about in regard to the record orientation. I think that a textbook on stable isotope geochemistry would help you with the subject. I recommend “Principles of Isotope Geology” by Gunter Faure.

      And Re: Kenneth Fritsch (#63), if you describe me as having a “bit of an edge”. …how would you describe the attitude of the comments on the “So How’s they do that?” thread?

      • M. Jeff
        Posted Dec 7, 2008 at 6:34 PM | Permalink | Reply

        Re: Jud Partin (#65),

        You say:

        You are confused …

        Seems a little edgy.

      • Geoff Sherrington
        Posted Dec 7, 2008 at 8:24 PM | Permalink | Reply

        Re: Jud Partin (#65),

        At the risk of repetition, as one who has worked with isotopes for decades, I repeat below part of post 46 of “Gavin Schmidt and Uniquely Oriented Speleothems” from CA Nov 30 2008.

        Maybe this is the first time I have seen scientists publish papers where the relation between two main variables can be positive or negative at will. That indicates to me an immature state of development, which further indicates that publications should be full of qualifications and not assertive or brashly deductive.

        With speleothems and carbon isotopes, we have at least non-CO2 carbon in air in the atmosphere, air in the cave, CO2 in the atmosphere, CO2 in the cave, calcite and related minerals in the solid phase, calcite precursors in the liquid phase adjacent, carbon in cave lakes sometimes.

        We have, at least, oxygen isotopes in air, CO2 and other atmospheric gases like NOx. We have them in rainwater, the seas/lakes etc that are the souce of the rainwater, in groundwater, in water proximal to calcite deposition, in calcite and related minerals.

        We have fractionation in the biomass which decays and produces soluble isotopes that might end up in the speleothem.

        We have mobility, mineral phases, fractionation, lagging, unpredictability of paths of gases and liquids and limited, good correlation data.

        The multiple ways in which these factors can be combined is complicated by an uncertain relation between (say) oxygen isotope ratios and temperature. Many isotope fractionation mechanisms are qualitative as observed, but quantitative by the time they are published, maybe with the help of circular logic to tie down temperature.

        As a geochemist, this subject of speleothems is a nightmare because of the abundance of uncontrollable variables. One of these is error-prone instrumental temperature reconstructions, whose close analysis indicates that they can be the first basic problem. I can show you records where the raw surface temperature data at isolated places has declined steadily over the past 50 years.

        Like many others, I welcome the chance for science to reconstruct the past with confidence. I am dismayed when methods fail after early promise (like dendrothermometry). That chance will not be realised by poor science that does not stand audit, even when raw data are graciously supplied.

  45. Steve McIntyre
    Posted Dec 7, 2008 at 4:36 PM | Permalink | Reply

    http://www.blueplanetbiomes.org/climate.htm defines “mid latitude” as 30-55 deg. So both Heshang and Wanxiang would be “mid-latitude”.

  46. Kenneth Fritsch
    Posted Dec 7, 2008 at 5:27 PM | Permalink | Reply

    DC and Jud Partin, you both seem to have a bit of edge when you start lecturing protocol and generalizing about CA, but as long as you can contribute to learning process here and perhaps learn yourselves, I personally welcome your (science-related) comments.

  47. RomanM
    Posted Dec 7, 2008 at 5:50 PM | Permalink | Reply

    DC, from your comments, it seems pretty clear that DC doesn’t stand for Deep Climate, but rather for Defender of the Consensus.

  48. Steve McIntyre
    Posted Dec 7, 2008 at 10:35 PM | Permalink | Reply

    #62. Nope, I’m not confused about isotopes.

    What I don’t understand is what I set out (with sufficient plainness) in the head post – how seemingly contradictory interpretations exist in prominent studies – why seemingly like dO18 data sets (the Chinese speleothems on the one hand) and Dasuopu ice core (and Mann’s Socotra) on the other can have opposite orientations resulting from mechanisms that are similar in relevant respects. That was the original issue and no one that’s been parsing the form of the question has even attempted to provide an answer in terms of temperature or amount of precipitation. I do not exclude the possibility of an answer to this question, but so far no one’s risen to the occasion. And I am certain that the answer to that particular question does not reside in an isotope textbook.

    Based on the discussion so far, the suggestion that I discussed above – Maher’s suggestion that Chinese speleothem dO18 (like Jud’s Borneo speleothem) is determined by changes in wind direction rather than either amount of precipitation or temperature.

    • Jud Partin
      Posted Dec 8, 2008 at 12:07 AM | Permalink | Reply

      Re: Steve McIntyre (#68), Mann did not measure the Socotra record.

      why seemingly like dO18 data sets (the Chinese speleothems on the one hand) and Dasuopu ice core (and Mann’s Socotra) on the other can have opposite orientations resulting from mechanisms that are similar in relevant respects

      Which mechanisms? Similar in which relevant respects?

      That was the original issue and no one that’s been parsing the form of the question has even attempted to provide an answer in terms of temperature or amount of precipitation. I do not exclude the possibility of an answer to this question, but so far no one’s risen to the occasion.

      That is not correct. I gave you a reference which describes why Socotra would be reversed.

    • Paul Dennis
      Posted Dec 8, 2008 at 2:40 AM | Permalink | Reply

      Re: Steve McIntyre (#68),

      I’m now quite confused, probably because I’m interested in the speleothem records and understanding them at a process level, but relatively insensitive to what Mann and others are using them for!

      It is very important to get to the fundamentals of what is going on in these archives and I’ll make this statement up front. The record from speleothems and ice core collected in monsoon regions (this includes tropical, sub tropical and mid-latitude areas and also, incidentally those tropical areas impacted by the ITCZ) record an amount effect in their oxygen isotope signal. This effect is relatively well understood in terms of processes. There is an inverse correlation between rainfall amount and the oxygen isotope composition. The greater the rainfall the more 18O depleted it becomes.

      The Wanxiang record and the Dasuopu ice core record this effect. At Dasuopu there have even been studies in the firn layer covering the last decade or so that show the monsoon effect. Strong summer and late summer monsoonal rainfall is depleted in 18O compared to the winter and spring precipitation which is weak. Interestingly the amount effect leads to a negative correlation between local temperature and 18O in rainfall or ice when measured on a seasonal basis.

      Taking this as a standpoint, at first run through the Dasuopu (and Socotra) and Wanxiang records over the past few centuries would seem to have opposite trends. i.e. Dasuopu and Socotra look like a weakening of the summer monsoon, whilst Wanxiang shows a marked strengthening. I believe that this is the point that Steve is making. Steve correct me if I’m wrong.

      To Parse the trends from Dasuopu (Socotra) and Wanxiang into global climate trends means one has to imply that the monsoon is weakening in Dasuopu and this is associated with a warming temperature, whilst it is strengthening in Wanxiang. I don’t think this is a tenable position unless one has very good process based evidence to back this up.

      Steve has picked up on the work of Maher in which she suggests that there is a subtle interplay between both the East Asian and Indian monsoons. That rather than changes in precipitation what we are seeing are changes in the relative contribution of the two different systems to the total precipitation at a single site. It is possible that this process also affects the high altitude glaciers as well.

      How can we test this. I’d suggest that with our present level of knowledge we can’t. Standing at a single location and looking back in time is like sitting in a small boat in the middle of the ocean, taking a sample from the bottom and then saying you understand oceanic circulation. You can put in all the models, all the caveats, all the arm waving you like but you are still, basically guessing.

      One way forward is to look at a large number of well dated proxies and construct areal time slices of, for example, palaeoprecipitation isotope composition, that will allow a much better understanding of the synoptic meteorology at any one time and then to test how this evolves.

      I think that Barbara Maher has made a very important point and counter hypothesis but don’t yet have the data to test it.

      I’ll just say one other thing about the 13C records. I’ve looked at Socotra again and some of the other records. Interpreting these records is at best arm waving. There could be a C3 to C4 transition, dissolution in the epikarst might have added more rock carbon, slower precipitation rates might lead to more evolved 13C isotope compositions in the speleothems. However, how you wrap this up in a mechanistic model that translates to local conditions, yet alone teleconnects to global climate is beyond me. Physics and chemistry it certainly is not!

  49. Geoff Sherrington
    Posted Dec 7, 2008 at 10:58 PM | Permalink | Reply

    Steve, the explanation probably exists in these words from

    The Holocene 9,6 (1999) pp. 659–669
    Calibration of the speleothem delta function: an absolute temperature record for the Holocene in northern Norway
    Stein-Erik Lauritzen1 and Joyce Lundberg2
    (1Department of Geology, Bergen University, Alle´gaten 41, N-5007 Bergen, Norway; 2Department of Geography, Carleton University, Ottawa, Ontario K1S 5B6, Canada)

    The components TFr and F(T, t, g) have different temperature sensitivities. TFr always has a negative response to temperature (i.e., heavier d18Oc values imply decreasing temperature), while F(T, t, g) may respond either positively or negatively, depending on regional meteorology. Consequently, the temperature response of speleothem carbonate is entirely dependent on the relative magnitude of TFr and F(T, t, g). The temperature response rate (m) of equation (3) is, in mathematical terms, its T-derivative (Lauritzen, 1995) defined as:
    (see equation) (6)
    which can, in principle, be negative, zero or positive. This can be
    further elucidated with respect to the relative magnitudes of T responses of (TFr) and of F(T, t, g) (Table 1).
    TFr alone will give mu less than 0 but in the cases of mu greater than 0 or mu much,much less than 0, variations in d18Ow of the precipitation dominate. The temperature sensitivity and therefore the interpretation of speleothem d18Oc is by no means straightforward and often ambiguous, but the problem may be overcome in ways that either aim at estimating F(T, t, g), or just estimating the sign of m:

    There, that’s not so hard and it’s half Canadian. (My emphasis)

  50. bernie
    Posted Dec 8, 2008 at 6:54 AM | Permalink | Reply

    Paul:
    Thank you for a very clear statement. However, let me make sure I fully understand you position: At the moment, you would not use speleothems as temperature proxies. They may turn out to be good paleoclimate temperature proxies but considerably more analysis of multiple (can you hazard a number for a given region?) well documented speleothems.
    Did I summarize your position correctly?

    • Paul Dennis
      Posted Dec 8, 2008 at 7:30 AM | Permalink | Reply

      Re: bernie (#73),

      For me the phrase temperature proxy has a very definite meaning that is associated with two questions: i) Does the isotope composition of the speleothem accurately reflect the cave temperature at the time it grew, and ii) can I explain the transfer function between isotope composition and temperature using equilibrium thermodynamics and understanding of the controls on the dripwater composition.

      In monsoonal areas the control on speleothem composition is not temperature but the amount and source region effects. I wouldn’t use such speleothems to measure local temperature.

      Neither do I think it is right to use them as a z-mometer as Steve and Hu put it. That is take a profile and wiggle match it to either local events, or teleconnect it to other events.

      On the other hand I do believe that speleothems, and in particular temperate and high latitude examples, do have the potential to be a very good thermometer but, and it’s a very big but, we need to understand the physics and chemistry of their growth. I believe that there is a lot of effort going into measuring speleothem profiles and rather too little in examining the fundamentals of the relationship between temperature and isotope composition.

      I don’t want to go on too much about this but the only unequivocal way to measure isotopic temperatures is to measure either i) both the drip water and coeval speleothem isotope composition, demonstrate isotopic equilibrium and then apply the correct equilibrium isotope fractionation factors to determine the temperature, or ii) use another internal thermometer such as the ordering of 18O and 13C in the carbonate anion. This latter approach is the so called ‘clumped isotope’ method (cf. Affek et al. 2008, Geochim Cosmochim Acta V72, 5351-5360). Both approaches are still fraught with great technical difficulties. This is why you will see just a handful of papers addressing these problems but very many more with fantastically detailed profiles.

      In terms of the number of well characterised studies then how long is a piece of string? We are barely scratching the surface and I would like to see us start within a single cave and demonstrate coherence between speleothems of similar age. Very few studies like this have been carried out even to check for the coherency of signal within a single hydrologic system.

  51. bernie
    Posted Dec 8, 2008 at 8:04 AM | Permalink | Reply

    Paul,
    Many thanks. As to the “length of a piece of string” – I guess, I will have to wait until you endorse some work in this area.

  52. Posted Dec 8, 2008 at 8:49 AM | Permalink | Reply

    Steve, #68, writes,

    What I don’t understand is what I set out (with sufficient plainness) in the head post – how seemingly contradictory interpretations exist in prominent studies – why seemingly like dO18 data sets (the Chinese speleothems on the one hand) and Dasuopu ice core (and Mann’s Socotra) on the other can have opposite orientations resulting from mechanisms that are similar in relevant respects.

    If my year of Caltech chemistry (back in the MWP or thereabouts) still serves me right, a big factor that may cause different behavior between ice cores and speleothems is that the chemistry involved is entirely different.

    There are 4 big sources of O involved — the gas O2, water H2O, carbonate CO3, and the gas CO2. Atmospheric O2 is presumably the norm from which the “standard” ratio of 16O to 18O is measured, but H2O, CO3, and CO2 can have their own ratios that may vary with the fractionation the sample has been through.

    Ice core O18 measures the O18 in pure H2O, which has precipitated from the atmosphere without reacting with O2 or CO3 in the process. This may indicate the seasonality of the precipitation, the altitude of the precipitation, or even the annual average temperature at the time of the precipitation, whence its potential as a (noisy) indicator of global temperature.

    Speleothem O18, on the other hand, must be from the calcite (or chemically equivalent aragonite) CaCO3, which contains no water of crystalization. There is a monohydrocalcite that does contain an H2O, but it occurs only occasionally in cave deposits, and would not be the building block of standard speleothems.

    While the groundwater that carries the Ca and CO3 ions contains O18 whose content was determined during precipitation of rainwater or snow, the O in H2O and the O in CO3 do not exchange in solution, so that there is no direct connection between the two sources of O18. The CO3 ions are primarily from dissolved limestone. The rate of dissolution may depend on the temperature of the groundwater and involve fractionation, but this is a totally different mechanism than precipitation. A little of the CO3 may be dissolved atmospheric CO2, but its dO18 is primarily unrelated to that of the water in the precipitation.

    The only place the cavewater H2O and CO3 can exchange their O’s is when CO3 outgasses (leaving behind an O in the water), and then is redissolved (acquiring an O that is unrelated to the one that was lost). This will happen continually in equilibrium, but must be a very slow process (albeit one that could itself cause fractionation, depending on temperature). Otherwise, O2, H2O, CO2 and CO3 are pretty stable and don’t do much exchanging of O’s, except through highly fractionating photosynthesis and biodegradation.

    When the speleothem is finally formed, the O18 is coming from the CO3 in the water, not the water itself. There will be further temperature-dependent fractionation as some CO3’s precipitate as CaCO3 and some don’t, but it would only be by accident that the end product would have the same O18 content as the groundwater.

    If I’m muddled, perhaps someone could please correct me.

    • Paul Dennis
      Posted Dec 8, 2008 at 9:00 AM | Permalink | Reply

      Re: Hu McCulloch (#76),
      Hu, here are a very few quick guide points. I have a meeting to go to so it will be quick;

      1) The norm for all isotope ratio measurements of water samples is the standard mean ocean or VSMOW (the V standing for Vienna) and this water has a composition of 0. All waters that are enriched in 18O wrt to this standard are positive and all that are depleted are negative. For carbonate samples it is VPDB. VPDB no longer exists but was a belemnite from the Cretaceaous Pee Dee formation. Instead of VPDB we actually use NBS-19 TS limestone which is assigned a value wrt to VPDB. Apocryphally TS stands for Toilet Seat…I can explain why some other time.

      2) The water cycle fractionates the isotopes through evaporation and condensation phase changes. Freezing has little effect on the isotope fractionation. Thus rainwater is depleted wrt to seawater.

      3) There is oxygen isotope exchange between dissolved inorganic carbon species and the water molecule. The mechanism and rate of exchange depends on the dominant speciation which in turn depends on the water pH.

      4) Carbonate precipitated in isotopic equilibrium with water is enriched by ca. 28 per mille in 18O when compared to its parent water.

      Phew…a very quick tour through the hydrologuic cycle and carbonate precipitation.

    • Jud Partin
      Posted Dec 8, 2008 at 9:05 AM | Permalink | Reply

      Re: Hu McCulloch (#76),

      While the groundwater that carries the Ca and CO3 ions contains O18 whose content was determined during precipitation of rainwater or snow, the O in H2O and the O in CO3 do not exchange in solution, so that there is no direct connection between the two sources of O18.

      No. HCO3- (the dominant species at groundwater pH) is constantly exchanging O molecules with H2O until they reach an equilibrium. Then the rate of forward reaction equals the rate of reverse reaction. Even then, they are still exchanging O molecules, but no net change.

      Speleothem calcite d18O reflects groundwater d18O which = rainwater d18O (if the calcite precipitates in isotopic equilibrium with the cave dripwater – Paul’s point)

  53. Steve McIntyre
    Posted Dec 8, 2008 at 9:51 AM | Permalink | Reply

    Paul, I appreciate your re-statement of my point and you’ve captured it precisely.

    Hu, Paul and Jud – my take on the literature for both ice cores and speleothems was that the interpretation of the core dO18 as a climate record was, in both cases, based on the assumption that the core dO18 more or less corresponded to the dO18 of the precipitation, notwithstanding the various things that happened at the ice/speleothem level.

    Also that for both the Dasuopu ice core and the Chinese speleothems in question, the local precipitation had less negative dO18 in wintertime and more negative dO18 in the summertime (the latter due to monsoon rainout in both cases.) It is also my understanding that this seasonality is opposite to polar regions, where winter dO18 is more negative than summer dO18.

    The interpretation of the Chinese speleothems, as I understand it, is that the dO18 in the speleothems (and presumably local precipitation) has become more negative during the 20th century, interpreted, as Paul observes, as a strengthening of the monsoon, and, in turn, in some circles, as a proxy for large-scale temperature increases.

    At Dasuopu, the dO18 trend in the 20th century core (and presumably local precipitation) is less negative in the 20th century, interpreted as powerful evidence of global warming by Lonnie Thompson. On a step-by-step basis, it would need to be interpreted, as Paul observes articulating my original point, as a weakening of the monsoon, requiring seemingly contradictory theories.

    I’m not in a position to assert that these are actually contradictory positions, as there may be other factors involved, but they are seemingly contradictory.

    Jud, I’ll comment separately on Fleitmann 2007.

    • Jud Partin
      Posted Dec 8, 2008 at 10:25 AM | Permalink | Reply

      Re: Steve McIntyre (#79), from Fleitmann 2007

      “Because the onset and termination of the summer monsoon on Socotra determines the end of the spring and start of the autumn precipitation season respectively (as it is in Eastern Africa; Camberlin and Okoola, 2003), it is conceivable that a gradual shortening in the length of the summer monsoon season resulted in an extension of pre- and post-monsoon rainy season precipitation on Socotra.”

      The seasonality of rainfall (annual cycle) at a site must also be considered. Not just that it’s in a monsoon area so they must all be alike. This is a point we make in the Cobb 2007 paper. Not that the speleothem record is a proxy for wind direction. It’s not an anemometer or a wind sock.

      In N. Borneo, there is a limited seasonal cycle in rainfall. However we observed a pronounced seasonal cycle in rainfall d18O. This cycle is most likely due to the wind reversal of the monsoon causing more rainout across Borneo when the winds are from the S. The most pronounced rainfall anomalies in the N. Borneo instrumental record are during ENSO events. Changes in rainfall d18O during ENSO events are controlled by the amount effect. Therefore, interannual rainfall d18O anomalies – and hence stalagmite d18O – are due to changes in the amount of precipitation.
      From Cobb EPSL 2007

      Interannual rainfall and dripwater δ18O anomalies at Gunung Mulu are linked to regional-scale changes in convective activity associated with ENSO variability, whereby increased precipitation drives more negative rainfall and dripwater δ18O, consistent with the ‘amount effect’.

      Maybe you accidentally left that out of your post.

  54. bernie
    Posted Dec 8, 2008 at 12:06 PM | Permalink | Reply

    Jud:
    Is the rainout hypothesis supported by measurements from speleothems in S. Borneo?

  55. Posted Dec 8, 2008 at 1:02 PM | Permalink | Reply

    Paul and Jud #77, 78, Thanks for setting me straight on the chemistry of CO3.

    So, if the O18 in H2O and CO3 in fact mixes well in ground water, and this is the primary source of the O18 variation in speleothems, Steve’s argument that ice cores and speleothems (at similar latitudes, anyway) should have a qualitatively similar response (if any) to air temperature is correct after all.

    Or, as Snoopy would say as his pet theory descended in flames, “Curse you, Red Baron!”

    I am still intrigued, though, by Paul’s statement,

    The water cycle fractionates the isotopes through evaporation and condensation phase changes. Freezing has little effect on the isotope fractionation.

    My understanding is that snow develops directly from water vapor, and not from vapor that has first condensed into the liquid state. Is the d18O variation in ice cores then primarily due to temperature variations at the time of evaporation (from the ocean or wherever), rather than at the time of freezing?

    Either way d18O might be a valid proxy for global temperature, and in fact the former would probably be more reliable. Yet Thompson and others seem to think it indicates the local temperature at which the snow formed.

    • Paul Dennis
      Posted Dec 8, 2008 at 2:05 PM | Permalink | Reply

      Re: Hu McCulloch (#82),
      Hu you are right that the snow develops directly from the vapour and there is an isotope fractionation between the vapour and the snow/ice. However, this fractionation would be nearly the same as if super cooled water formed rather than snow. Hence little fractionation between liquid and solid water.

      In polar ice the isotope composition is closely related to the local temperature and not the temperature in the source region. See my comment and explanation above in post 20. Note that it is not the effect of local temperature on the fractionation between vapour and liquid that is the primary control. Rather one has to consider the history of an air mass and its vapour from source to precipitation site.

      In low latitude, high altitude sites such as Dasuopu the isotope composition is an amount effect related to monsoon rain patterns. It is not easy to interpret these records in terms of temperature. Whilst an inverse relationship between temeprature and precipitation isotope composition might be observed on a seasonal basis, e.g. winter and spring rain might be enriched in 18O and summer depleted in 18O, one can’t translate this to long term records.

    • Kenneth Fritsch
      Posted Dec 8, 2008 at 3:36 PM | Permalink | Reply

      Re: Hu McCulloch (#82),

      So, if the O18 in H2O and CO3 in fact mixes well in ground water, and this is the primary source of the O18 variation in speleothems

      Hu, best to call that CO3 either HCO3- or CO3–. I do not believe that CO3 would be a very stable compound. Seriously, as a practicing inorganic chemist from many, many years ago I have enjoyed this discussion. Now that I look at it, I guess O18 would not be a very stable compound either.

  56. Craig Loehle
    Posted Dec 8, 2008 at 2:54 PM | Permalink | Reply

    Thanks to all for clarification. What I take home from this is that the isotopic composition of specleothem d18O can change based on the 1) source region temperature (fractionation from ocean water) 2) the amount that has rained out before reaching the site of interest and 3) events happening in the cave (at least). Thus to make a good spel-ometer, one must assume that monsoons have not strengthened/weakened over 1000 years, or winds changed direction, etc. It seems to me these are quite restrictive assumptions.

  57. Steve McIntyre
    Posted Dec 8, 2008 at 3:25 PM | Permalink | Reply

    Mann et al 2008:

    One advantage of the non-tree-ring proxy series used is that, in most cases, there is little reason to believe a priori that there are any problems with the series that are likely to eliminate the reliability of multicentury to millennial time scale information.

    • Paul Dennis
      Posted Dec 8, 2008 at 3:39 PM | Permalink | Reply

      Re: Steve McIntyre (#85),
      That comment comes from someone who has very little understanding of speleothems!

    • Jud Partin
      Posted Dec 8, 2008 at 4:13 PM | Permalink | Reply

      Re: Steve McIntyre (#85), I agree with Mann’s statement.

      Sediment cores (marine and lacustrine), speleothems, ice cores, corals etc. reliably record multicentury to millennial-scale information. Unlike a tree-ring where the tree has a lifetime and experiences vital effects that may alter the low frequency variability.

      What is difficult is the interpretation or transfer function for the speleothem record to a climate variable. That is an area of active research.

      • Paul Dennis
        Posted Dec 8, 2008 at 4:35 PM | Permalink | Reply

        Re: Jud Partin (#90),
        I think this is where Jud and I are not in complete agreement. Of course trees experience vital effects associated with temperature, response to water availability, age etc.

        However I’m not quite so sanguine that the problems associated with the main terrestrial archives have been and can be quite so easily resolved. They do offer the potential for decadal, centennial and millenial scale records but our understanding of the controlling processes is still rudimentary in some cases. This is particularly so when considering the temporal stability of transfer functions.

        I do applaud all those working in this area though. It is not easy and progress is being made.

  58. Steve McIntyre
    Posted Dec 8, 2008 at 3:45 PM | Permalink | Reply

    And who used the Tiljander sediments upside down :)

  59. Posted Dec 8, 2008 at 3:49 PM | Permalink | Reply

    Re Paul Dennis #83,
    Thanks for the clarification. I guess I was confused by the word “freeze”, which I thought meant transition to the solid state. On reading up, I see that it just means transition from liquid to solid. In fact, the direct transition from gas to solid is instead called “deposition”, the opposite of “sublimation”.

    Thus snow in fact never freezes, it just deposits (or deposes??). And although freezing per se doesn’t cause much O18 fractionation to speak of, deposition does, as does condensation per se.

    Your comment #20 above is also very helpful. I do have one question, though, about your statement

    In speleothems the situation is more complicated because we are not measuring the water isotope composition, but the isotope composition of the calcite or aragonite. Here we need to take account of several effects: i) the local recharge water composition. For temperate and high latitude sites this will be controlled by very similar processes to that for polar ice core.; ii) The local cave temperature and it’s effect on the 18O fractionation between calcite and water, and iii) the degree of isotopic equilibrium. Not with standing these observations I would expect to find that more depleted 18O composition of speleothems would indicate cooler climatic conditions.

    If the 18O in the water and carbonate/bicarb ions are already equal, or at least in equilibrium, why should the temperature of the cave itself matter?

    • Paul Dennis
      Posted Dec 8, 2008 at 4:22 PM | Permalink | Reply

      Re: Hu McCulloch (#89),
      Hu the 18O concentration (measured as the 18O/16O ratio)in the water and carbonate/bicarbonate ions is not equal. There is an equilibrium isotope partitioning such that at 25 degrees C:

      alpha = ((18O/16O)carbonate)/(18O/16O)water) = 1.028 and varies by about -0.0002 per degree C.

      These are small numbers and close to 1, but we can measure the ratio to better than +/-0.00005.

      This is the basis of the carbonate-water isotope thermometer. If we measure the speleothem carbonate and it’s parent water isotope composition we can measure the temperature of growth.

      Of course measuring the parent water composition is not easy (fluid inclusions) and there are no a-priori ways of estimating it.

  60. bernie
    Posted Dec 8, 2008 at 4:13 PM | Permalink | Reply

    #87
    Succinct and priceless!

  61. Steve McIntyre
    Posted Dec 8, 2008 at 4:51 PM | Permalink | Reply

    #90. Jud, I take your point. In other contexts, I’ve commented favorably on high-resolution ocean sediments and have been criticized by dendros for being too soft on non-dendros. The use of inorganic proxies has a lot of attraction. The dendros have an advantage on dating tho. The problem comes in the jump to Mannian multiproxy where weird multivariate methods are applied to data without worrying about consistency. I do realize that you are not opining on sausage manufacture, but that’s our starting point.

    For example, I think that people working in (say) monsoon-affected areas need to develop a regional O18 reconstruction model that reconciles the various proxies before throwing them all into a Mannomatic sausage-machine for global temperature. Similarly with dC13 – if that’s usably coherent, which seems pretty dubious.

  62. Posted Dec 8, 2008 at 4:54 PM | Permalink | Reply

    Re Paul Dennis, #92,

    There is an equilibrium isotope partitioning such that at 25 degrees C:

    alpha = ((18O/16O)carbonate)/(18O/16O)water) = 1.028 and varies by about -0.0002 per degree C.

    These are small numbers and close to 1, but we can measure the ratio to better than +/-0.00005.

    This is the basis of the carbonate-water isotope thermometer.

    But this means that holding groundwater 18O constant, speleothem 18O will be negatively correlated with cave and therefore average surface temperature. This works opposite the potential non-low-latitude precipitation effect on groundwater 18O.

    So in the end is the net effect ambiguous in theory and just an empirical matter that could go either way? Over in the 11/30 “Gavin Schmidt …” thread, Ken is finding slopes on the same order of magnitude for Wanxiang, but positive. (.0002 parts/dC = 0.2 parts per 1000 per dC.)

  63. Paul Dennis
    Posted Dec 8, 2008 at 5:20 PM | Permalink | Reply

    But this means that holding groundwater 18O constant, speleothem 18O will be negatively correlated with cave and therefore average surface temperature. This works opposite the potential non-low-latitude precipitation effect on groundwater 18O.

    Hu you are absolutely correct and the net effect can be ambiguous. In high latitude sites the temperature effect on rainfall and groundwater isotope composition is generally on the order of +0.4 to +0.7 parts per thousand per degree C. This is why we expect to see a net positive relationship in speleothems associated with temperate latitude caves. However, net negative relationships have been observed c.f. Mo-i-rana in Norway and some Alpine examples. This is unexpected and would seem to point towards some unexplained hydrology.

    It’s precisely this ambiguity, or under defined nature of the system that has led me down the route of trying to measure both fluid inclusion (parent water) and speleothem compositions, and now 18O-13C ordering (clumped isotopes) to determine cave and surface temperatures. Effectively I’m decoupling the temperature estimate from the water composition and potentially able to see how both T and precipitation isotope composition varies through time and space.

    Kenneth’s results sound very interesting and I want to go back and consider them in some detail. However, I caution against interpreting the correlation between temperature at Wudu and isotope composition mechanistically. I understand that the Science paper interprets this as a change from natural forcing of the monsoon system to an anthropogenic control. However, it is also possible that what we are seeing is a change in the source of the mositure or changing conditions in the source region.

    Below is the abstract from: Liu et al., 2008, Asian summer monsoon precipitation recorded by
    stalagmite oxygen isotopic composition in the western Loess Plateau during AD1875―2003 and its linkage with ocean-atmosphere system, Chinese Science Bulletin | July 2008 | vol. 53 | no. 13 | 2041-2049. I’m amazed that this paper on the same speleothem as the Science article and published just a few months before is not cited in the science article. In this paper there is a strong correpsondence between the rainfall amount at Wudu and isotope composition right up to the present day!

    ABSTRACT

    Based on 5 high-precision 230Th dates and 103 stable oxygen isotope ratios (δ18O) obtained from the top
    16 mm of a stalagmite collected from Wanxiang Cave, Wudu, Gansu, variation of monsoonal precipitation
    in the modern Asian Monsoon (AM) marginal zone over the past 100 years was reconstructed.
    Comparison of the speleothem δ18O record with instrumental precipitation data at Wudu in the past 50
    years indicates a high parallelism between the two curves, suggesting that the speleothem δ18O is a
    good proxy for the AM strength and associated precipitation, controlled by “amount effect” of the precipitation.
    Variation of the monsoonal precipitation during the past 100 years can be divided into three
    stages, increasing from AD 1875 to 1900, then decreasing from AD 1901 to 1946, and increasing again
    thereafter. This variation is quite similar to that of the Drought/Flooding index archived from Chinese
    historical documents. This speleothem-derived AM record shows a close association with the Pacific
    Decadal Oscillation (PDO) between AD 1875 and 1977, with higher monsoonal precipitation corresponding
    to cold PDO phase and vice versa at decadal timescale. The monsoonal precipitation variation
    is out of phase with the PDO after AD 1977, probably resulting from the decadal climate jump in the
    north Pacific occurring at around AD 1976/77. These results demonstrate a strong linkage between the
    AM and associated precipitation and the Pacific Ocean via ocean/atmosphere interaction. This relationship
    will aid to forecast future hydrological cycle for the AM monsoon region, and to improve
    forecasting potential of climatic model with observation data from cave.

  64. Posted Dec 9, 2008 at 1:41 PM | Permalink | Reply

    Paul Dennis writes in #96,

    In high latitude sites the temperature effect on rainfall and groundwater isotope composition is generally on the order of +0.4 to +0.7 parts per thousand per degree C. This is why we expect to see a net positive relationship in speleothems associated with temperate latitude caves. However, net negative relationships have been observed c.f. Mo-i-rana in Norway and some Alpine examples. This is unexpected and would seem to point towards some unexplained hydrology.

    Lonnie Thompson’s 6 ice cores from his CC03 paper show a much stronger effect of temperature on d18O than this, in the 4 cases when the correlation is significant (ppk = parts per 1000, aka o/oo):

    Himalayan cores:
    Guliya (35.30 dN) 2.73 ppk/dC, p = .0348
    Dasuopu (28.38 dN) 3.32 ppk/dC, p = .0056

    Andean cores:
    Quelccaya Summit (13.93 dS) 2.33 ppk/dC, p = .0364
    Huascaran (9.12 dS) 1.91 ppk/dC, p = .0794 (only weakly significant)

    These regressions are against HadCRU3v GL, constrained by Thompson’s archived data to be decadal. All DWs are greater than 2, so positive serial correlation is not an issue at this decadal frequency. His other two cores, Dunde (38 dN) and Sajama (18.10 dS), both have positive but insignificant coefficients less than unity.

    Since the O18 in H2O and dissolved HCO3- mingles (per #77, 78), wouldn’t groundwater that has been sitting in pores in limestone bedrock for days, months, or even years have O18 that reflects that in the limestone rather than that in the source rainwater or snowmelt? Or at least be an attenuated version of the original? Could this the source of the discrepancy?

    • Paul Dennis
      Posted Dec 10, 2008 at 2:09 AM | Permalink | Reply

      Re: Hu McCulloch (#98),
      Hu, an interesting observation wrt to Lonnie Thompson’s ice cores. The temperature effect on the isotope composition is extreme. It’s not possible to produce such effects from a simple rayleigh distillation model of an adiabatically cooling atmosphere. I would think that there is a possibility that there is a weakening amount (monsoon effect) or a change in source region in some of these cores. It may be that the amount effect is positively correlated with some climate or temperature change. However it is not the temperature change that is directly controlling the isotope composition of the precipitation.

      Now for your question about the 18O composition of water and speleothems. The answer is the system is buffered by the water isotope composition. The amount of bicarbonate in solution is small compared to the amount of water. In most of these systems the water rock ratio is also high. I’m not aware of any carbonate aquifer where the rock buffers, or exerts any influence on the isotopic composition of the dissolved bicarbonate.

      There are systems in which the oxygen isotope composition of the groudwater/porewater is heavily influenced by the rock isotope composition. However, these are in geothermal areas. Here the high temperatures promote water-rock interactions and the water composition changes to be in equilibrium with the aquifer at higher temeprature. There are many examples of geothermal areas where this is seen. The most significant one is the system of mid-ocean ridges where hydrothermal reaction between sea water and hot rocks exerts a significant control ohn the oceans isotopic composition.

      Of course the carbon system is another story entirely. Here the only two sources of carbon are either atmospheric, organic, or the rock derived carbon.

  65. Posted Dec 10, 2008 at 9:00 AM | Permalink | Reply

    Re Paul Dennis, #99
    Thanks, as always, Paul, for the valuable info! You write,

    Hu, an interesting observation wrt to Lonnie Thompson’s ice cores. The temperature effect on the isotope composition is extreme. It’s not possible to produce such effects from a simple rayleigh distillation model of an adiabatically cooling atmosphere. I would think that there is a possibility that there is a weakening amount (monsoon effect) or a change in source region in some of these cores. It may be that the amount effect is positively correlated with some climate or temperature change. However it is not the temperature change that is directly controlling the isotope composition of the precipitation.

    Very interesting! Some of his cores have well-defined annual layers that might permit the rate of accumulation to be measured and controlled for. But we still wouldn’t know what the seasonality of the precipitation was.

    Now for your question about the 18O composition of water and speleothems. The answer is the system is buffered by the water isotope composition. The amount of bicarbonate in solution is small compared to the amount of water. In most of these systems the water rock ratio is also high. I’m not aware of any carbonate aquifer where the rock buffers, or exerts any influence on the isotopic composition of the dissolved bicarbonate.

    Its true that the amount of bicarb in solution at any moment is trivial in comparison to the amount of water, but the amount of water in limestone bedrock is also trivial in comparison to the amount of limestone. The water must have lingered in the limestone long enough to become saturated, otherwise it would be dissolving the speleothems instead of building them. But even when it’s saturated, it continues to exchange carbs and therefore Os with the limestone at the same rate as before. Isn’t it just a matter of local hydrology how long this continues and how complete it becomes?

    There are systems in which the oxygen isotope composition of the groudwater/porewater is heavily influenced by the rock isotope composition. However, these are in geothermal areas.

    High temperatures would accelerate the exchange, but how do we know this groundwater hasn’t been sitting around for years or even centuries?

    • Paul Dennis
      Posted Dec 10, 2008 at 9:53 AM | Permalink | Reply

      Re: Hu McCulloch (#100),
      Hu as always very interesting questions and I’ll do my best to answer them starting from the question about groundwater residence times.

      There are a number of ways we can estimate the age of recharge of groundwater. Some of these are radiometric dating techniques. For young waters, say <100 years, we can use tritium. There is both a large bomb induced spike that is apparent in some groundwaters. This is as a result of atmospheric thermonuclear bomb testing in the late 1950’s and 1960’s. Whilst there has been a degree of dispersion in the spike, many aquifer systems can be modelled by piston flow and the spike still identified. This gives us a measure of flow velocity and recharge rates to an aquifer.

      We can extend the tritium method using T-3He techniques. T decays to 3He and measuring both in groundwater allows explicit solution of the radiometric decay equation.

      For older waters we can use 14C in the dissolved inorganic carbon load. Large corrections have to be made for dead carbon from the source aquifer but none the less reasonable dates can be determined.

      Okay now for your first question. Well it is a common misunderstanding that water:rock ratios are low. Yes the pore volume to rock ratio is low, but the factor we need to consider is the volume flux of water through any unit volume of the aquifer. This can be very large in some systems. I returned from an active karst region in Ireland in September with a group of undergraduates. Here the recharge rate is about 1 m per annum. Most of this flow is focussed in a joint and fracture system with low storativity in the aquifer. This means that 1 cubic metre of water passes through every cubic metre of rock a year. In 100 years the water rock ratio is 100:1, over 1000 years it is 1000:1. It has a pH of about 4 on entering the soil zone, rapidly dissolves CO2 and then carbonate reaching saturation at elevated CO2 pressures very quickly. From then on the only possible mechanism for further exchange with the aquifer is dissolution and reprecipitation, for which there has to be a driving force (pressure solution, dissolution of high Mg calcite and aragonite and precipitation of low Mg calcite etc). or solid state diffusion in the carbonate lattice. At low temperatures solid state diffusion is out which leaves dissolution and reprecipitation.

      With the large water:rock ratios that we see in these aquifers if dissolution and reprecipitation were ongoing we would see it in the oxygen isotope signature of the host carbonate which would tend towards a meteoric/freshwater value. We don’t see this. Many aquifers, despite 10’s-100’s of thousands of years and in some cases millions of years of being active still retain their original marine isotope signature in the carbonate.

  66. Steve McIntyre
    Posted Dec 10, 2008 at 9:46 AM | Permalink | Reply

    Holdsworth (JGR 2008) has an interpretation of ice core dO18 that is very much worth looking into.

    He did ice cores on Mt Logan (recently updated by Fisher and others) which have more negative dO!8 in the 20th century than 19th century – contrary to the Lonnie Thompson model. They attribute this to changes in source region – an increasingly prevalent theme it seems for everyone except Lonnie Thompson.

    I think that it would be prudent to keep an asterisk on the dating of Dasuopu. Some high-accumulation glaciers turn over their inventory very rapidly and have negligible information past a few centuries. Dasuopu is a high-accumulation glacier, which appears to me to have a surprisingly long time series relative to comparable glaciers (but this is a long exegesis and I’ve just dipped my toe in the data.) This would be an issue worth visiting in Thompson’s glaciers. Unfortunately Thompson has refused to provide a proper data archive so it’s a bit of a dead end.

    • Paul Dennis
      Posted Dec 10, 2008 at 11:58 AM | Permalink | Reply

      Re: Steve McIntyre (#101),
      Steve,

      I’ve not read the Holdsworth paper yet. However, from recollection the Mt Logan record has both d18O and d2H at high resolution. The advanatge of this is that it is that measuring both oxygen and hydrogen isotopes allows one to calculate the deuterium excess value, d, as:

      d = delta(2H) – (8 * delta(18O))

      The deuterium excess value is sensitive to processes that occur in the source region and should therefore be strongly indicative of changes in either the location of the source region, or changes in processes in the source region such as relative humidity. Again from memory the deuterium excess at Mt Logan undergoes a rapid shift during the 19th century that would indicate significant changes in the source region for the water vapour.

      It is to be regretted that we don’t have good deuterium data for many more ice cores and notably those at low latitude and high altitude.

      Out of interest I’m just completing the oxygen isotope measurements on an ice core from the Antarctic Peninsula. These have a very high temporal resolution, better than monthly, for the past 150 years (2500+ samples). I’ve completed some preliminary deuterium measurements and have many more to do. Early indications are that the deuterium excess has remained constant.

  67. Steve McIntyre
    Posted Dec 10, 2008 at 10:06 AM | Permalink | Reply

    As an aside, I was watching an old episode of CSI last night, where the investigators were trying to identify a body about which little was known. One of the CSIs decided to do an oxygen isotope analysis on the corpse, declaring that he could determine where the person was from by the oxygen isotopes of the corpse – a result that they proposed to publish in the Journal of Forensics. A short time later, the CSI reported that, from the oxygen isotope analysis, he could tell that the person had been in southeast Asia.

    As I mentioned, it was an old re-run. I believe that the CSI on the show re-trained and now publishes multiproxy temperature reconstructions.

    • Sam Urbinto
      Posted Dec 10, 2008 at 6:41 PM | Permalink | Reply

      Re: Steve McIntyre (#103),

      I think I saw that episode also. It was named “Night of the teleconnections” wasn’t it?

  68. bernie
    Posted Dec 10, 2008 at 12:21 PM | Permalink | Reply

    Paul, is there a proven theory about the link between deuterium excess and temperature or is this relationship still being explored?

    • Paul Dennis
      Posted Dec 10, 2008 at 3:21 PM | Permalink | Reply

      Re: bernie (#105),
      Deuterium excess is much more sensitive to parameters such as the relative humidity of the source region rather than the temperature. To understand the origin of the deuterium excess one needs to consider kinetic effects on isotopic fractionation during evaporation.

Post a Comment

Required fields are marked *

*
*

Follow

Get every new post delivered to your Inbox.

Join 3,132 other followers

%d bloggers like this: