## New Thompson Article at PNAS

There’s a new Lonnie Thompson article online at PNAS (thanks to Steve Bloom for reference). It has some "supporting data" – "supporting data to Thompson means only digital versions of the graphics, rather than detailed archives such as Majewski provided for the Everest ice core. Just to annoy anyone who was actually interested in the data and create a obstacle, Thompson archived this limited data as pdf rather than ascii. You can convert it to ascii, but it took me a couple of hours by the time that I’d sorted out different number of entries on each line.

Thompson mentions the 5000-year old plants. I’ll discuss this on another occasion. Here I only note that Thompson does not provide any detailed information on the exact location of the plant deposit or whether this was the only organics recovered.

Thompson takes note of the increasing doubt about whether $dO_{18}$ is a temperature proxy, and, in his closing, argues that anomalous $dO_{18}$ is evidence of anomalous climate, even if it’s not directly associated with temperature. However, for the most part, he directly equates $dO_{18}$ levels and temperature. He discusses both Tibetan and Andean results. Here I’m just going to talk briefly about Andean results.

The graph that carries forward to his Tropical Summary is his Andean "summary" shown below:

From Thompson et al Figure 5D

The blue curve is accumulation; the red curve is $dO_{18}$. (The only contributor to the blue curve is Quelccaya as flows at the other sites appear to have prevented annual dating. I haven’t checked exactly what’s going on with dating at these other sites.

From the above graphic, it appears that the Andean $dO_{18}$ levels are closing at a fairly loud two-sigma departure. Seemingly powerful evidence of anomalous climate. But this is the Hockey Team, so it never does any harm to check the individual series.

The underlying 5-year averages for Thompson’s three Andean sites are shown in Thompson’s Figure 5A, shown below. Now, I don’t know about you, I have trouble discerning a strong and robust common warming trend in the these three series. Squinting at the Huascaran series in particular, I’d say that, if $dO_{18}$ is supposed to be a thermometer, it doesn’t show anything anomalous about the last half of the 20th century. Same with the Sajama $dO_{18}$ series. No values are shown for Huascaran in the 1990s (I’m not sure why) and none for Sajama in the last part of the 1990s. Now Quelccaya was updated in 2003. It seems to be the active ingredient in the overall average, so I guess it must show a big 20th century trend. In this case, there have been 3 holes, 2 drilled in the mid 1980s (with archived results) and the 2003 update, so it must tell quite a story.

From Thompson et al Figure 5A

Thompson has archived annual $dO_{18}$ from the 2003 hole from 1604 (but no earlier) to 2002 in connection with the PNAS publication. I’ve compared this data to previous results for two 1984 holes in the graph below, showing 20th century results. A small preliminary point before I discuss the obvious: the very elevated values: although the 2003 hole tracks the earlier holes fairly closely, for some reason the very elevated $dO_{18}$ values associated with the 1983 El Nino are not shown in the 2003 results. Thompson does not discuss this. Possibly the values of adjacent years have blended together because of melting or diffusion or something, but the 2003 values are much smoother through the 1983 El Nino than the earlier results.

Quelccaya $dO_{18}$ for 20th century. Compiled from previously archived data and from NAS archive. Black – 2003 core; blue, red 1984 cores.

Maybe it’s just me, but I have trouble discerning a strong 20th century trend in this data. If I didn’t know that this data seet was suppsoed to end two sigmas on the hot side, I’d also be inclined to say that the last few years of the record end up pretty close to the average value. Yes, there was a strong El Nino reflected in elevated 1997 $dO_{18}$ values, but it looks like there was strong El Nino in 1941 or so as well.

Now the last few years of this record do not make a 5-year average and I’m sure that was the only reason why Thompson didn’t use them in his summary graphic. But don’t you wonder just a little a bit, if the up-to-date $dO_{18}$ values had been off the chart, that Thompson might have figured out a way to get them into the graphic.

But there’s interesting data news: here we have an up-to-date proxy record, right up to 2002. On the multiproxy theory of the world, this proxy record should have been off the chart because of late 20th century warming. I have trouble discerning an increase above levels in the 1930s (as with Greenland). But maybe that’s just me.

1. John A
Posted Jul 6, 2006 at 2:37 PM | Permalink

Here is the ENSO record for the last 127 years. Can you see a strong El Nino in the 1940s? I can.

2. jae
Posted Jul 6, 2006 at 2:40 PM | Permalink

I have trouble discerning an increase above levels in the 1930s (as with Greenland). But maybe that’s just me.

It’s not just you.

3. TCO
Posted Jul 6, 2006 at 2:42 PM | Permalink

No. Given that I don’t know if strong equates to blue or red.

4. Peter Hearnden
Posted Jul 6, 2006 at 2:51 PM | Permalink

Just so I know, is calling people a ‘prick’ ok?

5. Steve McIntyre
Posted Jul 6, 2006 at 2:56 PM | Permalink

#4. No, it’s not. I’ll change it. But I was really irritated at having to spend so much time converting the pdf into data, when there was no conceivable reason to file it in pdf except to be annoying.

6. Peter Hearnden
Posted Jul 6, 2006 at 3:03 PM | Permalink

Now you’re still ascribing an ulteriour motive to Thompson, that he set out to annoy. You don’t know he did, so you should withdraw that too.

7. jae
Posted Jul 6, 2006 at 3:11 PM | Permalink

Peter H. You are the prick here.

8. J. Sperry
Posted Jul 6, 2006 at 3:18 PM | Permalink

Re #3 (TCO):
The way I interpret it, “strong” is either blue or red. That is, large red is strong El Nino, large blue is strong La Nina (note the font color in the title). Not sure what the units are.

9. jae
Posted Jul 6, 2006 at 3:20 PM | Permalink

8. What’s with the blue below the “zero” line in the late 40’s?

10. John A
Posted Jul 6, 2006 at 3:26 PM | Permalink

Re #8

The units are millibars since the ENSO record is the difference in barometric pressure between Darwin and Tahiti. It’s averaged out to a full year but you can see the effect.

11. Ken Robinson
Posted Jul 6, 2006 at 3:28 PM | Permalink

Re: 4, 5, 6, 7

To quote from the blog rules:

1. Refrain from personal abuse and swearing.
2. Never attribute ulterior motives to another participant.
3. Be patient with people who know less science or maths than you do yourself.

Steve’s the boss of his own blog, but I’m sure he’ll agree that he should hold himself to the same standards he expects of others. Technically, Thompson’s not a participant and so Rule 2 doesn’t necessarily apply. That said, comments regarding motivations carry much less weight than the rigorous analysis of facts in evidence.

It would also be appropriate to occasionally remind posters that intelligent people can disagree and still be civilized about it.

Regards;

12. jae
Posted Jul 6, 2006 at 3:31 PM | Permalink

Sorry, Steve and Ken. I agree with the rules, but the temptations are sometimes overwhelming.

13. Peter Hearnden
Posted Jul 6, 2006 at 3:39 PM | Permalink

Re #7: noted.

14. Dave Dardinger
Posted Jul 6, 2006 at 4:09 PM | Permalink

Everyone seems to have forgotten that Peter is a classic troll and thus to be ignored. He’s not here to advance the discussion but to disrupt it. Placating him in any way just encourages him.

15. JSP
Posted Jul 6, 2006 at 5:20 PM | Permalink

Steve,

I know you are frustrated by the archived data, but never attribute to malice that which can be explained adequately by stupidity.

In this case, T_______ may have just not known better.

I’ll admit that storing data in a .PDF format is strange, but as someone who began analyzing data on a counter-sorter using punched cards in 1969, I’ve seen stranger. For example, a Ph.D. in Urban Planning from the University of Michigan had his research data punched up deliniated by commas instead of fixed fields. His proofing problems were massive to say the least.

JSP

16. TCO
Posted Jul 6, 2006 at 5:22 PM | Permalink

Filing stuff in a pdf form is incredibly unhelpful. He knew what he was doing. And Pete knows that he is too butt-stupid to engage on the content.

17. David Smith
Posted Jul 6, 2006 at 5:41 PM | Permalink

I know that Steve has often struggled to get the data used in papers. I am beginning to wonder what other publicly-funded data (ice cores from other spots, tree rings, etc) never make it to publication whatsoever, due to their failure to support the author’s view.

Interesting link for BINCs (Believers in Non-Catastrophic warming), like me

18. Steve McIntyre
Posted Jul 6, 2006 at 6:00 PM | Permalink

#17. My #1 candidate right now is Hughes’ update of the bristlecoens at Sheep Mountain in 2002. It is inconceivable to me that he wouldn’t have published them if they were on his side. I’m really attuned to this issue from all my experience in mineral experience. When someone has bad drill results, they want to hold off releasing the results just in case the next hole brings the program offside. Wise investors can smell delays and they are NEVER a sign of good drill holes. If a promoer has a good drill hole, he’ll announce it as soon as he can.

19. welikerocks
Posted Jul 6, 2006 at 7:13 PM | Permalink

#19

“Hence the argument constantly based on WHO is speaking, rather than what is being said, within the “greenhouse platform’s theatre”

That was the best.
LOL :)

20. Michael Jankowski
Posted Jul 7, 2006 at 5:13 AM | Permalink

I tried to give him the benefit of the doubt and say the PNAS archived it that way, but I see other articles where the supplemntary info is in text form.

21. beng
Posted Jul 7, 2006 at 7:51 AM | Permalink

RE 18:

I heard a rumor about secret shipments of fertilizer, water, & irrigation piping made to the Sheep Mtn area recently. Hughes will have new, updated cores in just a few more yrs…

22. Ken Fritsch
Posted Jul 7, 2006 at 11:23 AM | Permalink

re Steve M’s comments on Thompson article at PNAS:

On the multiproxy theory of the world, this proxy record should have been off the chart because of late 20th century warming. I have trouble discerning an increase above levels in the 1930s (as with Greenland). But maybe that’s just me.

This layperson’s view of the Thompson article takes away two major points that I believe will draw a goodly amount of media attention. Firstly, the composite oxygen isotope 18 ice cores graphs match well, on first look, the HS and over more than a millennium time span. Although this study is limited to tropical glaciers that limitation could be “overlooked” in the haste to get the millennial HS back in play and spread the word.

My second point is to note the throw away line for the media from this article that I predict will be the one I have listed below and concerns recent accelerating retreat of glaciers that could be connected to a tipping point for AGW.

The recent, rapid, and accelerating retreat of glaciers on a near-global scale suggests that the current increase in the Earth’s globally averaged temperature (Fig. 6D; refs. 49 and 50) may now have prematurely interrupted the natural progression of cooling in the late Holocene. These observations suggest that within a century human activities may have nudged global-scale climate conditions closer to those that prevailed before 5,000 yr ago, during the early Holocene. If this is the case, then Earth’s currently retreating glaciers may signal that the climate system has exceeded a critical threshold and that most low-latitude, high-altitude glaciers are likely to disappear in the near future.

The message of this article and those of the hockey stick are quite clear in reference to an “anomalous” warming in recent times being AGW that may well be showing a tipping point that forewarns, if not predicts, accelerated warming into the near future. (A prediction of a tipping point with accelerated temperature increases would, of course, be more readily tested with out-of-sample results than a gradual and constant rate of increase starting in the 1750 to early 1900s period could.)

A third point to which I doubt the lay public will be exposed is that looking at the individual oxygen 18 isotope graphs in the Thompson report does not support the picture of recent accelerating temperature increases and might even help explain why the oxygen 18 isotope proxy could miss some of the past variations in temperature/precipitation/climate. Thompson and his colleagues seem to want to make a case for the oxygen 18 isotope being a proxy restricted to temperature variations, but are willing to concede some other climate variables could complicate this use just so long as it can be said that it shows an anomalous climate in recent times.

On the surface and the levels to which one would anticipate the media reviewers will dig, this article could be viewed as a timely effort to publically, if not scientifically, repair the HS.

23. jae
Posted Jul 7, 2006 at 11:38 AM | Permalink

One problem with the “tipping point” concept is that temperatures have not gone up much since 1998. It seems to me that if we are nearing some “tipping point,” there should be some sort of exponential increase.

24. Andre
Posted Jul 7, 2006 at 12:36 PM | Permalink

Re 22,

Thompson and his colleagues seem to want to make a case for the oxygen 18 isotope being a proxy restricted to temperature variations, but are willing to concede some other climate variables could complicate this use just so long as it can be said that it shows an anomalous climate in recent times.

That is because all of them are very aware of the anomalous isotope behavior in the tropics, as I showed in the other thread, comment #9. So the best he could say is “climate change”, a warming statement would have been too dangerous and easily refutable. Apparantly it was the idea that the alarmist press would have filled in the blanks here with the no-brainer obvious. A well known and successful strategy.

25. Andre
Posted Jul 7, 2006 at 12:56 PM | Permalink

Sorry erratum, re #24 comment #9 should read #8

26. Paul Dennis
Posted Jul 24, 2006 at 6:31 PM | Permalink

I’ve just picked up on this thread relating to tropical ice sheets in the Andes and the Tibetan Plateau. In combining the data the authors use a parameter known as the ‘z-score’. Can anyone enlighten me as to what the ‘z-score’ is. My first stab is to say that it is a normalised variance of some sort, perhaps (value – mean)/std. dev. ?

I must say it’s hard to see how the composite figures relate to the individual data sets.

27. John Baltutis
Posted Jul 24, 2006 at 7:18 PM | Permalink

A google search for “z-score” brings up many hits, beginining with http://www.animatedsoftware.com/statglos/sgzscore.htm

28. Steve McIntyre
Posted Jul 24, 2006 at 8:15 PM | Permalink

#26. yes. The individual series are all over the place. The 20th century uptick in the average is mainly due to Dasuopu, a precipitaiton proxy.

I’ve seen at least 4 different and inconsistent versions of the Dunde series in grey literature. The first archive of any sort of the Dunde data came in 2004 as a result of my complaints to Climatic Change and that was only dO18 on a decadal basis. One needs to see a complete archive of all samples with all chemistry.

Lonnie Thompson is one of the worst archiving offenders in paleoclimate, and that’s a real beauty contest.

29. Paul Dennis
Posted Jul 25, 2006 at 1:14 AM | Permalink

#27, #28 John, thanks for the URL. It seems my first stab at what the z-score is was correct.

Steve, I agree with you about the differing versions of some of these series popping up in all sorts of different places.

I’ve only skimmed the text of this present paper. It was 1.30am when I staretd to read it! All sorts of concerns were triggered immediately and now I’ll spend more time going through it in detail.

30. Paul Dennis
Posted Jul 25, 2006 at 4:22 PM | Permalink

I’ve read the recent Thompson et al. paper (2006 PNAS, vol 103, 10536-10543). I’ve also downloaded the additional data, though as Steve points out, in PDF format it is going to be a time consuming chore to convert it to anything useable! I’ve managed to put the 5 year average data into an Excel spreadsheet.

However, my first comment concerns the evidence for unprecedented glacier retreat based on several radiocarbon dates from two different laboratories for a single rooted sub-glacial plant deposit. The important point to note is that a single plant deposit was found, from which several specimens were taken. Unfortunately the paper gives no details about the location of the find, the nature of the sub-glacial geology and topography so it is difficult to draw any further judgement. However, I’ve no doubt the dates are correct at 5,138 (+/- 45) yr BP. That this indicates that the current retreat of Quelccaya is unprecedented for the past 5 millenia is not based on any logic what so ever.

What the find shows is that at 5.1ky the glacier had retreated sufficiently for the plant to grow. It says nothing about subsequent retreats that may have exposed this plant, and also allowed others to grow. What is needed is a thorough survey for rooted ‘fossil’ plant material and a large number of radiocarbon dates on a large number of specimens collected from a large number of plant deposits. This will clearly identify if windows in the past several millenia occured when plants were growing and will provide a history of glacier retreat and advance. A single plant deposit, whilst tempting to speculate, only provides evidence for a retreat 5.1ky ago and says absolutely nothing about possible retreats in the intervening period.

The isotope data are interesting, but the graphs so compressed it is very difficult to say very much about them. Leaving aside questions of methodology with respect to data handling I would suggest that there are no distinct trends observable in the Andean glaciers. Huascaran shows a pronounced period of light precipitation prior to 1750, from which there was a sharp increase over a period of 20 or 30 years to compositions that have persisted to the present day with no apparent trend in composition. Sajama shows no apparent trend over the past 4 hundred years. Quelccaya shows a slight upward trend towards increasingle enriched delat 18-O compositions in the 19th and 20th centuries. I did a quick correlation between the z-scores for the 5 year average isotope composition for Quelccaya, Huscaran and Sajama, with correlation coefficients of +0.12 between Quelccaya and Huascaran and +0.02 between Quelccaya and Sajama. I guess all this confirms is that the Sajama data set looks random with no trends and the Huascaran dta set is random with two regions: one before 1700 and one after 1700. The only trending data is Quelccaya, and it is weak at that.

Given this observation I doubt any more sophisticated statistics or data analysis is really warranted.

31. Lee
Posted Jul 25, 2006 at 4:50 PM | Permalink

Paul, I dotn agree with your analyusis of the roted-plant observation. I do agree it woud lbe nice tto see mroe details about its loctins, etc, though.

IF it was uncovered in that interval, it could only have been breifly at best. Dead plant matter exposed to the weather simply doesnt have much of a chance of lasting long, even at high altitudes. It oxidizes, it becomes subject to attack by various organisms, it dessicates and is destroyed by wind, and hydrates and is destroyed by frost, and so on. While we cant definitively rule out a very brief interim exposure, I think we can rule out more than that.

I don’t consider this as strong evidence as, for example, the surface melt on quelcayya, but I dont think its all that weak, either. What will be interesting is to see what kind of similar finds are uncovered as the glacier retreats further – and we should be lookng.

32. Paul Dennis
Posted Jul 25, 2006 at 5:33 PM | Permalink

Lee these are good points, and may explain exactly why plant material is rare. At least I’m assuming it’s rare from the description in the paper. Plants have not been preserved due to glacial action and exposure during this, and perhaps, unknown other glacial retreats. However, there are plenty of examples of plant material surviving for long periods, especially if they become dessicated, and more especially if they become incorporated in the sub glacial sediment which is possible with an advancing and retreating system. It might be worthwhile looking for seeds, grains, and more woody stems etc which tend to be more resistant to oxidation. The point I am making is that the evidence as presented is not definitive. I agree with you completely we should be looking. It is a tease and the obvious conclusion to draw is that a full survey needs to be done. The hypothesis is that the glacial retreat is unprecedented. To test this hypotheses it is necessary to look for plant remains that might be younger than the 5.1 ka example. It’s going to be difficult for all the reasons you have stated but the importance of the problem demands that it is done.

I’m not up on the surface melt on Quelccaya. Thompson et al. don’t talk about surface melt in their paper, they talk about the surface being subject to warming conditions. I note that the annual structure is preserved which indicates that a seasonal delta 18-O cycle is still seen, even in the youngest layers. If melting is occuring then it is very short lived and doesn’t penetrate to any depth.

If melting has occured then it could lead to changes in the isotopic composition of the surface water. Evaporation will fractionate the isotopes and result in the residual water becoming enriched in oxygen 18. It’s likely that evaporation is not important, but I suspect that the area is one of low humidity which would enhance evaporative effects. It would be useful to have the deuterium data. Generally speaking an evaporative system has a very distictive trajectory in delta 2-H versus delta 18-O space, as opposed to un evaporated precipitation.

Is there a paper on this?

Cheers,

Paul

33. jae
Posted Jul 25, 2006 at 5:55 PM | Permalink

Lee, just how do you ignore the numerous studies that indicate that we can’t yet tell much from glaciers?

34. Lee
Posted Jul 25, 2006 at 6:07 PM | Permalink

Paul, to be sure I’m being clier – I’m usign ‘melt’ more loosely than I should. Teh report what effects of higher temps sufficient to cause loss of resolutin not previosly seen. I’ll try to dig out the story when I get some time – its just about time for dinner.

35. Lee
Posted Jul 25, 2006 at 6:08 PM | Permalink

jae, since I’m not talking (here) about mass balance analyses, none of those are relevant.

36. Steve Bloom
Posted Jul 25, 2006 at 8:45 PM | Permalink

Re 33/35: I suppose “irrelevant” is a polite way of referring to any given citation from the Idsos. jae might consider instead a resort to some actual science. It turns out that globally glaciers are on a pretty steady downtrend for the last 25 years. OTOH, who knows, any day now a farmhouse could be appearing out from under one of them.

37. welikerocks
Posted Jul 25, 2006 at 9:19 PM | Permalink

Each advance of ice is popularly known in the press as an “ice age” but it is important to note that these multiple events are just variations of the same glacial epoch.➠The retreat of ice during a glacial epoch is called an inter-glacial period and this is our PRESENT DAY CLIMATE system..

*The current Plio-Pleistocene Glacial Epoch had it’s beginning about 3.2 million years ago and is probably linked to the tectonic construction of the Isthmus of Panama which prevented the circulation of Atlantic and Pacific waters and ultimately triggered a slow sequence of events that eventually led to cooling of the atmosphere and the formation of new ice fields by about 2.5 million years ago.

*So far we have had around 15 to 20 individual major advances and subsequent retreats of the ice field in our current glacial epoch.➠The last major advance of glacial ice peaked about 18,000 years ago and since that time the ice has generally been retreating (albeit with some short term interruptions).

**It is worth remembering that our warm present day inter-glacial climate is the exception, not the rule during a glacial epoch.➠For as much as 90% of the last 2 million years the ice fields on earth have been more extensive than they are today.

**On the other hand, our the current glacial epoch and ice on earth and for the most part is also an abnormality.➠Our present-day Arctic Ocean is about 10-15°C cooler than it was at the time of the dinosaurs for almost all of the time from about 2 to at least 200 million years ago (Ma) the surface temperature exceeded that of today.

***Climatic can change more rapidly than previously thought. The Gulf Stream plays an important role carrying heat from the equator poleward. ➠When this current is disturbed, dramatic climatic changes can occur over a short period of time.
****Not to mention the Milankovitch cycles.
http://www.lakepowell.net/sciencecenter/paleoclimate.htm
And this graph to illustrate:
http://tinyurl.com/zm49w

38. mark
Posted Jul 25, 2006 at 9:55 PM | Permalink

Careful there welikerocks, you may confuse the issue with REAL science.

Mark

39. Paul Dennis
Posted Jul 26, 2006 at 6:06 AM | Permalink

Considerable time and effort is expended on sophisticated statistical analyses of proxy data sets. I’ve been looking at the Thompson et al. (2006 PNAS, vol 103, 10536-10543) Andean glacier oxygen isotope data and have commented on it above. My view was that there were no significant long term trends observeable in the data. Whilst driving into the lab I was thinking if I could back this up with some objective, and simple statistics. At the same time I was thinking about how we quality control isotope analyses in the lab. One of the procedures we use is the cumulative sum of differences (cusum) chart. This is an extraordinarily sensitive and effective way of identifying changes in the analytical means. Moreover the gradient of the cusum chart enables an estimate of the mean. It struck me that I could use a similar plot with the Huascaran, Sajama and Quelccaya 5 year data to see if I could identify any changes in mean compositions and long term trends.

For this exercise I took a mean of the data between 1600 and the present day for each series. Then at each time increment t (starting from 1600) the cumulative sum of deviations about the mean is given by S(t) = SUM(p

40. Paul Dennis
Posted Jul 26, 2006 at 6:07 AM | Permalink

Oops! I don’t know what happened to the rest of my comment? Any suggestions John A?

41. Posted Jul 26, 2006 at 6:14 AM | Permalink

Paul: My guess is you typed a less-than sign, which is also the beginning of an HTML tag. You need to use &lt; to create one of these: <

42. Paul Dennis
Posted Jul 26, 2006 at 6:42 AM | Permalink

Many thanks Nicholas. You’re absolutely right about the less than sign! I’ll try again.

43. Paul Dennis
Posted Jul 26, 2006 at 6:42 AM | Permalink

Considerable time and effort is expended on sophisticated statistical analyses of proxy data sets. I’ve been looking at the Thompson et al. (2006 PNAS, vol 103, 10536-10543) Andean glacier oxygen isotope data and have commented on it above. My view was that there were no significant long term trends observeable in the data. Whilst driving into the lab I was thinking if I could back this up with some objective, and simple statistics. At the same time I was thinking about how we quality control isotope analyses in the lab. One of the procedures we use is the cumulative sum of differences (cusum) chart. This is an extraordinarily sensitive and effective way of identifying changes in the analytical means. Moreover the gradient of the cusum chart enables an estimate of the mean. It struck me that I could use a similar plot with the Huascaran, Sajama and Quelccaya 5 year data to see if I could identify any changes in mean compositions and long term trends.

For this exercise I took a mean of the data between 1600 and the present day for each series. Then at each time increment t (starting from 1600) the cumulative sum of deviations about the mean is given by S(t) = SUM(p&lt=t)(X(p) – mean). Please excuse the clumsy notation!

If the data are stationary about a long term mean they will plot as straight lines on the cusum chart. If the trajectory is horizontal then the data oscillates about the long term (1600-present day) mean. A negative gradient indicates a lower mean, and a positive gradient a higher mean. A consistent trend, e.g. increasing delta 18-O will be seen as slope with an increasingly positive gradient.

If I can find out how to post the graphs on the site I will do so. Here is a summary of the results:

1) Sajama – The data plot along the horizontal indicating no trend in the oxygen isotope composition away from the long term mean.

2) Huascaran – The data plot on two linear segments. 1600 – 1750 The data plot with a negative gradient indicating a mean that is stationary, but below the 1600 – 2000 mean. Post 1750 all the data plot on a linear trajectory with a positive gradient. This indicates another period of stationary mean which is higher then the long term mean.

3) Quelccaya – 1540 – 1660 The cumulative sum of differences plot on a horizontal trajectory indicating a mean delta 18-O composition in accord with the long term mean. 1660 – 1860. Again another linear trajectory, with a negative slope indicating a stationary mean slightly below the long term average. Post 1860 – another linear trajectory with a positive gradient. i.e. a stationary mean slightly higher than the long term average.

This may seem like a trivial use of statistics but in reality provides a useful graphic tool when looking for subtle variations. The conclusion is that there are no discernible trends in the Andean data. Rather there are sharp transitions between temporal periods of stationary average isotope composition.

I believe, a trivial use of statistics would be to try and force a linear regression, higher polynomial or spline function on the data in an attempt to identify trends. Such an approach would indicate long term trends in both the Huascaran and Quelccaya data sets which are not there.

44. Posted Jul 26, 2006 at 8:34 AM | Permalink

Paul: Looks like you missed the semicolon at the end (note: “&lt;”), but no matter, I can read your comment, and what you say is interesting. I’d like to see the graphs. I note that in the preview, if you do &lt then it appears as a less-than, but when you hit “post” it does not work, you need the semicolon at the end.

I believe you can post images here but they have to be hosted somewhere. Can you render your graphs to .jpg or .gif files? If so you’re welcome to e-mail them to me (hb AT, ca DOT, x256 DOT, org) and I will attempt to post them here.

While I think less than (<) is the only symbol which you MUST escape when posting here, it’s probably worth escaping some others:

&amp; – ampersand (&)
&lt; – less than (<)
&gt; – greater than (>)
&ldquo; – left quote (“)
&rdquo; – right quote (”)
&mdash; – em dash (—)
&ndash; – en dash (–)

45. Steve McIntyre
Posted Jul 26, 2006 at 10:18 AM | Permalink

Paul – if you email the graphics and text, I’ll put this up as a post. It’s nice stuff. I’m a bit distracted right now for reasons that you can guess at, but I’ll get back to this.

46. Posted Jul 26, 2006 at 10:40 AM | Permalink

Well, I don’t know if Paul sent this to Steve, if he did I suppose you can delete this comment, but here’s the image:

47. beng
Posted Jul 26, 2006 at 11:29 AM | Permalink

RE 43: Paul Dennis

The d-18O issue in general brings up a question that’s at least been briefly discussed here before — the use of it in wood core-samples and how that might represent a temp proxy? It would obviously be a more complex relation than in ice-cores.

48. Henry
Posted Jul 26, 2006 at 5:45 PM | Permalink

A minor curiosity comparing the 1983 and 2003 Quelccaya cores (Figure 4/Supporting Data Set 1): there seems to be a reasonable match btween the two sets of data in the 18th and 19th centuries, but in the 20th century changes in the later core seems to lag the earlier ones by a year. In Steve’s third chart the black dots are slight to the right of the red and blue ones. A priori, I would have expected the reverse to be more likely: a missed year or an erroneous insertion might produce errors for distant time (assuming cores are dated from the top, i.e. “now”).

49. Paul Dennis
Posted Jul 27, 2006 at 3:07 AM | Permalink

#45 Steve, don’t worry I quite understand.

#46 Nicholas thankyou for posting the graph for me. The graph should be looked at with my comment #43.

#47 Beng, there are a few groups working on the isotopic composition of tree ring cellulose. Here they have the option of measuring delta 13-C, delta 18-O and delta 2-H. These kind of measurements have been going on for several decades now, including some work about 10 years ago on Bristlecone Pines. Sam Epstein, one of the great historic pioneers of isotope geochemistry, had done a great deal of work on tree ring isotope compositions. With respect to the carbon story and Bristlecones, all the trees they studied contain a high frequency signal which is controlled by precipitation and a long term trend controlled by the atmospheric CO2 level. This led Epstein and co-workers to suggest Bristlecones could be used as a proxy for past atmospheric CO2 levels. I don’t know wether or not this work has ever been followed up (See Feng and Epstein, Geochimica et Cosmochimica Acta, V 59, 2599-2608, 1995).

With respect to hydrogen and oxygen isotopes you might look at Yapp and Epstein, 1977, Earth and Planetary Science Letters, V34, 333-350, or for a summary Epstein and Krishnamurthy, 1990, Environmental information in the isotopic record in trees. Phil. Trans, Roy. Soc. Lond. VA330, 427-439.

I think the fact that there are so few published studies relying on tree ring isotopic proxies tells us the problems associated with interpreting the data. The physical and chemical processes are complicated and the signal controlled by several factors other than temperature. A bit like ring widths in general!

#48 Well spotted Henry. yes there are some anomalies that look like a time shift of 1 year between the 1983 and 2003 cores. I’m not sure how they counted the near surface layers. Presumeably these were in snow pack, or as Lee has mentioned there may have been some melting going on. This could cause problems in one of, or both cores that lead to erroneous dates. In the overall scheme of things I don’t think it matters. Certainly, after taking the 5 year averages there are only very small differences between the two data sets.

50. Paul Dennis
Posted Jul 28, 2006 at 3:34 AM | Permalink

Further to posting the graph of the cumulative sum of differences and using it to estimate mean isotope compositions we can describe the Andean data thus:

1) Sajama: The data show no overall trend with a mean composition of -17.4 per mille (SMOW)

2) Huascaran: There are 3 regions. a) 1600 – 1740 Mean delta 18-O = -19.2 per mille
b) 1740 – 1780 Mean delta 18-O = -18.1 per mille
c) 1780 – 1985 Mean delta 18-O = -17.4 per mille

3) Quelccaya: There are 4 regions. a) 1540 – 1640 Mean delta 18-O = -18.3 per mille
b) 1640 – 1820 Mean delta 18-O = -18.8 per mille
c) 1820 – 1900 Mean delta 18-O = -18.3 per mille
d) 1900 – 1995 Mean delta 18-O = -17.4 per mille

According to the cumulative sum of differences plot the transitions between the regions are sharp, occuring over a short time interval. I think you can appreciate this by superposing these means, with their time ranges onto the Thompson 5 year mean plots (Figure 5 A of the PNAS article).

To me this indicates that each ice sheet is responding differently. Huascaran and Quelccaya have similar patterns but with a distinct phase lag of 60 to 80 years, Quelccaya running behind Huascaran. Sajama shows no overall pattern of behaviour.

Thompson suggests that both Huascaran and Quelccaya ‘….record a consistent 20th century 18-O enrichment….’ What he neglects to say is that Huascaran also records a consistent 19th and latter part of the 18th century enrichment. He has also suggested that strong linkages exist between sea surface temperatures across the equatorial Pacific and the delta 18-O composition in ice cores from the tropical Andes and Dasuopo glacier on the Tibetan Plateau. This presumeably represents a short period correlation. It certainly is not supported by the long term (1600-present day) data which shows markedly different behaviour amongst the three tropical Andean glaciers.

A final comment, stacking the data for the three glaciers together will produce a profile which shows a pattern of enrichment beginning in 1740 and increasing in steps at 1780, 1820 and 1900 just like the composite displayed in Figure 5C of the PNAS article. One has to ask the question – given the different response of 3 glaciers in relative close proximity to each other vis-a-vis equatorial Pacific surface temperatures, or any other regional climate variable, is the stacking/averaging of z-scores amongs the 3 sites a reasonable thing to do?

51. Terry
Posted Jul 29, 2006 at 9:15 AM | Permalink

The graphs in #46 are extremely helpful. Thanks.

The graphs in #46 also tell us something very interesting about the RATE of temperature change, which is a very hot item lately as pro-AGW folks are currently claiming that the rate of temperature change is unprecedented and therefore strong evidence of AGW. The graphs allow us to compare the rate of change before the claimed AGW and after claimed AGW began.

The Huascaran plot shows a higher absolute rate of change pre AGW than post AGW (ignoring for a minute the problem that the “AGW” temperature change begins in the 1700s). Hence, the Huascaran rate of change is not anomalous post-AGW.

The Quelccaya plot shows a higher absolute rate of change post-AGW than pre-AGW (again, ignoring the proble that the “AGW” change begins in the 1800s). While it is larger post-AGW, it is still smaller than the pre-AGW Huascaran rate of change, so it cannot be said that the post-AGW change at Quelccaya is anomalous.

The Sajama plot, of course, shows no trends, so theres is no anomalous post-AGW rate of change either.

Bottom line: Thompson’s article provides evidence that the rate of temperature change in the twentieth century is not anomalous.

I look forward to 1) Thompson enthusiastically emphasisizing this valuable contribution to the science, 2) Steve Bloom citing to this paper whenever the topic of anomalous twentieth century temperature changes comes up and chiding any posters who refuse to acknowledge the science on the issue.

52. Steve McIntyre
Posted Jul 29, 2006 at 9:46 AM | Permalink

Christy said (by recollection) at the committee hearing that the panel’s view was that the panel’s view was that isotopes and temperatures were related at the poles, but that the relationship wore off towards the tropics, where the tropical glacier records had little in common.

It’s hard to say what the meaning is of just adding such disparate records. One of the nice outcomes of the current round is that Wegman and Said are puzzled by the Thompson graphic as well.

53. Paul Dennis
Posted Jul 29, 2006 at 11:27 AM | Permalink

There is now a very large, though still somewhat disjointed, data set for the isotope compsotion of precipitation, precipitation amount and air temperature. What Christy said is generally correct. Towards high latitudes there is a strong correlation between average annual precipitation isotope composition and mean temperature. This correlation also holds for monthly data, but is stronger for some months than others. The correlation is weakest at the single event level.

At lower latitudes the correlation between isotope composition and precipitation tends to be weak. There is a much stronger correlation between rainfall amount and isotope composition. However, we should be careful about generalising too much. Most of the low latitude data is collected on ocean island sites, with low elevation. The Andean glacier and Tibetan Plateau data are for high elevation sites where there is a very strong orographic effect.

I suspect that there are very few high altitude stations in the tropics that form part of the GNIP network. I’m sure that the correlation Thompson et al draw between equatorial Pacific temperatures and ice isotope composition is based on a short data set.

I’m very concered about the Tibetan Plateau data and hope to look at this in some detail over the next week. I’ll keep you posted on my views.

It’s interesting that Wegman and Said were equally puzzled by the Thompson graphic. Am I right in thinking that the 1983 date were incorporated in MBH 98, 99 and other multi proxy studies?

54. TCO
Posted Jul 29, 2006 at 12:08 PM | Permalink

An easy statistics type thing to do is to look at the differnce among the 3 records and then infer what the likely error from the true mean is based on your sample n/mean/sigma.

55. Paul Dennis
Posted Aug 2, 2006 at 4:01 AM | Permalink

There is an interesting paper in volume 103 of the PNAS:

Polissar, P.J., Abbott, M.P., Wolfe, A.P., Bezada, M., Rull, V. and Bradley, R.S., 2006, Solar modulation of Little Ice Age climate in the Tropical Andes, PNAS, v103, 8937-8942

I haven’t read the paper in detail but quote from the abstract:

‘Here we report a 1,500 yearreconstruction of climate history and glaciation in the Venezualan Andes using Lake sediments. Four glacial advances occurred between anno Domini (A.D.) 1250 and 1810, coincident with solar activity minima. Temperature declines of -3.2 +/- 1.4 degrees C and precipitation increases of ~20% are required ty produce the observed galcial responses. These results highlight the sensitivity of high altitude tropical regions to relatively small changes in radiative forcing, implying even greater probable responses to future anthropogenic forcing.’

The authors studied two sites and give evdience that glacial activity did not start until 1100 A.D. The onset of glacial activity co-incides with an extended period of low solar activity, with 4 glacial advances matching minima in the solar activity. These minima are identified using solar irradiance reconstructions based on cosmogenic isotopes (14-C and 10-Be) by other workers. Since 1820 the glaciers have been in retreat. This period co-incides with enhanced solar activity.

The authors have done a spectral analysis of the magnetic susceptibility of the lake sediments that shows significant peaks at 227 and 125 year periodicities. The magnetic susceptibility is linked to the glacial activity in the area through increased sedimentation in the pro-glacial lakes. The periodicity identified by the spectral analysis matches closely the de Vreis and Gleissberg oscillations of the solar irradiance reconstructions.

Here we have clear evidence of a dynamic tropical glacier system in the Andes responding to solar irradiance. It is interesting to note that the Venezualan glaciers start to retreat around 1820, exactly the same time that Quelccaya starts to show some sign of an increased delat 18-O signal. Meanwhile Huascaran shoes evidence of a shift to more 18-O enriched signals at ar close to 1720 and 1780. 1720 is the end period of a major minima in the solar irradiance. One needs to ask the question are the more southerly Andean glaciers also responding in some way to solar irradiance. Certainly the timing of shifts in 18-O with respect to irradiance minima and maxima is suggestive. It needs a more detailed analysis.

Not withstanding this here we have evidence of a dynamic glacier system in the Andes. The watershed did not develop a glacier until about 1150. It then shows 4 periods of advance and retreat, culminating with a modern retreat beginning in 1820.

Despite the warnings in the abstract and conclusion about AGW, there is not one bit of data that gives any support to AGW. To be fair to the authors, neither do they draw such a conclusion. UNfortunately, though they do feel compelled to mention AGW. This is rather symptomatic of the genre these days!

56. Steve McIntyre
Posted Aug 2, 2006 at 6:01 AM | Permalink

#55. Paul, here at climateaudir we aim to please. Here is my own review of the Pollissar et al article on the Venezuela Alps. They stated that the glaciers did NOT exist there in the MWP. Hormes et al 2001 recovered 5000-year old organics from retreating glaciers in the 1990s, but did NOT conclude that that showed that the retreat was unique. Quite the opposite, they identified organics from several more recent periods as well from which they concluded that there had been “Green Alps” at various times within the past 5000 years. Thompson’s reporting detail is so poor that I do not believe that it is possible to draw any firm conclusions from anything that he presents – particularly since he appears very anxious to get to a conclusion. I’m going to post up on Hormes et al 2001 to show some of the quality differences to Thompson.

57. Paul Dennis
Posted Aug 2, 2006 at 6:13 AM | Permalink

Steve,

Thanks for pointing me towards this thread. The Polissar paper is a good piece of work and well documented. I never bought into the interpretation of the 5000 year old organics at Quelccaya. Given such an important conclusion it deserved rigorous and detailed documentation by the authors.

58. beng
Posted Aug 2, 2006 at 7:59 AM | Permalink

RE 55:

Not withstanding this here we have evidence of a dynamic glacier system in the Andes. The watershed did not develop a glacier until about 1150. It then shows 4 periods of advance and retreat, culminating with a modern retreat beginning in 1820.

Other tropical areas would have there own characteristic responses, but this might suggest that Kiliminjara (Africa) formed at a similar, geologically recent time.

59. Paul Dennis
Posted Aug 2, 2006 at 8:09 AM | Permalink

Beng, I agree. What I’ll do now is go back and look at the Kenyan glacier data. I’m sure Steve is also doing this.

60. Steve McIntyre
Posted Aug 2, 2006 at 8:32 AM | Permalink

Paul, here’s a post from last year on tarns from Mount Kenya. It would be worth re-reading in light of the approach from Pollissar.

61. Paul Dennis
Posted Aug 2, 2006 at 8:48 AM | Permalink

Steve, thanks for this intro into the African glacier research. The Barker et al paper looks interesting and a quick look at the graphs suggests there is something interesting and dynamic going on. My first reaction is to suggest that the delta 18-O changes recorded by the diatoms probably represent changes in precipitation amount rather than temperature.

It’s a curse that there have been very few systematic studies done relating precipitation composition to climate conditions, rainfall amount etc. at these sites!

Anyway I’ll keep you posted.

62. Paul Dennis
Posted Aug 2, 2006 at 9:03 AM | Permalink

Steve, I’m struck by your closing comment on these tarns:

‘The negative excursions are surprisingly episodic- why is that? Since the negative excursions are associated with high turbidity, it looks like they are associated with periods of glacial melt (as well as with high snowfall – the negative dO18). I can think of two contrasting opposite states: no glacier and no melt; and a frozen glacier and no melt. Would it be possible that there are episodic periods of very high snowfall creating transient glaciers – with the long periods of high dO18 in the tarns representing absent glaciers? What accounts for the two excursions around ~6000 BP?’

The negative delta 18-O excursions, combined with high magnetic susceptibility could well indicate periods of high snowfall (depleted delta 18-O), and active glaciation. The periods in between, with increased delta 18-O and a very low and uniform magnetic susceptibility represent periods when there is little sediment input to the tarns and a reduced snowfall. It could be periods when a glacier did not exist.

This is the basis of the interpretation of magnetic susceptibility in the Polissar et al paper on the Cordillera de Merida. If the interpretation is correct then there are long periods where a glacier was absent.

63. Steve McIntyre
Posted Aug 2, 2006 at 11:01 AM | Permalink

Sonds like 2+2 =4 here. It’s a pretty logical interpretation. I’d be interested in your thoughts on the organics in Kilimanjaro some info on which I posted up here.

BTW one of the few places where you can download sample-by-sample dO18 from any ice core is here from a couple of Kilimanjaro cores, that I got from Science.

64. J Pearson
Posted Aug 3, 2006 at 3:45 PM | Permalink

I don’t understand the complaints about pdf. It took me about 30 seconds to copy the data into an ascii file. Admittedly if I wanted to plot it I’d have to spend another 5 minutes removing the annotation.

Data shown in Figure 4
Annual average àŽⲱ8O for the 1983 and 2003 Quelccaya cores
2003 1983 2003 1983 2003 1983
Year core core Year core core Year core core
àŽⲱ8O àŽⲱ8O àŽⲱ8O àŽⲱ8O àŽⲱ8O àŽⲱ8O
2002 -16.66 1957 -15.82 -15.92 1912 -17.84 -16.89
2001 -17.04 1956 -17.71 -16.52 1911 -19.69 -18.00
2000 -20.25 1955 -18.44 -18.46 1910 -18.68 -19.56
1999 -17.81 1954 -19.67 -18.38 1909 -19.31 -19.29
1998 -16.57 1953 -18.09 -19.23 1908 -18.63 -19.76
1997 -15.09 1952 -18.69 -17.51 1907 -15.07 -18.32
1996 -15.46 1951 -19.11 -18.42 1906 -14.82 -15.34
1995 -16.03 1950 -18.01 -18.84 1905 -16.61 -14.75
1994 -16.23 1949 -18.28 -18.07 1904 -17.38 -18.37
1993 -16.55 1948 -16.80 -17.89 1903 -18.63 -18.80
1992 -16.75 1947 -18.90 -16.85 1902 -18.45 -17.56
1991 -16.67 1946 -18.70 -18.34 1901 -16.77 -15.39
1990 -16.78 1945 -16.44 -18.32 1900 -16.60 -16.16
1989 -16.45 1944 -17.66 -15.50 1899 -18.43 -18.16
1988 -16.86 1943 -15.93 -17.39 1898 -19.07 -18.47
1987 -17.40 1942 -15.93 -15.57 1897 -17.34 -17.18
1986 -17.01 1941 -13.95 -16.32 1896 -16.64 -16.45
1985 -17.30 1940 -16.32 -14.47 1895 -18.55 -18.77
1984 -19.10 1939 -16.51 -14.86 1894 -20.38 -19.89
1983 -18.01 1938 -15.79 -15.79 1893 -19.43 -20.02
1982 -17.29 -19.22 1937 -15.65 -15.66 1892 -18.34 -18.07
1981 -16.95 -17.44 1936 -17.20 -15.54 1891 -18.25 -18.54
1980 -16.92 -16.01 1935 -18.66 -16.25 1890 -19.09 -18.68
1979 -17.37 -17.38 1934 -19.03 -17.80 1889 -17.21 -18.63
1978 -17.73 -17.27 1933 -19.79 -18.47 1888 -15.97 -16.81
1977 -18.05 -17.94 1932 -20.20 -18.87 1887 -17.65 -17.07
1976 -18.22 -17.87 1931 -17.96 -19.14 1886 -19.31 -18.82
1975 -19.20 -18.75 1930 -17.65 -17.42 1885 -18.42 -18.73
1974 -19.66 -19.75 1929 -17.36 -17.94 1884 -18.64 -18.98
1973 -19.04 -19.63 1928 -17.17 -17.09 1883 -17.59 -17.87
1972 -19.06 -19.01 1927 -16.05 -17.54 1882 -17.85 -18.15
1971 -19.01 -18.82 1926 -16.41 -15.80 1881 -16.76 -16.47
1970 -17.22 -18.97 1925 -16.39 -16.46 1880 -18.19 -17.69
1969 -15.81 -16.32 1924 -17.53 -16.72 1879 -19.14 -18.40
1968 -16.13 -15.39 1923 -18.66 -17.33 1878 -15.78 -15.28
1967 -15.95 -16.07 1922 -19.92 -18.79 1877 -18.75 -18.48
1966 -16.23 -16.11 1921 -19.79 -20.14 1876 -17.95 -18.43
1965 -16.14 -16.29 1920 -18.19 -19.36 1875 -20.52 -20.10
1964 -16.77 -16.55 1919 -19.01 -17.65 1874 -19.84 -20.20
1963 -17.73 -17.01 1918 -18.42 -19.65 1873 -20.53 -19.21
1962 -18.13 -18.06 1917 -18.29 -19.16 1872 -19.53 -19.33
1961 -17.33 -17.96 1916 -17.29 -17.79 1871 -19.11 -19.76
1960 -17.93 -17.62 1915 -15.73 -17.28 1870 -18.96 -19.04
1959 -16.74 -17.79 1914 -17.59 -15.58 1869 -17.98 -18.09
1958 -16.06 -15.91 1913 -16.32 -17.40 1868 -20.30 -19.66
1
Data shown in Figure 4
Annual average àŽⲱ8O for the 1983 and 2003 Quelccaya cores
2003 1983 2003 1983 2003 1983
Year core core Year core core Year core core
àŽⲱ8O àŽⲱ8O àŽⲱ8O àŽⲱ8O àŽⲱ8O àŽⲱ8O
1867 -19.05 -19.23 1822 -18.68 -19.06 1777 -19.72 -19.62
1866 -15.24 -15.47 1821 -17.34 -17.39 1776 -21.36 -20.48
1865 -17.07 -17.40 1820 -24.35 -23.71 1775 -20.09 -18.50
1864 -21.27 -19.86 1819 -20.08 -19.61 1774 -17.37 -17.87
1863 -18.41 -18.08 1818 -20.52 -20.22 1773 -18.11 -18.79
1862 -17.68 -17.28 1817 -21.06 -20.83 1772 -18.05 -17.58
1861 -20.40 -20.33 1816 -20.13 -20.76 1771 -17.49 -17.61
1860 -19.58 -19.03 1815 -19.84 -19.81 1770 -18.53 -18.63
1859 -18.79 -18.66 1814 -18.34 -19.02 1769 -20.35 -19.93
1858 -18.69 -19.15 1813 -19.80 -19.56 1768 -20.78 -19.72
1857 -17.66 -17.76 1812 -17.41 -18.08 1767 -17.10 -16.78
1856 -16.28 -16.59 1811 -20.27 -19.20 1766 -18.74 -18.67
1855 -17.53 -17.61 1810 -19.77 -20.42 1765 -18.33 -19.15
1854 -18.54 -18.20 1809 -19.65 -18.85 1764 -17.91 -18.29
1853 -18.78 -18.32 1808 -18.51 -18.87 1763 -18.17 -17.97
1852 -17.65 -18.04 1807 -20.86 -21.21 1762 -19.90 -19.40
1851 -19.73 -19.44 1806 -18.77 -17.36 1761 -19.49 -18.75
1850 -18.87 -18.53 1805 -14.78 -15.67 1760 -17.57 -18.21
1849 -20.16 -19.21 1804 -17.42 -17.57 1759 -20.14 -19.90
1848 -17.75 -19.02 1803 -19.06 -19.44 1758 -21.63 -22.75
1847 -15.70 -16.29 1802 -20.31 -19.30 1757 -20.89 -20.21
1846 -18.23 -17.99 1801 -18.84 -17.39 1756 -18.10 -18.80
1845 -17.31 -17.36 1800 -16.47 -17.16 1755 -18.93 -18.77
1844 -16.39 -17.11 1799 -19.93 -20.50 1754 -19.74 -19.79
1843 -18.99 -18.19 1798 -19.68 -19.72 1753 -19.82 -18.87
1842 -18.01 -18.82 1797 -21.21 -19.83 1752 -19.58 -18.68
1841 -16.86 -16.37 1796 -16.94 -17.15 1751 -17.58 -17.63
1840 -18.67 -18.12 1795 -17.73 -17.16 1750 -17.96 -18.57
1839 -16.71 -17.15 1794 -16.05 -16.33 1749 -16.11 -17.30
1838 -20.40 -18.90 1793 -15.76 -16.49 1748 -15.12 -14.86
1837 -18.61 -19.03 1792 -16.87 -17.00 1747 -19.81 -19.11
1836 -16.49 -17.70 1791 -18.51 -18.70 1746 -20.62 -20.40
1835 -19.89 -18.98 1790 -19.99 -19.64 1745 -18.17 -18.66
1834 -15.57 -16.60 1789 -19.78 -18.55 1744 -20.93 -20.43
1833 -19.02 -18.18 1788 -17.76 -18.75 1743 -20.65 -20.64
1832 -18.07 -17.93 1787 -19.71 -19.44 1742 -19.26 -19.54
1831 -18.42 -19.07 1786 -19.65 -19.57 1741 -16.94 -16.55
1830 -17.72 -17.60 1785 -19.33 -17.39 1740 -17.69 -17.17
1829 -16.55 -16.06 1784 -14.63 -16.56 1739 -17.40 -17.93
1828 -17.49 -17.42 1783 -19.38 -18.11 1738 -17.03 -16.63
1827 -17.40 -17.31 1782 -19.15 -18.02 1737 -19.08 -18.64
1826 -15.80 -14.47 1781 -17.49 -18.27 1736 -19.27 -19.17
1825 -18.26 -19.08 1780 -20.09 -20.92 1735 -18.16 -18.57
1824 -20.22 -19.50 1779 -21.26 -20.24 1734 -20.80 -21.16
1823 -18.46 -18.24 1778 -18.03 -18.40 1733 -19.56 -19.43
2
Data shown in Figure 4
Annual average àŽⲱ8O for the 1983 and 2003 Quelccaya cores
2003 1983 2003 1983 2003 1983
Year core core Year core core Year core core
àŽⲱ8O àŽⲱ8O àŽⲱ8O àŽⲱ8O àŽⲱ8O àŽⲱ8O
1732 -22.02 -22.90 1687 -17.31 -16.51 1642 -18.41 -22.40
1731 -19.86 -19.94 1686 -20.89 -19.75 1641 -21.29 -19.15
1730 -19.43 -19.52 1685 -19.06 -19.48 1640 -19.64 -17.29
1729 -16.03 -16.32 1684 -18.41 -18.31 1639 -17.65 -19.78
1728 -17.65 -17.35 1683 -16.65 -19.03 1638 -20.88 -19.59
1727 -19.30 -19.15 1682 -19.97 -17.01 1637 -18.59 -17.66
1726 -20.97 -20.07 1681 -17.51 -18.79 1636 -17.97 -13.41
1725 -23.02 -21.57 1680 -17.99 -20.37 1635 -13.57 -15.93
1724 -14.53 -14.96 1679 -21.03 -20.20 1634 -15.56 -18.03
1723 -15.70 -15.36 1678 -20.32 -19.84 1633 -17.62 -17.65
1722 -16.13 -16.09 1677 -19.36 -17.72 1632 -17.21 -19.09
1721 -18.19 -18.87 1676 -18.26 -19.85 1631 -19.09 -20.31
1720 -17.14 -17.32 1675 -20.38 -18.55 1630 -20.42 -18.89
1719 -17.90 -18.28 1674 -19.19 -19.47 1629 -19.06 -21.34
1718 -20.48 -19.38 1673 -18.83 -20.29 1628 -20.12 -19.47
1717 -20.12 -19.94 1672 -19.96 -19.24 1627 -19.84 -19.95
1716 -20.09 -20.62 1671 -19.16 -17.20 1626 -19.08 -19.62
1715 -18.56 -19.87 1670 -17.50 -19.14 1625 -19.86 -17.53
1714 -18.19 -18.76 1669 -19.28 -18.88 1624 -17.26 -17.35
1713 -14.76 -16.20 1668 -17.66 -18.68 1623 -17.07 -19.85
1712 -18.84 -17.21 1667 -20.71 -19.48 1622 -20.04 -18.79
1711 -17.18 -17.04 1666 -18.97 -18.36 1621 -18.39 -17.25
1710 -18.86 -17.21 1665 -18.57 -22.48 1620 -17.41 -17.34
1709 -15.81 -16.00 1664 -21.91 -20.09 1619 -17.35 -16.91
1708 -20.83 -20.64 1663 -19.99 -20.56 1618 -16.98 -18.32
1707 -18.36 -18.11 1662 -20.38 -15.81 1617 -18.03 -18.70
1706 -21.20 -19.82 1661 -14.34 -18.58 1616 -18.09 -18.71
1705 -19.95 -19.58 1660 -20.23 -20.28 1615 -18.79 -18.37
1704 -19.97 -19.67 1659 -19.41 -17.55 1614 -18.31 -18.87
1703 -18.76 -19.82 1658 -17.87 -17.12 1613 -19.92 -18.29
1702 -18.83 -18.90 1657 -17.25 -14.48 1612 -18.06 -18.73
1701 -20.98 -19.75 1656 -15.58 -17.37 1611 -19.58 -19.79
1700 -18.63 -18.49 1655 -17.12 -18.35 1610 -16.57 -17.88
1699 -19.76 -19.49 1654 -19.55 -19.68 1609 -16.67 -17.16
1698 -20.01 -21.77 1653 -18.65 -18.13 1608 -17.21 -16.83
1697 -19.97 -19.28 1652 -18.32 -16.71 1607 -17.48 -16.94
1696 -18.21 -18.60 1651 -17.25 -19.84 1606 -16.94 -18.27
1695 -18.02 -18.10 1650 -19.93 -17.99 1605 -18.89 -18.84
1694 -16.19 -16.39 1649 -18.03 -18.33 1604 -19.72 -19.02
1693 -18.10 -17.98 1648 -18.88 -16.00 1603 -18.68 -20.13
1692 -19.79 -19.73 1647 -16.58 -16.86 1602 -19.72 -20.68
1691 -16.34 -16.52 1646 -17.31 -18.23 1601 -20.86 -20.87
1690 -21.67 -20.76 1645 -18.69 -17.15 1600 -20.69 -18.78
1689 -19.83 -20.83 1644 -17.59 -20.59 1599 -18.17 -18.67
1688 -17.15 -17.44 1643 -21.67 -18.71 1598 -18.25 -17.92
3
Data shown in Figure 4
Annual average àŽⲱ8O for the 1983 and 2003 Quelccaya cores
2003 1983 2003 1983
Year core core Year core core
àŽⲱ8O àŽⲱ8O àŽⲱ8O àŽⲱ8O
1597 -17.33 -17.69 1553 -20.81 -20.40
1596 -17.71 -16.81 1552 -18.26 -18.50
1595 -16.31 -16.53 1551 -16.38 -17.03
1594 -16.22 -16.94 1550 -18.30 -17.94
1593 -17.38 -17.34 1549 -20.05 -18.77
1592 -17.00 -17.66 1548 -20.80 -19.90
1591 -17.88 -18.16 1547 -18.38 -18.73
1590 -18.28 -18.36 1546 -16.77 -17.50
1589 -18.25 -18.69 1545 -20.30 -18.57
1588 -18.89 -19.74 1544 -17.97 -19.09
1587 -19.75 -18.87 1543 -18.75 -18.28
1586 -18.14 -19.78 1542 -17.69 -18.60
1585 -19.80 -16.87 1541 -19.97 -19.19
1584 -15.87 -15.75 1540 -16.82 -18.28
1583 -16.03 -17.69
1582 -17.69 -18.03 Notes:
1581 -18.06 -18.90 Quelccaya 1983 core data
1580 -19.03 -18.03 average of annual data from Summit core and Core 1
1579 -18.14 -19.49 Quelccaya 2003 core data
1578 -19.86 -21.08 annual data from Summit Dome core
1577 -21.50 -19.13
1576 -19.45 -19.36
1575 -19.66 -18.72
1574 -18.94 -17.17
1573 -17.55 -20.10
1572 -19.68 -21.29
1571 -20.87 -16.00
1570 -16.64 -19.27
1569 -19.19 -20.43
1568 -20.49 -19.34
1567 -19.20 -19.07
1566 -14.22 -13.04
1565 -17.47 -17.57
1564 -17.06 -17.01
1563 -18.31 -18.09
1562 -20.10 -19.89
1561 -18.13 -18.29
1560 -19.78 -20.11
1559 -18.52 -17.93
1558 -18.88 -19.48
1557 -17.11 -17.52
1556 -18.43 -18.38
1555 -17.72 -17.70
1554 -18.45 -17.85
4

65. Steve McIntyre
Posted Aug 3, 2006 at 4:17 PM | Permalink

It took me the same length of time to copy it into an ascii file but it’s not a readable data file yet.

The number of entries on each line changes and you have different years on the same line. When I tried to transcribe it into an organized ascii data file to be a coherent matrix, I’m embarrassed to say how long it took me to make a correct ascii data file. Of course you can do it and of course it’s not all that hard. But it does take a little time, especially if you make any mis-steps along the way, which I did, and you then have to find them.

There was no reason whatever to put the data into a pdf file other than to be annoying.