Talk:Ice core

From Wikipedia, the free encyclopedia

This is the talk page for discussing improvements to the Ice core article.
Article policies
WikiProject Glaciers

This article is within the scope of the Glaciers WikiProject, a collaborative WikiProject related to glaciers and glaciology worldwide. It may include the Glacier infobox. If you would like to participate, you can choose to edit the article attached to this page (see Wikipedia:Contributing FAQ for more information).
B This article has been rated as B-Class on the assessment scale.


Earth sciences Portal
Ice core is part of WikiProject Geology, an attempt at creating a standardized, informative, comprehensive and easy-to-use geology resource. If you would like to participate, you can choose to edit this article, or visit the project page for more information.
B This article has been rated as B-class on the quality scale.
Mid This article has been rated as mid-importance on the importance scale.

Contents

[edit] Lost Squadron Discussion and Request

Came here looking for information on the Lost Squadron and wanted to know if any work as been done on this subject in relation to ice layers. The Lost Squadron is a group of P-38's that made and emergency landing on the eastern coast of Greenland on July 15th, 1942. The pilots were rescued after several days but the planes were abandoned. They were found again in 1992 under 268 feet (~75 Meters) of ice. The planes are now in a museum. The team that recovered the planes reported seeing hundreds and hundreds of visible layers of ice. Since a known time frame can be associated with the formation of the ice layers on the top of the planes, has any work been done to correlate the layers with other Ice Core projects? First ever Wikipedia edit so please forgive or correct my mistakes and I am not nearly knowledgeable enough to edit the main article or any Wikipedia article. I have been unable to find any work on this subject thus far. If any has been done wouldn't it be appropriate to add it to this article as it gives a known date and time frame for the formation of this thickness of ice with x amount of precipitation and also known climate conditions. Pnpjunction 23:56, 13 November 2007 (UTC)pnpjunction

[edit] Invitation

Work is currently in progress on a page entitled Views of Creationists and mainstream scientists compared. Also currently being worked upon is Wikipedia: NPOV (Comparison of views in science) giving guidelines for this type of page. It is meant to be a set of guidelines for NPOV in this setting. People knowledgable in many areas of science and the philosophy of science are greatly needed here. And all are needed to ensure the guidelines correctly represent NPOV in this setting.  :) Barnaby dawson 21:33, 27 Dec 2004 (UTC)

[edit] Leaded gasoline

Anyone know if Antarctic bases are still using gasoline with lead additives? (SEWilco 20:55, 18 May 2005 (UTC))

(William M. Connolley 21:17, 18 May 2005 (UTC)) For the UK, I'd doubt it, since its banned here (is it banned? Its certainly little used).
Leaded gasoline is also little used in the USA, although sparsely available for vehicles (such as collector cars) which require it. That's why it took me a while to realize that there was the possibility of local lead emissions around the bases. Although I'm not underestimating the sensitivity of modern equipment and being able to detect a penguin rubbing lead off rocks. (SEWilco 03:17, 19 May 2005 (UTC))
I've been told that there is no leaded gasoline at McMurdo, and that there is in fact no gasoline at South Pole Station, where all the equipment is modified to run on either diesel fuel or jet fuel. Dragons flight 21:59, May 19, 2005 (UTC)

(William M. Connolley 22:34, 19 May 2005 (UTC)) Is this anything to do with The levels of lead on the outside of ice cores is much higher than on the inside.[2]. That reads like a rather odd statement. Ice cores are plastic-bagged as they are taken (at least nowadays) so I really can't see how this could occur.

Yes, that is what got me wondering about the source. I found one set of numbers: Antarctic Peninsula 1984-1985: Pb 4.7, Cu 1.0, Zn 6.1, Cd 0.06 (atmospheric pg m&sub3;). [1] But with an ice core, a closer possible source is lead in the coring machinery (traditionally, lead in solder) or drilling fluid (including fluid containers). Looks also like penguins and rocks may be less of a source than their droppings. [2]. Oh, but never mind lead; the sulfur emissions look more interesting. [3] (SEWilco 04:50, 20 May 2005 (UTC))
Stumbled across more while looking for something else. [4] (SEWilco 05:31, 20 May 2005 (UTC))
  • "In the 1940s in Antarctica the lead concentration in snow was already approximately ten times greater than the one present in ice some thousands years ago. Between 1940 and 1980 its concentration has further increased, showing that the man contribution to the lead emissions into the atmosphere has also been significant in the austral hemisphere, even if on a smaller scale than in the north hemisphere. Our recent studies have proved that even in the Antarctic snow the lead concentration has quickly decreased during the 1980s, showing that the reduction of this toxic metal in petrol has dramatically improved the air quality on a global level."
I hadn't seen the plastic mentioned. If you know when that started, please update article. (SEWilco 04:50, 20 May 2005 (UTC))
Speaking of plastic bags, I do hope they are not the type of white plastic bags which use white lead paint, and ditto for the white boxes used for transport. :-) (SEWilco 05:05, 20 May 2005 (UTC))
Groan. Lead has often been used as a stabilizer in PVC. Vinyl, including wire insulation, too. Clear plastics seem to tend to use tin compounds rather than lead. Looks like cores rest on straight sheets of something, and I assume those are foam sheets, and I don't immediately find mention of lead in foams. (SEWilco 06:07, 20 May 2005 (UTC))

(William M. Connolley 14:02, 20 May 2005 (UTC)) I've revised the text of the page to something that seems more in tune with current practice. I'm not quite sure what you're getting at with all of the above. There seems to be a lack of a certain basic assume-people-are-competent.

There are a number of studies I have by Eric Wolff about lead in Antarctic snow, demonstrating the extreme care required to analyse the trace amounts and determine the increase to 1980 and somewhat decline since. See-also [5]. Pb is about 4 ng/kg in Coats land. Not sure where it would fit in though, since its snow not ice core.

Lead has been found in low-density polyethylene bottles. I haven't found mention of lead content in poly bags. But at least one study found LDPE levels were too low to be source of ice core contamination, and this study was examining cores stored in LDPE bottles rather than poly bags. (SEWilco 16:18, 23 May 2005 (UTC))
  • Vallelonga P.; Van de Velde K.; Candelone J.P.; Ly C.; Rosman K.J.R.; Boutron C.F.; Morgan V.I.; Mackey D.J. (2002). "Recent advances in measurement of Pb isotopes in polar ice and snow at sub-picogram per gram concentrations using thermal ionisation mass spectrometry". Analytica Chimica Acta 453 (1): 1-12.  doi:10.1016/S0003-2670(01)01490-8
I moved your material to the "Core processing" section, where there is already material about the current practices. As for atmospheric lead levels: Maybe it doesn't fit here; here so far all we have is curiosity about the ice lead levels. Recent lead levels won't be in ice cores yet. The firn description from the lead study may be of interest, as the depth and pattern of lead should be useful as a recent tracer, particularly if lead particles are involved rather than a diffusable gas. The article Lead doesn't really cover this subtopic, but such info may be relevant to Tetra-ethyl lead or Lead poisoning. Or create a "Lead in the environment" section in Lead, as I recently had to do in Mercury (element) when cleaning up a dental amalgam POV mess. (SEWilco 14:48, 20 May 2005 (UTC))

(William M. Connolley 18:39, 20 May 2005 (UTC)) The core contamination bit is out of context and misleading - it refers to early cores. I'll remove it once I've got the refs for it. Putting it back wasn't helpful.

Oh, it was helpful and relevant. (SEWilco 02:43, 21 May 2005 (UTC))


[J. Daniels, 25 November 2005] I believe it is particularly important in this politically charged area of science to maintain a clearer separation between the data and the interpretation of the data. In this article, the focus is apparently on the methods and practices of data gathering. That is, only to determine the constituent substances contained in the ice and other situational data about the ice, like geographic locations, depths, etc. A discussion of the potential for contamination of the data is also probably relevant in a general article about ice core samples. And while some mention of the types of conclusions that may potentially be drawn from the data, such as air temperature, other climatic conditions, solar activity, etc., is interesting and worthy of mention for completeness and to convey a basic understanding of the intent and thrust of the experiment as a whole, it is probably best separated wholly from the main body of the article, as it should be separated in the scientific process. When conclusions are incorporated into the process of data gathering or even into the discussions of data gathering, subsequent conclusions become more open to doubt than they would have otherwise been due to the potential for observational bias.

I think that the “Ice Core Data” section ought to be further sub-divided to break out conclusion-bearing statements from the specifics of core processing.

Statements like:

"Analysis of the Upper Fremont Glacier ice core showed large levels of chlorine-36 that definitely correspond to the production of that isotope during atmospheric testing of nuclear weapons."

and

"Dust in the core can be linked to increased desert area or wind speed."

ought to be in a separate analysis or theory section that links them directly to the studies or discussions that drew those connections: a “What is being done with ice core data” section, if you will.

Thanks for the comment. We don't have any ice core experts here, so feel free to try to improve the article. As you see, it hasn't been updated for a while William M. Connolley 20:07, 25 November 2005 (UTC).

[edit] Glaciers

however since the 1980s glacier ice has been found that is older, some cores dating as far back as 400,000 years.

Is there a ref for that? It sounds unlikely William M. Connolley 10:14, 20 February 2006 (UTC)

Well I was also surprised by this assertion, but it does indeed seem that Lonnie Thompson's group are claiming ca. 750,000 years for their oldest glacier ice core to date (cited on many web pages but mostly originating from a single press release). I don't know how they date their ice (haven't been keeping up), but here is a paper that contains this claim in its abstract [6]. Can't access Web of Science right now so can't find any more than this. Actinide 12:43, 20 February 2006 (UTC)
..originating from a single press release.. Here is the Science Magazine abstract for the first report of the Guliya glacier core. Tropical Climate Instability: The Last Glacial Cycle from a Qinghai-Tibetan Ice Core. -- Stbalbach 15:44, 20 February 2006 (UTC)
Hmm, I think the real point may be the useable climate signal. Very old ice is one thing; in sequence is quite another. William M. Connolley 13:46, 20 February 2006 (UTC).
Information is from Thin Ice: Unlocking the Secrets of Climate in the World's Highest Mountains, a fascinating account of Lonnies career and discoveries. Once the layer count runs out, they use "flow models", chlorine-36 dating and carbon dating, correlation with other records, some other methods to piece together a chronology, cross checking each other. In the end they concur with the results from the poles (with some important differences). Actually Actinide is right it's around 760,000 is the oldest, found at Guliya in 1997. The field is so focused on the poles that most people don't realize Guliya even exists. But its importance is not so much the age but the data they have found in the tropics which has been important in our understanding of climate change (which is for another day to explain!). -- Stbalbach 15:40, 20 February 2006 (UTC)
The science paper says:
An ice core record from the Guliya ice cap on the Qinghai-Tibetan Plateau provides evidence of regional climatic conditions over the last glacial cycle. 36Cl data suggest that the deepest 20 meters of the core may be more than 500,000 years old. The delta 18O change across Termination I is ~5.4 per mil, similar to that in the Huascarán (Peru) and polar ice cores. Three Guliya interstadials (Stages 3, 5a, and 5c) are marked by increases in delta 18O values similar to that of the Holocene and Eemian (~124,000 years ago). The similarity of this pattern to that of CH4 records from polar ice cores indicates that global CH4 levels and the tropical hydrological cycle are linked. The Late Glacial Stage record contains numerous 200-year oscillations in delta 18O values and in dust, NH4+, and NO3- levels.
"may be more than 500 kyr" - not 760 kyr. William M. Connolley 17:27, 20 February 2006 (UTC).

Yes, but there is a reason for that: the tropical cores aren't usable very far down. There is also a problem with their isotope thermometry, since there is no eq-to-pole fractionation in them William M. Connolley 16:52, 20 February 2006 (UTC).

A reason for what? "usable" in what sense? -- Stbalbach 17:00, 20 February 2006 (UTC)
Well, what do people use ice cores for? Primarily, to produce climate records. AS in the pic on the ice core page. Those long records come from Antarctica, where the accumulation rate is low. If your accumulation rate is higher, then the lower ice gets squashed very very thin - it must. Which means you don't get a usable signal from the older ice. Even the last interglacial isn't very usable from the mountain glaciers/icecaps (or even much from GRIP/GISP, if it comes to that). William M. Connolley 17:11, 20 February 2006 (UTC).

OTOH, this:

Glacier ice cores are more compact than polar cores, packing more years into a length, thus requiring shorter cores.[7]

is garbled. Glacier cores *are* shorter; but this is not a virtue. They are shorter just for geometrical reasons: mid-lat and tropical ice just never gets as thick as polar ice. "packing more years into a length" is close to false. Because they are high accumulation, they pack *fewer* years into a given length near the top of the core. Only at the very compressed stuff at the bottom do they have lots of years-per-length; and this is then a Bad Thing. William M. Connolley 17:16, 20 February 2006 (UTC).

Your the expert. Ill leave it to you. I just read Thin Ice and there's much more to the story, there is usable climate data in the older ice. I'm in no position to summarize it here. BTW the Bowen ref is for the whole section. -- Stbalbach 17:32, 20 February 2006 (UTC)

I'm only a humble climate modeller, but I did go ask an ice corer :-) I have the science paper now, so I'll try to add some more from that. I suspect that mid-lat/tropical cores are quite diverse. Is Guliya the pre-eminent deep core? William M. Connolley 17:41, 20 February 2006 (UTC).
Well your the Real Climate guy which is pretty cool in my book (or Bloglines account). Ill finish Thin Ice in the next day or two and have a better handle on where things stand, Lonnie went to Kenya and another mt in the Andes which I have not gotten to yet. -- Stbalbach 17:57, 20 February 2006 (UTC)
How kind, thanks. I've finished the Science paper; its interesting. I'm not sure I'm ready to edit much yet, I may even think about it... Some interesting points: the Guliya core is very much dated by comparing it to the GRIP2 core. It would be undatable, probably, without the polar cores. Also the delta-O-18 signal in it is... curious. It seems to me that there are some issues the Science paper evades. The DO18 from GISP2, for example, is clearly lower in the last glacial, with dansgaard-oeschger event spikes; but even the spikes are clearly below the Holocene values. Yet at G, DO18 was higher 30kyr ago (and at other glacial times) than now! There is no clear plateau for the holocene. The obvious interpretation (partly supported by conversation with a friend) is that much of the G DO18 signal is not temperature, but is probably the "amount effect", ie precipitation (there is some techical stuff hidden there in the way equ-pole fractionation occurs, which I don't fully understand).
The climate signal at G "only" goes back about 110 kyr. The very old ice is dated, indeed, from Cl36, and apparently there is so little Cl36 at the bottom that half-life arguments indicate that its > 500 kyr old. But... there is no evidence in the Science paper, at least, that this old ice is useful for anything! Apart from knowing the ice cap has been there that long, but I doubt that is new info. William M. Connolley 23:07, 20 February 2006 (UTC)

Also: I changed "glacier" into "non-polar". I would be fairly sure that most of them were done into ice-caps: the Guliya and Dunde certainly were. Which makes sense: the flow you get in a glacier would make the core nearly unusable. William M. Connolley 23:15, 20 February 2006 (UTC).

Also, I created an LT stub. If there is any decent biog info in the book, could you put some in? His home page is *not* informative... William M. Connolley 23:31, 20 February 2006 (UTC)

[edit] Gas age / pore close off

I removed the existing table, which was clearly wrong. Some of it was based on misinterpretation of [8]. The table asserted that Estimated air ages at firn closeoff was 6 years for GISP2. I don't believe that. Firn depth at GISP2 is 80m; accumulation must be about 0.1 m/y; which suggests a value more like 800 y. The above ref also sez At Vostok, for example, the bubble closeoff zone is about 8 m thick. A single layer of ice traps bubbles throughout the ≈300 yr it moves through this zone. so the previously claimed max of 156 years must be wrong (as well as far too precise).

If anyone knows this better, please comment William M. Connolley 20:28, 24 February 2006 (UTC)

Ah, browsing the same I found The gas age–ice age difference ([Delta]age) is as great as 7 kyr in glacial ice from Vostok. William M. Connolley 20:29, 24 February 2006 (UTC)
Ah #2: the ref sez The diffusivity of the firn is such that air at the base of the stagnant column today has a “CO2 age” ranging from about 6 yr for the GISP2 core (central Greenland) to about 40 years at Vostok (East Antarctica). which I think someone has misinterpreted: it doesn't mean the gas-ice age diff is 6yr; it means the air is 6yr old relative to the sfc, and therefore is much younger than the surrounding ice. William M. Connolley 20:33, 24 February 2006 (UTC).

[edit] Melting cores/Core Sampling Rate?

I'm not sure where this question belongs, so I'll ask it here and maybe someone can move it to the correct location when they answer it. If there is significant melting of ice caused by global warming resulting from increased CO2 and other greenhouse gases, how can that quantity of CO2 be taken from ice cores if there was no ice to trap the air bubble? Wouldn't it stand to reason that the bubbles trapped are the ones that could be trapped once the temperature levels dropped to a point that ice could reform? I'm sure there are ways that they can age a bubble by examining something, but how can they know what CO2 levels were in a period where ice was melting? Take what's happening currently. If the poles keep melting as scientists predict, how would scientists 400,000 years from now know what levels of CO2 we actually reached? What if all the ice on earth melted and then it refroze? Is there a way they can tell that from the ice core samples? I'd like to see this explained in the article. Thanks. Jimberg98 22:00, 4 January 2007 (UTC)

Cores are only usable from locations where there is (little or no) melting. This is true for all the major cores from Greenland and Antarctica. As for the future... we currently have much better mreasurements from direct atmospheric sampling William M. Connolley 22:08, 4 January 2007 (UTC)
So there are some locations where we know they never melt? Ever? I understand that we take do direct atmospheric sampling now. That's because we can. If ice is melting now, or 100 years from now, or 500 years from now, no data is being gathered by the ice, right? I guess an easier question would be, what is the sampling rate of ice core data? Is it constant? Thanks. Jimberg98 22:29, 4 January 2007 (UTC)
Yes (ie, that they haven't in the past). Certainly the ones in the interior of Antarctica. If the ice is melting, then the core won't be much use. Sampling rate depends on precipitation rate. But since there is the pore close-off problem, the resolution for gases is poor - decades, depending on the site William M. Connolley 22:50, 4 January 2007 (UTC)
When you say decades, are you talking 20 years, 40 years, 80 years? What about rates of sublimation in the locations that never melted in the past? How does one determine that ice has never melted or sublimated in the past if the evidence could have melted or sublimated? I'm not a skeptic about global climate change. I would expect that the earth would continue on a warming trend after coming out of an ice age just like it has always done in the past. I'm just wondering if we're like creatures that live for 30 minutes wondering about how we're causing summer and how to prevent it. How can we compare data with a daily sampling rate to data having an uncertain rate of decades? What's the margin of error for ice core samples? Jimberg98 15:15, 5 January 2007 (UTC)
Ice that has melted and refrozen has a very different crystal structure than that formed from compacted snow. If you are familiar with the problem, you can tell if there was melt just from looking at the core. On any given measurement, the uncertainty is likely to be few percent in age (for samples older than a couple thousand years) and a few to several ppm in concentration. Changes of 100 ppm (such as has occurred since 1800) are many times larger than the uncertainty in past concentration. Dragons flight 15:33, 5 January 2007 (UTC)
What about the crystal structure of ice that sublimated? Jimberg98 17:55, 5 January 2007 (UTC)

Maybe I asked about resolution in the wrong way so I just want to confirm what you're saying. I went to download the CO2 data collected from the Vostok cores and noticed that there are a couple thousand years between some sample ages. Some seemed to be 6 years. What is the average sample rate of data stretching back 400,000 years? When you answered about resolution, was that the accuracy of aging a sample? Thanks. Jimberg98 22:12, 16 January 2007 (UTC)


Random comment that may/not have to do with this: I don't have the exact cited sources, but the pattern of rings in an ice core is NOT summer-winter-summer-winter. It's warm-cold-warm-cold. I live in Muchigan. In January, we can have five of those warm-cold periods in a week. So any 'evidence' collected using the summer-winter thing would be wrong, if the warm-cold thing is correct.Could somebody maybe do some research on this? I don't have any time, but I think it's worth adding to the article if it's a valid theory. 204.38.47.171 16:21, 26 October 2007 (UTC)

[edit] Vostok Ice Core Graph

I've been reading a lot on global warming lately and I always come across this Vostok graph which is supposed to show how our current warming trend and increase in atmospheric CO2 concentrations is unprecedented in history. I went to the NOAA web site to find the source data for this graph and found that many of the samples are 1000 or more years apart, yet the points are connected as though there's a linear increase between those to points. This is very misleading unless the data I'm looking at is incomplete. Shouldn't the graph be a scatter plot instead of a linear graph?Jimberg98 00:02, 20 January 2007 (UTC)

I doubt it makes much difference. You can see he higher-rez EPICA data at ftp://ftp.ncdc.noaa.gov/pub/data/paleo/icecore/antarctica/epica_domec/edc-ca-na.txt if you like William M. Connolley 12:00, 20 January 2007 (UTC)
That's sodium and calcium data. Is that the data used to determine CO2 levels? Jimberg98 04:51, 21 January 2007 (UTC)
Oops, sorry, how about ftp://ftp.ncdc.noaa.gov/pub/data/paleo/icecore/antarctica/epica_domec/domec_co2.txt ? William M. Connolley 12:47, 21 January 2007 (UTC)
No problem and thanks. The new data only covers about 10,000 years with the majority of the numbers spanning 90 years or more. Some spans are over 200 years. Again, I think it's misleading to show the data as though it's linear. My biggest concern, obviously, is that I read a bunch of articles where the authors draw comparisons between this data and data that was directly observed and truly is linear. Would a CD sound the same if you only used only 490 samples per second instead of 44100?Jimberg98 00:12, 22 January 2007 (UTC)
Technically you're correct - it's misleading to use a solid line which suggests continuous measurement. However, this is palaeoclimate data. Consequently, and regardless of the archive (ice, marine or lacustrine sediments etc..) it's not possible to avoid sampling to some level. Given the length of the record it's unlikely that any significant changes haven't been picked up within the intervals used. Additionally, a solid line helps those not used to interpreting graphed data pick up any trends. Making comparisons with modern records is perfectly valid so long as the source of the data is explicit (as in scientific papers). I'm not sure how else we could to do it. NickW 09:42, 22 January 2007 (UTC)
My problem is that if you go over to the global warming page, you see directly measured data included in the graph that has samples that can be as far apart as 6000 years. Who's to say that there weren't periods in that missing data that are comparable to today's measurements?Jimberg98 17:31, 26 January 2007 (UTC)
hypothising non-observed continuities in basically chaotic system, hmm, the simplest model is usually the best (at least most easily applicable), but concerning this i'm not so certain. 91.153.52.32 10:41, 27 September 2007 (UTC)

[edit] References

Reference 1 (Kaspers, Karsten Adriaan) points to Dutch text, please clarify location 81.217.19.215 15:11, 18 November 2006 (UTC)

Try searching for the name of the document. (SEWilco 06:21, 21 November 2006 (UTC))

[edit] No evidence of a greater temperature cycle in relation to Vostok Ice Core data.

According to the Vostok Ice Core Data, Antartica was ice free around half a million years ago, even though their own data shows no greater temperature cycle in play. Where is the evidence the Earth was vastly warmer half a million years ago when ice began accumulating in the Antartic?

If there was no ice accumulation, then there would be no ice core data. The Vostok Ice Core Data is about to hit rock bottom. Only 300 meters to go. Or maybe it already has.

Half a million years ago, Antartic is still smack dab in the middle of the Southern Pole, receiving just as little sunlight as it is today. What changed?

Where is the evidence of a greater temperature cycle?

How was it so vastly warmer half a million years ago that ice began to accumulate at the Vostok Lake basin?

Why is it that over the last half a million years to present, that ice NEVER melted, giving perfect little layers for the Ice Core gatherers to gather Ice Cores?

70.176.5.79 19:45, 8 February 2007 (UTC)

[edit] What is the margin of Error of Ice Core Samples?

every measurement has a Margin of Error, but my google-fu can not locate this...--65.107.88.154 14:10, 26 October 2007 (UTC)

The margin of error is probably not very small, just because these are sampled measurements, and for that reason the Nyquist–Shannon sampling theorem applies. This means that the process should have been filtered for variations above half the sample frequency. If this is not done in advance of the sampling, aliasing will occur and distort the dynamic information of the measurements. And since there is no possibility of filtering in advance of sampling, the measurements are by nature distorted. Unfortunately, it is not possible to tell how big this distortion is, because there is no way of knowing how influential the dynamic behavior above half the sampling frequency is, compared to the frequencies below this frequency. But typically, the Vostok core is sampled at a rate of one sample per 1000 years (approx.), so if filtered correctly the Vostok core would have given dynamic information of variations up to 1 per 2000 years. It is not realistic to assume that earth's weather system could not have changed considarably within periods of 2000 years. Therefore, the error could be considerable. Arconada (talk) 20:48, 6 May 2008 (UTC)
There are at least two different questions here. First, what is the margin of error of an individual measurement of, say, d-o-18. Second, what processes are missed by only having a sample rate of, say, 1 per kyr. For the first, I think that many cores are repeat sampled (at least the recent ones?) and the error bars are much smaller than the variation displayed. For the second, there is diffusion within the cores (certainly of the gases included, perhaps rather less of the ice content itself) so smaller subsampling wouldn't help. Actually, that was probably a third... William M. Connolley (talk) 21:12, 6 May 2008 (UTC)
I cannot give you the margin of error of an individual measurement, but the fact that many cores are repeatedly sampled, as you said, does not impress me: similar measurements result in similar errors. It is simply necessary to find an entirely different and totally independent way to confirm the measurements of the climate down to 400.000 years ago. To your second question, what processes are missed by having a sample rate of 1 per 1000 years? Well, all the processes, random or cyclic that are above the frequency 1 per 2000 years. And looking at the last 2000 years, a lot of variation can occur in this time span. Actually, the processes are not really missed, but irreversably mixed into the frequency spectrum of 0 to 1 per 2000 years. To your third remark, I suppose there is diffusion in the ice core, and it acts as a filter, but that is not enough for two reasons. You’re a mathematician so you will probably know that aliasing is not prevented by ‘some’ form of filtering, but by a rigorous form of filtering. But even if you could design a filter to eliminate all the frequencies above the Nyquist frequency in the ice core, or the other way around, if you could sample at a high enough sample-rate(whatever it may be), theoretically it would not help, because aliasing is part of the process: aliasing already took place at the very beginning. Some measurements or processes are sampled by its nature. A sample of the atmosphere was taken at intervals every time a snow-shower took place. Maybe a day passed before another shower, maybe a week, or a month or even a year. The bottom line is, the ice core is a sampled measurement of the atmosphere at that time. The variations of the atmosphere were not recorded in between samples, therefore aliasing took already place then. This is not an uncommon paradox. Many measurements take place without being able to filter for aliasing, many processes are typically sampled. What to think of opinion polls, where every month voters are asked to say what they will vote in the next elections. We all know that the time-line that is produced in this way, does not predict the result of the election, and this is partly caused by the fact that we are unable to prevent aliasing. Or what about the stock market? The price of a commodity is only determined at the exact moment when a buyer and a seller agree about the price, and a trade takes place. In between trades the price could be anything: we don’t know what the price is, because there is no agreement. The point is, the price is sampled at trade-moments, so in the graph of the price the aliasing error is included because we are not able to filter against it anymore. Maybe this is one of the reasons why we perceive some processes as chaotic. But back to the question, measuring the climate by ice is a process of sampling, from the very beginning to the end. Aliasing is an important and unpredictable factor. So in response to the original question, I said the margin of error is probably not very small. Arconada (talk) 15:05, 7 May 2008 (UTC)
It is worth keeping in mind that not all techniques are the same. For recent cores, some of the chemistry (e.g. sulfate) and stable isotopes (deuterium) are gathered through a nearly continuous sampling process that generates data much more densely, and rarely skips more than a few hundred years. Some physical measurements (e.g. conductivity [proxy for ion concentration] and optical scattering [proxy for dust]) are even done truly continuously so that every inch of core is scanned. Unfortunately the gas data (e.g. CO2) is much more complicated to collect and hence is often sampled with skips of a few thousand years. To the extent that the more densely studied proxies are tightly correlated to gases and don't show large fluctations, it suggests (but does not prove) that there is not a lot being missed in the sampling of gases. Also, for the record, gaseous diffision of the interesting species operates at a rate of tens of centimeters per 100,000 years, and for most of the core would not have a major role when gases are sampled every couple meters. Dragons flight (talk) 15:58, 7 May 2008 (UTC)
Still, sampling related errors do occur. "Nearly continuous sampling", as you said, is sampling, not a continuous measurement. And even if the conductivity and optical measurements are truly continuous, the precipitation isn't. The ice is sheet is created by discontinuous snow falls, samples. So the history of the atmosphere is sampled and aliasing is part of that. It’s a paradox. Arconada (talk) 10:09, 17 May 2008 (UTC)

[edit] What exactly does the proxy temperature data represent?

When using isotopes to determine temperatures from the ice cores, what does it represent? Average over a certain amount of time? Peak temperatures? 150.228.40.142 (talk) 23:52, 17 December 2007 (UTC)


[edit] How Many Annual Cycles Can Be Visually Identified?

Could anyone tell how many annual cycles can be visually counted from the top of the ice core downward? Is it on the order of hundreds, thousands, or tens of thousands,or more?

Thank you!

--Roland 07:47, 4 March 2008 (UTC)

It depends on the area where the core was laid down. In greenland, they can layer-count 10,000 years and possibly more. This is because accumulation rates are (comparatively) high so you get a lot of layer to count each year. In Antarctica, rates are typically lower, and you can't count so far (but in exchange you get cores that cover a longer timespan). See for example the most recent RC [9]. BTW, I'm fairly sure that what is counted are not visual layers, but seasonally varying chemical traces William M. Connolley (talk) 08:23, 4 March 2008 (UTC)
Actually, layer counting is often based on a combination of annual variations in chemistry, optical layers(influenced by ice crystal size, and ultimately temperature at time of precipitation), and electrical conductivity. The use of multiple proxies helps to resolve layers when a single proxy may be ambiguous. As can be seen in images like Image:GISP2D1837 crop.jpg, annual layers do have a distinctive visual trace. At GISP2, the layer counters claimed to reach over 40,000 years if I recall correctly (with increasing uncertainty in the deep portions). Dragons flight (talk) 16:06, 7 May 2008 (UTC)