It’s true that intra-site comparisons should give useful information about the reliability (or lack thereof) of either core or of their average.

There is a fine line between valid empiricism and invalid data mining that makes this a much boader scientific/statistical issue. No easy answers at present!

It seems to this layman that a great deal of testing and analysis would have to be done before one could think that the two differing Quelccaya cores could either be averaged or used (either) alone with confidence that one had a reliable, calibrated relationship with temperature.

HU (above):
“OK, Pete, thanks to Ken’s comment I now see what you mean in Ken’s 10-yr moving average accumulation graphs. The two Quelccaya cores do look very different before about 1500 AD….”

It would appear that one could put limits on some of these effects, like diffusion, and thus I would expect that uncertainty bounds could be attached to the results.

CO2 and other gases trapped in ice core bubbles are more problematic vis a vis smearing of the record than O18 ratios that reside with the water molecules which are part of the solid ice. The water vapor in the ice cores trapped bubbles is, I am assuming, a small fraction of the water in solid ice. Of course, that situation leaves the water vapor to migrate into the ice and the water molecules to diffuse through the solid ice. Most of the work I have viewed in the literature talks about CO2 and gases trapped in the bubbles and much less about O18 in water diffusion in the ice.

What I have noticed about the reporting of the Thompson ice core results is that while detecting millennial temperature differences seem rather secondary in the published journal material, the press reports zero in on any proxies that show late series warming and without any precautions about the uncertainties of the results or explanations of why there are uncertainties. It would also appear that those using these proxies in temperature reconstructions do not bother to obtain any independent view on the proxy’s validity as a temperature responder.

Below is a link to a paper on the Guliya Thompson ice core. By registering you can obtain a free copy of the full article.

http://www.sciencemag.org/content/276/5320/1821.abstract

Below I have a couple of links and excerpts that would indicate that issue of gases in ice core bubbles is not a closed issue.

http://www.geocraft.com/WVFossils/Reference_Docs/CO2_diffusion_in_polar_ice_2008.pdf

“ABSTRACT. One common assumption in interpreting ice-core CO2 records is that diffusion in the ice
does not affect the concentration profile. However, this assumption remains untested because the
extremely small CO2 diffusion coefficient in ice has not been accurately determined in the laboratory. In this study we take advantage of high levels of CO2 associated with refrozen layers in an ice core from Siple Dome, Antarctica, to study CO2 diffusion rates. We use noble gases (Xe/Ar and Kr/Ar), electrical conductivity and Ca2+ ion concentrations to show that substantial CO2 diffusion may occur in ice on timescales of thousands of years. We estimate the permeation coefficient for CO2 in ice is _4_10–21 molm–1 s–1 Pa–1 at –238C in the top 287m (corresponding to 2.74 kyr). Smoothing of the CO2 record by diffusion at this depth/age is one or two orders of magnitude smaller than the smoothing in the firn. However, simulations for depths of _930–950m (_60–70 kyr) indicate that smoothing of the CO2 record by diffusion in deep ice is comparable to smoothing in the firn. Other types of diffusion (e.g. via liquid in ice grain boundaries or veins) may also be important but their influence has not been quantified.

http://icebubbles.ucsd.edu/Publications/closeoff_EPSL.pdf

“Gas ratios in air withdrawn from polar firn (snowpack) show systematic enrichments of Ne/N2, O2/N2 and Ar/N2, in the firn–ice transition region where bubbles are closing off. Air from the bubbles in polar ice is correspondingly depleted in these ratios, after accounting for gravitational effects. Gas in the bubbles becomes fractionated during the process of bubble close-off and fractionation may continue as ice cores are stored prior to analysis.”

===|==============/ Keith DeHavelle

http://adsabs.harvard.edu/abs/2005E&PSL.229..183I

As an aside here, if I use the molecular diffusivity of H2O, and here specifically H2O with O18 in ice, of 3×10^-10 meters per second and look at a time period of 1000 years (3.15 x10^10 seconds), I have migrations of O18 in H2O in ice on the order of a meter -providing my assumptions and calculations are reasonable.

http://www.climate.unibe.ch/~stocker/papers/bereiter09grl.pdf
Change in CO 2 concentration and O 2 /N 2 ratio in ice cores due to molecular diffusion

“Equation (1) governs solely the part of the gas that is dissolved in the ice which corresponds to less than 1% of the total air content (over 99% are kept in inclusions, either in air bubbles or clathrates). The two reservoirs (gas dissolved in ice and kept in inclusions) are assumed to be locally in equilibrium at all times (i.e., for the numerical case within each layer).”

http://adsabs.harvard.edu/abs/2005E&PSL.229..183I

“Enrichment of nitrogen gas has been found from gas analyses of ice cores retrieved from deep parts of Antarctica. Neither climate change nor gas loss through ice cracks explain the enrichment. In order to investigate the mechanism of the gas composition change, we develop a model of gas loss caused by molecular diffusion from clathrate hydrates toward the ice-core surface through ice crystal. We apply the model to interpret the data on the gas composition change in the Dome Fuji ice core during the storage for 3 years at 248 K. The mass transfer coefficients determined using the model are 1.4×10-9 and 4.3×10-9 m•s-1 at 248 K for N2 and O2, respectively. The difference in the coefficient between N2 and O2 causes the change in the O2/N2 ratio of the trapped gas in the ice core during the storage. During the storage period of 1000 days at 248 K, the O2/N2 ratio changes from -9.9‰ to-20.5‰. The effect of the gas loss decreases as the storage temperature decreases. The results have important implications for the accurate reconstructions of the paleo-atmosphere from polar ice cores.”

Here is where I have trouble understanding how much H2O vapor and CO2 capture in ice differ the final results for CO2 levels and O18 ratios. In the case of CO2 the authors always talk about the formation of bubbles in the firn and ice that trap the CO2 from further diffusion or at least more rapid diffusion and they appear to me to indicate that the most recently laid layers of snow cannot be used to determine an annually resolved CO2 level. With O18 we know that the most recent layer of snow can be, or at least are, used for determining annual ratios of O18 in the water precipitated as snow. It appears obvious that difference between CO2 and O18 levels in snow and ice are that CO2 is first and always a gas while the O18 in water is in the form of a solid in snow and with some lesser amount in the form of a gas entrainment depending on the density of the snow. I would suppose that the entrapment of gaseous H20 and CO2 in bubbles would be nearly the same but the big difference being that CO2 levels would be obtained primarily in gas form in the bubbles and secondarily from any dissolved CO2 in the ice. O18 from H2O is going to be primarily from the ice itself and secondarily from the H20 vapor in the bubbles. I do not know how much the O2 trapped in bubbles and dissolved in ice would interfere with the O18 in H20 measurement.

I am currently of the opinion that O18 would be affected mainly by molecular diffusion of H2O in ice at same long residence time and some great compaction of ice into thin layers and that CO2 is going to be affected by gaseous diffusion. After looking again at the diffusion constants for molecular H2O, and salts for that matter, in ice and realizing that I have to take a square root and divide by pi to obtain meters diffused in seconds, I think the slow diffusion process eluded to in the link on O18 was referring to molecular diffusion. I need to investigate this further, but at this point my supposition could agree with your point on the smearing of the annual O18 ratios at depths in an ice core – given enough compaction of the ice and sufficiently long periods of time.

The diffusion you cite for H2O in ice itself is too slow to mess things up much by itself — 1.3×10^-18 m^2/s equals 4.1×10^-8 m^2/millenium, in other words 0.2mm in 1 millenium or 2.0mm in 100 millenia, not enough to erase the signal even in the Vostok cores.

However, I am still concerned about the absorption of CO2 into ice. I assume that newly fallen snow, unlike rain, is essentially free of CO2. However, under high pressure and given enough millenia, can ice absorb gasseous (or liquid) CO2, if only near the surface? Since a microscopic air bubble is all surface, this could greatly change its CO2 content.

I have in mind the ancient “cementation” or “calamine” process for making brass: zinc boils below the melting point of copper, so you can’ just put them both in a pot and melt them together. However, hot copper that has not yet melted will absorb zinc vapor into its surface (say the first 1 mm), even though it is still “solid”. Therefore if you hammer copper into thin sheets that are all surface, and place these sheets in a crucible with zinc that is smelting from calamine into the vapor state, the copper sheets will turn to brass with up to 28% zinc content.

Likewise, even though ice is “solid”, it may be able to absorb gasses like CO2, if only very near the surface. Especially at the highish pressures at the bottom of an ice core. (Nothing like the pressures needed for the protonic diffusion you mention, I think.)

It is usually assumed that all the bubbles seal right at 70 or 80 m of overburden, but obviously this is not true — If that’s the average, some must seal at 30 m and some at 100 m, adding a further element of uncertainty to the dating of the air. Back when he was speaking to me, Lonnie Thompson once cautioned me against making too much of the relative timing of the CO2 and d18O in the Vostok cores, because of the great uncertainty in the air/ice lag. Sounds reasonable.

I’ll take a look at the articles you cite when I get some time.

http://debunkhouse.wordpress.com/2011/01/05/antarctic-ice-cores-diffusion-confusion/

“The age of the layers of ice can be fairly easily and accurately determined. The age of the air trapped in the ice is not so easily or accurately determined. Currently the most common method for aging the air is through the use of “firn densification models” (FDM). Firn is more dense than snow; but less dense than ice. As the layers of snow and ice are buried, they are compressed into firn and then ice. The depth at which the pore space in the firn closes off and traps gas can vary greatly… So the delta between the age of the ice and the age of the air can vary from as little as 30 years to more than 2,000 years.

The DE08 core from Law Dome core has a delta of 30 years. When the core was drilled in 1992 pores didn’t close off until a depth of 83 m, in ice that formed in 1939. According to the firn densification model, air from 1969 was trapped at that depth in ice that was deposited in 1939.

It doesn’t seem reasonable to assume that ”1969″ air was trapped at 83 m in “1939″ ice It seems to me that at depth, there would be a mixture of air permeating downward, in situ air, and older air that had migrated upward before the ice fully “lithified.” The air trapped in the 1939 layer should be a blend of air from 1909 to 1969. At the time that the 1939 layer was deposited, the ice crystals above 1909 would not have “lithified” yet. In 1939, the air within the interstitial pore space would be a mixture of 1909 to 1939 air. By the time the 1969 layer was deposited and the 1939 layer “lithified,” the air at the 1939 layer would have been a blend of 1909 to 1969 air.”

The diffusion problem of O18 in H2O in ice cores is explained in the link and excerpt below and involves H2O vapor diffusion and not perceptible liquid or solid ice diffusion – at least as I interpret what I read. It appears that the article here talks about the limits of using O18 to determine the annual layer and not about the smearing of O18 concentrations/ratios we are talking about. The process as described in this article would ,however, agree with your thinking about diffusion in deeper ice tending to smear the annual O18 concentrations/ratios.

http://www.iceandclimate.nbi.ku.dk/research/strat_dating/annual_layer_count/diffusion/

“Snow is slowly compressed into ice in the upper 80 meters of an ice sheet (read more about the process here). During this process, water vapour can move relative to the ice in the open pores between the snow grains, thereby smoothing the annual δ18O cycles. This diffusion process smoothes the δ18O signal and even erases the annual signal if the annual layers are thinner than 15-20 cm. In ice cores from sites with less than 15 cm of precipitation (measured in equivalents of compacted ice, not snow) per year, the annual cycle in δ18O will be obliterated, and dating based on annual δ18O oscillations is therefore not possible. This is the case for areas in north-eastern Greenland where the annual precipitation rate is significantly lower than 20 cm. For ice cores drilled in areas with about or slightly more than 20 cm of precipitation, diffusion will also blur the annual cycles, but it is possible to retrieve the annual cycle using diffusion correction techniques.

Very slow diffusive processes also take place deeper in the ice sheets. These processes slowly weaken the annual δ18O oscillations as the ice gets older and the layers thin due to the flow of ice.

Due to the diffusion processes, the limit of safe annual layer detection using δ18O / δD measurements is about 8500 years ago in the DYE-3 ice core. More favourable conditions at the summit of the Greenland ice sheet has permitted successful identification of annual layers from δ18O data in more than 14,000 year old ice from the GRIP ice core, while the NGRIP and NEEM ice cores cannot in general be dated using δ18O data alone.”

The diffusion of anions, cations and molecules in ice would appear to a much slower process and though I have not made any calculations I am guessing that those processes are not a problem in ice cores – or at least a much smaller one. The links below gave me some ballpark rates.

http://www.seas.upenn.edu/~biophys/cv_files/diffusion.pdf

“We report molecular simulation studies of the diffusion processes in ice and CO2 clathrate hydrates performed using classical potential models of water ~SPC/E and carbon dioxide ~EPM2. The
diffusivity of H2O in ice is calculated to be 1.3×10^-18 meter^2/seconds at 200 K using molecular dynamics simulations, a result in good agreement with experimental data. ”

http://www.sciencemag.org/content/295/5558/1264.abstract

“Near ambient pressures, molecular diffusion dominates protonic diffusion in ice. Theoretical studies have predicted that protonic diffusion will dominate at high pressures in ice. We measured the protonic diffusion coefficient for the highest temperature molecular phase of ice VII at 400 kelvin over its entire stable pressure region. The values ranged from 10−17 to 10−15 square meters per second at pressures of 10 to 63 gigapascals. The diffusion coefficients extrapolated to high temperatures close to the ice VII melting curve were less by a factor of 102 to 103 than a superionic criterion of ∼10−8square meters per second, at which protons would diffuse freely.”