What Does New Research Reveal About the End of the Last Glacial Period?

[Andre]

Today I'd like to draw the attention to the end of the last glacial period. Those who have looked at ice core graphs may recall that the Antarctic ice cores appeared to react first around 17,000 years ago and then with a relative mild change in isotope values (interpreted as temperature change) while the Greenland Ice cores jumped up suddenly with an unprecedent pace around 14,700 years ago, known as the Bolling event (of the http://home.wanadoo.nl/bijkerk/grootes.GIF ):

http://home.wanadoo.nl/bijkerk/vostok-gisp.GIF

So it appeared and was accepted as general paradigm that the southern hemisphere got out the glacial period first, followed by the northern hemisphere, (edit: despite the close correlation between hemispheric temperatures nowadays).

However, then it seemed to be time for another ugly fact as in: "The great tragedy of science -- the slaying of a beautiful hypothesis by an ugly fact. (Thomas Huxley).

So this study appeared:

http://www.sciencemag.org/cgi/content/abstract/312/5779/1510

Schaefer, J.M. et al 2006; Near-Synchronous Interhemispheric Termination of the Last Glacial Maximum in Mid-Latitudes Science 9 June 2006: Vol. 312. no. 5779, pp. 1510 - 1513

Isotopic records from polar ice cores imply globally asynchronous warming at the end of the last glaciation. However, 10Be exposure dates show that large-scale retreat of mid-latitude Last Glacial Maximum glaciers commenced at about the same time in both hemispheres. The timing of retreat is consistent with the onset of temperature and atmospheric CO2 increases in Antarctic ice cores.

We suggest that a global trend of rising summer temperatures at the end of the Last Glacial Maximum was obscured in North Atlantic regions by hyper cold winters associated with unusually extensive winter sea ice.

More elaboration here.

Now after the initial shock, let's discuss this abstract. This is very typical and logical development. The first paragraph is the actual description of the study in a nutshell. The results of months of work of obtaining and thorougly processing data with high skill etc.

But the result is not what is expected and moreover it challenges one of the most firm paradigms of the ice age, that is that the isotopes of the ice cores reflect the temperatures of the past. So you need an ad-hoc hypothesis in an attempt to harmonize the results being:

We suggest that a global trend of rising summer temperatures at the end of the Last Glacial Maximum was obscured in North Atlantic regions by hyper cold winters associated with unusually extensive winter sea ice.

Now I have not seen the study itself yet and I wonder if this statement is backed by any physical evedence. It could be but I've seen lots of studies where such an hypothesis is also the last sentence of the conclusion not even with reference to (recent) work.

So let us investigate the merits of that idea and we find http://web.awi-bremerhaven.de/Publications/Dol2002a.pdf , addressing Summer Sea Surface Temperatures (SSST) in de Nordic sea (Greenland Iceland Norway area).

I assume that "unusually extensive winter ice" would also have considerable effect on the max possible SSST. After all, freezing and thawing of the sea surface takes a while limiting the time for warming up in the summer time.

Hypotheses need predictions. So you would expect low SSST before that Bolling Allerod (14,700-12,800 Calendar years ago). That's essential, then perhaps a sharp increase next matching the ice core spike and a sharp decrease at the onset of the Younger Dryas and finally a sharp increase at the Preboreal some 11,600 years ago. Because that's how science worx, isn't it, verifying the hypothesis by investigating logical predictions.

Note, (table 2 on page 5), that the earliest recorded SSST of 13,400 14C BP (is about 16,000 Calendar Years BP) is already a blazing hot 11.6 degrees C compatible with the general modern Holocene values. Not much sea ice there -I would say- at the Norway coastal area well before the onset of the Bolling period around 14,700 Calendar years ago.

Furthermore, it’s a bit surprising that the SSST drops down more or less steadily (fig 3 page4) despite the apparent warmth of the Bolling Allerod, well before the assumed cool Younger Dryas. Only one thing is correct of our little predictions, the incredible spike initiated around 11,500 years Calendar years (10,070 14C years), which is consistent with some pet idea.

So what is the verdict:

was the end of the Last Glacial Maximum obscured in North Atlantic regions by hypercold winters associated with unusually extensive winter sea ice?

Nope, not really. The ocean had warmed already as well before the onset of the Bolling Allerod consistent with the global warming signal of Schaefer et al; actually there is no warming signal at all during that B-A onset more a cooling signal considering the summer sea surface temperutes.

As the temperatures go everywhere except where they are thought to go, it would appear to falsify the temperature explanation of the strong Greenland ice core isotope spikes at the last glacial transition.

So, there is a whole lot more work to do.

 

[Andre]

So let's start then.

Again, let's assume that combination of Schaefer et al and Dolven et al means the refutation of the big (d18O) isotope spikes in the Greenland ice cores ( http://home.wanadoo.nl/bijkerk/vostok-gisp.GIF ) then what else could those spikes signify?

To support the next narration I was planning to use publications of Georg Hoffmann here, http://www.ipsl.jussieu.fr/GLACIO/hoffmann/hoffmannengl.html

with whom I discussed the matter long before the Schaefer publication. We obviously do not agree but that's another story.

Anyway, the site seems to be down for me at the moment, so the refs will follow later, in which it is recognized that isotope behavior of ice cores is mainly dependent on three elements: the local temperature during condensation, rayleigh effect and seasonality in precipitation. Several publications go in great length to conclude that the isotopes of Greenland ice cores are indeed mostly temperature. But obviously it is not.

So, let's forget about rayleigh effect and think seasonality, the third factor. The reasons for that:

http://www.geol.lu.se/personal/seb/Geology.pdf.pdf

Bjork et al 2002 Anomalously mild Younger Dryas summer conditions in southern Greenland GSA Geology May 2002

The first late-glacial lake sediments found in Greenland were analyzed with respect to a variety of environmental variables. The analyzed sequence covers the time span between 14 400 and 10 500 calendar yr B.P., and the data imply that the conditions in southernmost Greenland during the Younger Dryas stadial, 12 800–11 550 calendar yr B.P., were characterized by an arid climate with cold winters and mild summers, preceded by humid conditions with cooler summers.

Climate models imply that such an anomaly may be explained by local climatic phenomenon caused by high insolation and Fohn effects. It shows that regional and local variations of Younger Dryas summer conditions in the North Atlantic region may have been larger than previously found from proxy data and modeling experiments.

So there was something fishy already here challenging the interpretation of the ice cores.

BTW the different period like the Younger Dryas can be seen http://home.wanadoo.nl/bijkerk/grootes.GIF

Anyway, see the same pattern as the Schaefer study? The real work is the analysis of lake sediments in South Greenland* the result cool wet summers before the Younger Dryas and mild but dry summers during the "cold" Younger dryas, surely a lot of seasonality change. Moreover, cool and mild are the wrong way around considering the temperature paradigm of the ice core. Hence Bjorck et al had no choice but inventing and modeling yet another ad hoc hypothesis to explain the apparent anomaly instead of assuming that the anomaly was actually normal.

So there was a distinct (summer) precipitation change in the South Greenland lake, how about precipitation changes at the ice core (GISPII)? Have a look at Alley 2000. The red graph is supposed to depict temperatures but is actually some math applied to the isotope ratio and indeed we see a strong correlation between precipitation and isotopes.

Before I go on and there is plenty more to follow, just for the record, I'm only combining publications here and showing that everything has been thought of before, only the wrong alternatives were chosen. There is no math or modeling so this is hardly theory development is it?

Anybody interested in the first place?




* (a lake on Greenland at the end of the ice age? Then surely it could not have been that extreme cold, could it?)

 

[Andre]

The site of Hoffmann is online for me again and http://www.ipsl.jussieu.fr/GLACIO/hoffmann/Texts/jouzelJGR1997.pdf is the publication I intended to refer to.

Unfortunately it does not OCR so please read the abstract. The seasonality of precipitation is discussed in chapter 4.3

 

[Andre]

So perhaps I should explain why seasonal precipitation changes have a major impact on the isotopes of the ice cores. Have a look at http://home.wanadoo.nl/bijkerk/GISP-2-site-15.GIF representing the uppermost part of the ice cores with fresh snow layers and several samples taken per year (x-scale in years AD). You can identify each individual year by a distinct summer winter variation of isotope value. You can also see that the spikes are the most distinct in the youngest years, as with age and accumulation, older layers getting compressed and the ice is getting mixed slowly, which progressive obscures the seasonal differences. At some 10,000 years the seasons are no longer discernable in Greenland, I've read somewhere in a textbook.

Now remember the Bjorck paper, cool wet summers and mild dry summers giving a distinct different precipitation pattern as seen in Alley 2000. So in the cool wet period there is a lot of "warm" summer snow and very little "cold" winter snow. If this mixes then the balance is towards the warm summer snow (weighted average). The other period, it doesn't matter if the summers are mild if there is only very little summer snow compared to the cold winter. The weighted average will make it colder anyway.

http://home.wanadoo.nl/bijkerk/YD-simulator.GIF simulates a part of the fresh ice core as it was 11,600 years simulating the end of the Younger Dryas with a distinct shift in summer snow but not in temperatures (amplitude remains the same). The X-axis is distance (accumulation of snow) and the Y axis d18O. The right graph is the same but after 11,500 years of compression with 1600 meters ice accumulated on top of it. Then only the weighted average remains, which shows an spurious "warm" jump towards the wetter summers which were actually cooler.

Although the researches are aware of this effect (Jouzel et al 1997 in my previous post) nevertheless -based on other assumptions and evidence they decided that the ice cores were about actual temperatures. But majority decisions are not necesarely equal to the truth.

 

[Andre]

Time for the coup de grace.

Now that we could consider the Bolling-Allerod spikes of Greenland no longer representing temperature adequately, we have yet another problem.

Those spikes are around everywhere in the North Atlantic:

http://www.lehigh.edu/~ziy2/pubs/YuGpQPreprint.pdf

publised here: http://sumaris.cbuc.es/cgis/sumari.cgi?issn=07057199&idsumari=A2001N000000V000055 (pp 171-180)

The authors compare the isotope spikes of GISP/GRIP with isotopes of the Ammersee in Germany, a new proxy in Canada (Crawford) and a renowned oceanic proxy, the Cariaco basin (Venezuela), which is also in our Mammoth extinction paper submitted to Quartenary International btw. And it shows that all spikes match with Greenland. But since the real warming preceeded those spikes with more than 2000 years, it appears that none of that is a clear climatal warming signal.

I wonder how long it will take, if at all, that the paleo-climate world realizes that we are in the midst of chapter VI and are approaching chapter VII of Thomas Kuhn's Structure of scientific revolution.

http://www.des.emory.edu/mfp/Kuhn.html

Because this is how science works, progress by failure.

 

[Andre]

So, one of the consequences of the Schaefer study, -warming not tied to the isotope spikes of the Northern Hemisphere- is undermining http://www.ldeo.columbia.edu/edu/dees/V1003/readings/NAS%20Abrupt%20Climate%20book/Exectutive%20Summary.pdf which bases quick and abrupt warming and cooling on the isotope jumps of the Bolling Allerod, Younger Dryas and Preboreal. Only one thing remains firm is quick and abrupt changes in precipitation and seasonality but not that much in temperature.

So no abrupt temperature changes (only gradual) associated with all kind of changes like the CO2 spike, which was the basis of the theorization of bi-stable climate and tipping points and the mainstay of "catastrophical" anthropogenic global warming.

 

[Andre]

Now let's add some more mystery.

Today I got hands on this chapter:

Herbert & Schuffert 2000; Proceedings of the Ocean Drilling Program, Scientific Results, Vol. 165 239 CH16

(sorry no link but if anybody insist I can mail it):

The paper is about a state of the art alkenone sea surface temperature reconstruction (in Celsius) of Ocean Drilling Program "hole" 1002 in the Cariaco Bassin in the Carribean.

I quote:

The Uk′37 (“alkenone”) approach estimates paleotemperatures by
measuring the unsaturation index of di- and triunsaturated C37 ketones
synthesized by several species of haptophyte algae, of which
Emiliania huxleyi and Gephyrocapsa oceanica are believed to be the
most important

So that's clear. Empirical labratory evidence has shown that this method is superior to other sea temperature reconstructions like isotopes or ion ratios (Ca/mg, Ca/Sr etc).

Now remember the "cold" -"hot"- yoyo of the Greenland Ice cores http://home.wanadoo.nl/bijkerk/grootes.GIF .

It could be observed that the world is reversed here. Warm already shortly after the Last Glacial Maximum ~20 ky ago, and during the "warm" Bolling Allerod, the sea surface temperature drops with an unprecedent rate to recover slowly again in the "cold" Younger Dryas and beyond.

Haven't been many people in the world with an explanation for that, yet it fits exactly to a certain pet idea.

 

[Andre]

Of course, research continues and occassionally new papers about the cold and glaciated Younger Dryas appear. A recent one:

Golledge N.R. 2007, An ice cap landsystem for palaeoglaciological reconstructions: characterizing the Younger Dryas in western Scotland Pages
Quartenary Science Reviews Volume 26, Issues 1-2 pp. 213-229 (January 2007)

It's #15 http://www.sciencedirect.com/science?_ob=PublicationURL&_tockey=%23TOC%235923%232007%23999739998%23640851%23FLA%23&_cdi=5923&_pubType=J&_auth=y&_acct=C000050221&_version=1&_urlVersion=0&_userid=10&md5=7d77d3361bbc1fd102c038e3e2ea61ee

...this landsystem tool predicts an extensive Younger Dryas ice cap with a maximum surface elevation of 900m above sea level, implying colder or wetter conditions than previously thought,...

So, after all that evidence, is my assessment about a much warmer Younger Dryas debunked by this? Let's see about the dates in the document. I only found these dates:

...Complete disappearance of the Younger Dryas ice mass by 10.6–10.4 ¹⁴C ka BP is inferred from dated basal organic sediments on Rannoch Moor (Lowe and Walker, 1976).

...The timing of glaciation in the study area is bracketed by organic deposits buried beneath subglacial till at Croftamie, southeast of Loch Lomond, suggesting a maximum glacier extent at 10,560 +/- 160 ¹⁴C ka BP, (Evans and Rose, 2003), and inferred ice-free conditions on Rannoch Moor by 10,660 +/- 240–10,390 +/- 200 ¹⁴C ka BP, (Lowe and Walker, 1976). Although the dates from Rannoch Moor may be 'too old' (Sissons, 1979), the Croftamie dates nonetheless imply that the Younger Dryas ice reached its maximum southerly extent very late in the Stadial, and that subsequent deglaciation was very rapid indeed...

So we have only these few carbon dates to go by. Carbon dating before the times of AMS around 1980 is too unreliable to depend on. So we are left with the boldface dates.

I may have suggested previously that the notion of an extreme cold Younger Dryas is probably caused by the discrepancy between carbon dates and calendar dates, caused by the variability of radiocarbon rates in the atmosphere. Nowadays we can calibrate those dates rather accurately with the INTCAL04 table (Reimer et al, 2004)

hence a "maximum glacier extent at 10,560 +/- 160 ¹⁴C ka BP" gives us

10,400 BP = 12,360 Cal BP
10,720 BP = 12,800 Cal BP

consequently, 10,560 +/- 160 ¹⁴C ka BP calibrates to 12,580 +/- 220 Calendar years before present.

Now, when was the Younger Dryas again?

http://www.ametsoc.org/amsedu/WES/glossary.html

http://highered.mcgraw-hill.com/sites/0072549386/student_view0/chapter9/glossary.html

between approximately 12,700 to 11,500 years Before Present

We see that the sources are a bit ambiguous, especially about the onset of the Younger Dryas, which may have been caused by a layer counting error either in the GRIP ice core (12,700) or the GISP II ice core (12,900). The newest ice core, ftp://ftp.ncdc.noaa.gov/pub/data/paleo/icecore/greenland/summit/ngrip/isotopes/ngrip-d18o-50yr.txt shows highest isotope values at the onset of the drop dated 12,800/12,750 cal BP and the end of the drop to the lowest values at 12,550 Cal BP.

All in all there are more suggestions that the 12,700 Cal BP start of the Younger dryas is more accurate than the 12,900 BP assumption

And our Scottisch glaciation reached the maximum at 12,580 +/- 220 Cal yrs BP, which (very) strongly suggests that the glaciation was not a feature of the Younger Dryas but from the preceding Bolling Allerod.

And this paper is peer reviewed science!