What caused the confusion surrounding the start date of the Younger Dryas?

[Andre]

It may not have gone unnoticed that I've looked at a paper or two concerning the Younger Dryas oscillation

The amount of data and records about large changes in weather patterns is dazzling, confusing and sometimes contradictory, as I demonstrated here

Recalling the enigmatic start date dispute of the Younger Dryas exhaustingly elaborated upon here I did some more checking and I may stumbled upon something that may explain those contradictions a bit more.

I believe that the erratic start date of 12,900 calendar years before present caused a lot of the confusion. But first, how did that error sneaked in? I can't know for sure but I believe it's about a misidentification of volcanic tracers in the GISP2 ice core on the summit of Greenland.

It's a popular thought that the dates in ice cores can be counted by annual layers, due to seasonal differences that remain visible in the ice, however that is only true to a certain depth. Progressing with the overlaying weight the compression of the ice becomes so strong that annual layers can no longer be discerned and dating has to be done with other techniques as decribed here by Svenson et al 2008.

Incidentely notice:

In the 40–60 ka interval, the new time scale has a discrepancy with the Meese-Sowers GISP2 time scale of up to 2.4 ka.

The GISP2 time scale is already a few decennia old and techniques may have been less develloped then. Anyway, one important dating technique is identifying volcanic markers. Volcanic ashes known as tephra are obviously isosynchrone layers, as well as sulphur in precipitation. However you have to identify the right one. Here are ftp://ftp.ncdc.noaa.gov/pub/data/paleo/icecore/greenland/summit/gisp2/chem/volcano.txt, which was used to contribute to the chronology.

Notice there is a huge spike at 12932 years. Now maybe -speculating- that they could have assumed that that must have been the Laacher See eruption in the Eiffel which was the biggest volcanic event in that time frame and which was dated around 12,900 years. However their Younger Dryas started soon after that spike fixing the onset of the Younger Dryas erroneously on 12,900 years. But there may be other explanations.

It was known from tree rings and mollusc scales that this volcanic event happened about 200 years before the onset of the Younger Dryas (Kaiser and Eicher 1987, Kaiser 1993). But google wasn't really active in that time.

Note that the bigger spike of volcanic markers at data points 13033 and 13038 may have been the real Laacher See event, if one would bring that forward 130 years, the GISP2 start of the Younger Dryas would have indeed be much closer to 12,700 calendar years before present. Just a thought though.

Much more to come.

Refs:
Kaiser, K.F., 1993. Beitraege zur Klimageschichte vom Hochglazial bis ins fruehe Holozaen, rekonstruiert mit Jahrringen und Molluskenschalen aus verschiedenen Vereisungsgebieten. Habilitation, Eidgenoessische Forschungsanstalt fuer Wald, Schnee und Landschaft.,
Ziegler Druck- und Verlags-AG, Winterthur, Birmensdorf.

Kaiser, K.F., Eicher, U., 1987. Fossil pollen, molluscs, and stable isotopes in the Daettnau valley Switzerland. Boreas 16, 293}303

 

[Andre]

Again, what's 200 years between friends? But there is a point here:



Ref and caption included. In the Netherlands (only) it's costumary to divide the Younger Dryas (GS1) into two separate biozones, 3a the coldest and very wet, 3b drier and warmer. But they used the erroneous GISP2 timescale here to align the different zones. If we realize that the GISP2 timescale should actually be shifted up about 200 years around the start of the YD, and then we see that the border between zone 2b and 3a gets pretty much lined up with the GI-1b event.

So the question seems valid if the last spike of the Allerod (GI-a) is actually the cold and wet biozone 3a. Interestingly that would dovetail nicely with Bjorck et al 2002 mentioned here.. Cold wet first and warmer dry later. In that case what would be normal and what would be an anomaly?

But that would require rigourous revisiting of all the dating and calibrating. As soon as I get some time I'll try to do that and I'll post the results.

 

[Andre]

Meanwhile let me tell something about carbon dating. It can be a great tool and it can be a tragical mess. Unfortunately we have to deal with the latter. One of the main problems is that we're dealing with variable concentration ratio's of unstable 14C atoms in the CO2 in the air. If you can count years with tree rings, coral rings, etc that you can carbon date, then you can make a calibration table. That has been going on for a few decades, we have seen INTCAL93,98,04 and now INTCAL09 calibration tables. INTCAL04 included the 12,900 error for the YD, INTCAL09 had it removed again. That had some major consequenses, which are seldomly recognised, I dare say, despite what 200 years could be between friends.

Anyway, maybe we should just try a good tool to calibrate carbon dates to get an idea of the problems. Just go http://calib.qub.ac.uk/calib/ to download Calib6 and start it up. First make sure to go under 'output' to 'calibration' to change output Cal AD/BC to Cal BP, (in line with our general life philosophy.)

Then just enter numbers in the sample proporties window like radio-carbon age: 11062 and age uncertainty 11 years. Wow that's pretty certain, but it's the result of numerous datings of macrofossils in the volcanic layer of the Laacher see eruption (Baales et al 2002 (mentioned in the previous thread). Now do 'calibrate' - go and then go 'view' - 'plot' and this is what you get:



On the Y-axis you see your razor sharp normal distribution of the carbon dates. The left down to right up diagonal lines represent the one sigma range of the calibration table and on the X-axis you see the probability distribution of the calibration range result. More than 300 years between two sigma ranges.

Bummer

More later.

 

[Andre]

The most exact possible dating, especially not that far in the geologic past, is if you could count annual layers somehow. To mind come tree rings, coral rings and annual sedimentation layers in lake beds (lacustrine varves), diversified by the seasons. This works out well, but often there is no fixation of the counts (floating chronologies), simply because it starts and ends somewhere in the blind. Often something in the record can be associated with other records, by corrolating variation in annual growth or it can be carbon dated, a chunk of wood in the tree rings or a macro fossil in the varves. After calibration you can fix the chronology and take it from there.

But the previous post shows exactly the problem with that, if you happen to date something on such a carbon dating platform, you introduce large errors. And also if the calibration tables change, you got another problem. So what we actually need is a couple of good continuous chronologies that starts at year zero (1950) all the way to our time frame of interest.

Such a record is the german pine/oak chronology http://www.pages.unibe.ch/products/scientific_foci/ql_dfg/friedrichabstract.html (oops, sciencedirect is down, so an alternate link)

the YD termination is clearly identified in the German pine chronology. Its absolute age of 11,570 BP appears synchronous, within the errors of the respective chronologies, to related signals in the Greenland ice cores (GRIP, GISP2) and in lacustrine varve sequences.

That looks good, an absolute chronology, however Svenson et al 2008 in the OP declare the Younger Dryas - Holocene boundary to be 11,703 +/- 50 b2k, in an attempt to change the standard present (1950) to y2k. So that would be 11,653 calendar years before present. Can we live with the 83 years difference? I don't know and I'm looking further but that's hard with sciencedirect.com down. So more later.

Oh I don't know where this is going to end. A lot of work has to be redone, but if that G1-b spike in GISP2 proves to be identical to biozone 3a in The Netherlands from the second post, then that would have some remarkable consequences.

 

[Andre]

Another annual layer counting tool is lacustrine (lakish) varve counting.

Varves are rhythmic sediments with an annual character, which can be formed in lacustrine and marine environments. The classic varve type consists of a silty, graded summer-layer and a winter-layer of clay and refers to laminated deposits in glacial lakes.

There are problems though, the lakebed must be anoxic or all kind of bottom dwellers (bioturbation) but also earthquakes may disturb the stratigraphy. Therefore there are many floating varve chronologies but very few absolute chronolgies as of year zero.

It was actually the first widely used attempt to date events but it appeared superseded when carbon dating became operational. However in the previous post it may have become clear that carbon dating may have even bigger problems. Especially for instance the Younger Dryas - Holocene boundary which may typically be dated like maybe 10020 +/- 20 radiocarbon year before present. But look what calibration does to that, increasing the one sigma range to 200 years:



That problem made varve counting a bit more popular again. However I have only found one instance with an absolute chronology from zero to well beyond the Younger Dryas Holocene boundary for now. Brauer et al 2001 found this boundary at 11,590 varve years in the Meerfelder Maar and 11,600 years in the Holzmaar.

The problems they talk about did not start before the Allerod, so it seems that the varves seem pretty consistent together with the 11,570 treerings from Friedrich et al, we could tentatively conclude that all absolutely counted records to the Younger Dryas Holocene boundary average 11,587 +/- 15 counted years. But we sure could use some more absolute chronologies and I'm currently looking at the http://www.suigetsu.org/embed.php?File=.

Edit, I observe now that this date is close to the beginning of that "radiocarbon platform" in the graph, tempting to ponder if both events are related.

 

[Andre]

Update on Lake Suigetsu, no dice, the upper part of the sediment is not varved. But looking forward to their chronologies.

Another record that I did not use, but maybe I should is the Cariaco basin grey scale record. The reason for that is that carbon dating oceanic sediments has a nasty complication: basin age -much more about that later-, this may have induced the error that threw the records off between about 12500 and 12900 cal BP which played a key part in INTCAL04. Hence this part was omitted in INTCAL09.

Anyway the graphs in the paper suggest that we have a reasonable estimate of the Younger Dryas to Holocene boundary at 11,587 years, so we can fix -still a bit tentatively- the floating chronologies to that age.

Using Baales et al 2002, linked to above we get:



Notice that the first record, the Meerfelder Maar is an (the only) absolute chronology independent of other datings, that did well in comparison to the others. The second and third series, Soppensee and Lake Gosziac are floating varve chronologies, so they only need a fix at the end of the Younger Dryas. The last two, ice cores of Greenland we are uncertain if these are really independent counted records, for problems mentioned before, moreover we see the use of otherwise dated records, especially carbon dated volcanic markers that may introduce errors.

Therefore I propose to use only record 1-3 to arrive at a Laacher See tephra (LST) date of 12,915 cal years BP, the onset of the Younger Dryas at 12,710 cal years BP and the Younger Dryas to Holocene boundary of 11,587 years. I'll use these data to check out the fitting of the Dutch biozones.

 

[Andre]

Laacher See tephra (LST) date of 12,915 cal years BP, the onset of the Younger Dryas at 12,710 cal years BP and the Younger Dryas to Holocene boundary of 11,587 years. I'll use these data to check out the fitting of the Dutch biozones.

And the result is a definite maybe.



Hoek and Bohnke define the border between biozone 3a and 2b at 10,950 radiocarbon years, obviously that intersects the calibation curves just within one sigma for the Allerod - Younger Dryas boundary, but it's also just outside the one sigma probability of the Laacher See Tephra dating.

Moreover they date the borders with 50 years increments. With these erratic calibration curves that may not be a good idea. It looks like all available carbon dates need to be calibrated first before we can do any statistics.

 

[Andre]

Meanwhile, Van Hoesel et al 2012 were researching the feasibility of the alleged extraterrestrial event. As the Usselo horizon appeared to be possible evidence for that, it became subject of scrutiny again. Interestingly they find that this carbon soot rich layer previously attributed to the Allerod, dates at..

...slightly younger average of 10,845 14C y B.P. (12,760–12,640 cal. y B.P.)

It occurs to me that my proprosal of the start of the Younger Dryas (12,710 cal BP) doesn't really deviate that much from the dead centre of that date.

 

[Andre]

I mailed the authors, to discuss these matters, but they are out of the office, probably to the conference 'Methods of Absolute Chronology'. The irony.