RealClimate logo

Technical Note: Sorry for the recent unanticipated down-time, we had to perform some necessary updates. Please let us know if you have any problems.

Younger Dry-as dust?

Filed under: — gavin @ 24 October 2007

The Younger Dryas is so called because it corresponds, in the pollen record from Europe, to the latest (i.e. youngest) appearance of the Dryas octopetala pollen, an alpine flower in regions that are now far from alpine. It marks a clear period towards the end of the last ice age when the warming trend of the deglaciation in Europe particularly was interrupted for a period of about 1300 years before it got going again. There were clear glacier advances during this time and the moraines can be seen very clearly all around Europe and Scandinavia.

The clues to what caused this remarkable, if temporary, turnaround have always lain in assessing its spatial extent, the exact timing and correspondence with other events. Two recent papers have shed some welcome and potentially controversial light on the subject.

To appreciate those papers though, you need a little background. It is clear from the Greenland ice cores that the Younger Dryas was a huge event in that region – 10 to 15ºC cooling at Summit – and this is confirmed by studies of ocean sediments in the North Atlantic which also show large temperature drops (a few degrees) over this period. Particularly clear records of climate impacts are seen off Portugal and as far south as the Cariaco Basin off Venezuela. New evidence from proxy circulation tracers suggest that the North Atlantic overturning decreased significantly during the YD, possibly shutting down completely. This has all lent support to the theory, first suggested over a decade ago, that glacial meltwater interfered with the N. Atl. circulation causing an interruption of the ocean heat flow to the North. This is of course the prototype of the “ocean circulation changes imply a new ice age” meme which has been so hard to get rid of in recent years.

But how far afield did this climate change reach? The event is been clearly seen in sediments off Santa Barbara (California) and in cave records from China, but both of these areas are still in the Northern Hemisphere, and exactly what is recorded (wind speed change and precipitation amounts?) is still a little ambiguous. But what about the south?

The initial results from Antarctic ice cores at first seemed to show something very similar – the long warming through the deglaciation was interrupted by a cold reversal half way along. The relative dating was not very good at that time and it was quite plausible that the two events in Greenland and Antarctica were one and the same. When glacial advances in New Zealand were found to be around the same time, it seemed clear that the YD cooling had extended the entire length of the Atlantic! The only problem was that the favored mechanism, an ocean circulation change, no longer matched the data. Models of these shutdowns found it very hard (actually impossible) to get a cooling in the North and South at the same time. Lots of other ideas were suggested, but none that were really convincing. So scientists went on thinking that it probably was the ocean, but always with a bit of unease about the southern hemisphere results or the models.

Clarity started to emerge when new techniques for lining up the ice cores in Antarctica and Greenland were developed. One technique used the very rapid changes in methane (which could be measured in both poles) to synchronise the chronologies. The thought being that methane changes are well mixed and so large changes in one hemisphere get transmitted very quickly to the other. With this came a big surprise – the Antarctic Cold Reversal started hundreds of years before the Younger Dryas! In fact, Antarctica stopped cooling just as the YD was getting started. This was evidence of a bi-polar see-saw in the ocean – something the models did seem happy to show.

But what about the New Zealand glaciers? How did they fit in? There had been some loosely constrained pollen data that didn’t show much cooling reported in 1999, but the result was still ambiguous. This is where the first of the new papers comes in. In it, Barrows et al show with improved dating that the New Zealand peak glacial advances actually were significantly younger than the YD. These dates seem more solid that the previous estimates and are supported by nearby ocean sediment evidence for a continued warming through the YD.

So now that the southern hemisphere oddities have left the scene, does that mean we now have a full understanding of the event? Not quite. The ocean circulation theory has indeed been strengthened in recent years, but the search for a trigger continues – why did it happen when it did? As always, many ideas have been put forward – a shift in drainage pathways for Lake Agassiz from the Mississippi to the St. Lawrence, a solar trigger or a tropical Pacific trigger and now we have a brand new idea – a cometary impact.

This has been suggested by a large group of researchers who have collectively been working on archaeological sites (Clovis) in North America and who noticed a layer of charcoal at about the same time as the YD at a number of disparate sites. They claim too that within this charcoal there is significant evidence of impact ejecta, and from this they suggest that the trigger for the YD was in fact an extraterrestrial impact. This doesn’t really undermine the ocean mechanism – the comet is hypothesised to have caused significant meltwater to flow into the Atlantic and the ocean circulation changed as would be expected in the standard view. Some suggestions were made at meetings that the direct impact due to dust and smoke forcing from fires actually caused the initial YD cooling, but this doesn’t seem quite as plausible (dust falls out of the air quickly).

The researchers have however tried to link the impact with everything that was previously linked in time to the Younger Dryas – mammoth extinctions, the disappearance of the Clovis culture etc. – but it is very difficult to disentangle a direct consequence of an impact from the indirect consequence of the subsequent climate change. But these ideas are quite intriguing and they made quite a splash when announced in a coordinated session at AGU in the spring.

There are three aspects of this work that will require independent confirmation to determine whether or not this is a viable explanation. Firstly, it should be possible to find the ejecta layer almost anywhere – peat bogs, lake sediment, ice cores etc. – wherever there is material of the right age. If that is indeed found (big if), then the first part of the hypothesis might be confirmed – that there was an impact at this time. The subsequent parts are much harder: depending on where any object landed or was centered, how can one show that it produced the meltwater that presumably caused the ocean circulation change? The source to the ocean of the meltwater (be it the Arctic, St Lawrence or Hudson Bay) has been unclear for many years. Finally, how can you show that the direct effect of the hypothesised comet was responsible for any impacts, rather than the indirect effect of the ocean change? These issues will, I suspect, take a long time to resolve.

There are still some more YD mysteries though. The ocean models might have won the Southern Hemisphere round, but they still have a hard time explaining why it lasted so long, and how the rapid warming (10 or so degrees in the space of a few decades in Greenland) at the end occurred. The fact that similar events occurred all through the glacial period (Dansgaard-Oscheger events) implies that they must be fundamental to the climate system rather than a one off. An impact event doesn’t impact those mysteries at all.

70 Responses to “Younger Dry-as dust?”

  1. 51
    SkyHunter says:

    Hi all,

    RealClimate is an invaluable resource for anyone interested in the truth about the science behind climate change. I would like to thank all the contributors for their efforts.

    On the MSNBC Environment forum someone linked an interview with a Dr. Jaworowski. He claims that since CO2 is soluble in water and that liquid water is present in the glaciers, even at -72C. Also that at 320mb that the CO2 in the ice has been “squeezed out”, therefore the CO2 levels recorded in the ice are lower than when the air was trapped. I pointed out that the ice is thousands of meters thick, therefore the CO2 could not be squeezed out of the glacier. Someone suggested that perhaps it migrates into other layers thereby contaminating the samples.

    This sounds preposterous to me, but I do not know the answer.

    Could someone explain why this is or is not occurring?

  2. 52
    Hank Roberts says:

    > Jaworowski

    Paste that name into the Search box (top of each page), and click on the links you get — the default is to search here at RealClimate.

    Within each linked page, put the name into the ‘Find on this page’ box in your browser.

    The very bottom one is James Annan’s mention of Jaworowski, one of the prominent skeptics who’s refused to actually bet on what he’s saying when challenged to do so.

    See what you think, once you look him up.

  3. 53
  4. 54
    SkyHunter says:

    Thank you ever so much Hank.

    That is very helpful.

  5. 55
    NileQueen says:

    Gavin et al.,
    A fly in the ointment perhaps for cooling Younger Dryas in the Northern hemisphere–this paper finds no YD extreme “cooling” offshore SE Greenland.

    Late Quaternary sedimentary processes and ocean circulation
    changes at the Southeast Greenland margin
    A. Kuijpers a;, S.R. Troelstra b, M.A. Prins b, K. Linthout b,
    A. Akhmetzhanov c, S. Bouryak c, M.F. Bachmann d, S. Lassen a,
    S. Rasmussen e, J.B. Jensen a
    Marine Geology 195, 2003.

    Also, a ‘warmer’ YD was found north of Iceland: Knudsen et al. 2004, Marine Micropaleontology 50, 263-305.

    Svante Bjorck’s paper shows a warming Greenland (perhaps regional)
    in Geology, May 2002; v. 30; no. 5; p. 427-430
    Anomalously mild Younger Dryas summer conditions in southern Greenland
    Svante Björck*,1, Ole Bennike*,2, Peter Rosén*,3, Camilla S. Andresen*,4, Sjoerd Bohncke*,5, Eigil Kaas*,6 and Daniel Conley*,7

    Re: the impact hypothesis, analysis of the black mat they discuss can be found in the 2007 book, Murray Springs, University of Arizona Press. Author C. Vance Haynes. I found it on Amazon.

    The impact is hypothesized to have occurred over the Great Lakes, and they consider Carolina Bays as oblique impact structures associated with the event.

  6. 56
    SteveSadlov says:

    Impacts are thought to occur ongoing, throughout geological history. Why does it appear that some have more of an effect than others? Take for example the late Permian extinction. What are the other factors which made phytoplankton in particular so vulnerable to effects of an impact? Could one of the factors be atmospheric mix? Dr. Doug Erwin has written extensively regarding the Late Permian Extinction and its implications regarding future impact events.

  7. 57
    Matt says:

    Lotta comments here, so if I missed my point above, I apologize.

    Perhaps, if an impact happened at this time, then maybe our orbit puts us in contact with asteroids at a proper timing to cause all the interglacial periods. After all, YD happened at an appropriate time, a glacial peak was due.

    If my reading of my ice core data is correct, then YD looks just like the other interglacial peaks, except it stopped short of a complete tipping point. It happened at the right time, but didn’t make the cut. If we assume an impact of this magnitude, and all other things being set up properly, then why didn’t the impact trigger an interglacial?

    I still hold to my amateur theory, we were doing no till farming, following the retreating ice, and we capture incrementally enough CO2 to stop the interglacial tipping. We did this because we started proto-farming about 12,000 years ago. It doesn’t take much, just a few million apes throwing out seed on the bare flood plains.

  8. 58
    Dan W says:

    Uhm, so I am somewhat confused. At the risk of sounding like a contrarian I wish some folks here would address the possibility presented by André (#46) and Nilequeen (#55) that the YD may not have been a time of significant NH cooling at all. (A mechanism to account for this is presented by Andre: Perhaps much of the Greenland d18O2 anomalies were caused by a seasonal precipitation change?)

    Could a further clue lies in this 10/4/07 paper in Science:

    Mixed-Layer Deepening During Heinrich Events: A Multi-Planktonic Foraminiferal 18O Approach
    Harunur Rashid and Edward A. Boyle

    Or perhaps not…

    Thanks folks.

    [Response: There are hundreds of records showing that the YD was cold - ocean sediment d18Oc, %pachyderma, pollen records, glacial moraine extensions, beetles in the UK, alkenones, Mg/Ca etc. etc. And because it was colder there was also less snow on Greenland. While there is much of interest still to discover about the YD, Andre's pet project is not one of them. - gavin]

  9. 59
    guthrie says:

    #56- Steve Sadlov- I imagine differing outcomes per impact would be due to location of impact (ice cap, mid continent, on weak plate joint with subsequent release of magma, mid-ocean) and energy and anlge of attack of incoming missile. Also time of year, I imagine that an impact in spring would have a slightly different effect on ecosystems than one in winter or autumn.

    On the other hand, I have no idea how to go around measuring these potential causes of variation.

  10. 60
    Eric Swanson says:

    The comet impact idea is an interesting hypothesis, however, I recall that there were Woolly Mammoths on Wrangle Island in the Chukchi Sea of the Arctic Ocean as late as 3,700 BP. Since these survived the YD, it would seem logical to assume that the other mammoths did not die out because of a sudden climate change at the beginning of the YD.

    There was a recent report in SCIENCE which may also be of interest.

    “Southern Hemisphere and Deep-Sea Warming Led Deglacial Atmospheric CO2 Rise and Tropical Warming”, Stott, Timmermann, Thunell, SCIENCE 318, p 435, 19 Oct 2007.

    Their work covers the period from 10,000 to 22,000 calendar yrs. BP, which includes the YD. There appears to be no obvious temperature reversal in their data from the Western Tropical Pacific, as seen in their Figure 3. There does appear to be a plateau in the record for the period, which may also be seen in the record from core MD98-2181, shown in Figure S2 of the supplemental data. Of course, getting the dates correctly synchronized is always a problem.

    E. S.

  11. 61
    Andre says:

    Re 58 Gavin response:

    “There are hundreds of records showing that the YD was cold …- ocean sediment d18Oc, %pachyderma, pollen records, glacial moraine extensions, beetles in the UK, alkenones, Mg/Ca etc…..”

    There are indeed hundreds of records showing those things roughly in the timefram 11,000-10,000 14C years BP. The original Younger Dryas boundaries. However scrutiny of all these records, especially on dating reveals a surprising picture. It were the last Allerod spike and the onset of the Preboreal; the Younger Dryas was innocent.

    [Response: This is completely backwards. The YD is defined as the last cold period in Europe coming out of the LGM. You can see the YD moraines (and date them - exposure ages etc.) all over Europe and it fits with the dates seen in the Greenland ice cores, and the independently dated Chinese speleothems - neither of which rely on 14C dating which has a plateau around this time. - gavin]

  12. 62
    Andre says:

    Re response #61, Yes, there is a radiocarbon platform at the end of the Younger Dryas and 10,020 carbon years yields a range of about 11,400-11,600, but we focus on the beginning of the YD, when there was a delta14C spike resulting in a steeper gradient than normal between carbon and calendar age.

    This makes carbon dating at the beginning a bit more accurate, as the carbon years count about 1.5 times as fast as the calendar years and that’s the main area of interest, considering a hypothical impact at around 12,900 Cal BP.

    A lot of glacier activity is rather accurately dated between 11.0 and 10.7 Ka BP (12.9 – 12.7 Ka cal BP) or earlier, which codates with the last Allerod isotope spike, as it is dated in high resoluting annually counted records (GRIP, NGRIP, Ammersee, Lake Gosziac, etc), except for the GISPII ice core.

    Hence the last Pleistocene glacial advance around 12,900 – 12,700 Cal years BP, usually associated with the onset of the Younger Dryas, is in reality late Allerod.

    Alternately if you would define the Younger Dryas as the last glacial advance, you’d have to assign the dating 12.9 – 12.7 ka Cal BP to that, which is evidently not the same as the Younger Dryas when defined to be the low isotope interval of 12.670 – 11.550 Cal BP.

    Same problem exists in the interval between the LGM and the Bolling spike, between 17.5 Ka and 14.5 Ka which is termed the Mystery Interval by Denton, Broecker and Alley (PAGES 2006), when the isotopes were low and both the biota proxies and the glaciers show warming.

  13. 63
    SteveF says:


    To a certain extent Andre may have a point with regards to glacial activity (in some areas). As I mentioned earlier, the latest Scottish dates (cosmogenic) for Loch Lomond glaciation (generally correlated with the Younger Dryas) suggest that ice built up pre-YD. The ice seems to reach a maximum pretty early during the Younger Dryas, probably around 400-500 years into the period. This isn’t massively surprising given how cold the Younger Dryas was.

    That aside, there are of course abundant, well dated, records showing a cold Younger Dryas. The Krakenes project is a fine example:

    Birks, H.H. et al. (2000) The development of the aquatic ecosystem at Kråkenes Lake, western Norway, during the late glacial and early Holocene – a synthesis. Journal of Palaeolimnology, 23, 91-114.

    Birks, H.H. and Ammann, B. (2000) Two terrestrial records of rapid climatic change during the glacial-Holocene transition (14,000- 9,000 calendar years B.P.) from Europe. PNAS, 97, 1390-1394.

  14. 64
    Chris C says:


    obviously the ice cores are showing us many different things: temperature, wind, precipitation, seasonality (and aridity), how much CO2 we had, etc. I will re-copy the second Richard Alley e-mail I put up in the MS forums:

    “I specifically noted that I was referring in this case to the ice-core data. If you find my book (two mile time machine), it unequivocally discusses that, as does the QSR paper on the Younger Dryas. The isotopic indicators do contain temperature information, and this is demonstrated with high confidence by physical indicators especially relying on thermal fractionation of gases that have nothing to do with the stable isotopic composition of the ice. I am guessing that your correspondent is arguing against cooling in Greenland during the Younger Dryas without having read the primary literature and especially Severinghaus (referenced in the book and the Younger Dryas paper); if so, I strongly recommend that your correspondent actually read what is out there, with an open mind, before making claims that are very easy to invalidate.”

    Denton, G.H., R.B. Alley, G.C. Comer and W.S. Broecker. The role of seasonality in abrupt climate change. Quaternary Science Reviews 24(10-11): 1159-1182.

    Alley, R.B. The Younger Dryas cold interval as viewed from central Greenland. Quaternary Science Reviews, 19, 213-226 (2000).

    I’m not sure if realclimate has an ice core expert, but it would be nice for one to comment to resolve the differences, becuase andre’s links do contain an abundance of information which he obviously believes to very strong evidence against the reliability of the paleothermometer. However, ice-core records are often used due to their high resolution of time (Greenland), multi-proxy nature, and highly confident paleoclimatic reconstruction.

  15. 65
    Chris C says:

    Also looks like the YD was flood-induced from water stored in Lake Aggasiz, though I guess there are other theories, but it was getting colder between 12,900 and ~11,500 years ago (see Broeker 2006 . Also, Alley’s paper “Wally Was Right: Predictive Ability of the North Atlantic “Conveyor Belt” Hypothesis for Abrupt Climate Change” Annual Review of Earth and Planetary Sciences, vol. 35, Issue 1, p.241-272 goes over some detail in these time intervals, though the paper is meant for details on the ocean conveyer belt, and not the YD in particular.

  16. 66
    Andre says:

    Re #63

    Thanks, Steve for pointing out the birks et al studies.

    Two main elements:

    Excelling confirmation of the YD boundary to the GRIP chronology instead of the GISP-II chronology.

    Biota abundance suggest a temp drop of the YD of about two degrees, not the alleged 10-15 degrees of the Greenland Summit, however the local isotope ratio changes of the same records (d18O) are in the order of magnitude of the Greenland ratio shifts, about 3-4 permil, suggesting a much stronger temperature drop. See here

    Biota abundance proxies like chironomids assume a direct relationship with temperature, but the water quality and the pace of the biologic cycle are also playing a role. The high seasonality due too the aridness, with warm summers and cold winters certainly affects both. It’s not clear if the researches have considered this. Therefore the result should be compared with other proxies, like for instance the ratio of cold and warm dwelling biota.

    Hubberten et al 2004 QSR-23, analyse insect assemblages on ratio of cold or warm insect species in the Kolyma area of Northern Siberia and they do not identify a Younger Dryas.

  17. 67
    Andre says:

    re #64 Chris,

    The question may not be: what caused the Younger Dryas? But instead: what stopped the Bolling Allerod? I do recall my original comment on PF that the deuterium excess correlation with d18O of GRIP show identical behavior of the Bolling Allerod and all Dansgaard-Oeschger events. Explaining that a bit here.

    This strongly suggest that if you can explain one, you have explained all, but I can’t imagine some 30 lake drainages or 30 extra terrestrial impacts in the last 100,000 years.

    Also note that the deuterium excess behavior appears to be opposite of the expectation, suggesting cooling SST’s (source areas) during the “warm” high isotope spikes

  18. 68
    SteveF says:


    First of all your strawman; no one thinks that the incredible temperature shift recorded in Greenland is going to be mirrored elsewhere. Secondly, your handwaving “explanation” of the chironomid results is noted. Thirdly, they did compare with other biota; it was a multiproxy study (pollen, cladocera, macrofossils and chironomids).

  19. 69
    Andre says:

    Steve, why did the incredible temperature shift on Greenland only shifted 4 mil on the isotopes and why did the mere 2 degrees shift in the lacustrine proxies also cause around 3-4 mil isotope shift? Would you care to explain or perhaps withdraw the stawman strawman.

    Have you noticed any unsual comment about contradictions on that Alpine multiproxy study? I’ll find it back. Give me a few hours.

    [Response: The temperature changes at the YD termination in Greenland are from nitrogen isotopes in the gas phase (Severinghaus and Brook), not water isotopes. You err in assuming that either isotope-temperature relationships must be universal in space or constant in time. For a lot of cases, that might be a good first assumption, but in many cases they have been shown to vary. Isotopes are physical recorders of climate change - they are their own measure and don't have to follow any simple relationship to anything else (see here for more details). - gavin]

  20. 70
    Andre says:

    Regardless of the validity of the 15N 40Ar method in Greenland, isn’t the long-term temperature gradient of d18O in meteoric water about 0.6 mil/oC (Rozanski et al., 1993)? This is easily verifiable with the GNIP database, showing the same general seasonal gradient on moderate latitudes. So, the ~4 mil jumps during the different glacial termination intervals would represent temperature changes of about 6-7 degrees in the lacustrine records of the different mid latitude lakes in Europe like the Gerzensee (Schwander et al 2000), Ammersee (Von Grafenstein 2002), etc. Yet the biota abundance proxies suggest no more than 2 degrees variation. Wouldn’t the difference exceed normal expectations?

    [Response: All you can get from GNIP is the spatial gradient (changes in T and d18Op at the same time but different places). What you really need for paleo work is the temporal gradient (same place, different times). If the reasons for change through time are not the same as the reasons for change in space, you can get significant variations between them. See Werner et al 2000 for a discussion related to Greenland. - gavin]

    Re #68 Steve,
    Considering the Swiss multiproxy studies, you may like to recheck for instance Brooks 2000 (PPP-159, 261-279) and Lotter et al 2000 (PPP-159, 349-361) to see them struggling with unusual, contradictory responses of several taxa to the inferred temperature changes. If the several proxies don’t add up, the ad hoc hypotheses, are merely just that, severely degrading the main hypothesis of sudden and extreme cooling during the Younger Dryas. You might just as well test the observations of a Bjorck et al 2002 scenario, featuring high aridity, and as a consequence warmer summers and colder winters, aggravated by the approaching northern hemisphere summer insolation maximum.

Switch to our mobile site