Out of the ice age, into the asteroid shower

21,000 years ago, at the peak of the last glacial, ice sheets covered most of Western Europe and North America; by 10,000 years ago, a warming climate had melted them away. Records from ice and sediment cores all agree that the ice’s retreat back to its strongholds in the Arctic Circle was not smooth, but jerky: the Greenland temperature record, calculated from oxygen isotope variations in the GISP2 ice core, indicates that episodes of more rapid warming and cooling were superimposed on the longer-term warming trend. Particularly prominent is a 1,500 year period between about 13,000 and 11,500 years ago, known as the Younger Dryas, when the mean annual temperature abruptly dropped to levels more typical of the Last Glacial Maximum.


There is evidence that the Younger Dryas cooling, and much of the other shorter-term variability seen earlier in the last deglaciation, are related to fluctuations in the strength of the thermohaline circulation, due to sudden additions of meltwater to its North Atlantic source region. But some scientists, such as Richard Firestone of the Lawrence Berkeley National Laboratory, have been wondering if the particular severity of the Younger Dryas cooling might be the result of a more exotic cause: an asteroid or comet exploding above North America 12,900 years ago.

A paper just out in PNAS by Firestone and a cast of dozens outlines the supporting evidence for an impact event at the beginning of the Younger Dryas. They focus on deposits like these, which have been uncovered at archaeological sites around North America (with at least one equivalent in Western Europe):


(source: Supplementary information)
The black horizon, rich in organic carbon, is always the same age – around 12,900 years – and therefore appears to represent a single event which correlates with the onset of the younger Dryas. It also marks a clear break in the archaeological and palaeontological record: below it, you find lots of distinctive stone tools made by the Clovis culture, some of the earliest human inhabitants of North America, and the bones of Pleistocene megafauna, including mammoths and ground sloths. Above it, there are no Clovis artefacts, and the North American megafauna have abruptly gone extinct. The question becomes, then, what produced the black horizon itself? A close look reveals lots of interesting things:

  • Organic carbon was mainly in forms produced in intense fires – charcoal, soot and polycyclic aromatic hydrocarbons.

  • Other forms of carbon sometimes found in the layer included nanodiamonds and fullerenes, which are often associated with impacts. The helium trapped in the buckyball cages had much higher proportions of Helium-3 than you would expect for a terrestrial source.

  • A large number of magnetic microspherules – interpreted to be formed from the molten ejecta of an impact.

  • A noticeable peak in the abundance of iridium, which is much more enriched in extraterrestrial bodies than the Earth’s crust (concentrations were generally below the limit of detection above and below the marker horizon).

Iridium peaks, and spikes of ammonium and nitrate that could have been produced by extensive burning of biomass, are also found in the Greenland ice cores. The authors therefore propose that the black horizon is the result of massive forest fires raging across the whole North American continent, triggered by the fiery impact of a meteorite or comet – or possibly a low altitude explosion like the Tunguska event. Enough soot from the fires, and dust from the impact, would have been thrown into the atmosphere to significantly cool the climate, causing enough environmental stress to wipe out both the Clovis people and the mammoths.
It all seems pretty convincing, which made me wonder how this new hypothesis stacks up against the one I was already familiar with, which attributed the Younger Dryas to a weakening of the thermohaline circulation (THC). So I revisited this paper by Piotrowski et al., which uses neodymium isotopes to estimate how much downwelling water in the North Atlantic was being transported into the deep Southern ocean (see this post for some more details). The plot below – their Figure 4 – compares inferred variations in THC strength with the Greenland ice core record.


I’m struck by the fact that there appears to be a good correlation between the two, with a stronger THC corresponding to warmer temperatures in Greenland and vice versa, except during the Younger Dryas; it is preceded by a weakening of the THC from 13-14 thousand years, but the minimum appears to be reached before the onset of the most rapid cooling at 12.9 thousand years. Perhaps, then, the two mechanisms are complementary, with the chilling effect of an extra-terrestrial impact adding to the temperature drop already caused by a weakening of the thermohaline circulation.

Categories: climate science, geochemistry, geohazards, paper reviews, past worlds, planets, Pleistocene

Comments (17)

  1. SteveF says:

    Part of the issue here could be chronological; radiocarbon isn’t especially precise or accurate during the deglaciation. Therefore reaching conclusions on the precise timing of events on the basis of a marine chronology (compared to GRIP which is probably more accurate) might not be appropriate.
    Roughly speaking, slowdown of the THC coincides with the onset of cooling into the YD and this is suggestive of a link. However, I’m not sure there is the scope to say much more – looking at the fine structure of these climate changes and trying to tie THC fluctuations into this is somewhat risky IMHO. I’d be very cautious about saying that the minimum of THC is reached before the onset of the YD (plus, eyeballing the graph, it doesn’t preceed, it coincides with the start of the YD – the little peak, a bit after 13,000 kyrBP).
    Having said all that, it would be reasonably to hypothesize that maybe THC slowdown and this meteor impact provided a cumulative effect. What would be really nice is a seriously well dated sequence, with evidence for impact as well as good climatic data, that gives us a better chance of linking meteor with the events of the YD.

  2. Hank Roberts says:

    Has anyone modeled what might have happened had such an impact been off by a few hours, over the Pacific or Europe or China?

  3. Thomas Palm says:

    In itself an impact would only cause a short cooling lasting several years. To get an YD-event you need to have longer lasting feedbacks. The paper suggests that the impact may have triggered the conventional explanation for the YD:
    “The largest potential effect would have been impact-related partial destabilization and/or melting of the ice sheet. In the short term, this would have suddenly released meltwater and rafts of icebergs into the North Atlantic and Arctic Oceans, lowering surface-ocean salinity with consequent
    surface cooling.”

  4. Stefan says:

    Hi Chris,
    thank you for this very interesting post!
    I have a question about the temperature scale used in the first figure: Which “mean annual temperature” is shown in this plot – it’s so low -, and did it really increase by about 15 centigrade between the last glacial maximum and the holocene?
    That’s an amazing step – I would never have expected anything that big.
    Best, Stefan

  5. BrianR says:

    Good post, Chris…an interesting topic, indeed.
    SteveF, you say: “radiocarbon isn’t especially precise or accurate during the deglaciation”
    Do you mean radiocarbon dating in general? From certain areas/archives? I am doing some research during this time period (18 ka-present) and haven’t come across a statement like that…what are the papers/studies that suggest the imprecision and/or inaccuracy of 14C dating during this time? Any info appreciated.

  6. Chris Rowan says:

    Stefan – the temperature record is from the top of the Greenland ice-sheet. The temperature changes indicated in less extreme environments are more modest.
    And the correlation issue is of course important – since the neodymium isotope record is from the Southern Ocean there is the problem of potential time-lag, as well. What we really need is a good record off Greenland somewhere…

  7. SteveF says:

    I probably should have phrased things a little better. Essentially what I meant was not that radiocarbon is completely useless during this time frame, just that it has limited utility when dealing with particularly rapid climate change. In the lateglacial, we are talking here about events on the decadal scale, which radiocarbon will really struggle to resolve accurately, both because of it’s own inherent errors and also the errors in calibration (calibration gets worse the further you go back, particularly as you get away from the dendro timescale). This is all accentuated as there were some pretty wacky things going on in the radiocarbon world during the lateglacial.

  8. BrianR says:

    SteveF…no worries…when one is getting close to submitting a paper in which radiocarbon plays a key role (for me, at the millennial scale during the Holocene), one gets a little nervous they missed something huge!
    thanks much for the clarification

  9. Stefan says:

    Chris – thnx for the explanation!

  10. SteveF says:

    Sounds interesting!
    John Lowe once told me that he reckons an decent age model for the lateglacial should really involve 100 dates! Also, I’ve heard along the grapevine that Johannes Van der Plicht doesn’t trust calibration much beyond dendro. Clearly there’s still a lot of work to be done.
    Anyway, a couple of papers you might find interesting (if you haven’t picked them up already) with regards to age model construction. Firstly, a paper dealing with a new Bayesian approach (mainly applied to the lateglacial):
    Blockley, S.P.E. et al. (2004) Bayesian analysis of radiocarbon chronologies: examples from the European Late-glacial. Journal of Quaternary Science, 19, 159-175.
    The oxcal package can be used for this:
    Also, we have a couple of provocative and interesting papers by Richard Telford:
    Telford, R.J. et al. (2004) All age-depth models are wrong: but how badly? Quaternary Science Reviews, 23, 1-5.
    Telford, R.J. et al. (2004) The intercept is a poor estimate of a calibrated radiocarbon age. Holocene, 14, 296-298.

  11. SteveF says:

    Funnily enough, another relevant paper has just popped up in the latest QSR:
    Blockley, S.P.E. et al. (2007) Building and testing age models for radiocarbon dates in Lateglacial and Early Holocene sediments. Quaternary Science Reviews, 26, 1915-1926.

  12. bigcitylib says:

    From what I have read, the fact of the impact is not being questioned as much as its effects. For example, a (dwarf) mammoth on the Aleutians must have survived the event.

  13. bigcitylib says:

    That should read “mammoth population”.

  14. BrianR says:

    SteveF…thanks for the refs.
    Yeah, as a researcher who applies age-dating methods to do geology, I obviously try and keep up with the latest. But, when it comes down to the nitty-gritty, I leave it to the experts. There comes a time when you have to accept the ages, knowing that they’ll be modified/refined down the road (since that’s science, after all), and publish your results. If those of us that apply these things wait for the “answer” from the experts, we would never get anywhere. We’ve been using the Stuiver et al. CALIB program for converting to calibrated calendar years, for better or worse. If (or more likely when) the methods for determining accurate ages changes, then we’ll all have to re-evaluate our interpretations and see how much it mattered. For some studies, little changes make huge differences…for other studies, it might be less significant.

  15. BrianR says:

    on another note…when is SEED gonna stop bragging about how many comments they got on the front page…it’s a little bit annoying (although I am in a surly mood at the moment)

  16. Chris Rowan says:

    After they’ve announced the winner of the competition, I think. Who knows, it may be one of you…

  17. Some ejecta I found that had been tossed around
    was waiting for someone to assess.
    That ‘twas up in the air before it got there.
    What tossed it, I had not a guess.
    Strange, and quite queer, no volcano was near,
    or round crater to account for the mess.
    How rude, and sore cruel; to break every rule,
    then lie in plain sight n’er the less?
    In northern Mexico you will find thousands of square miles of ejecta, breccias, melt basins, and pyroclastic materials in pristine, unweathered condition like they happened yesterday. When you follow the flow of those materials back to their respective sources you find no crater, and no volcanic vents. Only patches of barren, smoothly melted stone. Or strangely shaped denuded mountains with all traces of alluvium blown away. or miles wide basins of blast melted rock that flowed like water. The ejecta is in such good shape that a child could follow it to it’s source. And when you do so you always get to a place where you are pointing straight up. And knowing beyond all shadow of doubt that the energy to melt and blow whole mountains away came from up there. The idea that our atmosphere somehow protects us from impacts is a childs fairy tale. And Tunguska was only rare in that it arrived alone. It was a tiny little thing compared to all the others.
    If you are still looking for round impact structures and craters you aren’t studying earth impacts, or the causes of the YD event, or megafaunal extinctions. You are only chasing butterflies in the playground. Go to Mexico and think “Hyper thermal impact firestorm” and you will begin to get some real science done.