How big was that asteroid? The latest geochemist/geophysicist smackdown

ResearchBlogging.orgGeophysicists estimate the size of asteroids associated with past impact events, such as the one associated with the Cretaceous-Tertiary extinction, by comparing the craters they form to the results of impact models. Geochemists estimate the size of these objects by measuring the amounts of exotic elements such as iridium that have been brought down to Earth with them. Because geophysicists and geochemists also love to disagree with each other (just ask them about mantle convection), it’s no surprise that these different methods disagree. In fact, as Francois Paquay of the University of Hawaii and his colleagues report in their recent Science paper, in the cases of the KT impact and one of several impacts in the late Eocene, they disagree by a factor of three or more. Who’s right?

The idea behind the use of iridium as an indicator of asteroid impacts is simple: iridium is an extremely scarce element in the Earth’s crust, because it is a ‘siderophile’ element that headed for the core with most of the iron when our planet first grew large enough to melt and differentiate due to internal heating. In contrast, most asteroids are fragments of bodies that never differentiated, and therefore still retain high concentrations of iridium. When large extra-terrestrial bodies slam into the Earth and are vaporised in the impact, this iridium is dispersed around the globe and – because the normal background concentrations are so low – clearly shows up in sediments desposited at the time of impact. The large iridium spike associated with the Cretaceous-Tertiary extinction event is the classic example of this.
Other siderophile elements can also be used in this way, and the one used by Francois et al., osmium, throws up an interesting additional twist: osmium is not only present in higher concentrations in an asteroid, but also in markedly different isotopic proportions. The ratio of 187Os to 188Os is much lower in asteroids, which is a consequence of the fact that over time new 187Os is produced by the decay of 187Re (rhenium). On Earth, where most of the original osmium is in the core, this radiogenic 187Os is the dominant osmium isotope, resulting in a very high 187Os/188Os in the crust and ocean water; in asteroids, the presence of much more of the original or primordial osmium (and hence more 188Os) results in a much lower ratio. Therefore, the extra 188Os-enriched osmium which is added to the Earth’s oceans by a large asteroid impact should noticeably change the isotopic composition of osmium laid down in sediments after the impact, producing large negative excursions in the marine 187Os/188Os record. Francois et al. show that such isotopic excursions are indeed present in Late Eocene sediments from two deep ocean cores (left – black dots are data for a core from the equatorial Pacific, red dots are for a core from the Southern Ocean) around the time of a proposed impact event, and also at the KT boundary (right – note that this data is from an earlier study and the sampling points are more widely spaced); these excursions occur at exactly the same point as an iridium peak, strongly suggesting that they are impact related.


One notable difference between the two records is the size of the excursion; the decrease in 187Os/188Os associated with the larger KT impact (from 0.6 to less than 0.2) is greater than the excursion for the Late Eocene impact (from 0.5 to 0.3). This is not unreasonable: the bigger the impacting asteroid, the more 188Os-enriched osmium is added to the ocean, so the larger the isotopic excursion should be. Francois et al. have turned this around, and asked: how big must the impacting asteroids have been to supply the osmium necessary to produce the measured isotopic excursions? Their calculations indicate that the chondrite responsible for the Popigai impact in the Late Eocene was 2.8-3 km in diameter, and the carbonaceous chondrite responsible for the KT Chixulub impact was 4.1-4.4 km in diameter. In contrast, the estimates from impact simulations suggest much larger diameters: 8 km for the Popigai crater, and 15-19 km for the Chixulub crater.
The possibility of the object that many think finished off the ammonites and the dinosaurs being much smaller than we thought is obviously of more than academic importance. At a stroke, our planetary neigbourhood would become a much more hazardous place, as previously innocuous-looking, smaller space rocks suddenly look a whole lot more menacing. Whether you should be worried really depends on which you trust the more: geochemistry, or geophysics? On the geochemistry side, the authors acknowledge that their osmium-inspired size estimate is in fact a minimum, for the simple reason that not all of the osmium in the impactor may have made it into the oceans; you also have to make assumptions about the original make-up of the colliding body. Nonetheless, to reconcile the osmium data with the impact simulations you would have to assume that less than 10% of the material in the impactors is making it into the oceans, which seems a bit of a stretch. Independent estimates of impactor size based on the iridium enrichment are slightly larger (4 km for Popigai, 6 km for Chixulub), but still much smaller than the impact simulations predict. As for the impact simulations, variables such as the angle or velocity of impact can have a large effect on the size and morphology of the crater, but such a large disparity between model and alleged reality is probably somewhat difficult to entirely explain in this way.
Clearly on one, or indeed both, sides of this discrepancy, some important factor is not being accounted for in the calculations. Finding what it is, and establishing whether the geochemists or the geophysicists will be able to claim bragging rights in this particular debate, may take a while, and an argument or seven. In the meantime, regardless of it’s utility in the size estimation game, measuring osmium istopes is clearly another effective method for detecting impacts. I’m also struck by the fact that the 187Os/188Os excursions in the figure above persist over a much wider core interval (representing tens of thousands of years) than the spikes in iridium, presumably as a result of osmium having a longer residence time (it hangs around in ocean water for longer before being incorporated into settling sedimentary particles). This would likely make them easier to detect – unless you’re analysing at ultra-high resolution, I’d imagine that it’s easy to miss the very narrowly distributed iridium peaks associated with unknown impacts. Furthermore, would this tool be of any use in settling the raging controversies about the size and frequency of Holocene impact events? I’m less sure about that one, but it might be worth thinking about.
Paquay, F.S., Ravizza, G.E., Dalai, T.K., Peucker-Ehrenbrink, B. (2008). Determining Chondritic Impactor Size from the Marine Osmium Isotope Record. Science, 320(5873), 214-218. DOI: 10.1126/science.1152860

Categories: geochemistry, geohazards, geology, planets

Comments (4)

  1. Greg Laden says:

    I would think that what really counts with an impact is the total energy, which is a function of mass and velocity. Neighboring rocks that might hit us may be better classified (in reference to how scared we should be of them) on the basis of what their total energy would be if they hit us. Or do I have this totally wrong?
    Great post, by the way.

  2. Maria says:

    Factor of 3? Pfft! That’s hardly a disagreement at all!

  3. Lab Lemming says:

    Maria, that’s a factor of 3 in radius, which becomes a factor of 27 for mass of uncompressed objects (or energy, assuming the same impact v).
    Similarly, their claim that Os isotopes and Ir agree to within 50% in radius means that they have a factor of 3-4 disagreement in mass.
    I can’t access the paper, so I was wondering- how rigorously did they test for homogeneity across ocean basins? I’d expect all PGE’s to preferentially precipitate in anoxic conditions. It’s a shame they didn’t find a non-pge siderophile proxy to compare with.
    Still, it is nice to know that these excursions are detectable.

  4. Chris says:

    Would the type of carbonceous chondrite have a large effect on the amount of Os and Ir? I’ve only done trace-element analysis on CV3- (and CK3)- CAIs (and AOAs), but the enrichment of some trace elements is several orders of magnitude above that of CI chondrites.
    For examples, wouldn’t a CV or CO chondrite with 10-15 weight % of CAIs have a radically different bulk trace element composition compared to a CI chondrite (<0.01 Wt % CAIs). I’ve not got a figure for any enrichments of Os or Ir, do you know if they vary as much as, say, the rare earth elements, from chondrite to chondrite?