Friday, March 22, 2013

Extinctions in the news

Scientists trying to date the end-Triassic extinction (hat tip to SciTech) say they've used uranium in zircon crystals from the rocks to bracket "massive volcanic eruptions from a large region known as the Central Atlantic Magmatic Province (CAMP)" to be able to say that they happened over a relatively short time scale: 40,000 years. Over a long time, eruptions don't hurt much of anything (except locally), but dump vast amounts of sulfur dioxide in the air in a short amount of time and you can have problems.

Blackburn et al combined different approaches to try to estimate the dating: the Earth's wobble means that sedimentary rock accumulates with a pattern that has a cycle of about 26,000 years. They compared estimates taken from this with the classic U←Pb decay studies of zircon crystals, and came up with an error estimate of 30,000 years for dating the "CAMP" materials. They dated not the main flow, but "feeders" (suitably sandwiched rocks weren't available elsewhere) which are "interbedded with strikingly cyclical lacustrine strata," looking to "compare the time durations between basalt flows estimated by orbitally tuning the sedimentary rocks with differences between the zircon U-Pb dates of the flows.

OK, they found 4 lava flows, spaced at about at, 60, 290, and 650kyear after the extinction event. From their Figure 2 I only count 3; the first and second either don't overlap (their count) or are too close to call (my count). (No, I am not going to try to give you a better picture; I suspect this is pushing fair use pretty hard already.)

It isn't obvious exactly when the extinction event shows up. The definition they use is an aquatic one: when a particular ammonite shows up. Translating that to land-based measurements contributes largely to their 40 kyear uncertainty in comparing the extinction to their measures (not their fault; that's just the way things are in paleontology). But one thing seems pretty clear from their results: "the biologic recovery associated with the TJB was underway even as subsequent CAMP eruptions ... were occurring" later. (TJB=Triassic/Jurassic Boundary)

Pretty impressive chronological work.

I'm a little more suspicious of the dino-killer comet claims. They compare iridium and osmium levels and estimate that the impacter was smaller, and therefore must have been much faster than everyday asteroids--which leaves comets. Except why would comets be enriched in iridium? True, they might collect some from cosmic dust, but I thought they were mostly ices. (The comparisons I've seen so far comparing comet dust and cosmic dust isotope ratios have been for low atomic number. There is a body of literature on comet dust, plenty of puzzles remain for budding cosmicgeologists.)

On the other hand, one estimate for the ratio of volatiles (ice, CO, etc) to silicates etc is about 1:1. I'm not sure where they got the estimate, and am positive it would change with repeated sweeps near the Sun. So if the non-volatiles are half the mass, and those non-volatiles have the same elements concentrations as asteriods, then maybe it works OK after all.

It turned out to be a bit harder than I expected to find some simple numbers on concentrations. From here I read that iridium is O(50) parts per trillion except at the KT boundary where is is O(10) parts per billion (and that it could take 100,000 years for that much to precipitate out, meaning the boundary might be thicker than a thin layer--something I hadn't thought of). Here I read that some chondrite meteorites have iridium concentrations between 300 and 650 parts per billion.

Something about that ratio bothered me. Be generous and let the dino-killer have 1 part per million iridium. A world-wide layer of 10 parts per billion means that the dino-killer contributed 1% of the whole planet's stuff sifting into the future rocks for that era. That seems like kind of a lot. If the thing was an asteroid 10km across, then that's about 500 cubic km, and according to that font of all wisdom erosion is about 75 billion tons/year (they make that 13-40x normal), so about 3 cubic km/year would have been going to the sea back then (wave hands violently). So it actually could account for it if the iridium settled out over about 10,000 years. The numbers seem to be closer to ballpark than my naive impression of them.

And the good old Deccan Trap might have mixed in some extra iridium. I still think a broken asteroid/comet hammering the future Carribean and "popping a pimple" on the other side of the Earth is a satisfying resolution of the volcano/meteor disputes. But maybe Blackburn et al will try their hand at that next and prove the events were too far apart.

3 comments:

Assistant Village Idiot said...

The ideas for circling around the answer are pretty clever. Are most of these well-known, or did they develop them themselves?

Texan99 said...

I read a book a couple of years ago about the Permian Extinction and was surprised to find there was still so much uncertainty and controversy about its cause and timing. Of course it was even longer ago, and therefore more challenging. Now you cast doubt even on the KT event! I knew, of course, that it's not completely settled, but I had filed it under "probably OK."

Like AVI, I'm constantly amazed at the clever ways people find of tentatively reconstructing what happened so long ago, all starting with some guys musing over why there were seashell fossils up in the mountains and revising their concepts of how long it took things like that to happen.

james said...

So am I. I don't know how long they were working on this, but I'd guess years. There were a lot of bits to assemble.

I'm not casting a lot of doubt on KT myself--the pros in the field do enough of that. I just thought of something odd and went looking for answers, which turned out to be a little harder to dig up than I thought (ie. I had to try several different google searches and look beyond the first pages. I shudder to think how long it would take without google.)