Saturday, October 25, 2008

Gasping for Paleo Air



Musings of a Beginning

It's funny. Sometimes you find that you have an interest in something when you didn't even know it existed and only just happen to stumble across it happenstance. I always have been fascinated by the paleo world, but during the 1990s I ran off into the purely techno end of my interests and, frankly, mostly within the computer or IT world. This culminated in me returning to the supercomputing world and my job at NERSC.

What do you do once you've attained what is probably the highest point you can with respect to computing? Well, I could branch out and either go into the very smallest computing devices instead of the largest (from HPC to peewee?) or I could go into theoretical bleeding edge CS with quantum computing...which, as it turns out, I am undereducated for at present (to be fixed, math and quantum physics need a little more advancing to work there). So I was left with the situation of what to do...

I ended up participating in the Tri Challenge at SC05 as my something to do. This was a competition where I put together a team that participated in the Bandwidth Challenge, StorCloud Challenge, and Analytics Challenge. We supported science - climate sims and the frequency of hurricanes. It was a blast and we were noted for the science and the innovative set-up we'd done. I looked what to do next as a project and since my paleo interests had been rekindled and especially the start of my passion for mass extinctions and the ecology just prior to the Great Dying, I thought swapping some teraflops for participation in their project and help with mine. That didn't turn out so well, but I learned a few thing along the way.

The Atmosphere of Argument

One of the most important - and contentious! - has been that whenever someone does a simulation of the past atmosphere, they really, really need to make sure that they don't cut'n'paste the current gas mix into their simulation. That mixture ratio really messes with the results. Easy Example: a different amount of carbon dioxide would change the amount of heat retained by the atmosphere which in turn would change the speed of the winds and the speed of evaporation and those together would change the amount of precipitation that would get transported into the continental interior. The gases are equally important, but for different reasons. One of the researchers that I listened to as they presented got eviscerated as they stated they used the modern gas ratios. Since it really impacts the results, it's something that's really important.

Another reason that the gas ratios are so important is that the evolution of life absolutely must have been impacted by the level of oxygen available at different time periods. We know that this was the case in the Carboniferous: giant insects were possible through the massively higher oxygen level than now (somewhere around 30% of the atmosphere). Additionally, it has been hypothesized that the ups and downs of oxygen percentiles have triggered diversification (ups) or innovation (downs) in evolution. This was Ward's central hypothesis in Out of Thin Air.

Taste That Original Stale Paleo Air!

Since its so important, how do you figure out the atmospheric content of the past? After all, you don't have samples you can go pick up! Right? In fact, you can occasionally get samples, but that's not the only way to track the amount of oxygen, carbon dioxide and other gases in the atmosphere. Let's take a moment and explore the ways that are

The first and most obvious way is to measure the ratios directly from samples of the paleo-atmosphere. Wuh?! How do you do that, you ask? There are two ways that we have found that seem to produce repeatable, reliable results. Both are from trapped gas bubbles. The first only works for relatively recent history: say from some point in the Pleistocene to present. That would be trapped gas bubbles in glacial ice from Greenland and Antarctica. The second source of trapped paleo atmospheric samples is in volcanic glass. In both cases, it's possible to check, albeit carefully, the ratios directly. So what about the cases where there are no trapped bubbles?

The evidence for atmospheric content from the Deep Time when not actually encased in various bubbles is only found indirectly. That indirect evidence is in the form of isotope ratios in the sediments. Different isotope measurements and chemical traces have different meanings. Sometimes the ratios reflect the temperature of the region where the sediment was laid down: the ratio of O16 to O18, for example. The ratios of carbon dioxide tell us a lot about the productivity of the ecosystems at the time. The amount oxidized - or unoxidizied - materials in the sediment, especially iron, tell us about the amount of oxygen present. Other chemicals have other meanings. Sulfur compounds are another indicator, frex.

By studying the sulfur compounds, especially, groups of researchers have made claims about the percentile of oxygen in the atmosphere at any given time. Some have claimed that during the Jurassic that the oxygen fraction was as low as six percent - 6%! - and that it would explain why the dinosaurs were the dominant megafauna. The airsack system of the bird-dinosaur clade was superior to anything else out there supposedly: nothing else could scale up large enough to be that large and still not suffer from hypoxia, not even our diaphram based system that we mammals, and possibly therapsids, share. More frequently, I hear the number 11% tossed around for the Jurassic oxygen minimum.

Enter GEOCARBSULF

However, one researcher and his collegues have been modeling the atmopshere extensively for the whole of the Phraenozoic at least as far as the gas fractions are concerned. Dr Berner and his crew's model is called GEOCARBSULF and has produced some interesting results. You can read the abstract here and the paper here (warning, pdf), but a much simplified explanation, perhaps overly so, is that the oxygen levels in the atmosphere traclk well with the rise and fall of the seas, plant cover of the terrestrial environment, and volcanic events such that the weathering of the rock gives a strong indicator of the oxygen level. They do some computer simulation work to try to match up the atmospheric content with their data collection points. Their most recent round of modeling puts a lower limit for the oxygen fraction taking place at the Permian Extinction at around 15% and possibly 12% during the Triassic. The upper limit was 30% during the early Permian(!), not the Carboniferous. This would imply, to me, that the fall of sea levels would expose more rock to be oxidize and, thus, regulate the O2 content of our atmosphere. Even so, there are other researchers that maintain the atmospheric oxygen fraction was far, far lower during the Mesozoic, especially the Jurassic. How do we independently check that if there are arguments using the sulfur data?

Flame Out?

One of the oft cited criticisms of the models for Carboniferous Period's oxygen fraction is that it is too high! After a certain point, even wet vegetation will burn at that oxygen concentration and that would indicate that wild fires were common in that point of the Paleozoic. So, to counter, some scientists went looking for evidence of more common wild fires during the Carboniferous oxygen maximum. Low, and behold, there was a plethora of evidence of burned carbon chunks in the sediments. The fires were very frequent at the time.

Others have then wondered if that couldn't be used as a litmus test for other time periods. Tests have been done with respect to the opposite end of the spectrum of oxygen content. How low can you go? And still have forest fires! In the past, others have tested and reported that it is possible to burn plant material at oxygen fractions as low at 11%. A recent paper attempted to repeat these experiments and found something rather different: it is only possible to get plant material to burn at fractions of 15% at the lowest and really to match the observed pattern in the fossil record that it most be closer 18%.

Why did they get such a higher percentile? It's a little odd, isn't it? The suggestion from the paper itself is that the experimental apparatus of the previous experiments was deficient: the lit materials were tossed in through an open door and allowed to burn under the imperfect oxygen environment. This time, the researchers developed a triggering mechanism that started the fires without ever contiminating the experiment with the modern atmosphere: it was a purely sealed experiment. That would be why they found rather different results.

What does that mean though? It means that so long as there is evidence of a forest fire, or fire in general, in the fossil record, it is patently impossible for there to have had the uber low oxygen level reported. 12%? Not possible if there are Triassic and Jurassic wildfires: that appears to be the case too. That's an exciting result. A very exciting result.

The reason for the excitement is that there has been the idea out there that the oxygen content levels have been one of the great driving forces in evolution. (see Ward above). Other times, it must be admitted, there was some thought that maybe the oxygen level was higher during the Mesozoic to explain the gigantism of the dinosaurs. That has been ruled out. Now its in question whether or not the O2 content was even something that impacted the evolution and dominance of the Mesozoic at all. This paper certainly adds fuel to the, erm, um fire...

A Few Thoughts

(Ward's oxygen vs paleobiotic events graph)

I have to admit that I have been skeptical of the idea that oxygen fraction has dropped that low and the consequences extrapolated there from. If you go to the altitudes where the equivalent partial pressure of oxygen is present today as was in the Mesozoic hypoxic periods, the birds - the extant dinosaurs - are hardly the dominant megafauna present. Wait, you say, where do those conditions exist? Mountains. Where are the large ground forms? Even moderate ones? Or flightless small ones?

Then there's the fact there were six to eight inch insects during the Triassic and, as the partial pressure of oxygen drops, just as inversely when you increased it during the Carboniferous, bug sizes scale down; yet, the Goliath Beetles and other large insects of the present day are hard pressed to get even close to 5 inches in length. Sooo did the isects of back then happen to be more efficient breathers? *uber skeptical look*

I also have to admit that I am unsure whether or not Ward's "retreat from the land" even took place. Yes, we have not seen much in the way of fossils of tetrapods during Romer's Gap. Yet at the same time there are, iirc, a nontrivial amount of fossil footprints from the period, just not bones. I have a feeling, but its not really backed by much, that the Gap is really just a case of a lack of sampling from the sediments where the tetrapods were. We shall see.


3 comments:

Zachary said...

Good stuff. I've always been a touch critical of the oxygen impact on more or less modern fauna during evolutionary history. I also didn't realize there was such a big discrepency between the proposed high and low O2 values!

Will Baird said...

30 to 32% is what I have seen for the high and 6% for the low since the late Paleozoic. I have my doubts.

Actually, I wonder if there is an altitudal limit for lacerta (or other lizards). That'd tell us a lot about what the limits are for critters to have survived, I'd think.

Zach said...

That's a good question. It wouldn't surprise me.