Physics trumps biology

Posted 27 February 2015 by Matt Young

Or, perhaps more precisely, Did dark matter kill the dinosaurs?, which is the way that an article in ScienceNOW put it. Readers of PT doubtless know that there have been a half-dozen or so mass extinctions in the history of the earth, and they appear with a periodicity on the order of 30 million years. You can see an early graph here. The vertical arrows are separated by approximately 30 million years. Not every vertical arrow points to a mass extinction, so it might be better to say that the first harmonic of the data set is 30 million years; that is, if the periodicity is real, it sometimes skips a beat. What is interesting is that some of the extinctions appear to have been caused by collisions with an asteroid, whereas others may be the result of long periods of extreme volcanism – yet all the extinctions occur with the same period of 30 million years. According to the ScienceNOW article by freelance journalist Sid Perkins, scientists have speculated that the mass extinctions occur when the solar system in its orbit around the galactic center crosses the galactic plane. At that time, presumably, the earth is bombarded by comets or asteroids, and these account for at least some of the mass extinctions. It is hard to see, however, how such a bombardment would trigger a long period of volcanism. Michael Rampino of New York University has thrown dark matter into the mix. Good form requires me to confess that we do not know what dark matter is, but there appears to be a lot of it, and it is affected by gravity, if nothing else. Dark matter is thought to be concentrated in the plane of the galaxy. According to Mr. Perkins, if we think of it as a thin disk, then there ought to be an areal density of roughly one solar mass per square light year. Professor Rampino suggests that the dark matter, besides possibly perturbing comet orbits, may well penetrate the earth and heat the core to the point where it rips the crust of the earth or causes a long period of extreme volcanism before the core cools. If the dark matter is unevenly distributed, like ordinary matter, then we would not necessarily expect a mass extinction every time the Earth passes through the plane of the galaxy, and we do not see one. Far-fetched? Maybe, but the hypothesis has at least one thing going for it. Mr. Perkins quotes Dennis Kent of the Lamont-Doherty Earth Observatory to the effect that many large impacts did not cause mass extinctions, which suggests that there is something special about crossing the galactic plane – either the collisions and the dark matter are both required, or the collisions are incidental and dark matter caused the extinctions. Indeed, some scientists are apparently questioning whether the asteroid killed the dinosaurs, or whether it was extreme volcanism instead. As for me, I haven't foggiest idea; as a physicist, though, I rather like the dark-matter hypothesis. But then that is to be expected: according to the link immediately above, paleontologists generally prefer volcanism, physicists prefer asteroid impacts, and geologists are evenly split. Maybe dark matter will bring the physicists into the paleontologists' camp. Acknowledgment. Thanks to Sid Perkins for illuminating (sorry) the distribution of the dark matter.

31 Comments

gdavidson418 · 27 February 2015

How long do they think it takes for the earth to heat up, then to cool? Well, I suppose extreme, fairly uniform heating "like a microwave" could heat things quickly, while requiring a huge amount of energy. But cooling is rather slow regardless, unless some magic way of doing that is also thrown into the grab bag of ad hoc causes. You can speed it up some, but not a lot using "normal processes." It's not a geologically quick on-off process, whereby you raise the core temperatures several hundred degrees, then it cools off just as quickly, geologically. Why has the earth cooled off a good deal over its history, if every 30 million years (or every other, roughly), or so, an enormous amount of energy is injected into the core? 30 or 60 million years would hardly effect any overall cooling of the earth.

I don't think any good evidence of such heating can be seen in the Moon, Mars, or Venus, either. The first two are too small? Really, I doubt it, not if earth's core is heated a couple hundred degrees or more every, say, 60 million years. I'd think both Mars and the Moon would be geologically humming along quite fine if such intense heating were to regularly occur.

It's pretty ad hoc and hard to "test" in any case, until you realize that their cooling rates for the entire earth aren't even close to believable.

Glen Davidson

Matt Young · 27 February 2015

I am not a geophysicist, and I have not read the original paper, but I do not think the earth is assumed to have cooled as quickly -- the volcanism associated with the Deccan Traps, for example, may have lasted tens of thousands of years, according to Wikipedia. The time constant for cooling the earth may be very long, however; I have in the back of my mind that much of the present internal heat is primordial.

Incidentally, I do not think this proposal is a joke, but one commenter to the ScienceNOW article observed, correctly, that the publication date is April 1.

callahanpb · 27 February 2015

It would help to see a graph of the claimed periodicity. There is a world of difference between a mean-time recurrence of 30 million (which could just be a 1/30000000 uniform probability per year) and an actual cycle. It may not even be possible to make such a distinction without a long series of recurrences.

Assuming it is periodic, it should be possible to determine what phase we're in right now and come up with a rough idea of the next time such a disaster will occur. Has anyone done that?

ksplawn · 27 February 2015

A couple of things from a non-physicist that may be completely wrong but come from my understanding of DM.

First, dark matter as we gropingly understand it only interacts gravitationally (and possibly through the weak force, but that would only come into play across vanishingly small distances of 10^-8 meters). This means it natrually resists "clumpiness," since there's very little that can make it lose momentum enough to clump and nothing but the weakest of fundamental forces to even bring it together. It can't shed energy as heat to cool off and slow down. This is why we don't expect to find any "dark matter planets," for example.
As a corollary, this makes the whole "falling to the Earth's core" part of the hypothesis pretty laughable. Dark matter would wiz straight through the Earth without accumulating in any meaningful sense, just like neutrinos do. They wouldn't be able to slow down enough to get trapped in the core because they couldn't interact with the baryonic matter enough to slow down and stick around.

Second, because of the first, dark matter shouldn't really be that much denser in the plane of the galaxy. From what we can tell, dark matter halos are more like spheres than discs. So the idea that there's more dark matter concentrated in the galactic plane doesn't seem all that viable.

Third, if dark matter were densely-packed enough to disrupt the Earth's tectonics and cause volcanic eruptions, we really should have seen evidence of this kind of thing elsewhere.

So all of this sounds super implausible to me.

Mike Elzinga · 27 February 2015

It's an interesting hypothesis, but I remain skeptical.

An area density of one solar mass per square light year within the galactic plane? How does that compare with the distribution of ordinary matter? And dark matter interacts so weakly with matter that we haven't been able to detect it yet with sensitive detectors that have been designed in accordance with current theories about how dark matter might interact with matter other than gravitationally. So the interaction has to be primarily by way of changing gravitational tidal forces as the solar system oscillates back and forth through the galactic plane. By how much could such additional interaction heat the Earth's core? What kind of gravitational tidal forces could it produce as the solar system passed through the galactic plane?

It is not clear whether dark matter in addition to the matter distribution within the galactic plane is what is making the difference. I could conceive of tidal massaging of the solar system by the increased distribution of matter within the galactic plane; but what percentage increase in tidal effects occurs because of dark matter?

There have been some suggestions that there is more dark matter than ordinary matter within galaxies; but how is it distributed relative to ordinary matter? The velocity distribution of ordinary matter as a function of distance from the center of the galaxy suggests that dark matter is more "spherically" distributed than the ordinary matter that lies primarily in a flat plane. I would think such a more spherical distribution of dark matter would be less likely to contribute to increasing the changing tidal forces upon passing through the galactic plane.

John Harshman · 27 February 2015

So far I'm unclear on the mechanism. Why would a concentration of dark matter at the core cause its temperature to rise?

Matt Young · 27 February 2015

It would help to see a graph of the claimed periodicity.

See the link above. Or search for mass extinctions timeline and look at the images. I did not reproduce any images because they were all subject to copyright. Here is another popular article about Prof. Rampino's work; it includes a nice sketch of the solar system oscillating with respect to the plane of the galactic disk. The article also refers to a recent article by Lisa Randall and Matthew Reece. I have read only the abstract, but they apparently hypothesize a dark-matter disk to account for the apparent periodicity of meteorite impacts.

CS · 27 February 2015

This guy appears to be a bit out of his element. The dark matter heating part of the paper seems to rely mainly on 30 year old references and ignores the substantial amount of work that gone into the field since then. I am amazed this paper made it past the refereeing process.

The heating idea is ruled out by experimental searches that have been performed in the meantime at Super-Kamiokande and IceCube, which would see the products of the annihilations that would still be occurring today. He assumes an interaction cross section that is more than 10 orders of magnitude larger than current experimental limits. The timescales for annihilating the dark matter is wrong: being weakly interacting, the dark matter that gets trapped in the Earth still takes millions or billions of years to annihilate. There are also problems with the dense clump argument in the paper as well.

The gravitational perturbations in the Oort cloud is a more interesting idea.

Mike Elzinga · 27 February 2015

Matt Young said:
It would help to see a graph of the claimed periodicity.
See the link above. Or search for mass extinctions timeline and look at the images. I did not reproduce any images because they were all subject to copyright. Here is another popular article about Prof. Rampino's work; it includes a nice sketch of the solar system oscillating with respect to the plane of the galactic disk. The article also refers to a recent article by Lisa Randall and Matthew Reece. I have read only the abstract, but they apparently hypothesize a dark-matter disk to account for the apparent periodicity of meteorite impacts.

The Randall and Reese paper in PRL is far more interesting in that it at least tries to set up a model that can actually be tested with data coming in from the Gaia satellite. This might give some hint about the existence of a slightly enhanced dark matter disk embedded within the baryonic matter in the galactic plane.

Henry J · 27 February 2015

On the bright side, the graph on the website indicated that we're approaching the midpoint between two of these events, which means we have a margin of 10 to 15 million years before we need to worry about it.

https://www.google.com/accounts/o8/id?id=AItOawnoPnMWQfeCANdXlQBv5Z2lEoL0IJ3d54k · 28 February 2015

Considering that we're in the midst of a great mass extinction right now, I think we need to worry about it. But not from dark matter.

Scott F · 28 February 2015

This is pure speculation as I'm no physicist or geologist, but would the Earth's core need to be heated up at all? I mean, we appear to have enough heat today to sustain the mid-Atlantic ridge, other ridges, and plenty of other volcanoes. Rather than "heating", what magnitude of gravitational tidal forces might be sufficient to "crack" or merely weaken the Earth's crust enough to cause large rifts and allow massive vulcanism? If all you have is gravitational interaction, that seems to be a more likely weak link.

bigdakine · 1 March 2015

"Professor Rampino suggests that the dark matter, besides possibly perturbing comet orbits, may well penetrate the earth and heat the core to the point where it rips the crust of the earth or causes a long period of extreme volcanism before the core cools. "

Complete and utter nonsense.

bigdakine · 1 March 2015

CS said: This guy appears to be a bit out of his element. The dark matter heating part of the paper seems to rely mainly on 30 year old references and ignores the substantial amount of work that gone into the field since then. I am amazed this paper made it past the refereeing process. The heating idea is ruled out by experimental searches that have been performed in the meantime at Super-Kamiokande and IceCube, which would see the products of the annihilations that would still be occurring today. He assumes an interaction cross section that is more than 10 orders of magnitude larger than current experimental limits. The timescales for annihilating the dark matter is wrong: being weakly interacting, the dark matter that gets trapped in the Earth still takes millions or billions of years to annihilate. There are also problems with the dense clump argument in the paper as well. The gravitational perturbations in the Oort cloud is a more interesting idea.

Totally agree.

Mark Sturtevant · 1 March 2015

The graph is very old, and it has been refined over the years. The updated version of extinction patterns (such as here) show that mass extinctions are spaced apart, but are they really periodic? Some researchers on the subject do not think so. One can see on the graph in the provided link to Wikipedia that the major and minor mass extinctions (the blue and yellow arrows) have some pretty different spans of time between them. I suppose one could argue that the wider expanses are occasions where a mass extinction was 'skipped', but again, not everyone thinks that the periodic pattern is really all that periodic.
I had thought that essentially random processes can still show stretches that seem periodic. At least I recall reading somewhere that the stock market was an example of that sort of thing.

mail.andrew.kelman · 1 March 2015

Also ignores the likelihood that some of the large extinctions are actually closely spaced smaller events, or that the end ordovician is a cooling event. That many of these large climatic events are tied to particular tectonic mechanisms that have nothing to do with whatever is going on in space. Physicists should know better that to invoke dark matter for palpable events since its hallmark feature is that is is Weakly Interacting matter. It cannot concentrate like baryonic mass and lacking electro- magnetic interaction, cannot generate heat.

Henry J · 1 March 2015

What if the periodic event is something that reduces the odds of fossilization occurring, rather than actually wiping out species in a (relatively) short time frame?

Mike Elzinga · 1 March 2015

mail.andrew.kelman said: Physicists should know better that to invoke dark matter for palpable events since its hallmark feature is that is is Weakly Interacting matter. It cannot concentrate like baryonic mass and lacking electro- magnetic interaction, cannot generate heat.

The PRL paper by Randall and Reese is speculative in that it is proposing an enhancement mechanism that would perturb the solar system as it passes through the galactic plane. As they make clear in their paper, they are well aware of the statistical weakness of such an established period; but they, like most good physics theorists, are thinking ahead to another possible means of detecting dark matter by its gravitational effects rather than by its proposed extremely weak interactions with baryonic matter. The basic idea is that, IF a periodic cycle of perturbing events can be correlated with the passage of the solar system back and forth through the galactic plane, and can be confirmed, then there are potential experiments that can detect the presence of dark matter. If one can account for all - or most - of the baryonic matter in the galactic plane by doing careful surveys, one can then calculate its perturbation effects on the solar system as the solar system passes through the plane. If those perturbations are not sufficient to account for the periodic meteor events - assuming this periodic phenomenon can be establish and correlated with passage through the galactic plane - then presumably one can build a model of how a concentration of dark matter in the galactic plane could account for the additional enhancement needed to explain the period. This is the way good theorists work. They don't just work on immediate issues that are hot; they also look ahead and work on models and propose experiments that will be sitting on the shelf in case they are needed at some future date. Much of this model-building activity involves not only deriving equations; it involves building computer algorithms for current and possible future computers that can do the complex calculations needed in the modeling. In fact, a lot of this activity also pushes the development of future computer architecture. Theorists and experimentalists alike are always looking ahead to see what will be needed for future experiments and calculating capacity. This particular paper is also nudging experimentalists to look for more data to possibly confirm whether or not there is really such a periodic cycle. It stimulates thinking about theoretical and experimental issues. That's why the paper made it into PRL.

Matt Young · 2 March 2015

It cannot concentrate like baryonic mass and lacking electro-magnetic interaction, cannot generate heat.

Good comment, but I do not think that the last sentence is correct -- can't it clump because of gravitational interaction? Also, can't (non-electromagnetic) tidal forces cause heating? I think both answers are "yes."

eric · 2 March 2015

Matt Young said: can't it clump because of gravitational interaction? Also, can't (non-electromagnetic) tidal forces cause heating? I think both answers are "yes."

All AIUI, but... It would not do either at anywhere near the rates or efficiencies of normal matter. Normal matter clumps (and heats up) due to inelastic scattering: things bump and move away but with less kinetic energy than they had to start with. Some of the energy 'gained' by the attraction between particles is converted into other forms such as heat. AFAIK, dark matter has a very low interaction cross-section even with other dark matter: instead of bumping, particles just pass right through each other as well as normal matter. So it would not clump (or convert gravitational attraction into other forms of energy) anywhere nearly as efficiently.

eric · 2 March 2015

I should also point out that we may be reversing some of the historical discussion here. It may be the case that physicists first observed a lack of heating and clumping (in things like the lensing around the bullet galaxy, or in how spiral galaxies retain their spirals) and then calculated DM's interaction cross-section from these observations. So rather than "it has this cross-sectional value, so it doesn't clump," the historical characterization of DM may have been "it doesn't clump, so it must have a low cross-sectional value." I'm not sure which occurred in what order. The whole characterization process is probably iterative between both observation and theory.

John Harshman · 2 March 2015

Matt Young said:
It cannot concentrate like baryonic mass and lacking electro-magnetic interaction, cannot generate heat.
Good comment, but I do not think that the last sentence is correct -- can't it clump because of gravitational interaction? Also, can't (non-electromagnetic) tidal forces cause heating? I think both answers are "yes."

How would pure gravitational interaction cause clumping? Is there something in a multi-body interaction that would do that? Certainly no two-body interaction would. Tidal forces might cause heating, but is that what the paper is proposing?

Mike Elzinga · 2 March 2015

John Harshman said:
Matt Young said:
It cannot concentrate like baryonic mass and lacking electro-magnetic interaction, cannot generate heat.
Good comment, but I do not think that the last sentence is correct -- can't it clump because of gravitational interaction? Also, can't (non-electromagnetic) tidal forces cause heating? I think both answers are "yes."
How would pure gravitational interaction cause clumping? Is there something in a multi-body interaction that would do that? Certainly no two-body interaction would. Tidal forces might cause heating, but is that what the paper is proposing?

The mechanism of dark matter clumping would be by "dissipative cooling" as it exchanges momentum gravitationally with baryonic matter. The primary dissipation would come from electromagnetic radiation by the baryonic matter as it is massaged by gravitational interactions with the dark matter. There may also be a small component of gravitational radiation from both. Then further interactions with baryonic matter subtract momentum and energy from the dark matter and it also begins to clump due to decreasing momentum and energy distributions. Getting this process started relies on small variations in baryonic matter and dark matter density to begin with; but that seems to have been the case since we already see galactic clumping.

Matt Young · 2 March 2015

Mike Elzinga answered the first question with vastly more authority than I could have mustered.

Tidal forces might cause heating, but is that what the paper is proposing?

No, I do not think so, but the point was that you do not need electromagnetic fields to cause heating. If I had to choose between tidal forces and dark matter--matter annihilation, come to think of it, I'd choose tidal forces.

bigdakine · 2 March 2015

Matt Young said:
It cannot concentrate like baryonic mass and lacking electro-magnetic interaction, cannot generate heat.
Good comment, but I do not think that the last sentence is correct -- can't it clump because of gravitational interaction? Also, can't (non-electromagnetic) tidal forces cause heating? I think both answers are "yes."

Anything that would affect the rate of tidal dissipation would also affect length of day. These is no 30 million year periodicity in Length of Day. Heating up the core to an extent that it results in increased volcanism would also affect the geomagnetic field; intensity and reversal frequency. No 30 million year period apparent there either. I don't think Rampino's paper is good for anything except shits and giggles.

ksplawn · 2 March 2015

Mike Elzinga said:
John Harshman said:
Matt Young said:
It cannot concentrate like baryonic mass and lacking electro-magnetic interaction, cannot generate heat.
Good comment, but I do not think that the last sentence is correct -- can't it clump because of gravitational interaction? Also, can't (non-electromagnetic) tidal forces cause heating? I think both answers are "yes."
How would pure gravitational interaction cause clumping? Is there something in a multi-body interaction that would do that? Certainly no two-body interaction would. Tidal forces might cause heating, but is that what the paper is proposing?
The mechanism of dark matter clumping would be by "dissipative cooling" as it exchanges momentum gravitationally with baryonic matter. The primary dissipation would come from electromagnetic radiation by the baryonic matter as it is massaged by gravitational interactions with the dark matter. There may also be a small component of gravitational radiation from both. Then further interactions with baryonic matter subtract momentum and energy from the dark matter and it also begins to clump due to decreasing momentum and energy distributions. Getting this process started relies on small variations in baryonic matter and dark matter density to begin with; but that seems to have been the case since we already see galactic clumping.

We see galactic-scale clumping, but still nowhere near what we see happening with baryonic matter. DM is still in a roughly spherical halo much larger than the visible galaxy and hasn't lost enough momentum to turn into a flattened disk. Despite the claims in the paper there are no significant signs of a DM disk overlapping with the visible galactic plane (older observations suggesting one have been mostly explained away with further observations of baryonic matter to account for the gravity, AFAIK).

Mike Elzinga · 2 March 2015

ksplawn said: We see galactic-scale clumping, but still nowhere near what we see happening with baryonic matter. DM is still in a roughly spherical halo much larger than the visible galaxy and hasn't lost enough momentum to turn into a flattened disk. Despite the claims in the paper there are no significant signs of a DM disk overlapping with the visible galactic plane (older observations suggesting one have been mostly explained away with further observations of baryonic matter to account for the gravity, AFAIK).

Yes; that is the caveat in the Randal-Reese paper. It's a really tough issue experimentally and computationally; with even the best super computers. Just cataloging the major baryonic perturbations is a big shore. Randal and Reese are proposing a somewhat smooth distribution of each in their proposed model. That makes for somewhat easier computer simulations of perturbation effects on the solar system; but the real kicker comes with the all rogue stuff that is very likely out there but hasn't been cataloged or even seen - or recognized - within the evidence we have available to us. Still, I can think of ways that such effects can be pulled out of the data and out of a simulation; but, wow, the computing power required is really daunting.

Mike Elzinga · 2 March 2015

Since my recent Branch Retinal Vein Occlusion in one of my eyes, I seem to be having a lot more problems seeing my typos and correcting them. The problem makes my vision extremely wavy when both eyes are reading, and one eye alone doesn't seem to catch everything. There are several typos in my last comment. Sorry.

At least I am learning about eyes and some really interesting scanning equipment.

outrigger · 3 March 2015

The thing about dark matter is, it rarely interact with normal matter (think neutrinos) or even energy, hence dark. It is said dark matter exerts gravity, which is used to explain the supposedly missing mass in the universe...

Nick Matzke · 3 March 2015

Yeah IIRC even the existence of a 30-my cyclicity in extinctions is mostly disbelieved. Especially as the time-resolution etc. in the data has improved, and the differences in sampling through time have become apparent -- there is even a debate if the "Five Big Mass Extinctions" were really mass extinctions. The only two everyone agrees on are the End-Permian 250 Ma (which might have been a double-tap anyway) and the End-Creteaceous 65.5 Ma

Matt Young · 3 March 2015

... the existence of a 30-my cyclicity in extinctions is mostly disbelieved.

Of course. That's only 11 d.