and then: volcanoes
May. 27th, 2021 10:24 pm![[personal profile]](https://www.dreamwidth.org/img/silk/identity/user.png){kind=small}
kaberettOkay, so, the other day I got linked to a tumblr post about volcanoes, to which my response was:
... whereupon people went "NO WAIT WHAT SLAB GRAVEYARDS???" and I did a special-interest infodump, transcribed and slightly cleaned up below the cut for your potential amusement.
The general scientific consensus (unless you are one of the ~3 people running mantleplumes.org) is that mantle plumes do exist and are a real thing and represent convective upwelling. So far so good - this is pretty much like the macaroni in the tumblr post.
The thing about mantle plumes (that my PhD is on) is, however, that they're compositionally weird -- they don't look like the stuff mid-ocean ridges churn out, in ways that can't be simply accounted for by "oh we melted more/less of it, so various elements are more/less diluted". (The reason we care about trace elements like thallium, which, again, exists at levels of like maybe 0.000002% by weight in a rock on a good day, is that the variations in how much or how little of it there is helps constrain the problem of "how much of this is due to melt extent, and how much of it is due to the composition of what you're melting?")
So mantle plumes are (1) convective upwellings and (2) compositionally weird, implying that they're made up of material that is NOT what mid-ocean ridges have access to -- there you just get melting because you're pulling solid plates apart, and the solid rock below rises to fill the gap, and it cools down slower than it rises i.e. than the pressure decreases, which is why it melts, in the same way that water boils at a lower temperature at lower atmospheric pressures, rocks melt at lower temperatures at lower pressures.
So the follow-up questions, which I am going to answer in reverse order, are (i) where does the compositionally weird shit COME FROM, and (ii) what actually kicks a plume off?
Regarding (ii), you may or may not ever have been shown the experiment of pouringchampagnefizzy soft drinks into a glass, and observing that rather than forming evenly all the way through the liquid, you get a steady stream of bubbles rising from particular points on the glass? That's because there's a bit of dirt or a slight unevenness in the surface or something at that point, in a way that makes it take less energy to form a bubble (which is hard to make happen in the middle of a liquid with nothing to nucleate on, ~because surface tension).
ALRIGHT we say, so PLUMES, eh, THEY presumably ALSO arise from CONVECTIVE INSTABILITY due to ????????the core-mantle boundary????????some other shit????????????
Anyway, the geophysicists are now getting good enough at taking pictures with earthquakes that we know that (a) some plumes originate from the mantle transition zone (where the pressure hits a point that the mineralogy changes), (b) some originate from the core-mantle boundary, (c) there is WEIRD SHIT going on down at the core-mantle boundary, consisting of a bunch of piles of shit that looks faster and hotter than most everything else ("Large Low-Velocity Shear Provinces", or LLSVPs), and (d) WE CAN SEE SUBDUCTING SLABS, which we KNOW recycle shit into the mantle because [beryllium isotopes, get me to talk about this separately if you want, this is already several lectures compressed into one very short space].
(At this point I diverted briefly into explaining to a biologist that "taking pictures with earthquakes" can usefully be thought of as very slow sonar.)
SO. Slabs subduct! And they do gradually break down and release a bunch of fluids and shit, and that's why the Pacific ring of fire exists, but in the same way that melting happens at mid-ocean ridges because rock is rising more rapidly than it can cool down, at subduction zones rock is descending (because it's old and cold enough that it can no longer keep floating on the top of the mantle) faster than it can warm up. So what slabs don't do is break down all the way, precisely because they're warming up so slowly.
Again, seismic imaging lets us see that some of them founder at the boundary between the upper and lower mantle, because for whatever reason (they've warmed up enough, etc etc etc) that they can't push through into the mineralogically denser lower mantle, and what we THINK might be happening is that place where the slab is foundering in the mid mantle, or reaching the core-mantle boundary at the base of the lower mantle, causes the kind of instability/nucleation point that lets plumes get started, in the same way that the speck of dirt in your champagne flute is what lets the bubbles form.
We call these areas where slabs go to die "slab graveyards", and we think that the weird hot-and-fast lumps (LLSVPs) in the lowermost mantle might be what happens when you do, like, a mass grave over billions of years (okay okay okay about 2.5 billion years).
And this is where we get up to the cutting-edge Everyone's Still Shouting About This, because fundamentally the seismic pictures are still a helluva lot grainier than pictures we get of THE SURFACE OF MARS and even fucking PLUTO, for fuck's sake, so we haven't actually got the resolution to say for sure whether plumes are actually nucleating at the edges of places where we think dead slabs are, and there's a lot of Stats Yelling happening over in the Maths Fandom subfield [see Davies et al. (2015)].
But -- if that IS what's happening -- this could help to explain the compositional weirdness of intraplate volcanism, in that plumes could be dragging up some of the subducted slab material with them. This material has all been processed to HELL and back during subduction, and basically my thesis is "can we use, specifically, thallium to trace recycling of surface materials to the deep mantle, and then return to the surface via plume-related lavas?"
And, well, the answer turns out to be "no, because basically all of it fucks off out of the slab during shallow subduction processing at arcs, and basically none of it even makes it as far as the mantle transition zone".
kaberett: oh goodness okay so I need to clarify
kaberett: that that post is wrong about the physics
kaberett: and rising plumes are not (for the most part) liquid, and nor is the mantle
kaberett: but that aside
kaberett: (whereas ~half the core IS actually liquid, which is slightly different to the core being "semifluid")
kaberett: (and also that's not the whole story of our current theory of How Plumes Initiate, which involves "slab graveyards", but I will stop there for now ;) )
... whereupon people went "NO WAIT WHAT SLAB GRAVEYARDS???" and I did a special-interest infodump, transcribed and slightly cleaned up below the cut for your potential amusement.
The general scientific consensus (unless you are one of the ~3 people running mantleplumes.org) is that mantle plumes do exist and are a real thing and represent convective upwelling. So far so good - this is pretty much like the macaroni in the tumblr post.
The thing about mantle plumes (that my PhD is on) is, however, that they're compositionally weird -- they don't look like the stuff mid-ocean ridges churn out, in ways that can't be simply accounted for by "oh we melted more/less of it, so various elements are more/less diluted". (The reason we care about trace elements like thallium, which, again, exists at levels of like maybe 0.000002% by weight in a rock on a good day, is that the variations in how much or how little of it there is helps constrain the problem of "how much of this is due to melt extent, and how much of it is due to the composition of what you're melting?")
So mantle plumes are (1) convective upwellings and (2) compositionally weird, implying that they're made up of material that is NOT what mid-ocean ridges have access to -- there you just get melting because you're pulling solid plates apart, and the solid rock below rises to fill the gap, and it cools down slower than it rises i.e. than the pressure decreases, which is why it melts, in the same way that water boils at a lower temperature at lower atmospheric pressures, rocks melt at lower temperatures at lower pressures.
So the follow-up questions, which I am going to answer in reverse order, are (i) where does the compositionally weird shit COME FROM, and (ii) what actually kicks a plume off?
Regarding (ii), you may or may not ever have been shown the experiment of pouring
ALRIGHT we say, so PLUMES, eh, THEY presumably ALSO arise from CONVECTIVE INSTABILITY due to ????????the core-mantle boundary????????some other shit????????????
Anyway, the geophysicists are now getting good enough at taking pictures with earthquakes that we know that (a) some plumes originate from the mantle transition zone (where the pressure hits a point that the mineralogy changes), (b) some originate from the core-mantle boundary, (c) there is WEIRD SHIT going on down at the core-mantle boundary, consisting of a bunch of piles of shit that looks faster and hotter than most everything else ("Large Low-Velocity Shear Provinces", or LLSVPs), and (d) WE CAN SEE SUBDUCTING SLABS, which we KNOW recycle shit into the mantle because [beryllium isotopes, get me to talk about this separately if you want, this is already several lectures compressed into one very short space].
(At this point I diverted briefly into explaining to a biologist that "taking pictures with earthquakes" can usefully be thought of as very slow sonar.)
SO. Slabs subduct! And they do gradually break down and release a bunch of fluids and shit, and that's why the Pacific ring of fire exists, but in the same way that melting happens at mid-ocean ridges because rock is rising more rapidly than it can cool down, at subduction zones rock is descending (because it's old and cold enough that it can no longer keep floating on the top of the mantle) faster than it can warm up. So what slabs don't do is break down all the way, precisely because they're warming up so slowly.
Again, seismic imaging lets us see that some of them founder at the boundary between the upper and lower mantle, because for whatever reason (they've warmed up enough, etc etc etc) that they can't push through into the mineralogically denser lower mantle, and what we THINK might be happening is that place where the slab is foundering in the mid mantle, or reaching the core-mantle boundary at the base of the lower mantle, causes the kind of instability/nucleation point that lets plumes get started, in the same way that the speck of dirt in your champagne flute is what lets the bubbles form.
We call these areas where slabs go to die "slab graveyards", and we think that the weird hot-and-fast lumps (LLSVPs) in the lowermost mantle might be what happens when you do, like, a mass grave over billions of years (okay okay okay about 2.5 billion years).
And this is where we get up to the cutting-edge Everyone's Still Shouting About This, because fundamentally the seismic pictures are still a helluva lot grainier than pictures we get of THE SURFACE OF MARS and even fucking PLUTO, for fuck's sake, so we haven't actually got the resolution to say for sure whether plumes are actually nucleating at the edges of places where we think dead slabs are, and there's a lot of Stats Yelling happening over in the Maths Fandom subfield [see Davies et al. (2015)].
But -- if that IS what's happening -- this could help to explain the compositional weirdness of intraplate volcanism, in that plumes could be dragging up some of the subducted slab material with them. This material has all been processed to HELL and back during subduction, and basically my thesis is "can we use, specifically, thallium to trace recycling of surface materials to the deep mantle, and then return to the surface via plume-related lavas?"
And, well, the answer turns out to be "no, because basically all of it fucks off out of the slab during shallow subduction processing at arcs, and basically none of it even makes it as far as the mantle transition zone".