The silver bullet that ricocheted
[W]e note that if [elastically accommodated grain-boundary sliding] were as ubiquitous as theory implies, then the interpretation of seismological observations of any hot, solid regions of Earth based on single crystal elasticity would require a significant revision.-Karato et al., 2015
This concluding sentence from a recent paper suggests that a lot of seismological interpretations out there are wrong. Fully understanding what is going on is worthwhile but takes a bit of background. Unfortunately their press release is so tied up in knots that it hides what could be a really significant contribution.
One of the key elements in plate tectonics is, not surprisingly, plates. While the bulk of the mantle convects as a viscous fluid, some of it near the surface cools enough to essentially remain undeformed. This mantle tends to stay attached to the crust above it; it deforms more simply as an elastic material than a viscous one. Together, that uppermost part of the mantle and the crust form the lithosphere. And the lithosphere is basically where the plates are [let us set aside tectosphere arguments for today]. This paper in essence explores the failure of a promising approach to figuring out the thickness of the lithosphere and in so doing might undercut a fair amount of current understanding of the physical state of the shallow mantle.
So, of course, you’d like to know where the boundary between the lithosphere and underlying mantle (usually the asthenosphere) lies. This is super tricky, and the result is that earth scientists have a bevy of definitions of lithosphere which really are not quite the same: mantle that hasn’t mixed with deeper mantle for a long time, mantle transferring heat by conduction, mantle behaving as an elastic material at geologic timescales, and mantle above a seismic low-velocity zone (where seismic wavespeeds decrease with depth). The thing is, none of these are easy to observe. The petrologic/geochemical definition requires obtaining samples of the lithosphere, which only happens occasionally when even deeper seated magma sources carry chunks of the lithosphere up or you can confidently say the melt itself is from the lithosphere. Similarly, estimating the temperature structure from the observations at the surface is tricky: you need to know the thermal conductivity of the rock all the way down and you need the thermal structure to have stayed the same for a long time. Figuring out the deepest part of the earth that will hold shear stresses over geologic time (in essence, figuring out the thickness of the plate from how it has deformed) depends on many parameters. This all tends to leave the heavy lifting to seismology.
Unfortunately the seismic characteristic is a troubling one. You see, when seismic velocities decrease with depth, seismic waves from the surface get bent to even greater depths. At the surface, you see seismic energy decrease with distance from a source (earthquake, blast, etc) until it might fade away altogether; eventually energy returns from even deeper levels. The bottom of the lithosphere is thought to be gradational, so energy doesn’t reflect from it.
The traditional tool seismologists have used is the study of surface waves. The neat things about surface waves is that they can sample low-velocity zones. Unfortunately, they don’t yield unique results: you can move the base of the lithosphere up and down some as you change the velocities of the material above and below the base of the lithosphere. And more recently, attention has focused on how seismic anisotropy affects these interpretations (seismic anisotropy refers to the fact that seismic waves traveling in different directions in a rock can travel at different velocities; you could plop two chunks of mantle one on top of the other and just have their orientations be different and seem to produce a low-velocity zone.
So the development of receiver functions looked like a nice way of attacking this problem. A receiver function is actually extracting a series of seismic arrivals created by boundaries near a seismometer from a seismogram of a distant earthquake [well, usually distant]. As the seismic wave from a distant earthquake encounters a relatively sharp boundary, the interaction not only passes that wave on through but will usually generate some additional waves. In the animation below, a P-wave traveling up will generate a P-wave above the boundary but also an S-wave. The difference in time between the P- and S-wave’s arrivals tells of the depth of the boundary,
Now for the “usual” application looking at arrivals after a P-wave, the lithosphere-asthenosphere boundary gets lost in stuff bouncing around in the crust. But if you look before the arrival of an S-wave, you might see S-to-P conversions (because the converted P-waves travel faster). So you don’t get tangled up in reverberations. Although there are issues in constructing S-to-P receiver functions, this looked like the best tool at finding the base of the lithosphere. And, indeed, early results from the western U.S. produced results that looked like what we might expect from other means. Here was the silver bullet to slay the ambiguity of the base of the lithosphere.
However, when this was applied to the eastern U.S., trouble emerged. The S-to-P conversion that looked like the one in the western U.S. was coming from too shallow a depth; all other means of estimating the thickness of the lithosphere put the bottom no less than 200-250 km down, but this conversion was coming from about 100 km deep. What was going on?
There have been any number of possible explanations, the most popular probably being that there was a pronounced change in anisotropy with depth. This seemed hard to swallow: why would such a change, presumably reflecting the history of the construction of the craton, yield a boundary that could be followed all over the place? Other explanations have included a boundary between water-rich and water-poor layers and issues with seismic attenuation, but these too felt forced.
So this latest paper suggests that what is happening is that a fundamental change in seismic wave propagation is happening at these depths. At shallow depths, crystals are locked together. If you know the properties of the individual crystals, you can combine them appropriately to get the seismic velocity of the rock. But at greater depth, this paper argues, the crystals start to slide a bit when jostled at seismic frequencies. This the authors term “elastically accommodated grain-boundary sliding” (or EAGBS, a term GG thinks unlikely to remain the preferred shorthand for this). The idea is that once the crystals can slide a little, seismic wavespeeds drop: melt or fluids are not needed, just a somewhat higher temperature. Of course, tossing in some fluid can add an extra wrinkle in this too.
The authors note that the physics of this are poorly known in rocks (the literature that is the basis for this proposal is dominantly concerned with materials at much lower pressures), so work would be needed. The lead author, Shun-ichiro Karato, has been a forceful advocate for looking beyond temperature as the sole variable in affecting seismological observations; over the years he has suggested a number of novel possibilities, some of which have stuck, some of which haven’t (at least, not yet). If this is born out, the interpretation of seismic velocities in the upper mantle will have to be revised. It seems likely that the repercussions could be substantial.
And so it comes to this: the magic bullet to solve the asthenosphere-lithosphere problem has basically failed, but in the process it might have wounded the traditional means of inferring physical properties of the mantle from seismic wavespeeds. And so it goes….