Tony Watts on Seamounts and the Strength of the Lithosphere
Transcript
Note, transcripts are not fully edited for grammar or spelling.
Oliver Strimpel
This is Geology Bites with Oliver Strimpel. When plate tectonics was adopted in the 1960s and early ‘70s, researchers quickly mapped out plate movements. It seemed the plates moved as rigid caps about a pole on the Earth's surface, but since then a lot of evidence has accumulated suggesting that plates are not, in fact, totally rigid. In fact, we can see them flex in response to stresses that are imposed on them. Such stresses can arise on plate boundaries, such as when two plates collide and one plate flexes down to subduct under the other. For example, we see a flexural bulge in northern India where the Indian plate bends down under the Eurasian plate. Similar bulges are seen at subduction zones, where the oceanic lithosphere flexes up before it bends down into a trench, such as off the eastern coast of Japan. Stresses can also be imposed in plate interiors when the plate is subjected to a load, such as a volcano or a sedimentary basin. So, to better understand how plate tectonics works, we need to go beyond the kinematics of plates and study their mechanics as well. Our guest today has pioneered an ingenious method of determining a key mechanical property of a plate, namely its flexural strength. The method uses the flexural sag of plates in response to the weight of seamounts, most of which were emplaced on their surfaces by mid-ocean eruptions. Tony Watts is professor of marine geology and geophysics at the University of Oxford. Tony Watts, welcome to Geology Bites.
Tony Watts
Thank you, Oliver. It's a pleasure to be here.
Oliver Strimpel
To get these deformations into perspective, how deformable are plates, in fact?
Tony Watts
Plates are highly deformed at their boundaries, especially where plates are coming together or converging. For example, in continent-continent collisions. But in the plate interiors in the region of large loads such as volcanoes, river deltas, thrusts, and folds, we see that they bend or flex over large distances, sometimes up to hundreds of kilometres, and by large amounts up to several kilometers.
Oliver Strimpel
Why did you pick seamounts and the flexure they cause as your model system, if you like, for determining plate strength.
Tony Watts
One problem at trenches was that the loads there that were flexing the plate were not directly visible to us. Seamount loads, in contrast, were clearly visible to us in sonar data. We could see how they rose up above the regional depths of the seafloor and how they were growing not only up, but also outwards, sometimes by up to 4 kilometres above the regional depth. They were also quite concentrated loads on the plates. They weren't distributed over a wide area, and, importantly, they had formed quite quickly, making them ideal studies for flexure modelling.
Oliver Strimpel
Did you select a particular seamount to study, or a chain of seamounts?
Tony Watts
We initially chose the Hawaiian Emperor Seamount chain in the Central Pacific because the volcanoes there make up a chain that progressively increases in its age away from the active Big Island of Hawaii. Yet the oceanic lithosphere beneath them is more or less the same. What this meant was that the young end of the chain, the Hawaiian end, had formed on relatively old sea floor, far from a mid-ocean ridge, while the older end of the chain, around the Emperor seamounts, had formed on much younger sea floor, on or very close to a transform fault or a mid- ocean ridge. We can obtain the age at the time of loading by taking the age of the sea floor and subtracting from it the age of the seamount, and these different tectonic settings for the seamounts in the chain were ideal and crucial to telling us something about the thermal mechanical behavior of the oceanic lithosphere.
Oliver Strimpel
Can you describe what these Hawaiian seamounts, and presumably the submerged ones as well, what they're like and how they evolve after they’re formed?
Tony Watts
It is a classic example of volcanoes formed by migration of a plate over a deep mantle hotspot. The chain itself comprises ocean islands, seamounts including atolls and guyots, both of which are former islands that have subsided. Atolls have managed to keep up to sea level because of carbonate reefs that, despite the subsidence, have been able to be productive enough to keep at or near sea level forming an island. But guyots, those same carbonate reefs have given up trying to keep up with the subsidence and so have sunk below sea level in some cases by 1 1/2 to 2 kilometres below present-day sea level. All the islands and seamounts in the chain are prone to collapse through large-scale sector sliding of blocks and slumps and landsliding, which creates debris avalanches that in some cases have traveled several 10s of kilometres away from the islands. So they're in a state of collapse, despite the high magma-generation rates that we see at the young end of the chain.
Oliver Strimpel
OK, let's now get to how you actually managed to use seamounts to work out how strong the lithosphere is. I presume the basic idea is that the seamount represents a load on the sea floor, and how much the sea floor sags in response to that load gives you a handle on its strength. To start off with, you need to know the mass of your seamount load. How do you estimate that?
Tony Watts
To estimate the mass, we need the volume and the density. We can estimate the volume from echo-sounder data and, importantly, what is called multi-beam or swath bathymetry sonar measurements of the shape of the sea floor in the vicinity of the seamount, and density we can estimate either from drilled or dredged rock samples from around the seamount, or from calculations from its seismic wave velocity. So the two things together, the volume and the density, we require in order to calculate the mass.
Oliver Strimpel
The next thing you need to find out is just how much it's flexing in response. So how do you measure the flexure of the lithosphere?
Tony Watts
We initially used Earth's gravitational field because it is sensitive both to the mass excess of the load and to the mass deficiency of the low-density crust that has been forced by flexuring downwards into the denser mantle. But since the early 1980s, seismic techniques have been used in which large air-gun arrays generate sound waves which travel down into the underlying crust and mantle and are reflected or refracted back to be recorded in long, towed hydrophone cables behind a ship. And seismic reflection profiling has been particularly useful in imaging the surfaces of flexure, which include both the top of the oceanic crust and the Moho, the base of the crust, which have both been flexed downwards by the weight of the seamount loads.
Oliver Strimpel
Why can't you just simply use bathymetry to measure the flexure?
Tony Watts
The reason we can't always rely on the bathymetry is that, unlike trenches where the sediment distribution in the sediment in the trench is quite thin and kept thin by corrosive bottom countercurrents, around seamounts, because they are in a state of collapse, the equivalent moat area is often infilled by sediment, and so the top of the sediment does not give us what we want, which is the depression of the oceanic crust by the weight of the load. There are certainly some seamounts, like on the north side of Oahu and the Big Island of Hawaii, where the moat is underfilled, so we get a glimpse of the flexure from the bathymetry, but it's only a glimpse. It's not the full depression due to the weight of the load.
Oliver Strimpel
Oh, I see. So you really need the seismic waves to penetrate that confounding layer of sediment that's obscuring the view of the actual oceanic crust below.
Tony Watts
Yes, exactly.
Oliver Strimpel
So, naively, I'd have thought that we're done, that you've got the mass of the load and you've got the flexure, and that should be able to tell you what the strength of the plate is. Is something missing from that?
Tony Watts
In many respects, it is a bit like a simple experiment you would do in a physics lab. You are, say, determining the Young’s modulus from measurements of the sag of a beam that you had supported at its edges and was bending in response to that load in air. The difference when we apply it to the Earth's tectonic plates is that we are bending a plate, but it's being bent on a weak foundation, something that we assume is flowing on the timescales of geological processes of emplacement, for example, of volcanoes. So, essentially, there are two main features that are supporting the weight of these volcanoes. One is the intrinsic strength of the plate, but the other is the buoyancy from the weak material that underlies the plate. So that is the missing element, if you like, in considering what is supporting a load that we didn't have in that physics experiment.
Oliver Strimpel
But when you're actually determining the strength of the plate, you're actually estimating the thickness of a plate that deforms elastically to support the load placed on top of it. So, is that the whole lithospheric plate that's behaving elastically?
Tony Watts
It gives us the thickness of the layer that is deforming elastically in response to the load. That layer is somewhere within the lithosphere, not necessarily at the surface of the lithosphere, and it's a mechanical concept. There may not be a physical boundary at the base of the elastic layer. What we have found is that in the oceanic lithosphere, elastic thicknesses are in the range of 5 to 40 kilometres. And, importantly, we have found that there's a relationship between the elastic thickness and age with the smallest values at seamounts that were formed on or near a mid-ocean ridge and the largest values formed at seamounts that formed well away from a mid-ocean ridge. The elastic thickness is much thinner than what we refer to as the seismological thickness of the plate, which is in the hundreds of kilometres. The seismic thickness reflects the response to loads of very short duration, such as those associated with seismic body waves or surface waves. While the elastic thickness reflects the thickness of the plate that is responding elastically on long time scales, time scales of interest to geological processes, such as spreading at the mid-ocean ridge and expansion or closing of ocean basins. Loading must therefore induce some sort of viscoelastic relaxation in a plate as it thins from its initial thick short-term, maybe seismic thickness to its long-term elastic thickness. And so stresses must migrate up from the hotter, lower part of the plate into the cold upper part of the plate as it undergoes this relaxation.
Oliver Strimpel
So I've got a kind of sandwich picture in my head then, with the bottom part of the lithosphere behaving viscoelastically as you said, which makes it effectively ductile on those timescales, and then above the layer, I suppose you have a brittle portion which is maybe faulted but doesn't really give you much strength on these time scales that we're talking about. So that's still a huge range from 5 kilometers to 40 kilometers, but we've got to bear in mind, as you said, that the overall standard thickness of our lithospheric plate that we're thinking of is in excess of 100 kilometers, so it's still only, well, less than half of it.
Tony Watts
Yes, that's correct. And this process of viscoelastic relaxation is essentially a weakening of the plate. But what we observe is that the plate increases its strength with age. In particular, the elastic thickness is given pretty much by the depth of the 450-degree isotherm. So we can explain that long-term elastic thickness by the fact that as a plate is created at the mid-ocean ridge, it cools, increases its density, and so subsides, and in the process it becomes stronger in the way that it resists loading. And the important thing is that the relaxation we spoke about earlier is very rapid initially. It turns out that that relaxation then slows as we get to the age, for example, it takes to form a seamount, which is about two million years or so. By that time, there is still a relaxation, but not enough to stop the strengthening with age. And this is very important, because it means that we can regard the elastic thickness as being, if you like, frozen in at the time of the seamount formation, even though it wasn't actually this great seismic thickness, it is a representative of the elastic thickness on long time scales, and this is the one that's important geologically. So it means that the plate, despite viscoelastic relaxation, which is kind of trying to weaken it, is actually strong enough and has a memory of that thermal mechanical setting at the time when it was first formed. And the reason this is important is that we can detect through the gravity anomaly and through seismic observations whether a seamount has been formed in a ridge-like setting, versus one that formed well away from a ridge. So the memory is a good thing. It enables us to track down the tectonic setting of an individual bathymetric feature
Oliver Strimpel
God, that's fascinating. So, effectively, the freezing in of the flexural response reflecting the strength of the lithosphere at the time, or for the 2,000,000-year duration, if you like, when the seamount was emplaced, gives us an insight as to where it was relative to the local spreading ridge at the time it was formed. But subsequent to that, you said that there are two things happening. Firstly, over long time scales, viscoelasticity causes this strain to migrate up if you like, so it becomes weaker. But at the same time, the plate’s progressively cooling and getting thicker, so it's getting stronger. So how can we disentangle these two effects that work against each other?
Tony Watts
Disentangling the two processes is quite difficult. We know that strengthening due to thermal cooling generally wins out, and this is good because it means that very large old volcanoes formed on or near a mid-ocean ridge, such as the Emperor Seamounts, can, despite their size and age, be supported by the strength of a plate. The actual path that that relaxation has taken, unfortunately, is unknown at the present time. We know it happens. We suffer because we don't have an intermediate age like, for example, an ice sheet that could interrogate the lithosphere on timescales of, say, a few thousands of years, like 10,000 years that we have in the continents, and that would be extremely useful if we had that. We might be able to test the relaxation in the subsidence and uplift history of ocean islands and seamounts. But this requires deep drilling, and, unfortunately, we have far too few sites where we have drilled more than a kilometre into an ocean island or a seamount, and we therefore lack a complete rheological model that describes the behavior of the lithosphere on both-short term and long-term.
Oliver Strimpel
In an earlier podcast, David Bercovici explained a mechanism at the small scale of individual mineral grains that could weaken oceanic lithosphere, perhaps facilitating the initiation of subduction. Does your study of lithospheric strength using seamounts shed any light on whether such a process occurs in the Pacific? Or do you see other mechanisms at work that might make plates subduct?
Tony Watts
This is one of the big questions of our time, is what causes an oceanic plate to subduct and go down at a trench. What we see from flexure studies is that at some trenches, the oceanic plate is arriving as strong material. However, subsequent studies have shown that locally in the very high curvature area of the bending of the plate into the trench that, irrespective of the age of the approaching oceanic lithosphere, there is a significant weakening of the plate through faulting in its upper surface, which has been measured seismically through seismic reflection data. So this means that there is potentially a weak zone within that approaching plate that could be enough to create additional curvature and send the plate downwards. So I think the flexure studies are of interest. I'm not sure that we've got entirely the reasons why they go down at a trench. It's a little easier to understand where there are other areas of high curvature. For example, in continental margins, where there are very large accumulations of sediment. For example, of the margins of the Central Atlantic, where sediment thicknesses reach 15 kilometers or so, we see the crust being bent down through very large curvatures, and it's quite easy mechanically to envisage that that is a large enough system of loading that will eventually force the crust to go under another piece of crust and essentially lead to the closing of an ocean through subduction and suductioninitiation under regions of very high sediment thickness. But I don't think we have the full answer at subduction initiation in the trench setting, which David and others have been trying to pursue.
Oliver Strimpel
I'm curious about how you go about acquiring the bathymetric seismic and gravity data that you need for your study.
Tony Watts
We acquire data on board research ships. The oceans have been, and still are, a challenging environment to make sensitive geophysical measurements, and the data that's acquired on a ship requires considerable post-cruise processing to remove water-motion-related artifacts such as what we call in sonar, dropouts, and in seismic, ringing. Gravity is a particular problem because of the very large horizontal and vertical accelerations that the ship experiences due to wave motion. These accelerations are well in excess of the gravitational acceleration that we're trying to measure. So, yes, we can acquire the data, but it's very challenging, especially in certain areas of the southern oceans and the North Pacific, where we have quite unpredictable weather patterns.
Oliver Strimpel
When we spoke earlier, I was really impressed that some of the sonar hydrophone arrays towed behind the research vessels are as long as 15 or even 25 kilometers, which gives you the ability to pick up seismic waves refracted and reflected from the base of the lithosphere. But this creates major logistical challenges, such as keeping out of the way of other ships and dealing with large quantities of plastic and fishing gear snagged by the streamers. And you also have to suspend operations when marine mammals approach.
Tony Watts
Yes, that's correct. Yes.
Oliver Strimpel
I can imagine it's a very powerful thing to know what the strength of the oceanic lithosphere is. What are the wider implications of this result?
Tony Watts
Yes, I think what we've learned from studying the oceanic lithosphere has been very helpful in understanding the strength of continental lithosphere and the role that the elastic thickness plays in understanding how mountains are supported, how sedimentary basins are created, and how it is that they have different stratigraphic architectures. For example, rift-type basins in continental margins essentially form at or near a mid-ocean ridge when the ocean is very narrow, and so it's likely that flexure will contribute to the stratigraphic patterns that occur within sedimentary basins. Other examples where oceanic flexure studies have been useful have been to understand the contributions from dynamic aspects of forming topography on earth surfaces associated, for example, with mantle convection. Flexure studies have also been important for landscape evolution models. We can demonstrate in some places that flexure controls drainage patterns, and to studies of other planets. There have been a number of examples where the results from oceanic lithosphere have been used to understand deformation of the lithosphere on Mars and Venus and other terrestrial planets.
Oliver Strimpel
If you had access to unlimited funding for your research, how would you use it?
Tony Watts
At present, we have only surveyed using multibeam bathymetry about 15% of the ocean floor. So I would map the ocean floor in its entirety to find all those seamounts that we know are out there. If there were funds still remaining after this enterprise, I would use them to support deep ocean drilling, preferably on the flexural bulges around ocean islands and seamounts. Possibly the flexural bulge sea would have a trench because we know that's where mantle is shallowest because the plate has bulged upwards in order to attain pristine samples of oceanic crust and mantle. And I would carry out more seismic experiments using ships equipped with large source arrays and very long streamers to image deep into the lithosphere, below the crust. And, if possible, the boundary between the base of the plates and the asthenosphere, essentially the lithosphere-asthenosphere boundary being an important target for us in the future.
Oliver Strimpel
Tony Watts, thank you very much.
Tony Watts
You're very welcome.
Oliver Strimpel