This transcript was generated automatically and lightly edited for readability.

Oliver Strimpel: This is Geology Bites with Oliver Strimpel. By the end of the 1960s, the main ideas of plate tectonics had been accepted by the great majority of geologists. The plates had been mapped, their motions described and the ocean floor dated, and yet the theory rested on an assumption nobody had actually tested. To rotate a slab of lithosphere thousands of kilometers across without it crumpling, the plate has to be strong. But how strong? And how would you ever measure the strength of something that large? In this episode, we look at how that question was answered by borrowing the analysis engineers use to calculate the bending of a steel beam under a load and applying it to the lithosphere. It turns out that if you can describe how a plate bends under a load, you can apply it to all sorts of loads: a volcano, an ice sheet, or a pile of sediment. And some of what geologists have been reading as a record of changing sea level may instead have resulted from a plate sagging beneath the weight of the sediments piled up on it.

Tony Watts is Emeritus Professor of Marine Geology and Geophysics at the University of Oxford. In the early 1970s, at the Lamont-Doherty Earth Observatory, he tested the idea that the lithosphere bends like an elastic beam, first at ocean trenches and then at seamounts. He returns to Geology Bites after his earlier episode on seamounts. Tony Watts, welcome back to Geology Bites.

Tony Watts: Thank you, Oliver. It's good to be back.

Oliver Strimpel: Plate tectonics came together over about four years, from 1965 to 1969. What were the key moments?

Tony Watts: There were four main moments. The first was the discovery by Tuzo Wilson of what are called a new class of faults, or transform faults, recorded in new bathymetric maps, which showed that the spreading ridges were offset, and they were offset in a direction that was counterintuitive to what the offset would be for a normal transcurrent fault. And they showed that plates either side of the fault were slipping laterally with respect to each other.

The second key moment was in late 1966, when a magnetic anomaly profile that was collected by the research ship USNS Eltanin showed what is called the "magic profile," which was that the Pacific–Antarctic Ridge was spreading almost precisely symmetrically about the ridge crest.

The third key moment was when Dan McKenzie and Bob Parker showed that the Pacific plate could be explained by rotation about a pole that in the present day is located in Alaska, and that this rotation could explain the convergence directions in earthquakes in the western part of the Pacific.

The fourth and final key moment, I believe, was the work of Jason Morgan and Xavier Le Pichon, who brought the observations of transform faults, spreading ridges and trenches together into a global framework, which became known first as block tectonics and then finally accepted as plate tectonics.

Oliver Strimpel: So when you finished your doctorate, you went to Lamont-Doherty Earth Observatory just north of New York City, arriving there in the early 1970s, right after the theory was widely accepted. Did it seem then as if everything was kind of sewn up?

Tony Watts: I remember well my first day at Lamont-Doherty. Lamont-Doherty, as you say, is located north of Manhattan on a 150-acre site. I walked into what is called the Oceanography Building, and I remember seeing Walter Pitman and Manik Talwani huddled over a light table, and I peeked through to see what they were doing. And they were putting the finishing touches to a paper on the magnetic anomaly structure of the central North Atlantic, showing in great detail the separation of North America and Africa and establishing the poles of rotation needed to fit the two continents together. To a young postdoc arriving, it was overwhelming. And I remember thinking at that time, wow, what else is left to do?

Oliver Strimpel: So plate tectonics assumed that these huge plates were rigid enough to rotate about a pole without breaking up. Had anyone put that assumption to the test?

Tony Watts: Not really. There simply weren't at that time enough observations to carry out tests. The pioneer in testing the strength of the lithosphere was Joseph Barrell, and he actually used the term lithosphere and described it as a very strong structure capable of storing stress for long periods of time. And he argued that the lithosphere overlay a weak underlying layer, which he called the asthenosphere, so that loads put on top of the lithosphere would be supported by the strength of the lithosphere and the buoyancy of this weak fluid. But the observations were very limited.

Oliver Strimpel: How did that lead you to focusing on the rigidity of the plates?

Tony Watts: Well, by the early '70s, plate tectonics was essentially a kinematic theory. It was a description of the way the plates had moved and their interactions with each other. There was still no mechanism, and there was very little information on their physical properties.

Oliver Strimpel: How did you come up with the idea of using an engineer's equation and applying it to something that was so different from the kind of elastic rod that they had initially thought about?

Tony Watts: A paper that I read from Dick Walcott pointed out that seaward of the Kuril Trench there was a shallowing of the bathymetry before the trench was reached. In other words, there was a bulge in the seafloor. And Tom Hanks, the seismologist, had written a paper pointing out that there was a compressional earthquake of great magnitude, about 7.5, seaward of the Kuril Trench. And it was Tom who compared the bathymetry of the seafloor seaward of a trench to the engineer's formula for a broken plate that was loaded at its free end by loads, both vertical and horizontal. And he showed that if you applied a horizontal load in addition to the vertical load that was needed to take the oceanic plate down to the great depths of a trench, then you could explain why the plate was in a state of compression and explain his earthquake.

At the end of the paper, Tom suggested that his model was simply a model and it should be verified by looking at the gravity anomaly. And for some reason that I still don't quite fully understand, I took that as a challenge, because I was looking at the gravity anomaly at island arcs and I knew Lamont-Doherty had a vast database of gravity data.

Oliver Strimpel: Can you just explain briefly what the gravity anomaly is actually telling us? I understand that the bathymetry shows the depth of the seafloor, and therefore you get the shape of the bulge and then the shape of the slab as it plunges down into the trench. But what does the gravity anomaly tell you?

Tony Watts: Well, what I found was that there was an outer rise seaward of the trench in both the bathymetry and the gravity anomaly. And importantly, I was able to determine the correlation between them; in other words, the number of gravity units for every kilometer in the region of the bulge. And what it showed is that the oceanic crust was uniform in thickness and had been warped upwards in the bulge region. So now we would call that flexure. So associated with the flexural depression at the trench is an outer rise, which is an uplift of the plate. And in fact, the engineer's formula predicted not one depression or one bulge, but a whole series of bulge-like ripples away from where the loads occurred. I did look for the second depression, but I couldn't find it.

Oliver Strimpel: In the principal bulge, did the shape and magnitude of the bulge match what these formulae predicted?

Tony Watts: It matched it perfectly. The formulae would predict a bulge because the formula was based on an elastic plate overlying a weak foundation, and if an elastic plate over a weak foundation is bent, it creates bulges and depressions of decaying amplitude away from where the load is emplaced.

Oliver Strimpel: And from that shape and the nature of the fit, you were able to get a handle on the strength of the lithosphere at that point near the trench.

Tony Watts: To some extent we were. The problem at the trenches was that we didn't know the magnitude of the horizontal loads that were applied to the plate. So we could in fact make a strong plate look weak simply by increasing the horizontal loads that were acting across what was the plate boundary. So there was some uncertainty about resolving how thick or how strong the plates were in the bulge region.

Oliver Strimpel: In your earlier episode of Geology Bites, we talked about looking at the response of the lithosphere to seamounts. Was what you found there consistent with what you found from the plate flexure at the trenches?

Tony Watts: Yes. Seamounts were concentrated volcanic loads on the surface of the seafloor. We could actually see the loads. They were visible to us. We could see the way that a seamount had grown up and out on top of the oceanic crust. We couldn't see such loads at the trench. At a seamount, we were thrilled, because we could see the loads and calculate the flexure of the plate, which again involved a depression and bulges, this time two bulges, one on each side of the seamount. So we had a double bulge, which was even more impressive to us.

We went on to show that the gravity anomaly at a seamount that formed on a mid-ocean ridge would be much smaller in amplitude and shorter in wavelength than that of the same size of seamount if it had been emplaced on old lithosphere away from a mid-ocean ridge. And this to us was a very important result, because it suggested that the rigidity of the plates was actually a function of the age of the lithosphere at the time of volcano emplacement. It meant simply that as the oceanic lithosphere is created at the mid-ocean ridge, it increases its density and it subsides as it cools. And, we could add, it increases its rigidity with time.

Oliver Strimpel: Seamounts aren't the only loads on the lithosphere. As I mentioned in the introduction, your work on sedimentary basins led to quite a big revision of estimates of past sea level rise. Can you start with a basin you studied?

Tony Watts: The first rift-type basin we studied was the Baltimore Canyon Trough off New York, on the eastern seaboard of the United States, and we had some industry well data that told us how the margin had subsided. In fact, this well data was acquired by a consortium of oil companies. And because they couldn't agree on what structures to drill on, they decided to drill where everyone agreed there was no structure. And so they drilled what became the most perfect stratigraphic test well.

And we found, using a method called backstripping, that the subsidence had two components. The first was the thermal subsidence from the rifting and the extension, but two-thirds of the subsidence was caused by sediment loading, meaning that as the thermal subsidence was occurring, even though it was small, it was able to provide the accommodation space for sediments to come in, load and flex the lithosphere, and then do this repeatedly over 100 million years or more. And in the end, in the Baltimore Canyon Trough, the sediment loading together with the thermal subsidence produced a thickness of nearly 15 kilometers of sediments.

Oliver Strimpel: How did that affect people's estimates of sea level rise and fall?

Tony Watts: Well, when we started to ask the question of how the sediment loading worked, we were guided by the results from seamounts. So it was quite easy to conclude that the early sediments, soon after the rifting process had ended, would have been deposited on hot, and therefore weak, lithosphere, whereas the later sediments, the youngest sediments, would have been received by a cold, strong lithosphere.

The moment we took that result, which essentially came from the seamounts, and applied it to sedimentary basins, we could see that at their edges they had what was called onlap. Onlap was the overstepping of strata at the edge of the basin, and it was progressive because the rigidity we had assumed for the underlying lithosphere increased with time. So the flexure was going to be quite narrow and confined early on, but later on the plate was going to bend over a much wider area. So we would create the onlap. And this at once triggered interest, because it was precisely coastal onlap that was being used to infer sea level rise, because sea level rise involves transgression of the sea onto the continent, and the expectation would be that that would create onlap.

Oliver Strimpel: So can we disentangle the two processes, then? Do we know how much of it was the mechanical kind that you just talked about, rather than a real change in sea level?

Tony Watts: The short answer is it's difficult. The problem is that to construct a true global eustatic sea level curve, we need to determine the onlap and offlap patterns, offlap being the opposite of onlap, from widely separated sedimentary basins, ideally basins with different ages of rifting. Our studies showed that if the continental margin was part of the breakup of a very large supercontinent, say like Pangaea, then if we did that exercise, we would end up with very similar sea level curves. The flexure contribution to the problem shows that this is exactly what we would expect from tectonics in the form of flexure. The pattern of onlap would be identical.

Oliver Strimpel: Were the two contributions to the onlap of the same order of magnitude?

Tony Watts: It was quite similar in places. In fact, we used the same methods that were being used to extract sea level from measurements of the onlap, and we simply did that on our models, whereas they had done it on their observations, and they agreed.

Oliver Strimpel: Let's talk a bit about ice sheets. They seem to be quite different kinds of loads, because they only last on the order of thousands or tens or hundreds of thousands of years, rather than many millions. Does that affect how the plate responds?

Tony Watts: Yes, there is a strong control by the age of the load or unload on the rigidity. What the waxing and waning of ice sheets have shown is that when the ice starts to melt, it forms what is similar to a foreland basin, or a depression in front of the receding ice sheet. These are the proglacial lakes that form in front of an ice sheet as it recedes. The width of these lakes is quite large and requires quite a high rigidity in order to explain them. The best examples of such lakes are the Great Lakes in North America, the remnants of which we can still see today. So they extend over quite a wide region, which would mean the plate is quite thick and rigid with regard to that short-term loading event; as you say, thousands of years rather than millions of years as in the case of sedimentary basins.

Once the ice started to recede, the lithospheric plate was being unloaded and rebounded up. Studies of the shorelines of islands that were once in the lake require a thick elastic plate, of the order of about 80 kilometers, which is much thicker than what we envisaged for the strong parts of the Pacific plate, for example, which is more like 30 to 35 kilometers. So it appeared that the short-term loads and unloads, such as the ice sheets, were associated with a thicker lithosphere than the long-term loads, the million-year loads of seamounts, for example.

Oliver Strimpel: What about other planets? Does flexure give us a way to determine the strength of a rigid surface layer on another planet?

Tony Watts: It turns out that flexure is quite universal. It doesn't just apply on our planet. The Moon, Mars and Venus all have flexural features, mainly in their topography and sometimes in their gravity anomalies. Perhaps the best example of flexure is on Mars, where if we look at Olympus Mons, for example, which is the largest volcano in the solar system, it has a crater on the top of it, and it has a gravity anomaly of 3,000 milligals. And models of that using the engineer's formula suggest that the thickness of the Martian lithosphere is of the order of 200 kilometers, which is far in excess of the thickness on Earth. And really, it is that thick, cold structure that is responsible for being able to support such an enormous volcano for very long periods of time.

Oliver Strimpel: Fifty years on, we still have some pretty fundamental open questions. For example, ice sheets give a different answer for the thickness of the plate than seamounts or trenches do. Do we understand why?

Tony Watts: We certainly need to recognize that the observations that we have today are only snapshots that capture the response of the plates to loading and unloading. It's a fact that the short-term loads, for example seismic body waves, see a very thick mechanical structure, maybe one that involves the lithosphere and much, if not all, of the mantle. Then we have intermediate loads like the waxing and waning of ice sheets, which are giving us essentially a thinner mechanical structure, but still thick; their response is about the thickness of the lithosphere. And then finally we have loads such as seamounts and trenches, which are much longer-term loads, and these yield very thin elastic plates.

So what seems to be happening is that there is perhaps some kind of relaxation going on within the planet as a whole, whereby the thickness of the mechanical layer that is supporting a load thins from a very thick value to a very thin value. The problem is, because we have snapshots, we can't see the joined-up continuum of this apparent relaxation. The idea about the relaxation is that it's tied to the thermal structure of the outer layers of the Earth. It's simply hotter in the deeper portions of the lithosphere, and therefore lower in viscosity, than it is at the surface of the lithosphere, which has a higher viscosity. So a load will trigger the low-viscosity region first, because it wants to relax first, and stresses would migrate upwards into the strong upper layer. So we can make a prediction of the path of the relaxation curve. But it's only a prediction, and we still do not understand exactly how it goes from one transient state to another. And I think this is a really important and fundamental question left over from plate tectonics that we still have not addressed.

Oliver Strimpel: I want to ask you about subduction zones. In his Geology Bites episode, David Bercovici suggested that subduction zones can initiate through microscale grain-size reduction that weakens a rock enough to localize the shear and cause it to rupture. Does that mesh with your view of the mechanics of the whole?

Tony Watts: Plate tectonics does indeed require localization, and the grain-size and damage mechanisms proposed by David Bercovici are certainly a plausible explanation for the weakening of the plates in narrow plate boundaries, for example. In fact, our most recent work shows that the oceanic plates in the region of hotspot-generated seamounts are actually quite weak, but we have shown that the weakness is greater than would be expected from the results of experimental rock mechanics on brittle failure and ductile flow within the plates. So grain-size weakening may well be a process that is working in addition to these more established weakening mechanisms associated with brittle failure and ductile flow.

We found that the 4,500-kilometer-long Louisville Ridge chain of seamounts in the Southwest Pacific, which connects two major plate boundaries, the Tonga–Kermadec Trench and the Pacific–Antarctic Ridge, is a weak zone within an otherwise strong Pacific plate. So I don't think it's too much of a stretch of the imagination to imagine that if some time in the future external forces are applied to the great Pacific plate, it would see these regions along the hotspot-generated seamount chains as a weakness. Remember that they already have very large downward flexures of the oceanic crust, so the weakness, together with the downward flexures, would make them plausible settings for future subduction initiation, breakup of the Pacific plate, and a readjustment in the global network of plate motions.

Oliver Strimpel: Tony Watts, thank you very much.

Tony Watts: Thank you, Oliver.

Oliver Strimpel: To see pictures and illustrations that support this podcast, go to geologybites.com, where you'll also find a subject matter index of all the episodes. There you can also give me feedback, which I welcome, as well as sign up to get my emails about new episodes.