This transcript was generated automatically and lightly edited for readability.

Oliver Strimpel: This is Geology Bites with Oliver Strimpel. The heat liberated during the formation of our planet created an ocean of magma. As it began to cool, the Earth differentiated into a dense metallic core surrounded by a less dense, rocky mantle. At some point, we know that the surface of the Earth must have formed itself into the rigid blocks we call plates, and that these plates began to move and interact with each other as part of the global process we call plate tectonics. But did the plates form, and did plate tectonics start, soon after the Earth differentiated into the metallic core and rocky mantle over four billion years ago, or was there a long interim period before the initiation of plate tectonics? There's an ongoing debate about this, with some finding evidence for plate tectonics as far back as the Hadean, over four billion years ago, and others suggesting that it was not until the Neoproterozoic, about 750 million years ago, that modern plate tectonics got going. Peter Cawood is a professor in the School of Earth, Atmosphere and Environment at Monash University in Melbourne. His research aims to discover the origin of the Earth's continents by studying the ancient interiors of continents, called cratons. It is these cratons that preserve the clues as to the conditions and processes that prevailed in the early Earth, which in turn can tell us when and how plate tectonics started. Peter Cawood, welcome to Geology Bites.

Peter Cawood: Thank you, Oliver.

Oliver Strimpel: The early Earth must have been a lot hotter than it is today, so naively, it would seem pretty likely that the Earth would have had to cool down quite a bit before modern-style plate tectonics could take hold. Can we say anything about what the Earth's surface was like during that period?

Peter Cawood: We think that prior to plate tectonics, there was a mechanism sometimes referred to as stagnant lid. What that means is that there was a crust on the Earth, and rather than heat being lost through focused igneous activity, as with plate tectonics, along mid-ocean ridges or along the so-called Ring of Fire around the Pacific, those volcanic chains, it was probably lost in a more random process. That could have taken a number of different forms. One could be scattered volcanic activity through this crust, and the classic example is the moon Io, in the outer solar system, where images have shown major volcanic activity as a way of losing heat into space. Another way is that there was catastrophic overturn of this crust. The crust was there as a relatively stable feature, but it progressively heated up through heat from the interior, and then it would be catastrophically recycled back into the mantle, and a new surface would form, and the process would build up again. These are basically, as I said, referred to as stagnant lid. We think they operated for some period of time, and, as you pointed out in your introduction, that period of time until plate tectonics commenced is uncertain.

Oliver Strimpel: So what might the surface of the Earth have been like as we were about to transition into the plate tectonic regime?

Peter Cawood: The pre-plate-tectonic regime, I think, gets back to this idea that the Earth was probably considerably hotter than it is today; estimates put it at 200 to 300 degrees hotter. So the convecting mantle temperature, which today we think is around 1,350 degrees, was probably closer to 1,500 or 1,600 degrees back in the Archean. That means that not only did the mantle have higher degrees of melt in its upper layers, but that heat was being transferred to the crust. And so the crust itself had a lot more igneous activity. The rocks we find in the Archean, the so-called granite–greenstone terranes that characterize the cratons, represent really major pulses of igneous activity, both mafic, in the greenstones, and felsic, in the granites. And so we think the crust was a lot more squishy. One of the key criteria of plate tectonics is that you have these plates, and those plates are rigid, and they have boundaries of focused deformation; today, that's represented by earthquakes. But the plate itself is rigid, and the deformation is focused at its margins. So we think in this preceding period, the crust itself was a lot more viscous, if you like, a lot more squishy, and so the surface wouldn't have been as rigid as it is today.

Oliver Strimpel: So if we wanted to figure out when plate tectonics started, we establish what the manifestations of plate tectonics would have been, as proxies for the existence of plate tectonics, and then go looking to see when those proxies appear in the geological record. Is that sort of the methodology?

Peter Cawood: Absolutely. People would argue about what's a valid proxy to recognize plate tectonics in the past, but that's certainly the approach I've taken.

Oliver Strimpel: So what are these proxies?

Peter Cawood: Well, I think there are about five or so really good proxies that together give us some indication of when plate tectonics may have started. They're things like rigid crust: plate tectonics has these rigid plates, lacking intraplate earthquakes, with deformation focused at the margins. Sedimentary basins: basins are a piece of evidence that gives us some indication of the mechanisms of subsidence, and they're related to tectonic mode. One of the most popular ones is igneous geochemistry and the types of environments it can tell us about. I like the metamorphic record as well. And finally, perhaps the gold standard in terms of plate tectonics is paleomagnetic data.

Oliver Strimpel: OK, let's take these one at a time. Let's start with plate rigidity.

Peter Cawood: Rigid plates are characteristic of plate tectonics, but as I also mentioned, on the Archean Earth, we think the mantle was a lot hotter, and the crust itself was a lot squishier, if you like. It was squishy because there was a lot of magmatism taking place. And so is there a time within that history of Archean craton formation when we think that process stopped, when the cratons cooled sufficiently to be rigid and show brittle deformation, or rigid deformation? I think there is, and one of the key pieces of evidence is dikes. Not just an individual dike in an outcrop, but if you look at the cratons of the Earth, you can see that across a number of them, late in their history, they are pervasively cut by major dikes. The classic one is the Great Dyke of Zimbabwe, in the Zimbabwe craton, which can be traced for something like 500 kilometers and is kilometers wide in places, and that's dated at about 2.57 billion years ago, so right at the end of the Archean. So that's a really key piece of evidence.

Oliver Strimpel: Are there any other lines of evidence that point to plates having become rigid?

Peter Cawood: I mentioned earlier sedimentary basins, and I think sedimentary basins do provide some key evidence. Looking at these cratons towards the end of the Archean, they became coherent blocks, and then they were susceptible to large-scale subsidence and sediment accumulation. So we see on these cratons that the granite–greenstone terranes, the basement, if you like, had by this stage stabilized, and they were uniformly subsiding, and sedimentary successions were accumulating. An example is the Pilbara Craton in Western Australia. It's overlain along its southern margin by a sedimentary succession called the Fortescue Group, dated at about 2.7 billion years, and the base of that succession can be traced for some 500 or 600 kilometers around the margins of that craton. So we know that by that time, 2.7, we had a coherent block, the whole of the Pilbara Craton, and it was subsiding, and these sediments were accumulating on top of it. So at the end of the Archean and the start of the Paleoproterozoic, we see that these Archean blocks had stabilized. They had become rigid, and they were now subsiding and allowing these sedimentary successions to accumulate right across them.

Oliver Strimpel: So I suppose the point here is that these cratons were rigid and strong enough to be able to support these thick, continuous successions of sedimentary rock, which otherwise, I suppose, if the crust was squishy, would just sink down and wouldn't accumulate in this coherent fashion. So it's an indication of the strength of the plate.

Peter Cawood: Yes. Continuing on from that point, you can look at sedimentary basin successions and see what types of basins are present on these cratons, and what that tells us. In the modern Earth, we have some basic mechanisms of sedimentary basin formation. One mechanism is extension. You extend the lithosphere; you pull it apart like an elastic band, and eventually it breaks. And so extension produces rift successions. There are lots of little faults, rift basins forming, half-grabens and grabens, et cetera, and eventually you'll reach a point where the continental basement extends and breaks, and you form a mid-ocean ridge between the pieces. So that's classic plate tectonics. Well, if that's the case, what are the oldest examples of the transition from this lithospheric stretching to breakup and passive margin subsidence? The answer is that the oldest ones we seem to recognize are around 3 billion years old, in the Singhbhum Craton in India, but we also have some other examples at around 2.75, in the late Archean.

And if you're having that lithospheric extension, then, since we're not on an expanding Earth, you know, the Earth is of constant radius, there must be areas of compression. The compressional-type sedimentary basins are called foreland basins, and they're produced by loading of the crust. So when India and Asia collide, as an example, the Indian crust is loaded by the collision with Asia, and you get a big foreland basin succession accumulating. We see that throughout history, with older basin successions in North America, inboard of the Appalachian orogen and inboard of the Cordilleran orogen. We see this loading and subsidence. And so what are the oldest compressional, foreland-type basins in the geological record? Then again, the Witwatersrand and the Zimbabwe cratons show examples of potential foreland basins that are younger than about 3 billion years. So we're seeing this consistent type of number, in the late Archean, for these major basin types forming. And just to re-emphasize the importance of these basins: they require lateral motion. If you're getting a passive margin forming, you've got evidence for lateral motion, the stretching of the lithosphere, the moving apart of two continental blocks. If you've got a convergent, foreland basin setting, then you've got two blocks of material coming together and one loading the other.

Oliver Strimpel: OK. So sedimentary basins are useful proxies for extensional plate margins and compressional plate margins, as well as the extensive sedimentary basins also being a proxy for the strength of a rigid plate to hold them up. Let's move on to the third category of proxy. How does geochemistry give us clues as to whether plate tectonics was operating or not?

Peter Cawood: On the modern Earth, at convergent plate margins, what we see are distinctive geochemical signatures, sometimes referred to as thorium–niobium ratios. That reflects the fact that at convergent plate margins, oceanic crust is subducted, and as it goes back down into the mantle, it's obviously got fluid trapped within it. It's been in the oceans, and fluid phases are present there, both in the pores and as alteration phases in the mineral assemblages. As it goes down, that water is released into the mantle wedge in the overriding plate, and that induces what we call fluid-flux melting. The fluids go in, they generate melts, some elements are carried up in those melts, some are retained in the mantle, and so thorium–niobium ratios reflect that fluid-flux process. Enhanced thorium–niobium ratios are a signature of subduction. So what one can do, then, is look at rocks in the geological past and see: do they have high thorium–niobium ratios? People have done that, and the oldest that we can see go back to around 3.8, in West Greenland and in northern Canada. But they're localized. And remember, the definition of plate tectonics is not only rigid plates, but that it's part of a globally linked system; the plate boundaries all join up around the Earth to delineate the margins of the plates. So geochemistry is important, but these are kind of spot data, you know, data from West Greenland or data from the Pilbara. So you have to bear that in mind as well.

That's the archive from mafic rocks, but one can also look at the felsic rocks, what we call TTGs, tonalites, trondhjemites and granodiorites, sodic granites. All the Archean cratons are characterized by this rock type. The interesting thing is that this magmatic activity, these granites that are emplaced into the greenstone belts, has a very long history, extending over hundreds of millions of years, in all the cratons. But at the end of that magmatic activity, it changes in composition. It shifts from these sodic granites into more potassic-type granites, monzogranites, et cetera. And so, again, we have this history of a change in composition sometime in the late Archean. It varies from craton to craton, but the changes are regular. They all show this sodic-to-potassic change, and so we're seeing the cratons change in composition. It probably reflects their increasing thickness and, indeed, their increasing rigidity, which is coming together in the late Archean.

Oliver Strimpel: I suppose this is quite connected now to the fourth line of reasoning, or fourth proxy, if you like, which is the metamorphism present in the rock record. How does that tell us anything about whether plate tectonics was operating?

Peter Cawood: There are three basic types of metamorphic assemblages, and they're related, as you might expect, to the temperatures and pressures that these rocks were subjected to. The three categories are, first, high-temperature, relatively low-pressure assemblages, and on the modern Earth, they're characteristic of areas like back-arc basins and mantle plumes. You've got high temperatures, you know, mantle upwelling due to thin lithosphere or whatever. So, high temperature and low pressure. Then you have another band of intermediate temperature and pressure, the classic Barrovian-type metamorphic assemblages, which are very characteristic of lots of metamorphic belts around the world, and which we often associate with convergent plate settings, that is, subduction and collision-type zones. And finally, you have low-temperature, high-pressure assemblages. They're interesting because the first two I talked about, high temperature and intermediate, are present throughout Earth's history, from the early Archean through to the present day, whereas the low-temperature, high-pressure assemblages really dominate from the Neoproterozoic onwards. It's around, you know, 750 million years or so that they become common, and they're formed in convergent settings; subduction zones are the characteristic environment where they're formed, and collisional zones as well.

So I said the first two, the high-temperature and the intermediate, occur throughout Earth's history, and they do. But if you look at the frequency at which they occur, they tend to show clusters of activity: a cluster at around 2.7, at around 2.1 to 1.85, at around 1 billion years ago, at around 500 million years ago, et cetera. So you can see there are these patterns of frequency in this dual assemblage of high-temperature and intermediate metamorphism. Both types occur together in those time frames, 2.7, 2.1 to 1.85, et cetera, and that corresponds with the timing of supercontinent assembly. So we would say that's evidence, perhaps slightly indirect, but evidence, for the convergence, the coming together, of different cratonic blocks, which had been stable by that time. They converged, then they collided and generated these dual metamorphic assemblages of intermediate- and high-temperature metamorphism.

Oliver Strimpel: OK, so let me see if I get this. There are certain clusters of high- and intermediate-temperature metamorphism, and the fact that these are roughly coincident with the assembly of supercontinents is taken to be indirect evidence for the large-scale movement of cratonic blocks that triggered metamorphism as the blocks amalgamated into the supercontinents. But we also see a new kind of metamorphism come in, in the Neoproterozoic, around 750 million years ago, which is your high-pressure, low-temperature metamorphism, characteristic of modern-day subduction zones, which might be another clue that that's when plate tectonics got going. Just to explain, you get high-pressure, low-temperature metamorphism when a slab of cold oceanic lithosphere sinks into the mantle, because it remains cooler than its environment as it sinks, because of its thermal inertia, even though it experiences the higher pressures associated with its increasing depth right away.

Peter Cawood: Yes. And indeed, for those who argue for plate tectonics starting relatively recently in Earth history, that is, in the Neoproterozoic, that is one of their key lines of evidence. They say, well, look, that's a slam dunk for subduction today, so that's when subduction started on the Earth, and these other assemblages formed perhaps in a stagnant-lid-type association. I think everyone would agree that subduction had started by 750, in the Neoproterozoic, at some stage, and that, yes, these are evidence for modern-style convergent plate interaction. I suppose those of us, and I'm one of them, who believe that subduction, or plate tectonics, started earlier would say that essentially the Earth had cooled sufficiently by the Neoproterozoic for these high-pressure, low-temperature assemblages to form, but that in the preceding time frame, in the Mesoproterozoic, the Paleoproterozoic and perhaps the late Archean, we still had plate tectonics going on, and we were generating these metamorphic assemblages, but the mantle was still warm enough that it didn't allow the formation of these low-temperature, high-pressure assemblages.

Oliver Strimpel: I wanted to come back to one very specific argument that we talked about earlier, when we spoke about how the high-pressure, low-temperature rocks that we see in some places, particularly in continental rocks, might tell us something about the temperature of the mantle and how powerful subduction was.

Peter Cawood: There's been some very neat numerical modeling done on the effects of changing mantle temperature on collision, and how that might impact the metamorphic record. What the numerical models have done is look at convergence between plates as the mantle cools. And what they have found is that if you have high mantle temperatures, say 200 degrees hotter than today, then subduction is not viable. The lithosphere is too squishy; you don't have rigid plates. The plate that is underthrust, the so-called oceanic lithosphere, isn't rigid, and also, once it starts to go down into the mantle, it heats up and breaks into fragments. So you don't have a coherent oceanic slab going down, as you do in the modern Earth. One of the consequences of that is that if you imagine some oceanic lithosphere with a continent attached to it, moving towards a trench, then as that closure takes place and the continent reaches the trench, the rest of the oceanic plate that's going down detaches and can't pull the continental plate down to undergo high-pressure, low-temperature metamorphism. It just breaks off, and so you can't generate those types of metamorphic assemblages. But what these models found was that once the mantle temperature got below about 100 degrees hotter than today, and we think that may have happened in the Neoproterozoic, the oceanic plate could remain attached to the continental plate as it was being subducted.

So imagine India coming towards Asia. The Tethys Ocean is closing, India reaches the trench, and the mantle is cool enough for the oceanic Tethyan lithosphere to remain attached to India, and it can pull India down towards the mantle, and it can stack up and thicken the Himalaya, et cetera. Well, we think that process is only possible once the mantle temperature is less than 100 degrees hotter than it is today. And so that's the major reason why we see these ultra-high-pressure, granulite-type assemblages in the geological record in the Neoproterozoic and younger.

Oliver Strimpel: OK, let's turn to the last kind of evidence, the paleomagnetic studies, which you referred to as the gold standard. In his podcast on supercontinents, David Evans explained how paleomagnetism works, but the basic principle is that we look for the remanent magnetism frozen into rocks at the time that they formed, as an indication of the latitude of those rocks at that time. Paleomagnetics was instrumental in really validating the concept of plate tectonics.

Peter Cawood: What's really exciting is that if we look at the cratons, a number of them do show relative changes in their positions, which can only be explained by lateral motion. The ones that we seem to have enough data for are the Superior in North America, the Kaapvaal in southern Africa, Kola–Karelia in Russia, and also the Pilbara. When we look at them, we can see this in the late Archean period. We don't have data for the earlier periods, the early Archean or the middle Archean; we only have sufficient data for the late Archean. And we can show that the Kaapvaal, the Superior and Kola–Karelia changed their relative positions in the time frame of around 2.7 to 2.4. And similarly, the Kaapvaal and the Pilbara, over a more restricted time frame, 2.8 to 2.7, also changed their positions, which indicates lateral motion between those bodies. So that's really exciting.

Oliver Strimpel: We've used a number of different proxies, which we think are manifestations of plate tectonics. So to what extent are we justified in using these proxies, and in assuming that the manifestations of plate tectonics that we see today also applied in the distant past?

Peter Cawood: Part of the problem is that people use different criteria to reach an interpretation. My view on this is that the more criteria you can bring to bear, the more confidence you have in your interpretation. Often people use individual data points, something from an individual craton, or they use an individual data type, geochemistry or sedimentary basins. The more data you can bring to bear, the more confidence you have. And so what I've tried to do in our talk here is outline not just one line of evidence but multiple lines of evidence: rigidity of plates, metamorphic conditions, magma compositions, divergent margins, passive margins, convergent margins, paleomagnetic evidence, evidence for recycling of differentiated crust. So, a number of different criteria. That doesn't mean that I'm right; it's impossible to prove you're right. But it means getting an interpretation that's viable with as many lines of data as possible.

Oliver Strimpel: So one of the things we understand as characterizing modern-day plate tectonics is that we have a global network of plates that are part of a single, unified plate tectonic system. Do we think that plate tectonics got going fairly suddenly, with some kind of, if you like, dramatic phase transition, when we went from a squishy or stagnant lid, flipped the switch and suddenly had full-blown global plate tectonics? Or do the models or the data show some kind of intermediate states, where maybe it was a kind of stop and start, a much more hesitant transition that may have lasted tens or hundreds of millions of years?

Peter Cawood: The data that I've outlined suggest that, look, there may have been episodic subduction, or localized subduction, in Greenland and in northern Canada at 3.8, which is what those data suggest. But we have other lines of evidence that other features didn't start until later. So we may have had attempted subduction taking place that failed, that didn't last; it was episodic. And then, progressively, we developed this globally linked network. And I think the data, particularly over the period between about 3.2 billion years ago and 2.5 billion years ago, are recording this widespread development, in Archean cratons, of rigidity, the development of passive margins, lithospheric stretching and lithospheric convergence. And so by the late Archean, if you had to pick a date, I would say it had developed into a global system.

Oliver Strimpel: So if you were to come down on a particular time when we can be pretty sure that we actually did have plate tectonics operating, when would that be?

Peter Cawood: I would say 3.2 to 2.5, this transitional period from a stagnant lid to full-blown plate tectonics.

Oliver Strimpel: I want to ask you about other bodies in the solar system that are, very roughly, the same size as the Earth, such as Mars and Venus, and the moons of Jupiter and Saturn. What do we know about how they cool, and do we see any evidence of plate tectonics on any of them?

Peter Cawood: Very good question. What's interesting, perhaps, is that the Earth seems to be the only body, the only planet, in the solar system that shows evidence for plate tectonics. The other bodies, whether they be our own Moon, or Venus, Mars, et cetera, show largely stagnant-lid mechanisms, with lots of disparate, randomly orientated volcanic activity across their surfaces. Each of those bodies has lost heat, we think, through a stagnant lid, not through plate tectonics. Now, there are some features on Mars that people have interpreted as possible examples of plate tectonics, but again, it's not plate tectonics as we know it. It's kind of localized evidence, which may have driven the volcanic activity of Olympus Mons on Mars, but it wasn't a globally linked system. And there are different types of stagnant lid there, perhaps, you know, with lots of volcanic activity on some of these bodies. Venus shows evidence, from its cratering record, that its craters are relatively young, 500 million years, unlike somewhere like the Moon, say, which has impact craters going back four billion years. And so what we think happens on Venus is that it's a stagnant lid: it's got this crust with no surface volcanic activity, but it's still hot internally, and that reaches a point where there is catastrophic failure of the outer surface, the Venusian crust, and it's rapidly recycled back into the mantle, and a new crust forms. We think that happened 500 million years ago; that's one explanation for this change in the cratering record of that body. So there are variations on a theme within stagnant lid, but nothing in terms of plate tectonics as we know it.

Oliver Strimpel: Peter Cawood, thank you very much.

Peter Cawood: Pleasure.

Oliver Strimpel: For more about Geology Bites, as well as pictures and illustrations that support this podcast, you can go to geologybites.com.