This transcript was generated automatically and lightly edited for readability.

Oliver Strimpel: This is Geology Bites with Oliver Strimpel. Many of us have heard about the most recent supercontinent, which is called Pangea, but there is strong evidence for others, with the earliest one now speculated to have formed in the Neoarchean Era, about 2.7 billion years ago. So what makes us think they existed? And if they are real, how did they form and then break up?

David Evans is professor of Earth and planetary sciences at Yale University. He's a true puzzle master, piecing together all the available clues, especially those locked up in the magnetic fields imprinted into rocks when they formed, so as to trace the formation and subsequent breakup of supercontinents. David Evans, welcome to Geology Bites.

David Evans: Thank you, Oliver. It's great to be here.

Oliver Strimpel: How would you define a supercontinent?

David Evans: A supercontinent is a collection of continents that have gathered together by plate tectonic processes to come together as a single landmass. And the motions of those continents are the surface expression on the planet of deep forces in the interior that move the rocks of the Earth's mantle around. We think that those forces, and the cycle, act over a period of about 500 to 700 million years from one cycle to the next.

Oliver Strimpel: So do we think that the formation and breakup of supercontinents is an inevitable consequence of plate tectonics?

David Evans: I think there is some inevitability. When continents assemble, they usually assemble because the oceans in between them are narrowing. The oceans in between them are narrowing because the seafloor is subducting down into the depths of the mantle. So when continents collide, if you think about that process in the third dimension, which is depth, you think about a piece of ocean lithosphere dropping down into the mantle like a curtain, a cold curtain going into the interior of the Earth. The two continents come together, and that curtain breaks off and keeps going, all the way down to 3,000 kilometers below our feet. But then, when those continents are glued together, they have now created a large landmass under which no subducting ocean lithosphere can penetrate the mantle underneath. Knowing that these descending slabs of ocean lithosphere cool the mantle, if the continents have shut off the subduction zone in between them, that's a whole area of the underlying mantle that cannot be cooled, and therefore it will naturally be part of a system that's heating up, and that gives rise to the eventual breakup of the landmass.

So this is a kind of self-organizing system of continents assembling over mantle downwellings at the expense of internal oceans, whose seafloor slabs drop down to the depths of the Earth. And then, after a while, the subduction migrates to the perimeter of that supercontinent, and what happens underneath a supercontinent is a large upwelling that turns around the mantle convection and splits the continents apart again. The time scale for this, about 600 million years, corresponds with computer models that people make of mantle convection, which suggest that kind of time scale for this great overturning of the Earth's convective system.

Oliver Strimpel: So what is the evidence that the supercontinents actually existed?

David Evans: There are two types of evidence. I subdivide them into two categories: one is what I call the direct evidence, and the other is indirect, or proxy, evidence. The direct evidence would be the distribution of rocks of certain ages across the surface of the Earth. There are certain rock types that indicate to us tectonic environments more similar to places in the world today that are breaking apart, such as the Red Sea or East Africa. These are typically volcanic rocks, lavas. They are associated with faults that generate rifts that we can recognize in the ancient rock record. And there are other kinds of rocks that are common today in areas like the Himalayas and Tibet, where continents are colliding, and those might be granites or metamorphic rocks that give us an indication that there was a collision of continents. When we look at Earth history, we see alternating periods of time when we have great abundances of the rift-related rocks versus the collision-related rocks.

We also have direct evidence from the magnetic signatures of the rocks, which is the kind of lab work that I do. From those magnetic signatures, we can see the motions of the continents through time, on the basis of changing magnetic directions in rocks of different ages from each continent, and we can start to piece the puzzle together by amalgamating all the records of the magnetic histories of different continents as they wandered the globe.

Oliver Strimpel: How exactly does the magnetic signature of a rock tell you where it was when it formed?

David Evans: The Earth's magnetic field is mainly dipolar, which means it has a north and south pole that we're all familiar with. As the field lines emanate from the interior of the Earth out into space, they cross the Earth's surface in a predictable pattern that allows us to recognize, when we measure an ancient magnetic direction in a rock, that a continent was once at a particular latitude and had a particular orientation.

Oliver Strimpel: So we can get the latitude, but we can't get the longitude?

David Evans: I guess the longitude has been problematic, because there is this axial symmetry, this doughnut-like symmetry of the magnetic field around the polar axis. There are intriguing ways in which we can estimate longitude, either from earlier geological connections or, for example, if we see land animals that were able to walk from one continent to another; then we would be able to say that the continents were close together in longitude. We also may be able to use fixed spots of volcanism that we call hotspots, or mantle plumes. That might give us a deep reference frame for charting not only the latitude but also the longitude.

Oliver Strimpel: You actually have your own paleomagnetism lab at Yale. How do you tease out the magnetic fields that you're looking for in these samples of ancient rocks?

David Evans: Ancient rocks have gone through a long history of periods of maybe metamorphism, periods of maybe weathering. Of course, in the field we can try to minimize the effects of metamorphism and weathering, but there may be some other effects as well, like a recent lightning strike on the ground, which can also partially remagnetize a rock. So in the laboratory, we try to design careful strategies to remove the partial secondary effects of these remagnetization events. The way that we do that involves a laborious procedure: heating the rocks in a magnetically shielded oven to a certain temperature, cooling them back down to room temperature, making a magnetic measurement, then putting them back in the zero-field oven at a slightly higher temperature, essentially demagnetizing a portion of the mineral grains, cooling them back down, measuring them again, and repeating the process at successively higher temperatures, to slowly strip away the complex history that the rocks have experienced through maybe billions of years. In doing so, when we strip away those secondary effects, we may be able to extract the original magnetization from the time the rock formed.

Oliver Strimpel: What happens if magnetic fields are completely absent from the rocks? Can you still piece together the puzzle of how the plates would have fit together, by trying to match either the ages or the geochemical fingerprints of the rocks?

David Evans: Yes, this is also in the category of direct evidence. Let's imagine that we had two or three continental pieces that were once together, and they broke apart. As they broke apart, there may have been a lava, or a series of lavas, erupted onto those two or three different blocks, and then, when the pieces all went off in different directions, little fragments of that lava field would remain behind on each one. So if we had robust dating of those lavas on all three blocks, we might be able to surmise that they were once connected together. You could do the same thing with certain kinds of fossil species of fauna, of animals; sometimes people use fish, fossil fish from terrestrial lakes, or ancient forests. So all of these lines of evidence can be used together to make inferences about past continental connections.

Oliver Strimpel: To reconstruct the supercontinent puzzle, we need to get the ages of the rocks, and the traditional way of doing that is using the mineral zircon, which gives us a very accurate radiogenic uranium-lead clock. But the trouble is that zircon doesn't occur in all kinds of rocks, and in particular not in mafic dikes that form during continental rifting and breakup. So are there some new techniques that enable us to date these mafic rocks?

David Evans: Yes, it's very exciting. In the field of geochronology, just in the last 30 or 40 years, geologists have developed a much more routine method for dating minerals aside from zircon. One of the most successful ones is baddeleyite, which is a zirconium oxide mineral, and this forms, as you mentioned, in mafic rocks. The advances in dating baddeleyite mean that we are now able to combine geochronology with paleomagnetism on the same suite of rocks. We were successful for many years with paleomagnetism; mafic rocks work better for paleomagnetism because they also have high concentrations of iron oxide minerals, which are the minerals that hold the magnetic signal that we can use to chart the motion of the continents. But a magnetic direction from a mafic rock is not useful for the solution to this supercontinent puzzle unless it is associated with an age. So the ability to date these baddeleyite crystals has now revolutionized our ability to pair paleomagnetic studies with precise ages that we can use to reconstruct the continents.

Mafic dikes are extremely useful for reconstructing ancient supercontinents, because they're like cracks that extend into the stable interiors of continental blocks. If you think about the map of North America, you see there's a stable interior, and on the east side you have the Appalachian Mountains, and on the west side you have the Rocky Mountains and associated mountain ranges. And you can think of many continents that have these mountain ranges on one side or the other. So if you imagine the continents moving around, forming supercontinents or breaking away from supercontinents, they can be involved in collisions at their margins, but the interiors can be quite stable. And these mafic dikes, these cracks that form as an extension of some of the rift episodes, can penetrate deep into the stable interiors of continents. Therefore those rocks can be very well protected and have an excellent record of the magnetic signals, which can persist to the present day. Whereas, by contrast, the rocks in the mountain ranges near the edge could very well get their ages reset, but, more confoundingly for the purposes of my work, they can get their magnetization reset.

Oliver Strimpel: OK. So what is the indirect, or proxy, evidence for supercontinents?

David Evans: The proxy evidence is inferential, in that we can sometimes expect different climates of the Earth, for example, that may accompany a time when supercontinents were together or when they were apart. One example of this is if you imagine when you might expect an ice age, thinking about ice ages in the last few hundred million years of Earth history. Of course, we have the ice ages that we're living among today, in the last few million years, and we're very comfortable and familiar with those. It's been suggested that those ice ages formed as a result of the India-Asia collision. This, of course, is a matter of large debate, but the line of logic is as follows: if you push up a huge mountain range, like the Himalayas or Tibet, and you weather those rocks, then the CO2 in the atmosphere can combine with the silicon in those rocks and eventually work its way down into sediment, combine with calcium and form calcium carbonate, and therefore lock up the carbon from the atmosphere in the geosphere. So the idea is that if you have a lot of continental collisions, maybe that's a time when you'd expect more of this weathering, more CO2 depletion from the atmosphere and more sequestering of carbon in sedimentary rocks. So the suggestion is that if we had a supercontinent forming, we might expect that to be accompanied by an ice age. And indeed we did have a substantial ice age on the Earth 300 million years ago, at the time of Pangea. So maybe when we had earlier ice ages, we might also have had earlier supercontinents.

Oliver Strimpel: Are there other types of useful indirect evidence for supercontinents?

David Evans: People have tried a lot of proxies. Some of them involve the chemistry of sedimentary rocks representing the ancient chemistry of the ocean. They can involve trace elements that are usually out of the spotlight, like samarium and neodymium, rubidium and strontium, and rhenium and osmium. Also, of course, we can just look at the migrations of ancient animals that once walked the Earth. For example, dinosaurs could at one point have easily walked around Pangea, and their fossils are now found on widely dispersed continents. We might expect that when we have a supercontinent, the animals that are normally on separate continents, living happy lives in their ecosystems, might collide together. Those ecosystems might collide, and that could lead to an extinction. So when you have a supercontinent, we could have a propensity for extinction. And so the Permian-Triassic extinction was at one point linked to the supercontinent of Pangea as a possible causative factor. Of course, now we know that the Siberian flood basalts also played a major role in that extinction.

Oliver Strimpel: OK, let's talk specifically now about the supercontinents that we think have existed, or may have existed, during Earth history. The most recent one is Pangea, which assembled about 335 million years ago, during the Carboniferous, and began to break apart about 175 million years ago, in the Jurassic.

David Evans: Pangea is wonderful for us to study as the prime example of a supercontinent. We're so confident in Pangea because we not only have some of the remaining records of its formation, like the Appalachian Mountain belt, but we also have the direct evidence of its breakup, from rift valleys that are still preserved in places like where I am right now in Connecticut: abandoned rift valleys from the initial opening of the Atlantic Ocean. And we also have the direct evidence on the seafloor of continents that were separating and leaving behind a record of that separation, through the progressive accumulation of seafloor filling up the space in between. So we have a very good record of Pangea and its breakup, and it's been very challenging to learn about the predecessors of Pangea, primarily because we don't have that intact seafloor record available to us.

Oliver Strimpel: So you really don't have any doubts that Pangea was real.

David Evans: I think of Pangea with not quite, but almost, the same kind of confidence that I have that the sun will rise tomorrow when the astronomers tell us it will.

Oliver Strimpel: And did it contain pretty much all the continental masses that we know about today?

David Evans: Pangea contained almost everything, but not exactly everything. There are probably a few places in eastern Asia that were not actually physically connected to Pangea. But the idea is that if you have more than, certainly more than, 90% of the Earth's landmasses together as a single plate, some people say 75%, then that should be enough to count as a supercontinent.

Oliver Strimpel: OK, so what supercontinent preceded Pangea, then?

David Evans: The supercontinent preceding Pangea sounds like the most straightforward question in this business, but it's actually the most debated. This comes to an entity which is called Pannotia: "pan" for all, and "notia" stems from southern. This is a putative landmass that contained most of Earth's continents throughout the Southern Hemisphere around the time of the evolution of animals, about 600 to 500 million years ago. Pannotia has been the problem child of supercontinent research for a long time, and that is because its existence, or its alleged existence, falls right about the time when we have all sorts of other problems in trying to solve the geography of the ancient Earth.

One of those problems is that the magnetic field of the Earth was behaving very strangely at the time, so it may be very difficult to apply standard paleomagnetic techniques to this time interval. It also seems to be a time when, we know from the geological record, there were lots of collisions and quite a few rifts as well. So we know just from looking at the direct evidence of the rock record that there was a lot of tectonic motion. It's also a time just before the evolution of diverse animal life. We're jumping off into the Precambrian, where we don't have a robust fossil record to be able to give us dates, so we have to rely on other dating techniques, such as the uranium-lead system, when we're lucky enough to find a volcanic ash bed preserved in the sediments. And it's also a time when you might think you'd be able to use an indirect record, such as a glacial deposit, to tell us that we were at the poles. But this is also a time when maybe we were having snowball Earth ice ages, when the planet froze over completely. So now, if you say, here's a glacial deposit, were we at the pole? Not necessarily; you could have been at the equator at the time. So all of these confounding factors mean that the solution to Earth's geography 600 million years ago is very, very difficult.

Oliver Strimpel: I noticed that you skipped over Gondwana. Does that not qualify as a supercontinent?

David Evans: Yes, I skipped over Gondwanaland, because we know, hierarchically, that it was a portion of Pangea. So I don't think we can use the same terminology for the child and its parent in terms of categorizing these landmasses. But I think Gondwanaland was actually very, very important in the supercontinent cycle, and I think many of the questions that people raise about the existence of Pannotia, which are largely due to the indirect evidence, looking at things like a cold climate or the radiation of life, might also be explained merely by the amalgamation of Gondwanaland. Gondwanaland has actually been given the status of megacontinent. It is made up of about half of Earth's continental mass. It includes South America, Africa, Arabia, India, Madagascar, Antarctica and Australia, all together for quite some time on supercontinental scales, from about 500 million years ago until Pangea itself broke up. So Gondwanaland, as a semi-supercontinent, or megacontinent, is big enough that it might start to have some effects on the broader circulation of Earth's mantle and the convection processes that lead to the assembly and dispersal of supercontinents.

Oliver Strimpel: OK, so let's keep working our way back through time, now going back past Pannotia, whether it existed or not.

David Evans: So before Pannotia, we come to Rodinia. Rodinia is the supercontinent that we think assembled around 900 million years ago and broke up about 700 million years ago. This would be just before the diversification of the major groups of animal life, and it would be a time when we're starting to enter these possible global snowball Earth ice ages. Ironically, we have a much better handle on Rodinia than we do on the era of possible Pannotia, and this is because the Earth's magnetic field seems to be back on track as we go back in time. The Earth's magnetic field had this unstable period that may or may not be coincidentally tied to the radiation of animals and the snowball Earth events. All sorts of catastrophes were breaking loose around 600 million years ago. We go older than that, and, counterintuitively, the world seems to become a little bit more simple again, and this has allowed for more robust reconstructions of Rodinia.

Oliver Strimpel: I expect, as we keep on going back in time, that the evidence gets more and more tenuous, as there's more and more time for geological processes and magnetic fields to overprint and confound the original evidence. But let's keep on going anyway.

David Evans: Sure. Before Rodinia, we have a supercontinent that has been called Nuna, which has an etymology from the Inuit meaning lands of the north, and it was indeed a continent that we think was in the Northern Hemisphere. It's also been called Columbia, after some connections of various continental blocks with the area around Washington State, around the Columbia River. Nuna, or Columbia, would have assembled around 1,600 million years ago and broken up about 1,400 million years ago. What's interesting about that is that, as you mentioned, it seems that as we go back in time, things become more speculative. But in this regard, we have an abundance of mafic dike swarms. That might be giving us an extra advantage for going older in time, because we might have gone back to a time in Earth history that was a little bit more volcanically vigorous, so we just have a lot of good rocks that we can sample for paleomagnetism and uranium-lead baddeleyite dating. So we've been piecing together Nuna quite rapidly in the last decade or so.

We think that Nuna had some features in common with Rodinia. There are some geoscientists who speculate that the transition between Nuna and Rodinia was not a complete breakup, maybe something like what I was describing between Gondwanaland and Pangea. But there are others, myself included, who think that there actually was a substantial rearrangement of some parts of the world between Nuna and Rodinia. This also comes to the question of the indirect proxy records of supercontinents, because of some of the hallmarks of a vigorous supercontinent cycle, such as changes to ocean chemistry and the cycling of carbon from the atmosphere to the biosphere to the geosphere. You might expect that a lot of tectonic activity would lead to a lot of cycling, but the time interval between Nuna and Rodinia is one of environmental stasis. The chemistry of carbon in the Earth system seems to have flatlined for nearly a billion years. So some people have suggested that this is the indirect record of tectonic quiescence and not much of a transition between Nuna and Rodinia, whereas I would counter that argument by pointing to the direct evidence, which actually shows quite a bit of tectonic activity. So there's still a disconnect between some of the indirect and direct records, and that's a frontier of research moving forward.

Oliver Strimpel: The direct record that you're referring to, is that really the dike swarms you mentioned?

David Evans: It's some of the dike swarms, but it's also one of the world's largest mountain belts that we know of in Earth history. It is slightly interior to the Appalachians. It's called the Grenville belt, and we can see it in scattered exposures of uplifted rock, from central Texas, the Llano Uplift, to little portions of the internal parts of the Appalachians, to the Adirondack Massif, to the Laurentian Highlands in southern Quebec, and all the way around to other blocks now scattered around the world. It's one of the most major and dramatic ancient mountain ranges that we know of. It happened about a billion years ago, and I think it is direct evidence for a substantial set of collisions to form Rodinia.

Oliver Strimpel: OK. So can we peer even further back in time to discern the existence of a supercontinent that preceded even Nuna?

David Evans: Right. So now we're going to the time period about 2 billion years and older in Earth history. We have an abundant record of rocks back to about two and a half billion years ago, and that record is just beginning to be explored with these techniques. The mafic dike swarms exist. They are being dated as we speak. There are paleomagnetists sampling them, probably as we speak, and it's a very exciting frontier of research. But with all these continental blocks having been juggled around just one more time, it's proving to be a little more challenging, when we get back before the time of Nuna, to reconstruct some of these blocks. Now, we may suspect that there was a full-fledged supercontinent older than Nuna. It has already been given a name, out of speculation: Kenorland, which is named after the Kenoran mountain-building event that was measured and observed from rocks in southern Ontario, around the town of Kenora. With Kenorland, we don't really know if it was a true supercontinent, or if, alternatively, at that time there may have just been a collection of the first amalgamations of continental blocks. Plate tectonics might have been a little bit different prior to Nuna. Maybe we didn't even have a supercontinent cycle before Nuna. Maybe we were just amalgamating smaller blocks into continent-sized blocks at that time.

Oliver Strimpel: Let's turn the clock in the other direction for a moment. Do you see the supercontinent cycle continuing into the future?

David Evans: There's every reason to suspect that we should be heading toward another supercontinent following the breakup of Pangea. Australia is currently moving northward at about 6 centimeters per year, and at that rate it should be colliding with Taiwan and Japan about 100 million years from now. North America is moving west, and there's some question about whether it will continue to move west, or whether it will stall and decide to turn around and move back east, or whether it will move north or south. But we do think that these tectonic processes are happening, and it's natural to think that further collisions will lead eventually to another supercontinent. And that supercontinent has already been given a name: since it's the conjunction of Asia with the Americas, if we put the Americas in front of Asia, we come up with the euphonious name of Amasia.

Oliver Strimpel: So, for my last question: if you had unlimited funds for your research, how would you use them?

David Evans: I don't always get asked this question, but it's always welcome. I think, tongue in cheek, I would say, can we hire some engineers to design a time machine? Then we could actually answer these questions through direct observation. And of course not just the past; we could also see whether Amasia is going to exist and what it will look like. But in reality, one of the main impediments to doing the kind of work that I do, which involves going out into the field in quite remote places, is just access to the best exposures of rocks to sample. So if I had a personal helicopter at my beck and call, with long range, that could take my teams to the field areas, we would do a lot of very, very efficient research.

Oliver Strimpel: David Evans, thank you very much.

David Evans: Thank you, Oliver. It's been a pleasure.

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