This transcript was generated automatically and lightly edited for readability.

Oliver Strimpel: This is Geology Bites with Oliver Strimpel. The Tibetan Plateau is a one-of-a-kind feature on the planet. It is vast, stretching about a thousand kilometers north to south and 2,500 kilometers east to west, an area about five times the size of France, and it has an average elevation of over 4,500 meters. The naive explanation as to how this plateau was formed is that it was jacked up as a result of its collision with the Indian plate, starting about 50 million years ago and continuing on to this day. But the evidence suggests that there is actually much more to the story, both as to topography generated before the collision with India, and as to elevation gain and loss following the collision.

Peter Molnar is Professor of Geological Sciences at the University of Colorado at Boulder. He has worked on a great many subjects, but especially on how mountain ranges are built, and how climate on geological time scales is affected by topography and crustal movements. His research on mountain ranges has focused on the high terrain in Asia, the Himalaya, Tibet and the Tian Shan, which he has studied extensively in the field. Peter Molnar, welcome to Geology Bites.

Peter Molnar: Thank you, Oliver.

Oliver Strimpel: What was the topography of Tibet like before the collision with India?

Peter Molnar: Well, a hundred million years ago, much of Tibet was covered by limestone. Coral reefs are the prime example of limestone; virtually all limestone forms under water, under oceans. So this rock, which you see in southern parts of Tibet, attests to a low elevation. Tibet was low, near sea level, below sea level in places, a hundred million years ago. But then, roughly 90 million years ago, it appears that the southern margin of Eurasia began to develop into something like the present-day Andes. A high mountain range grew, with volcanoes on top of the high mountains, all along the southern edge. And so before India arrived, there already was something resembling a plateau, a narrow plateau. Today, this region is what's sometimes called the Transhimalaya. The city of Lhasa lies within it, and granite, the igneous rock that forms beneath the volcanoes, crops out all along the Transhimalaya, from east of Lhasa to far west of Lhasa.

Oliver Strimpel: So how do we know that there was a high mountain range there at the time?

Peter Molnar: There are a couple of facts that point in this direction. One fact is that the rock there is folded up. There's rock that's around 90 million years old, 80 million years old, and it's tightly folded, and capping that rock is lava that poured out on the top. The lava is more like 60 million years old. So we see evidence of folding of the surface rock, which is what happens when you're making mountains, and then we see rock on top of it that attests to this folding being before the collision.

But there's a second set of facts. One of the techniques, actually a multitude of techniques that have been developed in the past 30 years, are methods of trying to determine past elevations. We call this paleoaltimetry: paleo for past, altimetry, obviously, for heights. And a number of these point towards this region being high. Some of the dates are as old as 50 million years ago, approximately when India collided; some are younger. But there seems little doubt that there was a high zone in there. The combination of this folding and faulting of the rock that predates India's arrival, and the existence of high terrain that we see at the time of the collision, is what we use to infer that.

Oliver Strimpel: Let's go north of that. What do we know about the elevation history of the region further north?

Peter Molnar: If you go north of the Transhimalaya, there's a belt of basins: sediment that's eroded from mountains both to the north and to the south and deposited in these lower areas; hence they were basins. And paleoaltimetry suggests they were quite low. But then north, in the middle of Tibet, there was apparently high terrain. We know that because we see this rock deformed. It's been thrust southward on top of the rock at the edge of these basins, and it's probably been thrust northward over the area to the north. We don't know how high it was, but there's no question that there was high terrain there. North of that area, there's sedimentary rock, very coarse material, deposited by rivers that came from the south; there's no doubt about that. And so you have coarse material carried by rivers from the south to an area that's now 4,500 meters high. So there must have been a high mountain range, it's called the Qiangtang, through the middle of Tibet, and that then shed sediment both north and south.

Oliver Strimpel: We've talked a bit about the bedrock clues. Are there any other lines of evidence that we use in figuring out the history of the land elevation?

Peter Molnar: Paleoaltimetry relies on some processes in the atmosphere that are different when it's high from when it's low. We all know it gets colder when you go from low to high elevation, as an example. So we rely on these properties that change with elevation, like temperature, and then these properties, of course, are manifested in different ways. One way, obviously, is with plants. You go up in elevation and the plants change. Eventually you hit a tree line, and there are no trees above it. Geology is not good at mapping tree lines, but it is good at mapping plants that are different. A particularly good example: the Chinese have recently found fossil palm fronds in sediment that's approximately 25 million years old, in one of these basins that I mentioned between the Transhimalaya and the Qiangtang.

Today, palms cannot survive freezing; palms are tropical plants. Now, of course, there is the possibility that there were palms that could survive freezing, and they've since evolved so that now they can't. We assume not. These palms are an example of a tool that's used to place a constraint on the altitude. As I said, they can't stand freezing. Well, that would put a limit on how high Tibet could have been, if temperatures didn't drop below freezing. There are other tools, and the wonderful part about the Transhimalaya is that many of these techniques have been brought together, and they give the same answers, so they give us confidence.

Oliver Strimpel: When you go north of the Transhimalaya, well, then the techniques don't agree quite so well, or disagree in some cases. But can we say anything with any confidence about the history of the elevation north of the Transhimalaya?

Peter Molnar: I think so. I mentioned those basins, and I mentioned the palms that have been found. There's a wealth of information that's come out: fossilized bugs of various kinds, fossilized leaves of different shapes and sizes. And most of this can be used in a quantitative way. There are statistical tools that relate these to modern climatic conditions. You then take an assemblage of the fossil plants and assume the same relationships held in the past. You can infer a temperature, and then you have to take that temperature and infer a height. But I think that there's little doubt that these basins that I mentioned, between the Transhimalaya and the Qiangtang, were quite low, of order 2,000 meters, back 25 million years ago, whereas they're now 4,500 meters. On the north side of the Qiangtang, the data are sparser and less convincing. I still think the data suggest that area has risen a thousand meters in the last 25 million years.

Oliver Strimpel: That's interesting. So we're fairly confident that some parts of what is today the Tibetan Plateau were actually quite low even after India had collided. So something must have caused quite substantial elevation well after the collision was already well underway. Before we talk about that, is there any other kind of evidence we can use to further pin down the elevation history of Tibet?

Peter Molnar: Yes. Of course, the point of view I bring to this is not shared by everybody, but yes, there is another way of doing this. To appreciate this, it helps to understand one of the more fundamental concepts in geology. It's called isostasy. The name was given in the 20th century, but it was recognized back in the middle of the 19th century, and the logic is similar to what you have when you have ice on water. When you see an iceberg, 90% of the ice is below the water, and if you add ice and make the ice stick up higher, that means it'll reach deeper into the water. Isostasy is often described as Archimedes' principle applied to the Earth. The Earth consists of a crust which is less dense than the mantle underneath, just as ice is less dense than the water in which it floats. To a first, but crude, approximation, we can think of the crust as floating on the mantle, and if you have thick crust, the result will be that the crust will stick up high and also reach down and form a root deep into the mantle.

So for Tibet, you have crust that's five kilometers high, but the bottom, the boundary between the crust and the mantle, is 70 kilometers below that. Whereas under India, you're near sea level, and the bottom of the crust is closer to 40 kilometers below you. So you have both high terrain and a crustal root underneath. And one of the ways you make mountains is by thickening the crust. The Alps are a good example. The Italian Peninsula and the rest of Europe collided with one another tens of millions of years ago; the horizontal dimension of the crust became smaller, the vertical dimension became greater, mountains stuck up high, and there was a root underneath.

Well, the converse works. What happens if you thin the crust? What happens if the crust stretches and becomes thinner? Well, you'd expect the elevation to drop and the root of the crust to come closer to the surface. The stretching of the crust would, in fact, lead to a lowering of the crust. And what we know from Tibet today is that the crust is thinning today. The east-west dimension of the plateau is growing. The eastern edge is moving at 20 millimeters a year relative to the western edge. Part of this is compensated by some north-south shortening, but most of it, or half of it at least, is leading to a thinning of the crust. Well, if you're thinning the crust, it should be going down.

If you take the present-day rate at which it's extending, which we can measure with GPS, and you extrapolate back 15 million years, for example, you calculate that the crust would have been seven kilometers thicker. With isostasy, that means the bottom would have been six kilometers deeper, and the surface would have been one kilometer higher. So instead of having 4,500-meter elevations, you would have had 5,500-meter elevations. And I chose that number of 15 million years because that's approximately the time when this stretching seems to have begun. So what's happened is that for a period you were thickening the crust. This is what happened starting before the collision and continuing into the collision. But now, long after the onset of collision, long after India first met the southern edge of Asia, the crust has started to thin. And if we extrapolate the present-day rate, we would say the surface has dropped one kilometer. So that's a way we use. It's not paleoaltimetry; it's based on a series of pieces of logic, but those of us who think about this find this logic irrefutable.

Oliver Strimpel: Well, this crustal thinning, and the normal faulting that goes with it, we would expect to be a result of these isostatic principles operating the way you described. But is there any direct evidence of the actual thickness of the crust and the mantle that we can look at today that supports this? In other words, is something like seismic tomography giving us an actual insight into what's going on?

Peter Molnar: There are a number of tomographic studies. We do know without any doubt that the crust is thicker under Tibet; that's been demonstrated time and again over the past several decades. But the deeper structure, I don't think we've gotten there with this technique yet. Now, there is one exception that I would point out, and that is that Tibet's surface is quite flat, 4,500 meters, as you said, but the depth to the crust-mantle interface is not flat. It's more like 75 kilometers in southern Tibet and only 60 kilometers in northern Tibet. And you could say, oh, who cares? What's the difference between 60 and 75? Well, 15 kilometers is a big, big difference. You don't have 15-kilometer differences in crustal thickness just everywhere around the world. This is rare.

So why is northern Tibet as high as southern Tibet, when the crust in northern Tibet is 15 kilometers thinner? Straight isostasy would say the crust in northern Tibet should be two kilometers lower than that in the south, and it's not; it's the same height. The simple explanation, and I think abundant evidence corroborates it, is that the upper mantle of the Earth in northern Tibet is hotter than the upper mantle in southern Tibet. Just about everything, if you heat it, expands. And so if you have hotter upper mantle under northern Tibet, you have less dense upper mantle. In other words, hot: low seismic wave speeds. So hot in the north, cold in the south. That's a part of what's going on, and when you heated northern Tibet surely played a role in when the present-day elevation was reached.

Oliver Strimpel: OK. Now that we've looked at the evidence, and we can infer that at least some of Tibet rose by as much as two to three thousand meters over the last 25 million years, or maybe even over as little as 15 million years, why doesn't the simple idea that India is being thrust under Tibet and jacking it up explain what's going on?

Peter Molnar: It provides part of the explanation. I mentioned the Qiangtang. We don't know how high it was, but it probably rose a bit since India collided. The northern edge of Tibet surely has risen in part because of that. We see evidence of what are called thrust faults, faults where one part of the crust is thrust on top, pushed on top of the other, so you thicken the crust. The northeastern part of Tibet, the Qilian Shan, is very young. That's only 10 million years old, and that's because of crustal thickening. So the collision and the penetration of India into Eurasia have certainly contributed to the present-day elevations. Whether they've contributed all of it or not is one of these controversial topics. I'm in the school that thinks not all of it. I'm in the school that thinks that this change in the upper mantle, or replacement of cold material in the upper mantle beneath northern Tibet by hot material, is responsible for a thousand meters, not 5,000 meters, but a thousand meters, of the present elevation.

So one has a combination of India penetrating into Asia, thickening crust in front of it, but also the mantle being involved: removal of mantle lithosphere and replacement by hot material. Lithosphere is the cold, strong outer part of the mantle and the crust above. So removal of some of that cold mantle and replacement with hot accounts for a part of Tibet's current elevation. And also, I should say, part of why it's flat: the bottom of the thick crust is probably very soft. It probably flows easily. Compared to fluids on Earth, it's very strong, of course, but it would flow. So you can smooth out topography at the bottom.

Oliver Strimpel: OK, so let me see if I've got this straight. You're saying that the northward movement of India post-collision is a factor, for sure, but there's something else going on that accounts for a significant part of the elevation, and you mentioned the removal of the lithospheric mantle, which is the lower layer of the lithosphere. So how do you remove the mantle? Tell me a bit about the processes that you think might actually be going on.

Peter Molnar: We divide the Earth's structure up using two different approaches. One is composition, chemical composition. The crust is different chemically from the mantle, and it's less dense. We also divide the Earth up into layers where the boundary is not sharp: into a strong, cold outer layer, where all of the heat is transferred simply by conduction through it, and underneath, the region where heat is transferred by convection. You can imagine a pot of water that's boiling: you're transferring the heat by the movement of the water. The water is overturning. That's happening in the mantle. It's not boiling, but the material is moving down there, and then you have a lid on top that's cold, and it doesn't move very much.

But you can thin that lid. That lithosphere is the term that's used, lithos from the Greek word meaning rock. It overlies the asthenosphere; asthenos, I think, in Greek means weak. And so you can thin that lithosphere. My image of this is that it's like painting the ceiling. You paint the ceiling with paint. Paint is denser than the air underneath, and it's more viscous than the air, yet the paint can flow and come together and make drips that will then drip into the air, or onto your head if you're not careful. And I imagine that's what's happening in Tibet: you've thickened this viscous layer, the lithosphere, the strong part, relative to the less viscous part that's underneath it. You've thickened it, and then it's coalescing into drips that are falling off as blobs, blobs of relatively cold material compared to what they sink into, but warm compared to the material they leave behind when they sink.

Oliver Strimpel: So we have this picture of a blob of heavy, cold lithospheric mantle, which is the lower layer of the lithosphere, dripping off the bottom of the lithosphere into the hotter asthenosphere below, causing what's left to bob up. Is there any other hypothesis that is invoked to explain the elevation of Tibet today?

Peter Molnar: There is another mechanism. It's called delamination. If you think of the crust and the mantle, particularly the lithospheric part, as being laminar, you can then imagine you might peel one of those away. You somehow make a break, and the mantle part would peel away from the crust and leave the crust behind, and again you'd replace the cold material that's peeled away with hotter material from underneath. This was proposed more than 40 years ago by Peter Bird.

You can imagine rising hot material, we would say plumes, like plumes of smoke rising. You can imagine that there might be plumes of hot material rising under Tibet that would then heat the bottom of Tibet's lithosphere and remove it. This sort of phenomenon happens under Hawaii. That's part of why Hawaii is high: not only do you have volcanoes in Hawaii, but the sea floor around Hawaii is also high, because it's buoyed up by this hot rising material. The same with Iceland, in fact.

And then there are people who think that it's just crust, that you don't have the mantle involved, but they then want the crust to flow easily, and I think this is partly right, actually. You can imagine that if the lower part of the crust is really soft and gooey, you can push it around, or pressure gradients underneath will push it around. And I do think this is quite likely what's happened under these basins between the Transhimalaya and the Qiangtang. But some people would say you fill up the whole of Tibet doing this. The term that's often used for this process is channel flow: there's a channel, and there's flow of material in the channel. The channel has the mantle underneath and some part of the crust above, so the channel is in the lowermost part of the crust, or a lower part of the crust, and that material moves horizontally. As I said, I do think that contributes, but others would have it be a much bigger player in the process than I would.

Oliver Strimpel: I don't understand the mechanism of how that would result in a kind of wholesale uplift of a plateau.

Peter Molnar: It wouldn't result in wholesale uplift; it would rearrange the material. So you can imagine, if the crust under the Qiangtang was thicker than the crust under the basins between the Qiangtang and the Transhimalaya, well, you would squeeze crust out from underneath the two high areas and push it underneath the low area, and then you would raise the area that's low.

Oliver Strimpel: You've done a lot of fieldwork in Tibet. Does the political situation currently allow foreigners to enter Tibet, and if so, are you planning to return? And what would you be studying there?

Peter Molnar: There are two ways of working in Tibet. One is going in as a tourist and sneaking around and sneaking out, and it's dangerous. If you get caught, you could be put in jail. Tibet is controlled by China. Most of the scientists in Tibet are Chinese. All of my work has been done in collaboration with Chinese scientists, and my Chinese colleagues have been just wonderful colleagues, but the Tibetans are not much involved, and it's clearly a colonial situation, where the Chinese control things and the Tibetans don't. I feel that if you're going to work in another country, it's best to collaborate with the local people and not take an imperialistic view on this.

And I definitely plan to go back, or I hope to go back. The goal will be to determine, as best we can, the history of elevation change in Tibet. I didn't mention this technique: we're going to use biogeochemical tracers that allow us to infer temperatures in the soil, in the rock of course, but in what was once soil, or surface sediment. Because temperature, as we've discussed, decreases with height, the hope is we can determine the temperatures that then allow us to determine the past elevations. The hope is we'll go both to the northern part, north of the Qiangtang, and into these basins to the south, where the fossil evidence suggests a warm environment. Putting a number on a temperature is hard with the fossil evidence, but with these biogeochemical techniques, we hope we can do that. And I'm hoping that this will tie down a bit of the history that enables us to decide whether removing the lithosphere makes sense or not.

Oliver Strimpel: Is the kind of uplift that we see in Tibet happening elsewhere in the world?

Peter Molnar: There are a couple of places where I think the same phenomenon can be seen. One of them is in eastern Turkey, in Anatolia. There's a marine limestone that's just six or ten million years old, and it's now a kilometer high. That means that area was under water, under the ocean, not long ago. The upper mantle has very low wave speeds, and therefore is presumably hot. There's a bit of volcanism. So it looks like the same process we imagine under Tibet has taken place there recently, arguably more recently than in Tibet, over a smaller area.

Perhaps a second place where we think we see this is in the central Andes: southern Peru, western Bolivia, northern Chile and northwestern Argentina, the high plateau, the Altiplano. This is an area that does seem to have risen recently. This is perhaps the best place that's been studied using some of these paleoaltimetric tools to determine a past history. Carmala Garzione was the leader in this. What she found was that the part of the Altiplano where she studied had been as low as 1, or maybe 2, kilometers only back 10 million years ago. By roughly 5 million years ago, it was at its current four kilometers high. In other words, it's risen 2 or 3 kilometers in a short period of time, just a few million years: very fast. And then a whole lot of other evidence came in. People saw that rivers had started incising at the same time, as if relief had been small, and then suddenly you had high terrain, and rivers just cut right in. So there have been a number of studies that have extended this, and I think this is a very good example of this process.

Oliver Strimpel: This may be just a bit of a red herring, but in thinking about what you said about the mantle removal, with the idea of denser mantle literally dropping off, it reminded me of slab break-off following subduction of an oceanic slab. A very different geological context, I know, but is that something that suggests this kind of massive rupture of large-scale geological blocks can happen? Does that bear on this question at all?

Peter Molnar: That's a good question, for several reasons, actually. In my opinion, there's no question that slab break-off happens, and we've known this for 50 years. It's my opinion that 9 out of 10 times geologists appeal to slab break-off, the evidence in no way requires that process. It's just that they get stuck, and they do it. And you can see, I think easily, that this is a hard problem to study, because it's instantaneous in geology. It's not a process that's sustained. You break the thing off and down it goes. This can happen in a couple of million years, and we have to be lucky enough to live in a period when this is taking place. I don't know of a place where we can see it happening today.

So yes, it bears on this. It's similar: you drop the load off, and the surface is going to come up, for the same reason as when dropping the blobs off. But do we have evidence that shows that that's what's happened? Or is it more a case of: we're stuck, let's appeal to slab break-off, and we'll get our paper published, and nobody's going to doubt it because everybody knows it happens? Well, maybe that's why I'm a dinosaur. You asked when these problems will be solved. Well, maybe when those of us who have ideas that no one else accepts die, then the problems will be solved.

Oliver Strimpel: Peter Molnar, thank you very much.

Peter Molnar: Thank you, Oliver. I hope your audience finds something interesting in this.

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