This transcript was generated automatically and lightly edited for readability.

Oliver Strimpel: This is Geology Bites with Oliver Strimpel. Two-thirds of the Earth's surface is covered by the oceans. Until recently, only about 20% of the seafloor had been mapped using sonar from ships. We actually knew the topography of Mars a hundred times better than that of our own oceans. But that all changed when we started using satellite altimetry to map the ocean surface, and from that to infer the topography of the underlying seafloor.

David Sandwell is a professor at Scripps Institution of Oceanography at the University of California, San Diego. He has been a leader in using satellite measurements to map the sea floor, which has given us our first global sea floor map. David Sandwell, welcome to Geology Bites.

David Sandwell: I'm really happy to contribute.

Oliver Strimpel: What exactly is satellite altimetry, and how is it used to map the ocean surface?

David Sandwell: Okay, satellite altimetry is an active remote sensing technique, and it's used to map the topography of the land, or the ice, or the ocean. Actually, it's conceptually very simple. You have the satellite in orbit around the Earth. It goes around 14 times a day, and it has a GPS receiver, so you can track the position of that satellite to a few centimeters. This satellite has a radar that sends radar pulses down to the surface of the Earth, and they reflect off the ocean, in our case, and you measure the two-way travel time. And from those two measurements, the altitude of the satellite and the position from the GPS, you can recover a track of sea surface topography as the satellite goes across the surface. Then, from thousands of these tracks over many years and many satellites, you can build up an accurate topography map of the ocean surface globally.

Oliver Strimpel: So the basic idea, though, is you're sending pulses of radiation down from the satellite to the ocean, measuring the time delay, and then computing how far away it is based upon the time delay and the speed of light.

David Sandwell: That's right, so you're measuring the exact shape of the surface of the ocean.

Oliver Strimpel: But what does that have to do with the ocean floor, which is often many kilometers below that?

David Sandwell: Well, to first approximation, the shape of the Earth is an ellipsoid. The Earth is spinning, so it has a centrifugal force at the equator, and so the equatorial radius of the Earth is a little bit larger than the polar radius. So that's our ideal shape of the Earth, and the ocean surface pretty much follows that ideal elliptical shape, because it's an equipotential surface of gravity. However, there are these bumps and dips in the ocean surface, and they're created by mass anomalies inside the Earth. For example, if you have a seamount on the sea floor, it has an extra gravitational attraction, and that causes the water to pile up over the top of that seamount, and it creates a bump in the ocean surface. And if you have some kind of a trench on the sea floor, like a deep ocean trench, it'll create a dip in the ocean surface. So we can map out these surface bumps and dips and then infer what's on the seafloor.

Oliver Strimpel: But the ocean is pretty deep, so wouldn't the gravitational effect of the bathymetry, that is, the topography of the seafloor several kilometers below the surface, be all smeared out at the surface?

David Sandwell: Yeah, exactly. The resolution of this mapping method is about equal to the regional ocean depth, and this is really based on Newton's law. A typical ocean depth is about 4 kilometers, so the spatial resolution of this technique is about 4 kilometers. For example, if you had two volcanoes on the sea floor that were two kilometers apart, their gravity effect on the ocean surface, 4 kilometers above the sea floor, would be smeared out into one bump.

Oliver Strimpel: Does that mean that you get better spatial resolution in shallow coastal areas than you would in deep sea areas?

David Sandwell: That's correct. If you go to, say, the mid-ocean ridges, where the plates are spreading apart, the average depth is about 2 to 3 kilometers, and you get 2 to 3 kilometers of resolution. But as these plates move apart on the seafloor over millions of years, they increase their density. They subside, the ocean depth increases to about 4 kilometers on average, and the resolution gets worse. And then if you go out to the deep ocean trenches, where the depth is about, say, 10 kilometers, the resolution is even worse. So it really depends on the ocean depth.

Oliver Strimpel: So you have a mountain underneath the ocean that's about one kilometer high. How much would that raise the surface of the water directly above it?

David Sandwell: The ratio of the sea surface topography, what we're measuring from the satellite, to the seafloor topography, the mountain, is about one over 10,000 on average. So a volcano on the seafloor a thousand meters tall would produce a bump in the ocean surface of about 10 centimeters, sometimes larger, sometimes smaller. These latest radar altimeters have an accuracy of about a few centimeters, so one theoretically could map all features on the seafloor a thousand meters or larger.

Oliver Strimpel: But that's pretty small. I mean, it's much smaller than the tidal variations or the effects of waves or ocean currents.

David Sandwell: Yeah, that's true. There are a lot of factors that change the height of the ocean surface, and I sort of glossed over that. For example, tides: they're plus or minus a meter or so. You have things like El Niño, where you have water piling up in the eastern Pacific. Atmospheric pressure variations can change the height of the surface. Even ocean waves, ocean swell. So all these things can have amplitudes greater than the 10 centimeters I was talking about. But you can remove those effects. For the tides, we have accurate tide models. The other signals change in time; oceanographic effects change in time. So if you average many, many topographic measurements from the altimeter over a long period of time, they will eventually average out, and you can get the time-invariant part. That's the part we're interested in, the part that doesn't change on human time scales.

Oliver Strimpel: And so you wind up with vertical resolutions of a centimeter on scales of about 4 kilometers spatial resolution?

David Sandwell: Yeah, maybe a little larger. I didn't talk yet about the other way to map the seafloor, and this is actually quite important. You mentioned that 20 percent of the sea floor has been mapped by ships. When ships go out to sea, they have something called a multibeam echo sounder. It provides a swath of topography on the seafloor about equal to twice the ocean depth, and the resolution of that method is about 100 meters. It's much, much better than you can do with the satellite altimeter, but as you said, only 20 percent of the sea floor has been mapped that way. So it's really the combination of these two methods. We use the ship data to provide details where we have it, and we have the altimeter data fill in the gaps, most of the area. And then we actually use the ratio of the ship bathymetry to the gravity field measured by the altimeter to calibrate the depths from the altimeter. So this is really important. Also, the ship bathymetry is really used to define the overall shape of the ocean basins at large scale, you know, where the large trenches and the large ridges are. You can't see those things from the satellite. We're really focused on a certain band of wavelengths here.

Oliver Strimpel: When I look at the sea floor on Google Earth, for example, does that contain both satellite data and ship sonar data?

David Sandwell: Yes, that's correct. We've provided that grid to Google. If you look carefully, open up Google Earth and zoom into some area, you'll see part of the sea floor. It's all blue, of course; they don't really highlight the oceans. But if you zoom into certain areas, you'll see very high resolution. You might see things called abyssal hills, or seamounts, things that are smaller than the 4-kilometer scale, and those are usually in swaths across the ocean. That's where a ship has collected the high-resolution data. And then you look in the blanks where there's no ship data, and you'll see it's very smooth. That's really the bathymetry, or the topography, that's been estimated from satellite altimetry. So you really need both of these techniques. I also didn't mention that when you predict bathymetry from altimetry, you can find a lot of interesting features, seamounts and ridges and so on, which we'll talk about later. And if you're going to design a new seagoing expedition, you can use that prediction to figure out where you want to drive your ship and what you want to measure, to get a really optimal use of your ship time.

Oliver Strimpel: Now, I don't know anything about how well your research is funded, but I'm pretty sure you don't get enough money to put up your own satellites. So how do you get this data? Or are there just satellites up there with spare time to make the measurements you want?

David Sandwell: I guess maybe I'll go through a little bit of the history of how we've gotten this data. It really started back about 40 years ago, with a satellite called Seasat, which was launched by NASA in 1978, and it was really a revolution. That's what I call the first revolution of satellite gravity. There was a group at Lamont-Doherty Earth Observatory that created some really amazing maps of the seafloor. They had sort of 30-kilometer resolution, but that was still better than what we had from the ships. Unfortunately, Seasat failed after only 90 days in orbit, but given its success, the US Navy then developed another satellite altimeter, called Geosat, which was launched in 1985. That was a classified mission to map the details in the pull of gravity, or the marine geoid, at high spatial resolution. What they wanted was not so much to learn about the topography, but to use it for guiding their submarines when they're submerged, using a thing called inertial navigation. To do that accurately, you really need to know the gravity field due to seamounts and things. Those data were also used for launching ballistic missiles. This was back in Cold War times. When you launch a ballistic missile from the sea surface, you need to know the launch site gravity field so it takes off in the right direction, and then you can hit your target on the other side of the planet.

And so Geosat really collected some wonderful data, but it was unavailable until 1995, when the European Space Agency launched their own satellite, called ERS-1. Basically they duplicated what Geosat had, so the Navy declassified the Geosat data, and this was really the second revolution in marine gravity. This was an amazing time. We got much higher resolution. You could see all kinds of things on the seafloor, really wonderful data. So it was a very exciting time, the mid-'90s, when that came out.

But I haven't really mentioned that there's another use of satellite altimetry, and that's by the physical oceanography community. They don't really care about the geoid, the part of the ocean that doesn't change with time. They want to know about the variations in topography associated with currents and El Niños and sea level rise, global sea level rise. How is that going? That's actually a much bigger community, and they were able to get a number of altimeter missions funded: TOPEX and Jason and so on. But these missions are in a different kind of orbit. They aren't in a mapping orbit like the one we're interested in. They're in an orbit that repeats every 10 to, say, 30 days, and they use that repeat because they want to cancel out the geoid part and look at the time-variable part. So the tracks of those satellites are widely spaced and not very good for mapping the gravity or the seafloor.

So then the really exciting stuff: the third revolution came in about 2010, when the European Space Agency launched the CryoSat-2 altimeter. With CryoSat, you'd think it's there to measure the ice caps, the changes in the ice caps and the sea ice, but they leave it on over the ocean. And so it's providing a really remarkable measurement of the ocean topography. It's been going since 2010 and is still going today. Lots and lots of really exciting data, and it's a very good altimeter. This has been combined with another altimeter, called AltiKa, which went into a mapping phase around 2016. And then, in addition to that, we got what I call hand-me-down altimeters from the physical oceanography community. When their altimeters begin to fail, they give them to our community of marine geology people and say, here, you can run this until it's dead. And we've used that to get additional data.

So it's really an exciting time, and we're now beating this resolution down to almost 4 kilometers. That's the horizontal resolution. The vertical is maybe a few centimeters. And then I should mention one more thing: there's another altimeter that's planned for launch in 2022. This is planned by NASA. It's called SWOT, S-W-O-T, and it's going to measure swaths of height. The normal altimeters just measure a single profile, but this will measure whole swaths. And that might give us another factor of five improvement in accuracy. So things are changing and getting much better. And that's sort of how we've gotten the data. It's a whole variety of ways.

Oliver Strimpel: Okay, I'd like to talk a bit now about what satellite altimetry has actually revealed about the ocean floor. But first, why is it important to map the ocean floor from a geological, as opposed to a purely geographical, point of view?

David Sandwell: Well, we live on a planet that's actually two planets in one. We're living on the continents, and the continents are above sea level. They're thick crust, and the crust is billions of years old. It's gone through multiple cycles of rifting, collision, mountain building, erosion and sedimentation, and the tectonics is very complicated. If you walk around on the continents mapping out these structures, you don't have a clear picture of what's going on in plate tectonics. Take things like transform faults. We have a good example here in California: the San Andreas Fault system is a transform fault. It goes across the continent, but it's very complicated, and it has multiple strands, so if you look at that, you wouldn't immediately see plate tectonics. Also, there are no subduction zones on the continents, just complicated compressional mountain-building areas like the Himalayas.

In contrast, if you go to the oceans, first of all, the age of the oceans is relatively young. Instead of being billions of years, it's maybe 70 million years on average, with zero-age plates forming at the ridges. And if you look at the topography, or all types of geophysical data, you see that the tectonics is very clear in the ocean basins. You see the seafloor spreading ridges, where the plates are coming apart. This spreading ridge boundary encircles the planet like the seams on a baseball. Then these ridges are offset by transform faults that are perpendicular to the ridges, and they form this beautiful segmentation: ridge, transform, ridge, transform, beautifully perpendicular, just like in the textbooks. It's really, really simple, clear plate tectonics. And then of course you've got the subduction zones, where the plates are diving back into the planet. They have deep trenches and large megathrust earthquakes. So, looking at all these things, you can basically just see plate tectonics in action. That's the basic idea of the oceans, and from the altimeters and the mapping from the ships, we could see all these things.

Oliver Strimpel: So we understand these things pretty well today, the basic ridges, transforms and subduction zones. But with your global sea floor map, were you able to see new regions or aspects of those classic ridges and subduction zones that we hadn't been able to map with ships before?

David Sandwell: Yeah. I mean, because the ship coverage is only 20 percent, even with the data from Geosat, and back with Seasat, we saw lots of new things: new fracture zones, new ridges and so on. But as time went on and these data got better and better, we saw these things in much better detail. One of the things I haven't pointed out is that when you want to do global reconstructions of the plates, to figure out where they went and how old the sea floor is in a particular place, you take a combination of things. One is where the fracture zones are. The fracture zones are the extensions of the transform faults that used to go all the way across, say, the Mid-Atlantic Ridge, from South America to Africa. There are beautiful fracture zone lineations that tell you how the plates moved apart over the past 200 million years. We get a lot of that from ship data, but we also get a lot from satellite data. And then the other thing is marine magnetic anomalies, which provide the timing of the opening and spreading and so on. So in the Geosat era, we had these two data sets, and these were really used to do the very detailed plate reconstructions that we have available today.

Oliver Strimpel: Okay. So in addition to seeing more of, and in higher resolution, the classic features of plate tectonics, such as the subduction zones and the ridges and the transform faults, did your map uncover features that we had not seen before?

David Sandwell: Yeah. With the latest data from CryoSat and AltiKa and so on, this improvement in resolution has allowed us to look at much smaller things, and the new things are really in the deep oceans and in the older oceans, away from where the ships have explored. Most of the shipboard surveys are concentrated on the seafloor spreading ridges, because that's where all the interesting biological activity is, and the thermal vents. But we're now able to look at what these features were like in the past, in other words, the different types of tectonics.

For example, there are things called microplates, and there are microplates active today along the East Pacific Rise axis, the Easter microplate, for example. But a lot of these microplates don't grow into large plates. A microplate might be the size of Texas or West Virginia. When they don't grow into a large plate, they get cast off onto the older seafloor, but they leave behind a signature in the tectonics, the topography and the gravity, and so we can now see all these microplates in the satellite altimetry data. There's one really nice example in the southern Indian Ocean that we can see in the CryoSat data. It was previously unmapped, and there's one ship track that goes across it, where you can sort of ground-truth the microplate.

The other interesting features are all at small scale: the abyssal hills. Abyssal hills are interesting because they're the most common landform on the planet, but people don't really know much about them. These are hills that are created during the seafloor spreading process. You get small hills that are parallel to the ridge axis. They move off the spreading ridge as you spread, and they leave behind this fabric, this corrugation, and these are everywhere. We can just now start to resolve the abyssal hills in the deep oceans, and that tells us a lot about the seafloor spreading direction and seafloor spreading rate, because they vary with rate and so on. So we can look at the abyssal fabric of the sea floor with altimetry.

The other thing is that the gravity field is sensitive to the topography, but if you have sediments that come down onto the sea floor and cover the topography, maybe a kilometer thickness of sediments flowing down from rivers or just raining down through the ocean, these cover the topography, and so the sea floor is flat. A lot of the old sea floor is flat. But the gravity field is sensitive to the density changes, and there's a density difference between the sediments and the underlying original seafloor topography. So you can see that underlying topography in the gravity field and learn about tectonics in the past and so on.

Oliver Strimpel: So you can see through even thick sediment layers?

David Sandwell: As the thickness grows, you're going to lose the resolution, because you get deeper, but two or three kilometers thick, even thicker. One thing: on the continental margins, for example, there's a distinction between oceanic crust and continental crust on the edges of the continents, and you can see that boundary in the altimetry and map it out.

One other thing that's sort of exciting about the new altimetry is that it's revealed something that we call seesaw propagators, and I'll try to explain this without any graphics. As I said previously, normal sea floor spreading has ridges and transform faults, and they're orthogonal. The ridges are spreading laterally, and the transform faults are sliding beside one another in a strike-slip motion.

Oliver Strimpel: Those leave behind these long fracture zones that go all the way across the ocean basins at right angles to the ridges.

David Sandwell: That's right. And that's the normal sea floor spreading fabric. But what we see now are these things called propagating rifts, seesaw propagators. A propagating rift is a transform fault, just like I described, but it's a shorter one. It doesn't have a long offset, so it moves in the direction parallel to the ridge axis, and rather than leaving behind a single long fracture zone, it leaves behind a V shape, like a motorboat wake. These propagators have been pretty well known since the early days of plate tectonics. But what we're seeing now is that the propagators will propagate in one direction. We can now look off-ridge and see the scars left behind by these propagators: they'll move in one direction, but then reverse direction, go the other way, reverse direction, go the other way, and they leave behind a W-shaped anomaly on the sea floor. That's what we call seesaw propagators. These are pretty much everywhere on the intermediate-rate spreading ridges. They account for almost a quarter of the length of all the fracture zones. So if you take the length of all the fracture zones, it's a quarter of that. So it's a really pervasive tectonic feature that really wasn't known before CryoSat, because we didn't have the off-axis data that we have now.

Oliver Strimpel: Do we understand why they propagate, and why they might go back and forth in the direction of propagation?

David Sandwell: Actually, no. If you look at the theories for propagating ridges that were put forward in the '90s, they explained that propagating ridges propagate down a regional topographic gradient. Seafloor spreading ridges can be higher at one end than the other, and the propagator will propagate down that gradient. But we see examples where the seesaw propagators are going in both directions: they're going down the gradient and they're going up the gradient. So that theory is out the window. And there are also theories that they propagate away from hotspots, but that can't be true, because they're going in both directions. So we're scratching our heads trying to come up with an understanding of why these things form.

Oliver Strimpel: Have some other puzzles emerged from the new global map of the seafloor?

David Sandwell: Yeah, one thing I didn't talk about, but there are lots of interesting things: seamounts. A seamount is an undersea volcano that's more than 1,000 meters tall, more than a kilometer tall, and seamounts are pervasive on the sea floor. There are a few really tall seamounts. Hawaii is a gigantic seamount, maybe eight kilometers tall. And then the number of seamounts increases as you go to smaller scales. With the old Geosat data and mapping from ships, you could see all the seamounts greater than about one and a half to two kilometers tall. There are about 25,000 of them. But as the accuracy and resolution of the altimetry increase, we might see double that number. We're going to see lots and lots more seamounts, and really this will help us try to understand why they form. They aren't part of the normal sea floor spreading process; they form off the ridge axis.

But the big question is where the magma comes from. To form a seamount, you have to have magma. So the question is, is there partial melt or magma everywhere beneath the ocean lithosphere? No one knows. That would be pretty interesting if there was. The other thing is you need to have a crack. You've got to have a way for the magma to get from the bottom of the plate up to the top. And so there's a debate going on. There's an interesting webpage called mantleplumes.org that you should all look at. There are people who don't believe in hotspots. Hotspots are one of the generators of large seamounts, like Hawaii, but some people don't believe in hotspots. An alternative mechanism is to have extensional stress in the plate that opens these cracks and allows the melt to come to the surface. So by mapping out these new small seamounts and looking at the tectonics, we may be able to sort this out. I mean, I'm sort of a believer that both mechanisms are going on. But then you could ask, well, which seamounts formed by plumes and which seamounts formed by extension? And so it's sort of a new area of research.

Oliver Strimpel: Which brings me to my last question, in fact, which is: if you had unlimited resources to play with, which problems would you address, and how would you go about it?

David Sandwell: Okay, that's a great question. Here are two problems I think we could address with unlimited funds. First, no surprise, I would like to map the whole ocean using ships. You could have a Mission Planet Ocean, sort of like the way NASA does missions to other planets. Nowadays you could do this with large unmanned, remotely operated ships. You can send these things out full of fuel, with multibeam echo sounders that can map the tectonics. You could also do physical oceanography mapping and biological oceanography and so on. They could do the standard things that ships do nowadays, but on a global basis. And I would focus on the Southern Hemisphere: the South Pacific, the southern Indian Ocean, the South Atlantic. These are areas where nobody goes anymore. In the old days of exploration, the people at Lamont and Scripps sent their ships out to these remote areas with really no understanding of what was going on. It was just pure exploration, and I think we need to do something like that again, assuming we had infinite funds. So that's the first thing.

The second thing is more related to hazards, and it's more mainstream, and that's understanding megathrust earthquakes. The largest earthquakes occur at the subduction zones, where the plates are diving back into the trench, into the Earth. For example, the 2004 Sumatra earthquake, magnitude 9.2, generated a tsunami that propagated all across the Indian Ocean basin, killing more than 200,000 people. We don't understand this process of generating tsunamis very well. The issue is that tsunamis are generated when the earthquake occurs and you get uplift of the sea floor, but that sea floor is way down deep in the trench. You can't get there. So we need to develop methods for exploring the deepest parts of these trenches with all kinds of instruments, you know, multibeam echo sounders, seismometers. But what we're really missing is a way to put GPS on the seafloor. In other words, we want to measure the motion between earthquakes. How is the stress building up? There are technologies being developed to measure motion acoustically, and with pressure sensors and so on. There's actually an initiative to do this offshore Cascadia. It's sort of a small initiative, but this kind of thing needs to be ramped up. You said we had infinite money: we would just need to build better instrumentation, deploy it, and try to understand this tsunami generation process.

Oliver Strimpel: David Sandwell, thank you very much.

David Sandwell: My pleasure.

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