This transcript was generated automatically and lightly edited for readability.

Oliver Strimpel: This is Geology Bites with Oliver Strimpel. One of my first geological field trips was to Siccar Point on the southeast coast of Scotland, the site of James Hutton's famous unconformity. The underlying rock there is a 440-million-year-old marine turbidite, and I soon discovered similar-looking turbidites on other trips, from Greenland to South Georgia. I was struck by their ubiquity in the geological record. So how do they form, and why are they so widespread in the record? The answer lies on the sea floor, in one of the most powerful and least witnessed phenomena in geology: turbidity currents, massive underwater avalanches of sediment that can travel hundreds of kilometers, snap cables, and reshape the deep ocean floor in a matter of hours. In 2006, a single earthquake off Taiwan triggered a cascade of turbidity currents that sequentially severed more than 20 undersea cables, cutting internet and phone connections across Southeast Asia for nearly two months. And yet, because these events happen in the deep ocean, often without warning, we've rarely observed them directly.

Today's guest is one of the few researchers who has, sometimes more directly than she'd have liked. Esther Sumner is an associate professor of geology and geophysics at the University of Southampton. She studies turbidity currents by combining direct sea floor monitoring with laboratory experiments and field studies of ancient turbidite outcrops. In our conversation, she describes what it's like to instrument an active submarine canyon, what these flows have revealed about the way sediment moves across the sea floor, and the day her team accidentally flew an underwater robot into a live turbidity current. Esther Sumner, welcome to Geology Bites.

Esther Sumner: Thank you very much for inviting me.

Oliver Strimpel: So what exactly are turbidites?

Esther Sumner: Turbidites are the deposits of a type of sea floor avalanche called a turbidity current. Turbidity currents are in the same family of flows as things like snow avalanches and pyroclastic flows, so that gives you an idea of what they look like. The types of material they carry range from clay-sized material to sand-sized material, and even up to boulders and bigger. These are really big flows that carry lots of sediment. They can travel big distances, hundreds of kilometers, sometimes even in excess of 1,000 kilometers, and they can have a range of durations, which can be hours or in some cases even days.

Oliver Strimpel: So when we look at a rock in the field, how can we distinguish a turbidite from other kinds of marine sedimentary rocks?

Esther Sumner: Generally, if you're looking at a sequence of turbidites, so a turbidite succession, you're going to be looking at a sequence of interbedded sandstones and mudstones. Now, if we're thinking about an individual turbidity current event, or an individual turbidite, then I guess classically we would identify this as a normally graded sand with a mud cap. Normal grading means big grains at the bottom and small grains at the top. Then there's a characteristic series of sedimentary structures, with structureless sand at the bottom, overlain by plane-laminated sand, overlain by ripple-scale cross-laminated sand. I think in reality it's actually quite rare to see exactly that, and deposits are much more varied. That probably reflects the fact that that deposit would form if we had a very simple, dilute, decelerating flow, and that's just not always the case.

Oliver Strimpel: What triggers a turbidity flow?

Esther Sumner: My simple answer would be that we know that things like earthquakes can trigger turbidity currents. Big storms too: if you've got a big storm, you might have storm waves basically pounding the continental shelf, resuspending sediment that can form a turbidity current. River discharge can cause turbidity currents, especially at deltas, for example. There it might be that the sediment-laden river water is actually plunging down into the ocean, forming a turbidity current, or it might be that the delta front, which is a relatively steep area, collapses to form a turbidity current. Pyroclastic flows entering the sea can generate turbidity currents, and so can things like fishing, if there are human impacts like trawling in an area. Basically, anything where you're destabilizing some sediment in water can form a turbidity current.

I think I should probably make the point that just because you have one of these triggers doesn't mean you'll get a turbidity current. Actually, many times when we're measuring turbidity currents, we don't really understand what triggered them. What I think we're starting to understand is that what might be equally important, or maybe even more important, than triggers is preconditioning. Preconditioning often just means what sediment is there, available to fail in the event of one of these triggers happening. So imagine you have an earthquake, but there isn't a lot of sediment available to be destabilized; then you won't necessarily generate a turbidity current. So I guess, in a simple way, it's this combination of preconditioning combined with some kind of trigger.

Oliver Strimpel: Is there a typical setting where turbidity currents happen, like on continental shelves or on the flanks of seamounts?

Esther Sumner: In terms of settings, anywhere really where you've got a sediment-water mixture going into another body of water. So you can form turbidity currents in lakes, you can form them in fjords, you can form them in the oceans. I think most commonly we think about them occurring in the oceans, and there they can flow all the way from the coastlines, across the continental shelves, right out into the deep sea. One of the interesting things about this is that they can actually do this on remarkably low gradients. When we talk about slopes in the ocean, these are normally really quite pathetic slopes. Let's say we're in a submarine canyon. There's probably a slope of, say, 2 degrees, maybe a little more, a little less. If we exit that canyon into deeper waters, so we go out onto a basin plain, you're looking at a slope of 1 degree, or even much less than a degree, maybe half a degree or less. I often say to people that, as far as I'm aware, something like a Premier League football pitch has that kind of slope, of, say, half a degree, for drainage. And we would think of that as flat. But these are the kinds of slopes that turbidity currents can travel along, and they can travel great distances, maybe hundreds of kilometers, on those slopes, going at meters per second.

Oliver Strimpel: That's interesting. So I guess this is all because these flows are happening underwater, where, because of the buoyancy effect, you can think of the gravity as being much lower. And so that can enable them to move down very gentle gradients and also have these much longer durations that you mentioned.

Esther Sumner: Yeah, certainly that has to play a part. You've got your sediment particles and they're settling in water. If we put the same particles in air, they fall out more quickly. The thing that's driving these flows is the density difference between the flow and the surrounding ambient fluid, in our case seawater. So that's one part of the story, but I think there's more to the story. There's some really interesting recent work showing that some turbidity currents can reach a steady state, and steady state means a more or less constant velocity. When they're doing that, they must be achieving a more or less constant density, because that's what's driving them. The way that we image them doing this is that, in the flows that do this, at the front of the flow there's a dense near-bed layer, and that dense layer seems to be able to erode sediment into itself. So it's effectively feeding the flow. Then two things can happen. If it keeps feeding itself, your flow is going to speed up and get faster. But what often seems to happen is that it must then chuck some of that sediment into the rear parts of the flow, where it can get deposited out. And so the flow as a whole is maintaining this constant density, and therefore constant velocity. The long durations are probably somewhat an effect of reaching this constant velocity. The other thing that's happening is that the flow head tends to accelerate away from the body of the flow, so your flow stretches out and gets longer. It's quite weird, but through time, your turbidity current is getting longer and longer. This gets called autosuspension. It's actually something which was theorized in the 1960s and 1980s, but very recently we have data from direct monitoring of turbidity currents that suggests it does indeed happen, and it might be quite an important reason for how they can travel such long distances.

Oliver Strimpel: And when we're talking about distances and lateral spatial scales, what are we talking about?

Esther Sumner: In terms of distances traveled, it depends on your turbidity current, right? Not all of them go these great distances, and that's quite interesting in itself. But the really long-duration flows can be traveling hundreds, or in excess of 1,000, kilometers. How far they spread laterally really depends on their setting. If they're confined within a submarine canyon, they can't spread laterally. Once you get out of the mouth of the canyon onto a submarine fan, where they can spread out, then it's somewhat going to depend on the characteristics of that submarine fan, because even in that setting you have some confinement from channels. But individual flows could be many kilometers wide. So these are big, big flows.

Oliver Strimpel: And how long can they last?

Esther Sumner: In terms of durations, when we're monitoring them, I think we have to say that the typical duration for a turbidity current is on the scale of hours. But in certain settings, Congo Canyon off West Africa being one of them, turbidity currents have been monitored for much, much longer durations, so days or even in excess of weeks. These are really long-duration flows, and these are the flows that are going really long distances.

Oliver Strimpel: You mentioned the flows off the mouth of the Congo River. In fact, there's a giant submarine fan in the Bay of Bengal at the mouth of the Ganges River, which I suppose is filling up with sediment eroded down from the Himalayas. Are those kinds of huge submarine fans also built up by repeated turbidity currents?

Esther Sumner: Yes, yes, absolutely. Basically you've got the aggradation from lots of turbidity currents over long time periods. With something like that system, you've got the Ganges and other rivers, which are effectively feeding sediment onto your coastline, and then that sediment ends up in submarine canyons. Submarine canyons are basically just giant canyons on the sea floor. If you imagine a canyon on land, they look somewhat like that, often on a slightly larger scale. So your sediment ends up in your canyon, gets destabilized and forms a turbidity current. When that turbidity current is in the canyon, it's quite an efficient place for it to be. It's confined, so it keeps going. Once it reaches the canyon mouth, suddenly it's unconfined and it can spread out, and so it slows down, loses energy, and starts depositing its sediment load. And so you end up with this basically cone-shaped pile of sediment at your canyon mouth, called a submarine fan.

This is forming over really long time scales, so you're looking at millions, if not tens of millions, of years. So lots and lots of turbidity currents are stacked up, and these are huge volumes of sediment. They're some of the largest sediment accumulations on Earth. The Bengal Fan that you mentioned is in excess of 3,000 kilometers long, 1,000 kilometers wide, and in places greater than 15 kilometers thick. So that's a massive amount of sediment. And you're saying this sediment comes from weathering of the Himalayas. That's quite interesting in itself: this big pile of sediment on the sea floor is basically an archive of how the Himalayas have eroded in response to changing tectonics and climate. So they're really interesting archives of tectonic and climate histories in the past. But if we think about the present day, one of the things that turbidity currents might be doing is feeding pollutants to that submarine fan. If we've got pollutants mixed up with our sediments on the continental shelf, and those might be things like plastics or pesticides, they also have the possibility of making their way all the way out to the submarine fan.

Oliver Strimpel: Let's talk about how you go about studying turbidity currents, both with your experimental modeling and with your work in the field.

Esther Sumner: I guess I'll start with the field. I may be a little bit weird, because when I talk about the field, sometimes I'm talking about on land and sometimes I'm talking about fieldwork at sea. On land, we're talking about hiking up mountains and looking at outcrops. On the sea floor, we can do something really similar by taking sediment cores. Basically, we throw a big metal tube with a weight on it at the sea floor. It sinks into the layers of sediment beneath the sea floor, and hopefully we pull out a tube of sediment, split it open, and look at the layers within. My reason for starting with field deposits is that I think it's a really important point to make that this is the biggest data set we have about turbidites, which tells us about their characteristics and their dynamics. And it's also a data set that spans geological time. In terms of how we actually look at turbidites in the geological record, it's really simple. We look at things like grading patterns, we look at things like sedimentary structures, and then we try to interpret what that means about how that flow was behaving however many million years ago.

Now, there's a slight issue in doing this, and this was once illustrated to me quite amusingly on a conference field trip. We were up in the Italian Apennines, where there are some really exquisite sequences of turbidites. All these people whose job is to research turbidites are standing around this outcrop. We're looking at this outcrop, we're making really similar observations to one another, but actually some of our interpretations are wildly different. I think this illustrates the problem with trying to interpret flow dynamics from deposits: you're trying to interpret the dynamics of something of which we actually have very few observations.

For me, this is where experiments can be really, really useful. There are different ways we can do experiments on turbidity currents. One way is to generate small-scale turbidity currents in the laboratory, so actually generating a little density current. The other thing we can do, which is what I've focused on more, is to look at aspects of their behavior. In my own experiments, what I've tended to do is generate fast-moving slurries of sand and clay and then change how the flow is behaving: artificially change the velocity of the flow, artificially decelerate the flow in different ways, and then look at how that information about the flow gets recorded in the deposit. Because then you can hopefully take that information into the field to help you make better interpretations. What I would say, as someone who's done quite a few experiments, is that they also make me slightly nervous, because you always have this slight worry about how well what I'm doing in the lab actually scales to these full-scale flows in the real world. So I guess where we're at now is that in the last couple of decades we've started being able to directly monitor turbidity currents in the oceans in some situations. I can talk a little bit more about that if you like.

Oliver Strimpel: Yeah, that would be great. I'd love to hear about how you monitor turbidity currents in real time in the ocean.

Esther Sumner: The main way we do this is using moorings. A mooring is something we put on the sea floor. It's got a big weight at the bottom, some flotation at the top, and basically a vertical wire in between, and we can attach instruments to that wire. The overall mooring is a couple of hundred meters high, and it sits on the sea floor. One of the main instruments that we use is called an acoustic Doppler current profiler, or ADCP. These are basically the speed traps of the ocean. They work in quite a similar way to, say, a police speed trap. But whereas the police officer is using radar, firing out radio waves that bounce off moving vehicles, we're using sound pulses. This instrument is basically looking down towards the sea floor, sending sound pulses down into the water column. They bounce off moving particles, and we can build up a really nice picture of the velocity structure of the flow.

One major issue we have with these instruments is that they don't penetrate very well into concentrated layers within flows, which we keep finding in the oceans. The other issue we sometimes have with them is that they actually get transported by the flow, because obviously these are destructive flows that we're trying to measure. So we have some other quite neat bits of kit to deal with this. We basically have smart boulders. Imagine a smart boulder is about the size of a basketball. It's got its own pressure sensors, so it knows the water depth it's at, and it's got an accelerometer in it. That accelerometer is basically telling us how the boulder is moving. Is it sliding in the flow? Is it rolling in the flow? So what's the nature of the flow that it's moving within? And other people use all sorts of instruments: hydrophones to listen for turbidity currents, seismometers to feel turbidity currents.

One thing in all of these cases, though, is that we have to go somewhere we know has frequent flows, because we have to leave those instruments on the sea floor for quite long periods of time and then go back and collect them. Or, in the case of a mooring, if we've been unlucky, it's been damaged, cut loose and floated to the surface. Or, in the case of some instruments like those smart boulders, they're actually designed so that with time they'll pop some weights out and float to the surface. If we're working in a coastal area, they float to the shoreline. They've got GPS on them, so we can go and collect them. In reality, beachcombers usually collect them for us. So those are the approaches to understanding flow dynamics. I think it's really this combination of fieldwork, experiments and, more recently, being able to start directly monitoring flows.

Oliver Strimpel: So, as I mentioned in the introduction, you actually use remotely operated vehicles, sort of underwater robots, to study turbidity currents, and you even flew into one once. Can you tell us about that?

Esther Sumner: We do use remotely operated vehicles to, for example, take sediment cores from the sea floor. There are methods by which we can attach coring devices to ROVs and collect turbidites from the sea floor. We did have quite an exciting day in the office, so to speak, back in 2013, when we did indeed get a remotely operated vehicle entrained in a turbidity current. At the time I was working for an institute in California called the Monterey Bay Aquarium Research Institute, and we were doing some work in Northern California. We were actually meant to be collecting samples of gas hydrates, but it was really bad weather. It was a really awful set of days, to be honest. What we realized is that there was basically one place we could go and do some work on this particular day: we could collect some cores using this remotely operated vehicle from Mendocino Canyon.

I should probably explain what a remotely operated vehicle is. This is basically an underwater robot. You can think of it as being very roughly the size of a car. It's an unmanned vehicle, so there's nobody inside it, which was good news on this day. It's attached to the ship by a wire, and it's controlled from a control room on board the ship. So imagine on board the ship you've got a room with pilots looking like they're playing video games, flying the ROV around on the sea floor, and scientists telling them what they'd like them to do.

So we've basically flown this remotely operated vehicle down into this canyon. We're sitting on the sea floor. Now, obviously it's very dark on the sea floor, but these vehicles have incredible lights, so normally we can see many meters ahead of us. That was our situation when we landed. But then suddenly it all goes dark, we're basically surrounded by sediment, and there's a strong current coming down the canyon. The pilots are complaining that it's really quite difficult to keep the ROV in position. So, in order to try and stabilize the vehicle, they lift it off the sea floor. The moment they lift it off the sea floor, the vehicle starts traveling down the canyon at nearly a meter per second, despite the fact that they're trying to fly it up the canyon at about a meter per second. So we got some very interesting observations, basically from the perspective of being a particle within the turbidity current. They did get the vehicle out of the turbidity current, and I think they described the situation quite well, because they described it as like flying an ROV in a tornado. So that's what it feels like to be a particle in a turbidity current.

There were some instruments on this vehicle, so we were actually able to get some quite nice data about the stratification in this flow. There were basically two layers to the flow: a layer at the bottom of the flow, which was relatively dense, and then a really dilute cloud on top of that. That near-bed layer was a couple of meters thick, whereas the cloud on top of it was maybe in excess of 100 meters thick. That actually ties in more generally with other studies that we've done with more conventional methods in other canyons, where we do see these dense layers at the base of flows, usually much denser than the one we measured with the ROV. And, as I said earlier, they're potentially a really, really important part of how these flows are driven.

Oliver Strimpel: When you were flying this ROV, what was the goal of the project? What were you trying to discover?

Esther Sumner: We were meant to be doing something entirely different. We were meant to be collecting gas hydrates from a different part of the coastline, but that just was not possible. What we'd been doing at that time, as part of a larger study, was collecting sediment cores from lots of canyons along the California coastline in order to look at the nature of turbidites, and therefore turbidity currents, within those different canyons. We've got some canyons where the heads are very close to the shoreline and some canyons where the heads are away from the shoreline, and we were looking at what the different processes acting in those canyons are. So really what we realized was, well, we don't want to do nothing on this day on this ship. It's really expensive running ships. So rather than doing nothing, we could go and collect some cores from this canyon. And it just so happened that we encountered a turbidity current. So I think it was quite fortuitous. The vehicle survived, and so did everybody.

Oliver Strimpel: At those speeds, a meter per second, is the water flowing in a turbulent way, or is it laminar flow?

Esther Sumner: The flow was probably going about two meters per second, because we were actually trying to fly up the canyon at about a meter per second and we were heading down at a meter per second. It's quite hard to say what that dense layer near the sea floor was doing exactly, because a lot of the time the vehicle was actually thicker than that layer. But certainly on the interface between that dense near-bed layer and the more dilute cloud on top, there was actually quite a stable waveform, between the layers of fluid of different densities. And then that upper layer was very turbulent, which relates to that comment about flying in a tornado. They were just getting thrown all over the place while they were trying to keep this vehicle stable.

Oliver Strimpel: And were your experiments in the lab roughly consistent with what you were finding in the field there?

Esther Sumner: I guess the interesting thing there is that, if you compare it to lab experiments, the flow we measured in the field looks quite like what you get if, in the lab, you generate a turbidity current and then intrude another turbidity current into it. So you've basically got pulses of turbidity current, which is probably not uncommon within submarine canyons. There was a storm going on at this point, so you may well have been destabilizing the sediments at the head of the canyon. You won't necessarily do that all in one go, so you may actually generate multiple turbidity currents that intrude into one another. The structure of the flow we measured in that canyon looked quite like the structure of flows seen in labs, although all the lab experiments that show that, that I'm aware of, are actually done with dense fluids, not with sediment-laden water. And I should say that the near-bed dense layer we saw in Mendocino Canyon is, I think, quite different from the near-bed dense layers we've seen in places like Monterey Canyon and Congo Canyon, which have much, much higher concentrations than what we saw with the ROV, thankfully. Because I think we wouldn't have had an ROV at the end of it if there had been those kinds of concentrations.

Oliver Strimpel: There's a great deal of human infrastructure on the sea floor, such as fiber-optic cables and oil and gas installations. Are they affected by turbidity currents?

Esther Sumner: Yeah, absolutely. This is a really important reason for understanding the dynamics of these flows. There's a lot of infrastructure on the sea floor, anything from fiber-optic cables to high-voltage cables and oil and gas pipelines. If we think about fiber-optic cables, they are basically the network that underpins the internet. Those fiber-optic cables typically aren't protected on the sea floor in deep water. They're just lying on the sea floor, and they're vulnerable to lots of things. They're vulnerable to things like ships' anchors or trawling from fishing boats, basically dragging them and breaking them. They're also vulnerable to turbidity currents. The issue we have with turbidity currents is that, whereas if an anchor drags and breaks a cable you break one cable, with a turbidity current you can break lots and lots of cables at the same time. And that causes further problems, because what you normally do when you break a cable is reroute the data traffic somewhere else around the network. If you've just broken a lot of the network, you can't do that.

A classic example of this is Taiwan in 2006. Taiwan experiences a lot of natural hazards, from earthquakes to typhoons, and both of these things can be responsible for generating turbidity currents, or certainly be associated with turbidity currents. In 2006 there was an earthquake, and in association with this earthquake, over 20 cables on the sea floor were sequentially broken. So you've just broken a massive bit of the network. Temporarily, communications from Southeast Asia were basically severed. Then, on a longer time scale, as they were trying to mend these cables, the internet was slowed down, and I think that was for nearly two months. And you see the impacts of that on things like financial markets. So yes, these cables are definitely vulnerable to turbidity currents, and there are similar issues with oil and gas pipelines. This is one of the really important reasons we study them.

I should say, and I should probably say it quietly, that there is a positive to these cables getting broken. It's not positive for many people, but it is positive if you study turbidity currents, because they give you quite a nice data set about the speed of some of these flows in the deeper ocean, say on submarine fans, which otherwise we aren't really monitoring. If you imagine you've got a series of cables, you know when they got broken and you know the distance between the cables, so you're getting an idea of the velocity of the flows that are breaking them. So there are some positives.

Oliver Strimpel: Does your work also help these companies figure out where not to put infrastructure?

Esther Sumner: Yeah, that's a really important part. They've always got to decide where they're going to route this pipeline or this cable. And I think it's a really difficult job to do, because you've got this trade-off between hazards to your cable and the length of your cable. So often what happens, if it's known that you've got, say, an active canyon-fan system with turbidity currents, is that the cable needs to be routed into deeper water, which is going to cost a lot more money. Or, for example, some oil and gas pipelines are actually directionally drilled into bedrock beneath sea floor canyons, which is a phenomenally expensive thing to do.

Oliver Strimpel: What are you working on at the moment?

Esther Sumner: I guess a couple of different areas. One of the really interesting things at the moment, as we get more data sets from direct monitoring in different places, is understanding the diversity of turbidity currents that exist. If you think about something like a river, we all know what a river looks like, but you also know that if I look at a bedrock river in the mountains versus, say, a river on a plain like the Mississippi River, those two rivers look quite different from one another. I think this is what we're starting to see with turbidity currents: there are some similarities when we look at them in different places, but there are also some differences. It would be great if we could predict their character with less need for monitoring, and that would really, really help with what you were talking about a moment ago, which is how we design things like cable routes. So that's one area of interest.

Another is understanding land-detached canyons. I don't think I've mentioned these yet, but we often divide sea floor canyons into being either land-attached or land-detached. A land-attached canyon would be something like Monterey Canyon, where the canyon head is really, really close to the shoreline. It's in really shallow water, it's almost up at the beach, and there's a really obvious sediment source for the turbidity currents. You have longshore drift that's moving sediment along the coastline, and that sediment effectively falls into the hole in the ground that is the canyon. But actually most canyons around the world are land-detached canyons, so their canyon head is detached from the shoreline. For a long time we assumed that these types of canyons were inactive at present-day sea level, so if they had turbidity currents in them, they'd be quite small turbidity currents and we didn't need to worry about them. But more recently we've realized that this actually isn't the case. They can have quite fast and destructive turbidity currents in them.

For example, this summer I'm part of an expedition which is going out into the Atlantic to a canyon called Whittard Canyon. Whittard Canyon is over 300 kilometers southwest of the UK, so it's very, very land-detached. There's a big distance between it and the shoreline. This is an expedition led by my colleague Mike Clare, who works at the National Oceanography Centre in Southampton, and as part of his project we're going to put an array of moorings into this canyon. We already know from previous work that we've got fast and destructive turbidity currents in there, but it's really work in progress to understand why: what's actually triggering them, how far down the canyon system they go, and how they actually behave as they do so.

Oliver Strimpel: Esther Sumner, thank you very much.

Esther Sumner: Thank you very much for having me. That was a lot of fun.

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