This transcript was generated automatically and lightly edited for readability.

Oliver Strimpel: This is Geology Bites with Oliver Strimpel. We've all heard of the San Andreas Fault that runs along the coast of California. It represents the boundary between the North American plate and the Pacific plate. Its 800-mile-long extent is visible from space, and it has caused some powerful earthquakes, one of which devastated San Francisco in 1906. Most faults leave little or no trace of their locations on the Earth's surface, but it is still vital to know where they are, for both hazard assessment and scientific reasons. Rufus Catchings has been hunting down Californian faults for over 40 years. He is a research geophysicist at the US Geological Survey. Rufus Catchings, welcome to Geology Bites.

Rufus Catchings: Thank you, Oliver. It's a pleasure to join you.

Oliver Strimpel: In your research, you've used many different approaches to pin down the locations of faults. Do these all rely on seismic waves, or do you use other methods as well?

Rufus Catchings: Most methods I use involve seismology, but I also use other methods as appropriate. For example, in many places we use geomorphology, or changes in the Earth's landforms or surface, to infer that there are likely faults there. Sometimes we can see them at the surface, and it's obvious, and we act accordingly. For example, using geomorphology: sometimes there's a long linear valley within a mountain chain, and we suspect that those are likely faults. There are geologists who specialize in geomorphology, and they also help me in my efforts.

Oliver Strimpel: OK, let's talk about the seismic methods that you use and how they work, and then we'll talk about the results. There are two kinds of seismic wave: primary, or pressure, waves, called P waves, which are akin to sound waves in air, and secondary, or shear, waves, called S waves. Both types of waves are affected by the presence of faults. Can you tell us how they're affected?

Rufus Catchings: Yes. Within the subsurface, S waves slow down once they encounter a fault zone. So relative to the rocks on either side, the velocity of the S waves is much slower in the faulted rock body than in the outside rocks. Faulting essentially damages the rock, and this changes the rigidity of the rock, which makes the velocity slower. It's the same for P waves, but one complicating factor is that in the near surface, groundwater can saturate these materials, the rocks as well as the faulted material, and that actually causes an increase in velocity, which makes it a little more complicated when we use just P waves alone.

Oliver Strimpel: OK, so how does this change in the velocity of the S waves and the P waves enable you to pin down the location of faults?

Rufus Catchings: What we typically do is put out an array of seismometers across the fault, and then we input some type of source into the ground so that we can measure these waves. From that, we measure the travel times, usually to within milliseconds, and then we can locate the faults themselves. Using methods like tomography, we can actually see, or get an image of, where these low velocities are located. We can do the same thing for both P and S waves, but one of the most powerful techniques is to use the ratio of the two, P waves to S waves. Because S waves are slower and P waves are faster, even if it's saturated, we get Vp/Vs ratios that are very high in fault zones, so that shows us very nicely where the fault is in the subsurface.

Also, a fault is typically a barrier to groundwater flow, and we use these seismic waves to map out the top of the groundwater. The top of the groundwater velocity contour is typically 1,500 meters per second in sediments. Since a fault is often a barrier to the flow of groundwater, on one side of the fault the groundwater will be lower than on the other, and using tomographic methods, we can actually see this step. That's another indicator of faulting. So there are multiple methods that we use. We use the low P-wave velocities, the low S-wave velocities, the high Vp/Vs ratios, and the sudden change in depth of the 1,500-meters-per-second velocity contour. All of these are strong evidence of faulting, and when you have them all in one place, you're quite certain of the location of the fault.

Oliver Strimpel: And are the waves that you're picking up typically waves that have been refracted by the fault zone, or reflected at the edges of the fault zone?

Rufus Catchings: For the tomographic methods, we're using waves that have been refracted as they move through the subsurface, but we do use the reflection method as well.

Oliver Strimpel: Tell us a bit about the reflection method.

Rufus Catchings: The reflection method is a very powerful method, and it's very precise. The oil industry has used this method for exploration for oil for many, many decades. If we do it at very high resolution, what we can see is that the seismic waves go down and reflect off surfaces. Faults also typically cause the layers of the Earth on either side of the fault to be offset somewhat, and the reflection method is very good at showing a visual image of this offset. But unfortunately, you can't use it everywhere. In many places, you don't have this layered subsurface.

Oliver Strimpel: What kind of spatial resolution can you get for these fault locations?

Rufus Catchings: Well, it really depends on the method that we're using. As I mentioned, reflection seismology is one of the most accurate, particularly at greater depths. If we use very dense spacing of our seismometers, and very high-frequency seismometers, and the energy that we input into the ground is high frequency, we can typically image quite deep, to within a meter or so of where the fault is. But as we get deeper, we lose frequency, and so the resolution becomes less, and our accuracy is probably several meters to several tens of meters at great depth. But there are other methods as well that we can use to get high accuracy, particularly in the shallow subsurface.

Oliver Strimpel: How deep can you track a fault? I mean, if a fault goes all the way down to the Moho, for example, could you see it all the way down?

Rufus Catchings: Yes, you can. The resolution, of course, once you get down to 30 or 40 kilometers, is off. But yes, you can see these faults quite deep.

Oliver Strimpel: I'm reminded of a recent episode of Geology Bites with Bob White, in which he described using a dense array of seismometers to pinpoint the locations of thousands of microquakes and track the movement of magma within the crust of Iceland. Do you also rely on naturally occurring earthquakes as the source of the seismic waves?

Rufus Catchings: Yes, we do. We use both naturally occurring earthquakes, or ambient noise, and artificial, man-made sources.

Oliver Strimpel: How do the man-made sources work?

Rufus Catchings: In the shallow subsurface, it's as simple as, say, striking the ground with a hammer. We just input energy into the ground, and that generates seismic waves. But there's also another way: we use what are called accelerated weight drops. It's essentially a heavy weight that we just drop on the ground, imparting energy into the ground. We use shotguns and rifles, sometimes shooting blanks into the ground; that generates a lot of energy. We can use small vibrators, essentially anything that imparts energy into the ground. But for the deeper imaging, say kilometers to tens of kilometers, we typically use repeated heavy weight drops, or some type of huge truck-mounted vibrator, or explosions. Any of those can generate the kind of energy that we need to map out the faults.

Oliver Strimpel: I suppose a big explosion is fine if you're in the middle of the desert, but that wouldn't go down too well in the middle of Los Angeles.

Rufus Catchings: That is correct. So we tend to use other methods when we're in urban areas.

Oliver Strimpel: Apart from the effect of faults on seismic wave velocities, you also use another property of fault zones, namely that they can act as waveguides. Can you explain how that works?

Rufus Catchings: Yes. As I mentioned earlier, in fault zones, the shearing of the fault, the movement on either side, really damages the rock. And so these are low-velocity zones relative to the surrounding rock, and that's effectively like a waveguide. Once you get energy into that fault zone, it acts like a waveguide, and energy moves about in the fault zone, reflecting off the internal surfaces of the fault as it moves forward. You can think of this like an electromagnetic waveguide, or think of a fault as very similar to a coaxial cable or fiber-optic cable, where this energy, once it gets in there, stays in there, and it travels discretely in there. So if we impart that energy, it travels at very high amplitudes, and it's typically much stronger than either the P or the S wave. We can take advantage of that, because it only travels within the fault zone. If we can map out that high-amplitude area, that shows us exactly where the fault is.

Oliver Strimpel: And is that another way that's useful for finding the location of the faults, or do you use it for other purposes as well?

Rufus Catchings: Yeah, it's a very powerful technique for getting the location of the faults. If we place our seismometers across the fault, and if we want a very precise location, we space our seismometers at a one-meter spacing as they go across the fault, and then that's usually the resolution we can get. If it's submeter, we can get submeter. So just by looking at those high amplitudes, or what I call the peak ground velocity of the wave, as it comes through, and I look at it after the S wave has passed, this gives us a lot of information about precisely where that fault is coming to the surface. And the nice thing that we find is that there's often more than just one strand, and we can see that using this technique. So it's a very powerful technique. It tells us precisely where the fault is near the surface, but it can also tell us, say, if you have two different fault strands, whether they're connected to each other. So we use it both for determining connectivity and for the individual location of the fault near the surface.

Oliver Strimpel: So how do you use this method to determine connectivity? Is it that if there's a break in a fault, the energy leaks out in between, and you wouldn't see the continuity from one fault zone to the other?

Rufus Catchings: Absolutely. Just like if you were to take a fiber-optic cable and you break that cable and it's separated, the energy no longer travels. So if we see energy, say from an earthquake, input into one fault, and there's another fault, say to the south of there, and we have an array across there and we don't see that same guided-wave energy traveling onto the second fault, we know they're not connected. But if we see that same energy traveling, and we generally know the speed at which it travels, we know that they are connected.

Oliver Strimpel: So are there some instances, then, when you've seen two faults which initially look like they might be two separate faults, you put the energy into one of them, find that it's traveled to the other one, and then learn for the first time that actually they're part of the same fault system? Has that happened to you?

Rufus Catchings: Yes, yes, that has, and we've done that in multiple places. You might want to say, why does it matter if two faults are connected or not? It makes a big difference, because a short fault, for example a fault that's only, say, 10 or 12 kilometers long, can only generate about a magnitude 6. But if you have one long, continuous fault, it can generate a magnitude 7 or 8 and above. So the length of the fault is very important.

Oliver Strimpel: You study faults in dense urban areas, where humans generate lots of vibrations, not least from 18-wheeler trucks that might be thundering along the freeways. How do you manage to subtract out what you rather charitably refer to as cultural noise?

Rufus Catchings: With some difficulty, often. In some cases, we can actually use the noise as an exploration tool itself, particularly if it's coming from many different directions, and we sometimes do that. However, in most cases, the cultural noise is a major headache, and it's something that we want to try to mitigate. For active-source imaging, we're actually inputting energy into the ground ourselves, with a weight drop, with vibrators, et cetera. We've found that if you input that source into the ground over and over again, say we're using a weight drop and we just keep dropping that same weight, it's the same signal over and over again, whereas the noise from all the cultural activity is kind of random. And so if we stack that seismic wave over and over again, the energy stacks in, all that noise stacks out, and we can get a beautiful signal, even in very noisy areas. One example was where we did this in the Hollywood and Los Angeles area. We just continuously dropped the weight twice a minute for an hour and a half. There was a freeway nearby, there was local traffic, there were tall buildings; there was all kinds of noise going on. By doing this stack, all of that freeway traffic just disappeared, all that other noise just disappeared, and we got this huge signal. So it really works well in certain environments.

Oliver Strimpel: If you're in a dense urban area with this thumping, is that something that somebody living in the neighborhood would be able to feel?

Rufus Catchings: They probably wouldn't feel it. They may hear some of it. We're not talking about dropping huge weights; maybe we're talking about a 500-pound weight that's dropping a couple of feet. That's not very much compared to vehicular traffic.

Oliver Strimpel: OK, so let's turn to what these various methods have revealed. Which faults have you principally studied?

Rufus Catchings: I have studied a very large number of faults. It would take me a long time to name them all; I guess I've never really counted the number of faults that we've studied. Just here in California, these would include the San Andreas Fault, the Hayward Fault, the Rodgers Creek Fault, West Napa, Silver Creek, Calaveras, Monte Vista, the Garlock Fault. There's a very, very long list, but we've studied a lot of these faults, and we see a lot of similarities no matter where we go.

Oliver Strimpel: Yeah, California's got no lack of faults.

Rufus Catchings: There is no shortage of faults. And I've also studied faults outside of California, in Oregon, Nevada, Washington, Arizona, Alaska, Hawaii, Virginia, Delaware, Tennessee, Arkansas, Mississippi and Missouri, and internationally too.

Oliver Strimpel: Let's focus on the San Andreas Fault for a moment, then. Naively, one might think that, as a strike-slip fault, the fault plane would be roughly vertical. Is that what you found?

Rufus Catchings: Well, generally it's vertical to subvertical. We've looked at the San Andreas pretty much along its entire length, from Southern California all the way up to Northern California, where it goes off to the Mendocino triple junction. And yes, there are places where the fault is subvertical, where it has a dip to it, and there are places where it's vertical. But what we also find is that frequently there are subparallel faults. There may be a main trace, and then there are many subparallel faults, and those often dip into the main fault, or sometimes they're just their own fault that runs parallel to it. Any of these can move during a major earthquake, and even if they don't move, they can carry that guided-wave energy that I explained, and they can be quite dangerous. California has something called the Alquist-Priolo Act, which prohibits building within a certain distance of a known active fault. And if you have all these subparallel faults, you often don't know if you're on an active trace of a main fault. So that's one of the reasons why we do so much of this imaging: to try to find every fault, so that we don't put something like a huge apartment building or a hospital on top of these faults, because it can be very dangerous.

Oliver Strimpel: So you said earlier that you can actually measure the fault location to some considerable depth. Does the subsurface geometry of a fault affect how violent the shaking is when it moves?

Rufus Catchings: Yes, it does. I should point out that there are multiple types of faults. The San Andreas is a strike-slip fault, where the two pieces are moving kind of parallel to each other. But there are also normal faults, where one side goes down relative to the other, and reverse faults, where one side goes up relative to the other, and that's often at an angle. And there are many combinations of those. Generally, among the most violently shaking for the same size of earthquake are the reverse faults. And sometimes there are places, even along the San Andreas, where there's reverse faulting going on, or strands of it are reverse. And so we really want to find that and know what type of fault it is and what type of movement is on it, because, again, the reverse faults tend to give us the strongest shaking. If you're not on the footwall, if you're on the hanging wall, that's where you're going to get the most violent shaking.

Oliver Strimpel: In 2014, I think you used guided waves resulting from an earthquake near San Francisco to learn that certain faults that were previously thought to be distinct from each other were actually connected. Can you tell us about that instance?

Rufus Catchings: Yes. In 2014, there was an earthquake in the Napa Valley, which is a region famous for wine. It was a magnitude 6 earthquake, and it occurred on a fault that ruptured the surface for about 12 kilometers. And there were a lot of questions about, well, is this just a single small fault? If it's only 12 kilometers long, that's about the maximum magnitude earthquake it can generate, magnitude 6. So we wanted to know, is it longer? So we did this guided-wave technique, where we put arrays across faults to the south and faults to the north, and we looked to see if the guided waves were continuously coming through onto those other faults, and we found that they were. In fact, instead of being just 12 to 14 kilometers long, we found that it was connected to a fault to the south, and that made the entire fault zone at least 128 kilometers long, which is about 10 times what we suspected originally. And because we know that some of those faults are also connected to some of the larger faults, for example to the Calaveras Fault, it's possible, and likely, that there's a continuous connection of faults as much as 300 kilometers long. To put that in perspective, the 1906 earthquake that destroyed San Francisco occurred along the San Andreas, along a stretch about 300 kilometers long. So that's saying that this other fault, which we thought was relatively minor, might be quite a major fault that could generate a very large magnitude earthquake if it had a continuous rupture, as the San Andreas did in 1906.

Oliver Strimpel: Wow, that finding really could have pretty serious implications. I'm wondering if, as a result of this kind of work, there are new regulations or building codes put in place in the danger zone that might not have been there before.

Rufus Catchings: All of these faults are in close proximity, and they are part of the San Andreas Fault system, so that's largely in place. But people may think, oh, I'm quite a ways away; I'm 30 kilometers from the San Andreas. It's going to generate a large magnitude earthquake, which can be very damaging, but at least it may not rupture the ground underneath me. Well, I think people have to start thinking a little differently when you have these long faults that are subparallel to the San Andreas and go for a long way. So people have considered it. The USGS makes probability maps of where earthquakes are likely to occur over the next 20 years, 30 years, and it certainly takes those kinds of things into consideration.

Oliver Strimpel: I've seen some of your results in which you overlay fault lines onto an aerial shot of an urban neighborhood, and I must say I'd be very disturbed if the red line delineating a fault ran right through my home. Do you recommend that people take some action if they live that close to, or right on top of, one of the fault splays that you've located?

Rufus Catchings: Absolutely. If you live in close proximity to an active fault, you should take every precaution, such as making sure you have water, food, shelter and the medicines that you need stocked up, and that you can have access to them. Be prepared to leave; you may not be able to live in your home. Have a plan: you and your family, where are you going to meet? How are you going to contact each other? Maybe you contact a relative across the country, and that relative contacts your other significant person. Also, if you live in one of these areas, retrofit your home. Even if the fault doesn't run directly beneath your home, the strong shaking from the fault, even if you are tens of kilometers away, can still destroy your home. If you live right on top of the fault itself, you may want to rethink where you're living. I would.

Oliver Strimpel: As you mentioned earlier, you've worked on faults outside California, both in the US and abroad. One of the areas that you studied, surprisingly, is located right in the middle of the North American plate. Can you tell us about that?

Rufus Catchings: Yes. For a very long time, this area called the New Madrid Seismic Zone, or, as the locals call it, the New MAD-rid Seismic Zone, has been known. It's perhaps the most seismically active area in the US east of the Rocky Mountains. In the winter of 1811–1812, it generated three large magnitude earthquakes, and they were so strong that they reportedly rang church bells in Boston, and in places it's reported that the Mississippi River actually ran backward for a while. It's mainly located along the Mississippi River, in Missouri, Arkansas and Tennessee, and it's believed to be the third leg, or the failed rift, of a rift system that goes down to the Gulf. The faults are believed to be reactivated due to slight changes in stress from the North American plate reorienting. But the true cause is not well known, and there is some speculation that there is a very dense lower crust in this area, and that its pull-down effect could also be adding stress to the system to generate these earthquakes.

It's important, because if we have a repeat of the 1811–1812 sequence, it could be devastating to the whole midcontinent area, for two reasons. One, seismic energy propagates much more efficiently east of the Rocky Mountains than, say, in California, so the area of damage and strong shaking would be much wider for the same size of earthquake there than, say, in California. And in 1811–1812, there was very little population there, so there weren't a lot of deaths related to this. But given the large population, the millions of people who live in this area now, it could be devastating. So this is a very important zone, and we at the USGS have been studying it for some time. In the seismic surveys I did in the area, we can see that very dense, high-velocity lower crust, and it actually wells up; it looks like a bell-shaped feature. We can also see that the fault zone is very wide; there are multiple, multiple faults in the subsurface. You tend to think of earthquakes in California primarily, or Alaska or some other places, but you don't tend to think of them that much in the center of the continent. But this is one very, very active area.

Oliver Strimpel: Yeah, it's surprising. You mentioned that it may be the remnant of a former rifting system, a failed rift. Do we understand the tectonic history that led to the existence of this fault zone?

Rufus Catchings: Yeah, there are many papers on it. A rifting system typically has three arms. The two successful arms form an ocean, so you can think of the Gulf like that. And with the third arm, you can't have all three be successful, so one does not quite make it as a rift. We've seen that in many places in the world, and that's likely what's going on there.

Oliver Strimpel: Is that like a geometric constraint?

Rufus Catchings: Yes, yes.

Oliver Strimpel: So if they split apart at one point, there's no room for it to do so in the third location.

Rufus Catchings: Correct, yes.

Oliver Strimpel: You've also worked on zones that are seismically active because of human activity, and in particular a zone in India's Deccan Traps, where a damaging earthquake of magnitude 6.3 was induced. Fracking has received the most publicity, but is that the human activity involved there?

Rufus Catchings: There are several types of human activity that are known to cause earthquakes. These include hydraulic fracking, wastewater disposal at high pressures, water reservoir impoundment, mining and geothermal extraction, among others. Basically, these activities alter the existing stresses in the Earth, and the induced earthquakes are generally rather small. However, in special situations, significant damaging earthquakes are known to have occurred. In the India case, the induced seismicity was caused by the impoundment of a reservoir. It started in the '60s. We refer to this type of activity as reservoir-induced seismicity. Shortly after the dam was put up in the 1960s, earthquakes started to occur. It didn't take very long at all, maybe six months or so after they first did it, and then they got progressively more and more frequent. In 1967, a magnitude 6.3 earthquake occurred, which was quite destructive and killed quite a few people. The earthquakes even continue to this day, particularly following large monsoons.

Our data indicate that the reservoir was impounded at a step-over between two fault systems, strike-slip-type faults. When that happens, you have many faults perpendicular to the two main fault systems. And when you put a lot of water on that, just the weight of the water puts stress on it. The water going down also gets into the pores of the rocks, and that changes the pore pressure. And so that can generate these earthquakes. We see that again after a very large monsoon: you will see, say, six months later, a flurry of activity. We've seen that over and over again. There are other areas in the US where we've seen similar things. If you have these step-overs in the faults, you can get some induced seismicity there, nothing the size of what we've seen in India, but it can occur. So you have to pay special attention to the geometry of the faults if you're doing certain activities.

Oliver Strimpel: We have reservoirs all over the world, and in fact hydroelectric power, if anything, is becoming more important as we go through this energy transition. So does this kind of work give more precise pointers to dam engineers and surveyors and so on as to where, and where not, to place dams?

Rufus Catchings: I would hope so, if the study is done ahead of time. That's not always the case, but I think we're learning a lot more about fault systems and how they're oriented, how they're shaped, relative to the stresses that are there. I should point out that most of the induced seismicity in the US, particularly in Oklahoma, Kansas, Ohio, Texas and Colorado, is caused by the injection of wastewater related to petroleum production. When you inject this wastewater at high pressures, it tends to change the stresses on the faults, and you can generate these earthquakes. For example, in 2011, there was a magnitude 5.7 in Oklahoma, which was quite a strong earthquake that was felt throughout much of the Midwest. And that was definitely an induced earthquake. There were two other magnitude 5 earthquakes in that sequence, and they all occurred very close to where they were injecting wastewater at very high pressures.

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

Rufus Catchings: I've been working on data that I acquired in Turkey after the earthquakes of February last year, 2023. There was a magnitude 7.8 earthquake, and that earthquake was followed by a magnitude 7.6 in the same province, the same general area in Turkey, and the shaking was devastating. It resulted in more than 60,000 deaths in eastern Turkey and northern Syria. Turkey has two major fault systems, and they're both very similar to the San Andreas. One's called the North Anatolian Fault, and the other the East Anatolian. This earthquake occurred on the East Anatolian. And so we need to understand what happened there. It will help the Turks, it will also help us to understand the San Andreas, and it will help people worldwide if we understand these things better, because such large terrestrial earthquakes, no matter where they occur, are going to be devastating. So I led a group of scientists. We went over and deployed a lot of seismometers there, to try to see in three dimensions what's going on in the subsurface, and to understand how this energy propagated so far away and killed so many people. We're working, of course, with the Turkish government and with our colleagues nationally and internationally. We have, I think, an incredible data set, and we hope to get a lot of information from it.

Oliver Strimpel: I'm reminded of an earlier episode of the podcast that I did with Romain Jolivet fairly soon after the earthquake. Although there were some preliminary conclusions, he was very much telling us about the need to do some serious follow-up studies on that very complex fault system. So it's great to see that you're doing just that. Rufus Catchings, thank you very much.

Rufus Catchings: Thank you. I enjoyed it.

Oliver Strimpel: To see pictures and illustrations that support this podcast, go to geologybites.com, where you'll also find transcripts and 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.