1 00:00:04,840 --> 00:00:07,720 This is Geology Bites with Oliver Strumpel. 2 00:00:09,120 --> 00:00:12,360 We've all heard of the San Andreas Fault that runs along 3 00:00:12,360 --> 00:00:15,960 the coast of California. It represents the boundary 4 00:00:15,960 --> 00:00:19,040 between the North American plate and the Pacific plate. 5 00:00:19,600 --> 00:00:24,200 It's 800 mile long extent is visible from space and it has 6 00:00:24,200 --> 00:00:27,040 caused some powerful earthquakes, one of which 7 00:00:27,040 --> 00:00:29,640 devastated San Francisco in 19 O 6. 8 00:00:30,720 --> 00:00:34,360 Most faults leave little or no trace of their localities on the 9 00:00:34,360 --> 00:00:38,680 Earth's surface, but it is still vital to know where they are for 10 00:00:38,680 --> 00:00:41,880 both hazard assessment and scientific reasons. 11 00:00:43,000 --> 00:00:46,600 Rufus Catchings has been hunting down Californian faults for over 12 00:00:46,600 --> 00:00:50,480 40 years. He is a research geophysicist at 13 00:00:50,480 --> 00:00:55,040 the US Geological Survey. Rufus Catchings, Welcome to 14 00:00:55,040 --> 00:00:57,080 Geology Bites. Thank you, Oliver. 15 00:00:57,200 --> 00:01:01,000 It's a pleasure to join you. In your research, you've used 16 00:01:01,000 --> 00:01:04,440 many different approaches to pin down the locations of faults. 17 00:01:05,120 --> 00:01:07,920 Do these all rely on seismic waves? 18 00:01:08,320 --> 00:01:10,120 Would you use other methods as well? 19 00:01:10,560 --> 00:01:15,000 Most methods I use involve seismology, but I also use other 20 00:01:15,000 --> 00:01:18,120 methods as appropriate. For example, in many places we 21 00:01:18,120 --> 00:01:22,520 use geomorphology or changes in the Earth's landform or surface 22 00:01:23,120 --> 00:01:25,760 to infer that there are likely faults there. 23 00:01:26,080 --> 00:01:29,680 Sometimes we can see them at the surface and it's obvious and we 24 00:01:29,760 --> 00:01:33,560 act accordingly there. For example, using 25 00:01:33,560 --> 00:01:37,200 geomorphology, sometimes there's a long linear valley within a 26 00:01:37,200 --> 00:01:41,880 mountain chain, and we suspect that those are likely faults. 27 00:01:42,240 --> 00:01:45,600 There are geologists who specialize in geomorphology and 28 00:01:45,600 --> 00:01:50,800 they also help me in my efforts. OK, let's talk about the seismic 29 00:01:50,800 --> 00:01:54,720 methods that you use and how they work, and then we'll talk 30 00:01:54,720 --> 00:01:58,280 about the results. There are two kinds of seismic 31 00:01:58,280 --> 00:02:02,880 wave, primary or pressure waves called P waves, which are akin 32 00:02:02,880 --> 00:02:07,720 to sound waves in air, and secondary or shear waves called 33 00:02:07,760 --> 00:02:11,520 S waves. Both types of waves are affected 34 00:02:11,560 --> 00:02:14,840 by the presence of faults. Can you tell us how they're 35 00:02:14,840 --> 00:02:18,320 affected? Yes, within the subsurface, S 36 00:02:18,320 --> 00:02:21,680 waves slow down once it encounters a fault zone. 37 00:02:22,080 --> 00:02:25,760 So relative to the rocks on either side, the velocity of the 38 00:02:26,080 --> 00:02:31,640 S waves are much slower in the faulted rock body than the 39 00:02:31,640 --> 00:02:35,960 outside rocks. So faulting essentially damages 40 00:02:36,080 --> 00:02:39,880 the rock, and this changes the rigidity of the rock, which 41 00:02:39,960 --> 00:02:43,520 makes the velocity slower. It's the same way for P waves, 42 00:02:43,520 --> 00:02:46,800 but one complicating factor is in the near surface. 43 00:02:47,120 --> 00:02:51,360 Groundwater can saturate these materials, the rocks as well as 44 00:02:51,400 --> 00:02:55,080 the faulted material, and that actually causes an increase in 45 00:02:55,080 --> 00:02:58,160 velocity, which makes it a little more complicated when we 46 00:02:58,160 --> 00:03:03,120 use just P waves alone. OK, so how does this change in 47 00:03:03,120 --> 00:03:08,400 velocity of the S waves and the P waves enable you to pin down 48 00:03:08,400 --> 00:03:12,440 the location of faults? So what we typically do is put 49 00:03:12,440 --> 00:03:17,800 out an array of seismometers across the fault and then we 50 00:03:17,800 --> 00:03:21,120 import some type of source into the ground so that we can 51 00:03:21,120 --> 00:03:24,600 measure these waves. And from that we measure the 52 00:03:24,600 --> 00:03:28,880 travel times, usually to within milliseconds, and then we can 53 00:03:29,000 --> 00:03:32,600 locate the fault themselves. Using methods like tomography, 54 00:03:32,600 --> 00:03:36,040 we can actually see or get an image of where these low 55 00:03:36,040 --> 00:03:39,480 velocities are located. We can do the same thing for 56 00:03:39,480 --> 00:03:43,880 both P&S waves, but one of the most powerful techniques is to 57 00:03:43,880 --> 00:03:47,400 use the ratio of the 2P waves to S waves. 58 00:03:47,800 --> 00:03:51,840 Because S waves are slower, P waves are faster. 59 00:03:51,960 --> 00:03:57,400 Even if it's saturated, we get VPVS ratios that are very high 60 00:03:57,520 --> 00:04:01,920 in fault zones, so that shows us very nicely where the fault is 61 00:04:01,920 --> 00:04:05,720 in the subsurface. Also, typically a fault is a 62 00:04:05,720 --> 00:04:10,840 barrier to groundwater flow, and we use these seismic waves to 63 00:04:11,200 --> 00:04:14,480 map out the top of groundwater. The top of groundwater's 64 00:04:14,480 --> 00:04:18,000 velocity contour is typically 1500 meters per second. 65 00:04:18,000 --> 00:04:22,520 In sediments, a fault is often a barrier to the flow of 66 00:04:22,520 --> 00:04:25,840 groundwater. So one side of the fault 67 00:04:26,280 --> 00:04:28,240 groundwater will be lower than the other. 68 00:04:28,920 --> 00:04:33,560 And using the tomographic methods, we can actually see 69 00:04:33,560 --> 00:04:37,000 this step and that's another indicator of faulting. 70 00:04:37,560 --> 00:04:40,920 So there are multiple methods that we use. 71 00:04:41,040 --> 00:04:44,680 We use the low P wave velocities, the low S wave 72 00:04:44,680 --> 00:04:49,200 velocities, the high VPVS ratios, the sudden change in 73 00:04:49,200 --> 00:04:52,520 depth of the 1500m per second velocity contour. 74 00:04:53,240 --> 00:04:56,520 All of these are strong evidence of faulting, and when you have 75 00:04:56,520 --> 00:05:00,360 them all in one place, you're quite certain of the location of 76 00:05:00,360 --> 00:05:02,240 the fault. And all the waves that you're 77 00:05:02,240 --> 00:05:07,200 picking up, typically waves that have been refracted by the fault 78 00:05:07,200 --> 00:05:09,880 zone or reflected at the edges of the fault zone. 79 00:05:10,160 --> 00:05:13,840 For the tomographic methods, we're using waves that have been 80 00:05:13,840 --> 00:05:18,040 refracted as they move through the subsurface, but we do use 81 00:05:18,040 --> 00:05:21,000 the reflection method as well. Tell us a bit about the 82 00:05:21,000 --> 00:05:24,560 reflection method as well. The reflection method is a very 83 00:05:24,560 --> 00:05:26,560 powerful method and it's very precise. 84 00:05:26,720 --> 00:05:30,600 The oil industry have used this method for exploration for oil 85 00:05:30,680 --> 00:05:35,400 for many, many decades. But if we do it at very high 86 00:05:35,400 --> 00:05:40,080 resolution, what we can see is the seismic waves go down. 87 00:05:40,080 --> 00:05:44,040 They reflect off surfaces. Faults typically also cause the 88 00:05:44,560 --> 00:05:48,480 layers of the earth on either side of the fault to be offset 89 00:05:48,480 --> 00:05:51,440 somewhat. And reflection methods very good 90 00:05:51,440 --> 00:05:54,480 at showing a visual image of this offset. 91 00:05:54,880 --> 00:05:57,480 But unfortunately you can't use it everywhere. 92 00:05:57,480 --> 00:06:01,000 Many places you don't have this layered subsurface. 93 00:06:01,600 --> 00:06:04,840 What kind of spatial resolutions can you get for these fault 94 00:06:04,840 --> 00:06:08,040 locations? Well, it really depends on the 95 00:06:08,040 --> 00:06:11,720 myth that we're using. And as I mentioned, reflection 96 00:06:11,720 --> 00:06:14,680 seismology is one of the most accurate, particularly at 97 00:06:14,680 --> 00:06:18,400 greater depths. And if we used very dense 98 00:06:18,400 --> 00:06:23,360 spacing on our seismometers, very high frequency seismometers 99 00:06:23,400 --> 00:06:26,920 and the energy that we input into the ground is high 100 00:06:26,920 --> 00:06:30,760 frequency, we can typically image quite deep to within a 101 00:06:30,760 --> 00:06:35,440 meter or so where the fault is. But as we get deeper, we lose 102 00:06:35,440 --> 00:06:41,080 frequency and so the resolution becomes less and we can probably 103 00:06:41,080 --> 00:06:45,600 go to several meters to several 10s of meters at great depth in 104 00:06:45,600 --> 00:06:47,960 our accuracy. But there are other methods as 105 00:06:47,960 --> 00:06:52,840 well that we can use to get high accuracy, particularly in the 106 00:06:52,840 --> 00:06:56,920 shallow subsurface. How deep can you track a fault? 107 00:06:56,960 --> 00:06:59,280 I mean, if a fault's going all the way down to the Moho, for 108 00:06:59,280 --> 00:07:01,880 example, could you see it all the way down? 109 00:07:02,080 --> 00:07:04,760 Yes, you can. The resolution of course once 110 00:07:04,760 --> 00:07:07,560 you get down, you know 3040 kilometers is off. 111 00:07:07,560 --> 00:07:09,920 But yes, you can see these faults quite deep. 112 00:07:10,760 --> 00:07:15,120 I'm reminded of a recent episode of Geology Bytes with Bob White 113 00:07:15,360 --> 00:07:19,320 in which he described using a dense array of seismometers to 114 00:07:19,320 --> 00:07:23,320 pinpoint the locations of thousands of micro quakes and 115 00:07:23,320 --> 00:07:26,440 track the movement of magma within the crusts of Iceland. 116 00:07:27,080 --> 00:07:31,360 Do you also rely on naturally occurring earthquakes as the 117 00:07:31,360 --> 00:07:34,120 source of the seismic waves? Yes, we do. 118 00:07:34,400 --> 00:07:38,560 We use both the naturally occurring earthquakes or ambient 119 00:07:38,560 --> 00:07:42,600 noise, and we also use artificial sources, man made 120 00:07:42,640 --> 00:07:45,000 sources. How did the man made sources 121 00:07:45,000 --> 00:07:47,520 work? In the shallow subsurface, it's 122 00:07:47,520 --> 00:07:51,960 as simple as, say, striking the ground with a hammer. 123 00:07:52,480 --> 00:07:55,840 We just input energy into the ground and that generates 124 00:07:55,840 --> 00:07:58,960 seismic waves. But there's also another way we 125 00:07:58,960 --> 00:08:01,400 use what's called accelerated weight drops. 126 00:08:01,400 --> 00:08:04,800 It's essentially a heavyweight. We just drop on the ground. 127 00:08:04,800 --> 00:08:06,640 It's imparting energy into the ground. 128 00:08:07,160 --> 00:08:11,960 We use shotguns and rifles, sometimes shooting blanks into 129 00:08:11,960 --> 00:08:14,000 the ground. That generates a lot of energy. 130 00:08:14,000 --> 00:08:17,800 We can use small vibrators, essentially anything that is 131 00:08:17,800 --> 00:08:19,400 imparting energy into the ground. 132 00:08:19,800 --> 00:08:24,600 But for the deeper imaging, say kilometers to 10s of kilometers, 133 00:08:25,040 --> 00:08:30,000 we typically use repeated heavyweight drops or some type 134 00:08:30,000 --> 00:08:34,600 of truck mounted huge vibrator or explosions. 135 00:08:34,960 --> 00:08:38,200 Any of those can generate the kind of energy that we need to 136 00:08:38,240 --> 00:08:41,240 map out the faults. I suppose a big explosion is 137 00:08:41,240 --> 00:08:43,600 fine if you're in the middle of the desert, but that wouldn't go 138 00:08:43,600 --> 00:08:45,320 down too well in the middle of Los Angeles. 139 00:08:45,320 --> 00:08:50,440 That is correct. So we tend to use other methods 140 00:08:50,480 --> 00:08:54,800 when we're in urban areas. Apart from the effect of faults 141 00:08:54,960 --> 00:09:00,000 on seismic wave velocities, you also use another property of 142 00:09:00,000 --> 00:09:03,160 fault zones, namely they can act as wave guides. 143 00:09:03,480 --> 00:09:07,440 Can you explain how that works? Yes, as I mentioned earlier, 144 00:09:07,480 --> 00:09:11,080 fault zones, due to the shearing of the fault, the movement on 145 00:09:11,080 --> 00:09:13,920 either side, it really damages the rock. 146 00:09:14,280 --> 00:09:18,600 And so these are low velocity zones relative to the 147 00:09:18,680 --> 00:09:22,240 surrounding rock. And so that's effectively like a 148 00:09:22,240 --> 00:09:25,240 waveguide. Once you get energy into that 149 00:09:25,400 --> 00:09:30,320 fault zone, it acts like a waveguide and energy moves about 150 00:09:30,320 --> 00:09:33,280 in the fault zone reflecting off the internal surfaces of the 151 00:09:33,280 --> 00:09:36,400 fault as it moves forward. You can think of this like an 152 00:09:36,400 --> 00:09:40,880 electromagnetic waveguide or think of a fault as very similar 153 00:09:40,880 --> 00:09:45,040 to a coaxial cable or fiber optic cable where this energy 154 00:09:45,040 --> 00:09:48,400 just once it gets in there, it stays in there and it travels 155 00:09:48,560 --> 00:09:53,960 discreetly in there. So if we impart that energy, it 156 00:09:53,960 --> 00:09:58,560 travels at very high amplitudes and it's typically much stronger 157 00:09:58,560 --> 00:10:02,360 than either the P or the S wave and we can take advantage of 158 00:10:02,360 --> 00:10:05,000 that because it only travels within default zone. 159 00:10:05,280 --> 00:10:10,240 Then if we can map out that high amplitude area, that shows us 160 00:10:10,240 --> 00:10:15,200 exactly where the fault is. And is that another way that's 161 00:10:15,280 --> 00:10:20,360 useful for finding the location of the faults, or do you use it 162 00:10:20,360 --> 00:10:23,280 for other purposes as well? Yeah, it's very powerful 163 00:10:23,280 --> 00:10:25,840 technique for getting the location of the faults. 164 00:10:25,840 --> 00:10:29,760 If we place those seismometers across, and if we want very 165 00:10:29,800 --> 00:10:34,840 precise location, we space our seismometers at a meter spacing 166 00:10:35,120 --> 00:10:38,080 as they go across the fault, then that's usually the 167 00:10:38,080 --> 00:10:41,720 resolution we can get. If it's submeter, we can get 168 00:10:41,720 --> 00:10:45,400 submeter. So just by looking at those high 169 00:10:45,400 --> 00:10:50,520 amplitudes or what I call the peak ground velocity of the wave 170 00:10:50,680 --> 00:10:55,200 as it comes through and I look at it after the S wave has 171 00:10:55,200 --> 00:10:59,560 passed, this gives us a lot of information about absolutely 172 00:10:59,560 --> 00:11:02,280 precisely where that fault is coming to the surface. 173 00:11:02,680 --> 00:11:07,680 And the nice thing that we find is that it's often that there's 174 00:11:07,680 --> 00:11:11,000 more than just one strand. And we can see that using this 175 00:11:11,000 --> 00:11:12,920 technique. So it's a very powerful 176 00:11:12,920 --> 00:11:16,400 technique. It tells us precisely where it 177 00:11:16,400 --> 00:11:20,240 is near the surface, but it can also tell us, say, if you have 178 00:11:20,240 --> 00:11:23,520 two different fault strands, if they're connected to each other. 179 00:11:23,880 --> 00:11:28,000 And we use it both for determining connectivity as well 180 00:11:28,000 --> 00:11:30,880 as the individual location of the fault near the surface. 181 00:11:31,760 --> 00:11:34,200 So how do you use this method to determine connectivity? 182 00:11:34,200 --> 00:11:36,960 Is it that like if there's a break in a fault, the energy 183 00:11:36,960 --> 00:11:40,760 leaks out in between and you wouldn't see the continuity from 184 00:11:40,760 --> 00:11:43,080 one fault zone to the other? Absolutely. 185 00:11:43,120 --> 00:11:46,040 Just like if you were to take a fiber optic cable and you break 186 00:11:46,040 --> 00:11:49,720 that cable and it's separated, the energy no longer travels. 187 00:11:49,720 --> 00:11:55,080 So if we see energy, say from an earthquake input into one fault 188 00:11:55,280 --> 00:11:59,000 and there's another fault, say to the South of there and we 189 00:11:59,000 --> 00:12:02,160 have an array across there and we don't see that same guided 190 00:12:02,160 --> 00:12:05,600 wave energy traveling onto the second fault, we know they're 191 00:12:05,600 --> 00:12:08,760 not connected. But if we see that same energy 192 00:12:08,760 --> 00:12:11,640 traveling, we generally know the speed at which it travels. 193 00:12:11,720 --> 00:12:15,600 We know that they are connected. So are there some instances then 194 00:12:15,600 --> 00:12:19,560 when you've seen two faults which initially look like they 195 00:12:19,560 --> 00:12:22,960 might be two separate faults, you put the energy into one of 196 00:12:22,960 --> 00:12:25,240 them and you find it's travelled to the other one, and then you 197 00:12:25,240 --> 00:12:27,320 learn for the first time that actually they're part of the 198 00:12:27,320 --> 00:12:29,320 same fault system? Has that happened to you? 199 00:12:29,360 --> 00:12:33,120 Yes, yes, that has. And we've done that in multiple 200 00:12:33,120 --> 00:12:35,240 places. You might want to say, why does 201 00:12:35,240 --> 00:12:39,000 it matter if there are two faults that are connected or 202 00:12:39,000 --> 00:12:40,880 not? And it makes a big difference, 203 00:12:41,480 --> 00:12:46,000 because a short fault, for example a fault that's only say 204 00:12:46,000 --> 00:12:49,320 10 or 12 kilometers long, can only generate about a magnitude 205 00:12:49,320 --> 00:12:52,520 6. But if you have one long 206 00:12:52,520 --> 00:12:57,800 continuous fault, it can generate magnitude 7-8 and 207 00:12:57,800 --> 00:13:00,480 above. So the length of the fault is 208 00:13:00,480 --> 00:13:04,160 very important. You study faults in dense urban 209 00:13:04,160 --> 00:13:09,480 areas where humans generate lots of vibrations, not least from 18 210 00:13:09,480 --> 00:13:12,240 Wheeler trucks that might be thundering along the freeways. 211 00:13:12,920 --> 00:13:16,560 How do you manage to subtract out what you rather charitably 212 00:13:16,560 --> 00:13:23,280 refer to as cultural noise? With some difficulty, often in 213 00:13:23,280 --> 00:13:27,040 some cases we can actually use the noise as an exploration tool 214 00:13:27,040 --> 00:13:29,720 itself, particularly if it's coming from many different 215 00:13:29,720 --> 00:13:31,720 directions, and we sometimes do that. 216 00:13:32,200 --> 00:13:34,800 However, in most cases, the cultural noise is a major 217 00:13:34,800 --> 00:13:39,240 headache and it's something that we want to try to mitigate. 218 00:13:39,920 --> 00:13:44,520 So for active source imaging, we're actually inputting energy 219 00:13:44,520 --> 00:13:48,240 into the ground ourselves with a weight drop, with vibrators, 220 00:13:48,240 --> 00:13:51,000 etcetera. We've found that if you 221 00:13:51,200 --> 00:13:55,800 continually input that source into the ground over and over 222 00:13:55,800 --> 00:13:59,000 again, say we're using a weight drop and we just keep dropping 223 00:13:59,000 --> 00:14:02,360 that same weight, it's the same signal over and over again. 224 00:14:03,080 --> 00:14:06,520 Where is the noise from? All the cultural activity is 225 00:14:06,520 --> 00:14:12,240 kind of random. And so if we stack that seismic 226 00:14:12,240 --> 00:14:18,200 wave over and over again, then the energy stacks in, all that 227 00:14:18,200 --> 00:14:23,240 noise stacks out and we can get a beautiful signal through even 228 00:14:23,240 --> 00:14:26,640 very noisy areas. One example was where we did 229 00:14:26,640 --> 00:14:30,000 this in Hollywood and Los Angeles area. 230 00:14:30,680 --> 00:14:34,480 We just continuously dropped the weight twice a minute for an 231 00:14:34,480 --> 00:14:37,480 hour and a half. And there was a freeway nearby, 232 00:14:37,480 --> 00:14:41,560 there were local traffic, there were tall buildings, there was 233 00:14:41,560 --> 00:14:45,800 all kinds of noise going on. So by doing this stack, all of 234 00:14:45,800 --> 00:14:48,720 that freeway traffic just disappeared, all that other 235 00:14:48,720 --> 00:14:51,640 noise just disappeared, and we got this huge signal. 236 00:14:51,640 --> 00:14:55,040 So it really works well in certain environments. 237 00:14:55,160 --> 00:14:58,440 If you're in a dense urban area with this thumping, is that 238 00:14:58,440 --> 00:15:00,360 something that somebody living in the neighborhood would be 239 00:15:00,360 --> 00:15:02,920 able to feel? They probably wouldn't feel it, 240 00:15:02,960 --> 00:15:06,360 they may hear some of it and we're not talking about dropping 241 00:15:06,360 --> 00:15:09,360 huge weights, maybe we're talking about a £500 weight 242 00:15:09,360 --> 00:15:13,160 that's dropping a couple feet. That's not very much compared to 243 00:15:13,160 --> 00:15:17,000 vehicular traffic. OK, so let's turn to what these 244 00:15:17,040 --> 00:15:20,520 various methods have revealed. Which faults have you 245 00:15:20,760 --> 00:15:26,800 principally studied? I have studied a very large 246 00:15:26,800 --> 00:15:29,800 number of faults. It would take me a long time to 247 00:15:29,800 --> 00:15:32,440 name them all. I guess I've never really 248 00:15:32,440 --> 00:15:34,760 counted the number of faults that we've studied. 249 00:15:34,880 --> 00:15:38,720 Just say here in in California, these would include the San 250 00:15:38,720 --> 00:15:42,640 Andreas Fault, the Hayward Fault, the Rogers Creek Fault, W 251 00:15:42,640 --> 00:15:49,240 Napa, Silver Creek, Calaveras, Monte Vista, Garlock Fault. 252 00:15:49,600 --> 00:15:55,680 There is a very, very long list, but we've studied a lot of these 253 00:15:55,680 --> 00:15:58,240 faults and we see a lot of similarities no matter where we 254 00:15:58,240 --> 00:16:01,040 go. Yeah, California's got no lack 255 00:16:01,040 --> 00:16:04,040 of faults. There is no shortage of faults 256 00:16:04,200 --> 00:16:07,240 and I've also studied faults outside of California and 257 00:16:07,640 --> 00:16:11,840 Oregon, Nevada, Washington, Arizona, Alaska, Hawaii, 258 00:16:12,200 --> 00:16:17,000 Virginia, Delaware, Tennessee, Arkansas, Mississippi, Missouri 259 00:16:18,160 --> 00:16:21,440 and internationally too. Let's focus on the San Andreas 260 00:16:21,440 --> 00:16:25,120 Fault for a moment, then. Naively, one might think that as 261 00:16:25,120 --> 00:16:29,320 a strike slip fault, the fault plane would be roughly vertical. 262 00:16:29,840 --> 00:16:35,320 Is that what you found? Well, generally it's vertical to 263 00:16:35,320 --> 00:16:37,520 sub vertical. And we've looked at the San 264 00:16:37,520 --> 00:16:40,560 Andreas pretty much along its entire length, Southern 265 00:16:40,560 --> 00:16:43,960 California all the way up to Northern California where it 266 00:16:43,960 --> 00:16:46,240 goes off to the Mendocino triple junction. 267 00:16:46,960 --> 00:16:51,080 And yes, there are places where the fault is sub vertical. 268 00:16:51,080 --> 00:16:54,160 It has a depth to it and there are places where it's vertical. 269 00:16:54,840 --> 00:16:58,440 But what we also find is that frequently there are sub 270 00:16:59,080 --> 00:17:02,280 parallel faults to it. There may be a main trace and 271 00:17:02,280 --> 00:17:07,359 then there are many sub parallel faults and those often dip into 272 00:17:07,359 --> 00:17:11,359 the main fault or sometimes they're just their own fault 273 00:17:11,359 --> 00:17:14,440 that goes spare a little to it. Any of these can move during a 274 00:17:14,440 --> 00:17:18,680 major earthquake and even if they don't move, they can carry 275 00:17:18,680 --> 00:17:23,160 that guided wave energy that I explained and they can be quite 276 00:17:23,160 --> 00:17:26,440 dangerous. California has something called 277 00:17:26,440 --> 00:17:32,320 the Alquispio Act, which prohibits building on a known 278 00:17:32,320 --> 00:17:35,880 fault within a certain distance of a known active fault. 279 00:17:36,840 --> 00:17:39,760 And if you have all these subparallel faults to it, you 280 00:17:39,760 --> 00:17:43,080 often don't know if you're on an active trace of a main fault. 281 00:17:43,640 --> 00:17:47,800 So that's one of the reasons why we do so much of this imaging is 282 00:17:47,800 --> 00:17:51,360 to try to find every fault so that we don't put something like 283 00:17:51,360 --> 00:17:54,360 a huge apartment building or a hospital or something like that 284 00:17:54,360 --> 00:17:57,160 on top of these faults because it can be very dangerous. 285 00:17:57,920 --> 00:18:00,120 So you said earlier that you can actually measure the fault 286 00:18:00,120 --> 00:18:02,120 location to some considerable depth. 287 00:18:02,760 --> 00:18:08,960 Does the subsurface geometry of a fault affect how violent the 288 00:18:08,960 --> 00:18:11,360 shaking is on the fault when it moves? 289 00:18:11,440 --> 00:18:13,320 Yes it does. I should point out that there 290 00:18:13,320 --> 00:18:16,240 are multiple types of faults. The San Andreas is a strike slip 291 00:18:16,240 --> 00:18:19,520 fault where the two pieces are moving kind of parallel to each 292 00:18:19,520 --> 00:18:22,960 other. But there are also normal faults 293 00:18:22,960 --> 00:18:26,560 where one fault goes down relative to the other, reverse 294 00:18:26,560 --> 00:18:29,560 faults where one fault goes up relative to the other. 295 00:18:29,560 --> 00:18:33,240 And that's often at an angle. And there are many combinations 296 00:18:33,240 --> 00:18:36,680 of those. And generally among the most 297 00:18:36,680 --> 00:18:40,280 violently shaking for the same size earthquake are the reverse 298 00:18:40,280 --> 00:18:43,120 faults. And sometimes there are places 299 00:18:43,120 --> 00:18:46,480 even along the San Andreas where there's reverse faulting going 300 00:18:46,480 --> 00:18:48,600 on, or strands of it are reverse. 301 00:18:49,000 --> 00:18:53,240 And so we really want to find that and know what type of 302 00:18:53,240 --> 00:18:57,520 fault, what type of movement on it, because again, the reverse 303 00:18:57,520 --> 00:18:59,760 faults tend to give us the strongest shaking. 304 00:18:59,800 --> 00:19:03,400 If you're not on the foot wall, if you're on the hanging wall, 305 00:19:03,400 --> 00:19:05,920 that's where you're going to get the most violent. 306 00:19:05,920 --> 00:19:10,640 Shaking. In 2014, I think you used guided 307 00:19:10,640 --> 00:19:14,320 waves resulting from an earthquake near San Francisco to 308 00:19:14,320 --> 00:19:17,680 learn that certain faults that were previously thought to be 309 00:19:17,680 --> 00:19:20,480 distinct from each other were actually connected. 310 00:19:20,600 --> 00:19:22,720 Can you tell us about that instance? 311 00:19:23,040 --> 00:19:25,880 Yes. In 2014, there was an earthquake 312 00:19:26,200 --> 00:19:30,680 in Napa Valley, which is a region famous for wine. 313 00:19:30,840 --> 00:19:34,640 And it was a magnitude 6 earthquake and it occurred on a 314 00:19:34,640 --> 00:19:39,560 fault that ruptured the surface for about 12 kilometers. 315 00:19:40,720 --> 00:19:44,920 And there was a lot of questions about, well, is this just a 316 00:19:45,800 --> 00:19:49,400 single small fault? And if it's only 12 kilometers 317 00:19:49,400 --> 00:19:52,120 long, that's about the maximum magnitude earthquake it can 318 00:19:52,120 --> 00:19:57,080 generate magnitude 6. So we wanted to know, well, is 319 00:19:57,080 --> 00:19:59,520 it longer? So we did this guided wave 320 00:19:59,520 --> 00:20:03,400 technique where we put a rays across faults to the South, 321 00:20:03,960 --> 00:20:08,040 faults to the north. And we looked to see if the 322 00:20:08,040 --> 00:20:11,680 guided waves were continuously coming through onto those other 323 00:20:11,680 --> 00:20:13,880 faults, and we found that they were. 324 00:20:14,040 --> 00:20:19,280 In fact, instead of being just 12 to 14 kilometers long, we 325 00:20:19,280 --> 00:20:21,920 found that it was connected to a fault to the South. 326 00:20:22,240 --> 00:20:26,160 And that made the entire fault zone at least 128 kilometers 327 00:20:26,160 --> 00:20:30,400 long, which is about 10 times what we suspected originally. 328 00:20:31,040 --> 00:20:35,280 And because we know that some of those faults are also connected 329 00:20:35,280 --> 00:20:39,440 to some of the larger faults, for example to Calaveras Fault, 330 00:20:39,640 --> 00:20:43,760 it's possible and likely that there's a continuous connection 331 00:20:43,760 --> 00:20:46,400 of faults that as much as 300 kilometers long. 332 00:20:47,000 --> 00:20:51,400 And to put that in perspective, the 19 O 6 earthquake that 333 00:20:51,400 --> 00:20:55,880 destroyed San Francisco occurred along the San Andreas along an 334 00:20:55,880 --> 00:21:00,280 area about 300 kilometers long. So that's saying that this other 335 00:21:00,280 --> 00:21:04,000 fault that we thought was relatively minor might be quite 336 00:21:04,280 --> 00:21:07,240 a major fault that could generate a very large magnitude 337 00:21:07,240 --> 00:21:10,640 earthquake if it did a continuous rupture as the San 338 00:21:10,640 --> 00:21:14,200 Andreas did in 19 O 6. Wow, that finding really could 339 00:21:14,200 --> 00:21:16,000 have pretty serious implications. 340 00:21:16,360 --> 00:21:21,080 I'm wondering if as a result of this kind of work, there are new 341 00:21:21,080 --> 00:21:25,320 regulations or building codes that are put in place in the 342 00:21:25,320 --> 00:21:28,040 danger zone that might not have been there before. 343 00:21:28,440 --> 00:21:32,080 All of these faults are within close proximity and they are 344 00:21:32,080 --> 00:21:37,200 part of the San Andreas Fault system that's largely in place. 345 00:21:37,680 --> 00:21:41,760 But for people say you're within 10 kilometers, you may think 346 00:21:41,760 --> 00:21:45,320 that oh, I'm quite a ways I'm 30 kilometers from the San Andreas. 347 00:21:45,320 --> 00:21:47,640 It's going to generate a large magnitude earthquake, can be 348 00:21:47,640 --> 00:21:51,440 very damaging, but at least it may not rupture the ground 349 00:21:51,440 --> 00:21:55,600 underneath me. Well, I think people have to 350 00:21:55,600 --> 00:21:57,960 start thinking a little differently when you have these 351 00:21:57,960 --> 00:22:01,000 long faults that are subparallel to the San Andreas and they go 352 00:22:01,000 --> 00:22:03,920 for a long way. So people have considered it. 353 00:22:03,920 --> 00:22:09,120 The USGS does this probability map or where earthquakes are 354 00:22:09,120 --> 00:22:14,000 likely to occur over the next 20 years, 30 years, and it 355 00:22:14,200 --> 00:22:16,720 certainly takes those kind of things into consideration. 356 00:22:17,440 --> 00:22:21,360 I've seen some of your results in which you overlay fault lines 357 00:22:21,360 --> 00:22:25,600 onto an aerial shot of an urban neighborhood, and I must say I'd 358 00:22:25,600 --> 00:22:30,720 be very disturbed if the red line delineating a fault ran 359 00:22:30,720 --> 00:22:33,840 right through my home. Do you recommend that people 360 00:22:33,840 --> 00:22:37,920 take some action if they live that close to or right on top of 361 00:22:37,920 --> 00:22:40,520 one of the fault splays that you've located? 362 00:22:40,840 --> 00:22:45,640 Absolutely, if you live in close proximity to an active fault, 363 00:22:46,360 --> 00:22:49,840 you should take every precaution such as making sure you have 364 00:22:49,840 --> 00:22:53,520 water, food, shelter, the medicines that you need stocked 365 00:22:53,520 --> 00:22:56,280 up and that you can have access to them. 366 00:22:56,360 --> 00:22:59,320 Be prepared to leave. You may not be able to live in 367 00:22:59,320 --> 00:23:01,000 your home. Have a plan. 368 00:23:01,080 --> 00:23:03,160 You and your family, where are you going to meet? 369 00:23:03,320 --> 00:23:05,000 Are you going to contact each other? 370 00:23:05,120 --> 00:23:08,240 And maybe you contacted a relative across the country and 371 00:23:08,240 --> 00:23:11,880 that relative across the country contacts your other significant 372 00:23:11,880 --> 00:23:13,880 person. Also, if you live in one of 373 00:23:13,880 --> 00:23:17,400 these areas, retrofit your home. Even if the fault doesn't run 374 00:23:17,400 --> 00:23:21,200 directly beneath your home, the strong shaking from the fault, 375 00:23:21,200 --> 00:23:24,040 even if you are 10s of kilometers away, can still 376 00:23:24,040 --> 00:23:27,200 destroy your home. If you live right on top of the 377 00:23:27,200 --> 00:23:29,720 fault itself. You may rethink where you're 378 00:23:29,720 --> 00:23:33,400 living, or I would. As you mentioned earlier, you've 379 00:23:33,400 --> 00:23:38,000 worked on faults outside California, both in the US and 380 00:23:38,000 --> 00:23:40,120 abroad. One of the areas that you 381 00:23:40,120 --> 00:23:43,560 studied, surprisingly, is located right in the middle of 382 00:23:43,560 --> 00:23:46,720 the North American plate. Can you tell us about that? 383 00:23:46,800 --> 00:23:51,320 Yes, for a very long time it's been known this area called New 384 00:23:51,320 --> 00:23:55,280 Madrid Seismic Zone, or the locals call it New Madrid 385 00:23:55,280 --> 00:23:57,440 Seismic Zone. It's perhaps the most 386 00:23:57,440 --> 00:24:01,360 seismically active area in the US east of the Rocky Mountains. 387 00:24:01,520 --> 00:24:07,280 In the winter of 18111812, it generated 3 large magnitude 388 00:24:07,280 --> 00:24:11,680 earthquakes and they were so strong that it reportedly rang 389 00:24:11,680 --> 00:24:16,120 church bells in Boston and in places it's reported that the 390 00:24:16,240 --> 00:24:19,160 Mississippi River actually ran backward for a while. 391 00:24:20,080 --> 00:24:23,880 It's mainly located along the Mississippi River in Missouri, 392 00:24:23,880 --> 00:24:27,760 Arkansas, Tennessee and it's believed to be the third leg or 393 00:24:27,760 --> 00:24:32,560 the failed rift of a rift system that goes down to the Gulf and 394 00:24:32,800 --> 00:24:36,160 the faults are believed to be reactivated due to slight 395 00:24:36,160 --> 00:24:40,040 changes in stress from the North American plate reorienting. 396 00:24:40,160 --> 00:24:44,040 But the true 'cause is not well known, but there are some 397 00:24:44,040 --> 00:24:49,840 speculation that there is a very dense lower crust in this area 398 00:24:49,920 --> 00:24:54,240 and the pull down effect could also be adding a stress to the 399 00:24:54,240 --> 00:24:56,720 system to generate these earthquakes. 400 00:24:56,800 --> 00:25:01,960 So it's important because if we have a repeat of the 18111812 401 00:25:01,960 --> 00:25:04,960 sequence, it could be devastating to the whole mid 402 00:25:04,960 --> 00:25:09,600 continent area for two reasons. One, seismic energy propagates 403 00:25:09,600 --> 00:25:13,440 much more efficiently east of the Rocky Mountains then say in 404 00:25:13,440 --> 00:25:17,400 California. So the area of damage and strong 405 00:25:17,400 --> 00:25:21,080 shaking would be much wider for the same size earthquake there 406 00:25:21,160 --> 00:25:27,120 than say in California. And in 18111812, there were very 407 00:25:27,120 --> 00:25:30,840 little population there. So there wasn't a lot of deaths 408 00:25:30,880 --> 00:25:34,160 related to this. But given the large population, 409 00:25:34,160 --> 00:25:37,360 the millions of people who live in this area now, it could be 410 00:25:37,360 --> 00:25:39,960 devastating. So this is a very important 411 00:25:39,960 --> 00:25:42,120 zone. We at USGS have been studying it 412 00:25:42,120 --> 00:25:44,840 for some time. The seismic surveys I did in the 413 00:25:44,840 --> 00:25:51,000 area, we can see that very dense high velocity lower crust and it 414 00:25:51,000 --> 00:25:55,080 actually wells up, looks like a bell shaped feature. 415 00:25:55,840 --> 00:26:00,520 And we can also see that the fault zone is very wide. 416 00:26:00,600 --> 00:26:03,400 There are multiple, multiple faults in the surface. 417 00:26:04,160 --> 00:26:07,440 So you tend to think of earthquakes in California 418 00:26:07,440 --> 00:26:12,040 primarily or Alaska or some other places, but you don't tend 419 00:26:12,040 --> 00:26:15,040 to think of it that much in the center of the continent. 420 00:26:15,040 --> 00:26:18,160 But this is one very, very active area. 421 00:26:18,720 --> 00:26:20,920 Yeah, it's surprising. You mentioned that it may be the 422 00:26:20,920 --> 00:26:23,960 remnants of a formal rifting system or failed rift. 423 00:26:24,120 --> 00:26:27,440 Do we understand that tectonic history that led to the 424 00:26:27,440 --> 00:26:30,560 existence of this fault zone? Yeah, there are many papers on 425 00:26:30,560 --> 00:26:32,640 it. A rifting system typically has 426 00:26:32,640 --> 00:26:36,800 three arms. The two successful arms form an 427 00:26:36,800 --> 00:26:38,680 ocean. So you can think of the gulf 428 00:26:38,680 --> 00:26:41,720 like that. And the third arm, you can't 429 00:26:41,720 --> 00:26:45,080 have all three be successful. So one that does not quite make 430 00:26:45,080 --> 00:26:48,440 it to a rift. We've seen that many places in 431 00:26:48,440 --> 00:26:51,960 the world, and that's likely what's going on there. 432 00:26:52,280 --> 00:26:54,240 Is that like a geometric constraint? 433 00:26:54,400 --> 00:26:57,640 Yes, yes. That split apart the pace at one 434 00:26:57,960 --> 00:27:01,200 point, there's no room for it to do so in the third location. 435 00:27:01,280 --> 00:27:02,440 Correct. Yes. 436 00:27:02,760 --> 00:27:06,640 You've also worked on zones sort of seismically active because of 437 00:27:06,640 --> 00:27:10,400 human activity, and in particular a zone in India's 438 00:27:10,400 --> 00:27:16,480 Deccan Traps where a damaging earthquake of magnitude 6.3 was 439 00:27:16,480 --> 00:27:19,720 induced. Fracking has received the most 440 00:27:19,720 --> 00:27:23,800 publicity, but is that the human activity involved there? 441 00:27:24,240 --> 00:27:27,880 There are several types of human activity that are known to cause 442 00:27:27,960 --> 00:27:30,280 earthquakes. That includes hydraulic 443 00:27:30,280 --> 00:27:34,720 fracking, wastewater disposal at high pressures, water reservoir 444 00:27:34,720 --> 00:27:38,400 impoundment, mining, geothermal extraction, among others. 445 00:27:39,120 --> 00:27:43,200 So basically these activities altered the existing stresses in 446 00:27:43,200 --> 00:27:47,600 the earth and the induced earthquakes are generally rather 447 00:27:47,600 --> 00:27:50,440 small. However, in Special Situations, 448 00:27:50,680 --> 00:27:54,000 significant damaging earthquakes are known to have occurred. 449 00:27:54,480 --> 00:27:59,000 And for the India area, the induced seismicity was caused by 450 00:27:59,000 --> 00:28:00,800 the impoundment of this reservoir. 451 00:28:01,160 --> 00:28:04,120 It started in the 60s. We refer to this type of 452 00:28:04,120 --> 00:28:06,560 activity as reservoir induced seismicity. 453 00:28:06,960 --> 00:28:11,640 Shortly after the dam was put up in the 1960s, earthquake started 454 00:28:11,640 --> 00:28:13,960 to occur. It didn't take very long at all, 455 00:28:14,360 --> 00:28:17,640 maybe six months or so after they first did it, and then they 456 00:28:17,640 --> 00:28:23,480 got progressively more and more. And in 1967 a magnitude 6.3 457 00:28:23,480 --> 00:28:26,480 earthquake occurred, which was quite destructive and killed 458 00:28:26,480 --> 00:28:29,400 quite a few people. The earthquake even continued to 459 00:28:29,400 --> 00:28:32,400 this day, particularly following large monsoons. 460 00:28:32,920 --> 00:28:37,040 So our data indicate that the reservoir was impounded at a 461 00:28:37,040 --> 00:28:40,640 step over between two fault systems, strike slip type 462 00:28:40,640 --> 00:28:43,640 faults. And when that happens, you have 463 00:28:43,640 --> 00:28:47,160 many perpendicular faults to the two main fault systems. 464 00:28:47,680 --> 00:28:51,520 And when you put a lot of water on that, just the weight of the 465 00:28:51,520 --> 00:28:55,680 water put stress on it. The water going down also gets 466 00:28:55,680 --> 00:28:59,120 into the pores of the rocks. That changes the pore pressure. 467 00:28:59,440 --> 00:29:02,040 And so that can generate these earthquakes. 468 00:29:02,120 --> 00:29:05,560 And we see that again after a very large monsoon. 469 00:29:06,120 --> 00:29:09,800 You will see say six months later, a flurry of activity. 470 00:29:09,800 --> 00:29:11,400 We've seen that over and over again. 471 00:29:12,280 --> 00:29:16,560 There are other areas in the US where we've seen similar things. 472 00:29:16,560 --> 00:29:20,000 If you have these step over in the faults, you can get some 473 00:29:20,160 --> 00:29:23,640 induced seismicity there. Nothing the size that we've seen 474 00:29:23,640 --> 00:29:27,360 from India, but they can occur. So you have to pay special 475 00:29:27,360 --> 00:29:30,440 attention to the geometry of the faults if you're doing certain 476 00:29:30,440 --> 00:29:33,560 activities. We have reservoirs all over the 477 00:29:33,560 --> 00:29:36,240 world, and in fact, hydroelectric powers, if 478 00:29:36,240 --> 00:29:39,360 anything, is becoming more important as we're going through 479 00:29:39,360 --> 00:29:44,840 this energy transition. So does this kind of work give 480 00:29:44,840 --> 00:29:50,360 us more precise pointers to dam engineers and surveyors and so 481 00:29:50,360 --> 00:29:55,040 on, as to where and where not to place dams? 482 00:29:55,400 --> 00:29:58,360 I would hope so if the study is done ahead of time. 483 00:29:58,480 --> 00:30:02,040 That's not always the case, but I think we're learning a lot 484 00:30:02,040 --> 00:30:06,040 more about fault systems and how they're oriented, how they're 485 00:30:06,040 --> 00:30:09,160 shaped and relative to the stresses that are there. 486 00:30:09,520 --> 00:30:13,400 I should point out that most of the induced sites, Misty, in the 487 00:30:13,640 --> 00:30:20,720 US, particularly in Oklahoma and Kansas, Ohio, Texas, Colorado, 488 00:30:21,320 --> 00:30:27,200 they're caused by the injection of wastewater related to 489 00:30:27,200 --> 00:30:30,840 petroleum production. So when you inject this 490 00:30:30,840 --> 00:30:34,840 wastewater at high pressures, it tends to change the stresses on 491 00:30:34,840 --> 00:30:37,200 the faults and you can generate these earthquakes. 492 00:30:37,200 --> 00:30:42,440 And for example, in 2011, there was a magnitude 5.7 in Oklahoma 493 00:30:42,600 --> 00:30:45,720 that was quite a strong earthquake that was felt all 494 00:30:45,720 --> 00:30:49,160 throughout much of the Midwest. And that was definitely an 495 00:30:49,160 --> 00:30:51,440 induced earthquake. And there were two other 496 00:30:51,440 --> 00:30:55,320 magnitude 5 earthquakes in that sequence that occurred and they 497 00:30:55,320 --> 00:30:58,840 all occurred very close to where they were injecting wastewater 498 00:30:58,840 --> 00:31:01,880 at very high pressures. What are you working on at the 499 00:31:01,880 --> 00:31:05,560 moment? I've been working on data that I 500 00:31:05,600 --> 00:31:11,200 acquired in Turkey in February of last year 2023. 501 00:31:11,200 --> 00:31:15,520 There was a magnitude 7.8 earthquake and that earthquake 502 00:31:15,520 --> 00:31:19,680 was followed by a magnitude 7.6 in the same province, same 503 00:31:19,680 --> 00:31:24,000 general area in Turkey and the shaking was devastating. 504 00:31:24,000 --> 00:31:29,120 It resulted in more than 60,000 deaths in eastern Turkey, in 505 00:31:29,120 --> 00:31:33,080 northern Syria. And Turkey has two major fault 506 00:31:33,080 --> 00:31:37,160 systems and they're both very similar to the San Andreas one's 507 00:31:37,160 --> 00:31:40,320 called the North Anatolian fault near the East Anatolian. 508 00:31:40,320 --> 00:31:42,680 This earthquake occurred on the East Anatolian. 509 00:31:43,440 --> 00:31:46,680 And so we need to understand what happened there. 510 00:31:46,760 --> 00:31:50,360 It will help the Turks. It will also help us to 511 00:31:50,360 --> 00:31:54,720 understand the San Andreas, and it will help people worldwide if 512 00:31:54,720 --> 00:31:58,560 we understand these things better, because such large 513 00:31:58,560 --> 00:32:01,520 terrestrial earthquakes, no matter where they occur, they're 514 00:32:01,520 --> 00:32:05,160 going to be devastating. So I LED a group of scientists. 515 00:32:05,160 --> 00:32:09,440 We went over and we deployed a lot of seismometers there to try 516 00:32:09,440 --> 00:32:12,520 just see in three dimensions what's going on in the 517 00:32:12,520 --> 00:32:16,880 subsurface, understand how this energy propagated so far away 518 00:32:16,880 --> 00:32:20,200 and killed so many people. And we're working, of course, 519 00:32:20,200 --> 00:32:23,160 with the Turkish government and with our colleagues nationally 520 00:32:23,160 --> 00:32:25,920 and internationally. We have, I think, an incredible 521 00:32:25,920 --> 00:32:29,640 data set and we hope to get a lot of information from that. 522 00:32:30,360 --> 00:32:33,800 I'm reminded of an early episode of the podcast that I did with 523 00:32:33,800 --> 00:32:37,120 Roman Jolivet fairly soon after the earthquake. 524 00:32:37,280 --> 00:32:40,000 And although there were some preliminary conclusions, he was 525 00:32:40,000 --> 00:32:43,520 very much telling us about the need to do some serious follow 526 00:32:43,520 --> 00:32:46,560 up studies on that very complex fault system there. 527 00:32:47,000 --> 00:32:50,360 So it's great to see that you're doing just that. 528 00:32:50,840 --> 00:32:52,680 Rufus Catchings, thank you very much. 529 00:32:52,720 --> 00:33:07,080 Thank you, I enjoy it. To see pictures and 530 00:33:07,080 --> 00:33:12,440 illustrations that support this podcast, go to geologybytes.com, 531 00:33:13,040 --> 00:33:16,320 where you'll also find transcripts and a subject matter 532 00:33:16,320 --> 00:33:19,640 index of all the episodes. There you can also give me 533 00:33:19,640 --> 00:33:23,600 feedback which I welcome, as well as sign up to get my emails 534 00:33:23,600 --> 00:33:24,840 about new episodes.