WEBVTT Kind: captions Language: en-US 00:00:00.429 --> 00:00:03.970 Okay. Hello, everyone. Welcome to our very first hybrid 00:00:03.970 --> 00:00:08.656 seminar. This will be a learning process, so please bear with us. 00:00:08.699 --> 00:00:12.531 Welcome to the ESC seminar for today. It’s May 11th. 00:00:12.531 --> 00:00:15.040 As we said, this is our first hybrid seminar. 00:00:15.040 --> 00:00:18.100 So the way that we’re going to try to work this is that we have – we have 00:00:18.100 --> 00:00:21.102 some people here in the room. We’re going to try to show you them. 00:00:21.102 --> 00:00:23.212 There really are people here. I mean it. 00:00:23.212 --> 00:00:28.626 And we have a bunch of people online. So Chad will introduce the speaker, 00:00:28.626 --> 00:00:30.423 and then she’ll give her talk. We’re going to save the questions 00:00:30.423 --> 00:00:32.970 for the end. Keep this a little simpler. 00:00:32.970 --> 00:00:35.980 So, if you’re online and you have a question, put it in the chat. 00:00:35.980 --> 00:00:39.460 And we will try to make sure that we can have people who are online 00:00:39.460 --> 00:00:42.197 be able to ask their questions themselves by unmuting themselves. 00:00:42.223 --> 00:00:45.401 We’ll see how that goes. Okay. So I’m going to hand it off to Chad. 00:00:45.401 --> 00:00:47.018 Go ahead. 00:00:48.697 --> 00:00:52.650 [silence] 00:00:52.650 --> 00:00:57.219 - So it’s my honor today to introduce Alba Rodriguez Padilla. 00:00:57.219 --> 00:01:00.500 Alba got her undergrad from the College of the Atlantic where she did research 00:01:00.500 --> 00:01:03.850 on a variety of tectonic and geomorphology problems including a 00:01:03.850 --> 00:01:08.350 thesis on the tectonic climatic evolution of the forearc of southern Peru, which – 00:01:08.350 --> 00:01:11.240 actually, some of which was just published, I believe. 00:01:11.240 --> 00:01:14.700 She then moved on to UC-Davis and joined Mike Oskin’s research group, 00:01:14.700 --> 00:01:18.720 which is where I first got to know her. Her Ph.D. work explores a variety of 00:01:18.720 --> 00:01:21.940 fault mechanics problems using a diverse array of techniques, 00:01:21.940 --> 00:01:24.830 including field work, landscape analysis with remote data, 00:01:24.830 --> 00:01:28.170 and finite element modeling. Most recently she was awarded 00:01:28.170 --> 00:01:31.450 a NASA fellowship to investigate off-fault inelastic deformation 00:01:31.450 --> 00:01:35.310 over multiple earthquake cycles. And, in addition to her Ph.D. research, 00:01:35.310 --> 00:01:37.909 she’s done some really good undergrad mentoring work 00:01:37.909 --> 00:01:43.130 doing research with undergrads primarily using data from the 00:01:43.130 --> 00:01:45.324 2019 Ridgecrest earthquake sequence. 00:01:45.324 --> 00:01:49.100 So, without further ado, I’ll turn it over to Alba tell her – to tell us 00:01:49.100 --> 00:01:53.144 about her research on multi-fault earthquakes in southern California. 00:01:53.144 --> 00:01:57.926 [applause] 00:01:58.723 --> 00:02:02.670 - Thank you for that introduction, Chad. I’m happy to be joined by 00:02:02.670 --> 00:02:04.840 a live audience today. This is really exciting. 00:02:04.840 --> 00:02:08.729 It’s been a long time from giving in-person talks. 00:02:08.729 --> 00:02:13.469 And today I’m going to be taking you on a tour of southern California from 00:02:13.469 --> 00:02:18.719 east to west, where the joining nexus of this talk is multi-fault earthquakes. 00:02:18.719 --> 00:02:23.529 We’re going to start east talking about the 2019 Ridgecrest earthquakes 00:02:23.529 --> 00:02:27.409 and how inelastic deformation was distributed for those events and the 00:02:27.409 --> 00:02:31.249 implications for how we think of the damage zone from those observations. 00:02:31.249 --> 00:02:34.489 Then I’m going to bring in observations from prior events in the Eastern 00:02:34.489 --> 00:02:36.489 California Shear Zone. And at the end, we will move 00:02:36.489 --> 00:02:40.359 all the way east to Cajon Pass just north of L.A. where the San Andreas 00:02:40.359 --> 00:02:42.930 and the San Jacinto Faults come together. 00:02:42.930 --> 00:02:46.769 This is work done in collaboration with my Ph.D. adviser Mike Oskin, 00:02:46.769 --> 00:02:52.578 and then with Chris Milliner [echoing sounds], Andreas [inaudible]. 00:02:52.578 --> 00:02:56.195 Okay, so without further ado [loud echoing] – well, actually, 00:02:56.195 --> 00:02:59.701 before I start, I want to acknowledge that I know several of you 00:02:59.701 --> 00:03:02.269 [echoing stops] have seen at least one-half of this talk before, 00:03:02.269 --> 00:03:04.910 so thank you for being with me again one more time. 00:03:04.910 --> 00:03:08.175 I hope these will still be interesting to see. And now actually, 00:03:08.175 --> 00:03:13.089 without further ado, I’m going to start talking about a few damage zone models. 00:03:13.089 --> 00:03:17.019 Why do we care about damage zones? Well, they are a key component of 00:03:17.019 --> 00:03:19.659 the earthquake energy budget in seismic hazard assessment because 00:03:19.659 --> 00:03:23.119 they constitute a permanent sink of strain energy. They modify the 00:03:23.119 --> 00:03:26.030 elastic properties of the shallow crust. They threaten lifelines. 00:03:26.030 --> 00:03:28.689 They amplify ground shaking during earthquakes. 00:03:28.689 --> 00:03:32.959 We all care about damage zones. This problem has attracted geologists, 00:03:32.959 --> 00:03:36.790 geophysicists, geodesists alike. But damage zones are defined 00:03:36.790 --> 00:03:40.059 slightly differently for these fields of geoscience. 00:03:40.059 --> 00:03:43.979 So geologically, we think of damage zones as the fractured and warped 00:03:43.979 --> 00:03:47.140 volume surrounding a fault. And within this volume, fracture 00:03:47.140 --> 00:03:49.889 density decays away from a principal slip surface until 00:03:49.889 --> 00:03:54.560 it reaches some background level. As an example – can you see my cursor? 00:03:54.560 --> 00:03:57.829 Yeah. Okay. Here’s an example from the field from Mitchell and Faulkner. 00:03:57.829 --> 00:04:01.689 You can see fracture density on the Y axis and distance away 00:04:01.689 --> 00:04:05.139 from the fault on the X axis. And you can see how that decays. 00:04:05.139 --> 00:04:07.780 The same approach has been taken with aftershocks where damage zone 00:04:07.780 --> 00:04:11.599 is defined as a volume where seismicity decreases until it reaches some 00:04:11.599 --> 00:04:16.590 background level. An example at the bottom from Powers and Jordan 2020. 00:04:16.590 --> 00:04:19.750 And different functional fits have been used to describe this decay. 00:04:19.750 --> 00:04:23.199 Typically prefer something like a power law or an exponential decay. 00:04:23.199 --> 00:04:26.441 And most of the knowledge from this approach to damage zone 00:04:26.441 --> 00:04:28.830 comes from strike-slip faults. 00:04:28.830 --> 00:04:32.650 Seismologically, damage zones are measured as a volume of decreased 00:04:32.650 --> 00:04:36.900 shear wave velocity surrounding a fault that can host guided waves. 00:04:36.900 --> 00:04:41.169 And geodetically, we characterize damage zones as areas of reduced shear 00:04:41.169 --> 00:04:44.729 rigidity, or sometimes when we directly image earthquake ruptures, the area 00:04:44.729 --> 00:04:49.360 where strain exceeds the elastic limit of the material, as shown from the work of 00:04:49.360 --> 00:04:54.289 Chelsea Scott on the top plot here. The black lines denote the area that’s 00:04:54.289 --> 00:04:58.830 in blue and yellow where strain has exceeded the elastic limit of rock. 00:04:59.508 --> 00:05:02.830 A question that’s been persistent in damage zone studies has been 00:05:02.830 --> 00:05:05.300 whether there are breaks in scaling within the damage zone. 00:05:05.300 --> 00:05:08.320 And the reason we care about that is because, if there are breaks in scaling, 00:05:08.320 --> 00:05:11.370 there may be different physics that are operating at different distances away 00:05:11.370 --> 00:05:16.669 from the fault, dominating deformation. So, from their seminal work in 2010, 00:05:16.669 --> 00:05:19.680 Powers and Jordan interpreted three zones from their decay and aftershock 00:05:19.680 --> 00:05:22.483 density with distance away from the fault. 00:05:22.483 --> 00:05:27.789 Close to the intercept, a zone where aftershock density stays 00:05:27.789 --> 00:05:33.229 constant until a certain distance. This is a break in scaling. 00:05:33.229 --> 00:05:37.789 This break in scaling is followed by an area where density decays with 00:05:37.789 --> 00:05:41.099 a constant slope that ends where the dashed line is, 00:05:41.099 --> 00:05:44.550 which is where this decay meets the background seismicity. 00:05:44.550 --> 00:05:48.780 And these plots inspire the idea of an inner damage zone, shown in 00:05:48.780 --> 00:05:52.450 the cartoons on the right, that contains one or several strands of fault core, 00:05:52.450 --> 00:05:55.199 and the width of that inner damage zone is determined by 00:05:55.199 --> 00:05:59.699 how many strands of fault core are within it. 00:05:59.699 --> 00:06:02.419 And this ends in a break in scaling between the inner and 00:06:02.419 --> 00:06:04.879 the outer damage zone. And some field geology studies have 00:06:04.879 --> 00:06:09.749 also supported this conceptual model of an inner versus and outer damage zone. 00:06:10.501 --> 00:06:16.300 A challenge in damage zone studies is that we typically think of damage zones 00:06:16.300 --> 00:06:20.499 from a single data set – for example, a fracture or an aftershock distribution. 00:06:20.499 --> 00:06:24.060 And these data sets often capture an integrated history of deformation 00:06:24.060 --> 00:06:25.860 that spans multiple earthquake cycles. 00:06:25.860 --> 00:06:29.064 So may have contributions from aseismic processes. 00:06:29.064 --> 00:06:32.250 The 2019 Ridgecrest earthquakes provide this unprecedented view 00:06:32.250 --> 00:06:35.819 of the distribution of inelastic deformation from an earthquake. 00:06:35.819 --> 00:06:41.190 We had so many data sets covering it. So we want to incorporate different 00:06:41.190 --> 00:06:45.970 data sets from the same earthquake to see if we can reconcile the view 00:06:45.970 --> 00:06:49.039 of the damage zone from a bunch of different spatial scales that were 00:06:49.039 --> 00:06:52.569 covered from these data sets. So we start by incorporating three 00:06:52.569 --> 00:06:55.460 different independent rupture mapping data sets. 00:06:55.460 --> 00:06:58.444 The first one is a rupture map of Ponti et al. 2020 that I know 00:06:58.469 --> 00:07:01.189 many in this group actually collaborated in making. 00:07:01.189 --> 00:07:04.139 It’s not shown in this slide but was included in this study. 00:07:04.139 --> 00:07:07.370 This map was generated from field and geodetic observations. 00:07:07.370 --> 00:07:10.620 The second one is our own map of the surface rupture that was generated 00:07:10.620 --> 00:07:14.550 from post-earthquake Lidar data. I’m showing you that on the top left. 00:07:14.550 --> 00:07:19.039 The north-south-trending features are related to the main shock broadly, 00:07:19.039 --> 00:07:26.379 and the east-west-trending features are related with the 6.4 foreshock. 00:07:26.379 --> 00:07:31.849 We supplement these two rupture maps with an additional rupture map 00:07:31.849 --> 00:07:35.580 that covers five sections that are highlighted in orange. 00:07:35.580 --> 00:07:39.159 And these sections were mapped from high-resolution aerial imagery either 00:07:39.159 --> 00:07:43.439 collected from a drone or from the same airplane that collected the Lidar data. 00:07:43.439 --> 00:07:47.110 And these are 2 to 20 centimeters-per-pixel resolution maps. 00:07:47.110 --> 00:07:52.228 An example of one of those sections is shown below these two maps on top. 00:07:52.228 --> 00:07:57.159 Okay. We also include two earthquake catalogs in our analysis. 00:07:57.159 --> 00:08:00.233 For those that are into earthquake catalogs, these are the QTM and the 00:08:00.233 --> 00:08:06.702 SESM catalogs, which are shown – the QTM catalog is shown on the top right. 00:08:06.740 --> 00:08:10.409 Okay, so let’s just take a closer look at one of these high-resolution maps. 00:08:10.409 --> 00:08:16.315 This is the middle of the magnitude 7.4 – 6.4, sorry, foreshock. And what I want to 00:08:16.315 --> 00:08:20.174 highlight here is that the rupture is pretty complex at this scale. 00:08:20.174 --> 00:08:24.680 You can see that there are two parallel fault strands, that things branch out. 00:08:24.680 --> 00:08:29.109 In some locations, the rupture goes blind and distributed, localizes again, 00:08:29.109 --> 00:08:32.440 distributes one more time. and I also want to highlight here 00:08:32.440 --> 00:08:36.089 that the cracks go all the way until the edge of this footprint. 00:08:36.089 --> 00:08:39.529 They’re not just limited to the primary rupture. 00:08:39.529 --> 00:08:42.214 So the whole area is pretty heavily fractured. 00:08:42.214 --> 00:08:44.340 This is just a field view of what that looks like. 00:08:44.340 --> 00:08:47.850 You can see a big fault scarp here and all of this distributed 00:08:47.850 --> 00:08:50.670 fracturing away from it. 00:08:50.670 --> 00:08:54.050 This is just a 3D view of the surface rupture from Lidar point clouds. 00:08:54.050 --> 00:08:58.000 I don’t know – the video doesn’t seem to want to play. 00:08:58.920 --> 00:09:00.797 Oh, it is playing now. Okay. 00:09:00.797 --> 00:09:03.420 So you can see there are two fault strands here. 00:09:03.420 --> 00:09:08.670 You can see a few mole tracks in this view as well. 00:09:08.670 --> 00:09:11.370 This is going to turn. You’re going to be able to see that there 00:09:11.370 --> 00:09:15.610 are some sections where the Lidar point cloud penetrated into cracks. 00:09:15.610 --> 00:09:18.655 So you can actually see the down-dip extent of some of these 00:09:18.655 --> 00:09:22.810 fractures pretty neatly from the Lidar point cloud. 00:09:22.810 --> 00:09:27.670 Okay, so I said we have three different surface rupture map data sets, 00:09:27.670 --> 00:09:32.279 two earthquake catalogs, and we’re going to supplement these with maps of 00:09:32.279 --> 00:09:36.540 the distribution of strain mapped from satellite imagery cross-correlation. 00:09:36.540 --> 00:09:40.160 And here we use fault-parallel shear strains that are measured from profiles 00:09:40.160 --> 00:09:43.730 that are oriented perpendicular to the primary rupture. 00:09:43.730 --> 00:09:47.029 And these are stacked over 138-meter distance. 00:09:47.029 --> 00:09:53.625 So each one of these – having a hard time moving the cursor here – okay. 00:09:53.657 --> 00:09:59.620 So the fault is somewhere around here, and each of these pixels has 138 meters 00:09:59.620 --> 00:10:03.160 along strike and 12-meter windows across strike. 00:10:03.160 --> 00:10:06.750 And, from these profiles, we see the transition from inelastic to 00:10:06.750 --> 00:10:10.570 predominantly elastic happens around 10 to 70 meters away from the fault, 00:10:10.570 --> 00:10:13.279 depending on location. And beyond a kilometer away 00:10:13.279 --> 00:10:15.949 from the fault, the strains are too small to resolve from noise. 00:10:15.949 --> 00:10:20.635 So we only use this data from zero to a few kilometers away from the fault. 00:10:20.635 --> 00:10:24.820 So, to put these data sets together to say something about the distribution of 00:10:24.820 --> 00:10:31.149 inelastic deformation from the fault, we are going to plot how fracture 00:10:31.149 --> 00:10:34.560 density, aftershock density, and strain intensity for stacked 00:10:34.560 --> 00:10:39.100 profiles across the rupture decay with fault-perpendicular distance. 00:10:39.100 --> 00:10:42.639 And I guess it’s worth here pausing for a second and defining what I mean 00:10:42.639 --> 00:10:45.128 every time I say “inelastic deformation.” 00:10:45.171 --> 00:10:48.730 When I say inelastic, I just mean a point at which a material has yielded. 00:10:48.730 --> 00:10:53.990 For most of these data sets, that is in the form of fracture from the rupture maps 00:10:53.990 --> 00:10:57.540 and from the aftershocks, but also the strain maps may contain additional 00:10:57.540 --> 00:11:01.223 processes that lead to yielding like warping or granular processes. 00:11:01.223 --> 00:11:04.300 That’s just worth keeping in mind. Okay. 00:11:04.300 --> 00:11:07.600 So here I’ve measured the nearest distance between each damage feature. 00:11:07.600 --> 00:11:11.680 So a fracture, an aftershock, or a [inaudible] strain map. 00:11:11.680 --> 00:11:15.300 And the – oh, whoops. That moved on its own. 00:11:15.300 --> 00:11:18.769 And the nearest point on the fault. That is done in 2D for the fractures 00:11:18.769 --> 00:11:22.699 and the strain maps, and that is done in 3D for the aftershocks. 00:11:22.699 --> 00:11:26.410 And we fit these decays following the expression in Powers and Jordan 2010, 00:11:26.410 --> 00:11:31.089 which is shown here on the slide. And the decay is defined by the fracture 00:11:31.089 --> 00:11:37.990 density, d-naught, at the intercept. D, which is a break in scaling at 00:11:37.990 --> 00:11:39.730 a given distance away from the fault. 00:11:39.730 --> 00:11:43.559 Here for the aftershocks, it’s about 50 meters away from the fault 00:11:43.559 --> 00:11:47.040 M, which is a parameter that defines the sharpness of that corner, 00:11:47.040 --> 00:11:49.649 and we hold that equal to 2. And it doesn’t really change the 00:11:49.649 --> 00:11:52.620 way the decay looks like, so to reduce the number of parameters 00:11:52.620 --> 00:11:56.871 and following Powers and Jordan, we hold it equal to 2. 00:11:56.871 --> 00:12:00.220 Okay, and then last, gamma, which is the decay – 00:12:00.220 --> 00:12:04.190 sorry, the slope of the decay beyond that [inaudible]. 00:12:04.190 --> 00:12:06.009 This is shown here in green for the aftershocks. 00:12:06.009 --> 00:12:08.490 I’m going to start slowly populating these plots. 00:12:08.490 --> 00:12:11.880 I’m going to add the decay for fracture density from the 00:12:11.880 --> 00:12:16.020 high-resolution imagery. And the thing that’s obvious here is that that 00:12:16.020 --> 00:12:19.960 [inaudible] has moved pretty far inward, matched closer to the fault than it 00:12:19.960 --> 00:12:23.759 was for the aftershock data. But the slope gamma is actually 00:12:23.759 --> 00:12:28.202 pretty similar for both the imagery and the aftershocks. 00:12:28.202 --> 00:12:31.540 Okay. I’ve populated this with all of the data sets now. 00:12:31.540 --> 00:12:36.150 I have aftershocks, the Lidar cracks, the field-plus cracks, which are from 00:12:36.150 --> 00:12:39.850 the Ponti map, the imagery, and the strain, which, instead of 00:12:39.850 --> 00:12:44.461 density for strain, it’s strain intensity on the Y axis. 00:12:44.461 --> 00:12:48.130 And you can see these look pretty similar. 00:12:48.130 --> 00:12:51.339 Look at the distribution of those parameters from the Powers and 00:12:51.339 --> 00:12:56.220 Jordan fit. Starting with Panel B here, I’m showing you the distribution of the 00:12:56.220 --> 00:12:59.880 maximum likelihood values of gamma, which is that slope. 00:12:59.880 --> 00:13:05.170 And it ranges from about 0.8 to 1.1, and you can see that these distributions 00:13:05.170 --> 00:13:07.477 are different, but they overlap within uncertainty. 00:13:07.477 --> 00:13:11.980 And we interpret the similar slopes to indicate that fractures, aftershocks, 00:13:11.980 --> 00:13:15.980 and other processes that are captured by the strain maps are controlled by 00:13:15.980 --> 00:13:20.079 the same mechanism and strain field given the overlap of these slopes. 00:13:20.079 --> 00:13:24.260 Let’s all take a look at that parameter d, which is the corner – the breaking 00:13:24.260 --> 00:13:28.560 scaling of the inner damage zone, which is shown in Panel C. 00:13:28.560 --> 00:13:31.019 Here I’m showing you again the maximum likelihood values, 00:13:31.019 --> 00:13:35.389 and I’ve added the uncertainty in the fault location at the surface and at depth 00:13:35.389 --> 00:13:39.230 in the plots, which is shown by the orange shading in the background. 00:13:39.230 --> 00:13:42.649 The corner position, d, is out farthest out for the aftershocks. 00:13:42.649 --> 00:13:46.259 It’s 50 meters away from the fault, but it sits between zero and 12 meters 00:13:46.259 --> 00:13:51.019 for the higher spatial resolution Lidar-, field-, and imagery-derived data sets. 00:13:51.019 --> 00:13:54.139 And it cannot be resolved for the strain data. 00:13:54.139 --> 00:13:58.500 And note that d is on the order of the spatial error of the main rupture trace 00:13:58.500 --> 00:14:02.540 location at the surface, and the fits allow for the absence of a corner. 00:14:02.540 --> 00:14:05.040 So d is allowed to be equal to zero. 00:14:05.040 --> 00:14:08.730 Now, for the aftershock data, d is comparable to the 00:14:08.730 --> 00:14:11.790 100-meter horizontal location uncertainty of events that are 00:14:11.790 --> 00:14:16.807 used to define the subsurface fault geometry. 00:14:16.807 --> 00:14:24.230 Okay. So, given that we have these distributions that span large spatial 00:14:24.230 --> 00:14:30.269 footprints, we want to compare them to prior studies that looked at 00:14:30.269 --> 00:14:33.350 inelastic deformation from the Ridgecrest earthquake. 00:14:33.350 --> 00:14:36.310 So here I’m going to be plotting the results from these studies 00:14:36.310 --> 00:14:40.329 at the bottom half of this plot. And what you want to look at here 00:14:40.329 --> 00:14:42.759 is not really the Y axis of these, but it’s the X axis. 00:14:42.759 --> 00:14:47.110 So the X axis denotes the maximum width of inelastic deformation 00:14:47.110 --> 00:14:50.220 from each of these studies. So I’m starting here with the strain 00:14:50.220 --> 00:14:55.254 maps from Chris Milliner’s work published in JGR last year. 00:14:55.254 --> 00:14:59.402 And here, they were looking at the second invariant of the strain tensor. 00:14:59.402 --> 00:15:06.286 And the width of inelastic deformation from that was calculated to be about 00:15:06.286 --> 00:15:12.700 30 to 100 meters away – wide, sorry – and that’s centered on the fault. 00:15:12.700 --> 00:15:15.529 Okay. There are more data sets we can look at. 00:15:15.529 --> 00:15:19.970 Here I’m plotting the relative surface displacements from the work of 00:15:19.970 --> 00:15:23.959 Solène Antoine, also published last year. And relative surface displacements 00:15:23.959 --> 00:15:29.029 just means fault zone width for a broad zone of deformation that receives 00:15:29.029 --> 00:15:33.370 contributions both from elastic and inelastic processes. 00:15:33.370 --> 00:15:37.639 You can see on the right plot each of the lines denotes the width 00:15:37.639 --> 00:15:40.920 of that zone of deformation that, again, receives contributions 00:15:40.920 --> 00:15:43.720 from elastic and inelastic processes. 00:15:43.720 --> 00:15:48.232 And the width of the zones is about 600 meters to 2 kilometers, depending 00:15:48.232 --> 00:15:55.779 on what section you’re looking at. And the black lines denote the mean 00:15:55.779 --> 00:15:59.060 for the foreshock and the main shock, respectively. 00:15:59.060 --> 00:16:03.050 We can also look at depth at the distribution of low-velocity zones 00:16:03.050 --> 00:16:07.319 that were imaged seismologically. So here I’ve added two different lines 00:16:07.319 --> 00:16:11.620 in green, both low-velocity zones. And then, within low-velocity zones, 00:16:11.620 --> 00:16:15.959 the width of trapped wave zones that are enclosed by those low-velocity zones. 00:16:15.959 --> 00:16:22.949 And you can see that those range from 100 meters to 2 kilometers in depth. 00:16:22.949 --> 00:16:27.069 And I want you to note here that the corner is d, so that break in the inner 00:16:27.069 --> 00:16:32.959 damage zone, that may or may not be allowed to be equal to zero, 00:16:32.959 --> 00:16:37.250 does not match the width of any of the extents of inelastic 00:16:37.250 --> 00:16:41.470 deformation that are imaged from these other data sets. 00:16:43.109 --> 00:16:50.920 So, from this, we propose that d is actually an artifact in data resolution. 00:16:50.920 --> 00:16:53.980 And the fault damage manifests in a continuum that doesn’t have a break 00:16:53.980 --> 00:17:00.420 in scaling. So, in other words, there is no inner damage zone. 00:17:00.420 --> 00:17:04.580 And something else worth pointing out here is that the decay of damage 00:17:04.580 --> 00:17:08.660 is consistent out to pretty far distances away from the fault. 00:17:08.660 --> 00:17:13.400 We can see cracks that are over 10 kilometers away from the fault. 00:17:13.400 --> 00:17:22.510 So, to recap, from our overlap in gamma, we propose that these data sets 00:17:22.510 --> 00:17:26.560 are showing that the damage zone is controlled by the same mechanism or 00:17:26.560 --> 00:17:30.130 stress field up until pretty far distances away from the fault. 00:17:30.130 --> 00:17:34.460 Now, we cannot use these decay exponents as the value of gamma alone 00:17:34.460 --> 00:17:38.070 to distinguish between static and dynamic triggering processes, 00:17:38.070 --> 00:17:41.680 but it looks like whatever is happening is pretty homogeneous up to pretty far 00:17:41.680 --> 00:17:46.110 distances away from the fault. And a few questions that naturally come 00:17:46.110 --> 00:17:50.708 out of this is, well, are these damage decays sensitive to lithology? 00:17:50.708 --> 00:17:54.330 So, to do this, to investigate this, we filter fractures cutting through 00:17:54.330 --> 00:17:57.981 sediment and bedrock. It is important to note that, as you 00:17:57.981 --> 00:18:02.850 can see from the map on the left, the bedrock area we sample is really small 00:18:02.850 --> 00:18:06.459 because the earthquake overwhelmingly ruptured through sediment. 00:18:06.459 --> 00:18:11.000 But, even for this, we find the fracture density in sediment is 3 to 10 times 00:18:11.000 --> 00:18:14.459 higher than that of bedrock at a given distance away from the fault. 00:18:14.459 --> 00:18:17.930 But, however, the decay in fracture density with distance is identical. 00:18:17.930 --> 00:18:21.366 So the slope of that curve is identical. 00:18:21.366 --> 00:18:23.790 And something that may be leading to these decreased 00:18:23.790 --> 00:18:27.920 fracture density in bedrock may be greater cohesion of the material. 00:18:27.920 --> 00:18:32.160 Again, those similar decay exponents support that the mechanism responsible 00:18:32.160 --> 00:18:35.820 for the nucleation of fractures operates under the same conditions 00:18:35.820 --> 00:18:39.090 irrespective of material. 00:18:39.090 --> 00:18:42.900 Another question that naturally follows from this is whether these damage 00:18:42.900 --> 00:18:46.939 decays are sensitive to the magnitude of on-fault slip. 00:18:46.939 --> 00:18:55.410 So, to evaluate this, we calculated the average slip and the slip variants, 00:18:55.410 --> 00:18:57.870 which are shown here for each of the sections where we have 00:18:57.870 --> 00:19:01.840 high-resolution imagery maps available. 00:19:01.840 --> 00:19:06.950 And then, for each of these sections, I’ve plotted the fracture density 00:19:06.950 --> 00:19:11.520 decays on the right. And what we find from this is that, 00:19:11.520 --> 00:19:17.460 despite variability in the intercept, so in the fracture density at the overall 00:19:17.460 --> 00:19:22.690 fracture density, we find no correlation between corner, slope, and overall 00:19:22.690 --> 00:19:26.450 fracture density with slip magnitude. The place that has the highest fracture 00:19:26.450 --> 00:19:30.190 density doesn’t have the highest on-fault slip and doesn’t have – 00:19:30.190 --> 00:19:34.120 or any correlation to slip variability. However, we find that the decays 00:19:34.120 --> 00:19:37.750 are relatively consistent throughout the rupture. 00:19:37.750 --> 00:19:42.340 So we’ve earlier noted that the edges of these zones of inelastic deformation 00:19:42.340 --> 00:19:45.600 that are defined by prior studies didn’t overlap with any of the 00:19:45.600 --> 00:19:49.610 corners that are different data sets we’re able to define. 00:19:49.610 --> 00:19:54.400 So we want to compare data sets in map view to see how these fracture 00:19:54.400 --> 00:19:59.840 maps versus these geodetically mapped fractures overlap. 00:19:59.840 --> 00:20:02.990 So here I’m showing you fracture density and strain intensity on the 00:20:02.990 --> 00:20:07.870 left and the right plots, respectively, for two different regions. 00:20:07.870 --> 00:20:12.510 What we find is typically the magnitude of strain, which here is conveyed by the 00:20:12.510 --> 00:20:17.050 second invariant of the strain tensor, and fracture density overlap pretty well. 00:20:17.050 --> 00:20:21.370 You can see that, as these zones narrows and widens, so does the zone mapped 00:20:21.370 --> 00:20:27.470 from fracture density. But occasionally we find areas of high fracture density 00:20:27.470 --> 00:20:32.490 that are disconnected from the main rupture – you can see a few of those 00:20:32.490 --> 00:20:37.420 here – that do not appear on the strain maps. And this is because 00:20:37.420 --> 00:20:40.250 the cumulative displacements of these fractures are too small. 00:20:40.250 --> 00:20:42.600 They’re under 20 centimeters. They’re not large enough to 00:20:42.600 --> 00:20:46.660 appear geodetically. But nevertheless, this deformation 00:20:46.660 --> 00:20:50.680 is there, and it’s important to account for from a probabilistic displacement 00:20:50.680 --> 00:20:55.030 hazard standpoint. So I just want to show you, to highlight the value of 00:20:55.030 --> 00:20:59.060 high-resolution orthophotography in identifying individual fractures, which 00:20:59.060 --> 00:21:03.836 cannot at all be discerned using these image correlation procedures. 00:21:03.836 --> 00:21:09.340 Okay. So, so far we’ve taken a damage zone model, we’ve said something about 00:21:09.340 --> 00:21:14.090 the shape of the distribution of damage from the Ridgecrest earthquakes. 00:21:14.090 --> 00:21:18.560 And we’re interested on whether these damage distributions can be translated 00:21:18.560 --> 00:21:23.876 into something that tells us about the physical properties of the damage zone. 00:21:23.876 --> 00:21:27.690 So, based on the relationship between fracture density and shear rigidity, 00:21:27.690 --> 00:21:31.240 we can estimate how much fracturing contributed to the rigidity decrease 00:21:31.240 --> 00:21:36.330 in the volume of rock surrounding the Ridgecrest Fault. And this requires 00:21:36.330 --> 00:21:40.470 both integrating the density and the length distribution of mapped fractures. 00:21:40.470 --> 00:21:44.580 Okay. So we do that. An important – this is based on 00:21:44.580 --> 00:21:49.890 the work of Budiansky and O’Connell. And an important part of this step is 00:21:49.890 --> 00:21:54.900 that this approach does not distinguish whether the fractures were reactivated 00:21:54.900 --> 00:21:57.600 during the earthquake or they were coseismically nucleated. 00:21:57.600 --> 00:22:01.590 It just says how many cracks are there in this volume, what are their length 00:22:01.590 --> 00:22:05.000 distributions, and how does that – what’s the shear rigidity decrease 00:22:05.000 --> 00:22:07.100 that emerges from that. 00:22:07.601 --> 00:22:13.030 So, from this, we estimate a decrease of about 40% in sediment immediately 00:22:13.030 --> 00:22:16.360 adjacent to the fault and a decrease of about 20% in bedrock. 00:22:16.360 --> 00:22:22.210 And these decline pretty fast to under 1% at 100 meters away from the fault. 00:22:22.210 --> 00:22:26.120 These estimates are within the rigidity decrease that was geodetically 00:22:26.120 --> 00:22:28.823 estimated by the work of Eric Xu. 00:22:28.823 --> 00:22:34.020 And our estimates are based on fracture density mapped at the surface. 00:22:34.020 --> 00:22:38.700 But the width of reduced rigidity that we predict compares very well 00:22:38.700 --> 00:22:42.120 to the width and velocity structure of the trapped wave zones that were 00:22:42.120 --> 00:22:46.090 observed postseismically at Ridgecrest. These zones typically extend 00:22:46.090 --> 00:22:50.890 3 to 5 kilometers down-dip and remain constant in width. 00:22:50.890 --> 00:22:56.590 And one thing that’s expected from this analysis is that there are low – 00:22:56.590 --> 00:23:00.370 we can predict that there are low levels of rigidity reduction far away from 00:23:00.370 --> 00:23:04.160 the fault. Because we see fractures extending pretty far away. 00:23:04.160 --> 00:23:06.900 And this finding is supported by regional reductions 00:23:06.900 --> 00:23:09.250 in shear wave velocity after the Ridgecrest earthquakes. 00:23:09.250 --> 00:23:11.150 These are measurements from the far-field. 00:23:11.150 --> 00:23:13.510 They’re averaging through a really large crustal volume. 00:23:13.510 --> 00:23:17.666 And they show a ubiquitous small – around 2% velocity reduction 00:23:17.666 --> 00:23:19.570 over this volume, which we think may be akin 00:23:19.570 --> 00:23:24.142 to the widespread damage we’re mapping at the surface. 00:23:24.142 --> 00:23:29.430 And these observations simultaneously reveal two things, which is, one, 00:23:29.430 --> 00:23:32.420 fracturing generates this intense near-field damage. 00:23:32.420 --> 00:23:35.790 And then widespread damage is accumulating in the volume 00:23:35.790 --> 00:23:40.060 surrounding the fault. These two behaviors emerge from 00:23:40.060 --> 00:23:43.700 the same nonlinear distribution. It’s the same power law distribution. 00:23:43.700 --> 00:23:47.430 You don’t need any break in scalings to make these intensity near-field 00:23:47.430 --> 00:23:51.990 damage and widespread rock damage on the overall volume. 00:23:51.990 --> 00:23:53.580 We know the damage zones are pretty heterogeneous. 00:23:53.580 --> 00:23:56.860 We’ve seen this from field observations and seismological constraints in 00:23:56.860 --> 00:23:59.350 low-velocity zones again and again for Ridgecrest. 00:23:59.350 --> 00:24:02.130 These low-velocity zones change along strike. 00:24:02.130 --> 00:24:05.950 We know there’s variability. But, within this heterogeneity, 00:24:05.950 --> 00:24:09.220 there emerge simple relationships that describe the average intensity 00:24:09.220 --> 00:24:11.220 of deformation. And these may be useful 00:24:11.220 --> 00:24:14.942 for developing generalized models of fault damage zones. 00:24:15.707 --> 00:24:21.400 And there are, I think, still many outstanding questions from this work, 00:24:21.400 --> 00:24:23.910 so I’m going to put a few of these out there for the room to see 00:24:23.910 --> 00:24:27.220 if you can help me answer them. So first one says, how much of 00:24:27.220 --> 00:24:31.250 the damage we map is inherited? How many of these cracks are 00:24:31.250 --> 00:24:35.330 getting reactivated during the Ridgecrest earthquake versus newly nucleated, 00:24:35.330 --> 00:24:39.900 and how does that distinction work for bedrock versus sediment? 00:24:39.900 --> 00:24:43.740 Kind of the flip side to that is, how much damage is recovered 00:24:43.740 --> 00:24:46.770 and healed after the earthquake? We see damage zones in the late 00:24:46.770 --> 00:24:50.740 interseismic stages of faults, so we know that a portion of the damage 00:24:50.740 --> 00:24:55.380 has to be permanent and irrecoverable. But how much of it is recovered 00:24:55.380 --> 00:24:58.090 after the event? A 40% shear modulus decrease 00:24:58.090 --> 00:25:00.510 is very high to maintain over long time scales, right? 00:25:00.510 --> 00:25:03.858 Like, the cracks would be soup after a couple of earthquakes. 00:25:03.858 --> 00:25:08.490 And then last, what happens to the crust in zones of dense faulting such as 00:25:08.490 --> 00:25:11.350 the Eastern California Shear Zone? So here we’re seeing that there are 00:25:11.350 --> 00:25:15.170 cracks that are 15 kilometers away from the main Ridgecrest Fault. 00:25:15.170 --> 00:25:19.000 But those cracks are also 10 kilometers away from the Garlock Fault. 00:25:19.000 --> 00:25:21.630 They’re 5 kilometers away from the next fault over. 00:25:21.630 --> 00:25:24.530 So what happens in these zones of dense faulting that are 00:25:24.530 --> 00:25:27.490 having frequent earthquakes, that are having inelastic 00:25:27.490 --> 00:25:30.685 deformation over these really large footprints? 00:25:31.380 --> 00:25:34.980 And then my last question is whether Ridgecrest is an exception. 00:25:34.980 --> 00:25:38.170 We know Ridgecrest is weird. We have this orthogonal cracks. 00:25:38.170 --> 00:25:41.300 We have this really large footprint that we were able to map. 00:25:41.300 --> 00:25:43.000 There are cracks everywhere. 00:25:43.000 --> 00:25:46.063 Is this a good model for thinking of earthquakes in general? 00:25:46.063 --> 00:25:50.266 And the answer is, well, it depends what earthquake, obviously. [laughs] 00:25:50.266 --> 00:25:52.710 So this is an example from the Madoi earthquake. 00:25:52.710 --> 00:25:55.270 This is work that’s being done by Jing Liu in her group 00:25:55.270 --> 00:25:58.202 at the China Earthquake Administration. 00:25:58.202 --> 00:26:03.790 And what we find is that the Powers and Jordan model fits their fracture 00:26:03.790 --> 00:26:07.230 density decays really well for subsections of the rupture. 00:26:07.230 --> 00:26:10.140 An example of that is on top. But when you look at all of the 00:26:10.140 --> 00:26:13.110 data collapsed together, it’s not a very good model. 00:26:13.110 --> 00:26:16.600 It looks more like there are two – actually two distinct zones. 00:26:16.600 --> 00:26:20.590 This is work that’s in the early stages, so it may change from now on, 00:26:20.590 --> 00:26:24.380 but we’re already seeing that what defines Ridgecrest very well may not 00:26:24.380 --> 00:26:27.820 be a good model for many other events. 00:26:27.820 --> 00:26:30.320 But what if we look at Ridgecrest neighbors? 00:26:30.320 --> 00:26:33.288 What if we look at the Eastern California Shear Zone? 00:26:33.288 --> 00:26:37.108 I’ve looked at the fracture density decays for Landers, Hector Mine, 00:26:37.108 --> 00:26:40.577 and El Mayor-Cucapah, together with Ridgecrest. 00:26:40.577 --> 00:26:43.510 So I’m showing you those decays here with distribution of gamma 00:26:43.510 --> 00:26:47.350 at the bottom. And we see that other Eastern California Shear Zone events 00:26:47.350 --> 00:26:51.350 also exhibit similar fault-perpendicular fracture density decays and 00:26:51.350 --> 00:26:53.940 widespread damage. So maybe this is a good 00:26:53.940 --> 00:26:59.000 damage zone model to think about immature fault zones. 00:26:59.000 --> 00:27:02.630 And I want to shift gears here and start talking a little bit more about the 00:27:02.630 --> 00:27:05.510 multi-fault aspect of these earthquakes. 00:27:05.510 --> 00:27:09.830 All these events have in common that they rupture several faults. 00:27:09.830 --> 00:27:14.950 The earthquake have to transfer faults at several locations. 00:27:14.950 --> 00:27:18.390 And multi-fault earthquakes are tricky. They are hard to identify in 00:27:18.390 --> 00:27:22.370 the paleoseismic record. They’re hard to correlate across faults. 00:27:22.370 --> 00:27:25.270 They are able to produce greater magnitude events and earthquakes 00:27:25.270 --> 00:27:29.080 confined to a single fault, obviously. And they lead to longer sustained 00:27:29.080 --> 00:27:33.340 more widespread strong ground motions and surface rupture. 00:27:33.340 --> 00:27:36.540 When thinking of them in the long-term, they alter the stress state or probability 00:27:36.540 --> 00:27:39.690 of events throughout the fault system, which makes them important 00:27:39.690 --> 00:27:42.705 to deal with in long-term simulators. 00:27:42.705 --> 00:27:45.910 For these events in the Eastern California Shear Zone, they happened 00:27:45.910 --> 00:27:52.630 in pretty lower-populated zones in faults that have pretty slow slip rates. 00:27:52.630 --> 00:27:57.410 But we know fault junctions that have faults that are slipping much faster and 00:27:57.410 --> 00:28:01.750 are much more widely populated. And an example of that is Cajon Pass, 00:28:01.750 --> 00:28:04.450 which is just north of L.A. where the San Andreas and the 00:28:04.450 --> 00:28:07.780 San Jacinto Faults come together. These faults carry the majority of the 00:28:07.780 --> 00:28:11.120 Pacific North America Plate motion. They’re fast-slipping. 00:28:11.120 --> 00:28:15.390 And one of the things that have motivated thinking about multi-fault 00:28:15.390 --> 00:28:21.230 events in the context of these two faults is the enigmatic 1812 event which has 00:28:21.230 --> 00:28:25.480 been compellingly shown to have involved both faults from dynamic 00:28:25.480 --> 00:28:29.754 models, but is there geologic evidence for this multi-fault rupture? 00:28:29.754 --> 00:28:32.460 Well, we’re not the first ones to look for this. 00:28:32.460 --> 00:28:36.510 And the challenge for identifying these joint earthquake paleoseismically 00:28:36.510 --> 00:28:44.040 is that the northern San Jacinto has a very challenging geomorphic setting. 00:28:44.040 --> 00:28:46.160 It looks like this. Where do you dig your trench? 00:28:46.160 --> 00:28:50.030 The fault is obviously at the bottom of the valley, but there are 00:28:50.030 --> 00:28:52.740 these huge boulders. We don’t really know where they are, 00:28:52.740 --> 00:28:59.230 and there’s no good fine-grain material deposited for a trench. 00:28:59.230 --> 00:29:03.791 And this has led to the northern section of the San Jacinto, which you 00:29:03.791 --> 00:29:06.870 can see here a map of the San Andreas and the San Jacinto in black. 00:29:06.870 --> 00:29:10.440 The dots represent different paleoseismic sites along strike 00:29:10.440 --> 00:29:15.670 of these faults. And you can see that, from Colton all the way to Cajon Pass, 00:29:15.670 --> 00:29:20.290 there’s these 30-kilometer gaps. You can fit a magnitude 6 event there, 00:29:20.290 --> 00:29:25.210 the surface rupture, and not have any paleoseismic record of that event. 00:29:25.210 --> 00:29:28.700 So today I’m going to be showing you some work regarding the Lytle Creek 00:29:28.700 --> 00:29:32.100 Ridge Fault, which is a low-angle normal fault that we believe has 00:29:32.100 --> 00:29:35.970 the potential to record multi-fault events through Cajon Pass. 00:29:35.970 --> 00:29:38.880 This is a low-angle shallow normal fault. 00:29:38.880 --> 00:29:42.250 It’s too shallow and too short to be seismogenic on its own 00:29:42.250 --> 00:29:46.460 or to act as a transfer structure during multi-fault earthquakes, but it is ideally 00:29:46.460 --> 00:29:50.060 positioned as a passive recorder of linked rupture. 00:29:50.060 --> 00:29:52.401 So the question I’m going to be addressing today is whether we can 00:29:52.401 --> 00:29:56.270 use these small faults to understand what the big players are doing at these 00:29:56.270 --> 00:30:00.270 fault junctions where earthquakes may be transferring between them. 00:30:00.270 --> 00:30:06.570 So let’s get to know Lytle Creek Ridge Fault, or LCRF, a little bit better. 00:30:06.570 --> 00:30:09.280 Here I’m showing you the Lytle Creek Ridge Fault cutting 00:30:09.280 --> 00:30:12.850 through the Pelona Schist. This is bedrock exposure. 00:30:12.850 --> 00:30:16.730 Here I’m showing you the Lytle Creek Ridge Fault exposed in our trench that 00:30:16.730 --> 00:30:20.410 I’m going to be talking about later. And last here’s the Lytle Creek Ridge 00:30:20.410 --> 00:30:25.940 Fault localized along a serpentine band. So it has a lot of flavors along strike. 00:30:25.940 --> 00:30:29.270 To test whether the LCRF is a good passive recorder, 00:30:29.270 --> 00:30:33.760 we need to have a good sense of when does this fault slip. 00:30:33.760 --> 00:30:37.990 So the LCRF is a low-angle normal fault, but it’s sitting on the middle 00:30:37.990 --> 00:30:42.860 of the Transverse Ranges. So the first thing we need for this 00:30:42.860 --> 00:30:46.660 low-angle normal fault to be activated in the strong regional north-south 00:30:46.660 --> 00:30:49.350 contraction of the Transverse Ranges 00:30:49.350 --> 00:30:54.900 is strong dilation that is able to overcome that contraction. 00:30:54.900 --> 00:30:58.860 So I’m going to be considering three different scenarios here. 00:30:58.860 --> 00:31:02.770 All of these are a combination of the suite of Coulomb stress models 00:31:02.770 --> 00:31:06.160 and finite element models. And what I’m showing you here – 00:31:06.160 --> 00:31:10.610 in orange, I’m going to have the fault that’s slipping in this model. 00:31:10.610 --> 00:31:15.480 On the top left, I have the slip vector on the LCRF. 00:31:15.480 --> 00:31:19.410 Underneath it, I have the amount of stress that’s triggered on the LCRF. 00:31:19.410 --> 00:31:23.401 And then, at the bottom of it, I have the amount of slip that’s triggered. 00:31:23.401 --> 00:31:30.450 Okay. So first – sorry, the first scenario here is rupture on the San Andreas 00:31:30.450 --> 00:31:33.970 Fault that goes past Cajon Pass, and it imposes mostly left-lateral 00:31:33.970 --> 00:31:37.570 motion on the LCRF. So that’s not really consistent with normal motion. 00:31:37.570 --> 00:31:39.660 And it triggers very little slip. 00:31:39.660 --> 00:31:44.030 If we have rupture on the San Jacinto Fault that’s traveling southwards, 00:31:44.030 --> 00:31:47.880 we impose normal sinistral motion on the LCRF, but it triggers 00:31:47.880 --> 00:31:52.179 a relatively small amount of stress and slip on the LCRF. 00:31:52.179 --> 00:31:55.591 Last, if we have a scenario that involves taper slip on both 00:31:55.591 --> 00:31:59.460 the San Andreas and the San Jacinto, this triggers double the stress change 00:31:59.460 --> 00:32:03.560 on the LCRF than the scenario that involves the San Jacinto Fault alone. 00:32:03.560 --> 00:32:08.849 So, from this, we hypothesize that the LCRF is most likely to host the large 00:32:08.849 --> 00:32:15.340 event with really large slip, like tens of centimeters, which, even if these 00:32:15.340 --> 00:32:19.943 events on the San Jacinto or on the San Andreas only are also stored – 00:32:19.943 --> 00:32:23.350 sorry, also triggered the LCRF, they impose too little slip 00:32:23.350 --> 00:32:27.570 for that to be recognizable in a paleoseismic trench. 00:32:27.570 --> 00:32:32.592 So we dug a trench in the middle of the San Gabriel Mountains. 00:32:32.592 --> 00:32:38.150 This is an 18 meters long, couple meters wide, 3 meters deep at a couple sections. 00:32:38.150 --> 00:32:41.480 And the fault zone is all concentrated in 3 meters within the fault. 00:32:41.480 --> 00:32:43.520 So most of that trench was actually useless to dig. 00:32:43.520 --> 00:32:45.421 It was just a good workout. [laughter] 00:32:45.421 --> 00:32:48.990 This is the trench record. We found – I’m speeding up here 00:32:48.990 --> 00:32:53.040 because I realize I’m going over time. We have three events. 00:32:53.040 --> 00:32:56.490 The event horizons are highlighted in yellow here. 00:32:56.490 --> 00:33:01.440 And each of these event actually have tens of centimeters of slip each. 00:33:01.440 --> 00:33:04.030 So they’re fairly large events. 00:33:04.030 --> 00:33:08.610 So we took charcoal samples from each of these different trench layers, 00:33:08.610 --> 00:33:10.750 and we dated them. And, from this, we were able to 00:33:10.750 --> 00:33:16.630 constrain ages for events number 1, number 2, and number 3. 00:33:16.630 --> 00:33:20.140 And I want to focus on the most – this is three events over 2,000 years, 00:33:20.140 --> 00:33:25.330 by the way, but I want to focus on the most recent event. 00:33:25.330 --> 00:33:29.440 And here the distributions from the charcoal dates get very multi-modal. 00:33:29.440 --> 00:33:33.320 So can we – can we refine these ages by adding constrains from invasive 00:33:33.320 --> 00:33:37.810 pollen species, which are – we know the introduction dates to California of 00:33:37.810 --> 00:33:41.370 several pollen species pretty well, so can we look at the different samples 00:33:41.370 --> 00:33:45.570 here and say what the most recent event on the LCRF is? 00:33:45.570 --> 00:33:48.590 And there are two candidates here. There’s 1812, which has been previously 00:33:48.590 --> 00:33:51.910 shown compellingly to have dynamically bridged the gap. 00:33:51.910 --> 00:33:55.680 So we expect 1812 to be there. But there’s also 1857, which traveled 00:33:55.680 --> 00:34:00.827 pretty far south on the San Andreas Fault, and it traveled south past 00:34:00.827 --> 00:34:03.790 Cajon Pass, shown in the work of Olaf Zielke. 00:34:03.790 --> 00:34:06.593 So which one of these two events is the one in the trench? 00:34:06.593 --> 00:34:12.080 Well, we were able to retrieve three different samples of significance 00:34:12.080 --> 00:34:15.980 that have species that we know with well-introduced dates. 00:34:15.980 --> 00:34:19.279 These are tamarisk, Spanish broom, and Russian thistle. 00:34:19.279 --> 00:34:22.940 And particularly, tamarisk and Spanish bloom are on the un-faulted 00:34:22.940 --> 00:34:27.070 unit that caps the most recent event. And those were introduced to 00:34:27.070 --> 00:34:32.734 California in 1839 and 1848, respectively, which really make 1812 00:34:32.734 --> 00:34:37.720 a really good candidate to be the most recent event in the trench. 00:34:37.720 --> 00:34:40.889 So the recurrence interval of the LCRF from these three events 00:34:40.889 --> 00:34:47.560 is about 660 years. And we can compare our data here to chronologies on the 00:34:47.560 --> 00:34:51.919 San Andreas and the San Jacinto Faults to – if we assume that the record 00:34:51.919 --> 00:34:55.919 in the trench only contains large events that represent joint rupture, 00:34:55.919 --> 00:34:59.000 we can look at the number of events in each of the paleoseismic sites 00:34:59.000 --> 00:35:02.000 north and south of Cajon Pass on the San Andreas and the San Jacinto, 00:35:02.000 --> 00:35:04.079 and we can compare the numbers. 00:35:04.079 --> 00:35:07.650 We cannot correlate specific events across the three faults because the 00:35:07.650 --> 00:35:10.720 probability density functions are very abrupt, particularly 00:35:10.720 --> 00:35:13.869 the ones on the LCRF. But we can get an understanding 00:35:13.869 --> 00:35:17.099 of the frequency of joint ruptures through these earthquakes. 00:35:17.099 --> 00:35:21.130 So if I count all of the earthquakes on the Wrightwood, Mystic Lake, 00:35:21.130 --> 00:35:25.544 and LCRF trenches, we end up finding that 20%, so three out of 15, 00:35:25.544 --> 00:35:29.140 and 23% – 3 out of 13 – of the events are shared respectively 00:35:29.140 --> 00:35:33.005 between these San Andreas the San Jacinto Faults. 00:35:33.248 --> 00:35:35.870 We did a bunch of modeling to determine the mechanics of these 00:35:35.870 --> 00:35:39.139 earthquakes, but I will only go there if people want to ask at the end 00:35:39.139 --> 00:35:43.279 because I don’t want to use your time. So, to conclude this part of the talk, 00:35:43.279 --> 00:35:46.420 the San Andreas and the San Jacinto Faults have ruptured together in the 00:35:46.420 --> 00:35:51.500 past 2,000 years three times, most recently in the historic 1812 event. 00:35:51.500 --> 00:35:56.230 And the Lytle Creek Ridge Fault ruptures passively every 660 years 00:35:56.230 --> 00:36:01.520 or so with 20% and 23% of the ruptures in the paleoseismic record of the 00:36:01.520 --> 00:36:05.359 San Andreas and the San Jacinto Fault respectively are being shared. 00:36:05.359 --> 00:36:06.789 And, with that, I’ll take any questions.