WEBVTT 00:00:03.000 --> 00:00:08.000 Woo welcome back everyone. It is fun. Do talks 3. 00:00:08.000 --> 00:00:15.000 Same as before, same as before, before. We still have more talks than you've ever had before. We have no time. 00:00:15.000 --> 00:00:17.000 Race. Take it away. 00:00:17.000 --> 00:00:21.000 Thanks, sar Thanks for tuning into our 3rd thunder. Talk session. 00:00:21.000 --> 00:00:32.000 Each talk will be 3 min, and each talk will segue directly to the next talk, based on the 1st 2 sessions. It seems like we might have time for one or 2 questions at the end, so please. 00:00:32.000 --> 00:00:38.000 Be ready with your questions, and with that i'll hand it over to sar beth. 00:00:38.000 --> 00:00:51.000 Hey, i'm sarbath, siebury, and i'm gonna talk about the way faults form in laboratory stepovers. So we know in the earth faults are not nice, flat, uniform planes, they're discontinuous, they have gaps, they have steps. 00:00:51.000 --> 00:00:57.000 And so to understand how new faults form, and also how these stepovers can affect ruptures. 00:00:57.000 --> 00:01:02.000 We came up with a novel laboratory procedure that allows us to have 2. 00:01:02.000 --> 00:01:08.000 Like weakened faults that are separated. By a region of intact rock. 00:01:08.000 --> 00:01:11.000 So on the left we have the picture, and then we have the cartoon. 00:01:11.000 --> 00:01:14.000 And so, if you look at the cartoon, the blue lines. 00:01:14.000 --> 00:01:20.000 Are the sealed fractures. This is a saw-cut notch that we backfilled with the cement paste. 00:01:20.000 --> 00:01:28.000 And then between that this red dash line shows the region of intact rock. 00:01:28.000 --> 00:01:33.000 So we load these samples up in a triaxial rig. So you have pressure in all directions. 00:01:33.000 --> 00:01:36.000 And then we press on the ends of the cylinder. 00:01:36.000 --> 00:01:43.000 Until the weakened faults break. And then eventually, until that center region of impact rock breaks. 00:01:43.000 --> 00:01:47.000 And what we end up seeing. So if you see the picture on the right. 00:01:47.000 --> 00:01:55.000 We have kind of these smooth, grayish surfaces. And so that's where we had the sealed fracture with the cement paste. 00:01:55.000 --> 00:02:03.000 And then. In the middle. This roughened brown area is the newly formed intact rock. 00:02:03.000 --> 00:02:09.000 So that's where rupture, where that area was loaded up, both by the. 00:02:09.000 --> 00:02:14.000 weakened faults, and also by the voting on the ends of the sample. 00:02:14.000 --> 00:02:23.000 Before breaking. And so we take our preexisting, weakened, sealed fractures, and we offset them. 00:02:23.000 --> 00:02:28.000 So we do a range of offsets from the one all the way. 00:02:28.000 --> 00:02:31.000 On the left is about. 00:02:31.000 --> 00:02:33.000 2 times the intact rock is 2 times larger than the. 00:02:33.000 --> 00:02:42.000 pre-existing faults. And all the way to the one on the right, where the faults are not offset. They just have a small gap between them. 00:02:42.000 --> 00:02:50.000 And we measure mechanical properties, voletric strain, and electrical conductivity. If you want to hear about the electrical conductivity and the vole metric strain. 00:02:50.000 --> 00:02:59.000 Ask me more questions, but to quickly smarize the mechanical properties. 00:02:59.000 --> 00:03:04.000 We see that they all follow a pretty nice elastic, loading curve. 00:03:04.000 --> 00:03:10.000 So here on, we have axial load on the y-axis and strain on the x-axis. 00:03:10.000 --> 00:03:15.000 And so they follow a nice elastic, loading curve, as we kind of expect in lab samples. 00:03:15.000 --> 00:03:20.000 Before reaching a critical point where the samples break. 00:03:20.000 --> 00:03:24.000 And we see that. The blue one. 00:03:24.000 --> 00:03:29.000 Shown here is the weakest, and that was our smallest offset. 00:03:29.000 --> 00:03:36.000 In comparison the. Largest offset was the strongest of the step over samples. 00:03:36.000 --> 00:03:41.000 And so this kind of shows that larger lab step overs can act as larger barriers. 00:03:41.000 --> 00:03:47.000 And then we also did some cool analysis of how much of this vole is really deforming and elastically. 00:03:47.000 --> 00:03:57.000 So feel free to ask me more questions. Thank you. 00:03:57.000 --> 00:04:18.000 right so i'm going to. Talk a little bit about rupture barriers, but in a different way. So last year i presented some work on doing a large set of nerical simulations to address whether, under what conditions, a zone of a seismic creep, can stop, rupture, and so 00:04:18.000 --> 00:04:38.000 You know it says i have twice as many simulations as last year, and so the motivation for this is just that, like many people have mentioned, creep is a major concern for earthquake hazard through northern california in particular, and yet there are very few events that we actually know about going through these things to analyze. And so that's where models can come in to help with this. 00:04:38.000 --> 00:04:50.000 so In these. I'm i'm testing a very simplified fault. Geometry and so i basically, it's just a planar strike. Slip fault with a central creeping patch of some variable length. 00:04:50.000 --> 00:05:01.000 And so this is a cartoon diagram of it, and within that creeping patch i'm varying its length. The amount of stress release due to creep the frictional properties of the creeping patch. 00:05:01.000 --> 00:05:21.000 Then also the overall stress amplitude on the whole system and the overall rupture velocity through the whole system. And the thing that's really new since last year is now i'm testing the effects of nucleation, site, or direct activity. So last year i was just nucleating at the far end of the fault. Maximize directivity, and the new set of models is. 00:05:21.000 --> 00:05:28.000 Nucleating right next to the creeping patch. And so these are fully dynamic, finite element simulations. I'm using fault, mud and its internal measure. 00:05:28.000 --> 00:05:46.000 And so last year i had over like i had about 1,500 simulations, and now i'm at over 3,000 and still counting. So this massive chart of things each different symbol represents one simulation, and the details represent the details of the rupture, propagation. 00:05:46.000 --> 00:05:57.000 But the really important thing is that blue is go, and red is stop, so anything that is shaded red is a set of conditions that prevented the rupture from reaching the other end of the fault. 00:05:57.000 --> 00:06:01.000 And if you'd like to see this more zoomed in just email me. 00:06:01.000 --> 00:06:19.000 So overall, i'm seeing basically the same things as last year, and even with twice as many models, which shows that this is robust. I'm seeing that really the the biggest control on whether rupture is able to keep going through other than the length of the creeping patch is how much stress is released due to creep. 00:06:19.000 --> 00:06:40.000 And actually the frictional difference really just controls how much slip is in that creeping zone. If rupture gets through or not. I also find that faster rupture velocities are able to promote throughgoing rupture, more than slower ones, regardless of the overall stress amplitude so low, stress versus high stress doesn't matter as much as high speed versus low speed. 00:06:40.000 --> 00:06:54.000 But the new results are related to that directivity. And so here, i'm zooming in on one of the the very center of my stress cases. I see with ruptures that start further away and have a larger directivity effect can get in. 00:06:54.000 --> 00:07:16.000 Further than one that start right next to it, although i do see some effect of my force nucleation zone. So i need to think for that later. And so the next steps are. Just. I might have to see what happens with even larger directivity to see how this scales, so maybe this will get to like 5,000, and then eventually, i'm going to be calculating passing probabilities based on these results. 00:07:16.000 --> 00:07:25.000 I do have to teach at one. So if you have questions, ask them quick, or ask them at 2 15. Thanks. 00:07:25.000 --> 00:07:38.000 Hi, everyone. I'm betsy madden. I'm at san jose state university, and one of the goals in our department at san jose state is to involve undergraduates in research, and we also train master students. 00:07:38.000 --> 00:07:44.000 So you'll see throughout the talk students working on different projects, highlighted at the bottom. 00:07:44.000 --> 00:07:59.000 There. I included this global strain rate model Because, of course, this is why we're all here these red and pink colors are really beautiful for valentine's day, but it's a lot harder to figure out how to live. 00:07:59.000 --> 00:08:16.000 on these high strain areas. And so we work basically toward 3 goals, one to better understand long-term fault bavior by modeling culative slip rates and offalt deformation with a focus on the bay area. 00:08:16.000 --> 00:08:24.000 My group also wants to understand short-term fault bavior by modeling kinematics and dynamics of individual earthquakes on bay area faults. 00:08:24.000 --> 00:08:32.000 And in the long term the the pie in the sky is to understand how these long-term baviors are related to these short-term. 00:08:32.000 --> 00:08:36.000 Baviors. So i'll talk through 00:08:36.000 --> 00:08:39.000 3 different projects that are working toward these goals. 00:08:39.000 --> 00:08:45.000 and you can see that the students involved in these projects are from our geology department, and. 00:08:45.000 --> 00:08:52.000 We also are inviting students from other departments to apply their skills in computer science and math. 00:08:52.000 --> 00:08:59.000 And data analytics to these geophysics, problems. 00:08:59.000 --> 00:09:03.000 Okay, so project one with a generous grant from the department of defense. 00:09:03.000 --> 00:09:11.000 kim blissniak and i were able to purchase. a drone, and 2 high resolution lidar systems. 00:09:11.000 --> 00:09:17.000 and what my students are using these for is to go out and image the landscape to better understand. 00:09:17.000 --> 00:09:27.000 Where faults are connected, and where they're segmented. And of course this is a very important question for seismic hazard. We've already seen it. 00:09:27.000 --> 00:09:31.000 In sar's talk as well as others throughout the workshop. 00:09:31.000 --> 00:09:39.000 And we've trained a lot of students to use this equipment through both class time as well as research projects. 00:09:39.000 --> 00:09:45.000 project 2. We are doing nerical modeling of long-term slip. 00:09:45.000 --> 00:09:51.000 taking both a forward modeling approach, as well as an inverse modeling approach. 00:09:51.000 --> 00:10:00.000 And the idea is to use these models to better understand how segmentation and connectivity, influence, long-term culative fault slip rates. 00:10:00.000 --> 00:10:09.000 and understand the role of all fault deformation in accoodating this strain here at this playboundary in the bay area. 00:10:09.000 --> 00:10:14.000 I mentioned that we're also modeling individual earthquakes, using dynamic rupture models. 00:10:14.000 --> 00:10:21.000 We have ongoing projects. Looking at the hayward, calaveris, roger's creek faults, as well as nascent project. 00:10:21.000 --> 00:10:28.000 modeling dynamic rupture along the conquered fault. And really curious about the role of creep. 00:10:28.000 --> 00:10:36.000 in accoodating again the total plate boundary strain and. 00:10:36.000 --> 00:10:39.000 In contrast, where things are taken up by earthquakes. 00:10:39.000 --> 00:10:43.000 Thanks a lot. 00:10:43.000 --> 00:11:01.000 Oh, hey, everyone, and thank you, betsy, for that great talk. So today i'm going to talk about my attempt to do some long-term earthquake simulations for the bay area region, and a lot of similar regions to those that betsy is also looking at. 00:11:01.000 --> 00:11:09.000 Ye So for this work, what i'm using is rsq SIM, which was developed by keith, richards, dinger, and jim dietrich. 00:11:09.000 --> 00:11:13.000 And the advantage of rsq SIM. Is. It relies on. 00:11:13.000 --> 00:11:21.000 it's a boundary, integral method, and it relies on this rate and state formulation of friction here. If we have. 00:11:21.000 --> 00:11:35.000 a minus b or b minus a being positive, then we have velocity weakening. This is what we have for our classical earthquake but we can also implement b minus a values that are negative, which would give us velocity, strengthening and creep. 00:11:35.000 --> 00:11:42.000 And, as we know, many of these faults in the bay area have large portions of them. Experiencing creep. 00:11:42.000 --> 00:12:01.000 Another big advantage of the methodology set up in rsq SIM is, we can generate many earthquakes Pretty computationally efficient, and that's because it relies on a three-state model for each patch within our framework. When it's at state 0. Nothing is slipping, and we can 00:12:01.000 --> 00:12:12.000 Know that the stresses are purely tectonic, so we can analytically solve for the time to failure. For each node, or each element in our model. 00:12:12.000 --> 00:12:16.000 And then have it nucleate from that. That element. 00:12:16.000 --> 00:12:21.000 And what i'm going to show you now is, if my slide's a little advance. 00:12:21.000 --> 00:12:38.000 There we go. Are some preliminary results. This is really early work. I did a lot of it actually last week. So there's there's more on top of this, but i would love some suggestions, and and any questions people have. But here are some example, earthquakes from a uniform model, where i keep. 00:12:38.000 --> 00:13:00.000 B, minus a uniform throughout the entire domain, and then a non-uniform model, where i allow that b, minus a value to vary based on studies that have shown this region to be creeping or not. And so here's 2 example. Earthquakes on the hayward fault in the uniform model. You can see. We get kind of this uniform slip distribution for this earthquake. 00:13:00.000 --> 00:13:06.000 Versus this non-uniform model. We can see that the the slip is heterogeneous. 00:13:06.000 --> 00:13:15.000 With regions kind of in the upper portion here, which have creeping, having a much lower stress drop and then a much lower split value. 00:13:15.000 --> 00:13:25.000 another goal of this project is to be able to use these to generate ground motions. So we can kind of get it some time. Dependent ground motions in the bay area. 00:13:25.000 --> 00:13:42.000 i please, ask questions and suggestions. Thank you. 00:13:42.000 --> 00:13:51.000 Well, i wouldn't on the schedule right now, but i'll take it away. So remember the pyrenew detacent. That was a topic. Yesterday morning. 00:13:51.000 --> 00:13:55.000 We're gonna go. Try and look for it in the next 3 min. So if you look on the bottom left. 00:13:55.000 --> 00:14:00.000 啊。We're going to use receiver function。I know soccerp。So this image is contrasts in rock fabric。 00:14:00.000 --> 00:14:04.000 We can get at the depth of the contrast, the fabric, orientation. 00:14:04.000 --> 00:14:11.000 In the fabric strength. So we're going to look for sheer fabric associated with the detacent. If you remember the 2 plots on the right. 00:14:11.000 --> 00:14:15.000 This thing is supposed to set at about 20 kilometer depth, so. 00:14:15.000 --> 00:14:22.000 On the left map. We calculate a receiver functions at these little blue stations, which are in seismicity. 00:14:22.000 --> 00:14:28.000 I'm showing the lfes from david's study that he talked about yesterday, and the little wedge is still over losty zone. 00:14:28.000 --> 00:14:34.000 I also marked the pioneer fragment and the direction that is traveling in underneath. 00:14:34.000 --> 00:14:41.000 North america. So let's look at 2 stations. Under the pony of fragra detacent. Well. 00:14:41.000 --> 00:14:47.000 On top of the detacent, and if you look at the waveforms, you see the green one is the standard receiver function. 00:14:47.000 --> 00:14:56.000 the one on the left in blue and red, is the 1st azimuthal harmonic, and that's where the sheer fabrics gonna show up and you see in both of them. 00:14:56.000 --> 00:15:01.000 You see a big bp close to 20 kilometers depth. 00:15:01.000 --> 00:15:06.000 Alright. So that should be. We're looking at a sheer contrast at about the right depths. Here. 00:15:06.000 --> 00:15:12.000 And if we look at orientations and at all stations now, so if you. 00:15:12.000 --> 00:15:20.000 First.st Look at the top, right? That's a little cartoon explaining the symbols. So the strike off the fabric is the red bar. 00:15:20.000 --> 00:15:30.000 And the size of that circle shows how strong the contrast is. Okay. So let's look at the middle map now that steps from 15 to 2020 kilometers. 00:15:30.000 --> 00:15:33.000 Each circle is a station, and you see that. 00:15:33.000 --> 00:15:38.000 There are contrasts at many of the stations in the area of the. 00:15:38.000 --> 00:15:48.000 Pioneer fragment. So in the northernmost outlined. Blue area. The the fabric shows strikes that about parallel to the. 00:15:48.000 --> 00:15:56.000 Motion, relative motion on a detacent. As he approached the plate boundary. You know the transition in the middle. 00:15:56.000 --> 00:16:00.000 Blue outline area. You see rotation of that fabric. 00:16:00.000 --> 00:16:04.000 And then, if you go further north. And cross the. 00:16:04.000 --> 00:16:08.000 You know, slab edge into the subduction area. Then you see. 00:16:08.000 --> 00:16:17.000 fabric that's parallel to the subduction. On the rightmost part. If you go below the depths of the proposed attacent, we don't see a whole lot of. 00:16:17.000 --> 00:16:27.000 Arrivals. So, in conclusion, it seems like. We do see some kind of signal from these parangut attacent, and. 00:16:27.000 --> 00:16:34.000 For the this is really early work. So you know, it would be really interesting to look at the expected deformation fabric. 00:16:34.000 --> 00:16:40.000 From modeling, perhaps, and compare that to our observations. So, if you have any questions, feel free to. 00:16:40.000 --> 00:16:54.000 Engage in the chat. Thanks. 00:16:54.000 --> 00:17:00.000 Hey? So in in this talk we're gonna have a goose at making really really simple models of earthquakes. So. 00:17:00.000 --> 00:17:11.000 Might simpler than what julian and betsy and evan were talking about, but the motivation in some ways is the same. It's can we get from observations to physical modelling to earthquake hazard. 00:17:11.000 --> 00:17:15.000 And the reason i want to think about physical models is because if you. 00:17:15.000 --> 00:17:21.000 Ask that question. You start. Try to think about a physics-based model physics based forecast. 00:17:21.000 --> 00:17:27.000 We know that lots of things on earth. Are important to whether or not an earthquake happens. 00:17:27.000 --> 00:17:33.000 We have lots of observations, so if we want to map between these 2, we need to know the physics make lots of observations. 00:17:33.000 --> 00:17:44.000 We need to figure out how to actually use this, and we need to make lots and lots of simulations, potentially millions or billions trillions of simulations. So let's make a really simple model of an earthquake. 00:17:44.000 --> 00:17:55.000 That we can run really fast. And so the simple model that i've chosen to use is Energy-based model. This will be familiar to a lot of people. It essentially says that. 00:17:55.000 --> 00:17:58.000 An earthquake is going to rupture in a stress field. 00:17:58.000 --> 00:18:09.000 That stress provides energy to the earthquake. So let's say we have an earthquake that starts in the red region. Here it starts spreading outward as long as there's enough stress to provide energy to rupture. The earthquake keeps going. 00:18:09.000 --> 00:18:14.000 If, however, there is not enough energy to rupture, the earthquake will stop. 00:18:14.000 --> 00:18:21.000 and tokyo were saying, you're on the bottom here. We're tracking how much energy is available to rupture. Initially it starts off. 00:18:21.000 --> 00:18:36.000 But eventually it stops. And so the nice thing about this simulation is, it's probably wrong. There's probably all sorts of caveats here, but it is extremely fast, and what that means is that we can run hundreds of thousands of simulations in an afternoon. 00:18:36.000 --> 00:18:39.000 And we can make a maps like this. So this is. 00:18:39.000 --> 00:18:45.000 For instance, if you put a particular mean stress, what is the likelihood of an earthquake at any different size. 00:18:45.000 --> 00:18:51.000 So we can map the physical parameters to lots of different outputs, including earthquake, size, distribution, stress drops. 00:18:51.000 --> 00:19:00.000 Directivity. Don't know whatever you like. We can mimic the simulation output with a neural network, and that lets us kind of probe the simulations. 00:19:00.000 --> 00:19:15.000 And So i think. Overall, this is relatively straightforward to do and it gives me a little bit more optimism that we have the potential to explore the hazard parameter space with simplified physical models. 00:19:15.000 --> 00:19:37.000 And also we might have the potential to test these physical models more robustly against lots of different collections of data sets. 00:19:37.000 --> 00:19:46.000 Alright question for the audience who's been asked whether the send andreas fault is overdue. You can put a hands up, or something like that. 00:19:46.000 --> 00:19:51.000 this is a bit of a perplexing question, right? Because. 00:19:51.000 --> 00:20:00.000 In the last 100 years we've had no earthquake. And yet the paleo-saismic record would suggest that there's an earthquake roughly every 30 years. 00:20:00.000 --> 00:20:04.000 So what's going on here? And just to really. 00:20:04.000 --> 00:20:10.000 Put a finer point on this whoop. To put a finer point on this issue. 00:20:10.000 --> 00:20:16.000 And kate shearer played this soup. Prospective game of asking if. 00:20:16.000 --> 00:20:24.000 We took a time machine went back a hundred years ago. What would be the probability of not having an earthquake in the next century, and they got a nber like. 00:20:24.000 --> 00:20:38.000 A point one to a 5% chance. So that's weird enough that we might want to start asking questions. Right is the data wrong. Maybe the paleo seismic data is misleading, or maybe the physics are wrong. 00:20:38.000 --> 00:20:45.000 Maybe our recurrence models are wrong. And sar minson is mentioning in the in the chat clarence law, that. 00:20:45.000 --> 00:20:50.000 Just. This seems to be a pretty universal bavior. 00:20:50.000 --> 00:20:57.000 So i'm gonna suggest option c. That we should expect to be in a boring time to study earthquakes. 00:20:57.000 --> 00:21:02.000 And i'm going to be leaning on the wait time, paradox, and this is a fairly well-known. 00:21:02.000 --> 00:21:10.000 statistical quirk of. Temporal point processes, which says that if a random observer is. 00:21:10.000 --> 00:21:17.000 Parachuted into a timeline. They'll tend to land in these longer intervals, rather than short intervals. 00:21:17.000 --> 00:21:25.000 And we can play this exact game with paleo seismic data. So we parachute randomly into the timeline, and lo and bold. 00:21:25.000 --> 00:21:31.000 的random。 00:21:31.000 --> 00:21:36.000 Time traveling parachutor. What experience is longer than the average recurrence interval, and that. 00:21:36.000 --> 00:21:42.000 Biases more severe if the coefficient of variations of the variability and the timing. 00:21:42.000 --> 00:21:48.000 Of earthquakes is larger. Alright. So we can. 00:21:48.000 --> 00:21:59.000 Account for this funny paradox and revisit those probabilities, and we end up with something much more tenable, something like a 50, 15% chance of a hundred year hiatus. 00:21:59.000 --> 00:22:07.000 On the san andreas fault system. Alright. This might seem like a simple reframing. 00:22:07.000 --> 00:22:14.000 But there's some really interesting implications for how that can arise from the wait. Time, paradox. 00:22:14.000 --> 00:22:24.000 So let's just consider one example. If there's any way in which the magnitude of the next earthquake depends on the acculated slip deficit. 00:22:24.000 --> 00:22:34.000 Then the implication that. We typically are in longer intervals is that the next earthquake would be bigger. And this is not just an and. 00:22:34.000 --> 00:22:38.000 An implication for the sun address fault. It's an implication, worldwide. 00:22:38.000 --> 00:22:47.000 So hopefully in these. In this brief 3 min. I've given you an intuition for this funny paradox, how it might resolve. 00:22:47.000 --> 00:22:56.000 Long standing debate about the san andreas fault. And how it has. It has the potential for a really strange implications, worldwide. 00:22:56.000 --> 00:22:59.000 Thank you. 00:22:59.000 --> 00:23:17.000 Hi, everyone, i'm bing how today i'm gonna talk about the scaling relationship. 00:23:17.000 --> 00:23:19.000 也不一定。Earthquakes are。 00:23:19.000 --> 00:23:36.000 As we understand. as the simplest groups of earthquakes. They are essential for testing our current knowledge of earthquake physics, especially on how earthquakes reoccur at the same place. 00:23:36.000 --> 00:23:41.000 Moreover, this repeating earthquakes are often practically used as quick meters. 00:23:41.000 --> 00:23:46.000 Adapt. Have. It depends on. 00:23:46.000 --> 00:23:54.000 Or the scalding relationships, especially to observable recurrence times, and their seismic moments. 00:23:54.000 --> 00:24:00.000 However, there has been. A decade long jp on this very important scaling relationship. 00:24:00.000 --> 00:24:06.000 On the one hand, the classic crack model with constant stress job and no a side to slip. 00:24:06.000 --> 00:24:12.000 Predicts that the recurrence time should be proportional to the moment, to the power of 1 3. 00:24:12.000 --> 00:24:17.000 On the other hand, the observed scaling in california. 00:24:17.000 --> 00:24:25.000 30 years ago, suggests a different scaling. Our relationship. That is 1, 6. 00:24:25.000 --> 00:24:34.000 This 1, 6, relation. Is believed to be consistent with the recent global completion. Down in 2,019. 00:24:34.000 --> 00:24:48.000 Here. We would like to review this scaling. Problem by including the latest global, repeating earthquake catalogs. Not really. There's a very nice, high quality, repeating earthquake catalog in northern california, constructed by. 00:24:48.000 --> 00:24:52.000 Woodhouse and sheffield in 21. These are the blue dots here. 00:24:52.000 --> 00:25:01.000 as we can see from this initial global compilation Now the recurrence time in that moment. 00:25:01.000 --> 00:25:06.000 Forms a highly scattered and probably also incomplete data cloud. 00:25:06.000 --> 00:25:12.000 That prevents any meaningful global benefit. So we really need to. 00:25:12.000 --> 00:25:18.000 Understand this origin of such scattered, and try to reduce the scatter in order to get a meaningful. 00:25:18.000 --> 00:25:29.000 Here's what we did. We simply select those. Simplest groups of repeating earthquakes, those that occur throughout the observation, time, period. 00:25:29.000 --> 00:25:34.000 And require them to also. Regularly in time. 00:25:34.000 --> 00:25:40.000 This selection reduces a large scale amount of a scatter in the data. Cloud. 00:25:40.000 --> 00:25:47.000 And the remaining data cloud reviews a triangular. Data 00:25:47.000 --> 00:25:52.000 Cloud, with clear, lower bound that features, a slope of 1, 3. 00:25:52.000 --> 00:25:57.000 As shown by the red line consistent with the classic crack model. 00:25:57.000 --> 00:26:07.000 Moreover, they are intercept. Corresponds to an apparent stress job. About 10 megapascals, consistent with independent estimates from seismology. 00:26:07.000 --> 00:26:16.000 In short, we argue that. Global repeating earthquakes may review a recurrence time in size, movement, relation. 00:26:16.000 --> 00:26:20.000 That it can be consistent with classic crack model in a complex for zone. 00:26:20.000 --> 00:26:38.000 Please see our prepaint for more information, and welcome any questions. Thank you. 00:26:38.000 --> 00:26:49.000 Good morning, everybody, or good afternoon. So this is Talk on calendar time, local or quick forecast. Say, do it yourself, method. You can do this on your laptop takes about. 00:26:49.000 --> 00:26:53.000 2 or 3 min to create a forecast, so you can create the movies, and i'm going to show you. 00:26:53.000 --> 00:27:01.000 In about 10 min on a laptop. okay, so let's see if i can advance the slide. Where do i advance it? 00:27:01.000 --> 00:27:08.000 Work. Press the button on the screen. Ye okay. 00:27:08.000 --> 00:27:15.000 papers have been published. The one on the left is on calendar time, forecasting our local natural time forecasting, which is. 00:27:15.000 --> 00:27:21.000 Events, so counts of small events between earthquakes. One on the right is their calendar time forecasting. 00:27:21.000 --> 00:27:26.000 Alright. So then the next slide shows you the the steps in this method you start off with. 00:27:26.000 --> 00:27:33.000 basically, you create an ensemble of states. We're looking at the 100 25 kilometer circle around los angeles. 00:27:33.000 --> 00:27:40.000 And we create an expanding rectangles of catalogs. And those are converted to time series of. 00:27:40.000 --> 00:27:48.000 Cycles of a magnitude. 6 earthquakes are larger. Using a what we call a natural time, or a natural time, or an outcast transform. 00:27:48.000 --> 00:27:55.000 So you get this ensemble of time series. And then what you can do is Compute the roc diagrams for this. 00:27:55.000 --> 00:27:59.000 this is the one in the upper left is after. 00:27:59.000 --> 00:28:02.000 No event. So just after the last 96th earthquake. 00:28:02.000 --> 00:28:07.000 The one on the upper right is after 150 small earthquakes in the ensemble. 00:28:07.000 --> 00:28:11.000 After 300 and after 4, 47, the red curve is. 00:28:11.000 --> 00:28:15.000 The mean of all the ensemble. Ro c curves. 00:28:15.000 --> 00:28:19.000 And then from directly from the ro. C diagram, you can compute the. 00:28:19.000 --> 00:28:27.000 forecast itself? Are the ones on the upper are for one year, forecasts or excuse. Ye One year forecasts. 00:28:27.000 --> 00:28:32.000 12 month forecasts, and the main effect is, you see, main shot clustering is the major effect. 00:28:32.000 --> 00:28:37.000 Again. The red line is the. Ensemble average of all the cyan curves. 00:28:37.000 --> 00:28:42.000 below, you see below that you see the 00:28:42.000 --> 00:28:47.000 5, 5 year forecast, and what you see there is the tectonic reloading effect. 00:28:47.000 --> 00:28:53.000 Okay, so the ones the ones on the left are for an ensemble of 30 events. The ones on the right are an ensemble of 60 events. 00:28:53.000 --> 00:29:02.000 So with that you can actually. Create a movie which i don't know if i will, it will play or not. Probably i don't have time to play it. 00:29:02.000 --> 00:29:09.000 Oh, there it goes. Okay. And so i don't really want to wait. 00:29:09.000 --> 00:29:12.000 Get to go, so i'm going to go to the next slide. 00:29:12.000 --> 00:29:28.000 Which is that and that, and that is You can do an exceedance probability for magnitudes. Okay? So The blue curve is after no events of no small events have occurred. The. 00:29:28.000 --> 00:29:36.000 red curve is the middle curve there the magenta curve is after current events since parkville, since northridge. 00:29:36.000 --> 00:29:40.000 And the one of the lower bottom, the red one is projected. 00:29:40.000 --> 00:29:47.000 Which this would be the mean expected for the next mean interval expected for the next magnitude. 6 earthquake. 00:29:47.000 --> 00:29:49.000 And you can see there are the results so. 00:29:49.000 --> 00:29:55.000 That's what i have. 00:29:55.000 --> 00:30:25.000 Hey? Hello, everyone. My name is brandon schuck. I'm an assistant professor at louisiana, state university. Happy early mardi gras to everyone who celebrates. If you have no idea who i am, i'm relatively new to this counity coming from more of a subduction zone Press inside of things. So ye for my postdoc. I worked in northern cascadia on explorer plate and that complex triple junction, and fortunate for the last few years, been working in southern cascadia and around the mendocino triple junction area 00:30:27.000 --> 00:30:49.000 and was fortunate to be funded by a knee herb grant to work in this area. So the goal of our project is to reprocess legacy, seismic data that was collected in the menacino triple junction area. I worked on casey 21 data further north, which is a deep, penetrating, active source. Seismic experiment along most of the cascade subduction, zone. 00:30:49.000 --> 00:31:04.000 It's given us a lot of incredible new insights, but unfortunately they couldn't collect data for theirself to complete the full subduction zone so the only crustace scale experiment is in the bottom left figure. That's shown by the green lines there, and the 19. 00:31:04.000 --> 00:31:34.000 90 s. On the mend the signal triple junction seismic experiment. So we're reprocessing this data to look at Tectonics in the region, and the geometry of the subducting gourda plate Okay, there, my slides. Go here's an example. The top is the southernmost. Casey, 21 profile offshore, southern oregon, on the bottom is the northernmost mendocino triple junction experiment offshore crescent city. That's published in the literature. 00:31:35.000 --> 00:31:45.000 And we'll see if this works on my slides here. It should go. 00:31:45.000 --> 00:32:00.000 Okay? Oh, there it goes. Okay, apologies, powerpoint. Animation is not working as intended. Essentially you're looking at the one of our reprocess images here, offshore crescent city. 00:32:00.000 --> 00:32:08.000 And showing that we can nicely track the geometry of the top of the subducting plate as well as the gorda mojo. 00:32:08.000 --> 00:32:21.000 Cordis lab mojo, and in terms of the overriding plate there's a lot of interesting variability in display faults. And this is an area where recent coring has shown that some of these blade faults may be cosismically active. 00:32:21.000 --> 00:32:34.000 During mega thrust earthquakes. I'll continue on here. This is an example of a deploying for their self, not the best image quality, as you can see, but we can track the top of the plate all the way to the shelf. 00:32:34.000 --> 00:32:47.000 And there's some interesting variations here within the forarch, at the transition between a young accretionary prism sediments and the older metamorphic basement, where we have fluid, expulsion, type, features, coming out. 00:32:47.000 --> 00:33:07.000 we'll skip ead for one of the most relevant pieces. This is an image here that is in the overriding plate, and it cross. We cross all the faults from the menacino triple junction up to big lagoon. And i'm highlighting some of those by the red arrows and the black arrows are pointing to the gourda slab geometry and. 00:33:07.000 --> 00:33:12.000 We have a map in the top right of 2 way. Travel time of the gourdish lab. Geometry. 00:33:12.000 --> 00:33:23.000 One of the biggest challenges right now is converting that to depth. Anyways, one of the big takeaways is that we're seeing changes in the plate. Geometry across these strikes lit faults, so there might be some connection between. 00:33:23.000 --> 00:33:29.000 The overriding plate falls in the mega, thrust geometry and still working on this, and we'll. 00:33:29.000 --> 00:33:40.000 ye Contribute this lab model for use and hazard estimates moving forward. Thanks. 00:33:40.000 --> 00:33:46.000 Ok. Hi, i'm clarion, my usgs colleagues, and i with support from pg. And e. 00:33:46.000 --> 00:33:52.000 Relocated the past 12 years of earthquakes near lake alminore, in northern california. 00:33:52.000 --> 00:34:01.000 And we found a fault structure at depth that is connecting 2 known mapped surface faults. 00:34:01.000 --> 00:34:11.000 Okay. So we're looking at this part of northeastern california, which is sparsely covered by the permanent seismic network, but it is an active area. 00:34:11.000 --> 00:34:24.000 And so in blue is like almondore. In 2023 there was a magnitude 5.5 earthquake, followed by a 5.2 a few hours later, and it's aftershocks are shown as purple dots. 00:34:24.000 --> 00:34:31.000 And then, 10 years before that. In 2013 there was a magnitude 5.7 earthquake just to the southwest. 00:34:31.000 --> 00:34:39.000 It's aftershots are these yellow and green dots. And this is concerning because pg. And e has 6 dams. 00:34:39.000 --> 00:34:44.000 And other critical facilities. These black diamonds near these earthquakes. 00:34:44.000 --> 00:34:49.000 And so our goal here was to really resolve the fault structure hosting these earthquakes at depth. 00:34:49.000 --> 00:34:58.000 To ultimately inform site, specific earthquake, hazard. 00:34:58.000 --> 00:35:08.000 So what we did was, we took all the permanent and seismic data around, like almanour, and then combined that with a usgs led and. 00:35:08.000 --> 00:35:12.000 Nodal deployment for 2 months. Following the 2023 earthquake. 00:35:12.000 --> 00:35:22.000 There's excellent nearby coverage as shown by the blue triangles and a double difference, a relocation of earthquakes from the past 12 years. 00:35:22.000 --> 00:35:34.000 Which spanned both of the sequences. And our goal was to resolve the fault, geometry, dimensions, and connections at depth. 00:35:34.000 --> 00:35:38.000 So here are the the results of our relocated earthquakes. 00:35:38.000 --> 00:35:44.000 Here in matview, on the top left, and then the depth cross sections at the bottom. 00:35:44.000 --> 00:35:50.000 What we found was that the 2013, the yellow and the 2023. The purple sequences. 00:35:50.000 --> 00:35:59.000 They occurred very close to each other, but they don't overlap much which suggests some kind of stress transfer to the in the northwest direction. 00:35:59.000 --> 00:36:13.000 We also see that both sequences occurred on the same larger, underlying fault structure. The strike is about 290 degrees, which is consistent with one of the nodal planes of the magnitude. 5 plus earthquake. A moment, tensors. 00:36:13.000 --> 00:36:21.000 The dip is about 70 degrees, and based on the dimensions of this fault. It looks like it's big enough for a magnitude. 6. 00:36:21.000 --> 00:36:33.000 And then stepping back and looking at the big picture. It looks like these sequences fill in a gap between 2 larger known map surface faults the aminore fault to the north skinner flat fault to the south. 00:36:33.000 --> 00:36:39.000 If all 3 faults were to ruptured together and one earthquake, it could be big enough for magnitude. 7. 00:36:39.000 --> 00:36:52.000 So that's it can see me. Ssa, for more info. Thank you. 00:36:52.000 --> 00:37:05.000 So hi, my name is vicki, and the divine misera asked me to talk about the setting of the magnitude 4.7 and 4.9 earthquakes that rbled to northeast california the last week of 2025. 00:37:05.000 --> 00:37:10.000 About 15 kilometers north of the town of susanville, and about 50 kilometers from where. 00:37:10.000 --> 00:37:16.000 Claire was just talking. These earthquakes warranted a mention in the san francisco clinical, which stated quote the. 00:37:16.000 --> 00:37:22.000 Area as well to the east of major california fault lines. But there are some faults such as the hat creek in the vicinity. 00:37:22.000 --> 00:37:28.000 The map on the left shows the regional setting of the earthquakes located east of lacent peak volcano. 00:37:28.000 --> 00:37:37.000 At the intersection of the walker lane, that belt of right lateral shear east of the sierra nevada, and the extending basement and range province. 00:37:37.000 --> 00:37:43.000 The map on the right is a closer look at the usgs. Recent earthquakes website with highlighted the. 00:37:43.000 --> 00:37:50.000 Magenta star, the location of the susenthal earthquakes and red lines being quaternary faults. Note that the website only shows. 00:37:50.000 --> 00:37:54.000 Those faults active in the last 130,000 years from the us. 00:37:54.000 --> 00:37:58.000 put your default database. I've labeled the named faults the honey lake. 00:37:58.000 --> 00:38:04.000 The hat creek, and the likely faults and the earthquakes are not obviously tied to any of these. 00:38:04.000 --> 00:38:10.000 There are 2 short earth, faults map near the eastern shore of eagle lake. 00:38:10.000 --> 00:38:14.000 Could one of those be the faults responsible? Or are there other candidates. 00:38:14.000 --> 00:38:17.000 但aren't shown on the。 00:38:17.000 --> 00:38:24.000 Website. So this is a geologic map of the susanville region, with the quaternary faults from the earthquake web page sun in red. 00:38:24.000 --> 00:38:34.000 And i've added the rest of the quaternary faults from the national database shown in black, and it's easy to see that the susanville earthquakes, likely ruptured in older quaternary fault. 00:38:34.000 --> 00:38:41.000 As the micro seismicity forms a northwest, elongated cluster about one kilometer northeast of an older quaternary fault. 00:38:41.000 --> 00:38:48.000 Consistent with the northwest, striking steeply northeast, dipping vocal mechanisms of the 4.7 and 4.9. 00:38:48.000 --> 00:38:55.000 Events. And so these earthquakes appeared to have ruptured along a northwest, striking fault that extends 20 kilometers to the southwest. 00:38:55.000 --> 00:39:01.000 Towards susanville. So if we overlay gravity, contrast shown by the blue lines. 00:39:01.000 --> 00:39:10.000 It indicates a low on the northeast side of the fault, where the earthquakes occurred, supporting a longer term down to the north vertical component of the slip. 00:39:10.000 --> 00:39:15.000 Along the southeastern part of the fault. There's another gravity low, this time showing down to the southwest. 00:39:15.000 --> 00:39:24.000 the honey lake fault strands marked the southwest margin of this gravity low, so this likely represents a right step, forming a step over basin. 00:39:24.000 --> 00:39:28.000 Beneath the town of susanville, and this argues that the 2,025 events. 00:39:28.000 --> 00:39:35.000 Are tied to the honey lake fault rather than the hat creek fault, so given that san francisco chronicle article. 00:39:35.000 --> 00:39:39.000 And the public will be looking at our recent earthquake's web page. 00:39:39.000 --> 00:39:44.000 Should we include the ability to show older cue faults. 00:39:44.000 --> 00:39:49.000 in the usgs, recent earthquakes, web page, especially for regions such as this. 00:39:49.000 --> 00:39:56.000 That have distributed vaulting and low slip rates, and note that this fault, that ruptured is actually shown as a proxy fault. 00:39:56.000 --> 00:40:10.000 In the 2,023 national seismic hazard model. So it's there, but we're not showing it to the public, so thank you for your time and attention. 00:40:10.000 --> 00:40:16.000 Thanks, mickey. I'm jared. I'm. Vicky supervisor. I guess. 00:40:16.000 --> 00:40:28.000 Working usgs and. In the thunder talk. I just wanted to dice a different way of looking about. How faults form and reorganize in relation to lithispheric structures. 00:40:28.000 --> 00:40:33.000 Looking at specific examples in northern california and the clear lake region. 00:40:33.000 --> 00:40:43.000 For context. Along the california coast is a series of long-lived volcanic centers that are assed to be related to slabwind of vulcanism and wake of northward migration of the mendocino, of triple junction. 00:40:43.000 --> 00:40:50.000 Of which the clearly volcanic. Field is the youngest in northernmost of these volcanic centers. 00:40:50.000 --> 00:40:54.000 These long-lived, volcanic fields occurred inboard of the san andreas fault. 00:40:54.000 --> 00:41:00.000 Namely, the plate, boundary and false, tend to reorganize around, and later. 00:41:00.000 --> 00:41:06.000 Through the fields. The clay lake volcanic field is bound to the west by the maxima fall, and to the. 00:41:06.000 --> 00:41:17.000 East by the bartlett springs, fall. Both accoodating right lateral motion, whilst creating extensional regime between the 2 supporting magma sent through dilation structures. 00:41:17.000 --> 00:41:25.000 I'm macmost is the presed plate boundary, and there is a bulge observed, and both faults around the clear lake field. 00:41:25.000 --> 00:41:34.000 Suggesting reaction to listed ferric structures. Next slide. Please. 00:41:34.000 --> 00:41:38.000 Now, here's a postulation about how this is lisa. Spheric structures came to be. 00:41:38.000 --> 00:41:43.000 So. Subduction of the gourdish lab includes complicated geometry of ridge segments. 00:41:43.000 --> 00:41:49.000 When each ridge bridge segment is subducted, it basically decouples into 3 separate pieces. The leading plate. 00:41:49.000 --> 00:41:55.000 The mid-ocean ridge, and. Which we call the mid ocean ridge, magma, supplier, or the morms. 00:41:55.000 --> 00:41:59.000 And the trailing plate. So the leading plate gravitationally sinks in. 00:41:59.000 --> 00:42:06.000 Sinks and flounders into the mantle, eventually creating a slabless window, initially promoting mantle, upwelling. 00:42:06.000 --> 00:42:11.000 The morms cools as it moves away from the sea source and subicreats to the north american plate. 00:42:11.000 --> 00:42:18.000 Where it becomes a competent listsospheric structure. Then the trailing plate now becomes the plate boundary. 00:42:18.000 --> 00:42:23.000 And transforms with the pacific plate northwards, and over time reorganizes with the san andreas. 00:42:23.000 --> 00:42:30.000 So the sub-accreated morms now enforces bottom-up control on volcanic field, and where faults occur in the crust. 00:42:30.000 --> 00:42:35.000 The markham of fault follows the edge. Western edge, and the bartless springs follows the eastern edge. 00:42:35.000 --> 00:42:42.000 Both being bulged by the plutonic rocks in place during melting events influenced by the morms. 00:42:42.000 --> 00:42:44.000 So next time you want to see the whole picture. 00:42:44.000 --> 00:42:47.000 Don't just look down, but look up as well. 00:42:47.000 --> 00:42:56.000 That's it. Thanks. 00:42:56.000 --> 00:43:02.000 Hi, i'm takasha taira from ucry. We are very happy to report to the progress. 00:43:02.000 --> 00:43:08.000 Of our nodal seismic experiment in the little more body. 00:43:08.000 --> 00:43:14.000 supported by the usgs, new hub project. 00:43:14.000 --> 00:43:26.000 Are to understanding to our seismic hazard. In this area are we are put up to 200 side. We have a two-hand side, where we put a 2 to 3 component, and notice. 00:43:26.000 --> 00:43:32.000 Seismometers for 2 months, starting to our october 2,002, 5. 00:43:32.000 --> 00:43:45.000 我们的counity people thank you so much again for for helping us to this experiment。 00:43:45.000 --> 00:43:55.000 So this is a map view over the our seismic experiment. We have our 2 r. 1 d. Linear ra. To doing for the 2D. Nba nodes tomography. 00:43:55.000 --> 00:44:01.000 Also we put to the widely distributed sensor to increase the laypas coverage. 00:44:01.000 --> 00:44:16.000 And also you can see that in the light inside, figure out art erosion. Thanks to him, he got the approved to the sensor. In the in on the campus, in the liverpool national, rabb and sunday. National level. So we have identified in this area. 00:44:16.000 --> 00:44:34.000 And then the finally, we have some discussion this morning, after we put to the our nodalex nodal sensor. 2 weeks later son was happening. So we put to the 10 additional sensor dedicating this area, which is showing to the orange cycle. Here. 00:44:34.000 --> 00:44:40.000 So we are still in shadow stage to looking on. Data. Quality are left the plot, showing you to. 00:44:40.000 --> 00:44:49.000 Particip. So you can see nice one microphasma signal to one to 10 second, which is really into it unbinders tomography. 00:44:49.000 --> 00:44:54.000 We also get to the telescopenic event signal for potentially using the receiver function. 00:44:54.000 --> 00:45:00.000 Of course we have a nice recording from the local earthquake, from salmon sequences. 00:45:00.000 --> 00:45:05.000 So our so far the data, our quality is very excellent. 00:45:05.000 --> 00:45:23.000 ph student in the university of ut walking on the noise, close correlation, and here is the one example we are very happy to see the signal from ambient noises nicely leakable for in particular green function. So what we are expecting to tomography for this data. 00:45:23.000 --> 00:45:31.000 Finally for touch the san ramon area. So here again we put it to the 10 station in the dedicating to san la monstrom area. 00:45:31.000 --> 00:45:35.000 We put it to the sensor here for end of the december. 00:45:35.000 --> 00:45:40.000 啊。Whatever reason，after we put the sensor，the swarm activity is a little bit less。 00:45:40.000 --> 00:45:47.000 So we don't really have captured the prime time. But hopefully, we are using those data to get to the more smarts. 00:45:47.000 --> 00:46:01.000 When you're locating our hopefully mechanism. 2. And of course, this data collecting is available to the all over you in this counity. So we are working together to distributing the data as as soon as possible. 00:46:01.000 --> 00:46:15.000 That's all i have. Thank you so much. 00:46:15.000 --> 00:46:21.000 Great great So when we think about fault, complexity typically that we think about the surface traces. 00:46:21.000 --> 00:46:27.000 But how that complexity that we see at the surface extends to depth is something that is still debated. 00:46:27.000 --> 00:46:32.000 Whether we have this tabular surface, or whether all it's actually converged in a flower type structure. 00:46:32.000 --> 00:46:44.000 Is not well resolved. So in this talk we're going to be going beyond 2D. And going into full 3D. 00:46:44.000 --> 00:46:50.000 So when we take a map view of of faults, it can be quite appealing to. 00:46:50.000 --> 00:47:00.000 Represent these as a single modified. scalar nber, which shauna, chu and others did a few years back. And what this is is. 00:47:00.000 --> 00:47:08.000 taking the projected lengths of faults within a specific area, and projecting them onto a rotating line that rotates in space. 00:47:08.000 --> 00:47:19.000 So. The important thing here is that. This value is low when faults are parallel, and and quite quite well baved, and it goes to one as we go into randomness. 00:47:19.000 --> 00:47:31.000 Now, why do we actually care about this. Well, this misalignment measurement has been shown to correlate with high frequency ground motions, and particularly the event term telling us that this is actually telling us something about. 00:47:31.000 --> 00:47:37.000 Complexity at depth. We also see that misalignment seems to correlate with the creep rate. 00:47:37.000 --> 00:47:47.000 Where low misalignment leads to locking bavior. And then, finally, aftershock productivity seems to be well correlated with this misalignment measurement. 00:47:47.000 --> 00:47:52.000 Which makes sense because we have an available area. Our region of. 00:47:52.000 --> 00:48:00.000 Of smaller faults at various orientations. So all of this makes sense. But this is these are measurements that are occurring at depth. 00:48:00.000 --> 00:48:06.000 A correlating with something that's at the surface. So this is kind of suggesting that we have a tabular. 00:48:06.000 --> 00:48:14.000 System, and this stands in direct contrast with the localization of a fault, zone, and a flower structure, as people coonly attribute it to. 00:48:14.000 --> 00:48:26.000 So we're trying to go beyond this and measure fault, complexity, and 3D. Using our recently published surf algorit Which measures 3 default orientations, using seismicity patterns. 00:48:26.000 --> 00:48:33.000 and so we determine orientations, using magnetic patterns and weight measurements by their seismicity, density. 00:48:33.000 --> 00:48:41.000 This value will be low when faults are parallel and well organized, and again be high. When we have random fault or orientations. 00:48:41.000 --> 00:48:44.000 So i'm going to skip directly to the results. 00:48:44.000 --> 00:48:54.000 On the left-hand side. You see the traditional Measurement based on on fault traces and on the right are new size, messedie-based measurement. 00:48:54.000 --> 00:49:01.000 And what stands out to me is that 1st order things in the eastern california shear zone and the walker lane. 00:49:01.000 --> 00:49:08.000 Aligned quite well, but that we see actually. A significant difference in the coast ranges. 00:49:08.000 --> 00:49:18.000 Where we see more complexity at depth than we do within the surface traces, suggesting that there may be more complexity at depth. 00:49:18.000 --> 00:49:32.000 So this is really new for us just getting started here. There's a lot more to do. I'm gonna leave it there. 00:49:32.000 --> 00:49:35.000 Hi, everyone. My name is well kansiv. Again. I'm with tblers. 00:49:35.000 --> 00:49:43.000 I'm gonna show you let's move. Zoom out from bay area and look at the western us as a whole, and. 00:49:43.000 --> 00:49:47.000 What can we say about the frequency of large earthquakes? 00:49:47.000 --> 00:49:53.000 In the region. I'm going to show you 3 data sets. One of them is the. 00:49:53.000 --> 00:49:56.000 Observation from 19. 00:49:56.000 --> 00:50:05.000 5, and we only used. Turkicks about the magnitude of cognitiveness, and what you are seen in the upper right panel. 00:50:05.000 --> 00:50:12.000 The magnitude of completeness. Drops with time, as expected. 00:50:12.000 --> 00:50:17.000 So let's plot the. Earthquakes, since 1,905. 00:50:17.000 --> 00:50:21.000 And a standard magnitude, frequency, plot. 00:50:21.000 --> 00:50:23.000 我们是。 00:50:23.000 --> 00:50:31.000 It's a very nice b-value equals one. Kind of plot, except 1,906 earthquake. 00:50:31.000 --> 00:50:35.000 At tamer. We also build a global stochastic model. 00:50:35.000 --> 00:50:43.000 And if you compare the observed. Earthquake occurrences with our model. 00:50:43.000 --> 00:50:51.000 What you are seeing is it fits the data very well, especially between 5 point magnitude 5.7. 00:50:51.000 --> 00:51:02.000 And 7.7. Uss also has a model 2,020 stochastic event set from the 2,023. 00:51:02.000 --> 00:51:08.000 we are stress model. It seems to be a. 00:51:08.000 --> 00:51:10.000 Overestimating. 00:51:10.000 --> 00:51:13.000 的，对的。 00:51:13.000 --> 00:51:17.000 And with that b-value. Point 8. That seems. 00:51:17.000 --> 00:51:24.000 Rather low for the area. this difference. 00:51:24.000 --> 00:51:28.000 Can have a huge impact, especially for the large earthquakes. 00:51:28.000 --> 00:51:34.000 Here are some examples, for example, the magnitude 7.8 in larger earthquakes. 00:51:34.000 --> 00:51:39.000 5 times larger than the. tandler model in the usgs. 00:51:39.000 --> 00:51:45.000 说，these big differences，large damage in our cake could have a huge impact。 00:51:45.000 --> 00:51:48.000 The thought that i want to leave you with is. 00:51:48.000 --> 00:51:53.000 We have a globally consistent. Independently tested model. 00:51:53.000 --> 00:51:59.000 That provides an alternative view in california, and seems to have a different view. 00:51:59.000 --> 00:52:05.000 And we need to understand. How? What's the reason of these differences? 00:52:05.000 --> 00:52:16.000 Thank you. 00:52:16.000 --> 00:52:20.000 Okay, thanks to all of our thunder, talk speakers in this session. 00:52:20.000 --> 00:52:27.000 We now have about 6 min for a. Live. Q. And a. So if you have questions for any of our speakers. 00:52:27.000 --> 00:52:33.000 Please raise your hand. and. 00:52:33.000 --> 00:52:40.000 I might put travis on the spot if he doesn't mind. I know he's in the 70 room. 00:52:40.000 --> 00:52:45.000 do you have plans, travis, to see if your. 00:52:45.000 --> 00:52:48.000 Metrics correlate with. 00:52:48.000 --> 00:52:54.000 Grand motion in the same way that some past complexity matrix have. 00:52:54.000 --> 00:52:55.000 Sorry. Can you repeat the question? I was walking to the phone. 00:52:55.000 --> 00:53:03.000 Sure i know you're getting situated there. Do you have plans to look at correlations with ground motions between your complexity metric. 00:53:03.000 --> 00:53:10.000 Metrics. The way. That's been done for past. 00:53:10.000 --> 00:53:11.000 I i haven't thought about it, but that would be something that. 00:53:11.000 --> 00:53:15.000 Metrics. 00:53:15.000 --> 00:53:25.000 Should naturally be done, and i'd. Love to work with. Someone that does know things about about ground motions and see if we can find some some true correlations. There. 00:53:25.000 --> 00:53:34.000 Thanks. And ross, do you have a question. 00:53:34.000 --> 00:53:35.000 Okay? 00:53:35.000 --> 00:53:40.000 Ye i thought killian's talk was fascinating, and morgan there was some exchange with morgan saying. 00:53:40.000 --> 00:53:44.000 That ned fields work also shows a larger nber. 00:53:44.000 --> 00:53:53.000 On the one hand, the nber is slightly larger than what glenn, biazi and kate chair found. On the other hand. These are still really small nbers. 00:53:53.000 --> 00:53:59.000 So i hope you're also considering the alternative. In some sense the paleo seismicity. 00:53:59.000 --> 00:54:11.000 Work isn't really calibrated to earthquake magnitudes. So there, you know, there's 2 ways to look at this, and i hope you're looking at both. 00:54:11.000 --> 00:54:17.000 Absolutely. I think. Within the recurrence models. 00:54:17.000 --> 00:54:21.000 There's variability, and one view on the variability is that. 00:54:21.000 --> 00:54:25.000 It's random, and we don't know about it. And another view is that. 00:54:25.000 --> 00:54:31.000 Their physical controls on it, and i think. I the conclusion of what i'm presenting. 00:54:31.000 --> 00:54:37.000 Is not. That that variability is not interesting, and that we shouldn't try to understand. 00:54:37.000 --> 00:54:40.000 In fact, i think it's very interesting. So maybe that's a. 00:54:40.000 --> 00:54:42.000 Thank you. 00:54:42.000 --> 00:54:46.000 Answered your question. 00:54:46.000 --> 00:54:51.000 Thanks, and it looks like le has a question for vulcan. 00:54:51.000 --> 00:54:52.000 Le do you want to ask your question, live? Or i can also. 00:54:52.000 --> 00:54:57.000 Ye 00:54:57.000 --> 00:55:07.000 Read it. If we can get them pulled up. I know john's working hard. 00:55:07.000 --> 00:55:14.000 Okay? The questions. What did you use for the usgs? 2023 event set in your last figure. 00:55:14.000 --> 00:55:20.000 Netfield created a stochastic event set from the 2,023 model. 00:55:20.000 --> 00:55:22.000 In beta form. 00:55:22.000 --> 00:55:27.000 And we are using that. 00:55:27.000 --> 00:55:31.000 Okay, and le gives a thbs up. 00:55:31.000 --> 00:55:33.000 Great. 00:55:33.000 --> 00:55:42.000 Okay, any more questions, please. Raise your hand, and we can spotlight you. 00:55:42.000 --> 00:55:48.000 i know there was lots of discussion. About brandon's talk, too. 00:55:48.000 --> 00:56:02.000 reprocessing the. legacy seismic data. And i think that's A great example of a project that. 00:56:02.000 --> 00:56:09.000 where the pipeline from from science to application is really important. Brandon, i don't know if you want to talk about. 00:56:09.000 --> 00:56:20.000 That, and how you plan to integrate the work you're doing into kind of the counity models that are starting to Be developed as part of the crescent or other organizations. 00:56:20.000 --> 00:56:29.000 Ye that's Ye Thank you for giving me the opportunity because i wanted to mention at the end, and i got flustered by the bell that came up. 00:56:29.000 --> 00:56:59.000 ye absolutely. We we spent a lot of work just recovering the data. It. It's a lot to work with this legacy data. Some of it is not well docented that you have corrupted files. So You know. That's a that's a big lift in terms of ye contributing to the counity. We're right now. We're working on trying to merge the preliminary surface of the slab in this area with that from the kc. 21 footprint further north, and we're gonna. 00:57:00.000 --> 00:57:15.000 Pay a little bit of careful attention around the triple junction area. And then there's also some other data that we're gonna try and analyze here the onshore offshore recordings from that experiment. To get better seismic velocity control so we can convert things into depth. 00:57:15.000 --> 00:57:30.000 So ye it seems like. It's it's very well aligned with crescent, and will be fed into the counity fault, model and counity velocity model as well as other products, i'm sure, from the earthquake hazard side. So ye 00:57:30.000 --> 00:57:37.000 Great. Thank you so much. Really interesting work. 00:57:37.000 --> 00:57:45.000 Okay, i will make a last call for questions for any of our thunder talk. Speakers. 00:57:45.000 --> 00:57:51.000 Otherwise maybe i'll turn it back to sar H And we can have an extra minute in our break. Oh, thank you, christy. Ye great. 00:57:51.000 --> 00:57:54.000 Okay. Christy's got a question. 00:57:54.000 --> 00:58:00.000 Thanks. This is for killian There's so few paleo-sismologists in the world. 00:58:00.000 --> 00:58:10.000 There's so little paleo seismic data compared to the nbers and segments of interesting faults like i know you've thought about that? How do you accoodate that in your status. 00:58:10.000 --> 00:58:11.000 6. 00:58:11.000 --> 00:58:19.000 I mean something i lost over is i. I want to give a lot of credit to mike oskin for letting me play around with like. 00:58:19.000 --> 00:58:24.000 Pretty big data set of. Paleo seismic sites. I. 00:58:24.000 --> 00:58:28.000 I guess the 1st thing i wanna say is that there was more than i expected. 00:58:28.000 --> 00:58:31.000 That being said, i think there's also a big effort. 00:58:31.000 --> 00:58:35.000 To acculate the existing records that needs to happen. 00:58:35.000 --> 00:58:42.000 And. i think, as those. 00:58:42.000 --> 00:58:53.000 Records get aggregated. We can start. Sort of pinning down those questions of like, what are the right recurrence models? What are the right statistical models? And i think something mike will show probably this afternoon is that. 00:58:53.000 --> 00:59:01.000 They're. They're cool things to do with it that data. So stay tuned this afternoon, and you'll see answers to your question. 00:59:01.000 --> 00:59:03.000 Awesome. Thank you. 00:59:03.000 --> 00:59:08.000 And can i just add, pop in here, and just, you know, say we can. 00:59:08.000 --> 00:59:11.000 We can come up with slip rates with mechanical models. 00:59:11.000 --> 00:59:18.000 We don't know if they're right, if we don't have data to compare them to. But maybe there's a place there to work together to try to build that data set. 00:59:18.000 --> 00:59:25.000 At least a synthetic one to test different. Ideas right? That could be. 00:59:25.000 --> 00:59:30.000 Easily easily done in a mechanical way. In a physics. Space, like consistent way how to faults interact. 00:59:30.000 --> 00:59:32.000 So. 00:59:32.000 --> 00:59:36.000 Absolutely. 00:59:36.000 --> 00:59:42.000 Excellent, and thank you, kellyan, for giving us that nice little plug for. 00:59:42.000 --> 00:59:46.000 The talks coming up this afternoon. However, right now. 00:59:46.000 --> 00:59:51.000 And oh, and huge, thanks to all of our wonderful thunder talks, because then grace on moderate oh, they are amazing. 00:59:51.000 --> 00:59:57.000 Now, right now we get to have a little break until 1, 45 pacific time, unless. 00:59:57.000 --> 01:00:02.000 You all in our next session. Our awesome talk about tsunamis. 01:00:02.000 --> 01:00:17.000 So if you all would. Orlandoff, louissauelo lynette. Laselle or ross to stick around so we can get you set up. Everyone else. Take 15, and we'll see you back. 01:00:17.000 --> 01:00:27.000 其他的。