WEBVTT Kind: captions Language: en-US 00:00:01.906 --> 00:00:15.781 [silence] 00:00:15.781 --> 00:00:17.438 Hey, everyone. Welcome to the Earthquake 00:00:17.438 --> 00:00:21.093 Science Center seminar for August 17, 2022. 00:00:21.093 --> 00:00:23.867 Please remember to mute your mic and turn off your camera. 00:00:23.867 --> 00:00:24.910 A quick announcement. 00:00:24.910 --> 00:00:28.650 We’ll be having a series of five special seminars over the next couple of weeks. 00:00:28.650 --> 00:00:30.730 You should have received an email about that yesterday. 00:00:30.730 --> 00:00:32.680 They’ll be hosted by Dave Lockner. 00:00:32.680 --> 00:00:35.900 And we’re going to be sending out email reminders about each of these talks 00:00:35.900 --> 00:00:38.480 a couple days in advance and right before the talk. 00:00:38.480 --> 00:00:43.960 So you can look out for those and remember to join by clicking the link in each email. 00:00:43.960 --> 00:00:47.650 As usual, for today’s talk, you can put questions into the chat box at any time. 00:00:47.650 --> 00:00:50.063 And we’ll be taking questions at the end, so at that point, you’ll have 00:00:50.063 --> 00:00:53.750 an opportunity to unmute yourself and ask questions directly to the speaker. 00:00:53.750 --> 00:00:56.070 Today’s speaker is Matt Herman. 00:00:56.070 --> 00:00:59.938 Matt completed his bachelor’s degree in geology and physics at Amherst, 00:00:59.939 --> 00:01:04.125 and then he went on to do a master’s and a Ph.D. at Penn State University 00:01:04.126 --> 00:01:06.719 under the supervision of Kevin Furlong. 00:01:06.719 --> 00:01:11.375 From there, he went to Utrecht where he did a postdoc under the supervision 00:01:11.375 --> 00:01:15.352 of Rob Govers. And he is now at Cal State-Bakersfield 00:01:15.353 --> 00:01:17.086 where he’s an assistant professor. 00:01:17.086 --> 00:01:19.313 Matt will be speaking with us today about earthquakes 00:01:19.313 --> 00:01:22.430 in an uncoupled subduction zone. Take it away, Matt. 00:01:22.430 --> 00:01:25.125 - Great. Thank you so much for the introduction, 00:01:25.125 --> 00:01:27.562 thank you for inviting me, and good morning, everyone. 00:01:27.563 --> 00:01:32.188 So today I’m going to talk about some of the work I’ve been doing with these 00:01:32.188 --> 00:01:35.376 collaborators on the recent sequence of earthquakes 00:01:35.377 --> 00:01:39.184 near the Shumagin Islands in Alaska. 00:01:39.695 --> 00:01:43.033 Okay, so – this works. Gotcha. 00:01:43.033 --> 00:01:47.563 Okay, I want to take you back and give you a little bit of historical context about the 00:01:47.563 --> 00:01:49.824 Shumagin Islands and why they’re important 00:01:49.824 --> 00:01:54.364 from a seismotectonic perspective. So we’re going to go back to 1979. 00:01:54.364 --> 00:01:57.614 Plate tectonics is brand new at this point, and people are just starting to 00:01:57.614 --> 00:02:02.501 think about the implications of plate tectonics for various geological 00:02:02.501 --> 00:02:06.103 and geophysical observations, particularly earthquakes. 00:02:06.103 --> 00:02:11.804 And so this paper back in 1979 thought about this idea of seismic gaps. 00:02:11.804 --> 00:02:15.610 These were places that had not had large earthquakes – 00:02:15.610 --> 00:02:17.134 they said for more than 30 years. 00:02:17.134 --> 00:02:20.485 We’ll just say for some longish amount of time. 00:02:20.485 --> 00:02:24.104 And they said that, hey, this place hasn’t had an earthquake for a long time. 00:02:24.104 --> 00:02:27.173 We know the plates are moving relative to each other. 00:02:27.173 --> 00:02:31.914 This is telling us there’s a likelihood to be a big earthquake in this region. 00:02:31.914 --> 00:02:35.544 So let me take you to a map instead of reading just their abstract. 00:02:35.544 --> 00:02:38.104 We can see this colored map of the world. 00:02:38.104 --> 00:02:43.013 The colors indicate where the authors said, okay, here’s where we’re likely to have 00:02:43.013 --> 00:02:46.444 big earthquakes. Here’s where we’re not likely to have big earthquakes. 00:02:46.444 --> 00:02:52.876 So, if we circle the Shumagin Islands region of the Aleutian subduction zone, 00:02:52.876 --> 00:02:55.173 they said the historic record is incomplete, 00:02:55.173 --> 00:02:58.793 but maybe it has the potential for large earthquakes, particularly because 00:02:58.793 --> 00:03:03.244 most of the rest of that margin had experienced large to 00:03:03.244 --> 00:03:07.983 extremely large earthquakes, some very recently. 00:03:07.983 --> 00:03:10.914 I just want to make the point here – I’m not really going to talk much about 00:03:10.914 --> 00:03:16.184 Cascadia, but, you know, in 1979, our knowledge was limited. [chuckles] 00:03:16.184 --> 00:03:18.904 They didn’t even color Cascadia. It wasn’t even considered as 00:03:18.904 --> 00:03:22.354 a seismic hazard. Now we know things a little bit better there. 00:03:22.354 --> 00:03:27.235 So just that little bit of trivia for some of you to note that Cascadia 00:03:27.235 --> 00:03:31.047 is relatively new in our understanding of seismic hazards. 00:03:31.047 --> 00:03:36.314 Okay, so just a couple years later, this paper came out focusing on 00:03:36.314 --> 00:03:39.924 the Shumagin seismic gap – this place where big earthquakes 00:03:39.924 --> 00:03:45.814 hadn’t occurred before, but the rest of this margin had basically ruptured in the 00:03:45.814 --> 00:03:49.233 20th century. And so the authors identified this region 00:03:49.233 --> 00:03:54.414 as having a high probability for a great earthquake, something magnitude 8 00:03:54.414 --> 00:03:57.454 or larger, during the next one to two decades. 00:03:57.454 --> 00:04:00.313 So we’re looking at this paper, published in 1981. 00:04:00.313 --> 00:04:02.513 They’re saying, before the turn of the millennium, 00:04:02.513 --> 00:04:05.394 there’s going to be a great earthquake here. 00:04:05.394 --> 00:04:08.614 But no one said earthquake prediction was easy. 00:04:08.614 --> 00:04:10.934 As we know, there hasn’t been a great earthquake there. 00:04:10.934 --> 00:04:14.704 In fact, 10 years after that paper was published, this one came out. 00:04:14.704 --> 00:04:19.196 And it was re-evaluating the seismic gap hypothesis. 00:04:19.196 --> 00:04:24.103 And, in fact, it argued that the increased earthquake potential after 00:04:24.103 --> 00:04:27.944 a long quiet period can be rejected with a large confidence. 00:04:27.944 --> 00:04:32.446 And one of the places they looked at, in fact, was the Shumagin seismic gap 00:04:32.446 --> 00:04:36.044 in here. And they said that, oh, look, this is a place that 00:04:36.044 --> 00:04:39.164 hasn’t had a large earthquake. That actually probably means 00:04:39.164 --> 00:04:42.104 it’s not going to be able to have a large earthquake. 00:04:42.104 --> 00:04:45.877 As opposed to these other places to the east and to the west of it, 00:04:45.877 --> 00:04:49.194 which just seemed to be more likely to have large earthquakes. 00:04:49.194 --> 00:04:53.914 So there’s this debate about whether we can even use these ideas of plate tectonics 00:04:53.914 --> 00:04:56.784 to say anything about seismic hazard, 00:04:56.784 --> 00:05:00.845 specifically in this area, but maybe globally as well. 00:05:00.845 --> 00:05:04.974 Okay, so this is kind of the picture today of the Shumagin seismic gap. 00:05:04.974 --> 00:05:09.634 It’s a fancy version of the pictures we’ve already looked at. 00:05:09.634 --> 00:05:14.844 The red areas represent the aftershock zones of great earthquakes. 00:05:14.844 --> 00:05:17.874 You can see most of these are in the 20th century. 00:05:17.874 --> 00:05:21.474 And most of this plate boundary has ruptured 00:05:21.474 --> 00:05:25.254 in a great megathrust earthquake in the 20th century. 00:05:25.254 --> 00:05:29.329 The Shumagin seismic gap is the notable area that hasn’t. 00:05:29.329 --> 00:05:32.814 The question is, does it mean it’s more likely or less likely to have 00:05:32.814 --> 00:05:37.517 a big earthquake. And, to answer this, we need a little bit more information. 00:05:37.517 --> 00:05:39.794 We’re going to start with a geodetic view. 00:05:39.794 --> 00:05:44.694 Okay, so when we think about whether an earthquake is going to occur 00:05:44.694 --> 00:05:48.344 in a subduction zone, we think of maybe coupled versus 00:05:48.344 --> 00:05:51.744 uncoupled subduction zones. I’m going to use this term a lot in here. 00:05:51.744 --> 00:05:55.564 And I’m going to try and explain what I mean on this slide. 00:05:55.564 --> 00:05:59.830 So we’ve got this cartoon image here of the time period between earthquakes, 00:05:59.830 --> 00:06:05.002 the interseismic time period in a subduction zone, and, somewhere along their 00:06:05.002 --> 00:06:09.254 plate interface, the two plates are locked together, or coupled together. 00:06:09.254 --> 00:06:14.189 As a result of that, the motion of the subducting plate drives the 00:06:14.189 --> 00:06:20.002 upper plate landward. Okay, so we can see that, if we go to Japan before 00:06:20.002 --> 00:06:25.440 March 11th, 2011, all of Japan is moving to the west because 00:06:25.440 --> 00:06:30.434 the subducting Pacific Plate here is driving Japan west, okay? 00:06:30.434 --> 00:06:32.434 And we sort of know the elastic rebound model. 00:06:32.434 --> 00:06:34.024 We know what’s going to happen. 00:06:34.024 --> 00:06:37.024 When the big earthquake occurs after a century or a millennium 00:06:37.024 --> 00:06:41.244 or something like that, the upper plate is going to rebound landward. 00:06:41.244 --> 00:06:45.704 So we’re going to see all of the motion in the upper plate 00:06:45.704 --> 00:06:48.164 back the other direction during the earthquake. 00:06:48.164 --> 00:06:52.080 So this is what we would expect to see at a coupled subduction zone. 00:06:52.080 --> 00:06:54.978 Okay, what do we see in the Shumagin Islands? 00:06:54.978 --> 00:06:57.234 Okay, I’m going to actually start over here on the right. 00:06:57.234 --> 00:07:01.714 I know we usually read left to right, but the data are over here on the right 00:07:01.714 --> 00:07:04.994 from this paper by Li and Freymueller in 2018. 00:07:04.994 --> 00:07:10.714 And they’ve got GPS stations on several of the islands in here, 00:07:10.714 --> 00:07:13.004 the peninsulas in here. And each of these arrows 00:07:13.004 --> 00:07:18.634 represents the speed that that station is moving with respect to the 00:07:18.634 --> 00:07:21.274 North American Plate fixed somewhere back here. 00:07:21.274 --> 00:07:26.824 Okay, so what we see is these large landward velocities over in the east. 00:07:26.824 --> 00:07:32.268 These indicate there must be pretty high coupling on the plate interface 00:07:32.268 --> 00:07:35.877 over here driving – as the Pacific Plate subducts, 00:07:35.878 --> 00:07:40.234 driving these islands to the northwest. 00:07:40.234 --> 00:07:43.994 As we go over to the circled area – that’s the Shumagin seismic gap – 00:07:43.994 --> 00:07:46.705 we see lower landward velocities. 00:07:46.705 --> 00:07:51.424 And that implies lower plate interface coupling in this region. 00:07:51.424 --> 00:07:53.814 So, you know, many of you are probably familiar. 00:07:53.814 --> 00:07:57.334 You can invert these onshore data for a distribution of coupling. 00:07:57.334 --> 00:08:01.114 And we see these red areas over here are highly coupled. 00:08:01.114 --> 00:08:04.484 That’s not surprising. There were large earthquakes. 00:08:04.484 --> 00:08:08.384 The sort of western end of which – the 1964 was over here. 00:08:08.384 --> 00:08:11.954 The 1938 earthquake occurred generally in an area of intermediate 00:08:11.954 --> 00:08:15.424 to high coupling. The Shumagin seismic gap over here 00:08:15.424 --> 00:08:19.134 is an area of intermediate to low coupling, okay? 00:08:19.134 --> 00:08:24.013 So this seems to be telling us that maybe the Shumagin seismic gap 00:08:24.013 --> 00:08:27.884 isn’t going to be a place that has large earthquakes. 00:08:27.884 --> 00:08:31.334 Now, interestingly – and I’m going to kind of punt on this a couple of times 00:08:31.334 --> 00:08:34.846 in this talk – the 1946 earthquake is over here. 00:08:34.846 --> 00:08:39.824 That was a big earthquake. Big tsunami. Large aftershock zone. 00:08:39.824 --> 00:08:43.004 And you’ll notice it’s in an area that has low coupling. 00:08:43.004 --> 00:08:46.463 We can talk about that later if people want, but as I said, 00:08:46.463 --> 00:08:50.834 I’m generally going to punt on 1946 throughout this talk. 00:08:51.534 --> 00:08:56.444 Okay, so that’s the sort of geodetic horizontal motion view. 00:08:56.444 --> 00:08:58.664 We can also take a geologic view. 00:08:58.664 --> 00:09:02.314 What do we expect to see at a coupled subduction zone? 00:09:02.314 --> 00:09:05.753 Throughout the earthquake cycle, we think we’re going to see the coast 00:09:05.753 --> 00:09:09.194 go up and down and up and down throughout the earthquake cycle. 00:09:09.194 --> 00:09:12.204 Now, that vertical motion is fairly complicated. 00:09:12.204 --> 00:09:17.190 And why it goes up in some places and why it goes down in some places is actually 00:09:17.190 --> 00:09:23.694 one of the unsolved mysteries, I would say, of various subduction zones. 00:09:23.694 --> 00:09:25.834 We think we know what’s going to happen coseismically. 00:09:25.834 --> 00:09:31.213 There’s going to be this red line, this pattern, of subsidence to uplift 00:09:31.213 --> 00:09:34.691 as you sort of cross the rupture zone, which is going to be generally 00:09:34.691 --> 00:09:39.604 in the coupled zone. But what happens interseismically in this green line. 00:09:39.604 --> 00:09:43.414 What’s going to happen postseismically. 00:09:43.414 --> 00:09:46.816 These are going to be a little more complicated and depend on things like 00:09:46.816 --> 00:09:50.816 the distribution of coupling, the distribution of rheology 00:09:50.816 --> 00:09:54.440 in the subduction system. In any case, we’re going to see things going up. 00:09:54.440 --> 00:09:56.253 We’re going to see things going down. 00:09:56.253 --> 00:10:00.003 And, in a coupled subduction zone, we’re going to look in this region 00:10:00.003 --> 00:10:03.274 and generally see sudden vertical motions. 00:10:03.274 --> 00:10:06.744 Okay, so what are we going to – what do we see in the Aleutian Islands? 00:10:06.744 --> 00:10:09.254 If we look over here, this is the part where 00:10:09.254 --> 00:10:13.124 the geodesy tells us is pretty well-coupled. 00:10:13.124 --> 00:10:17.753 And this is not my area of expertise, so please don’t ask me too much about it, 00:10:17.753 --> 00:10:23.414 but if you look at the geology on Sitkinak Island, 00:10:23.414 --> 00:10:28.628 there is evidence that the coastline has gone up rapidly and down rapidly 00:10:28.628 --> 00:10:32.261 at various times. We can see this in the lithology. 00:10:32.261 --> 00:10:36.614 You can see this in the critters that live on the coast at different water depths. 00:10:36.614 --> 00:10:41.144 And so there’s several times over the last few thousand years 00:10:41.144 --> 00:10:44.324 where the coast has gone up and down, indicating, 00:10:44.324 --> 00:10:49.124 oh, yeah, earthquakes must be happening in and around this region consistent with this 00:10:49.124 --> 00:10:52.776 being a coupled part of the subduction zone. 00:10:52.776 --> 00:10:55.964 Okay, well, what about in the Shumagin Islands? 00:10:55.964 --> 00:11:00.794 Is there a comparable signal in the geologic record of the Shumagin Islands? 00:11:00.794 --> 00:11:04.724 Apparently not. Okay, so this paper by Rob Witter and 00:11:04.724 --> 00:11:09.964 folks argue that there’s little Holocene strain accumulation and release 00:11:09.964 --> 00:11:16.574 on this part of the megathrust. Okay? In fact, they say, if the rupture of the 00:11:16.574 --> 00:11:20.129 megathrust produced great earthquakes, then coseismic uplift or subsidence 00:11:20.129 --> 00:11:23.034 was too small to perturb the onshore geologic record. 00:11:23.034 --> 00:11:29.589 Okay, no stratigraphic or geomorphic evidence for sudden vertical motions. Okay? 00:11:29.589 --> 00:11:32.994 So a conclusion there could be that there’s just not big earthquakes. 00:11:32.994 --> 00:11:36.684 Or maybe these earthquakes are somehow hidden. 00:11:36.684 --> 00:11:41.941 So this is the picture of the Shumagin gap before 2020, that it’s poorly coupled 00:11:41.941 --> 00:11:45.614 and aseismic. And I’ve highlighted the Shumagin gap region. 00:11:45.614 --> 00:11:52.713 Really what we’re talking about is this white dashed area, okay? 00:11:52.713 --> 00:11:57.604 We saw the 1938 region had a big earthquake in 1938. 00:11:57.604 --> 00:12:00.534 The geodetic data suggests it’s coupled. 00:12:00.534 --> 00:12:03.863 The geologic data also suggests it’s coupled. 00:12:03.863 --> 00:12:06.042 Questions remain about 1946. 00:12:06.042 --> 00:12:08.316 As I said, I’m just not going to get into that too much right now, 00:12:08.316 --> 00:12:12.154 but we can talk about it. We do see some seismicity in these various regions. 00:12:12.154 --> 00:12:18.254 I didn’t put the depths on here, but we see seismicity clustering in and 00:12:18.254 --> 00:12:20.114 around the asperities and actually at the base 00:12:20.114 --> 00:12:24.574 of the seismogenic zone of the Shumagin gap. 00:12:24.574 --> 00:12:31.504 Okay, so just kind of give a cartoon image summarizing what we think 00:12:31.504 --> 00:12:33.954 the mechanical state of the plate boundary is. 00:12:33.954 --> 00:12:36.264 We’ve got an uncoupled Shumagin gap. 00:12:36.264 --> 00:12:41.317 We’ve got asperities, perhaps on either side, but certainly over here 00:12:41.317 --> 00:12:43.894 on the east side of it. Okay, so I’m reading this. 00:12:43.894 --> 00:12:47.434 I’m reading these papers. And I’m saying, okay, sort of makes sense. 00:12:47.434 --> 00:12:50.224 I sort of understand it. End of story. 00:12:50.224 --> 00:12:51.944 We’re going to get earthquakes maybe around the Shumagin gap, 00:12:51.944 --> 00:12:54.824 but it’s just never going to have an earthquake. 00:12:54.824 --> 00:12:58.629 I wasn’t thinking too deeply about the implications yet, but I have since then. 00:12:58.629 --> 00:13:01.847 Okay, so July 21st, 2020. 00:13:01.847 --> 00:13:04.043 Magnitude 7.8 earthquake. 00:13:04.043 --> 00:13:10.534 And it occurs right here in this area that may or may not be part of the Shumagin gap. 00:13:10.534 --> 00:13:14.354 So the initial stories about this paper – the initial reports 00:13:14.354 --> 00:13:17.914 said things like, hey, scientists predicted an earthquake like this 00:13:17.914 --> 00:13:22.514 would happen in 1981. We looked at that paper. Okay? 00:13:22.514 --> 00:13:25.874 But, you know, news that comes out in the first few days 00:13:25.874 --> 00:13:29.418 after an earthquake should be taken with large grains of salt. 00:13:29.418 --> 00:13:31.994 Science that comes out in the first few days after an earthquake 00:13:31.994 --> 00:13:34.384 should also be taken with a large grain of salt. 00:13:34.384 --> 00:13:39.014 So we wanted to ask, did this earthquake actually break the Shumagin gap? 00:13:39.014 --> 00:13:42.804 Okay, that’s kind of a fundamental question. 00:13:42.804 --> 00:13:46.574 And, in fact, the earthquake epicenter was located over here 00:13:46.574 --> 00:13:51.957 in the western corner of the 1938 aftershock region. 00:13:51.957 --> 00:13:59.692 Now, we actually think the behavior of this earthquake represents a transition 00:13:59.692 --> 00:14:05.834 from rupture occurring in an area where we have an asperity, 00:14:05.834 --> 00:14:10.614 things are strongly coupled, and then, pop, you know, it goes off in an earthquake. 00:14:10.614 --> 00:14:15.004 And then a rupture crossing into maybe a lower coupling area. 00:14:15.004 --> 00:14:17.463 So I want to talk a little bit about this. 00:14:17.463 --> 00:14:22.755 If we focus on seismic analyses of this event, they locate slip 00:14:22.755 --> 00:14:26.094 right near the epicenter. So if we look at the USGS finite 00:14:26.094 --> 00:14:32.544 fault model, the one that only uses seismic waves, they see a concentration of slip 00:14:32.544 --> 00:14:34.713 right around the star here. That’s the epicenter. 00:14:34.713 --> 00:14:38.393 Okay, so, you know, [chuckles] that would be sort of what 00:14:38.393 --> 00:14:41.834 a lot of people would expect for an event like this, okay? 00:14:41.834 --> 00:14:43.824 Decent size, 7.8, but not enormous. 00:14:43.824 --> 00:14:46.804 So just large slip in and around the epicenter. 00:14:46.804 --> 00:14:51.304 But it turns out, when we added local information to the inversion – 00:14:51.304 --> 00:14:56.567 so when you did geodetic and seismic analyses, the slip started here 00:14:56.567 --> 00:14:59.824 at the epicenter – and there was plenty of slip over here, 00:14:59.824 --> 00:15:04.403 but it actually propagated over to the west. 00:15:04.403 --> 00:15:09.590 So, in fact, the centroid of this event was west of the epicenter. 00:15:09.590 --> 00:15:11.234 Okay, so how do we interpret this? 00:15:11.234 --> 00:15:15.713 Well, the fast slip – the stuff that goes – you know, pop, the asperity unlocks and 00:15:15.713 --> 00:15:20.194 produces shaking, that is happening over in the coupled zone. 00:15:20.194 --> 00:15:23.514 We still have slip over here west of the coupled zone, 00:15:23.514 --> 00:15:26.704 but maybe it’s a little slower – less seismically productive. 00:15:26.704 --> 00:15:30.264 So we can see it with local displacements. It’s still slip. 00:15:30.264 --> 00:15:35.145 But it’s not something that is as easy to see teleseismically. 00:15:35.145 --> 00:15:40.255 Okay, so, in fact, that transition might be indicating where we’re going 00:15:40.255 --> 00:15:43.248 from coupling to something else. 00:15:43.270 --> 00:15:47.163 But still, we haven’t really ruptured the Shumagin gap here. 00:15:47.164 --> 00:15:50.804 So the main shock here appears to have broken across the northwest edge 00:15:50.804 --> 00:15:54.153 of the 1938 asperity. We’ve got these aftershocks 00:15:54.153 --> 00:15:58.604 and afterslip mainly down-dip of the Shumagin gap. 00:15:58.604 --> 00:16:02.494 But the Shumagin gap itself is mainly aseismic – a couple of aftershocks 00:16:02.494 --> 00:16:06.374 here and there, but nothing major in here, okay? 00:16:06.374 --> 00:16:11.333 So still looks aseismic from a megathrust perspective. 00:16:11.333 --> 00:16:16.463 Fast forward a couple of months. Boom. Magnitude 7.6 earthquake 00:16:16.463 --> 00:16:20.273 in the Shumagin gap. So, without even looking at any details 00:16:20.273 --> 00:16:23.674 of this event, I looked at this and said, okay, well, I guess a lot of people 00:16:23.674 --> 00:16:28.844 were wrong about the Shumagin gap. It’s not aseismic after all. 00:16:28.844 --> 00:16:33.974 But it turns out this was not the expected event. 00:16:33.974 --> 00:16:37.734 So I think most people, when they’re looking at subduction zones, 00:16:37.734 --> 00:16:40.974 and they’re looking at the big earthquakes in subduction zones, 00:16:40.974 --> 00:16:43.734 they are expecting megathrust earthquakes – 00:16:43.734 --> 00:16:48.354 these shallow thrust-faulting earthquakes on the plate interface. 00:16:48.354 --> 00:16:51.135 This was not what happened. So, first of all, the mechanism 00:16:51.135 --> 00:16:55.653 is not thrust faulting at all. It’s oblique strike-slip. 00:16:55.653 --> 00:17:00.443 Also, when we look at the locations of these events, particularly evident 00:17:00.443 --> 00:17:04.443 in cross-section here, so going from the Pacific Plate over to North America, 00:17:04.443 --> 00:17:09.505 Pacific Plate over to North America, we see the aftershocks are pretty much 00:17:09.505 --> 00:17:12.864 exclusively below the plate interface here. 00:17:12.864 --> 00:17:19.183 So it appears that this is an intra-slab, right-lateral strike-slip – 00:17:19.183 --> 00:17:21.493 oblique strike-slip earthquake. 00:17:21.493 --> 00:17:25.256 And new information has come out suggesting this might be a little more complicated 00:17:25.256 --> 00:17:28.631 than that, but, you know, all of the analyses show this 00:17:28.632 --> 00:17:32.208 dominantly strike-slip character of this event. 00:17:32.208 --> 00:17:36.162 So, just to give you a little bit of context to why this is weird, 00:17:36.162 --> 00:17:40.005 okay, first of all, how you get it. It’s in the seismic gap. 00:17:40.005 --> 00:17:43.133 But the seismic gap, I guess, only refers to megathrust earthquakes. 00:17:43.133 --> 00:17:45.365 This is intra-slab strike-slip. 00:17:45.365 --> 00:17:50.564 There have been big, you know, oceanic intraplate earthquakes before. 00:17:50.564 --> 00:17:53.944 About 10 magnitude greater than 7.5 since 1980. 00:17:53.944 --> 00:17:56.643 This is the only one that happened in the subduction zone. 00:17:56.643 --> 00:18:02.201 Everything else is kind of outer rise or even further offshore regions. 00:18:02.201 --> 00:18:05.482 Okay, so this is an unusual event. This is weird. 00:18:05.482 --> 00:18:09.193 And, to answer some of these questions and think about some of these issues, 00:18:09.193 --> 00:18:14.503 we asked ourselves, okay, let’s assume the Shumagin gap is, in fact, uncoupled. 00:18:14.503 --> 00:18:17.818 What stresses would accumulate within the subducting slab where 00:18:17.818 --> 00:18:22.740 this weird oblique strike-slip earthquake happened? 00:18:22.740 --> 00:18:26.256 Okay, so just to give you a little bit of sense of the modeling we do 00:18:26.256 --> 00:18:32.204 to think about this, remember we’ve got this picture of GPS velocities on 00:18:32.204 --> 00:18:35.993 the upper plate. You know, high velocities indicate high coupling. 00:18:35.993 --> 00:18:38.694 Low velocities indicate lower coupling. 00:18:38.694 --> 00:18:42.094 Got their Shumagin gap here. Okay, well, what is coupling? 00:18:42.094 --> 00:18:46.944 Coupling is some kind of mechanical connection between the subducting 00:18:46.944 --> 00:18:50.803 Pacific Plate and the overriding North American Plate – 00:18:50.804 --> 00:18:53.014 so Alaska and the Aleutian Islands. 00:18:53.014 --> 00:18:57.854 So we’ve developed some very simple locking models for subduction zones. 00:18:57.854 --> 00:19:01.099 We first published this work back in 2018. 00:19:01.099 --> 00:19:04.474 We’ve published a few things on it since then looking at some more details. 00:19:04.474 --> 00:19:08.923 But, in any case, we’ve got this picture here of the Pacific Plate subducting. 00:19:08.923 --> 00:19:13.543 Something is locked on the plate interface between the Pacific and the Alaska Plates. 00:19:13.543 --> 00:19:17.381 And we kept these models very geometrically simple because we don’t want to 00:19:17.382 --> 00:19:21.704 deal with all the complications of geometry and rheology and things like that. 00:19:21.704 --> 00:19:25.060 We just want to focus on the patterns of locking. 00:19:25.060 --> 00:19:27.993 Okay, so this is what our models look like. 00:19:27.993 --> 00:19:31.256 And some of the details are a little different from this initial paper and 00:19:31.256 --> 00:19:35.903 the ones we did for the Shumagin area, but in general, 00:19:35.903 --> 00:19:38.344 all these models have a very similar setup. 00:19:38.344 --> 00:19:40.854 We’ve got this triangular wedge of an upper plate. 00:19:40.854 --> 00:19:44.498 We’ve got a subducting plate that’s just planar. 00:19:44.498 --> 00:19:48.753 We add displacements to the upper plate to move it. 00:19:48.753 --> 00:19:52.084 And we put a backstop fixing the back of the upper plate 00:19:52.084 --> 00:19:54.701 to have relative motion. 00:19:54.701 --> 00:19:59.454 Okay, so, you know, relatively straightforward geometry. 00:19:59.454 --> 00:20:01.783 Purely elastic for now. 00:20:02.389 --> 00:20:05.084 Now, I guess the most important part is how we’re going to define locking 00:20:05.084 --> 00:20:09.204 on the plate interface. And we define areas – rectangular areas 00:20:09.204 --> 00:20:13.974 on the plate interface as locked – as saying, you cannot have any slip. 00:20:13.974 --> 00:20:16.733 We make the dramatically simplifying assumption that 00:20:16.733 --> 00:20:20.353 everything outside of locked areas is just free to slide. 00:20:20.353 --> 00:20:25.293 You can think of it like Teflon. It’s totally frictionless. 00:20:25.293 --> 00:20:30.513 And the only areas, then, in the context of this model, that can start earthquakes 00:20:30.513 --> 00:20:34.248 are these locked zones – these so-called asperities. 00:20:34.248 --> 00:20:38.474 Okay, so when we run this model, interseismically, what do we get? 00:20:38.474 --> 00:20:41.334 Here is that sort of perspective view of the model. 00:20:41.334 --> 00:20:45.314 On the right here, we’re sort of looking perpendicularly down onto 00:20:45.314 --> 00:20:49.124 the plate interface. We’ve got the trench over here on the left. 00:20:49.124 --> 00:20:55.209 Over here on the right, this is the plate interface down at 185-ish kilometers. 00:20:55.209 --> 00:21:02.543 Okay? So, in the areas we [audio cuts out] the colors indicate the amount of slip. 00:21:02.543 --> 00:21:06.319 Black indicates no slip. It would be kind of surprising if there were slip 00:21:06.320 --> 00:21:08.943 in areas we designated as having no slip. 00:21:08.944 --> 00:21:10.413 That would be mean our models weren’t working. 00:21:10.413 --> 00:21:12.459 So these are the locked areas. 00:21:12.459 --> 00:21:15.553 When we go far from the locked areas, everything is white. 00:21:15.553 --> 00:21:20.616 The plates are just happily cruising along at the relative motion. 00:21:20.616 --> 00:21:22.884 In this case, it’s 1 meter of relative motion, 00:21:22.884 --> 00:21:26.569 but because it’s elastic, we can scale it to plate motions. 00:21:26.569 --> 00:21:31.624 Now, there’s these areas in between the locked zones that, rheologically, 00:21:31.624 --> 00:21:34.923 mechanically, they could slip at the plate rate. 00:21:34.923 --> 00:21:38.507 But, because the locked zones are holding them back, 00:21:38.507 --> 00:21:41.413 they actually don’t slip at the full plate rate. 00:21:41.413 --> 00:21:43.494 We call this pseudo-coupling because 00:21:43.494 --> 00:21:46.974 it looks just like something that’s mechanically coupled, but it’s not. 00:21:46.974 --> 00:21:50.757 As I said, if these asperities, if these locked zones were not there, 00:21:50.757 --> 00:21:54.324 they would be happy to move at the full plate rate. 00:21:54.324 --> 00:21:57.663 Okay? So this produces a pattern 00:21:57.663 --> 00:22:01.423 of surface motions that would be seen in geodetic data. 00:22:01.423 --> 00:22:04.864 And we can say, hey, we’ve got this pattern of locking. 00:22:04.864 --> 00:22:07.754 What would this do to stresses in the plates around it? 00:22:07.754 --> 00:22:11.084 We don’t just have to look at the kinematics on the plate interface 00:22:11.084 --> 00:22:15.383 at the surface. We can probe this model for stresses as well. 00:22:15.383 --> 00:22:20.959 And that’s what we did. We set up this model for the Shumagin gap region, 00:22:20.960 --> 00:22:23.793 and we asked what the stresses would be in the [inaudible]. 00:22:23.793 --> 00:22:27.733 So, just to give you a little context, we’ve got our subduction trench 00:22:27.733 --> 00:22:32.053 right here. So we’re – got the Pacific Plate to the south. 00:22:32.053 --> 00:22:36.046 We’ve got the North American Plate to the north. 00:22:36.046 --> 00:22:41.632 We’re looking at the colors in here represent the stresses along the surface 00:22:41.632 --> 00:22:46.613 of the subducting Pacific Plate. We put a locked zone here in the east. 00:22:46.613 --> 00:22:50.084 That’s like the 1938 asperity. 00:22:50.084 --> 00:22:54.454 And everything else on the plate interface – the Shumagin gap region is uncoupled. 00:22:54.454 --> 00:22:59.194 Okay, again, think of it like Teflon. It’s just allowed to go freely. 00:22:59.194 --> 00:23:03.493 Okay, the colors in the background represent the stresses resolved 00:23:03.493 --> 00:23:09.324 on right-lateral strike-slip faulting perpendicular to the trench – 00:23:09.324 --> 00:23:13.834 so of the kind that occurred in October 2020. 00:23:13.834 --> 00:23:18.445 And we see that this is about the location of the earthquake in the context 00:23:18.445 --> 00:23:22.133 of the model. That’s squarely in the red to dark red area. 00:23:22.133 --> 00:23:26.964 So this earthquake is strongly favored by the stresses that accumulate 00:23:26.964 --> 00:23:32.763 in the subducting slab. And the reason why is really easy to see. 00:23:32.763 --> 00:23:38.993 So, if we’re out here over in the western part of the model, the subducting 00:23:38.993 --> 00:23:43.533 Pacific Plate is just happily cruising, you know, down the subduction zone. 00:23:43.533 --> 00:23:48.033 There’s no coupling, so it’s just, you know, going down without any problem. 00:23:48.033 --> 00:23:51.403 These arrows represent the motion of the subducting Pacific Plate. 00:23:51.403 --> 00:23:53.513 They’re large over here. 00:23:53.513 --> 00:23:59.258 In the locked zone, the upper plate acts as a buttress preventing the subducting 00:23:59.258 --> 00:24:02.113 Pacific slab from going down the subduction zone. 00:24:02.113 --> 00:24:06.613 So there’s no arrows in here because it’s being – it’s locked. 00:24:06.613 --> 00:24:08.824 It’s coupled to the overriding plate. 00:24:08.824 --> 00:24:13.944 So there’s a huge right-lateral shear as you cross from the locked zone 00:24:13.944 --> 00:24:15.653 into the unlocked zone. 00:24:15.653 --> 00:24:21.103 Okay, this was a right-lateral strike-slip fault strongly favor in the transition from 00:24:21.103 --> 00:24:26.603 locked to unlocked. So exactly where the October earthquake occurred. 00:24:26.603 --> 00:24:28.834 So our conclusion is that an uncoupled Shumagin gap 00:24:28.834 --> 00:24:31.914 strongly favors the type of faulting that occurred. 00:24:31.914 --> 00:24:35.554 Now, remember, there was a series of earthquakes and aftershocks and afterslip 00:24:35.554 --> 00:24:38.913 that also occurred. So we can ask how they affected 00:24:38.913 --> 00:24:45.133 the stress conditions at the October epicenter. 00:24:45.133 --> 00:24:50.133 We do this by calculating the Coulomb static stress change, thinking about the 00:24:50.133 --> 00:24:53.884 shear stresses on the fault, thinking about the normal stresses on the fault. 00:24:53.884 --> 00:24:57.383 We just do this in an elastic half-space, which is pretty consistent with 00:24:57.383 --> 00:25:01.614 the more complicated interseismic model I just showed you. 00:25:01.614 --> 00:25:07.133 And the conclusion – the take-home here is that the stress changes from 00:25:07.133 --> 00:25:13.013 the magnitude 7.8, particularly when you add this down-dip afterslip, 00:25:13.013 --> 00:25:19.883 strongly promote this type of faulting, okay, this oblique trench-perpendicular 00:25:19.883 --> 00:25:23.224 right-lateral strike-slip faulting, okay? 00:25:23.224 --> 00:25:27.064 So you can see the location is in the orange dot here. 00:25:27.064 --> 00:25:31.023 That is nice and red, particularly when you add the afterslip. 00:25:31.023 --> 00:25:33.354 Even without the afterslip, it’s generally red. 00:25:33.354 --> 00:25:36.253 But, when you add the afterslip, it goes squarely in the dark red. 00:25:36.253 --> 00:25:40.883 So that means high stresses promoting this type of faulting. 00:25:40.883 --> 00:25:43.043 Okay, so we’ve got two things going on here. 00:25:43.043 --> 00:25:48.314 We’ve got the interseismic stresses, which are strongly favoring this 00:25:48.314 --> 00:25:50.844 October type of mechanism. 00:25:50.844 --> 00:25:57.753 We’ve got the coseismic stress changes from the megathrust plus afterslip. 00:25:57.753 --> 00:26:03.633 So this is triggered faulting occurring in the Pacific slab. 00:26:03.633 --> 00:26:09.133 All right. So we’ve got this picture now of a normal – I should say typical – 00:26:09.133 --> 00:26:12.613 normal has other meanings – megathrust event here 00:26:12.613 --> 00:26:14.744 with its afterslip and aftershocks. 00:26:14.744 --> 00:26:19.321 We’ve got this unusual event that now seems to make a little bit of sense in 00:26:19.321 --> 00:26:22.064 an uncoupled to coupled transition. 00:26:22.064 --> 00:26:25.364 And we can put this in our sort of cartoon of what’s going on. 00:26:25.364 --> 00:26:28.824 We’ve got the locked asperity area. 00:26:28.824 --> 00:26:32.433 Okay, that’s what produces the megathrust event. 00:26:32.433 --> 00:26:35.863 This zone of pseudo coupling – so kinematic coupling here. 00:26:35.863 --> 00:26:39.509 That’s why it kind of looks like it has intermediate coupling even though 00:26:39.509 --> 00:26:42.149 the whole area is mechanically unlocked. 00:26:42.149 --> 00:26:46.571 But, in that transition is where the October 2020 strike-slip 00:26:46.571 --> 00:26:49.043 earthquake occurred. Okay. 00:26:49.043 --> 00:26:50.773 So there’s a little bit of an elephant in the room. 00:26:50.773 --> 00:26:54.407 Some of you may be thinking about it already. 00:26:54.407 --> 00:26:58.149 The question is, a 7.6 is not a small earthquake. 00:26:58.149 --> 00:27:02.714 It’s not like this is just first breaking the slab in a 7.6 earthquake. 00:27:02.714 --> 00:27:07.084 So where did the strike-slip fault come from? 00:27:07.084 --> 00:27:09.464 The answer is not totally clear. 00:27:09.464 --> 00:27:13.883 We have some hypotheses, but we’re not entirely sure. 00:27:13.883 --> 00:27:20.196 There do seem to be some magnetic anomalies over here that might indicate there are 00:27:20.196 --> 00:27:24.123 fracture zones oriented roughly north-south in this region. 00:27:24.123 --> 00:27:27.579 So it could be activating a fracture zone. 00:27:27.579 --> 00:27:32.613 It’s possible, I guess, that it could also be some kind of outer rise, but those would 00:27:32.613 --> 00:27:37.344 be nearly perpendicular to the actual strike-slip faulting that occurred. 00:27:37.344 --> 00:27:40.913 So it would be a little surprising to be an outer rise fault. 00:27:40.913 --> 00:27:44.944 The slab undergoes some curvature in this region along strike, 00:27:44.944 --> 00:27:48.523 so there could be stresses associated with that as well. 00:27:48.523 --> 00:27:51.803 But our guess at the moment, and really is a guess, 00:27:51.803 --> 00:27:57.314 is that it’s related to subducting fracture zones that are being re-activated 00:27:57.314 --> 00:28:02.360 in the stress field associated with the transition and coupling. 00:28:02.360 --> 00:28:06.734 Okay, so we’ve got some conclusions. Through June 2020, 00:28:06.734 --> 00:28:10.944 these earthquakes actually paradoxically, maybe, provide evidence for low coupling 00:28:10.944 --> 00:28:13.373 on the plate interface in the Shumagin gap. 00:28:13.373 --> 00:28:16.584 But low coupling doesn’t mean aseismic, okay? 00:28:16.584 --> 00:28:20.423 We’ve got strike-slip faulting favored in this transition 00:28:20.423 --> 00:28:23.324 from coupled to uncoupled. 00:28:23.324 --> 00:28:26.947 We see that, if that can be triggered, as you have this sort of typical event 00:28:26.947 --> 00:28:31.944 triggering the atypical event. And, in fact, the atypical event 00:28:31.944 --> 00:28:35.759 occurred in a region thought to be aseismic but really should not be 00:28:35.759 --> 00:28:38.834 thought of as aseismic. In fact, the tsunami from the 00:28:38.834 --> 00:28:42.947 October 19th event was substantially larger than the tsunami from 00:28:42.947 --> 00:28:46.663 the July 21st event, which I think surprised me and a lot of people. 00:28:46.663 --> 00:28:50.947 I’m not a tsunami scientist, but I would have certainly thought a thrust-faulting 00:28:50.947 --> 00:28:54.235 event would have produced a bigger tsunami, but it didn’t. 00:28:54.235 --> 00:28:56.126 I’m not going to get too much into the tsunami, 00:28:56.126 --> 00:28:58.283 but I’m happy to talk about it if people want to. 00:28:58.283 --> 00:29:02.613 Okay, so that could be the end of the talk, but I got a little more time. 00:29:02.613 --> 00:29:06.454 And you may be thinking, hey, wait, wasn’t there another earthquake there? 00:29:06.454 --> 00:29:09.324 And you’d be right. Here’s my title slide again. 00:29:09.324 --> 00:29:12.665 And we didn’t talk at all about the green earthquakes. 00:29:12.665 --> 00:29:16.884 And we probably should talk about the green earthquakes, because, in fact, 00:29:16.884 --> 00:29:22.403 the earthquake in July 2021 was the largest event in the sequence. 00:29:22.403 --> 00:29:24.143 Magnitude 8.2. 00:29:24.143 --> 00:29:30.214 It occurred on the megathrust in the 1938 asperity region along its down-dip edge. 00:29:30.214 --> 00:29:35.446 So it didn’t seem to actually break the entire 1938 asperity. 00:29:35.446 --> 00:29:40.814 The earthquake epicenter is here, so it started right next to the July 2020 event, 00:29:40.814 --> 00:29:43.673 but most of the slip, instead of propagating to the west, 00:29:43.673 --> 00:29:46.814 propagated to the east of the epicenter. 00:29:46.814 --> 00:29:50.233 Okay, so the first thing we ask when this sort of thing happens – 00:29:50.233 --> 00:29:52.354 or, I ask, at least, when this sort of thing happens, 00:29:52.354 --> 00:29:56.293 is did the previous earthquakes perhaps trigger the big one. 00:29:56.293 --> 00:30:00.510 And those of you who have thought about stress changes are probably, like, 00:30:00.510 --> 00:30:03.853 well, yeah, this is going to be an easy one to model. 00:30:03.853 --> 00:30:08.853 We see that the previous events strongly load the epicenter of the magnitude 8.2. 00:30:08.853 --> 00:30:12.572 And this is expected behavior for the region right off the edge 00:30:12.572 --> 00:30:16.385 of a previous earthquake on the same fault. Okay, so we see, 00:30:16.385 --> 00:30:21.806 here is the rough rupture area of the 8.2. Here’s its afterslip zone. 00:30:21.806 --> 00:30:26.793 And here is the – sorry, this is the 7.8. This is then the 7.6. 00:30:26.793 --> 00:30:31.964 And we see that those strongly favor the magnitude 8.2 thrust faulting there. 00:30:31.964 --> 00:30:36.194 Okay, so the triggering story is relatively straightforward, but we actually think that 00:30:36.194 --> 00:30:39.084 this earthquake tells us something about the coupling 00:30:39.084 --> 00:30:41.205 in the Shumagin gap. 00:30:41.205 --> 00:30:46.834 And we had to think about this in a little more nuanced way to make this argument. 00:30:46.834 --> 00:30:50.955 So let’s start with the earthquake slip. 00:30:50.955 --> 00:30:54.260 Models of the earthquake showed that the peak slip was 00:30:54.260 --> 00:30:58.009 somewhere in the 4 to 5-ish meters range. 00:30:58.009 --> 00:31:03.704 Okay, so that’s fine. That’s pretty typical for a magnitude 8. 00:31:03.704 --> 00:31:09.611 The slip deficit that accumulated since 1938 was roughly 5 to 5-1/2 meters. 00:31:09.611 --> 00:31:17.416 So, if you take this 65 millimeters a year, you multiply it by 2022 minus 1938, 00:31:17.416 --> 00:31:19.824 [chuckles] so, you know, you do that subtraction in your head, 00:31:19.824 --> 00:31:27.383 and you get about 5.4 meters. Okay, so this may be normal to everybody. 00:31:27.383 --> 00:31:29.883 Everybody may look at this and say, well, what’s surprising about that? 00:31:29.883 --> 00:31:34.885 But it turns out, in many of our models we looked at, for an earthquake 00:31:34.885 --> 00:31:37.954 that’s smaller than about 8.5, we actually would have expected 00:31:37.954 --> 00:31:39.683 a little lower peak slip. 00:31:39.683 --> 00:31:45.385 So this was actually surprisingly large peak slip accounting for most 00:31:45.385 --> 00:31:49.447 of the slip deficit accumulated since 1938. 00:31:49.447 --> 00:31:53.373 And this is only from doing those models I showed you earlier when we were 00:31:53.373 --> 00:31:58.444 looking at the kinematics of coupling and what happens when you unlock them. 00:31:58.444 --> 00:32:00.963 So we decided to dig in this a little bit deeper. 00:32:00.963 --> 00:32:05.364 We took basically the same types of coupling models as before. 00:32:05.364 --> 00:32:10.416 So we’ve got our Shumagin gap region in here that’s uncoupled. 00:32:10.416 --> 00:32:15.724 We’ve got the 1938 asperity region in here – the red area that is coupled. 00:32:15.724 --> 00:32:20.198 And all we’re going to do here is run this model interseismically, 00:32:20.198 --> 00:32:24.864 then we’re going to unlock the western edge of the 1938 asperity region. 00:32:24.864 --> 00:32:30.663 And, when we unlock it, it’s going to slip in response to all these stresses that have 00:32:30.663 --> 00:32:36.315 accumulated in the plates around it. Okay, so it’s going to rebound elastically. 00:32:36.315 --> 00:32:37.963 And we’re going to run two models. 00:32:37.963 --> 00:32:42.908 We’re going to assume, first, that there is a coupled asperity in this Shumagin gap. 00:32:42.908 --> 00:32:46.573 Okay, and we’re going to run another model where there is no coupling 00:32:46.573 --> 00:32:50.104 in the Shumagin gap. Okay, so let’s take a look at this. 00:32:50.104 --> 00:32:53.314 Now we’re looking straight down again on the megathrust. 00:32:53.314 --> 00:32:57.163 We’ve got our trench down here at the bottom now. 00:32:57.163 --> 00:33:00.903 We’ve got the deep part of the megathrust up here. 00:33:00.903 --> 00:33:07.369 So we see this is probably at about 125 kilometers’ depth over here. 00:33:07.369 --> 00:33:12.383 So, if we unlock this region of the 1938 – this is the western part 00:33:12.383 --> 00:33:17.573 of the 1938 asperity, we have a Shumagin gap region in here that in fact is 00:33:17.573 --> 00:33:22.574 coupled – okay, so if we lock it, that actually reduces the maximum 00:33:22.574 --> 00:33:27.963 potential slip magnitude in the 2021 earthquake to about 2-1/2 meters. 00:33:27.963 --> 00:33:32.633 Okay, why does this happen? It’s the same reason we get that pseudo coupling effect. 00:33:32.633 --> 00:33:37.473 Locked zones restrict slip on the plate interface around them from occurring. 00:33:37.473 --> 00:33:42.314 So, if you have a locked zone on the east side and a locked zone on the west side. 00:33:42.314 --> 00:33:48.033 They both hold back the plate interface, reducing the amount of slip that occurs. 00:33:48.033 --> 00:33:52.393 Okay, so now what happens if the Shumagin gap is uncoupled? 00:33:52.393 --> 00:33:57.073 Well, perhaps unsurprisingly, when this area is unlocked, you can get 00:33:57.073 --> 00:34:01.886 a lot more slip, particularly on the side of the rupture 00:34:01.886 --> 00:34:04.194 closest to the unlocked zone. Because that’s free. 00:34:04.194 --> 00:34:07.362 It’s free to go and slip however much it wants. 00:34:07.362 --> 00:34:11.354 Okay, but max slip now is 4-1/2 meters. 00:34:11.354 --> 00:34:15.636 If we, in fact, add up all this moment here, it’s equivalent to about 00:34:15.636 --> 00:34:22.184 a magnitude 8.1, okay, for the amount of relative motion since 1938. 00:34:22.184 --> 00:34:26.654 So this is looking fairly similar to the observations. 00:34:26.654 --> 00:34:30.744 Just to remind you, since 1938, there were 5.4 meters of convergence. 00:34:30.744 --> 00:34:34.074 We ran this model with 5.4 meters of convergence. 00:34:34.074 --> 00:34:39.324 And the earthquake recovered most of this motion as we see in this model here 00:34:39.324 --> 00:34:44.244 with an unlocked Shumagin gap. Okay, so this seems like a nice story. 00:34:44.244 --> 00:34:47.449 Is there maybe any other evidence that we’ve got close to 00:34:47.449 --> 00:34:52.134 full strain release from this event? 00:34:52.134 --> 00:34:54.954 We looked a little bit at the aftershocks here. 00:34:54.954 --> 00:34:56.784 And maybe they can tell us something. 00:34:56.784 --> 00:35:01.684 So the aftershocks do show this pretty typical decay over time. 00:35:01.684 --> 00:35:02.834 Not really surprising. 00:35:02.834 --> 00:35:06.511 But, when we looked at the magnitudes of the aftershocks, or particularly 00:35:06.511 --> 00:35:10.814 the cumulative moment over time, we noticed that the aftershocks 00:35:10.814 --> 00:35:15.074 had kind of low magnitudes compared to the main shock. 00:35:15.074 --> 00:35:18.924 Not, you know, sort of egregiously low, but somewhat low. 00:35:18.924 --> 00:35:23.441 So we decided to quantify this a little more precisely. 00:35:23.441 --> 00:35:28.594 We found that the aftershock sequences of the megathrust events, in this case, 00:35:28.594 --> 00:35:31.844 both the 7.8 and the 8.2, actually, only released 00:35:31.844 --> 00:35:35.154 about 1% of the main shock’s seismic moment. 00:35:35.154 --> 00:35:39.424 Now, this is lower than is typically seen for thrust-faulting aftershock sequences, 00:35:39.424 --> 00:35:44.264 which generally release about 5% of the main shock moment. 00:35:44.264 --> 00:35:45.994 So what is this telling us? 00:35:45.994 --> 00:35:50.816 This seems like we are getting [audio cuts out] … 00:35:50.816 --> 00:35:58.808 [silence] 00:35:58.808 --> 00:36:02.524 … relatively larger aftershocks occurring. 00:36:03.480 --> 00:36:08.887 Okay, so the last thing we want to do is probably the hardest part of 00:36:08.887 --> 00:36:12.824 subduction zone science in general, and that’s to answer the question, 00:36:12.824 --> 00:36:15.542 what the heck is happening up-dip? 00:36:15.542 --> 00:36:20.014 Okay, this is a notoriously difficult place to observe because it’s underwater. 00:36:20.014 --> 00:36:22.384 Our observations are typically onshore. 00:36:22.384 --> 00:36:28.527 So we often have 100 or more kilometers between the nearest observation 00:36:28.527 --> 00:36:32.164 and the action going on up here. 00:36:32.164 --> 00:36:37.574 There are a lot of great marine geophysics, marine geology with cores 00:36:37.574 --> 00:36:45.234 and things like that, but it’s hard to get sort of a time span of observations 00:36:45.234 --> 00:36:47.784 in the shallow part of the subduction zone. 00:36:47.784 --> 00:36:51.887 There are some new observation techniques that are a little expensive, 00:36:51.887 --> 00:36:53.984 but it sounds like some of them are just getting funded, 00:36:53.984 --> 00:36:58.164 even in the last few days, to do GPS acoustic and things like that. 00:36:58.164 --> 00:37:00.384 So those data sets are going to be really, really valuable 00:37:00.384 --> 00:37:03.035 for answering these kinds of questions. 00:37:03.035 --> 00:37:07.454 But, in any case, when I did this study, and when I put this talk together, 00:37:07.454 --> 00:37:10.637 there really isn’t a lot of information about the shallow part 00:37:10.637 --> 00:37:13.204 of the plate interface. And this is really important. 00:37:13.204 --> 00:37:17.566 This is the part where tsunamis that are generated are going to be the largest. 00:37:17.566 --> 00:37:22.434 If we have slip in the shallow part, that’s going to move the seafloor the most. 00:37:22.434 --> 00:37:25.856 So we want to know what’s happening up here in terms of its coupling state, 00:37:25.856 --> 00:37:28.754 potential for large earthquakes, potential for slip, etc. 00:37:28.754 --> 00:37:33.325 Okay, so, in the 1938 asperity region, we’ve actually seen 00:37:33.325 --> 00:37:37.504 relatively little slip occurring in 2020 and 2021. 00:37:37.504 --> 00:37:38.774 Why do we think that’s the case? 00:37:38.774 --> 00:37:41.154 Well, we’ve got these slip models here. 00:37:41.154 --> 00:37:45.512 These are based on geodetic data but also on seismic data, 00:37:45.512 --> 00:37:47.884 which are showing relatively little slip up here. 00:37:47.884 --> 00:37:51.637 The tsunami is consistent with the earthquake being buried 00:37:51.637 --> 00:37:54.244 relatively deep on the plate interface. 00:37:54.244 --> 00:37:57.354 So also very few aftershocks in this shallow region too. 00:37:57.354 --> 00:38:01.274 So very low coseismic slip occurring here. 00:38:01.274 --> 00:38:05.512 As I said, it’s just really hard to resolve the coupling state here on 00:38:05.512 --> 00:38:09.494 the shallow megathrust, whether it’s locked and building up to the next earthquake, 00:38:09.494 --> 00:38:12.575 whether it’s unlocked, but maybe still has slip deficit because of 00:38:12.575 --> 00:38:16.164 our pseudo coupling story, or something else entirely. 00:38:16.164 --> 00:38:19.124 So we’re actually going to go back to our locking models. 00:38:19.124 --> 00:38:23.473 And this is kind of the best we can do right now is sort of model 00:38:23.473 --> 00:38:28.707 what’s going on here and say, hey, is this consistent with other observations? 00:38:28.707 --> 00:38:31.954 Okay, so we took the locking model again. 00:38:31.954 --> 00:38:34.964 We started with an unlocked Shumagin gap. 00:38:34.964 --> 00:38:39.284 Because, again, this seems to be most consistent with 00:38:39.284 --> 00:38:45.734 the slip occurring in 2021 when we unlocked this region. 00:38:45.734 --> 00:38:49.584 And now we said, okay, let’s lock the shallow plate interface. 00:38:49.584 --> 00:38:52.704 So this is the model I showed you before with 4-1/2 meters of slip. 00:38:52.704 --> 00:38:57.144 Now we’re going to lock this part of the plate boundary. 00:38:57.144 --> 00:39:01.293 And we see this has the exact same effect as locking the Shumagin gap. 00:39:01.293 --> 00:39:04.414 It changes the spatial pattern, but the general effect is 00:39:04.414 --> 00:39:07.674 reducing the magnitude of coseismic slip, okay? 00:39:07.674 --> 00:39:14.535 So the magnitude equivalent of this slip distribution is about magnitude 7.7. 00:39:14.535 --> 00:39:21.645 Okay? So it seems like locking the shallow plate interface mechanically completely 00:39:21.645 --> 00:39:27.145 is not consistent with what was observed in the 2021 earthquake. 00:39:27.145 --> 00:39:30.404 We would have seen a smaller earthquake occur then. 00:39:30.404 --> 00:39:32.814 Unless, of course, it broke that part. 00:39:32.814 --> 00:39:38.514 Okay, so now, you know, what happens if this area is unlocked? 00:39:38.514 --> 00:39:41.263 Now, in the original model, we actually said, well, 00:39:41.263 --> 00:39:46.604 we should slip this shallow part, okay? But what if it’s unlocked interseismically, 00:39:46.604 --> 00:39:49.624 but it’s velocity-strengthening? 00:39:49.624 --> 00:39:54.494 At the high strain rates of an earthquake, it actually kind of seizes up. 00:39:54.494 --> 00:39:57.524 It doesn’t slip coseismically. 00:39:57.524 --> 00:40:00.674 This is maybe not surprising to some people here. 00:40:00.674 --> 00:40:06.051 There are materials that do this – that get stronger as slip rates increase. 00:40:06.051 --> 00:40:12.036 Okay, so this area is unlocked interseismically but locks up seismically. 00:40:12.036 --> 00:40:16.674 When that occurs, actually we get significant slip occurring 00:40:16.674 --> 00:40:22.700 still in the down-dip part of the rupture, the part that went off in 2021. 00:40:22.700 --> 00:40:25.731 Okay, in fact, the peak slip gets even higher. 00:40:25.731 --> 00:40:30.325 I have to probe this model a little bit more to see why that’s the case, 00:40:30.325 --> 00:40:37.356 but it actually seems to focus slip down at the western end of the rupture here. 00:40:37.356 --> 00:40:40.154 Okay, so that’s kind of an interesting thing. 00:40:40.154 --> 00:40:43.653 But then we sort of push the question a little bit further. 00:40:43.653 --> 00:40:49.466 Okay, there’s no coseismic slip, but this area is unlocked interseismically. 00:40:49.466 --> 00:40:53.154 So, at some point, it seems like this area is going to slip. 00:40:53.154 --> 00:40:56.872 Does it do it shortly thereafter? Is it aseismic? 00:40:56.872 --> 00:41:00.034 Is it going to be rapid enough to produce a tsunami? 00:41:00.034 --> 00:41:03.824 These are some really important questions that I don’t actually have the answers to 00:41:03.824 --> 00:41:08.194 right now but is part of the thing that I’m working on 00:41:08.194 --> 00:41:13.106 moving forward with these kinds of subduction zones. 00:41:13.106 --> 00:41:17.574 Okay, so just to kind of wrap up a little bit, and here’s the sort of picture. 00:41:17.574 --> 00:41:24.263 You can see the July 2021 earthquake has sort of been just plastered on top 00:41:24.263 --> 00:41:27.434 of this picture. We’ve got the subduction zone here. 00:41:27.434 --> 00:41:32.076 We’ve got coupling in the area that produced the 1938 earthquake but also 00:41:32.076 --> 00:41:35.466 now the 2020 and 2021 megathrust earthquakes. 00:41:35.466 --> 00:41:40.414 There’s a transition to an uncoupled Shumagin gap here, okay? 00:41:40.414 --> 00:41:45.384 This produces all sorts of things that we’ve seen in these recent sequences. 00:41:45.384 --> 00:41:50.564 It produced this intraplate trench-perpendicular strike-slip fault. 00:41:50.564 --> 00:41:54.443 Okay, which had a pretty significant [audio cuts out] … 00:41:54.443 --> 00:41:59.224 [silence] 00:41:59.224 --> 00:42:04.254 … aseismic. We see that, in the 2020 and 2021 megathrust earthquakes, 00:42:04.254 --> 00:42:06.896 there was near-complete strain release. 00:42:06.896 --> 00:42:10.074 Okay? We see this from the amount of slip relative to the slip deficit. 00:42:10.074 --> 00:42:13.354 We also see this in the aftershocks. 00:42:13.354 --> 00:42:22.004 So our models and our observations of these events are suggesting this poorly coupled 00:42:22.004 --> 00:42:29.114 Shumagin gap doesn’t slip by itself as a megathrust earthquake, 00:42:29.114 --> 00:42:31.844 but can cause all other kinds of seismicity. 00:42:31.844 --> 00:42:33.744 And then, of course, the shallow part, 00:42:33.744 --> 00:42:38.614 our models are suggesting that this area is not coupled interseismically, 00:42:38.614 --> 00:42:41.584 but also isn’t slipping coseismically. 00:42:41.584 --> 00:42:43.854 So a few questions we have. 00:42:43.854 --> 00:42:48.644 As I said, I totally just said 1946, I’m not going to deal with that. 00:42:48.644 --> 00:42:53.144 But there was a large earthquake with a very large tsunami in 1946. 00:42:53.144 --> 00:42:57.683 So is it an asperity that we can’t resolve because it’s shallow? 00:42:57.683 --> 00:42:59.604 Or is something else going on there? 00:42:59.604 --> 00:43:04.244 Is it some different kind of plate interface rheology? 00:43:04.244 --> 00:43:06.184 When does slip occur near the trench, right? 00:43:06.184 --> 00:43:11.154 I said, you know, there’s this zone up here that should be unlocked at this point. 00:43:11.154 --> 00:43:15.094 Is it slowly, gradually releasing its slip deficit? 00:43:15.094 --> 00:43:20.076 Or is it holding on for some reason and will get maybe a tsunami earthquake 00:43:20.076 --> 00:43:22.670 or a big, shallow earthquake, or something like that? 00:43:22.670 --> 00:43:25.984 I don’t know. I don’t have the answer to that right now. 00:43:25.984 --> 00:43:29.701 And then we might be looking at the seismic signals, 00:43:29.701 --> 00:43:32.234 or maybe the geologic records of this coupling transition. 00:43:32.234 --> 00:43:35.920 Now that we’ve seen some of the complexity that can occur here, 00:43:35.921 --> 00:43:38.524 maybe we can start looking in a little more detail 00:43:38.524 --> 00:43:42.874 at some of the islands over here – the Shumagin Islands over here. 00:43:42.874 --> 00:43:46.639 Just to remind you, you know, there’s this sequence of islands over here 00:43:46.639 --> 00:43:52.232 that might have records of this transition from coupled to uncoupled. 00:43:52.232 --> 00:43:56.834 Okay, so that’s about 45 minutes. Thank you so much for having me. 00:43:56.834 --> 00:43:59.077 I’m going to leave my conclusion slide up, 00:43:59.077 --> 00:44:02.639 and I am happy to take any questions you might have. 00:44:05.004 --> 00:44:13.131 [silence] 00:44:13.131 --> 00:44:16.024 - Thanks so much, Matt. That was really awesome. 00:44:16.024 --> 00:44:18.164 We’re going to take questions. 00:44:18.164 --> 00:44:23.202 So, as usual, you can put your question in the chat or you can raise your hand 00:44:23.202 --> 00:44:24.664 and we’ll call on you. 00:44:24.664 --> 00:44:26.814 And we already have a bunch of questions in the chat and a bunch of raised hands. 00:44:26.814 --> 00:44:28.184 I’m going to start off with the one 00:44:28.184 --> 00:44:31.214 that was put in the chat during your talk from Natalie Culhane. 00:44:31.214 --> 00:44:36.064 It says, what type of information about the plates did you need to put into the model 00:44:36.064 --> 00:44:40.434 to output the resultant slip magnitude, lithology, material strength, etc.? 00:44:40.434 --> 00:44:42.534 - Yeah. Good question. 00:44:42.534 --> 00:44:51.404 It turns out – I’m just going to go back to just this model for now. 00:44:51.404 --> 00:44:55.756 So this model here is actually purely elastic. 00:44:55.756 --> 00:45:00.204 So, if we look at sort of the view here, the upper plate and 00:45:00.204 --> 00:45:05.534 the subducting plate are identical, purely elastic materials. 00:45:05.534 --> 00:45:11.434 Now, we gave them a sort of realistic shear modulus. 00:45:11.434 --> 00:45:15.304 So the stresses here would be scaled appropriately, 00:45:15.304 --> 00:45:19.504 but these stresses would scale linearly to the shear modulus. 00:45:19.504 --> 00:45:24.154 We have tried models where we contrast the slab in the upper plate. 00:45:24.154 --> 00:45:29.494 And again, it sort of scales the stresses, but not the patterns of what’s going on. 00:45:29.494 --> 00:45:36.600 This is a pretty robust result irrespective of the specific properties we use. 00:45:36.600 --> 00:45:41.889 Now, where it gets interesting is when you put in a little more rheological realism. 00:45:41.889 --> 00:45:49.764 So, if you have an upper plate that’s relatively thin over a viscoelastic mantle, 00:45:49.764 --> 00:45:55.144 again, this is sort of shallow enough that it’s just elastic interactions. 00:45:55.144 --> 00:45:59.314 But there’s some changes to the details of what’s going on. 00:45:59.314 --> 00:46:04.202 So, because this is in mostly the elastic part of the system, 00:46:04.202 --> 00:46:09.631 treating it as a purely elastic problem has been pretty robust. 00:46:11.561 --> 00:46:16.522 - Thanks. Awesome. I’m going to take a raised hand. 00:46:16.522 --> 00:46:20.674 Let’s go with Ben Brooks since you’re already here. 00:46:21.608 --> 00:46:23.494 - Hi. Can you hear me? 00:46:23.494 --> 00:46:24.866 - Yes, I can. 00:46:24.866 --> 00:46:28.244 - Great. Thanks for a great talk, Matt. 00:46:28.244 --> 00:46:30.890 And I’ve enjoyed very much reading your papers on the topic 00:46:30.890 --> 00:46:33.600 over the past couple years. - Thanks. 00:46:33.600 --> 00:46:37.424 - Kind of a comment and a – and a related question. 00:46:37.424 --> 00:46:41.424 With respect to the Chignik earthquake, at least, some of the work that 00:46:41.424 --> 00:46:45.204 we’ve been doing recently – well, it’s important in terms of 00:46:45.204 --> 00:46:50.249 the coseismic slip models, at least the ones that are published, all share 00:46:50.249 --> 00:46:55.974 the exact same modeling methodology – the Chen Xi modeling methodology. 00:46:55.974 --> 00:46:59.327 And they all result in high peak slip values. 00:46:59.327 --> 00:47:02.744 That’s very much not required by the data. [chuckles] 00:47:02.744 --> 00:47:06.203 So there can absolutely be – if you – if you take a model that’s 00:47:06.203 --> 00:47:10.364 a little bit more harmonious with the data, these models tend to have 00:47:10.364 --> 00:47:13.405 16-kilometer- by-16-kilometer patch sizes. - Right. 00:47:13.405 --> 00:47:17.814 - Which is not necessarily justified by the very, very sporadic data. 00:47:17.814 --> 00:47:23.034 It leads to, I think, 1,000 unknowns and 45 knowns. [laughs] 00:47:23.034 --> 00:47:24.964 And there’s lots of smoothing, as you’re aware. 00:47:24.964 --> 00:47:27.434 So that’s kind of a comment. 00:47:27.434 --> 00:47:31.515 And then, related is that some of the modeling that we’ve been doing 00:47:31.515 --> 00:47:34.453 shows that you could have 2 to 3 meters’ peak slip, no problem, 00:47:34.454 --> 00:47:36.664 and satisfy the data just as well. 00:47:36.664 --> 00:47:41.284 So that leads into the question, then, how would you think about Chignik 00:47:41.284 --> 00:47:44.890 if it were a 3-meter peak slip event? 00:47:44.890 --> 00:47:47.544 - Yeah. I think that’s a really great question, 00:47:47.544 --> 00:47:50.364 and it’s something I’m talked with Gavin Hayes about as well. 00:47:50.364 --> 00:47:55.132 He said, oh, peak slip is just sub-fault size-dependent. 00:47:55.132 --> 00:47:58.904 And so, you know, he’s, like, never believe peak slip. 00:47:58.904 --> 00:48:03.265 Like, okay, well, I got to use it somehow or – you know, I don’t have to, 00:48:03.265 --> 00:48:05.994 but I like to use it because it’s there in the model. 00:48:05.994 --> 00:48:08.491 Otherwise, don’t do these models. 00:48:08.491 --> 00:48:13.224 [laughs] But, in any case, yeah, if it’s lower peak slip, you know, 00:48:13.224 --> 00:48:15.204 that’s something we have to think about a little bit. 00:48:15.204 --> 00:48:22.084 But I would say it’s still probably been – there’s other, I think, better evidence 00:48:22.084 --> 00:48:26.054 for the Shumagin gap being uncoupled, but then we have to start considering 00:48:26.054 --> 00:48:31.324 the fact that maybe the shallow part has an asperity, too, 00:48:31.324 --> 00:48:33.844 reducing the amount of slip on the plate interface. 00:48:33.844 --> 00:48:38.694 There actually are some details here that I didn’t get into involved with multiple 00:48:38.694 --> 00:48:42.144 earthquake cycles at the base of the seismogenic zone. 00:48:42.144 --> 00:48:45.953 And you see these in some of the more, I would say, dynamic earthquake cycle 00:48:45.953 --> 00:48:51.289 models that are dealing with sort of, you know, rupture processes 00:48:51.289 --> 00:48:55.874 and sort of frictional properties in a way we’re not doing. 00:48:55.874 --> 00:48:59.054 I’ve got the interfaces locked or unlocked. 00:48:59.054 --> 00:49:04.891 There’s some things that can sort of change particularly the down-dip magnitude 00:49:04.891 --> 00:49:07.484 of slip that I just – I can’t resolve in these 00:49:07.484 --> 00:49:09.742 relatively simple models. 00:49:09.742 --> 00:49:13.944 But, if I were being sort of naive about it, [chuckles] in a simplistic way, 00:49:13.944 --> 00:49:16.204 I would say, if it were lower slip, 00:49:16.204 --> 00:49:22.014 that would imply there has to be something locked elsewhere holding it back. 00:49:22.014 --> 00:49:26.766 Where can you put that locked zone? Probably up-dip of the rupture. 00:49:28.129 --> 00:49:30.641 - Cool. Thanks. - Yeah. 00:49:34.874 --> 00:49:37.219 - We can go to another question in the chat. 00:49:37.219 --> 00:49:40.039 And I think you already touched on this in your answer to Natalie’s question, 00:49:40.039 --> 00:49:44.054 but Ruth Harris wrote, hi, Matt, I’m enjoying your really good talk. 00:49:44.054 --> 00:49:46.890 I realize that the models you’re using are simple – my favorite, 00:49:46.890 --> 00:49:50.453 but I’m wondering if pore pressure variations or other unmodeled geophysical 00:49:50.453 --> 00:49:54.109 or geological parameters might be affecting the earthquake rupture locations. 00:49:54.109 --> 00:49:57.844 - Yeah, absolutely. I mean, that’s totally the case. 00:49:57.844 --> 00:50:01.266 So, as I’ve just mentioned in my response to Ben, 00:50:01.266 --> 00:50:05.204 we treat the plate interface as being super simple. 00:50:05.204 --> 00:50:09.784 We treat these stresses in the plates as being, you know, purely elastic. 00:50:09.784 --> 00:50:14.375 So there are all sorts of complexities here that could 00:50:14.375 --> 00:50:17.617 change the details of what’s going on. 00:50:17.617 --> 00:50:22.204 I have not played with poroelasticity at all, like, in my entire career. 00:50:22.204 --> 00:50:26.574 So I have no intuition for how that would change things here. 00:50:26.574 --> 00:50:35.664 My speculation is that it would change things more in detail than in essence. 00:50:35.664 --> 00:50:38.514 But that is pure speculation on my part. 00:50:38.514 --> 00:50:43.084 I honestly, as I said, don’t have good insight there. 00:50:43.084 --> 00:50:49.574 So I know that, you know, as – there is – you know, there’s processes by which 00:50:49.574 --> 00:50:53.664 the plate interface can strengthen and plate interface can weaken. 00:50:53.664 --> 00:50:58.204 That can change things, but it seems like these locked/unlocked models 00:50:58.204 --> 00:51:03.766 are doing a pretty good job of capturing, as I said, the essence of what’s going on. 00:51:03.766 --> 00:51:09.579 But, yeah, unfortunately I don’t have a much better answer for you than that. 00:51:11.547 --> 00:51:13.703 - That sounds good. 00:51:13.703 --> 00:51:15.141 Ruth says thanks. 00:51:15.141 --> 00:51:16.854 - You’re welcome. 00:51:17.594 --> 00:51:23.284 - So Alex Hatem, you want to ask a question? 00:51:24.649 --> 00:51:27.234 - Hey, Matt. Really awesome talk. Thank you. 00:51:27.234 --> 00:51:30.784 So I have a question about the strike-slip faulting that you talked about 00:51:30.784 --> 00:51:34.614 with the magnitude 7.6 event. 00:51:34.614 --> 00:51:37.642 And this is probably an unanswerable question, but it got me thinking 00:51:37.642 --> 00:51:42.224 during your talk. So you pointed out really clearly with your Coulomb modeling 00:51:42.224 --> 00:51:46.004 that the right-lateral strike-slip faulting is totally encouraged by 00:51:46.004 --> 00:51:49.234 having that locked patch and unlocked patch. 00:51:49.234 --> 00:51:52.504 And I guess I’m wondering, what is the chicken and the egg here? 00:51:52.504 --> 00:51:59.294 Is the strike-slip fault causing that locked fault – or, locked patch boundary? 00:51:59.294 --> 00:52:03.754 Or is the strike-slip fault forming because of that locked patch? 00:52:03.754 --> 00:52:07.446 And how are you thinking about that? 00:52:07.446 --> 00:52:09.164 - Yeah. 00:52:09.164 --> 00:52:14.514 So I am of the opinion – although I don’t have a great – 00:52:14.514 --> 00:52:18.664 you know, I don’t have great evidence for it – I’m of the opinion that 00:52:18.664 --> 00:52:27.754 the transition in coupling is kind of what exists and is activating these faults. 00:52:27.754 --> 00:52:33.517 That it’s not something specifically about the fault that is causing the 00:52:33.517 --> 00:52:36.084 transition in coupling. And one of the reasons – 00:52:36.084 --> 00:52:38.524 actually, no, maybe there is a little bit of evidence for this. 00:52:38.524 --> 00:52:44.142 I don’t have the mechanism on here unfortunately, but it turns out this event 00:52:44.142 --> 00:52:47.914 is actually the same mechanism as this event. 00:52:47.914 --> 00:52:49.864 It is – this is not a thrust-faulting event. 00:52:49.864 --> 00:52:54.624 This is an inter-slab, nearly trench-perpendicular, strike-slip event. 00:52:54.624 --> 00:52:57.324 So you’re looking at this. You’re, like, wait, wait, wait, wait. 00:52:57.324 --> 00:53:00.224 Why is it over here in the coupled zone? 00:53:00.224 --> 00:53:05.540 But it kind of turns out, if you look at our – if you go back here, 00:53:05.540 --> 00:53:11.705 and you look at these models, this zone over here is strongly favored 00:53:11.705 --> 00:53:18.251 with right-lateral strike-slip, down-dip – at the down-dip edge of the locked zone. 00:53:18.251 --> 00:53:22.774 Now, we haven’t looked at Coulomb stress changes from the magnitude 8.2 main shock 00:53:22.774 --> 00:53:26.814 or anything like that to see if it, in fact, is favored. 00:53:26.814 --> 00:53:30.142 But we think that the coupling is more fundamental, 00:53:30.142 --> 00:53:34.334 and it’s these faults probably – as I said, I think it’s on fracture zones, 00:53:34.334 --> 00:53:39.744 but it could be something else, are being activated in that 00:53:39.744 --> 00:53:46.267 sort of stress regime setup by the coupling transition. 00:53:46.924 --> 00:53:49.234 - So a follow-on, then. 00:53:49.234 --> 00:53:53.464 Does that, then, help you determine that these patches are locked in the long-term? 00:53:53.464 --> 00:53:55.344 Like, these are always been locked. 00:53:55.344 --> 00:53:59.955 If it’s an – if it’s locked enough to cause a strike-slip fault to grow 00:53:59.955 --> 00:54:03.282 and to have these really big earthquakes, it’s … 00:54:03.282 --> 00:54:04.189 - Yeah. 00:54:04.189 --> 00:54:07.474 - It means that this patch has been locked for a really long time in our history. 00:54:07.474 --> 00:54:09.744 - Right. Well, and then that’s the question, right? 00:54:09.744 --> 00:54:13.224 Is this activating a big structure that exists already? 00:54:13.224 --> 00:54:15.259 That’s my opinion. 00:54:15.259 --> 00:54:18.624 If it were producing these big structures, yeah, then I’d expect 00:54:18.624 --> 00:54:22.625 to see more surface signal of it. Now, [chuckles] unfortunately, 00:54:22.625 --> 00:54:24.384 these earthquakes are happening in the slab. 00:54:24.384 --> 00:54:26.939 If these were happening in the upper plate, it’d be, like, oh, yeah, 00:54:26.939 --> 00:54:30.154 probably could see them really easily. 00:54:30.154 --> 00:54:34.204 We might see a complementary signal in the upper plate. 00:54:34.204 --> 00:54:38.611 Yeah. I think, you know, there’s no reason to expect why it – 00:54:38.611 --> 00:54:41.364 so, if we go – well, we can just look at this. 00:54:41.364 --> 00:54:44.214 If we drew the arrows for the upper plate relative to the subducting plate, 00:54:44.214 --> 00:54:48.654 they’d be exactly the opposite. We’d see left-lateral shearing, okay? 00:54:48.654 --> 00:54:54.374 So maybe there is, you know, a periodic left-lateral 00:54:54.374 --> 00:54:57.714 big earthquake going on in here. 00:54:57.714 --> 00:55:02.044 Would it be big enough to produce a geomorphic signal? 00:55:02.044 --> 00:55:06.518 Would it be big enough to exceed the signal that we get from 00:55:06.518 --> 00:55:10.074 subduction zone coupling – vertical motions and things like that there? 00:55:10.074 --> 00:55:11.923 I don’t know. 00:55:11.923 --> 00:55:16.393 It’s very likely, in my mind, that the type of earthquake we had in 00:55:16.393 --> 00:55:20.194 October, that strike-slip earthquake, doesn’t always happen 00:55:20.194 --> 00:55:23.877 throughout the megathrust earthquake cycle. 00:55:23.877 --> 00:55:26.564 That’s why I think the triggering story is kind of important. 00:55:26.564 --> 00:55:34.064 It was this right set of circumstances that allowed this big event to happen. 00:55:34.064 --> 00:55:37.955 But, you know, again, this is – this is all in the realm of 00:55:37.955 --> 00:55:41.334 sort of [laughs] speculation at this point. I don’t have a good argument. 00:55:41.334 --> 00:55:43.984 Remember, all these are also happening offshore too. 00:55:43.984 --> 00:55:50.580 So, as we get more onshore, where we might be able to see these things, 00:55:50.580 --> 00:55:55.044 the likelihood, I think, of getting a trench-perpendicular strike-slip earthquake 00:55:55.044 --> 00:55:58.518 in the coupling transition, it’s just less likely, for the same reason 00:55:58.518 --> 00:56:03.713 we can’t observe the offshore stuff. That’s where the coupling action is. 00:56:03.713 --> 00:56:07.404 - Cool. Thank you. - Yeah. Thanks for the question. 00:56:09.721 --> 00:56:12.205 - Tom, you want to go ahead? 00:56:18.353 --> 00:56:22.963 - Yes. Could you go to the slide where you show the aftershocks and, 00:56:22.963 --> 00:56:27.354 in particular, the yellow earthquake and its aftershocks? 00:56:27.354 --> 00:56:29.456 - Yeah. - Or your title slide, I think, 00:56:29.456 --> 00:56:31.119 shows it as well. 00:56:31.119 --> 00:56:36.743 Yeah. You know, that’s – it’s one thing, you know, 00:56:36.743 --> 00:56:41.684 where you talked about mainly the – on the onshore, 00:56:41.684 --> 00:56:44.658 where you have these strike-slip mechanisms. 00:56:44.658 --> 00:56:48.742 I guess you could say there’s a – there’s a linear feature there. 00:56:48.783 --> 00:56:55.376 But then, if you propagate it to the – towards the trench and beyond, 00:56:55.377 --> 00:57:01.174 it’s more of an arc than a linear feature. 00:57:01.174 --> 00:57:04.624 And then the mechanisms change … - Yep. 00:57:04.624 --> 00:57:07.418 - … as well. 00:57:08.024 --> 00:57:10.104 And then the question is, 00:57:10.104 --> 00:57:17.143 so how are they associated with each other, and is there anything that you 00:57:17.143 --> 00:57:27.374 might be able to see in the topography of the ocean floor offshore 00:57:27.374 --> 00:57:35.334 of the trench that might indicate a fault or sequence of fault zones? 00:57:35.334 --> 00:57:39.404 - I think that’s a great question. We have been thinking about it. 00:57:39.404 --> 00:57:45.096 We didn’t come up with a convincing answer, even though these sort of normal faulting 00:57:45.096 --> 00:57:49.940 seemingly outer rise events were occurring when we were writing our paper on this. 00:57:49.940 --> 00:57:52.384 That’s why it didn’t make it into the paper. 00:57:52.384 --> 00:57:55.456 But we’ve been looking at these and thinking, okay, well, 00:57:55.456 --> 00:58:00.299 this is not just coincidence that these events are happening here. 00:58:00.299 --> 00:58:04.519 Probably not coincidence that they’re spatially related to what’s going on 00:58:04.519 --> 00:58:06.434 with the strike-slip events. 00:58:06.434 --> 00:58:13.094 Now, I think this gap here is probably about 40 kilometers or so. 00:58:13.094 --> 00:58:15.914 I suspect that these are actually different structures. 00:58:15.914 --> 00:58:19.994 I do not think – although it does appear like it’s a curved structure going 00:58:19.994 --> 00:58:26.664 all the way to the trench, I suspect that there actually is a gap here. 00:58:26.664 --> 00:58:33.594 But apparently whatever has happened with the megathrust, the afterslip, 00:58:33.594 --> 00:58:40.706 and now the oblique strike-slip is also sufficient to trigger activity, 00:58:40.706 --> 00:58:46.073 which I would – I guess I would probably consider outer rise activity. 00:58:46.073 --> 00:58:49.831 And whether that is a consistent relationship or not, 00:58:49.831 --> 00:58:54.644 I don’t have good insight for, but I think it’s a really important observation that, 00:58:54.644 --> 00:59:02.754 you know, these are really only active after the recent events. 00:59:02.754 --> 00:59:06.307 So I don’t think the structures are connected. 00:59:06.307 --> 00:59:11.331 But, you know, it’s possible they are and we just can’t see it. 00:59:13.956 --> 00:59:16.948 - Okay. Thank you. - Yeah. Thank you. 00:59:19.815 --> 00:59:22.255 - We’re just about at the end of the hour, but we have a couple more questions 00:59:22.255 --> 00:59:25.654 in the chat, so I’ll try to read maybe one of them for a start. 00:59:25.654 --> 00:59:29.620 Morgan Page asked, how do we know that the 1938 and 1946 events 00:59:29.620 --> 00:59:32.034 had shallow slip approaching the trench? 00:59:32.034 --> 00:59:36.294 - Ah, yes. Because both of them had substantially larger tsunamis 00:59:36.294 --> 00:59:39.354 than anything we’ve seen recently. 00:59:39.354 --> 00:59:42.832 So, first of all, they’re larger earthquakes, as we can tell 00:59:42.832 --> 00:59:46.714 by their magnitudes and their aftershock areas and things like that. 00:59:46.714 --> 00:59:49.414 But they have substantially larger tsunamis. 00:59:49.414 --> 00:59:54.004 And at places where we can make repeat measurements. 00:59:54.004 --> 01:00:00.207 So we know 1946 – I’m not – again, I’m not a tsunami expert, 01:00:00.207 --> 01:00:02.574 but 1946 is a really anomalous event. 01:00:02.574 --> 01:00:07.174 I think it’s considered one of these slow tsunami earthquakes. 01:00:07.174 --> 01:00:11.614 And then 1938 is just a big earthquake that appears to have ruptured close to 01:00:11.614 --> 01:00:16.294 the trench, again, because of its big tsunami, much larger than 2021. 01:00:17.995 --> 01:00:20.764 - Well, I like that answer. - [laughs] 01:00:21.706 --> 01:00:24.844 - So Nicholas Cunetta wrote, great talk. 01:00:24.844 --> 01:00:28.050 What was the depth of the magnitude 7.6 right-lateral event? 01:00:28.050 --> 01:00:32.034 I’m curious how this coupling transition, slash, oblique earthquake relate to 01:00:32.034 --> 01:00:35.730 subduction accretion, most of which occurs at about 30 kilometers’ depth. 01:00:35.730 --> 01:00:39.956 - Ah, okay. Yeah. So let me show this cross-section here. 01:00:39.956 --> 01:00:47.488 So, yeah, the earthquake occurred just below the megathrust. 01:00:47.488 --> 01:00:50.597 And so I’d put that roughly at 30 kilometers. 01:00:50.597 --> 01:00:53.874 I don’t – I don’t have the depth off the top of my head, but you can see 01:00:53.874 --> 01:00:56.794 in this cross-section here, you know, roughly 30 kilometers. 01:00:56.794 --> 01:01:01.214 The plate interface here is probably at about 25 kilometers. 01:01:01.214 --> 01:01:08.253 And you can see all of these events are happening right below the plate interface. 01:01:08.253 --> 01:01:12.134 So, yeah, how that’s related to subduction accretion, ooh, 01:01:12.134 --> 01:01:14.324 that’s not something I’ve thought about in a long time. [laughs] 01:01:14.324 --> 01:01:17.174 And certainly not in the context of this sequence. 01:01:17.174 --> 01:01:20.243 So I think that is an interesting question. 01:01:21.918 --> 01:01:24.082 - That’s fair. 01:01:24.082 --> 01:01:26.511 You got a lot of thank-yous in the chat. 01:01:26.511 --> 01:01:29.004 Thanks for the enthusiastic talk. 01:01:29.004 --> 01:01:31.954 Clara Yoon has a question in the chat. Thanks for the great talk, Matt. 01:01:31.954 --> 01:01:34.824 What’s your opinion about the earthquake hazard in the Shumagin gap? 01:01:34.824 --> 01:01:37.243 Since you found it to be aseismic, slash, uncoupled, 01:01:37.243 --> 01:01:40.144 how concerned should we be? For example, might a future 01:01:40.144 --> 01:01:43.404 large rupture on either side of the gap extend all the way across the gap? 01:01:43.404 --> 01:01:46.484 Or would the uncoupling boundary reliably stop it from getting 01:01:46.484 --> 01:01:49.124 through the gap? And thinking about dynamic weakening effects. 01:01:49.124 --> 01:01:52.684 - Yeah. I think that’s a great question, and I don’t have 01:01:52.684 --> 01:01:56.564 a great answer for it, unfortunately. 01:01:56.564 --> 01:02:01.801 Our argument is that, if we go back to – oop, wrong direction, sorry. 01:02:01.801 --> 01:02:04.784 Let me maybe scroll this way a little more. 01:02:04.784 --> 01:02:07.332 Find my mouse. There’s my mouse. 01:02:08.114 --> 01:02:12.724 If we go to, say, these models, and we say, how much can any of this slip? 01:02:12.724 --> 01:02:17.614 In our – in the context of our models, anything with slip deficit can slip. 01:02:17.614 --> 01:02:22.374 So, you know, this is not a representation of the Shumagin gap, 01:02:22.374 --> 01:02:27.324 but in these relatively simple models, from the edge of an asperity 01:02:27.324 --> 01:02:32.654 at 600 kilometers along the Y axis, all the way to 700 kilometers on the Y axis 01:02:32.654 --> 01:02:36.824 could have slip. Now, the slip over here would be much less. 01:02:36.824 --> 01:02:42.520 But, you know, I think that we’ve seen some earthquakes with sort of 01:02:42.520 --> 01:02:46.474 very long tails [chuckles] of relatively low slip. 01:02:46.474 --> 01:02:50.604 And maybe, you know, my argument might be those aren’t rupturing asperities. 01:02:50.604 --> 01:02:54.484 Those are rupturing the edges of the pseudo coupled zones. 01:02:54.484 --> 01:03:02.208 You know, I don’t have great reasoning for that, but – other than it seems to 01:03:02.208 --> 01:03:07.520 be consistent with the magnitudes of slip deficit we’re getting in these models. 01:03:07.520 --> 01:03:11.145 But whether and earthquake can specifically cross the Shumagin gap, you know, 01:03:11.145 --> 01:03:13.834 I think there we just have to look at the history 01:03:13.834 --> 01:03:18.184 of earthquakes there – history of the geological evidence. 01:03:18.184 --> 01:03:21.574 And we just don’t see that happen very often. 01:03:21.574 --> 01:03:24.104 So perhaps it can occur. 01:03:24.104 --> 01:03:28.274 Perhaps there are earthquakes that can extend into the Shumagin gap partially. 01:03:28.274 --> 01:03:32.544 But it just doesn’t seem to cross the gap all that often. 01:03:32.544 --> 01:03:37.854 Otherwise we’d probably see evidence of significant tsunamis and vertical motions 01:03:37.854 --> 01:03:41.774 and things like that. So it appears to be a place where, 01:03:41.774 --> 01:03:48.583 although, you know, I completely agree dynamic weakening can be invoked, 01:03:48.583 --> 01:03:52.143 we’re not – it doesn’t seem to be happening much there. 01:03:52.168 --> 01:03:57.443 It seems to be a place that sort of impedes megathrust rupture, generally. 01:03:57.444 --> 01:04:02.770 Now, as far as just seismic hazard there, what we’re seeing from this is 01:04:02.770 --> 01:04:05.714 seismic hazard doesn’t equal megathrust earthquake hazard. 01:04:05.714 --> 01:04:09.934 And I think that’s important to recognize in subduction zones that, 01:04:09.934 --> 01:04:13.458 you know, we tend to – I tend to get this tunnel vision 01:04:13.459 --> 01:04:18.034 where it’s just megathrust earthquake, megathrust tsunamis, but, you know, 01:04:18.034 --> 01:04:22.794 any earthquake that displaces the ocean floor produces a tsunami. 01:04:22.794 --> 01:04:24.964 Any big earthquake produces strong shaking. 01:04:24.964 --> 01:04:29.344 And so those don’t have to be plate interface events, necessarily. 01:04:29.344 --> 01:04:35.146 So I think that’s a really good question with a lot of different answers. [laughs] 01:04:36.911 --> 01:04:39.872 - Okay. Well, thanks so much for the awesome talk. 01:04:39.872 --> 01:04:44.744 Let’s thank Matt one more time and also for giving very thorough awesome answers 01:04:44.744 --> 01:04:47.934 to all of our questions. - Thanks so much for having me. 01:04:47.934 --> 01:04:50.059 This was a lot of fun. 01:04:50.059 --> 01:04:51.630 - Okay. We’re glad you enjoyed. 01:04:51.630 --> 01:04:54.644 We’re going to end the recording since we’re a little over the end of the hour. 01:04:54.644 --> 01:04:56.844 I don’t know if Matt can stick around for a minute or two.