WEBVTT 00:00:04.000 --> 00:00:08.000 Woohoo, so this is our late breaking Ferndale session late in the day. 00:00:08.000 --> 00:00:28.000 Thank you for coming. A really big special thanks to Lori and Bob, and Jay, and Peggy, who, right in the middle of an earthquake, and sometimes not having power or whatever was willing to actually organize a session about that earthquake, so thank you so much. 00:00:28.000 --> 00:00:38.000 And so the format for the session is, we're just gonna go talk, talk, talk, talk, and then at the end we'll stop and have questions about everything, but as always feel free to get into the chat. 00:00:38.000 --> 00:00:44.000 These talks will be live, they will be hosted from John's computer. 00:00:44.000 --> 00:00:48.000 Everyone has already sent their PowerPoints to John, so he's going to run them off of his computer. 00:00:48.000 --> 00:00:49.000 But everyone is going to get remote control of John's computer. 00:00:49.000 --> 00:00:56.000 They will see a little a set of arrows in the bottom left corner. 00:00:56.000 --> 00:01:06.000 You press right arrow to advance the slide, you press left arrow to end a slide, so our speakers are responsible for unmuting themselves and advancing their own slides. 00:01:06.000 --> 00:01:19.000 So unless there is any more questions, thoughts, queries, before we take the show into high gear. 00:01:19.000 --> 00:01:28.000 Okay, then let's do this thing. We're going to start things off with our one and only pre-recorded talk, except maybe Hamid as well. 00:01:28.000 --> 00:01:39.000 Brought to you by Bob McPherson. Woohoo! 00:01:39.000 --> 00:01:50.000 Hello, everyone. My name is Bob McPherson, and I'd like to thank Sarah and Keith, and everyone else on the committee at the USGS for having this workshop. 00:01:50.000 --> 00:02:09.000 Thank you very much. Today, I'd like to show that the 1975 Rio Dell earthquake and the 2022 Ferndale earthquake occurred on the same slip patch within the Gorda plate, and to do this i'll show damage, geographic 00:02:09.000 --> 00:02:20.000 distribution of damage, aftershock trends, hypocentral locations, and focal mechanisms. 00:02:20.000 --> 00:02:25.000 So over here on this photo that I took in 1975 00:02:25.000 --> 00:02:29.000 it shows the Scotia Bluffs by Rio Dell 00:02:29.000 --> 00:02:31.000 that collapsed and wiped out the North Pacific Railroad line shown here at the base of the cliff. 00:02:31.000 --> 00:02:40.000 Here's the trestle, and you can see some rails going down into the Eel River. 00:02:40.000 --> 00:03:02.000 These same cliffs had minor rock fall in 2022, and Rio Dell in this panel here shows that a lot of the chimneys went down in 1975, and seeing in Fortuna, and likewise in 2022 those chimneys, that remain after all these 00:03:02.000 --> 00:03:22.000 earthquakes also suffered damage in both Rio Dell and Fortuna. A large plate glass windows in downtown Fortuna and downtown Rio Dell suffered extensive damage in 1975 and again in 2022. So just in a 00:03:22.000 --> 00:03:26.000 preliminary field excursion I did a day after the earthquake, and I'm sure we'll hear more about it in the workshop that 00:03:26.000 --> 00:03:35.000 the damage seems very similar in both quakes 00:03:35.000 --> 00:03:38.000 If we look at Gorda plate earthquakes, as shown in this map, we made in 1975 00:03:38.000 --> 00:03:45.000 this is a small portion of it. Here's the coastline. 00:03:45.000 --> 00:03:50.000 Here's False Cave, Cape Mendocino, Mendocino fault for reference. 00:03:50.000 --> 00:03:55.000 Here's the 75 sequence. Here's the main shock. 00:03:55.000 --> 00:04:02.000 Notice that the aftershocks trend off to the southwest 00:04:02.000 --> 00:04:10.000 If we look at a focal mechanism in that sequence, the main shock is in the upper hemisphere, that are hand plotted in 1975. 00:04:10.000 --> 00:04:14.000 Here's our choice of the fault, nor 70 East Matching the epicenters. 00:04:14.000 --> 00:04:22.000 And it matches very well with the moment tensor and the focal mechanism of the 2022 earthquake. 00:04:22.000 --> 00:04:27.000 All of these ENE plains are within 5 degrees. 00:04:27.000 --> 00:04:31.000 By the way, these are lower hemisphere projections. 00:04:31.000 --> 00:04:35.000 So the source was the same 00:04:35.000 --> 00:04:41.000 If we look at a cross section west to east vertical 10 to 20, 30 kilometers. 00:04:41.000 --> 00:04:46.000 So here comes the corner. Play down, descending down underneath North America. 00:04:46.000 --> 00:04:51.000 This is one of the first slides showing this. Here's the main shock. 00:04:51.000 --> 00:04:59.000 And again the aftershocks occurring to the west, and all the earthquakes within the upper Gorda Plate. 00:04:59.000 --> 00:05:10.000 If we look at the 2022 sequence here's the main shock rupturing to the northeast longest aftershock trend, I would point out, notice this intense cluster just east of the main shock 00:05:10.000 --> 00:05:27.000 I would say this 12 kilometer long, approximately grouping is where the slip occurred, and both finite fault models that I'm aware of 00:05:27.000 --> 00:05:32.000 the slip occurs directly beneath this mapping of epicenters. 00:05:32.000 --> 00:05:37.000 So now on this slide we'll plot the 1975. 00:05:37.000 --> 00:05:47.000 And here's the 1975 right here. Remember it started here and broke to the west, and the 6.4 started here and broke to the east. 00:05:47.000 --> 00:05:52.000 On both sides of this alleged patch 00:05:52.000 --> 00:06:05.000 If we look at the earthquakes in cross-section shown by this Anthony Lomax slide the earthquakes in red, or the 2022, here's the dense patch within the Gorda Plate. Here's the Gorda Plate 03:59:59.000 --> 04:00:05.00 descending down underneath North America. 00:06:11.000 --> 00:06:18.000 So in summary, the damage, geographic distribution of damage was the same. 00:06:18.000 --> 00:06:20.000 The aftershocks aligned in the same direction. 00:06:20.000 --> 00:06:23.000 The hypocenters were all within the Gorda Plate. 00:06:23.000 --> 00:06:27.000 The main shock were at either end of the patch. 00:06:27.000 --> 00:06:49.000 The mechanisms were identical. So we conclude that the 1975 in the 2022 ruptured the same slip patch within the Gorda Plate. 00:06:49.000 --> 00:06:50.000 Excellent. Thank you so much, Bob. And now we're going to go to our live, 00:06:50.000 --> 00:06:55.000 there are live talks, and start with Jay Patton. 00:06:55.000 --> 00:06:56.000 Please 00:06:56.000 --> 00:06:58.000 take it away, Jack. 00:06:58.000 --> 00:07:04.000 Alright! Thank you for having me here to present the results from our review of field evidence for the 00:07:04.000 --> 00:07:08.000 M6.4 Ferndale earthquake. We use the CGS 00:07:08.000 --> 00:07:11.000 and USGS jointly developed, post earthquake digital observation schema. And this time we enlisted 00:07:11.000 --> 00:07:32.000 a desktop survey of social media posts as input for the database. This is just to remind everyone of the CSV where these oceanic plates subduct where the earthquake cycle causes vertical land motion and we're megathrust earthquakes cause 00:07:32.000 --> 00:07:36.000 tsunamis. But we're gonna focus on the Gorda plate. 00:07:36.000 --> 00:07:47.000 The Gorda plate is formed at the Gorda Ridge, where normal faults are oriented in parallel to the ridge in southern Gorda, in the Mendocino deformation zone known as the triangle of doom as a plate is 00:07:47.000 --> 00:07:58.000 effectively shortens north to south. These faults rotate in a clockwise fashion, and are reactivated as left lateral strike-slip fault as evidenced by these earthquake mechanisms. 00:07:58.000 --> 00:08:05.000 The M6.4 Ferndale earthquake was most likely along one of these left-lateral strike-slip faults. 00:08:05.000 --> 00:08:08.000 Here's a brief view of the geology of the upper plate 00:08:08.000 --> 00:08:17.000 including CDMG and Bob Mclaughlin's mapping, and note the new gene to Quaternary Eel River 00:08:17.000 --> 00:08:25.000 sedimentary basin is folded in a syncline, and the latest Quaternary to modern Eel River fluvial systems is inset 00:08:25.000 --> 00:08:31.000 within this older basin, and these two things might have something to do with the ground motions and the 00:08:31.000 --> 00:08:36.000 M6.4 seismicity trend is oblique to the geologic fabric. 00:08:36.000 --> 00:08:46.000 So these are the mechanisms from many of the earthquakes and aftershocks, and they're all most all either sinistral or normal. 00:08:46.000 --> 00:08:57.000 There's one reverse mechanism, and then M5.4 sequence is dextral, and the next slide will look at seismicity that are plotted along the Gorda 00:08:57.000 --> 00:09:03.000 2021 profile B2B prime, which is represented by that black line. 00:09:03.000 --> 00:09:15.000 So this shows how the hyppcenters plot nicely within the Gorda slab. The M6.4 earthquakes are blue, and the M5.4 earthquakes are green. 00:09:15.000 --> 00:09:19.000 Here, [clears throat] excuse me, here I've outline that maybe three main faults are involved pretty hypothetical. 00:09:19.000 --> 00:09:40.000 Note how the projected in the upper right corner is a projected magnetic anomaly from Wilson, and how that magnetic anomaly is perfectly aligned with the M6.4 seismicity and also note how the M5.4 mechanism is aligned with a 1992 00:09:40.000 --> 00:09:50.000 Cape Mendocino triggered, and M6.6 and M6.5 earthquakes. And so now we'll start with the field observations. 00:09:50.000 --> 00:09:59.000 This map shows the USGS intensity model and M7 in the middle; red dots show social media reports of damage; 00:09:59.000 --> 00:10:12.000 blue triangles show boots on the ground; and in the next slides I'll show where the location is for an observation, and here we see the observation of road-fill failure. 00:10:12.000 --> 00:10:35.000 Here we see along the side of Humboldt Bay. Road buckling. Here we see along the toll road. South of Ferndale, a cutbank failure in the road. And here we see along the road north of Lolita, more cut bank failures and then also from the M5.4 00:10:35.000 --> 00:10:45.000 earthquake. This is the intensity map we saw a landslide triggered by that earthquake. So there are more observations in the database. 00:10:45.000 --> 00:10:47.000 Thank you very much. 00:10:47.000 --> 00:10:52.000 Okay, this is Kevin Furlong. Thanks for putting together this session. 00:10:52.000 --> 00:11:03.000 Today, I want to talk about how the earthquakes that are occurring this cluster in the vicinity of Mendocino Triple Junction, fit into the tectonic setting of the area. 00:11:03.000 --> 00:11:12.000 The key point I want you to take away from this talk is that this dense cluster of earthquakes that we've seen recently and over the past decades occur in the vicinity of 00:11:12.000 --> 00:11:19.000 the Mendocino Triple Junction, and they primarily occur within the overlying North American Franciscan crust. 00:11:19.000 --> 00:11:30.000 So, although they're located near the southern edge of the Cascadia slab, they do not appear to be directly subduction related, but I think are clearly plate tectonics related in the area. 00:11:30.000 --> 00:11:31.000 So what is the tectonic setting of these events? 00:11:31.000 --> 00:11:42.000 Well what I think is pretty obvious is they are not associated with upper plate faulting or known map faults in the upper plate 00:11:42.000 --> 00:11:57.000 they blow right across that. But with some new seismic tomography that, in conjunction with Harley Benz and Antonio Villa Señor, we have put together in the region we can place these into their setting within the crustal and upper mantle setting of the area. Map on the right 00:11:57.000 --> 00:11:58.000 just shows the tomography at a shallow depth which maps surface geology. 00:11:58.000 --> 00:12:14.000 If we go to the depth of these earthquakes, we see on the left the 18 km depth which would be the depth of the 2022 event, the right, about 25 km, steps the depth of the 2021 event we see that in both 00:12:14.000 --> 00:12:31.000 cases. And in this area in general, there is a nose of thickened North American crust that extends out into the Cape Mendocino area that sits on top of the Gorda slab, which has been known for quite a while is bent to the south at its southern edge, and so if we 00:12:31.000 --> 00:12:49.000 zoom in on these two areas, we see that the 2022 event lies basically, has a strike up that is basically parallel with the strike of the boundary between the border and the overlying North American crust. 00:12:49.000 --> 00:13:07.000 We look at the cross-section on the right, and you can see that it sits about 10 km above the Gorda slab, based on the tomography within this thickened North American crossed the left lateral mechanism says that north is to the west, and the south 00:13:07.000 --> 00:13:25.000 is to the east. If we now move to the fault from last year the previous year's event, it is deeper, but it is still within this thickened wedge of North American crust that we see overlying the boundary between the Gorda and Pioneer fragment which is the eastern 00:13:25.000 --> 00:13:26.000 extent of the Pacific plate in the area. 00:13:26.000 --> 00:13:39.000 Again north to the west, south, to the east. So how these faults are produced, and what drives them is sort of the $64,000 question. 00:13:39.000 --> 00:13:55.000 If we look at the situation, what we can say is, we know based on the plate motions that relative to a North American reference frame the Gorda and Pacific plates are moving to the north northwest Allah, the Mendocino Triple Junction and in doing that space is created 00:13:55.000 --> 00:14:14.000 on the south end of the Gorda slab. That would be need to be filled by this thickening North American crust that would lead to left-lateral motions along the boundary between this thickneing wedge of North American crust and the Gorda plate so we think that these are 00:14:14.000 --> 00:14:23.000 upper plate North American centric events associated with the large-scale plate tectonics. Thank you! 00:14:23.000 --> 00:14:26.000 Great. Hi! Everyone! I'm Maggie Ortiz-Milan. 00:14:26.000 --> 00:14:32.000 I'm the director of programs at EERI, and I'll be giving an overview of the Clearinghouse response. 00:14:32.000 --> 00:14:36.000 And then also an overview of the damage from the earthquake 00:14:36.000 --> 00:14:40.000 So EERI serves as the vice chair of the California Earthquake Clearinghouse in the California Geological Survey serves as the chair, and the USGS 00:14:40.000 --> 00:14:53.000 and the California Seismic Safety Commission and the California Office of Governors Office of Emergency Services sit on the Management Committee. 00:14:53.000 --> 00:15:01.000 So for major earthquakes in California, the Clearinghouse activates to help coordinate the scientific and technical response to the earthquakes. 00:15:01.000 --> 00:15:05.000 So before I get into the the Ferndale impacts, they just want to say a few words about EERI. 00:15:05.000 --> 00:15:11.000 If you're not familiar, EERI is a nonprofit membership association dedicated to advancing earthquake resilience. 00:15:11.000 --> 00:15:15.000 So I encourage you to learn more about EERI and and learn some more about some of our membership options. 00:15:15.000 --> 00:15:24.000 If you're interested in joining our community. So for the Ferndale earthquake, this is a timeline of the Clearinghouse activation. 00:15:24.000 --> 00:15:25.000 So the earthquake happened at 2:34 a.m. in the morning. 00:15:25.000 --> 00:15:34.000 So Cindy Pridmore at CGS is the chair, and she was very soon after, in touch with CalOES 00:15:34.000 --> 00:15:43.000 And then around 7 a.m., CGS and EERI I met and decided that a virtual activation of the Clearing House was warranted for this earthquake. 00:15:43.000 --> 00:15:48.000 So that afternoon a website was set up for the earthquake, and we held our first Clearinghouse briefing that evening. 00:15:48.000 --> 00:15:55.000 The next day an update was sent out about the activation with resources that had been posted to the website and notes from that first call. 00:15:55.000 --> 00:16:10.000 We held a second briefing on December 22nd, and then paused a little bit for the holidays, and then all came back together on January 6th, for a third and final briefing, and then the Clearinghouse was deactivated following that third briefing. So a 00:16:10.000 --> 00:16:14.000 Summary of the activation. It was a 10 day virtual clearinghouse activation. 00:16:14.000 --> 00:16:22.000 There were three Clearinghouse briefing calls with over 170 participants, and there were over 2,700 views of the virtual Clearinghouse website. 00:16:22.000 --> 00:16:23.000 So the URL is on the slide. I'll also stick it in the chat. 00:16:23.000 --> 00:16:41.000 We have some data, photos, and a bunch of resources, including the notes from the briefing calls, all post on that website which will stay alive and be publicly available as an archive of the information that was collected for this earthquake. 00:16:41.000 --> 00:16:56.000 So before I get into the damage, I thought that this fact sheet that was developed by Humboldt County Sheriff's Office is helpful to kind of set the stage summarizing some of the impacts 170 plus residents displaced more than 90 structures deemed 00:16:56.000 --> 00:17:02.000 unsafe more than 31 million dollars in economic damages. 00:17:02.000 --> 00:17:06.000 So I'm gonna briefly give an overview of impacts to a few different areas. 00:17:06.000 --> 00:17:10.000 And all of this information is drawn ports that were given during the clearing House so roads and bridges, several local road closures, power, disruption to signals. 00:17:10.000 --> 00:17:24.000 There was some observed damage to seismic, resisting features, most notably that Fern Bridge was close for repairs, and then also close for a short time for inspections. 00:17:24.000 --> 00:17:30.000 There were several health care facilities that were inspected in green tagged. 00:17:30.000 --> 00:17:38.000 They were performed as expected power outages triggered backup power systems which all worked as planned. 00:17:38.000 --> 00:17:39.000 There were a lot of impacts to housing, especially in Rio Dell. 00:17:39.000 --> 00:17:43.000 There were chimney failures, 69 red tech buildings, 111 yellow tag buildings at about 1,400. 00:17:43.000 --> 00:17:57.000 Units, housing Units in Rio Dell, and also many impacts to mobile homes. 100 mobile homes in 3 parks, 2 of the older parts. 00:17:57.000 --> 00:17:58.000 Many of the homes did not have any lateral force resisting systems power outages were widely reported. 00:17:58.000 --> 00:18:06.000 So over 70,000 causes without power, aerial inspections were impaired by cloud cover, which required slower on the ground. 00:18:06.000 --> 00:18:20.000 inspections, and finally damaged to water and wastewater systems in Rio Dell, and there were also boil water notices issued for Rio Dell, and parts of Fortuna. 00:18:20.000 --> 00:18:23.000 So this is very quick overview of damages. 00:18:23.000 --> 00:18:27.000 If you like to learn more, we're hosting a webinar in February 9th, to go into some more detail about these topics. 00:18:27.000 --> 00:18:35.000 I'll also stick the registration URL in the chat for that. 00:18:35.000 --> 00:18:36.000 [noise] 00:18:36.000 --> 00:18:42.000 If you want to learn more about the California Clearinghouse, we're having a workshop and exercise in April at our meeting in San Francisco. 00:18:43.000 --> 00:18:49.000 And so finally to end, thanking a few people who we're sharing information on the Clearinghouse calls, Andrew Lozano and 00:18:49.000 --> 00:18:59.000 Chris Traina of Caltrans, Chris Tomas and Hussain Bhatia, HCAI, Bruce Maison structural engineer, and Megan Stanton of PG&E 00:19:00.000 --> 00:19:05.000 Hi! This is John Eidinger I' gonna talk about the power outages in the earthquake. 00:19:05.000 --> 00:19:19.000 The red line here shows the actual number of customers, Pacific Gas and Electric customers, that lost power as a function of time during the earthquake, blue line shows 50,000. Seventy-one thousand 00:19:19.000 --> 00:19:28.000 total where one customers, one billing account, and that represents every customer in Eureka, Arcata, McKinleyville, Samoa Peninsula, Fortuna, Rio Dell, Trinidad, 00:19:28.000 --> 00:19:33.000 and some outline customers as far as Garberville and Ukiah. 00:19:33.000 --> 00:19:34.000 So I went up after the earthquake with PG&E 00:19:34.000 --> 00:19:44.000 we inspected all the substations. We looked at all the computer logs, and I'll just highlight why PG&E had the outages 00:19:44.000 --> 00:19:58.000 and what caused them. So of these, all these power outages, 5% or 37 locations were due to damage overhead 12 Kilovolt, 12 Kv. wire systems. 00:19:58.000 --> 00:20:03.000 Those are the three line wires systems on wood polls that you see on every street. 00:20:03.000 --> 00:20:04.000 Basically the polls vibrated, the wires get tight, 00:20:04.000 --> 00:20:16.000 in some cases copper wires broke, and the wires ended up down on the ground that represents 5% of all the power outages. 00:20:16.000 --> 00:20:28.000 The other 95% was caused due to phase-to-phase and phase-to-ground faults on the basically, the 60 Kv and 115 kv, that's the high voltage transmission system. 00:20:28.000 --> 00:20:38.000 In every case except two locations the wires came together faulted, causing the circuit breakers to open, but afterwards by patrols. 00:20:38.000 --> 00:20:41.000 Inspector found no damage at all on the lines. 00:20:41.000 --> 00:20:48.000 There were two locations which actually had physical damage, which were repaired quite quickly 00:20:48.000 --> 00:20:56.000 At one substation there was a transform bank that was anchored was built in the 1940s, pretty vintage. 00:20:56.000 --> 00:21:04.000 This is in Fortuna. This site is estimated to have gotten peak grounds acceleration around point 0.05g. 00:21:04.000 --> 00:21:11.000 These transformers four of them here were anchored with 2-half inch diameter bolts. 00:21:11.000 --> 00:21:16.000 They were grossly overloaded and the bolts, failed. 00:21:16.000 --> 00:21:27.000 Once the bolts failed, these transformers rocked around their base and then they broke their bushings on top of the transformers as they interacted with the bus work up above. 00:21:27.000 --> 00:21:28.000 These are photos of the actual damage to the bolts. 00:21:28.000 --> 00:21:38.000 There were largely, very largely undesigned by modern standards, and represents what was done in 1940s. 00:21:38.000 --> 00:21:48.000 There are plenty of utilities in California, Oregon, Washington, Utah, and other places that have similarly weak anchor bolts. 00:21:48.000 --> 00:21:54.000 If you actually see 0.05g expect transformers to break the anchorage rock, and then break the bushings up above. 00:21:54.000 --> 00:21:55.000 Fortunately this substation is another transform bank which carried all the load. 00:21:55.000 --> 00:22:10.000 So this damage caused 0 power outages. So summary, the primary cause of outages were wire swayed and the 60 to 115 kV overheads. 00:22:10.000 --> 00:22:11.000 There were some outages in the local distribution system. 00:22:11.000 --> 00:22:23.000 Most of the duration of the outages reflects a time needed patrols, in other words, physically see that there was no other damage on the lines. 00:22:23.000 --> 00:22:31.000 Most of the patrols found 0 damage. The total number of outages was 79 customer minutes, 79 million customer minutes that's the area under the red line 00:22:31.000 --> 00:22:38.000 we showed first. It's just like a big winter storm that occurs once or twice or 3 times a year. 00:22:38.000 --> 00:22:41.000 This is the largest power outage in the PG&E system 00:22:41.000 --> 00:22:43.000 since the 1989 Loma Prieta earthquake; that one produced about 1 billion customer minutes. 00:22:43.000 --> 00:22:50.000 of outages, or about 15 times as much. Another point is everything that PG&E design installed to the latest standards; 00:22:50.000 --> 00:22:57.000 IBI triple is 693. 00:22:57.000 --> 00:23:00.000 Everything that was installed to the latest standards worked perfectly 100% success. 00:23:00.000 --> 00:23:06.000 That's good. We know the old, old stuff, V=0.25 W. 00:23:06.000 --> 00:23:24.000 Which is about double what the earthquake code requires for regular buildings is still not good enough on Anchorage transformers are totally substandard, and the last point is in terms of power plants, whether we use an importance factor 1.0 or 1.5 design for 4.75 or 00:23:24.000 --> 00:23:27.000 2475, five-year motions make absolutely no difference in terms of power 00:23:27.000 --> 00:23:36.000 outages. I've reviewed 25 different earthquakes in California since 1952. 00:23:36.000 --> 00:23:49.000 Everyone. There's not been a single earthquake where damage of power plants caused outages 00:23:49.000 --> 00:24:01.000 Hello! I'm gonna sort of take a slightly different take on this. Every earthquake has lessons, and they're really important. 00:24:01.000 --> 00:24:09.000 when you're trying to communicate information to people who haven't been thinking about earthquakes for a while. 00:24:09.000 --> 00:24:14.000 So I'm gonna touch on five points. 00:24:14.000 --> 00:24:19.000 It didn't feel like a M6.4. It seemed much bigger. 00:24:19.000 --> 00:24:24.000 Almost everyone, confuses magnitude with the strength of shaking. 00:24:24.000 --> 00:24:28.000 I'm just showing the strong motion record from the Rio Dell Painter Street. 00:24:28.000 --> 00:24:54.000 site. We're talking about arguably, the second or the third strongest ground motion ever recorded in a California earthquake, and in the top 15 of accelerations globally, this was really strong, and a lot of people who know that we have a threat of magnitude, 7.0, 8.0 and even 9.0's are 00:24:54.000 --> 00:24:57.000 terrified that the thought of a future earthquake being a 100 times stronger than what they felt. 00:24:57.000 --> 00:25:11.000 So it's really important to reiterate that accelerations don't relate specifically to magnitude. 00:25:11.000 --> 00:25:14.000 And for the folks in Rio Dell. It's pretty unlikely that they will ever experience stronger accelerations; 00:25:14.000 --> 00:25:25.000 they could be longer, but not stronger. 00:25:25.000 --> 00:25:32.000 So many confusion in different ways about the Triple Junction faults. 00:25:32.000 --> 00:25:35.000 What fault was this on? They want it to be on one fault. 00:25:35.000 --> 00:25:39.000 Is that same fault as 2021 or 1990? 00:25:39.000 --> 00:25:51.000 2, and we have to go over again and again that what we see at the surface is not what you see at depth, and I use the example of making a jigsaw puzzle. 00:25:51.000 --> 00:25:52.000 I love jigsaw puzzles. You put a puzzle together 00:25:52.000 --> 00:26:05.000 it has all these boundaries, complex shapes. And then someone else puts another puzzle right on top of the one 00:26:05.000 --> 00:26:10.000 you just made with different shapes and different edges. 00:26:10.000 --> 00:26:11.000 How can I figure out what's happening beneath the upper puzzle? 00:26:11.000 --> 00:26:24.000 Well, of course there are many geophysical tools but it's the earthquakes themselves that give us the most intriguing information. 00:26:24.000 --> 00:26:41.000 And if I start tweeking, one of the deep puzzle pieces, it's going to affect the other puzzle pieces both beneath and potentially on the surface. 00:26:41.000 --> 00:26:45.000 I'm always told that I predicted a big earthquake. 00:26:45.000 --> 00:26:58.000 I didn't. But people continue to be confused about aftershock forecasts. And it's really important to be super c 00:26:58.000 --> 00:27:14.000 My shake was problematic. We'll be looking at this in more detail, but when an earthquake wakes you up in the middle of the night, and you've already felt the shaking. My shake is not very useful. 00:27:14.000 --> 00:27:20.000 And finally having a toolbox on hand, is critically important. 00:27:20.000 --> 00:27:26.000 We had fortunately finished a virtual tour of the Mendocino Triple Junction about 8 months ago. 00:27:26.000 --> 00:27:45.000 It was great for telling people. We have lots of pre-existing material, and we have established community contacts, established media contacts, and a sustained outreach program with consistent messaging that you don't do it after the earthquake. 00:27:45.000 --> 00:27:56.000 You have to have it in place before. Thanks. 00:27:56.000 --> 00:28:17.000 Hi, everyone! I'm Bob de Root. Greetings from the Blue Lake Rancheria on the River Coast. I just wanted to spend a few minutes talking a little bit about the successes with ShakeAlert alert delivery to end users during actually 00:28:17.000 --> 00:28:25.000 this earthquake, that's being discussed now, but also some context with the earthquakes in the Fall of 2022, and of course into the early 2023. 00:28:25.000 --> 00:28:32.000 I just wanna mention really quickly, a little bit about our work. 00:28:32.000 --> 00:28:44.000 Of course, the delivery of ShakeAlert powered alerts to phones has been live in California since October 17th of 2020. 00:28:44.000 --> 00:28:56.000 Excuse me, 2019, in Oregon and Washington, on October 11th and May 4th, 2021, respectively. 00:28:56.000 --> 00:29:05.000 And let me actually revise that actually March 11th and May 4th of 2021. 00:29:05.000 --> 00:29:16.000 We currently have five ways of bringing alerts to people's cell phones, including the Wireless Emergency Alert System delivered over through FEMA's integrated [indiscernible] and warning system. 00:29:16.000 --> 00:29:36.000 And it's not only about Shake or power alerts that are being delivered through cell phones, but we all have 12 licensed operators, or we refer to them as LTOs, that are the ones who distribute and or sell ShakeAlert power products and services. Some of these 00:29:36.000 --> 00:29:46.000 include, Metrolink in Southern California. They're doing enforce slowing of trains and also BART in the Bay Area, slowing down trains when they receive ShakeAlert powered information from USGS. 00:29:46.000 --> 00:29:57.000 Or customers of LTOS, such as as LA Metro in Southern California. 00:29:57.000 --> 00:29:58.000 So this is just a little bit of an overview, I think. 00:29:58.000 --> 00:30:01.000 A lot of what we talk about is sort of the background. 00:30:01.000 --> 00:30:21.000 [technical problems] 00:30:21.000 --> 00:30:27.000 Bob, we can't hear you anymore 00:30:27.000 --> 00:30:28.000 Now 00:30:28.000 --> 00:30:36.000 Can you hear me? Now. Good! I have a bad connection here, so I actually have a backup. 00:30:36.000 --> 00:30:37.000 Yes. 00:30:37.000 --> 00:30:38.000 So you can hear me now? Right. Perfect. Backup is good, so let me go to the next slide. 00:30:38.000 --> 00:30:46.000 I don't think I can control. If someone could advance to the next slide for me, if possible. 00:30:46.000 --> 00:30:55.000 Please. I don't have any... Thank you very much. Sorry about all the issues here. Oh, the slides gone 00:30:55.000 --> 00:30:56.000 [noise] 00:30:56.000 --> 00:31:00.000 Let's go to the second slide. Please. 00:31:00.000 --> 00:31:01.000 Great, so just super quickly. I know my time is extremely limited. 00:31:01.000 --> 00:31:20.000 Only have about 2 minutes left, but I just wanted to quickly mention our particular ShakeAlert powered delivery for the earthquakes in the Fall of 2022 going into 2023, and there... it was a Fall of success. 00:31:20.000 --> 00:31:33.000 And so one of the, sort of the alert delivery for the Santa Rosa earthquake across the board is in the order of 400,000 phones received ShakeAlert powered alerts. 00:31:33.000 --> 00:31:43.000 For Alum Rock, over 2.2 million and Ferndale over 3 million and Rio Dell over 240,000 phones. 00:31:43.000 --> 00:31:55.000 And the really great thing about this is that we are now getting situations where millions of people are receiving ShakeAlert power alerts on their phones. 00:31:55.000 --> 00:31:56.000 And so we have a lot more information about people's experiences 00:31:56.000 --> 00:32:05.000 We can learn from what they're, what they like, what they didn't like, how we can improve things. 00:32:05.000 --> 00:32:12.000 We also have a very active Twitter account over 37,000 followers, and we are learning a lot from these folks, including, you know, did they get the alert? 00:32:12.000 --> 00:32:23.000 Did they not get the alert? What was their experience in terms of the content of the information that they got? 00:32:23.000 --> 00:32:28.000 We're doing all kinds of work on the social sciences around this whole question. 00:32:28.000 --> 00:32:44.000 So I know my time is extremely short, but I think that now that the ShakeAlert power alert delivery is up and running and active across the West Coast. We have a a lot to learn from the end users that people are actually getting the alerts. 00:32:44.000 --> 00:32:55.000 I apologize for all the connectivity issues today, but I'm glad I had 5 minutes to tell you all a little bit about this, and I'm sure there'll be a lot of great discussion on this. 00:32:55.000 --> 00:33:00.000 Thank you for your attention. 00:33:00.000 --> 00:33:06.000 Hello, everyone. My name is Jessie Saunders, and I'm going to be highlighting the technical performance 00:33:06.000 --> 00:33:13.000 of several earthquake early morning algorithms during the Ferndale earthquake sequence. 00:33:13.000 --> 00:33:20.000 So, as we just heard earthquake early warning is all about alerting for incoming shaking, and the ShakeAlert system 00:33:20.000 --> 00:33:26.000 does this by estimating distributions of Modified Mercalli Intensity, or MMI, using a source characterization based approach shown in the simplified flow chart here. 00:33:26.000 --> 00:33:42.000 I'm going to be talking about what the current operational ShakeAlert system did as well as touch on the performance of two other early warning algorithms that are undergoing internal real-time testing on the ShakeAlert data streams 00:33:42.000 --> 00:33:47.000 and these are the GFAST algorithm which uses GNSS 00:33:47.000 --> 00:33:58.000 peak ground displacements to estimate magnitude. And then the PLUM algorithm which estimates MMI distributions directly from the station observations. 00:33:58.000 --> 00:34:05.000 I'm going to be focusing on the M6.4 mainshock as well as the M5.4 aftershock. 00:34:05.000 --> 00:34:18.000 I'm going to show several maps of the alert regions associated with the first alert as well as the maximum alert extent, and in the background will be maps like these, and these showed the MMI distributions from ShakeMap as well as from the Did You Feel It? 00:34:18.000 --> 00:34:26.000 intensity reports which are shown in the squares here. Primary public alert thresholds in ShakeAlert our MMI 00:34:26.000 --> 00:34:32.000 M4.5, M3.5, and M2.5, and I have highlighted these contours on these background maps 00:34:32.000 --> 00:34:35.000 here. 00:34:35.000 --> 00:34:50.000 So, let's begin with the mainshock. The map on the left, shows the first alert regions and the map on the right shows the maximum magnitude alert, which corresponds to the maximum extent of these different alert regions. ShakeAlert had an initial magnitude 00:34:50.000 --> 00:34:56.000 estimate of magnitude 5.6 at 8.5 seconds after earthquake origin time; and then it had a maximum magnitude 00:34:56.000 --> 00:35:00.000 estimate of magnitude 6.6 at almost 18 seconds after origin 00:35:00.000 --> 00:35:09.000 time. This earthquake was the first event where the new alert pause feature contributed to the alert region, and this is shown by the dashed octagon on the left map. 00:35:09.000 --> 00:35:16.000 This feature restricts the size of the alert region to within a 100 kilometers 00:35:16.000 --> 00:35:21.000 for the first 5 seconds of the alert, and this was established to reduce the chances of over alerting, 00:35:21.000 --> 00:35:24.000 if the initial magnitude is overestimated 00:35:24.000 --> 00:35:26.000 for some reason. Looking at the maximum magnitude alert on the right and because the magnitude was over 00:35:26.000 --> 00:35:36.000 magnitude 6 the FinDer fault line source was incorporated into the higher MMI alert regions which we can see in the green MMI. 00:35:36.000 --> 00:35:41.000 4.5 contour on this right map. The line source. 00:35:41.000 --> 00:35:53.000 by FinDER is shown by the pink line, and while it doesn't match up exactly with the USGS finite fault model, I think FinDER still did a pretty good job for this event. 00:35:53.000 --> 00:35:58.000 Okay. So now let's look at the magnitude 5.4 aftershock. 00:35:58.000 --> 00:36:06.000 This earthquake was a bit deeper, so there's a chance that ShakeAlert could have detected this earthquake before the sways reached the surface. 00:36:06.000 --> 00:36:09.000 ShakeAlert also had a very good magnitude estimate for this event. 00:36:09.000 --> 00:36:17.000 The maximum magnitude of magnitude 5.6 was estimated right around that 5 second window for the alert pause. 00:36:17.000 --> 00:36:23.000 But the MMI M2.5 alert region did extend beyond this 100 kilometer boundary. 00:36:23.000 --> 00:36:38.000 Okay. Let's not take a quick look at how PLUM did these maps show the maximum alert extends for the magnitude 6.4 mainshock on the left, and then the magnitude 5.4 aftershock on the right using a new configuration of PLUM called 00:36:38.000 --> 00:36:44.000 APPLES which incorporates MMI distance, attenuation into its MMI predictions. 00:36:44.000 --> 00:36:54.000 APPLES did a really good job for these events, where the peak, alert times are close to the maximum magnitude alerts by ShakeAlert and then the first alert for the mainshock, is also quite similar. 00:36:54.000 --> 00:37:07.000 The alert, contours for APPLES also match up pretty well with the observed MMI contours, particularly for the aftershock, and then we can see that APPLES takes a pretty conservative approach to the MMI 00:37:07.000 --> 00:37:11.000 2.5 region, particularly for the mainshock. 00:37:11.000 --> 00:37:17.000 So while it's a bit smaller than what ShakeAlert had, and what we can see on ShakeMap 00:37:17.000 --> 00:37:19.000 it does do a pretty good job of matching up with the 00:37:19.000 --> 00:37:25.000 Did you feel it? intensities. So finally, I wanted to highlight GFAST performance during the mainshock. 00:37:25.000 --> 00:37:32.000 The GFAST implementation that's undergoing testing for ShakeAlert is designed to contribute magnitude estimates during larger magnitude 00:37:32.000 --> 00:37:34.000 events of around magnitude 7 and higher, and this is because GNSS data can be a bit noisy 00:37:34.000 --> 00:37:41.000 and signals from smaller magnitude earthquakes can be obscured. However, while GFAST 00:37:41.000 --> 00:37:45.000 would not have contributed to ShakeAlert for this event, 00:37:45.000 --> 00:37:47.000 some of the after-effect analysis using the real-time GNSS 00:37:47.000 --> 00:37:48.000 solutions done by Brendan Crowell, shown here, shows that GFAST 00:37:48.000 --> 00:37:58.000 can obtain a magnitude estimate of magnitude 6.3, using both 20 second and 30 second windows 00:37:58.000 --> 00:38:02.000 and so these plots here show the 30 seconds window results. 00:38:02.000 --> 00:38:30.000 And so with that, I just would like to sum up by saying that the Earthquake Early Warning algorithms were very successful for both of these events. Thank you. 00:38:30.000 --> 00:38:32.000 If Hamid is not here, maybe we play the video? 00:38:32.000 --> 00:38:40.000 Yeah, I just had Jay just beat me to it. So Hamid wasn't sure he'd be able to make it. 00:38:41.000 --> 00:38:42.000 So he sent a video instead. If you can find that in the folder John. 00:38:42.000 --> 00:38:44.000 Is it? Okay? If we go out of order? 00:38:44.000 --> 00:38:47.000 Sure. 00:38:47.000 --> 00:38:50.000 Thanks. 00:38:50.000 --> 00:38:51.000 Hey, everyone, I'm Grace Parker, at the USGS 00:38:51.000 --> 00:39:01.000 Earthquake Science Center, and I'll be discussing regional trends in ground motion observations from the recent Ferndale earthquakes. 00:39:01.000 --> 00:39:15.000 So I've been looking at ground motions from three earthquakes, the magnitude 6.4 mainshock, and two aftershocks, including the January 1st event, and I've compiled and processed ground motions from a number of networks using the USGS open 00:39:15.000 --> 00:39:18.000 source software GM process, which is available via GIT lab. 00:39:18.000 --> 00:39:22.000 And here I'm showing the ROT D50 peak ground motions. 00:39:22.000 --> 00:39:27.000 PGA on the left, and PGB on the right for the three earthquakes with bend means and standard errors 00:39:27.000 --> 00:39:38.000 along with the board. All 2014 ground motion model in the dash lines for comparison, and you can see that for the mainshock and for PGV 00:39:38.000 --> 00:39:55.000 in particular, the model is close to the average, but for the rest of the earthquakes, and particularly for PGA, there's stronger attenuation as a function of distance than the model predicts, and you can also see the record that recorded the horizontal component 00:39:55.000 --> 00:39:59.000 PGA. Well, above 1g. I can't point to it using this configuration. 00:39:59.000 --> 00:40:11.000 But one question we'd like to answer is, we're like these large amplitudes driven by directivity, site response, or a combination of those factors or something else. 00:40:11.000 --> 00:40:27.000 And that's one reason why I've compiled records for multiple earthquakes, because we can start to get a sense of what features were event-specific versus repeatable across multiple earthquakes like site response and one way to do that is through a process called residuals partitioning 00:40:27.000 --> 00:40:29.000 Where we compute residuals relative to a reference ground motion model. 00:40:29.000 --> 00:40:37.000 The equation in the first line there, and try to attribute portions of those residuals to physical processes. 00:40:37.000 --> 00:40:44.000 And so we can first estimate an event term or the average misfit of the ground motion model for each earthquake. 00:40:44.000 --> 00:40:45.000 And here I've done that only using stations within 50 kilometers of the source. 00:40:45.000 --> 00:41:07.000 As suggested by a Baltay and others 2020 to avoid mapping that distance misfit we saw on the previous slide, and from that we can generate the plot on the right and see that the ground motion model is relatively centered for the main shock, and the smaller aftershock but over predicts the 00:41:07.000 --> 00:41:13.000 Magnitude 5.4 at short periods, and under predicts the motions for all earthquakes. 00:41:13.000 --> 00:41:16.000 at the long periods. 00:41:16.000 --> 00:41:20.000 And then we can also go through a similar procedure to partition 00:41:20.000 --> 00:41:40.000 the event adjusted residuals into the repeatable site response and the remaining residual epsilon. And these are maps that show those site terms in the near source region, and these are relevant to or in addition to the site response model of the ground motion model and you 00:41:40.000 --> 00:42:01.000 can see here that the Rio Dell and the Eel River Valley areas have larger positive deviations from the ground motion model indicating under prediction, whereas the rest of the sites are closer to 0, indicating that the ground motion model is performing better and you can see that CE.89462 the 00:42:01.000 --> 00:42:08.000 station in Rio Dell does have a very large site response factor, especially at PGA, a factor of 4.2 00:42:08.000 --> 00:42:09.000 In addition to what was predicted by the ground motion model. 00:42:09.000 --> 00:42:25.000 So certainly, and ask, the relative amplification is very consistent across all three earthquakes. So certainly some of the large motions in this area were driven by that local site response. 00:42:25.000 --> 00:42:30.000 And then we can also use the residual slipper directivity effects by looking at the room remaining residual. 00:42:30.000 --> 00:42:33.000 The epsilon for the mainshock, which is shown here in contrast to the PGB 00:42:33.000 --> 00:42:43.000 site terms on the left. So that's basically just the zoomed out version of the plot on the previous slide. 00:42:43.000 --> 00:42:51.000 And then on the right hand side we see the portion of the ground motions for the mainshock that we're not explained by the ground motion model 00:42:51.000 --> 00:43:03.000 the event bias or the site terms. And here you can see a clear trend, a clear spatial trend with larger ground motions to the east, northeast of the rupture and lower ground motions 00:43:03.000 --> 00:43:17.000 to the north, and somewhat also to the south, which could be related to rupture directivity, although it's always hard to tell when you are missing kind of half of the observations to the west at the center. 00:43:17.000 --> 00:43:23.000 All, in all these results point to a combination of site, response, and directivity, driving the observed trends in the ground 00:43:23.000 --> 00:43:33.000 motions. Thank you. 00:43:33.000 --> 00:43:53.000 Hi! So as we've heard a couple of times today, this region of California, or has one of the highest seismicity rates. So we've had a number of earthquakes occurring and have actually been fairly well recorded. Back almost 30 years ago, one of the first projects I worked 00:43:53.000 --> 00:44:02.000 on out of graduate school was to look at the ground motions from the 1992 magnitude 7.2 00:44:02.000 --> 00:44:08.000 I believe Cape Mendocino earthquake, and in particular, in the Eel River River Valley 00:44:08.000 --> 00:44:16.000 there was very strong amplification, longer periods, and so I wanted to model that. 00:44:16.000 --> 00:44:26.000 So on the slide on the left here is actually a geologic map from 1980, and you could see a number of blue lines there. 00:44:26.000 --> 00:44:32.000 Those are cross sections that were developed in 1953. 00:44:32.000 --> 00:44:37.000 In this publication by Ogle, the cover page of that is shown on the right. 00:44:37.000 --> 00:44:53.000 I wanted to use those to try to develop a model that I could do, use perform simulations just for reference on this map, on the left, that green star is where the epicenter for the most recent magnitude 00:44:53.000 --> 00:45:05.000 6.4 was located. So just looking at these cross sections, a little bit more detail, I produced them here. 00:45:05.000 --> 00:45:16.000 These are again photocopies or scans of photocopies that I produce or made 30 years ago, so the resolution here isn't fantastic. 00:45:16.000 --> 00:45:20.000 But basically, I use these cross-sections to delineate the basin basement interface. 00:45:20.000 --> 00:45:41.000 And there's Quaternary-Tertiary sediments above that interface, and then much older Jurassic race, Cretaceous, Metamorphic walks beneath it, and just as an example on this cross section the it's showing on the upper right that dash 00:45:41.000 --> 00:45:50.000 Blue line is that interface and the depth there is about 3 kilometers at its deepest extent. 00:45:50.000 --> 00:46:12.000 So using those cross sections, I digitize them, I put them into, or I developed a 2D surface fitting to that, using generic mapping tools and the contours of that are showing on this slide on the left the deepest extent is just beneath Fortuna, 00:46:12.000 --> 00:46:13.000 3 kilometers, and again the green star 00:46:13.000 --> 00:46:21.000 there is the epicenter of the M6.4 for reference. 00:46:21.000 --> 00:46:34.000 So using this model, which I pulled out of my old backup tapes, I wanted to see what we could do in a rapid simulation of of this earthquake. 00:46:34.000 --> 00:46:46.000 So on the upper right, is just some assume velocity, seismic velocities kind of a generic stochastic slip model that I'm showing on the bottom right. 00:46:46.000 --> 00:46:52.000 Put those into a rapid simulation, and hopefully, that will be showing here. 00:46:52.000 --> 00:46:53.000 Yes, a shake movie. And you can see a couple of the effects that Grace talked about. 00:46:53.000 --> 00:47:19.000 So we had an event which ruptured from the west to the east, so we had strong eastward directivity, and that was directed right into the Eel River Basin, so we had this coupling of amplified motions from directivity as well as amplification within the sedimentary 00:47:19.000 --> 00:47:27.000 Basin. The results we're still processing these, Evan Hirakawa has been helping. 00:47:27.000 --> 00:47:32.000 We see some general qualitative agreement, amplified motions in the basins relative to those outside. 00:47:32.000 --> 00:47:43.000 But we need to do a lot more work in terms of getting the refined velocity structure and that's part of our next steps here. 00:47:43.000 --> 00:47:48.000 So we're going to continue and analyze the recorded and simulated motions. 00:47:48.000 --> 00:47:54.000 We also want to look at additional data for constraining the basin structuring velocities 00:47:54.000 --> 00:48:06.000 I'm sure there's a lot more information that has come out in the last 30 years, and then ultimately, we'd like to incorporate this structure into the USGS Bay Area of philosophy model. 00:48:06.000 --> 00:48:13.000 And that's all I have. Thank you very much. 00:48:13.000 --> 00:48:29.000 Hi Everybody! I'm Dara Goldberg with the National Earthquake Information Center and I'm gonna do an overview of the rapid source modeling that was done at NEIC in the hours after the event, this is probably familiar to many of you as the model is currently posted on 00:48:29.000 --> 00:48:33.000 the Ferndale event page. Let's get into it. 00:48:33.000 --> 00:48:40.000 So the NEIC routinely publishes a distributed slip model for earthquakes bigger then magnitude 7.0. 00:48:40.000 --> 00:48:41.000 And that's because broadband teleseismic data is of limited value for moderate magnitude 00:48:41.000 --> 00:48:57.000 events. However, if there are local observations, we can try to model smaller events of interest, and in this case there was an almost overwhelming number of near source 00:48:57.000 --> 00:49:03.000 strong motion stations, and many high-rate GNSS 00:49:03.000 --> 00:49:07.000 observations available in real time. We also had static GNSS 00:49:07.000 --> 00:49:16.000 estimates available within hours and updated regularly. Unfortunately, the deformation signal was too small for InSAR. 00:49:16.000 --> 00:49:28.000 So the model currently posted includes just a few broadband teleseismic observations with signal above the noise and strong motion accelerometer data with high rate and static GNSS. 00:49:28.000 --> 00:49:29.000 So just a quick note on the high rate GNSS 00:49:29.000 --> 00:49:35.000 And I can go through this incredibly quickly because Jessie already showed some of this. 00:49:35.000 --> 00:49:39.000 This is what the data looked like coming into the NEIC. 00:49:39.000 --> 00:49:40.000 So this is north end up components of stations with sorted by their distance from the source. 00:49:40.000 --> 00:50:05.000 So, despite the relatively low magnitude, geodetically speaking, the real-time GNSS peak ground displacement, observations did more or less align to a magnitude of around M6.5, as Jessie stated. I also wanted to show how GNSS data improves over 00:50:05.000 --> 00:50:06.000 time. So on this next slide on the left are three stations. 00:50:06.000 --> 00:50:17.000 They're being sorted northeast up from top to bottom, northeast up, northeast up, northeast up for three different stations. 00:50:17.000 --> 00:50:41.000 So we have 1Hz real time data in black and 5Hz observations using rapid orbits rather than ultra rapid in blue, so this demonstrates the value of improved satellite orbits, and also having higher rate data so you can see 7 distinct 00:50:41.000 --> 00:50:50.000 aliasing features there, compared to the 1Hz data. So we ended up using this 5Hz data in the model 00:50:50.000 --> 00:51:03.000 and then in the middle here you can see how the static GNSS office sets that were estimated by Jerry at ESC, how they evolved with time. On the right 00:51:03.000 --> 00:51:11.000 here's how those compare to the estimates from additional processing facilities, and these are very small offsets, and I didn't include the uncertainties just for spatial clarity. 00:51:11.000 --> 00:51:29.000 But you can see that estimating GNSS offsets is a bit finicky, so now for the meat of it, here are the results of the slipping version that are posted to the event page. 00:51:29.000 --> 00:51:33.000 So on the map you have the local station locations and static GNSS. 00:51:33.000 --> 00:51:39.000 Data fits the slip distribution in the top right and the simple source time function on the bottom. 00:51:39.000 --> 00:51:55.000 I also wanted to show how we estimate the shake map polygon from these slip versions, and this is the finite fault dimensions that then get fed into the Shakemap product, and then on the right here some currently unused figures that we produce with 00:51:55.000 --> 00:52:04.000 these models that show a simple [indiscernible] displacement model in the horizontal on top and the vertical directions on the bottom. 00:52:04.000 --> 00:52:08.000 I'm gonna kind of blow through these model fits, 00:52:08.000 --> 00:52:09.000 data is in black synthetics are in red. 00:52:09.000 --> 00:52:26.000 So you can see that we did a pretty good job fitting the data and then I'll just wrap up with one of our key findings, which is that the model with unconstrained depth clearly prefers shallow slip all the way up to about 5 kilometers 00:52:26.000 --> 00:52:28.000 depth. I've heard a lot of feedback 00:52:28.000 --> 00:52:48.000 that some people think it's more likely that the rupture may be deeper. So in the coming weeks I'll be working on adjusting the fault location and orientation to limit slip to within this lab and see if a viable model comes out of that thanks 00:52:48.000 --> 00:52:56.000 Hey! Hi everybody! I'm presenting a little bit more on the exploration of the coseismic static 00:52:56.000 --> 00:53:07.000 GNSS offsets from the Ferndale earthquake. Before I do that I wanted to display a little bit about the context. 00:53:07.000 --> 00:53:12.000 We've seen figures like this from other talks. 00:53:12.000 --> 00:53:18.000 This is one that I made showing a lot of background information about seismicity in this area. 00:53:18.000 --> 00:53:26.000 And unfortunately, Bob, it goes only back to 1976, so we missed the 1975 earthquake on this plot. 00:53:26.000 --> 00:53:35.000 What we oh, what we see on the right is the 00:53:35.000 --> 00:53:52.000 seismicity showing a slab profile approximately across the latitude of the Ferndale earthquake, and it shows that the estimated location of the Gorda slab around the coastline is somewhere in the 15 to kilometer 15 to 20 00:53:52.000 --> 00:54:03.000 kilometer depth range. Within this context the Ferndale earthquake happened right around here, as we've seen. This earthquake produced 00:54:03.000 --> 00:54:11.000 GNSS offsets at around a dozen GPS stations in northern California. 00:54:11.000 --> 00:54:17.000 So I'm showing here next the 00:54:17.000 --> 00:54:31.000 map of the static offsets and the aftershocks made by Jerry Svarc, and the important thing to note is that most of the offsets at these the static offsets at the GPS stations are 5 to 10mm. 00:54:31.000 --> 00:54:50.000 But there is one station, P161, that's about 25mm in total displacement and so you can see it a little bit north of the hypocenter. And that station having moved 25mm prompts us to ask 00:54:50.000 --> 00:54:54.000 some questions related to Dara's talk 00:54:54.000 --> 00:55:14.000 just a moment ago. So 25mm is kind of a big displacement for an an earthquake around this depth at 20 kilometers depth, and it prompted us to ask questions as to whether that could be believed. So the first thing that I want to show is 00:55:14.000 --> 00:55:22.000 some plots investigating whether the displacement at P161 can be believed. 00:55:22.000 --> 00:55:23.000 So this is the full-time series of that station more than 15 years. 00:55:23.000 --> 00:55:29.000 Black lines are the the earthquakes have been removed 00:55:29.000 --> 00:55:37.000 this is in the North American reference frame. The black lines have the trends, and the red lines have been detrended. 00:55:37.000 --> 00:55:41.000 And so to me. This this is quite a stable time series. 00:55:41.000 --> 00:55:42.000 This is showing the overall interseismic motion relative to North America is mostly North. 00:55:42.000 --> 00:55:59.000 The time series looks very good, and when we zoom in on the last year, there's offsets that are really about 20mm 00:55:59.000 --> 00:56:06.000 maybe more. So we ask a question, how can we explain this from a modeling perspective? 00:56:06.000 --> 00:56:22.000 As one hypothesis I am showing Dara's model, and the other hypothesis is to take the shallow is what we call the Gorda scenario, where we take the shallow part of the slip distribution and we move it 10 kilometers deeper it's an 00:56:22.000 --> 00:56:25.000 arbitrary perturbation of the model, but it's potentially interesting. 00:56:25.000 --> 00:56:43.000 We wanna see how this Gorda scenario fits the data, and particularly the P161. So, I forward modeled both of these, and the results are shown here where the USGS finite fault model is blue and the modified model, the Gorda scenario where everything in the 00:56:43.000 --> 00:56:51.000 shallow subsurface is 10 kilometers deeper is shown in red, and they have given that they have the same moment 00:56:51.000 --> 00:56:59.000 they have similar predictions in the far field, but they're different in the near field where P161 and P159 00:56:59.000 --> 00:57:06.000 have differences in how they fit. The modified model does not fit P161 as well as well as the original model does. 00:57:06.000 --> 00:57:12.000 This is very preliminary, and more than anything 00:57:12.000 --> 00:57:25.000 it suggest that inverse modeling, as Dara suggested, is warranted. Because off the bat, the large displacement at P161 is suggestive of shallower, of slip in the shallower subsurface. 00:57:25.000 --> 00:57:38.000 But more inverse modeling could help us determine if a deeper only variety of models could still fit as well. 00:57:38.000 --> 00:57:44.000 Thank you. 00:57:44.000 --> 00:57:57.000 Hi! I'm Danielle Lindsay, and I'm here to pick up where Katherine's left off, and we have to defend InSAR and say, we can see stuff, and to also be like, P161 is real, so here 00:57:57.000 --> 00:58:06.000 we go. So I have made some small stacks of interferograms that are spanning the time period from October to January. 00:58:06.000 --> 00:58:11.000 And I've used two different data sets. So on the top here, I'm showing examples from the Sentinel 1 00:58:11.000 --> 00:58:20.000 and I used the on-demand product for [indiscernible]. On the bottom I'm showing examples from ALOS-2, which the key difference here is that it's a long wavelength signal 00:58:20.000 --> 00:58:22.000 so it maintains coherence in vegetated areas. 00:58:22.000 --> 00:58:40.000 And we received this data as part of the agreement between [indisernible] and NASA so for each of these I can connected one date to the next 4, and then created a stack and created it so for the time series and then I'm extracting just the step. Sorry. The coseismic step 00:58:40.000 --> 00:58:48.000 from there so I just apply a linear velocity and allow for one step, and then I mask out the pixels based on a temporal coherence. And then the next slide 00:58:48.000 --> 00:58:54.000 you can see I'm going to compare these to coseismic displacements from Kang Wang, and then the finite fault model inside predictions. 00:58:54.000 --> 00:59:03.000 So essentially what Katherine just showed, but interpolated everywhere as if it was InSAR. 00:59:03.000 --> 00:59:11.000 So then I compare what her models predicted inside displacement from Dara's model. 00:59:11.000 --> 00:59:16.000 And then what the InSAR observations are actually showing. So let's jump right in. 00:59:16.000 --> 00:59:21.000 Here we go. So in the ascending this is what the sentinel looks like. 00:59:21.000 --> 00:59:36.000 So this is just the coseismic step, and I've compared it with the GPS coseismic displacements for pixels that are within 500 meters of each of these points, and so, if we just compare on the right the InSAR and the GPS, we have a pretty 00:59:36.000 --> 00:59:43.000 good first order, agreement, considering that we only have one post earthquake interferogram for this. 00:59:43.000 --> 00:59:49.000 So we're seeing kind of displacements on the order of 20-25 mm which is similar to what we see 00:59:49.000 --> 00:59:58.000 on the GPS coseismic. So I guess the first thing that also jumped out to me is after seeing everybody doing the finite fault modeling and P161. 00:59:58.000 --> 00:59:59.000 one was really hard to fit. From the InSAR it looks like a spatially coherent signal. 00:59:59.000 --> 01:00:07.000 It's not a random signal at this particular site. So I think this is real. 01:00:07.000 --> 01:00:08.000 So that's the ascending. This is what the descending 01:00:08.000 --> 01:00:17.000 look like. This is the again, the lack of data in the back field is because of the vegetated areas. 01:00:17.000 --> 01:00:21.000 So again for the descending with've got this like pretty good first order 01:00:21.000 --> 01:00:29.000 agreement between the InSAR and the... Oh, hello! Somebody's texting me. Yeah. 01:00:29.000 --> 01:00:37.000 Good first order agreement, but then, if we start to look and compare this to the ALOS-2, we have much, greater spatial coverage of the data. 01:00:37.000 --> 01:00:44.000 So again, this is the descending as well, and so I've got more kind of coseismic to compare between the GPS and the line of site from the InSAR. 01:00:44.000 --> 01:00:54.000 This data is much noisier. So again, I only have the one interferogram after the earthquake which was taken on the 26th of December. 01:00:54.000 --> 01:01:06.000 And so we can say that there's like a lot of residual topographically correlated noise in here, which, with further acquisitions of data that we should be able to get out of there but this is like promising for having constrained the shell how shallow the slip 01:01:06.000 --> 01:01:09.000 actually was. Because if we can like, look at the gradient between P160 01:01:09.000 --> 01:01:14.000 and P161. Hopefully, this can add to those modeling efforts. 01:01:14.000 --> 01:01:17.000 So yeah, I was like, does this make sense? Is this too much noise? 01:01:17.000 --> 01:01:22.000 And then Katherine was like, "hey, look at these finite fault model predictions," and so this is her model project into the line of site. 01:01:22.000 --> 01:01:32.000 So comparing. This is the ascending track, so like first order for patents, we can see the kind of like negative I mean, there we go, the logs kind of match where it's below. 01:01:32.000 --> 01:01:40.000 It's red, it's red. This looks good, and the ace ending it looks like pretty good as well. 01:01:40.000 --> 01:01:46.000 We've got this like positive flow, and then, when we look at the all state, it starts to fill in that gap even more. 01:01:46.000 --> 01:01:52.000 So it's just starting to make sense. And then the other cool things we found were signals from moving landslides. 01:01:52.000 --> 01:01:58.000 So the closest slow, moving land side to the if you center was less than 5 kilometers, and so we've got this every 14 day. 01:01:58.000 --> 01:02:01.000 Acquisition for the last 18 months. So we're starting to look at what's their triggered landslides? 01:02:01.000 --> 01:02:06.000 What does this tell us about how landslides behave, and and shaking? 01:02:06.000 --> 01:02:10.000 And yeah, hopefully, I can pass on my results to others to constrain models. 01:02:10.000 --> 01:02:15.000 And then also look at the behavior of slow moving landslides and triggering. 01:02:15.000 --> 01:02:19.000 Thank you. 01:02:19.000 --> 01:02:32.000 So I'm going to talk about the earthquakes recently, and the Mendocino Triple Junction region and I am interested in the mechanisms, fault plane solutions and moment tensors that we have. 01:02:32.000 --> 01:02:48.000 There are quite a few with moment tensors ranging in magnitude up from about 3.0 to big and fault plane solutions for whatever was good enough to publish, and I don't have very many slides, and I don't have very many words. 01:02:48.000 --> 01:03:13.000 This was the 2023 earthquake sequence, with the mainshock and the Rio Dell aftershock now in white, and these are the moment tensors that we have for both the 2021 and the 2022 biggest, bigger earthquakes there is no moment 01:03:13.000 --> 01:03:18.000 tensor for the foreshocks from 2021, because the mainshock was 12. 01:03:18.000 --> 01:03:23.000 seconds later and bigger. And so we don't have enough data to do a full way for moment. 01:03:23.000 --> 01:03:34.000 tensor. I'm going to be looking at how the mechanisms match. 01:03:34.000 --> 01:03:40.000 If we add the fault plate solutions for the bigger earthquakes, there are a lot more. 01:03:40.000 --> 01:03:51.000 Then we can clearly see the Mendocino Fault Zone, and we can clearly see that there are a lot of mechanisms that are around and interesting to look at. 01:03:51.000 --> 01:03:55.000 So the gray ones in this case are from the 2021 01:03:55.000 --> 01:03:58.000 sequence, and the yellowish ones from the 2022 01:03:58.000 --> 01:04:01.000 sequence, and I still have a lot more work to do. 01:04:01.000 --> 01:04:08.000 Thank you. 01:04:08.000 --> 01:04:17.000 Well, I think the unquestionably the biggest news of this earthquake is that it is bad news for fault-based hazard models. 01:04:17.000 --> 01:04:19.000 Because clearly this earthquake occurred on a fault in a very well mapped area, and it has nothing to do with the surface 01:04:19.000 --> 01:04:34.000 faulting, and it joins the magnitude 7.1 Ridgecrest earthquake and being yet another large California earthquake that occurs on a fault, we didn't know about it. 01:04:34.000 --> 01:04:39.000 So let's be careful when we hitch our wagon to mapped faults. 01:04:39.000 --> 01:04:40.000 Now, what is this earthquake about in terms of its position and location? 01:04:40.000 --> 01:05:05.000 As Sarah and others have pointed out, this is probably a feature of the Gorda or Juan de Fuca slab, as it subducts underneath California, and there have been similar earthquakes to it, although it's a little shallow and as we've heard in this discussion 01:05:05.000 --> 01:05:11.000 maybe it could be shoved a little deeper or maybe, as Kevin Furlong suggests, it actually isn't in the slab. 01:05:11.000 --> 01:05:15.000 So that's a fascinating question. 01:05:15.000 --> 01:05:18.000 What do we have here? We have the Mendocino Triple Junction. 01:05:18.000 --> 01:05:22.000 It's an unhappy junction because of distortion in the plate 01:05:22.000 --> 01:05:26.000 so there's a network of surface faults that accommodate that. 01:05:26.000 --> 01:05:42.000 And what's interesting is that in the Temblors global hazard model, this is the site with the largest magnitude earthquake that occurs at 1% per year in the entire United States, including the Alaska. 01:05:42.000 --> 01:05:43.000 So what this means is, this is the least surprising earthquake you can imagine. 01:05:43.000 --> 01:05:53.000 It's an area that could have and has had much larger earthquakes in the same area. 01:05:53.000 --> 01:05:59.000 Well given that what is our expectation? For how it's affected 01:05:59.000 --> 01:06:04.000 the surrounding fault systems? So we've done some Coulomb analysis. 01:06:04.000 --> 01:06:13.000 And and it has mildly promoted slip on the Cascadia Mega thrust as you see, in the upper left, and it's promoted 01:06:13.000 --> 01:06:28.000 slip on portions of the little salmon and Bear River faults, as you can see below, and it's increased stress on the eastern end of the Mendocino fault, but it is inhibited stress there's the good news on the northern tip of the San 01:06:28.000 --> 01:06:34.000 Andreas. So in these calculations, we're resolving stress more or less on the planes of interest. 01:06:34.000 --> 01:06:42.000 There's another way to do this where we simply look at use existing photo mechanisms, where we actually know there's a fault. 01:06:42.000 --> 01:06:57.000 And look at what has happened there, and you can see again that the eastern end of the Mendocino fault is promoted, and the northern tip of the San Andreas is inhibited. We don't have any focal mechanisms that capture the Cascadia 01:06:57.000 --> 01:07:02.000 Mega thrust surface so we can't get that information from here. 01:07:02.000 --> 01:07:10.000 Now we did have a large aftershock, and, as far as we can tell, this is not promoted by the rupture. 01:07:10.000 --> 01:07:15.000 One aftershock neither confirms nor refutes 01:07:15.000 --> 01:07:22.000 A Coulomb analysis, but you can see particularly on the right that it's sensitive to the location of this earthquake. 01:07:22.000 --> 01:07:33.000 Now, I'd like to close with one other thought, which is the remarkable association of strong shaking with highly sight amplifying material. 01:07:33.000 --> 01:07:53.000 In fact, it barely looks like the shaking, diminishes with distance instead, all of the sites, as you see on the right, that have strong shaking, are in the red Zone in Temblor's 100 meter resolution side amplification map, and unfortunately, that's where people live so getting site amplification right is 01:07:53.000 --> 01:08:04.000 clearly in in understanding and forecasting strong shaking. Thank you. 01:08:04.000 --> 01:08:08.000 Okay. So I'm gonna talk about our aftershock forecasts. 01:08:08.000 --> 01:08:13.000 This is a USGS operational aftershock forecast in the plots I'm gonna show you today which were created with a software written by Gay Paris, who is an intern with us. 01:08:13.000 --> 01:08:29.000 We're done on January 21st, but I looked at things this morning, and largely results would look very much the same if I could have submitted new slides. 01:08:29.000 --> 01:08:36.000 So, the first thing that stands out about the sequences is here's a map of the 2022 to 2023 aftershocks. 01:08:37.000 --> 01:08:51.000 And this is the dark red dash line is the extent of the aftershock zone we're using in the calculations and we're having to use a much larger radius than we would expect, which is shown in black for a magnitude 6.4 earthquake. 01:08:51.000 --> 01:08:56.000 And this could be due to, you know, strong directivity and dynamic triggering off the end. 01:08:56.000 --> 01:09:00.000 But it was also true in the 2021 sequence 01:09:00.000 --> 01:09:08.000 may have been extended by the 12 second foreshock that Piggy mentioned. 01:09:08.000 --> 01:09:16.000 We look at a magnitude time plot on the X-axis we have days after the mainshock, and why X is magnitude. 01:09:16.000 --> 01:09:22.000 Each gray dot is an aftershock. The blue line is a time varying magnitude of completeness. 01:09:22.000 --> 01:09:29.000 The red solid line is the observed cumulative number of earthquakes greater than that magnitude completeness. 01:09:29.000 --> 01:09:33.000 And then we show the fit to a Reasenburg, Jones, aftershock model. 01:09:33.000 --> 01:09:37.000 And right now that model is fitting things really quite well. 01:09:37.000 --> 01:09:45.000 We jump back to the same plot, but we're gonna look now at a much longer period of time for the 2021 mainshock. 01:09:45.000 --> 01:09:54.000 Really things fit very well through about a 140 days, and then sort of almost a linear background seismicity took over. 01:09:54.000 --> 01:10:01.000 And then, of course, you see that the model fits extraordinary poorly, once we get the 2022 mainshock. 01:10:01.000 --> 01:10:18.000 So this is something we're gonna have to keep track of for this current sequence is when the aftershocks start being dominated by the very active background seismicity in this region. I'm gonna go very quickly on this slide just say that we use a default B equals one for the 01:10:18.000 --> 01:10:23.000 forecast, and that fits the observations for the sequence very well. 01:10:23.000 --> 01:10:26.000 Okay, so get a couple of other parameters for the sequence. 01:10:26.000 --> 01:10:32.000 We have time to the start of each forecast on a log scale on the X-axis. 01:10:32.000 --> 01:10:36.000 The top plot shows the Reasenburg and Jones A parameter. 01:10:36.000 --> 01:10:45.000 That's a log productivity parameter. The purple line show the generic or prior model we use that comes from Hardebeck et al. 01:10:45.000 --> 01:11:03.000 in 2019. In general, we find this region tends to be low productivity compared to the rest of California, but these aftershock sequences, the green shows, are, when we go to a Bayesian forecast after a few hours, and the red are purely sequence specific forecasts we have about 01:11:03.000 --> 01:11:08.000 two times of productivity that we expect for this region. 01:11:08.000 --> 01:11:21.000 That was also true in 2021, and on the bottom we show the more p-value, the decay exponent, and once we've free it up for the sequence specific model, we see that it, you know, there's uncertainty. 01:11:21.000 --> 01:11:25.000 But really the p-value is about what we expected. 01:11:25.000 --> 01:11:29.000 Finally, I want to show a success plot for magnitude 3 and greater forecasts. 01:11:29.000 --> 01:11:42.000 This is a log, log plot of time since the mainshock number of earthquakes of magnitude 3. On day one, for example, we forecast that there would be 0 to 6 magnitude 3 aftershocks the next day. 01:11:42.000 --> 01:11:50.000 There's already been 13 aftershocks and some forecast that there should be 13 to 19 magnitude 3's 01:11:50.000 --> 01:12:00.000 by the end of the second day, and we can plot that as a gate that goes from 13+ 0 to 13 +6. 01:12:00.000 --> 01:12:15.000 and since the observations intersect that gate, the forecast was successful. I can stick on the rest of the one day and one week forecasts, and we see that really we're doing quite well in terms of the success of our forecasts. 01:12:15.000 --> 01:12:16.000 So basically, I don't need to go over this very much. 01:12:16.000 --> 01:12:28.000 We have a very large spatial extent, models fitting well, at least for this early parts but we have to keep an eye on it we've got this over the next 10 days. 01:12:28.000 --> 01:12:32.000 I'm sorry to see hints that maybe we're starting to see background seismicity. 01:12:32.000 --> 01:12:37.000 The important, and but so far the forecasts are still being successful. 01:12:37.000 --> 01:12:41.000 Thank you! 01:12:41.000 --> 01:12:54.000 Hi, everyone! I'm Claire Yoon from USGS Pasadena and together with David Shelley from USGS Golden, we've developed enhanced relocated catalogs for Ferndale aftershocks. 01:12:54.000 --> 01:13:00.000 Enhance earthquake catalog is often out of many more lower magnitude events and precise relocations. 01:13:00.000 --> 01:13:04.000 They're really useful for illuminating active fault structures. 01:13:04.000 --> 01:13:11.000 I'll address two key questions in my talk. The first one is, how to enhance catalogs from different methods 01:13:11.000 --> 01:13:16.000 compare. So I've developed a deep learning based enhanced catalog. 01:13:16.000 --> 01:13:23.000 I'm looking to apply it to characterize aftershocks in a rapid response mode automatically, I'm using the EQTransformer event 01:13:23.000 --> 01:13:35.000 detection. A face picking model, and then I couldn't EikoNet+HypoSVI for location, followed by a HYPODD double difference relocation. 01:13:35.000 --> 01:13:52.000 Now David has independently developed a template matching catalogue where he's taking known a catalogue earthquakes, waveforms, and cross-correlating them with continuous seismic data to detect previously unknown smaller events and then relocating them 01:13:52.000 --> 01:13:59.000 With HYPODD, and then the second question is, what does this enhance catalog 01:13:59.000 --> 01:14:05.000 tell us about the Ferndale aftershocks? 01:14:05.000 --> 01:14:23.000 So on this slide, I'm comparing the the two enhanced catalogs; the one from deep learning, which had a total of about 3,500 events versus the one from template-matching, which had about total of 7,000 events, and this was for the first 2 weeks of 01:14:23.000 --> 01:14:35.000 the aftershocks, both in enhanced catalogs are much more sensitive than the Northern California/ComCat catalog, which had only about 300 events here shown in white. 01:14:35.000 --> 01:14:36.000 One thing that sticks out is template-matching is definitely a more sensitive detector than the deep-learning. 01:14:36.000 --> 01:14:45.000 It's able to find the smallest earthquakes down to magnitude 01:14:45.000 --> 01:14:53.000 0 and 1 shown in light blue. There is a fair amount of overlap between the 2 enhanced catalogs. 01:14:53.000 --> 01:14:54.000 I do want to point out that deep-learning is adding a lot of value. 01:14:54.000 --> 01:15:16.000 Finds over a 1,000 near earthquakes, mostly with magnitudes between 1 and 2, as shown in the red that are missed by templates, and so idea what we want to do is, you know, combine both methods, and to produce an ideal enhance catalog and do it quickly enough so that 01:15:16.000 --> 01:15:28.000 it's useful for earthquake response. You know, ideally, we'd like to do it in real time, but more practically, I think we can get it to work every few hours. 01:15:28.000 --> 01:15:46.000 So here I'm showing a map of the deep learning enhance the relocated catalog with the earthquakes colored by time over the the two time period and moment tensors are plotted for the larger earthquake. David will talk more about the template-matching catalog later in his 01:15:46.000 --> 01:15:56.000 talk, and most of these friendly aftershocks were between the 15 and 25 kilometer depth contours of the downgoing Gorda plate slab. 01:15:56.000 --> 01:16:00.000 Just want to point out to the west, near the mainshock on the left- 01:16:00.000 --> 01:16:07.000 lateral fault. The aftershocks are kind of tightly clustered, you know, further to the east, aftershocks are more distributed, this magnitude 5.4 01:16:07.000 --> 01:16:12.000 aftershock on January 1st, it is interesting. It's later, 01:16:12.000 --> 01:16:16.000 it's deeper on a different right-lateral strikes slip fault 01:16:16.000 --> 01:16:24.000 and it's far from the other aftershocks. Ross Stein's Columb stress transfer doesn't explain why it happened so it's an open question. 01:16:24.000 --> 01:16:35.000 And then, you know, looking at what's happening at depth with some depth cross-sections it a prime here on the bottom, and then cross fault cross-sections to the right. 01:16:35.000 --> 01:16:43.000 The main takeaway is that most of the Ferndale aftershocks were located in the crust of the downgoing Gorda plate slab, including the mainshock and hypocenter. 01:16:43.000 --> 01:16:57.000 The only exception was the 5.4 aftershock, which was deeper in the flap mantle and aftershock's may have occurred where the fluid pressure was higher. Here's a summary please contact, me 01:16:57.000 --> 01:17:04.000 if you have any questions or want to work together. Thanks. 01:17:04.000 --> 01:17:15.000 Hello, folks! I'm here as a representative of the Gorda DAS experiment, and it turns out to be a two-phase project. 01:17:15.000 --> 01:17:37.000 The first phase we started planning feels like a couple of years ago following one of these meetings, and that concluded over the summer, we basically found a section of telecommunications fiber running from roughly downtown Arcada to a PG&E 01:17:37.000 --> 01:17:53.000 substation in Eureka area. And I say representative, because really the success of this project depended critically on the enthusiasm and hard work of our friends at CalPoly 01:17:53.000 --> 01:17:59.000 Humboldt, City of Arcadia Bureau Networks, and and some other folks. 01:17:59.000 --> 01:18:11.000 What this is is a distributed acoustic, sensing project where we turn that to telecommunications fiber into thousands of measurements of dynamic strain. 01:18:11.000 --> 01:18:26.000 So this is about a 15 kilometers section we're sampling every 2 meters at about 250 Hz, and at this, along the path we deployed nodal seismometers every 3 to 500 meters. 01:18:26.000 --> 01:18:36.000 We set up an earthquake detection algorithm to take earthquakes in real time and notify us. 01:18:36.000 --> 01:18:56.000 So we concluded the first phase over the summer like I mentioned, and then Ferndale mainshock occurred, and I don't see control of the slides, so if someone could advance it. I should just say this this is the type of signals that we were recording 01:18:56.000 --> 01:19:02.000 In phase one, we see earthquakes like in the bottom left. 01:19:02.000 --> 01:19:07.000 And we see cars. So in the night time it's relatively quiet. 01:19:07.000 --> 01:19:11.000 Not many people, on the road you can see who's driving and their speed; 01:19:11.000 --> 01:19:16.000 basically the diagonal lines. And in the daytime it gets pretty noisy. 01:19:16.000 --> 01:19:34.000 These are kind of the main things we see, and so when the Ferndale earthquake occurred, we rushed up there, and within a few days we're basically measuring the same thing with the same detection 01:19:34.000 --> 01:19:55.000 system. And so here's an example. Well, I should say this is the map of the region showing Clara's catalog that she just talked about, and I'm just filling in the circles where the events are after the interrogator unit went online, and really again, this was the success of 01:19:55.000 --> 01:20:05.000 this aftershock deployment was, you know, depending critically on our local contacts and relationships. 01:20:05.000 --> 01:20:25.000 Let's see, here we go. So this is the events that we've detected so far are roughly 4 to 5 times the number in ComCat, and that's the the black squares that are kind of increasing over time. 01:20:25.000 --> 01:20:37.000 There's Clara's catalog in the blue, so we're starting to approach the the detection capabilities of an enhanced catalog, and we've caught, you know, the biggest aftershock 01:20:37.000 --> 01:20:44.000 so far the 5.4, and I'll get to that in a second. 01:20:44.000 --> 01:20:57.000 And one thing we're starting to notice is the reliability of these string late forms is really good, so this is an example of a pair of events closely spaced with the same focal mechanism. 01:20:57.000 --> 01:20:58.000 Showing up really well in terms of their waveform similarity, 01:20:58.000 --> 01:21:20.000 so we're starting to explore some EGF-type approaches there on the left the scatterplots are showing you the node velocity compared to the peak strain that we've measured for just a small sample of the events and finally, one 01:21:20.000 --> 01:21:30.000 One of the main issues that's plague DAS is whether or not they have the dynamic range to capture big earthquakes in the near to intermediate field. 01:21:30.000 --> 01:21:38.000 So this is evidence that the DAS system can function properly as long as it's deployed properly. 01:21:38.000 --> 01:21:42.000 So, this is the 5.4, showing no evidence. Clipping. 01:21:42.000 --> 01:21:47.000 We're seeing some interesting converted phases, possibly from the slab interface we're seeing amplitudes that are consistent with prior scaling relationships. 01:21:47.000 --> 01:21:53.000 So this is still ongoing, and we have a lot more to to dig through. 01:21:53.000 --> 01:22:01.000 So thank you. 01:22:01.000 --> 01:22:06.000 Hey everybody! This is following up on the talk that Claire gave a few minutes ago. 01:22:06.000 --> 01:22:08.000 So we're looking at trying to see what we can find in the high resolution 01:22:08.000 --> 01:22:15.000 aftershocks of the Ferndale sequence. 01:22:15.000 --> 01:22:24.000 So, thanks to my collaborators here. Let me just start, before I launch into the 2022 sequence 01:22:24.000 --> 01:22:27.000 let's take a step back and talk for just a moment about the 2021 01:22:27.000 --> 01:22:31.000 sequence. There was also an earthquake on December 20th, 01:22:31.000 --> 01:22:38.000 this one was in 2021. This figure on the right is a paper from the paper by Beck et al. 01:22:38.000 --> 01:22:48.000 That's in review. But basically, what we think happened is that there was a magnitude 6.1 or so earthquake on this pink offshore Mendocino transform fault. 01:22:48.000 --> 01:22:53.000 About 11 seconds later there was a magnitude approximately 6 01:22:53.000 --> 01:22:59.000 earthquake onshore, and the purple was within the Gorda slab. 01:22:59.000 --> 01:23:05.000 And then to put in context, this year's earthquake is this green star in the green focal 01:23:05.000 --> 01:23:15.000 mechanism there. So just a bit north of where the 2021 sequence occurred. 01:23:15.000 --> 01:23:32.000 So as Claire introduced, we're gonna be using a template-matching approach we leverage the 285 earthquakes that are in the NCSN catalog for this time period in the region we're able to get more than 01:23:32.000 --> 01:23:38.000 7,000 events that we can detect and relocate using HYPERDD. 01:23:38.000 --> 01:23:44.000 So this is a plot with both the 2022 and the 2021 sequences. 01:23:44.000 --> 01:23:48.000 They're not relocated together; they're separately relocated at this point, 01:23:48.000 --> 01:24:00.000 but just to give you a sense of where things are. This is the moment tensor from the 2022 magnitude 6.4 mainshock here. 01:24:00.000 --> 01:24:01.000 I think it's been pointed out before that 01:24:01.000 --> 01:24:11.000 there's well known fabric of these left-lateral faults within the within the Gorda slab. 01:24:11.000 --> 01:24:20.000 Okay. So the first thing that it's always exciting for me to look at an aftershock sequence, because you never know what you're gonna see. 01:24:20.000 --> 01:24:28.000 In this case we see lots of interesting structures that appear to be activated, you know, not just a single boring strike slip fault. 01:24:28.000 --> 01:24:36.000 So, you know, if we want to be speculative, we can start to draw what may be the primary magnitude 01:24:36.000 --> 01:24:49.000 6.4 rupture. We see a hint of the subduction interface in the cross-section down at the bottom, and we see lots of other little faults being activated as part of this sequence. 01:24:49.000 --> 01:24:51.000 Now in the next slide, I'm going to zoom in on this western zone that has a lot of seismicity both in map view and the cross-section 01:24:51.000 --> 01:24:56.000 there. Here's what that looks like. And again, 01:24:56.000 --> 01:25:16.000 there's the 6.4 mainshock moment tensor. And you know again, you can see lots of different fault structures that appear to be activated, including what may be the subduction interface in the cross-section there. 01:25:16.000 --> 01:25:20.000 Okay, so just quickly, to try to give you a sense of the complexity of these 01:25:20.000 --> 01:25:25.000 this is a 3D animation, showing the whole aftershock sequence. 01:25:25.000 --> 01:25:32.000 So this is the westend here, currently, on the right, eastend, currently on the left. 01:25:32.000 --> 01:25:39.000 And well, you get the idea. This is a zoom view of the western seismicity 01:25:39.000 --> 01:25:49.000 here, so I'll spin this around the same way, and you can see there's a bunch of false structures here to the south of what's probably the main rupture 01:25:49.000 --> 01:25:56.000 they seem to be dipping to the west, or sorry dipping to the east. 01:25:56.000 --> 01:26:04.000 Okay, so preliminary conclusions are the aftershock seems to be primarily confined within the Gorda slab, and perhaps on its surface 01:26:04.000 --> 01:26:18.000 and they're mostly south of what you might infer as the main magnitude 6.4 rupture. Aftershocks show numerous activated faults, and as shown previously, a lot of them have normal faulting mechanisms. 01:26:18.000 --> 01:26:24.000 So we're hoping to keep working on this sequence, and in particular, you know, understand how 01:26:24.000 --> 01:26:39.000 this aftershock complexity, and transfers into what happened in this earthquake, the processes involved, as well as combining it with finite fault results. Thanks very much. 01:26:39.000 --> 01:26:40.000 Alright. Hi everyone! I'm opting for something completely different. 01:26:40.000 --> 01:26:44.000 I'm Jenna Hill. I'm mostly a sedimentary 01:26:44.000 --> 01:26:45.000 geographer, and I work on the portion of the land 01:26:45.000 --> 01:26:51.000 that's underwater to the west of the part that most of you have been focusing on. 01:26:51.000 --> 01:27:00.000 And we in the Pacific Coastal & Marine Hazard Group, we have been focusing a lot on the Cascadia Subduction Zone and thinking a lot about what we can learn in the offshore 01:27:00.000 --> 01:27:08.000 in response to areas that actually have seismicity unlike the the Megathrust. 01:27:08.000 --> 01:27:18.000 And so we just wanted to to take a minute and sort of think of what we can learn from non-megathrust earthquake and trying to sort out the geologic record here. 01:27:18.000 --> 01:27:37.000 So we are as I mentioned, we are thinking about the Cascadia Megathrust, and most of you probably know this, but the primary optional record for earthquake recurrence on the Megathrust comes from the marine turbidite record and correlation of the marine turbidite 01:27:37.000 --> 01:27:51.000 record suggests that we have earthquake recurrence in the Megathrust every 500 years but there's a number of these turbidite deposits of the basin slope that there's a lot of extra ones in the southern portion in the 01:27:51.000 --> 01:27:52.000 Gorda region that have been inferred by Chris Goldfinger et al. 01:27:52.000 --> 01:27:57.000 who's done most of the work on these turbidites here to represent more frequent ruptures of the Megathrust. 01:27:57.000 --> 01:28:18.000 Mostly these come from different types of turbidites, but one of the things we're mostly interested in here is, you know, can some or some portion, or all, or any number of these more frequent turbidites in the study be explained by other earthquake sources other than the megathrust things like upper 01:28:18.000 --> 01:28:21.000 lower plates events that are much more common around the MTJ. 01:28:21.000 --> 01:28:38.000 Or events like these Ferndale sequence earthquakes. So we've been doing a lot of work in the offshore in the last couple of years to learn a lot about turbidite sequences and earthquake currents, landslides out here, I'm not going to get into it too much but it's mostly to say that 01:28:38.000 --> 01:28:39.000 we, you know, learned a lot of where the abyssal turbidite record is coming from. 01:28:39.000 --> 01:28:48.000 It's coming from slope failures at the steep portion of the outer wedge near the very basin slope 01:28:48.000 --> 01:29:05.000 [indiscernible] close to the abyssal plane. We've got tons and tons of cores from this region, where we can look at the earthquake record in there but one thing that we notice, and similar to what the [indiscernible] team is there's numerous extra events as you get closer to 01:29:05.000 --> 01:29:11.000 the Mendocino Triple Junction, and so, you know, what we're here to talk about today is what we can learn from that. 01:29:11.000 --> 01:29:21.000 So zoom in here's showing showing just the ShakeMap from the the 2022 Ferndale event and sort of we're kind of proposing some work that we think you know would be helpful here. 01:29:21.000 --> 01:29:27.000 One of those would be pre-imposed event sampling. 01:29:27.000 --> 01:29:36.000 We have cores from this region from 2021, 2020 as well as the number of cores from the last couple of years before that and I'm showing an example. 01:29:36.000 --> 01:29:40.000 of cores from the base of the Trinidad Canyon 01:29:40.000 --> 01:29:46.000 here, and essentially the correlated events here, these cores are really close together, only a couple 100 meters apart. 01:29:46.000 --> 01:29:53.000 Correlated red events or megathrust events. The correlated yellow events have been suggested to be historical, [indiscernible]. 01:29:53.000 --> 01:29:56.000 earthquakes, like the 1982 Petrolia event. 01:29:56.000 --> 01:30:03.000 And then these orange ones are just extra events, and we don't really understand if those represent ruptures of the megathrust or what the threshold is for making those kinds of events. 01:30:03.000 --> 01:30:16.000 So our side is there a distinguishable event from these kinds of Ferndale earthquakes that we could sample and compare pre-impose 01:30:16.000 --> 01:30:19.000 event samples and see you know what the the threshold is. 01:30:19.000 --> 01:30:24.000 for this. The other piece of this that we're really interested in is improving the ShakeMap in the offshore. 01:30:24.000 --> 01:30:25.000 These concentric rings of MMI and offshore make us 01:30:25.000 --> 01:30:34.000 a little crazy. You know we all know they're not realistic, and it's really hard to look at. 01:30:34.000 --> 01:30:49.000 You know what's the threshold of shaking if you don't really understand what the shaking is predicted to be offshore, so there's a lot of exciting work that's being done to improve offshore velocity models from multitudinal seismic surveys 01:30:49.000 --> 01:30:53.000 that were conducted on OBS deployments in Oregon and Washington, and now we really need to do is verify these things with institute measurements of site response in the offshore. 01:30:53.000 --> 01:31:03.000 This could be through OBS deployments or borehole observations, DAS or other things like this that would really help us answer these questions. 01:31:03.000 --> 01:31:18.000 That was is the shaking threshold to slope failure, and to produce these turbidites that we see in the offshore record, and I think you know, we want to work with more seismologists on these kinds of things. 01:31:18.000 --> 01:31:21.000 More in Northern California earthquake geologists, and these kinds of things. 01:31:21.000 --> 01:31:30.000 And I think we are a really nice opportunity with the academic community and the [indiscernible] going forward to potentially do some kind of deployment like this. 01:31:30.000 --> 01:31:37.000 So that's what we're asking. Thank you. 01:31:37.000 --> 01:31:56.000 Good afternoon everyone, first of all, I like to thank the organizers for adding my last minute presentation so at the Berkeley Seismological Laboratory we've been routinely estimating source parameters of earthquakes, Peggy Hellwig showed some of the 01:31:56.000 --> 01:32:06.000 moment tensor results in earlier talk, and for the larger earthquakes we routinely estimate finite source models and that's what I like to show today. 01:32:06.000 --> 01:32:14.000 Okay. So, I'm going be presenting results for finite source model. 01:32:14.000 --> 01:32:17.000 We're using the method of Hartzell and Heaton (1983); 01:32:17.000 --> 01:32:18.000 it's a multiple time window approach. We use inverted displacement waveforms from the local strong motion stations 01:32:18.000 --> 01:32:28.000 and here it's hard updated GPS model from January 6th. 01:32:28.000 --> 01:32:35.000 The waveform data is broadband. Only a high pass filter would apply to 20 second period. 01:32:35.000 --> 01:32:37.000 Only 0.05 seconds period and waveform instead of Green's 01:32:37.000 --> 01:32:48.000 functions where computed with a common layered elastic model to GIL7 model, which is used for waveform modeling in the region 01:32:48.000 --> 01:32:51.000 As I mentioned, it's a multiple time window approach. 01:32:51.000 --> 01:32:55.000 So we have summarize time variability in a rupture blockchain variability. 01:32:55.000 --> 01:32:59.000 The hypocenter was placed at the USGS reported depth. 01:32:59.000 --> 01:33:00.000 So on the left we're looking at the 10 recording stations that we've used so far. 01:33:00.000 --> 01:33:16.000 We're using free component data and on the right are the, 21 observations from Jerry Svarc data set. 01:33:16.000 --> 01:33:17.000 And I guess this is sort of a conclusion slide as well. 01:33:17.000 --> 01:33:25.000 The earthquake epicenter is located here, and we saw in previous talk. 01:33:25.000 --> 01:33:28.000 it's primarily unilateral to the northeast. 01:33:28.000 --> 01:33:41.000 Pretty strong directivity effect producing strong ground motions, and this east, northeast direction. 01:33:41.000 --> 01:33:51.000 This shows the fits of the waveform data. This is broadband data. 01:33:52.000 --> 01:33:58.000 The records are quite impulsive. The station 89255 has a peak displacement of about 10 cm, and that's on the right-hand side, and some of the other stations 01:33:58.000 --> 01:34:08.000 KCT, for example, is in a position broadside to the rupture, where you can see a sequence of perhaps three subevents on the north-south component. 01:34:08.000 --> 01:34:14.000 It's overall pretty good. This is a cross-sectional view of a model. 01:34:14.000 --> 01:34:18.000 The hypocenter is the small black circle. 01:34:18.000 --> 01:34:20.000 The rupture is predominantly unilateral 01:34:20.000 --> 01:34:24.000 To the northeast it is confined to deeper depth. 01:34:24.000 --> 01:34:28.000 The hypothetical depth is 17.9 kilometers. 01:34:28.000 --> 01:34:32.000 The main moment release or slip was about 20 kilometers depth. 01:34:32.000 --> 01:34:43.000 A common feature in the model is truncation of about 16 kilometers in all of the models that I produced. 01:34:43.000 --> 01:34:47.000 This was a common feature, now there's low levels slip both to the west and to the east. 01:34:47.000 --> 01:34:54.000 I don't have very much confidence in that. 01:34:54.000 --> 01:35:12.000 From the slip model we're able to compute the average of the 7 stress change on the fault and the average and peak stress drops are 7.3 and 44 MPa from the higher stress drop end which would be consistent with the level strong shaking the 01:35:12.000 --> 01:35:18.000 rupture speed is about 78%. So maybe 90% of the shear wave velocity 01:35:18.000 --> 01:35:25.000 and that explains the strong directivity effect that's been observed. 01:35:25.000 --> 01:35:29.000 This is a plot that's showing the seismicity distribution from Anthony Lowmax. 01:35:29.000 --> 01:35:33.000 We saw this earlier about McPherson's talk. 01:35:33.000 --> 01:35:45.000 The red circles are showing the aftershock and if I'm superimposed with flip on the model, it correlates well with the Eastern most aftershock zone. 01:35:45.000 --> 01:35:55.000 And now, while these models are non unique, it is intriguing that the main slip patch is located in area where there is a hole in the aftershock 01:35:55.000 --> 01:35:59.000 distribution, of course more work needs to be done on that. 01:35:59.000 --> 01:36:07.000 This slide compares the model that we obtain from the local strong motion data and the GPS data and DARS USGS 01:36:07.000 --> 01:36:15.000 result. So there is something to be worked out with respect to the depth of the main slip main moment 01:36:15.000 --> 01:36:21.000 releasing this and I think I'll just skip to the conclusions. 01:36:21.000 --> 01:36:27.000 So the three most important ones are the last three. There's some discussion about station, 01:36:27.000 --> 01:36:35.000 P161. We can fit that probably about a half of the total displacement. 01:36:35.000 --> 01:36:40.000 So there's a lot more work to be done. I have tried models that allow for shallower slip. 01:36:40.000 --> 01:36:45.000 I can fit this station, this record much better. 01:36:45.000 --> 01:36:55.000 However, the slip is so far away from the hypocenter, and late in time, that it's completely unconstrained by the seismic waveform data. 01:36:55.000 --> 01:37:15.000 The current model suggested the rupture was in the subducted Gorda slab, and I'm planning to update the modeling, incorporating the InSAR observations that Danny showed. Thank you. 01:37:15.000 --> 01:37:22.000 That was wonderful. Thank you. Everyone. Wow! That was a great series of talks. 01:37:22.000 --> 01:37:30.000 Yeah. 01:37:30.000 --> 01:37:31.000 Oh! 01:37:31.000 --> 01:37:33.000 So, jay I have. I have been trying to write notes on questions from the chat Jay may have been doing that as well, and people who have live questions. 01:37:33.000 --> 01:37:36.000 Please put your hand up and we will call on you. Yeah. 01:37:36.000 --> 01:37:40.000 Sarah, do? Do we have a means video. 01:37:40.000 --> 01:37:45.000 So how many city uploaded it? And I'm looking through the. And John, I look at the Google drive, and I don't see it. 01:37:45.000 --> 01:37:46.000 Okay. 01:37:46.000 --> 01:37:47.000 Did you see? 01:37:47.000 --> 01:37:49.000 I didn't see it either, so I just thought 01:37:49.000 --> 01:37:57.000 Okay, so I think what we should do is we should send him an email and and ask him if he can come on on Wednesday all day, and we can definitely vitamin somewhere. 01:37:57.000 --> 01:37:58.000 Great Idea. 01:37:58.000 --> 01:38:08.000 Okay, so thank you, everyone. And hopefully start putting your hands. 01:38:08.000 --> 01:38:33.000 I would say, I'll the 4 series of questions I saw Will asking kind of related to Bob Mcfilson's talk about how similar this earthquake was to the 1,975 earthquake, including damage despite the fact that that the directivity was a voice between the 2 of them bob do you have any thoughts about that 01:38:33.000 --> 01:38:36.000 No. 01:38:36.000 --> 01:38:37.000 Good. Yeah. 01:38:37.000 --> 01:38:52.000 I actually, I thought the the the models, the star motion models, and the waves, and that one particular movie was quite fascinating. 01:38:52.000 --> 01:39:01.000 And Ross Stein's predictions of where strong sake shaking should take place. 01:39:01.000 --> 01:39:06.000 I I I think, because of the strike slip. 01:39:06.000 --> 01:39:17.000 I'm not sure I I I I'm not sure it looks like the 5, 4, 5, 2 in 1975. 01:39:17.000 --> 01:39:24.000 Didn't have near as strong shaking as the 6 4 as far as the strong motion records. 01:39:24.000 --> 01:39:33.000 So maybe directivity was less than 75 01:39:33.000 --> 01:39:34.000 I see? Yeah. 01:39:34.000 --> 01:39:41.000 Yeah. Ruth had asked. Sorry Ruth had asked that question, and I thought maybe that, you know like like. 01:39:41.000 --> 01:39:52.000 So Bob posed how the 6, 4 main shock, and then the slip moved in one direction, whereas 5, the 75 that moved to the end of the direction. 01:39:52.000 --> 01:39:53.000 But maybe that's just the where the slip is. 01:39:53.000 --> 01:40:09.000 Migrating, and not necessarily the orientation where the ground, the strong ground motions are moving, so that may be a way of disentangling those 2 things 01:40:09.000 --> 01:40:19.000 Well, I I think, too, that they have shown today in many videos, the boundaries between sedimentary basins and mountain ranges have a profound effect. 01:40:19.000 --> 01:40:23.000 And so I think it's 75. You have that same boundary. 01:40:23.000 --> 01:40:29.000 And so you you focus shaking, and Rio Dell without directivity. 01:40:29.000 --> 01:40:38.000 Good point. And that's why I included the cross-section showing the new gene Eel River sedimentary basin. 01:40:38.000 --> 01:40:51.000 It looks like Bruce has a hand raised. Unmute yourself, Bruce. 01:40:51.000 --> 01:40:57.000 Yes, my name is Bruce Mason. I'm a structural engineer that's documenting damage in Rio Dell. 01:40:57.000 --> 01:41:09.000 The question I have is for Bob Mcpherson. I was fascinated that he said that the damage from this earthquake was similar to the magnitude 6.4 event. 01:41:09.000 --> 01:41:14.000 So my question is, do you have a document that you can point me to? 01:41:14.000 --> 01:41:21.000 That talks about visual descriptions of the damage that was observed in 1,975. 01:41:21.000 --> 01:41:24.000 Right now. Of course, we've observed chimneys that fell. 01:41:24.000 --> 01:41:25.000 Of course we've observed much damage at post and peer construction. 01:41:25.000 --> 01:41:35.000 I'd like to know what you observe. Back in 1975 01:41:35.000 --> 01:41:36.000 Please comment. 01:41:36.000 --> 01:41:44.000 Okay, yeah, I maybe could find documents in my unorganized office. 01:41:44.000 --> 01:41:54.000 And certainly we had descriptions. We wrote to Pacific as an electric, but I can. 01:41:54.000 --> 01:41:56.000 I can look for those, and we just we need to commute 01:41:56.000 --> 01:42:00.000 Okay. Okay. For sake of time, I'll contact you via email. 01:42:00.000 --> 01:42:06.000 I have 2 other questions, and I'll be quick. The second is, for who was the gentleman that had the Eel River Basin slide? 01:42:06.000 --> 01:42:09.000 I missed his name. 01:42:09.000 --> 01:42:13.000 That was Robert Graves 01:42:13.000 --> 01:42:14.000 Yeah. 01:42:14.000 --> 01:42:15.000 Robert Graves. This question is for him, Robert. 01:42:15.000 --> 01:42:20.000 As we know, there is very high intensity shaking in Rio Dell. 01:42:20.000 --> 01:42:32.000 Am I correct to understand that that is, from like a double whammy, that is, we had a basin effect, and then we had combined with that a directivity is my understanding correct 01:42:32.000 --> 01:42:39.000 I I would say yes, that's a possibility, and there could also be additional sites. 01:42:39.000 --> 01:43:02.000 Response, effects, that that we're not included in in my model per se, and I haven't looked in detail at real Dell, but I think the the work that Grace Parker had shown where she looked at multiple events kinda separating out source versus site response and and she could see that 01:43:02.000 --> 01:43:08.000 Rio Dell lights up right, even if so, you kind of remove directivity from it. 01:43:08.000 --> 01:43:15.000 For example, or maybe some of these larger scale basin features, Rio Dell tends to light up. 01:43:15.000 --> 01:43:24.000 So I think that may be a location that in and of itself does get does tend to amplify motions. 01:43:24.000 --> 01:43:29.000 The directivity, I think, certainly contributed, though in the in the main shot 01:43:29.000 --> 01:43:30.000 Okay, good. So, maybe I'll correspond with Grace. 01:43:30.000 --> 01:43:36.000 She was one of the speakers. Final question, quick! This is for Ross Stein. 01:43:36.000 --> 01:43:43.000 You know I'm an instructional engineer, and most you people deal with seismology as a structural engineer. 01:43:43.000 --> 01:43:48.000 It seems like, after every significant earthquake, the seismologists are surprised. 01:43:48.000 --> 01:43:54.000 Ridge crest, now this Ferndale event. 01:43:54.000 --> 01:44:12.000 So my question to you is this, we, as engineers many times do something called site specific type, analyses where we look at the seismicity and source links, and then come up with the motions for the design of a critical structure, say, I this is an informal opinion i'd like from you but do you think 01:44:12.000 --> 01:44:15.000 We're really, that's just an illusion. Considering. 01:44:15.000 --> 01:44:20.000 You guys are always surprised after these major earthquakes. Please comment 01:44:20.000 --> 01:44:28.000 Okay. Well, first of all, I said that this is the most expectable place in the United States for a large earthquake, and in so in that sense, there's no surprise. 01:44:28.000 --> 01:44:33.000 This is the way surprising location you could possibly pick out of a hat. 01:44:33.000 --> 01:44:44.000 But what I did say is that the problem is that almost all probabilistic seismicism analyses from which you'll base your local site. 01:44:44.000 --> 01:44:47.000 Characterization are based on mapped, active faults, and they can mislead us because these guys don't want to be caught. 01:44:47.000 --> 01:45:04.000 You know we keep getting surprised and not just in California, in Japan, in New Zealand, basically, all the places that do a really good job of mapping faults and California's been spectacular. 01:45:04.000 --> 01:45:12.000 We still have this problem that we miss faults. So faults are not the only way you can care about the site. 01:45:12.000 --> 01:45:17.000 We now have 30,000 GPS receivers continuously quoting around the world. 01:45:17.000 --> 01:45:21.000 We can measure, strain. You can't get an earthquake without a strain. 01:45:21.000 --> 01:45:30.000 Buildup. So measuring string, looking at the last century of earthquakes are other tools to help us understand seismic hazard 01:45:30.000 --> 01:45:47.000 Okay, can I? Interject Rawson, sir? So I just wanna make sure it's clear since we have hundreds of people listening at the national seismic hazard model has all these included smooth seismicity which in which will give a high seismic rate and signal in this spot and we do 01:45:47.000 --> 01:46:02.000 actually, include, now, at least in the 2023 update, I think someone in 2,018 deformation models so false are only one part of it, and if that it falls for all of it, this would be a huge problem, but you know, it's not the only thing 01:46:02.000 --> 01:46:07.000 we're using to estimate here earthquake rupture, forecast 01:46:07.000 --> 01:46:11.000 Okay. Thank you. 01:46:11.000 --> 01:46:15.000 Thanks for that, Andy. 01:46:15.000 --> 01:46:16.000 Ross, did you have your hand up all you 01:46:16.000 --> 01:46:17.000 Ross said, Yeah. 01:46:17.000 --> 01:46:24.000 I just wanted to say to Doug Jger that was fascinating, that you were able to shove this earthquake so much deeper, almost twice as deep as the Usgs. 01:46:24.000 --> 01:46:30.000 5 fault model. I hope you put your model as soon as you can into the finite fault databases. 01:46:30.000 --> 01:46:45.000 So, for example, we can do a Coulomb analysis and see if now that magnitude 5.4 after shock lights up and of course it'll change all the calculations of what this earthquake might have done to the surrounding region and to the Cascadia Mega thrust surface because 01:46:45.000 --> 01:46:52.000 it's quite a bit closer to it. 01:46:52.000 --> 01:47:03.000 Yeah, thanks for all. So I'd be happy to do that. 01:47:03.000 --> 01:47:11.000 Thanks, and now we have how oh, cool, pronounce your name correctly. 01:47:11.000 --> 01:47:12.000 He's Esther. Yes. 01:47:12.000 --> 01:47:14.000 Can you hear me? Yeah, this is how go from University of Wisconsin? 01:47:14.000 --> 01:47:24.000 Madison. It's great to hear so many great talks in this about this earthquakes. 01:47:24.000 --> 01:47:35.000 I just I saw some argues in throughout the talks and some confusion about where there may shock, and afterwards are so I think I is great. 01:47:35.000 --> 01:47:48.000 If I can show my figure on the cross-sections of my threed locations in the context of my recent Austin models to help people, can I share my screen? 01:47:48.000 --> 01:47:58.000 Just one figure. 01:47:58.000 --> 01:47:59.000 Yes, we can see it. 01:47:59.000 --> 01:48:02.000 Okay, can you smash screen? Okay, so this is some this figure on the top. 01:48:02.000 --> 01:48:04.000 And this is Matthew off the earthquake locations. 01:48:04.000 --> 01:48:17.000 So on. This star is the Meshak relocations, and these black dots are the upshot locations scaled by the Af. 01:48:17.000 --> 01:48:24.000 Shark Magnus. Order on the meshalk, and abstracts are relocated by the 3D velocity model. 01:48:24.000 --> 01:48:31.000 So the 3 panels in in the bottom shows the Mono cross sections along along the profile. 01:48:31.000 --> 01:48:37.000 A prime show here from a a prime from here to here. 01:48:37.000 --> 01:48:41.000 So this is the Vp model cross section. This is Vs model cross session. 01:48:41.000 --> 01:48:44.000 And this is the Pbs ratio model cross section. 01:48:44.000 --> 01:48:49.000 So we see there's some lines. So the the dash line is the Mccarthy. 01:48:49.000 --> 01:48:50.000 At our 2025 step, interface model, and the so our Vp. 01:48:50.000 --> 01:49:00.000 My Vs ratio model shows there is a hypothesis layer at the top of the slab. 01:49:00.000 --> 01:49:04.000 So which is our line spine is 2 solid lines. 01:49:04.000 --> 01:49:11.000 So we think on this hypvis layer. So we interpret to be the sl across the annual. 01:49:11.000 --> 01:49:17.000 Maybe some sediments at the top of the slack, so you can't see. 01:49:17.000 --> 01:49:35.000 So this again, this direct, the rest are, and the the black dots are their Mayhawk and abstract, so you can see the all the earth, all the automation can options are are are in the step, cross. 01:49:35.000 --> 01:49:42.000 Very few above the stepcraft above a step, and some in a few inner stemming, STEM. 01:49:42.000 --> 01:49:46.000 Mental and the earthquakes. Meshawk and the upstream years. 01:49:46.000 --> 01:50:02.000 Earthquakes are as at the very top layer of the step, crust and the upshocks to the turn economy to northeast of the Michigan, and distributed in both the the upper and the lower State. Cross. 01:50:02.000 --> 01:50:16.000 Yeah, so someone, my result. So my locations using a three-d, a regular high restriction, 3D model shows that this Michel kind of most app shocks are within a step crust. 01:50:16.000 --> 01:50:18.000 So I hope this help with some people's confusion. 01:50:18.000 --> 01:50:23.000 Thanks. 01:50:23.000 --> 01:50:30.000 Thank you so much for showing that actually presented your 2021 cross section in my presentation. 01:50:30.000 --> 01:50:36.000 But this is much more helpful, and people are thanking you in the chat. 01:50:36.000 --> 01:51:02.000 And that was actually what I was gonna mention. This, the one of the biggest conversation points that that has come up is, you know whether and how the depth of this earthquake, and whether or not it was in the slab does anyone have anything to say in response to this 01:51:02.000 --> 01:51:03.000 I have a comment 01:51:03.000 --> 01:51:05.000 Alright, Bob. 01:51:05.000 --> 01:51:10.000 I I find the finite fault models really interesting. 01:51:10.000 --> 01:51:31.000 Both. Doug. Thank you always for sharing your information. And Dora, who I don't know, but in that region Gary Carver, Bob Mclaughlin, and Kenn Alto have always argued that the Mega thrust is bent vertically and comes up or right at 01:51:31.000 --> 01:51:37.000 The false key, where this aftershock sequence plots on the map view. 01:51:37.000 --> 01:51:41.000 So. I love the see how you guys resolve this. 01:51:41.000 --> 01:51:47.000 Shallow model versus deep model, because it's a very interesting area, and there is a vertical share zone there. 01:51:47.000 --> 01:52:01.000 So I'm looking forward to you guys doking it out scientifically. 01:52:01.000 --> 01:52:08.000 Bob Mclaughlin has his hand raised. Unmute yourself 01:52:08.000 --> 01:52:12.000 Okay, yeah, yeah, I, 01:52:12.000 --> 01:52:36.000 This is a very interesting point, that that that Jay and and Bob raised, and I think I I as soon as I I saw where the offshore epicenter was on the front Dale earthquake it it hit me that that this it's the the fault that is that 01:52:36.000 --> 01:53:00.000 This is occurring on is now is on a north east trending structure that is going right underneath these northwest trending structures in the Franciscan at the north end of the of the San Andreas system, and going around into the Mendocino Fracture 01:53:00.000 --> 01:53:22.000 Zone, and and the the seismicity associated with the earthquake has nothing to do with those structures, so I guess the question I I would pose, and is how, to the extent that that that the structures were seeing at the surface are associated with the 01:53:22.000 --> 01:53:30.000 Menndocino and the San Andreas. What is, and we're seeing those structures at the surface. 01:53:30.000 --> 01:53:44.000 Some of those structures we know are are pretty young, and our sizeically active, but we're seeing this seismicity down in the Gorda plate going right across those structures. 01:53:44.000 --> 01:54:02.000 So it seems like there must be some sort of interface between those structures at depth in the Gorda plate and the the what we're seeing at the surface, and I don't understand it myself. 01:54:02.000 --> 01:54:16.000 But I think it's an important question to pose in terms of the superposition of of the active seismic activity that's going on between the the well North and South of the well in the trip triple junction area. 01:54:16.000 --> 01:54:35.000 It seems like we're seeing seismicity superposed on on the subducting or or the seismicity in the, in the subducting. 01:54:35.000 --> 01:54:37.000 Go to plate is is going on underneath. 01:54:37.000 --> 01:54:47.000 Active seismicity associated with the transform system. 01:54:47.000 --> 01:54:55.000 And I'm not sure I understand the kinematics of all that. But I think it's a very important question to ask 01:54:55.000 --> 01:54:57.000 Yeah, I'm I'm glad you brought that up. 01:54:57.000 --> 01:55:13.000 If we look further to the north and the outer in and outer wedges of Kesadia, there are a number of those oblique strike slip faults that are in the I want to fucking plate that are also in the accretionary prism and Chris 01:55:13.000 --> 01:55:19.000 Explained that as it has something to do with the false couple there, and so this, that fault is transferred. 01:55:19.000 --> 01:55:20.000 Yeah. 01:55:20.000 --> 01:55:26.000 Yet we don't. We lack evidence for that in this area related to the 6.4 sequence. 01:55:26.000 --> 01:55:27.000 We don't have geologic structures in the North. 01:55:27.000 --> 01:55:36.000 70 east orientation. So you know, maybe this is telling us something about the like couple. 01:55:36.000 --> 01:55:53.000 Excitement, Jan, coupling along the Mega thrust, and in this area that you know, cause if if we're seeing in 1975 and 22, there have been, you know, there probably been many more earthquakes, just like that, but yet they haven't 01:55:53.000 --> 01:56:01.000 Generated the structure in the upper place, so 01:56:01.000 --> 01:56:02.000 Yeah. 01:56:02.000 --> 01:56:04.000 Well, something like like that is going on. I you know I I almost wish you well. 01:56:04.000 --> 01:56:07.000 I I do wish the ken auto was W. W. 01:56:07.000 --> 01:56:08.000 Yeah. 01:56:08.000 --> 01:56:26.000 Was, was here to to comment, because Ken introduce me to the false Cape sheer zone, and and that is probably one of the the most intensely deformed play places in the Mendocino triple junction area that we see on shore. 01:56:26.000 --> 01:56:32.000 And again, this, these structures, northeast trending structures, are going right right underneath that. 01:56:32.000 --> 01:56:51.000 And so I'm wondering if yeah, yep, in some way or form the the kinematics associated with with the faulting in the Gore to play underneath is perhaps driving, not not just faulting at the surface, but perhaps uplift diaperism and those sorts 01:56:51.000 --> 01:56:53.000 Alright! 01:56:53.000 --> 01:57:00.000 Of structures, which is kind of what the false cap share zone sort of looks like. 01:57:00.000 --> 01:57:01.000 Alright! 01:57:01.000 --> 01:57:04.000 It's sort of a dive. Eric's and a formal sort of structure. 01:57:04.000 --> 01:57:12.000 And could could that be getting driven by by movement down in the underlying, Gorda plate? 01:57:12.000 --> 01:57:18.000 That's not actually rupturing the surface, but producing that kind of deformation in the overlying crust 01:57:18.000 --> 01:57:21.000 Oh, my! God! I'm so glad you just said that that was really great. 01:57:21.000 --> 01:57:22.000 Yeah. 01:57:22.000 --> 01:57:26.000 John, do you have something to say 01:57:26.000 --> 01:57:30.000 John Idinger, a structural engineer, couple of open questions to everybody. 01:57:30.000 --> 01:57:38.000 Here first is the instrument on the at Real Dell Painter Street, recorded PGA. 01:57:38.000 --> 01:57:42.000 1.4G. And a corresponding response. Spectra. 01:57:42.000 --> 01:57:43.000 We went up there and took a look at the instrument. 01:57:43.000 --> 01:57:45.000 It's just a couple of feet from the bridge. 01:57:45.000 --> 01:58:04.000 We're pretty well, at least, I'm pretty convinced the instruments recording the ground motion affected by the bridge not free field, and should probably be discounted from all these attenuation models that people use second question is the entire real eel river area along the Eel 01:58:04.000 --> 01:58:08.000 Rivers map is hi liquefiable. We went around all over the place, and we did see some liquor faction, but very sporadic. 01:58:08.000 --> 01:58:17.000 Does anybody have? Or is anybody gonna take a crack at mapping? 01:58:17.000 --> 01:58:26.000 The geology and geotechnical features correlating to this earthquake as to what did and did not liquidify 01:58:26.000 --> 01:58:32.000 That's a great question. Before I answered, I just wanna 7 s to say we are dangerously close to 5 Pm. 01:58:32.000 --> 01:58:35.000 Now we are not gonna kick anyone out. This is a fun discussion. 01:58:35.000 --> 01:58:41.000 This is an important north grade, but also, like officially, we ended 5. 01:58:41.000 --> 01:58:49.000 I know this is brutally late for people in other time zone so definitely feel free to go and let us know if there's something else you need like. 01:58:49.000 --> 01:58:52.000 If you need to have a ton of break into a discussion section on an on another day. 01:58:52.000 --> 01:58:59.000 But we're gonna keep this going. But also be feel free to call it a day, since it's 5 pm. 01:58:59.000 --> 01:59:03.000 At best in the Pacific time zone 01:59:03.000 --> 01:59:07.000 Thanks, Sarah. I didn't know edit event, but I'd love. 01:59:07.000 --> 01:59:11.000 I'll try to connect with you, John, later, to coordinate our observations with yours. 01:59:11.000 --> 01:59:16.000 Pretty good. 01:59:16.000 --> 01:59:18.000 Ross, did you have something 01:59:18.000 --> 01:59:28.000 Yeah, I, I want to respond to the explosion of color in the chat box about my, you know the perception that I'm taking down fault-based models. 01:59:28.000 --> 01:59:32.000 First of all, what I'm advocating for is humility. 01:59:32.000 --> 01:59:38.000 We are building models. We are building fault-based, characteristic earthquake hazard models. 01:59:38.000 --> 01:59:46.000 Yes, there are areas, sources. There are other things that go into those models, but that is the trunk of the tree. 01:59:46.000 --> 01:59:58.000 The logic tree that we're building, and when we have earthquakes like Darfield, magnitude 7.1 in New Zealand, obviously ridiculous. 01:59:58.000 --> 02:00:10.000 7.1 Kikora, 7.8, which broke all our rules and wasn't not allowed to happen because they were faults separated as much as 15 kilometers apart, ruptured together. 02:00:10.000 --> 02:00:15.000 Christ Church, 6.3 small but damaging, Iwata Miyagi, 6.9. 02:00:15.000 --> 02:00:29.000 I think we have to say to ourselves that we have some issues to deal with, that we could be misled by even the best surface mapping in the world in the countries of Japan, New Zealand, in the Us. 02:00:29.000 --> 02:00:32.000 That's my point. 02:00:32.000 --> 02:00:34.000 Some. 02:00:34.000 --> 02:00:42.000 Thanks, thanks, Ross. I think we have a lot to learn from each other. 02:00:42.000 --> 02:00:54.000 You know it is 50'clock 02:00:54.000 --> 02:01:07.000 Are there any last minute things that people want to raise before we end the day? 02:01:07.000 --> 02:01:10.000 I just wanted to say, that's a it's been a fun session. 02:01:10.000 --> 02:01:31.000 I enjoyed hearing all of all of the contributions that especially the folks that were up there in Humboldt during the earthquake, and all the data that they've collected, and Laurie Dangler and Bob Mcpherson and Jay thanks for for really 02:01:31.000 --> 02:01:34.000 Entertaining, afternoon 02:01:34.000 --> 02:01:35.000 Awesome. Thank you. 02:01:35.000 --> 02:01:40.000 And I just want to add that this was an absolutely amazing session of absolutely amazing research on an earthquake that happened a month ago. 02:01:40.000 --> 02:01:45.000 This is a month's fourth of walk. This is amazing. 02:01:45.000 --> 02:01:48.000 Alright! 02:01:48.000 --> 02:01:55.000 Totally rapid response. 02:01:55.000 --> 02:01:59.000 Awesome. 02:01:59.000 --> 02:02:05.000 Okay. Then let's just give a huge hand to all of our presenters and to Jay our wonderful moderator tonight. 02:02:05.000 --> 02:02:06.000 Thank you. 02:02:06.000 --> 02:02:07.000 Excellent, co-moderating. 02:02:07.000 --> 02:02:10.000 Good. 02:02:10.000 --> 02:02:15.000 And have a fabulous evening. See you again. Same time, same place. 02:02:15.000 --> 02:02:45.000 Tomorrow morning, 9 30, Pacific. 02:02:55.000 --> 02:02:59.000 Am I the only one that wants to rock out right now? 02:02:59.000 --> 02:03:01.000 Thank you. John. 02:03:01.000 --> 02:03:07.000 That was great. All the speakers did a phenomenal job that was intense. 02:03:07.000 --> 02:03:11.000 Yeah. Asking for 5 min. Actually, is it really good? Good. Move 02:03:11.000 --> 02:03:13.000 It was supposed to be 3 min actually 02:03:13.000 --> 02:03:32.000 That would have killed us with time. Thanks for the invitation 02:03:32.000 --> 02:03:41.000 Remember. 02:03:41.000 --> 02:03:55.000 Great job. Today, everybody 02:03:55.000 --> 02:04:08.000 We all tomorrow 02:04:08.000 --> 02:04:17.000 I was muted, I was saying. It shows that it was a very intense day, because Key started with a big sweater and we removed, and he was just had a T-shirt. 02:04:17.000 --> 02:04:18.000 Yeah. 02:04:18.000 --> 02:04:20.000 But he missed my comment. But thank you, thank you so much, John. 02:04:20.000 --> 02:04:24.000 And everyone 02:04:24.000 --> 02:04:25.000 Yeah. 02:04:25.000 --> 02:04:26.000 Yeah. Great job that was that was fun. And thanks for the side comments. 02:04:26.000 --> 02:04:31.000 I did not realize at all. And once you said I was like, Well, yeah, of course that happened because it 02:04:31.000 --> 02:04:33.000 Yeah to screens. And yeah, no, that's no big deal. 02:04:33.000 --> 02:04:36.000 Again, it's just it's just to help right? 02:04:36.000 --> 02:04:37.000 It's not to create. Yeah. 02:04:37.000 --> 02:04:40.000 Yeah, I know I appreciate it, cause I was then I thought about it. 02:04:40.000 --> 02:04:41.000 And it actually, it happened one more time, and then the other ones didn't. 02:04:41.000 --> 02:04:42.000 Yeah. 02:04:42.000 --> 02:04:46.000 And all this fine 02:04:46.000 --> 02:04:47.000 Yeah, thank you. Yup. 02:04:47.000 --> 02:04:55.000 Yeah, exactly. All good. Thank you so much. Great job. See you tomorrow, bye. 02:04:55.000 --> 02:05:07.000 Alright and for everybody else. This is good night. The sun is setting, so I'm gonna end the meeting. 02:04:55.000 --> 02:05:07.000 We'll see you tomorrow.