WEBVTT 00:00:00.000 --> 00:00:04.000 [Yep. Record now.] 00:00:04.000 --> 00:00:12.000 Hello, everyone. Welcome back. We have one more fantastic session to finish off the day. 00:00:12.000 --> 00:00:28.000 And to this journey up back to the Mendocino Triple Junction and back into the yester, yester year of 2022. Here to lead us is Bob and Judy [woohoo!] 00:00:28.000 --> 00:00:31.000 Thanks. Yeah. 00:00:31.000 --> 00:00:35.000 Well, thanks, everyone. Bob, you want to go? 00:00:35.000 --> 00:00:45.000 I guess I'm going first. So I would like to introduce our first speaker, Jay Patton. 00:00:45.000 --> 00:00:57.000 And he's going to talk about the 1975 and the 2022 Gorda Plate earthquakes. 00:00:57.000 --> 00:01:05.000 Thank you for having us here to present an overview of two earthquake sequences that we posit slipped the Gorda intraplate fault. 00:01:05.000 --> 00:01:15.000 Listed here on the left are the team players. 00:01:15.000 --> 00:01:23.000 This presentation includes an overview of Gorda deformation zone intraplate seismicity. 00:01:23.000 --> 00:01:39.000 The southernmost part of the Cascadia Subduction Zone. 00:01:39.000 --> 00:01:54.000 As a reminder, the 2022 magnitude 6.4 Ferndale earthquake sequence seismicity had an east-northeast trend. The mainshock is in yellow and the aftershocks are in orange. 00:01:54.000 --> 00:02:04.000 North is to the right. Here is a view of the geomorphology of the Mendocino Triple Junction. We can see the structural grain of the Gorda Plate offshore. 00:02:04.000 --> 00:02:15.000 We can see the north-northwest oriented structural grain onshore as evidenced by the USGS quaternary faults in white. 00:02:15.000 --> 00:02:28.000 An early observation made before sunrise on 20 December 2022 was that the 2022 seismicity cuts across the grain of the geology of the North America plate. 00:02:28.000 --> 00:02:34.000 However, if we overlay the magnetic anomaly model from Doug Wilson. 00:02:34.000 --> 00:02:41.000 We can see that the 2022 seismicity trend actually aligns along a magnetic anomaly boundary. 00:02:41.000 --> 00:02:50.000 These observations combined provide strong evidence for Gorda structural control of the 2022 seismicity. 00:02:50.000 --> 00:02:56.000 Bob McPherson ran the Humboldt Bay Seismic Network for Terra Corporation. 00:02:56.000 --> 00:03:08.000 We plot seismicity from a little more than a week from June 1975 which crudely aligns along an east-northeast trend. 00:03:08.000 --> 00:03:20.000 On 6 June 1975, There was a magnitude md5.7 earthquake sequence near Ferndale in the same location is the 2022 sequence. 00:03:20.000 --> 00:03:32.000 Here we plot two days of December 22, 2022 seismicity relocated by Anthony Lomax showing the east-northeast trend. 00:03:32.000 --> 00:03:38.000 And here we see both together 2022 in blue and 1975 in green. 00:03:38.000 --> 00:03:45.000 Note how the two mainshocks plot along the 2022 fault trend. 00:03:45.000 --> 00:03:55.000 These seismicity cross-sections are views looking either parallel or perpendicular to the fault plane defined by the Lomax relocations. 00:03:55.000 --> 00:04:02.000 The plots on the left are views looking parallel to the fault left to right is south to north. 00:04:02.000 --> 00:04:09.000 The plots on the right are views looking perpendicular to the fault Left to right is west to east. 00:04:09.000 --> 00:04:15.000 The top row are the 1975 data, the bottom row are the 2022 data. 00:04:15.000 --> 00:04:29.000 And the middle row are both data combined. In the middle row on the left, we can see that the 1975 in green and 2022 in blue mainshocks are along the same fault plane. 00:04:29.000 --> 00:04:42.000 In the middle row on the right, we can see that the 1975 sequence ruptured the lower part of the fault and the 2022 sequence ruptured the upper part of the fault. 00:04:42.000 --> 00:04:52.000 We conclude that the 1975 and 2022 earthquake sequences are one of several examples of repeated slip on Gorda Intraplate fault. 00:04:52.000 --> 00:05:04.000 Here we see these seismicity data with the Wilson magnetic anomaly model as an overlay. 00:05:04.000 --> 00:05:14.000 Thank you for your time. 00:05:14.000 --> 00:05:20.000 Hey, thanks, Jay. 00:05:20.000 --> 00:05:24.000 [Wait, should I wait until this slide finishes.] 00:05:24.000 --> 00:05:43.000 I will. Looks like I'm ready to introduce our next speaker, Jimmy Adderholt at the USGS, who is going to give a talk about structural variability in the Gorda slab with converted phases measured by a fiber optic array. 00:05:43.000 --> 00:05:45.000 Apparently +M7.0 Sneak Peek. 00:05:45.000 --> 00:05:57.000 Hi, everyone. Today I want to talk about some of our results looking at Gorda slab geometry using a distributed acoustic sensing array And towards the end of the talk, I'll talk about some preliminary results. 00:05:57.000 --> 00:06:10.000 Imaging this magnitude 7 Mendocino rupture using that same fiber. So for a bit of motivation, I just wanted to say that Sometimes slab structure varies quite rapidly. 00:06:10.000 --> 00:06:17.000 Both in terms of geometry and composition. So on the left, I'm showing the study from Kirkpatrick et al. 2020. 00:06:17.000 --> 00:06:23.000 Which is a 3D image of the megathrust at shallow depths off the coast of Costa Rica. 00:06:23.000 --> 00:06:33.000 Generated using offshore active source experiments. And what this model shows is that slab geometry varies over many scales. 00:06:33.000 --> 00:06:43.000 And sometimes the megathrust varies at scales that are much finer than the resolution of slab models that are typically used in simulations and inversions. 00:06:43.000 --> 00:06:54.000 And this bumpiness or roughness is important to capture. Because it expectedly plays a large role in megathrust phenomenology. And so we want to come up with ways to characterize it. 00:06:54.000 --> 00:06:59.000 Additionally, on the right, I'm showing this tomography model from Guo et al. 2021. 00:06:59.000 --> 00:07:05.000 Which is a tomography model of the Gorda slab region shown in cross-sectional view. 00:07:05.000 --> 00:07:16.000 And the primary result of this paper is that there's this high VPVS ratio region with depths that correspond to the expected depths of the Gorda slab crust. 00:07:16.000 --> 00:07:21.000 And so the interpretation is that these high VPBS ratios correspond to fluids. 00:07:21.000 --> 00:07:30.000 Which correspond to increased poor fluid pressure within the slab crust and therefore lower effective stress near the megathrust. 00:07:30.000 --> 00:07:47.000 And so this result is important and interesting. But I also think this demography model showcases some of the limitations of our current ways of imaging slabs at deeper depths, depths greater than five kilometers depths greater than the resolution of this Patrick et al. Model. 00:07:47.000 --> 00:07:55.000 And so, for example. This Hy-V PBS ratio and this tomography model shows all sorts of variability. It's quite splotchy. 00:07:55.000 --> 00:08:07.000 But tomography models have limited resolution and artificial smoothing So it's difficult to interpret these difficult second order characteristics of this high VPVS zone. 00:08:07.000 --> 00:08:25.000 Additionally, this slab geometry shown in black that's superimposed on this thermography model is very, very smooth. And this is typical for slab models at these depths Simply because it's very difficult to access the megathrust and the interior of the slab with high frequencies 00:08:25.000 --> 00:08:34.000 At depths greater than five kilometers. And so there's motivation to come up with alternative ways to characterize slab geometries. 00:08:34.000 --> 00:08:45.000 At depth and in onshore regions And so we're going to try and do that in this talk today using distributed acoustic sensing or DAS. 00:08:45.000 --> 00:08:53.000 And so DAS is this emergent technology and seismology that transforms fiber optic cables into dense arrays of strain meters. 00:08:53.000 --> 00:09:00.000 And it works using phase intervarometry of backscattered light. And this technique is quite powerful for a few reasons. 00:09:00.000 --> 00:09:05.000 For one, it can yield meter scale coverage over tens of kilometers. 00:09:05.000 --> 00:09:12.000 Additionally, it allows for these very long-term deployment of dense arrays, which allows you to passively record a diversity of phenomena? 00:09:12.000 --> 00:09:22.000 And finally, it can leverage existing telecommunications infrastructure And this infrastructure is increasingly ubiquitous, both onshore and offshore. 00:09:22.000 --> 00:09:32.000 And so we wanted to see if we could use DAS to study the megathrust and slab structure Which is new for fiber optic seismology. 00:09:32.000 --> 00:09:47.000 So to do this, we're going to use this Das Array in Northern California that was deployed a few years ago by a team at the USGS and collaborating institutions And so this DAS array probes a fiber that spans a path between Arcata california 00:09:47.000 --> 00:09:57.000 And Eureka, California. It's about 15 kilometers long, 3,000 channels. And it's just north of the Mendocino Triple Junction. It's above the Gorda Slab. 00:09:57.000 --> 00:10:01.000 And so I don't need to tell this audience that this is a structurally quite complex region. 00:10:01.000 --> 00:10:09.000 With a lot of seismicity. And so this array has recorded many earthquakes and a subset of those earthquakes are in the interior of the Gordis lab. 00:10:09.000 --> 00:10:15.000 These so-called intraslab earthquakes are interesting to us because they generate these very high frequency wave fields. 00:10:15.000 --> 00:10:22.000 That are transmitted through the slab interface. And sometimes they generate these phase conversions that are observable at the surface. 00:10:22.000 --> 00:10:38.000 And so we thought perhaps combining this dense array geometry from our DAS array with these high frequency converted phase observations, we could image slab structure at depths that are difficult to reach with other methods at fine scale resolution. 00:10:38.000 --> 00:10:44.000 And so the first question we had was, can we observe these converted phases with DAS? And the answer is yes. 00:10:44.000 --> 00:10:53.000 And so on the right are two earthquake wave fields one from a magnitude 5.4, one from a magnitude 4.8. They're very close. They have the same source mechanism. 00:10:53.000 --> 00:10:58.000 And in the 4.8, we don't really see anything. The wave field is very simple, I should say. 00:10:58.000 --> 00:11:04.000 But in the magnitude 5.4, we see this very strong secondary onset of energy following P but before S. 00:11:04.000 --> 00:11:09.000 And so we might suspect that this is a phase conversion at the slab interface. 00:11:09.000 --> 00:11:23.000 But we need to verify it. And to do that. Well, one thing we can do is we can compare the expected arrival times of slab converted phases to the onset time of the secondary energy. 00:11:23.000 --> 00:11:40.000 And to do that, we apply this 3D raytracing approach, and I won't go into the details of this, but essentially, we combine this 3D velocity model from Guo et al., 2021 with the slab to Cascadia slab geometry. 00:11:40.000 --> 00:11:46.000 And so when we do this, we can compute expected arrival times for SP and PS off the slab interface. 00:11:46.000 --> 00:11:52.000 And we see that these expected arrival times match up quite well with the onset time of the secondary energy. 00:11:52.000 --> 00:11:59.000 And so this is good evidence that we're observing these slab converted phases in the magnitude 5.4. 00:11:59.000 --> 00:12:05.000 Now, we don't observe these converted phases in the M4.8, which is interesting. 00:12:05.000 --> 00:12:15.000 Part of the reason could be that just using gray theory to compute converted phases shows a diversity of conversion efficiency. 00:12:15.000 --> 00:12:21.000 Even for the same model, but these sources are close enough where that can't explain everything. 00:12:21.000 --> 00:12:29.000 And so another way to explain this is there's fine scale structural variability between the conversion points from these two earthquakes. 00:12:29.000 --> 00:12:36.000 And one way you can have this variability in conversion efficiency is by changing the fluid composition of the structure quite rapidly. 00:12:36.000 --> 00:12:47.000 And so it could be that this magnitude 5.4 has a converted phase in a region with high VPDS 00:12:47.000 --> 00:12:52.000 ratios either within the slab or the basal for our crust. 00:12:52.000 --> 00:12:57.000 That's contributing to this very strong conversion and the same structure does not exist for the magnitude of 4.8. 00:12:57.000 --> 00:13:04.000 So nonetheless, we've observed a converted phase for the magnitude 5.4. 00:13:04.000 --> 00:13:12.000 And so now we want to use it. And so the question of how do we use these converted phases is non-trivial, and I'll show you how we do it through an example. 00:13:12.000 --> 00:13:17.000 And so on the left is an example of another intraslab earthquake. 00:13:17.000 --> 00:13:24.000 And on the right is its corresponding earthquake wavefield. This wavefield is much messier than the wavefield I was showing in the last slide. 00:13:24.000 --> 00:13:36.000 It's got all sorts of scattered phases superimposed on it. It's very difficult to see any sort of clear secondary phases in the P-wave CODA and it's also very noisy. 00:13:36.000 --> 00:13:41.000 And so we're looking for these really subtle perturbations following the P-first arrival. 00:13:41.000 --> 00:13:46.000 And so we're not going to get really far looking at the raw waveforms. 00:13:46.000 --> 00:13:53.000 However, great success has been found in seismology using correlations to extract small signals in noisy data. 00:13:53.000 --> 00:14:05.000 And so we apply this autocorrelation based approach that actually extracts the differential arrival time between these secondary phases and the P-wave first arrival and so that's what you're seeing on the right. That's a chorelogram. 00:14:05.000 --> 00:14:08.000 And the wave field on the right is much simpler and cleaner. 00:14:08.000 --> 00:14:12.000 And what we see is that in the first 5 seconds following the P-wave first arrival. 00:14:12.000 --> 00:14:19.000 We see two really strong secondary arrivals at 3-4 seconds respectively. 00:14:19.000 --> 00:14:31.000 And when we do 3D ray tracing. Assuming this source location and slab geometry and slab two geometry we find that these secondary phases align very well with the expected arrival times of PS. 00:14:31.000 --> 00:14:40.000 Both at the slab interface and the slab Moho. However, our observations are slightly offset from the expected arrival times. 00:14:40.000 --> 00:14:47.000 And we can actually pick these observations with high precision just by choosing a local maximum and the cryogram. 00:14:47.000 --> 00:14:56.000 And so we can use this difference between our observations and the expected arrival times to invert from modified slab geometry and so that's what you're seeing on the bottom right there. 00:14:56.000 --> 00:14:59.000 As distance aligned the mega thrust on the horizontal, interface depth on the vertical 00:14:59.000 --> 00:15:05.000 And we find in order to fit our observations, we need to change the absolute slab depth by about a kilometer and a half. 00:15:05.000 --> 00:15:13.000 And we need to impose bumpiness on the flab interface. This bumpiness might correspond to this roughness parameter that we're quite interested in 00:15:13.000 --> 00:15:19.000 that I was talking about at the beginning. And so I just want to emphasize what this image on the bottom right is. 00:15:19.000 --> 00:15:27.000 This is a transect of the megathrust fault at depth and onshore generated by a single earthquake with sub-kilometer resolution. 00:15:27.000 --> 00:15:36.000 And so this is potentially quite powerful if applied to DAS arrays that are more expensive across a mega thrust. 00:15:36.000 --> 00:15:42.000 Additionally, we have these we have many observations of converted phases from a number of events. 00:15:42.000 --> 00:15:50.000 In a large distribution of regions and so we can start to use all of these observations to look at distributions of parameters. 00:15:50.000 --> 00:15:54.000 To see how we would amend the overall Gorda slab geometry. 00:15:54.000 --> 00:16:01.000 And so I'm showing a couple on the right here. One is how we would amend the absolute slab depth compared to the slab two model. 00:16:01.000 --> 00:16:09.000 And we find that in this region where we have sensitivity, the SLAB 2 model is likely systematically too deep by about a kilometer or two. 00:16:09.000 --> 00:16:21.000 Additionally, in some regions we have co-located measurements of conversions off the slab interface and the slab Moho and so using these we can infer crustal thickness of the slab. 00:16:21.000 --> 00:16:27.000 And we find that the crustal thickness is between 7 and 9 kilometers. 00:16:27.000 --> 00:16:40.000 And so I wanted to show this map on the right here because we are using this DAS array with a kind of small footprint. It's about 15 kilometers, and an interrogator can probe over up to 200 kilometers of fiber 00:16:40.000 --> 00:16:47.000 with the technology that exists now, and so here are some fibers that are in the ground that we know about in interesting regions. 00:16:47.000 --> 00:16:51.000 This is likely a small proportion of the fibers that are actually in the ground. 00:16:51.000 --> 00:17:04.000 And so we can start to imagine the potential here of probing hundreds of kilometers of fibers in subduction zones produce enhanced slab models in certain regions. 00:17:04.000 --> 00:17:10.000 And so this has the potential to be quite powerful. I also wanted to show this image here, 00:17:10.000 --> 00:17:17.000 because the USGS is thinking or is planning on expanding its DAS array substantially. 00:17:17.000 --> 00:17:22.000 in the next year and so on the left is the current slab geometry shown in white there. 00:17:22.000 --> 00:17:39.000 And then on the right is the planned fiber array geometry. And you can see that the future fiber array will cover a much, much larger region that actually covers a substantial portion of the high-slip deficit region of the GorDAS. 00:17:39.000 --> 00:17:45.000 And so the future using the stats array to image slab structure is quite exciting. 00:17:45.000 --> 00:18:00.000 And I just kind of want to make an abrupt transition here to talk a little bit about the GorDAS records of the magnitude 7.0 Mendocino earthquake, and so our fiber did record this earthquake quite well, actually. 00:18:00.000 --> 00:18:14.000 So on the left is a map of the setting here you can see the magnitude 7.0 origin location and our DAS array. So they're about 90 kilometers apart. And on the right is the corresponding earthquake wavefield so that distance along the ray and the horizontal 00:18:14.000 --> 00:18:32.000 time on the vertical. And you can see we recorded this earthquake with very high fidelity. And it's compared to a local broadband station just to show you that our measurements align quite well with the expected measurements from a high-quality broadband station. However, we have several thousand of them. 00:18:32.000 --> 00:18:36.000 And so you can see just looking at this source receiver distance. 00:18:36.000 --> 00:18:45.000 Compared to our array geometry that the array geometry's footprint is quite small compared to the source receiver distance. 00:18:45.000 --> 00:19:01.000 And so using these channels as individual recording stations you're likely not going to get us much extra information. However, if we use array processing techniques, maybe we can learn something new about this earthquake that wouldn't be possible with other data sets. 00:19:01.000 --> 00:19:10.000 And so one array processing technique is seismic beamforming and I won't go into the details of this methodology, but I'll just kind of give you a rough outline. 00:19:10.000 --> 00:19:19.000 So essentially, we take the strain wavefield and we convert it to particle velocity which has the effect of simplifying the wavefield for array processing techniques further on. 00:19:19.000 --> 00:19:30.000 And then we apply seismic beamforming and this is just a way to extract the azimuth and apparent velocity of incoming energy in a seismic wavefield. 00:19:30.000 --> 00:19:40.000 And so what we can do is we can just perform seismic beamforming through time to see how the azimuth of incoming energy relative to the array changes through time. 00:19:40.000 --> 00:19:50.000 And so we take these narrow time windows relative to the P-wave and we see that in these heat map plots, which along the perimeter of these heat map plots 00:19:50.000 --> 00:19:59.000 yields the azimuth of the incoming energy and then further from the center of these heat map plots gives you the apparent velocity. 00:19:59.000 --> 00:20:14.000 And so in these heat map plots progressing through time, we see that the azimuth of incoming energy is systematically rotating clockwise as we get further and further through the counterclockwise as we get further and further through the wavefield. And this counterclockwise rotation 00:20:14.000 --> 00:20:32.000 corresponds to an eastward propagating rupture. And indeed, we can project these azimuths through time on a reasonable about geometry and that yields this plot on the right. So that's time on the horizontal distance from the origin of the earthquake on the vertical 00:20:32.000 --> 00:20:38.000 relative in the eastward direction. And what we see is that we've imaged this unilateral rupture to the east. 00:20:38.000 --> 00:20:53.000 And it's quite a simple rupture with an average velocity of 3.2 kilometers per second. And so we're working to refine this technique But for now, we've imaged a rupture that's 45 kilometers long with unilateral rupture propagation, which is quite interesting. 00:20:53.000 --> 00:21:13.000 And so I will just end with these conclusions so we imaged the Gorda slab geometry and a narrow footprint, but there's a potential to expand this geometry inversion substantially using unprobed fibers. And also we were able to image the magnitude 7.0 rupture propagation with high temporal frequency. 00:21:13.000 --> 00:21:22.000 So I'll end there. Thank you. 00:21:22.000 --> 00:21:30.000 Hey, thanks a lot, Jimmy. Our next speaker 00:21:30.000 --> 00:21:41.000 is Clara Yoon from the USGS and she will be speaking about distinct yet adjacent earthquake sequences near the Mendocino Triple Junction. 00:21:41.000 --> 00:21:46.000 2021 Petrolia and 2022 Ferndale. 00:21:46.000 --> 00:21:52.000 Hi, I'm Claire Yoon, seismologist at the USGS Earthquake Science Center in Pasadena. 00:21:52.000 --> 00:22:01.000 With David Shelley, who's also from USGS, we took a closer look at the recent Petrolia and Ferndale earthquake sequences. 00:22:01.000 --> 00:22:08.000 Near the Mendocino Triple Junction and found them to be distinct spatially adjacent. 00:22:08.000 --> 00:22:15.000 This is a very seismically active part of Northern California where three tectonic plates meet 00:22:15.000 --> 00:22:27.000 Gorda Pacific and North American plates and where the Southern Cascadia Subduction Zone meets the Northern San Andreas Fault. 00:22:27.000 --> 00:22:43.000 On December 20th, 2021, the Petrolia earthquake sequence started with an offshore magnitude 6.1 earthquake shown by the blue strike-slip mechanism along the Mendocino Fault Zone. 00:22:43.000 --> 00:22:51.000 Eleven seconds later, a second magnitude 6.0 earthquake shown by the green strike-slip mechanism. 00:22:51.000 --> 00:22:59.000 Happened onshore 30 kilometers away, which was described by Yak et al., 2023 as a triggered earthquake. 00:22:59.000 --> 00:23:11.000 One year later, on December 20th, 2022, the magnitude 6.4 Ferndale earthquake happened nearby, which is shown by the black strike slope mechanism. 00:23:11.000 --> 00:23:19.000 The questions we seek to answer today are first, how are these two earthquake sequences related, if at all? 00:23:19.000 --> 00:23:33.000 And what faults did they rupture within this complicated region? To answer these questions, we wanted to look at Mendocino earthquake activity in high resolution. 00:23:33.000 --> 00:23:48.000 Our approach was to create an enhanced and relocated earthquake catalog for the Mendocino region using an entirely automatic workflow, which started as a rapid response to the Ferndale earthquake. 00:23:48.000 --> 00:24:04.000 Input to this workflow was continuous seismic data at 66 stations shown by these triangles from December 1st, 2021, to June 1st, 2023, spanning both the Petrolia and Ferndale sequences. 00:24:04.000 --> 00:24:16.000 The first step in the workflow was event detection and face picking at each station with a widely used EQ transformer deep learning model which was trained on a global data set. 00:24:16.000 --> 00:24:23.000 We then combine picks at different stations into events with real grid search associator. 00:24:23.000 --> 00:24:31.000 Next, we located events with a hypo-SVI deep learning based probabilistic event location method. 00:24:31.000 --> 00:24:45.000 We then computed local magnitudes from peak amplitudes and automatically removed false detections. We finally used GrowClust to get more precise relative event locations. 00:24:45.000 --> 00:24:53.000 And we located events with a 1D velocity model used by the Northern California Seismic System. 00:24:53.000 --> 00:25:06.000 We check the quality of our deep learning enhanced catalog By comparing it against the authoritative ComCat catalog contributed by the Northern California Seismic System. 00:25:06.000 --> 00:25:13.000 The deep learning enhanced catalog found almost all ComCat events as shown in dark blue. 00:25:13.000 --> 00:25:19.000 While missing only eight out of 860 ComCat events shown in red. 00:25:19.000 --> 00:25:36.000 The enhanced catalog really contributes a more complete view of both sequences, having newly detected over 14,000 smaller events, mostly with magnitude zero to two, as shown in cyan. 00:25:36.000 --> 00:25:48.000 We also compare our deep learning enhanced catalog against independently derived template matching catalogs, which are available for both the Petrolia and the Ferndale sequences. 00:25:48.000 --> 00:26:06.000 Template matching is a widely used method in seismology. Where a known earthquake waveform colored red and blue here in the top right figure is scanned against continuous seismic data to detect unknown smaller earthquakes with similar waveforms. 00:26:06.000 --> 00:26:16.000 So here we look at three categories of events. The first one, the dark blue events are events common to both the deep learning and the template matching catalogs. 00:26:16.000 --> 00:26:31.000 Second category is the cyan events, which are only in the deep learning catalog, but not in the template matching catalog; you see, these are generally lower in magnitude and occur kind of throughout all times that we looked at. 00:26:31.000 --> 00:26:39.000 In contrast, the third category, magenta events, were only in the template matching catalog, but not in the deep learning. 00:26:39.000 --> 00:26:47.000 These generally include the smallest magnitude less than zero events since template matching is extremely sensitive. 00:26:47.000 --> 00:26:52.000 These results show that deep learning and template matching methods are complementary. 00:26:52.000 --> 00:27:01.000 And ideally, both methods should be used together to get the most complete earthquake catalog possible. 00:27:01.000 --> 00:27:08.000 We also looked at the locations of events in each of the three categories from the previous slide. 00:27:08.000 --> 00:27:17.000 The center panel here shows locations of earthquakes that are common to both the deep learning and the template matching catalogs. 00:27:17.000 --> 00:27:32.000 The left panel shows events that are only in the template matching catalog. We see that these are located close to the known catalog events and it includes the smallest magnitude less than zero events. 00:27:32.000 --> 00:27:37.000 On the other hand, events only in the deep learning catalog, which are shown in the right panel 00:27:37.000 --> 00:27:49.000 are much more evenly distributed in space, which suggests better performance for identifying background seismicity all over the place. 00:27:49.000 --> 00:28:00.000 Now let's look at what the deep learning enhanced relocated catalog reveals about the Petrolia and Ferndale earthquake sequences. 00:28:00.000 --> 00:28:06.000 Here on this slide, the left panel shows only earthquakes in the Petrolia sequence. 00:28:06.000 --> 00:28:18.000 Up to just before when the Ferndale sequence started. The center panel shows only earthquakes in the Ferndale sequence and the right panel shows all earthquakes in both of the sequences 00:28:18.000 --> 00:28:25.000 during the time that we studied. And on this slide, all the earthquakes are colored by depth according to this color scale. 00:28:25.000 --> 00:28:33.000 The most striking observation here is that the Petrolia and Ferndale sequences are very distinct in space. 00:28:33.000 --> 00:28:41.000 They have very little overlap. But they are located spatially adjacent to each other with a very small gap. 00:28:41.000 --> 00:28:55.000 Both of these sequences ruptured mostly strike-slip faults deeper than 15 kilometers As we can see from the mechanisms, which are oriented kind of to the east northeast direction onshore. 00:28:55.000 --> 00:29:07.000 On the right, we note that the automatic deep learning workflow pushed the offshore magnitude 6.1 at Petrolia hypocenter too far to the west, which is incorrect 00:29:07.000 --> 00:29:12.000 and that's because it could not cleanly separate the overlapping waveforms. 00:29:12.000 --> 00:29:21.000 From the offshore 6.1 and the onshore 6.0 11 seconds later. So you really need some manual review. 00:29:21.000 --> 00:29:28.000 So now let's look at the depth dimension. On this slide, the earthquakes are now colored by time. 00:29:28.000 --> 00:29:43.000 With a Petrolia sequence in red. And the Ferndale sequence in purple to green colors. We also show two depth cross sections. There's A to A' prime from west to east, kind of along the Ferndale aftershock zone. 00:29:43.000 --> 00:29:52.000 In the top right. And then there's E to E' prime from north to south and the bottom right. 00:29:52.000 --> 00:29:59.000 We see that both of the Petrolia and the Ferndale sequences happened in the down going Gorda Plate slab. 00:29:59.000 --> 00:30:20.000 But at different depths where the material conditions can be very different and the fault orientations we see from the seismicity are very different from the surface fault trace orientations and shown by these black lines in the overriding North American plate. 00:30:20.000 --> 00:30:38.000 The Ferndale sequence happened in the uppermost of Gorda Plate slab kind of a depths between 15 to 20 kilometers year, but not on the subduction interface with the later aftershocks moving to the east towards more populated areas. 00:30:38.000 --> 00:30:43.000 Guo et al. found evidence of high fluid pressure and low effective stress. 00:30:43.000 --> 00:30:56.000 At these steps where the Ferndale aftershocks occurred and Hellwig et al. reported that slip from the magnitude 6.4 mainshock was confined to the Gorda slab. 00:30:56.000 --> 00:31:15.000 In contrast, the onshore Petrolia sequence occurred deeper in the mantle at depths of 25 to 35 kilometers. There's a very small gap of less than 5 kilometers that separates the Petrolia and Ferndale sequences showing that they are distinct in depth yet spatially adjacent 00:31:15.000 --> 00:31:32.000 but it's an open question why that gap is there. I also note the largest Ferndale aftershock was a magnitude 5.4 on January 1st, 2023, and that was located here about 20 plus kilometers away from the rest of the Ferndale aftershocks 00:31:32.000 --> 00:31:43.000 is deeper at 35 kilometers. The Petrolia and Ferndale sequences were distinct in time as well, 00:31:43.000 --> 00:31:50.000 but much more widely separated by a year. We count the number of events per day from the Petrolia and Ferndale sequences, 00:31:50.000 --> 00:31:56.000 and overlay them, looking at the number of days since the mainshock here on the bottom left. 00:31:56.000 --> 00:32:10.000 We see that both sequences have similar fast decay rates and low aftershock productivity Which is typical of the Mendocino region, possibly because the Gorda Plate crust is very young and warm. 00:32:10.000 --> 00:32:23.000 Now zooming into within two days of the mainshock, we see that the petroleum sequence decayed especially rapidly, pretty much within the first two hours. 00:32:23.000 --> 00:32:34.000 Now, stepping back, taking a broader view, the Petrolia and Ferndale sequences are really examples of diverse earthquake sequence and rupture behavior in the Mendocino Triple Junction. 00:32:34.000 --> 00:32:44.000 If we look at the space-time separation plot considering Petrolia and Ferndale as distinct sequences shown by the green circle. 00:32:44.000 --> 00:32:52.000 We see that they're separated by a longer time of one year, but a very short distance of five kilometers. 00:32:52.000 --> 00:33:06.000 In contrast, if we think about the Petrolia Offshore 6.1, and onshore 6.0 ruptures shown by the pink star, they were separated by a short time of 11 seconds been a much larger distance of 30 kilometers. 00:33:06.000 --> 00:33:15.000 The 1992 magnitude 7.2 Cape Mendocino earthquake sequence shown by the black diamond was also complex. 00:33:15.000 --> 00:33:21.000 Magnitude 7.2 mainshock was on a thrust vault here in the overriding North American plate. 00:33:21.000 --> 00:33:33.000 But one day later, magnitude 6.6 and 6.5 aftershocks occurred 30 kilometers away offshore and deeper in the Gorda Plate. 00:33:33.000 --> 00:33:44.000 So we really need to consider the full range of possible earthquake ruptures and sequences in the Mendocino area, pretty much anywhere on the space-time plot. 00:33:44.000 --> 00:33:48.000 Adding to the complexity are a seismic fault interactions, which I didn't really get into. 00:33:48.000 --> 00:33:57.000 Here, such as post seismic slip after the Ferndale earthquake on the subduction interface inferred by Shelley et al. 2024. 00:33:57.000 --> 00:34:07.000 This motivates future work in earthquake modeling to understand faulting, stress, and earthquake interactions in the Mendocino area. 00:34:07.000 --> 00:34:15.000 To summarize, we have developed a deep learning enhancement, relocated Mendocino earthquake catalog completely automatically. 00:34:15.000 --> 00:34:26.000 This enhanced catalog revealed that the 2021 Petrolia and 2022 Ferndale sequences were distinct in space and time, but spatially adjacent. 00:34:26.000 --> 00:34:35.000 And the Petrolia Ferndale and the 1992 Cape Mendocino earthquake sequences are all examples of diverse and complex earthquake behavior. 00:34:35.000 --> 00:34:44.000 In the Mendocino Triple Junction. For more details, please see our paper, Yoon and Shelley 2024 in the Seismic Record. 00:34:44.000 --> 00:34:52.000 You can get it at this QR code, our enhanced catalogs are also on Zenodo for you to use. 00:34:52.000 --> 00:34:58.000 Now up to this point, I've really given you a limited view of the Petrolia and Ferndale earthquakes. 00:34:58.000 --> 00:35:07.000 Much more recent earthquake catalog research has gone way beyond our study for understanding these recent Mendocino earthquakes. 00:35:07.000 --> 00:35:23.000 And I'll just mention a few examples here. First, Guo et al 2024 relocated the Ferndale sequence with a 3D velocity model, and they examined stress conditions in the upper Gorda Plate where the Ferndale sequence occurred. 00:35:23.000 --> 00:35:37.000 Hellweig et al. 2024 gives the most comprehensive view of the recent Petrolia and Ferndale earthquakes that I've seen so far, including a 40-year relocated Mendocino earthquake catalog. 00:35:37.000 --> 00:35:49.000 And then Rong and Zhu from their 2020 for EGU talk has shown that they've developed a 20-plus year deep learning enhanced and relocated catalog for the Mendocino region. 00:35:49.000 --> 00:36:02.000 This really provides a striking view into active fault structures. So I think the future looks really promising for unraveling the complexity of faulting and earthquake hazards in the Mendocino Triple Junction. 00:36:02.000 --> 00:36:14.000 Thank you, and I'm happy to take your questions. 00:36:14.000 --> 00:36:39.000 Okay. Thank you very much, Clara. Really nice. And so our last talk of the day is going to be by Peggy Hellweg talking about the 1954, I bugged people for years to revisit this earthquake and Peggy was willing to put together 00:36:39.000 --> 00:36:45.000 Team 54. And so here's her story. 00:36:45.000 --> 00:36:57.000 Shortly after I retired. Bob McPherson and Lori Dangler asked me if Berkeley might not have any information about the 1954, December 21 00:36:57.000 --> 00:37:14.000 earthquake that occurred just east of Arcata and I did find some information and I've been working now with a group of people, each of whom has contributed important work toward our project. 00:37:14.000 --> 00:37:19.000 So what do we know about the tectonics of the Mendocino Triple Junction? 00:37:19.000 --> 00:37:26.000 There are earthquakes on the Mendocino Fault. There are earthquakes in the Gorda Plate Oceanic mantle. 00:37:26.000 --> 00:37:35.000 Those are both strike slip type earthquakes. There are earthquakes on the subduction interface. Those are thrust type earthquakes, of course. 00:37:35.000 --> 00:37:45.000 And then there are earthquakes in the North American plate, which are also thrust type earthquakes based on the faults that have been mapped. 00:37:45.000 --> 00:37:50.000 This earthquake occurred. Here. 00:37:50.000 --> 00:38:06.000 Just to the east of the city of Arcata and in the literature there are 14. 00:38:06.000 --> 00:38:15.000 In the literature, there are 14 earthquake locations given, 00:38:15.000 --> 00:38:27.000 in the various bulletins and in various studies that were using this earthquake. The depth was not determined. The mechanism was not determined, 00:38:27.000 --> 00:38:37.000 and we exclude the Mendocino fault as a source because the earthquake was way too far north to be on the Mendocino Fault. 00:38:37.000 --> 00:38:58.000 So the first step was to determine a new location. I found the Berkeley reading sheets instead of using the picks given in the Berkeley Catalog, the Berkeley Bulletin. I used the information from the reading sheets that information also included 00:38:58.000 --> 00:39:19.000 picks from California Seismom Lab stations. There was some information in Don Tosher's thesis and in some BSSA papers and actually what's new for this article, for this work, is that Lind Gee found records from two U.S. Coast and Geodetic Survey Stations, USCGS stations 00:39:19.000 --> 00:39:29.000 in Eureka and Ferndale, which were triggered accelerometers the P-wave is not on them, but you can clearly see the S- wave arrival 00:39:29.000 --> 00:39:38.000 and using information about this instrument, and assuming that the P-arrival triggered the records. 00:39:38.000 --> 00:39:43.000 We can determine a value of S-minus-P for these two stations. 00:39:43.000 --> 00:39:49.000 S-minus-P has been measured in these various articles for Arcata 00:39:49.000 --> 00:39:54.000 from the aftershocks ranging between M2.6 and M2.9 00:39:54.000 --> 00:39:59.000 and we take that value, but it's not actually used in the relocation. 00:39:59.000 --> 00:40:18.000 Our assumptions are that all station operators, because it's not only the Berkeley stations and the stations from the California Seismo Lab all station operators use the same time base and make time corrections consistently and correctly in this same proper way. 00:40:18.000 --> 00:40:30.000 That we could use the the S-minus P time for Arcata from the aftershocks as a determine consistency. 00:40:30.000 --> 00:40:41.000 And that Eureka and Ferndale triggered on the P-wave And there's a 0.5 second turn on time for the accelerometers as described in there. 00:40:41.000 --> 00:41:02.000 In there documentation. So their analysis was done by Anthony Lomax with the nonlinear location hybrid method and using the hybrid velocity using only picks out to 600 kilometers. We tried global picks too but it was not as consistent. 00:41:02.000 --> 00:41:09.000 So this is a map of the red dots are the probability locations. 00:41:09.000 --> 00:41:14.000 The plus sign is the location with the highest probability. 00:41:14.000 --> 00:41:32.000 You can see the station's Arcata, Eureka, and Ferndale here on the map, exactly which they are. They are the three closest stations this kind of oval-shaped cloud on map view is also a relatively oval shaped cloud on the view from the east. 00:41:32.000 --> 00:41:38.000 You can see that after a little below 20 kilometers, the probability is very low. 00:41:38.000 --> 00:41:53.000 But if you view this cloud from the south, you can see that the cloud is tilted, again, an oval but tilted with tilted higher probabilities to the east of Arcata, farther to the east of Arcata 00:41:53.000 --> 00:42:02.000 and increasing probabilities with depth as you go to the west. 00:42:02.000 --> 00:42:20.000 And the primary reason for that is that the S-minus-P times S and the pick from Arcata and the S-minus-P times for Eureka and Ferndale require that earthquake be 00:42:20.000 --> 00:42:31.000 A certain distance away from these stations, and in order to get that distance from the S-minus-P times you have to go deeper as you go to the west. 00:42:31.000 --> 00:42:42.000 So the best location for this analysis is that this earthquake is 11 kilometers deep and it's about 3 kilometers east of Arcata, 00:42:42.000 --> 00:42:54.000 and you can see all these surface faults here. This strand of the surface faulting is the Fickle Hill fault. 00:42:54.000 --> 00:43:06.000 So the next step was to look for a mechanism we read the polarities from the scan records of the UC Berkeley station, all of them that we could find. 00:43:06.000 --> 00:43:11.000 For the location Corvallis there was no record for the mainshock. 00:43:11.000 --> 00:43:24.000 I'm sorry, but there was a record for the largest aftershock on January 30th, and we assume that the first motion of the largest aftershock is the same as for the mainshock. 00:43:24.000 --> 00:43:41.000 We use some polarity from the Tocher thesis and the BSSA papers and then we use the digitized waveforms from Eureka and Ferndale and the S-polarities that we could read off of those records. 00:43:41.000 --> 00:43:47.000 Again, a number of, this is the record from the record Eureka, 00:43:47.000 --> 00:43:57.000 and I have to look up which this one is. So again, we have a lot of assumptions we assume that the instrument polarities are reported correctly 00:43:57.000 --> 00:44:03.000 that the first motion of the largest aftershock at Corvallis is the same as for the mainshock, 00:44:03.000 --> 00:44:27.000 and there were no vertical instruments like at Arcata, the closest station, we infer the vertical P-polarities for stations for the stations with only horizontals so at Ricata both the north and the east polarities were negative and that means that if the source of the earthquake the 00:44:27.000 --> 00:44:35.000 hypocenter is to the east of Arcata. The polarity of the vertical would be up. 00:44:35.000 --> 00:44:39.000 So what did we come up with? This is the mechanism. 00:44:39.000 --> 00:44:44.000 You can see all of the polarities plotted on this mechanism. 00:44:44.000 --> 00:45:01.000 They're all positive and there's no way that we could have a strike-slip mechanism reasonable for the Mendocino area either from Gorda or from the Mendocino fault with no negative polarities. 00:45:01.000 --> 00:45:12.000 Even though all of the stations accept Arcata are essentially to the east of and to the south of the earthquake 00:45:12.000 --> 00:45:17.000 except for Corvallis, which is to the north. 00:45:17.000 --> 00:45:32.000 So we tested possible mechanisms for most of the hypocenters. It requires there be a thrust mechanism and the best solution came up For a depth of 18 kilometers 00:45:32.000 --> 00:45:42.000 and a strike of 350 degrees, which was actually very consistent with subduction interface. 00:45:42.000 --> 00:45:59.000 Once we had solution, part of the solution was to determine a mechanism and the second part of the solution was to produce models for the waveforms to record it at Eureka and Ferndale. 00:45:59.000 --> 00:46:04.000 And you can see these are not perfect matches at Eureka and Ferndale, 00:46:04.000 --> 00:46:09.000 but the timing and the direction of the waveforms is actually pretty good. 00:46:09.000 --> 00:46:30.000 So we are very happy with this solution. At the same time, as we were working on this, Sue Hough was working on collecting and updating the reports of the intensities that the U.S. Coast and Geodetic Survey collected. They're all 00:46:30.000 --> 00:46:41.000 available, what the us is produced in the intensity map from the U.S. Coast and Geodetic Survey is much, many fewer points they were compressed. 00:46:41.000 --> 00:46:57.000 Lori and Bob have been collecting reports from newspapers and local people and this shake map actually includes the maximum intensities from Eureka and Ferndale from the U.S. Coast of Geodetic Survey Stations. 00:46:57.000 --> 00:47:06.000 So this is what the intensity map looks like with a much more detailed view of the reports. 00:47:06.000 --> 00:47:13.000 This is how it matches with U.S. Coastal and Geodetic Survey intensity report. 00:47:13.000 --> 00:47:32.000 [And whoops.] And if we zoom in on the epicentral area, you can see there are some values here that actually reach eight and even a little bit in this region close to our best epicenter, which is the star. 00:47:32.000 --> 00:47:37.000 And this is how this region matches the Coastal and Geodetic Survey, 00:47:37.000 --> 00:47:44.000 map which says in all of this area the highest intensities were separate. 00:47:44.000 --> 00:47:59.000 So what new insights have we gained from our work? The earthquake happened and essentially in the same area as all of these others, but we think we have a much better location and more realistic location. 00:47:59.000 --> 00:48:12.000 The best location is just east of the station Arcata at a depth of 11 kilometers, the mechanism is a thrust earthquake. 00:48:12.000 --> 00:48:28.000 It could be from this Fickle Hill earthquake, but it's certainly under Fickle Hill, which was why we've renamed the event. The best mechanism occurs somewhat deeper than the best Location, but as we know 00:48:28.000 --> 00:48:40.000 the mechanism, the location for the mechanism is centroid for the movement as opposed to the hypocenter, which comes from the pics. 00:48:40.000 --> 00:48:55.000 So we can rule out Gorda because it's not a strike-slip event. It could be on the subduction interface or it could be on the North American one of the North American plate thrusts, but they tend more southwest-northeast, 00:48:55.000 --> 00:49:07.000 and so we think this is unlikely and the event is more likely to be somewhat deeper and on the subduction interface. 00:49:07.000 --> 00:49:14.000 Thank you very much. 00:49:14.000 --> 00:49:25.000 Okay, thank you very much, Peggy. And thank you to all the speakers. 00:49:25.000 --> 00:49:33.000 Seem to have quite a bit of activity in the chat. 00:49:33.000 --> 00:49:34.000 Let's see here. 00:49:34.000 --> 00:49:40.000 Oh, look, I missed it all. 00:49:40.000 --> 00:49:46.000 Yeah, I'm trying to see if there are any questions that didn't get answered in the chat. 00:49:46.000 --> 00:49:55.000 Well, I think we could start with Dave Shelley. Do you want to ask your questions about You questioned Jay. 00:49:55.000 --> 00:50:05.000 In Jay's talk about the slip patch you want to start there. 00:50:05.000 --> 00:50:19.000 Sure. Yeah, thanks for the talk. Yeah, it's really interesting to see. I was just trying to understand, I mean, you described it as a repeat of the 2022 or the 1975 and 2022 is a repeat but it sounds like in detail you're talking about 00:50:19.000 --> 00:50:25.000 complimentary slip segmented in depth that's happening on the same fault. Is that right? 00:50:25.000 --> 00:50:33.000 Yeah, so we think that they both happened on the same fault. So that's sort of like the repeating part. It's the same fault. 00:50:33.000 --> 00:50:54.000 But yeah, it looks like based on the you know, the depths you know, after, you know, plotted you know looking perpendicular to the fault. Yeah, it looks like that the mainshocks and the larger magnitude aftershocks happened on different parts of the fault. 00:50:54.000 --> 00:51:01.000 And the reason why this is important, you know, Bob, you know. 00:51:01.000 --> 00:51:20.000 remember, Bob is the person who, you know. He and I were talking about the 2022 earthquake he's the one that ran the Terra Seismic Network and he's like, hey, we had an earthquake in 1975 and so you know, going through the reports from the 1975 earthquake 00:51:20.000 --> 00:51:33.000 you know, we've, you know, he hand copied the earthquake information from the report and so that we could plot them up. And so it's just really fascinating. 00:51:33.000 --> 00:52:02.000 And the important, you know, one of the reasons why this is important is that, you know, for databases like the National Seismic Hazard Mineral or CFMs like SCAC and Crescent is that these faults are sources of seismic hazard and so having aftershocks like this 00:52:02.000 --> 00:52:10.000 allows us to maybe characterize them a little bit better and try to include them in future versions. 00:52:10.000 --> 00:52:11.000 So. 00:52:11.000 --> 00:52:24.000 Yes, it's really exciting. I mean, I think we don't think about faults offshore the same way, but do you think these kinds of faults have consistent long-term slip-rates like we tend to assume for onshore faults. 00:52:24.000 --> 00:52:34.000 Seems like they would. They're earthquake faults but we probably don't know as much about them as we know about other faults. 00:52:34.000 --> 00:52:47.000 Like we know a lot about the Hayward Fault. We know that, you know, it does a seismic and seismic you know and so and, you know, uh. 00:52:47.000 --> 00:52:56.000 Clara's work showing that there's a seismic slip on the Mendocino fault. So it looks like the Mendocino fault also behaves like that. 00:52:56.000 --> 00:52:57.000 So, yeah. 00:52:57.000 --> 00:53:02.000 What? 00:53:02.000 --> 00:53:12.000 You know, Jimmy, I have a question for you. 00:53:12.000 --> 00:53:13.000 Yeah. 00:53:13.000 --> 00:53:30.000 If I could use Jimmy, the familiar. There was a discussion in the chat about clipping and so actually I have two questions can you have two questions and Andy or whoever discussed the clipping issue And you stayed on scale in the magnitude seven. 00:53:30.000 --> 00:53:37.000 And also, could you elucidate us on switching from strain to particle velocity? 00:53:37.000 --> 00:53:42.000 And why you do that. 00:53:42.000 --> 00:53:57.000 Yeah, so the clipping issue with DAS is like a big problem that A lot of people are trying to deal with And so we're using phase interferometry And so we're measuring the difference in two-way travel time between backscattered light 00:53:57.000 --> 00:54:06.000 And so if that change is too fast, then you have this 2-Pi error where you get phase wrapping or phase wrapping issue. 00:54:06.000 --> 00:54:11.000 And so that can lead your measurement into 00:54:11.000 --> 00:54:22.000 regimes that are very difficult to interpret and so when this clipping will happen is dependent on properties of the interrogator or like parameters that you set. 00:54:22.000 --> 00:54:38.000 And so you have to set these things carefully. And so this is a big problem for earthquake early warning with bass because we want to record big events But we're concerned that big events will clip interrogator records. 00:54:38.000 --> 00:54:45.000 It was good news that the magnitude 7 didn't clip our machine and so We were quite happy about that. 00:54:45.000 --> 00:54:48.000 And sorry, what was the second question? 00:54:48.000 --> 00:54:57.000 Well, did you know how close it came to clipping? I mean, is that a I mean, were you well within not clipping or where you write it? 00:54:57.000 --> 00:55:07.000 We were close, so it was within a factor of 10. So there's an equation has been pretty well defined with some empirical values in there. 00:55:07.000 --> 00:55:13.000 And so when we compute the peak strain that we recorded and compared it to where we would expect to have clipped. 00:55:13.000 --> 00:55:28.000 It's like a factor of five smaller. 00:55:28.000 --> 00:55:29.000 Interesting comment from Bob in the chat. Sorry. 00:55:29.000 --> 00:55:32.000 Judy, you got anything? 00:55:32.000 --> 00:55:34.000 Go ahead. Go ahead and answer it then. Okay. 00:55:34.000 --> 00:55:42.000 No, I can't answer it, but it'd be great for Bob to read his comment. 00:55:42.000 --> 00:55:46.000 If he can unmute himself. I'll read it for you, Bob. 00:55:46.000 --> 00:55:50.000 I'm unmuted, but… 00:55:50.000 --> 00:55:57.000 But this is not my what I wrote. This is what Bob wrote. A little bit of geologic context for the Fickle Hill event. 00:55:57.000 --> 00:56:21.000 Basement is Franciscan central belt, much but of importance is that Fickle Hill includes some unusual Silic volcanic rocks that have fission track ages, only ages from zircons that are oligocene, which does not fit slab window model. That's interesting. And are likely more related to young slab structure, being the Fickle Hill volcanics with uranium lead 00:56:21.000 --> 00:56:27.000 Would be very interesting for relating to Cascadia history. I don't know if anyone else has anything to add to that. 00:56:27.000 --> 00:56:30.000 That's important context, I think. Yeah, Bob. 00:56:30.000 --> 00:56:39.000 Oh, that, Bob. 00:56:39.000 --> 00:56:45.000 Yeah, I don't think you could write that, Bob. I mean, the first Bob, the silent Bob. 00:56:45.000 --> 00:56:46.000 Oh, wait, no. Yes. 00:56:46.000 --> 00:56:56.000 Okay, I've unmuted here. Can you hear me? Okay. Yeah, I was trying to find… I'm not familiar with this new Zoom screen here. 00:56:56.000 --> 00:57:17.000 Yeah, yeah, I actually have sort of wanted to talk to Bob and maybe Harvey about um going to the digging out the old granite Butte quarry just east of the university there because that's where these rocks that Charlie Meyer 00:57:17.000 --> 00:57:34.000 Many years ago did a master's on it at Humboldt and He did fish and track work on the zircons out of out of these rocks and they're very fresh There's a quarry in there and it would be very interesting 00:57:34.000 --> 00:57:56.000 To get a uranium lead age on those zircons. They have somewhat unusual As I recall, I think… amphibles. And it's a very unusual rock. It would make a nice study for a good student study and it would be very interesting with regard to 00:57:56.000 --> 00:58:20.000 Where it fits in the Cascadia context because these Volcanics are like you know right at the right at the continental margin, they're way too to be ARC related and But they're also a way too old to old be easily relatable to the slab window model. 00:58:20.000 --> 00:58:30.000 You know, they're 30, 35, 37 million years old. So, and it looks like, you know. 00:58:30.000 --> 00:58:36.000 The Fickle Hill event was like right under that area. 00:58:36.000 --> 00:58:39.000 It kind of grabbed my attention. 00:58:39.000 --> 00:58:40.000 Yeah, actually the fickle Hill event is right under my damn house. 00:58:40.000 --> 00:58:44.000 Interesting. 00:58:44.000 --> 00:58:46.000 It's your fault. 00:58:46.000 --> 00:58:50.000 Sorry, Bob. 00:58:50.000 --> 00:59:04.000 Pegging can you can you… say something about how confident you are in the mechanism that we arrived at. 00:59:04.000 --> 00:59:13.000 So… So it's impossible that it be a strike slip earthquake. 00:59:13.000 --> 00:59:25.000 And… mechanism, I don't think changes terribly much depending on whether it's a little bit shallow or a little bit deeper. I mean. 00:59:25.000 --> 00:59:42.000 One of the interesting questions is our hypocenter is relatively shallow at 12 kilometers and the the best fit mechanism is much but all of these earthquakes for the mechanism, it's the point of moment release, which is very different we know from the 00:59:42.000 --> 00:59:58.000 Hypocenter. And it's It all looks consistent to me in the location process. We use the p and the The p-values and the S minus p values for the stations that we had. 00:59:58.000 --> 01:00:16.000 Part of the nonlinear location process is that you can have a… You associate an uncertainty with each P and S wave and the uncertainty for the P waves was plus or minus a second. For the S waves, it was plus or minus two seconds. 01:00:16.000 --> 01:00:33.000 And so then you go through and you have all of your stations and you calculate all of the various possibilities where could this earthquake be if I change the p-value by of all of each of them by so much of them so much 01:00:33.000 --> 01:00:46.000 And you try all of this possibilities. And the result is you get this cloud of possibilities, which are the red dots And the center of those possibilities is just east of Arcata. 01:00:46.000 --> 01:00:57.000 And so I'm convinced that this is about the best you can do until somebody comes up with something better. 01:00:57.000 --> 01:01:04.000 Sue, you just made a comment. Shuhuff, would you like to follow up on that? 01:01:04.000 --> 01:01:11.000 Well, I've talked with Peggy and the group before. There is an indication that it was a somewhat high stress drop. 01:01:11.000 --> 01:01:24.000 Event, start my video yeah so i can Which doesn't really fit preconceptions about plate interface events. 01:01:24.000 --> 01:01:43.000 But I've been, you know, I'm the one who knows the least about tectonic interpretation. I've been persuaded that the plate interface event is the the preferred solution. So it is a little bit of a a question in my mind, but the macro seismic data 01:01:43.000 --> 01:02:01.000 To support both the location and the depth. And if anything, it might push the depth a little bit deeper but it it's really not consistent with a more shallow source because you would have seen higher intensities near the field near in the near field. 01:02:01.000 --> 01:02:13.000 And there were people. Out there. We have the reports of laurie talked about. So if you had a really shallow fall with magnitude six and a half. 01:02:13.000 --> 01:02:21.000 I think you would have seen higher intensities according to all the intensity prediction equations. 01:02:21.000 --> 01:02:22.000 So this tool. 01:02:22.000 --> 01:02:32.000 Yeah, and… Sorry, the magnitude So when Doug was doing the mechanism, Doug Gregor was doing the mechanism Because everybody did stuff except for me. 01:02:32.000 --> 01:02:53.000 I only digitized everything. When Doug Dragger was doing the mechanism, he used 01:02:53.000 --> 01:03:00.000 And it does hang together with the with the intensities so 01:03:00.000 --> 01:03:15.000 Switching presenters clara I was wondering if you had the chance to discuss the clusters that you guys located in your paper uh with Bo. 01:03:15.000 --> 01:03:25.000 And if you could say anything about those clusters and maybe fluid 01:03:25.000 --> 01:03:28.000 Do you have a comment on that? 01:03:28.000 --> 01:03:34.000 Yeah, okay. So if you're referring to Bo's talk yesterday. 01:03:34.000 --> 01:03:39.000 The 2020… Yeah. 01:03:39.000 --> 01:03:40.000 Yeah. 01:03:40.000 --> 01:03:58.000 About the Fortuna cluster. So I do see that the location of the Fortuna cluster was active in the 2022 Ferndale Aftershock sequence and it was it was pretty close to the um subduction interface, but we don't think it was on it. We do think that it was within 01:03:58.000 --> 01:04:17.000 The Gorda Plate. So we do see it but I don't see anything out of the ordinary. You know, I haven't had a chance to compare my catalog against Bose more directly, like do a direct comparison yet because her results are very new 01:04:17.000 --> 01:04:22.000 But yeah, that would be interesting to do. 01:04:22.000 --> 01:04:23.000 I was wondering, didn't you… Judy. 01:04:23.000 --> 01:04:43.000 Can I add some mute? Connect at some point actually i look at the uh compared the Fortuna cluster and also the aftershock sequence of the Ferndale earthquake So we found that actually the strike direction of the Ferndalendar earthquake is kind of uh 01:04:43.000 --> 01:04:53.000 Has some angle with the Fortuna cluster, it's kind of conjured gate And at the Ferndale earthquake is a little bit north of the Fortuna cluster. 01:04:53.000 --> 01:05:09.000 And for the but i think for at depth they are quite close but they have different strike directions. 01:05:09.000 --> 01:05:21.000 We had a… a couple of people talking at once for a second there. Did somebody want to speak and ask a question. 01:05:21.000 --> 01:05:22.000 Anyone else? 01:05:22.000 --> 01:05:48.000 Oh, I guess I just wanted to ask a naive question getting off topic so um You know, I mean, is there a consensus about what happened, like where the rupture was in 1992 you know was it on the overriding North American plate or was it on the subduction interface? You know, I thought it was on the overriding North American plate because that's what I've been reading everywhere but 01:05:48.000 --> 01:06:00.000 Apparently there's debate. And so, you know, maybe that's a hotly controversial topic. But I was just curious about what the status of that is. 01:06:00.000 --> 01:06:05.000 Robert? 01:06:05.000 --> 01:06:15.000 Yeah, Claire, I think it's in debate right now since this last earthquake and With every earthquake, we learn something new, right? 01:06:15.000 --> 01:06:26.000 And um it was interesting in your talk, you showed the Oppenheimer paper and Dave told me that the warranty run out ran out on that paper. 01:06:26.000 --> 01:06:35.000 But it was interesting to see the cartoon that was drawn in 1992 And it's an interplate earthquake. 01:06:35.000 --> 01:06:52.000 In other words, it was between Gorda and North America. And then over time, we realized that that plane the plane kind of moved around in 3d And it moved away from the top of the garter plate into the North American plate. 01:06:52.000 --> 01:06:59.000 So there's separation. Of the 92 and the top of the Gorda Plate. 01:06:59.000 --> 01:07:19.000 And then this last magnitude seven earthquake the aftershocks, which I showed Keith Newsom said cartoonerie I showed it in the cartoon that the earthquakes go all the way up to that 92 surface again. 01:07:19.000 --> 01:07:24.000 And that surface projects exactly to the trench. 01:07:24.000 --> 01:07:50.000 So… with all that arm waving and speculation um it's still, I think we're still all debating it is anyone that works there and um And with Jimmy's and the new projection of the DOS fiber experiment getting bigger and bigger and covering more ground. 01:07:50.000 --> 01:07:55.000 And being a bigger being, maybe some of this stuff will be resolved. 01:07:55.000 --> 01:08:01.000 In my lifetime. 01:08:01.000 --> 01:08:02.000 Yeah, so I think that the 92 surface is the mega thrust. 01:08:02.000 --> 01:08:07.000 Yeah, okay, okay, thanks for that. Yeah, it looks like 01:08:07.000 --> 01:08:13.000 that's my bias, because again 01:08:13.000 --> 01:08:24.000 I'm a geologist by nature. So if you follow a reverse fault from the trench down under North America, what are you following? 01:08:24.000 --> 01:08:35.000 Well, the question remains, you know, like it projects to the trench, but, you know, there's lots of variability in what you would project using your eyes. 01:08:35.000 --> 01:08:38.000 So I'm not sure we need to see. 01:08:38.000 --> 01:08:43.000 Well, the earthquakes, it doesn't just project. The earthquakes go to the trench. 01:08:43.000 --> 01:08:44.000 Yeah, anyway. 01:08:44.000 --> 01:08:59.000 Yeah. I look forward to your talk at next year's Northern California workshop meeting where we see that presented. 01:08:59.000 --> 01:09:12.000 Jimmy, I had a question and this is switching Totally, topics. But because I only somewhat understood you were doing. 01:09:12.000 --> 01:09:33.000 But you were… You've used it for evaluating the subduction zone and you talked about subduction zones, but you're showing the map of available fiber optic tables you showed all along the coast of california and i'm wondering about how you use it or what other kinds of other 01:09:33.000 --> 01:09:40.000 Questioned you. Consider answering using this methodology. 01:09:40.000 --> 01:09:45.000 Yeah, so it's a it's a pretty flexible. 01:09:45.000 --> 01:09:55.000 And seismology, really anything that'll generate a reflection or a converted phase will produce these phases on a fiber optic cable. 01:09:55.000 --> 01:10:13.000 And so you can apply the same thing to all of those things which are It can range from like a sedimentary basin depth to the Moho And I actually did this, the first study on this was, I was looking at the Moho and so 01:10:13.000 --> 01:10:25.000 We were able to map the moho through Southern California. And also Crustle falls if they offset enough mythology and you get this strong impedance contrast there, then you can image crustal faults as well. 01:10:25.000 --> 01:10:37.000 So anything you can imagine that generates some impedance contrast in the lithosphere you can use this for. 01:10:37.000 --> 01:10:42.000 Can you speak on the expansion of the cable or the fiber optic? 01:10:42.000 --> 01:10:51.000 Optic network. Are you talking with Andy and Jeff? What are… what are your guys ideas on the expansion? 01:10:51.000 --> 01:10:54.000 Going forward if if there's money 01:10:54.000 --> 01:11:17.000 Yeah, so there is a planned expansion So we're really leveraging So California is building this middle mile network to get high-speed internet to more of California. And so We're working with people deploying these fibers to get leases on fibers that are more extensive. 01:11:17.000 --> 01:11:37.000 And so just this year, we should be probing If you guys know where Highway 299 is, maybe this audience knows that that goes sort of down dip of the subduction interface to the east And then there's also a fiber that's going around Humboldt bay and the other way and so 01:11:37.000 --> 01:11:42.000 We uh we have leases on both these fibers and we should be probing them quite soon. 01:11:42.000 --> 01:11:49.000 And so that'll make the array Or that'll upgrade the array from 15 kilometers to 200 kilometers. 01:11:49.000 --> 01:12:07.000 Yeah. Any other fibers that are available, I think we're interested in. It's kind of hard to make the connections that you need to be able to probe these fibers, but Yeah, if you guys know anyone or have any connections, we'd be interested. 01:12:07.000 --> 01:12:16.000 So extending out 299 for those people that don't know That's connecting Arcata to Redding. 01:12:16.000 --> 01:12:33.000 So will that fiber then that would give a lot of information of the down dip. In other words, it'll go over the top of the the slow earthquakes and slow earthquakes the transition to totally ductal. 01:12:33.000 --> 01:12:34.000 So, yeah, so that'll be pretty exciting. 01:12:34.000 --> 01:12:43.000 Yeah, yeah, yeah. It is very exciting. Yeah. And I mean, we'll see how effective it is, sort of what the earthquake distribution is. 01:12:43.000 --> 01:12:56.000 So one limitation of this methodology is you need good earthquakes that generate converted phases Which a lot of subduction zones have but um So yeah, so I mean, the hope is that we can get sort of a sub kilometer 01:12:56.000 --> 01:13:09.000 Picture all the way down dip past the slow slip into the ductal zone. 01:13:09.000 --> 01:13:15.000 Any questions from the audience? 01:13:15.000 --> 01:13:20.000 Yeah, thereafter. You're up for a… 01:13:20.000 --> 01:13:28.000 Yeah, that was a really nice talk and some really amazing images. Thanks for that. I was wondering. 01:13:28.000 --> 01:13:33.000 How do you disentangle variations like undulations on the subduction interface? 01:13:33.000 --> 01:13:39.000 With velocity variations on the path from the interface back to your array. 01:13:39.000 --> 01:13:41.000 Back to your cable. 01:13:41.000 --> 01:13:52.000 Yeah, so there is some non-uniqueness there. So we use a 3D velocity model to try and correct for this as well as we can, but that's obviously not perfect. 01:13:52.000 --> 01:14:06.000 And so one of the benefits of So I don't know if it was clear in the talk, but we're using differential times of these converted phases and so Especially for PS, we're looking at P to PS differential times and so 01:14:06.000 --> 01:14:12.000 That removes the uncertainty coming from or removes a lot of the uncertainty coming from the source location. 01:14:12.000 --> 01:14:25.000 But yeah, and then the in the four arc crust there's still a trade-off between especially the VPDS ratio in the forearc crust And the slab depth, the absolute slab depth. 01:14:25.000 --> 01:14:28.000 And so that's hard. I mean, we can't do it perfectly. 01:14:28.000 --> 01:14:34.000 We use Hou's 3D model, which is particularly interested in VPBS variability. 01:14:34.000 --> 01:14:47.000 And so we do as best we can. But yeah, there's there's some inherent uniqueness there. 01:14:47.000 --> 01:14:48.000 I saw a comment in the chat, but I lost it. 01:14:48.000 --> 01:14:51.000 More cool. 01:14:51.000 --> 01:15:02.000 About the 1992 earthquake Now, being… in the overlying plate. 01:15:02.000 --> 01:15:07.000 Someone want to speak to that? I've lost the comment. 01:15:07.000 --> 01:15:12.000 I think we covered that. I think Claire just asked that one. 01:15:12.000 --> 01:15:16.000 Maybe. 01:15:16.000 --> 01:15:17.000 So. Yeah, there it is. 01:15:17.000 --> 01:15:21.000 Well, also how put in a figure showing the the plot. 01:15:21.000 --> 01:15:22.000 Whoa. 01:15:22.000 --> 01:15:40.000 From the paper and from the paper and It shows 92 clearly in the north america But it also shows that the top of Gorda and the 92 seismicity both point towards the trench. 01:15:40.000 --> 01:15:41.000 Yep. That's true. 01:15:41.000 --> 01:15:47.000 So… 01:15:47.000 --> 01:16:06.000 But there's lots of variation in where that line is drawn so So I don't know if the 92 seismicity, you know, if we can actually say it has enough control on the positions to say that it points towards the trench. 01:16:06.000 --> 01:16:16.000 So Jimmy, what kind of relief were you getting on the um the top of the Gorda Plate. Did I see like one kilometer? 01:16:16.000 --> 01:16:19.000 Between peak and Valley. 01:16:19.000 --> 01:16:32.000 Yeah, so it's like a few hundred meters. I'd say like 400 meters is what we resolve Assuming there's not some sort of really sharp variability in the velocity model that's contributing to it. 01:16:32.000 --> 01:16:46.000 And so that's like a reasonable height for a reasonable height subducting the seamount or any sort of topographic relief on the sea floor. 01:16:46.000 --> 01:16:59.000 Anyone else in the audience have… Questions for any of the speakers? 01:16:59.000 --> 01:17:09.000 Chris, what were you? Chris Goldfinger, yes. 01:17:09.000 --> 01:17:10.000 Oh, I didn't have a question. I just put a note in the chat. 01:17:10.000 --> 01:17:12.000 Oh, okay. 01:17:12.000 --> 01:17:18.000 Okay. 01:17:18.000 --> 01:17:25.000 Anyone else? 01:17:25.000 --> 01:17:44.000 Clara, I guess I'll just, this is just, this is probably like a dumb question but um you were showing the relationship between these two earthquakes but Do you think they I mean, do you think they actually have one? 01:17:44.000 --> 01:17:57.000 Well, other than just two random events that are in kind of the same area I mean… can you cocktail. 01:17:57.000 --> 01:18:18.000 Yeah, so in the you know in our the seismic record paper, we just kind of laid out the observations that stopped short of saying hey like These earthquakes are 2021 sequence like triggered 2022 and you know like we kind of stopped short of saying that and 01:18:18.000 --> 01:18:40.000 You know, I really think it's an open question. Why? When I wrote that paper, it kind of left me unsatisfied, you know, with more questions than answers And, you know, I was like, why did 2021 kind of jump from offshore to you know trigger this onshore event and then all these aftershocks 01:18:40.000 --> 01:18:49.000 Kind of stopped short of where the fern deal aftershocks a year later eventually occurred. 01:18:49.000 --> 01:19:09.000 And, you know, I think ultimately somehow They are related, but we didn't do any stress modeling or anything to show that there is a relationship so Yeah, I think that's an open question. Like, is there some kind of structural barrier or 01:19:09.000 --> 01:19:18.000 You know, was there just not quite enough in that particular location where there was a gap, like what exactly is going on? 01:19:18.000 --> 01:19:28.000 Like what's changing throughout the medium there's so many open questions. So we just kind of left it there. 01:19:28.000 --> 01:19:43.000 Okay, yeah, thanks. 01:19:43.000 --> 01:19:45.000 Anyone else? Peggy. 01:19:45.000 --> 01:19:51.000 Oh, Peggy. We can't hear her. 01:19:51.000 --> 01:19:58.000 And we can't hear you, Peggy. Maybe put it in the chat. 01:19:58.000 --> 01:20:18.000 I'm stuck. 01:20:18.000 --> 01:20:19.000 She's writing it. Yes. 01:20:19.000 --> 01:20:23.000 Peggy, can you nod if you're writing? Okay. 01:20:23.000 --> 01:20:25.000 All right, she just did. 01:20:25.000 --> 01:20:39.000 2021 onshore that earthquake and aftershock were a cloud. Did you find any structure that might have ruptured? 01:20:39.000 --> 01:20:40.000 I think that's probably for you, Claire. 01:20:40.000 --> 01:20:41.000 Yeah. 01:20:41.000 --> 01:20:54.000 Oh, okay. Yeah, you know. It was hard to find like a single definitive structure that was obvious. 01:20:54.000 --> 01:21:14.000 We generally kind of saw these aftershocks like from 2021, they were deeper but then deeper it seemed like they were like getting shallower you know kind of closer to eventually where the 2022 Ferndale aftershocks happened But yeah. 01:21:14.000 --> 01:21:19.000 There was no like single fault line that you can see. I think there were multiple structures. 01:21:19.000 --> 01:21:31.000 That were activated and you know after even after the relocation, it was hard to see anything obvious that was just like a single fall line. 01:21:31.000 --> 01:21:33.000 Now I can hear you. 01:21:33.000 --> 01:21:35.000 Yeah, yeah, I can hear you now. 01:21:35.000 --> 01:21:36.000 You're in use. 01:21:36.000 --> 01:21:37.000 Thanks. Did you hear me now? Good. So I changed it from microphone from the computer to microphone from the computer and now you can hear me. Sorry about that. 01:21:37.000 --> 01:21:46.000 Yeah. 01:21:46.000 --> 01:21:47.000 So… 01:21:47.000 --> 01:21:56.000 And just for Sarah, in case nobody read, it's called the Fickle Hill Earthquake now because it's under Fickle Hill. 01:21:56.000 --> 01:22:09.000 Bob lives on the side of Fickle Hill. And the Fickle Hill Fault runs as a strand along there, but it's at a different It's at a different azimuth than the mechanism that we got. 01:22:09.000 --> 01:22:10.000 But I wanted to know which way is it called Fickle Hill? 01:22:10.000 --> 01:22:13.000 Pickle Hill, that's what, yeah. 01:22:13.000 --> 01:22:18.000 Hawaii is Fickle Hill called Fickle Hill? I don't know, Bob, do you know? 01:22:18.000 --> 01:22:24.000 Back to the famous explorer peter fickle. I have no idea. 01:22:24.000 --> 01:22:29.000 I feel like that. You made that up. 01:22:29.000 --> 01:22:30.000 Right. 01:22:30.000 --> 01:22:47.000 I'm unmuted. Maybe I can ask a different question. How much would we have to bribe you to get you to look at a lot of the previous earthquakes and get us a long time look at high resolution locations. And do you accept vibes in the form of cookies? 01:22:47.000 --> 01:22:51.000 And how can we put this into our catalog. 01:22:51.000 --> 01:22:52.000 Yes. 01:22:52.000 --> 01:23:15.000 Um. Well, actually, I'd like to say that Bo Rong has already done most of the legwork to get us there, at least in terms of a high resolution deep learning enhanced earthquake catalog that's relocated. So, you know, and in our talk yesterday, she showed some really amazing results where you can see a lot of fault structure. So I think 01:23:15.000 --> 01:23:35.000 That would be a great starting point for a much more comprehensive and systematic study of kind of the history of earthquakes um and the Mendocino area, you know, maybe that could be a starting point for uh CFM and you know 01:23:35.000 --> 01:23:52.000 For the Mendocino area, you know, and we got to have the subduction, you know, interface and we got to have the main plate boundaries in place. But, you know, perhaps we can have also have a lot of these other very active you know fault structures interest labels 01:23:52.000 --> 01:24:08.000 That do create damaging earthquakes. So I think, and then once we have this felt model, we can use this as the basis for kind of stress modeling, stress interaction, earthquake interactions, you know, like the talk that Vulcan gave um 01:24:08.000 --> 01:24:27.000 Yesterday saying, hey, the, you know, the Mendocino magnitude 7 last month um stress the tip of the northern San Andreas felt like i think it'd be interesting to do more like earthquake interaction type studies and then You know, I think there's a lot of 01:24:27.000 --> 01:24:40.000 Kind of just open ground for dynamic rupture models or earthquake cycle models i think it's going to be incredibly complicated but incredibly complicated but I think there's a lot to learn. 01:24:40.000 --> 01:24:56.000 That's going to be a big group effort, you know, not just something that I do, but I think really leveraging what everybody here has done and kind of putting our knowledge together, I think that'll really help us understand Mendocino going forward. 01:24:56.000 --> 01:25:00.000 Oh, and I also want to like emphasize the fiber optic dash arrays. 01:25:00.000 --> 01:25:15.000 That are, you know, they're currently recording continuously in the the Arcada array. And then Jimmy said there's going to be expansion And I think we're going to learn so much more from that as well. 01:25:15.000 --> 01:25:23.000 I think the future is very promising, as I said in my talk to learn 01:25:23.000 --> 01:25:33.000 So Bo, Jimmy, the Bay Area Fault modelers, the Skek fault modelers. Claire has thrown down the gauntlet and I'll provide the cookies. 01:25:33.000 --> 01:25:38.000 Yeah, so I'm kind of offloading the responsibility from myself onto everybody else. 01:25:38.000 --> 01:26:05.000 Who knows so much more about tectonic history than I do. I think we really need kind of a long-term view To me, it just felt like I studied 2021 and 2022 but that's really not enough. You know, you really got to like look at the past history of where all these earthquakes have happened in Mendocino. 01:26:05.000 --> 01:26:09.000 All right, well, if… Oh, Harvey. 01:26:09.000 --> 01:26:12.000 Okay. 01:26:12.000 --> 01:26:15.000 Harry, take it away. 01:26:15.000 --> 01:26:24.000 Yeah, well, I just want to say to follow up what Claire said one aspect that hasn't been mentioned. There's lots of different mentions. 01:26:24.000 --> 01:26:35.000 But our knowledge of onland faulting in the Mendocino triple checking area and How that could contribute to a cfm Most of this is great literature, but it's important great literature. 01:26:35.000 --> 01:26:48.000 When the nuclear plant was evaluated by Woodruffly and published lengthy report was in 1980. There was a tremendous amount of seismicity and borehole data that came out of that. 01:26:48.000 --> 01:27:06.000 And that information relates directly to how structures such as the Little Salmon Fault, which is the only published fault in terms of slip rates, how that fault transitions to the offshore There's also other quite active structures south of Little Salmon Fault. 01:27:06.000 --> 01:27:18.000 In theses and great literature. And I just would encourage it If a serious effort is going to make to a cumulative model in this area that, in addition to the wonderful recent seismicity. 01:27:18.000 --> 01:27:34.000 That a close look is taken at the the onland geology both in on The few have published… reports, really sort of quality grade literature. 01:27:34.000 --> 01:27:35.000 Absolutely. 01:27:35.000 --> 01:27:42.000 That's true, Harvey. There is a lot of information out there on our surface faults. 01:27:42.000 --> 01:28:02.000 And one of the things that was pointed out in these earthquake sequences is the contrast when people look at these intraplate earthquakes, they cut across the surface faults that jay showed and other people showed up It's a really interesting area and you got to know where you are 01:28:02.000 --> 01:28:17.000 What plate you're sitting on. And to understand the falls. A lot of people have a problem imagining a fault under North America within a subducting plate. 01:28:17.000 --> 01:28:29.000 Very different than what's very different than what all the faults is harvey has mapped for years and Bob McLaughlin over there Gary. 01:28:29.000 --> 01:28:34.000 Are all all different. 01:28:34.000 --> 01:28:40.000 Yeah, can I raise my hand here? This is Bob McLaughlin. Do you hear me? 01:28:40.000 --> 01:28:43.000 Yes, we hear you. 01:28:43.000 --> 01:29:02.000 Yeah, I just wanted to second Harvey's comments with regard to it would probably be at this point in time with all this new high-end seismicity you guys are doing up there in the triple junction area and seeing different clusters 01:29:02.000 --> 01:29:12.000 Related to different structures. You… There's some recent work done. 01:29:12.000 --> 01:29:29.000 On the Mendocino. Fracture zone by the law of the sea people there is a law there is a… document that's a document that that I think was published by the State Department. 01:29:29.000 --> 01:29:41.000 That's out there. That looks in great detail at the history of the Mendocino bridge, fracture zone, geomorphology, bathymetry. 01:29:41.000 --> 01:30:02.000 And Doug Wilson was involved in this study and we have looked at the relationship of the Mendocino Ridge with relation to, for instance, the king range And there's some very important relationships of the Mendocino Ridge. 01:30:02.000 --> 01:30:16.000 To the geology of the King Range that link rocks in the king range actually to rocks on Mendocino ridge For instance. 01:30:16.000 --> 01:30:29.000 With regard to Harvey's comment about, yeah, let's look more closely at the onshore geology and faulting Now, in relation to this new data set. 01:30:29.000 --> 01:30:35.000 And then and then the data that's um that's coming out. 01:30:35.000 --> 01:30:54.000 I think it's very important to take a look at what's also gone on in the last 20 years with the onshore geology and try to try to make some sense with regard to that geology. 01:30:54.000 --> 01:30:58.000 Okay. Interesting. 01:30:58.000 --> 01:31:09.000 Thanks, everybody, for that. I think we're out of time and I don't see any hands raised, so we might want to wrap it up now. 01:31:09.000 --> 01:31:13.000 Sarah, probably sending it to you. 01:31:13.000 --> 01:31:21.000 Finally finish on time. 01:31:21.000 --> 01:31:28.000 That was amazing. Thank you, everyone. Hey, I don't have anything to add to that. We have learned a lot. 01:31:28.000 --> 01:31:38.000 Everybody, please have a restful evening and come back bright and early tomorrow, same bath time, same bat channel. 01:31:38.000 --> 01:31:44.000 And enjoy yet more Northern California hazards goodness. Thank you.