WEBVTT Kind: captions Language: en-US 00:00:02.389 --> 00:00:08.920 Okay. Thank you, everybody, for being here. [laughter] 00:00:08.920 --> 00:00:12.860 I know it’s been really crazy with the move, so I appreciate 00:00:12.860 --> 00:00:17.540 that all of you came here to watch Christine’s talk. 00:00:18.000 --> 00:00:23.699 There’s been many, many emails about the move with lots of information. 00:00:23.699 --> 00:00:28.019 So the main announcement with that regarding the seminar 00:00:28.019 --> 00:00:31.890 is that there will be a conference room in Moffett 00:00:31.890 --> 00:00:35.149 for us to have seminars there, but it’s not ready yet. 00:00:35.149 --> 00:00:41.260 So, for the time being, all of the seminars will still be here. [laughs] 00:00:41.260 --> 00:00:46.340 So please come back here when the seminars happen. 00:00:46.340 --> 00:00:48.780 And next week’s seminar will be by Evan. 00:00:48.780 --> 00:00:52.940 So especially come back for that because all of you know him. [laughs] 00:00:52.940 --> 00:00:56.560 And, with that, I will hand the mic to Annemarie. 00:00:56.560 --> 00:00:58.680 - All right. Thanks, Jessie. Thanks for doing the logistics. 00:00:58.690 --> 00:01:01.570 So today I’m really excited to have Christine Ruhl here. 00:01:01.570 --> 00:01:05.670 She got her B.S. in geology from Radford University in Virginia, 00:01:05.670 --> 00:01:09.430 an M.S. from New Mexico Tech, and then did her Ph.D. at UNR 00:01:09.430 --> 00:01:13.070 on earthquake source physics and the statistics of earthquake swarms 00:01:13.070 --> 00:01:19.500 and microseismicity, particularly in eastern California and western Nevada, 00:01:19.500 --> 00:01:22.530 which is, I think, what she will mostly be talking about today. 00:01:22.530 --> 00:01:25.990 After her Ph.D., though, she spent a while at Berkeley as a postdoc 00:01:25.990 --> 00:01:30.850 working primarily on early warning and working on the geodetic algorithms 00:01:30.850 --> 00:01:35.340 doing some GPS work for development and testing for early warning. 00:01:35.340 --> 00:01:38.439 So she won’t be covering that today. If you want to talk to her about that, 00:01:38.439 --> 00:01:42.540 or any of this work, she has time – she’ll be around the rest of today. 00:01:42.540 --> 00:01:47.109 So find her and come chat. And with that, Christine. 00:01:47.109 --> 00:01:49.000 - Thanks, Annemarie. Is this working? 00:01:49.000 --> 00:01:50.760 - I think you’re good, yeah. 00:01:50.760 --> 00:01:55.060 - Yes. So, like Annemarie said, I’m going to be talking about 00:01:55.060 --> 00:01:59.240 an extension of some of the work I did in my Ph.D. and some 00:01:59.240 --> 00:02:03.000 new work that we’ve done on it. And particularly just this 00:02:03.000 --> 00:02:05.689 Mogul earthquake swarm, which is just a really rich and 00:02:05.689 --> 00:02:09.720 beautiful data set – a very shallow earthquake swarm. 00:02:09.720 --> 00:02:13.880 So I’m going to start the talk by going over sort of the regional tectonics 00:02:13.890 --> 00:02:18.069 in the Walker Lane and talking about some historical seismicity there as well 00:02:18.069 --> 00:02:20.959 as drawing some comparisons to the Ridgecrest earthquakes 00:02:20.959 --> 00:02:23.400 that just happened. And then I’ll focus on the 00:02:23.400 --> 00:02:28.760 previous work on Mogul and continue with my four-dimensional directivity 00:02:28.760 --> 00:02:33.700 analysis. So basically, I’ve looked at directivity through time. 00:02:33.700 --> 00:02:36.080 And in three dimensions. 00:02:36.090 --> 00:02:39.590 So that’s what we’re going to be talking about today. 00:02:39.590 --> 00:02:44.349 So the Walker Lane is a wide zone of discontinuous faulting on the eastern 00:02:44.349 --> 00:02:46.599 side of the Sierra Nevada Mountains. 00:02:46.599 --> 00:02:50.510 It’s highlighted by these two white lines in between there. 00:02:50.510 --> 00:02:53.379 And you can see there’s quite a lot of seismicity happening there. 00:02:53.380 --> 00:02:56.530 So this is 12 years of magnitude 4-plus earthquakes. 00:02:56.540 --> 00:03:01.880 The Ridgecrest events are here. 00:03:01.880 --> 00:03:07.080 And Mogul is up here, north of Lake Tahoe, in the Reno area. 00:03:07.099 --> 00:03:11.740 This zone accommodates about 10 millimeters per year of plate motion. 00:03:11.740 --> 00:03:17.470 And so it’s an overprinting of the Basin and Range with dextral slip. 00:03:17.470 --> 00:03:21.010 So all of these en echelon normal faults are being 00:03:21.010 --> 00:03:24.830 reorganized to accommodate strike-slip. 00:03:24.830 --> 00:03:27.400 And because of that, we get abundant microseismicity and 00:03:27.400 --> 00:03:30.520 quite a large number of earthquake swarms and sort of 00:03:30.520 --> 00:03:33.599 foreshock-main shock- aftershock sequences. 00:03:33.599 --> 00:03:37.769 So this is David Shelly’s relocations of the Ridgecrest earthquake. 00:03:37.769 --> 00:03:40.989 And I wanted to include it because, not only is it in the Walker Lane, 00:03:40.989 --> 00:03:45.670 but it had a foreshock about a day and a half before the main shock. 00:03:45.670 --> 00:03:49.610 And the fault plane appears to have these big holes in it. 00:03:49.610 --> 00:03:54.360 And it’s been suggested that those holes might be patches of large slip in the 00:03:54.360 --> 00:04:01.519 main shocks – so in the magnitude 6.4 foreshock and then the 7.1 main shock. 00:04:01.519 --> 00:04:06.630 And this seems to be – it’s been observed in a lot of places for larger 00:04:06.630 --> 00:04:09.440 earthquakes and things like that, but it seems to be a very common 00:04:09.440 --> 00:04:13.260 feature in the Basin and Range and in the Walker Lane. 00:04:13.260 --> 00:04:17.890 So here’s an example of the 1986 Chalfant sequence on the left. 00:04:17.890 --> 00:04:22.500 And you can see this sort of hole where the aftershock – where the 00:04:22.500 --> 00:04:27.640 aftershocks are occurring around the rupture of the main shock. 00:04:27.650 --> 00:04:30.530 And then this is the 1984 Round Valley sequence. 00:04:30.530 --> 00:04:35.640 And both of these also had foreshocks either a few seconds before or about 00:04:35.640 --> 00:04:39.200 a day before the main shock. And again, here we see this sort of 00:04:39.200 --> 00:04:43.100 Mogi doughnut behavior – this hole with seismicity around it. 00:04:43.100 --> 00:04:47.910 And actually, in the Round Valley, Ken Smith also found a correlation 00:04:47.910 --> 00:04:51.560 of high stress drops around the boundary of the – of the inferred 00:04:51.560 --> 00:04:55.550 rupture zone in that bottom-right panel. And so I just want you to keep that 00:04:55.550 --> 00:04:58.680 in mind as we continue to go through this talk. 00:04:58.680 --> 00:05:02.110 Because, in my opinion, this sort of foreshocks and fault zone heterogeneity 00:05:02.110 --> 00:05:07.820 is a – is a self-similar process from laboratory to megathrust scale. 00:05:07.820 --> 00:05:11.480 So this bottom left image is of a laboratory experiment, 00:05:11.480 --> 00:05:14.980 and these are all the foreshocks that are occurring before the main shock 00:05:14.980 --> 00:05:19.900 rupture for that small cylinder. And then I have my Mogul events 00:05:19.900 --> 00:05:24.510 in here as a magnitude 5. And then the Ridgecrest is 6 and 7. 00:05:24.510 --> 00:05:29.470 And if you go all the way up to the Tokachi 2003 Japanese event, 00:05:29.470 --> 00:05:32.220 you see the same behavior on a megathrust scale. 00:05:32.220 --> 00:05:36.950 So not all earthquakes have foreshocks, but when they do, it’s not just the 00:05:36.950 --> 00:05:38.800 big ones or not just the small ones. 00:05:38.800 --> 00:05:40.980 I think that a lot of them have this behavior. 00:05:40.980 --> 00:05:43.990 And because there are so many more small earthquakes, it’s easier to 00:05:43.990 --> 00:05:50.560 investigate these kinds of rupture – these sort of properties on the fault zone. 00:05:50.560 --> 00:05:54.060 So I also wanted to draw your attention to some of the analysis that’s going on 00:05:54.060 --> 00:05:57.520 for the Nevada earthquake swarms that we’ve been studying over 00:05:57.520 --> 00:06:01.060 the last maybe five years. So some of the work I did in my Ph.D. 00:06:01.060 --> 00:06:04.660 has been continued on by Rachel Hatch, shown here on the bottom left. 00:06:04.660 --> 00:06:08.300 And she’s been publishing a few papers and has some in work about these 00:06:08.300 --> 00:06:12.530 different earthquake swarms that have happened in the Reno/Tahoe area. 00:06:12.530 --> 00:06:15.160 So Mogul is that yellow dot here. 00:06:15.160 --> 00:06:21.970 And then some of these others are the other sequences that she’s worked on. 00:06:21.970 --> 00:06:25.580 And I just wanted to point that out because, again, she sees, for these 00:06:25.580 --> 00:06:30.470 small magnitude 3 and 4 events, this sort of Mogi doughnut shape 00:06:30.470 --> 00:06:32.870 where all of these events are happening on the same fault plane. 00:06:32.870 --> 00:06:36.130 And she’s done directivity analysis using essentially the same exact method 00:06:36.130 --> 00:06:38.080 that I’m going to be talking about here. 00:06:38.080 --> 00:06:41.460 And she sees that the foreshock and the main shock for this event, 00:06:41.460 --> 00:06:45.120 which happened about seven minutes apart, rupture towards each other 00:06:45.120 --> 00:06:48.620 and cover the same kind of part of the fault plane a little bit. 00:06:48.620 --> 00:06:52.650 So this is the kind of detail that we’re getting about these 00:06:52.650 --> 00:06:55.910 little earthquake swarms, and I’m open to plenty of ideas 00:06:55.910 --> 00:07:00.580 of what we can do with this kind of high-quality data. 00:07:00.580 --> 00:07:04.990 And additionally, she’s working on this Virginia City sequence, which is sort of 00:07:04.990 --> 00:07:09.680 like a – if you folded a piece of paper into a tri-fold, and then sort of held it up 00:07:09.680 --> 00:07:13.420 with a vertical fault splitting two normal faults, that’s what you see here. 00:07:13.430 --> 00:07:19.180 So the blue and the yellow are normal faulting, or obliquely slipping planes, 00:07:19.180 --> 00:07:21.870 that are split by this vertical strike-slip plane. 00:07:21.870 --> 00:07:25.860 So we see lots of interaction of multiple fault planes in these swarms, 00:07:25.860 --> 00:07:28.030 which I think makes a lot of sense. 00:07:28.030 --> 00:07:32.140 So the Mogul 2008 swarm is a unique case because it was extremely shallow. 00:07:32.140 --> 00:07:34.450 It was between 2 and 5 kilometers’ depth, 00:07:34.450 --> 00:07:36.850 and it had over two months of foreshocks. 00:07:36.850 --> 00:07:41.190 During those foreshocks, people started feeling them because the events were 00:07:41.190 --> 00:07:44.380 between about 2 and 5 kilometers’ depth, and there’s a 00:07:44.380 --> 00:07:48.500 neighborhood right above this sequence. And it also was very convenient because 00:07:48.510 --> 00:07:52.520 it’s about 10 kilometers west of the Nevada Seismological Laboratory. 00:07:52.520 --> 00:07:57.060 So very rapidly, we put out seismic stations – temporary seismic stations. 00:07:57.060 --> 00:08:01.290 And this map shows about 10 of the stations in a 8-kilometer-square 00:08:01.290 --> 00:08:06.870 area right above the sequence. And the duration of those stations 00:08:06.870 --> 00:08:09.380 that are – that were put in temporarily are shown by 00:08:09.380 --> 00:08:13.070 those gray lines in the top-right plot right here. 00:08:13.070 --> 00:08:17.650 So our relocations are really good. We have a lot of coverage right on top 00:08:17.650 --> 00:08:22.320 plus a pretty dense regional network because it’s in the Reno/Tahoe area. 00:08:22.320 --> 00:08:24.820 So some previous work that’s been done in this sequence. 00:08:24.820 --> 00:08:28.420 John Anderson, who lived right above this sequence and felt these earthquakes 00:08:28.420 --> 00:08:32.419 in the Mogul neighborhood, he rapidly published this study 00:08:32.419 --> 00:08:35.020 about the exceptional ground motions of the main shock. 00:08:35.020 --> 00:08:40.840 So it was really, really strongly felt, and it had ground motions in excess of 1g. 00:08:40.840 --> 00:08:45.130 David von Seggern did some ground truthing of the main shock depth, 00:08:45.130 --> 00:08:49.180 and he found that it was around 3 kilometers’ depth, or 2.6 kilometers’ 00:08:49.180 --> 00:08:54.860 depth, I think is his preferred. So he reconfirmed how shallow it is. 00:08:54.860 --> 00:08:59.020 And, in addition, this main shock actually produced a significant amount 00:08:59.020 --> 00:09:03.110 of displacement, which was measured on GPS and InSAR. 00:09:03.110 --> 00:09:08.470 So Bell et al. in 2012 did an InSAR model, and he found that that white 00:09:08.470 --> 00:09:15.250 fault plane that you see on this figure is the fault model from InSAR. 00:09:15.250 --> 00:09:19.370 And you can see it matches really well with the seismicity relocations 00:09:19.370 --> 00:09:22.880 that I have done in previous work. And, in addition, he found that there 00:09:22.880 --> 00:09:26.450 was a significant amount of aseismic slip after the main event. 00:09:26.450 --> 00:09:29.730 Almost equivalent or exceeding the moment magnitude of the main shock 00:09:29.730 --> 00:09:33.650 itself. And that happened in about a month after the main shock. 00:09:33.650 --> 00:09:37.360 Because of the timing of windows for the InSAR, we can’t quite tell 00:09:37.360 --> 00:09:40.360 if there was any aseismic slip in the foreshock period. 00:09:40.360 --> 00:09:44.200 But, in my talk, I might suggest that there is. 00:09:45.160 --> 00:09:50.340 So I’ve done relocations for this complex, long-lasting sequence. 00:09:50.340 --> 00:09:55.060 And we found that the initial migration seems to fit a pore pressure diffusion 00:09:55.070 --> 00:09:58.210 related to fluids, which is often what people assign as the driver 00:09:58.210 --> 00:10:01.570 of these earthquake swarms. Fluids going through a highly fractured 00:10:01.570 --> 00:10:06.050 subsurface, and they cause seismicity. And that seismicity spreads with time. 00:10:06.050 --> 00:10:10.050 in addition, I did over 1,000 focal mechanisms by hand 00:10:10.050 --> 00:10:14.530 and about 10 moment tensors. And you can see that we have this 00:10:14.530 --> 00:10:18.620 really complex fault geometry between about 2 and 5 kilometers’ depth. 00:10:18.620 --> 00:10:21.940 It’s a sort of curved or kinked fault plane. 00:10:21.940 --> 00:10:25.090 And then these bottom two plots are showing sort of the aftershock 00:10:25.090 --> 00:10:29.060 distributions that are off the main fault. These are only foreshocks. 00:10:29.060 --> 00:10:31.480 So these are all the foreshocks, and they basically highlight the 00:10:31.480 --> 00:10:36.270 entire main shock fault plane before the main shock happens. 00:10:36.270 --> 00:10:39.440 So one thing to mention that I’m not going to talk about a lot 00:10:39.440 --> 00:10:43.530 is that I statistically clustered these by using Ilya Zaliapin’s method 00:10:43.530 --> 00:10:46.760 and looking at nearest neighbor space-time-magnitude distances. 00:10:46.760 --> 00:10:50.450 I randomized the catalog, and then I subdivided that into clusters. 00:10:50.450 --> 00:10:52.810 That’s important because I’m using that to investigate 00:10:52.810 --> 00:10:57.590 the events around each of the target events that we got directivity for. 00:10:57.590 --> 00:10:59.750 And those are based on the statistical clusters – 00:10:59.750 --> 00:11:03.980 based on their – this HypoDD catalog. 00:11:04.580 --> 00:11:07.930 Jansen et al. also did some modeling where he assumed fluids were coming 00:11:07.930 --> 00:11:11.780 up from some impermeable seal, either from recharge or 00:11:11.780 --> 00:11:15.320 maybe magmatic fluids at depth that have migrated upward. 00:11:15.320 --> 00:11:19.340 And he found, through modeling, that it could explain the initiation 00:11:19.340 --> 00:11:22.330 of the seismicity. And we also see some downward 00:11:22.330 --> 00:11:26.010 migration, which isn’t what you might expect for fluids that are 00:11:26.010 --> 00:11:29.120 coming upward, but he’s able to explain that with his model. 00:11:29.120 --> 00:11:33.230 What he’s not able to explain is this rapid migration of foreshocks in the 00:11:33.230 --> 00:11:37.400 days just prior to the main shock, which is what I’m going to focus on in this talk. 00:11:37.400 --> 00:11:42.900 So, for the first 60 days, the seismicity spreads with time. 00:11:42.910 --> 00:11:45.300 And then, all the sudden, it just goes crazy. 00:11:45.300 --> 00:11:48.070 And the number of events, the magnitude of events, 00:11:48.070 --> 00:11:53.070 and the migration of events all behaves very differently during this period. 00:11:53.070 --> 00:11:57.330 So I’m investigating that sort of acceleration into the main shock 00:11:57.330 --> 00:12:00.940 using the stress drop analysis that I’ve presented here previously 00:12:00.940 --> 00:12:03.140 and our new work with Rachel Abercrombie 00:12:03.150 --> 00:12:07.580 doing the rupture directivities of 87 of these events. 00:12:07.580 --> 00:12:12.340 So I just want to say again, with the statistical clusters we did, 00:12:12.340 --> 00:12:16.450 most of them reached their full extent really quickly, 00:12:16.450 --> 00:12:18.800 which is what these top two plots are showing. 00:12:18.800 --> 00:12:22.900 And the largest event in the sequence is usually first. 00:12:22.900 --> 00:12:27.240 And they follow Bath’s law, so the second-largest event 00:12:27.250 --> 00:12:30.950 is typically about 1 magnitude unit smaller than the first event. 00:12:30.950 --> 00:12:33.500 So we interpret that each of these sub-clusters that we identified 00:12:33.500 --> 00:12:36.490 is basically acting like a main shock-aftershock sequence. 00:12:36.490 --> 00:12:41.610 That’s true of all of the clusters except for this C18, which is the largest 00:12:41.610 --> 00:12:44.960 foreshock cluster before the main shock, which is when I think that there 00:12:44.960 --> 00:12:47.279 might have been some aseismic slip happening. 00:12:47.279 --> 00:12:50.380 And that’s because the seismicity here reaches the full extent of these 00:12:50.380 --> 00:12:54.040 two magnitude 4 co-main shocks, and then it spreads away from that 00:12:54.040 --> 00:12:57.420 with time at a rate of about 1 kilometer per hour. 00:12:57.420 --> 00:12:59.190 And that’s much too fast for fluid diffusion. 00:12:59.190 --> 00:13:03.880 It’s more likely that that rate previously has been attributed to aseismic slip. 00:13:03.880 --> 00:13:07.890 So I think potentially there was aseismic slip in a day and a half before the 00:13:07.890 --> 00:13:11.300 main shock, and because of the timing of InSAR, we aren’t able to resolve that. 00:13:11.300 --> 00:13:14.620 But maybe it’s not true. 00:13:14.620 --> 00:13:20.490 So I did a EGF analysis where I took each event, and I looked at a bunch of 00:13:20.490 --> 00:13:25.450 small events near it, and I used those to correct for the site and source and – 00:13:25.450 --> 00:13:30.300 or, I mean, the site and path effects. And it just leaves me with the source. 00:13:30.300 --> 00:13:35.220 And so we fit that source-time – we fit that source spectra – sorry – the 00:13:35.220 --> 00:13:39.500 spectral ratio here for corner frequency. And then we convert that corner 00:13:39.500 --> 00:13:44.040 frequency into stress drop using the moment magnitudes. 00:13:45.100 --> 00:13:47.640 We have extremely high-quality data here. 00:13:47.640 --> 00:13:51.430 These spectral ratios are shown for one event. 00:13:51.430 --> 00:13:56.470 They’re all greater than 0.8 correlation – cross-correlation. 00:13:56.470 --> 00:13:59.830 And so we stacked them and fit them, and we do a grid search to get 00:13:59.830 --> 00:14:03.610 uncertainties on that corner frequency, and then we get these stress drops. 00:14:03.610 --> 00:14:07.830 To make this study even better, or more robust, we did this 00:14:07.830 --> 00:14:10.850 independently for the P waves and for the S waves. 00:14:10.850 --> 00:14:13.490 So we get separate results, and they correlate really well. 00:14:13.490 --> 00:14:17.140 And we only kept the ones that were the highest quality. 00:14:18.100 --> 00:14:20.260 Regardless of having really high quality, we see 00:14:20.279 --> 00:14:24.090 a large scatter and a large variation. So this sort of green spread 00:14:24.090 --> 00:14:29.070 are my Mogul stress drops. And I want to just make a note 00:14:29.070 --> 00:14:32.680 that this suggests that these shallow earthquakes are not 00:14:32.680 --> 00:14:37.920 low stress drop compared, you know, to high – to deeper earthquakes. 00:14:37.920 --> 00:14:42.440 But what does this variation mean? You know, they’re normal. 00:14:42.440 --> 00:14:46.850 they’re plotting, you know, on scale with all of these other events and – 00:14:46.850 --> 00:14:51.089 or sequences, and all of the sequences have a huge amount of variation. 00:14:51.089 --> 00:14:52.720 So where does that variation come from? 00:14:52.720 --> 00:14:55.970 How can we explain it, and can we believe it? 00:14:55.970 --> 00:14:59.410 So I want to say yes, we can believe it. And that’s because, if you look at 00:14:59.410 --> 00:15:03.680 these two events, one is a stress drop of about 3 megapascals on the left. 00:15:03.680 --> 00:15:06.230 And one is a stress drop of about 30 or more megapascals. 00:15:06.230 --> 00:15:09.040 So that’s a very high stress drop. 00:15:09.040 --> 00:15:13.170 And if you look at them, all of the ratios are very good. 00:15:13.170 --> 00:15:17.080 This is over a very close distance – 250 meters. 00:15:17.080 --> 00:15:22.779 So it’s really only well-correlated, really similar EGFs that we’re using 00:15:22.780 --> 00:15:25.700 here – or empirical Green’s functions – to make that correction. 00:15:25.700 --> 00:15:28.960 And the corner frequencies of these two events, even though they’re both 00:15:28.960 --> 00:15:33.440 magnitude 3.1, could not be the same. And if you look at the waveforms on 00:15:33.440 --> 00:15:37.700 these two close-in temporary stations, you can see that the low stress drop 00:15:37.700 --> 00:15:40.840 event is sort of longer period than this high stress drop event 00:15:40.850 --> 00:15:43.350 if you look at the wave train. 00:15:43.350 --> 00:15:47.520 So we believe the variation, but why is it there? 00:15:47.520 --> 00:15:49.580 What does it mean? Where is it coming from? 00:15:49.580 --> 00:15:53.740 So we think it’s coming from the fault. 00:15:53.750 --> 00:15:57.580 That’s what we’ve proposed previously is that there’s variation or heterogeneity 00:15:57.580 --> 00:16:01.460 on the fault which is causing these variations in stress drop. 00:16:01.460 --> 00:16:07.040 And part of the reason that we think that is because of the pattern of 00:16:07.050 --> 00:16:10.970 seismicity that we see on the fault plane. So you can see all of these black circles 00:16:10.970 --> 00:16:14.830 are the foreshocks, and the gray are the aftershocks. 00:16:14.830 --> 00:16:18.140 And they’re highlighting this really dense ring, and there’s 00:16:18.140 --> 00:16:21.710 pretty much no seismicity within that ring. 00:16:21.710 --> 00:16:24.681 And so we think this is sort of a Mogi doughnut, and maybe that was a 00:16:24.681 --> 00:16:31.270 large area of slip during the main shock. When we plot the rupture dimensions 00:16:31.270 --> 00:16:34.560 here, colored by stress drop, you can see that we get a concentration of 00:16:34.560 --> 00:16:38.470 high stress drops around that ring. So I think that makes sense with it being 00:16:38.470 --> 00:16:42.110 sort of a locked patch where it’s trying to slip on the edges, but it can’t quite 00:16:42.110 --> 00:16:45.630 get there, and we’re getting a lot of stress in the transition zones between 00:16:45.630 --> 00:16:47.940 whatever other property is going on on the other parts of the 00:16:47.940 --> 00:16:51.910 fault on this patch. But maybe that’s true. Maybe it’s not. 00:16:51.910 --> 00:16:56.560 So we suggested previously that this might be a high slip patch, 00:16:56.560 --> 00:16:59.950 or an area of high-slip during the main shock. 00:16:59.950 --> 00:17:02.700 And it’s not re-ruptured in the aftershock period. 00:17:02.700 --> 00:17:07.280 And so that makes sense, and that fits this idea. 00:17:08.329 --> 00:17:10.800 So did the main shock rupture multiple patches? 00:17:10.800 --> 00:17:13.910 Well, our complex spectra that we have for the main shock suggests 00:17:13.910 --> 00:17:17.620 that the answer could be yes. We see this double-pulse source-time 00:17:17.620 --> 00:17:21.640 function in the P wave, and it’s a little less noticeable, but you sort of see 00:17:21.650 --> 00:17:25.130 complexity in the S wave source-time function as well. 00:17:25.130 --> 00:17:26.280 And when there’s multiple pulses, 00:17:26.280 --> 00:17:29.760 that suggests that it’s rupturing multiple patches. 00:17:29.760 --> 00:17:35.140 Eleven out of 148 events that we got stress drops for we identify as complex. 00:17:35.140 --> 00:17:39.929 And most of the largest events in the sequence are complex. 00:17:39.929 --> 00:17:41.620 So keep that in mind. 00:17:41.620 --> 00:17:45.560 So we tried to do a slip inversion, but the main shock clipped at a lot of 00:17:45.560 --> 00:17:49.760 the short-period stations, and it’s very, very high frequency, and we didn’t fit it 00:17:49.760 --> 00:17:54.470 well with synthetics, and I didn’t go through and do an EGF-based 00:17:54.470 --> 00:17:56.540 source inversion. So instead, we tried 00:17:56.540 --> 00:17:59.070 another approach since we have all these source-time functions, 00:17:59.070 --> 00:18:01.990 and we actually also calculated them azimuthally. 00:18:01.990 --> 00:18:05.080 So we’re looking at the rupture directivity to try to identify the 00:18:05.080 --> 00:18:08.250 fault plane and see which way each event ruptured. 00:18:08.250 --> 00:18:12.660 So the idea here is that, if you have a fault, and it ruptures in one direction, 00:18:12.660 --> 00:18:16.720 the source-time functions are going to be narrower, or shorter in duration, 00:18:16.720 --> 00:18:19.600 with a higher amplitude in the direction of propagation, 00:18:19.600 --> 00:18:20.950 or the direction of the rupture. 00:18:20.950 --> 00:18:26.300 And, on the sides, you’re going to have a lower amplitude with a longer 00:18:26.300 --> 00:18:30.340 duration. And in the back direction, you would have the longest duration 00:18:30.340 --> 00:18:36.720 and lowest amplitude. So this azimuthal variation of source- 00:18:36.720 --> 00:18:41.260 time functions can tell you about the directivity and about the fault plane. 00:18:41.260 --> 00:18:46.280 So we’re using this method developed by Prieto et al. in 2017, 00:18:46.280 --> 00:18:50.210 which is to stretch the time axis at each station azimuthally 00:18:50.210 --> 00:18:53.530 to find the best correlation with the other source-time functions. 00:18:53.530 --> 00:18:56.080 The reason we stretched the whole waveform is because it can be really 00:18:56.080 --> 00:19:00.360 hard to pick the start and end points. But if you just take this P wave, 00:19:00.360 --> 00:19:04.120 and then you stretch it a little bit, or squish it, I guess – compress it a bit 00:19:04.120 --> 00:19:07.780 to fit that one and then compress it a bit more to fit here, and then you fit the 00:19:07.780 --> 00:19:12.300 azimuthal variation of those stretching, you can actually get the line source – 00:19:12.300 --> 00:19:14.480 the 3D line source for this event. 00:19:14.480 --> 00:19:18.700 So we do that for the whole waveform for these source-time functions, and then 00:19:18.700 --> 00:19:22.450 we end up getting a rupture directivity, and it may or may not match the 00:19:22.450 --> 00:19:25.390 focal mechanism or the seismicity alignments that we have. 00:19:25.390 --> 00:19:27.380 So that’s the kind of thing that we’re trying to figure out. 00:19:27.380 --> 00:19:29.620 Are these rupture directivities aligning with the fault planes? 00:19:29.620 --> 00:19:33.660 Can they be used to identify which faults ruptured? That kind of thing. 00:19:34.310 --> 00:19:39.420 So, for each event, we search for the strike dip and rupture velocity 00:19:39.420 --> 00:19:42.470 that best fits the observed variation in stretching. 00:19:42.470 --> 00:19:46.100 We test it with three models – a line source – I mean, sorry – 00:19:46.100 --> 00:19:49.140 a unilateral source, where it starts at one end and ruptures completely 00:19:49.140 --> 00:19:53.720 to the other side, a bilateral, and then a 2-to-1 bilateral. 00:19:55.060 --> 00:19:57.620 And here’s an example of that fitting. 00:19:57.630 --> 00:20:00.150 So this is one of the Mogul magnitude 3 earthquakes. 00:20:00.150 --> 00:20:04.240 I have my P source-time functions, my S source-time functions. 00:20:04.240 --> 00:20:07.040 We measure them both and combine those results. 00:20:07.040 --> 00:20:10.240 And the red line is showing the modeled fit of the azimuthal 00:20:10.240 --> 00:20:12.800 stretching of those source-time functions. 00:20:12.800 --> 00:20:18.460 And then we do a grid search to find the 5% and 10% variance regions of the 00:20:18.470 --> 00:20:24.490 lowest or the best-fit rupture direction. And for this event, here’s the lower 00:20:24.490 --> 00:20:27.960 hemisphere fault planes, and here’s the upper hemisphere fault planes. 00:20:27.960 --> 00:20:31.960 and you can see that this rupture directivity is basically landing 00:20:31.960 --> 00:20:36.240 exactly on the fault plane in the upper hemisphere. 00:20:36.250 --> 00:20:39.540 So the Mogul directivity results. The majority of our events are 00:20:39.540 --> 00:20:44.420 best fit by unilateral directivity. But it is of note that the 2-to-1 bilateral 00:20:44.420 --> 00:20:48.000 results are often identical, and you can’t tell which was best. 00:20:48.000 --> 00:20:53.200 So for us, these 2-to-1, essentially we’re calling it unilateral. 00:20:55.480 --> 00:20:59.180 We did this for 87 earthquakes. So some of them didn’t have good 00:20:59.180 --> 00:21:03.470 enough azimuthal variation – or, azimuthal coverage, 00:21:03.470 --> 00:21:07.520 or they just weren’t well-fit. So we did 87 of them, and 36 00:21:07.520 --> 00:21:11.750 of those are before the main shock. Many of the line sources align with 00:21:11.750 --> 00:21:14.510 the clustered seismicity and the focal mechanism. 00:21:14.510 --> 00:21:17.780 And we observe a relationship between the rupture directions 00:21:17.780 --> 00:21:20.150 and the migration of foreshocks along the main shock fault, 00:21:20.150 --> 00:21:22.700 which is what the title of my talk was about. 00:21:22.700 --> 00:21:26.400 So, in order to get into this and show you guys our – what I think are 00:21:26.400 --> 00:21:30.830 really cool results, I want to sort of break down how I’m looking at this. 00:21:30.830 --> 00:21:35.860 So I think this is a new sort of way of looking at sequence evolution, 00:21:35.860 --> 00:21:39.600 and it’s a little bit complicated. So what we’re doing is we’re 00:21:39.600 --> 00:21:42.510 plotting a focal sphere here. And we plot the focal mechanism, 00:21:42.510 --> 00:21:45.460 if it exists, on the focal sphere, in both the upper hemisphere and 00:21:45.460 --> 00:21:48.299 lower hemisphere projections. So the black lines are the lower 00:21:48.299 --> 00:21:51.550 hemisphere, or the down – if you were looking down the plane. 00:21:51.550 --> 00:21:54.280 And the gray lines here are the upper hemisphere. 00:21:54.280 --> 00:21:57.110 In the center of the sphere is the event that we’re interested in – 00:21:57.110 --> 00:22:00.830 the target event. Then we plot the foreshocks on there as squares. 00:22:00.830 --> 00:22:03.210 And they’re a little bit light, but you can see them here, 00:22:03.210 --> 00:22:06.730 lying exactly on the fault plane. So the question we want to answer – 00:22:06.730 --> 00:22:10.440 well, did this event rupture towards the foreshocks? Away from the foreshocks? 00:22:10.440 --> 00:22:15.110 How did they interact? Was it even on the same plane as the foreshocks? 00:22:15.110 --> 00:22:18.210 So the next thing that we do is we plot the rupture directivity with 00:22:18.210 --> 00:22:23.560 the 5% and 10% variance regions. And you should note here, I plot 00:22:23.560 --> 00:22:27.590 an X if it’s in the upper hemisphere – so if it’s rupturing upward. 00:22:27.590 --> 00:22:31.340 And I plot a plus sign if it’s rupturing downward. 00:22:31.340 --> 00:22:35.320 And then we add the aftershocks. So this is the third plot here. 00:22:35.330 --> 00:22:38.220 And the aftershocks are circles. So the foreshocks are squares 00:22:38.220 --> 00:22:42.240 that are a little bit lighter. And the aftershocks are these circles. 00:22:42.240 --> 00:22:45.760 The foreshocks and aftershocks are also closed if they’re in the lower 00:22:45.760 --> 00:22:49.100 hemisphere and open if they’re in the upper hemisphere. 00:22:49.100 --> 00:22:53.970 So it’s – yeah, it’s a little complicated, but I think it’s worth it. 00:22:53.970 --> 00:22:57.570 So here’s our example event. And also it’s colored by time, 00:22:57.570 --> 00:23:00.250 and there will be a time scale bar on the plots that I’m going to show. 00:23:00.250 --> 00:23:04.190 So these go from red, being the – the darkest red would be the foreshocks. 00:23:04.190 --> 00:23:07.480 Here’s the red main shock, or target event, and then these 00:23:07.480 --> 00:23:10.880 other colors are showing the aftershocks through time. 00:23:11.520 --> 00:23:14.660 So this is the first phase of foreshocks, 00:23:14.660 --> 00:23:18.950 which we don’t have included in our clustered seismicity. 00:23:18.950 --> 00:23:21.970 So this wasn’t – didn’t behave like a main shock-aftershock. 00:23:21.970 --> 00:23:25.150 This is the period in the first couple of weeks where it was really getting going, 00:23:25.150 --> 00:23:27.520 and it was sort of spreading with time. 00:23:28.240 --> 00:23:32.020 Yeah, okay, let me take a minute. So this is a map [chuckles] up here. 00:23:32.020 --> 00:23:35.080 This is my stereonet that you’re going to see here. 00:23:35.080 --> 00:23:37.750 This is a cross-fault cross-section that’s plotted here 00:23:37.750 --> 00:23:40.300 from southwest to northeast. 00:23:40.300 --> 00:23:44.200 This is the fault-parallel cross-section, which is not shown because it kind of 00:23:44.200 --> 00:23:46.679 goes through this early phase of foreshocks before it moves on 00:23:46.679 --> 00:23:51.510 to the main shock fault plane. So, for simplicity’s sake, I left it off the map. 00:23:51.510 --> 00:23:55.770 The gray lines are the surface faults, which are all primarily normal faults. 00:23:55.770 --> 00:23:58.240 This sequence was primarily strike-slip. 00:23:58.240 --> 00:24:02.200 And this northwest-to-southeast line is this one here. 00:24:02.200 --> 00:24:04.830 So, for the first few events, we’re going to be looking at this 00:24:04.830 --> 00:24:07.309 fault plane, but this is where the cool stuff is going to happen. 00:24:07.309 --> 00:24:10.000 So look at this one as I go through these. 00:24:10.000 --> 00:24:11.950 So this period lasted about two weeks. 00:24:11.950 --> 00:24:16.340 It’s what I call Phase I foreshocks of unclustered seismicity. 00:24:16.340 --> 00:24:22.320 Included in this Phase I, we start to get some earthquake ruptures happening. 00:24:22.320 --> 00:24:25.320 Maybe the locations weren’t that good at this time because we had temporary 00:24:25.320 --> 00:24:29.700 stations that were coming in after we started feeling these magnitude 3s. 00:24:29.700 --> 00:24:32.580 But, more or less, you know, maybe. 00:24:32.580 --> 00:24:35.419 Maybe it ruptured two planes. Who knows? 00:24:35.419 --> 00:24:38.470 You know, and it’s showing that it ruptured down. 00:24:38.470 --> 00:24:43.710 But also note the quality. So I’ve labeled this as low-quality. 00:24:43.710 --> 00:24:45.309 So maybe you don’t believe that one. 00:24:45.309 --> 00:24:47.980 This one, however, we think was quite high-quality. 00:24:47.980 --> 00:24:53.880 And, again, the seismicity around the event, maybe not so convincing. 00:24:53.880 --> 00:24:57.420 As we go farther into the sequence, however, you’ll start to see that the 00:24:57.420 --> 00:25:02.080 correlation and the quality of the relocations and of the stress drops 00:25:02.080 --> 00:25:06.890 and everything is being improved by the addition of seismic stations. 00:25:06.890 --> 00:25:10.100 So, for this event, in map view, you can see that the directivity 00:25:10.100 --> 00:25:15.210 perfectly aligns with the seismicity. And, in my stereonet plot, 00:25:15.210 --> 00:25:18.669 you can definitely tell which fault plane this event is happening on, 00:25:18.669 --> 00:25:22.160 both from the seismicity and from the directivity. 00:25:22.160 --> 00:25:25.860 And we have it here in cross-section, and, eh, it looks pretty good. 00:25:25.870 --> 00:25:28.410 Maybe the depths are a little bit worse. 00:25:28.410 --> 00:25:30.380 So we keep going. 00:25:31.700 --> 00:25:35.580 And – okay, so this is the last one of the Phase 6 foreshocks. 00:25:35.580 --> 00:25:38.960 We’ve just gotten this fault fracture mesh highlighted. 00:25:38.960 --> 00:25:40.540 And if you’re interested in the fault fracture mesh, 00:25:40.540 --> 00:25:43.320 I talked a lot about it in my 2016 paper. 00:25:43.320 --> 00:25:45.520 Then we jump onto the main shock fault plane. 00:25:45.530 --> 00:25:47.600 And this is when things get pretty cool. 00:25:47.600 --> 00:25:51.130 So this sequence lasts for about 60 hours. 00:25:51.130 --> 00:25:54.560 And here’s where the events happen. 00:25:54.560 --> 00:25:58.090 So this directivity ruptures along the fault. 00:25:58.090 --> 00:26:02.100 And then all of these dark filled-in circles are happening down-dip, 00:26:02.100 --> 00:26:06.940 or below this event and most likely on the lower edge of this fault plane 00:26:06.940 --> 00:26:10.650 because they’re aligning on it. If we keep going, we see that 00:26:10.650 --> 00:26:13.380 this event ruptures there. And, again, in the cross-section, 00:26:13.380 --> 00:26:16.760 you can see it’s being – it’s aligned really well, both in map view, 00:26:16.760 --> 00:26:19.540 in the cross-section, and on this focal sphere. 00:26:19.540 --> 00:26:23.220 You can see that these events are happening on the down-dip edge 00:26:23.220 --> 00:26:27.200 of the – of the northwest-striking fault plane. 00:26:28.950 --> 00:26:32.480 So I’ve been talking a lot about patches in the beginning of this talk. 00:26:32.480 --> 00:26:35.309 And it’s going to appear right here where my mouse is. 00:26:35.309 --> 00:26:39.030 So all of the seismicity is starting to rupture around the patch. 00:26:39.030 --> 00:26:41.510 And the rupture directivities actually move around the patch 00:26:41.510 --> 00:26:44.760 and very rarely rupture into the patch. 00:26:45.970 --> 00:26:49.460 So if we keep going in the sequence, now we get another event. 00:26:49.460 --> 00:26:53.710 It matches really well. So those are all the foreshocks when I go back. 00:26:53.710 --> 00:26:55.799 All those light squares are the foreshocks. It happens 00:26:55.799 --> 00:26:58.780 right in the middle of them. It ruptures to the northwest. 00:26:58.780 --> 00:27:02.980 And then these aftershocks occur behind the rupture. 00:27:03.560 --> 00:27:06.960 And then it loops back down. So if you see the colors on the 00:27:06.970 --> 00:27:09.670 fault plane, it’s ruptured up and around, and now it’s 00:27:09.670 --> 00:27:12.620 rupturing back down and around. 00:27:17.120 --> 00:27:19.010 So I’ll pause here again. 00:27:19.010 --> 00:27:22.380 So this entire patch now – and you can sort of see a little hole – 00:27:22.380 --> 00:27:25.850 like this is an ellipse up there – is being ruptured. 00:27:25.850 --> 00:27:29.559 It’s a dipping plane that’s dipping towards you in this view. 00:27:29.560 --> 00:27:35.440 You can see in the cross-section here, it’s dipping down to the west. 00:27:35.440 --> 00:27:40.790 And it’s quite nicely defined. And so, if we keep going, we have a few 00:27:40.790 --> 00:27:47.600 more clusters that aren’t that intriguing, directivity or related seismicity-wise. 00:27:47.600 --> 00:27:51.280 But they continue to highlight the edges of this fault plane. 00:27:51.280 --> 00:27:54.920 Again, here we see this nice one. In the cross-section, it looks pretty good. 00:27:54.920 --> 00:27:59.250 In the stereonet, maybe not so nice. But there’s uncertainty in 00:27:59.250 --> 00:28:02.660 the focal mechanisms as well. So we’re trying to resolve all this focal – 00:28:02.660 --> 00:28:07.320 all this uncertainty by combining all of these robust observations. 00:28:07.330 --> 00:28:11.301 So, in the Phase II, we continue to rupture that patch. 00:28:11.301 --> 00:28:14.299 And then the seismicity jumps down to the southeast. 00:28:14.299 --> 00:28:18.990 It’s happening in this XSEC 1, or this Cross-section 1, which is 00:28:18.990 --> 00:28:22.360 more west of the cross-section plotted on the map. 00:28:22.360 --> 00:28:24.920 And that’s because the dip – or, the strike – I’m sorry – 00:28:24.920 --> 00:28:28.010 of the fault plane changes. And so, right around where 00:28:28.010 --> 00:28:31.350 my mouse is right now, the strike changes a little bit. 00:28:31.350 --> 00:28:36.120 So some of these events end up plotting on this guy instead. 00:28:39.240 --> 00:28:43.360 Yep. So we get some action over there, but no patch, necessarily, really being 00:28:43.370 --> 00:28:46.690 highlighted – not 100% sure what’s going on. 00:28:46.690 --> 00:28:51.960 And then we get to Phase 3. So this is the Cluster 18 – the one where 00:28:51.960 --> 00:28:56.220 we see 1 kilometer-per-hour migration, where I think that there was likely 00:28:56.220 --> 00:29:01.340 aseismic slip that was driving this really rapid migration of foreshocks. 00:29:01.340 --> 00:29:05.240 And they start to form, again, around that same patch 00:29:05.240 --> 00:29:08.320 that was ruptured by Phase 2. 00:29:09.660 --> 00:29:13.220 The seismicity in the stereonet seems to align quite nicely. 00:29:13.220 --> 00:29:16.240 I’m going to stop here because, again, you see there’s a lot of seismicity 00:29:16.240 --> 00:29:19.740 in the back direction. And it’s rupturing sort of away from those. 00:29:19.740 --> 00:29:22.350 And we see that a lot. I haven’t quantified it yet, 00:29:22.350 --> 00:29:25.160 but I would say there’s more cases where the aftershocks 00:29:25.160 --> 00:29:28.450 occur in the back direction than in the fore direction. 00:29:28.450 --> 00:29:32.940 What does that mean? Something to – something to investigate. 00:29:32.940 --> 00:29:37.700 So if we keep going along this, we see a lot of vertical ruptures as well, 00:29:37.700 --> 00:29:43.059 or not horizontal ruptures. So they tend to, a lot of times, have strong dips. 00:29:43.059 --> 00:29:45.160 And that tends to correlate with the fault plane. 00:29:45.160 --> 00:29:49.230 So, not only is there a change in the strike of the fault, there’s a change 00:29:49.230 --> 00:29:51.559 in the dip of the fault. So here you can see it’s sort of 00:29:51.559 --> 00:29:54.669 dipping to the southwest below about 4 kilometers. 00:29:54.669 --> 00:30:01.990 And it’s dipping to the northeast, north – or, upward of 4 kilometers. 00:30:01.990 --> 00:30:04.000 And so there’s a kink in the fault right here. 00:30:04.000 --> 00:30:06.799 And you could imagine a fault kink might get a lot of stress there. 00:30:06.799 --> 00:30:11.309 So keep that in mind as I play up through this sequence. 00:30:11.309 --> 00:30:15.620 So we see more directivity occurring, highlighting these patches, 00:30:15.620 --> 00:30:19.240 rupturing sort of in a figure 8 around the patches. 00:30:22.920 --> 00:30:26.870 And if you keep going – here is one where this rupture activity 00:30:26.870 --> 00:30:30.380 is fault-perpendicular. But if you look at the map view, 00:30:30.380 --> 00:30:35.039 you can see that maybe this location isn’t quite right, but all of this seismicity 00:30:35.039 --> 00:30:38.350 is associated with that event as the aftershocks of that event. 00:30:38.350 --> 00:30:44.090 So it is highly likely that this event did rupture that northeast-striking plane 00:30:44.090 --> 00:30:48.970 instead of the main fault plane. So we’re getting a lot of concentration 00:30:48.970 --> 00:30:52.840 of events in these sort of – you could think of them as step-overs or I would 00:30:52.840 --> 00:30:57.059 say more like little slight changes in the strike of the – of the fault plane, 00:30:57.059 --> 00:31:00.059 or these little bends in the fault. And those are primarily in this 00:31:00.059 --> 00:31:03.570 area here, where we get a lot of events concentrating, and here. 00:31:03.570 --> 00:31:06.350 And a lot of those directivities are fault-perpendicular. 00:31:06.350 --> 00:31:08.580 So if you looked at a map, you would say those directivities 00:31:08.580 --> 00:31:13.020 couldn’t possibly be right. But, in this 3D view, you can see 00:31:13.020 --> 00:31:15.870 that it is happening on one of the fault planes that it makes sense. 00:31:15.870 --> 00:31:19.659 And a lot of times, the perpendicular ones are rupturing either down or up 00:31:19.660 --> 00:31:25.460 these dipping fault planes. So in map view, it can be a little bit deceiving. 00:31:26.660 --> 00:31:30.840 So then that event is followed by an opposite rupture back, again in this 00:31:30.840 --> 00:31:35.460 second kink in the fault up here where we see another change in the dip. 00:31:35.460 --> 00:31:39.570 And those are the aftershocks of it. And again, we get another one 00:31:39.570 --> 00:31:44.020 rupturing fault-perpendicular, but back in the opposite direction. 00:31:45.560 --> 00:31:50.680 After C18, the remaining clusters are significantly shallower – 00:31:50.690 --> 00:31:53.600 not significantly, but they’re shallower than 4 kilometers. 00:31:53.600 --> 00:31:55.789 So they’re shallower than the part of the fault plane that 00:31:55.789 --> 00:32:00.850 we’ve seen activated so far. And they primarily rupture upwards. 00:32:00.850 --> 00:32:04.700 So we see this event in the bottom-left cross-section here. 00:32:04.700 --> 00:32:07.581 The arrow is covered by these dots because it’s aligning so well, 00:32:07.581 --> 00:32:09.830 and it’s rupturing up this fault plane. 00:32:09.830 --> 00:32:12.740 So, in our 3D view, you can see this is an X. 00:32:12.740 --> 00:32:15.880 That means the rupture directivity is going upward. 00:32:15.880 --> 00:32:21.360 And all of these open circles mean that those are all aftershocks that are 00:32:21.360 --> 00:32:25.130 happening upward of this event, or in the direction of rupture directivity. 00:32:25.130 --> 00:32:28.020 So those are all of these shallower events here. 00:32:30.000 --> 00:32:33.620 And, in the last couple of hours, right before the main shock, 00:32:33.620 --> 00:32:36.980 we just got a bunch of seismicity between 2 and 4 kilometers’ depth 00:32:36.980 --> 00:32:40.750 in the hanging wall of this fault. 00:32:40.750 --> 00:32:44.400 So the summary of the foreshock evolution is that we see a Phase I 00:32:44.400 --> 00:32:47.940 where it’s spreading with time, primarily driven by fluids. 00:32:47.940 --> 00:32:52.419 We see Phase II, where multiple clusters circle around that first patch – 00:32:52.419 --> 00:32:57.050 what I call Patch 1. And it includes many high stress drop events. 00:32:57.050 --> 00:33:00.480 So I didn’t mention it when I was going through, but a lot of those ones 00:33:00.480 --> 00:33:03.830 on the edges like you saw in the cross-section of stress drops 00:33:03.830 --> 00:33:06.880 I showed before are really high stress drop events. 00:33:08.020 --> 00:33:12.680 The seismicity also in Phase II activates at the other end of the rupture, 00:33:12.680 --> 00:33:14.810 or the other end of the main shock fault. 00:33:14.810 --> 00:33:18.549 And then, in Phase III, that’s Cluster 18, where the seismicity accelerates into 00:33:18.549 --> 00:33:22.440 the main shock over a day and a half. We see migration that’s consistent 00:33:22.440 --> 00:33:25.260 with the directivity and that’s consistent with about a 1 kilometer 00:33:25.260 --> 00:33:30.220 per hour away from the fault plane. And, in Phase IV, we see multiple 00:33:30.220 --> 00:33:32.140 upward ruptures, shallow aftershocks, 00:33:32.140 --> 00:33:37.770 persisting in the – above the kink in the fault plane. 00:33:37.770 --> 00:33:42.080 So, overall, the observations that I’ve made so far are that the 00:33:42.080 --> 00:33:44.580 events migrate repeatedly around these seismicity voids 00:33:44.580 --> 00:33:48.700 that we’re inferring or interpreting to be fault patches. 00:33:48.700 --> 00:33:51.059 And that we see both high stress drop events and events with 00:33:51.059 --> 00:33:55.320 fault-perpendicular directivity, primarily in these sort of edges of 00:33:55.320 --> 00:34:01.250 the patches or in these zones where the fault orientation changes slightly. 00:34:01.250 --> 00:34:03.160 So then what happens with the main shock? 00:34:03.160 --> 00:34:07.090 Well, there was a foreshock that happened about 10 seconds before. 00:34:07.090 --> 00:34:12.900 So that’s this event here on the map. And it’s rupturing to the southwest. 00:34:12.900 --> 00:34:17.580 And then it has an aftershock, which is the main shock, 00:34:17.589 --> 00:34:20.259 which is this big circle that happens right above it. 00:34:20.259 --> 00:34:22.770 And, if you look in this view, you can see that the main shock 00:34:22.770 --> 00:34:25.239 is essentially right in the kink of the fault. 00:34:25.240 --> 00:34:30.840 So it’s happening in that intersection of these two dipping normal faults. 00:34:32.249 --> 00:34:36.080 Here are the immediate aftershocks of the main shock rupture, 00:34:36.089 --> 00:34:40.319 which was to the northwest. The main shock is not one of our 00:34:40.320 --> 00:34:46.900 best well-resolved events because it did clip at a lot of the short-period stations, 00:34:46.900 --> 00:34:50.920 and the azimuthal coverage because of that is a little bit lower. 00:34:52.100 --> 00:34:55.440 But it’s very high stress drop, which I think makes sense 00:34:55.450 --> 00:34:58.259 with the story of it being in the kink of this fault. 00:34:58.259 --> 00:35:01.490 Here are the immediate aftershocks, which happen on both of these fault- 00:35:01.490 --> 00:35:07.560 parallel cross-sections, because it’s sort of a curved, doubly kinked fault plane. 00:35:08.380 --> 00:35:12.840 So I’m just sort of going to skip through these ones quickly, but you can see, 00:35:12.849 --> 00:35:17.130 as the events in the immediate aftershock period start filling in 00:35:17.130 --> 00:35:21.660 the fault plane, there’s no seismicity in the previous patch. 00:35:21.660 --> 00:35:26.130 And you’ll see two new patches sort of form up north of the – 00:35:26.130 --> 00:35:28.740 or, I don’t know why I keep saying “north” – upward in depth – 00:35:28.740 --> 00:35:32.799 shallower in depth than the main shock. 00:35:32.799 --> 00:35:36.650 And here you can see there are some events rupturing that it 00:35:36.650 --> 00:35:41.480 looks like fault-perpendicular in the cross-section. 00:35:41.480 --> 00:35:46.320 And maybe those are cross-faults that are happening. It’s unclear. 00:35:52.190 --> 00:35:56.740 So this sort of point cloud is going to be bigger than this one because of 00:35:56.749 --> 00:36:00.420 the way that the fault plane is doubly kinked or curved. 00:36:00.420 --> 00:36:04.589 And so we’re getting the farther-out edges, and then this would be inward. 00:36:04.589 --> 00:36:06.710 And the main shock actually doesn’t plot on here because the kink 00:36:06.710 --> 00:36:11.619 would be back one layer on another fault-parallel cross-section. 00:36:11.619 --> 00:36:16.599 So these are all of the aftershocks remaining in the main shock rupture. 00:36:16.599 --> 00:36:20.479 And it gets pretty hairy. You know, how can we interpret 00:36:20.479 --> 00:36:24.710 which fault planes actually ruptured during this sequence? 00:36:24.710 --> 00:36:28.200 You know, if you look at this view, it looks like, you know, it’s probably 00:36:28.200 --> 00:36:32.190 this guy going up to there. And then maybe there’s two faults here. 00:36:32.190 --> 00:36:38.200 But, again, there’s some conjugates that were also activated in this aftershock 00:36:38.200 --> 00:36:41.400 sequence, and before the main shock. So it makes the cross-sections 00:36:41.400 --> 00:36:47.829 a little bit – a little bit hard to interpret. But what I do see here are these sort of 00:36:47.829 --> 00:36:52.130 patches that are highlighted by this M shape on the main shock fault plane, 00:36:52.130 --> 00:36:55.049 and that’s where the highest concentration of aftershocks are. 00:36:55.049 --> 00:36:57.859 And then we see nothing over here. 00:36:57.859 --> 00:37:02.430 So, to get back to my original question, you know, could the main shock 00:37:02.430 --> 00:37:05.430 have ruptured into the patch that we previously identified? 00:37:05.430 --> 00:37:12.480 Well, just for convention’s sake, let’s name this P1, P2, P3, and P4. 00:37:12.480 --> 00:37:17.200 Well, if we look at the source-time function, it has definitely two pulses. 00:37:17.200 --> 00:37:20.460 There’s maybe another third little blip here. 00:37:21.440 --> 00:37:24.820 It would locate on this plot essentially right on this kink 00:37:24.820 --> 00:37:27.360 of these two fault planes, so right there. 00:37:27.360 --> 00:37:33.000 And the primary rupture is to the left side of this plot, or to the northwest. 00:37:33.010 --> 00:37:38.799 So the little pulse is a little patch, right? The big pulse is a big patch. 00:37:38.799 --> 00:37:43.160 So maybe the little pulse was 4. The big pulse was P1. 00:37:43.160 --> 00:37:46.140 Maybe the blip is P3. Unclear. 00:37:46.140 --> 00:37:50.119 But I think it’s highly likely that this event did rupture multiple fault planes, 00:37:50.119 --> 00:37:56.500 and that’s why we’re seeing this sort of behavior in the source-time function. 00:37:56.500 --> 00:37:59.340 So my summary of the directivity – oops. 00:37:59.340 --> 00:38:02.300 Hold on. I just knocked off my microphone. 00:38:05.060 --> 00:38:08.140 So the summary of the directivity analysis and the spatiotemporal 00:38:08.140 --> 00:38:12.940 evolution of both the foreshock and the aftershock period is right here. 00:38:12.940 --> 00:38:16.819 So a lot of times, the seismicity aligns with the fault planes and with 00:38:16.819 --> 00:38:21.530 the rupture directivity, but not always. And data’s messy. And so maybe 00:38:21.530 --> 00:38:24.079 we should only pick the really high-quality ones. 00:38:24.079 --> 00:38:29.020 Or should we spend time trying to explain the ones that don’t match? 00:38:29.020 --> 00:38:32.220 It seems to be more cases, like I said, where the aftershocks are in the back 00:38:32.220 --> 00:38:36.780 direction, or in the opposite direction of the rupture propagation. 00:38:36.780 --> 00:38:39.809 We previously hypothesized that the main shock ruptured the foreshock 00:38:39.809 --> 00:38:43.470 outlined patches, but it’s clear that there are similar smaller patches 00:38:43.470 --> 00:38:47.969 within the main shock cluster itself. So now I’m not so sure. 00:38:47.969 --> 00:38:51.880 Perhaps those patches are areas of hidden aseismic slip that drove 00:38:51.880 --> 00:38:55.609 the foreshock acceleration. Or perhaps they were ruptured by 00:38:55.609 --> 00:39:00.280 the aseismic slip after the main shock. Or maybe it’s something else altogether. 00:39:00.280 --> 00:39:03.849 What I do know is that the unilateral directivity is well-resolved, 00:39:03.849 --> 00:39:07.320 and it correctly identifies the fault planes for many, many 00:39:07.320 --> 00:39:11.510 of these events as low as magnitude 2.4. 00:39:11.510 --> 00:39:15.640 So the rest of the talk, I’m going to kind of just do a thought experiment on 00:39:15.640 --> 00:39:19.920 what else – what kind of questions can we try to answer with that knowledge? 00:39:19.920 --> 00:39:24.200 So one question I’m trying to answer is, how many aftershocks occur in the 00:39:24.200 --> 00:39:27.099 fore direction and the back direction? And I’d like to quantify that. 00:39:27.099 --> 00:39:30.440 So something I’ve tried doing is plotting all of the events on here 00:39:30.440 --> 00:39:34.280 and doing a density of it. So you can see clearly the fault plane is here. 00:39:34.280 --> 00:39:36.120 This red is the directivity. 00:39:36.120 --> 00:39:39.390 And it’s occurring right on the fault plane that we see. 00:39:39.390 --> 00:39:44.430 And then I’ve done some histograms – rose diagrams of the orientations. 00:39:44.430 --> 00:39:48.690 And trying to see if the foreshocks and aftershocks show any 00:39:48.690 --> 00:39:53.670 sort of meaningful, you know, statistics relative to the directivity. 00:39:53.670 --> 00:39:57.970 Another thing I’m interested in is, is there a relationship between 00:39:57.970 --> 00:40:00.430 the foreshock location and the rupture direction? 00:40:00.430 --> 00:40:04.549 In Rachel Hatch’s work, she saw that the foreshock ruptured towards where 00:40:04.549 --> 00:40:07.150 the eventual main shock happened, and the main shock ruptured back. 00:40:07.150 --> 00:40:11.150 Do we see that relationship a lot? I’ve seen a handful of cases, 00:40:11.150 --> 00:40:14.539 but I’d like to quantify that. How often do we see that they rupture towards 00:40:14.540 --> 00:40:18.820 each other – an immediate foreshock and then immediate aftershock? 00:40:18.820 --> 00:40:23.119 Additionally, this new sort of 4D stereonet analysis allows us 00:40:23.119 --> 00:40:27.079 to see fault plane complexity. So here you can see the open circles are 00:40:27.079 --> 00:40:32.359 completely overlapping the dark circles. So, in theory, if this was a perfectly 00:40:32.359 --> 00:40:34.870 planar fault, we would see all of the open circles on the 00:40:34.870 --> 00:40:38.459 other side of the focal sphere. But because they’re overlapping, 00:40:38.459 --> 00:40:42.279 that implies that this is a kinked fault and that we aren’t getting the other 00:40:42.279 --> 00:40:46.130 extension of the – of the northeast side of this rupture. 00:40:46.130 --> 00:40:50.319 So this is showing us a kinked fault. And, like I showed with the main shock, 00:40:50.319 --> 00:40:52.729 you could then go back and look at the source-time function of this and 00:40:52.729 --> 00:40:56.839 see if it looks like a complex rupture. And, like I mentioned earlier, 00:40:56.839 --> 00:41:01.690 most of the large events in our sequence are complex ruptures. 00:41:01.690 --> 00:41:03.759 So why are the largest earthquakes complex? 00:41:03.759 --> 00:41:05.690 Well, I think it’s because they’re rupturing multiple 00:41:05.690 --> 00:41:10.840 fault planes in this really highly fractured unorganized medium. 00:41:10.840 --> 00:41:15.400 Here’s an example of an aftershock cluster. 00:41:15.400 --> 00:41:19.680 And these are, again, the statistically clustered events that occurred with 00:41:19.680 --> 00:41:24.549 this event. It seems to rupture the kinked fault plane here. 00:41:24.549 --> 00:41:31.140 And then a fault – or, a fault conjugate or perpendicular fault here. 00:41:31.140 --> 00:41:36.900 And you can see that all of these are open and upward, and all of these 00:41:36.900 --> 00:41:39.019 ones are downward, and the rupture directivity, 00:41:39.019 --> 00:41:44.200 because it’s a plus, is downward. So I think that this kind of analysis 00:41:44.200 --> 00:41:47.710 could tell us some really interesting things about source physics and the 00:41:47.710 --> 00:41:54.560 variability that we see when we look at stress drops and this kind of a analysis. 00:41:56.400 --> 00:42:00.260 Right. And then this is the main – this is the source-time function for that event. 00:42:00.260 --> 00:42:03.440 And, again, you can see it’s clearly two significant pulses 00:42:03.440 --> 00:42:07.480 that are almost separated in time. 00:42:07.480 --> 00:42:10.670 So the last thing I’m going to talk about is this very preliminary 00:42:10.670 --> 00:42:15.220 Coulomb stress analysis. So, a lot of times, when I talk about this sequence, 00:42:15.220 --> 00:42:18.019 people are always asking me, like, well, have you tried Coulomb on it? 00:42:18.019 --> 00:42:19.369 Why haven’t you done Coulomb? 00:42:19.369 --> 00:42:22.150 And it’s because, for Coulomb, you need to know the fault planes, and 00:42:22.150 --> 00:42:25.509 you have to make a lot of assumptions. But, now that I’ve done this directivity 00:42:25.509 --> 00:42:29.329 and these stereonets, I think I kind of know with some confidence which fault 00:42:29.329 --> 00:42:32.719 plane ruptured in each of these events. So what I did was, I chose the 00:42:32.719 --> 00:42:35.400 most likely fault plane based on those stereonets. 00:42:35.400 --> 00:42:38.089 And, for the events that didn’t have a focal mechanism, 00:42:38.089 --> 00:42:41.420 I just took the best-fit plane of the previous event. 00:42:41.420 --> 00:42:47.719 And I fit those – or, not fit those – I built little faults for each of those 00:42:47.719 --> 00:42:50.869 based on the rupture dimensions that we got with source – with the 00:42:50.869 --> 00:42:57.910 stress drop, so our source dimensions. And I scaled that. And then I put slip 00:42:57.910 --> 00:43:06.650 on that by calculating the – yeah, by calculating using our seismic 00:43:06.650 --> 00:43:09.709 moments and the rupture dimension, I calculated how much slip you 00:43:09.709 --> 00:43:12.619 would expect on each little fault. And then, using the fault orientations, 00:43:12.619 --> 00:43:16.940 I built them all and put it into Coulomb. And then I ran it so that each fault 00:43:16.940 --> 00:43:20.230 ruptured one at a time and caused a Coulomb stress change. 00:43:20.230 --> 00:43:24.740 And I resolved that Coulomb stress change on all of the other faults. 00:43:28.150 --> 00:43:31.240 So these are my calculated displacements. 00:43:31.249 --> 00:43:35.440 I put Wells and Coppersmith relationship on here. 00:43:35.440 --> 00:43:38.309 It’s not designed for earthquakes this small, which is why I think 00:43:38.309 --> 00:43:42.839 it doesn’t fit the data. But if I just shift it a little bit down, 00:43:42.839 --> 00:43:48.589 then my data, more or less, is following this trend of displacement 00:43:48.589 --> 00:43:51.979 increasing with magnitude. The colors here are stress drop. 00:43:51.979 --> 00:43:54.099 So, like you would expect, the higher stress drop events 00:43:54.099 --> 00:43:57.650 have more slip for the same magnitude. 00:43:57.650 --> 00:43:59.240 These are the fault planes that I created, 00:43:59.240 --> 00:44:02.469 and they’re colored here by their stress drop. 00:44:02.469 --> 00:44:05.529 So you can see a lot of these perpendicular, or fault conjugate, 00:44:05.529 --> 00:44:10.559 or perpendicular faults are red. So that means they’re high stress drop. 00:44:10.559 --> 00:44:13.999 The main shock is high stress drop. And a lot of these ones on the ends 00:44:13.999 --> 00:44:18.579 and off-fault are either average or lower stress drop. 00:44:18.579 --> 00:44:20.599 And I put all that information into Coulomb. 00:44:20.599 --> 00:44:23.210 And so here’s an example for the first event. 00:44:23.210 --> 00:44:26.620 This is the Coulomb stress change from that first earthquake. 00:44:26.620 --> 00:44:32.780 And the plot on the right is the stress change on each fault from that event. 00:44:32.789 --> 00:44:35.200 And you can see that the maximum stress change occurs 00:44:35.200 --> 00:44:38.130 on the second fault to rupture chronologically. 00:44:38.130 --> 00:44:41.739 So I’m interpreting this, very preliminary, is, like, 00:44:41.739 --> 00:44:45.410 that event just triggered this other event with static stress change. 00:44:45.410 --> 00:44:49.249 A Coulomb stress change explains the triggering of these two events. 00:44:49.249 --> 00:44:52.239 So I did that for all of the events, and it turns out that six out of eight 00:44:52.239 --> 00:44:56.819 of the first events directly trigger, assuming that the maximum Coulomb 00:44:56.819 --> 00:45:01.369 stress change is going to trigger the next event – six out of eight of them 00:45:01.369 --> 00:45:05.190 actually do predict the next event to rupture, or the next fault plane 00:45:05.190 --> 00:45:08.319 to rupture. But once you get onto the main shock fault plane, only three 00:45:08.319 --> 00:45:17.749 out of 29 of the – of the foreshocks directly trigger the next event to rupture. 00:45:17.749 --> 00:45:21.220 So what does it mean? One possible interpretation is that static stress 00:45:21.220 --> 00:45:27.259 triggering dominates in the Phase I evolution, and then things change. 00:45:27.259 --> 00:45:29.900 Why would they change? Maybe because of aseismic slip. 00:45:29.900 --> 00:45:32.430 Maybe because of the rupture directivities, and dynamic triggering 00:45:32.430 --> 00:45:36.360 really takes over once you’re sort of on the same fault plane. 00:45:36.360 --> 00:45:38.779 There’s a lot of other things I need to test here. 00:45:38.779 --> 00:45:41.079 For instance, I could include the nodal planes that I didn’t pick 00:45:41.079 --> 00:45:44.099 and make sure that the one that is most stressed is the fault plane that 00:45:44.099 --> 00:45:47.950 I selected and the one that ruptured. I could include aftershock fault planes 00:45:47.950 --> 00:45:52.460 and see if the foreshocks are still the ones that are most likely to rupture. 00:45:52.460 --> 00:45:54.700 And I need to improve the fault plane area calculation. 00:45:54.700 --> 00:45:58.190 Right now, I just made square faults. So, in the Coulomb analysis, 00:45:58.190 --> 00:46:00.849 their stress drop is actually too low because the area of a square 00:46:00.849 --> 00:46:04.620 is bigger than a circle, and these are circle sources. 00:46:05.789 --> 00:46:10.220 So the conclusions of my talk are that we resolved robust 3D directivity 00:46:10.221 --> 00:46:13.029 for events as low as magnitude 2.4. 00:46:13.029 --> 00:46:17.410 They confirmed faults identified by a seismicity and focal mechanisms. 00:46:17.410 --> 00:46:20.430 We observed this rapid migration of foreshocks around the patches on the 00:46:20.430 --> 00:46:23.920 main shock fault plane that’s consistent with our rupture directivities. 00:46:23.920 --> 00:46:28.680 And potentially, the static stress changes, or the Coulomb stress changes, 00:46:28.690 --> 00:46:30.859 could explain triggering in the first part of the sequence, 00:46:30.859 --> 00:46:37.599 but not in the later part of the sequence. And, again, in Jansen et al. paper, 00:46:37.599 --> 00:46:43.099 we showed that fluids and fluid diffusion could explain the first phase 00:46:43.099 --> 00:46:47.980 of foreshocks, but not the, like, last phase right before the main shock. 00:46:47.980 --> 00:46:51.700 So something else happened there. What it is or how we can actually get at 00:46:51.700 --> 00:46:55.349 that, you know, I’m open to ideas. And I would love to discuss this 00:46:55.349 --> 00:46:58.709 with anybody who’s interested or has ideas for how I could 00:46:58.709 --> 00:47:03.549 turn this into, like, some manageable digestible papers. 00:47:03.549 --> 00:47:04.910 So thank you very much for listening, 00:47:04.910 --> 00:47:08.000 and hopefully there’s time for questions and interest. 00:47:08.000 --> 00:47:12.040 [Applause] 00:47:12.040 --> 00:47:15.080 - Thanks, Christine. Does anyone have a question? 00:47:17.640 --> 00:47:19.060 - Sure. 00:47:19.940 --> 00:47:23.060 [Silence] 00:47:23.900 --> 00:47:26.820 - That was a great talk. - Thank you. 00:47:26.820 --> 00:47:31.980 - I was wondering if there’s a way to detect the hidden aseismic slip. 00:47:31.999 --> 00:47:35.721 Because the static stress changes didn’t work well, it sounds like maybe 00:47:35.721 --> 00:47:39.520 dynamic wouldn’t either because the aftershocks were in the wrong direction. 00:47:39.520 --> 00:47:42.800 But it seems like the hidden aseismic slip could be a key. 00:47:42.800 --> 00:47:47.239 - Right. Yeah. I don’t know. I’m – anybody who does that 00:47:47.239 --> 00:47:50.349 kind of thing, maybe they could re-analyze the InSAR. 00:47:50.349 --> 00:47:53.309 Maybe – there are GPS. There were four stations, basically, 00:47:53.309 --> 00:47:56.180 right around this. And that data maybe could be 00:47:56.180 --> 00:47:59.749 reprocessed by someone who’s an expert in that to look for these sort of things. 00:47:59.749 --> 00:48:04.319 I did try to convince the geodesists and UNR to look for it, and they could 00:48:04.319 --> 00:48:07.329 only confirm that, if there was aseismic slip, it was within 00:48:07.329 --> 00:48:11.809 the uncertainty of the GPS data, which is a handful of millimeters 00:48:11.809 --> 00:48:14.740 or something like that. So it would be a very small amount. 00:48:14.740 --> 00:48:17.400 And, yeah, maybe it happened. Maybe not. Unclear. 00:48:17.410 --> 00:48:20.459 But there are a lot of papers coming out now that show that aseismic slip 00:48:20.459 --> 00:48:24.259 and fluids and all these things interact together, and they all play a role 00:48:24.259 --> 00:48:26.440 in the evolution of these sequences. 00:48:26.440 --> 00:48:30.660 So, yeah, I don’t know if we’ll ever know. [laughs] 00:48:31.940 --> 00:48:35.079 - Hi, Christine. That was a really interesting talk with a lot to think about. 00:48:35.079 --> 00:48:39.640 And I was just wondering if you could maybe elaborate a little bit on the 00:48:39.640 --> 00:48:46.059 Phase I – you suggested both fluid diffusion and static stress changes 00:48:46.059 --> 00:48:50.190 as possible driving mechanisms. And maybe you have some thoughts on 00:48:50.190 --> 00:48:54.620 how those things might be working in concert to produce what you saw. 00:48:54.620 --> 00:49:04.180 - Yeah, okay. So I didn’t do as much of the modeling of fluids as Jansen, 00:49:04.180 --> 00:49:06.809 my colleague in Switzerland, did – Gunnar Jansen. 00:49:06.809 --> 00:49:11.869 So, yeah, I don’t know how those two things would play together. 00:49:11.869 --> 00:49:15.420 Essentially, the way he did his analysis was he put all the fault planes in there, 00:49:15.420 --> 00:49:18.199 kind of the same way that I did with Coulomb, and then he had fluids 00:49:18.199 --> 00:49:25.599 kind of coming up at depth rapidly. And he resolved, essentially, Coulomb 00:49:25.600 --> 00:49:28.600 stresses to calculate the failures of them. So I think the fact that 00:49:28.600 --> 00:49:31.680 that's similar is probably – has something to do with the calculations. 00:49:31.680 --> 00:49:35.140 But this Coulomb stuff I’ve just done over the last week. 00:49:35.140 --> 00:49:38.020 And I’m not prepared to, like, defend it very strongly. 00:49:38.020 --> 00:49:41.589 I just thought, wow, this is kind of a cool result, and maybe I’ll include it 00:49:41.589 --> 00:49:45.150 as something to think about. But, I mean, obviously static 00:49:45.150 --> 00:49:47.940 stress changes are always going to have some role. 00:49:47.940 --> 00:49:50.990 And I think, how could you get at the difference between that – 00:49:50.990 --> 00:49:52.949 I think you’d have to ask someone who’s more of 00:49:52.949 --> 00:49:56.180 a modeler and not an observationalist, probably. 00:49:57.620 --> 00:50:00.560 [Silence] 00:50:01.340 --> 00:50:04.560 - A simple clarification question when you’re plotting your patches 00:50:04.569 --> 00:50:09.089 and your circles. Are your circles scaled to your estimated 00:50:09.089 --> 00:50:14.160 source dimensions in all those plots? - No. So … 00:50:14.160 --> 00:50:17.760 - Because it – then you can get very misleading interpretations. 00:50:17.760 --> 00:50:19.339 - Yes. 00:50:19.340 --> 00:50:24.240 So let me go back to this. So this plot, they are scaled by source dimension. 00:50:26.400 --> 00:50:28.800 Yeah. So here they’re scaled by source dimensions. 00:50:28.819 --> 00:50:31.699 So here you would say, oh, I don’t see a patch. 00:50:31.699 --> 00:50:36.219 But I would argue that the only events to rupture into the patch 00:50:36.219 --> 00:50:39.829 are very low stress drop events. And that because of that, you could 00:50:39.829 --> 00:50:45.250 say that potentially those very low stress drop events didn’t rupture it 00:50:45.250 --> 00:50:50.319 significantly enough that – or maybe the foreshocks did just rupture it there. 00:50:50.320 --> 00:50:54.380 Maybe those low stress drop events are related to the hidden aseismic slip. 00:50:54.380 --> 00:50:58.319 It’s unclear. But, yeah, I agree with what you’re saying. 00:50:58.319 --> 00:51:03.109 But, like, fine. In the foreshock period, maybe the patch was ruptured. 00:51:03.109 --> 00:51:05.739 In the aftershock period, even with the rupture dimensions, 00:51:05.740 --> 00:51:10.660 I can say clearly that no aftershocks ruptured this patch here. 00:51:10.660 --> 00:51:16.020 - So your – one of your last slides that had P1, P2, P3, and P4 … 00:51:16.020 --> 00:51:18.540 - Yep. - Can you sort of go back to that one 00:51:18.559 --> 00:51:21.959 and try and explain … - Yeah. 00:51:21.960 --> 00:51:27.560 - … how big those patches really are relative to your circles? 00:51:31.240 --> 00:51:35.520 - The mouse is – the scroll is backwards. It’s messing me up. Sorry. 00:51:37.240 --> 00:51:39.140 Where was that? 00:51:42.880 --> 00:51:44.280 Here. 00:51:48.940 --> 00:51:52.839 Right. So these patches would be very small. 00:51:52.839 --> 00:51:58.349 I mean, the dimension of this one is probably – I don’t know. 00:51:58.349 --> 00:52:02.279 The diameter is probably half a kilometer or something like that, or less. 00:52:02.279 --> 00:52:05.839 So they’re quite small patches, which would also support our 00:52:05.839 --> 00:52:07.390 very high stress drop main shock. 00:52:07.390 --> 00:52:13.359 It had a stress drop of around 35 or 40 megapascals or something like that. 00:52:13.359 --> 00:52:18.170 So if there were extremely tiny patches that were ruptured by this M5, 00:52:18.170 --> 00:52:21.060 that could help explain it. 00:52:21.900 --> 00:52:25.519 But, yeah, like here, a lot of these source dimensions might overlap 00:52:25.520 --> 00:52:30.319 those patches if I plotted it in a similar way. Yeah. 00:52:31.300 --> 00:52:34.640 Yeah, I don’t know. I mean, maybe they’re not exactly 00:52:34.640 --> 00:52:37.779 patches so much as the different fault planes. You know what I mean? 00:52:37.779 --> 00:52:40.930 And they’re being re-ruptured. So it’s not that they’re not being 00:52:40.930 --> 00:52:46.200 ruptured at all or anything, but I think they’re distinctly identifying sort of the 00:52:46.200 --> 00:52:52.440 four wings of this sort of kinked fault. So, yeah. But to say that they’re 00:52:52.440 --> 00:52:54.779 absolutely locked patches, I wouldn’t go that far. 00:52:54.779 --> 00:53:00.709 I guess I’m using patches more to describe differential stresses 00:53:00.709 --> 00:53:03.940 or differential roughness compared to the places where 00:53:03.940 --> 00:53:08.269 all the seismicity is occurring. You know, if the inside, where I have 00:53:08.269 --> 00:53:15.280 labeled P1 was exactly the same as, you know, 250 meters outside of P1, 00:53:15.280 --> 00:53:19.839 why didn’t we get as much seismicity in the location of P1 as outside? 00:53:19.839 --> 00:53:22.650 You know? So even if they’re not strictly patches 00:53:22.650 --> 00:53:26.859 like in the modeling sense or, like, that we see on a really, really 00:53:26.859 --> 00:53:30.849 big megathrust scale, I still think that they’re similar. 00:53:30.849 --> 00:53:34.249 And I think that, because of the quality of locations here, we see a lot more 00:53:34.249 --> 00:53:37.890 events that maybe are happening inside of them than they do in subduction 00:53:37.890 --> 00:53:40.459 zones or in other places where the data quality is much lower, 00:53:40.460 --> 00:53:43.980 and they’re not including really tiny events. 00:53:43.980 --> 00:53:48.059 But, like this catalog is complete down to, you know, 00:53:48.059 --> 00:53:51.869 negative 0.5 magnitude units. So I don’t think we’re missing 00:53:51.869 --> 00:53:56.589 a lot of seismicity in those zones. So the fact that nothing nucleates there, 00:53:56.589 --> 00:54:00.299 I think still says something about those areas rather than 00:54:00.300 --> 00:54:04.040 maybe they were or weren’t ruptured. I don’t know. 00:54:09.460 --> 00:54:16.000 - So, regarding the Coulomb, you have a distribution of faults. 00:54:16.000 --> 00:54:19.400 I mean, I guess the whole area is fairly small, but – and you have 00:54:19.400 --> 00:54:24.039 some distribution in time – maybe one more – yeah, here. 00:54:24.039 --> 00:54:27.720 So I think what you’re showing is just, you looked sequentially, if you have 00:54:27.720 --> 00:54:29.560 a stress change here, does it trigger the next fault? 00:54:29.560 --> 00:54:34.999 But did you – or you could consider the fact that they’re not just ordered in time, 00:54:34.999 --> 00:54:37.789 but there’s some time separation. There’s some space separation, right? 00:54:37.789 --> 00:54:43.280 So maybe they’re triggering on different sides of the fault or something so it’s not 00:54:43.280 --> 00:54:49.660 strictly, you know, in some time order. Maybe I’m not making sense. 00:54:49.670 --> 00:54:55.420 But, like you said, for the first set, six out of eight or whatever, the faults 00:54:55.420 --> 00:54:58.479 triggered in the correct order, where there’s all – like, in a small region. 00:54:58.479 --> 00:55:01.930 And then, when you go to the other … - No. So the first – the first handful 00:55:01.930 --> 00:55:04.409 of events are these ones here. - Right. Right. 00:55:04.409 --> 00:55:07.269 - And those are all triggered, like, directly. 00:55:07.269 --> 00:55:11.119 But then there are some where the most increased stress is one 00:55:11.119 --> 00:55:13.200 that’s already ruptured. And it doesn’t re-rupture again. 00:55:13.200 --> 00:55:15.200 - Oh, okay. - So we do see that later on, 00:55:15.200 --> 00:55:17.609 but not in the first eight. - So you – I was also asking – 00:55:17.609 --> 00:55:19.910 maybe Ruth was getting at this as well with the dynamic triggering, 00:55:19.910 --> 00:55:22.560 but if you have a Coulomb stress change, and then it doesn’t 00:55:22.560 --> 00:55:25.479 trigger the very next fault, but could it have triggered the one … 00:55:25.479 --> 00:55:26.859 - Right. - … two or three after that 00:55:26.859 --> 00:55:29.430 that are still close? - So I included – so I took any – 00:55:29.430 --> 00:55:34.339 if the – if the next one triggered was any of the top three, I included it. 00:55:34.339 --> 00:55:36.039 - Oh, so okay. - So it’s very arbitrary. 00:55:36.039 --> 00:55:40.589 But, yeah, I just – back of the hand – so, in this one, it’s exactly 00:55:40.589 --> 00:55:42.260 the most triggered. But if it was … 00:55:42.260 --> 00:55:43.989 - I see. - If the next one chronologically 00:55:43.989 --> 00:55:47.369 was any of the top three, I considered it a positive prediction. 00:55:47.369 --> 00:55:49.949 But, again, this is very, very preliminary. 00:55:49.949 --> 00:55:52.089 And I just thought it was interesting, you know? 00:55:52.089 --> 00:55:55.489 If anybody’s a Coulomb expert, this is the first time I’ve used it. 00:55:55.489 --> 00:55:58.740 I’m happy to take advice. - So I also had a question on your – 00:55:58.740 --> 00:56:02.520 I think one of your summary stress drop plots where you had the magnitude and 00:56:02.520 --> 00:56:04.780 the – that one – the blue and red one. - Okay. 00:56:04.780 --> 00:56:06.900 - I think you told me this the other day, but the red ones are … 00:56:06.900 --> 00:56:08.320 - Complex. - Complex. 00:56:08.320 --> 00:56:11.000 - Yeah. - So are you suggesting that, 00:56:11.000 --> 00:56:14.400 for the larger – the larger magnitude events tend to be more complex, 00:56:14.400 --> 00:56:18.769 but you also see smaller events that are complex that have small stress drops? 00:56:18.769 --> 00:56:21.150 - Yes. And those ones, I think, are tending to happen in the kinks 00:56:21.150 --> 00:56:24.440 or in these bends of the fault zone. So they’re happening in complex areas 00:56:24.440 --> 00:56:27.589 of the fault. So I think that all of the source complexity we’re seeing 00:56:27.589 --> 00:56:31.819 is related to the fault geometry. Maybe, you know, variable stress 00:56:31.819 --> 00:56:35.859 state on the faults to begin with. If there’s fluid still messing with 00:56:35.860 --> 00:56:40.440 part of it, maybe that’s affecting it. You know, it’s – a lot going on. 00:56:41.720 --> 00:56:44.320 But I think definitely, for these larger ones, you could explain 00:56:44.329 --> 00:56:48.930 the complexity here by them rupturing multiple sort of planes, 00:56:48.930 --> 00:56:52.260 which we see in the locations as well. 00:56:54.540 --> 00:56:56.780 - Any additional questions? 00:57:00.140 --> 00:57:02.259 Okay. Let’s thank our speaker again. 00:57:02.260 --> 00:57:04.640 [Applause] 00:57:04.640 --> 00:57:06.880 And we’ll be taking Christine out to lunch, so if anyone 00:57:06.880 --> 00:57:10.319 wants to join, please do. And she’ll be around this afternoon. 00:57:10.319 --> 00:57:13.269 I think she still has space in her schedule for meetings. 00:57:13.269 --> 00:57:16.600 So if you’re interested in that, please come talk to her. 00:57:16.600 --> 00:57:18.120 - Thank you. 00:57:20.720 --> 00:57:26.780 [Silence]