WEBVTT 00:00:00.000 --> 00:00:03.530 Yifang Cheng, Tongji at the University of Shanghai, who is giving her talk at 1:30 a.m. in the morning, 00:00:03.540 --> 00:00:04.210 her time. 00:00:04.220 --> 00:00:08.930 So thanks for being so accommodating. 00:00:09.360 --> 00:00:18.470 Yifang completed her BS in geophysics at Peking University before coming to the U.S. to earn a masters at the University of Oklahoma. 00:00:18.720 --> 00:00:31.680 She went onto obtain a PhD at University of southern California, where she worked with Yehuda Ben Zion, and she recently completed a postdoc at UC Berkeley working with Richard Allen, Roland Bergman, and others. 00:00:32.190 --> 00:00:36.400 She just began a new appointment at Tongji University in Shanghai. 00:00:36.470 --> 00:00:43.640 Her research has focused on earthquake source properties, fault mechanics, earthquake interaction and triggering, array seismology, 00:00:43.650 --> 00:00:48.160 and subsurface monitoring with applications to geothermal operations, 00:00:48.170 --> 00:00:56.320 wastewater injection, and carbon storage. In her recent work, she has applied machine learning to develop new focal mechanism catalogs. 00:00:56.630 --> 00:01:03.450 Today she will present a talk entitled "Unraveling multi-scale fault zone behaviors with small earthquake focal mechanisms. 00:01:04.650 --> 00:01:05.030 Thank you. 00:01:08.290 --> 00:01:23.940 Thank you Jessica for the great introduction and thanks for providing me such a kind of a great opportunity to share my recent work in collaboration with my colleagues at Berkeley Seismological Lab and unraveling multi-scale fault zone behaviors with small earthquake focal mechanisms. 00:01:23.950 --> 00:01:44.180 Earthquake is the main topic about for brackets a small lab and the USGS because what we leave right on the San Andreas Fault System and there are many larger historic earthquake occurring on this system. 00:01:44.390 --> 00:02:01.310 And here is a spatial and temporal distribution of the historic larger earthquakes in California, and we can see that even though the large earthquakes are rare in time, but they can be really destructive. 00:02:01.740 --> 00:02:12.300 As an example, I will show you the photo taken right after the most recent big one 1906 San Francisco earthquake 00:02:13.420 --> 00:02:17.350 near the Ferry Building and California Street, 00:02:17.440 --> 00:02:36.610 and what it looks like in 2006. As you can see, it's really destructive and it's also very important for us to better understand the seismic potential and to better assess the seismic hazard for the future major earthquakes. 00:02:38.080 --> 00:02:50.110 Ideally, if the crustal strain and the stress loading rate are constant, then we would expect the really constant recurrence interval between these major earthquakes. 00:02:50.520 --> 00:03:00.530 However, as we can see from the top figure, they don't really have the regular recurrence interval. 00:03:01.100 --> 00:03:05.790 Instead, the crustal strains and the stress loading 00:03:05.920 --> 00:03:07.200 they are not constant 00:03:07.670 --> 00:03:08.190 oin time. 00:03:08.670 --> 00:03:21.980 As a result, it's really important for us to understand how stress accumulates in the fault zone during the inter-seismic period and what is the current stress status. 00:03:21.990 --> 00:03:29.490 Of course, the fault system in California and how the stress releases in the fault zone during the earthquake rupture. 00:03:31.720 --> 00:03:41.530 So even though the major earthquakes are rare, we have a small earthquakes everywhere and they occur very frequently in California. 00:03:41.820 --> 00:03:49.510 And with this smaller earthquakes, it can provide us more information about the fault zone processes. 00:03:49.700 --> 00:03:59.290 So in this talk, I will show you how we can use small earthquake focal mechanisms to unravel the fault zone behaviors in California. 00:04:00.240 --> 00:04:28.290 I will first start with the introduction about the tools for fossil monitoring and then fault zone monitoring for the focal mechanism estimation and then I will show you how we utilize the smaller earthquake focal mechanisms to improve our understanding about the stress accumulation, the stress status, and the stress release processes in the fault zone. 00:04:30.910 --> 00:04:38.650 So fault zone consists of the highly deformed fault core last, [uh, let me see]. 00:04:41.280 --> 00:04:51.410 highlight the fault core last deformed damage zone and the surrounding host rock and we can observe it directly from the field. 00:04:51.910 --> 00:05:05.230 However, most part of the fault zone has been at the surface adapts, so it's harder to observe it from the surface and the surface deformation only shows the field 00:05:05.240 --> 00:05:08.780 observation only shows the deformation at a time it deforms. 00:05:08.970 --> 00:05:12.800 It cannot help us to track the status of the fossil, adapts. 00:05:13.610 --> 00:05:29.450 And we can also use the geodetic observations like the GPS and the InSAR to trace the surface displacement, and it can provide the temporary or spatially dense observations. 00:05:30.440 --> 00:05:32.450 However, there's no depth resolution. 00:05:32.460 --> 00:05:42.810 We need to assume a certain fault model adapts and by which to infer how the fault zone 00:05:42.870 --> 00:05:43.300 behaves, 00:05:43.850 --> 00:05:47.010 adapts. 00:05:47.020 --> 00:05:49.770 And is there any way that we can directly [indiscernible] 00:05:49.780 --> 00:05:50.590 adapts? 00:05:51.060 --> 00:05:52.350 The answer is yes, 00:05:52.360 --> 00:05:59.760 sure, we can dig the hole so that we can get the direct measurements inside properly measurements. 00:06:00.040 --> 00:06:10.690 However, it's really expensive and very local compared with such kind of larger scale fault zone system in California. 00:06:12.560 --> 00:06:30.590 And lucky, we have a lot of small earthquakes adapts and they occur very frequently in time so that we can make use of the spatial temporal distribution of this assessment to trace the bottom processes adapt. 00:06:31.400 --> 00:06:49.390 But if we only use the time, location, and magnitude of earthquakes, we can only perform the statistical analysis and it's really challenging to link the statistical parameters with the physical processes of the full zone adapts. 00:06:50.350 --> 00:06:55.100 Therefore, in this study away use the small earthquake focal mechanisms. 00:06:55.770 --> 00:06:58.700 But what can we learn from focal mechanism? 00:07:00.440 --> 00:07:00.730 The focal 00:07:01.120 --> 00:07:13.850 mechanism describes the photo orientation and the sleep direction on which the earthquake occurred, so it can tell us the information about the fault orientation and the slip direction. 00:07:14.320 --> 00:07:29.630 And moreover, if we know a certain stress field and we know the photo orientation under this stress field, then we can have the expectation of about how this photo will move. 00:07:29.740 --> 00:07:41.460 For example, if the fault is normal to the maximum horizontal stress orientation with by the reverse faulting rupture of the fault. 00:07:42.090 --> 00:07:42.640 And if 00:07:43.070 --> 00:07:52.040 the fault is parallel to the maximum horizontal stress orientation, we will expect normal faulting rupture. 00:07:52.330 --> 00:08:02.540 But if the fault shows a certain angle to the maximum horizontal stress orientation, we may expect stress slip deformation. 00:08:02.850 --> 00:08:12.330 Therefore, if we know the stress and the fault orientation, we can estimates possible slip direction of this fault. 00:08:13.570 --> 00:08:14.220 But we 00:08:14.230 --> 00:08:16.480 don't know the stress in the crust, 00:08:17.250 --> 00:08:33.720 and we have a focal mechanism with fault orientation and slip direction by which we can do it in the other way. We can use the information from the focal mechanism to inverse the stress in the crust. 00:08:34.270 --> 00:08:49.930 Therefore, from the focal mechanism we can get to the geometric property and the displacement and the stress information of the active fault adapts, which, yeah, it can be, 00:08:49.940 --> 00:08:57.140 it could be a very powerful tool for us to monitor the physical processes of the fault zone adapts. 00:08:59.090 --> 00:09:02.610 How can we estimate the focal mechanism of earthquakes? 00:09:05.330 --> 00:09:08.190 Suppose there's an earthquake rupture adapts 00:09:08.580 --> 00:09:31.690 then it will cause compressive strain on the direction where the fault moves and the tense strain on the opposite side, and then this information will propagate via the seismic wave in which the surface recorded by the seismic stations at the surface. So 00:09:33.700 --> 00:09:34.110 for example, 00:09:34.120 --> 00:09:41.230 this can cause different first emotion polarity at the surface station. 00:09:41.700 --> 00:09:42.210 The Reasenberg, 1985 00:09:42.220 --> 00:09:53.620 utilized this polarity and the first emotion polarity and the back projected to the focal sphere around the source and inverse 00:09:53.680 --> 00:10:23.650 the focal mechanism and then later on in 2003 Hardback and Shearer incorporated the P-wave and S-wave amplitude ratio to further constrain the focal mechanism because for the same earthquake the P-wave and the S-wave share the similar paths and the side effect but near the nodal point the P-wave amplitude is really small. 00:10:23.720 --> 00:10:32.920 But S-wave amplitude is really large, so we can utilize the amplitude ratio between these two to constrain the location of the nodal plane. 00:10:35.240 --> 00:10:50.210 And if we have even a pair with collocated even pair, then the amplitude ratio between these two even pair because they share similar paths and study effect the amplitude ratio between these two events 00:10:50.850 --> 00:10:57.460 is mainly related to their magnitude differences and their differences in the focal mechanism. 00:10:58.090 --> 00:11:39.920 So, similar to the idea of utilizing differential time to constrain the location to do the relocation from the original location, we can first use the hash to obtain the original focal mechanism and then use the inter-event amplitude ratio to further constrain the focal mechanism from the original focal mechanism estimation such we propose the [indiscernible] method in 2003 and then we utilize this applied this method and sorry and then here is the comparison between this [indiscernible]. 00:11:41.090 --> 00:11:51.240 If a large earthquake, if we have a high quality waveform, we can directly perform the waveform fitting and get the moment tensor with really high quality. 00:11:51.610 --> 00:12:02.000 However, for small earthquake we can only use this waveform features and waste more and more waveform features incorporated into the focal mechanism calculation. 00:12:02.250 --> 00:12:06.710 We can get more focal mechanisms with higher quality. 00:12:09.140 --> 00:12:12.170 And then we 00:12:12.210 --> 00:12:32.350 applied this [indiscernible] method to over 1.7 million earthquakes in California since 1980, and we obtained over 800,000 focal mechanisms with at least 8 polarities 00:12:32.520 --> 00:12:40.010 and uncertainty <45 degree, which is over 50 percent of all catalog earthquakes. 00:12:40.200 --> 00:13:08.040 So here are the focal mechanisms colored by the 14th style and the 50% of the catalog events means that whatever the statistical analysis or spatial temporal analysis we have applied to the traditional earthquake location catalog, we can also apply them to the focal mechanism catalog and extract more information about the focal processes. 00:13:11.170 --> 00:13:21.830 So we'll first use this focal mechanism to understand how the stress accumulates in the fault zone during the inter-seismic period. 00:13:23.360 --> 00:13:27.110 So for the inter-seismic period deformation 00:13:27.560 --> 00:13:55.540 one important concept is for coupling, which is defined as the slip deficit rate versus long-term slip rate. For the locked area we have the slip rate approaches 0, which means we have a high slip deficit rate and a high fault coupling value. For the creeping area, there's limited slip deficit rate and the low for coupling the value. 00:13:56.940 --> 00:14:09.490 Traditional study use the geodetic observations from the surface and the seismicity analysis to understand the fault coupling adapts. 00:14:11.410 --> 00:14:30.900 However, they usually assume that what slip along the same direction and only analyze the statistical properties of the earthquake occurrences without considering an on-fault and off-fault deformation and their interactions. 00:14:32.380 --> 00:14:44.010 Here in our study, we incorporate the focal mechanism into analysis, which can help us to solve the local heterogeneity of the fault slip direction. 00:14:44.660 --> 00:15:04.020 It can also provide us more physical properties about the fault adapts like the geometry and sleep motion and stress in the IT can also help us to identify the office structures and the related kinematics here in this figure. 00:15:04.030 --> 00:15:05.580 This blue beach balls. 00:15:05.590 --> 00:15:21.600 There's no they have no nodal plan along while with the major fault orientation, so they can only be off fault and their uh distributions and properties represents this awkward structure and kinematics. 00:15:25.120 --> 00:15:32.460 So here we choose the centuries Angeles for the area to further enter. 00:15:33.200 --> 00:16:09.950 Apply the focal mechanism and the other analysis to understand the intercessory deformation, because in this area here in the on the left side, this Gray colored region, or the region which the photo creep occurs and the photo central Sanders for the area, it's about 200 kilometer long and only about a 500 meter wide, very localized long for zone like with varying for coupling and very obvious for creep. 00:16:10.450 --> 00:16:12.960 And it also has many smaller squeaks. 00:16:12.970 --> 00:16:16.530 Ways very different. 00:16:16.540 --> 00:16:34.170 Diverse focal mechanisms, so it provide us a great opportunity to study the photon behavior and the more importantly each is bounded by many in some of the larger historic earthquake. 00:16:34.380 --> 00:16:41.110 But along around this 4th segment, this no observed historic larger quake. 00:16:41.420 --> 00:16:55.390 So it's also important for us to understand that whether umm, to understand the behavior of this creeping section and its implication to the assessment hazard along the San Andreas Fault. 00:16:59.980 --> 00:17:17.260 So here is the distribution of the earthquakes along this centuries and just fought area and the Blue Beach ball and white also shows the magnitude larger than 4 earthquakes and repeating earthquakes respectively. 00:17:18.230 --> 00:17:35.380 And we can see that they show very localized distribution with very, uh, consistent focal mechanism showing the like with the strike along for the nodal plan and the right letter structurally motion. 00:17:36.090 --> 00:17:44.980 So we assume that both the magnitude larger than four square and repeating earthquakes as unfolds events, but for the magnitude. 00:17:45.930 --> 00:17:59.580 In contrast, the magnitude larger than one just makes they are like shows like a widely distributed in this falzone, and also they have a highly diverse for Commack kingdoms. 00:17:59.650 --> 00:18:22.870 So we assume that they occur both on the fault and offer fault, and based on the distribution of this assessment, see we proposed a photo coupling model where this magnitude larger than four earthquakes are around the boundary of this locked patch and this. 00:18:25.910 --> 00:18:41.390 Uh, they about for this Fort is loaded by a stable equipping a lower layer in the way performed as a the modeling, uh we model the intercept make deformation based on this 4th coupling model. 00:18:42.410 --> 00:19:04.770 And here is what we got under this for coupling model, we would expect to see that we see that we based the model result show that it has a high equip rate near the in the low coupling area and the low corporate near the locked patch which is expected. 00:19:05.640 --> 00:19:06.000 Umm. 00:19:07.160 --> 00:19:25.230 And also for the quick direction near the locked patch, if the fault move from the left to the right when it matters when it meets the the lock, the patch it cannot move horizontally. 00:19:25.240 --> 00:19:52.960 Instead, it need to move around this log patch, which will cause an upward equipped direction on the left side of the light patch and downward uh moving crypto action on the right side of the log patch and the this for coupling heterogeneity can and can also cause uh, the heterogeneity in the stress orientation. 00:19:53.110 --> 00:20:39.430 And then we can see a highly the the changing just as each Max orientation along the fault some area has the low angle to the SH Max to the major fault and some areas showing high angle SHHH Max to the major fault and wait for their compare these model results with our observation of from we observe the surface cooperate from the entire data and the inside the crypt rate from repeating occurrence rate inside the fault crypt direction from Retyping earthquake focal mechanisms and the stress field obtained from the stress inversion using magnitude larger than one for. 00:20:39.500 --> 00:20:40.330 Call back in apps. 00:20:40.640 --> 00:20:41.470 They all show. 00:20:41.480 --> 00:20:43.610 Really considerable accurate. 00:20:43.990 --> 00:20:53.620 Of course, a correlation coefficient suggesting that since the modeling results are consistent with multiple unfold and unfold observations. 00:20:55.840 --> 00:21:29.050 And besides the unforged properties in the then this obtained for company model, we further compare it with previous model obtained from the geodetic data and it shows that because we added more constraint utilizing the seismicity and focal mechanisms adapts it provide more features, adapts that cannot be seen from the other models that mainly use the show Daddy your service operations. 00:21:32.560 --> 00:21:35.210 And besides, they offered property. 00:21:35.480 --> 00:22:03.180 Uh, the focal mechanism can also help us to better understand the output structure and kinematics here shows the ohm those maximum stress orientation versus the percentage of transpressional events which we can go larger than zero and each dial just shows the value on each grade of each grade on the floor plan. 00:22:04.940 --> 00:22:29.950 When the stress show high angle to the fault and though we see a larger percentage of transtensional events on with optimally oriented direction showing high angle to the major faults, when the stress show up is about a 45 degree to the major faults we see comparable number of transformational and transactional events. 00:22:31.000 --> 00:22:46.370 When the stress show low angle to the major faults in the, we see a larger percentage of transpressional events with optimally oriented for orientation showing also show high end go to the major fault. 00:22:46.780 --> 00:23:04.620 Therefore, we find that instead of most of the events occurring along the for parallel direction, there are so many fun skill faults activated high angle to the main fault and they tend to deform along the for the parallel direction. 00:23:06.290 --> 00:23:20.420 Where is company, so suggesting that it is a highly localized narrow week for zone in the that is much easier to deform compared with the surrounding strong hostile arc. 00:23:23.010 --> 00:23:46.190 So the takeaway message from the first part is we have observed the lot of fine skill structure and kinematics for this central central sword and this diverse features can all be reconciled into a simple for coupling model and the narrow mechanically weak zone. 00:23:48.090 --> 00:24:03.420 And the based on this uh for coupling model, we can further estimate uh those stored the sesmic potential, the stored moment over 150 years is equivalent to a magnitude 7.5 earthquakes. 00:24:04.450 --> 00:24:13.460 And because of this, for coupling heterogeneity, it also can cause significant stress rotation. 00:24:13.470 --> 00:24:18.050 Of course, fault and this for company. 00:24:18.270 --> 00:24:18.900 That's what you need. 00:24:18.910 --> 00:24:29.300 You can also cause the local stress concentration on some part of the faults and hence the further affect the multi scale force rupture. 00:24:34.340 --> 00:24:48.740 So after the analysis about the interest sesame deformation so the the intercessory deformation, we analyze, the only shows the average the behavior during the whole intercement period. 00:24:49.040 --> 00:24:57.040 But we don't really know how the what is the current status of the fault in California? 00:24:57.230 --> 00:25:00.900 So how can we use the focal mechanism to understand it? 00:25:01.760 --> 00:25:04.130 Umm, before uh. 00:25:04.140 --> 00:25:09.490 Talk about the stress status of the the full system of California. 00:25:10.260 --> 00:25:18.190 I just want to briefly introduce the California's current for system and its deformation. 00:25:19.020 --> 00:25:31.400 So currently the California is located at the transform boundary between the Pacific Plate and the North America Plate, and the most of the relatively motion are commodate. 00:25:31.520 --> 00:25:42.220 To buy those Andreas fault and partially accommodated by worker learn, share zone and Eastern California Shear Zone in the in Southern California. 00:25:42.290 --> 00:26:00.180 The Big Bend of Los Angeles, Fort and the Transverse Range developers in association with the opening of Golf, California on the right side, shows the corresponding string field related to the current force system. 00:26:00.770 --> 00:26:30.050 And it's important to note that this string field, the geodetically, observe the string field only shows the temporal fluctuation of the cross defamation, while the observe the stress field shows the stress accumulated across the long term across the deformation and it provide us the information about the geological history in this area and also the current status of the stress. 00:26:30.620 --> 00:26:35.430 Umm can provide us some information about the assessment hazard. 00:26:36.670 --> 00:26:53.360 So when you utilize the focal mechanism catalog we developed to perform the stressing version, use stress inverse program and with 0.5 * 0.05 degree weight. 00:26:53.630 --> 00:26:58.000 And here shows the result hit the stress field. 00:26:58.470 --> 00:27:03.900 The left side shows the the my view of the stress field colored by the 14 style. 00:27:04.030 --> 00:27:25.260 The right side shows uh, so the maximum horizontal stress as meals in the for the previous study shows that the for the large scale variation of the stress field in California is mainly with can be. 00:27:26.900 --> 00:27:31.040 Explained by the gravitational potential energy and the relative plate motion. 00:27:33.150 --> 00:27:53.400 But the uh this because benefiting from or large volume of focal mechanisms we can solve more detailed local scale stress rotation and we find that this local scale stress heterogeneity shows a strong association with the major faults. 00:27:54.830 --> 00:28:07.220 Some area has the the stress rotation caused by the Inter fault interaction, like the area between near the intersection of the San Andreas Fault and the Sangria Fault. 00:28:07.530 --> 00:28:39.360 Because of the relative motion between these two fault, it can cause the the local normal 14 stress regime near this Fort wage and also counterclockwise rotation of the just just uh orange maximum horizontal stress orientation and also the changing angle of the San Andreas Fault and lead to this folk band and the compressional stress regime near the Santa Cruz Mountain. 00:28:42.960 --> 00:29:03.950 And in Southern California, umm, the interaction between the left and lateral faults in the Transverse Ranges and the right lateral faults stress the faults of the elsinoe, San Jacinto and the San Andreas Fault caused the mass inflow. 00:29:03.990 --> 00:29:28.190 Near the end, the reverse faulting strategy gene in the eastern Transverse Ranges, but mass outflow and the normal 14 stress regime in the western Transverse Ranges and near and for this the interaction of this subparallel right lateral stressfully faults and the this this falls near the wedge. 00:29:28.200 --> 00:29:38.770 It can also cause the counterclockwise stress rotation between these separate legal Taurus Jack sleep faults. 00:29:40.280 --> 00:30:14.350 Some stress rotation can also be caused by the sleep partitioning, like the central Centers for the area, even though overall the stress regime is reverse sporting style and but because of the highly high sleep rate along the central Angels fought, it shows that the strict sleep stress regime and the counterclockwise rotation of the Ashish Max near the central San Andreas Fault. 00:30:14.390 --> 00:30:37.130 Because of this high strain rate near the central central spot, and similarly in the area near the Panamint Valley Fort not to the Garlock Fault, this high strain rate near this major fault also cause this clockwise rotation around the faults. 00:30:40.010 --> 00:30:53.860 In the worker lunch years on Ohh we also observe a a strong association between the minimum horizontal stress orientation and the and the major for the orientation. 00:30:54.030 --> 00:31:38.590 The minimum stress orientation always show very high angle to the major faults orientation and it shows us the normal 14 stress regime, suggesting that in this part uh it's the it's mental this this is stress regime is mainly related to the crystal stretching of the basin and Range province because the worker that she is always located as the western boundary of the basin and range even though it currently accommodate one force of the relative plate motion but still or observe the stress field suggesting that it's mainly related to the I still stretching. 00:31:41.750 --> 00:31:59.630 In the on the one hand, the fault orientation can cause this local stress rotation and on the other hand the stress can also help us to understand the the current fault status wait. 00:31:59.740 --> 00:32:02.170 We use the stress tensor. 00:32:02.360 --> 00:32:04.110 We obtained an uh. 00:32:04.120 --> 00:32:20.630 Combine it with the major photo orientation to compute the normal stress and shear stress on the floor plan and then projected to the more circle to see whether this fault is optimally oriented to fail or not. 00:32:21.460 --> 00:32:30.560 If there's a project point is near the critical line, it means that this folder is optimally oriented to fail, and that's easy. 00:32:30.710 --> 00:32:34.940 If we put a certain stress, it's it's more likely to fail. 00:32:35.370 --> 00:32:44.700 But if this point is away from this critical line, it will have a a low photo instability and it's less likely to fail. 00:32:44.730 --> 00:32:51.510 Under the current stress regime in the way computed, this 14 stability. 00:32:51.520 --> 00:33:05.440 Of course, the whole California most area have the high fault instability, which means that this forge are optimally oriented to fail, and there's just the current stress field. 00:33:06.200 --> 00:33:18.500 But some region shows the low 40 instability some of because of the limited stress accumulation like the Garlock Fault and the white Wolf fault area. 00:33:18.630 --> 00:33:54.890 This fault orientation are shown really high angle to the relative plate motion direction so that they are not low like the the retaining plate motion doesn't doesn't really load these fault a lot and for the area northeast to the Clear Lake area, even though it along well with the Rock Hill played motion direction but it's located between the San Andreas Fault and the Mendocino Triple Junction area which means this part the played motion direction can be. 00:33:57.060 --> 00:34:06.440 Different and the more complicated compared with uh, the the relative compared with the the other part of the transform boundary. 00:34:06.520 --> 00:34:25.490 So that even though this major faults are oriented aligned while with other part of the Sanders for the better, maybe in this area they are also show a certain angle to the plate motion direction and it is not optimally oriented tool fail. 00:34:27.580 --> 00:34:35.430 Some region, as we discussed before, they have a the local stress rotation like the Centrals angels for the area. 00:34:35.960 --> 00:34:43.470 Uh, this stress rotation caused by the sleep partitioning and the sudden Sandra's for the area is local. 00:34:43.480 --> 00:34:46.490 Stressful situation caused by the interface interaction. 00:34:46.680 --> 00:34:59.720 So even though this fault are alike, while with the relative plate motion direction, but because of the low cost strategy rotation, this water are also not optimally oriented to fail. 00:35:01.050 --> 00:35:15.460 And for the Eastern California Shear Zone, it has the most, the very recent, uh, major ruptures like the lenders and hackmann earthquakes, which has released a considerable amount of the stress in this area. 00:35:15.670 --> 00:35:21.950 And making this region or less optimally oriented to fail. 00:35:24.440 --> 00:35:25.150 So. 00:35:27.620 --> 00:35:51.320 The takeaway message from this part is that based on the stress map of California, we find a very strong forty and stress association on the one hand, the major fault kinematics can cause the local structure rotation by the 14 interactions, the petition and the the full geometry orientation. 00:35:51.800 --> 00:35:56.450 On the other hand, the stress loading can also affect the major faults. 00:35:56.460 --> 00:36:04.160 Instability for the area with limited stress loading and local stress rotation and recent stress release. 00:36:04.550 --> 00:36:16.610 There are last like for this part of the major force has low instability and the less likely to fail, and there's the the they're the stress orientation. 00:36:19.580 --> 00:36:27.230 So we have a analyze the intersection, make deformations and also. 00:36:27.970 --> 00:36:31.240 On about the kindness, just status. 00:36:31.630 --> 00:36:41.660 But when the stress reaches a certain reaches, the crystal strains it will cause the earthquake rupture and release the energy. 00:36:42.800 --> 00:36:56.220 So how the smart earthquake focal mechanisms can help us to understand the stress release in the frozone during the earthquake rupture ohm, which means uh? 00:36:56.230 --> 00:37:09.110 How it's important for us to better solve the earthquake source properties and understand how these earthquakes release the stress in the crust. 00:37:11.680 --> 00:37:20.210 Here I'll show you the overview of the different uh of the earthquake source property estimations. 00:37:20.220 --> 00:38:07.220 Of course, different magnitude for the large earthquakes because we eat has the large amplitude and the longer duration waveforms so that we can have a plenty of high quality waveforms for detailed spatial temporal analysis and for get a more detailed physical parameters for this large earthquakes like the finance source inversion, we can get the physical properties on the 4th plan and backprojection can trace the the kinematics of the the earthquakes and how they energy points moves in time and for the multiple moment tensor inversion it can it. 00:38:07.460 --> 00:38:15.290 It could be very useful for the multiple fought segment rupture during a single large earthquakes. 00:38:17.490 --> 00:38:23.560 And the the days uh method can only be applied to a very large earthquake. 00:38:23.570 --> 00:38:52.270 Let's say magnitude larger than six or seven earthquakes, but for the moderate magnitude earthquake they have, like maintained from 3 to 6 earthquake, they have a the umm relatively higher frequency records and smaller amplitude so that it's more challenging to solve this as detailed as those larger squares. 00:38:53.540 --> 00:38:59.110 So we can only get the last source parameters or less spatial temporal informations. 00:38:59.800 --> 00:39:07.700 For example, instead of a finance sourcing version, we can only get a second or moment with the simple source radius. 00:39:08.170 --> 00:39:18.560 People velocity and sleep direction for the earthquake and instead of multiple moment tensor inversion we can only perform the multiple source inversion. 00:39:18.870 --> 00:39:34.250 To trace how this earthquake evolves, and also we cannot get to the multiple moment tensor, maybe the observation can be only plenty for a single moment tensor inversion. 00:39:36.470 --> 00:39:40.800 And if we go even smaller for magnitude or smaller than three or quake. 00:39:41.900 --> 00:39:51.990 Uh, it's also really challenging to solve the source time function when it's really challenging to solve it in time domain because it they have really high frequency. 00:39:52.700 --> 00:40:10.140 So usually we extract some way from features and by analyzing this waveform features we extract some source properties like for the stress job and search radius we use the earthquake source Spectra. 00:40:12.070 --> 00:40:25.140 And the for earthquake rupture directivity, we use the small variation of the source spectral and focal mechanism with the polarity and amplitude and for the photo orientation. 00:40:25.150 --> 00:40:36.920 If we have a group of earthquakes, we can analyze their alignment and to see and estimate the photo orientation based on the earthquake clusters. 00:40:38.320 --> 00:40:48.640 However, there's no integrated solution for the smaller quakes, and there's no way to estimate the single fault orientation. 00:40:49.240 --> 00:41:05.540 Therefore, here we proposed that we can combine uh, the information of the polarity, amplitude and spectral to jointly inverse all the small earthquake source properties together. 00:41:05.600 --> 00:41:05.800 Sure. 00:41:07.530 --> 00:41:08.050 Here. 00:41:08.830 --> 00:41:09.620 I will show you. 00:41:09.630 --> 00:41:12.210 Here is the reason why we want to do so. 00:41:13.260 --> 00:41:24.170 Umm, so for the smaller script rupture, if we solve it independently for the focal mechanism, we can just two nodal plan. 00:41:24.340 --> 00:41:32.530 We cannot tell which ones the full plan, but if we ohm combine the ruptured directivity estimation. 00:41:32.870 --> 00:41:46.220 Uh into the the focal mechanism estimation we can use the rupture directivity to see like which nodal plane can be, uh. 00:41:46.260 --> 00:42:02.200 More suitable to explain the observe the rupture directivity and use it to pick the one model plan formula two to tell the smaller scale photo orientation and the for the search radius and the stress job estimation. 00:42:02.440 --> 00:42:15.690 You usually people use the stacked Spectra from all the spectrum from different stations and for that process they ignore the effect of the photo orientation and the rupture directivity. 00:42:16.390 --> 00:42:26.200 But if we combine the full mechanism and the rupture directivity into the inversion processes, we can like incorporate this. 00:42:26.310 --> 00:42:34.180 We can remove this effect during the time we estimate the stress job and the for the rupture directivity. 00:42:35.070 --> 00:42:41.680 Umm if we solve it independently, we need to search the rupture directivity orientation. 00:42:42.590 --> 00:43:04.910 I am in a 3D volume, but if we you have the focal making solution we can search the rupture directivity only along these two nodal plan which can significantly decrease our search volume and help to solve a 3D rupture directivity with lower uncertainty. 00:43:06.080 --> 00:43:17.350 So in this way we can get a better source property estimation for each small earthquakes in the fall zone so that we can get. 00:43:18.850 --> 00:43:33.820 More better understanding of the fun scale of Falzone behaviors and also if we get an integrated solution, this solution can be comparable to the other. 00:43:34.330 --> 00:43:49.000 Umm for the parameter uh solutions from from other, uh larger magnitude events and by which we can get a better understanding of the earthquake source processes. 00:43:49.010 --> 00:44:18.000 Of course, different scales, so this is the currently ongoing work and we are trying to uh develop this method and further apply it to this larger small earthquake data sets in California and hopefully it can help us to get a more new understanding about the earthquake of and for the earthquake and of four to mechanics. 00:44:20.630 --> 00:44:25.250 So here is a summary about what I have. 00:44:25.260 --> 00:44:25.590 Uh. 00:44:25.600 --> 00:44:38.660 Shown today, I'll show you the how the small earthquakes can help to understand the intercessor maker for the deformation and how it can help us to uh. 00:44:38.910 --> 00:44:53.350 Batcher analyze the current stress status and how you can help to solve the smaller squiggle rupture properties in the better help us to understand the the earthquake rupture processes. 00:44:55.420 --> 00:44:59.830 It then also it's also promising to apply. 00:44:59.910 --> 00:45:08.830 Uh, the the understanding from this different aspect to the practical seismic hazard assessments. 00:45:09.380 --> 00:45:27.080 For example, the Belcher constrained intercepts make deformation model can be incorporated into the user and the the for the estimation of the current stress status, especially the for the fault instability. 00:45:27.290 --> 00:45:44.150 Maybe we can incorporate that into the earthquake rate model for better for optimizing the estimation of the that's a possible, uh, large magnitude events UM. 00:45:44.780 --> 00:46:06.040 And also if we get to the better, UM, smaller square because it's properties and because we have larger volume of small earthquakes and their ground motions, we can better understand how the source term uhm play their Joe in the observe the ground motion and optimize the ground motion model. 00:46:06.770 --> 00:46:11.740 And in this way, it's in the future it's promising to. 00:46:13.450 --> 00:46:20.950 Incorporate the smaller strike focal mechanism data set to optimize or assessment assessment. 00:46:24.080 --> 00:46:35.990 So here hope I have shown you that it's a small earthquakes can help us to see to better and rival the full zone behaviors. 00:46:36.160 --> 00:46:39.810 Course, different scales and uh. 00:46:41.500 --> 00:46:58.940 Welcome questions to in in the if you have a, if you want to discuss with me more about the potential applications and you can also email me and we can discuss and thanks for your listening. 00:46:59.230 --> 00:47:00.800 Welcome questions. 00:47:02.440 --> 00:47:02.880 Thank you. 00:47:02.890 --> 00:47:03.290 Thank you. 00:47:04.210 --> 00:47:05.090 What's the time? 00:47:08.780 --> 00:47:09.550 Uh, so? 00:47:09.210 --> 00:47:10.030 Uh, so? 00:47:17.920 --> 00:47:18.320 OK. 00:47:17.430 --> 00:47:20.210 We'll take questions from. 00:47:20.540 --> 00:47:21.440 Let's start with online. 00:47:23.090 --> 00:47:27.730 So, uh, Fred, if you want to go first, that sounds great. 00:47:30.150 --> 00:47:30.800 Oh, hi. 00:47:30.810 --> 00:47:32.630 Thanks for a really fascinating talk. 00:47:41.040 --> 00:47:41.200 No. 00:47:33.810 --> 00:47:43.100 Umm, I wanted to return to the disagreement between stress orientations and strain orientations, like how severe was that? 00:47:43.110 --> 00:47:45.660 Are there any places where there's a strong disagreement? 00:47:47.410 --> 00:47:51.250 Well, uh, thanks for the question actually. 00:47:51.260 --> 00:47:51.540 Uh. 00:47:51.550 --> 00:48:05.200 The stress and strain they are showing totally different things, so it's a it's really hard to as you can see here it only shows the high strain rate or or. 00:48:05.610 --> 00:48:18.070 Ohh yeah, because I didn't compare the orientation yet, but but definitely for example for the the work line share zone based on the string field that you can see it has me too. 00:48:18.180 --> 00:48:20.130 It's mainly accommodated. 00:48:20.140 --> 00:48:21.560 The relatively the motion. 00:48:22.690 --> 00:48:23.820 Ohh, like it. 00:48:23.830 --> 00:48:26.300 Accommodate about 1/4 of the relative plate motion. 00:48:27.140 --> 00:48:41.190 Umm, but from the stress field it shows this mainly dominated by the normal 14 stress regime and the with the the minimum horizontal stress orientation showing high angle to the major faults. 00:48:41.400 --> 00:48:55.450 So it's mainly represented the crystal stretching in basin and Range Province instead of the transform boundary of this between the Pacific Plate and North America plate. 00:48:56.810 --> 00:48:57.240 OK. 00:48:57.250 --> 00:48:57.940 Yeah. 00:48:55.500 --> 00:48:58.420 So saying that's. 00:48:57.990 --> 00:49:03.730 Yeah, it would be great to compare them in more detail and see maybe how it varies with distance from the fault. 00:49:04.990 --> 00:49:05.790 Yes, we are. 00:49:05.800 --> 00:49:17.120 We are trying to find a a good uh string field model that can can better, uh compare between or stress model and the the string field model. 00:49:18.850 --> 00:49:19.040 OK. 00:49:17.410 --> 00:49:19.680 Yes, thanks for the suggestion. 00:49:20.890 --> 00:49:21.020 Yeah. 00:49:21.230 --> 00:49:21.530 Thank you. 00:49:24.530 --> 00:49:24.940 Alright. 00:49:24.950 --> 00:49:27.070 And I think there's another question from yens. 00:49:29.500 --> 00:49:32.110 I thank you for a very interesting talk. 00:49:32.120 --> 00:49:45.760 Those are some impressive methods and results that are I I may have missed this, but earlier the sort of towards the conclusions of your first section, you were talking about calculating the moment accumulation rate over I think the last 100 years. 00:49:46.290 --> 00:49:46.430 Yes. 00:49:46.040 --> 00:49:47.530 And I can. 00:49:47.540 --> 00:49:51.630 I may have missed how you did that, but I'm curious how you approach that. 00:49:54.280 --> 00:49:55.490 Oh, OK. 00:49:55.800 --> 00:49:56.430 Yes. 00:49:56.210 --> 00:49:56.610 50 years. 00:49:56.500 --> 00:49:59.050 Uh, so for, for, for this. 00:49:59.760 --> 00:50:16.540 Because we said this for coupling model and they will perform the this in like modeling based on this for coupling model to to see how other four deforms during the interesse maker period and then we get this model to creep rate. 00:50:24.380 --> 00:50:25.450 OK, so the creeper. 00:50:37.320 --> 00:50:38.100 OK, got it. 00:50:19.310 --> 00:50:46.210 So the model creep rate over 150 years in the versus, yeah, if if the whole segment is fully freely creeping without a lock patch, it's supposed to be 34 millimeter per year, so that uh, wait, can get the deficit rate and use that to estimate the stored energy, yeah. 00:50:46.070 --> 00:50:47.440 OK, I I misunderstood. 00:50:47.450 --> 00:50:50.740 I thought I was making from the stuff, but another question is sort of more general. 00:50:50.750 --> 00:50:52.260 One is there's a lot of evidence. 00:50:53.610 --> 00:51:11.510 Think some from Lavender mints of the USGS that some of the false along the San Andreas system have local efficiency friction, and I'm wondering if that if that's something you're considering when it comes to, for example, thinking about how, how well oriented or, you know, misoriented various faults. 00:51:13.140 --> 00:51:16.740 You, you, you, you mentioned the sum of Forza has the friction. 00:51:29.580 --> 00:51:30.970 Outlook Ohio. 00:51:18.550 --> 00:51:40.380 Some of those folks segments based on a few other evidence, including laboratory measurements at the USGS, have shown low coefficients of friction along some part of the San Andreas system that are maybe lower than you know, which local firing stirring thoughts in the interpolated area. 00:51:40.910 --> 00:51:41.050 Yes. 00:51:40.690 --> 00:51:45.260 And so I'm just wondering if that is something that you know is is worth considering for. 00:51:47.070 --> 00:51:52.310 When you calculate how misoriented those faults are, because it might be different in the different location friction. 00:51:54.690 --> 00:51:57.560 Umm and I'm missing. 00:51:59.960 --> 00:52:02.730 Right, that, that might be the problem. 00:52:18.100 --> 00:52:18.520 OK. 00:52:05.030 --> 00:52:27.010 Yes, uh, I think for the current for the instability, I'm using the friction coefficient that I obtained a based on the stress inversion because the for the stress inversion I need to search over our possible friction coefficient and all possible stress orientation to find the best fitting one. 00:52:27.380 --> 00:52:33.220 So for each grade we also have a local friction coefficient, but the. 00:52:35.950 --> 00:52:37.770 Got yeah, but I understand. 00:52:37.780 --> 00:52:38.790 I know this messengers. 00:52:34.460 --> 00:52:41.530 Yeah, maybe that's that's all that to the major photo friction coefficient. 00:52:41.540 --> 00:52:49.330 But uh, that can kind of indicate the local drug status, yeah. 00:52:49.080 --> 00:52:50.220 You think you that depends on. 00:52:51.470 --> 00:52:51.970 Thank you. 00:52:55.290 --> 00:52:58.800 And we have a couple questions in the room here at Moffett. 00:52:59.310 --> 00:53:00.660 Rob, if you wanna go. 00:53:00.770 --> 00:53:01.040 Yeah. 00:53:01.050 --> 00:53:02.280 Thanks a lot for a great talk. 00:53:02.910 --> 00:53:04.120 A lot of work clearly. 00:53:04.190 --> 00:53:04.900 You put into this. 00:53:11.900 --> 00:53:12.020 Yes. 00:53:05.550 --> 00:53:12.580 I think you said that you had about 800,000 earthquakes that meet your criteria of eight blurry measurements, but you know, really what the? 00:53:12.930 --> 00:53:16.400 How many of those earthquakes produce dependable mechanisms? 00:53:16.460 --> 00:53:19.530 You know, in terms of the quality, how many of those are A or B quality? 00:53:20.650 --> 00:53:21.480 Uh. 00:53:21.830 --> 00:53:22.890 A RB quality. 00:53:26.910 --> 00:53:27.680 You mean here? 00:53:29.360 --> 00:53:46.730 Actually AOB quality thing, it's a lot of parameter incorporated in the AM B qualities and yeah, I also personally I I redefined it in in my paper about the Southern California focal mechanism catalog. 00:53:47.600 --> 00:53:51.890 Umm, so it it really depends. 00:53:51.900 --> 00:53:56.400 Especially there's some criteria about the station coverage. 00:53:56.710 --> 00:54:07.770 If the take off angle gap or ask meals or gap is large it it will be treated as a low quality for call mechanism. 00:54:08.240 --> 00:54:30.070 So in that way, it's really hard to use those parameter to to judge the performance of the focal mechanism calculation and the the, the the improvements of the solutions because of the station is there, it's it cannot get more station, get better coverage, yeah, so. 00:54:32.060 --> 00:54:35.370 It it I I think it really depends. 00:54:35.380 --> 00:54:47.170 But uh yeah, for the quality RB, I will say it should be around the 30% or 2530%. Mm-hmm. 00:54:48.030 --> 00:54:48.350 Thank you. 00:54:49.400 --> 00:54:49.550 Yeah. 00:54:49.750 --> 00:54:51.960 Alright, one more question from the room. 00:54:53.520 --> 00:55:01.220 So maybe you could go back to the slide that was discussing the creep distribution and the central San Andreas Fault. 00:55:04.030 --> 00:55:05.420 You mean where he? 00:55:08.450 --> 00:55:08.710 OK. 00:55:05.220 --> 00:55:27.900 Yeah, this or that one maybe or maybe one back it not too important, but I'm curious if you had some thoughts as to what your results might tell us about the possibility of a large rupture going through the creeping section which have has been discussed in some of these other studies that you are showing as a comparison. 00:55:28.190 --> 00:55:30.330 I'm wondering if you have some new insights about that. 00:55:32.510 --> 00:55:34.740 Ohh said, thanks for the for. 00:55:34.750 --> 00:55:35.740 Thanks for the question. 00:55:36.170 --> 00:55:37.640 Uh, at least. 00:55:37.650 --> 00:55:39.000 Uh umm. 00:55:39.070 --> 00:55:49.660 So for previous study, as you can see for for some previous for coupling models they see a very highly creeping part in the central area. 00:55:50.170 --> 00:55:57.360 So but which it can be a barrier because the there's a limited stress accumulated in the central area. 00:55:57.650 --> 00:56:08.230 It's harder to rupture through it, but here we we have a some local locked patch and the these are distributed. 00:56:08.240 --> 00:56:09.420 Of course, the fault. 00:56:10.260 --> 00:56:25.130 So if we have some some logs we have in the central area, it's indeed will increase the possibility for the major earthquake rupture to rupture through date compared with other folk coupling model. 00:56:25.420 --> 00:56:34.330 But on the other hand, because they have some, there's some local very local log pattern instead of the large logs patch. 00:56:34.490 --> 00:56:47.920 As a result, it's also increase the possibility for this local logged patch to rupture themselves as a moderate earthquake or the after sleep of war, the larger earthquake. 00:56:49.040 --> 00:57:00.120 So it it it's it's harder to have a a single sided conclusion based on our observations. 00:57:00.160 --> 00:57:08.950 So yeah, but that's the thing we we can tell from here, it's more heterogeneous and more distributed, small like the Hatch. 00:57:10.010 --> 00:57:10.480 Yeah. 00:57:10.490 --> 00:57:19.270 And the currently we indeed we don't really see any and the observation showing that the IT can rupture through it. 00:57:19.560 --> 00:57:30.520 But we can use this new for coupling model to do some to do some rupture modeling to see whether it can increase or decrease the possibility. 00:57:30.960 --> 00:57:36.130 Yeah, that's all I get from it, yeah. 00:57:36.670 --> 00:57:36.980 OK. 00:57:36.990 --> 00:57:37.420 Thank you. 00:57:38.830 --> 00:57:39.940 Thank you so much. 00:57:39.950 --> 00:57:42.460 We appreciate you staying up so late to give this talk. 00:57:42.470 --> 00:57:45.080 And with that, we'll thank our speaker with a round of applause. 00:57:46.280 --> 00:57:46.640 Thank you.