WEBVTT Kind: captions Language: en-US 00:00:00.950 --> 00:00:02.360 All right. Hello, everybody. 00:00:02.360 --> 00:00:08.260 Thank you for being here for another edition of the Earthquake seminar. 00:00:08.260 --> 00:00:11.860 Apologize for last week’s seminar getting cancelled on very short notice. 00:00:11.860 --> 00:00:15.700 We are working to reschedule Noha’s talk. 00:00:15.700 --> 00:00:21.080 Next week will be Kris Pankow from University of Utah. 00:00:21.080 --> 00:00:27.740 And today’s speaker is one of our own, Walter Mooney. 00:00:27.750 --> 00:00:31.810 Walter did his undergrad at Cornell and Ph.D. at 00:00:31.810 --> 00:00:35.220 University of Wisconsin-Madison, right? 00:00:35.220 --> 00:00:42.820 And he has focused a lot of his career on imaging of the continental lithosphere. 00:00:42.820 --> 00:00:46.260 And that’s, of course, taken him all over the world. 00:00:46.260 --> 00:00:50.989 So Walter is a – you know, one of our finest examples of international 00:00:50.989 --> 00:00:56.170 scientific collaboration here at USGS. He has a lot of collaborations. 00:00:56.170 --> 00:00:59.760 He travels a lot, and he also, to our benefit, brings a lot of 00:00:59.760 --> 00:01:03.500 interesting people from all over the world here to Menlo Park. 00:01:03.500 --> 00:01:07.980 And today, he’s going to be talking to us about imaging the 00:01:07.980 --> 00:01:11.940 North American upper mantle. - Thank you, Jeanne. 00:01:14.160 --> 00:01:22.540 [Silence] 00:01:22.549 --> 00:01:26.770 So I thought I’d at least wear my hat up to the podium. Now I can take it off. 00:01:26.770 --> 00:01:30.500 You wouldn’t know it was me if I didn’t have the hat on. 00:01:31.900 --> 00:01:34.100 Imaging the North American upper mantle. 00:01:34.100 --> 00:01:40.329 This is a really exciting new topic that’s been made possible by 00:01:40.329 --> 00:01:47.329 amazing data sets that have been coming in from IRIS and others that – 00:01:47.329 --> 00:01:51.709 you know, a lot more seismometers. And the talk I’m giving today is 00:01:51.709 --> 00:01:58.049 a paper that’s in review in JGR. The reviews just came back. 00:01:58.049 --> 00:02:05.200 The first author is my colleague Rainer Kind, who I first met in Germany 00:02:05.200 --> 00:02:09.520 when I was a graduate student in 1976. So we go back 43 years. 00:02:09.520 --> 00:02:12.459 So we’ve been working on this paper for about 43 years, 00:02:12.460 --> 00:02:14.820 I guess you could say. [laughter] 00:02:14.820 --> 00:02:22.640 And Yuan Xiaohui is a colleague originally from Germany who is 00:02:22.640 --> 00:02:29.930 now working also in Potsdam at the so-called GFZ, GeoForschungsZentrum, 00:02:29.930 --> 00:02:34.480 which is, of course, a big, powerful institution in Germany. 00:02:34.980 --> 00:02:41.120 Thank you, Jeanne and Sara for putting together all these seminars. 00:02:41.130 --> 00:02:45.340 You know, last week’s seminar at NOAA was cancelled. 00:02:45.340 --> 00:02:49.440 And I was thinking to myself, if I were about to speak, and my seminar were 00:02:49.440 --> 00:02:54.239 cancelled, I wonder how I could deal with that psychologically. [laughs] 00:02:54.240 --> 00:02:58.820 Can you cancel one day before, you know? Is there a message? 00:02:58.820 --> 00:03:03.079 Anyway, I’d like to also acknowledge David P. Hill, 00:03:03.079 --> 00:03:07.890 who died about three months ago. David was a dear friend to 00:03:07.890 --> 00:03:12.799 almost everyone in this audience. He was a real gentleman, and he was 00:03:12.799 --> 00:03:20.209 the person who hired me at the Survey in October of 1978. 00:03:20.209 --> 00:03:23.329 Two images of Dave. The one on the left is what most of us 00:03:23.329 --> 00:03:27.400 think of him – remember him as. I don’t remember that often – 00:03:27.400 --> 00:03:32.760 seeing Dave that often in a tie. He was a real Western-style fellow. 00:03:32.760 --> 00:03:40.440 And I learned an awful lot from Dave Hill about science and humanity. 00:03:41.480 --> 00:03:44.540 So what are my – what’s all this noise? 00:03:44.540 --> 00:03:48.300 What am I going to be telling you about? 00:03:49.760 --> 00:03:53.670 Try to give you a new view of the structure and evolution of the upper 00:03:53.670 --> 00:04:00.300 mantle of North America. And, in fact, about the upper mantle in general. 00:04:00.300 --> 00:04:04.239 And it would be a fair question to ask, well, why worry about the upper mantle 00:04:04.239 --> 00:04:11.129 if we’re focused on earthquake hazards and sometimes, in this group, 00:04:11.129 --> 00:04:16.829 also volcano hazards? Well, it turns out that the seismicity 00:04:16.829 --> 00:04:21.960 of North America is very well correlated with mantle structure. 00:04:21.960 --> 00:04:27.220 Basically, regions of weak rheology in the upper mantle are areas 00:04:27.230 --> 00:04:32.700 where earthquakes get focused. And Fred Pollitz and I, 00:04:32.700 --> 00:04:36.600 and my other colleagues, have written some papers recently. 00:04:36.600 --> 00:04:38.700 Turns out both of them are in EPSL. 00:04:38.700 --> 00:04:44.890 I just want to point out that the upper mantle is in the title of both of these. 00:04:44.890 --> 00:04:49.150 Crustal seismicity and earthquake catalog – the maximum moment 00:04:49.150 --> 00:04:52.810 magnitude for stable continents. Correlation with the seismic 00:04:52.810 --> 00:04:56.950 velocity of the lithosphere. So it’s more than the crust. 00:04:56.950 --> 00:05:01.780 If you have weak lithosphere, it deforms more rapidly and seismicity 00:05:01.780 --> 00:05:07.970 is focused there – concentrated. The second paper with Fred Pollitz – 00:05:07.970 --> 00:05:11.960 seismic structure of central U.S. crust in shallow upper mantle. 00:05:11.960 --> 00:05:17.280 The uniqueness of the Reelfoot Rift. You know, Seth Stein has postulated 00:05:17.280 --> 00:05:22.110 that the seismicity of North America is basically random. 00:05:22.110 --> 00:05:26.130 That New Madrid turned on bright and then next – the next place 00:05:26.130 --> 00:05:28.970 you could have a lot of earthquakes might be Minnesota. 00:05:28.970 --> 00:05:33.590 And our point in this paper – Fred and I – is that that’s not true. 00:05:33.590 --> 00:05:38.360 That the New Madrid region is quite unique because it has 00:05:38.360 --> 00:05:44.450 special weak subcrustal lithosphere. 00:05:44.450 --> 00:05:48.070 So anyway, getting back to the motivation for today’s talk, looking – 00:05:48.070 --> 00:05:50.590 sometimes looking beneath the crust is worthwhile, 00:05:50.590 --> 00:05:55.410 even if you’re looking at questions such as earthquake hazards. 00:05:55.410 --> 00:06:01.720 Well, seismologists have been studying the upper mantle for quite a long time. 00:06:01.720 --> 00:06:08.860 And two big contributors were Sir Harold Jeffreys and Keith Bullen. 00:06:08.860 --> 00:06:15.940 Keith Bullen was a assistant of Jeffreys. Keith Bullen was from New Zealand 00:06:15.940 --> 00:06:21.250 and went to Cambridge University to help Jeffreys compile the earthquake 00:06:21.250 --> 00:06:25.000 travel times and determine the interior of the Earth. 00:06:25.000 --> 00:06:27.750 As you can see, Jeffreys lasted quite a long time. 00:06:27.750 --> 00:06:34.120 He lived to the ripe age of 98 and was famous not only for studying 00:06:34.120 --> 00:06:37.840 the Earth’s interior, but for his opposition to plate tectonics. 00:06:37.840 --> 00:06:41.840 But that’s not the topic of today’s talk. So they came up with 00:06:41.840 --> 00:06:47.090 a standard model of the Earth. Their first model was published in the 00:06:47.090 --> 00:06:52.650 Transactions of the Royal Society, 1936, and they showed that the Earth 00:06:52.650 --> 00:06:59.580 has boundaries such as the inner and outer core, the lower mantle, 00:06:59.580 --> 00:07:02.690 the core-mantle boundary, and then the transition zones – 00:07:02.690 --> 00:07:07.660 the famous transition zones at 410 and 660. 00:07:07.660 --> 00:07:10.229 Should I be looking somewhere for an electronic pointer, or … 00:07:10.229 --> 00:07:13.750 - You could use the mouse. - Oh, use the mouse. Okay. 00:07:13.750 --> 00:07:16.030 That works. All right. Yeah, there we are. 00:07:16.030 --> 00:07:23.770 So here’s the so-called mantle transition zone between 410 and 660. 00:07:23.770 --> 00:07:26.520 And then another boundary, which is much debated, 00:07:26.520 --> 00:07:31.820 the 220 discontinuity below the lithosphere. 00:07:32.960 --> 00:07:38.040 Inge Lehmann. Inga Lehmann was a Danish seismologist. 00:07:38.050 --> 00:07:44.110 As you can see, she lived to be 105. And I’d like to point out that, if you use 00:07:44.110 --> 00:07:48.610 Jeffreys and Inge Lehmann as examples, seismology is a very good field 00:07:48.610 --> 00:07:53.240 for a long life. [laughter] Geochemists don’t last very long. 00:07:53.240 --> 00:07:59.170 They fight with each other constantly, and seismologists do really much better. 00:07:59.170 --> 00:08:03.800 So Inge Lehmann found the Earth’s inner core and suggested 00:08:03.800 --> 00:08:07.610 the Lehmann Discontinuity. Let’s talk about the inner core first. 00:08:07.610 --> 00:08:13.039 She looked at this data shown here – on the upper trace, the zed component. 00:08:13.039 --> 00:08:17.710 And it was a really big phase that was only evident on the – 00:08:17.710 --> 00:08:20.970 on the P wave component. And it doesn’t – because the 00:08:20.970 --> 00:08:24.170 outer core is liquid, there’s no S wave transmission. 00:08:24.170 --> 00:08:30.950 So she realized in her 1936 paper that there was a solid inner core. 00:08:30.950 --> 00:08:36.810 And the second contribution that Inge Lehmann made was to postulate what 00:08:36.810 --> 00:08:42.060 she called the 20-degree discontinuity. It’s a paper in the BSSA. 00:08:42.060 --> 00:08:47.090 And it’s a velocity increase below about 200 kilometers where both P and 00:08:47.090 --> 00:08:55.240 S undergo a rapid increase. So this is the so-called Lehmann Discontinuity. 00:08:55.240 --> 00:08:59.490 I’ll talk about the lithosphere- asthenosphere boundary in a minute. 00:08:59.490 --> 00:09:06.610 But so the key points are that there are two very strong discontinuities in the 00:09:06.610 --> 00:09:11.230 upper mantle – the 410 and the 660. And these are – these are strong 00:09:11.230 --> 00:09:16.460 phase boundaries. They’re quite sharp. And I think when John Vidale was here 00:09:16.460 --> 00:09:22.820 working at the USGS in the mid-1990s, he worked with Harley Benz using 00:09:22.820 --> 00:09:29.550 the California network to define the sharpness of the 410 and 660. 00:09:29.550 --> 00:09:35.430 And it’s about 5 to 10 kilometers. They are about 5 to 10 kilometers thick. 00:09:35.430 --> 00:09:42.670 So there are no other phase – olivine is the most abundant constituent mineral 00:09:42.670 --> 00:09:47.260 of the upper mantle, and there are no other phase boundaries above the 410, 00:09:47.260 --> 00:09:51.990 so we shouldn’t be seeing any sharp phase boundaries above 410. 00:09:51.990 --> 00:09:56.830 Now, we do have, of course, orthopyroxene and clinopyroxene 00:09:56.830 --> 00:09:59.900 in the upper mantle. And pyroxene goes to garnet 00:09:59.900 --> 00:10:05.340 with a gradual velocity increase, and that can be shown here. 00:10:05.340 --> 00:10:10.140 I’m going to show a close-up of this in just a second. 00:10:10.150 --> 00:10:14.170 On the – what we’re looking at is the minerology of the mantle. 00:10:14.170 --> 00:10:22.450 The upper scale is the modal proportion going to, of course, 100%. 00:10:22.450 --> 00:10:26.830 As a function of depth, we have pressure and depth and temperature 00:10:26.830 --> 00:10:28.730 in the lower scale. So let’s look at the – 00:10:28.730 --> 00:10:34.020 first at the transition zone – the properties of the transition zone. 00:10:34.020 --> 00:10:39.380 So at 410, olivine transforms to wadsleyite, 00:10:39.380 --> 00:10:43.480 and at about 520, to ringwoodite. 00:10:43.480 --> 00:10:50.500 And the pyroxene increasingly is transformed into garnet, 00:10:50.500 --> 00:10:55.730 which is a higher velocity as well. But the sharp – the sharp discontinuities 00:10:55.730 --> 00:11:00.380 are at 410 and 660, and sometimes people can see 00:11:00.380 --> 00:11:03.850 the intermediate discontinuity in between. 00:11:03.850 --> 00:11:06.660 Now, what is the Lehmann Discontinuity? 00:11:06.660 --> 00:11:11.120 Well, the Lehmann Discontinuity correlates with the disappearance 00:11:11.120 --> 00:11:16.920 of OPX and the increasing amounts of garnet. 00:11:17.880 --> 00:11:25.279 Well, exploring the upper mantle took a big step forward when 00:11:25.279 --> 00:11:31.370 Adam Dziewonski and John Woodhouse produced the first global seismic surface 00:11:31.370 --> 00:11:37.360 wave models doing a 3D inversion for the structure of the Earth. 00:11:37.360 --> 00:11:46.520 And their first image of the Earth is shown here, which I’ll describe now. 00:11:46.520 --> 00:11:50.920 So we’re centered on the Pacific Ocean. The big blue in the Pacific Ocean – 00:11:50.920 --> 00:11:54.520 I’ll see if I can get this cursor – yeah, there we are. 00:11:54.520 --> 00:11:58.870 This big blue is old Pacific crust, which has – is about – 00:11:58.870 --> 00:12:03.560 more than 100 million years old. It’s old and cold, and so it has 00:12:03.560 --> 00:12:07.600 a high S wave velocity. This is S wave velocity. 00:12:07.600 --> 00:12:12.260 Blue is high S wave velocity. Red is low. 00:12:12.260 --> 00:12:15.860 East Pacific Rise is quite clear. But the pattern that I would like to 00:12:15.860 --> 00:12:21.640 draw your attention to for the purposes of today’s talk are that the continents 00:12:21.640 --> 00:12:28.660 are underlain by high S wave velocities under the cratonic regions, such as 00:12:28.660 --> 00:12:34.210 western – central and western Australia, much of the – of Africa, which is largely 00:12:34.210 --> 00:12:41.210 Precambrian, the Baltic and Siberian Shield, and Canada and North America. 00:12:41.210 --> 00:12:44.520 Now, eastern China is one – a really fascinating region 00:12:44.520 --> 00:12:48.770 because it’s Precambrian – it’s Archean and Proterozoic in age, 00:12:48.770 --> 00:12:52.800 but it lacks a lithospheric root. And that has been a big topic 00:12:52.800 --> 00:12:58.520 of research, which, again, is not one that I have time to go into today. 00:12:58.520 --> 00:13:03.440 So this is the first – this is the first view of the – of the Earth’s mantle 00:13:03.440 --> 00:13:08.180 produced on a 3D scale in 1984. Since that time, there have 00:13:08.180 --> 00:13:13.420 been many, many models. I like this S wave model from Steve Grand. 00:13:13.430 --> 00:13:18.120 And we’re focused today on North America, so I’ll kind of emphasize this. 00:13:18.120 --> 00:13:23.940 At 100-kilometer depth, the Archean Superior Province in the neighboring 00:13:23.940 --> 00:13:28.160 Proterozoic provinces have high S wave velocity. 00:13:28.160 --> 00:13:34.240 Where we’re sitting today is really very anomalous – extremely thin lithospheric 00:13:34.240 --> 00:13:38.800 thickness – less than 100 kilometers. We’re already out of the lithospheric – 00:13:38.800 --> 00:13:42.850 the root of the continent in the western United States. 00:13:42.850 --> 00:13:46.740 And the asthenosphere is really very shallow here. 00:13:46.740 --> 00:13:50.560 And I’ll show some images of the – of the lithosphere-asthenosphere 00:13:50.560 --> 00:13:55.700 boundary in the western United States. Otherwise, the continents are very well 00:13:55.700 --> 00:14:02.500 demarked – West Africa, the Congo craton, the Amazon craton, the Baltic 00:14:02.500 --> 00:14:10.680 Shield, western Australia – all of these show up as thick lithospheric roots. 00:14:11.430 --> 00:14:16.660 I mentioned that the exception is eastern China. 00:14:16.660 --> 00:14:21.380 The Sino-Korean craton is not underlain by a lithospheric root. 00:14:21.380 --> 00:14:28.500 So presumably, this has delaminated – the blue, cold, rigid and high S wave 00:14:28.500 --> 00:14:32.300 velocity material has delaminated beneath eastern China. 00:14:32.300 --> 00:14:38.360 If we go down to 200 kilometers, then we begin to focus in on the – on the 00:14:38.360 --> 00:14:44.490 truly Archean provinces – the most ancient portions of the continents. 00:14:44.490 --> 00:14:49.620 West Africa is very pronounced. Superior Province, the Slave Province, 00:14:49.620 --> 00:14:52.880 the Nain Province in Greenland. Of course, the Baltic Shield – 00:14:52.880 --> 00:14:57.250 very pronounced. And, in the western United States, 00:14:57.250 --> 00:15:03.240 we have no evidence for a lithospheric root, and therefore 00:15:03.240 --> 00:15:08.040 the lithosphere-asthenosphere boundary is – as I said earlier, is quite shallow. 00:15:08.040 --> 00:15:12.650 If we go to 250 kilometers, then pretty much all bets are off. 00:15:12.650 --> 00:15:16.980 You’re basically just at the bottom of the – of the lithospheric root 00:15:16.980 --> 00:15:22.240 that underlies continents. There is, of course, some trace of it 00:15:22.240 --> 00:15:27.670 under West Africa and the Congo. But in general, you’re moving out of 00:15:27.670 --> 00:15:32.280 the lithosphere – the part that moves with the plates as part of plate tectonics. 00:15:32.280 --> 00:15:36.320 Moving out of the lithosphere into the convecting mantle, and you can 00:15:36.340 --> 00:15:41.420 see the very high temperatures under the oceanic regime. 00:15:42.900 --> 00:15:48.820 So lithosphere-asthenosphere boundary is very thick under Archean cratons, 00:15:48.820 --> 00:15:52.680 and it varies a lot in the – in other regions. 00:15:52.680 --> 00:15:57.050 So, as I mentioned, under tectonic North America, the lithospheric root 00:15:57.050 --> 00:16:00.420 is very, very thin. You have low velocities right – 00:16:00.420 --> 00:16:04.670 almost right beneath the crust. And the stable North America – 00:16:04.670 --> 00:16:11.420 like in Illinois, the lithosphere is about 160 kilometers thick, 00:16:11.420 --> 00:16:13.870 and then you go into a low-velocity zone. 00:16:13.870 --> 00:16:17.560 And that is the asthenosphere under Illinois, for example, and then 00:16:17.560 --> 00:16:22.029 the Lehmann Discontinuity comes in, and you get a velocity increase. 00:16:22.029 --> 00:16:26.080 At the Atlantic, in New York and New Jersey, it’s intermediate. 00:16:26.080 --> 00:16:29.650 You have a lithospheric thickness of about 100 kilometers. 00:16:29.650 --> 00:16:33.330 And then a pronounced low-velocity zone below it. 00:16:33.330 --> 00:16:38.870 As I said, all of this has something to do with the occurrence of earthquakes 00:16:38.870 --> 00:16:47.860 as well, and I wrote about that in this paper in EPSL where, what we found 00:16:47.860 --> 00:16:53.310 in our study of the correlation of the lithospheric structure and earthquakes 00:16:53.310 --> 00:16:59.440 was that there were three primary regions where intra-plate earthquakes 00:16:59.440 --> 00:17:04.340 occurred – that is, not plate boundary earthquakes like the San Andreas, but 00:17:04.340 --> 00:17:09.180 for the central and eastern United States. And those were at the rifted margin, 00:17:09.180 --> 00:17:15.700 where the continents have drifted apart. Interior rifts, like New Madrid. 00:17:15.709 --> 00:17:19.329 And then we found an interesting phenomenon that there is increased 00:17:19.329 --> 00:17:23.269 seismicity between the Archean craton – the boundary between the 00:17:23.269 --> 00:17:25.919 Archeaon craton and the Proterozoic Mobile Belt, 00:17:25.919 --> 00:17:31.220 where you get stress concentration at this – at this suture. 00:17:31.220 --> 00:17:36.629 This cross-section is basically from the Archeon Superior Province 00:17:36.629 --> 00:17:41.659 to the Atlantic coast – this area here. And today, I’ll talk more about 00:17:41.659 --> 00:17:47.460 the structure of this – of this region from seismology. 00:17:47.460 --> 00:17:54.120 So this cartoon depicts the lithosphere-asthenosphere boundary – 00:17:54.120 --> 00:18:00.300 LAB – as varying from about 240 kilometers in the Archean 00:18:00.309 --> 00:18:05.320 of Minnesota coming to about 140 in the central U.S., and then becoming 00:18:05.320 --> 00:18:11.799 quite thin – only 80 kilometers – right at the – right at the coastline. 00:18:11.799 --> 00:18:18.119 Well, now we can test these ideas because the Transportable Array has 00:18:18.120 --> 00:18:22.980 marched its way across North America. Now it’s working up in Alaska. 00:18:22.980 --> 00:18:29.800 And there are lots and lots of networks. So there – shown in yellow of 00:18:29.809 --> 00:18:35.149 different kinds. So there are thousands and thousands of seismogenic stations 00:18:35.149 --> 00:18:39.450 that have recorded teleseismic earthquakes, and you can play all 00:18:39.450 --> 00:18:43.119 kinds of games, like ambient noise, teleseismic tomography, 00:18:43.119 --> 00:18:49.169 and other things. So I’d like to acknowledge the data. 00:18:49.169 --> 00:18:53.999 If we can fly, like these imaginary creatures here, as scientists, 00:18:53.999 --> 00:18:56.259 it’s because we have – we have good data. 00:18:56.259 --> 00:19:03.659 And the – it’s going to take me five slides to show you the sources of the – 00:19:03.659 --> 00:19:08.580 of the data that we’ve processed to get the images that I’m about to show you. 00:19:08.580 --> 00:19:13.630 I will not read all the networks, but you can see Oklahoma Seismic Network, 00:19:13.630 --> 00:19:21.239 Miami University Seismic – there are a huge amount of networks in addition to 00:19:21.239 --> 00:19:27.240 the Transportable Array, all of which contribute data that can be processed. 00:19:27.240 --> 00:19:32.080 Now, if you want to find boundaries inside the Earth, a good way of doing 00:19:32.080 --> 00:19:36.950 that is to look for converted phases. That’s where a P wave hits a boundary, 00:19:36.950 --> 00:19:41.460 like the Moho or the LAB – a P wave hits the boundary, and as you know from 00:19:41.460 --> 00:19:46.309 physics, you get some S wave energy – some transmitted S wave energy and 00:19:46.309 --> 00:19:50.970 some transmitted P wave energy. And these are called P wave receiver 00:19:50.970 --> 00:19:55.840 functions. So looking at this particular example, here’s a boundary. 00:19:55.840 --> 00:19:58.840 It might be the Moho. I could be another boundary. 00:19:58.840 --> 00:20:05.159 A teleseismic P wave comes in, and it’s refracted upward and 00:20:05.160 --> 00:20:10.280 comes in as the primary arrival. But there’s also a later arrival, which is 00:20:10.280 --> 00:20:15.980 a P-to-S conversion, and it’s shown here. And you can imagine, if I have – 00:20:15.980 --> 00:20:21.200 if I have many teleseisms and a lot of seismic stations, and if I just stack the 00:20:21.200 --> 00:20:27.619 data – keep stacking the recordings, these S-to-P conversions – 00:20:27.619 --> 00:20:32.340 these P-to-S conversions are going to become really quite strong. 00:20:32.340 --> 00:20:36.909 And you’ll get quite a good signal. You can see, in this particular signal, 00:20:36.909 --> 00:20:43.380 there’s also another boundary above that’s even closer to the P wave. 00:20:43.380 --> 00:20:49.400 So, for a P-to-S receiver function, the first arrival, of course, is the P wave. 00:20:49.400 --> 00:20:55.100 And S waves, which travel slower, come in as a – as secondary arrivals. 00:20:55.100 --> 00:21:00.340 Now, Alex Blanchette is sitting here somewhere, and he was looking at the 00:21:00.340 --> 00:21:06.759 area in western Saudi Arabia, where there’s a nice network, 00:21:06.759 --> 00:21:12.759 and taking all of these – a couple thousand teleseismic earthquakes. 00:21:12.759 --> 00:21:18.539 You can sum everything up at various stations and make a – like, a data cube. 00:21:18.539 --> 00:21:22.389 And you get the following image. Well, Alex got the following image. 00:21:22.389 --> 00:21:25.499 Now, this is a little bit hard to look at because your eyes are going to 00:21:25.499 --> 00:21:28.879 shift a little bit, but let me just walk you through it. 00:21:28.879 --> 00:21:33.380 So red means the velocity increases downward. 00:21:33.380 --> 00:21:39.000 So you – the velocity increase is determined by the polarity. 00:21:39.000 --> 00:21:42.399 And that’s the Moho. You go from crust to mantle, 00:21:42.399 --> 00:21:45.529 and there’s a velocity increase. And you can see, on this side of 00:21:45.529 --> 00:21:50.059 the cube, you have a positive. On this side, likewise. 00:21:50.059 --> 00:21:54.970 And if you take a slice at 38 kilometers, you get almost all red. 00:21:54.970 --> 00:22:00.380 So we’re very confident that we know where the crust-mantle boundary is. 00:22:00.380 --> 00:22:07.220 And this is an unusual area because the lithosphere is very thin because the 00:22:07.220 --> 00:22:12.259 Red Sea has rifted open the continent, and you get an LAB at a depth of only 00:22:12.259 --> 00:22:18.559 60 kilometers. So blue is a low velocity. So, in this particular case, you can see 00:22:18.559 --> 00:22:22.230 both the mantle and the LAB with the P waves. 00:22:22.230 --> 00:22:25.409 So that’s called a P wave receiver function. 00:22:25.409 --> 00:22:29.820 Now, the unfortunate thing about this technique is that you 00:22:29.820 --> 00:22:34.769 get a lot of multiples that are very annoying that are mixed in 00:22:34.769 --> 00:22:37.259 with the good signal that you want to get. 00:22:37.259 --> 00:22:41.590 You have scattering and multiple reflections after the P wave. 00:22:41.590 --> 00:22:46.940 So people like to use S wave receiver functions. 00:22:46.940 --> 00:22:51.090 That would be where the primary S wave goes – hits a boundary, 00:22:51.090 --> 00:22:56.369 and some of the S wave energy converts to a P wave, which comes in early. 00:22:56.369 --> 00:23:01.840 So the beauty of S-to-P wave receiver functions is that you’re looking for 00:23:01.840 --> 00:23:07.859 a precursor. And the precursor comes in the quiet part of your seismogram. 00:23:07.859 --> 00:23:12.830 So I need to explain a couple of things from some of the people 00:23:12.830 --> 00:23:18.350 who do processing of data. And this is shown here. 00:23:18.350 --> 00:23:21.480 What we are doing is taking the S wave – the S wave 00:23:21.480 --> 00:23:26.129 comes to a seismographic station. And instead of looking at the data 00:23:26.129 --> 00:23:32.769 from the point of view of the radial and transverse, we do a location of 00:23:32.769 --> 00:23:37.110 the coordinate system to the – what’s called the L and Q coordinate system, 00:23:37.110 --> 00:23:43.019 where L is along the S wave ray path and Q is perpendicular. 00:23:43.019 --> 00:23:47.830 So all the S wave – the strong S wave vibration is concentrated 00:23:47.830 --> 00:23:53.710 on the Q component. And if you plot the L component, you won’t even see 00:23:53.710 --> 00:23:59.240 the S wave. All you’ll see is the P wave. That’s this guy over here. 00:23:59.240 --> 00:24:06.150 So the S wave also comes in, converts at any interface to a S-to-P conversion, 00:24:06.150 --> 00:24:09.429 and that shows up really nicely on the L component. 00:24:09.429 --> 00:24:15.940 So instead of using the usual radial-transverse coordinate rotation. 00:24:15.940 --> 00:24:18.020 So this is what you get. 00:24:18.020 --> 00:24:21.840 Okay, it’s not very complicated. You have an S wave that comes in. 00:24:21.840 --> 00:24:24.889 It hits an interface and converts to a P wave. 00:24:24.889 --> 00:24:30.990 And at the Moho, it produces a pulse. And at the LAB, it produces a pulse. 00:24:30.990 --> 00:24:34.200 And there is no S wave here because I’ve rotated my 00:24:34.200 --> 00:24:37.960 coordinate system so there’s no S wave vibration. 00:24:37.960 --> 00:24:42.779 Now, in order to make things look the way we think as human beings, 00:24:42.779 --> 00:24:46.739 what you do is you flip the trace. You reverse the trace. 00:24:46.739 --> 00:24:52.979 And so now this trace has been reversed in time so that the Moho is late and the 00:24:52.979 --> 00:24:58.990 LAB is late, and also we flip it in polarity. So now we have S waves and P waves 00:24:58.990 --> 00:25:03.330 that look exactly the same, and we feel comfortable. 00:25:03.330 --> 00:25:10.960 We can see that the trace – the Moho follows the first arrival, 00:25:10.960 --> 00:25:15.570 and the LAB follows even later. So here’s the – here is our S wave 00:25:15.570 --> 00:25:19.379 receiver function showing the Moho and the LAB. 00:25:19.379 --> 00:25:22.149 And these signals are precursors to the S wave. 00:25:22.149 --> 00:25:25.769 So they’re coming in in a very quiet part of the seismogram. 00:25:25.769 --> 00:25:31.240 And seismologists like quiet data with high signal-to-noise ratio. 00:25:31.240 --> 00:25:35.220 So this has been done by many people. Here’s a list of papers that have been 00:25:35.220 --> 00:25:40.700 written using this technique. And what I’m going to show today 00:25:40.700 --> 00:25:45.139 is that we believe that this method can be improved 00:25:45.140 --> 00:25:47.900 beyond what has been done so far. 00:25:49.220 --> 00:25:57.840 So, in 2015, my co-authors, Rainer Kind and others, looked at the structure of the 00:25:57.840 --> 00:26:01.680 upper mantle in the western and central United States with S wave receiver 00:26:01.680 --> 00:26:06.840 functions, and they used the Transportable Array that existed 00:26:06.840 --> 00:26:11.860 at that time – lots and lots of stations in a couple thousand locations. 00:26:11.860 --> 00:26:17.980 And they studied the structure of the upper mantle from the western 00:26:17.980 --> 00:26:24.480 United States into the Archean province of Minnesota and Wisconsin. 00:26:24.480 --> 00:26:30.309 So, as I tried to lay out, the answer is, under the western United States, 00:26:30.309 --> 00:26:34.470 there’s going to be a shallow lithosphere-asthenosphere boundary – 00:26:34.470 --> 00:26:41.500 about 60 to 80. And under the Archean, it should go to about 240 kilometers. 00:26:41.500 --> 00:26:43.660 But that’s not what they found. 00:26:45.400 --> 00:26:47.640 Problems in River City. 00:26:48.660 --> 00:26:51.960 What happened was that, under the Moho – the red is the Moho. 00:26:51.970 --> 00:26:55.980 Under the Moho, there’s a low-velocity zone under the western United States 00:26:55.980 --> 00:27:00.309 just as expected, but under the central United States, there’s also 00:27:00.309 --> 00:27:04.590 these other boundaries here. Blue is a low velocity. 00:27:04.590 --> 00:27:08.480 And if you look on the north-south profile, all through – right under – 00:27:08.480 --> 00:27:13.840 going all the way up to Chicago, there is this low velocity below the 00:27:13.840 --> 00:27:19.169 Moho at a depth of about 80 kilometers. So this is the well-known 00:27:19.169 --> 00:27:23.080 lithosphere-asthenosphere boundary of the western United States. 00:27:23.080 --> 00:27:27.440 But this is an unexpected mid-lithosphere discontinuity, 00:27:27.440 --> 00:27:30.970 which is contradicted by the work of Adam Dziewonski and 00:27:30.970 --> 00:27:37.240 John Woodhouse and all the – Steve Grand. 00:27:38.049 --> 00:27:43.340 So what we’re finding with this S wave receiver function is there’s 00:27:43.350 --> 00:27:47.269 plenty of evidence for a mid-lithosphere discontinuity, 00:27:47.269 --> 00:27:53.149 but we can’t find the doggone LAB. So it’s like many things in science. 00:27:53.149 --> 00:27:57.169 As soon as you try to do something, you don’t get the result you expect. 00:27:57.169 --> 00:28:05.020 And so this led to a big contradiction that surface waves were telling us that 00:28:05.020 --> 00:28:08.940 continents have roots that are about 200 kilometers thick. 00:28:08.940 --> 00:28:13.549 But the S-to-P receiver functions are saying, no, no way. 00:28:13.549 --> 00:28:16.440 They’re – continents are thin. And Rainer Kind published 00:28:16.440 --> 00:28:19.320 a paper showing the lithosphere- asthenosphere boundary at 00:28:19.320 --> 00:28:24.720 80 kilometers for all of North America. He’s my co-author, and when 00:28:24.720 --> 00:28:29.729 I’m implying that he did something wrong, he’s seen the light, 00:28:29.729 --> 00:28:33.600 and he’s – we have a better idea today. 00:28:33.600 --> 00:28:38.340 Okay. Let’s just take a standard example of one of those 50 papers 00:28:38.340 --> 00:28:41.529 that I – I showed you previous work. There’s about 50 papers. 00:28:41.529 --> 00:28:46.759 This is just a recent one talking about S-to-P converted phases 00:28:46.759 --> 00:28:49.470 in the contiguous United States. 00:28:49.470 --> 00:28:54.350 So they are using an extended-time multitaper cross-correlation method 00:28:54.350 --> 00:29:00.720 to get the receiver functions – let’s see – to deconvolve. 00:29:00.720 --> 00:29:04.989 They’re using deconvolution to get the receiver functions. 00:29:04.989 --> 00:29:08.379 And I’m going to show you, then, a couple of cross-sections 00:29:08.379 --> 00:29:11.549 across here from these data. And here’s what you get. 00:29:11.549 --> 00:29:15.419 Okay, so this is a very standard publication. 00:29:15.419 --> 00:29:21.489 This one is in GQ, but you can find them everywhere, that shows the Moho going 00:29:21.489 --> 00:29:26.649 from the western United States – Moho is in red. Goes all the way across. 00:29:26.649 --> 00:29:32.280 And right below the Moho is this blue mid-lithosphere discontinuity – MLD. 00:29:32.280 --> 00:29:35.520 Here is the real LAB. We know that this is here because 00:29:35.529 --> 00:29:38.639 we have surface waves that show us really, really clearly that this is a very 00:29:38.639 --> 00:29:46.759 strong boundary. But we cannot find – these authors cannot find the LAB. 00:29:46.759 --> 00:29:51.429 So the lithosphere-asthenosphere – the base of the continents is invisible, 00:29:51.429 --> 00:29:55.849 but something inside the continents just below the Moho is very clear. 00:29:55.849 --> 00:30:00.399 Likewise on the lower section. 00:30:00.399 --> 00:30:06.440 So that means that this boundary here is not imaged, 00:30:06.440 --> 00:30:10.980 and a boundary in the middle is imaged. 00:30:12.540 --> 00:30:16.980 So the problem seems to be the use of deconvolution. 00:30:16.980 --> 00:30:24.279 People are processing data and using a very strong processing technique. 00:30:24.279 --> 00:30:27.860 When an earthquake occurs, it has a – it has a rupture history. 00:30:27.860 --> 00:30:35.679 It has – it has a signature – a source signature, which has multiple pulses. 00:30:35.679 --> 00:30:41.409 And if you take the recorded signal and deconvolve the source 00:30:41.409 --> 00:30:47.559 signal from the recorded signal, you hope to get simple spikes. 00:30:47.559 --> 00:30:52.580 You get a cross-correlation every time you hit a signal that came from the – 00:30:52.580 --> 00:30:55.950 from the earthquake. However, this comes at a cost. 00:30:55.950 --> 00:31:01.230 Because deconvolution has side lobes. And it also – the actual wave that 00:31:01.230 --> 00:31:06.420 comes to the recording station undergoes the phase transformations. 00:31:06.420 --> 00:31:11.389 It’s a very complicated story. So what we suggest is that, 00:31:11.389 --> 00:31:15.659 instead of doing a lot of processing to the data, now that we have the 00:31:15.659 --> 00:31:20.169 U.S. array and thousands of seismic stations and many earthquakes, 00:31:20.169 --> 00:31:25.620 why don’t we leave the data alone? Just stack the pure data. 00:31:25.620 --> 00:31:30.450 Don’t do a lot of processing. So that means that we’re suggesting 00:31:30.450 --> 00:31:35.549 that you take a signal like this, and instead of deconvolving the source from 00:31:35.549 --> 00:31:42.840 the Moho and deconvolving the source from this weak LAB signal, just stack it. 00:31:42.840 --> 00:31:45.679 Now, there are a couple of tricks that you have to play. 00:31:45.679 --> 00:31:51.729 One is, if the earthquake source has a strong maximum that’s negative, 00:31:51.729 --> 00:31:55.999 we flip that one and make it – make the strong signal positive. 00:31:55.999 --> 00:31:59.450 And if the earthquake source has a strong positive, we leave it. 00:31:59.450 --> 00:32:05.239 So we do a normalization by flipping to get all of the strong signals lined up. 00:32:05.239 --> 00:32:08.720 Otherwise, you know, you’re adding signals together, 00:32:08.720 --> 00:32:12.369 and you have destructive interference. 00:32:12.369 --> 00:32:17.369 So all we’re suggesting is don’t mess with your data too much. 00:32:17.369 --> 00:32:20.619 Just keep the processing simple. 00:32:20.619 --> 00:32:26.779 And this is the original processing that was suggested by Lev Vinnik 00:32:26.779 --> 00:32:30.580 in Moscow when the – when the S-to-P receiver function was 00:32:30.580 --> 00:32:35.519 first developed by these people. At that time, Lev Vinnik didn’t 00:32:35.519 --> 00:32:38.809 have a lot of big supercomputers. So the easiest thing for him 00:32:38.809 --> 00:32:42.749 to do was just stack the data. But then Chuck Langston came along, 00:32:42.749 --> 00:32:46.090 and Chuck said, well, we can do a better job by deconvolving. 00:32:46.090 --> 00:32:51.090 And it went downhill from there. Too much processing. Okay. 00:32:51.090 --> 00:32:55.830 So if you take Region A over here and look at some of the stations, 00:32:55.830 --> 00:33:00.649 here’s a comparison of Region A. Here’s the S wave coming in. 00:33:00.649 --> 00:33:04.440 You line up – there are three traces here – red, blue, and black. 00:33:04.440 --> 00:33:08.399 Black is deconvolved. Whoop. Black is deconvolved. 00:33:08.399 --> 00:33:12.759 And the red and blue are just stack – simple stacking, either on the onset 00:33:12.760 --> 00:33:17.680 of Sv or on the maximum. Now I’m going to blow this up in the next ... 00:33:18.129 --> 00:33:21.960 So you see what happens is, when you do the deconvolution, 00:33:21.960 --> 00:33:25.499 you get this pulse after the Moho. 00:33:25.500 --> 00:33:31.100 That’s the side lobe. And that pulse is the MLD. 00:33:32.929 --> 00:33:37.999 So the feature that is being analyzed is actually something which is 00:33:37.999 --> 00:33:41.809 the product of deconvolution. If you get rid of the deconvolution, 00:33:41.809 --> 00:33:50.359 you have the red and blue traces where this big, strong pulse is missing. 00:33:50.359 --> 00:33:57.529 So what I’m saying is that the phase which is following just below the 00:33:57.529 --> 00:34:02.710 Moho is a side lobe to the Moho that’s introduced by deconvolution. 00:34:02.710 --> 00:34:07.309 And when we take the – when we take the data from the U.S. array 00:34:07.309 --> 00:34:12.200 and process it the same way using deconvolution, this is what we get. 00:34:12.200 --> 00:34:16.309 So you get this strong scattering, and you can’t – you can’t see anything – 00:34:16.309 --> 00:34:19.840 this is the same depth scale down to 300 kilometers. 00:34:19.840 --> 00:34:21.660 But if you get rid of deconvolution, 00:34:21.660 --> 00:34:25.840 and you do a simple stack, here is the comparison. 00:34:25.850 --> 00:34:30.860 Okay, this is North America. This is the Mid-Continent Rift. 00:34:30.860 --> 00:34:35.950 This is the Great Lakes. Nova Scotia. Baja, California. 00:34:35.950 --> 00:34:39.339 And we’re summing the seismic traces between the black lines – 00:34:39.340 --> 00:34:44.560 the two black lines over about 10 degrees. This is no deconvolution. 00:34:44.560 --> 00:34:47.940 Here is the LAB under the western United States. 00:34:47.940 --> 00:34:50.400 Here is something under the Wyoming Province. 00:34:50.419 --> 00:34:52.799 And there’s a dipping boundary that goes between 00:34:52.799 --> 00:34:55.909 the Proterozoic and the Archean. 00:34:55.909 --> 00:35:01.890 This is the low – red is the low velocity. This is the LAB. 00:35:01.890 --> 00:35:06.410 This is the Lehmann Discontinuity. This is the 410, which is a positive step. 00:35:06.410 --> 00:35:09.720 And, above the 410, there’s a low-velocity zone. 00:35:09.720 --> 00:35:14.100 If we apply seismic deconvolution to the same traces – 00:35:14.100 --> 00:35:18.100 these are the exact same traces – you get this. 00:35:19.160 --> 00:35:23.340 And that then highlights mainly what’s called the MLD. 00:35:23.340 --> 00:35:29.240 And you don’t get the – you no longer can see the LAB. 00:35:29.710 --> 00:35:38.640 So my co-authors, Rainer Kind and Yuan Xiaohui, have themselves 00:35:38.650 --> 00:35:43.230 published papers with deconvolution, and they’re saying, we’d rather 00:35:43.230 --> 00:35:48.490 go back to pure stacking of the – of the traces and get rid of 00:35:48.490 --> 00:35:51.650 this kind of data processing. 00:35:51.650 --> 00:35:55.090 So I say, good riddance, mid-lithosphere discontinuity. 00:35:55.090 --> 00:36:00.500 It’s been nice having you for 20 years, but goodbye. 00:36:00.500 --> 00:36:06.660 So this is what the traces look like. This is what the image looks like 00:36:06.660 --> 00:36:11.980 if you just stack the data. Now, I admit that it’s pretty 00:36:11.980 --> 00:36:15.180 low-resolution because we’re stacking over, you know, 00:36:15.180 --> 00:36:21.000 10 or 15 degrees between these two lines to get a signal-to-noise 00:36:21.000 --> 00:36:24.589 ratio that’s large enough. And maybe if someone’s even more 00:36:24.589 --> 00:36:30.020 clever, they can start finding ways of getting really good images from S-to-P 00:36:30.020 --> 00:36:36.000 that are, you know, very narrow and more closely aligned with geology. 00:36:36.000 --> 00:36:39.060 Well, what we think is going on is pretty exciting. 00:36:39.070 --> 00:36:43.109 So this boundary here, which is – which we see going from the 00:36:43.109 --> 00:36:49.930 Proterozoic Yavapai/Mazatzal into the Archean Province, this is, 00:36:49.930 --> 00:36:55.880 for the first time ever seen, is the record of the docking 00:36:55.880 --> 00:37:00.619 of the Proterozoic continents against the Archean, as shown here, 00:37:00.619 --> 00:37:02.980 and we’re seeing a dipping boundary. 00:37:02.980 --> 00:37:07.020 This dipping boundary here is this one here. 00:37:07.020 --> 00:37:13.640 So, yes, the lithosphere is complicated. The lithosphere does have internal 00:37:13.640 --> 00:37:17.700 discontinuities. It does have a geology, and it does have a history. 00:37:17.700 --> 00:37:21.510 It just doesn’t have that phase, which is right below the Moho, 00:37:21.510 --> 00:37:28.329 which is a, you know, flat MLD. Instead, what it has is a tectonic history. 00:37:28.329 --> 00:37:33.310 And the tectonic history is the underplating of the oceanic 00:37:33.310 --> 00:37:38.710 lithosphere beneath the Archean core. So the Archean core is here. 00:37:38.710 --> 00:37:43.079 And this feature here is the Proterozoic lithosphere. 00:37:43.079 --> 00:37:46.250 In the south – now, I really need to orient you here because 00:37:46.250 --> 00:37:49.849 this is pretty hard to find. This is Texas here. 00:37:49.849 --> 00:37:53.670 This is Louisiana. You can see the delta. 00:37:53.670 --> 00:37:56.960 And we’re going from the Gulf Coast to the north. 00:37:56.960 --> 00:38:01.450 North is on your right. South is on your left. 00:38:01.450 --> 00:38:03.339 And what we see is this dipping boundary. 00:38:03.339 --> 00:38:06.789 We have the Moho above here, and you have this dipping boundary. 00:38:06.789 --> 00:38:12.539 Well, this is very exciting because this is, as predicted – this little piece 00:38:12.539 --> 00:38:15.369 of material here between this dipping boundary and the Moho, 00:38:15.369 --> 00:38:19.980 this is a piece of Africa that was left behind. 00:38:19.980 --> 00:38:21.900 When you made Pangea a supercontinent, 00:38:21.900 --> 00:38:26.730 you brought Gondwana together with Laurentia in a suture. 00:38:26.730 --> 00:38:30.160 And then later, that suture was rifted – was itself rifted. 00:38:30.160 --> 00:38:36.430 So what we’re talking about is the history of Laurentia and Gondwana. 00:38:36.430 --> 00:38:39.550 They collide at a suture zone here. 00:38:39.550 --> 00:38:43.510 And then, of course, we open up the Atlantic in the Triassic 00:38:43.510 --> 00:38:48.970 at about 160 million years ago. But the opening of the Atlantic is 00:38:48.970 --> 00:38:53.420 not exactly at the suture between Laurentia and Gondwana. 00:38:53.420 --> 00:38:58.731 And what it leaves behind is this – Africa and North America 00:38:58.731 --> 00:39:03.500 come together, and you rift it under the letter F or R. 00:39:03.500 --> 00:39:07.890 You rift it here, and what you see behind – leaving behind is the 00:39:07.890 --> 00:39:11.270 upper boundary, which is the Moho, and the dipping boundary, 00:39:11.270 --> 00:39:18.220 which is the suture zone. So what we see in the – under the Gulf Coast is we see 00:39:18.220 --> 00:39:25.380 a fragment of Africa that was left behind when it – when the rifting took place. 00:39:25.380 --> 00:39:31.049 Now, another interesting feature, to put it mildly, is this 00:39:31.049 --> 00:39:37.859 low-velocity layer above the 410. The 410 discontinuity is that first 00:39:37.859 --> 00:39:41.609 major phase boundary where you go to wadsleyite. 00:39:41.609 --> 00:39:44.770 And we’re looking deep into the upper mantle. 00:39:44.770 --> 00:39:48.430 And this is a really prominent, prominent phase here, 00:39:48.430 --> 00:39:54.549 which shows up sitting on top of the 410 – a low-velocity boundary. 00:39:54.549 --> 00:40:00.900 Here’s a cross-section that’s from Baja, California, up to the north. 00:40:00.900 --> 00:40:04.040 And it shows up – it shows up everywhere. 00:40:04.040 --> 00:40:05.820 So what is this? 00:40:05.820 --> 00:40:11.359 Well, it turns out that, like most things that you do in science, somebody has 00:40:11.359 --> 00:40:15.880 already done it before and talked about it, and then they were forgotten. 00:40:15.880 --> 00:40:20.360 And then they are remembered when the topic gets hot again. 00:40:20.360 --> 00:40:26.401 And Justin Revenaugh – and a lot of you know Stuart Sipkin from USGS in 00:40:26.401 --> 00:40:32.359 Golden – a 1994 paper in Nature. They took the seismographic stations in 00:40:32.359 --> 00:40:37.920 China that were installed by USGS and the China Earthquake Administration – 00:40:37.920 --> 00:40:42.619 there are 10 stations. And they looked at all of the S wave 00:40:42.619 --> 00:40:48.980 reverberations from earthquakes in the Japan Sea going to those stations. 00:40:48.980 --> 00:40:53.109 And they found a really interesting feature. 00:40:53.109 --> 00:40:57.109 Of course, they found the 410. You’re not going to get a Nature paper 00:40:57.109 --> 00:41:01.789 for discovering the 410. I told you Jeffreys found that in 1936. 00:41:01.789 --> 00:41:04.750 But if you find a low-velocity zone above 410, you can probably 00:41:04.750 --> 00:41:07.349 get a Nature paper. And that’s what they did. 00:41:07.349 --> 00:41:10.460 Here it is there. Here is the low-velocity – 00:41:10.460 --> 00:41:13.670 here’s the low-velocity here as well. 00:41:13.670 --> 00:41:19.890 So they already suggested, in 1994, that there was a low-velocity zone 00:41:19.890 --> 00:41:24.039 above the 410. And people have, since that time, searched for it 00:41:24.039 --> 00:41:29.200 with mixed results. Sometimes they find it, and sometimes they don’t. 00:41:29.200 --> 00:41:33.140 But there’s a guy named Ken Dueker out at the University of Wyoming, and 00:41:33.140 --> 00:41:41.180 he and his student, Jasbinsek, worked on looking at the structure under Wyoming. 00:41:41.180 --> 00:41:44.750 And they did find a low-velocity zone above the 410. 00:41:44.750 --> 00:41:49.410 And they did a lot of work modeling that discontinuity, and they suggest it’s 00:41:49.410 --> 00:41:55.089 about 20 kilometers thick with a velocity reduction of about 8%. 00:41:55.089 --> 00:42:00.279 And it’s pretty well-developed. So this is a poorly understood 00:42:00.279 --> 00:42:05.320 low-velocity feature on top of the 410. Some French scientists have been 00:42:05.320 --> 00:42:13.269 working on this as well. Tauzin – Benoit Tauzin wrote a paper with Eric Debayle. 00:42:13.269 --> 00:42:17.350 This is 410 – they began to say that there’s seismic evidence 00:42:17.350 --> 00:42:22.170 for a global low-velocity layer. And they write, within the upper mantle, 00:42:22.170 --> 00:42:26.029 the seismic discontinuity at 410 kilometers marks the top of 00:42:26.029 --> 00:42:30.289 the transition zone and is attributed to a phase-induced transformation 00:42:30.289 --> 00:42:35.660 of olivine to wadsleyite. Just above the 410, a layer characterized 00:42:35.660 --> 00:42:41.080 by low wave velocities has been identified regionally, they go on to say. 00:42:41.080 --> 00:42:44.280 What we’re – what we’re finding is we find it everywhere under 00:42:44.280 --> 00:42:49.400 North America. So we think that the reason it’s been regional is that it’s just 00:42:49.400 --> 00:42:58.320 been hard to identify because it’s deep and you need really good data to see it. 00:42:58.320 --> 00:43:02.650 So what’s going on? Well, I am not, by any stretch of the 00:43:02.650 --> 00:43:08.500 imagination, a expert on the Earth’s – on the transition zone of the mantle. 00:43:08.500 --> 00:43:13.119 You know, there are a lot of really good high-pressure mineral physicists 00:43:13.119 --> 00:43:18.190 who have looked into this question. But I think the general impression is 00:43:18.190 --> 00:43:22.990 that the ringwoodite and wadsleyite, especially ringwoodite, and I think – 00:43:22.990 --> 00:43:28.710 contains water. And this water is introduced by subducting slabs 00:43:28.710 --> 00:43:34.880 that pass through the transition zone. And some of it leaks out and forms 00:43:34.880 --> 00:43:40.859 a layer just above the 410 that can produce this low-velocity layer. 00:43:40.859 --> 00:43:48.089 So Bercovici and Karato at Yale University hinted at this early on – 00:43:48.089 --> 00:43:54.920 as early as 2003, where they say – they suggested a melt layer above the 410. 00:43:54.920 --> 00:43:59.349 It could also be just a hydrated layer. 00:43:59.349 --> 00:44:07.240 But the water is being introduced by this subduction process. 00:44:07.240 --> 00:44:10.380 Here it is again. Here is another couple of views of 00:44:10.380 --> 00:44:16.109 this low-velocity layer above the 410. And it’s really very, very prominent. 00:44:16.109 --> 00:44:20.740 And this has been an opportunity to really get into the physics 00:44:20.740 --> 00:44:26.660 of the transition zone. So let me conclude. 00:44:26.660 --> 00:44:30.079 What we are finding when we just simply stack the data rather 00:44:30.079 --> 00:44:36.040 than use heavy processing, filtering, and deconvolution, is the lithosphere- 00:44:36.040 --> 00:44:40.630 asthenosphere boundary is consistently visible, but it’s weak. 00:44:40.630 --> 00:44:48.099 So I would modify my original figure from 2012 and show the LAB as being 00:44:48.099 --> 00:44:53.769 this kind of transition zone rather than a sharp boundary as I showed it – 00:44:53.769 --> 00:44:58.400 showed it before. We confirmed the existence of the 410. 00:44:58.400 --> 00:45:05.680 Well, of course, that’s no discovery, as I said, but the 410, then, is below this. 00:45:05.680 --> 00:45:10.560 A strong low-velocity zone is found above the 410. 00:45:10.560 --> 00:45:15.420 It may be due to dewatering. And we depict this here. 00:45:15.420 --> 00:45:21.820 So we have a new view of the upper mantle that includes this kind of detail. 00:45:21.820 --> 00:45:24.820 And we image the plate tectonic evolution of the North American 00:45:24.820 --> 00:45:30.540 continent, including paleo plates that have been accreted to the bottom. 00:45:30.540 --> 00:45:34.880 And, as I mentioned here, previously identified elsewhere 00:45:34.880 --> 00:45:38.650 in seismic reflection data. I’m referring to the paper by 00:45:38.650 --> 00:45:43.349 Andy Calvert, where he saw a dipping plate in the mantle under 00:45:43.349 --> 00:45:49.440 the Abitibi Province and a paper by Fred Cook in the Slave Province 00:45:49.440 --> 00:45:56.400 in northern Canada where they also see low-angle dipping paleo plates. 00:45:56.400 --> 00:46:00.480 So the way the upper – the way the whole picture really looks would 00:46:00.480 --> 00:46:06.630 be to have this Proterozoic terrane underplating the Archean craton. 00:46:06.630 --> 00:46:12.010 And finally, we also see the Lehmann Discontinuity at 00:46:12.010 --> 00:46:18.940 a depth of about 270 kilometers. And this is, again, due to the 00:46:18.940 --> 00:46:22.960 phase change that goes from the pyroxene into garnet. 00:46:22.960 --> 00:46:29.480 So, in conclusion, this is our basic image that we’ve obtained. 00:46:29.480 --> 00:46:34.520 And we’ve gotten this by just simplifying the processing of our data, 00:46:34.530 --> 00:46:39.809 relying on the very large abundance of information, of traces we have, 00:46:39.809 --> 00:46:46.340 so we can – we can avoid introducing artifacts into our data. 00:46:46.340 --> 00:46:53.060 So the paper should be available for you to read in a couple months. 00:46:53.060 --> 00:46:55.720 And I thank you for your attention. 00:46:55.720 --> 00:47:01.320 [Applause] 00:47:01.320 --> 00:47:03.400 Great. Thank you, Walter, for a really interesting talk. 00:47:03.400 --> 00:47:05.820 Do we have any questions? 00:47:07.700 --> 00:47:10.080 I guess part of the problem with Walter giving the talk is there’s no Walter 00:47:10.080 --> 00:47:13.289 in the audience asking the questions. [laughter] 00:47:13.289 --> 00:47:16.200 All right. Here we have some questions. 00:47:18.060 --> 00:47:20.032 - You can rely on Wayne to have something to say. 00:47:20.040 --> 00:47:25.780 - [inaudible] compliment the talk first, so it’s like a Walter question. [laughter] 00:47:25.780 --> 00:47:31.600 I nearly always compliment the talks, and this time … [laughter] 00:47:31.600 --> 00:47:36.609 - You have to pay it back. - Really very provocative and 00:47:36.609 --> 00:47:42.140 interesting, especially the issue about the mid-lithosphere discontinuity. 00:47:43.060 --> 00:47:50.200 In your receiver function stack – maybe it was the previous slide? Yeah. 00:47:50.200 --> 00:47:56.540 Corresponding to distance of, what, 2,000 to 3,000? 00:47:56.550 --> 00:48:03.190 The – all of the – all of the discontinuities are weaker, 00:48:03.190 --> 00:48:09.760 and then there’s nothing much – excuse me – that corresponds to 00:48:09.760 --> 00:48:16.220 what the LAB is in that – in that distance range. What’s going on? 00:48:16.220 --> 00:48:23.789 - Yeah. So we do find that there are sections of our images 00:48:23.789 --> 00:48:31.220 which are less clear than others. I made the pitch that the low-velocity 00:48:31.220 --> 00:48:36.480 zone above 410 is present everywhere, but if you follow along in here, 00:48:36.480 --> 00:48:40.760 it really varies a lot. And, as you just got done pointing out, 00:48:40.760 --> 00:48:44.690 that’s true also for the Lehmann and elsewhere. 00:48:44.690 --> 00:48:52.760 So there is somewhat lower density of seismic – a lot of it depends on the 00:48:52.760 --> 00:48:58.170 additional arrays – those yellow stations that are complementing the central U.S. 00:48:58.170 --> 00:49:04.500 Doesn’t have a lot of extra arrays that are present in the western U.S. 00:49:05.240 --> 00:49:07.840 I don’t know. I don’t think I can give you a satisfactory answer 00:49:07.840 --> 00:49:13.720 for why the image varies regionally. 00:49:13.720 --> 00:49:19.200 - Okay. I’ll ask you another question. Maybe you have an answer for it. 00:49:19.210 --> 00:49:22.779 Many people who have done this receiver function work 00:49:22.779 --> 00:49:31.069 in western United States, say that the amplitude of the precursor that they 00:49:31.069 --> 00:49:40.559 identify as the LAB is so large that there’s something like an 8 or 10% 00:49:40.560 --> 00:49:43.960 velocity decrease at the putative LAB. 00:49:43.960 --> 00:49:50.289 - Right here. - And attribute it to partial melt. 00:49:50.289 --> 00:49:55.210 They don’t think that there’s any other way in which you can produce that. 00:49:55.210 --> 00:50:00.640 In the central and eastern U.S., the LAB boundary, even with 00:50:00.640 --> 00:50:04.440 your improved processing, is a little bit obscure. 00:50:04.440 --> 00:50:09.400 What’s your view about this partial melt issue? 00:50:09.400 --> 00:50:16.000 - Well, based on the scattered incidences of volcanism in the western 00:50:16.000 --> 00:50:20.160 United States, partial melt really sounds quite reasonable. 00:50:20.160 --> 00:50:25.009 And the depth for that kind of – much of the volcanism is basaltic 00:50:25.009 --> 00:50:29.619 and would be consistent with a depth – coming from a depth of about 00:50:29.619 --> 00:50:37.430 60 to 80 kilometers. So I think partial melting is quite reasonable. 00:50:37.430 --> 00:50:40.250 I wouldn’t say a high degree of – it’s not, you know – not like 00:50:40.250 --> 00:50:45.190 a mid-ocean ridge, necessarily, but it’s quite reasonable to have partial melting 00:50:45.190 --> 00:50:49.080 based on the volcanic evidence. - Thanks so much. 00:50:49.900 --> 00:50:52.560 - I think Tom Sisson had a question. 00:50:53.540 --> 00:50:57.360 She’s going to bring you a microphone. - So I have to run across the room. 00:51:03.940 --> 00:51:07.099 - This is probably covered in the literature that I’m not familiar with, 00:51:07.100 --> 00:51:13.280 but would silicate melt on top of the 410 be density-stable? 00:51:14.520 --> 00:51:17.600 - Would silica melt on top of the 410 – I don’t think so. 00:51:17.600 --> 00:51:21.640 I think it would be continually – you would – I would expect it to be 00:51:21.640 --> 00:51:28.630 continuously rising up and having to be, you know, replenished because … 00:51:28.630 --> 00:51:32.579 - I mean, there was this work that – his name escapes me right now – 00:51:32.579 --> 00:51:37.230 who worked with Dave Walker, indicating that, with increasing pressure, 00:51:37.230 --> 00:51:41.430 silicate melts go from positively buoyant to negatively buoyant. 00:51:41.430 --> 00:51:44.790 - Mm-hmm. - So I mean, I was – is it possible 00:51:44.790 --> 00:51:48.140 that the melt just sits there? It’s not being replenished. 00:51:48.140 --> 00:51:52.800 It just ponds on top of the 410. 00:51:54.600 --> 00:51:56.560 Sounds like you don’t know. - Well, I don’t know. 00:51:56.560 --> 00:51:58.380 - Okay. - And it’s constantly being replenished 00:51:58.380 --> 00:52:02.260 because we know that slabs are constantly, you know, 00:52:02.260 --> 00:52:05.300 being re-introduced into the lower – into the upper mantle and 00:52:05.300 --> 00:52:08.900 into the transition zone. So there’s a constant flux. 00:52:08.900 --> 00:52:17.140 So it would seem that it would be hard to keep accumulating it over – 00:52:17.140 --> 00:52:20.180 as long as we’ve had plate tectonics. And I would say we’ve had 00:52:20.180 --> 00:52:23.820 plate tectonics for at least 2 billion years, so … 00:52:25.900 --> 00:52:31.340 [Silence] 00:52:31.640 --> 00:52:35.460 - Getting back to Wayne’s first question, go back to that slide you were 00:52:35.460 --> 00:52:39.280 just showing. What would it take to put uncertainties on your image? 00:52:39.280 --> 00:52:41.720 Because that’s really going to tell you what you are resolving 00:52:41.720 --> 00:52:44.680 and what you may not be resolving. 00:52:45.780 --> 00:52:52.319 - Well, okay, so putting an uncertainty on a converted wave image. 00:52:52.320 --> 00:52:54.940 I wonder how I would do that. 00:52:58.840 --> 00:53:03.349 You can see the uncertainty, to some extent, by the blurriness of the image. 00:53:03.349 --> 00:53:08.059 I mean, if you – what we’re thinking of doing – I guess the best answer 00:53:08.059 --> 00:53:11.829 I can give you is that we’re thinking of starting with forward modeling and 00:53:11.829 --> 00:53:16.490 putting in a hypothetical structure and looking at what we should be seeing if 00:53:16.490 --> 00:53:22.589 everything were – if the – if the Earth really looked like that, what would – 00:53:22.589 --> 00:53:26.859 what would be the recovered image, and see the difference between 00:53:26.859 --> 00:53:32.130 the idealized forward model and the observed. 00:53:32.130 --> 00:53:39.670 But I’m not quite sure how I would calculate the uncertainty in another 00:53:39.670 --> 00:53:44.160 sense. Do you have a suggestion? What did you have in mind for … 00:53:45.280 --> 00:53:47.840 - Well, I mean, it seems like this whole controversy started with, 00:53:47.840 --> 00:53:52.420 how accurate is the data processing you’re doing and whether the 00:53:52.420 --> 00:53:56.599 deconvolution is sort of, you know, introducing these artifacts and 00:53:56.599 --> 00:54:00.420 removing – sort of introducing greater uncertainty as compared to stacking. 00:54:00.420 --> 00:54:05.490 And if you were able to really quantify the errors that you’re introducing or 00:54:05.490 --> 00:54:10.720 be able to sort of say, these are the sort of features that I can – that you 00:54:10.720 --> 00:54:13.569 can actually resolve, it would sort of help resolve the discrepancy 00:54:13.569 --> 00:54:17.720 between the two methods in a more quantitative way. 00:54:17.720 --> 00:54:22.190 - Yeah. So all we’ve done is this kind of quantitative modeling where we’ve 00:54:22.190 --> 00:54:26.940 taken the same traces and processed them either with deconvolution or 00:54:26.940 --> 00:54:35.160 without. And that’s the result you get, which clearly shows – there’s 00:54:35.160 --> 00:54:41.059 something seen here that is not evident in the simple stack. 00:54:41.060 --> 00:54:44.349 And then, likewise, these two. 00:54:46.640 --> 00:54:51.520 Yeah, it would be great if I could think of a way of quantifying 00:54:51.529 --> 00:54:55.490 the uncertainty more rigorously. 00:54:55.490 --> 00:54:59.140 As I said, what we were planning on doing was doing forward modeling. 00:55:01.140 --> 00:55:10.480 - Walter, I’m not sure I can see whether it’s on the – within your domain of 00:55:10.480 --> 00:55:17.580 interest here, but there’s been a lot of talk about a subducted slab under the 00:55:17.589 --> 00:55:23.580 Colorado Plateau and the contributions that it may or may not have to uplift of 00:55:23.580 --> 00:55:29.380 the Colorado Plateau, you know, in the last tens of millions of years. 00:55:29.380 --> 00:55:37.000 Is that in there? Or is it – just am I too far away [chuckles] to see – 00:55:37.000 --> 00:55:39.160 you’re down to 30 degrees south, I think – 00:55:39.160 --> 00:55:40.780 I mean, 30 degrees north, right? 00:55:40.780 --> 00:55:45.220 - This is – Colorado Plateau would be – would be right in here. 00:55:45.220 --> 00:55:51.060 - Okay. - And – blank spot. [laughs] 00:55:54.260 --> 00:55:56.039 - I think … - I don’t know. I think it would – 00:55:56.039 --> 00:55:59.090 to answer your question, what we would like to do, I think, would be 00:55:59.090 --> 00:56:02.150 to make a bunch of transects that are at different azimuths and 00:56:02.150 --> 00:56:07.529 see if we see a consistent pattern. We do see something under 00:56:07.529 --> 00:56:12.660 the Wyoming Province. We have this dipping boundary here, 00:56:12.660 --> 00:56:16.279 but also this seems quite unique. And the question is whether or not 00:56:16.279 --> 00:56:22.990 we’re looking at the beginning of the delamination of Wyoming Province. 00:56:22.990 --> 00:56:25.799 But what you’re talking about in the Colorado Plateau, it’s pretty much 00:56:25.799 --> 00:56:30.170 a blank spot, Tom. I don’t – we got to look at more transects. 00:56:30.170 --> 00:56:36.400 - Did Steve Grand write papers on, you know, finding – excuse me – 00:56:36.400 --> 00:56:40.539 his observations for a slab underneath the Colorado Plateau or other people? 00:56:40.539 --> 00:56:43.300 - He focused on the history of the Farallon Plate. 00:56:43.300 --> 00:56:46.019 - Okay. - And the Farallon – that, of course, 00:56:46.019 --> 00:56:50.210 would be going back to the Laramide about 70 million years ago. 00:56:50.210 --> 00:56:55.240 And the Farallon Plate now is down – in central U.S., it’s down below 00:56:55.240 --> 00:56:58.770 the transition zone. It’s in the upper – it’s in the mantle. 00:56:58.770 --> 00:57:03.430 So Steve Grand focused on the Farallon Plate, not the delamination 00:57:03.430 --> 00:57:06.840 or the plate – or plate under the Colorado Plateau. 00:57:10.580 --> 00:57:14.819 - Okay. But shouldn’t there be part of that still underneath 00:57:14.820 --> 00:57:17.300 the Colorado Plateau? 00:57:20.540 --> 00:57:22.660 - Does anyone remember where the Farallon Plate is – 00:57:22.660 --> 00:57:24.980 Gary, do you remember where the Farallon Plate is under 00:57:24.980 --> 00:57:30.720 the Colorado Plateau, or Keith? I think it’s deeper. 00:57:31.800 --> 00:57:35.200 It’s not – it’s not in this depth range. - Yeah, well … 00:57:35.200 --> 00:57:39.020 - It’s significantly deeper than this depth here. 00:57:40.360 --> 00:57:44.100 I mean, that’s where the plate was at 70 million years, and since that time, 00:57:44.100 --> 00:57:49.800 it’s fallen into the – fallen into the upper mantle. 00:57:53.160 --> 00:57:55.240 Gary has a question. 00:57:56.960 --> 00:58:05.440 [Silence] 00:58:06.040 --> 00:58:11.000 - Hi, Walter. Very intriguing thought – talk. 00:58:11.019 --> 00:58:13.660 I have a pretty specific question. 00:58:13.660 --> 00:58:18.460 I probably should talk to you offline about this, but I’ll just throw it out there. 00:58:18.460 --> 00:58:24.220 Southern California. We’ve got a P wave drip 00:58:24.220 --> 00:58:30.680 under the Transverse Ranges. It goes down to 150 to 250 kilometers. 00:58:30.680 --> 00:58:35.730 I haven’t seen that plotted together with the LAB. 00:58:35.730 --> 00:58:39.560 And I wonder how those – first, what it looks like. 00:58:39.560 --> 00:58:45.080 Who did it? What is Brandon Schmandt or Humphreys or – and secondly, 00:58:45.099 --> 00:58:51.460 is there a way to rationalize those two pictures in southern California? 00:58:51.460 --> 00:58:54.970 Yeah. You know, compared to what we’re doing here – we’re summing 00:58:54.970 --> 00:58:59.010 over – we’re summing over a very large distance – 10 or 15 degrees. 00:58:59.010 --> 00:59:04.700 And the Transverse Ranges, you know, has a width of about 2 degrees. 00:59:04.700 --> 00:59:09.030 So it would be kind of blurred in our – any of our images. 00:59:09.030 --> 00:59:15.119 I mean, we – what we really see is the average existence of this 00:59:15.119 --> 00:59:19.930 very shallow LAB under the western United States here and here. 00:59:19.930 --> 00:59:24.369 But as far as localized features, when you – when you’re summing 00:59:24.369 --> 00:59:29.360 over such a large distance, a lot of that would be blurred out. 00:59:29.360 --> 00:59:33.049 If we can do a better job processing, we can start bringing these bounds 00:59:33.049 --> 00:59:37.080 in and start looking for more localized features. 00:59:38.460 --> 00:59:43.420 Perhaps it’s surprising that we even find this paleo plate here because that – 00:59:43.430 --> 00:59:48.640 I’m surprised that – that comes – that comes out really very clearly, this here. 00:59:48.640 --> 00:59:52.940 And something here under – as I said, under Wyoming Province. 00:59:55.380 --> 00:59:57.680 - Any more questions? 00:59:59.580 --> 01:00:02.100 All right. Well, let’s give Walter another round of applause. 01:00:02.100 --> 01:00:07.200 [Applause] 01:00:08.280 --> 01:00:12.080 [Silence] 01:00:12.080 --> 01:00:15.480 Okay. Hey, how you doing? [inaudible voices] 01:00:15.480 --> 01:00:19.580 [Silence]