Xiaodong Song (PhD '94), Geophysicist

Xiaodong Song is a leading researcher in seismic wave propagation - the study of how waves move in a radial fashion from their origin point in an earthquake. He has worked on applying seismic observations in modeling and he has completed important work on imaging the structure and dynamics of the Earth's interior with particular focus on the deep mantle and the core, as well as the lithosphere structure and tectonics of east Asia.

Song completed his undergraduate degree in geophysics at the University of Science and Technology of China before coming to Caltech's Seismo Lab, where he focused on the Earth's core under the supervision of Don Helmberger. After completing his PhD, Song went on to become a postdoctoral scholar Lamont-Doherty Earth Observatory, and he then joined the faculty at the University of Illinois at Urbana-Champaign. His current focus on Asian plate tectonics began at Illinois, and he continues this work in his current position at Peking University.

DAVID ZIERLER: This is David Zierler, Director of the Caltech Heritage Project. It's Tuesday, April 26, 2022. I am delighted to be here with Professor Xiaodong Song. Xiaodong, it's great to be with you. Thank you for joining me today.

XIAODONG SONG: It's great to be here, and thank you for the opportunity.

ZIERLER: To start, would you tell me your current title and institutional affiliation?

SONG: I'm a Chair Professor at Peking University in Beijing, China right now.

ZIERLER: Is that your most recent affiliation since the University of Illinois? Is that where you went next?

SONG: Yes. I left Illinois a couple years ago, in 2020. This is my most recent place.

ZIERLER: Tell me about your decision to return to China.

SONG: I'd been in the US for 30 years, and my family in the US were grown up and all independent. I thought maybe something different could be interesting. It was a very simple decision to find something different. This is where I grew up, lots of friends and a lot of people who helped me. Also, a lot of students and some who I took on when I was in the US. It was a combination of different factors, but it all seemed to be a natural change where I would like to find something different.

ZIERLER: A few overall questions as they relate to your research. First, there's seismology, there's geophysics, there's plate tectonics. What is the broadest field that encompasses all of the research that you do?

SONG: The broadest field, depending how I phrase it, obviously, we're doing earth sciences, and we're studying the earth and how it works. To be more specific, I'm focusing on seismology, studying earthquakes and earthquake waves, seismic waves. Then, I use them to image, to reveal what the interior of the earth looks like.

ZIERLER: Tell me about the role of interferometry in your research. What do you use that tool for?

SONG: That's relatively more recent. In seismology, we study earthquakes. We essentially are more interested in the signals, the waves that are generated by earthquakes. Those are the signals. But starting nearly 20 years ago, 2005 or so, people started using noise rather than signals, noise that's in the background, and using interference, correlation of coherent signals within the noise to generate waves from one place to another. So "noise" is "signal" in this case and we can actually extract signals between two receivers, for example, rather than from the source (earthquake) to the receiver. I wasn't the first one who started on this, but after it became clear that the phenomenon was very important, I started looking into that and using it more as another tool, another wave generated by the earth and to image the interior structure. Since then, it's become a booming field. A lot of people are doing it. My main use of it is to try to image the lithosphere structure. You mentioned plate tectonics. That's what makes the plates. And interestingly, in one study, we were able to use interferometry to detect the wave that goes straight through the center of the earth. That was an interesting twist. It turned out that the signal was larger than any signal you could observe from the largest earthquake in the world.

ZIERLER: Tell me about some of your work on lithosphere structure. What have been some of your key findings?

SONG: I think lots of it is about getting a better image of the lithosphere. There are certain key issues, for example, I studied quite a bit the Tibetan Plateau. Since the early days of plate tectonics, how continents collide with continents and what happens afterwards have been challenging issues, and even in the early days of plate tectonics, people understood that it's not a perfect theory, and in terms of continents in particular, it's quite different. We don't have a sharp plate boundary. Normally, the deformation zone is a much larger region with a lot of earthquakes that spread away from the collision front. How the continents collide, and how it changes the deformation, how they change the land's shape, have been key issues. In the case of the Tibetan Plateau, why, in collision zones, do we have normal faults? We have rifts. We also have earthquakes that happen in the mantle. Those mantle earthquakes normally happen in the subduction zones in the ocean slab.

Those issues have been sort of intriguing, and people have different theories. I think that might be one of the interesting results in recent years by imaging the lithosphere structure of the Tibetan Plateau, and we've come up with a model that's sort of sandwiched between a typical plate subduction, like between the ocean and continents, and also a very simple model of continental collision, namely that in the case of continental collision, you still have plates, which have different strengths. Like, the Indian Plate would be stronger, the Tibetan lithosphere would be weaker, but you can still have subductions, so you have a situation which happens in between a very typical subduction and continental collision, where you don't expect any subduction. In such a model, you could explain the mysterious phenomenon I mentioned.

Plate Tectonics in Asia

ZIERLER: I'm curious, your focus on plate tectonics in East Asia and China. Is there anything that's unique about that region of the world? And what might you be able to extrapolate globally from your focus in East Asia?

SONG: That would go back to my graduate study years at Caltech, but for quite some time, I dove deeply into the earth and its core. But at a certain point of my career, I thought it would be interesting to look at the shallow part as well because we live at the surface and there are just more people who are interested in the shallow earth. China and East Asia are very active in terms of earthquakes. The planet's very active with volcanoes and earthquakes. These are the manifestations of our dynamic planet. And being a seismologist, obviously, earthquakes are an important phenomenon to study. In terms of China and East Asia, what really stands out for me are the earthquakes in the high mountains. They're very active, they're still deforming, you can measure it with GPS, and that sort of thing. In terms of China, there's another very important aspect. There are so-called intra-plate earthquakes, which happen well inside the plate. That's also something that doesn't fit into our standard theory of plate tectonics where we normally expect earthquakes to happen along a plate boundary or near a plate boundary.

The plates are rigid, so you don't expect too many earthquakes inside the plate. I spent a lot of time, a little over 20 years, at Illinois and in the Midwest, and there are earthquakes within the plate as well. New Madrid is the most prominent one. And I've always been intrigued. These earthquakes are well inside a plate, well inside a continent, and where a lot of population is. Beyond the question of science, earthquake hazard is a very important issue. In China, for example, you have a devastating earthquake like the 1976 Tangshan earthquake and several other ones that were within the plate and that were large, strong, and damaging. And these are all well within the plate boundaries. What causes those earthquakes is very interesting scientifically.

Places like the Himalayas are irreplaceable. But it could help understanding how, over time, the landforms have changed in other places as well, from the Central Asia and the Middle East, for example, Iranian all the way to the Alps. This type of continental collision is an important problem around the world, and it happens on land, which means that there are populations. And in the interior as well, and how it happens. I think those are two key questions facing seismologists, geologists, and people who are interested in how the planet evolves, which have implications for societies.

ZIERLER: I've come to appreciate that in Japan, the history of seismology and geophysics goes back very far. Do you have a sense if the same is true in China?

SONG: Not so much, depending on how you compare. China has a long history, obviously, and people realized the phenomenon of earthquakes a long time ago. The first seismometer, or you may not call it a seismometer, but more like an earthquake detector, a seismoscope, was invented in China more than 2,000 years ago (Note: at 132 BC by Zhang Heng). It was written in the history books that it could detect an earthquake and the direction where it came from. The seismoscope used the principle of inertia, which is still the basic principle for modern seismometers. It was a remarkable achievement for the time, even though that's very far away from modern seismometers. You mentioned Japan, which certainly was one of the early societies, together with Europe, U.S. and Russia, which developed modern seismology. China, in a way, could be important in seismology. That is, the written language, written records of earthquakes happened in China, even though I don't think people were able to detect the ground motion like with modern seismometers.

But China has very good written records of natural phenomena. After earthquakes, for example, there are records of damage, and those records have been kept regarding where, when, and how big, just from the intensity (survey of the damage). There are actually these types of records in history since 780 BC in China, which help us to understand the earthquake activity. Large earthquakes have happened on land, but they don't happen too often, so those are valuable records. But in terms of modern seismology, really, it was quite late. Obviously, there were other things going on because of world history in the mid-19th century, all kinds of wars, to mid-20th century. But that's when that modern seismology started, in the late 19th century. I would say that in terms of modern seismology in China, it was very much influenced by the Western countries, particularly the United States, the people from China who went to the US and the Europe to study.

ZIERLER: What have been some of your major findings on the core-mantle boundary? What have been the tools you've used, and what are some of the arguments and debates around where the core ends and where the mantle begins?

SONG: Since my PhD and thesis work, I've been working on the core continuously. I was lucky enough to connect with it from my advisor, Don Helmberger, at Caltech, which attracted me to this problem. I'm very lucky now to be able to continue on this project, both with funding and with the positions I've held, as well as the students I was able to work with. But in terms of the core, at the beginning, we knew very little. Because it's deep, and it's not easy to sample. At the very beginning, not too many people had studied the problem. Also, the tools were limited, even with observation, with seismograms, for example. You need tools in order to decipher what's embedded in the signal. In that way, I was very lucky that Don Helmberger was the pioneer who had a lot of expertise and was able to decipher what was embedded in the seismic wiggles. Normally, people study the travel time, the time it travels from the earthquake to the station. That's normally the easiest information to extract. But Don looked beyond that, looking at the waveforms, the shapes. And not just the one wave.

You could have multiple waves arriving at the same station from the same earthquake. That actually gives you keys about the earth's structure. Just like if you have different types of lenses, for example, you can focus the waves or divert the waves. That tells you a lot about what type of lens you have. Not just one wave passing through, in which case, you might not be able to tell what happens because a structure could happen any place along the wave path. This is something we call waveform modeling. We tried to come up with a structure that would be able to match the waveform you observe because you can look at the shape, look at the multiple arrivals. That structure would be much more deterministic than otherwise. That's a general background of the approach in the early days, when I studied the core for my PhD. And the study of the core, particularly the inner core, has taken up a great deal of my time up until this day. I think the findings about the inner core in particular have been interesting and different from the past. We found that the inner core is not just one uniform sphere, that it has complex structure.

Anisotropy was proposed, and my colleague and I were able to confirm the structure and to determine the structure, gradually. We found that the inner core has its own layers, that the shallow part has a one-layer structure, then the innermost part has another layer structure. Later on, we found that the inner core is rotating relative to the mantle. These are new structures that people didn't realize before. You asked about the core-mantle boundary, and that's very interesting. This boundary obviously is very important. It's a chemical boundary separating the iron core and rocky mantle, and in many ways a sharper boundary than between the solid earth and the atmosphere in terms of density, temperature, and velocity differences. In fact, at the beginning, Don's interest, and my first project when we talked about potential project in my first year of grad study, his interest was actually the core-mantle boundary. Before my time, a lot of people had studied this, but mainly on the S-wave, shear wave. The earth has different properties, P-wave, S-wave. They're all related to earth's properties. But shear waves, in the early days, was easier to study, and the phenomenon was clearer in terms of the structures that happen at the core-mantle boundary. But Don was also interested in the P-wave. We started looking at P-waves that pass through the core to study the core-mantle boundary.

But the project changed very quickly before it really got started. It turned out that studying P-waves through the core is different than studying the S-waves. For S-waves, you can pretty much ignore the inner core. But for the P-wave, because it dives down much more steeply, it has to pass through the inner core. That's when we found that we had to study the inner core, and it turned out to be something I've continued for the rest of my life. But the initial interest was on the core-mantle boundary. With respect to the core-mantle boundary, I did a bit of work, in particular, with Don, because that was where I started. In fact, at one point, I had my sabbatical, when I went back to Caltech and connected back with Don again. I said, "Hey, this is something I started. I probably should do something in order to meet your expectation at the beginning."

But obviously, it's a question that's very interesting to start with. Essentially, the key question is, what is the nature of the core-mantle boundary? What kind of material is it made of? What really causes such a deep structure and other phenomena that manifest at the core-mantle boundary? For example, you talk about superplumes. The origin, where they come from. There are large-scale structures that people have imaged. What is the nature? Is that just coming out of the temperature? Is there some kind of violent chemical reaction? What really happens? What we've tried to do is to characterize the structure. What's the boundary, what's the wave speed, what's the heterogeneity, and that sort of thing? And as I mentioned, more people have been focused on the shear wave because the observations are more prominent, and they can ignore the core and inner core pretty much. My work mainly focused on the P-wave as a complementary property. One interesting study is under Central America, where we found a very prominent P structure as well, and that correlates quite well with the shear wave. And that helps us understand the nature of the heterogeneity of the structure. That's just one study I'll mention.

Rotation of the Inner Core

ZIERLER: You mentioned previously the rotation of the earth's inner core. First, are the terms rotation and motion interchangeable, or are those two different ways of describing how the earth's inner core moves?

SONG: I, myself, use them interchangeably, but they do have distinct meanings. Motion obviously means movement. It doesn't have to be rotation. It could be some other type of motion, for example translation. But rotation normally just means that it's rotating around an axis, like the earth's rotation. Right now, people are still trying to characterize the type of change coming out of the inner core. I think the key thing is that we and other people have found that the inner core is changing, and we can observe it. I think right now, it's nearly 100% for sure that we have that kind of evidence showing that the inner core is changing, but what type of change the inner core is experiencing is something that's still under debate. Rotation is one proposal. Within rotation, there are other categories. For example, it could be uniform rotation, steady rotation, or non-steady rotate, stop and go. Or you could have oscillations, rotate and come back. There's another type of proposal which essentially says that the inner core as a whole doesn't rotate, doesn't move, but you can have localized changes. Like at the inner-core boundary, for example, the boundary could increase and decrease, changing the topography. That would be another proposal, and there has been a lot of debate. Even a couple years ago, we had a couple of papers. But I think it's clear now that it looks like some type of rotation, could be oscillation, is happening. That's what we believe at the moment.

ZIERLER: You mentioned some of the debates. Is that to say that there are some seismologists and geophysicists who don't think the earth's inner core rotates?

SONG: Yes. In terms of rotation, I think there are still a lot of people who don't believe. It may not be a majority, but certainly some people. The debate has gone through different stages. The first stage is the fact that initially, we observed the travel time of seismic waves change with time, with years and with decades. At the beginning, people thought, "That's not real." It went through the stage of trying to determine whether the temporal change of seismic waves through the inner core was real or not, and that took almost a decade or so. It was very difficult, and the signal is kind of delicate, there could be artifacts, and so on. But that debate, I think, has been well-resolved. I think very few people doubt that such a phenomenon, seismic waves through the inner core or some part of the inner core changed over time. Another part of the debate is the interpretation. What causes the change of the travel time? And I mentioned two categories of interpretation. One is some type of rotation. The other one is no rotation, but some type of change. For example, the change of the inner core topography. And this debate is still going on rigorously right now over the last couple years. We actually have a paper under review right now, and there's a paper from one of my colleagues coming out of Science Advances, saying the inner core oscillates. This is still under debate and may take more time, but I'm proud that part of my observation until this day has been quite consistent, remarkably, and holds up to a great extent.

ZIERLER: Without the ability to drill down into the earth's inner core, and your focus on 3D anisotropic techniques, how can we settle some of these debates about the nature of earth's inner core?

SONG: It's difficult. Science is always a process, as you know. But in such a continual pursuit, surprises come up. In terms of the inner core, I'm very hopeful. Modern seismology, in a way, is only a little over 100 years old. Plate tectonics is just a little over 50 years old, established in the late 60s, how much we've learned about how earth works, the earthquakes, the structure of the interior, and so on. From the beginning of modern seismology until the early to mid-1950s, people discovered very quickly the major layers of the earth. The inner core itself, the new phenomena we've observed are also relatively recent, the rotation, the 3D structure, 20 or 30 years old. One way to try to settle some of these debates would be with more seismometers. Even though we have a lot of seismometers, we're still missing key observations. For example, we have very few stations under the ocean and in high latitudes. The Arctic and Antarctica are not well-instrumented. But for 3D anisotropy -- essentially, the structure of the inner core is three-dimensional and anisotropic -- you do need observations from different angles and different places. Those key observation are still to come.

Once they're available, in terms of the deepest interior, it would be revolutionary, I think. Another part would certainly be advances in terms of tools and methods, such as fuller information of waveforms. Even though we looked at the waveforms in the early days, the tool we used was only approximate. Nowadays, it's more powerful with computational tools. And we will be able to model, to image, to use the information in a waveform more fully, and that would also be very helpful. Obviously, there are other advances, for example, in terms of experiment, in terms of the predictions of the material properties of the inner core, which also can serve as a guide for structures to look for. For example, what really causes such a 3D anisotropic structure? What kind of material, what kind of phases, and how has it evolved over time under extreme pressure and temperature to arrive at the structure that we have for the inner core today? I think all those advances will be important in the coming years and decades.

ZIERLER: In order to set the context prior to your arrival at Caltech, tell me about your undergraduate education at the University of Science and Technology of China.

SONG: I entered my undergrad in 1981, and I graduated in 1986. That's five years. That was fairly unique. Later on, they changed it to four years, like all the other universities, to be competitive. But USTC was very unique at the beginning. It was affiliated with the Chinese Academy of Sciences. It was supposed to be the university which trained scientists. Later on, continuing a research career is one of the main focuses, which is why I think about it in comparison with Caltech. There were a lot of similarities. Proportionally, Caltech trained a lot of people who went onto an academic track. And USTC tried to train scientists with four years of normal undergrad curriculum plus one more year of undergrad research experience, where students would go to different research institutions, mainly, within the Chinese Academy of Sciences, but it could be some other research institutions, to gain research experience. It was one of the top universities, and by my time there, China had opened up to the world after the cultural revolution, and science and tech were some of the main drives of the country. People wanted to study science and tech in order to modernize China. I'd say the best students in the country were attracted to this field and this university in particular. With such a strong student body, it was obviously very competitive. I grew up in a rural area, and even though I did well in my high school, it's obviously a different environment in the undergrad. The training was fairly typical of modern curriculum, even compared with the US, in terms of science courses, liberal arts courses, and so on. What was different about this university was that we also gained more research experience than at some other colleges in China, but not so much compared to the US system.

ZIERLER: Was there an applied component to your undergraduate education? Were you specifically training with the idea of doing oil exploration?

SONG: No. In undergrad, it was pretty much an all-around education, with physics, math, chemistry, foreign language, some liberal arts courses. It was a pretty liberal curriculum that was not specialized, for example, in oil exploration. I didn't know anything about oil exploration until I went to grad school, when I had a very short exposure to oil exploration. That was an interesting twist. You may have heard that I've had several interesting twists in my life. [Laugh] I would say that from undergrad to my master's program, oil exploration was an interesting twist. Essentially, after I graduated, I could enter scientific research, academic research in the Chinese Academy of Sciences, China Earthquake Administration, or go abroad, but I chose to study oil exploration because I thought with all the knowledge, it might be interesting to do something practical. But I didn't train for that purpose.

ZIERLER: Politically, was it difficult to leave China, to pursue a graduate education at Caltech?

SONG: Depending on the time periods. And I was right at the transition. You might know that from the late 80s, a lot of students were beginning to go abroad to study from China. Then, after 1989, probably around 1990, it became more difficult. Later on, it became easy again. Nowadays, it's really opened up. I was really at the transition that people were starting to go abroad to study. I arrived at Caltech on July 7, 1989. There was a little bit of a hurdle I needed to get over, but I got a lot of help, so there was no problem.

ZIERLER: Was it specifically Caltech you were interested in going to? Did you apply elsewhere in the United States?

SONG: I didn't apply to too many schools. At the time, we were able to actually get a waiver in terms of the application fee and that sort of thing. That was a consideration. You have to realize, at the time, we were very poor students. I'd say I knew very little about the universities in the US and Europe. What I knew was pretty much from other people and Peterson's, for example. You'd read the data and so on. I applied to several schools, and Caltech actually was very interesting. I would say the way I eventually went to Caltech was by chance, to a great extent, because I didn't request an application form. Some other people gave me the form and said, "Hey, with you record, you should apply here." They gave me the form and urged me to apply. I applied pretty late, but I got accepted. Eventually, I had a very difficult time deciding between Caltech and Stanford. I was accepted at some other schools, but my decision was mainly between these two schools.

ZIERLER: When you got to the Seismo Lab, how well-formed were your plans? Did you know specifically what you wanted to study and who you wanted to study with?

SONG: No, as I mentioned, I knew very little about US universities. I think one of the key things was recommendations from people who had worked or studied at the Seismo Lab before. There were a couple people. One was sort of my class supervisor, who had done a short-term visit for a couple years at the Seismo Lab. Then, there was another researcher who also spent time at the Seismo Lab as well as USC. I got recommendations from them, and I just asked around and compared different schools. The people who had spent time at Caltech were able to tell me a little bit about the Seismo Lab, and they thought based on my strengths that it would be a good match. That was one thing. Another aspect of my ignorance was that I really didn't know what research was about, even though I had a little bit of research experience, master's studies, and so on. But I really didn't know what research was about and how to go about it. After I arrived at the Seismo Lab, as was tradition then, we were asked to do two projects with two different faculty, and we had to pass our oral after a year doing these projects. That was when I really started looking into what the PhD was really about. Obviously, I had some colleagues and professors to talk to. The first year, maybe two years, it was really crucial for me to try to find out what is going on. Talking with people, with faculty, but also with peers, was very helpful.

ZIERLER: Just on a personal note, how was your English when you arrived in California?

SONG: I would say it was pretty good. I think somehow, I had a talent for language. It's interesting that I didn't have much training in English in China. I was in a rural area, and English instruction was very limited. I had three years of English training in high school, but only partly. In undergrad, it was one year, but more oriented towards science and tech training. The rest of it was pretty much self-study. But there was one key person in my undergrad who also visited the US, San Diego, and returned to China after a couple of years. She was a professor who I happened to know and mentioned that English was very important and told me to pay attention. That was the message I got. And obviously, at the time, we were able to read some papers, professional journals in undergrad.

Also, novels. We were able to borrow novels in the library. Learning language was one thing, but it was very interesting that through the language, you were able to see the world outside. And by then, we had radios to listen to some of the English broadcasts. And English had been in fashion at the time. But in terms of training, it was pretty much self-study. I landed at LAX, and I arrived at the Seismo Lab about 5:30 or 6 pm on a Friday. I went to the division beer right away and was able to strike up conversations with other people without any difficulty. That was another issue in terms of language. Perhaps, that was one advantage compared to other Chinese students ahead of me, who apparently had to struggle at the beginning.

ZIERLER: Once you got comfortable in the Seismo Lab and got a sense of all the different possible projects to work on, how did you decide your focus? Was it based on a class, a professor? How did you go about determining that process?

SONG: Very much because of Don and the problem he proposed. Obviously, there were many choices, and people did switch from one professor to another. The Seismo Lab is very unique. There was so much expertise in seismology. You could do a lot of things, work with a lot of different people. Even if you work with one professor, you could still interact with other professors. It was very open. For example, I took six seismology classes with different professors, including one by a visiting professor. [Laugh] And they were all different. But Don was very unique. Unfortunately, he passed away. We have this special issue that Thorne Lay and I, as guest editors, have tried to put together, all with his former students and post-docs, a special collection of memorial articles to remember him. He was very unique in terms of being able to get students motivated and find things interesting to do. But I did have two projects. One project was with Don on the core, and the other project, which was also interesting but unfortunately didn't have a chance to pursue later on. After all these years, and trying to learn something, science, all the formulae, all the physics, you suddenly find something where you go, "Oh, this is cool, this is interesting," where you can do something useful that other people don't know. That was almost like love at first sight in some way. You sort of just get attached to it, the appeal of something other people don't know that you can learn something interesting about, a structure that could match with the data. The problem and the personalities were really important in determining which project to work on.

Waveform Modeling at Caltech

ZIERLER: Tell me how all of this specifically got narrowed down into what would become your thesis research.

SONG: That was sort of a natural, relatively easy process. Once I found something interesting, I tried to learn the tools, the waveform modeling tools. It turned out Don had used this tool for many studies, including a study of the core-mantle boundary. But he didn't use this tool to study the core. It needed to be modified a little bit for the core. I spent some time trying to modify it and trying to generate the so-called synthetic seismograms, seismograms generated by humans. But the idea was that once you generated a synthetic seismogram, you'd be able to match the real seismogram, the observed seismogram. I spent some time trying to modify the tools. Just to give you one example how Don would motivate a student, after we finally got it to work, Don would always refer the codes to be "your codes". Really, I did very little. But it's just one way he was so good at motivating students, which made you feel that you are doing something useful. You just tried to understand the waveforms, and in the process, we found out that the models of the inner core boundary vary greatly. You had all kinds of models. Not all of them could be right. It just meant that at the time, people hadn't really had a good handle on it, and data were limited. So, we started to model it with waveforms to be able to have a better constraint about it. It was my very first scientific publication and I consider it to be one of my best papers. One figure was later used in a classic textbook.

As we went on, we started to realize that there were more questions. Another question at the time that became obvious was the question about anisotropy. That idea had been proposed, but there were different controversies. Then, we tried to find data and see whether we could add some new observations and new constraints into such a debate. Essentially, once you dive into this new field, you start to find problems. Some problems are small, and some problems are major. You choose one to work with, determine the right tool, then you try to find observations, and you move on. In terms of development of my thesis, it was quite natural to me. The key was finding the right problem and the right approach. Also, having the right people to talk to. One more thing I'll mention about Don is that different professors have different styles, but he was very hands-on. Some other professor were very hands-off. Don liked to talk to students. In fact, quite often, there was a line for students to talk to. At the beginning, my office was a little bit further away from his office, and I would have to go to his office, peek in, and see whether he was available or talking with some other students. Sometimes I'd come back at a later time.

Sometimes he'd have a note that said, "See me," which meant he had time available, or that he was interested in what was going on, or he had new ideas on something. I think for someone like me, who was less experienced, it was very effective in terms of really getting onto the right track with this type of hands-on experience. I mentioned all the other things, like other professors and peers. Those were also very helpful. We'd have coffee hours twice a day, morning and afternoon. People would talk about what they thought and interesting things. Also, you knew what other people were doing, more senior students, how they were doing research. There were seminars going on constantly. I think those are all natural, and that was the environment the Seismo Lab provided, which all helped students to grow how they needed to put together a decent thesis. I would say once you were on that track, if you wanted to do some outstanding things, that's different. But if you just wanted to do cutting-edge and decent PhD research, it seemed to be a pretty natural process. I mentioned that different people have different personalities. And we did have students who failed exams or struggle. Personality sometimes plays a role in that. For me, I just like to learn, and I like to talk to people. I think for me, it was quite helpful.

ZIERLER: What role did computers play, and specifically simulation, in your thesis research?

SONG: A great deal. At that time, computing was going through a revolution, and the internet came about around that time. In terms of my own thesis work, I had to generate synthetic seismograms, so I had to use a computer. The first time I met Rob Clayton, a professor at the Lab, he gave me a book about C programming. The message was not so much that it could be a useful tool, C programming. The message was that computing was important and something we would need to pay attention to. In China, in undergrad, I did have some exposure to computing and programming languages. And I did use a computer, even though it was not very fast or powerful. For me, Caltech was special also that in fact I got a computer science minor out of my PhD major. I did several computer classes at Caltech in order to meet the requirement of the minor.

The thought was, the field of computing was very important. If I knew more about it, I could be more efficient. The courses I took were not so much useful in terms of my thesis research. Eventually, I could do without those courses in terms of research. I could have spent more time actually doing coding. A lot of people did that, and that was just fine. What I found more useful later on was that, because I knew more about it, I was able to manage computing resources in my lab with my students because they didn't know too much about it. You may not be able to find staff members to solve a problem right away. Also, understanding code was important. If I couldn't write code, but my students could, I would've been short-handed. It's important, but I guess it would be still natural that what I needed to know is to be able to write and understand codes, and generate seismograms, and put together figures, and that sort of thing. But it's just part of your PhD process.

ZIERLER: What were some of the central conclusions of your thesis research?

SONG: I would say maybe just two parts. One, that we proposed a new model, a refined model for the core, the inner core in particular. And the other one, we confirmed the existence of the anisotropy of the inner core and put some new constraint on its structure.

ZIERLER: Who was on your committee besides Don?

SONG: There are two committees. One's the oral exam committee, and the other one is the defense committee. On my oral exam, I remember Toshiro Tanimoto, Dave Stevenson and Peter Wyllie. Wyllie was my chair. And I remember Wyllie in particular because after the oral exam, he wrote a very short but very influential and very nice letter. He essentially pointed out that I need to broaden the scope of my interest. And that was very good advice. For my PhD defense, it was a different committee. I had Hiroo Kanamori, and perhaps Toshiro Tanimoto, Tom Ahrens, and Don Anderson (Note: I have the records but most of them were left in the U.S. and inaccessible due to the Covid). Kanamori, I remember well. I really liked his advice. He was someone I would've liked to have interact with more, but I didn't have too much interaction with him during my PhD. I thought that was my last chance. [Laugh]

ZIERLER: With your thesis complete, where did you see yourself situated in the field? In other words, where did you see your research contributing to some of the big debates that were happening at that point?

SONG: I would say with my thesis completed, I thought I had a good grasp of the problems about the core from a seismology perspective. I thought I was well-equipped to continue research in this area. I was sort of one of the people who could do the cutting-edge research in the world. It wasn't so much about specifics. There were certainly specifics, for example, the debate about anisotropy. By the time I graduated, its existence was about settled. But still, you can do more work. I think the bigger question was it was less studied, which means less known. Simply looking at new data, more data, new tools, more work could reveal something different. By the time I graduated, we already knew that the inner core was not where people thought decades before then. That was sort of where I was.

ZIERLER: When you arrived at the Lamont-Doherty Earth Observatory in Columbia for your post-doc, in what ways institutionally was it similar, and in what ways was it different from the Seismo Lab?

SONG: It's actually quite different. Lamont obviously is much bigger than the Seismo Lab. The Seismo Lab is very concentrated. A lot of people are doing a lot of different things, but more focused on earthquake studies, the earth's interior, and so on. At Lamont, seismology and tectonics is one unit. You have geochemistry, then ocean science as well. Obviously, in the division at Caltech, we have those types of people as well, and we do interact with other people outside the Seismo Lab in the division. But I would say less so at the Seismo Lab because so much is going on. I think people tend to interact within the Seismo Lab a lot more. I guess at Lamont, even though they also have a unit of SG&T, Seismology, Geology, and Tectonics, there are other units as well, and I think people tend to interact with other units a lot more. People tend to see other people a lot more. By the time I was a post-doc, I guess I was even more open to talk to people in different groups. The other thing that's different is, Lamont has a lot of independent researchers. People who aren't full-time faculty but are full-time researchers. They just concentrate on research. As a post-doc, that's a good environment to develop further how to be an independent researcher. Not just doing research, but how to secure funding and that sort of thing. [Laugh]

ZIERLER: As a result of being in such a different environment, in what ways did your research focus change as a post-doc?

SONG: One thing I really benefitted a lot from the PhD, because I did a reasonable job, I think people knew my work. Lamont allowed me, a post-doc, to write proposals, so I wrote one, and I was able to get funded. And that was pretty much out of my PhD work. Then, I think a major change was the discovery of the temporal change of the seismic waves through the inner core with Paul Richards at Lamont, which were interpreted as inner-core rotation. That discovery became so popular that it helped greatly in terms of future research. At the time, because of that work, I was able to continue pretty much working on the core because of the funding and because there were enough questions to study at Lamont. But at the time, I did start to think about some other questions out of necessity of diversifying funding, but also just thinking more broadly about what I could contribute in terms of seismology and earth studies long into the future.

ZIERLER: Tell me about the opportunity at Illinois. How did that become available for you?

SONG: Essentially, Lamont is a very stimulating environment. I could have stayed on, and I struggled a bit on whether to move on. There were faculty positions opening up now and then, and they were rare, not so easy to come by. I suddenly paid attention to those opportunities. Why not? But I really didn't know anything about Illinois. I knew engineering was really good. There were some people from Illinois in earth science, geology, and all that. But I knew very little, even by the time I was a post-doc. But one faculty from Illinois contacted me and said, "Hey, there's this position available, and you should apply." I said, "Hm. Why not?" [Laugh] kind of a normal process. But I did well enough, I guess. I decided to leave Lamont and take on this position as another opportunity to have a tenured position and be able to supervise with students. Teaching is also something I had been interested in, being able to pass knowledge and experiences down through generations, and particularly there were interactions over the years with other professors, both in undergrad and my PhD. They're certainly excellent examples to guide my career.

ZIERLER: In Illinois, the department was of geology. Is that to say that geophysics and seismology were located within that department? Or were they separate disciplines?

SONG: They were within the department. Our department is Department of Geology, and that hasn't changed. It's also different from Lamont and Caltech. It's a small department, but we have faculty in all various disciplines. At the Seismo Lab, there was really a lot of expertise in seismology, tectonics, and earth interior. But Illinois's very different. We have all the major disciplines. Part of it is that we need to teach a large body of our students various different aspects of geology as a state university. Seismology is within the Department of Geology, but the department covers, very broadly, different areas of geosciences.

ZIERLER: What were some of the advances in instrumentation and computation during your time at Illinois that were relevant for your research?

SONG: I would say one was the explosion of seismic data, both in deployment and in terms of availability with the internet and all that. It started around 1989, 1990, right about that time, there was TerraScope at Caltech. But it continues, the data has accumulated, worldwide. In my time at Illinois, nationwide, you have this Earthscope project with the US and other countries as well. There's been an explosion of seismic data. In terms of the data I use a lot, there's global high-quality network, which is critical in terms of understanding, imaging, and exploring the interior of the earth, and then regional studies, which are important for understanding plate tectonics, questions which cannot be explained by standard plate tectonic theory. At that time, I started working on China and East Asia. One of the key reasons is that there was a lot of data that became available at the time. It may not have been available openly, but I had good collaborations. These regional and global revolution of data were really critical. The other thing is computation. Illinois hosts one of the supercomputers in the country, so the computational resources are really good. And that's really helpful in terms of some of my more computationally costly projects. You mentioned interferometry. It was during my time at Illinois that I started to take on interferometry. For that, you really needed computational resources in terms of computational power, speed and time to compute, but also in terms of storage. Because you have continuous data, the amount of data is huge, very different from the earthquake data.

Returning to China

ZIERLER: Just on a personal level, were you happy at Illinois? Did you think you would spend your long-term career there?

SONG: I was very happy. But it was different, in the middle of a small town compared to Pasadena, which was close to metropolitan LA. There wasn't much to do recreationally and that sort of thing. But it's one of the excellent institutions in terms of research and education, and people are very friendly, and you can collaborate with people very easily. In fact, I collaborated with Richard Weaver in the Physics Department of Illinois on the interferometry, and that was very profitable to me. In fact, I didn't really go into this project until I realized that there was a world expert just a block away in the physics department. I didn't know he was doing this sort of thing until I saw his Science piece in 2005. We started with another professor at Purdue, which is very close and also very open. And we had a joint class, two people from Illinois and one from Purdue. We started as a class talking about this phenomenon, then we started this project. Scientifically, it's a very stimulating place, easy to collaborate. In terms of personal life, even though there's really not much to do, the science took a lot of my time, and I would be happy working 24 hours a day, if I could. But my family is very supportive of me. My wife stays at home taking care of the family. Both of my kids, son and daughter, grew up in Illinois. In that aspect, Illinois is a fantastic place to raise a family. You can have a very big house with a basement for the kids at a very affordable price. There's a lot to do, museums, recreational parks, close circle of friends, and so on. And because there's not much to do in terms of spending money in a small town, you can save it and travel around the world during holiday breaks, summer, and so on. [Laugh] It was wonderful.

ZIERLER: Returning to China, in what ways has the field in geophysics and seismology advanced since you left initially to go to Caltech?

SONG: Oh, there's a lot from '89 until now. First, there have been continual groups of scientists who initially visited abroad who have returned to China. I mentioned that I think modern seismology and geophysics really started as a field when well-trained geophysicists returned to China in the first half of the 20th century, including one of the earliest Chinese graduates from the Seismo Lab (Dr. Fu Cheng Yi, Ph.D. 1944). He became one of the pioneers in China who trained a lot of students. After I left China, there had been a continuous stream of students who have studied with well-trained professors and visitors who have gone to the US, Europe, and other countries. The world became more integrated and open, and China became richer. You have people who attend international conferences, a lot of visitors coming to China, and all that. All these scientific exchanges have been very helpful and all the training of the new generations has been very important.

But I'd say that within the last ten years, there's really been a flood of well-trained geophysicists who have otherwise stayed in the US. Not necessarily at faculty positions because they're rarer, but people just stay and work in the US in the oil industry, in the financial industry, in the IT business, for example. But they returned to China, and that has really made the geophysics, seismology, for example, more exciting and more competitive. It became a boom in a way. You just have more excellent people whereas in the past, there may have only been a few. They were always excellent, but now there's more in quantity. There's a more vibrant community, and in that way, it's been a big change. The other big change is investment, the amount of money put into basic research. In seismology and the understanding of earth, you really require a lot of fundamental research, where there's also been increasing investment over the years. And that actually makes a lot of difference in terms of different types of projects people are able to try and do.

ZIERLER: Just to bring our conversation right up to the present, what are you currently working on?

SONG: I'm currently working on three projects. One is something dear to my heart on the center of the earth, the core. That's one project. Another project is something I started in Illinois, mentioned earlier, on the lithosphere and tectonics of East Asia and China. And the third one is really about the earthquake itself, looking at the so-called transient phenomena, whether there's something associated in the process of earthquake genesis, earthquake nucleation, and earthquake ruptures. In other words, whether there's a certain changing phenomena associated with major earthquakes, something we can either detect before the earthquake or something we can try to understand what happened after the earthquake, how a big earthquake is nucleated, and then after the earthquake, how the fault is healed. These types of processes are interesting to me. Those are the three major directions I'm interested in.

ZIERLER: For the last part of our talk, I'd like to ask a few retrospective questions about your time at the Seismo Lab, and then we'll end looking to the future. First, with all that you've been able to accomplish, all of your professional success, what are some of the things that stay with you from your time at Seismo Lab, the way that you learned to approach the science, the way you learned to analyze the data, the way you learned to work with collaborators? What are some of the secrets to your success from Caltech?

SONG: Well, I would say many. The Caltech experience certainly is a life-changing experience. Without it, I would be a very different person doing very different things, I'm sure. In terms of my academic career, the way you learn to do research has been very fundamental. This is more on a philosophical level, and this is something I try to instill in terms of educating my students. The way I learned to do research from advisors and professors is something that's stayed with me over my career. The way you think, how to really find a problem and find a way to tackle it. This is still about research, but different aspects of doing research. I would also say the humiliation you feel when you have that kind of experience at the Seismo Lab is important and stays with you. Because you're working with the best scientists in the world, and your peers are the smartest, most competitive people in the world. Even though you feel talented and smart, you really have that humiliation when you consider what they've been able to do and come up with that you didn't or haven't.

I think that sense of how small we are in this world and how much challenge you face is a very important notion that makes you feel that there's so much to do in terms of understanding the planet. On the other side is the sense of pride and confidence you gain from your experience at the Seismo Lab. These are sort of two sides of the coin. You see how the best people in the world work, how you can match up with that level of ability to do cutting-edge problems. You realize you can do it. You work hard and find a way. And that kind of confidence, I think, is immeasurable compared to things like coding and that sort of thing. It's very important and has stayed with me, and has stayed with other people, I'm sure. Certainly, in terms of special techniques and so on, you do inherit them. Some of the graphic tools the student and faculty have developed at the Seismo Lab, I'm still using it. It doesn't mean I couldn't use something else, but some of the tools are very effective and still useful. Certainly, the way you look at seismograms over coffee break, talking with Don, those small things that seem small, but if you don't have that experience, you might not feel how these things actually make a difference and stay with you. I could continue forever. [Laugh]

ZIERLER: Have you maintained relations with the Seismo Lab? Do you have a sense of whether it's retained its stature and excellence in the field over the years?

SONG: I do have the confidence, and I did maintain contact with the Seismo Lab, but not as much as I wanted. Mainly, through the faculty over there and the students who study or have graduated from there. I've visited a few times after I graduated, but not as much as I would like. And I certainly would love to have the opportunity, once this pandemic is over, to visit the Seismo Lab in Pasadena again. I think the Lab is doing well. The world's changing, and everything has to change. But I think it'll still maintain the highest level of reputation around the world. And I'd say it will still attract the best students, in the U.S. or out of China or elsewhere, and will be competitive with any other institutions in both the United States and the world in terms of attracting student talents.

ZIERLER: Last question, looking to the future. What areas of research have you not yet tackled? What do you want to do that hasn't been covered yet?

SONG: Oh, boy. It's hard to do a lot of different things. The three things I outlined for you, I hope, will be enough for me to work on in the future. If I can make a little bit of difference in those areas, I'd be very happy. Research is a career, in a way, but I'd say it's more like a hobby. It's something that can be very addictive. In the future, even after I fully retire, I may still be looking at some interesting questions, maybe some new questions, new phenomena that have come up. If I have the ability and resources to work on them, I'll certainly do it. But right now, those areas, in terms of earth structure, in terms of tectonics, in terms of earthquakes themselves will be enough to keep me busy, at least for several years to come.

ZIERLER: Xiaodong, it's been a great pleasure spending this time with you. I'm so glad we were able to capture your insights and your contributions from the Seismo Lab all the way up to the present. I'd like to thank you so much.

SONG: Thank you very much.

[End]