This project focuses on the study of intermediate-depth earthquakes, those between approximately 50 and 300 km depth. The collaborative research applies machine-learning and waveform-based matched filter techniques to detect acoustic emissions events recorded during high-temperature high-pressure laboratory experiments. In addition, the lab plans to derive source properties (focal mechanisms, seismic moment and corner frequencies, stress drops) and statistical features (Guttenberg-Richter a and b values, aftershock productivity and decay constants), in order to gain more insights into the failure process at laboratory scales. These results will be compared with an ongoing work at subduction zone scales beneath Japan, and can provide important constraints for numerical modeling (stress state, fault geometry, etc) done at collaborative institutions. I will be responsible for managing the continuous waveforms recorded by broadband transducers during high-temperature high-pressure laboratory experiments. I will also apply both deep-learning and matched filter techniques to detect acoustic emissions events, and derive additional source and statistical parameters during laboratory experiments.
This week was a little bit (lot) more intense than the first. Outside of the internship has been a whirlwind, as I got to go to my first NBA playoffs game (go Hawks!) and got a little sick. As far as work goes, I have really been enjoying the IRIS Seismology Skill Building Workshop, especially because whenever I don't know how to do something lab-related I can do a tutorial and still feel productive. So far it feels like everytime I slowly, painstakingly learn something new in lab, it shows up on the next tutorial and makes perfect sense. I'm trying to be more ahead of the game on the tutorials now, so that I can hopefully switch this and learn new tools before I need them in lab! My work has ramped up a bit, and I'm starting to write small bashrc, MATLAB, and Python scripts for various SAC files. My big project this week has been writing a script to add 13,884 P-wave arrival-times to SAC headers on various frequency-time plots. It's starting to look like it'll be next week's project too....
My project is slowly becoming clearer. In a very short summary, I am looking at continuous waveform files from lab experiments that are simulating intermediate-depth earthquakes (IDEQs). Most of my work is processing data, turning it into readable audio files, adding to SAC headers, locating and identifying events, and applying frequency filters to separate data from noise. Really I am just generally assisting in making the data more manageable. Meanwhile, Dr. Peng's lab is attempting to clarify whether these laboratory experiments are adequate representations of IDEQs. As the summer progresses and I learn to understand my data more, my goal is to be involved in this part of the process as well. Besides this, I am working on a side project, which is turning seismic data into extremely slowed-down videos that depict fractures forming while playing audio of the cracking. This really excites me because it's such a different way of viewing earthquakes, and a lot more intuitive than the typical squiggles. I think it will have very large educational applications.
The most important part of my job will be applying filters to distinguish noise from data. The laboratory experiments require the use of electronic equipment placed milimeters from the extremely sensitive sensors, so there is significant noise. Right now, I am learning about the different types of filters, how to automate them, and how to ensure the data I am left with is correct. This data is coming from lab work done at the University of Chicago. Because of this, I do not get to see the experiments happening. While I've seen lots of diagrams and read countless readme files outlining the experiment, it's still a little confusing. A funny moment was on Monday, when I asked Dr. Peng why two machines were placed closer to the sample, and the ensuing conversation confused both of us to the point he emailed the lab PI at U of C to ask. In Dr. Peng's words, "Qiushi (his grad student) and I had this exact same conversation yesterday and we determined that it made sense, but now that you ask it again I'm unsure." Here is a diagram of the experiment:
IRIS provided us with a self-reflection guide that covers various skills we can gain this summer, in order for us to evaluate ourselves before and after the internship. Looking at the list of skills, I believe the most important skill I can gain this summer is the ability to analyze and evaluate research results. I am good at processing data. Give me a project to turn one thing into something else and I will find a way to make it happen. (Hey, code doesn't have to be pretty, right?) Projects like these always have an end goal that I am aware of. I know what my completed product should look like or do. However, research at its core is about solving the unknown, and those in charge of the larger project don't have the luxury of knowing the end result—that's what they're trying to find. This summer, I would like to take a step towards this mentality. How do you approach a problem without any idea of its answer? Therefore, along with my data processing work, it is important for me to look forward to the potential results. I know I am not a lab PI and this isn't really my problem to worry about, but this is a critical component of research and one that I want my final poster to accurately reflect.
This week I started my internship in Dr. Peng's computational geophysics lab at Georgia Tech. On Sunday I made the 9-hour drive from STL to ATL, and moved into the apartment above the garage of my dad's house. I was a little nervous about this living arrangement at first, but it's actually been very nice. I am able to operate on my own schedule, but can still see my little brother and my dogs when I need to destress. The only thing I've had trouble with so far is waking up in the morning, but that's nothing new! Overall, the transition to Atlanta and working life has been easier than I thought it would be, if not perfect.
Due to covid, the entire lab seems to be in a state of transition. Most members are still remote, which has been a blessing and a curse for me. On one hand, it has given me lots of time to work with Dr. Peng one-on-one, and I have not been overwhelmed with new names and faces. On the other hand, however, I have not yet met any of the other lab members I will be working with. Also, the FIRE ALARM went off on my very first day!! As we walked out, Dr. Peng shared an anecdote with me about a previous fire alarm, when some professors got caught hiding behind their desks trying to continue getting work done, and every time I picture this I laugh. For my work in the lab, Dr. Peng introduced me to something he calls "Earthquake Music" which is basically a way of depicting the acoustic measurements his lab is recording. I find this so fascinating, since it's such an extreme break from the typical squiggly line graphs you see from seismometers, and it really emphasizes the power behind 'quakes! So far, I've been able to take some example Seismic Analysis Code (SAC) and transform it into readable .wav files. The first time I played the file on my computer, it was extremely quiet so I turned my volume all the way up and leaned in. All of a sudden, BANG!!!! the AE signal blared through my speakers. That's a mistake you definitely only make once. Other than that, this week has been about getting logged into the various servers, getting access to the building, and other administrative stuff. It's definitely been a slow start, especially as far as learning new concepts goes. However, that's largely because Dr. Peng is saving most of the project introductions for when the other IRIS intern, Mandy, gets here next week. I'm very glad I ended up in one of the sites with two interns, if only so it won't just be me learning everything for the first time around a group of seasoned seismologists. (Also I want someone to explore Atlanta with, so hurry up Mandy!!)
This is my first internship, so I want to use this chance to adapt to life outside of school. I am looking forward to better understanding what a job in this lab looks like, as well as just getting used to pulling a regular 9-5, packing my lunches, etc.
Second third of the internship:
After I've had some time to adjust to the lab setting, I want to drastically improve my coding. To be completely honest, when Dr. Peng was getting me set up with remote access to the servers using ssh, I had to look up how to open terminal on my mac....... The only real coding experience I have is with MATLAB, and even then it's more that I have memorized and can regurgitate simple commands. I want to understand why I do certian things, plus really know what they are doing.
Final third of the internship:
Finally, after having almost two months to get comfortable in the lab environment and with coding, I want to take on a bigger role in my project. By this point, I will be thinking about my AGU presentation, and this is something I want to be proud to present as my own. It's not enough to be given some instructions that I blindly follow; I want to own my research that I present. Therefore, for the final third I really want to be in control of where my project is going, and decide how to interpret and present my findings.
Some final, random goals:
Throughout the summer, I also want to talk to the grad students in the lab about their grad school experience. I want to make connections in this field and at Tech, and I want to understand how people decided their careers. I also want to become better at regulating my schedule, so that I get a healthy amount of sleep, eat at normal meal times, and have time for exercise.
That's all for this week, thanks for tuning in! Don't go anywhere - we'll be right back next week with more updates on the life of an IRIS intern, right after this commercial break.