I will work with Dr. Joseph Byrnes this summer, in Flagstaff, Arizona, to use receiver functions to image shallow base fault zones, using the high frequencies captured from a short-term nodal array deployment. We are using Dr. Byrnes's code, which uses the rjMCMC approach, and focusing on the Ridgecrest earthquake data, August 2019, using the Southern California Earthquake Catalog (SCEC).
The research that I have been doing this summer with Dr. Byrnes has been running smoothly, aside from when the code wants to throw errors occasionally. We have been focusing on using high-frequency data collected from nodes to image shallow earth structures. The tests that I have run show that the rjMCMC method produces consistent results across different events, frequencies, and time windows. With this knowledge, we will be moving on to testing the rjMCMC method on the nodes to hopefully produce consistent results with teleseismic events.
This figure shows results from one of the local events that we ran through station B501 from the 3J nodal system. This event was one of many that showed a double polarity, the second positive peak around time 0.1 sec, which is a strong indicator of the possibility of a fault. To make sure that the double polarity was real we tested different time windows, and we can see that the receiver function was stable across the different windows. The only window that showed disruption was the far left one and it only had a time window of 0.75 sec, which we believe did not hold enough data.
Throughout the summer I have read many papers that deal with receiver functions and the methods that are used, one of the papers that has been helpful in this research is "Thickness and Structure of the martian crust from InSight data", by Brigitte Knapmeyer-Endrun. In this paper, they are using the same method that we are using in our research to create the receiver functions for Mars. This method helps when there are many unknown variables/ dimensions, which helped greatly in the InSight data from Mars where the crustal thickness is still being identified.
One challenge that I face is troubleshooting the code that I am working with. Since I did not get involved in coding until a couple of months ago I struggle with understanding some of the variables and how they run. I feel that I constantly have to ask my mentor how to fix it, which I know is okay, but it is still slightly frazzling every time it throws errors at me. I also struggle with just reminding myself that every failure/error is a learning experience and not something to get hung up on. One success that has occurred is that when the code does run smoothly we have been able to see consistency in the receiver functions of local events. This indicates that the receiver function method that we are working with, rjMCMC, is successful in the high-frequency and shallow depths, which it is not commonly used for.
For the map of my area of research that I created, I used MatLab. I listed the latitude and longitude of the station I am focusing on along with the latitudes and longitudes of the local events that are around the station with similar depths and locations. I then used the geoplot variable to plot the locations of the events and the stations, followed by adding limits to the boundaries of my figure and adding the geobase map topography. We are studying this area to see if we can image the shallow crustal structure and possibly identify shallow fault zones. The data set that we have was optimal for this as this is the Ridgecrest, California area, and during 2019 there was plenty of local seismicity occurring to be useful in our research.
For this summer I am working with Dr. Byrnes code in MatLab and we are working with the Southern California Earthquake Catalog, SCEC, for the year of 2019, as the original focus was on the nodes deployed during the Ridgecrest earthquakes in August and September of 2019, however, we are trying out other stations deployed in the surrounding area. I believe that this data set is widely used but we are trying to use this data to collect the high frequencies from the local earthquakes to image the shallow depths to see possible shallow fault zones, which I believe is not what the data set is usually used for. The strengths of this data set are that they are relatively quiet as they are further away from the bigger cities in Southern California, but they are still slightly noisy. This leads to the weakness of this data set, as we are trying to focus on the higher frequencies to image the shallow layers, it is harder to pick out the P waves at certain stations as most anthropogenic noise is ~20-45Hz and we are currently sampling our data within 1-40Hz. We are starting off with raw data and running it through the code but the rjMCMC method it is helping predict the p-wave arrival and showing us possible changes in the possible layer differences.
I think that the skill that I need to focus on from the mentoring rubric is producing a written explanation of the research that I am completing this summer. This is an important skill as this will be something I will use throughout the rest of my career, be it for grants, research papers, or progress write-ups to employers. To demonstrate proficiency in this area is by writing up a paper and it is understood by my peers but also by people outside our field and the general public. The writing should be informational but not overly saturated with jargon to where the general public gets lost and interested in the topic. This is important as this will allow us to share our information and help grow our scientific community, by sharing our findings, etc.
I started my internship on May 15, 2023, we began by doing some readings on what receiver functions were, as I have yet to take a formal seismology class, and how they would be used for this project. Then I was given the code that we will be working with for the summer to look over and familiarize myself with how it runs, but this is where I started to get nervous. My coding abilities are still in the very beginner stages and I had worked with Matlab a total of two times before this internship. So, this was a little daunting, but with help from my mentor and many questions later I started to get the overall concept of it. My goals for the rest of the summer are to fine-tune my coding abilities, practice my scientific writing, and complete the poster for AGU so that I can practice for the conference and also focus on my studies as this is my heaviest load of classes before I graduate.
First third of summer:
Understand how the code runs: do this by running and manipulating different parts of the code to see how the product has changed, keep notes of all the things that have been changed, and if it benefited the run time at all.
Begin to get desired results/ field work: still a little unsure of what to look for, but we will be selecting certain local events and running them through the code to try and identify possible shallow base fault zones. Will also be going to California to conduct some field observations.
Will compile all data and results and begin to create a poster for AGU.