I can't believe it really, but this summer's over. I left Stanford last Monday and headed back to Houston to prepare for and work at Rice's freshmen orientation week. My project didn't exactly turn out the way I wanted it to. After doing all the splitting measurements and running them through an inversion written by a previous PhD student, the results were inconsistent, unlikely, and unconstrained :( So, I am struggling a little to write an abstract and poster that describes this result in a positive and informative way and also including more general information about the project (to take up space haha). It's going to be a lot harder to work on this away from Mairi and Simon but I will get it done.
Looking back on this summer and the goals I had for myself, I couldn't have asked for a better experience. I have learned so much about the research process, the amount of time it takes, the steps you have to go through, and that things don't always work out. One of my goals was to learn to use Splitlab and Matlab, and while I can't say I really used Matlab for anything, I am a Splitlab pro. Okay, maybe not pro but highly capable. In general, I feel like my knowledge of programming has improved, thanks in part to some generous lessons from other interns and grad students in my office. The field work was actually one of the best parts of the summer. Spending 3 weeks in Elko is not something I would like to do again any time soon but I did learn a lot and got the chance to bond through manual labor with the people I would be working with the rest of the summer. I learned how seismic equipment is installed, saw the amount of planning and work that goes into it, saw some beautiful places, fulfilled my fieldwork requirement, and learned how to change a tire (or a few).
I think the most valuable thing I learned this summer came from conversations with other grad students and interns about grad school. They gave me some great advice about choosing a grad school, taking the GRE, applying, meeting advisors, etc. That's something that I didn't have any clue about and to get insight from people who had just gone through the process and ended up at a great school like Stanford was very helpful. I also felt very reassured that this summer project, despite the fact that my results weren't groundbreaking, would be a huge advantage for me when I apply.
My goal for myself as of now is to continue to keep this project as a priority so that I can create a quality poster and abstract for AGU.
I would like to say thank you to Michael and IRIS for giving me this great learning opportunity, to Rob for all his help and hard work, and to Mairi Litherland and Dr. Simon Klemperer for their support and guidance. I can't wait to see everyone again in December!
Well I've been making some serious steps forward and some minor steps back. After hours of fiddling around with it and a few emails from the creator, we managed to crack the code on Splitlab. It turned out that to make it work I had to rename all the sac files into a specific format, search all the earthquakes in the CMT catalog within the same distance that I had requested on my sod file, and then associate the sac files I had with those earthquakes before I could look at the seismograms. After that, there was a minor issue with calculating the phase arrivals but that turned out to be a permissions issue with a java file. So, once I got them all read in, I started to do some actual splitting. As I previously said, a lot of the signals from this stations were very noisy and I couldn't actually see arrivals on about half of them. The ones that I did do were difficult to get good results. The splitting relies on the fact that SKS waves are totally radially polarized after they go through the core. This means that all of the energy is initially carried on the radial component of the wave. However, when it goes through the anisotropic zone, the wave is split into the fast and slow directions. Once the program calculates the best fit for the dt and phi, it tries to 'undo' the splitting. If the dt and phi are exactly right, the transverse component would be flat (no energy) and all the energy would be on the radial component, as it originally was. It was often impossible to achieve this with my data. Once corrected, the particle motion should change from being elliptical to linear (also, didn't look that great). I did the splitting anyway as best I could for the permanent station, ELK, in Elko. The results had some subtle trends: the azimuth of the fast direction seemed to be somewhere between -60 and -90 degrees. The dt values were significantly scattered and showed no obvious trends. We decided to increase the distance window of the earthquakes in the hopes of increasing the backazimuthal variation. It is necessary in two-layer shear waves splitting to have a wide variety of backazimuths because there are 4 (instead of 2) directions that you must avoid because they will lead to 0 splitting in that layer. We also finally figured out how to apply a filter in Splitlab. It was amazing how much a .05-.2 Hz filter improved the look of the data. In some cases it went from being completely unusable to a seismogram that produced a good split.
All in all, it took a long time doing things wrong to figure out how to do them right (or at least better). I guess that's how these things go, though. So tomorrow I am starting over on my ELK data to hopefully produce a more convincing result.
In other news, my friend from Rice came to visit me and we spent Saturday in San Francisco and Sunday in Santa Cruz. It was awesome to see her and despite some car trouble (our rental car broke and wouldn't accelerate..or reverse!) we had a great time. Here are some pics and I will try to put some pictures from Splitlab up tomorrow.
Example of a good splitting result. Note that the waveforms corrected for delay time line up nicely, the corrected radial and transverse waves show minimal energy on the transverse component, and the particle motion has gone from being elliptical to linear. Parameters are reasonably well-constrained.
The house from Full House!!!!
Golden Gate Bridge
This past week I finished making measurements on all my station data. Simon came back from his trip on Monday to check up on our progress and look at the results we had gotten so far. We decided that the rotation-correlation method I had been looking at primarily to make the measurements was suspect because all of the fast-axis directions fell along a line 45 degrees off of the null direction. We decided to switch over to using the Minimum Energy method, instead, which determines the splitting parameters by finding those that minimize the energy on the transverse wave upon correction. For me, this meant going through the splits I had already done and finding the ones that had a good ME result as well as RC result and using only those. This significantly reduced the amount of data I had to work with.
After I had refined the data for my 5 stations, we started to compare it to results obtained by other people in the region. In particular we compared it to a paper published by West, Fouch et. al (2009) that used data from the same stations (plus more) but only came up with a 1 layer anisotropic model. I plotted up their results quickly and made a rough conglomeration of my stations using Illustrator (Splitlab does not allow you to plot and model multiple stations) to compare the two. Our results appear to be similar, however their results look more like our Rotation-Correlation results than our Minimum Energy results. Simon asked me to prepare some figures and a short explanation to send to one of his former students, Kris Walker, who is apparently the expert on shear wave splitting. We want to get his opinion on whether our data is high enough quality to justify fitting a 2 layer model (fingers crossed he says yes).
In the meantime, Mairi has finished preparing the data from her stations (the ones we recently deployed in NV) and says that somewhere between 6-8 earthquakes occurred during the short time frame she has. We are working on getting the files in the right format so that I can perform splitting measurements to add to our data collection.
I can't believe it but I only have about a week left here. I don't know where the summer went but I am definitely feeling the pressure to produce a complete project.
Example of one possible fit for just one station (ELK). We are not sure yet if this model will fit the combined data from all the stations.
So it's been a very long week. Mairi and I stayed behind in Elko to finish 24 service runs. We definitely ran into some major problems: the first day by ourselves we got stuck in a huge mud slick (I maayy have been driving). After a good 45 minutes of shoving rocks under the tires, shoveling out the mud, and some just general tire spinning, we got a guy to pull us out. In the process we got a flat tire. We only got 2 stations done that day. The next day was more successful: we got 9 stations done but ended the day with two flat tires in the middle of nowhere. We drove on one flat to a tiny town, found 3 old guys drinking beer outside of their trailer home and paid them 15 dollars plus tip to fix our tire. In the end we finished all the stations and made the drive back to Stanford.
Wednesday started my actual work on my project. Simon gave me three different things to focus on: 1. Acquiring the data from a permanent station in Elko as well as from 4 transportable array stations to begin splitting measurements on. 2. Figure out how to use the Splitlab program to perform the splitting and ultimately assess the anisotropy in the region and 3. to attempt to open the RAMP data given to us by Glen Biasi at UNR that was acquired following the Wells earthquake. I started on the first two goals and ran into major issues with both. I gathered the station data using a SOD file which bbasically allows you to specify the criteria of the data you want such as the network name, station name, the distance of the event (in our case, 85-110deg is optimal for the splitting), the time frame of the event, magnitude, etc. Problems occured however when looking at the .sac files because the requested phases are not always picked (P & SKS) and the signal is often so noisy that it is difficult if not impossible to detect any earthquake arrival at all. Filtering proved to be little help. I have received a lot of help with this but as of now am completely stalled.
The second project, to begin to work in Splitlab, has also reached a dead end. Splitlab is a program that allows you to analyze seismic waves for splitting due to anisotropy. The main principle is that when a wave travels through an anisotropic medium, that is: one in which the crystals have a preferred orientation, the wave splits into two orthogonal parts: a fast direction and a slow direction. The fast component will arrive first and the delay time between the fast and the slow reflects the thickness of the anisotropic layer. This program searches through all the possible values of phi (the azimuthal orientation of the fast component) and dt so that the splitting is completely removed when it is back-calculated. The program displays the components corrected for anisotropy (they should coinside) the energy of the radial and transverse waves (transverse wave energy should be minimized when phi and dt are correct) and particle motion plot (should be linear after correction). I ran through the practice project with little problem but encountered major problems when I tried to load some of the .sac files I had downloaded from IRIS the problem being that they could not be read in. This program seems to be kind of shotty and it may be necessary to find a different one. As of today I am very frustrated at these dead ends...I need data for my project! BUT it's the weekend so I guess it'll have to wait until Monday. Happy 4th of July everyone!!
Today was our fourth day of deployment in the Ruby Mountain area. We have been working in teams of 3-4 and most teams get about 2 stations installed each day. Today was especially nice for my team as we got to install a station along the summit. The dirt road leading to our site was pretty treacherous, winding, steep, and very bumpy, but when we got to our site it was absolutely gorgeous (see pic below). It also had amazing soil that allowed for the easiest hole digging I've had so far. Installing a stations is fairly easy once you know what you're doing. So far it's been easiest to assemble the solar panel first and align it facing south and then plan out the layout of the site. When we can, we've been building our fence off of an existing fence but we usually make a triangle when this isn't impossible. Of course there is a lot of hole digging to be done (my least favorite part), and then a lot of protecting the cables and instruments from weather. We are doing a direct burial for this project which means that the sensor is not in any kind of vault. Instead, we place it in a long plastic bag and then put that bag with the sensor inside another nylon bag to protect it and then put the hole thing in the ground. Leveling and orienting the sensor is the trickiest and most important part. Actually today we messed up on one of our sites because the declination on our compass was set in the wrong direction and we weren't reading the compass correctly. As a result, we ended up orienting our sensor in the direction of magnetic north instead of true north. Still don't know if we are going to have to go back and dig it up or if they will just rotate the data (I hear this is a pain). Anyway, the finishing touch is to build the barbed wire fence. By the time you get to the fence, you really just want to get it over with, which may be why most of our fences are less than perfect. Also, I have never been so dirty in my entire life. I took a picture to document the filth (see below).
In other news, I had my 21st birthday in Elko, NV. We all went out to dinner as a group and I got the most gigantic piece of carrot cake I've ever seen in my entire life. We literally had to pass it around the table 10 times w/ everyone taking a bite in order to finish it... soooo good though! Then a few of us hit up the Elko downtown. Let's just say, you only turn 21 once so I couldn't really go easy just because I had to wake up at 6 am the next day to do physical labor in the desert. We stumbled (literally) into a bar that had karaoke and my graduate student advisor and I belted out Don't Stop Believin' to a bunch of biker men. My encore to that was "I Like Big Butts". I think they were pretty confused. More shenanigans took place but overall it was an epic night. The next day: not so epic. I pretty much died. While I managed to get up and dressed and head out with my team, I found the desert sun and my massive hangover didn't go well together so I had to sit out for some of the first station. Luckily my team had four people so it wasn't that bad. By 4 o'clock I recovered and helped with our last station.
More station installs tomorrow (and every day after that for a week) but also a free Uncle Kracker concert at the Elko motorcycle rally! Should be a good time!
Working on the action packer
The two people from PASSCAL, Greg and Derry, came on Saturday night and we had to immediately begin setting up the first 25 of our sensors for the huddle test. I learned how to connect the sensors, the DAS, the break-out box, the gps, the battery, and the power box and how to program it with the Clie. We are using the electrical systems and technology (or something like that) building at Great Basin College. We ran into some trouble when we realized that the gps wouldn't lock with the metal roof so we had to move all the equipment to just inside the garage door and run the gps cables underneath it. It was a little extra lifting but in the end it worked out. We connected everything and then went out to a GREAT dinner at a Basque restaurant (Elko has a lot of Basque restaurants which is strange). It was awesome family style dining and we had so much food leftover that some people took home 2 whole steaks in little doggie bags to eat for breakfast. After dinner it was 2 more hours of work to program all the DAS's. We had a slight set-back when we realized that all the Firmware had to be updated, but we ended up finishing around 11 PM.
The next morning we came in, took out all the disks and checked the data to make sure each unit was recording and the data was consistent. Everything looked good so we took down those 25 and set up the next 25. This morning we checked them out and found one bad sensor that was showing consistent little 'pings' on the signal so we had to send for another one from Socorro. We took everything down (by this time we had gotten pretty fast at it) and then went to lunch and then on an extreme shopping trip where we tested the limits of our Dodge. We hauled about 75 bags of sand, then got 260 fence posts, 8 bales of barbedwire, and of course went on another trip to Home Depot (I have come to hate that place with a passion).
The rest of the group got in this evening, mostly Stanford grads and a couple undergrads. We went out for pizza and then on a giant Smith's excursion for food. Tomorrow we go out as a group to two training sites so Derry and Greg can make sure we know what we're doing. Should be fun, but gotta get up bright and early!
Our more organized, huddle test #2.
So much has happened in the past week that it's hard to remember. But let me take a jog down memory lane and see what I find...
I got into San Jose, CA and almost immediately began feeling like an idiot. This was because I had mistakenly told Mairi, the grad student in charge of my project who had graciously volunteered to pick me up from the airport, that my flight got in at 2 pm when it actually got in at 4 pm. Oops! She was very nice and it luckily hadn't inconvenienced her too much so it ended up okay. Talking with her on the car ride to Stanford, I found out that we had a lot in common. Not only did we both go to Rice, but she also had doubts about going to grad school, whether or not she should get her PhD, and what kind of career and life she wants. It's been very helpful and encouraging to talk to someone who has a mentality that's similar to mine. She had arrranged for me to stay in her roommate's room for the 4 days before we had to drive out to Nevada to start setting up the project. I settled into my room and we had pizza that night with some of her friends.
The next day was when it began... supply shopping. With the help of the friendly Home Depot staff, we found out what "nut drivers" are and then literally bought out the whole store. We must have spent a solid hour and a half in there and spent about $500...and we still weren't done. We went to 2 more hardware stores after that! I had no idea how expensive it is to do a project like this. Monday we went to the Geophysics building and I met Dr. Klemperer. He is British, which I wasn't expecting, and seems very nice, which I was hoping for. Mairi informs me however (and I already got this vibe) that he can be very intimidating. I also laid to claim to one of the four identical desks in a room with no windows. I definitely got the best one, though. Sorry, Daniel.
On Tuesday night we went to a reception in San Francisco for a conference on the Himalayas that Dr. Klemperer invited us to. I was expecting an ugly hotel convention center with a bunch of people standing around shmoosing. Instead, it was at an amazing science museum that had a rainforest and an aquarium. We walked around looking at the exhibits and sampling the food, talked to some professors, and even though I wore painfully impractical shoes, it was a really fun night.
Yesterday, Mairi and I made the 9 hour drive out to Elko, NV. She drove the whole way since I am not allowed to drive until I turn 21 :( I felt bad but apparently she is used to making long drives. We stopped halfway to talk to a Dr. Viasse (?) at the University of Nevada in Reno. He gave us some dvds of data from 6 seismic stations near the Ruby Mountains that I will be using to study the anisotropy in the region. He also inspired us both to learn a c-shell language, Pearl in particular...I'll get right on that. When we finally got into Elko we found out the motel had lost our reservation and thanks to the mining convention in town this weekend, most places were full. The lady who lost our rez felt bad I guess because she called around town until she found us a room and then drove us to it so we didn't get lost (it was honestly 3 blocks away). In the end the motel was fine and we ate a late dinner at (a contestant for) one of the top 100 Chinese Restaurants in the Country!!!
Anyway, today we received the seismometers that were shipped from PASSCAL and of course had to check out the Home Depe in Elko. We looked into buying all the fence posts, barbed wire and sand needed but decided we should wait for some more muscle to arrive before we do that. After running errands all morning, we settled in to our quite nice arrangements at the Great Basin College, ate some Velveeta Shell's & Cheese, and watched Rush Hour. Tomorrow, a couple things to do but the real work doesn't start until everyone else arrives on Monday. I'm excited and nervous for the real work to begin. We are installing the stations in teams of 2-3 so I will definitely get to learn everything about the installation process. The Ruby Mountains are beautiful but pretttty snow-covered...might need to purchase a sweatshirt.
This week has been great so far. It's been helpful to learn about the very basics of seismology but i think the best part has been to hear about grad school and how successful earth scientists got where they are. I hope this continues throughout the week. New Mexican food is legit and ultimate frisbee in the dark is hard.