Daniel Bowden is a student at U. of California - San Diego currently completing his research at Stanford University under Simon Klemperer.
I am working on a joint-effort project between Simon Klemperer at Stanford University and a team at the USGS in Menlo Park. Several stations throughout the SF Bay Area monitor Magnetotelluric signals alongside a seismometer. The goal is to collect enough data over a many year period to determine any correlation between E&M anomalies and earthquakes, or plate tectonics in general. There is much debate over whether or not such a signal exists, and its our duty as good scientists to accurately and rigorously record data at multiple stations over multiple years -- whether or not any signal exists, we'll try to find it. Aside from spending time in the field helping with the installation and maintenance of these dual stations, I intend to design an coil to create my own magnetic field that might help test the accuracy and calibrate the magnetic coils already in place. Then, its all up to the wonders of signal analysis to see what kind of data is coming in!
Its been a few weeks since the last post! I've had a little time off: went on a long weekend backpacking trip with my girlfriend, and then last week was out sick for a couple days. With time to make that up, I'll be still around here for another 2 and a half weeks! Which is a relief, because there's still plenty to keep me busy. I've spent the last couple weeks in the following ways:
-- Working on my coil and circuitry. The frequency at which the current in my coil oscillates is controlled by an IC chip that's basically a glorified RC circuit. It actually has a resistor and capacitor on the front end. I wanted to be able to step through several different frequencies, and I'm doing that by changing the R value. (For those who know/care about electronics, I'm doing this with my own 'digital potentiometer': a 555 timer sends pulses to a counter, the value of the counter goes to a demultiplexer that adds resistors to the line as the count goes up.)
-- First full field testing of my equipment!! Went out to the field, set my coil on the ground above one of the sensors, and flipped it on. This was a 'practice run' so to speak. We wanted to determine the working frequency range, play with different variables such as the orientation, distance away, to include mag-foil shielding or not, etc.
-- With data I created on the records, I got to start playing with Fourier Transforms. I wrote up matlab script to show a whole block of data, and then "pick" the time ranges of interest, fourier transform it, plot it, save it, and move on. I put pick in quotes, because I got the idea and code from Alistair Harding's (Scripps prof.) codes to pick seismic arrival times.
Now, I have some things to tune up, ideas where I can make things more convenient. I hope to get out to all three of the Bay Area sites next week and throw my coil at each of the three component sensors. Finally, I'll have the data that I can bring together. I just have to hope it works out and was worth the effort!
First test of my coil this afternoon over at the USGS! Not without some troubles, but still its rewarding to see some kind of result for a solid few weeks of work. I've learned alot about basic circuitry and power amplification through this, if nothing else. I started the project with a couple fundamental misunderstandings of what an op-amp and transistor would / could / will do to my signal. Since the last blog though, I now have a coil... it runs between .5 and 2.5 volts with a resistance of about 2.5 ohms -- you can do this math, that's about an Amp of current at its peak. There's challenges ahead to keep me busy:
-Reducing noise in my created signal (designing a better circuit with shorter wires and more twisted pairs to cut down on self-induced pick-ups)
-Need to take it to the field and characterize responses given that my sensor is buried about a foot deep. What happens when we don't know exactly how deep, or even exactly where its buried!?
- Want to be able to automatically step through different frequencies. I've designed it to be variable by hand (turn a screw on the Pot out front), but we want to be able to automate changing through a series of known frequencies.
On the other side of things, still working on getting traces of a co-seismic signal. That is, when the ground is actually shaking, it should show a trace on our magnetic sensors ( think of it as a pseudo-seismometer ). There doesn't seem to be a strong correlation, "strong" being relative to the daily noise from other things such as our BART muni-system, so its still a little frustrating. We haven't given up hope though, we just need to know exactly what to look for. In this whole process I since learned the ins and outs of shell scripting to specify exactly which seismic events SOD should download, automating that process to speed up the search.
I'm starting to feel the pressure as a deadline is approaching. I know some of you are even closer to a finale, so I can understand the excitement / anxiousness going around. Despite things picking up here though, I'm excited to say I'm taking a couple days off at the end of this week to get some backpacking in with the girlfriend. First couple days off for me since March! (weekends and the IRIS orientation week not-included, though that was pretty fun)
Got to go to the electronics shop yesterday! I've designed my circuit, now I get to hope that it all works right...
To better explain: the time varying magnetic fields at our several MT-monitoring stations have been varying in time. That's not saying much. But! They show different long term trends at different locations -- So aside from normal changes in the earth's total magnetic field, there might be something to be learned about the plate tectonics here along the San Andreas Fault. While this is something I will be thinking about and dabbling in, my main goal right now (as discussed and set out by my advisors) is to make sure that our magnetic recording equipment is calibrated equally across the stations. To do that, I'm working on a circuit that will output a varying current at a given sine-wave frequency (low, 1-10Hz), then hooked up to a large coil to produce a magnetic field. Since my magnetic field is a known quantity (or will be known, once I test it), I can then see if the magnetic coils buried in the field are picking up the same thing.
Another possible method for this suggested by Simon, is to look at an incoming teleseism (from far away). Since shaking the ground moves our monitoring coils through the earth's magnetic field, we get a signal (a pseudo-seismometer). Since the teleseism should affect all of them the same, unlike all the rest of the data that comes from very local noise, we can find out if the stations' data are at least on the same page (same order of magnitude). We'll see! So far I have only just begun by playing with downloading seismic data both with breq_fast and SOD (with help from grad-students and our very own Kelsey).
The second week for me here is almost completed, and although things are pretty quiet here I've been keeping plenty busy! We finally got our matlab software GUI to download and convert all our MT data from Berkeley (magneto-telluric: refers to both magnetic and electric fields measured). Its been a pretty steep learning curve in UNIX and networking and public key cryptography. The network administrator for the crustal geophysics group is a grad student who, go figure, is out in the field (thanks again, Kelsey ; ) ). But one learns more by doing it themselves right? Then there's the actual signal analysis, which I've only begun to scratch the surface of... All the electric and magnetic data is buried in noise for the 22 hours every day that BART is running. BART is the local Bay Area transit system, which powers their trains by sending current on the 'third rail.' Go figure, sending huge amounts of electric current along the ground at high voltage without any shielding would throw off our highly sensitive measuring equipment (We're talking about hundreds of square kilometers of area that this affects!).
On another note, I've (hopefully) settled on ideas for a project that I can call my own over the summer. While the seismometers have been well calibrated and tested, the magnetic coils in the ground haven't been checked in awhile. MT equipment is very well developed and used in industry as an exploratory technique, but not as often permanently installed to look at long term trends. We fully understand how they work and how well they should respond to a local magnetic field, but we're not sure if some of the connections have degraded over time, nor do we understand all the finer points of its setup (ex: is it better to bury the coil, or mount it on a concrete slab? Both will measure local variations, but what to what magnitude in SI units?). So, I've been challenged to design an induction coil to create our own magnetic field -- start with a 9V car battery and make a low-freq (~1Hz) magnetic field. Then we take it to the sites and see if what is measured is what we expect. I'm excited by this project because it should provide a good range of activities, from designing basic circuitry, to field deployment, to computer analysis (it'll take a little work to filter out the field we create from all the other noise, and then compare magnitudes). And all very hands on, which I definitely appreciate!
On a more personal note, I'm having a good time settling in to the bay area! I've been getting to know my undergrad counterpart (a stanford undergrad working on the same project), as well as the other students in our offices. I'm staying with family, which is always nice, and my girlfriend is in the area working for the summer at a German summer school/camp. I signed up for membership at a local rock climbing gym, and also just picked up a used road-bike to help with the daily commute -- should be solid couple hours of exercise each day.
Tomorrow should be my first day really getting dirty out in the field, shovel in hand. We'll be driving up to the Marin Headlands (just north of the golden gate bridge). We need to pull up one of the buried magnetic coils there and swap it with a working one. Time to get to work...
Finals in San Diego are finished, and every personal belonging moved to the north. I reported for my first day yesterday at the USGS here in Menlo Park (right close to the Stanford Campus), and met the crew there. Darcy was very warm and welcoming, as was everyone I met. Looks like there's alot of different aspects to this whole project, so while Simon Klemperer is still my go-to advisor, I'll be going-to lots of other people thoughout the summer. The E&M half of things is rather underdeveloped as compared the huge world of seismology, but that just means I get to learn about litteraly EVERY step of the way from collection to processing from the very same people who developed it.
The project, to describe as simply as I can, is to compare E&M data collected by the magnetic coils and electrode pairs that we maintain to the seismic broadband station immediately adjacent (technically maintained by the Berkeley guys). The project tech, Aaron, showed me two of the sites yesterday, one right by Stanford and another close to Berkeley (have you ever been out to the BRIB station Rob?). So, in addition to whatever research aspect I choose to pursue, a couple days a week I'll be helping Aaron and the other undergad on the project out in the field fixing, improving, testing, etc. etc. A student at Berkeley just custom designed his own Analog to Digital converter specifically for our purposes -- no small task, and now it'll be a team effort to install and test the component (which is actually the main focus of my undergrad counterpart on the project, Henry). Turns out tech Aaron was a UCSD physics undergrad back in the day (exactly the same as myself), so we'll have plenty to talk about every time we get stuck in Bay Area traffic like we did yesterday. He'll be teaching me alot of about the electronics, and seeing as how he had the same professor for the same electronics class I just finished a week ago, he has a pretty good idea where my background is.
Right now I'm working with Unix and matlab... there's a program freshly written to help download the data off the Berkeley server that should be pretty straightforward and user-friendly, once its installed and running properly. But therein lies the trouble, or fun, depending on how you see it.
Goals for the summer:
First third: Learn everything I can about everything. This is not realistic in itself, but the mentality I'm taking on... Specifically:
-Matlab and Unix (something that will come out of necessity as I go rather than through tutorials)
-History of the project -- want to be able to clearly define its purpose, methods, and goals in my own words.
-How the equipment works. I'm very interested in how the seismographs and E&M stations actually measure a property of the earth and get it to a table of values that we can make graphs with. This is an approach I think every scientist who works with data should undestand. Maybe I won't ever use need to design my own digitizer, or ever have to rely on the basics of signal processing, but this is my opportunity to learn how it works nonetheless.
-Choose a more focused contribution I can bring to the table. I've been given a few options as to what I might hunker down with, one being a more computer focused analysis of data and another to designing my own coil w/ power source to help calibrate and test the equipment. I need to explore these different tracks and talk to the researchers more to understand what they all mean.
- Get to work. Make something. More focused project details to come (as soon as I know them, believe me!)
- Continue to help in the field with the maintenance of these sites and the field testing of the new digitizer.
- Regardless of which project I choose, I'll be pouring over the collected data both for my own practice with computers and for my own curiosity!
- Finish up the project, start forming it all into a poster that I bring to AGU. I've gotta represent this project as whole and the team working on it, so I'll be hard pressed to do well.
The orientation program here is almost coming to a close -- we've had quite a whirlwind of talks, labs, in-field examples, etc. Professors who are very specialized in their field have been here just to show us all the various avenues avaliable to us. Its been pretty interesting to observe the social dynamics of the week, as us interns grew more comfortable with eachother and with the instructors around us. At the moment our minds are all spinning with anticipation. The following is a download of thoughts, expectations, questions, anxieties:
Step 1 for me is to go finish finals at UCSD. I'm the only one of the group who still has school to think about, though I haven't been thinking about it much... I'm fairly confident that my preparation in the weeks before will set me up for a smooth finish (we'll just have to see). I also have some social goals, saying goodbye to those fellow seniors who are graduating on time.
Then its up to the bay area. I understand that my main advisor, Simon Klemperer, will be out of town for the first couple weeks (Kelsey is stealing him from me). That being said, I will be meeting my co-advisor, Darcy McPhee, and other mentors at the USGS. I'm really not sure what that will entail, what the work will be, what my role will be, etc. I do know directions to the USGS office. That's a good start. I also know that this program has been going awhile, and IRIS interns have been coming back year after year. Looking at their work and blogs gives me some idea of where my summer will go, but also just means the bar has been set (I will be aiming to raise it).
I also understand there's another Stanford student who will be working on the same project... I'm quite curious as to what our interaction will be. Will we be working together the entire time? What's his/her background? Will there be any way I can share what I've seen this week, or will he/she be helping me around?
There's alot I want to learn -- I hope to buff up my programming / matlab skills substantially. I'll also admit I'm pretty interested just in the electronics and process of how the equipment works. I'm ready to get hands on and finally apply what I've been learning about in physics, electronics, and geophys. My goal for the summer is to ask as many questions as possible, and get involved with as much as possible. In the past this mindset has even been a detriment to actual productivity on any one single topic. But how can someone be content with only one piece of the puzzle?! In general I prefer to thoroughly understand something (in this case, the equipment, data collection and processing, general theory and principles, etc) before jumping right into working with the pieces. Am I capable of understanding all of this? Is that mindset reasonable for "big-time" projects? I guess there's only way to find out, and that's to show up on the USGS front door in about a week.
Gotta run for now, career panel is starting! ~Daniel
First day. Here we are in New Mexico, getting everything up and running. Up up and away!
So throughout the summer this webpage/blog should reflect all progress, thoughts, experiences, frustrations, questions, etc. New Mexico is beautiful, the professors and Michael are great, and us interns are all pretty cool too.
Keeping it short for now, we've gotta get back to being oriented. More to come soon enough!