Get ready for a long last blog post. I have some serious reflecting and sharing to do.
I've spent my last week here sick. Most of campus has a cold, and in my attempts to spend time with as many people as I can in my last stretch of time here, it was inevitable that I too would get sick. Though it has slowed me down, it isn't the worst thing and I'm still smiling as much as ever. I had my final goodbye lunch with Rick, Nicole, and Rob today at the Rainbow Cafe (the same place we went on my first day - symbolic? definitely) to talk about the fall and the future. Rick gave me a signed copy of his Parameter Estimations and Inverse Problems textbook - a very thoughtful, and applicable, gift from the summer. Tonight and tomorrow I have all of my goodbyes with friends. Being a very sentimental person who is not the best at goodbyes, I've gotten cards and small silly gifts for people (example - a $4 giftcard to DQ for my friend Emily, just enough for a celebratory blizzard for the climbs I know she'll keep crushing after I leave). I am just so thankful for the people I've become good friends with - all motivated, hilarious, and incredibly kind people. They made my summer here unforgettable. All that is left is the packing. It will be too quick and easy, just a couple tubs of clothes and gear to shove into my car with my bike and crashpad, then on the open road.
Now to reflect on my project and this summer as an IRIS Intern --
We are supposed to evaluate our progress and reread our goals. During the first third, there was a lot of waiting. But I did feel like I had that time to understand the research and methodologies that I was going to pursue. I did not learn that much about different signals and characterizing them, but working with Nicole on OK earthquakes, though a different type completely, did help me understand what I would be doing. I created maps and tested my understanding with our communication project and writing the abstract for AGU with my advisor. I do feel like I have become familiar and proficient in my data processing, though I do still rely on Nicole for coding assistance and to talk through my plans. Similarly, I feel more confident about being self-guided and asking for help when I need it. I think I was kind of shocked into how necessary it is to develop that skill, especially when you are so new to research. I still feel it necessary to talk through a solid "to do list" with my advisor, but beyond that I enjoy being independent. I have certainly made some professional contacts, and know will only continue to make more. And finally, I have become much more informed about my post internship choices and feel much more confident about how I will pursue them.
Overall, I was a bit over-ambitious with hoping that everything would be done at the end and I would be reflecting and outlining papers. Definitely not the case. Along those lines, I will be continuing my research into the fall, completing my poster, focusing on my work with Marie Byrd Land, and then possibly beginning investigation into triggered icequakes if time allows (but if these 13 weeks are any indication, I doubt the 10+ weeks of fall quarter will be enough to do all I hope). Rick and I have also discussed writing up a paper from our summer research, which seems daunting right now with little produced. I think just using this technique to review a modern study, seeing it's application and ability to expand upon a catalog (especially one as patchy as in Antarctica), is quite impressive.
Reading some of the other intern's retrospective blog posts, we all seem to have a common feeling about our futures and grad school: it's a big decision. I think this summer really showed us how incredible, but also how much work, research is. Not just in processing data and all the steps along the way, but how research is not inherently easy or clearly laid out for you. You do struggle. And you ultimately do need to have a certain desire motivating you. Therefore, a PhD is a 6 year intense commitment that I know I am not ready to make. After talking to many people about pursuing a PhD, I've discovered it is something that you need to know you want. Neither of my parents went to college, so though academics became very important to me, I never thought of going as far as getting my PhD. I never dreamed about being Dr. Myers. And right now, it makes me much more nauseated than excited. It brings on some pretty undeniable stress and feelings of unease. That's a pretty clear sign it's not for me, at least not right now. However, I have loved it here. There are moments where I am thrilled by my research and the research being done by others, especially working with the NEIC. It's still intimidating but the excitment I feel is definitely real. And so, I have had two developments arise towards the end of my research this summer...
Dr. Harley Benz, my advisor at the NEIC/USGS, proposed that I return in February, after I graduate in December and take some time off for myself in January, for an internship with them. I would be continuing similar work to my summer research, but also be working on two new areas of data processing and multiple event location. I couldn't say no to this opportunity to get more experience, to work with Harley again (who has been just a riot and so supportive), and to return to Colorado so soon! Golden will be a fun place to live, smaller than Fort Collins, but with the same outdoor opportunities and only an hour and a half away from the wonderful friends I have made this summer. Paperwork is underway, and though there is a slight chance it could fall through with government budget, we are all quite confident this will fall into place!
I also met with my advisor at CSU, Dr. Rick Aster, about graduate school possibilities a couple weeks ago. After discussing my fears of diving straight into a PhD, he fully supported the option of going for Master's first. That can then lead to many choices - stopping after completing my MS, changing institutions for my PhD, or continuing on for my PhD at the same institution. Hearing his support, and being on the CSU campus as school started back up again, made me incredibly excited about grad school again. 3 years is much more manageable. So this fall I will be applying to a few schools and I am happy to say that CSU will be one of them! The opportunity to move back here next summer after my internship in Golden is something that appeals greatly to me. The area, the research, and the people are all things I know I want to be a part of. And the fact that this decision feels much more right, less daunting, makes me less nervous about my future.
And so it comes to a close. What is it they say about when one door closes...? I have never felt like the end of something could bring as much opportunity as this. I am so grateful and can't help wondering, is this the end or the beginning?
Where has the time gone?! Tuesday was the end of my 11th week. Two weeks from Wednesday I will be headed out on my drive home - with the specific pit stop at Arches National Park with one of my dear friends Katie. I told her if she bought a one way ticket to Denver that I would drive us back to Washington - good deal right?
As the last intern left, I'm finally understanding what they all talked about - it surely, and sadly, flies by at the end. Things have been chugging along as normal, spurts of intense work blended into the more constant stream of slow progress.
Nicole and I had a WebEx meeting with our advisor down at the NEIC Wednesday to talk about progress and plans for the next steps. It seems like it is always easier to make a list of things to do than to actually cross things off. It seems like we can always add to it, and that each component is actually made up of a small list itself.
Our current plan is as follows:
1. Use the velocity model and traveltime picks from Amanda's (Lough et al) Maire Byrd Land paper, and relocate her existing catalog useing BayesLoc. BayesLoc is based upon Bayesian statistics (I've been reading a lot about, and understanding much less), which involves probabilistic computations from the velocity model (travel time predictions), the phase picks, and the station data (all prior constraints) to generate a sample of locations consistent with observations while also evaluating the correctness of utilized models and providing error bars to those locations. Relocating her catalog provides a good benchmark for the events that I will be detecting with the subspace detector code.
2. Each YT station is processed with particular parameters. These were discussed in my last blog entry... I do not have too much to change about what I discussed. I did settle on using Amanda's filtering parameters so that this will be a more consistent comparison/extension of her project. Template length does indeed change with each station, with the distance from the source, so that is chosen individually. Correlation threshold, however, has been improved! We have been trying to create a more statistically robust reasoning behind our choice of threshold, and Harley has developed something that I think is quite wonderful! We are still comparing signal to noise with randomly generated, Gaussian noise, but we are looking more critically at the statistics of the distribution. I'm still tryign to figure out how to explain it (must mean I don't understand it all quite yet), but it's related to Gaussian distributions having a mean centered at zero and the number of standard distributions (sigma) to represent that distribution... That was a poor attempt, I'll try to improve my AGU. Anyway, looking at an example day of our data, it appears that 5 sigma is a statistically consistent choice that defines a threshold adding little noise, but picking up most signal. This will mean that we will want to look at the false alarm rate, when an event picked is truly noise. However we expect this to be quite small, and the work it would take to write a code to fix this is beyond our scope right now.
3. Use the subspace detector to find all detections for each of the 6 stations over Amanda's 2010-2011 time period. (Don't be fooled by the little I wrote on this step - this is big plug and chug, exciting step!)
4. All station results collated into a single file, new events are associated with the previous catalog to be relocated. Relocation again? Well yes! With many more events, we will want to see how these distribute across the locations we find in the first BayesLoc run. Associating the events will run them across each station to see at how many stations the same event is picked.
I should be getting Amanda's pick tables this afternoon, that way I can get everything ready for processing at each station, while also learning how to generate the inputs for BayesLoc. Harley keeps saying the program is "stupidly easy" - and it does appear that way, but it's always understanding the background that complicates things.
I'm getting pretty sad about leaving Fort Collins, and Colorado in general. I have definitely grown to love it here. Each week only got better and better. Highlights from these past two? Well I went to Rifle for an amazing clinic with Glenwood Climbing Guides and the author of The Vertical Mind on the mental side of climbing, managing to get match my hardest sport onsight from last season then immediately do my newest hardest onsight yet, all while in the company of some truly wonderful people. I know I will go back to Rifle for some more limestone climbing in the future. The next day I was back at Rocky Mountain National Park. I sent my second v6, which is my hardest bouldering grade thus far, with my good friends Jed and Emily. It was both of their first v6's and I was incredibly glad to be there to see them succeed. Emily and I also sent our first v7's this past weekend at the park - on a day that I didn't think I'd climb let alone send my hardest boulder problem yet. I adore climbing with her, another short crusher girl with my same goofy humor and positive attitude, competition free, just out to have fun - a friend for the ages. A new friend, Ryan, who came out to climb with us at the park, explained that days like these had to be celebrated with a DQ blizzard - so we all sat on the curb, devouring our well deserved Pumpkin Pie (Apple for Ryan) blizzards in the last bit of summer sun. This past weekend was also New Belgium's Tour de Fat, which we dressed up for and biked downtown with 25,000 other crazies for ample beer and dancing. The next day, I took myself out on some belated hikes near Fort Collins - Horsetooth Rock and Grayrock. Both of which were beautiful, and I was able to run most of each trail. A long, but invigorating day with gorgeous views (Grayrock's meadow trail pictured above). It all ended with a fun barbecue at my friend Mo's house. So many new climbing friends, all incredibly supportive and fun people - it makes me sad to leave when I feel like I only just met a lot of them... At least it won't be long before I see them again... (What does that mean? Stay tuned for the next blog post - one you won't want to miss!)
The ebb and flow of research continues. I'm coming to learn we are all trying to figure it out together, you begin by pursuing one path, then find something interesting and follow it, usually stopped by having to make some difficult decisions. The easy part is getting carried away, thinking about all the things you want to do with your data, once those vital steps have been taken. At a certain point, you have to just make a decision, and move on. What excatly do I mean? Well for my project, creating a catalog of the subspace detected events requires specific parameters to be chosen: filtering, correlation thresholds, template lengths, the reference event/configuration file of events, etc. Each of these change the amount of detections, the number of templates, and the correlation coefficient values - so it's a troublesome to try to weigh what is important, while also what is statistically relevant and proper representation of the data. What is worse is further complications by the fact that many of these values should not be set permanently, but change with the frequency of events, the station (distance from sites), etc.
We are trying to look at noise (blue) vs. event signal (red) to find where the best high and low frequency cut offs should be for filtering (roughly where they cross). Too small of a window and everything looks alike, too low and we miss the high frequency events, etc. Above is a plot of events added together for an average look at the signal to noise, and it appears the best range is aboiut 1-11 Hz, with some random peaks inbetween. For my eye, it is easiest to see events clearly around 2-4Hz filtering in gsac (a great tool for plotting seismic data). But I can't just pick what "looks best" I have to have a reason for why I choose the filtering I do. Honestly, I think I may use the filtering parameters from the Lough et al paper that I'm basing a lot of this primary research on (1.5-4Hz, 4 poles).
Correlation Threshold/Scaling Factor
So similar to looking at noise vs event relationships, Harley and Rick had been looking at empirical noise for determining a correlation threshold. This is the coefficient for which anything above we will consider a true, honest event, and for all below, is likely noise, or so lost in the noise that it will be hard to tell. Here are to images of days of activity during one of the swarms in Marie Byrd Land I'm looking at. The black is just correlation coefficients plotted with time, showing the highly active, many event day on 1/22 and then showing the much fewer distinct events on 2/11. The red is also correlation with time, but not of the raw data, instead Harley has developed the code to generate a random noise stream that still has the spectral shape of true data. We can see that it does fit the noise floor very well, only spiking up into the events occasionally. So this is pretty amazing - whereas before we were picking a correlation coefficient as a default for the whole time, this clearly shows that the threshold needs to change with time. The code will then run through, and for example, I set it to run 10 min of every hour - it finds the maximum value for the noise correlation coefficient, multiplies it by a scaling factor so that we set the threshold above the noise, and then applies that as the new threshold for the rest of the hour, recalculating at the beginning of each hour. This is especially important for the highly active days, or for farther, noisier stations. Seems pretty easy huh? What is there to decide? Well, I need to decide by what value I want the maximum threshold recorded to be multiplied - the "scaling factor." Too high and I will be missing what are real events, too low and I will risk taking in too much noise... Right now we are playing with a scaling factor of 1.5, but I think we can do lower.
Here, I am to decided the length of the templates. My stations are quite far from the source area, and so I must do a longer template. If I only did a couple seconds, almost anythign could be picked up as an "event," so I must think of incorporating the important parts that define an event as a true event - phase arrivals. It must be long enough to include both the P and S wave arrivals, which are shown above. Including noise before and into the coda is also important, because to have events highly correlate that are truly the same, the event would be correlated through until the coda dies back into the noise. I have chosen to not go to that far, including the important phase arrivals and some of the coda. I think this can be justified as sufficient, and I think I will be sticking with this length (at least one decision made!).
Finally, these past two weeks have been great fun outside! I have made it out to Rocky Mountain National Park to boulder a few times and highly recommend it as a must see for anyone passing through. There are hoards of tourists, and they love to ask about the "funny mattresses" we are carrying into the woods. My friend Jenny came to visit this past weekend, we climbed some sport at Clear Creek Canyon (shown above) and then got out to the park on Sunday. It was nice to have her around, even though it was a whirlwind weekend.
Only one month left! Hoping to start cranking things out with these decisions, and begin work on my poster soon. With the other interns mostly done, save a few, I hope I won't be blogging to no one. Regardless, I will be able to share everything at AGU in December!
Okay so I have to admit, when I found out that I was going to be working on Antarctica-related research, this is what popped into my head. I apologize ahead of time - this WILL get stuck in your head. I played it for Nicole this morning, and I think she may not help me for the rest of the day... That chorus is just so catchy! If (WHEN) I go down to Antarctica someday, this will no doubt be on my travel playlist. And along with Antarctica-themed updates, with Nicole looking at data on Mount Erebus, and another student in our computer lab looking at seismic reflection data in Antarctica, I felt like I had the right to edit the sign on the door for the summer...
So my research has finally picked up a bit. Still comes in waves of productivity, especially since both Rick and Harley are out of town this week. But I'm really excited about everything I've been exploring so far. And I definitely think that's the best way to describe this kind of research - exploration. You start pursuing one thing, then likely go over some hurdles as well as find some new aspects to your project that you didn't expect. For example, the focus of my research, at least to begin, is to review areas of two previously published papers: Lough et al 2013, "Seismic detection of an active subglacial magmatic complex in Marie Byrd Land, Antarctica" and Peng et al 2014 "Antarctic icequakse triggered by the 2010 Maule earthquake in Chile" (which may still be in the process of publication). I am starting my research looking at Amanda Lough's catalog of events (many of which she devotedly hand picked!), using the subspace detector to enhance the catalog with previously undetected events, and hopefully increase the characterization of these swarms past the 2010-2011 time frame she analyzed. I am excited to look at how the b-value changes with time (relationship between the number of small vs large events - a look at energy release), perhaps how templates change (how the source mechanism may be moving or changing), and go from there. The induced icequakes are also another big focus - especially with triggering being such a hot topic right now with fracking, waste water injection, and the increasing seismicity in the central US. However, like I said, there always seem to be exciting new aspects to your research that you don't see coming. For me, it was that these two papers did not end up being so disconnected. As I looked into the MBL data, trying to decide on filtering parameters, correlation thresholds, and all that nitty gritty - Rick and I stumbled upon something quite exciting. Let me show you some figures and explain...
So what do you see? Well you can see my red phase picks on one of the largest events of Amanda's catalog - but it looks pretty lost in that big thing right? What is that you ask? It's a teleseism! It's a large remote earthquake that has been picked up at station ST08 in Anatarctica! And what's the date? March 11, 2011. Sound familiar? It's the M9.0 Tohoku earthquake in Japan! Okay so now let's filter this a bit and try to find our event in there...
Now you can see a whole bunch of smaller events! They were lost in the tail of the teleseism! Rick and I had started with Amanda's largest events, and noticed the largest four all arrived in a two hour period on this day - when we took a closer look (or perhaps a broader one), we see that they are all in the tail of this teleseism. Are you thinking what we are thinking? Could this possibly be... TRIGGERING?! So we went to the next biggest event in her catalog on a different day - and it too appeared in a teleseism! So this got us thinking... Of course we can see that not all of these events occur in teleseisms. They are also not like the triggered icequakes, as Amanda's paper wonderfully describes that these are Deep Long Period earthquakes, typical of magmatic processes. So things just got a little more complicated, and a whole heck of a lot more interesting. There is no reason this couldn't be a swarm of magmatic activity, which started independent of remote triggering, but is susceptible to being "kicked" or excited by remote, large scale event triggering, taking advantage of the system when it is in high stress conditions and creating some of the larger events of the swarms, like we see with the March events.
So here I took a snapshot of a rough plot of global events against the event count figure from Amanda's study (Lough et al 2013). There are some interesting coinciding peaks that I definitely want to explore - April 2010, June 2010, October 2011, etc. And I can even look into more recent events and see if this triggering/swarm relationship continues or changes. Perhaps as the swarm is less prominant, it is more likely to be affected by triggering, whereas when the swarm is in full swing, like Jan/Feb 2010, it is less obviously affected by outside sources. Very exciting!
Now outside of internship stuff - two of my best friends have visited in the last two weeks! Highlights from both:
Zack and I climbed the Bastille Crack - a bit of an epic with thunder and rain, but with gorgeous clear skies at the top of the 5 pitch climb. Then we drove late at night down to Independence Pass, sleeping in a parking lot in Leadville for the night. Some great climbing and total shenanigans with friends down there. Sad to see my friend leave, but we already have plans to do some seriously climbing back home - finally getting up the Chief in Squamish (the rain has foiled that plan a couple times now...)
Rachel visited the following weekend. So wonderful to be reunited. We are often asked if we are sisters, when we don't quite look alike - but I think being smaller, athletic, continuously giggling girls who can both eat on par with most grown men, there is probably an unconscious attitude exuded by us that make us seem like not just sisters, but twins. We enjoyed a bit too much sun at a lake, some climbing down near Golden, lots of food, a ghost tour of Fort Collins with Nicole, and Rick and his family, some nice movies in bed, and just being together again. The highlight however was going to Boulder to watch the Iron Man finishers this past Sunday. Talk about inspiring! No shame - we both almost cried when the first woman crossed the finish line. Thank goodness for sunglasses. It sparked us to plan to do a half Iron Man next summer (I need to start swimming more...), and just had us taking off talking about our futures. This summer has really shown me how growing up can be a struggle, but it is also so exciting and full of wonderful experiences that you don't even know will happen yet.
This is belated. I wrote this on Friday, when I meant to post it, and I'll just pretend that it's still last week...
You know how I said I was hoping to have great things to share? Well... Not so much. But as my title alludes, there was at least one high point. Let me explain...
On Monday Nicole and I went to the NEIC for our normal day. Harley took us to another meeting, this time with the Colorado Geological Survey, where members of CGS met with seismologists of the NEIC, of CSU (my advisor), CU, and School of Mines to talk with the COGCC (Colorado Oil and Gas Conservation Commission) to talk about the current state of induced seismicity specific to Colorado. This meeting was pretty fascinating, mostly because it was the first time where I got to see the Oil and Gas side involved in the discussion, and let me tell you, they definitely are actively involved. The topics of the meeting were mostly about policy and what sorts of ideas can be implemented and pursued for the future: expanding the Colorado seismograph network by a great deal, instating a state seismologist to monitor, manage date, archive, and be readily available to respond rather than having multiple people with other responsibilities, and creating a more consistent advising network for the COGCC. I was impressed by how much they do want to understand what needs to happen, and having advice from those who really are proficient in the state seismology offers a great amount of substance behind the decisions that the COGCC makes and advocates. Nicole and I briefly explained our research, and for one the first times I felt quite confident explaining at least what I hope to do. I sincerely appreciate that Harley takes us to all of these meetings. It is nice to see what a job in earthquake hazard entails and meetings like these are a big part of it.
After the meeting, I had an hour of breakthrough (this is the elation part!) - I could look at my data! Harley wanted me to find an event that had been cataloged in my data stream to make sure that everything was looking correct. To stress just how patchy the intraplate Antarctic network is, there are only two events in the USGS archive for 2011 and just one for 2010, which is the time period I'm starting with on the POLENET network. Both of the 2011 events were close to only one station each, but the 2010 M4.2 was perfectly in the middle of four! I was able to find all three, picked P & S arrivals for the 2010 event and Harley showed me how to do a rough location with sac2eloc, comparing its location to the one posted in the archive. Then when we tried to carry it further in our processing, we arrived at some problems with how the data is loaded. Oh well. I had a sudden feeling of monumental success for that hour! I had another PhD student at CSU tell me this about research science: 95% percent of it is frustration and 5% is exultation. Perhaps a bit harsh, but definitely experienced some brief exultation!
Another big part of this week: AGU ABSTRACTS! Nicole and I had to finish ours by today so that we could submit them to USGS for internal review before going to AGU. Writing an abstract when you have no results does feel a bit weird, and the editing process was a bit of an intense shock - but it makes sense that Rick, my advisor, would be able to truly explain our methodology more completely than I. It was definitely a learning process, but with a great deal of his input we finished it pretty quickly. It feels pretty crazy to have it "out there" and done, but also really exciting. I'll likely be on three total abstracts - mine, Nicole's (been helping her with some data processing - which is definitely prepping me to be really efficient and competent when my own data is process-able), and likely another with Nicole doing an education experiment with kids (details to come!).
I'm in Golden again today... We started off meeting with Harley and Dan. When your advisor begins by saying "How attached are you to this data?" - you know something is up. We are all starting to realize that Antarctica does not provide the best data. This isn't surprising considering the conditions. But the NEIC is not used to such messy signals or intense gaps. As an international earthquake force, they are are usually able to get some good looking data. However, we did not quit, instead returned to "hand to hand combat" with the network and before lunch Harley and I managed to find a small earthquake, make a template, and run it through our program without any problems. Almost eerie how well it worked, that all throughout lunch I'm just hoping it's running smoothly by the time I get back. Dan took Nicole and I out for lunch, let us vent about research frustrations and gave us some much needed and appreciated guidance and advice. I'm realizing that most people have the same experiences with research, and are usually more than willing to help if asked. Maybe I've just met some really wonderful people at the NEIC, but I like to think it's a universal quality.
After lunch and a little waiting, suddenly my internship was back into a full swing of elation once more! WE HAVE CORRELATIONS (shown in the figure below)! Of course their correlation coefficient isn't wonderful with the template we chose, but they do show remarkable consistency between events. This is good news for my future work with the data! We were all giddy with excitement about seeing tangible results - we could even see the events decreasing in detection/correlation as we moved towards the end of this known swarm. In another part of the building, someone was working with my data on Hydra, attempting to find locations with events in the stream, and of course we get a phone call about the results while we are all excitedly looking at our correlations. The room goes silent... We watch Harley for a reaction as he says "Okay so you've found locations... Are they events or are they crap?" And he looks at me, grins and shouts "Emma, they're not crap!" Antarctica - you truly brought the goods today! I have an actual to do list for next week and am excited to begin. Of course, only after my weekend plans... Last night my best friend Zack came to town for a visit, we did a taster tray at Odell Brewery, got chicken and waffles at 415 (an amazing restaurant), and planned our weekend of climbing. This afternoon after work we will go straight to Eldorado Canyon and climb the Bastille Crack (super classic easy trad) then head down to Independence Pass for the rest of the weekend with some friends of mine from Fort Collins. I'm incredibly happy to be reunited with my favorite climbing partner - this weekend will be a celebration of all we've been through and our future adventures as friends. I will definitely post some pictures!
Last entry was a bit of a downer, so let's try to improve with this week's.
I still have not processed data. I know, shocking right?! However, this week I feel much more acceptance about that fact. We have uploaded it, and I can see it, taunting me, available to use, however, there are of course a few more things we have to do at the NEIC before I can dive in. Oh, and a new twist for you all - my abstract is no longer due on the the 30th, it's due next Friday. Yes, that's right, 1 week from today. And today is almost over.
BUT, I have a plan! I have done a lot of thinking about what I need and want to do with my data this past week from working with Nicole on the same processes in OK, from meetings we have had with our NEIC advisor Harley, Dan McNamara of the NEIC/USGS, and now talking with Rick. We were also assigned to make a powerpoint this week explaining our research to a hypothetical audiance of 15-18 year olds with only basic intro knowledge of earthquakes. Not having much research to talk about, it really made me reflect on the background of my project, what has been done and what I will do.
Since I have no real research leaps and bounds to discuss, I'll just tell you about highlights from my week...
Tuesday Nicole and I planned to go down to Golden for a day at the NEIC, so I left Monday night to meet my good friend Laura for an evening climb at Earth Treks, her local gym down there. It was incredible. This gym is really new, with an obviously insane budget for holds, and the sport climbing walls are both steep and long - perfect for an evening of girls crushing on some plastic. Tuesday at the NEIC was pretty mellow. Harley and I made plans for what data was priority to download next, and things began to make headway there. We went to lunch at the Sherpa House again - third time is not the charm. I think we need to try something new... However, Wes Thelen of the Hawaii Volcano Observatory joined us. A former student of University of Washington, it was really nice to talk to someone who spent some time in the northwest, knew my advisor Jackie, and knew the research that had made me interested in pursuing seismology in the first place (Kate Allstadt's research with icequakes on Rainier and Jackie's work with signals from rock-ice avalanches on Iliamna). Wes and I joked about how much these women enjoy and excel at studying these less than classical earthquake signals that I have a growing interest in - it's like the "one man's trash is another man's treasure" except with seismology... and that trash sounds like a harsh way to describe these events... because they truly are intriguing. I'm excited to delve more into icequakes with my own research this summer. The highlight of Tuesday though was Nicole and my meeting with Dan McNamara. He told us about all he'd been doing with OK earthquakes, doing multiple event relocation, as well as looking at b-values and swarm characterization with this program in Matlab called MapSeis. What he was doing revived my excitement for this research. I suddenly started thinking of all the things I could do with my data - how I could do a lot of the same things he did to understand swarms myself.
Wednesday was a pretty lackluster day at CSU for Nicole and I. Got some work done, but nothing exciting. I met a new friend for lunch, after which we were going to go climb in Poudre Canyon. However, it was drizzling and the idea of climbing on wet rock is not the most appealing. Yet, like it was meant to be, I locked my keys in my car, so we had to wait for AAA to come rescue me, and by the time that was done it was starting to get sunny! Off to the Canyon we went! We managed to get in 3 pitches of sport, then it poured. I mean really started raining. And I still had to clean the draws off our last climb. A really cool moderate 5.10 that I wanted to climb, but with water starting to pool in crimps, I opted to top rope it so we could get out of there faster and with no needless falls. So we finished our climb in the gym. But the craziness of that afternoon, and just getting out on the rock for those two pitches each made Wednesday one of my favorite days in Fort Collins so far.
Thursday, yesterday, we were back at the NEIC. In dresses. Yes. Nicole and I dressed up (literally. ha. ha.) for a meeting with the Colorado Earthquake Hazard Mitigation Council, where Dr. Anne Sheehan of CU Boulder gave a talk about the earthquakes in Greeley earlier this year. It was really interesting to see the methods that we are pursuing put to use with recent events, especially with events as controversial as induced seismicity. They are making some incredible policy decisions in that area, and it's exciting to see the development of these regulations occur in realtime with the methods Nicole and I are using! We also went to a talk on earthquake induced landslides, part of a summer seminar series at the USGS. After taking a landslides class this winter, and learning more about seismology, and the difficulties of research, I felt like I could understand a lot of what was being discussed, and the fellow USGS/NEIC interns and I talked about what we thought of the predictive programs being used.
All in all, a long post for a long, yet much improved, week. Don't have much substantial to show for it, but do have a new peace of mind. Let's hope next week goes off without a hitch and I will have lots to share.
Let's just start by saying this post is going to be really honest.
This week has been really hard for me.
I still have not been able to begin processing my data. It's not that I'm worried about time, I still have almost 10 weeks left and will likely be continuing work into the fall, but with an abstract looming, due in less than a month (a draft by July 30th for me because Rick will be gone the beginning of August), and with other interns being so far along with understanding their research, it's hard not to get incredibly discouraged and ultimately feel truly incompetent. I know I'm much earlier into my internship than the others, that we all have different backgrounds coming into this, and well, I would love to hear from others that this is a shared feeling, but it has been pretty difficult not knowing even the basics of coding, the upper division math that is the basis of my project's processing, and really how to do most things on my own. Nicole has been extremely helpful with learning code as well as the steps to using the processing code, and I do feel like I've learned a lot. But sitting around waiting has this nack of making you think a lot, and of course we usually go to thinking a lot about what we don't know, sinking into the struggle of collaborative scientific research. Because ultimately I've downloaded all my data, I am just waiting for it to be uploaded in the format/location we use for processing, and also for the processing code to be fixed. The general feeling is a bit of helplessness mixed with a lot of incompetence.
After talking on the phone with my wonderful advisor back home, Jackie Caplan-Auerbach, I got some reassurance that I'm not on the path to failure, and that I can still have a future in geophysics (yes, my week was so bad that it brought me to those thoughts). She reminded me that I was chosen for this internship for a reason. And I know that - perhaps that's also where my struggle is, because I know I am somoene who is very dedicated and determined when it comes to learning, and feels annoyance when there is a lack of progress or understanding. I still have a lot to learn, and that's okay. That's what grad school is for. That's what talking to your advisor is for. So Rick and I talked, and he was very understanding and wants to help refocus on what I don't know, taking this waiting time to really build up my background knowledge of this research in the nitty gritty details. I had read papers on past reserach of the area, of the techniques I will use, and got the jist, but to really understand it is taking a lot more work. Today I'm reading a linear algebra textbook (one that my advisor Rick wrote!), we have gone over the basics, talking about how they relate to seismology and tomography. It does feel good to start to understand some things, but it's dense stuff let me tell you! For someone who has only taken one class past calculus in college (multi-variable), and that was four years ago when I was a freshman, this is quite the leap. Anyway, we will discuss what I've read, any issues, and then take it further into subspace detection and singular value decomposition (which is what my research is based on). Harley is here from the NEIC to catch us all up on where we are with the projects, and hopefully we can all get on the same page about what I will do next.
Updates to come.
Today is the last day of my second week at CSU! As I'm starting to get more involved with my project here are some more details, the nitty gritty about the data and the processing that will be used...
The dataset that I will be working with this summer includes seismic recordings from various networks/stations throughout Antarctica. We are in the process of downloading data from several networks, each with it's own functioning time period: POLENET YT (2007 to present), Erebus ER (2002 to present), Southberg XV (2003 - 2006), AGAP ZM (2007 to 2012), as well as a few other stations that will give us data from time periods before those above. Each of these networks has many stations within them, and those then have many channels. As you can tell, there is just a whole heck of a lot of data! And as we began processing it, we realized that some of it was just too large, taking too long, and ultimately wasn't necessary for our studies. So within some networks we would only use data stations, not conditions stations (temperature, pressure, etc). We would also only select the highest sampling rate for each station, because that would give us the most complete dataset. What has been the greatest hurdle is making sure we are getting all the data from POLENET. Some of it is restricted and it's hard to tell if you've actually got all the data or are missing some. Therefore a lot of this data is available to the public, however some is not. I doubt we are the first people to look at most of this data, but perhaps some more recent POLENET information is less analyzed. The data is all raw as far as I'm aware. We wil be cross correlating events with an already made code by the NEIC/USGS, but I'm sure I will have to do some manual picking for alignment purposes.
Tools that I will use to do my summer research are many! Just within terminal I am using a lot of different programs, gsac, sac, pico, etc. I am also using Java a great deal, which is where we run our subspace detector coding and made our getfetch requests of the data. Matlab has been the best resource for making my basemap and interpreting the background seismicity so far with Gutenberg-Richter relations and the magnitude of completeness. I am sure it will continue to be a vital program - we even installed the student edition on my personal laptop. I'm excited to play around with it; maybe someday get as good at generating scripts as Nicole... GMT will definitely be needed later on for figures. But due to my lack of current enjoyment with that program, we will just say that and move on...
Outside of the internship, I have officially made friends! I guess it's not that surprising that many of the people I have met are also summer interns or part of an REU (Reserach Experiences for Undergraduates). We come from all over, France, Puerto Rico, throughout the US, etc etc and yet seem to understand the process we've all been going through this summer. Two of the students from France, Robin and Tony, invited me to go bouldering last Thursday at Carter Lake and then up to Vedauwoo, near Laramie, WY, for the weekend. I have already had such a blast with these two. Brushing up on my French, I have found people I can joke around with and also climb hard with. It was a fun weekend of getting lost, drinking wine, making ratatouille at our campsite on a tiny stove, accidently getting on a 12c/d sport route when we meant to be climbing the 11b next to it, learning the basics of dancing salsa from Tony, midnight roof crack bouldering, and even later dancing under the stars and a disco ball that their friend Ryan just happens to keep in his car for moments like these. The Brewer's Festival was this weekend, but I have to say that the adventures I had in Wyoming could not compare and I'm not too bummed I missed it. Instead, I am excited to get back up there to climb more.
Hoping to get some data to start playing around with in the next couple days. I need to feel like I'm making progress again... Soon enough! I need to not worry and just remind myself I am only just finishing week two, and I have plenty of time left!
Nicole and I went down to Golden, CO yesterday to visit the NEIC. I am still in the works of getting personal access to the building, but was quickly accepted to the group of wonderful college-aged interns working at the USGS/NEIC this summer. As residents of "FoCo" (Fort Collins), USGS intern Sam dubbed Golden "G Den," unpleased that there was no associated abbreviation with Golden. The intern office is very quiet and motivated, so Nicole and I forced them out of the office and out to lunch, something they had yet to do together. All nine of us went to the Sherpa House and had the most filling, and delicious $10 Nepali/Indian buffet. After seconds, and thirds for some, it was agreed that this lunch was infinitely better than any pb&j.
But now down to business! Harley Benz, whom Nicole is working with on Oklahoma data, started by giving us a tour of the basement computer rooms. I felt like I was staring up at many versions of the great black monolith in 2001: A Space Odyssey. Their computer setup is amazing, and so incredibly backed up against any failure. This is necessary since they are such a national and global force in earthquake response. Harley then showed us the difference between portable and permanent seismic stations. He then explained how they process and share earthquake information, the scary steps that go into publishing this vital information to hundreds of thousands of people with a click of a button; the White House, the press, the public, etc etc. It was pretty amazing to see how quickly they can provide so much information. Much of the data that I'd been introduced to in my geophysics classes at school, through the earthquakes.usgs.gov website, is processed, analyzed, reassessed, and published within tens of minutes. Very impressive to say the least. Harley explained this is quite the norm around the NEIC. Even with the big 7.9 in Alaska on Monday, he didn't have to even leave his office to handle this process. This kind of visit really made me want to continue my involvement with the NEIC even after this internship is over. What they do is truly amazing.
Now I didn't get much done at the NEIC. I did get some new event and station data for my basemap. But with the discovery of Bedmap2, GMT has been abandoned for this assignment (thank goodness), and I pursued putting all of my initial data onto this reference map. Bedmap2 is truly an amazing toolbox used through Matlab that makes easy maps of Antarctica. Who knew that the south pole was such a pesky place to plot.... Anyway! Here is my initial map! I wanted to add a legend for the symbols (Triangles are stations - magenta are POLENET and yellow are OTHER, Circles are past events), but since the symbols were ploted on top of the bedmap, it is likely a task to do later in Illustrator, as Nicole and Rick tell me.
We have also been given the okay to start downloading our data! So after I made my map for us and we bought a new 2TB external drive for all of data, we began the slow process of downloading it all. Station by station. Year by year. It's going to be slow. That's okay. We've got pandora open, a webinar to watch, and plenty to do. Our hopes are to finish the downloading by Friday, when we return to the NEIC, so that we can give it Harley to put it into a format that we can use in our codes. It's all very exciting to finally be producing things, like this silly little map and files of data for future processing. We will need to create a script to run the downloading even after we leave the lab today... We'll see how it goes!
This is my last day of my first week in Fort Collins, and I’d like to provide a bit more information about my internship, my goals, and what I’ve been up to so far!
Above you can read my project summary, which is essentially how I understand the project thus far. Unfortunately, I have yet to delve into this data myself, but have been reading up a lot on past research and the methodologies that we will be implementing, as well as practicing with those methods on other data and with Nicole on her research of the induced seismicity in Oklahoma in the past year.
Anyway, here are my goals for the summer, both general and specific...
My general goals for the internship:
- Become familiar, proficient, and largely independent in my data processing and program handling.
- Personally become more confident about being self-guided, trusting my intuition/ideas. While at the same time, be confident in asking questions and for guidance when needed.
- Make some professional contacts in the field of seismology so that I have a network of people who know who I am and what I am doing for support with my project as well as in future endeavors.
- Become more informed about my post internship choices from my advisor’s own experience/advice, as well as others who assist me this summer, and my own experiences; these include grad school, industry, research, etc.
More specific goals related to each third of my internship (which will last 13 weeks total, June 17th – September 17th!)
- The first third – I would like to begin with really obtaining a thorough understanding of the area (Antarctica), past resereach, and the methodologies that we will be using to reassess the area and develop upon that past resreach. With that regard, it seems appropriate to learn about different seismic signals/sources and characterizing them (earthquakes vs icequakes).
- The second third – With more of a background in the codes/methods I will be using, I hope to begin retrieving and preparing data for analysis, then applying those developed codes for event detection/catalogue creation. Hopefully I can create some maps and catalogs of data, as well as expand my understanding of my project in my abstract, all things that will help me prepare for AGU.
- Final third – At this point, my abstract for AGU should be finalized and sent in (deadline is August 6th online), and then assess and prioritize my time left for what results will be presented and need follow-up. This will largely be a time of final reflection and analysis, outlining and beginning scientific papers.
*These are tweaked from what my advisor has outlined for me, and I am certain that they will quickly blend together throughout the summer, as I already feel like I have become well informed about relevant publications and begun understanding and applying the use of subspace detectors and other correlation/detection techniques we will be using. Also, I will be continuing my internship for more than a month after AGU abstracts will be due, so I hope to be able to really process a lot of information, and provide substantial work towards this project.
So far, with a lack of data, I have been working to become familiar with the many programs I hadn’t used before, such as coding in Terminal, using bash and editing text files with pico, pulling in metadata, processing it with gsac and sac, and then taking it further with applications of the methodologies of subspace detectors we will be using in Java, scripting in Matlab, and presentation in GMT. I must say GMT has been the biggest struggle for me personally. It is difficult to figure even where to start. The Internet is a good resource for base projects that you can tweak, but there is very poor explanation for why someone made his or her code a specific way. I have made one successful map of Oklahoma and the earthquakes of the past few years in Nicole’s research area, as well as a histogram of the magnitudes. This was pretty easy compared to my next step - I will be attempting to understand the program more this week by generating a good base Antarctica map from a file my advisor gave me of the Ross Ice Shelf, expanding the fields over the entire continent so that I can later zoom in and use whatever specific areas I need for the results I find. I hope I can get a bit further with it than I did in my struggles last week. I know this all just takes time and repetitive application to truly become proficient. And as Nicole and my advisor remind me, that’s what this summer and being an intern are all about.
Tomorrow I head to the NEIC (National Earthquake Information Center) in Golden, CO with Nicole to meet with her advisor and my contact Harley Benz. He has finished the last bit of code we need to process data, and while I wait for government access and my data to be provided to me, I can at least continue practicing with Nicole’s Oklahoma data in the meantime. I am excited to go to the NEIC and get a look into what it’s like to work for a nationally, if not globally, influential place in the field of seismology.
As for personal experiences, I've essentially been embracing the solitude. Still not knowing many people here, I've taken it upon myself to really explore Fort Collins, especially the outdoors. Taking a walk downtown as well as biking, trail running, just laying out catching up on some fun reading (100 Years of Solitude - appropriate in the title at least?) at Horsetooth Reservoir, and enjoying a beer at Odell Brewery. I have met a few people at the school climbing gym, and am looking forward to climbing with them outside sometime this week.
More to come in the next few days. Hopefully I will have struggled through a good basemap of Antarctica to use, as well as a map with all of Antarctica's historic background sesimicity...
Anyway, something to look forward to: The Brewer's Festival is this weekend and sure to be a good time!
After driving about 20 hours from Bellingham, WA with a stop in Missoula, MT, I have finally made it to Fort Collins, CO for my summer internship! Numerous rest area stops, a couple nervous, fully leaned forward in my seat coastings off exit ramps to gas stations, about a dozen Snap Judgement and ClimbingBeta podcasts, and just about endless singing. I must say that the independent road trip may be a struggle, but it definitely is fun to see how you make it through.
I am officially on day 2 of the internship, and though maybe less nervous than I was initially, I am still very excited. The general explanation of my internship is that I am working with Dr. Rick Aster at Colorado State University and his PhD student Nicole McMahon assessing the seismicity of Antartica. I will also be working down in Golden, CO at the NEIC (National Earthquake Information Center) for this project. Though I haven't seen any of my actual data, I have already been thrown into immense amounts of learning with Nicole using Java, SAC, GMT, Bash, etc etc, practicing the sort of tasks I will be doing with data of West Antarctica. I have a lot to learn, but I'm very thankful that Nicole is eager to help me, will answer any of my silly questions, and is essentially learning along with me on many fronts. It makes the transition to a completely new place, where I know no one, and have no idea where anything is much less intimidating. We started my first day about with a quick chat, tour, and a little computer time before lunch at the Rainbow Cafe across campus. I can already tell that Fort Collins is presenting me with the wonderful challenge of trying a lot of amazing food and perhaps a beer or two - I accept. I'm excited to get to know my advisor and student mentor a bit more. They are both incredibly smart, but I can also tell they will be a lot of fun. We've already discussed climbing together (getting Nicole on the ropes for her second time, and I'm eager to check out Poudre Canyon and Lumpy Ridge), Nicole teaching me a thing or two about basketball (let it be known I'm 5' and very uncoordinated, think I'll learn to dunk?), and we've even discussed getting Rick to try Zumba sometime with us. Let's just say it's going to be a fun summer and we'll all have plenty to learn and enjoy! I will definitely keep you updated, hopefully with documentation of our joys and struggles.
Want to know more?! I will post more about my project, my goals, and my general thoughts about the academic side of this internship in my next blog entry at the end of this first week, and of course the nonacademic experiences that arise as I settle into the sunny "Fort Fun" - looking forward to maybe exploring some mountain biking trails this afternoon!
At the IRIS orientation for all of the interns, discussing our blog entries and how to stay in touch this summer.