Theresa Sawi

August 16th, 2017

My time is over at SMU, and I will reflect a bit on my experience.

Differences between being an intern and an undergrad:

1. A lot more individual attention is given to the intern. Just compare an undergraduate class of at least 30 versus a 6-person reasearch team. I had to get used to people looking over my shoulder as I worked, them asking me for frequent updates and specifics about what I was doing. This kept me more focused and aware of my tasks.

2. For an intern, learning is about being shown something once, taking the best notes you can, and then being able to do the task yourself. There’s no busy work like in undergrad classes; no exercises to develop skills, no exams. What can you do, how effectively, how clearly can you explain your choices – these are the metrics by which to evaluate one's skills.

3. Interns have more freedom. There aren’t hard deadlines for every single task– but we’re working in a group and you don’t want to be the one holding up progress.

4. Interns have more responsibility for their projects, and not knowing whether things will work out can be emotionally taxing. But it's also more exciting. As an undergrad, you know there is an answer to the problem given. As an intern, your findings could confirm or contradict previous studies, but as long as the science is sound, then you’re contributing something.

I personally enjoyed all of these changes.I hope returning as an undergrad I'll have a broader perspective of how my learning can be applied beyond means to get a good grade.

I've had a great experience at SMU. I feel that I can speak intelligently about an important scientific topic, that I can generate ideas for future research projects, and that I have a handle on the technical skills needed to carry out those projects. But most importantly for me, I feel closer to being in a scientific community with a common goal: to understand why and how injecting fluids underground induces earthquakes in some places (but not others) and to be able to suggest ways to make fluid injection safer and more reliable. Thank you IRIS for giving me this opportunity! Now to find a grad school - I'm accepting all applications to be my reserach advisor... lol. 

Some parting words :

Man muss das Unmöglich versuchen um das Möglich zu erreichern. = One must attempt the impossible to achieve the possible. (Herman Hesse). I encourage us all to attempt our impossible dreams.

~ Theresa

July 26th, 2017

Have you ever heard an earthquake? I haven't. Some people in Texas have claimed to have heard to hear earthquakes. One man I talked to said he thought someone had driven into his house. People probably aren't used to hearing their house shake, so I would guess that's why they say they can hear earthquakes. Earthquakes do make low vibrations, however, and those vibrations can be transposed into a pitch that people can hear. That's exactly what seismologist Alicia Hotovec-Ellis has done with the 2009 eruption of Redoubt volcano in Alaska. She captured what is called "gliding". As the shaking becomes more intense and high frequency, the pitch reaches a...uh...fever pitch (sorry) and then a hits a slight repose and then....boom. Listen for thineself:

https://www.wired.com/images_blogs/wiredscience/2013/07/volcano_1.mp3

Cool right? It's like some junkyard symphony. The audio and a complete article were published in Wire in 2013. Dr. Hotovec-Ellis's full paper from Journal of Volcanology and Geothermal Research can be found here.

But let's move from cool sounds right on to pretty pictures. I'm currently going through about six months of data, finding earthquakes and creating an earthquake catalog, which basically means looking at wiggles and choosing where the wiggle begins. There are thousands of earthquakes in this dataset, however, so we needed to automate the process by using "template matching". This is where we take a sample of waveforms and then find other waveforms that match the shape of those samples. Since many of these earthquakes are coming from the same fault, they have the very similar shape. This allows us to stack them and then pick the arrival times (that is, find the beginning of the wiggle), for all of the earthquakes at once.

Some aligned and stacked waveforms -- Wow, they stack nicely, it's quite satisfying!

So wiggly...

Here's GMT (generic mapping tool) generated map of the Wavefields experimental site and the earthquakes used for the sample group:

Yup! No legend, it's a work in progress. Stars of all colors are earthquakes (red ones were used in the sample for template matching), grey triangles and little blue dots are seismic stations, yellow circles are injection wells scaled to volume injected. North is up : P

Below is a graph of the injection volumes for the three years leading up to the Wavefields experiment. There's a data gap and some other funniness, but you get the gist. This is a humble graph, but making it involved writing a script that could download and read hundreds of annoyingly diverse PDF files...yeah this was a behemoth of a task, the magnitude of which is not quite expressed by the output. And, yes, the data gap and other funniness irk me quite a bit *snaps pencil*

Time here is wrapping up quickly, I hope to share preliminary results before long. Bis nächtes mal,

~Theresa

July 13th, 2017

We all know what an elevator speech is, yeah? You’re in an elevator with your superior and you have to pitch your idea to them in the time it takes to go from the 1st to the 15th floor…that sort of thing. We interns will be presenting our work in a conference attended by over 20,000 geoscientists this December, and having a clear and concise summary of our research will be vital to networking at this event. It's like speed dating for your career. Additionally, if we want to convince our political representatives that our research is worth supporting, we have three sentences to get our point across-- that is what is allowed to be left in a voicemail that will get logged by a staffer and tallied up at the end of the day. So, here it goes:

I’m studying induced seismicity in Oklahoma caused by wastewater injection that is associated with hydraulic fracking. These earthquakes have already injured people and likely caused millions of dollars’ worth of damage. We want to understand the processes by which these earthquakes are created in order to protect the safety of Oklahomans and reduce the risk of operating high-volume injection wells near populated areas.

This would be something for a policy maker. Note how I do not say that we are trying to shut down fracking. A lot of Oklahomans work in the fracking industry, and the state makes a lot of money from their operations. Opening with some line threatening to their industry is a good way to tank the conversation. I want to find common ground…oil companies don’t want to hurt people and destroy buildings; it makes them look bad and could be a big financial liability. Policy makers like protecting people and saving money. Everybody wins.

At a scientific conference, I would probably say something like:

I’m studying induced seismicity from wastewater injection in Oklahoma. We’ve been analyzing the Wavefields community dataset from a 6-month deployment of 18 broadband stations and 370 nodes in northern Oklahoma, and have found over XX thousand earthquakes, complete up to magnitude XX, through cross-correlation template matching. Using double-difference location, we have been able to image several previously unmapped faults in the vicinity of XX high-volume injection wells.

Something jargony like that. Scientists love jargon : )

Bis später meine Lieblings xo

June 22nd, 2017

In my research, I’ll be looking at data from a suite of seismic monitors that were deployed in northern Oklahoma for the last six months of 2016. I’m not the first to explore this data, known as the Wavefield community dataset. Actually, anybody can access it; it’s free and available to the public. Scientists at IRIS, the fine folks who facilitate research within the seismology community (and organization supporting my internship, full disclosure) will eventually be presenting preliminary findings from the Wavefield experiment.

But how did we get here? Oklahoma is known for many things (cattle drives, pioneers, professional sports and the Rodgers and Hammerstein musical, to name a few), but house-shaking earthquakes would normally be at the bottom of the list. And yet, earthquakes in Oklahoma made national headlines last year as magnitude 5.1 and 5.8 earthquakes rocked Fairview and Pawnee, respectively; two cities that had never experienced quakes that size. Scientists have concluded that recent seismicity in Oklahoma is caused by injecting wastewater produced by oil and gas extraction deep underground. Earthquakes have become so prevalent in Oklahoma that the state now holds the dubious honor of hosting the most magnitude 3 earthquakes in the US…. Sorry, California.

But before the ground started shaking in the Sooner State, their neighbor to the south began feeling unfamiliar and unwanted tremors. Dallas, Texas, where I now write this blog, experienced a magnitude 3.5 on January 5, 2015. Texas Rangers pitcher Derek Holland reportedly fell in the shower. Thankfully, he survived to tweet, “This earthquake crap needs to stop. #imfine.”

With windows breaking and house foundations cracking in numerous cities and small towns across Northern Texas, including a crack in one town’s city hall, scientists and oil company executives alike began realizing the magnitude (pun intended) of the problem and acted accordingly: scientists at Southern Methodist University (SMU) published a paper that concluded that, apart from the obvious correlations in time and space between wastewater injection and seismicity (injection rates increased 1000 fold in the years preceding the onset of seismicity, and earthquakes occur primarily near the wells), the underlying physics supported the hypothesis that high-volume, high-pressure injection can lubricate faults and make them slip. Oil executives also acted. They paid tens of thousands of dollars in campaign contributions to the officials elected to regulate them and hired (IMO) shills of seismologists to cast doubts upon the SMU scientists’ claims. I’m proud to say that I’m now working alongside the former of these two groups.

A riveting, Hollywood-ready story detailing the showdown between scientists and the Texas Railroad Commission (the agency charged with overseeing oil operations) was published in the Dallas Morning News last year. To me, the story parallels the fight between climate scientists and the government: broad scientific consensus, well-grounded reseach-based facts, and even common sense break like water on the brick facade of a government agency that is not-so-secretly bankrolled by oil companies and their interests.

But, as Texas residents now know, brick facades do not hold up in earthquakes.

The authors of the aforementioned article, Steve Thompson and Anna Kuchment, lay out the story beautifully. I’ve liberally borrowed fun facts from their piece, but please check it out yourself.  My favorite quotes are as follows:

"The fluids went down Hillcrest and took a right on Lovers Lane and then the pressures arrived at 12:07 on this part of the fault - we're never going to be able to do that,"

            -  Brian Stump, SMU scientist, on the oil company’s claims that SMU’s research was unsound.

"The only reason that criticism comes up is because people perceive there to be different interests between oil company executives… and the general public. That's a misconception."

            - Ryan Sitton, commissioner for the Texas Railroad Commission. Accepted over $700,000 in legal campaign contributions, mostly from oil and gas company executives. He plans to run for reelection in 2020.

Wish us luck,

~Theresa

PS. Another earthquake (magnitude 2.8) struck near the town of Azle, in north Texas yesterday.

June 8th, 2017

Greetings y'all from Dallas, Texas. I've almost finished up the first week of my summer internship at Southern Methodist University studying induced seismicity in Texas and Oklahoma.

Beautiful SMU campus:

Man-made earthquakes are actually not anything new in Texas, and in fact there's practically no debate about whether oil and gas operations cause earthquakes. As early as 1926 a town called Gooseville on the Gulf of Mexico experienced earthquakes when the land sunk after some oil drills began operating nearby (that's known as subsidence). The land sunk by over a meter, actually causing it to go below sea level. The state of Texas sued the oil company, figuring that, because the land was now underwater,  the oil company has lost "land rights" and should share the oil profits with the state. The courts disagrees, saying that because the subsidence was man-made, that the land still belonged to the oil company. Boom - 1926, the first legal precendence for human-induced seismicity. *drops geophone*

Anyway, about my internship. I've become a "Daily Analyst," which means that every morning I look at seismic recordings and try to find earthquakes that the autodetector missed -- may not sound exciting but I kind of love it. It's wild how full-circle this feels; I used to stare at waveforms for hours as a music engineer. A seismologist actually invented the software for autotune/pitch correction, so I guess the similarities are inescapable.

I might have to remix that magnitude 5.8 from Pawnee last year...

It feels great to be in a research group; my colleagues are incredibly nice and helpful (including one former IRIS intern!). I get my own desk in a lab, we have frequent group meetings and I have gotten one-on-one crash courses in bash shell scripting/command line (read: coding), database management (again, from command line), and Antelope, the software used to interpret seismic waveforms (also ran in part from command line!). All this and I've read about seven papers on earthquakes in Texas (and one on ambient noise seismology). I have a stack of 10 for Oklahoma (one is coauthored by my home institution research supervisor, I miss you Berkeley!). One of our stations went offline suddenly, so we'll be doing some emergency field work tomorrow. Broken antennae? Stolen solar panel? Rats in the wires? Taking all bets.

Oh, right, I'm supposed to list some goals for the summer:

1) I wanna rock command line. I just want to rock it.

2) The group wants to transition from Matlab to Python (these are coding languages), and so do I!! This summer would be a great opportunity to take some initiative and translate their code from our beloved Matlab to Python, the language of the immediate future.

3) I've never seen the Gulf of Mexico -- let's do that!

Until next time.