This summer we will be looking at seismic noise data from the Alaskan Transportable Array, attempting to characterize and filter noise from different sources (wind, pressure, ground-freeze thaw) with a focus on noise originating from sea ice. This research will provide greater understanding of noise associated with sea ice, specifically with the convergence and breakup of sea ice.
Hey Sports Fans!
This week's exciting edition of Lucas's blog features a VIP, sneak peak at member content, yet to be released to the public -- Brag to all your friends about seeing such exclusive plots!
Below is a figure showing the ice concentration (red) and normalized seismic power (blue), in decibels, seen at the 1-2s period range for stations A21K, C36M, and Q23K. Sea Ice Concentration represents the ice concentration as measured by the NSIDC in the nearest 25km² grid point. The normalized power is the mean of daily power as reported by IRIS for 1-2s. Normalized power measures the amount of daily seismic noise relative to the stations normal noise levels. A21K and C36M are within close enough proximity to sea ice to experience noise attenuation. Q23K acts as a control for ordinary seismic noise levels.
Below there is a large anticorrelation between sea ice concentration and seismic noise for stations A21K and C36M. Ordinarily, seismic noise experiences a seasonal fluctuation in seismic noise with higher noise levels in the winter -- a trend seen at Q23K. However, this fluctuation is reversed at C36M and A21K, a reversal we believe is caused by the seasonal accrual of sea ice. This relationship is strongly exhibited in the figure.
Our work so far has closely followed the work of this paper written, in part, by one of our collaborators Rick:
Links between atmosphere, ocean, and cryosphere from two decades of microseism observations on the Antarctic Peninsula
Overview of Key points:
- explores the relationship between microseism noise and sea ice in Antarctica
- Antarctica has high amplitude waves and high frequency of storms, causing a very strong microseism signal
- They find significant attenuation of microseism noise by sea ice,
- with the strongest relationship in the 2.5-5s “short” microseism range
- They made correlation maps showing the correlation of sea ice in that location with the noise measured using 25km2 grid points from entire NSIDC dataset
- This paper was extremely helpful in identifying some of the first steps to take in our project, and we followed similar methods in the preliminary stages of our project.
Otherwise, to keep you all up to date on my lunch habits, I have moved on from tacos to focus my hunger on cheese, utilizing various mediums and pairings to consume as much of the dairy goodness as possible. Last week I consumed an entire wheel of brie and two blocks of cabot cheddar cheese.
With much cheese,
Hello my avid followers,
This week I made many matlab plots of daily ice concentration and daily average power spectral densities for different alaskan TA stations; these allow us to better examine the relationship between the two at different locations in Alaska. this went well once I had the code working for one station it was pretty easy to modify it for other stations. The only issue is that because we are looking at a lot of data my computer has diffculty handling it and matlab often crashes or causes everything to freeze - which has made me more mindful of saving my work often. We also had our first call with Rick Aster, whom I will be working with in colorado for a couple weeks later in the summer. It went well I think; he seemed to think I made good progress over the first few weeks, so that is good to hear. We also had a call with Alice Bradley, our collaborator from Williams College, who gave us a quick introduction to sea ice and several of the relevant processes for our project.
Oh and here is a map of Alaska and the arctic with the locations of the first three stations we have been looking at. Sea ice concentration is marked in yellow and the data for that is available from the NSIDC. To make this map I used a matlab package called arctic sea ice that we are using to get our daily ice concentration data from; this package made making this map super easy, only involving downoading the package and a few commands.
Now to explain why we are studying this area:
Essentially at seismic stations around the world there are specfic frequency bands, usually within the 2.5-20s period range, where there is a seasonal change in the power of seismic noise detected. These frequency bands are called microseisms and generally increase in power duing the winter months and decrease in the summer months. These power changes in the primary and secondary microseisms are likely caused by the effect of ocean waves colliding and reflecting off of continental coasts, as well as the pressure gradient changes on the sea floor due to wave setup. This explanation explains the seasonal power changes well because in winter there are generally more storms and higher amplitude waves.
However, in the arctic something strangee happens at microseism frequencies; rather than the ordinary increase in power that occurs during winter at nearly every other place in the world, there is a decrease. We believe this decrease in power is due to the attenuation of waves by the formation of sea ice, causing a drop off in power each year. The data from the Alaskan TA gives us a unique, high resolution image to investigate this process.
I spent some time making my project elevator speech this week, a short about minute-long description of my project this summer. I find this a useful exercise as it reminds of the bigger picture of what the research is about, as well as got me thinking about how I present myself to different people: my elevator speech changes depending on who I am talking to and how much they know about Seismology/sea ice. Someone recently asked me about what I was doing this summer and what my project was and I definitely modified it by emphasizing the broader impacts and consequences of the research, going less into the technical aspects. As such, I feel like I really need 3 or 4 elevator speeches geared toward different types of people.
I do feel like my project may shift depending on what direction I decide to take it in, so my pitch may evolve as I continue forward this summer. I think it is very important to be able to convey my work in a clear and concise way especially if I am in the sciences where things tend to be esoteric and highly specific.
In terms of actual research, I felt this week I really started doing actually useful tasks; the past few weeks were all about gaining the tools to work effectively. Now, I feel that I am applying those tools and getting somewhere with the project. I also work independently for the most part and then check in with Kasey for short bursts to show her what I'm doing. This system has made me think critically and make a plan for each day; I think it is a good system and teaching me to think ahead more so and decide what attainable goals are for the day.
Other than work this week, I went climbing, did a few nice runs where I stumbled my way to roosevelt Island, and got a flat tire on my bike back from work yesterday. Luckily, I fixed it and was able to ride home, but need to get a new tire as my current one now has a hole in it :(. Oh, and I only had 6 tacos this week!
.-.. ..- -.-. .- ... / --- ..- -!
This week was fun because a friend of mine, Marshall, from Williams came to DC to work on this project with me. He is working with a professor at Williams who is a collaborator with Kasey, so we got the chance to work together and explore how to access the datasets we are using.
We will primarily use data from the Alaskan Transportable Array, an extension of the Earthscope project, installing nearly 200 seismic stations in Alaska. Installs started in 2014 and continued until 2017, so there is data available as early as 2014 for some stations, but all stations have at least a year or more of data available and are continuing to collect.
For the most part the data is really great. Because it was an extension of the Lower 48 Transportable Array the installations were very methodical and experienced teams handled all the installs, resulting in mostly very high quality data.
However, like any dataset, there are issues. For one, bears seem to be quite curious about these stations and particularly enjoy digging up and chewing on cables, including the ones that connect the station to power…. Due to this curiosity and the harsh weather and exposure of Alaskan winters, there are some data gaps at particular stations.
Additionally, while most stations are quite uniform in setup, some of the early stations are different because they needed to test out different emplacements that would be suitable for Alaska. These differences include having different depths for the borehole; different casing types; and slightly different equipment (wind turbine, solar panels, hut style, etc.). As of now, we are unsure whether these differences will affect the noise signals we are studying.
To actually get the data, IRIS has many neat tools to do so that anyone can access. For quick looks at a station one can look at a variety of graphs through IRIS Mustang, which lets us look at PSDs, PDFs, Spectrograms, and Seismograms right in the browser. This is really helpful for quick analysis or tests of what data I might want to download.
For downloading the raw data there are several options: iris_fetch is a matlab package that lets you download data quite easily, but is not good at large quantities of data; downloading from one’s browser is also an option, but is somewhat monotonous; the best option I have found is to use matlab and the terminal curl command to automatically download the data to matlab.
I initially had some issues with this because windows is not linux based and it was a whole issue, but luckily I also have Ubuntu dual booted on my laptop, so I redownloaded Matlab on that partition and everything has worked swimmingly.
So far Kasey has been really helpful in giving me the tools I need to work with this data. This week I spent some time creating a rough plan of how I am going to approach my research project and then talked to Kasey about the plan and what steps were more/less important and what I should do first. I feel that with research the hardest skill to learn is what the next step is going to be. Once I have a task and know what I’m doing everything just chugs along, but once I finish and need to decide what is next, I am indecisive. Talking to Kasey about my ideas really helped me make a plan for at least the immediate future, and hopefully as I get further immersed into the project I will become better at deciding what my next task should be.
Other than work, I’ve been doing a lot of running around the DC area; a friend took me on a nice run in Rock Creek Park and on the Billy Goat (?) Trail. I’ve been going to a climbing gym a couple nights a week, gone mountain biking, and explored both the Natural History and Air and Space Museum — definitely would like to go back to the Natural History Museum though. Oh, and I’ve eaten A LOT of tacos, like more than one probably should...
Over and Out!
My first week in DC is going well! I explored several bike routes to work, allowing me to see more of the city; I found the building locker room after a several day search; and I ran along the canal path many times.
At work, I have been working on a few different tasks: reading research materials on the TA array stations and sea ice noise, accessing real time data on sea ice concentration from the NSIDC and plotting it up in matlab, and a data task.
The data task was the most fun so far; I was supposed to collect data on each of the 196 Alaskan TA Stations from a text file of written construction reports, so we would have a centralized excel file to keep track of important details of each station. I decided to write a program in python to sift through each construction report and gather the info we wanted into a csv file, and it worked!! Mostly…. Because the written report did not have a completely uniform format it did not work for all stations, so I did some manual entry, but it was significantly less painful than doing it all manually. Now, I have to look through photos of the stations and categorize them, a process which will need to be done manually... oh well.
Some goals for this summer:
Write up my AGU abstract, even though we won’t actually be done with the research
Third…. Third??/ Longer Term Goals:
I think my first and second third goals are concrete enough that I will know when I have accomplished them; my last third will probably take more self reflection, and will be something I need to decide whether I succeeded.
That’s all folks! ‘Til next time.