To best represent my data, I chose a Schuster plot for the Geysers geothermal area because it was my first time in MATLAB writing my own code. It’s much easier to write code now and have done so for remote triggering, wellhead injection volume, etc. I used a program called Schuster and edited the time period intervals to fit my own analysis. I wrote a code that converted the earthquake matlabdate periods in hour periods on the x-axis., and looped the same code into all of the other sites I am studying. I used the title function, axis label, and the legend (which was complicated for a while). I need to read up on how to put in a caption, but I will explain the image below.
What this image shows is a range of time periods (in hours), and how well these periods correlate to the earthquake times in a catalog. For example, if earthquakes only occur in the morning on a 24 hour basis, you would see a huge spike at 24 hours, because in that situation earthquakes are correlated to the 24 hour period. What I am testing to see if earthquakes are more likely to occur when the earth tides are at their peak, which makes sense intuitively because peak times are when earth tides exert the greatest force on faults.
The Geysers catalog I’m using is from 2012-2014 and includes over 40,000 earthquakes in a small area around the Geysers geothermal field. The colored lines show different tidal periods at different times, and if there is any correlation between induced seismicity and earth tides, the periods at the colored lines should see a spike. The most important line is the black line M2 and represents the strongest tide at about 12.4 hours. The M2 does pass the 99% confidence interval, so according this this graph earth tides can trigger earthquakes.
For some reason the colored lines change back to gray after they pass the 99% confidence interval, so I need to still figure out why that happens.
This has been a good week so far. Writing code in matlab is becoming an easier task and I can do most of it on my own now. I am also learning how to make plots very neat so when the time comes for making my AGU poster, I will be ready for that. The 4th of July is tomorrow so I will probably see what my roommates are doing and plan something for the weekend. Last week I went to Newberry and saw an awesome shield volcano. Pics
Rounding out the 4th week, I have made progress analyzing the earthquake data in MATLAB. So far, the coding has been challenging in some respects, but I think I am getting the hang of it. This coding includes plotting the data on lat/lon maps, creating Schuster plots, filtering the data with respect to space and time, and filtering the data with respect to other earthquake locations.
The project overall coming along pretty well. I am spending extra time organizing data so I can quickly remember what exactly I did over the summer when time comes for AGU. In addition, the way I am organizing data is useful because I can create a loop in MATLAB that filters out all the data locations at once (I am studying different places around the country). This is a good way to save time when it comes to long coding blocks that can take hours to complete.
My living situation is great. There is a convenient Safeway only four blocks away from where I live. That store has surely been a life saver.
This week I met some other REU students for the materials science (physics and chemistry). Monday was their first day and all of them are in Oregon for the summer, at U of O. They invited me to take some field trips with them throughout the summer, so we’ll see how that goes.
This week we are getting into analyzing the earthquake data. Right now we will see if there is any correlation with earth tides triggering earthquakes. The data we are studying has been filtered based on the time the earthquakes occured over roughly a 30 year period. What we are expecting to see is some evidence earthquakes occur more frequently when the tides are strong (every 12 hours, depending on the tide we are studying) and plot this data on a Schuster plot. We have not factored in the pressure data from nearby injection wells, or fault orientations, but this will give us an idea if earthquakes can be triggered by earth tides.
Over the summer, I will be working with ten datasets that contain information on earthquakes that occurred in various places around the country (ten regions total). This information includes the magnitude of earthquakes, location of the earthquakes, time interval, fault orientation, and wellhead pressures from wells near the earthquake epicenters. The time interval in each of the regions is roughly from 2011-2013 when the data was gathered. Because of this, I don't have to start with raw data as different research articles have already compiled the data for me.
We had to contact some of the authors in order to get information we needed on some of the data if we could not find it within the paper. We will process these datasets in MATLAB to analyze any correlation between waste injection/hydraulic fracturing and induced seismicity. We will also use MATLAB to analyze tidal modulation of induced seismicity as well. GMT may be used for map making if needed.
My first entry is right after a matlab exercise using sine_waves and fourier transforms. Lunix was used instead of the familiar Windows 7.