This was the final week of my internship and I was able to wrap everything up quite nicely. I finished writing up a paper about our findings with permeability as well as finishing up a coding project about the Gutenberg – Richter relationship. Working alongside Luca and Sophia, the other high school interns, has been incredibly rewarding. Our time in Salt Lake City, collaborating at the University of Utah under the guidance of Dr. Kevin McCormack, has truly been an amazing experience.
Throughout Monday, Tuesday, and Wednesday, our focus was primarily on finalizing the research paper. I dedicated a significant amount of time to tasks like unit conversion, algebraic manipulations, and creating graphs and spreadsheets. This was my first experience writing a scientific research paper and I found the process pretty enjoyable. One of the larger problems I experienced while writing the paper was the propagation of uncertainty. The fact that most of our data was in cm/min after calculating saturated hydraulic conductivity led to a significant number of conversions to get the data into centimeters squared after then calculating permeability using water’s viscosity in the form of Pa * s. Because of the units expanding into (kg*m*s)/(m^2*s^2) we needed the hydraulic conductivity to be in m/s in order for most of the units to cancel. I was not very familiar with the propagation of uncertainty, but it made a lot of sense as it seemed to be related with some linear algebra and orthogonal projections. Since every operation on a number with uncertainty affects it, this math was just a little tedious. While there several challenges I faced, this was probably the most annoying one – I discovered that I do not enjoy typing out math equations.
I just made a few changes to the code analyzing the effect of sample size on the Gutenberg – Richter relationship. The largest improvement was converting from geodetic (latitude, longitude, depth) to Cartesian coordinates (x, y, z). Originally I had just assumed the error was negligible due to the area of interest being small in comparison with Earth’s curvature. However, this seemed to completely change the positive linear correlation between the number of events vs sample volume. Rather than being scattered with a somewhat positive slope, the distribution became much more neat. While there was a bit of multivariable calculus in the conversion from spherical to cartesian coordinates, it was mostly based in basic trigonometry so it was very easy to understand.
Loading...
Reload document 
Open in new tab
Loading...
There are no comments published yet.