An Independent School • Grades 5-12

by Tanvi G. ’22

My project in the Lakeside Summer Research Institute (LSRI) this summer was to analyze temperature readings from temperature sensor arrays that had been deployed on Mount Baker from July 2018 to July 2019. The goal of my project was to determine snow presence on the mountain as a function of altitude. Mount Baker is particularly vulnerable to climate change as it is at a high elevation, so data on snow presence will help us monitor the effects of climate change over time. I began the summer by understanding the work that Kimberly L. ‘21 did last summer. She had developed an algorithm that searched for low and stable temperatures based on thresholds for daily mean temperature and daily standard deviation. Kimberly chose appropriate thresholds for both, and I extended this idea by plotting all the data on a jointplot (see Figure 1) and determining thresholds for snow presence on Mount Baker from those boundaries (see Figure 2). The more I learned about snow insulation and scientific data analysis, the more I realized that this method was hand-wavy and the way I was determining thresholds was in ways arbitrary. 

So, in addition to gaining experience with data analysis in Python, I also had the opportunity to research existing data sets and research papers on a correlation between snow depth and standard deviation. I didn’t know if data existed with both soil temperature and snow depth data, so I started by looking at snow telemetry (SNOTEL) sites near Mount Baker to see if I could make a correlation across sites. In my internet searching, I was excited to find a series of SNOTEL sites that measured both soil temperature and snow depth. My next challenge was to understand the way the SNOTEL website was organizing its data and if there was a way for me to get the data tables as opposed to the graphs. I eventually found a page that would allow me to select the variables and years I wanted and download an Excel file. During this process, I had to learn what all the different variables meant and what their abbreviations were. For example, I was searching for snow depth to approximate the amount of snow, and thus, the snow insulation. But from my searching, I found that SNOTEL sites also measure snow water equivalent (SWE) which is the amount of liquid content you would get if you melted down the snow. After speaking with Dr. Town, I learned that this was a measurement of the snow density and was actually a better measurement of snow depth. 

I didn’t have the opportunity to create any plots with these data sets; however, I learned a lot just from the process of filling in the gaps in my knowledge as I researched. It is possible we might not find a correlation due to the limited accuracy of the sensors, but I hope the next LSRI student will be able to use these data sets to reach some conclusions. Thank you Dr. Town and my fellow LSRI Summer 2020 Cohort for an engaging and insightful research experience.

Figure 1: Jointplot of daily mean temperature and daily standard deviation of all Mount Baker data. NOTE: the entire temperature range and variability is not shown here. 

Figure 2: Days with snow at each of the 6 different Mount Baker sites from July 2018 to July 2019 using thresholds (-2.75, 2.75) °C for daily mean temperature and < 0.45 °C for daily standard deviation. The elevation of the summit of Mount Baker is 3286m.