Dreaming About Data

Tuesday, August 4, 2020 - 1:34pm


London Bernier

To imagine my time with Henry’s Fork Foundation is coming to a close makes me a combination of incredulous, disappointed, grateful, and excited. How all those mixed emotions fit together reflects the incredible, and far too fleeting, experience I have gained from my internship this summer. Learning from and being immersed in such a high impact non-profit has taught me the power of genuine collaboration, both within and across organizations. I have also learned an incredible amount about watershed science and management. Beyond the watershed specific knowledge, I have been exposed to opportunities that have grown my ability to effectively communicate science and have actually begun to enjoy data analytics. In fact, my mentor, Christina, and I both admitted to dreaming about coding in R… So, as I count down the days until my project presentation and the end of my time at HFF, I am excited to take with me the many lessons learned, and the hope that I may be back some day.

Speaking of data analytics, I have been spending much of my time working through projects that align with Christina’s work on habitat-streamflow relationships. I started out working on streamflow data and diversion data, information on how much water is being diverted from the river for uses like irrigation or other beneficial uses. Within Christina’s study area, there are four major diversions that impact streamflow. You might have read in Rob’s water reports about low-flow indicators; a low-flow indicator is an established lowest streamflow target during irrigation that aims to minimize draft of Island Park Reservoir, while also providing sufficient flow for downstream uses. A low-flow indicator of 1000 cfs was set at the top of her reach at St. Anthony, but that 1000 cfs flow quickly became 300 cfs after it passed by these four major diversions. From her research, a new low-flow indicator of 350 cfs was established for Parker, the bottom of that reach, to account for those diversions. Christina wanted to understand over the past 40 years when the diversions begin and end during the irrigation season, how many days Parker fell below 350 cfs, and whether the days that fell below 350 cfs were consecutive or not. Through much trial and a lot of errors, I ended up extracting and manipulating data into a table to understand how many days Parker fell below 350 cfs, when those days started and ended, and how many total days were below 350 cfs. 

I jumped from this project to learning a new type of modeling with the help of Rob: time series analysis. Time series is a type of statistical analysis that looks at trends over time. The goal is to better understand the relationship between reach gains and diversions for four spatial and temporal scenarios. The Henry’s Fork loses and gains water from the Eastern Snake Plain Aquifer and we want to get a reach specific idea of any trends over time. I am looking at the Ashton to St. Anthony reach from 1978-2019, the Lower Watershed from 1978-2019, the Teton River from St. Anthony to its North and South Fork gages with modern data (2004-2019), and the Lower Watershed from 2004-2019. Essentially, we know there has been a significant watershed wide drop in reach gains and diversions from 1978 until now, approximately 200,000 acre-feet, and that drop occurred in mostly one big jump in 2000 to 2001. Our analysis will help us figure out how much of that 200,000 acre-feet drop in reach gains and diversions occurred in these different reaches.

A screenshot of RStudio, the environment used to write and run R code. The plot on the bottom right shows gains for the Ashton to St. Anthony Reach from 1798-2019. 

While challenging, these data analysis projects have allowed me to apply my previous experience with statistics to real-world situations. I have come to see that data and statistics support the very core of HFF and guide impactful science-based collaboration. Still not sure if my coding dreams are dreams or nightmares, though…

I could not have imagined a better way to spend my summer and am extremely grateful to have worked with the HFF team. This internship has solidified my desire to work in the non-profit sector, although HFF is going to be hard to beat. 

A day of data collection using the ADCP, the orange device in the boat.

This is the Trestle Bridge, located within Christina's study area. Compare the streamflow in this photo to the next to see how different stream habitat can be with low flows!

Christina and I were talking about how vastly different these two images are. She told me this change can happen in a matter of a day or two!

Ventured out on my favorite day hike for a solo adventure last weekend.