About this project: The prompt was something I missed from home - and for the first few weeks that I was home for the holidays, I remember simply marveling at the rain. Every time we passed Stevens Creek Reservoir (pictured above), it felt like the water levels had increased. Driving through the mountains, we could hear the creek’s water rushing. So for this project, I decided to take a look at the creek, the reservoir and the South Bay’s rainfall from December 1, 2021 to January 15, 2022 to see how much it actually rained.
Even though it seemed to rain for many days, it turns out that most of the area’s rainfall came on a single day: December 13, when an atmospheric river brought 2.23 inches of rain to the area. Rainfall from the rest of the time period, all added together, contributed an additional 2.12 inches of rain.
That's about two inches more than the average December in the area gets, making this past December the 17th rainiest one in the last century.
Beginning on December 13, the rainfall filled Stevens Creek. A creek’s flow rate reflects how much water is moving from the watershed into the creek, with more water meaning faster flow. In the sections of Stevens Creek that ran above Stevens Creek Reservoir and above Highway 85, this meant that the flow rate went from nearly nothing to more than 100 cubic feet per second.
As with the creeks, the atmospheric river kicked off the initial increase in water levels in the Stevens Creek Reservoir. But water flowing in from Stevens Creek above the reservoir and more rainfall later in the month continued to fill the reservoir. Only 10% of the reservoir’s 3,138 acre-feet capacity was filled at the beginning of the month. But by January, it was more than 50% full.
To be fair, that means it’s still about half empty, but it was nice nonetheless to see previously dry areas of the reservoir submerged again. The reservoir's capacity peaked on January 4 at 53.7% full - its highest capacity since July 1, 2020, which is as far back as ValleyWater's online historical data goes.
This project started with reservoir data from ValleyWater's website. After experimenting with several ways of getting this data (like using the API, which turned out to have different fields than the website), I ended up using Selenium and BeautifulSoup to scrape the data on reservoir and the creek's sections into tables.
Because I couldn't find county-specific rainfall data, I turned to the National Weather Service for daily data on rainfall. This data came in PDFs by month, which I then copy-pasted into a spreadsheet for further cleaning. For example, precipitation in the National Weather Service data was sometimes indicated as “T,” which indicates a barely measurable trace of rain. These values were replaced with zeroes.
Finally, for visualization, I used ggplot2 in R for the static charts and Datawrapper for the interactive ones. All the code and data for this project is available on GitHub.