by Advik E. ’23
One of my favorite novels, Desmond Bagley’s “The Snow Tiger,” has an avalanche scientist protagonist who quotes the great skier Mathias Zdarsky: “Snow is not a wolf in sheep’s clothing — it is a tiger in lamb’s clothing.” And what was this “tiger” in reference to in the novel? Why, an avalanche, of course!
Avalanche forecasting is an extremely crucial tool to protect lives in mountainous environments, such as our very own Cascades. Sol P. and I hoped to supplement the forecasts of the Northwest Avalanche Forecasting Center (NWAC) by getting Swiss avalanche-forecasting model SNOWPACK to run off of meteorological outputs from the Weather Research and Forecasting Model (WRF) for the Pacific Northwest — something that, to our knowledge, had never been done before. This was a project originally begun by Sid H., and continued by Olivia W. and Zelia E. last year. This summer, Sol P. and I attempted to continue and expand upon their work.
While Sol worked on getting SNOWPACK to actually run, I focused on converting the outputs from WRF into the two input files needed for SNOWPACK — a .SMET (a time series of data for meteorological parameters such as temperature, incoming solar radiation, and precipitation), and a .SNO (information about the initial state of the snowpack). Though the Swiss developers had their own method of retrieving weather data, I wrote my own Python program that pulled this information from WRF datasets, for each of three locations on Mt. Baker we ran the model for.
A graph showing WRF’s forecasted snow height (m) for each of three weather stations on Mt. Baker during February 2021. This data, along with numerous other variables, was fed into SNOWPACK. There are still unresolved issues with this data, but our goal was just to get SNOWPACK to successfully run with WRF data.
However, some data we needed wasn’t present as available WRF outputs. In those cases, I had to get the data from alternative sources such as SNOTEL snow data observations. Once I curated the data, the program synthesized it into the file formats we needed, and I handed it off to Sol. Amazingly, I got the formatting of the files right and the model actually ran, sparking celebrations and joy. For the first time ever, the SNOWPACK model ran with WRF. Yes, many questions and issues still remain. However, even if we didn’t manage to tame the SNOWPACK tiger, we were still able to gain a greater understanding of it. Hopefully, our work this summer will enable students in future years to subdue that fearsome beast.