Publication Date


Date of Final Oral Examination (Defense)


Type of Culminating Activity


Degree Title

Master of Science in Geophysics



Major Advisor

Hans-Peter Marshall, Ph.D.


John Bradford, Ph.D.


James P. McNamara, Ph.D.


Snow accounts for the majority of precipitation in many areas of the Western United States, and accurate measurements of the amount of water contained in the snowpack, known as snow water equivalent (SWE), are therefore important for water resource managers. The National Resources Conservation Service Snow Telemetry (SNOTEL) sites are the current standard remote measurement of SWE, with approximately 800 sites across the Western United States. Measurements at these sites are made by snow pillows, which weigh the overburden pressure of a snowpack, and are relatively expensive to install and maintain. Spring runoff is modeled using a 30-year average of SNOTEL SWE values, and recent years are increasingly diverging from the historical record as climate change impacts both the timing and amount of runoff. Additional measurements of in-situ SWE would increase model performance, but the current technology is several decades old and has limited range for site expansion. Radar has been proven to effectively measure SWE since the 1970s, but has not been developed as an operational sensor because the technology has been expensive and the data processing has not been developed for real-time applications necessary for remote sites.

This study applies a novel automatic processing algorithm, which inputs raw radar data and outputs SWE values available for transmission, to newly available hardware. The combination of automatic processing and new, high-resolution hardware allows radar to continuously measure SWE at remote sites, which have the potential to make radar the next generation of SWE sensor technology.

The accuracy of the radar was first determined by a series of focused, 1-2m radar profiles over subsequently excavated manual snow pits, with accuracy of 7% in SWE compared to manual measurements. A network of eight radars was deployed at remote sites in Idaho, Montana and Colorado. Three of the eight remotely deployed radars were located at sites with independent SWE or precipitation measurements: Bogus Basin SNOTEL, Banner Summit SNOTEL and Garden Mountain weather station. Automatically processed radar SWE values are compared to the traditional snow pillow SWE values at the Bogus Basin and Banner Summit SNOTEL sites, and to a precipitation gauge at the Garden Mountain weather station. Radar-derived SWE values were highly correlated with SNOTEL SWE values, as well as with the precipitation gauge values of water equivalent. The combination of new hardware and an automatic processing algorithm has proven that radar can be an effective sensor for remotely measuring SWE in a range of alpine snowpacks.