Capturing the evolution of mountain snow in the Western U.S. using Synthetic Aperture Radar

Check out the narration of my AGU poster on this project here!

Background and Motivation

Knowledge of mountain snow is important for water supply and hazards. However, snowmelt is highly variable in time and space. Capturing this variation can improve water forecasting and knowledge of snowmelt variability over glaciers can help us improve our understanding of glacier mass balance and water delivered to the glacier bed.

Motivating question

What is the spatial and temporal variability of snowmelt timing on mountain glaciers and in mountain environments?

Methods

This work focuses on capturing the spatial and temporal variability of snowmelt timing in mountain environments using Synthetic Aperture Radar (SAR). We’ve built our analysis on scalable tools for future time series comparisons between similar mountain environments. To capture snowmelt runoff onset, we calculate the minima of a SAR backscatter time series as described in Marin et. al 2020. We estimate the spatial variability of snowmelt by calculating these minima pixel by pixel. Finally, we investigate controls of elevation and a diurnal anisotropic heating index on snowmelt timing using regression analysis.

Results

So far, we’ve identified SAR as a powerful tool for assessing the evolution of mountain snow, and with even greater potential when paired with optical data. Elevation and diurnal anisotropic heating index alone can explain 33.4% of the variance in SAR derived runoff onset dates. For every 100 meters of elevation gain, the runoff onset date at that location is predicted 4.28 days later.

Figure 1: SAR backscatter derived runoff onset day of year (DOY) estimates with vegetation masked out using a Normalized Difference Vegetation Index (NDVI) with a threshold of 0.2. The darker blues indicate a runoff date early in the year, and the bright yellows indicate a runoff date late in the year.

Funding

I am supported on a NSF GRFP fellowship.

References