A Hyperspectral Method for Remotely Sensing the Grain Size of Snow

We have developed a robust, accurate inversion techare useful indicators of thermodynamic processes in the snowpack. Changes in snow grain size can help identify nique for estimating the grain size in a snowpack’s surface ice sheet surface features, such as melt areas, snow dunes, layer from imaging spectrometer data. Using a radiative and blue ice regions, and often indicate changes in snowtransfer model, the method relates an ice absorption feapack energy balance (Nolin and Stroeve, 1997). ture, centered at k51.03 lm, to the optically equivalent In this paper, we present an inversion technique that snow grain size. Because the interpretation is based on can retrieve quantitative estimates of snow grain size from the area—not depth—of the absorption feature scaled to near-infrared reflectance data, using the wavelengths around absolute reflectance, the method is insensitive to instrument the ice absorption feature centered at k51.03 lm. We apply noise and does not require a topographic correction. We the inversion method to airborne and in situ spectral reflectested the method using Airborne Visible/Infrared Imaging tance data, and we validate the results over a wide range Spectrometer data over the eastern Sierra Nevada, Califorof grain sizes using field spectrometer data and measurenia, and we validated it with a combination of ground-based ments of grain size from snow samples. Finally, we examine spectrometer data and grain size measurements. Elsethe robustness and error sources of the method, considervier Science Inc., 2000 ing snow conditions, topography, and instrument noise.