Grain Size, Mammoth Mountain, 4/5/94
Grain Size, Mammoth Mountain, 4/5/94
Snowmelt, snow chemistry, and climate modeling all require input of spatial and temporal distributions of snow grain size. Detection of grain size and change in grain size permits characterization of the thermal state of the snowpack, which in turn facilitates estimates of the timing and distribution of snowmelt. Size distributions of the grains and pore spaces are critical for modeling solute movement and concentration enhancement. Grain size is the primary control on the spectrally integrated albedo of snow which, because it is the highest reflecting natural surface, is a crucial component in the global radiation budget.
Qualitative estimates of grain size were made by Dozier and Marks with the Landsat Thematic Mapper band 4 (0.78-0.90um), classifying into clean new snow and older metamorphosed snow. With the advent of AVIRIS came the spectral resolution necessary for quantitative estimates. Nolin developed a technique which exploits the grain size dependent absorption feature at wavelength=1.03um.
The depth of this absorption feature directly corresponds to the diameter of ice particles in the surface layer of the snowpack. However, a single band estimate is suspect due to detector noise and radiometric calibration noise. Assuming the reflectance spectrum noise to be approximately random, Gaussian noise, then calculating the area of the absorption feature rather than the band depth minimizes noise effects. Nolin calculated a continuum- scaled band depth (as per Clark and Roush, 1984) for each AVIRIS channel in the absorption feature and used the trapezoid rule to calculate the absorption feature area. Using a two stream model, reflectance values and modeled band areas are generated for a wide range of grain sizes, thus providing the inversion to optically equivalent grain size. The continuum- scaling of the band depth removes virtually all topographic dependence of quantities derived from the band depth. The opening image of this page is a product of applying Nolin's technique to an April 5, 1994 AVIRIS cube of Mammoth Mountain, CA.
Current work is under way by Rob Green to further refine grain size estimates from AVIRIS data and will be presented here soon.
* A. Nolin,1993,"Radiative heating in alpine snow",PhD thesis, University of California, Santa Barbara.
* R.N. Clark and T.L. Roush, 1984,"Reflectance spectroscopy: quantitative analysis techniques for remote sensing applications",Journal of Geophysical Research, v89,pp 6329-6340.