By exploiting the sensitivity of near-infrared snow reflectance to snow grain size, hyperspectral data were used to recover grain size using regression of Airborne Visible and Infrared Imaging Spectrometer (AVIRIS) observations and modeled spectra [ Nolin and Dozier, 1993]. Comparisons of satellite and ground measurements of snow reflectance highlight the variability in satellite-derived reflectance due to grain size and surface contamination [ Winther, 1992].
AVHRR data have been routinely used for binary classification of snow covered vs. snow-free area [ Xu et al., 1993]. Because of the low (1 km pixel size) resolution of AVHRR data, the estimation of snow-covered area would benefit from adoption of mixed-pixel image-processing methods. Such a method was implemented for AVIRIS imagery by Nolin et al. [1993], and for TM imagery by Davis et al. [1993 a].
Measurement of snow albedo is useful because of the importance of the snowpack energy budget in snowmelt calculations. Efforts to validate reflectances derived from TM imagery using ground-acquired spectrometry were carried out in the Canadian Arctic, highlighting the importance of solar zenith angle in determining reflectances [ Hall et al., 1992; Hall et al., 1993]. For off-nadir look angles in the forward-scattering direction, reflectances consistently exceeded 1.0 due to enhanced forward scattering. Because of the importance of bidirectional effects in reflectance of snow, spectral-reflectance curves and the nature of the directional anisotropy were found to be more useful in discriminating different snow properties. Hall et al. [1990] reported improvements in estimating the reflectance of snow-covered surfaces, which will be quite important in the global albedo estimates that should result from the planned NASA Earth Observing System (EOS) Moderate Resolution Imaging Spectrometer (MODIS). Spatial resolution and terrain both affect the calculation of surface energy budgets. TM imagery has been used in conjunction with digital-elevation data to derive surface albedo and brightness temperature. In mountainous terrain, irradiance calculations must account for reflection of nearby terrain features; calculations thus carried out yield irradiances within about 10% of pyranometer-measured irradiance [ Gratton et al., 1993]. Use of TM imagery to determine albedo in improved surface energy balance modeling was reported by Duguay [1993]. Improvements in radiative-transfer calculations were reported for the microwave [ Kuga et al., 1991] and for the visible and infrared, using both ground-based and airborne instruments [ Nolin et al., 1990]. The spectral reflectance of snow was modeled by Davis et al. [1993] and compared favorably with field observed reflectances.