Effects of sampling density on estimations of snowpack characteristics

While remote sensing offers great potential for snowmelt modeling, many of the techniques being proposed need to be tested against ground truth (actual snow measurements), but how does one determine the number of measurements required? Unfortunately, the number may depend on the parameter being measured and on the variability of the snowpack. One of the problems in determining the density or intensity of measurements needed is the lack of validating information. In this study, samples obtained on a 30 by 30 meter grid over a 26 hectare wind swept basin in southwestern Idaho are used to determine the number of samples needed to adequately describe snowpack characteristics under conditions of extreme spatial and temporal variability. Estimates of snow covered area obtained from sample measurements are compared with estimates based on aerial photos. Estimates of snow water equivalent and snow depth are determined for a range of grid and sample spacings.