Spatial scales of snow texture as indicator for snow class

Abstract

A method for automated and fast classification of snow texture would be useful for applications where snow structure must be quantified. Large numbers of field measurements were carried out on natural snow in order to investigate small-scale variations of the micro-penetration force. Snow characterization was done for snow from the Himalaya and the Alps, using a high-resolution snow penetrometer (SnowMicroPen). Measurements of snow resistance at equal intervals of 4 mm were geostatistically evaluated. The range parameter (correlation length, or CL) of penetration force was estimated for all major snow classes from the sample semivariogram. Average CL was lowest for new snow and highest for melt-freeze snow. For major snow classes, CL was found to increase with snow density. Ground-perpendicular and ground-parallel snow profiles were also obtained for homogeneous snow, and CL was estimated along these directions. New snow showed larger CL in the ground-parallel direction, and depth-hoar snow showed larger CL in the ground-perpendicular direction. Based on CL, the directional anisotropy was calculated. An attempt was also made to show the relationship between CL and texture index. The semivariogram was used to estimate the fractal dimension. Both CL and fractal dimension were found to be potential parameters to describe snow.