Influence of Snow Temperature Interpolation Algorithm and Dielectric Mixing-Model Coefficient on Density and Liquid Water Content Determination in a Cold Seasonal Snow Pack

Abstract

The SNOWPOWER measurement method combining high- and low-frequency measurements has been tested under Canadian cold temperate climate for determination of liquid water content and density of a snow pack. Since, cold and shallow seasonal snow pack are very often found in northern environment, it was shown that both snow temperature and exponent coefficient of common dielectric mixing models have to be considered to improve measurement accuracy. Indeed, Looyenga’s mixing-model coefficient at the fourth order together with the weighted average snow temperature measured along the sensor cable and calculated by linear or and cubic spline interpolation provided the most accurate results and good correspondence with manual snow density and snow liquid water content measurements. The best accuracy for snow density shows a relative error of 0.70% and an absolute error of 2.5 kg/m−3 compared to the reliable reference measurements. The mean relative error on the liquid water content was 0.62% and the absolute error 0.15%. The SNOWPOWER sensor system seems to have promising chances to become an operational tool for large-scale monitoring of the snow cover properties.