Estimation of the total amount of water stored as snow in a catchment area during the winter season is a major driver for successful modeling and managing of water resources as well as for accurate predictions of mass balances and changes thereof on glaciated areas. As a comprehensive measurement of the entire catchment is usually impossible, the main difficulty is to link scales. Point measurements of snow depth and density must be combined to estimate the distribution of snow water equivalent (SWE) in a slope, and various slopes are combined to estimate in the average amount of SWE in a catchment. However, especially in mountainous areas, wind redistribution in combination with variable precipitation and complex surface topography, reduce the representativeness of single point data of SWE to sometimes less than a few meters. Therefore, the estimated variability pattern will highly depend on the applied measurement grid and its spatial resolution. For the present study, we employed radar technology to increase the resolution of measurement points to tens of centimeters and less. These radar measurements were performed at three different locations: (i) a relatively low slope, high Alpine glacier in Tirol, Austria, (ii) a non glaciated, high Alpine site in SW Colorado, USA and (iii) a highly wind influenced middle elevation site in Idaho, USA. A regular grid of circles subdivides the respective measurement area in several parts. The variability patterns of the two-way travel time (TWT) of the radar signal are analyzed for each circle separately utilizing geostatistical methods. These patterns are compared with the results using different spatial resolutions and to the results of the respective probings in the circles. At site (i) the observed snow depths were very homogeneous on a scale of hundreds of meters, and the variability patterns of the radar data stay fairly constant and correspond well with the probings. Site (ii) and (iii), however, are characterized by high variabilities in snow depth on a relatively small spatial scale. Therefore, the variability pattern changed significantly with varying spatial resolutions and the probings don't correspond to the radar measurements.