In Alpine regions the varying snow cover is of importance for men and nature. Remote sensing techniques allow studying quantitatively the spatial distribution and the changes of snow in time during the snowmelt season. A multitemporal sequence of high-resolution satellite images (Landsat-TM) has been acquired for the Rhône/Sion basin (3371 km2, 491-4634 m a.s.l.) in Switzerland for the snowmelt seasons 1985 and 1998. The images have been converted to snow cover maps. Each pixel of the maps is showing the affiliation to one of the categories: snow covered, transition zone (=50% snow covered) and snowfree (=aper). In addition, each pixel of areas obscured by clouds has been assigned in a postprocessing step to one of the above mentioned classes using a GIS based extrapolation algorithm. In a next step, for each pixel of a time series the date has been determined when it became snowfree. From those dates Snow Cover Duration Maps (SCDM) for the seasons 1985 and 1998 have been compiled. The maps show visually how long the individual pixels remained snow covered in the corresponding season. The SCDMs serve as a basic entity in a GIS allowing detailed analyses in the region with respect to snow reliability.