Dynamics of snow in semi-arid mountains are poorly investigated despite the fact that snow may represent an important source of water for downstream populations especially in the spring and early summer. Data acquired by space-borne optical sensors (i.e. reflectance and derived snow indices) may be suitable for spatial and temporal monitoring of snow cover. However, due to prevailing terrain and climatic conditions, the use of satellite sensor data to monitor snow dynamics is not trivial over such regions. Snowfall as well as precipitation are characterized by strong space–time variability. Indeed snow can fall and melt within one week. Under such conditions, appropriate monitoring of snow dynamics requires space instruments that provide data with high spatial and high temporal resolutions. In this context we developed a new approach based on the combination of two types of instruments: low spatial and high temporal resolution (Syste`me Probatoire de l'Observation de la Terre (SPOT)- VEGETATION) and high spatial and low temporal resolution (Landsat Thematic Mapper). This new approach improves the relationship between snow index and snow area. The method is validated against snow maps derived from classification of high spatial resolution data on the Atlas range, in Morocco. It is then applied to a one-year series of SPOT-4 VEGETATION images allowing to derive a temporal snow cover profile at 1km spatial resolution over the entire Atlas. The yearly snow profile obtained is of great interest for hydrological modelling.