Monique Bernier

Potential of Ers-1 Sar Data for Snow Cover Monitoring

The aim of this study is to develop a methodology to map the snow cover parameters (extent, state, liquid water content, snow water equivalent) from an orbiting SAR sensor (ERS-1 and eventually RADARSAT) and to use the resulting information as input to an hydrological simulation model. During the winter of 1992, ERS-1 scenes were ac uired over an agricultural region located to the EAST of Qut%ec City, in Eastern Canada. Snow cover and soil state were characterized by ground measurements over a number of sites in order to interpret quantitatively the SAR data. The pro osed methodology is based on the results of a previous study &ne at INRS-Eau. Those results suggest that a model based on the thermal properties of the snow cover could be developed to estimate the Snow Water Equivalent (SWE) with SAR data using a minimum of ground data. The preliminary analysis of the ground data indicates that a relationship between the surface temperature of the soil and the thermal resistance of the snow cover, similar to that found before, can be obtained. However, it is expected that a specific relationship could be obtained between the signal ratio and the thermal resistance of the snowpack for each soil type. Preliminary results of the SAR data analysis will be presented at the Symposium.

Estimating snow cover at the sub-pixel level over the French Alps from VEGETATION images

With spectral bands in the visible, near and middle infrared, the VEGETATION sensor can be used to detect snow cover because of large differences between reflectance from snow covered and snowfree surfaces. At the same time, it allows separation between snow and clouds. As the pixel size is 1km x 1km, a VGT pixel may be partially covered by snow, particularly in Alpine areas, where snow may not be present in valleys lying at lower altitudes. Also, variation of reflectance due to differential sunlit effects as a function of slope and aspect, as well as bidirectional effects may be present in images. In this communication, we will first explain the methodology which has been developed first over rolling hills and plain areas in Southern Québec and later over mountainous areas in Lebanon. Application of this methodology to the French Alps is then presented with the procedure used to prepare full snow cover and snowfree reference maps over mountainous areas, together with proposed solutions to diminish the importance of differential sunlit effects on the estimated percentage of snow over on each VGT pixel.