Etude de la variation spatio-temporelle du couvert nival par teledetection micro-ondes passives et validation du Modele Regional Canadien du Climat (MRCC) (French text)

The spatial/temporal variation of the snow cover is simulated by many models such as the Canadian Regional Climate Model (CRCM) developed by Université du Quebec à Montréal that simulates the snow water equivalent (SWE), a key parameter for hydrological cycle investigation. Better understanding of snow cover dynamics and the validation of these models suffer from the space observational record available and passive microwave satellite data and appears as a very useful tool for such an objective. The data are derived from the daily Special Sensor Microwave/Imager (SSM/I) data from the DMSP satellite series and also in the Northern Hemisphere Weekly Snow Cover and Ice Extent from the NSIDC in the EASE-Grid format. A threshold adapted for four vegetation density classes (Canadian Landcover 1km-resolution image) is applied to the normalized difference brightness temperature signal between 37GHz and 19GHz to extract snow cover. This satellite database is available since 1987, however, the analysis has been carried out between August 1992 and June 1995 over Eastern Canada corresponding the CRCM run simulation driven by NCEP atmospheric objective analysis. The results shows that the model is underestimating snow cover, the winter season tends to arrive later when compared to SSM/I and spring is arriving too early. For the study area, the CRCM is modelling snow cover with an error of 9.3%. Locally, this error can be as large as 30–40%. This thesis shows the potential of satellite microwave data for comprehensive spatial and temporal evaluation of model behavior.