Combining ENVISAT ASAR and MODIS data to enable improved snow cover maps

Abstract

Information about snow covered area is important for several purposes, and this information can be found by detecting reflection of optical waves using optical sensors or by using active radars such as SAR. This thesis is explaining how information from the measurements are used to make snow maps. Optival sensors cannot detect snow when the area is covered by clouds, and this is a problem in the melting season in Norway. Microwaves can penetrate clouds, and detect wet snow. It cannot detect dry snow, but we will make snow maps based on the probability of dry snow. When we combine snow maps made from an optical instrument (MODIS) and a SAR (ENVISAT ASAR) we are able to get improved snow cover maps in times when the optical instrument cannot see the surface because of clouds. We will study a moving average filter and the Kalman filter. These filters are used to be able to filter the information about the snow cover to get an estimate of the snow covered area for each day. When comparing the results from the moving average filter with the Kalman filter, the results indicate that the Kalman filter gives a better estimation based on the measurements. We will also study relation between measured snow cover and the topography, and also if there are any relation between land cover and snow cover. The results indicate that the data from the optical sensors is underestimating the snow cover in northern slopes. The results also indicates that the melting of snow is correlated to the elevation.