Author(s): Kjetil Melvold; Zelalem Mengistu; Nils Kristian Orthe
Linked Author(s):
Keywords: No Keywords
Abstract: During the 2008 and 2009 winter season we have used Earth Observation data to estimate thaw dates of a lakes with area between 100 km and 10 km in Norway. The primary aim of the study is to establish an operational data processing chain in order to monitor frozen lake or open water surface (OWS) for lakes larger then about 10 km. We have established a production line for near real time automated download of geocoded Radarsat (ScanSAR) images from KSAT and automated download and geocoding of Terra MODIS images (both reflectance and snow products). Multi-temporal information has been used to merge optical and radar data in order to fill gaps in optical or radar data acquisition. Due to significant differences in surface roughness of various types of land (land, snow on frozen lakes, glaciers, forest etc) the spectral signature or backscatter coefficient will responds differently in time and space. Therefore images for OWS study has been masked out toremove all the pixels that do not belong to the water classes. Different algorithms have been used in order to map OWS using the RADARSAT images. The initial results obtained from a study in the central southern Norway mountain area for the two years of data are promising. For the relative small and narrow lakes in Norway MODIS data with a spatial resolution of250m x 250m greatly improved the visually mapping of the presence or absence of ice cover (snow cover on top of ice) compared to the 500m x 500m resolution.
Year: 2010