Interferogram quality versus digital elevation model characteristics

Discusses the use of radar remote sensing for the derivation of land surface topography. CNES has always been in favor of using Digital Elevation Models (DEMs), when available, for the production of topography free interferograms. The reason is that they are available on numerous world wide areas, and that they allow the direct production of an interpretable interferogram from 2 data takes. Any differential method will always require 3 data takes at minimum, and will put further requirements on them. DEM types are numerous and can be defined by the following characteristics: its reference Earth ellipsoid-its type of coordinates-so called “geographical”: each point is referenced by its latitude and longitude-so called “cartographic”: each point is expressed in meters, representing the projection of the terrain on a geometrical surface (e.g. UTM and Lambert projections)-its mesh size: DEMs are data files, organized in records of fixed data point intervals. The DEMs used until now at CNES have had data point intervals ranging from 20 m (SPOT satellite generated DEMs) to 200 m (DEMs generated from large scale maps). Although one cannot quantify the relation between the DEM mesh size and the elevation precision, it is obvious that narrow cell DEMs always have a more accurate elevation than large mesh size ones