Assessment of a Near-Global 30-meter Resolution DEM Derived from the Publicly Available SRTM Data Set for Use in Orthorectification of Satellite SAR Imagery

The Shuttle Radar Topography Mission (SRTM) utilized an interferometric synthetic aperture radar (InSAR) flown onboard the space shuttle Endeavour to obtain high resolution elevation data of Earth's land surface. Virtually all land surface between +/- 60 degrees latitude was mapped. Regions within these bounds contain some data gaps but this represents less than 0.2 % of the coverage. Standard publicly-available data sets from SRTM include a 3 arc-second (~90 meter) resolution Digital Elevation Model (DEM) with absolute average global vertical accuracy of approximately 4 to 5 meters. A 1 arc-second (~30 meter) resolution DEM has also been developed, but only the portion of the data set covering the United States is publicly available. The finished version of these products has been edited for pixel-level errors and delineation of coastlines and water bodies, although some data voids are still present. Utilizing such DEMs of appropriate resolution in a common framework with satellite synthetic aperture radar (SAR) data allows robust ortho-rectification and geo-referencing of the SAR data sets. We have derived a 1 arc-second resolution DEM over the entire domain of the SRTM coverage using a 3- dimensional interpolation scheme applied to the 3 arc-second SRTM DEM. Development of this product involves (1) translation of SRTM products into the WGS84 datum, (2) interpolation of the lower resolution DEMs to 1 arc- second, and (3) assembly of the global-scale 1 arc-second DEM. We assess effectiveness of this interpolation scheme through comparative statistical analysis of the 3 arc-second finished product, the 1 arc-second finished product, and the 1 arc-second interpolated product over selected test regions within the USA where all products are available. Comparisons are also made to standard GTOPO30 products for regions inside and outside of the USA. Comparisons are presented for regions representative of gentle and complex terrain. Ortho-rectification of SAR data such as those obtained from the Japanese Earth Resources Satellite (JERS) and ALOS PALSAR allows for an accurate representation of these data, providing crucial information accounting for effects of topography on geophysical retrievals. This work was carried out at the Jet Propulsion Laboratory, California Institute of Technology, under contract to the National Aeronautics and Space Administration.