Removed Version 1 data because of a number of image processing problems that were fixed in Version 2 data. Version 2 data benefits from the following error corrections and other updates:

• RAMS software was modified to address the antenna pattern banding problem, which resulted in blocks with bright centers and dark edges; thus, a radial banding pattern across the Version 1 mosaic. This banding effect is greatly reduced in Version 2, although faint banding caused by drops in sigma-naught with decreasing incidence angle remains.



• ASF provided corrected noise floor vector data to reduce noise in low-backscatter areas.



• Ghosting was only a minor problem in the Version 1 mosaic, and was further reduced in Version 2. Ghosting is an effect of side-lobe detection of bright features. When the main antenna lobe is imaging very low-backscatter targets, it is possible that bright targets passing beneath a side-lobe of the antenna will scatter enough energy back toward the antenna to be detected and mapped erroneously at the location the main lobe was viewing. As a result, bright coastal features can appear as ghost images in regions of very dark snow near the South Pole. The problem has been addressed by selection of alternate passes without ghost images.



• Version 2 of the RAMP DEM was used in processing the Version 2 mosaic. Swath data were reprojected onto the new DEM surface, improving the georegistration of the mosaic.



• Shadow and layover problems were addressed by covering all mountainous areas with ST7 data, such as RADARSAT-1 data acquired using a different standard beam looking mode. ST7 data are radiometrically much darker then the ST2 data used for most of the mosaic, so their use resulted in a few radiometric artifacts.



• As with Version 2 of the RAMP DEM, Version 2 of the RAMP mosaic extends farther over the sea ice surrounding Antarctica.




The effects of Automatic Gain Control (AGC) used in initial data gathering by the satellite remain in the Version 2 images. In collecting the data, a region of each swath is sampled for intensity so that the gain of the processed data may be set optimally for increased contrast over the surface to be imaged. Within swaths that include abrupt changes in surface type, for example, near the coast or near rock outcrops, this initial sample may include only water or rock outcrop, and thus the gain is set for these surfaces rather than for ice. The section of the swath is processed with gains that are not optimal for the ice sheet mosaic. The affected regions of the ice sheet show up as slightly darker blocks or corners within the mosaic.