No formal ASTER GDEM validation has been performed over Antarctica, but it is evident that there are significant errors within the tiles throughout the Antarctic Peninsula region, introduced as a direct result of high reflectance and lack of features on snow-covered plateaus. Since conventional data correction techniques do not work in this region, a new approach was used to significantly improve the data set. Processing for the new DEM converted the GDEM into contours and removed the erroneous contours, producing a smooth and realistic new DEM from the remaining contours. This method was applied only to poor-quality regions of the GDEM, and the resulting DEMs were merged with the unaltered high-quality GDEM regions. See Figure 2 for the edited regions. The method is made possible by the fact that the high-artefact regions are those where the real-surface (as observed on the Landsat Image Mosaic of Antarctica [LIMA]) is smooth and has a low slope angle.

Processing Steps

ASTER GDEM tiles were downloaded from ASTER GDEM site and mosaicked according to each latitudinal degree across the Antarctic Peninsula. Each mosaic was projected onto a reference system suitable for minimizing distortions in scale and for preserving angles locally. In this case, Lambert Conformal Conic (LCC) projection was used with standard parallels and other parameters according to latitude. The subsequent methodology was then applied separately to each latitudinal degree raster between 63 degrees and 70 degrees South.

Using a Geographic Information System (GIS), contours were generated automatically from the GDEM at 20 m intervals. A new file was then created by digitizing around regions of erroneous contours, based on the GDEM stacking number file and manual corrections using visible band imagery. These "noisy" regions of the DEM were then resampled to 200 m to simplify and remove gross errors, and sinks in the DEM were filled and smoothed. Contours at 20 m intervals were then created for this generalized DEM.

In order to correct these contours a slope model was created and contours that fell within a slope angle of greater than 20 degrees were removed. Secondly, contours less than 1 km in length were deleted, as these came from spikes or pits in the DEM. The surface topography was reconstructed by interpolation between the remaining contours, which represented consistent elevation values between the anomalies. The corrected regions of the DEM were then mosaicked with the remaining high-quality regions to create a topographically consistent DEM. The final step of the process involved reprojecting all tiles to a common reference system (Polar Stereographic projection with a standard latitude of 71 degrees South and a central meridian of 0 degrees), mosaicking all tiles and clipping to remove any remaining artifacts along the coast. The outer limit is based on the LIMA coastline with a 500 m buffer.

Error Sources

The new DEM has a mean elevation difference of - 4 m (±25 m Root Mean Square Error [RMSE]) from ICESat, and a horizontal error of less than 2 pixels, although elevation accuracies are lower on mountain peaks and steep-sided slopes. Some anomalies along the coast have been removed, resulting in small gaps, and the DEM has a small number of remaining artifacts. It only covers regions included in ASTER GDEM, in which there are inherent gaps and some missing islands.