A digital elevation model (DEM) and orthophoto of Shirley Island, Windmill Islands, Antarctica

Full description

This dataset includes:
(i) a 2 metre resolution digital elevation model (DEM) of Shirley Island, Windmill Islands, Antarctica;
(ii) reliability data for the DEM;
(iii) contours interpolated from the DEM; and
(iv) an orthophoto created using the DEM.

The data are stored in the UTM zone 49 map projection.
The horizontal datum is WGS84.

The data were created by Robert Anders, Centre for Spatial Information Science, University of Tasmania, Australia to support the postgraduate research of Phillipa Bricher into the nesting sites of Adelie Penguins.

See a related URL below for a map showing Shirley island.

Lineage

Progress Code: completed

Statement: Digital Photogrammetric Software, Virtuozo v3.6, was used in the production of the DEM and orthophoto. http://www.supresoft.com The DEM was derived using photogrammetry techniques from aerial photography flown on 4/01/01. The photography used was from the Australian Antarctic Data Centre's aerial photography collection: Film ANTC1099 Run 1- Frames 2, 3, and 4 Run 2 - Frames 2, 3, 4, 5, 6, 7, and 8 Run 3 - Frames 5 and 6 See a related URL below to search the the Australian Antarctic Data Centre's Aerial Photography Catalogue. Ground control for the photography was measured by Phillippa Bricher and an AAD employee in the 2005/2006 season using differential GPS. Orthometric heights were supplied, some with negative values (max -2.84). A +3.000 m factor was added to all ground control heights to give positive values, a requirement for the Virtuozo software. Ground Control Limitations: Results from the post processing of the GPS data indicate that the average precisions of the ground control coordinates are: Horizontal 0.68 m plus or minus 0.15 (s.d.) (range 0.113 - 1.020 m) Vertical 1.33 m plus or minus 0.32 (s.d.) (range 0.14 - 2.07 m) In addition to the accuracy of the GPS coordinates, there is an error associated with the identification of ground control points, estimated to be in the 0.5 - 1.0 m range. This error is primarily due to some poorly defined natural targets being used. In forming the DEM, individual stereo models were formed and a polynomial block adjustment was used (there has been no bundle adjustment) for the absolute orientation. Data accuracy: The stated accuracy of the DEM (see below) is limited to areas of terrain that is not covered by snow or ice. Areas covered by snow or ice were only edited to the extent of removing significant gross errors by using a linear interpolation algorithm based on surrounding elevation data. This problem is due to difficulties in not being able to view non-textured surfaces in stereo, and consequently not being able to edit the DEM. Similarly given the transient nature of snow/ice, this data is unreliable. Accompanying this DEM is a polygon shapefile of the snow/ice layer which can be used as a data uncertainty layer - see a related URL below. Similarly, areas that have had high concentrations of penguins also exhibit a non-textured surface. Given the small relative size difference to the snow/ice areas, elevation data in these areas would be expected have a greater reliability. Absolute height of the DEM remains uncertain, given the accuracy of the ground control. The issue of the +3.000 m adjustment applied to the height of the ground control has not been resolved, in particular the position of the sea-level boundary. Viewing the DEM or its derived products may show areas of 'smoothness' in contrast to areas showing detailed relief. These smooth areas typically occur in the regions between the overlapping boundaries of the individual stereomodel DEM's, an artefact of the merging of stereomodel DEMs. An examination of the DEM in these areas is consistent with ground control accuracies, and the Australian Antarctic Data Centre's 5 metre interval contour data. A kriging algorithm was applied to the DEM to reduce this effect. In summary, subject to the explanations and limitations noted above: Orthophoto - The planimetric location of a well defined point is expected to be within 1 metre of its true value (90% certainty). DEM - The height associated with a planimetric position for regions unaffected by snow, ice, or penguin guano is expected to be within 2 metres (90% certainty) of its true value.