GlobDEM50 FAQ
What is the accuracy of the GlobDEM50 elevation data?
The coast lines in GlobDEM50 are far from reality! Are the elevation data that bad?
Can I use the elevation files *.aag directly as inputs to AUSTAL2000?
Can I use elevation data files *.xyz, which I have build with the tool “aag_To_xyz” directly as input to AUSTAL2000?
I have downloaded data files from the maps in GlobDEM50. But the extensions of files on the hard disc changed from *.aag to something like *.htm or *.txt.
I can not convert data files with the tools “aag_To_xyz”, “aag_To_grd” and “aag_To_a2k”. The protocol file (*.rpt) has an error message “Error reading data header...”. When I load the file *.aag into a text editor, I can see some strange character in the data header.
In some subareas elevation data look unrealistically (especially in shaded relief plots).GlobDEM50 FAQ
What is the accuracy of the GlobDEM50 elevation data?
Principally GlobDEM50 elevation data have the same accuracy as the raw data they are based on. The absolute accuracy of the SRTM raw data are said to be around 6 to 8 m in height. Corresponding to other publications and own studies normally the data are better. Bias and mean square error to official German elevation data are in the range of some metres. Correlations are mostly at 0.99. Because the data are based on radar measurements they are less accurate for water surfaces where they may have errors of more than +/- 10m. See also the documentation of GlobDEM50.
The coast lines in GlobDEM50 are far from reality! Are the elevation data that bad?
Far from it! Normally, GlobDEM50 elevation data have a high quality. The only reason for “bad coast lines” is that GlobDEM50 does not show any coast lines! In GlobDEM50 blue areas do not mean sea areas! All heights below 1m (in some regions several metres) are shown blue, heights above 1m are green. The line between blue and green colors does not correspond to the coast line because many land areas near the coast are below sea level. Due to the data accuracy even sea areas may have (small) heights of some metres above / below sea level.
Can I use the elevation files *.aag directly as inputs to AUSTAL2000?
*.aag files have exactly the data format which is called “Arcinfo GRIDASCII” in the documentation of AUSTAL2000. However, because AUSTAL2000 assumes the co-ordinates “xllcorner” and “yllcorner” to be the “lower left corner of the lower left mesh cell” and GlobDEM50 defines both co-ordinates as the “centre of the lower left mesh cell” you should subtract 25m from both co-ordinates in the header of each *.aag file.To convert multiple *.aag files you can use the tool aag_to_a2k.exe.
Please note: If you are going to convert *.aag files with the tools “aag_to_xyz” or “aag_to_grd” you must NOT modify the headers!
Can I use elevation data files *.xyz, which I have build with the tool “aag_To_xyz” directly as input to AUSTAL2000?
Yes. But you must NOT have modified the headers of the *.aag files before! Please read this too.
I have downloaded data files from the maps in GlobDEM50. But the extensions of files on the hard disc changed from *.aag to something like *.htm or *.txt.
The extension of files which you downloaded from CD/DVD depends on your system and/or browser configuration. To download data files with their original extension please follow these steps: Right click the mouse and choose “save target as...”. In the next dialog choose “All files” as the file type.
I can not convert data files with the tools “aag_To_xyz”, “aag_To_grd” and “aag_To_a2k”. The protocol file (*.rpt) shows “Error reading data header...”. When I load the file *.aag into a text editor, I can see some strange chars.
May be you first have loaded data files into the browser (with a left mouse click) and then saved the file with “File | Save as ...”. To download data files and save them unmodified to the hard disc please read here.
In some subareas elevation data look unrealistically (especially in shaded relief plots).
In all probability these subareas have data gaps and thus have been fillied up with lower resolution data from other sources. Subareas of lower resolution within larger areas of high resolution look like rough cuboids. Please also read the documentation.

