ul 2004 
lipsoid. 
le (see 
3S all 
raphic, 
ly with 
in the 
to be 
)rocess 
'orking 
surface 
curacy 
^et" al 
to the 
:RASS 
sibility 
sphere 
c and 
itasets, 
c. The 
Istance 
tion of 
biaxial 
>r data 
)proxi- 
herical 
luence 
gitude 
ysoidal 
small 
herical 
at the 
to 1.2 
errors 
ointed 
s give 
For à 
versus 
two 
place, 
ribute 
riment 
;t (see 
Mars, 
their 
pu 
Equat. Polar y; | Ret. 
ID| Dataset 1 Latitu- | Radius | Radius s 
; de [km] [km] t 
Goole: graphic| 4493 40] 3375.73 [14 
(1)] gic Map 
MDIM graphic 3393 40 | 3375.73 176. [14] 
(2) | 646 
MDIM graphic | ,, ie 176. | [6] 
G) 20 3396.00 | 3376.80 725 
MSSS" graphic : 176. |[15] 
(4) Atlas 3396.00 | 3376.80 725 
MOLA | IAU | centric [11] 
International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B4. Istanbul 2004 
  
  
  
  
  
  
  
(5) MEGDR| 2000 3396.00 | 3396.00 | 176.63 
*** 
  
  
  
  
  
  
  
  
  
  
  
  
TES IAU | centric [16] 
(6) 2000 3396.00 | 3396.00 | 176.63 
MDIM 
Gi 21 | [AU ; 
(except 2000 centric | 3396.19 | 3396.19 | 176.63] [1] 
the 
poles) 
MDIM AU | 
(8| 2.1 2000 centric | 3376.20 | 3376.20 | 176.63 [1] 
(poles) 
  
Table 2: Geodetic properties of raster datasets (mosaics, 
geology, heights) implemented in DLR GRASS GIS. 
As a consequence of the slightly different coordinate system 
definitions stated in chapter 3.1.2, we chose to utilize a 
spherical coordinate system on a sphere of r-3396 km to 
conform with the MGCWG recommendation (see chapter 2.1). 
A unique Wy, of 176.630?, according to IAU2000, was selected. 
GRASS provides a module to directly import raster data in raw 
format. Registration information are latitude and longitude of 
the four edges of the file. 16 MOLA MEGDR (5, see Table2) 
topography tiles were imported in raw format using the 
(converted) registration information from the accompanying 
Planetary Data System (PDS) labels. TES (6, see Table2) multi- 
band mineral data in raw format were divided into single band 
files prior to import. As we used a sphere as the reference body, 
no resampling of the planetocentric MOLA and TES data was 
necessary. MDIM2.1 (7 and 8, see Table2) compiled with 
latitude definition as well, so the import of this dataset was also 
straightforward. 
Datasets with a different definition of the prime meridian were 
corrected by shifting the longitudes prior to import into 
GRASS. In the same step, all registration information was 
converted to the GRASS longitude range (see chapter 3.1.2). 
Subsequent to W, correction and removal of the attached PDS 
header, all 28 tiles of the MSSS atlas where read as raw data. 
GRASS' ability to reproject data from the ellipsoid (4, see 
Table 2) to the sphere was used to transfer the tiles to the 
planetocentric reference frame of the database. The reprojection 
from ellipsoid to sphere was also necessary for the geologic 
map (1, see Table 2) after W, correction and import. 
Nevertheless, as it was compiled on a much older base, this 
  
. Mars Digital Image Mosaic 
… Malin Space Science Systems 
Thermal Emission Spectrometer: 
Various Minerals, Surface Emissivity, Albedo 
dataset does not register very well to recent data. Still it is the 
only global geologic data available. 
Point data (e.g. MOLA) were read from standard ASCII tables 
along with associated attribute information. 
5. APPLICATION 
Once the database was created as a foundation, this data base 
found a wide variety of applications. To name one example, we 
used GRASS to compile a dataset of tectonic surface faults. 
MOLA maps artificially lit from varying azimuth angles were 
used as a basis for the mapping. Hence, contrary to mapping 
from an image base, where the light conditions have to be taken 
as is, we were able to avoid any sampling bias due to 
illumination geometry. 
On the MOLA base map, the surface faults data were 
conveniently extracted by visual interpretation, stored, and 
analyzed. Each surface fault was stored in a number of equally- 
spaced (250m) points. Thus, we collected a total set of 3642 
thrust faults and 3746 normal faults, ranging from lengths 
between 8 and 1445 km. The total length of all faults was 
approx. 600,000 km. Using the geologic map, each fault was 
then assigned to a geologic region and a specific surface age 
(see Figure 1). 
We expect that our data set, globally more homogeneous than 
that of previous studies, which can now be examined under a 
variety of aspects: spatial variations of fault patterns, length 
statistics, correlations of surface faults with age, or sequence of 
formation. Specifically, we intend to use this digital data base 
to generate synthetic sets of Mars quake catalogs for 
simulations of the performances of future seismometer network 
on Mars. 
In addition to the already mentioned import, projection and 
vector editing tools, GRASS provides a vast amount of other 
data im-/export, processing, analysis and visualization routines. 
The reader is referred to [7] and [13] for more information. 
6. CONCLUSIONS 
The number and diversity of available datasets render GIS 
technology an especially suitable tool for scientific studies on 
Mars. Keeping the geodetic properties of the individual Mars 
datasets and GRASS’ coordinate system definitions in mind, the 
application of GRASS in planetary research turns out to be 
straightforward. 
ACKNOWLEDGEMENTS 
The authors greatly acknowledge helpful discussions with R. 
Kirk (USGS). 
This study was supported by the German Science Foundation 
(Deutsche Forschungsgemeinschaft, DFG). 
813