ed 
he 
at 
ert 
de 
dy 
0° 
no 
le 
ne 
or 
al 
es 
surface is not only a physical figure of mars, but it is 
also a long-range planning for accommodating high- 
resolution (1 - 2 km) and high-precision (1 m) 
topographic data from future Mars missions. 
4.1 Gravity field of Mars 
Because the gravity field of Mars is rough, a large 
number of coefficients is required in the spheric 
harmonic representation of the gravity field. But, 
Jordan and Lorell (1972) discovered that high-degree 
spherical harmonics are not stable and only the low- 
degree harmonics have demonstrated stability, 
especially those coefficients of the fourth degree. 
Therefore, the  fourth-order and  fourth-degree 
spherical harmonics derived for the gravity potential 
surface are used for computing elevations on the 
Marian surface. Incidently, a fourth-order and forth- 
degree spheric expansion had been considered to be 
used for representing the world-wide gravitational 
potential of the Earth when disagreement was found 
between both the northern and southern hemispheres 
(Utila, 1962a, 1962b). 
The fourth-degree and fourth-order gravity coefficients 
of Mars are: J2=19.65, J3=0.36, J4=-0.29, C22=- 
0.548, S22=0.31, C31=-0.48, C32=-0.55, C33=0.048, 
S31=0.26, S32=0.026, S33=0.035. Values were 
obtained from the results of the celestial mechanics 
experiment of Mariner 9 (Jordan and Lorell, 1975). In 
the spheric harmonics expression, the associated 
Legendre polynomials are used and the mean radius 
of Mars is 3,382.946 km which was obtained by fitting 
the 6.1-millibar occultation pressure data from Mariner 
9. Detail derivation of the spheric harmonics is 
discussed in Wu's publication (Wu, 1975). 
4.2 The Figure of Mars 
For many years, by optical measurements, 
astronomers observed a large flattening value of about 
0.011 for Mars (Dollfus, 1972). However, by dynamic 
measurements from observations of the two satellites 
of Mars, a value of approximately 0.0052 was 
obtained. The radio occultation data from Mariner 9 
indicated that the average terrain elevation in the 
northern hemisphere is about 3 to 4 km lower than that 
in the southern hemisphere. This mans that the 
northern hemisphere is closer to the center of mass of 
mars and Mars is also asymmetric in the north-south 
direction (Kliore, et al., 1972, 1973), therefore an 
oblate spheroid is not adequate to represent the 
planet. 
Using the fourth-order and fourth- degree gravity 
coefficients and the mean radius of the 6.1-millibar 
pressure surface, a triaxial ellipsoid of Mars can be 
obtained to geometrically establish the topographic 
datum based on the gravity field of Mars 
967 
(Christensen,1975), which are: A = 3,394.6 km, B = 
3,393.3 km, C= 3,376.3 km, 6 = 105°. These results 
confirm the Mars dynamic flattening of 0.0052 which 
was arrived in the early days by observing the two 
satellites of Mars. 
4.3 Topographic Datum of Mars 
The radius of the Mars mean sphere has a 6.1- 
millibar pressure (Christensen, 1975). As C.A. Berth 
suggested that a meaningful normalization on Mars is 
to choose the zero altitude at the triple-point of water 
(6.105 millibar). The radius of 6.1 millibar pressure 
surface is 3,382.946 km. Therefore, the Mars 
topographic datum is defined by equations derived in 
Wu's publication (Wu, 1975.) 
5. PLANETOID TOPOGRAPHIC CONTROL 
NETWORK OF MARS 
Using high-altitude Viking Orbiter images, the United 
States Geological Survey in Flagstaff has established 
a Mars planetwide control network by analytical. 
aerotriangulation (using the USGS GIANT program.) 
Primary controls used for the adjustment include: (1) 
camera positions and orientations derived from the 
tracking data of Viking missions; (2) occultation 
measurements from the S-band radio experiment on 
both Mariner 9 and Viking missions (Kliore, et al, 1970, 
1972, 1973); (3) elevation derived from Earth-based 
radar observations of Mars (Downs, et al, 1971, 1973, 
1975; Goldstein, et al, 1970); and (4) horizontal 
coordinates from the control network by davies (Davies 
et al, 1978). This control network not only provides 
control for systematic mapping topography of Mars, 
but also valuable for: (1) improving the horizontal 
locations of occultation points on both the Mariner 9 
and Viking missions; (2) calibrating locations of earth- 
based radar profiles; (3) improving elevations of 
Davies horizontal control network; and (4) adjusting 
the internal inconsistency of camera positions and 
orientations. In fact, the adjusted internal consistency 
of camera stations is the only means by which the 
extreme narrow field-of-view of Viking images can be 
used to establish stereomodels on the analytical 
stereoplotters for photogrammetric compilation. The 
adjusted camera parameters have been used to 
enhance the SEDR data to be called "Wu's Version" in 
the SPICE file. 
The GIANT program uses an interactive least-square 
technique which requires initial approximation for each 
of unknown parameters. All parameters are treated as 
weighted parameters, ranging from known to 
unknown. This is particularly helpful for the Mars 
control network since we don't have good ground 
control points except occultation points and radar 
observations with limited precision. Using SEDR data, 
International Archives of Photogrammetry and Remote Sensing. Vol. XXXI, Part B4. Vienna 1996