each ground station. The number of these measurement parameters exceeded 3000 for the
14 station network, and 5000 for a 24 station network. A pass algorithm (Rim et aI.,
1995) was employed to estimate these measurement parameters efficiently. A phase bias
in the z-coordinate of the TIP satellite centered coordinate system was also adjusted for
GPSIDR data.
When the information equations were combined with the JGM-1 information equations,
the relative weights of the JGM-1 information equations were fixed, while the relative
weight of the new satellite information was changed for the JGM-3 solutions. The
relative weight of 0.05 was chosen for the GPS data.
GRAVITY FIELD EVALUATIONS
In order to evaluate the gravity fields, six JGM fields were analyzed. Those were JGM-1,
JGM-2, JGM-3, JGM-3A(no TIP GPS), JGM-3B (no TIP SLRlDORIS), and JGM-3D. It
should be noted when evaluating the contribution of the GPS data by comparing JGM-3A
and JGM-3B, that JGM-3A was generated using 20 Cycles of the TIP SLRlDORIS data,
while JGM-3B included only 8 Cycles of the TIP GPS data. However, these comparisons
will give some indication of how the GPS data affected the gravity solution in
comparison with the SLRlDORIS data.
Table 1 summarizes the performance of these fields on the TIP SLRlDORIS data fit for
Cycle 45, which was not included in the gravity solution. Eight hour Ct and daily 1-cpr T
and N acceleration were estimated for these runs. The SLR fit improved 18 mm from
JGM-1 to various JGM-3 fields. However, there was no difference in the SLR and
DORIS fits for JGM-3A, JGM-3B, and JGM-3D. Also, the crossover statistics did not
change for the various JGM-3 fields. Comparison with the JPL's reduced dynamic orbit
shows that the agreement with the JPL's reduced dynamic orbit degraded by 3 mm, 8 mm,
and 4 mm in RTN components, respectively, by excluding the TIP GPS data in the
gravity solution. However, withholding the TIP SLRlDORIS data d.id not change the
agreement, except for a 3 mm degradation in the T component.
Table 2a and 2b compares the performance of the JGM fields using tracking data from
several other satellites including Ajisai, Starlette, Lageos-1, Lageos-2, and Stella. Table
2a summarizes the data fits when 1-cpr parameters were not adjusted, while Table 2b
shows the data fits when 1-cpr parameters were estimated. JGM-3B and JGM-3D fit the
data slightly better than JGM-3 and JGM-3A for Ajisai and Lageos-1. However, Starlette,
Lageos-2, and Stella fits were the same for various JGM-3 fields for both
parameterizations.
Table 1. TOPEXIPOSEIDON orbit comparison (cycle 45).
Model
SLR
DORIS
Xover
JPL(cm)
Z-bias
(cm)
(mm/s)
(cm)
R
T
N
(cm)
JGM-1
3.7
5.6
9.2
4.1
11.1
8.9
-0.1
JGM-2
2.4
5.5
8.8
2.6
7.3
7.5
-0.6
JGM-3
2.0
5.5
8.5
1.5
4.4
5.6
-0.3
JGM-3A
1.9
5.5
8.5
1.7
5.2
6.1
-0.4
JGM-3B
1.9
5.5
8.5
1.4
4.7
5.7
-0.3
JGM-3D
1.9
5.5
8.5
1.4
4.4
5.7
-0.3
Note:
1) JPL's new reduced dynamic orbits used for comparison
2) 8-hour C t and daily 1-cpr T and N estimated
44
14 station network, and 5000 for a 24 station network. A pass algorithm (Rim et aI.,
1995) was employed to estimate these measurement parameters efficiently. A phase bias
in the z-coordinate of the TIP satellite centered coordinate system was also adjusted for
GPSIDR data.
When the information equations were combined with the JGM-1 information equations,
the relative weights of the JGM-1 information equations were fixed, while the relative
weight of the new satellite information was changed for the JGM-3 solutions. The
relative weight of 0.05 was chosen for the GPS data.
GRAVITY FIELD EVALUATIONS
In order to evaluate the gravity fields, six JGM fields were analyzed. Those were JGM-1,
JGM-2, JGM-3, JGM-3A(no TIP GPS), JGM-3B (no TIP SLRlDORIS), and JGM-3D. It
should be noted when evaluating the contribution of the GPS data by comparing JGM-3A
and JGM-3B, that JGM-3A was generated using 20 Cycles of the TIP SLRlDORIS data,
while JGM-3B included only 8 Cycles of the TIP GPS data. However, these comparisons
will give some indication of how the GPS data affected the gravity solution in
comparison with the SLRlDORIS data.
Table 1 summarizes the performance of these fields on the TIP SLRlDORIS data fit for
Cycle 45, which was not included in the gravity solution. Eight hour Ct and daily 1-cpr T
and N acceleration were estimated for these runs. The SLR fit improved 18 mm from
JGM-1 to various JGM-3 fields. However, there was no difference in the SLR and
DORIS fits for JGM-3A, JGM-3B, and JGM-3D. Also, the crossover statistics did not
change for the various JGM-3 fields. Comparison with the JPL's reduced dynamic orbit
shows that the agreement with the JPL's reduced dynamic orbit degraded by 3 mm, 8 mm,
and 4 mm in RTN components, respectively, by excluding the TIP GPS data in the
gravity solution. However, withholding the TIP SLRlDORIS data d.id not change the
agreement, except for a 3 mm degradation in the T component.
Table 2a and 2b compares the performance of the JGM fields using tracking data from
several other satellites including Ajisai, Starlette, Lageos-1, Lageos-2, and Stella. Table
2a summarizes the data fits when 1-cpr parameters were not adjusted, while Table 2b
shows the data fits when 1-cpr parameters were estimated. JGM-3B and JGM-3D fit the
data slightly better than JGM-3 and JGM-3A for Ajisai and Lageos-1. However, Starlette,
Lageos-2, and Stella fits were the same for various JGM-3 fields for both
parameterizations.
Table 1. TOPEXIPOSEIDON orbit comparison (cycle 45).
Model
SLR
DORIS
Xover
JPL(cm)
Z-bias
(cm)
(mm/s)
(cm)
R
T
N
(cm)
JGM-1
3.7
5.6
9.2
4.1
11.1
8.9
-0.1
JGM-2
2.4
5.5
8.8
2.6
7.3
7.5
-0.6
JGM-3
2.0
5.5
8.5
1.5
4.4
5.6
-0.3
JGM-3A
1.9
5.5
8.5
1.7
5.2
6.1
-0.4
JGM-3B
1.9
5.5
8.5
1.4
4.7
5.7
-0.3
JGM-3D
1.9
5.5
8.5
1.4
4.4
5.7
-0.3
Note:
1) JPL's new reduced dynamic orbits used for comparison
2) 8-hour C t and daily 1-cpr T and N estimated
44