[IGSMAIL-0469] IGS Orbit/clock combination
Analysis
Analysis
Wed Feb 2 13:47:32 PST 1994
******************************************************************************
IGS Electronic Mail Wed Feb 2 13:47:32 PST 1994 Message Number 0469
******************************************************************************
Author: Analysis Centre Coordinator/ Geodetic Survey of Canada
Subject: IGS orbit/clock combination
All IGS orbits and clocks are now being combined according to the
recommendations of the Ottawa IGS Analysis Centre Workshop. Currently the IGS
RAPID service orbits (based on the IERS Rapid Service Bull. A) have been
completed and archived at the IGS data centres for the GPS weeks 723-732.
For the convenience of IGS products users three types of IGS
product files are generated and archived (WWWW - week#. e.g. 0732):
IGSWWWW[0-6].SP3 IGS orbits & clocks in sp3 format, including orbit accuracy
codes (no clock accuracy codes available in sp3 format)
IGSWWWW7.ERP IERS pole to be used with the above orbits (IERS Bull. A
or Bull. B)
IGSWWWW7.SUM Complete summary file including combination statistics
(on IGS orbits/clocks as well as a feedback to Analysis
Centres). See example bellow.
Currently only the Rapid orbits based on the current IERS Bull. A pole have been
generated. The orbit type (i.e. RAPID or FINAL) is coded in all the three files
above.
The FINAL IGS orbits will be generated for all the 1994 IGS (Rapid) orbits when
the Final IERS Bull. B will be available. Then, all the above files will
be replaced on CDDIS (and other IGS data centres). The old Rapid orbits may be
archived under a different name (e.g. 'IGR') and an IGSMAIL will also be
issued. In this way only the best orbit/clock products will be available under
the IGS designation, which may be important for uninitiated or infrequent IGS
product users.
The naming convention above is only provisional and is subject to change
pending the outcome of the Paris IERS/IGS Workshop (March 21-25, 1994) and the
IGS GB meeting.
------------ABBREVIATED EXAMPLE of IGS07327.SUM on IGS DATA CENTRES---------
Author: Analysis Centres Coordinator / Geodetic Survey of Canada
Rapid Service IGS Orbit Combination - Week 0732 (Jan 16, 1994 - Jan 22, 1994)
Contacts: F. Lahaye (lahaye at geod.emr.ca), J. Kouba (kouba at geod.emr.ca)
COMBINATION STRATEGY:
Each centre ephemeris and satellite clock corrections are first transformed to
common reference. The differences between each centre X and Y pole solutions
and the IERS Bull. A are used to rotate the corresponding ephemeris. If the
pole solution is not available for a particular centre, no such rotations are
applied. For each centre, a clock offset and drift, common for all satellites,
are estimated with respect to broadcast clock corrections. These offsets and
drifts are used to align each centre satellite clock corrections to GPS time.
The resulting ephemerides and clock corrections are then combined as weighted
averages computed over all centres. Each centre is given a position weight and
a clock weight. Currently the weights are computed from the RMS fit to the
unweighted average orbit and satellite clocks, after the estimation of a
7-parameter Helmert transformation (in the case of ephemeris) and offsets and
drifts (in the case of clocks). The above estimations are performed using the
L1-norm robust scheme. Each centre ephemeris are also evaluated individually
and independently from the orbit combination by using week long arcs.
Like the clock corrections broadcast in the navigation message,
the combined satellite clocks are corrected for the periodical relativistic
correction (see e.g. GPS interface Control document ICD-GPS-200, 1991, p.73):
-2*(X.V)/c,
where X and V are the sv state and velocity vectors and c is the speed of
light. Also for compatibility, no satellite L1/L2 delay calibration should be
applied.
SOFTWARE:
The weighted average software used was developed by T. Springer and G. Beutler
during 1993. The GPS long arc orbit analysis software used here for individual
centre ephemeris evaluation was developed by Beutler et al. at the Astronomical
Institute of the University of Bern (AIUB). For more details on the orbit
combination and complete references for both software see:
Beutler, Kouba, Springer, 1993, Combining orbits of the IGS processing
centres, in IGS Analysis Centre Workshop, October 12-14, Ottawa, pp 19-55.
PRODUCTS GENERATED:
The resulting average ephemeris and clocks are output to daily files in SP3
format. The satellite accuracy codes displayed in the header are based on the
weighted RMS for each satellite. Note that each file contains all the
satellites that appear in any of the centre ephemeris files. If a satellite or
centre was excluded from combination, a note in the REMARKS section of this
report will identify the exclusion and the reason.
The Rapid Service Orbit files generated and uploaded into CDDIS @ NASA/GSFC
are:
IGS07320-6.SP3 IGS GPS ephemeris files for days 0-6, week 0732 in SP3 format.
IGS07327.ERP Earth rotation parameters (ERP) associated with ephemerides.
IGS07327.SUM Complete report.
REMARKS on IGS orbit combination:
1. Starting with week 0731, the above satellite clock corrections combination
was put in place. The centres used in the combination are EMR, ESA, GFZ and
JPL. The other centres are excluded because they either provide broadcast
clocks (COD) or clock corrections are not provided (NGS & SIO). The clock
information not used is still compared to combination for completeness.
2. Only the GFZ "exact" clock corrections (one epoch out of two) are included
in the combination and RMS computation (no "interpolation" values).
3. ESA satellite clock corrections were not incuded in the combination since
they exhibit systematic effects at the 10 nano-seconds level due to not
accounting for relativistic effects.
4. GFZ clock corrections show offset changes on day 5 (Jan 21) at 6:00 and
12:00 on all satellites.
5. EMR and JPL contain on day 6 (Jan 22) outlier clock corrections for PRN02,
PRN16 and PRN26.
6. The broadcast clock corrections for PRN03 on day 6 (Jan 22) did reflect
a large (14 micro-s) change between 14:00 and 20:00 (ref NANU 028-94024).
7. A maneuvre for PRN17 on days 4 and 5 (Jan 20-21) (ref NANU 003-94007).
8. In the long arc analysis, a state vector reset was modeled for PRN17 on
day 5 (Jan 21) and a momentum dump was modeled for PRN22 on day 3 (Jan 19).
ORBIT COMBINATION AND EVALUATION STATISTICS:
Most combination statistics are with respect to the combined orbits and consist
of three parts. The first, Table 1, summarizes all statistics for the week. It
contains means and standard deviations over the week of satellite ephemeris and
clock correction transformation parameters for each centre (table 1.0732.a).
It also displays the computed satellite position sigmas for the combined orbits
(table 1.0732.b). The second part, Table 2, contains the complete ephemeris
and clock correction transformation parameters (includes both the apriori and
a posteriori values) for each centre and each day (tables 2.0732.0-6). The
third and final part, Table 3, contains two types of daily RMS for each
satellite and centre, along with the computed satellite position sigmas for the
combined orbits (tables 3.0732.0-6). The first type of satellite RMS is based
on orbit combination and the second RMS type is a result of the 7-day arc
evaluation of each centre ephemerides.
In all tables, the most important statistic which is meaningful for intra
centre/satellite analysis is labeled WRMS. It is a weighted RMS where
normalized weights are based on accuracy codes given in the SP3 files, and
should not be affected by downweighted/problem satellites when acknowledged by
individual centres in their SP3 file.
Table 1: Week 0732 summary. (Jan 16, 1994 - Jan 22, 1994)
Table 1.0732.a GPS week: 0732 MJD: 49368.0-49374.0
Mean and standard deviations of transformation parameters.
WRMS - orbit RMS weighted by the SP3 header accuracy codes.
Units: meters, mas, ppb, nano-sec, nano-sec/day.
CENT| DX DY DZ RX RY RZ SCL RMS WRMS | TOFT TDRFT RMS
----|---------------------------------------------------|------------------
cod | .01 .01 .02 1.54 .11 -.03 .1 .93 .59 | .2 -.8 74.6
| .01 .01 .01 .19 .67 .25 .2 | 1.5 1.9
emr | .01 -.01 -.01 2.59 .56 -.14 -.3 .12 .12 |-294.1 -13.1 .9
| .01 .01 .00 .52 .48 .25 .2 | 305.1 19.7
esa | .01 .00 .01 1.80 .06 -.84 -.1 20.11 10.64 | 1.2 -3.5 14.0
| .01 .01 .01 .23 .75 .62 .2 | 18.8 2.7
gfz | -.04 .00 .01 1.64 1.36 -.82 -.4 .13 .13 | 35.6 -11.3 9.5
| .00 .01 .01 .40 .49 .23 .2 | 35.5 28.3
jpl | -.01 .04 .00 1.54 .11 -.20 .2 .12 .12 | 35.8 -10.8 1.0
| .01 .01 .01 .25 .53 .37 .1 | 36.1 30.4
ngs | .02 -.06 .01 1.47 -.12 .41 2.0 .95 .57 | .0 .0 .0
| .01 .02 .03 .29 .82 .26 .5 | .0 .0
sio | .02 -.05 -.05 2.48 .54 2.54 .5 .92 .93 | .0 .0 .0
| .01 .01 .01 .31 .31 1.22 .3 | .0 .0
Table 1.0732.b GPS week : 0732 MJD: 49368.0-49374.0
Satellite position sigmas for the combined orbits
(coded in the IGS SP3 orbit files).
"E" flags eclipsing satellites.
Units: cm.
Day of GPS week
PRN | 0 1 2 3 4 5 6 | Remarks
------|-----------------------------|------------------------------------------
1 | 6 7 4 5 4 4 5 |
2E | 7 8 8 9 6 9 9 |
3 | 4 5 5 5 4 6 7 |
4E | 7 6 5 6 4 6 7 |
5E | 8 9 7 10 6 7 9 |
7 | 5 5 5 5 3 4 6 |
9 | 5 5 4 5 3 5 4 |
12 | 5 4 5 4 3 4 4 |
13 | 6 5 5 6 2 4 6 |
14 | 5 5 4 4 4 5 4 |
15E | 5 4 4 5 4 6 5 |
16 | 6 7 5 7 4 5 6 |
17E | 5 5 5 7 >>> 10 6 | Maneuvres on days 4 and 5 (Jan 20-21).
18 | 4 4 4 4 3 4 4 |
19 | 4 4 4 5 3 5 5 |
20E | 7 6 7 6 5 8 8 |
21 | 5 5 4 5 3 4 7 |
22E | 6 7 5 7 4 6 8 |
23 | 7 5 7 4 3 7 6 |
24E | 7 6 5 6 5 6 7 |
25 | 6 6 6 7 4 5 7 |
26 | 6 5 5 6 3 5 6 |
27 | 4 4 4 6 3 5 5 |
28 | 4 5 4 5 4 4 4 |
29 | 4 4 4 4 3 4 4 |
31 | 6 6 6 5 3 5 4 |
Table 2: Daily transformation of each centre to the combined IGS orbit and
clocks for each day of the week. The transformation displayed is the
sum of a priori and estimated transformations.
WRMS - RMS weighted by the centres SP3 header accuracy codes.
Units: meters, mas, ppb, nano-sec, nano-sec/day.
Table 2.0732.0 GPS week: 0732 Day: 0 MJD: 49368.0
CENT| DX DY DZ RX RY RZ SCL RMS WRMS | TOFT TDRFT RMS
----|---------------------------------------------------|------------------
cod | .02 .01 .02 1.27 .84 .22 -.1 .11 .11 | 2.1 1.6 76.4
emr | .00 -.02 -.01 2.98 1.27 -.19 -.2 .12 .12 |-593.8 -14.0 .5
esa | .01 .01 .02 1.79 .96 .10 -.3 .18 .18 | -29.1 -2.5 11.8
gfz | -.04 .00 .01 1.84 2.14 -.40 -.4 .12 .11 | 85.8 -23.4 .4
jpl | -.01 .05 .00 1.44 .74 -.34 .2 .12 .13 | 85.9 -21.3 .6
ngs | .01 -.08 -.04 1.24 .25 .45 1.3 .29 .28 | .0 .0 .0
sio | .01 -.05 -.05 2.40 1.15 .78 .3 .20 .20 | .0 .0 .0
** For complete Table 2 see the IGS07327.sum file on IGS data centres **
.
.
Table 3: Daily fit for each satellite and each centre resulting from weighted
average combination and from long arc orbit dynamics (7-day). The
first IGS column gives resulting satellite accuracy measures found in
SP3 header. The last IGS column gives the RMS fit of the IGS orbit
when processed through the 7-day arc evaluation.
"E" flags eclipsing satellites.
Units: cm.
Table 3.0732.0 GPS week: 0732 Day: 0 MJD: 49368.0
Weighted Average Orbit Dynamics (7 days)
PRN | cod emr esa gfz jpl ngs sio IGS | cod emr esa gfz jpl ngs sio IGS
------|----------------------------------|----------------------------------
1 | 4 7 26 10 13 43 16 6 | 11 26 15 20 15 44 29 7
2E | 14 20 22 18 17 24 21 7 | 33 43 33 51 31 36 43 31
3 | 5 12 12 6 11 15 9 4 | 12 21 16 21 16 18 14 7
4E | 10 11 10 11 15 49 32 7 | 18 29 17 24 17 31 23 10
5E | 14 10 32 15 14 30 34 8 | 23 32 41 35 20 40 40 27
7 | 11 14 11 11 12 22 14 5 | 17 21 19 24 12 21 22 10
9 | 9 13 11 14 13 17 18 5 | 13 23 14 19 14 20 23 9
12 | 5 14 11 6 10 18 24 5 | 10 23 13 14 16 20 25 7
13 | 14 12 5 8 17 26 24 6 | 14 23 18 24 18 15 24 8
14 | 6 11 13 12 9 32 17 5 | 18 28 22 18 14 21 21 12
15E | 9 12 14 16 10 16 11 5 | 12 25 21 20 15 22 19 13
16 | 16 12 16 13 11 33 11 6 | 21 25 23 22 13 36 25 10
17E | 8 11 14 10 11 28 15 5 | 9 23 20 26 9 30 27 10
18 | 7 9 18 6 10 23 15 4 | 12 22 12 18 13 19 20 6
19 | 4 9 15 9 9 17 11 4 | 14 26 17 25 15 21 22 9
20E | 13 10 19 20 18 44 16 7 | 13 26 25 27 9 27 26 16
21 | 10 10 15 5 10 31 17 5 | 17 22 13 22 15 18 25 8
22E | 17 9 14 10 10 31 21 6 | 16 29 18 23 12 21 21 12
23 | 19 14 23 12 11 23 22 7 | 19 47 22 45 39 33 48 27
24E | 9 11 19 15 13 46 29 7 | 17 29 24 21 13 36 28 9
25 | 13 12 22 12 10 41 21 6 | 16 27 14 24 14 27 30 9
26 | 8 11 14 12 11 33 24 6 | 14 25 23 18 17 28 20 7
27 | 4 8 15 7 8 16 19 4 | 14 27 17 22 10 22 23 9
28 | 8 10 20 5 9 15 12 4 | 17 20 15 18 10 19 19 10
29 | 5 9 18 7 8 22 9 4 | 12 22 14 18 12 18 19 5
31 | 11 10 30 11 13 17 17 6 | 14 21 16 20 11 17 24 9
------|----------------------------------|----------------------------------
RMS | 10 11 18 11 12 29 19 | 16 27 20 25 16 26 26
WRMS | 10 11 17 11 12 28 20 |
** For complete Table 3 see the IGS07327.sum file on IGS data centres **
.
.
-------------------END OF EXAMPLE--------------------------------------------
More information about the IGSMail
mailing list