[IGSMAIL-37] Preliminary look at GPS polar motion

Dennis Dennis
Tue Jul 7 09:07:35 PDT 1992


***********************************************************************
IGS Electronic Mail        9-JUL-1992 09:34:53       Message Number 37
***********************************************************************
 
==============================================================================
                   THIS FILE CONTAINS 3 MESSAGES
==============================================================================

>From:     Dennis McCarthy
Subject:  Preliminary look at GPS polar motion
          ------------------------------------


An updated preliminary look at Earth orientation data submitted to the IERS
follows.  Units are milliseconds of arc for x and y, and
milliseconds of time for UT1-UTC.


Contributor Data Span Points       Mean          Standard Deviation
                                (NEOS-GPS)  
                              x     y   UT1-UTC   x     y   UT1-UTC
U. of Texas  48794.5     7  -0.56  0.40          0.43  1.81  
             -48800.5

Scripps      48781.0    19  -0.10  1.39          1.48  1.04
             -48799.0     

U. of Berne  48792.5     8   0.02  0.40  -1.53   2.12  1.80  0.16 
             -48800.5

Plots are available by FAX to anyone interested.

Regards,

Dennis McCarthy
  dmc at maia.usno.navy.mil


==============================================================================


>From:    Dr. Detlef Angermann
Subject: Antenna Height Problems
         -----------------------

1) We detected discrepancies for the antenna height of
   station TROM (Tromsoe). In IGS-Mail Message Nr. 33
   the antenna height was given to 2.473 m. This value
   was identical with the RINEX file information until
   day 174. For the day 175 the antenna height in the
   RINEX file is 0.000 m.
   What is the reason for this change????

2) Furthermore we detected discrepancies for the antenna 
   height of the following IGS stations between the RINEX
   files and the Message Nr. 33:

              RINEX (DAY175)     Message Nr. 33
   - FAIR       0.000 m            0.116 m
   - KOKB       0.000 m            0.093 m
   - YAR1       0.000 m            0.073 m

3) Regarding to the IGS-Mail Message Nr. 33 we would like
   to know if the antenna height of the stations NYAL, KOSG,
   FAIR and SANT have been checked or confirmed already.

In order to process the data properly we request an immediate
answer to the antenna height problems !!!! 

regards
Detlef Angermann


==============================================================================

>From:    James F. Zumberge
Subject: JPL analysis results
         --------------------

Submitted by

J F Zumberge, F H Webb, U J Lindqwister, D C Jefferson, M B Heflin, G Blewitt

all at the Jet Propulsion Laboratory, California Institute of Technology, 
Pasadena, CA  91109.

This first JPL submission to the IGS campaign contains a descriptive narrative 
and polar motion results for the week of June 21. Precise GPS ephemerides have 
also been computed for that week in file jpl06507.eph, and will be available 
soon by anonymous ftp from the CDDIS as well as from the BODHI archive at JPL 
(contact starr at logos.jpl.nasa.gov for details).

Tracking data from the global network of Rogue GPS receivers were analyzed using
the GIPSY/OASIS-II software, described in JPL internal documents. (These 
documents are available on request.)

Data were analyzed in daily batches with 24-hour satellite arcs.

The measurements consist of undifferenced dual-frequency carrier phase decimated
to 6 minutes, and carrier-smoothed dual-frequency P-code pseudorange. For both 
carrier phase and P-code, linear combinations of the individual frequencies 
provide the ionosphere-free phase and pseudorange. Data below 15 degrees 
elevation were not used. 

Estimates of a 9-component state vector for each GPS satellite (3 parameters 
each for position, velocity, and solar radiation pressure) were made, with 
nominal values of position and velocity from the broadcast ephemeris. Weak a 
priori constraints of 1 km and 10 mm/s for position and velocity, respectively, 
were imposed.

Also estimated were 3-component position vectors for each terrestrial site. 
Those for Fairbanks (Alaska,US), Algonquin Park (Ontario, Canada), and Madrid 
(Spain) had a priori constraints of 0.1 mm, effectively making them fiducial 
stations (see tables below, where they are identified by FAIR, ALGO, and MADR).
 
The ITRF'90 coordinates for these fiducial sites were realized by applying a 7-
parameter similarity transformation of our GPS solution for the GIG'91 
experiment into the ITRF'90 reference frame (see Blewitt et al., GRL vol. 19, 
no. 9, May 4, 1992]. This transformation was estimated using 12 globally-
distributed stations, and resulted in a WRMS of 15 mm. These coordinates were 
then mapped to epoch 1992.5 using VLBI-derived velocities of solution GLB718 
[Ma, Ryan and Caprette, private comm., 1991] with the additional constraint that
the vertical velocities are zero (see 'Model B' in Blewitt et al. [1992]). Since
the fiducial coordinates define the reference frame of the orbits, they are 
listed here. 

Station     Monument         X [m]             Y [m]            Z [m]

 ALGO      ARO 883160     918129.5841     -4346071.2094     4561977.7947


 FAIR	   Gilmore      -2281621.3243     -1453595.7491     5756961.9449
         Creek NCMN

 MADR	   Top of        4849202.5864      -360329.2017     4114913.0940
         choke ring

For the fiducial stations, the offsets of the TOP of the antenna choke ring from
the monument are:

Station      East [m]    North [m]     Up [m]

ALGO           0.000        0.000      0.1840
FAIR           0.000        0.000      0.1860
MADR           0.000        0.000      0.0000

The assumed phase center offsets from the TOP of the choke ring for ROGUE Dorne-
Margolin/Choke Ring antennas are:

Band         East [m]    North [m]     Up [m]

L1            0.0000       0.0000      0.0079	
L2            0.0000       0.0000      0.0264	
LC            0.0000       0.0000     -0.0207


Coordinates for all other sites were loosely (1 km) constrained. The other sites
are Alberthead (Canada), Canberra (Australia), Goldstone (US), Hartebeesthoek 
(S. Africa), Herstmonceaux (UK), JPL (US), Kootwijk (Netherlands), Madrid 
(Spain), Matera (Italy), McMurdo (Antarctica), Metshovi (Finland), Ny Allesund 
(Norway), Onsala (Sweden), Tahiti, Penticton (Canada), Pacific Geoscience Centre
(Canada), Pinyon Flat (US), Santiago (Chile), Scripps (US), St. Johns (Canada), 
Taiwan, Tromso (Norway), Usuda (Japan), Wettzell (Germany), Yaragadee 
(Australia), and Yellowknife (Canada).

The wet troposphere zenith delays at each site were modeled as a random walk in 
time with a 1 cm^2/hour variance derivative, and were mapped to satellite 
elevations using the Lanyi mapping function. Their nominal values were 0.1 m at 
every site and were estimated with 500-m a priori constraints.

The zenith dry delays were assumed fixed at 2 m for every site, and were also 
mapped to satellite elevations using the Lanyi mapping function. (Because the 
wet and dry delays are nearly degenerate parameters, the estimated wet delay in 
fact represents the combined delay which deviates from the nominal 2.1 m.)

Clocks for each station and satellite (except for the reference clock, taken as 
the maser-based clock at Fairbanks, and believed accurate to ~1 microsecond or 
better) were estimated as a white-noise process with updates every 6 minutes.

The carrier phase ambiguities were estimated as real-valued parameters.

Nominal values for X and Y polar motion and UT1R-UTC were obtained from 
interpolation in the IERS Bulletin B predicted values. Estimates of polar motion
had weak (5-m) constraints, while UT1R-UTC was constrained (0.01 mm) to be the 
nominal.

The Williams solid Earth tide model was used. Pole tide and ocean loading were 
not modeled.

The Earth's gravity field was described by the GEM-T3 multipole expansion using 
terms up through degree and order 8. (IERS-recommended values for C21 and S21 of
-0.17E-09 and 1.19E-09 were also used.) The value of GM was taken as 398600.4415
km^3/s^2.


Description of the Terrestrial System for Solution EOP(JPL)92P01
========================================================================
(1)  Technique:                       GPS
(2)  Analysis Center:                 JPL
(3)  Solution identifier:             EOP(JPL)92P01
(4)  Software used:                   GIPSY/OASIS-II
(5)  Relativity scale:                Local Earth
(6)  Permanent tide correction:       No
(7)  Tectonic plate model:            None (single epoch solution)
(8)  Velocity of light                299792458 m/sec
(9)  GM                               3.986004415E+14 m^3/s^2
(10) Reference epoch                  1992.5
(11) Adjusted parameters:             Station X, Y, Z at 1992.5
                                      PMX, PMY every 24 hours;
                                      GPS epoch state and 3 solar
                                      radiation biases every 24 hours;
                                      Random walk zenith tropospheres;
                                      White noise clocks (Fairbanks
                                      H-maser is the reference clock);
                                      Carrier phase ambiguities as
                                      real valued (not bias-fixed)
(12) Definition of origin             Defined though a realization of ITRF90
                                      and by fixing gravity field to GEM-T3.
(13) Definition of orientation        Defined through a realization of ITRF90 
(14) Constraint for time-evolution    No time evolution.

==============================================================================

Submission Format
==============================================================================
Field Contents (Value of 0 indicates parameter not available)
  1   MJD of the measurement
  2   x of the pole (")
  3   y of the pole (")
  4   UT1-UTC (s)
  5   uncertainty on x (")
  6   uncertainty on y (")
  7   uncertainty on UT (s)
  8   rms residual of the least squares solution providing the EOP
  9   correlation coefficient : x, y
 10   correlation coefficient : x, UT1
 11   correlation coefficient : y, UT1
 12   number of contributing stations
 13   number of contributing satellites
 14   number of passes

  MJD       X        Y       UT1R-UTC XSIG    YSIG   UTSIG  RMS  XYC  XUC  YUC 
STA SAT PASS
           ''        ''        sec     ''     ''      sec
 48794.5  -.15231   .35740   .00000  .00025  .00033  .00000  .0  .16  .00  .00
25 17  0
 48795.5  -.15059   .35946   .00000  .00023  .00029  .00000  .0  .23  .00  .00
27 17  0
 48796.5  -.15110   .35929   .00000  .00024  .00029  .00000  .0  .20  .00  .00
27 17  0
 48797.5  -.15060   .36119   .00000  .00032  .00036  .00000  .0  .04  .00  .00
25 17  0
 48798.5  -.14894   .36726   .00000  .00027  .00033  .00000  .0  .22  .00  .00
22 17  0 
 48799.5  -.14740   .37000   .00000  .00024  .00029  .00000  .0  .29  .00  .00
24 17  0
 48800.5  -.14558   .37239   .00000  .00023  .00029  .00000  .0  .29  .00  .00
27 17  0
======================================================================

The precise ephemeris file for the week of June 21 contains GPS satellite 
coordinates as a function of time (every 15 minutes) in the ITRF90 coordinate 
system, based on the propagation of the estimated 9-element state vector. Users 
of these ephemerides should take care to fix one of their network's receivers to
coordinates compatible with ITRF90, otherwise systematic errors will result. 
The Fairbanks receiver (attached to a H-maser) was used to define reference 
time, which is known to be aligned with GPS time to better than 1 microsecond (<
4 mm of orbital motion). Since satellite clock solutions are not distributed in 
these precise ephemerides, users of these orbits must be especially sure that 
their reference time is consistent with GPS time. The ephemerides are given in 
the NGS sp1 format which contains tabulated positions and velocities every 15 
minutes.

For each day and satellite, a measure of orbit quality is calculated by 
computing the satellite's position as a function of time in two independent 
ways. The first uses tracking data from the same day, while the second uses 
tracking data from either the previous or following day, whichever is closer to 
the time of the position estimate. The 3-dimensional rms difference of these two
estimates over the first three and last three hours of each UTC day is computed 
for each satellite. (This 6-hour comparison period is half an orbital period).

The following diagram illustrates this method for the "second day". The scale is
such that each character represents one hour, and time increases from left to 
right.

                           |||                  |||
   <-------first day------><-------second day-----><-------third day------>
...111111111111111111111111111111111111
               222222222222222222222222222222222222222222222222
                                       333333333333333333333333333333333333...

Times marked with "1" contain estimates of the satellite's position based on 
data from the first day (ellipses on the left are used to indicate that such 
estimates extend prior to the beginning of the first day). Times marked with "2"
contain estimates of the satellite's position based on data from the second day,
and similarly for times marked with "3" and data from the third day.

The differences in estimates at times marked with a vertical bar ("|", estimate 2 minus estimate 1 for the first three hours of the second day, and estimate
2 minus estimate 3 for the last three hours of the second day) are used to
calculate the 3-d rms for the second day.

Since this method involves orbit prediction, the results represent pessimistic 
estimates of the accuracy of the tabulated orbits. Note that a "bad" solution 
on, say, day 3 would give a large rms value for days 2 and 4 and a larger rms
for day 3 (assuming days 2 and 4 are good). Users of our precise orbits may use 
these tables to decide whether to reject certain satellite data, or to derive 
relative data weights.

      92  6 21   92  6 22   92  6 23   92  6 24   92  6 25   92  6 26   92  6 27
prn    rms (m)
  2       1.00       1.06       1.35       1.10       0.70       0.61       0.9
  3       1.16       1.12       1.51       0.80       0.82       0.68       0.6
 11       1.46       1.84       5.42       4.58       2.77       0.54       0.4
 12       3.98       1.80       4.13       4.21       0.93       0.59       0.5
 13       0.51       0.77       1.00       1.40       0.58       0.56       0.7
 14       0.55       0.48       0.90       0.64       0.63       0.27       0.3
 15       0.81       1.53       1.37       1.55       1.55       1.13       1.5
 16       0.39       0.55       0.63       0.85       0.40       0.32       0.3
 17       1.21       0.88       1.23       1.24       1.19       1.06       0.5
 18       0.48       0.76       1.26       0.91       0.49       0.57       0.3
 19       2.49       2.02       0.62       0.62       0.45       0.78       0.2
 20       0.73       0.88       1.15       0.47       0.56       0.58       0.8
 21       0.57       0.64       1.92       1.27       0.68       0.62       0.3
 23       0.38       0.37       0.29       0.76       0.91       0.29       0.2
 24       0.80       0.87       0.92       0.64       0.47       0.59       0.7
 25       0.90       1.97       1.87       2.32       3.59       1.51       1.2
 28       1.99       2.81       3.68       1.60       2.29       1.39       1.2

(Because June 28 has not yet been processed, the above column for June 27 is the
result of a 3-hour comparison with an extrapolation from June 26 only. This is 
indicated by the single decimal place in the table.)

Finally, the following table compares our precise orbits with those from the
broadcast ephemeris (3-dimensional rms over one UTC day, which is not the same 
definition of "rms" as in the previous table).


      92  6 21   92  6 22   92  6 23   92  6 24   92  6 25   92  6 26   92  6 27
prn    rms (m)
  2       4.3        6.4        6.1        8.0        4.5        6.0        6.1
  3      12.2       13.3       14.8       14.2       10.7       12.8       11.8
 11      24.6       15.3       15.6       28.1       31.9        9.1        6.0
 12       6.1       28.7       11.4       22.2       10.2        6.3        7.4
 13       7.5        7.2        4.6        4.4        6.0        6.4        6.6
 14       3.2        3.1        4.2        3.3        4.0        3.9        3.5
 15       4.8        5.0        6.8        7.6        4.3        5.7        5.8
 16       5.6        5.1        4.1        5.1        6.7        6.7        4.9
 17       7.8        5.8        6.1        5.6        6.9        7.4       10.3
 18       8.0        7.9        6.3        9.1        7.7        6.8        6.9
 19      13.6       16.0       12.3       11.8        9.7        8.9       11.0
 20       5.6        4.3        5.9        6.0        7.4        6.0        6.1
 21       5.7        4.8        5.3        4.4        3.3        3.1        3.8
 23       2.4        3.5        5.5        3.6        4.1        4.0        5.2
 24       9.2       10.2       10.1        9.8       10.5       14.2       13.9
 25       9.9        7.7        5.6        9.3       13.7       18.0       24.3
 28       4.6        4.5        4.9        5.8        5.2        5.6        6.2




More information about the IGSMail mailing list