[IGSMAIL-306] Replacement of CODE orbits

[M.]

IGS Electronic Mail       10-AUG-1993 18:52:46       Message Number 306
***********************************************************************
 

>From: M. Rothacher
Subj: Replacement of CODE orbits
--------------------------------



Corrections to 16 CODE Precise Orbit Files
------------------------------------------

In the last few months new GPS orbit models were implemented and tested at
the CODE processing center.  In particular a program ORBIMP, interpreting
the satellite positions in the precise IGS orbit files as pseudo-
observations, was written.  The orbit model allows to use arc-length up to,
let us say three weeks, where the rms of the satellite position after the
orbit determination still is of the order of 10 - 30 cm; 15 orbit
parameters (six initial conditions and nine parameters related to radiation
pressure are necessary for such a performance).  More information may be
found in (Beutler et al., 1993).

One application of ORBIMP was to perform an extensive quality check with
the CODE orbits delivered to the IGS network centers:  Starting with week
677 (27 Dec 92-2 Jan 93) up to the current week 707 (25 Jul-31 Jul 93) the
satellite positions in the seven files COD0WWWi.EPH, i=0,1,..,6 (WWW=week
number) were fitted with one arc of seven days, where the 15 parameters
(per satellite) mentioned were adjusted for each week separately.  This
quality check will now be performed each week before the results are
released to the IGS network centers.

It was one of the objectives of this quality check to produce a reliable
table of "unusual" events (manoeuvres, sudden velocity changes due to
momentum dumps etc).  Unfortunately we also found strange events on some
days for some of the eclipsing satellites, which had nothing to do with the
satellite system but with our software system.  In view of the quality of
the other data the problem really showed up very clearly when the residuals
of the one week arcs were inspected:  it could happen that on a single day
the satellite residuals started moving in along track direction from zero
to about 5 m (!) to be again of the order of about 10-20 cm for the
following days.  The bug was introduced with a software release in November
1992.  It could affect only the second (and final) iteration of our
processing.  After the detection of the problem it was possible to tell
exactly which results were contaminated.  Because the results of the first
iteration are already very close to the final result (the second iteration
is so to speak a "security-round") it was in principle easy to correct the
contaminated files.  Table 1 shows the contaminated CODE orbit files
together with the satellites which were affected for GPS weeks 677 to 699
(after week 700 onwards the problem did not show up or was already
removed).  The satellites in brackets (...) were very slightly affected and
only corrected because the file had to be replaced anyway because of the
satellite(s) listed first.

                       Table 1
                       -------
     Satellites and Files Affected by Software Bug
            (CODE Processing Center)
***********************************************************
|   MJD   WEEK DD DOY    | Satellite       | File         |
|*********************************************************|
| 48985.0  677  2 364    | 12(,25,19,27,24)| COD06772.EPH |
| 48986.0  677  3 365    | 19(,24,25,27)   | COD06773.EPH |
| 48988.0  677  5 001    | 15(,17,24)      | COD06775.EPH |
| 49006.0  680  2 019    | 24(,2,12)       | COD06802.EPH |
| 49022.0  682  4 035    | 20              | COD06824.EPH |
| 49035.0  684  3 048    | 2,23,(14,21)    | COD06843.EPH |
| 49047.0  686  1 060    | 23(,20,21)      | COD06861.EPH |
| 49049.0  686  3 062    | 21(,20)         | COD06863.EPH |
| 49050.0  686  4 064    | 20              | COD06864.EPH |
| 49086.0  691  5 099    | 29(,18,26)      | COD06915.EPH |
| 49106.0  694  4 119    | 26(,18)         | COD06944.EPH |
| 49124.0  697  1 137    | 28(,31)         | COD06971.EPH |
| 49132.0  698  2 145    | 3(,13)          | COD06982.EPH |
| 49134.0  698  4 147    | 13              | COD06984.EPH |
| 49140.0  699  3 153    | 25,31(,19,27,28)| COD06993.EPH |
| 49141.0  699  4 154    | 19,27           | COD06994.EPH |
***********************************************************

The sixteen files in Table 1 were sent to CDDIS and IGN on 10 August 1993.
Our 1993 data should now be clean, the quality is of the order of 10 cm to
30 cm rms per coordinate.  There still may be problems for some of the
eclipsing satellites in GPS weeks 671-676.  We will check these weeks as
soon as possible.  We excuse ourselves for any inconvenience caused by this
problem and advise everybody using some of the precise orbit files listed
in Table 1 to replace them by the files now at CDDIS, IGN, or IfAG.

                       For the CODE Processing and Development Team

                           Gerhard Beutler and Markus Rothacher

REFERENCE :

Beutler, G., E. Brockmann, W. Gurtner, U. Hugentobler, L. Mervart, M.
Rothacher, A. Verdun (1993). "Extended Orbit Modelling Techniques at the
CODE Processing Center of the IGS: Theory and Initial Results." submitted
for publication.