[IGSMAIL-1705] CODE Analysis Changes
T.A.
T.A.
Fri Oct 17 05:57:26 PDT 1997
******************************************************************************
IGS Electronic Mail Fri Oct 17 5:57:26 PDT 1997 Message Number 1705
******************************************************************************
Author: T.A. Springer
Subject: CODE Analysis Changes
Dear Colleagues,
Starting with GPS week 0926 the CODE Analysis Center has made essential
changes to its routine processing:
1) The elevation cut-off was changed from 20 to 10 degrees.
2) An elevation-dependent weighting of the observations is applied
(1/cos(zenith)**2).
3) The Niell mapping function (dry) is used to map the tropospheric
zenith delays.
4) We adopted the solid Earth tide model as given in the IERS 1996
conventions. Only ocean loading is not yet applied.
We are currently updating our "Analysis Center Questionnaire" and will soon
submit the updated version to the IGSCB.
Although the change was mainly based on tests which we performed with our
European regional (EUREF) solutions we, of course, studied the influence
on our global (IGS) products using a couple of days worth of data.
Observed changes in the orbit and EOPs:
---------------------------------------
Orbit EOPs
-----------------------------------
rot x. 0.03 mas x: 0.06 mas
rot y. -0.07 mas y: 0.02 mas
rot z. -0.01 mas ut: 0.01 us/day
tra z. 3 mm
-----------------------------------
>From the above table we conclude that there are no significant systematic
changes in both our orbit and EOP estimates.
For the station coordinates we see the following relatively large changes:
- scale decreased by 0.8 ppb. This is caused mainly by the change of the
tropospheric mapping function.
It must be noted that the change in cut-off elevation (20 to 10 degrees)
changes the scale significantly but the elevation-dependent weighting
counteracts this effect.
- repeatability of the coordinates improves significantly,
especially the height component.
- Unfortunately some station coordinates will show a jump in our time-series!
This is, in our opinion, mainly caused by multipath and antenna phase
center variations. This means that those stations for which we see a
jump are probably in a bad multipath environment or have a very different
antenna setup. For the following stations we see a systematic
change in the coordinates larger than 15 mm (the values in brackets are
the observed systematic height changes in mm.):
MDVO (+30), TAIW (-20), IRKT (+25), MCMU (-20)
It should be noted that MDVO has a "special" antenna without a ground
plane. So it is very likely to have strong multipath effects. Furthermore
we have no antenna phase center variations for this antenna which will
cause a bias as well.
The change of the cut-off elevation angle was initiated by the considerable
improvements found in our European regional network solution (for EUREF).
There are several publication about the results of our European processing
which may be of interest to you:
1) Processing Strategies for Regional GPS Networks
M. Rothacher, T.A. Springer, S. Schaer, G. Beutler
to be published in the proceedings of the IAG General Assembly in
Rio de Janeiro in September 1997.
2) The Role of GPS in the Study of Global change
M. Rothacher, G. Beutler
submitted to Physics and Chemistry of the Earth
3) EUREF Activities at the CODE Analysis Center
T.A. Springer, W. Gurtner, M. Rothacher, S. Schaer
to be published in proceedings of the International Seminar on GPS in
Central Europe, Penc, Hungary.
NOTE: papers 1 and 3 can also be found on our anonymous ftp in the directory
"PAPERS" with the names "RIOPAP97.PS" and "EUREF97.PS"
In the above publications we have shown that lowering the cut-off elevation
significantly improves the repeatability of the estimated station coordinates.
In general the height component significantly improves whereas the
horizontal components get slightly worse. The elevation-dependent weighting
retains the improvement of the height component but leads to an
improved horizontal repeatability.
Furthermore we have shown (1) that with a cut-off angle of 5 degrees the
estimation of tropospheric gradients improves the horizontal repeatability
by about a factor of 2!
However, there were two factors which made us decided to lower our cut-off
elevation to "only" 10 degrees and not to estimated tropospheric gradients
for the global network:
1) The official IGS list with antenna phase center variations only
contains values down to 10 degrees. Since especially the DORNE MARGOLIN
antennas have a large change in the phase center position at low
elevation angles this would lead to systematic errors in our coordinate
estimates.
2) There are very few stations tracking below the 10 degrees.
Note that it was already requested once to lower the cut-off elevation
to 5 degrees (IGSMAIL 1586), especially for the European stations.
The following stations do not track below 11 degrees (the list
contains surprisingly many European sites):
OBSERVED ELEVATION
STATION MIN MAX
--------------------------
GOPE 11502M002 14.9 89.7
LAMA 12209M001 12.0 89.5
MASP 31303M002C 12.3 89.0
UPAD 12750M002 14.0 89.3
VILL 13406M001 13.5 89.0
CRO1 43201M001 13.5 86.6
KIRU 10403M002 11.0 73.9
LPGS 41510M001 11.7 89.3
BRAZ 41606M001 12.6 88.4
--------------------------
*******************************************************************************
*******************************************************************************
** **
** We would like to STRONGLY recommend that ALL STATION MANAGERS lower the **
** cut-off elevation in the receivers to 5 degrees or lower! **
** **
*******************************************************************************
*******************************************************************************
NOTE FOR THE GNAACS: The scale of our covariance matrix may have changed!
------------------- Although the change should be small please be aware.
Furthermore we have adopted the DOMES numbers as provided by Zuheir Altamimi
for all sites in our Network (only for the site MOIN there is no DOMES number).
Kind regards,
The CODE Analysis Team:
Markus Rothacher
Stefan Schaer
Tim Springer
[Mailed From: Tim Springer <SPRINGER at aiub.unibe.ch>]
More information about the IGSMail
mailing list