Continuous processing results
NOTE: Suominet net sites SA09 and SA13 added 30 Jan 2002, see Processing
changes at bottom of page.
The time series of station position estimates are generated
in two analyses using globk/glorg from GAMIT runs which process
both networks simulataneosly. The autcln
command file is let unchanged each data and the
sestbl. file is used to control the GAMIT
runs. The main features here are use of 30 second data for cleaning
which is decimated to 2 minutes for the final runs. The Use N-file
feature in the sestbl. file set the use of a postfit, elevation angle
dependent phase data noise model. The sittbl. template
sets the apriori constraints on the station coordinates during the
GAMIT run. These constraints are removed before processing in globk.
During processing the atmospheric pressure, temperature and relative
humidity are replaced by values computed from NCEP analysis fields.
We used the Scripps Orbit and Permanent Array Center
(SOPAC) final g-files obtained
from the SOPAC combinations
ftp area by GPS week number.
For the Californian network
glorg is used to estimate daily translations of the network with
the orienataion constrained to the IERS Bulletin A values. Over the
125 km extent of this network, 1 mas error in Earth orientation
should change the relative coordinates by <1 mm.
The California globk command file and
glorg command file can be viewed
with these hyper-links. globk.usesite contains
the name of the sites used in the globk solution (used to eliminate
un-wanted sites). glorg.cali.usesite
and glorg.cont.usesite contain
the names of the sites used for the origin constraint in the California
and Central Asia glorg runs.
The current California RMS scatters
are updated daily.
For the Central Asia network
which spans 1800 km we estimate only
translation in glorg (same as the smaller California network).
The Central Asia globk command file and
glorg command file can be viewed
with these hyper-links. The same rotational constraints are applied
as California. (These constraints were previously 10-times larger).
The current Central Asia RMS scatters
are updated daily. These summary files include both the orginal larger
Central Asia network (with translation and rotation) and the
smaller network (with only translation).
In both analyses, the SIO orbits are constrained to +-100 mm in
position and 0.01 mm/sec in velocity. Radiation parameter adjustments
are constrained to 1% of the direct radiation pressure.
The data from Central Asia is collected in collaboration with the
Russian Academy of Sceinces Institute for High Temperature Physics (IVTAN).
The IVTAN web page contains more information
about the continuous sites and the GPS campaigns that have been carried out
in the region.
The results shown on these pages were made possible by support
from the National Science Foundation
(NSF), the National Aeronautics and
(NASA), the Southern California Earthquake Center
and the University Navstar Consortium
Day 096, 1998
||Introduced site-dependent data noise computed from postfit
residuals with an elevation angle dependent model.
Day 119, 1998
|Introduced estimation of atmospheric gradients with constraint
of +-10 mm at 10 deg elevation angle.
Day 334, 1998
|Changed the stations in the Central Asia Network to cover an 1800-km
region only. (The original 6000-km network can be seen
Day 072, 1999
|Introduced site-dependent daily-mean-atmospheric delays into GAMIT processing.
These delays are computed from the NCEP global forecast models.
A new set of PNGs are included for each station showing the atmospheric delays
since the beginning of 1999. The red-line on these plots shows the apriori
delay being used in the processing. The green line is the RMS scatter of the
phase residuals. The lower plot shows the results for the last 30 days.
Day 101, 1999
|Reduced the elevation-angle cutoff from 15 degrees to 10 degrees. The
elevation angle dependence of the phase noise means that the effective cutoff
angle is still somewhat higher than 10 degrees.
Day 309, 1999
|No data processing changes. We converted all of our GIF images
to PNG (Portable Network Graphics) format so as not be in violation of the
Unisys/Compuserve patent on GIF. Use of GIF images, without paying a
licensing fee can result in $5000 fines. NOTE: Many of our images still
use the .gif extension but are PNG format. Netscape shows the correct
format type; Microsoft Internet Explorer, while correctly displaying the
PNG image, reports the image as being a gif image. Viewing the page
source of the images shows the PNG header.
Day 310, 1999
|Modified the bias fixing sigma in sestbl. Increased values
from 0.15 cycles to 0.25 cycles for both the WL and NL fixing. First
analysis effected is day 292, 1999.
Day 314, 1999
|Changed the "ionospheric constraint" in sestbl. to 8 ppm due
to increasing solar activity. First analysis effected is day 297, 1999.
Day 161, 2000
|Started to process with IGS Final orbit rather than the SOPAC final
gfile. Also converted time series from days since 1998 to years.
Day 224, 2000
|Started to apply ocean tide models based on Scherneck's tables to
the processing. Stations not the standard tabular form are interpolated from
Day 030, 2002
|Added two stations from the SuomiNet
network. The sites are Trimble 4700 recievers with SA09 having a TRM33429.00+GP, L1/L2 microcentered
Compact antenna with Ground plane, antenna and SA13 having a TRM29659.00 Choke ring antenna. These
sites are distant from the California network we process
and so geodetic quality of the positions
is not our primary aim. Being at the end of a large lever aim to the rest of the network their position
estimates are senstive to small Earth orientation errors.
The atmospheric delay plots for these sites show the 6-hourly estimates from the NCEP
analysis field (in red; other sites show the daily mean value). We save the statistics of the RMS scatter
of the phase residuals and the number of phase measurements as a function of elevation angle.
RMS phase residuals by date and elevation angles
Number of phase residuals by date and elevation angles
(Multiply fractional part of year by 365 to get day-of-year)
Day 189, 2002
|Added a "Time Series Data" link to the results page to allow access to the time series in ascii