NMT GAGE Processing Strategy

Mark H. Murray, AC Director
January 31, 2014
----------------------------

The New Mexico Tech (NMT) GAGE Analysis Center uses the GAMIT/GLOBK GPS analysis software.

http://facility.unavco.org/kb/categories/Software/Processing+Software/

provides a brief description of GAMIT/GLOBK, and information regarding access and support for
the software package. NMT processing currently uses GAMIT/GLOBK Version 10.5. Results
prior to January 2, 2014 were generated using Version 10.4.

This README file provides a summary of the processing strategy used by NMT to provide
SINEX file products containing daily station position estimates for the GAGE project.
Included with this file are standard format station and model information files used by
GAMIT/GLOBK to process RINEX GPS station observation files and convert the estimated 
parameter results to SINEX.

NMT uses the shell scripts provided in the standard GAMIT/GLOBK distribution to call the 
executables used in the processing. It is assumed that the user is familiar with the
general procedures used by GAMIT/GLOBK to process the observations, which are well 
documented in manuals and examples provided in the distribution. This README file  
gives the specific calls made to the shell scripts that would enable the user to 
reproduce the NMT products.

STEP 1: Obtaining the RINEX files

The SINEX files provide a list of the stations processed for each day, which 
represent data available from the master site list given in the sites.defaults file.

RINEX files for these stations are downloaded from three archives in the following 
order: unavco, cors, and sopac.

Here are calls for day 100 of year 2000, which will be used as an example throughout this
README file.

sh_get_rinex -archive unavco -yr 2000 -doy 100 -sd sites.defaults -expt [rapd|rap1|rap2]
sh_get_rinex -archive cors -yr 2000 -doy 100 -sd sites.defaults -expt [rapd|rap1|rap2] -ftp_prog wget
sh_get_rinex -archive sopac -yr 2000 -doy 100 -sd sites.defaults -expt [rapd|rap1|rap2] 

Due to csh limitations, the list is divided into 3 "experiments" (rapd, rap1, rap2) so the
above calls need to be performed for each case.

STEP 2: Obtaining orbits

NMT uses the IGS rapid and final orbit products, obtained by:

sh_get_orbits -archive cddis -yr 2000 -doy 100 -eop_ser [usno|usnd] -max_fit_tol 1

Rapid orbit solutions (using the usnd rapid earth orientation parameter files) are 
generated every day. Final orbit solutions (using the usno eop files) are generated
weekly after the IGS final orbit products become available. These are supplemented by
additional 12- and 26-week old solutions that acquire data that became available after 
the initial rapid and final solutions were performed.

The orbit parameters are held fixed to the IGS values by setting 
"Choice of Experiment = BASELINE" in the sestbl. file, and 
by fixing these parameters in the GLOBK combinations.

STEP 3: Defining subnets

GAMIT places a limit on the number stations that can be simultaneously processed.
The entire GAGE network of > 1800 stations is therefore divided into subnets that 
share common sites between nearby subnets to "tie" the solutions together. 
A "base" subnet containing at least 1 station from ever other subnet is also constructed
to provided additional stability to the entire network combination. The GAMIT/GLOBK 
utility netsel is used to define the subnets dynamically for each day.

netsel -f rinex.list -v lfile.vel -n 45 -t 1 -s station.info > sites.defaults.2000.100

where 
  rinex.list is a list of rinex files generated with ls -s <rinex files>
  lfile.vel is a GLOBK velocity file generated with apr_to_vel pbo_itrf08.apr lfile.vel
  45 = nominal number of sites in each subnet
  1 = number of tie sites (must be > 0)

The output sites.defaults file now includes experiments for each subnet called
ne01, ne02, etc.
  
The nominal number of 45 sites in each subnet with 1 tie site are typical values
for the current size of the entire network. These numbers have varied in the past,
particularly in pre-2005 solutions where the entire network had considerably fewer
stations and including more tie sites was useful for improving stability.

STEP 4: Running GAMIT

Processing directories (e.g., $procdir) for each subnet are created, and set up by:

sh_setup -yr 2000 -update links -topt all

and the additional necessary processing files supplied with this README file 
are copied into the tables directory. For the sestbl., the files are renamed:

cp sestbl.rapid $procdir/tables/sestbl. # used for rapid solutions
cp sestbl.final $procdir/tables/sestbl. # used for final solutions

Note, the final solutions use VMF1 mapping function grid information, which is 
typically not available in time for rapid solutions.

GAMIT processing of each subnet is then invoked, for example, by:

sh_gamit -d 2000 100 -expt ne01 -netext ne01 -yrext -orbit IGSF -metutil Z 

Note, the metutil option is used to output tropospheric estimates for each station.

STEP 5: Running GLOBK

After all the subnets have been processed, GLOBK is run to combine all the hfile results 
into a single binary GLX solution file.

sh_glred -d 2000 100 -yrext -net $nets -expt nmtbr -opt R LA H G

where
   $nets is a list of all subnet experiments (i.e., ne01 ne02 etc.)
   the b in "nmtbr" is a single character denoting the type of solution (a=rapid, b=final, etc.)
   the r in "nmtbr" is the usual GAMIT/GLOBK convention for free (r) vs. fixed (x) solutions

This creates the combined binary solution file H000409_nmtbr.GLX

Until summer 2013, the NMT processing strategy used "Choice of Experiment = RELAX." in the sestbl.,
so that this GLX file contained loosely constrained orbit parameters. These solutions were 
then converted to fixed-IGS solutions using globk:

ls H000409_nmtbr.GLX > L000409_nmtbx.gdl
globk 6 globk_nmtbx_00100.prt globk_nmtbx_00100.log L000409_nmtbx.gdl globk_fix.cmd

This creates the combined binary solution file H000409_nmtbx.GLX

This second step is still performed, although it should be unnecessary when using 
"Choice of Experiment = BASELINE" in the sestbl. file.

STEP 6: Create products

Three products are created: 

1) SINEX of daily site positions and covariance

Conversion of the combined GLX file to SINEX is straightforward:

   glbtosnx . H000409_nmtbx.GLX nmt10570.20000409.b.snx
    
2) Phase RMS files

Phase RMS files are created from autcln.post.sum files archived in the $adir directory:

   sh_exstats -d $adir/2000_100 -o nmt10570.20000409.b.rms
   
3) Tropospheric parameter files

Tropospheric parameter files are created from met_* files archived in the $adir directory, 
which were created using the sh_gamit -metutil option above:

   sh_met2pbo -o nmt10570.20000409.b.met -f $adir/2000_100*/met_*
   
These 3 product files are then uploaded to the Analysis Center Coordinator (ACC) at MIT,
where the final combined products are then created and made available on the UNAVCO 
Facility web site.

Files included with this README
-------------------------------  
pbo_itrf08.apr		Apriori coordinates file
process.defaults	Example process.defaults files to be edited for local processing
sestbl.final		sestbl. for final processing
sestbl.rapid		sestbl. for rapid processing
sites.defaults		all sites used in GAGE processing
sittbl.			Apriori constraints on site coordinates
station.info		station metadate used in GAGE processing
globk_comb.cmd		globk command file used in sh_glred
glorg_comb.cmd		glorg command file used in sh_glred (standard)
globk_fix.cmd		globk command file used to fix orbit parameters
