12345678901234567890123456789012345678901234567890123456789012345678901234567890

This is an example of how to generate time series statistics of temporal
correlations using tsfit's "real_sigma" option (see Floyd and Herring, 2020) and
sh_gen_stats.

Floyd, M. A., and T. A. Herring (2020), Fast Statistical Approaches to Geodetic
  Time Series Analysis, in Montillet, J.-P., and M. Bos (eds), Geodetic Time
  Series Analysis in Earth Sciences, Springer Geophysics, Springer, Cham,
  Switz., doi:10.1007/978-3-030-21718-1_5.


1. Download a .pos-file if you do not have one available from your own
   processing, e.g.

curl -s -f -L -R -O https://data.unavco.org/archive/gnss/products/position/P617/P617.cwu.final_igs14.pos


2. Write a simple tsfit command file, including the "real_sigma" option, e.g.

cat << END > tsfit.cmd
 eq_file ~/gg/tables/igb14_comb.eq
 eq_file ~/gg/tables/All_NOTA_ants.eq
 eq_file ~/gg/tables/All_NOTA_unkn.eq
 eq_file ~/gg/tables/All_NOTA_eqs.eq
 real_sigma
END


3. Run tsfit, e.g.

tsfit tsfit.cmd tsfit.sum P617.cwu.final_igs14.pos >& tsfit.log


4. Run sh_gen_stats, e.g.

sh_gen_stats -ts tsfit.sum -or demo


You will then have a series of *_demo.* files, the most commonly used of which
is rw_demo.rw, which contains "mar_neu" commands for globk and can be used by
including "source rw_demo.rw" in globk.cmd. Do this for all sites in your
processed network simultaneously, not one-by-one, when generating a full
velocity solution.


  Michael Floyd (2021/08/10, MIT)

12345678901234567890123456789012345678901234567890123456789012345678901234567890