Empirical Site Model for GPS Processing

Create an empirical model from GPS undifferenced phase residuals obtained from a GAMIT processing run to mitigate multipath and site-specific effects.

#==============================================================================

#============================================================================== gamit > 10.5 python 2.6 or greater numpy scipy > 0.12 matplotlib

#==============================================================================

#============================================================================== sklearn > 0.15 (needed for functions in esmML.py) - available from http://scikit-learn.org/stable/index.html - or use pip install sklearn

#==============================================================================

#============================================================================== esm.py esmML.py GamitStationFile.py antenna.py residuals.py gpsTime.py

#==============================================================================

#==============================================================================

Copy python files into ~/gg/com cp esm.py ~/gg/com/ cp GamitStationFile.py ~/gg/com/ cp antenna.py ~/gg/com/ cp residuals.py ~/gg/com/ cp gpsTime.py ~/gg/com/

Set your python path to include this directory, for csh:

setenv PYTHONPATH ~/gg/com.

#==============================================================================

#==============================================================================

In order to run the ESM program, GAMIT needs to be configured correctly.

In your sestbl. ensure the following options are selected:

In your autcln.cmd file we recommend using:

POST  phs_res_root DPH SINGLE

This will output all of the DPH one way residuals files for each 
satelllite into one single file for each site processed.

#=======

To use the sitemodels generated by this program in GAMIT Copy the sitemodels antmod.SSSS to project/tables/

In Sestbl. :  

Use site-specific antenna model = Y;

We had to update two files:

gg/gamit/model/open.f
gg/gamit/model/get_antinfo.f
~/gg/com/links.day needs the following line added:
ls -1 ../tables/antmod.???? | xargs -l1 -I % ln -s % ./ >&! /dev/null

#==============================================================================

#==============================================================================

Overview of workflow:

Assuming the station file and antmod file is available in ~/gg/tables/ (if not use --station_file and --antex to pint to the right path)

-Convert the GAMIT .DPH files into a consolidated file that contains the epoch, azimuth, elevation, LC resiudal and prn

python ~/gg/com/esm.py --trav /short/dk5/repro2/2012/ --site YAR2 --network yyyy_dddnN --save_file /short/dk5/repro2/sitemodel/2012/YAR2.2012.DPH.gz

--trav the GAMIT project directory structure to look through to find DPH 
       files

--network how the GAMIT project sub directories are formated, currently 
          supports:
          yyyy_dddnN (2012_001n2/)
          ddd        (001) (not tested yet)

--site 4 character if of the station to create an ESM for (upper case or 
         lower case, no worries)

--save_file filename to save the consolidated residuals to, I use and 
  extension of .CL3 to distinguish them from the GAMIT .DPH files (but 
  this is not required)

-Create the ESM from the consolidated file and save it in antex format:

python ~/gg/com/esm.py --model --site yar2 -f ./t/YAR2.2012.CL3

--site 4 character if of the station to create an ESM for (upper case or 
         lower case, no worries)

--model    create the esm model

-f         path to the consolidated phase residual file
 
This will read the consolidate residual file, do some basic outlier 
detection, then do a blockmedian on 0.5 grid for all of the phase residuals,    and then save the result to antmod.sss