HELMERT ESTIMATION: DATA AND SCRIPTS
====================================



1. DATASET INFORMATION
----------------------

This dataset accompanies the journal article: 
	Sensitivity of Antarctic GPS velocity estimates to terrestrial reference frame transformation strategies. 
	Under review in Journal of Geodesy.

Authors:	Jenan Rajavarathan, Matt King, and Christopher Watson
Affiliation: 	School of Geography, Planning, and Spatial Sciences, University of Tasmania, Hobart, Tasmania, Australia.

Description: 	This dataset provides daily coordinate and covariance inputs from 01-01-1995 to 31-12-2025, Python scripts, and previously generated Helmert transformation output files.



2. FILE ORGANISATION
---------------------

2.1 nf_frame_files.tar

This archive contains the daily coordinate inputs, organised as follows:
    nf_frame_files/
    +-- YYYY/
        +-- YYYY-MM-DD/
            |-- predicted_frame_YYYY-MM-DD.gdcov
            +-- projected_YYYY-MM-DD_nf.gdcov

YYYY-MM-DD is the date (four-digit year, two-digit month, two-digit day).

GDCOV files (\*.gdcov) are geodetic covariance files used by GipsyX.

  predicted frame files: 
    Contain the reference coordinates and their covariance information, derived from the ITRF2020-u2024 SINEX files using the ITRF2020 trajectory model.
	
  projected non-fiducial files: 
    contain the daily JPL non-fiducial GNSS network coordinates and their associated covariance information, projected onto a minimally constrained datum.


2.2 xfiles_generated.tar.gz

The archive contains .x files for each strategy in separate folders according to the transformation strategy. 

These files were generated using the Python scripts provided in this dataset and the corresponding daily coordinate and covariance inputs described in Section 2.1.


2.3 Python scripts

The following scripts are provided and should be located in the same directory:
	
  generate_xfiles.py 
    Estimates daily Helmert transformation parameters and generates `.x` output files. Users only need to run this script to test outputs.
	
  student_t_extension.py
    Implements the Student-t robust reweighting method.
	
  gdcov_to_vcv.py	
    Converts GDCOV files to variance-covariance files.


2.4 Additional input data requirement

The SINEX files corresponding to the ITRF2020-u2024 solution are required to reproduce the results presented in the paper. 
SINEX files are available from https://files.igs.org/pub/station/coord/IGc20/.



3. RERUNNING THE HELMERT ESTIMATION
-----------------------------------

3.1 Preparing input data

Extract nf_frame_files.tar into the directory containing the scripts. The extracted nf_frame_files folder must be located beside generate_xfiles.py.

Input and output paths are already configured relative to the script location. 
The processing period is set in generate_xfiles.py:

  START_DATE = "1995-01-01"
  END_DATE   = "2025-12-31"


3.2 Selecting the transformation strategy

Select the required strategy in the USER CONFIGURATION section of generate_xfiles.py. 
Use the settings below and leave the other tuning parameters/settings unchanged.
	
  Strategy 1: Seven-parameter WLS
    NO_SCALE = False
    ROBUST = False
    APPLY_HEIGHT_WEIGHTING = False
    APPLY_SPATIAL_WEIGHTING = False

  Strategy 2: Six-parameter WLS
    NO_SCALE = True
    ROBUST = False
    APPLY_HEIGHT_WEIGHTING = False
    APPLY_SPATIAL_WEIGHTING = False
 
  Strategy 3: Height-reweighted six-parameter WLS
    NO_SCALE = True
    ROBUST = False
    APPLY_HEIGHT_WEIGHTING = True
    HEIGHT_FACTOR = 1.5
    APPLY_SPATIAL_WEIGHTING = False

  Strategy 4: Huber estimation
    NO_SCALE = True
    ROBUST = True
    ROBUST_METHOD = "huber"
    APPLY_HEIGHT_WEIGHTING = False
    APPLY_SPATIAL_WEIGHTING = False

  Strategy 5: Student-t estimation
    NO_SCALE = True
    ROBUST = True
    ROBUST_METHOD = "student_t"
    APPLY_HEIGHT_WEIGHTING = False
    APPLY_SPATIAL_WEIGHTING = False

  Strategy 6: Spatial-reweighted six-parameter WLS
    NO_SCALE = True
    ROBUST = False
    APPLY_HEIGHT_WEIGHTING = False
    APPLY_SPATIAL_WEIGHTING = True

After selecting the required strategy, save the configuration and execute: 
  
  python3 generate_xfiles.py.

Estimated daily Helmert transformation parameters are saved as .x files in strategy-specific output directories.



4. COMPARING RESULTS
--------------------
The provided .x files can be compared by extracting xfiles\_generated.tar.gz into a separate directory.