Technical description of the rapid/SINEX solution opa2021a

1. Origin of input data: databases from IVS data center, version 4 or more.

2. Authentic processing steps after data retrieval: use VGOSDB format.
   Pre-2018 data were originally stored in MK3 format in Paris Observatory
   disks. These data were transformed to VGOSDB. Some of them (oldest ones)
   were preliminary re-calced with Calc 11.0. Post-2018 data were directly
   retrieved in VGOSDB format and processed with no modification.

3. Origin of meteorological data: use database content.

4. Use numerical weather model: ECMWF for computing VMF1.

5. Setting of mapping functions: VMF1.

6. Origin of cable cal data: use database content.

7. Application of cable cal flags: use database settings.

8. Selection/settings of geophysical models: replace database settings.

   8.1. Nutation/precession: IAU 2000A nutation and IAU 2006 precession.
   
   8.2. Short-period tidal variations in x, y, UT1: IERS 2010.
   
   8.3. Troposphere mapping functions: VMF1.
   
   8.4. A priori mean gradients: from VLBI data or DAO weather model.
 
   8.5. Solid Earth tide: IERS 2010.
 
   8.6. Ocean loading displacements: FES 2004.

   8.7. Atmospheric loading displacements: APLO.
   
   8.8. Permanent tide correction: applied.

   8.9. Thermal expansion: Nothnagel (2008).
   
   8.10. A priori station positions and velocities: ITRF 2014 or database
         content for new stations.
   
   8.11. A priori source coordinates: positions computed in the opa2021a
         global solution.

   8.12. Galactic aberration: applied with amplitude of 5.8 micros/yr and
         epoch 2015.01.01.

9. Parameterization

   9.1. Clock: quadratic + continuous piece-wise linear with 60-mn interval.

   9.2. Atmosphere: continuous piece-wise linear 20-mn interval.

   9.3. Gradients: estimated at 6-hr interval at all sites.
     
   9.4. Earth orientation: polar motion and rate, UT1 and rate, nutation 
        offsets.
        
   9.5. Station positions: estimated.
        
   9.6. Station velocities: not estimated.
   
   9.7. Source coordinates: estimated.
        
   9.8. Antenna axis offsets: fixed.
 
10. Outlier elimination: use database settings.

11. Constraints (only of interest for direct EOP estimates)

   11.1. Sources: loose no-net rotation is uniformly applied to all sources.

   11.2. Stations: loose no-net rotation/translation is uniformly applied to
         all stations except those affected by earthquakes or nonlinear motions.
         
12. Generation of pre-reduced datum-free normal equations: yes.

13. Further details

   13.1. Data type: group delays.

   13.2. Data editing: 5 degree elevation cutoff.

   13.3. Data weighting: weights are defined as follows: 1/sqrt(f**2+a**2) 
         where "f" is the formal uncertainty of the ionosphere free linear
         combination of group delays at X- and S-band obtained by fringe fitting
         on the base of achieved signal to noise ratio. Station-dependent 
         parameter "a" was computed for each session by an iterative procedure 
         such that the ratio of the sum of squares of the weighted residuals to 
         the estimate of their mathematical expectation is about unity.
 
   13.4. Standard errors reported: all errors are derived from least-squares 
         estimation propagated from the data weights and the constraints applied 
         to the troposphere, clock and EOP parameters.
         
   13.5. Software: Calc 11.0, Solve revision date 2018.xx.xx.

   13.6. Other informations about the solution are reported at
         http://ivsopar.obspm.fr.