[Spice_discussion] precision of sun-earth radial velocity from spice?

[Hans-Peter Doerr]

Dear Navigators,

is there a way how I could estimate the accuracy of velocities and 
distances derived from SPICE?

I need very precise radial velocities (sub meter/second if possible) 
between an earth based observatory and an arbitrary position on the sun 
(neglecting solar rotation for the moment). I stumbled on SPICE and 
thanks to the excellent documentation I soon was able to hack up a C 
code that (most probably) does what I want.

The procedure I use is

  * compute state 'O' of the observer wrt solar barycenter in J2000 
frame (using spekzr)
  * compute state of the target position on the solar surface 'T' wrt 
solar barycenter
  * project velocity component of the observer-target state S = O + T 
onto the Observer-Target line-of-sight unit vector to get the 
line-of-sight velocity

I use the ITRF earth rotation kernel, in our case the difference is less 
than 1 m/s in velocity.

I now wonder which absolute accuracy I can expect for such a calculation.


Regards

  Hans-Peter