[Difx-users] ITRF realisations {External}

[Leonid Petrov]

Dear Chris,

> but the sched catalogue as what most people use as the positions used 
> for Calc.

   Pity!

   Coordinates are not measurable quantities. The is is the
essence of geodesy. Path delay is invariant to a group
of rotations and translations. As a result, one cannot determine
station positions using only observations. One needs to apply
constraint equations with some arbitrary constants right hand sides.
You can select these constants whichever way you like, for set to
a telephone number of your spouse, but usually, these constants
are chosen in such a way that the origin of the coordinate system
is not too far from the place where nobody traveled -- I mean the
geocenter, and orientation is not too different then the EOP time
series maintained by Christian Bizouard. "Too far", "too different"
sounds subjective. In my view, 10 cm is not too far.

   Reference frame is just a synonym to station position catalogue.
If one determines position of ATCA pad XXX with respect, say ATCA-104,
using total station, then that geodesist uses solution gsf_2025b,
then that guy can write: position of ATCA pad XXX in the gsfc_2025b
reference frame is A.

   ITRF is not consistent with anything. The latter versions are worse
than the previous one. I do not recommend of using it all.

   I know users AIPS got used to a concept that the correlator would
provide them a perfect model and then they can work with residuals.
At NASA we have never used that approach at least since 1975.
Instead, we compute our model that is orders of magnitude more accurate
than that the correlator used and add to theoreticals the quantity
new_path_delay minus old_path_delay. And we do not care about
TRF, EOP used by the correlator. I offered 5 or 6 times to the AIPS
maintainer to add this feature to AIPS and always heard "no!".
Personally, I have PIMA for VLBI data analysis instead of AIPS and
I am not affected by this problem.

> So what frame used likely does not matter, but for normal astronomical
> use cases my worry is that we as a whole are not careful enough.

   If you really want to have a precise theoretical model, use 
open-source
NASA SGDASS that has station positions, source positions, EOP, 
non-linear
station motion updated on a quarterly basis, various loading  applied,
the ionospheric model, path delay through the atmosphere computed
by integration of equations of wave propagation using the output of NASA
numerical model, etc. The right place for this refined model is
the post-processing stage, not correlation. Though, you can run difxvtd
if you like.

Leonid
2025.07.04_22:46:50


> On 2025-07-04 22:26, Phillips, Chris (S&A, Marsfield) wrote:
> Hi Leonid,
> 
> My concern with sched was that the catalogue have no frame info
> recorded into the positions. Sched does not do much with positions, so
> I agree for sched it does not matter but the sched catalogue as what
> most people use as the positions used for Calc.
> 
> I have chatted to Cormac offline, who points out that the absolute
> difference between frames would only affect stuff like ionospheric
> corrections, which will make difference at all. However the different
> frames have many cm difference between them so if inconsistent frames
> are used in the catalogues, that *will* have an effect.
> 
> So what frame used likely does not matter, but for normal astronomical
> use cases my worry is that we as a whole are not careful enough.
> 
> Cheers
> 
> Chris
> 
> From: Leonid Petrov <Leonid.Petrov at lpetrov.net>
> Date: Friday, 4 July 2025 at 19:49
> To: Phillips, Chris (S&A, Marsfield) <Chris.Phillips at csiro.au>
> Cc: Difx-users <difx-users at listmgr.nrao.edu>
> Subject: Re: [Difx-users] ITRF realisations {External}
> 
> Chris,
> 
>    I looked at sched-12.0. I found that internally it uses UTC(t)
> function as time. It assumes UT1(t) = UTC(t), see for instance
> schgeo.f. This can introduce an error that is equivalent to an
> error in position of mid-latitude sites at a level of several
> hundreds meters.
> 
>    In that context difference in several centimeters between different
> station catalogues is irrelevant.
> 
> Leonid
> 2025.07.04_08:47:11
> 
>> On 2025-07-04 01:02, Phillips, Chris (S&A, Marsfield) via Difx-users
> 
>> wrote:
>> Hi all,
>> 
>> I have just learned that ITRF (which we use as the XYZ coordinate
>> system of our telescopes) has multiple realisations (ITRF2000,
>> ITRF2014 etc).
>> 
>> There was approx a 9cm change between around ITRF2005, which is
> huge.
>> 
>> The “Frame” detail in Sched is light on details and vex2difx
> make
>> no mention of frame.
>> 
>> Does anyone know what ITRF frame Sched expects?
>> 
>> Thanks
>> 
>> Chris
>> _______________________________________________
>> Difx-users mailing list
>> Difx-users at listmgr.nrao.edu
>> https://listmgr.nrao.edu/mailman/listinfo/difx-users