[PROJ] ETRFxx to national ETRS89 realizations (e.g. ETRS89-NLD [AGRS2010], RGF93 v2/v2b, ETRS89-HUN [ETRF2000], etc.) (was Re: Time dependency in ETRFxx to ETRFyy coordinate transformations?)

[Greg Troxel]

Even Rouault <even.rouault at spatialys.com> writes:

> https://docs.ogc.org/as/18-005r8/18-005r8.pdf:
>
> """datum anchor epoch:  epoch at which a static reference frame
> matches a dynamic reference frame from which it has been
> derived. Note: Not to be confused with the frame reference epoch of
> dynamic geodetic and dynamic vertical reference frames. Nor with the
> epoch at which a reference frame is defined to be aligned with another
> reference frame; this information should be included in the datum
> anchor definition."

Thanks.   So:

  CRS A: "NAD83(2011)" is a dynamic datum and has a frame reference epoch of 2010.0.

  CRS B: "NAD83(2011) epoch 2010.0" has a datum anchor epoch of 2010.0,
  but this value is logically distinct from the reference epoch 2010.0
  Had it been 2015.0 instead, then coordinates in A with a missing time
  would be different from coordinates in B.   But the values are the
  same, so A-missing-t and B use the same coordinates.


It would seem madness to pick a different datum anchor epoch than the
underlying frames frame reference epoch, but I see that AGRS2010 is an
annual solution for a reference network and thus I guess only expected
to be used by a select few with high clue.  Or maybe not, if it's the
frame of the transmitted reference data.

It seems that there is an implied null transform:

  [B] to [A at B's datum anchor epoch]

which if the epochs match is

  [B] to [A at A's frame reference epoch]

which is also

  [B] to [A with epoch not specified]

and these are definitional, not aproximate.  But also that this is
probably not what's going on in this case.



These static-derived-from-dynamic frames seem highly useful for
operators of RTK networks.  Consider running a network: what values do
you plug in for reference station coordinates and how do you update
them?  With a dynamic datum, you have to either put in velocities and
hope the software computes new coordinates at some reasonable frequency
and that receivers pick up the new coordinate, or you need to just
change the coordinates say monthly.  (It's interesting to note that GPS
changes the coordinates ~yearly as I understand it.)   With a
static-derived-from-dynamic frame, you just put the coordinates in once,
being careful not make mistakes, and then you don't touch them, avoiding
future mistakes.