[PROJ] Fwd: some help with cm-level GRS80 to WGS-84 transformation?

[Greg Troxel]

Joseph Boardman <boardman at aigllc.com> writes:

> OK, I'll take my licks and ask again.
>
> How would I use Proj to do a precision transform from a particular
> NAD83 (GRS80) realization and a particular WGS-84 realization?

There are three practical problems to address:

  - The EPSG database has historically not dealt well with WGS84
    realizations.  (Try to find a datum code for WGS84(G2139).)  More
    importantly, there is an ensemble concept which results in null
    transforms when, if the people asking for the transform instead
    understood what they were trying to do and asked for a different
    transform, it would be more accurate.  That's perhaps fair in a
    theoretical sense, but it's not helpful to led the history
    WGS84(TRANSIT) affect transforms among modern realizations -- nobody
    has any remotely accurate data in WGS84(TRANSIT).

  - WGS84-GXXXX and ITRF-XXXX are dynamic and you need to understand
    what you are doing about epoch.   NAD83(2011) starts to look dynamic
    at 10 years and 1 cm, and it's more so on the Pacific plate.

  - There are close relationships between each (reasonably recent) WGS84
    and some ITRF, and you can use an ITRF as a proxy.  But then you
    aren't really doing what you are saying you are doing, so you need
    to have a story about why.

If you are really talking cm-level, then the only thing that I think
makes sense is to transform from

  EPSG:6319	NAD83(2011) epoch 2010.0

to something like

  EPSG:7912	ITRF2014 (which I am 99% sure uses default epoch 2010.0)

if you mean WGS84(G2139).

I'm pretty sure ITRF2008 is the right answer if you mean WGS84(G1762),
and it seems there is EPSG:7665 for that.

Then, you'll need to think about what epoch you want the ITRF2014
coordinates in and why.


----------------------------------------

If you mean some NAD83 other than NAD83(2011) epoch 2010.0 then you
should be using HTDP/NCAT from NGS to get to EPSG:6319 first.  And, the
idea that previous NAD83 realizations have internal accuracy at the 1 cm
level is one that I'd be highly skeptical of.

Definitely look at the results very carefully, and cross check with OPUS
results in both NAD383 and ITRFxxx for the same solution, as very few
people are really trying to deal with these transforms at cm level and
you'll be a bit of a pioneer.  (If you aren't doing 24h occupations with
dual-frequency receivers and static processing, you probably don't have
1 cm coordinates to start with.)

----------------------------------------

However, there is a more serious problem, which is that the very concept
of cm-level coordinates in WGS84 is hard to talk about until you equate
it to some ITRF and accept that you are really dealing e.g. with
ITRF2014 instead of WGS84G(xxxx).

So let me ask: are you going to be able to establish precise
WGS84(GXXXXX) coordinates for some point, and how are you doing to do
it?  How could you determine if coordinates you got via transform are
correct or not?

Basically, I am challenging you to explain how you will access the
WGS84(G2139) datum at better than meter level.  (Other than accessing
ITRF2014 or IGS08 and saying they are equivalent.)
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 194 bytes
Desc: not available
URL: <http://lists.osgeo.org/pipermail/proj/attachments/20220119/311777d0/attachment-0001.sig>