[PROJ] Vertical deflection: of interest for PROJ ?

[Even Rouault]

> Will E. Featherstone & Jean M. Rüeger (2000) The importance of using
> deviations of the vertical for the reduction of survey data to a geocentric
> datum, Australian Surveyor, 45:2, 46-61, DOI:
> 10.1080/00050326.2000.10440341 (available from
> https://www.tandfonline.com/doi/abs/10.1080/00050326.2000.10440341 or by
> request to the author, perhaps), describes these issues (and more) in
> detail.

Unofficial link to the paper:
http://fgg-web.fgg.uni-lj.si/~/mkuhar/Pouk/SG/Seminar/Azimut/Featherstone_dev_vert.pdf

> The paper discusses the importance for the SURVEY community

I don't perceive PROJ as being used to establish geodetic surveys, but more at
transforming coordinates once surveys have been established.

> in applications
> including: 1. transformation between astronomical coordinates and geodetic
> coordinates; 2. conversion between astronomic or gyro azimuths and geodetic
> azimuths; 3. reduction of measured horizontal directions (and angles) to
> the ellipsoid; 4. reduction of measured zenith angles to the ellipsoid;
>      5. reduction of slope electronic distance measurements (EDM) to the
> ellipsoid using zenith angles; and [See table 10]
>      6. determination of height differences from zenith angles and slope
> distances. [See table 16]
> 
> Some of the above (e.g. #1, #5, #6) are directly related to your coordinate
> transformation question. 

I'm not so sure. At least not in obvious ways regarding to what PROJ API and tools
currently offer (but users are often very creative in extending the intended use)

1. Astronomical <--> geodetic coordinates: not covered by PROJ AFAIK
5. reduction of slope electronic distance measurements to the ellipsoid: idem.
   I'd expect this to be something done internally in total stations
6. determination of height differences from zenith angles and slope distances: idem

The use case I mentionned would require a new PROJ operation method, something like

+proj=add_ortho_height +grids=dov_geoid_XXX.tif +dh=YYY

And probably with angles w.r.t the plumb line to make it more general than just along
the plumb line.

Or perhaps a more realistic example as Nyall mentionned would be to transform coordinates
between a CAD/BIM local 3D cartesian coordinate system, whose horizontal plane would
be perpendicular to the plumb line (assuming they work in such systems!) to
geodetic/geocentric coordinates.

My main point was to make sure that we're not missing anything in not using vertical
deflection for an operation like vgridshift which affects only the Z value.

Even

-- 
Spatialys - Geospatial professional services
http://www.spatialys.com