[PROJ] Ellipsoidal height VCRS in WKT/EPSG

[Even Rouault]

Javier,

> Does exist any EPSG or WKT for a Vertical CRS that means "ellipsoidal
> height"?

No, and this is by design. Ellipsoidal height doesn't make sense, from a referencing point of 
view, if the longitude,latitude of the point where you measure it is not specified as well, 
hence the need for a Geographic3D CRS. This is contrary to a gravity-based height, where if 
someone gives you just the the elevation at 2 points, you can for example know from which 
point towards which one water would flow.

See http://docs.opengeospatial.org/is/18-010r7/18-010r7.html
"""
3.1.19 ellipsoidal height

    distance of a point from the reference ellipsoid along the perpendicular from the 
reference ellipsoid to this point, positive if upwards or outside of the reference ellipsoid

    Note 1 to entry: Only used as part of a three-dimensional ellipsoidal coordinate system or 
as part of a three-dimensional Cartesian coordinate system in a three-dimensional projected 
coordinate reference system, but never on its own.
"""

Very early versions of the EPSG dataset (like 10 years ago or more) add codes for vertical CRS 
with ellipsoidal height, but this was removed in later versions. Some remains of that are still 
found in the GeoTIFF 1.0 specification that hard-coded such dataset, but they have 
disappeared since, and are no longer allowed in GeoTIFF 1.1. I've seen recently also some 
WKT-1 string for LAS files that were using a ellipsoidal vertical CRS as part of a compound 
CRS, but this is no longer considered as geodetic-savy practice.

> I don't want to "promoteTo3D" the 2D CRS  adding the third axis, because it
> removes the authority.

For projected CRS, yes, since there is currently no Projected 3D CRS in EPSG
For geographic CRS, to the extent where there's a 3D version of the one you promote to 2D 
registered, promoteTo3D should attach the code of the 3D CRS:


$ projinfo EPSG:4326 --3d
[....snip ...]
    ID["EPSG",4979]]

Even

-- 
Spatialys - Geospatial professional services
http://www.spatialys.com
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.osgeo.org/pipermail/proj/attachments/20200901/e74fbfcb/attachment-0001.html>