[PROJ] Swiss geoid file (now free)

[Even Rouault]

Hi,

> -- License --

Agreed with Bas that they should really consider using a known license. From 
the rights and requirements they indicate, it would seem to me that CC-BY 4.0 
would be a good fit.

> -- Source CRS --
> The geoid file is provided in
> https://www.swisstopo.admin.ch/en/knowledge-facts/surveying-geodesy/geoid.ht
> ml in 4 different "flavors": LV95/LHN95, LV95/LN02, ETRS89/LHN95,
> ETRS89/LN02 (I will now concentrate on LHN95, as LN02 is older).
> 
> LV95 (short version of "CH1903+ / LV95") is a projected CRS (EPSG:2056).
> The geographic system "under" it is CH1903+ (EPSG:4150)
> I was expecting a version like "CH1903+/LHN95" (the Swiss geographic CRS
> with the Swiss vertical CRS), but there is not AFAIK.
> 
> What should we include in PROJ-data?
> a) ETRS89/LHN95
> b) LV95/LHN95
> c) CH1903+/LHN95 computed from LV95/LHN95
> d) something else
> e) multiple of the above

No special opinion regarding this. Yes currently we'd have to derive a grid in 
geographic coordinates. We should ultimately provide a way of dealing with 
grids in projected coordinates (UK uses that too). Not necessarily a huge deal 
of work, but has to be done...

> -- Transformation --
> There is no such a transformation in EPSG like "ETRS89 to LHN95 height" or
> "LV95 to LHN95 height" or "CH1903+ to LHN95 height". I was considering
> asking Swisstopo to "create" them (via a proper EPSG request), but, which
> one(s)? (this is probably related to the previous question, I guess).

That's a bit their call, but yes it would be good to ask them to register the 
transformations with EPSG to make them more easily usable. I'm not sure if 
EPSG includes geoid-like grid transformations where the interpolation CRS is a 
projected one currently, so they might be only able to register the ETRS89 to 
LHN95 height currently.

Even

-- 
http://www.spatialys.com