[PROJ] [EXTERNAL] Re: package 'geosphere' -- transforming spherical-based geographic coordinates to WGS84

[Even Rouault]

On mercredi 22 avril 2020 21:15:56 CEST Fischbach, Anthony S wrote:
> Even,
> 
> Your indication that EPSG is favored over use of proj raises a burning
> question. Our study area is in the Chukchi Sea, which spans the
> anti-meridian and is far north.  To minimize spatial distortion, we have
> used an equal area projection centered on our study area (+proj=laea
> +lat_0=70 +lon_0=-170 +x_0=0 +y_0=0 +datum=WGS84 +units=m +no_defs
> +ellps=WGS84 +towgs84=0,0,0). Can you advise on the merits of using a
> custom projection specified in +proj terms, given that no EPSG projection
> is published for this region.

The issue with PROJ.4 string is that their expressivity in terms of datum specification was 
limited to:
- using +datum with a few hardcoded datum names which didn't catch up with later additions 
in the EPSG catalog,
- or an ellipsoid + towgs84
- or an elliposid + nadgrids.

Hence the recommendation to use an EPSG code when there is one, since PROJ will be able 
to apply more relevant datum transformations when there's one registered in the database, 
and not necessarily by using the WGS84 pivot as it did before PROJ 6.

If your datum is WGS84 (or for your level of accuracy can be considered equivalent to it), you 
can use a PROJ string just fine. In your above string, as you use +datum=WGS84, then the 
+ellps and +towgs84 are implied and can be safely removed.

Even

-- 
Spatialys - Geospatial professional services
http://www.spatialys.com
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.osgeo.org/pipermail/proj/attachments/20200422/d0cfade1/attachment.html>