I have been trying to make a netCDF dataset compliant with the CF conventions. The dataset is currently in the EPSG 3411 NSIDC Sea Ice Polar Stereographic projection (i.e., x and y coordinate variables, no Lat/Lon auxiliary coordinate variables). I am trying to make sure the grid mapping variable is set up properly, but I'm not sure which grid mapping to use.
I assumed I should use the Polar Stereographic projection at first, and was trying to make it work, but I am confused by the requirement that the
latitude_of_projection_origin attribute should be either +90 or -90, because the corresponding value (I think) for EPSG 3411 in the OGC WKT string is 70. Also, for the third "Map Parameter" listed, it appears that only two EPSG codes (9829 and 9810) might be acceptable, although that's not made explicit.
I looked a bit further and found there is also a plain "Stereographic" grid mapping available, and I wonder if I should use that one instead.
pyproj.crs.CRS (3.0.1) and the
to_wkt() method, I've found that the EPSG 3411 projection uses the "Polar Stereographic (variant B)" (EPSG 9829) for the "CONVERSION" parameter.
I've also realized that
pyproj.crs.CRS has a
to_cf() method that creates a dict automatically from the epsg code, and when used with
CRS.from_epsg(3411), it follows the convention for the "Polar Stereographic" grid mapping.
So that answers my original question, but there is still room to clear things up:
- Why doesn't the OGC WKT entry for EPSG 3411 at spatialreference.org include anything about this coordinate system relying on EPSG 9829?
- Why isn't the "Latitude of origin" included in that WKT string either?