CDO

Hi-

I'm running CDO 1.6.2.

I have a gaussian grid on an OPeNDAP server:

http://hydra.fsl.noaa.gov/thredds/dodsC/ana4MIPs/20CRv2/NCEP-GFS2008ex/ensemble_mean/tas/tas_Amon_reanalysis_20CRv2_187101-201212.nc

which has lat/lon defined as coordinate variables with the correct standard names, e.g.:

double lat(lat=94);
      :units = "degrees_north";
      :long_name = "Latitude";
      :standard_name = "latitude";
      :axis = "Y";
      :coordinate_defines = "point";
      :actual_range = -88.542f, 88.542f; // float
      :_CoordinateAxisType = "Lat"

When I run the following on this:

cdo -L -s -O -timmean -selmon,1 -selyear,1980,1981 http://hydra.fsl.noaa.gov/thredds/dodsC/ana4MIPs/20CRv2/NCEP-GFS2008ex/ensemble_mean/tas/tas_Amon_reanalysis_20CRv2_187101-201212.nc output.nc

the standard names of latitude and longitude get changed to "grid_latitude" and "grid_longitude", e.g.:

double lat(lat) ;
        lat:standard_name = "grid_latitude" ;
        lat:long_name = "Latitude" ;
        lat:units = "degrees_north" ;
        lat:axis = "Y" ;

Which are used in the CF conventions for a rotate pole grid mapping:

"The rotated latitude and longitude coordinates are identified by the standard_name attribute values grid_latitude and grid_longitude respectively."

(http://cfconventions.org/1.6.html#appendix-grid-mappings)

and:

"A CF compliant application can determine that rlon and rlat are longitude and latitude values in the rotated grid by recognizing the standard names grid_longitude and grid_latitude."

(http://cfconventions.org/1.6.html#idp5605040)

It seems that the standard_names for these coordinates in this grid should be latitude and longitude since this is not a rotated pole grid.

Currently, this causes problems when reading this grid with the netcdf-java library which is interpreting this as a rotated pole grid based on the CF convention standard names. I've been in contact with Unidata on this issue to see if they can adjust their logic to disambiguate the conflicting info, but I also wanted to let you know that the incorrect CF standard names are being used in this case.

Thanks for your great software.

Don Murray