Hi Kyle,
The "earth_radius" attribute was added to CF in version 1.2 (May 2008).
The "spherical_earth_radius_meters" attribute pre-dates that CF version
and our GRIB reading code hasn't been updated to reflect that change.
We'll get that fixed and give our GRIB to CF code another look. We'll
try to get any fixes on this into our next netCDF-java release (which
will hopefully be in the next few days). If they don't make it into that
release, they should be in the next one.
Thanks,
Ethan
On 2/16/2010 1:31 PM, Kyle Shannon wrote:
> Is there any reason why the radius tag is: "spherical_earth_radius_meters"
> instead of the CF-1.0 tag:
> "earth_radius"
> ?
>
> kss
> # ===================
> Kyle Shannon
> Physical Science Technician
> RMRS Fire Sciences Lab
> Fire, Fuels & Smoke - RWU 4405
> 5775 Highway 10 W.
> Missoula, MT 59808
> (406) 829-6954
> kshannon@xxxxxxxxx <mailto:kshannon@xxxxxxxxx>
> # ===================
>
>
> On Mon, Feb 15, 2010 at 9:52 AM, John Caron <caron@xxxxxxxxxxxxxxxx
> <mailto:caron@xxxxxxxxxxxxxxxx>> wrote:
>
> Robb Kambic wrote:
>
> On Tue, 9 Feb 2010, Kyle Shannon wrote:
>
> Hello all,
> I am using the netCDF subsetting service on ucar's
> motherlode thredds
> server. I am having an issue with the projection of the
> subset.
>
> I run ncdump on the entire ndfd grid (full spatial extents)
> and get:
>
> kyle@Lucky13:~/nc_issue$ ncdump -h ndfd_all.nc
> <http://ndfd_all.nc> | grep Lambert_Conformal
> Maximum_temperature:grid_mapping = "Lambert_Conformal" ;
> char Lambert_Conformal ;
> Lambert_Conformal:grid_mapping_name =
> "lambert_conformal_conic" ;
> Lambert_Conformal:standard_parallel = 25. ;
> Lambert_Conformal:longitude_of_central_meridian = 265. ;
> Lambert_Conformal:latitude_of_projection_origin = 25. ;
> Lambert_Conformal:earth_shape = "Earth spherical with
> radius
> specified by producer in m" ;
> Lambert_Conformal:spherical_earth_radius_meters =
> 6371200. ;
> Lambert_Conformal:GRIB_param_Dx = 5079.361 ;
> Lambert_Conformal:GRIB_param_Dy = 5079.472 ;
> Lambert_Conformal:GRIB_param_GDSkey = 801839731 ;
> *Lambert_Conformal:GRIB_param_La1 = 20.192 ;
> Lambert_Conformal:GRIB_param_LaD = 25. ;
> Lambert_Conformal:GRIB_param_Latin1 = 25. ;
> Lambert_Conformal:GRIB_param_Latin2 = 25. ;
> *Lambert_Conformal:GRIB_param_Lo1 = 238.446 ;
> Lambert_Conformal:GRIB_param_LoV = 265. ;
> Lambert_Conformal:GRIB_param_NpProj = "true" ;
> Lambert_Conformal:GRIB_param_Nx = 1073 ;
> Lambert_Conformal:GRIB_param_Ny = 689 ;
> Lambert_Conformal:GRIB_param_ProjFlag = 0 ;
> Lambert_Conformal:GRIB_param_Quasi = "false" ;
> Lambert_Conformal:GRIB_param_ResCompFlag = 8 ;
> Lambert_Conformal:GRIB_param_SpLat = -90. ;
> Lambert_Conformal:GRIB_param_SpLon = 0. ;
> Lambert_Conformal:GRIB_param_VectorComponentFlag =
> "gridRelative" ;
> Lambert_Conformal:GRIB_param_Winds = "Relative" ;
> Lambert_Conformal:GRIB_param_grid_name =
> "Lambert_Conformal" ;
>
> Lambert_Conformal:GRIB_param_grid_radius_spherical_earth =
> 6371200.
> ;
> Lambert_Conformal:GRIB_param_grid_shape = "Earth
> spherical with
> radius specified by producer in m" ;
> Lambert_Conformal:GRIB_param_grid_shape_code = 1 ;
> Lambert_Conformal:GRIB_param_grid_type = 30 ;
> Lambert_Conformal:GRIB_param_grid_units = "m" ;
> Lambert_Conformal:GRIB_param_scanning_mode = 64 ;
> Lambert_Conformal:_CoordinateTransformType =
> "Projection" ;
> Lambert_Conformal:_CoordinateAxisTypes = "GeoX GeoY" ;
>
> and when I run it on the subsetted data, I get:
>
> kyle@Lucky13:~/nc_issue$ ncdump -h ndfd.nc <http://ndfd.nc>
> | grep Lambert_Conformal
> Maximum_temperature:grid_mapping = "Lambert_Conformal" ;
> char Lambert_Conformal ;
> Lambert_Conformal:grid_mapping_name =
> "lambert_conformal_conic" ;
> Lambert_Conformal:standard_parallel = 25. ;
> Lambert_Conformal:longitude_of_central_meridian = 265. ;
> Lambert_Conformal:latitude_of_projection_origin = 25. ;
> Lambert_Conformal:earth_shape = "Earth spherical with
> radius
> specified by producer in m" ;
> Lambert_Conformal:spherical_earth_radius_meters =
> 6371200. ;
> Lambert_Conformal:GRIB_param_Dx = 5079.361 ;
> Lambert_Conformal:GRIB_param_Dy = 5079.472 ;
> Lambert_Conformal:GRIB_param_GDSkey = 801839731 ;
> Lambert_Conformal:GRIB_param_La1 = 20.192 ;
> Lambert_Conformal:GRIB_param_LaD = 25. ;
> Lambert_Conformal:GRIB_param_Latin1 = 25. ;
> Lambert_Conformal:GRIB_param_Latin2 = 25. ;
> Lambert_Conformal:GRIB_param_Lo1 = 238.446 ;
> Lambert_Conformal:GRIB_param_LoV = 265. ;
> Lambert_Conformal:GRIB_param_NpProj = "true" ;
> Lambert_Conformal:GRIB_param_Nx = 1073 ;
> Lambert_Conformal:GRIB_param_Ny = 689 ;
> Lambert_Conformal:GRIB_param_ProjFlag = 0 ;
> Lambert_Conformal:GRIB_param_Quasi = "false" ;
> Lambert_Conformal:GRIB_param_ResCompFlag = 8 ;
> Lambert_Conformal:GRIB_param_SpLat = -90. ;
> Lambert_Conformal:GRIB_param_SpLon = 0. ;
> Lambert_Conformal:GRIB_param_VectorComponentFlag =
> "gridRelative" ;
> Lambert_Conformal:GRIB_param_Winds = "Relative" ;
> Lambert_Conformal:GRIB_param_grid_name =
> "Lambert_Conformal" ;
>
> Lambert_Conformal:GRIB_param_grid_radius_spherical_earth =
> 6371200.
> ;
> Lambert_Conformal:GRIB_param_grid_shape = "Earth
> spherical with
> radius specified by producer in m" ;
> Lambert_Conformal:GRIB_param_grid_shape_code = 1 ;
> Lambert_Conformal:GRIB_param_grid_type = 30 ;
> Lambert_Conformal:GRIB_param_grid_units = "m" ;
> Lambert_Conformal:GRIB_param_scanning_mode = 64 ;
> Lambert_Conformal:_CoordinateTransformType =
> "Projection" ;
> Lambert_Conformal:_CoordinateAxisTypes = "GeoX GeoY" ;
> Total_cloud_cover:grid_mapping = "Lambert_Conformal" ;
> Wind_direction_from_which_blowing:grid_mapping =
> "Lambert_Conformal"
> ;
> Wind_speed:grid_mapping = "Lambert_Conformal" ;
>
> My question is, how can the latitude and longitude of the
> 1st cell not
> change after the subsetting? It appears that the meta data
> from the full
> dataset is just copied to the subset data. Is this correct?
>
>
> Kyle,
>
>
> As Robb says, all the Lambert_Conformal:GRIB_XXXX attributes are
> simply copied over to the netcdf file from the GRIB GDS. We dont use
> these GRIB parameters for anything. Robb failed to mention that the
> actual coordinates are stored in the x and y coordinate variables in
> the netcdf file. In this case these are projection coordinates,
> which are converted to lat/lon values through the lambert conformal
> projection. If you check the "include lat/lon" box, the 2D lat/lon
> coordinates will also be included in the file, as Robb indicates.
> This is needed to be in strict compliance with CF conventions, but
> is not needed by the CDM software.
>
>
>
> Yes. The GDS parameters that you are looking at are the
> parameters for the model, so they are copied over to the
> resulting file. Usually users make the subsetting file name
> contain the latitude and longitude information of the subset.
>
> How can I
>
> figure out where my subset data is on the world?
>
>
> Click the check box on the sub-set page for "Add Lat/Lon to file"
> to the resulting file. Now the ncdump on the Lat and Lon
> variables will be available.
>
> When you open/display the file the information will also be
> available.
>
> On a programming level:
>
> If you open the file as a GridDataset, you should be able to
> access the needed information. Here's an url about GridDatasets:
>
>
> http://www.unidata.ucar.edu/software/netcdf-java/tutorial/GridDatatype.html
>
> You will also need to to look at the java-docs on the GridDataset.
>
>
> RObb...
>
>
> Thanks, sorry for the messy message.
