CDO

Resuelto.
grid descriptor in root dir.

Hello:

I have a similar question about transform from unstructured to regular grid under Linux.
The error information always showed like that:

"cdo remapbic (Abort): Bilinear/bicubic interpolation doesn't support unstructured source grids!"

Does it mean ICON data couldn't be interpolated Lat/Lon grid like GFS or just something need to be modified?

Thanks,
Leon

Leon Lin wrote:

Hello:

I have a similar question about transform from unstructured to regular grid under Linux.
The error information always showed like that:

"cdo remapbic (Abort): Bilinear/bicubic interpolation doesn't support unstructured source grids!"

Does it mean ICON data couldn't be interpolated Lat/Lon grid like GFS or just something need to be modified?

Thanks,
Leon

BTW, here is the command I type:

cdo -f grb2 remapbic,zzzgrid ICON_iko_single_level_elements_world_CLCT_MOD_2017111912_180.grib2 zzz

GRIB2 files come without coordinates of the horizontal grid. You have to add these information in order to use the interpolation functions in CDO. You can add the grid coordinates with the CDO operator setgrid. Here is an example:

cdo  remapnn,eragrid  -setgrid,<gridfile>  icon.nc  outfile.nc

The CDO operators remapbil and remapbic can't be used for ICON grids, use remapdis, remapnn or remapycon instead.

Hello all:

After following your comments, it's converted to lat/lon grid already.
Thanks a lot you all.

Dear all,
I have a similar situation,

I have ICON global data, I managed to interp all to lon/lat except the invariant package.

I have the source_grid, the target_grid and the weights. It gives me a constant error :
"
cdo remap (Warning): Reference to horizontal grid not available!

cdo remap (Abort): Reference to source grid not found!
"
The command I run is
cdo -f grb2 remap,target_grid_world_025.txt,weights_icogl2world_025.nc -setgrid,icon_grid_0026_R03B07_G.nc ICON_iko_invar_package_world_2018011800.grib2 temp.grib2

Any help will be much appreciated
Regards
J

There is a guide available called "Guideline for the transformation of GRIB2 datasets from the
original triangular to the regular lat/lon grid": https://www.dwd.de/DE/leistungen/opendata/help/modelle/Opendata_cdo_EN.pdf?__blob=publicationFile&v=3

I follow the guidelines from DWD and everything goes fine. But when I view the converted file it looks very strange and not like a nap.

Help please!

please let us know:

input file (link)
output grid
cdo version
cdo call or a script

Thank you for the fast reply!
I think I figured it out, the problem was that the target grid file form DWD wasn't well extracted which resulted in missing information.

I was wondering though. In the documentation of ICON the created weights file is used to convert it to Latlon. However in the presented explanation on this forum the weight file is not used for conversion. Is there a reason for this?

It works by your comments but wanted to know this.

which forum message do you refer to?

The fourth reply on this forum page.
---------------------
GRIB2 files come without coordinates of the horizontal grid. You have to add these information in order to use the interpolation functions in CDO. You can add the grid coordinates with the CDO operator setgrid. Here is an example:

cdo remapnn,eragrid -setgrid,<gridfile> icon.nc outfile.nc
<gridfile> is a file which needs to contain all the grid coordinates. For ICON data, you will get this file from your data provider. Lots of DWD grids are available on the ICON grid file server (http://icon-downloads.zmaw.de).

ah, ok

with CDO you can to interpolation in two way

1. like in the DWD document:
   1. create the weights file (e.g. gencon operator for conservative remapping)
   2. apply the weights file with remap
2. avoid the extra step of generating a weights file and do both in ONE step (e.g. remapcon for convervative remapping)

the first option is preferable if you want to interpolate multiple files, hence use the weights file multiple times. The second version is good for one-time-interpolations. This is the reason, why there are genXYZ and corresponding remapXYZ opeators.

hth
ralf

Dear Colleagues,

I'm trying to produce a synoptic chart using the data PMSL from ICON Model, which are available on website from DWD.
But now I've been facing problems with a high resolution the Model above the Andes Mountain. And in order to have this product ready I have to have a smooth line of pressure. I've already tryed a cdo command (cdo smooth9 infile outfile ; and cdo smooth infile outfile), but I didn't have any good affect.

I would like to know if you have any idea how Can I procedure? What command should I used to have a smooth result?

Thanks in advance!
Best Regards,
Alana.

hi Alana!

I think, smooth is the good choice. you might check the different parameters of smooth to tune the area around grid point, that is taken into account for smoothing. check the docu with

cdo -h smooth

If you post the link to the file, I could give it a try myself.

cheers
ralf

I wrote a little docu about smoothing here

Ralf Mueller wrote:

I wrote a little docu about smoothing here

Hello Ralf Mueller, First of all, Thanks for reply my question. Here I've attached the data example of PMSL already interpolated to regular grid (lat/lon). I really apreciate
if you teach me how can I procede to smooth the pressure lines in order to create my synoptic chart.

Thanks for help us.
Best Regards,
Alana.