CDO abort due to different number of layers

[aearamos]

Hi,

I'm running a CMIP6 simulation with the last version of ece3-postproc but I'm getting an error related to the number of layers in my output. Here it is:

`cdo setzaxis: Processed 1729 values from 1 variable over 1 timestep ( 0.04s )

• [[ 0 == 1 ]]
• cdo -f nc div -sqr -sub /gpfs/scratch/bsc32/bsc32436/a295/19900101/fc00/runtime/../tmp/ece3post/ecmean_a295_nZeNWw/new_T.nc /gpfs/scratch/bsc32/bsc32436/a295/auto-ecearth3/ece3-postproc/ECmean/Climate_netcdf/climate_ERA40_T_zonal.nc /gpfs/scratch/bsc32/bsc32436/a295/auto-ecearth3/ece3-postproc/ECmean/Climate_netcdf/variance_ERA40_T_zonal.nc /gpfs/scratch/bsc32/bsc32436/a295/19900101/fc00/runtime/../tmp/ece3post/ecmean_a295_nZeNWw/temp_T.nc
  cdo div: Started child process "sqr -sub /gpfs/scratch/bsc32/bsc32436/a295/19900101/fc00/runtime/../tmp/ece3post/ecmean_a295_nZeNWw/new_T.nc /gpfs/scratch/bsc32/bsc32436/a295/auto-ecearth3/ece3-postproc/ECmean/Climate_netcdf/climate_ERA40_T_zonal.nc (pipe1.1)".
  cdo(2) sqr: Started child process "sub /gpfs/scratch/bsc32/bsc32436/a295/19900101/fc00/runtime/../tmp/ece3post/ecmean_a295_nZeNWw/new_T.nc /gpfs/scratch/bsc32/bsc32436/a295/auto-ecearth3/ece3-postproc/ECmean/Climate_netcdf/climate_ERA40_T_zonal.nc (pipe2.1)".
  cdo(3) sub (Abort): Input streams have different number of layers per timestep!
  `
  My file has 19 layers (plev19) and 3hr and 6hr output. I noticed the files inside Climate_netcdf have 11 vertical levels (lev 11). Is this what's causing the error? If yes, how to fix it? From our side would be nice to have 19 layers.

@etiennesky @plesager
Thanks

[plesager]

Which tool is the problem? Is it hiresclim2, EC-mean, ...? For hiresclim I am using the -6 option to filter the IFS output. Check the filter I used on the Rhino platform:
conf/rhino/conf_hiresclim_rhino.sh. We do not process the 3D output in EC-mean.

Some of our output includes model levels output. For those, I used the filter in the cordex branch (rhino platform again).

[aearamos]

Hi @plesager

The problem was with EC-mean. I think it's similar to the problem Paolo had when filtering the PRIMAVERA output, since it had more layers than climatology. I turned off the 3D output and now it's working, We should be fine now.