reproducing_EFP_sum(_2d)

[Parag0206]

complete error:

FATAL: NaN in input field of reproducing_EFP_sum(_2d).

This definitely means some NaN is there in some input file

[nmathewa]

• which run is this?
• if this is with the forcing dataset you must fill nan values. One python based script in the scripts folder
• Or @cpneema knows how to do it in ferret

[Parag0206]

• this is a test run
• yes, there must be some NaN values in forcings, I will check

[Parag0206]

• Using NCO command

  ncks --chk_nan infile.nc

  to check for NaN values in input files.

• Also, this error is specifically because of NaN in some 2D field. The 2D fields files are:

  • horizontal_grid.nc
  • topog.nc
  • atmos_mosaic_tile1Xland_mosaic_tile1.nc
  • atmos_mosaic_tile1Xocean_mosaic_tile1.nc
  • land_mosaic_tile1Xocean_mosaic_tile1
  • land_mask.nc
  • ocean_mask.nc
  • ocean_mosaic.nc

[Parag0206]

Found no NaN in any of the 2D field or 3D field input files

[Parag0206]

Also, this error occurs only after the initial time step, maybe there is some problem in the initial_condition file.

[nmathewa]

• It is not possible NAN in the domain files or grid files
• Check initial conditions and forcing

[Parag0206]

• The error is in the initial conditions file.
• As mentioned above, NaN could not be present in domain/grid files generated using FRE-NCTOOL
• The IC file is from NEMO(CMEMS) website which is a 75 layer file. The bottom most layer has a depth of 5902 m and the values are "nan", which is not picked up by "ncks --chk_nan" command. This was found using "cdo infon" command.
• Removed the last layer to get an IC file containing 74 layers
• Python3 code :

  dmin = z[0]
  dmax = z[-2]
  ic = ic.sel(depth=slice(dmin, dmax))

[Parag0206]

Use raphael dussin's HCtFlood python library (https://github.com/raphaeldussin/HCtFlood) for making initial conditions file