Question of thickness in the boundary files

[favorliao]

The current PyCNAL uses the interface height of each grid cell to compute the layer thickness. However, the examples only use the cell center depth instead of cell interface height as input for the PyCNAL to computer layer thickness. This may cause issues to layer thickness in boundary files.

In the PyCNAL, depth is computed using interface height
lib_ioncdf.py, line 348: dz[:-1, :, :] = vc[1:, :, :] - vc[:-1, :, :]

In the PyCNAL examples, only cell center depth is provided:
Creating_Initial_and_Boundary_conditions_from_SODA3.py, line 59:
salt_domain.interpolate_from(sodaic,'salt',frame=0,depthname='st_ocean',coord_names=['xt_ocean','yt_ocean'],

[raphaeldussin]

good catch! Do we have s_w ocean in the files?

A more comprehensive fix would be (and I think we should move towards this) to operate on xarray datasets and not on files, and that would make it much easier to use. Wish I knew xarray back then :(

[favorliao]

There is st_ocean in the files (SODA files). I wrote a raw one for my own case using XESMF and xarray. If you want to take a look, https://github.com/favorliao/Reginal-MOM6/blob/master/obc_initial_mking/make_obc_mom_z.py