add_t_cell_cutoff_function

[ezhilsabareesh8]

This PR adds a subroutine to remove smaller T cells, particularly in regions near the tripole where smaller grid cells contribute to instability. These cells were previously removed in OM2 to enhance stability, and this change has now been incorporated into domain-tools in the current PR.

[micaeljtoliveira]

@ezhilsabareesh8 Just a few comments, mostly about style. Otherwise code looks fine.

[micaeljtoliveira]

One last comment: there's actually no need to store the full dy_t array if you combine the two loops over all elements of dy_t. Something like this:

    do i = 1, ny_len / 2
      do j = 1, (nxp_len - 1) / 2
          if (dy(2 * i - 1, 2 * j) + dy(2 * i, 2 * j) < cutoff*1000.0) then  !Input cutoff in Kilometers covert it to meters
            this%depth(i, j) = MISSING_VALUE  ! Set values below cutoff to zero or another value as needed
          end if
        end do
    end do

[ezhilsabareesh8]

One last comment: there's actually no need to store the full dy_t array if you combine the two loops over all elements of dy_t. Something like this:

Thanks @micaeljtoliveira I have removed the intermediate array in this commit Updated in this commit 63bc14f

[micaeljtoliveira]

Note that you also need to add the new case here:

  select case (name)
  case ('gen_topo', 'deseas', 'min_max_depth', 'fill_fraction', 'fix_nonadvective', 'mask')
    file_out = sget('output')
    if (file_out == 'unset') then
      write(error_unit,'(a)') 'ERROR: no output file specified'
      error stop
    end if
end select

so that the code stops code if no output file was specified.

[micaeljtoliveira]

@ezhilsabareesh8 Thanks for the changes! Just a couple more comments. We are almost there! :)

[aekiss]

Thanks @ezhilsabareesh8, looks good but I've made some suggestions to hopefully improve clarity.

[aekiss]

Also, would it be better to avoid unit conversion for <cutoff_value>, ie use the same units as ocean_hgrid.nc (m) instead of km? Some advantages are:

• a user familiar withocean_hgrid.nc would know that the units are m and would probably assume that the units for cut_off_T_cells are the same
• MOM uses SI base units throughout, so it could trip up users to remember to use km in this one case
• making no assumptions about the units used for dy makes it more general (ie fewer changes needed for it could work for grids of other models using other units) - the documentation could simply say "<cutoff_value> is the minimum T-cell y-size in the units used for dy in ocean_hgrid.nc"

[micaeljtoliveira]

Thanks @ezhilsabareesh8 for the latest changes! All good from my side, so I'll approve.

[ezhilsabareesh8]

Thanks @micaeljtoliveira and @aekiss, all the comments have been addressed now and I checked the output, it looks fine. I will merge this PR if there's no further comments or suggestions.

[aekiss]

Thanks @ezhilsabareesh8, I've suggested some tweaks to the documentation. Otherwise it looks great.

[ezhilsabareesh8]

Thanks @aekiss, I have made the suggested changes.

[aekiss]

Thanks @ezhilsabareesh8, I'm happy for this to be merged

[ezhilsabareesh8]

Thanks @aekiss, Could you please approve this PR?