MITgcm periodic boundary

[ah-dinh]

Hi,

I am trying to use periodic (longitude) boundary on an MITgcm dataset.

XG[0] = -180
XG[-1] = 179.95834

I get a "FieldOutOfBoundError" when I try to sample the last column of the dataset, basically anything from 179.95834 to 180.

I see in the tutorial that a halo and a periodic boundary kernel is need for an A-grid dataset. MITgcm is C-grid, but do I need to do something similar as well to get this to work?

[erikvansebille]

Hi @ah-dinh, thanks for asking. While normally we only advice to use fieldset.add_periodic_halo() for A-grid, there is no a-priori reason it wouldn't work on C-grids.

Perhaps the confusion lies in the point that we advise against using halo's for ORCA (NEMO) grids, because these already include a halo (i.e. XG[-1] > XG[0] in your notation above).

If that is not the case in your MITgcm grid, then using a halo would be a good fix I reckon. Alternatively, you can also try the 'hycom-solution' (adding an extra column of longitudes at the end of the longitude array) proposed in the second cell of the periodic boundaries tutorial.