Halo misalignment for ocean surface velocity and slope

[aekiss]

Just preserving @russfiedler's Slack comment before it disappears
https://arccss.slack.com/archives/C9Q7Y1400/p1613975875009400

[aekiss]

Russ wrote:

I think I may have found the source of the tripole shear problem. Ocean velocity ( ssuo, ssvo) is supplied to CICE on the u grid but its halos are updated as if it was at the cell centre. https://github.com/COSIMA/cice5/blob/master/drivers/auscom/cpl_interface.F90#L662-L663 (the slope of the ocean is also misaligned). This means that the northern halo is treated as -u(M-m+1,N) rather than -u(M-m,N-1). Relative velocity is used in the elastoviscous solver amongst other things. I'm pretty sure it isn't regridded before this. Can anyone check? The call to start updating the halos is here https://github.com/COSIMA/cice5/blob/master/drivers/auscom/CICE_RunMod.F90#L180

and

On further investigation it doesn't look like velocities in the halos are used so it's probably ok. The calls should be fixed though to field_loc_NEcorner for the velocities https://github.com/COSIMA/cice5/blob/master/drivers/auscom/ice_constants.F90#L199-L206

[aekiss]

This is the tripole shear bug Russ referred to: COSIMA/access-om2#86