[MITgcm-devel] negative H in calc_r_star.F

[Dimitris Menemenlis]

Jean-Michel, yes, that's exactly how I understand problem,
except that I am not familiar with the thin wall issues.

Although no water can be exchanged between these
two columns, the columns can exchange water with
other adjacent columns.  With hFacMinDr=50
we did not have any such cases but with
hFacMinDr=0, many such cells appeared, the blue
cells in Hong's lower-right panel of ppt he sent earlier.

(Hong, we will be reprimanded tomorrow morning by
Martin for sending a ppt file --- Martin to Hong's defense,
he generates his slides using Open Office ;-)

Cheers

Dimitris Menemenlis


On 05/30/2012 12:42 PM, Jean-Michel Campin wrote:
> Hi Dimitris and Martin,
>
> If I follow what has been written here:
> - the code stop (when using rStar) but the negative H is
>    independant of rStar.
> - it happens where the bottom of the ice-shelf is lower than
>    the bottom of the ocean in the neighbour column.
>
> It's an unusual case (no water can be exchanged between the
> 2 columns, but none of them is dry), but it should be possible to
> handle this case (might need to check a couple of things).
> In some ways, it's similar to "thin-walls" capability (but we know
> that several pieces of code don't work with thin-walls).
>
> Martin, I am not sure that pkg/shelfice has been written with this
> case in mind. Might need to revise shelfice_update_masks.F for
> this special case.
>
> Do you agree on this ?
>
> Cheers,
> Jean-Michel