[MITgcm-support] passive constant homogeneous salinity blows up

[Madeline Miller]

Hi support,

I have been dealing with an interesting MITgcm issue in which the salinity
value causes the model to blow up over a single timestep (100s, 50s, 20s,
10s...). It is interesting because the salinity is passive (linear EOS with
sbeta=0), constant, and homogeneous, and is not forced by any external
packages/fields.

The problem seems to come from gS accumulating rounding error over a long
integration time. I also found a 13-year-old (!) email to the list from
Martin about a similar issue:
http://mitgcm.org/pipermail/mitgcm-support/2003-September/001697.html

It has been so long that the code is significantly different, and gS is no
longer reset in thermodynamics.F

Can anyone suggest an appropriate place in the code to reset gS so that the
drift doesn't run away with itself?

Best,

Madeline
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mitgcm.org/pipermail/mitgcm-support/attachments/20160610/97c185bf/attachment.htm>