[Mitgcm-support] Open boundary troubles

[mitgcm-support at dev.mitgcm.org]

Hi,

I'm having some problems with open boundary conditions with the
MIT GCM.  I'm trying to use ecco_c44_e24.

I've been doing test runs with a single layer, in an open box (1000m
deep everywhere) with a spatially sinusoidally varying input current
(fixed in time). I also doing some with more realistic bathymetry. All
runs are at 1/12 degree. The problem I'm having is that :
1) Unless I use implicit viscosity the model crashes after 2 days. If I
use implicit diffusion the temperature and salinity (appear?) to go to
zero after 1 timestep. I can operate the model with implicit viscosity
and explicit diffusion, however.
2)  The SSH is smooth, but within the domain it varies spatially
unrealistically on the ~10-100m for open boundary currents of ~0.1 m/s.
(perhaps I've got the units wrong?)

Have I made an obvious error somewhere? Any idea what I need to do to
make it work?

Cheers

Daniel Lea
Johns Hopkins Applied Physics Lab