[MITgcm-devel] advcfl_W_hf_max

[Dimitris Menemenlis]

Alistair, I just tried to call to clear things up but you were out.

The code you sent me is for biharmonic momentum/viscosity and it works
fine except that the scaling at the cube corners does not seem quite
right.  They go unstable first and they set an upper bound for
non-dimensional viscA4 to about 4e-3.

Biharmonic diffusion for salt and theta did not work when I tried it so
I turned it off.  But maybe it was the values that I picked?  Altix is
hung right now.  I'll get you the data files as soon as it comes back
up.

>> Also, on what basis are you editing topography? It seems
>> incredible to us that you know how to change topography in
>> order to fix a blow-up. 

Pretty ad hoc but it works.  I locate the region where the code blows up
and then smooth the bathymetry with a 121 filter for a few points around
the problem grid to get rid of deep, narrow channels.  We had to do this
three or four times in the 1/4-deg code.  For the cube sphere there is
only one change (three points on the figure I sent in an earlier
message).  I wish I had some recipe for doing this all at once, but
without getting rid of Gibraltar, etc.

D.

-- 
Dimitris Menemenlis <menemenlis at jpl.nasa.gov>
Jet Propulsion Lab, California Insitute of Technology
MS 300-323, 4800 Oak Grove Dr, Pasadena CA 91109-8099
tel: 818-354-1656;  fax: 818-393-6720