[MITgcm-support] Internal wave
Jody Klymak
jklymak at uvic.ca
Fri Aug 17 13:01:04 EDT 2007
Hi Abbas,
I am hardly a great expert on the MITgcm, but I've used it for
similar cases to what you seem to be after.
On Aug 17, 2007, at 9:20 AM, Abbas Dorostkar wrote:
> I am going to set up the MITgcm for Cayuga Lake, NY. I have run
> internal wave test case but I would like to clarify following
> questions for myself.
> 1) The nonhydrostatic flag is false and model is using semi-
> implicit solver (Am I right?). I am guessing in internal wave on
> slope, nonhydrostatic process could be more important especially in
> small scale (like lab scale)? Have anybody testes this model for
> smaller cases?
Non-hydrostatic should work fine. Just be aware that it will be much
slower. You could try hydrostatic first to get things looking right,
and then move to non-hydrostatic after. You may need to change
deltaT to something smaller as well depending on what mixing scheme
you decide to use.
> 2) I saw a flag "implicitIntGravWave" (default=false) in the list
> of parameters but I could not find any documentation. Can I use it
> in internal wave test case? If yes, Should I use in conjunction
> with nonhydrostatic= ‘True’?
I'm not sure about this one, but I do IW simulations with this set to
default of false.
> 3) I wonder how we can define boundary condition on the slope in
> this case. Is it the same as bottom boundary condition?
The slope would be called the bottom if you are asking about
something like where "no_slip_bottom" would be applied. You want the
friction to be along in the vertical direction. BTW, I run this as
FALSE, primarily because I don't have enough resolution to resolve
the bottom boundary layer. If you want there to be bottom drag you
can also specify bottomDragQuadratic=0.003.
> 4) I am not sure which kind of OB is used in this test as it is not
> defined in data.obcs file but by seeing obcs_calc.F , I guess
> Orlanski?
I don't think that example uses Orlanski, or any other radiative
scheme. The obcs_calc.F is making there be a sinusoidal mode-1 wave
in U,V, and temperature on the western boundary. This wave is
allowed to radiate in, onto the shelf. Anything that back radiates
is allowed to bounce around.
> 5) Instead of open boundary forcing, could I use a close boundary
> with a two layer prescribed density profile?
Yes, just set the ends of your box to be depth=0.
> 6) In packages.conf of this test, there is a gfd package. What does
> it do? I could not find this package in pkg folder of MItgcm.
gfd is a short form for a whole bunch of packages. When you run
genmake2 it tells you what they are.
I hope this helps. More complicated questions may be beyond my
expertise!
Cheers, Jody
--
Jody Klymak
http://web.uvic.ca/~jklymak/
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mitgcm.org/pipermail/mitgcm-support/attachments/20070817/0f730597/attachment.htm>
More information about the MITgcm-support
mailing list