[MITgcm-support] Internal wave

[Jody Klymak]

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>