[MITgcm-support] internal tide beam in nonuniform stratification

[Martin Losch]

Hi Sherry,

this is a quick guess:
with your resolution you are in a regime, where non-hydrostatic dynamics start to have an impact (still, the nh-pressure is small).
You could try turning on the non-hydrostatic code, but I would use only very few iterations for the 3D solver (cg3dMaxIter = 40 or so), in order to reduce the computational cost. The least that will happen is that you add dissipation to the vertical velocity field that may stabilize your simulation.

M.

> On 02 Feb 2016, at 07:31, Sherry <schou at hawaii.edu> wrote:
> 
> Hi everyone,
> 
> I am a grad student trying to model internal tide beams as they propagate up towards the surface after being generated in the open ocean. 
> 
> I started with the internal wave example in verifications and modified obcs_calc.F to make an upward beam that's the sum of 10 modes.
> I've been encountering a barotropic instability that starts in w, which is sometimes worsened by decreasing Uinflow or deltaT, or increasing the number of points per process. (Decreasing the forcing frequency also led to increased instability, and switching to rigid lid improved this.)
> 
> Right now I have a (seemingly) stable configuration for H=4200m (dz=20m) and Lx=120km (dx=120m), with uniform stratification, that ran for more than 22 wave periods (M2 period = 12.42 hrs). I am interested in the effect of different stratification profiles, and would like to start with a simple Gaussian shape where the maximum bvsq is ~1e-4. The attached gendata.py (it is written in python) makes a temperature profile that has max bvsq ~2e-5. When I increase max bvsq slightly to ~3e-5 (by increasing the drho_rel parameter to 0.2*3.6e-3), w-velocity becomes unstable after 1 wave period. 
> 
> Any kind of help for this specific configuration, or for learning how to deal with the model blowing up in general, would be much appreciated.
> 
> Thank you,
> 
> Sherry Chou
> schou at hawaii.edu
> 
> Department of Oceanography
> UH Manoa, Hawaii
> 
> <image.png>
> <obcs_calc.F><data><gendata.py><Tref.be>_______________________________________________
> MITgcm-support mailing list
> MITgcm-support at mitgcm.org
> http://mitgcm.org/mailman/listinfo/mitgcm-support