[MITgcm-support] anelastic configuation for Jupiter

[Roland Young]

Hi Eli,

I also had a similar problem of small-scale velocities near the poles when running a shallow Jupiter configuration (both with the MITgcm and with another model, but I only managed to fix it for the MITgcm). I’m not running the anelastic version (only goes down to 18 bars) and the density won’t change as much as in your case. I found I was able to fix it by using a 4th order Shapiro filter with the damping tau set to one time step. Any tau longer than that I got small-scale noise near the poles. Also I have the zonal filter activated, but I assume you have that already included as otherwise it should crash very quickly? 

My model goes all the way to the poles - it may interest you to know that before I solved the problem I did try putting the latitude boundaries at +/- 80 degrees (following Yohai’s configuration) but even with the Shapiro filter it did not fix it - I had to push it to 90 degrees to get the small-scale noise to go away. If you can solve it for a non-global domain I would be interested to know how (I think there were problems with the boundary conditions, even with a fudge using bathymetry,  but I forget now as it was a while ago). In any case, in my opinion there is not much to be learned near the poles in the giant planet configuration - the multiple numerical fudges you have to use near the poles make it hard for me to believe anything I see there.

Glad to see someone is continuing the deep giant planet work - it is very interesting.

Roland