[MITgcm-devel] [acondron at whoi.edu: Re: advection of freshwater from ice]

[Martin Losch]

Advection in pkg/seaice (without pkg/thsice) works the same on both B  
and C-grid.
For the default advection scheme (2), you need to have some  
horizontal diffusion turned on (which is at DIFF1=0.004 by default).  
For the flux-limited schemes (33, 77) and for the 1st order upwind  
scheme (1), setting DIFF1=0. is probably a good idea.

advection of snow is NOT turned on by default.

I am using SEAICEadvScheme = 1, 2, and 33 for my coarse resolution  
runs, but 77 and 20 should work, too (all flux limited schemes,  
except for 7, which is not implemented for seaice as far as I know).

Dimitris: 77 is not the flux limited version of 7, but a 2nd order  
(central differences) scheme with flux limiter (superbee by default).

Recommendation: I would use a flux limited scheme for seaice, e.g. 77  
or 33.

For thsice the story is different, and we only have 1 and 77 working  
so far, right, Jean-Michel?

Martin

On 11 Jun 2007, at 20:19, Dimitris Menemenlis wrote:

> Jean-Michel and Martin, I have same questions as Alan.
>
> Which SEAICEadvScheme is it OK to use with pkg/seaice B-grid and C- 
> grid code and which is it recommended to use?
>
> I also mostly use the default SEAICEadvScheme=2.  I once  
> experimented with "1" on the CS510 grid and the solution seemed way  
> too smooth.
>
> Does 33 and or 77 work with either the B-grid or the C-grid solvers?
>
> Finally, in latest series of CS510 tests, I am using
> 7th Order One Step method with Monotonicity Preserving Limiter
>  tempAdvScheme=7,
>  saltAdvScheme=7,
>
> Should I instead be using the flux-limited variant: "77"?
>
> Dimitris
> _______________________________________________
> MITgcm-devel mailing list
> MITgcm-devel at mitgcm.org
> http://mitgcm.org/mailman/listinfo/mitgcm-devel