[MITgcm-support] external&internal mode, boundary

Martin Losch Martin.Losch at awi.de
Wed Jan 17 15:48:45 EST 2007


Dimitris,

First of all, please excuse my English, how can one make so many  
mistakes in just one sentence.
Here's an excerpt from OBCS.h:
> C For convenience, negative values for Jnorth/Ieast refer to
> C points relative to the Northern/Eastern edges of the model
> C eg. OB_Jnorth(3)=-1  means that the point (3,Ny) is a northern O-B.
And that's how I understood it.
Here is code from obcs_readparms.F:
>        IF (OB_Ieast(J).lt.0) OB_Ieast(J)=OB_Ieast(J)+Nx+1
> [...]
>        IF (OB_Jnorth(I).lt.0) OB_Jnorth(I)=OB_Jnorth(I)+Ny+1
which supports my interpretation (in better English now):
OB_Ieast(5) = -1 turns the last grid point (with index nx) in x- 
direction (of a "row" with index j=5) into an open boundary point,  
just as OB_east(5) = (value of)nx would. That's the point with the  
indices (nx,5). Likewise, OB_Jnorth(11)=-1 makes the point (11,ny) an  
open boundary point.

Martin


On 17 Jan 2007, at 18:07, Dimitris Menemenlis wrote:

> Martin, Patrick, or JM, just to make sure that I understand this  
> correctly:
> "OB_Ieast = 0 make to the last (nx) grid point in x-direction a  
> boundary value"
> Not "OB_Ieast = -1".  Right?
>
>> I don't understand the question. OB_Ieast = -1 make to the last  
>> (nx) grid
>> point in x-direction a boundary value, just a, OB_Ieast = (value  
>> of)nx would.
>> OB_Iwest(1) makes the first grid point in x-direction a boundary  
>> point etc.
>> In the code sometime the values are applied to more than one grid,  
>> but only
>> on the outside of the domain.
> _______________________________________________
> MITgcm-support mailing list
> MITgcm-support at mitgcm.org
> http://mitgcm.org/mailman/listinfo/mitgcm-support




More information about the MITgcm-support mailing list