periodic boundary condition

[jifoyi]

I am now trying to add a periodic boundary condition in xbeach. But I
find it is difficult to do as I can't understand where is the true
boundary of this model.
The uu is updated in (2:nx-1 & 2:ny) while vv is updated in (2:nx &
2:ny-1) in excluding mpi solution. But why uu(nx,:) and vv(:,ny) not
updated. Is this because of the mpi solution??
What is the real use of line nx+1 and ny+1 ??
If I want to add periodic boundary condition in mpi excluded solution,
do you think copy value of zs,uu,vv in 2 to ny and ny-1 to 1 would be
ok this is because vv(:,ny) don't update.


From Robert:

The reason that uu(nx,:) and vv(:,ny) are not updated in the flow
solver is because these are the model boundaries (expect in case of
MPI simulations) for u and v. uu(nx+1,:) and vv(:,ny+1) are dummy
points to ensure grid symmetry and are not used in any calculations.
If you want to impose periodic boundary conditions you would indeed
have to copy zs,uu,vv from 2 to ny and from ny-1 to 1.

A word of warning if you are using periodic boundary conditions: these
boundaries force/reinforce certain longshore wave lengths, and will
not allow other longshore wave lengths to exist in the model. This is
why we normally apply neumann boundary conditions.