[Mitgcm-support] Re: some progress

[mitgcm-support at dev.mitgcm.org]

Hello Alistair,

> There was a reason you wanted to move the filters - do you remember
> what?

I remember quiet well:
with topography, u*,v* are far from beeing representative of
a true velocity field since eta have strong graqdient associated
with topography.
But this was tested with Shapiro only, also at this time I moved 
the 2 filters (FFT & Shapiro).

Since FFT has a much bigger impact on the divergence of 
the flow than Shapiro, it's possible what we end up with 
a different position of the 2 filter in the code.

What is clear is that with rigid-lid, only u*,v* can be
filtered.
With flat topo and free surface, I would like to compare the 
2 options:
* filtering u,v and use exact-conserv,
* filtering u*,v* 
and make the test with FFT on.

See you,

Jean-Michel