[MITgcm-devel] Visca4 diagnostics

[Andrea Molod]

hi alistair,

> 
> > Andrea/Ed,
> > 
> > The diagnostics we added yesterday are producing numbers but seem
> > wrong. The problem we had yesterday is because mom_vecin never got 
> > properly updated to use PACKAGES_CONFIG.h. It is now.
> > 
> > For the visc.*.data output:
> > 
> >  Level 1 seems OK
> >  Level 2 seems shifted by one point in the Y direction and has funny
> > stuff at i=1,N  Level 3,4 are zeros
> >  level 5 seems shift by one point in the Y direction
> > 
> > I'm not sure whether the arguments to fill_ are the problem or perhaps
> > the internal call of mdswritefield_new() is wrong.
> > 



indeed, there was a case not handled in fill_diag - the case of
sending an overlapped, local single-level array to fill a multi-level
diagnostic. that has now been added, and is flagged with a NEGATIVE
level number -- i also changed the call sequence to send the negative
level in mom_vi_hdissip --- the numbers look more coherent now, at
least to me - can you check it?

andrea