[MITgcm-devel] conservation Problem with KPP viscosity

[Dimitris Menemenlis]

Jean-Michel,

> OK like this ?

Of course!  Thank you for fixing bug.  Strange though.  I still do not 
understand how an error in viscosity can impact T/S conservation.  Is it because 
there will be some difference in the velocity near the edge of a tile and that 
in the halo of the adjacent tile?

Way back I remember doing some halo tests with KPP, i.e., making sure that I get 
the same answer with different number of tiles and also with or without 
FRUGAL_KPP on.  But that was before the appearance of exch2 and of the cube.

D.