[MITgcm-devel] Re: [MITgcm-cvs] MITgcm/model/src CVS Commit

[Dimitris Menemenlis]

JM, what is "Pb in indices"?

You are right it could be costly.  Should I transfer this stuff to ini_mixing 
with "defined ALLOW_3D_DIFFKR".  Or do you prefer that I leave it in 
MITgcm_contrib/high_res_cube, where I intend to use it.

In the high-res integrations we need to vary background vertical diffusivity 
horizontally in order to get thermocline correctly both at high latitudes and at 
the Equator.  The stuff I checked in is pretty similar to what MOM has already 
been doing for years and is also supported by theoretical and observational 
evidence.

D.

> Hi,
> Apart from Pb in indices, computing the cos of lat at every grid point
> at every level and for every tracer and at every iteration might cost 
> some cpu time on some platforms. 
> So, should be avoided if no intention to use this feature.
> Jean-Michel