[MITgcm-devel] Re: recent changes not tested

[Jean-Michel Campin]

Hello Martin,

Thank for your answer.
A simple verification, where you can see what are the parameter
to use and demonstrate how it works would be great
(in the same spirit as front_relax for GM).
And from what you say, it's exactly what we need for now 

For a practical point of view, you can, within a single 
verification exp, test different schemes (like global_ocean.cs32x15, 
standard one + thsice one) as long as you use the same executable.

For the other thing, I am working on this modification of 
the implementation (specially for ptracers), but there is
some issues related to the adjoint.

But regarding the background value, just to remain simple:
if you replace, e.g. in pp81_calc_diff.F
KappaRS(i,j,k) = KappaRS + (PPdiffKr - diffKrS)
with
KappaRS(i,j,k) = KappaRS + (PPdiffKr - diffKrT)
you will keep the possibility to use a background value 
of diffKrS that is different from diffKrT.
And, of course, nothing change if you set them to the same value.

Jean-Michel