[MITgcm-devel] ad test on lab_sea exp.

[Dimitris Menemenlis]

Patrick, sorry for KPP mess.  Changes are as discussed with JM last week
and have to do with treatment of background vertical diffusivity.

Specifically:

1) kpp_calc calls calc_3d_diffusivity with trUseKPP=.FALSE. in order to compute 
background diffusivity.

2) background diffusivity is temporarily stored in the KPPdiffKzS/T arrays and 
passed to subroutine kppmix, where PBL, convective, and shear instability 
contributions are added.

3) calc_3d_diffusivity is called a second time by thermodynamics with 
trUseKPP=.TRUE. to transfer the KPP values to KappaRTr