[Mitgcm-support] kpp update

[mitgcm-support at dev.mitgcm.org]

Patrick, there is a KPP update in
/ocean/data4/dimitri/mitgcmuv.jpl.c19_natl.tar.gz
The pkg/kpp routines, with the exception of diagnostics
can be imported into c31+ as is.

The main changes are that:

a) cpp option FRUGAL_KPP is now compatible with horizontal
   smoothing, so that horizontal smoothing can
   be used with OLx=OLy=1 when FRUGAL_KPP turned on,

b) that kpp_check checks OLx and OLy so that you don't
   run into im1/ip1 array range subscript problems again, and

c) that a bunch of test and development code has been
   removed to improve readability.

>> why are there differences between c30 and c31+,
>> and why are there even differences when switching off
>> KPP in different ways (not even compiling it vs.
>> compiling it but switching useKPP to .FALSE.

I have verified that c31, with modifications I sent
you a couple of weeks ago gives identical results whether
KPP is turned off using cpp or runtime flags (as does the
JPL code).  Did you run into this problem with c30, c31,
or both?

I had a quick look at ecco_c30_e1.  Do you really want
to use this code as a common basis?  I hope not!
If you do, then KPP units, surface forcing variable names,
and relaxation, shortwave components have to be changed
yet again.

D.