[Mitgcm-support] Re: c27

[mitgcm-support at dev.mitgcm.org]

OK, last message I write today on this topic.  I will try to cover everything
that was directly or indirectly addressed to me, but let me know if I miss
something and I'll get back to you all tomorrow.

Chris, I did talk with Alistair.  The agreement is to test everything using
Patrick's SubAtl.  That way we can compare Classic, JPL, and new UV code
directly.  I will make this a priority here at JPL.

One remaining issue with KPP, is how to make it a bit less dependent on
external forcing particulars.  Right now total external forcing (in JPL's case
flux + relaxation) is computed twice, once by external_forcing_T/S and once by
kpp_transport_T/S.  We can agree on a cleaner way to do this if you like.

Detlef, the KPP values that we do "tune" at JPL are the background (internal
wave) mixing coefficients.

Other diagnostics which we output at JPL are the various diffusion and
advection contributions to heat and salt.  But they would be quite costly for
a 1/6 degree integration.

Dimitris