[MITgcm-devel] Bug in KPP

[Dimitris Menemenlis]

David you are right and I am wrong.  Ocean surface velocity is not
used for determining boundary layer depth in stable conditions.
I was not pulling your leg.  I replied without bothering to check
code or the Large et al paper and it's been a while since I went
through either.

Still puzzled by your situation though.  Rereading the Large et al 1994
paper (and now I do have it in front of me), it seems to make sense
that for stable forcing the boundary layer depth should be less than
the Monin Obukhov length and the Ekman depth.  In your situation
however, you are introducing destabilizing heat flux not from the
surface but from the side.  Maybe the condition for computing
the stability criterion should be modified to account for lateral fluxes
in the model's surface level in addition to surface fluxes proper?
I am a bit uncomfortable at dropping the LimitHblStable condition
altogether, except for very particular situations, e.g., your channel  
experiment.

Cheers, D.

On Jun 23, 2009, at 1:13 PM, David Ferreira wrote:

> Are you pulling my legs ?
> You acknowledge that the problem is due to the HBL not being a  
> function
> of Ri_c (under stable
> conditions), but you still think it has to do with the vertical
> resolution, right ?
> Seriously, I don't see where you are going. How the vertical  
> resolution
> would remove the
> static instability here ?
> david