[MITgcm-devel] Hack to increase mixing near bottom/surface

[Jean-Michel Campin]

Hi Dan,

Thanks for your comments. I will try to answer the 3 points below.

On Mon, Dec 21, 2015 at 05:28:53PM -0500, Daniel Goldberg wrote:
> Hi J-M and Martin:
> 
> JM, this is very interesting and looks great. m44 and m43 show very smooth
> profiles, not sure if there is a rationale for one over the other. With
> regard to the differences in melt rate magnitude, it is difficult to say
> but as Martin mentions the melt rate will depend strongly on u*. So while
> there may be differences in the temperature and salinity used in the melt
> rate calculation (which will differ between shelficeBoundaryLayer and the
> various mixing parameterisations) a difference in the way u* is calculated
> might make a large difference. Possibly u* is larger in magnitude for wet
> point averaging because in the original u* formulation wet cells are being
> averaged with dry cells. I saw this as well in the formulation that i
> showed you. Due to all the different settings, though, it is difficult to
> say which differences are due to the differing u* formulation and presence
> of the boundary layer.
> 
> Some statements/questions:
> 
> 1) Is mixSurf a new parameter? that is, based on your notes
> http://mitgcm.org/~jmc/pCellMix/pCellMix.notes,
> i thought that pCellMix_select=40 had the same effect as mixSurf=4
mixSurf is not a new parameter, but it's a shortcut for the "tens digit"
part of pCellMix_select, which control the surface BC effect of pCellMix.
So yes, pCellMix_select=40 means mixSurf=4

> 2) Have you set interDiffKr_pCell=T in any of these tests? I thought this
> gave a more accurate vertical discretisation near the surface. Or should it
> not impact this at all?
I did not check again, but I think that the effect of interDiffKr_pCell=T
is masked when I set mixSurf=1 or higher, because the increase of mixing due 
to mixSurf is larger than the one due to interDiffKr_pCell.

> 3) In this test, hFacSurfC <= 1, which was the original intention for
> shelfice. But for coupling we are interested in the case where hFacC can be
> slightly larger than 1, though we try to limit this to ~1.3.
This (hFac > 1) would need to be tested. Right now, pCellMix is not doing
anything when hFac > 1, but it might just be OK.

Cheers,
Jean-Michel

> 
> Dan
> 
> 
> On Sun, Dec 20, 2015 at 10:28 PM, Jean-Michel Campin <jmc at ocean.mit.edu>
> wrote:
> 
> > Hi Martin,
> >
> > Thanks for your explanations.
> >
> > I updated my "pCellMix" hacked code (http://mitgcm.org/~jmc/pCellMix/)
> > which now allows for larger increase of visc & diffK (~ 1/hFac^n,
> > n=1,2,3,4).
> >
> > I repeated some of these tests with Dan's 2-d set-up but modified so that
> > it contains some smaller hFac (down to 0.05), and the improvement seen
> > before
> > (with pCellMix_select=30 or 40) remains.
> > I guess to test this in a set-up with rotation, one could use one of the
> > isomip experiment ?
> >
> > Cheers,
> > Jean-Michel
> >
> > On Fri, Dec 18, 2015 at 01:36:08PM +0100, Martin Losch wrote:
> > > Hi Jean-Michel,
> > >
> > > I am not sure if I can contribute  to this, but I can speculate about
> > your question:
> > >
> > > > I wonder if this large sensitivity of melt-rate when using
> > SHELFICEuseGammaFrict=T
> > > > is common ? or is it due to no-rotation ? or maybe there is something
> > wrong in
> > > > my wet-point averaging code (now cheked-in) ?
> > >
> > > when using SHELFICEuseGammaFrict the melt rates become much more
> > sensitive to the details of the flow regime. With constant exchange
> > coefficients (SHELFICEuseGammaFrict=F), the heat flux/melt rate is
> > determined by the temperature difference to the freezing point, i.e. the
> > deeper, the more melting with the same potential temperature, and the
> > warmer the temperature the more freezing, and it???s just a matter of where
> > the warm water is brought into contact with the ice shelf bottom. With
> > SHELFICEuseGammaFrict=T, the highest melt rates are found where the
> > under-ice currents are larges (and uStar is largest). I assume that
> > changing the vertical mixing with your ???hack??? is enough to
> > substantially change the circulation and hence the strength of the boundary
> > currents under the ice; or modifying uStar by averaging changes the actual
> > strength of the exchange and hence the melting.
> > >
> > > Am I stating the obvious?
> > >
> > > M
> > > _______________________________________________
> > > MITgcm-devel mailing list
> > > MITgcm-devel at mitgcm.org
> > > http://mitgcm.org/mailman/listinfo/mitgcm-devel
> >
> > _______________________________________________
> > MITgcm-devel mailing list
> > MITgcm-devel at mitgcm.org
> > http://mitgcm.org/mailman/listinfo/mitgcm-devel
> >
> 
> 
> 
> -- 
> 
> Daniel Goldberg, PhD
> Lecturer in Glaciology
> School of Geosciences, University of Edinburgh
> Geography Building, Drummond Street, Edinburgh EH8 9XP
> 
> 
> em: D <dgoldber at mit.edu>an.Goldberg at ed.ac.uk
> web: http://ocean.mit.edu/~dgoldberg

> _______________________________________________
> MITgcm-devel mailing list
> MITgcm-devel at mitgcm.org
> http://mitgcm.org/mailman/listinfo/mitgcm-devel