[MITgcm-support] eta in the atmospheric model

[Jean-Michel Campin]

Hi Roland,

Eq. 1.55 and description of "eta" in 3.6.3.1 are correct,
the surface pressure is: ps = Ro_surf + Eta,
where Ro_surf is fixed in time, from GRID.h: 
> surface reference (at rest) position, r_unit.

Now if you don't have any mountain then Ro_surf will be uniform
and equal to rF(1) (= Ro_SeaLevel, but it's now depreciated
since it's generally more convenient to set "top_Pres" = rF(Nr+1)).

Regarding question (2), it depends if you are using linear-free surface
(in this case, the geometry does not change and the bottom of
the lowest grid cell is at Ro_surf) or if you are using non-linear
free-surface (nonlinFreeSurf > 0), and in this case, the bottom of the 
lowest grid-cell is at ps.

And regarding question 3, it depends if you are using non-linear FreeSurf
in p coordinate or in p* coordinate (select_rStar=2).
I would suggest to read: Adcroft & Campin, Ocean Modelling (2014),
with, in appendix B, the atmosphere set of equations.

Cheers,
Jean-Michel

On Fri, Nov 09, 2018 at 11:49:07AM +0100, Roland Young wrote:
> Hi,
> 
> I???m a bit confused about what eta is in the atmospheric model (partly because the units in the output file are for the ocean model). In particular the relationship between the surface pressure (which appears to change) and the model cell centres and faces (which appear to be fixed).
> 
> The surface boundary condition given by Eq. 1.55 is p = ps, and (I thought, at least until now) that ps is constant and specified in the ???data??? file as Ro_SeaLevel.
> 
> But Sect. 3.6.3.1 describes eta as surface pressure anomaly, which implies that the surface pressure changes in time.
> 
> So (1) in the atmospheric model, is ps = Ro_SeaLevel + eta?
> 
> (2) Is the lowest grid cell face at Ro_SeaLevel (fixed), or ps (varying)?
> 
> (3) The vertical grid cell faces are at rF, and rF[1] = Ro_SeaLevel. When eta<0, does this mean that the lowest cell face is below the model surface? If so, how are you supposed to interpret the vertical velocity there? Similarly, if eta is large enough such that it is more negative than rC[1] - rF[1], how should I interpret the values of the prognostic variables at the first cell centre?
> 
> Thanks,
> 
> Roland
> 
> 
> 
> _______________________________________________
> MITgcm-support mailing list
> MITgcm-support at mitgcm.org
> http://mailman.mitgcm.org/mailman/listinfo/mitgcm-support