[MITgcm-support] Re: nonHydrostatic

[Dimitris Menemenlis]

Christopher, thank you for explanation.  This is then the answer to Geoffroy's 
conundrum.  That is, since the only bathymetric feature in his domain is a thin 
vertical wall that does not come all the way to the surface, salinity is 
everywhere uniform, and temperature is stably stratified and horizontally 
homogeneous, I can imagine that his circulation would be marginally stable, 
meaning that it may be possible to relax to a uniform stratification simply via 
the explicit diffusion terms, horizontal and vertical, without exciting any 
motion.  The slightest disturbance to this equilibrium, however, e.g., 
truncation errors in the salinity diffusion equation, could initiate vertical 
convection events and currents, exactly as Geoffroy observed.  D.

> Dimitris,
> 
> This behavior is entirely physical, but I wouldn't expect to see it in the
> real ocean. Even if you start out with a stable stratification, there's going
> to be diffusive fluxes as the system tries to adjust to a uniform
> temperature. If the bottom is flat then the bottom boundary condition is
> horizontally uniform and the density remains horizontally uniform throughout
> the adjustment process. With topography, location of the bottom boundary
> condition is no longer horizontally uniform, so horizontal density gradients
> develop and the velocity becomes non-zero.
> 
> Also, if your topography pokes up into a region of strong stratification,
> there's going to be some rapid local adjustment to satisfy the bottom
> boundary condition T_z = 0, which will likely create horizontal
> inhomogeneities in the temperature field, leading to flow.
> 
> This shouldn't have anything to do with whether the model is hydrostatic or
> not, although the details of the adjustment might.
> 
> Christopher