barotropic equation in ROMS

[aryapaul]

When ROMS is run in the barotropic mode, what is the explicit form of the barotropic equation for SSH that is being solved ?

[kate]

It is equation 30 in the manual.

[aryapaul]

Thanks Kate for the reply.
But what's the forcing term look like ? Is it curl of the wind stress/H where H is the bottom topography ? Does it have any bottom friction term ? Is there no dependence on beta ?

[kate]

The forcing terms are in the momentum equations. The equation for SSH simply follows from mass convergence/divergence based on the velocities. The momentum equations get the wind stress and the bottom friction, also Coriolis.

[rduran]

ROMS doesn't solve vorticity equations, solve primitive equations.

[aryapaul]

Thank you for your response. I got the equations. However if I want to run the model in the barotropic mode dropping all the non-linear terms from the equations for some controlled experiments, how do I do it ?
Also, how is it possible to generate barotropic boundary conditions ?

[kate]

For the linear equations, turn off the UV_ADV option.

Could you be more specific about your boundary condition question? You need to provide ubar, vbar, zeta on all open boundaries. Is this an idealized problem with ana_m2obc? Something more realistic with say global HYCOM to extract from?

[aryapaul]

We are trying to run roms in the barotropic mode for the Indian Ocean wherein there is open boundary in South, East & West. I was wondering how to generate daily realistic boundary condition for ubar, vbar and zeta for any particular year ( say 2011 ).

[kate]

There's more than one way to do it. I have Python scripts for (a) downloading a subset of global HYCOM fields for a specific region and (b) interpolating them to a ROMS grid. They provide zeta, u, v. The scripts will compute ubar,vbar from u,v. If you are allergic to Python there's likely something similar in Matlab - check out the Matlab tools from the ROMS site.

You probably want to build your grid file first.

[aryapaul]

If we put NDTFAST > DT, does that mean that the ocean state will be purely baroclinic in nature and all barotropic contributions will be filtered out ?

[kate]

NDTFAST is the ratio between the barotropic and baroclinic timesteps, usually set to something like 20. If you set it to 1, the timesteps are the same - but I've never run that way and the ROMS machinery will do its thing anyway, possibly messing it all up. A lot of thought has gone into the ROMS split timestepping and I don't think you can turn it off to run with NDTFAST=1.

[aryapaul]

Thanks for the reply Kate. Actually, I wanted to know if ROMS can be run only in the baroclinic mode thereby avoiding all barotropic contributions ?

[kate]

Nope. ROMS had a grandparent named SPEM which had a rigid lid which might be closer to what you are asking for.

[aryapaul]

Thanks. Actually we are trying to run a series of controlled experiments. For example, I was wondering if I put a constant f ( Coriolis parameter ) in the grid file, will this imply that there will be no beta-effect ? Or are there more simpler way to introduce f = constant in governing equations ?

[kate]

ROMS will read f from the grid file or compute it in ana_grid - those are your only options.

[aryapaul]

We are trying to completely eliminate wave reflections from the boundaries which are closed.
We have used SPONGE in the *.h file and made changes in ana_hmixcoef.h. We have only modified cff2 to 100 and
Iwrk to be 8. But still we are unable to cancel the reflections from the boundaries.
Are we doing the right thing? Is there a better and sure way to eliminate all reflections from these closed boundaries ? Please suggest.


For your kind information we have defined these functions in *.h.

#define SPONGE
#define UV_LDRAG
#define UV_VIS2
#define UV_COR
#define MASKING

[kate]

The nature of a closed boundary is to be reflective. Why do you insist on closing your boundaries of that's not what you want? I've had the best luck with boundaries when getting daily values from a larger domain, then using RadNud as the BC for all 3D fields.