ROMS reduced open boundary conditions

[lanerolle]

If only water levels/elevation/harmonic consituents are available then, we can perform tides in ROMS for example in the following two ways:

#define XXXX_FSCLAMPED
#define FSOBC_REDUCED
#define XXXX_M2FLATHER

or

#define XXXX_FSCLAMPED
#define FSOBC_REDUCED
#define XXXX_M2REDUCED

where XXXX=EAST, WEST, NORTH or SOUTH.

It appears that FSOBC_REDUCED + XXXX_FLATHER gives a modified version of the normal Flather BCs and XXXX_M2REDUCED is a new kind of BC added to ROMS.

I have tried both and the second set of options appears to work better but we are unsure about how to treat the baroclinic BCs when using these BCs (for the barotropic variables). Currently, we are using XXXX_M3RADIATION for the baroclinic velocities but then there appear to be instabilities in the vicinity of open boundaries for some scenarios. I am not sure about the consistency of the radiation BCs for the baroclinic velocities with the above BCs for the barotropic variables.

Does anybody know how the above barotropic boundary conditions were derived and where they came from? Are there any references/publications showing their derivation, validity and consistency for open ocean boundary conditions in ocean models via say model problems or realistic applications?

[m.hadfield]

I wrote the xxxx_M2REDUCED code (Nonlinear model only ,but Hernan has adapted it for the Adjoint) when I had no luck with the FSOBC_REDUCED + xxxx_M2FLATHER combination.

Setting aside the question of which M3 scheme is best, the M2REDUCED code is not suitable for 3D simulations, as it does not include the vertical integral of the pressure gradient force. There's no fundamental reason why it couldn't include this, but the relevant code hasn't been written yet.

I can't reqally advise you on FSOBC_REDUCED + xxxx_M2FLATHER becasue (as I said) it never worked for me.

Your best bet might be to run a tide-only 2D simulation with xxxx_M2REDUCED, process the velocity data to get barotropic tidal information at the boundaries, then re-run the model in 3D with xxxx_M2FLATHER, using the tidal barotropic data, plus whatever you know about density-driven 3D flows. This is basically what FSOBC_REDUCED + xxxx_M2FLATHER does (I think) but in two stages.

This advice is free, and worth every cent!