offline-nested BCs vanish!

[johnluick]

Can anyone suggest why the sea level and velocity boundary conditions, which are read in by my nested model (nested in the "offline" sense), appear to be ignored when I add tides to it?

Here are the relevant switches used prior to adding tides:
#undef SSH_TIDES
#undef UV_TIDES
#define WEST_FSCHAPMAN
#define WEST_M2FLATHER
and here are they are when adding tides:
#define SSH_TIDES
#define UV_TIDES
#define WEST_FSCHAPMAN
#define WEST_M2FLATHER

The boundary conditions are read in from a boundary file whose values are taken from a larger grid. For what its worth, the larger grid has no tides, and it's a 2D grid driven only by a coastal trapped wave paddle on the W boundary, and winds.

Just to make it clear, the nested model works fine unless I add tides to it. When tide switches are turned on, the CTW variability that previously came in at the western boundary of the nested model no longer appears.

The tidal forcing is specified by tidal ellipses etc. the way ROMS wants it - ie, it is not imposed at the open boundaries.

Thanks

John

[jivica]

if you are adding tides to the open boundaries then you have to do something like:

/* if you are adding tides on turn on SSH_TIDES and UV_TIDES */
# define SSH_TIDES
# define UV_TIDES

# ifdef SSH_TIDES
# define ADD_FSOBC
# define WEST_FSCHAPMAN
# else
# define WEST_FSCHAPMAN
# endif

# if defined UV_TIDES
# define ADD_M2OBC
# define WEST_M2FLATHER
# else
# define WEST_M2FLATHER
# endif

/* the rest is up to type of boundaries that you use */
# define WEST_M3NUDGING
# define WEST_M3RADIATION
# define WEST_TRADIATION
# define WEST_TNUDGING

[johnluick]

jivica, thanks very much, that fixed the problem...which is strange because I am not, in fact, adding tides at the boundary. If you read my email carefully, you will see that things were fine, it was just that when I turned the tides on, the tides were fine, but the subinertial waves disappeared. I will read some more and try to figure out why your suggestion worked. But clearly, I probably should have "defined" add_fsobc and add_m2obc regardless.

For those that are interested in open boundary conditions, (i.e. an identical set to the set of people who are currently having problems with them) I might add that my method as described in the first post is an effective way of getting the FSCHAPMAN/M2FlATHER OBCs to radiate worth a damn. (The only really effective way, in my experience.) Before running your 3D coastal model, create a slightly larger grid which is identical except it has added five or ten gridpoints around the boundaries. Run the bigger grid in 2D (#undef SOLVE3D), forced by winds (and in my case, coastal trapped wave paddle based on an analytical model plus lowpass filtered coastal tide gauge data). The sea levels and velocities from the 2D gridpoints which are at the same locations as the 3D boundary points are then stuck in a file and used as the "bry" (input boundary conditions) for the 3D model. That's what some people call offline nesting. For tides, just turn on the tides for the 3D model.

Thanks again

John