boundary file and tides

[asujrpv]

I am trying to simulate a case where the domain is a rectangle, the boundaries are all open and the flow is driven by tides. I have created two files, grid and forcing by tides.

Is it possible to have a *.h file that reads only these two files without requiring to read a boundary file?

In order to read successfully my simulation, I have to add ANA_M2OBC which is an option to set 2D momentum open boundary conditions using analytical expressions. But I am not using an analytical expression, I am using a tide forcing that is interpolated onto the boundary.

My simulation runs using ANA_M2OBC and ANA_FOBC, but the results show the influence of the walls, which is something should not occur.

Any suggestions will be greatly appreciated.

thanks
Rafael

[kate]

This should be either an FAQ or a bug report. The problem happens in the globaldefs.h logic, where certain assumptions are made for complex domains. These assumptions aren't always true for the idealized domains. Well, I think you can work with it by setting ADD_M2OBC and ANA_M2OBC, with the ubar/vbar set to zero. That way, you're adding zero and not affecting anything.

So, what other boundary options do you have? Is it the Flather/Chapman combo?

[asujrpv]

Kate,
I have done exactly what you said in the past, adding the M2OBC and FSBC with the Flather/Chapman combo?

In the ANA_M2OBC I set the values to zero, as you suggested, but still I have problems with the boundaries...

What can I do?