Dear ROMS users;
I have been implementing a ROMS/SWAN coupled simulation for the South Shore of Oahu - Hawaii, based on a previous ROMS simulation that has been running without any major issues for several years. I am using a regional implementation of the WW3 to provide the spectral boundary conditions to my regional domain.
While the ROMS model is very maleable in what concerns the computational grid, SWAN doesn't seems to follow the same standards.
The SWAN manual says: "Note that curvilinear grids can be used for nested computations but the boundaries should always be rectangular."
What I understand fomr this phrase is that I can't use my ROMS curvilinear grid in SWAN - (note: I did that and model runs without problems, but the results are not what I expected - seems to have problems with the wave directions). Is that right?
If so, I know that in previous versions of the coupled system both wave and circulation models had to be in the same grid. Is it still truth or can I just setup a retangular SWAN grid, that would implicate in points where I have wave info but no currents and vice-versa?
Thank you
Joao