Hello,
As the title of my post says, I'm working with a rotated domain but I've obtained a few weird signals and I want to discard some rotation-related problem in my simulation. I've reviewed the posted answers before the preparation of this post (so that Hernan doesn't tell me off) but I'd like to have a more recent answer (in case something changed recently in the model), and/or just moral support. Below I present what I've done and please take a look and tell me if it's OK or I'm doing something wrong.
1. I'm using a rotated grid (and defined the angle and proper metrics), so I activate the CURVGRID flag.
2. In preparation for the forcing and boundary files, I use the grid parameter "angle" rotate ALL the vector variables: sustr, svstr, u, v, ubar, vbar, Uwind, Vwind.
3. I use "angle" in the rotation but in the opposite direction; for example:
ubar (xi,eta) = real ( ( ubar + i*vbar ).*exp( - i*angle) ) ,
vbar (xi,eta) = imag ( ( ubar + i*vbar ).*exp( - i*angle) ) ,
where "ubar(xi,eta)" and "vbar(xi,eta)" are the variables used in the input files; NOTICE the minus sign (-) in the "exp" function. Then, if "angle"=30 degrees, positive, defined counterclockwise, the rotation has to be clockwise, an angle "angle"=-30 degrees. Needless to say, "angle" is defined in radians and not degrees.
4. To take the vector variables from the output files, now I have to rotate them in the direction opposite to that used in the input files, that is, now the rotation is using "angle" with no minus sign.
Thank you very much for your attention.
Best regards,
DAVID