To create ROMS boundary conditions, the matlab package gives an example from Mercator data set named 'd_obc_mercator.m'.
Looking through the script, I noticed that it uses the same Land/Sea mask in interpolating temp, u and v:
358 % Interpolate temperature and salinity.
359
360 temp=obc_mercator('temp',S,Temp,Tlon,Tlat,Rmask3d,Tdepth);
361 salt=obc_mercator('salt',S,Salt,Tlon,Tlat,Rmask3d,Tdepth);
362 Urho=obc_mercator('u' ,S,Uvel,Ulon,Ulat,Rmask3d,Udepth);
363 Vrho=obc_mercator('v' ,S,Vvel,Vlon,Vlat,Rmask3d,Vdepth);
since Rmask3d is defined as:
329 Rmask3d=ones(size(Temp));
330 ind=find(Temp == 0);
331 Rmask3d(ind)=0;
Will this be a problem on considering NEMO also uses Arakawa C-grid, in which case u and v have different Land/Sea masks from temp does?
Thanks for your comments!