Hi, I ran the I4D-Var module with my 10km model and now have an optimized initial condition. Now I have a 3km version of the model featuring the exact same domain.
My question is could I run my 3km model with the optimal IC from my 10km model if I interpolate it?
With my DA process I run from the new IC for the duration of the DA window so it would be like using a his output from the coarse resolution model as an initial condition for my 3km model? I run both model with HYCOM boundaries although the boundary adjustment from the DA process would occur over this period so I guess that's where the discrepancy might creep in, however what if I interpolate that part aswell? [u_obc, v_obc etc)
Or do I have to bite the bullet and set up a nested grid within the 10km model? I think this is the traditional approach, i.e. run larger domain coarse model DA -> use the output to run finer smaller domain. I'd quite like to maintain as much of the domain as possible really.