Compute initial streamfunction from initial velocity about veros HOT 5 CLOSED

I agree that this is counter-intuitive. It may take me some time to investigate what's going on. We have an experimental surface pressure solver where this doesn't happen, so it definitely seems connected to the stream function.

from veros.

Cool, yes, I'm used to a surface pressure solver in the MITgcm. Though that can be expensive and hard to parallelize.

I'm not sure what I would do if I wanted to initialize a shear flow with non-zero barotropic flow. Remove the barotropic and put that in as a stream function and just initialize u and v with the shear?

Here, where I wanted a barotropic initial flow, the solution to a mean flow definitely seemed to be not to specify vs.u and only specify vs.psi.

from veros.

I'm not sure what I would do if I wanted to initialize a shear flow with non-zero barotropic flow. Remove the barotropic and put that in as a stream function and just initialize u and v with the shear?

Yes that seems to be it. The streamfunction solver removes any barotropic mode from u and v, and adds it back based on psi.

That's a bit awkward of course. I think I will pull in the first version of the surface pressure solver with linear free surface soon (identical to the MITgcm implementation). That should make things easier for you.

from veros.

I think, even working within the confines of the stream function solver we might consider calculating the initial steam function from the initial velocity field as an initial step. Most of your demos start from rest, but I would expect people to be more likely to know u, v rather than psi as an initial condition?

from veros.

Yes, that's a good suggestion.

from veros.