alternating flow in constant depth channel.

[jklinck]

I am trying to set up roms to represent alternating flow in a uniform depth channel (x-z). It is 40 km long and periodic in y. density is uniform. vertical viscosity is constant. The boundaries are open on east and west ends. I want the flow to have ubar = uo sin(2 pi t/period) for different choices of uo.

I have tried using ana_m2obc to set boundary values of ubar to the desired flow. I have also set ana_fsobc to have the desired changing elevation at either end. I finally tried to set ana_m3obc to the steady vertical frictional structure. OBC choices are

LBC(isFsur) == Cha Per Cha Per ! free-surface
LBC(isUbar) == Fla Per Fla Per ! 2D U-momentum
LBC(isVbar) == Fla Per Fla Per ! 2D V-momentum
LBC(isUvel) == Rad Per Rad Per ! 3D U-momentum
LBC(isVvel) == Rad Per Rad Per ! 3D V-momentum
LBC(isMtke) == Gra Per Gra Per ! mixing TKE

Because of the friction, the analytical solution is complicated (although possible). Do I really need to give zeta, ubar and u(z) at either boundary from an analytical solution?

Is there a way to simply set ubar at the ends and have the model calculate zeta and u(z)? I thought about using ana_m2clima and nudging, but it is unclear how to have a time dependent climatology using ana_m2clima.

Has anyone done such a model? I thought that this would be easy to set up. I am clearly not understanding something about ROMS. Hints and suggestions are appreciated.

[kate]

It's tricky, for sure. Since you have it periodic in y, you can't support a north-south surface slope in balance with the ubar. This used to be easy with a walled channel and the streamfunction solver.

If you want to apply the ana_m3obc, you will have to pick a different LBC for M3, say RadNud.

Why wouldn't the ana_m2clima be the same as your ana_m2obc? Are you trying to propagate something from one end to the other or is it all simultaneous?