LES simulation

[owen14121022]

In the Langmuir Simulation example, are the resolved velocity fields (u,v,w) the Lagrangian-mean velocity fields?

[ali-ramadhan]

Yes I believe so but I'll let @glwagner confirm since he wrote that example and the SurfaceWaves module (now called StokesDrift as of yesterday).

He also wrote this in the example:

In other words, model.velocities.u is the Lagrangian-mean x-velocity beneath surface waves.

[navidcy]

@owen14121022 I'm also pretty sure the answer to your questions is yes. See also here: https://clima.github.io/OceananigansDocumentation/stable/physics/surface_gravity_waves/

[glwagner]

Yes, model.velocities.u, etc, are the Lagrangian-mean velocities when surface_waves / stokes_drift is used.

[owen14121022]

Yes, model.velocities.u, etc, are the Lagrangian-mean velocities when surface_waves / stokes_drift is used.

Thank you for your reply！Have you compared the model result (for example，horizontal and time-mean horizontal velocity profiles) with that in the McWilliams 1997.

[glwagner]

I did that in a paper (not published, but in press for JPO) for a high-resolution version of the example.

This is figure 2 from McWilliams et al. 1997:

and figure 6:

My preprint contains some discussion (as well as demonstrating via its analysis) why the Lagrangian-mean velocity field is a more appropriate and physically-intuitive prognostic variable than the Eulerian-mean velocity field.

In appendix A, I also argue that using the Lagrangian-mean rate of strain in turbulence closures for LES is more appropriate for modeling downscale transfers of kinetic energy than a closure formulated in terms of the Eulerian-mean rate of strain --- because closures formulated in terms of the Eulerian-mean rate of strain do not always dissipate mean kinetic energy.