search

ikeshwani / HorizontalConvection

2 stars 1 forks source link

Add `horizontal_diffusion.jl` problem to verify diagnosis of buoyancy dissipation against the analytical solution #2

Closed hdrake closed 1 year ago

hdrake commented 1 year ago

Renamed the variables so that it is clear this was not to be interpreted as a Nusselt number, but instead as an "Equilibration ratio"

hdrake commented 1 year ago

Commit https://github.com/ikeshwani/HorizontalConvection/pull/2/commits/04650d02e4df2be358245fe6d74a419c38267559 also adds the basic Project.toml environment for reproducibility and reverts the problem parameters and boundary conditions to be exactly that of the Oceananigans.jl example. I also increased the timestep, the advective CFL condition (which normally constrains the timestep but does not matter since advection is turned off), and the TimeInterval for saving outputs.

hdrake commented 1 year ago

@ikeshwani, if you run the diffusive_convection.jl script here, you should find that the Equilibration ratio equilibrates around a value of 6, which means something is wrong with either our simulation analysis or the analytical solution. I think the problem has to do with the high buoyancy dissipation along the insulating boundaries in the simulation output (see the buoyancy dissipation movie), which I think is erroneous and dominates the volume-integral dissipation in the numerator of the equilibration ratio.