Henri Drake
> The paper that I cited used periodic boundary conditions for the deviations in the horizontal. Not physically meaningful but it is one way to avoid the walls. Agreed. But...
> > Because no boundary conditions are applied to the background field, there is an implied diffusive flux n^⋅(−κ∇b¯)=±κN2cosθ across these boundaries. > > How can this be? There is...
> As for clarity, we can add a line to the page on background fields: > > https://clima.github.io/OceananigansDocumentation/stable/model_setup/background_fields/ > > that says something like "note that boundary conditions apply to...
I propose two changes: 1) An example that re-implements the existing 1D diffusion example but with a background field (modifying the boundary conditions on the perturbations to make the two...
> > There actually should be a term in the equations, but with a constant diffusivity and background stratification this term drops out in the interior. This is how all...
Good point. A simple illustration in the docs for background field will suffice.
I guess a more general point is that we could replace some of the existing ad-hoc parsing functions with the f90nml package. I was planning on just using `parse_namelist(file, silence_errors=True)`...
> Which parsing functions are you referring to? We don't currently have any namelist parsing in xmitgcm. We only parse `.meta` files, which, afaik, are only used in MITgcm and...
> The overall goal was to implement a "open_folder" function which would parse `data.diagnostics` and open everything in a dataset or dict of datasets. I have a version of this...
Sure, I'd be happy to take charge of this and make a PR with: - a refactored parsing function based on `f90nml` - some tests using both 1) generated "toy"...