Navid C. Constantinou

I agree. I would even propose a way that the user doesn't even need to care! Why do we expect the user to call `Q = on_grid(Q, grid)`? For example,...

Who can edit? I guess those with repo write access, right?

is buoyancy in each layer `bₙ = - g ρₙ / ρ̄`, where `ρ̄` is the mean density?

there is another issue similar to bathymetry: what happens when an interior layer outcrops

@glwagner & @dhruvbhagtani why don't we have a zoom and see how we can get things started and then continue the discussion either here or in a PR?

Regarding vanishing layers: what MOM6 does (I think) is that layers never vanish but they are restricted to have height of 1e-10.

After #2522 me and @dhruvbhagtani will start working on the multi-layer. The plan is to extend the `ShallowWaterModel` to allow non-flat `z` dimensions with `Nz` the number of fluid layers....

Yeap, I agree. So are you thinking a config similar to the examples (e.g., rectilinear), start with mean state + some prescribed perturbation and run for up to, e.g., t=20...

ok, now that #2522 is merged we can start thinking of multiple layers :) I put an overleaf doc with eqs at https://www.overleaf.com/read/mtyjxnnrjpqv

Any chance we use vertically stretched grid?