Add tracer budget closure test

[navidcy]

Add tracer conservation tests for hydrostatic and nonhydrostatic model.

The test checks whether the total tracer changes in agreement with how much the surface flux of the tracer is.

w @taimoorsohail

[simone-silvestri]

We already have some conservation tests for the hydrostatic model in
https://github.com/CliMA/Oceananigans.jl/blob/main/test/test_zstar_conservation_explicit.jl
https://github.com/CliMA/Oceananigans.jl/blob/main/test/test_zstar_conservation_implicit.jl
https://github.com/CliMA/Oceananigans.jl/blob/main/test/test_zstar_conservation_tripolar.jl
maybe we could add to those (small caveat for tracer conservation in a hydrostatic model zstar is required, on the non-hydrostatic it is not a problem since we use a rigid lid approximation)

[navidcy]

Gotcha. I’ll move them there/combine. Will also perturb the free surface so that zstar is required for.

[simone-silvestri]

apparently tracers are not conserved?

[navidcy]

@simone-silvestri, I perturbed the free surface but locally still the test passes, although we are not using $z$-star. how come?

[navidcy]

I guess z-star is required if we were testing point-wise conservation, not conservation in an integral sense?

[glwagner]

I think its the opposite

[navidcy]

Perhaps I don’t know what “conservation” means? Here I am testing the tracer budget and that the total tracer content changes according to the surface forcing. Seems that it doesn’t require z-star unless my test case is too simple?