High resolution no-motion profile needed for N2 and eta_true calculation

[JeffreyEarly]

Perhaps we should build back in the high resolution version of the hydrostatic vertical modes, so that we can quickly construct a high resolution version of rho_nm at the current time from the mda. We can simultaneously construct a version of N2.

With these, in an interpolation algorithm, I think we can compute a closed version of the potential enstrophy nonlinear advection flux, which doesn't appear to close if we use a mismatched form of N2 and eta.

[JeffreyEarly]

As far as naming goes, here's an idea:

rho_bar is the current mean density
rho_nm is the current no-motion density
rho_nm_t0 is the initial no-motion density or maybe rho_nm0?