Penalize isopycnal depths and T and S on isopycnals?

[ifenty]

Suggestion from J. Toole: perhaps we should write cost functions to constrain 1) isopycnal depth and 2) the salinity - or temperature - on isopycnals?

"The variance of Theta/S - below the surface waters - is far less than T or S on pressure surfaces, which should translate into an up-weighting in the cost function."

It true, much of our large in-situ variance is associated with isopycnal displacements in the presence of vertical T and S gradients.

[gaelforget]

It true, much of our large in-situ variance is associated with isopycnal displacements in the presence of vertical T and S gradients.

is true in general I think

Suggestion from J. Toole: ...

Thanks John!! Much appreciated

We might have talked about this at a climode meeting or something like that in fact. K. Haines & a few others have brought it up a few times too I think.

A well integrated version of this in https://github.com/MITgcm/MITgcm could be a nice contribution, no?