ocean depth for ocean cavity model

[matthiasmengel]

Ice shelves have diverse thickness, and the bottom topography below them also varies.
At which depth do we then take the ocean temperatures and salinities as input for the cavity model?
Do these depths spatially vary and can we diagnose these?
Does it then mean, that we need to read in 3d ocean fields (if we want to perfect this)?
What would POEM need?

[ronjareese]

For POEM, 2d ocean input fields to PISM are fine. They should represent the temperature and salinity on top of the continental shelf, and, as a first guess, extended at that depth into the open ocean.

[ronjareese]

In order to allow the grounding line to advance up to the continental shelf edge, I think that we should compute the ocean means over the 2d field described above within a 100km (?) belt in front of the ice shelf. This should replace the current way to compute the means over all ocean areas which are above a certain threshold (-continental_shelf_depth, default -800). Because in the latter case, if the grounding line advances to the continental shelf break, we might have no ocean cells to compute the means over. Additionally, the continental shelf is in places deeped than -800m or flatter.
What do you think?