Bug in case of iceberg module activated and embedded sea ice

[mathiot]

Context

In case of embedded sea and iceberg module, in sea ice area, the sea surface slope used to compute the icebergs dynamics should be the gradient of the 'equivalent' sea surface in 'leads' as it is done for the sea-ice dynamics, instead of the slope of the sea surface at the sea ice base.

Analysis

...

Fix

As the 'equivalent' ssh is already computed in SI3 (zpice), I suggest:

• define and allocate pice in ice.F90
• change zpice by pice in icbdyn_rhg_evp.F90
• use pice in icbutl.F90 to define ssh_e if key_si3 is defined (precision icb_stp is call after ice_stp).

By doing this my ~20 m ice-pillar in 7 months of run disappear (neighbour of a calving point)

[clem]

I think you should keep local calculation of zpice (which would better be called zsshice or something like that) because it changes with ice and snow mass, so all the time.
Or you write a small function with ssh and snow+ice mass as inputs and zpice as output and you call it in icedyn_rhg_evp and iceberg

[mathiot]

After discussion with Clement, we agreed on:

• Put the computation of the equivalent ssh in leads in a function call ice_var_sshdyn in ice_var.F90
• zpice is renamed zssh_lead_m
• ice_var_sshdyn is called in icbdyn_rhg_evp.F90 and icbutl.F90.

[mathiot]

In 10332:

Error: Failed to load processor CommitTicketReference

No macro or processor named 'CommitTicketReference' found