SI3 salt drainage and flushing

The task consists in implementing gravity drainage and flushing in SI3.

Desalination in sea-ice is very crudely represented in the model by simple relaxation schemes. The vertical profile is supposed linear and reconstructed from the bulk salinity (see Vancoppenolle et al. 2007). Here we propose to implement more realistic parameterizations. For gravity drainage, we will base on Griewank and Notz 2013 and/or Rees Jones and Worster 2014 (see Thomas et al. 2020 for the framework). For flushing, we will base on Vancoppenolle et al. 2007.

This is a change of paradigm since salt was treated as a bulk salinity while the new parameterizations impose to treat salt as temperature (hence each layer must be tracked). Here will be the equivalent arrays (most probably):

• t_i <=> sz_i (4D)
• tm_i <=> sm_i (2D)
• et_i <=> st_i (2D)
• e_i <=> szv_i (4D, new variable = sz_i * v_i / nlay_i)
• • s_i and sv_i (3D)

Gravity drainage will be activated when the atm. is colder than the ocean. In other words, when salinity of the brines is larger than the salinity of the ocean (maxval(Sbr) > Sw). Flushing will be activated when ice/snow surface is melting

Implementation steps:

1. Code gravity drainage and flushing in icethd_sal.F90 (namelist nn_icesal=4 ?)
2. Set new ice salinity (in both bottom accretion and open water) as 75% of ocean salinity (75% must be a namelist parameter). Therefore remove Cox and Weeks 1988 param for nn_icesal=4
3. Thermo: melting/freezing, change s_i into sz_i. Maybe s_i needs to be recalculated afterwards
4. Thermo: snow-ice formation, ocean salinity comes at the top the ice. Hence treat salt as temperature with zs_old = sss*zfmdt... and remove max on sea-ice salt if it exists
5. Thermo: at the end of icethd_dh, do a remapping as for temperature
6. Thermo: open water formation, change sv_i into szv_i and do a remapping
7. Dyn: advection of szv_i instead of sv_i (hence no more bounds)
8. Dyn: ridging/rafting and itd, treat szv_i as e_i with entrapement of salt (sss*vsw)
9. Others: set min(sz_i)=0.1, max(sz_i)=sss or 75%*sss
10. Others: in case we need to reconstruct a salinity profile from a bulk salinity (for bdy, init, restarts...), use the linear profile from SI3 (icevar_salprof)

Let's go

Thomas, M., Vancoppenolle, M., France, J. L., Sturges, W. T., Bakker, D. C. E., Kaiser, J., & von Glasow, R. (2020). Tracer measurements in growing sea ice support convective gravity drainage parameterizations. Journal of Geophysical Research: Oceans, 125, e2019JC015791. https://doi.org/10. 1029/2019JC015791