Update to latest Consortium/main

[DeniseWorthen]

Updates to include the following PRs which are not expected to change baselines:

CICE-Consortium#778. This PR simplifies the initialization of the dynamics and removes unnecessary code or allocations.

CICE-Consortium#790. This PR renames cicedynB to cicedyn (with link) and fixes some metadata in netcdf files for C-grid.

The following PRs change answers as noted:

CICE-Consortium#802. This PR changes calculation of stress12T to give non-zero values for the C-grid. It will change answers only for the C-grid test.

CICE-Consortium#799. This PR contains a bug fix for the accumulation of salt and freshwater fluxes when ktherm=2 (ie, mushy thermo) as well as a new feature to use the actual bulk-salinity of sea ice to calculate the fluxes. Without the saltflux feature turned on, sea ice is assumed to have a constant bulk salinity of 4psu. The new feature allows the actual sea ice salinity prognostically determined by the mushy therm scheme to determine the salt flux. The bug fix changes all UFS baselines which use ktherm=2.

The relevant section of code in icepack/columnphysics/icepack_therm_itd.F90 which contained the bug is:

      if (update_ocn_f) then
         if (ktherm <= 1) then
            dfresh = -rhoi*vi0new/dt
            dfsalt = ice_ref_salinity*p001*dfresh
            fresh  = fresh + dfresh
            fsalt  = fsalt + dfsalt
         ! elseif (ktherm == 2) the fluxes are added elsewhere
         endif

See Icepack issue CICE-Consortium/Icepack#390

Baseline changes for the NG-GODAS tests are due to code refactoring associated with the saltflux feature, since the debug test case reproduces.