Fix for specified ice restart issue #107

- Variable Ice%sCS%OSS%SST_C is not an ice state and hence is not in ice_model.res.nc
  However for SPECIFIED_ICE=True it should be updated after restart, otherwise the model
  would have a restart issue (#107)
  In specified ice this variable is read from data_table every time step. So that is also
  the easiest way to update it after a restart.
- This update resolves issue#107.

- Adding some indication that SIS2 is being restored from restart files.
  With input_filename being 'n' or 'r' having no effect there is no
  indication in stdout whether restart files are being read or not.

- cleanup

src/ice_model.F90

@@ -2500,11 +2500,10 @@ subroutine ice_model_init(Ice, Time_Init, Time, Time_step_fast, Time_step_slow,
       !However it is not an ice state variable and hence is not in ice restarts.
       !Updating it from specified_ice data avoids restart issues for such models.
       if (specified_ice) then
-        allocate(dummy2d(sG%isc:sG%iec,sG%jsc:sG%jec))
+        allocate(dummy2d(isc:iec,jsc:jec))
         allocate(dummy3d(isc:iec,jsc:jec,2))
         call get_sea_surface(Ice%sCS%Time, Ice%sCS%OSS%SST_C(isc:iec,jsc:jec), &
-                             dummy3d(isc:iec,jsc:jec,:), &
-                             dummy2d(sG%isc:sG%iec,sG%jsc:sG%jec), ice_domain=Ice%slow_domain_NH, ts_in_K=.false. )
+                             dummy3d,dummy2d, ice_domain=Ice%slow_domain_NH, ts_in_K=.false. )
         deallocate(dummy2d,dummy3d)
       endif