Replaced group_pass of h with pass_var of h

  Reverted to the simpler pass_var routines for the single 3-d variable h.  This
avoids the issue that triggered the recent pull request #520, and it should have
no performance impacts.  (In fact, by letting each call have its own halo size,
it might lead to performance improvements in subsequent commits.)  All answers
are bitwise identical.

src/core/MOM.F90

@@ -413,7 +413,6 @@ module MOM
 
   ! These are used for group halo updates.
   type(group_pass_type) :: pass_tau_ustar_psurf
-  type(group_pass_type) :: pass_h
   type(group_pass_type) :: pass_ray
   type(group_pass_type) :: pass_bbl_thick_kv_bbl
   type(group_pass_type) :: pass_T_S_h
@@ -568,8 +567,6 @@ subroutine step_MOM(fluxes, state, Time_start, time_interval, CS)
     call create_group_pass(CS%pass_tau_ustar_psurf, fluxes%ustar(:,:), G%Domain)
   if (ASSOCIATED(fluxes%p_surf)) &
     call create_group_pass(CS%pass_tau_ustar_psurf, fluxes%p_surf(:,:), G%Domain)
-  if ((CS%thickness_diffuse .and. (.not.CS%thickness_diffuse_first .or. CS%t_dyn_rel_adv == 0)) .OR. CS%mixedlayer_restrat) &
-    call create_group_pass(CS%pass_h, h, G%Domain) !###, halo=max(2,cont_stensil))
 
   do_pass_Ray = .FALSE.
   if ((.not.G%Domain%symmetric) .and. &
@@ -758,7 +755,7 @@ subroutine step_MOM(fluxes, state, Time_start, time_interval, CS)
                                CS%MEKE, CS%VarMix, CS%CDp, CS%thickness_diffuse_CSp)
         call cpu_clock_end(id_clock_thick_diff)
         call cpu_clock_begin(id_clock_pass)
-        call do_group_pass(CS%pass_h, G%Domain)
+        call pass_var(h, G%Domain) !###, halo=max(2,cont_stensil))
         call cpu_clock_end(id_clock_pass)
         call disable_averaging(CS%diag)
         if (showCallTree) call callTree_waypoint("finished thickness_diffuse_first (step_MOM)")
@@ -898,7 +895,7 @@ subroutine step_MOM(fluxes, state, Time_start, time_interval, CS)
       if (CS%debug) call hchksum(h,"Post-thickness_diffuse h", G%HI, haloshift=1, scale=GV%H_to_m)
       call cpu_clock_end(id_clock_thick_diff)
       call cpu_clock_begin(id_clock_pass)
-      call do_group_pass(CS%pass_h, G%Domain)
+      call pass_var(h, G%Domain) !###, halo=max(2,cont_stensil))
       call cpu_clock_end(id_clock_pass)
       if (showCallTree) call callTree_waypoint("finished thickness_diffuse (step_MOM)")
     endif
@@ -915,7 +912,7 @@ subroutine step_MOM(fluxes, state, Time_start, time_interval, CS)
                               G, GV, CS%mixedlayer_restrat_CSp)
       call cpu_clock_end(id_clock_ml_restrat)
       call cpu_clock_begin(id_clock_pass)
-      call do_group_pass(CS%pass_h, G%Domain)
+      call pass_var(h, G%Domain) !###, halo=max(2,cont_stensil))
       call cpu_clock_end(id_clock_pass)
       if (CS%debug) then
         call hchksum(h,"Post-mixedlayer_restrat h", G%HI, haloshift=1, scale=GV%H_to_m)
@@ -1396,7 +1393,7 @@ subroutine step_tracers(fluxes, state, Time_start, time_interval, CS)
     fluxes%netMassOut(:,:) = CS%offline_CSp%netMassOut(:,:)
     call offline_diabatic_ale(fluxes, Time_start, Time_end, time_interval, CS%offline_CSp, &
         CS%h, eatr, ebtr)
-    call pass_var(CS%h,G%Domain)
+    call pass_var(CS%h, G%Domain)
 
     ! Do the transport, the final ALE remappings,  horizontal diffusion if it is
     ! the last iteration
@@ -1415,9 +1412,7 @@ subroutine step_tracers(fluxes, state, Time_start, time_interval, CS)
       call cpu_clock_begin(id_clock_ALE)
       call ALE_offline_tracer_final( G, GV, CS%h, h_end, CS%tracer_Reg, CS%ALE_CSp)
       call cpu_clock_end(id_clock_ALE)
-      call pass_var(CS%h,G%Domain)
-
-
+      call pass_var(CS%h, G%Domain)
     endif
 
   else ! NON-ALE MODE...NOT WELL TESTED
@@ -1445,9 +1440,9 @@ subroutine step_tracers(fluxes, state, Time_start, time_interval, CS)
     CS%tv%S = salt_mean
     CS%h = h_end
 
-    call pass_var(CS%tv%T,G%Domain)
-    call pass_var(CS%tv%S,G%Domain)
-    call pass_var(CS%h,G%Domain)
+    call pass_var(CS%tv%T, G%Domain)
+    call pass_var(CS%tv%S, G%Domain)
+    call pass_var(CS%h, G%Domain)
 
   endif