Changeset 2758 for palm/trunk

Timestamp:

Jan 17, 2018 12:55:21 PM (7 years ago)

Author:

suehring

Message:

Bugfix, no initial masking of wind velocity at first prognostic grid level in case of land- or urban-surface spin-up.

Location:

palm/trunk/SOURCE

Files:

• init_3d_model.f90 (modified) (3 diffs)
• time_integration_spinup.f90 (modified) (2 diffs)

palm/trunk/SOURCE/init_3d_model.f90

@@ -25 +25 @@
 ! -----------------
 ! $Id$
+! In case of spinup of land- and urban-surface model, do not mask wind velocity
+! at first computational grid level
+! 
+! 2746 2018-01-15 12:06:04Z suehring
 ! Move flag plant canopy to modules
 ! 
@@ -1253 +1257 @@
 !
 !--       Set velocities back to zero
-          DO  i = nxlg, nxrg
-             DO  j = nysg, nyng
-                DO  k = nzb, nzt
-                   u(k,j,i) = MERGE( u(k,j,i), 0.0_wp,                      &
-                                     BTEST( wall_flags_0(k,j,i), 1 ) )
-                   v(k,j,i) = MERGE( v(k,j,i), 0.0_wp,                      &
-                                     BTEST( wall_flags_0(k,j,i), 2 ) )
-                ENDDO
-             ENDDO
-          ENDDO
-          
+          u = MERGE( u, 0.0_wp, BTEST( wall_flags_0, 1 ) )
+          v = MERGE( v, 0.0_wp, BTEST( wall_flags_0, 2 ) )         
 !
 !--       WARNING: The extra boundary conditions set after running the
@@ -1309 +1304 @@
              ENDDO
           ENDDO
-
+!
+!--       Mask topography
+          u = MERGE( u, 0.0_wp, BTEST( wall_flags_0, 1 ) )
+          v = MERGE( v, 0.0_wp, BTEST( wall_flags_0, 2 ) )
 !
 !--       Set initial horizontal velocities at the lowest computational grid 
 !--       levels to zero in order to avoid too small time steps caused by the 
 !--       diffusion limit in the initial phase of a run (at k=1, dz/2 occurs
-!--       in the limiting formula!).
-          IF ( ibc_uv_b /= 1 )  THEN
+!--       in the limiting formula!). 
+!--       Please note, in case land- or urban-surface model is used and a 
+!--       spinup is applied, masking the lowest computational level is not 
+!--       possible as MOST as well as energy-balance parametrizations will not 
+!--       work with zero wind velocity. 
+          IF ( ibc_uv_b /= 1  .AND.  .NOT.  spinup )  THEN
              DO  i = nxlg, nxrg
                 DO  j = nysg, nyng

palm/trunk/SOURCE/time_integration_spinup.f90

@@ -25 +25 @@
 ! -----------------
 ! $Id$
+! Comment out homogeneous setting of wind velocity as this will lead to zero
+! friction velocity and cause problems in MOST relationships.
+! 
+! 2728 2018-01-09 07:03:53Z maronga
 ! Set velocity componenets to homogeneous values during spinup
 ! 
@@ -170 +174 @@
 !-- otherwise. The precise reason is still unknown. A minimum velocity of 0.1
 !-- m/s is used to maintain turbulent transfer at the surface.
-    u = SIGN(1.0_wp,u) * MAX(ug_surface,0.1_wp)
-    v = SIGN(1.0_wp,v) * MAX(vg_surface,0.1_wp)
+!     u = SIGN(1.0_wp,u) * MAX(ug_surface,0.1_wp)
+!     v = SIGN(1.0_wp,v) * MAX(vg_surface,0.1_wp)
 
     dt_save = dt_3d