Changeset 2765 for palm/trunk/SOURCE/land_surface_model_mod.f90

Timestamp:

Jan 22, 2018 11:34:58 AM (6 years ago)

Author:

maronga

Message:

major bugfix in calculation of aerodynamic resistance for vertical surface elements

File:

• palm/trunk/SOURCE/land_surface_model_mod.f90 (modified) (3 diffs)

palm/trunk/SOURCE/land_surface_model_mod.f90

@@ -25 +25 @@
 ! -----------------
 ! $Id$
+! Major bugfix in calculation of f_shf for vertical surfaces
+! 
+! 2735 2018-01-11 12:01:27Z suehring
 ! output of r_a moved from land-surface to consider also urban-type surfaces
 ! 
@@ -1770 +1773 @@
 !        ENDIF
 !
-!--    Calculate aerodynamical resistance. For horizontal and vertical 
-!--    surfaces MOST is applied. Moreover, for vertical surfaces, resistance
-!--    can be obtain via parameterization of Mason (2000) / 
-!--    Krayenhoff and Voogt (2006).
+!--     Calculation of r_a for vertical surfaces
+!--
+!--     heat transfer coefficient for forced convection along vertical walls
+!--     follows formulation in TUF3d model (Krayenhoff & Voogt, 2006)
+!--            
+!--       H = httc (Tsfc - Tair)
+!--       httc = rw * (11.8 + 4.2 * Ueff) - 4.0
+!--            
+!--             rw: wall patch roughness relative to 1.0 for concrete
+!--             Ueff: effective wind speed
+!--             - 4.0 is a reduction of Rowley et al (1930) formulation based on
+!--             Cole and Sturrock (1977)
+!--           
+!--             Ucan: Canyon wind speed
+!--             wstar: convective velocity
+!--             Qs: surface heat flux
+!--             zH: height of the convective layer
+!--             wstar = (g/Tcan*Qs*zH)**(1./3.)
+                
+!--    Effective velocity components must always 
+!--    be defined at scalar grid point. The wall normal component is 
+!--    obtained by simple linear interpolation. ( An alternative would
+!--    be an logarithmic interpolation. )
+!--    A roughness lenght of 0.001 is assumed for concrete (the inverse,
+!--    1000 is used in the nominator for scaling)
 !--    To do: detailed investigation which approach gives more reliable results!
 !--    Please note, in case of very small friction velocity, e.g. in little 
@@ -1782 +1806 @@
                              ( surf%ts(m) * surf%us(m) + 1.0E-20_wp ) )
        ELSE
-          surf%r_a(m) = 1.0_wp / ( 11.8_wp + 4.2_wp *                          &
+          surf%r_a(m) = rho_cp / ( surf%z0(m) * 1000.0_wp                      &
+                        * ( 11.8_wp + 4.2_wp *                                 &
                         SQRT( MAX( ( ( u(k,j,i) + u(k,j,i+1) ) * 0.5_wp )**2 + &
                                    ( ( v(k,j,i) + v(k,j+1,i) ) * 0.5_wp )**2 + &
                                    ( ( w(k,j,i) + w(k-1,j,i) ) * 0.5_wp )**2,  &
                               0.01_wp ) )                                      &
-                                 )
+                           )  - 4.0_wp  ) 
        ENDIF
 !