Changeset 626 for trunk/LMDZ.MARS/libf/phymars/growthrate.F

Timestamp:

Apr 18, 2012, 10:17:12 AM (13 years ago)

Author:

tnavarro

Message:

New scheme for the clouds, no more sub-timestep. Clouds sedimentation is done with the dust one in callsedim.F like it was before. Added latent heat for sublimating ground ice. Bugs corrected. THIS VERSION OF THE WATER CYCLE SHOULD NOT BE USED WITH THERMALS DUE TO NEGATIVE TRACERS ISSUES.

File:

• trunk/LMDZ.MARS/libf/phymars/growthrate.F (modified) (6 diffs)

trunk/LMDZ.MARS/libf/phymars/growthrate.F

@@ -1 @@
-      subroutine growthrate(timestep,temp,pmid,ph2o,psat,
-     &                      seq,rcrystal,growth)
+      subroutine growthrate(temp,pmid,psat,seq,rcrystal,coeff1,coeff2)
 
       IMPLICIT NONE
@@ -10 +9 @@
 c     Authors: F. Montmessin
 c       Adapted for the LMD/GCM by J.-B. Madeleine (October 2011)
+c       Use of resistances in the analytical function 
+c            instead of growth rate - T. Navarro (2012)
 c     
 c=======================================================================
@@ -27 +28 @@
 c   ----------
 
-      REAL timestep
+c     Input
       REAL temp     ! temperature in the middle of the layer (K)
       REAL pmid     ! pressure in the middle of the layer (K)
-      REAL*8 ph2o   ! water vapor partial pressure (Pa)
       REAL*8 psat   ! water vapor saturation pressure (Pa) 
+      REAL*8 seq    ! Equilibrium saturation ratio
       REAL rcrystal ! crystal radius before condensation (m)
-      REAL*8 seq    ! Equilibrium saturation ratio
-      REAL*8 growth 
+
+c     Output
+      REAL coeff1,coeff2     ! coefficients for the analytical relation between time and radius
+
 
 c   local:
@@ -43 +46 @@
       REAL*8 afactor,Dv,lambda       ! Intermediate computations for growth rate
       REAL*8 Rk,Rd
+      
+      
 
 c-----------------------------------------------------------------------
@@ -73 +78 @@
      &   ( pi * pmid * (molco2+molh2o)*(molco2+molh2o) 
      &        * sqrt(1.+mh2o/mco2) )
-
+      
       knudsen = temp / pmid * afactor / rcrystal
       lambda  = (1.333+0.71/knudsen) / (1.+1./knudsen)
-      Dv      = Dv / (1. + lambda * knudsen)
+      
+c      Dv is not corrected. Instead, we use below coefficients coeff1, coeff2
+c      Dv      = Dv / (1. + lambda * knudsen)
 
 c     - Compute Rk
@@ -82 +89 @@
 c     - Compute Rd
       Rd = rgp * temp *rho_ice / (Dv*psat*mh2o)
-
+      
+      
+      coeff1 = real(Rk + Rd*(1. + lambda * knudsen))
+      coeff2 = real(Rk + Rd*(1. - lambda * knudsen))
+      
+c Below are growth rate used for other schemes
 c     - Compute growth=rdr/dt, then r(t+1)= sqrt(r(t)**2.+2.*growth*dt)
-      growth = 1. / (seq*Rk+Rd)
-c       growth = (ph2o/psat-seq) / (seq*Rk+Rd)
+c      growth = 1. / (seq*Rk+Rd)
+c      growth = (ph2o/psat-seq) / (seq*Rk+Rd)
 c      rf   = sqrt( max( r**2.+2.*growth*timestep , 0. ) )
 c      dr   = rf-r