Changeset 1621 for trunk/LMDZ.MARS/libf/dynphy_lonlat

Timestamp:

Oct 25, 2016, 9:23:21 AM (9 years ago)

Author:

emillour

Message:

Further work on full dynamics/physics separation.

LMDZ.COMMON:

• added phy_common/vertical_layers_mod.F90 to store information on vertical grid. This is where routines in the physics should get the information.
• The contents of vertical_layers_mod intialized via dynphy_lonlat/inigeomphy_mod.F90.

LMDZ.MARS:

• physics now completely decoupled from dynamics; the physics package may now be compiled as a library (-libphy option of makelmdz_fcm).
• created an "ini_tracer_mod" routine in module "tracer_mod" for a cleaner initialization of the later.
• removed some purely dynamics-related outputs (etot0, zoom parameters, etc.) from diagfi.nc and stats.nc outputs as these informations are not available in the physics.

LMDZ.GENERIC:

• physics now completely decoupled from dynamics; the physics package may now be compiled as a library (-libphy option of makelmdz_fcm).
• added nqtot to tracer_h.F90.
• removed some purely dynamics-related outputs (etot0, zoom parameters, etc.) from diagfi.nc and stats.nc outputs as these informations are not available in the physics.

LMDZ.VENUS:

• physics now completely decoupled from dynamics; the physics package may now be compiled as a library (-libphy option of makelmdz_fcm).
• added infotrac_phy.F90 to store information on tracers in the physics. Initialized via iniphysiq.
• added cpdet_phy_mod.F90 to store t2tpot etc. functions to be used in the physics. Initialized via iniphysiq. IMPORTANT: there are some hard-coded constants! These should match what is in cpdet_mod.F90 in the dynamics.
• got rid of references to moyzon_mod module within the physics. The required variables (tmoy, plevmoy) are passed to the physics as arguments to physiq.

LMDZ.TITAN:

• added infotrac_phy.F90 to store information on tracers in the physics. Initialized via iniphysiq.
• added cpdet_phy_mod.F90 to store t2tpot etc. functions to be used in the physics.
• Extra work required to completely decouple physics and dynamics: moyzon_mod should be cleaned up and information passed from dynamics to physics as as arguments. Likewise moyzon_ch and moyzon_mu should not be queried from logic_mod (which is in the dynamics).

EM

Location:

trunk/LMDZ.MARS/libf/dynphy_lonlat

Files:

• inigeomphy_mod.F90 (modified) (2 diffs)
• phymars/iniphysiq_mod.F90 (modified) (2 diffs)

trunk/LMDZ.MARS/libf/dynphy_lonlat/inigeomphy_mod.F90

@@ -21 +21 @@
   USE mod_interface_dyn_phys, ONLY :  init_interface_dyn_phys
   USE nrtype, ONLY: pi
+  USE comvert_mod, ONLY: preff, ap, bp, aps, bps, presnivs, &
+                         scaleheight, pseudoalt
+  USE vertical_layers_mod, ONLY: init_vertical_layers
   IMPLICIT NONE
 
@@ -213 +216 @@
                      airefi,cufi,cvfi)
 
+  ! copy over preff , ap(), bp(), etc 
+  CALL init_vertical_layers(nlayer,preff,scaleheight, &
+                            ap,bp,aps,bps,presnivs,pseudoalt)
+
 !$OMP END PARALLEL
 

trunk/LMDZ.MARS/libf/dynphy_lonlat/phymars/iniphysiq_mod.F90

@@ -10 +10 @@
                      prad,pg,pr,pcpp,iflag_phys)
 
-use infotrac, only : nqtot ! number of advected tracers
+use infotrac, only : nqtot, & ! number of advected tracers
+                     tname ! tracer names
 use comgeomfi_h, only: ini_fillgeom
 use temps_mod, only: day_ini, hour_ini
@@ -68 +69 @@
 call init_dimphy(klon_omp,nlayer)
 
-call phys_state_var_init(klon_omp,nlayer,nqtot, &
+call phys_state_var_init(klon_omp,nlayer,nqtot,tname, &
                          day_ini,hour_ini,punjours,ptimestep, &
                          prad,pg,pr,pcpp)