Add better handling of fcm path from makelmdz_fcm script to build_gcm script; and some translation of output messages to English.
EM
50 lines of code changed in 4 files:
"-auto" reduces virtual memory.
1 lines of code changed in 1 file:
Version testing basée sur r1745
---------------------------------
Testing release based on r1745
4432 lines of code changed in 77 files:
Added handling of Newtonian (and Shallow Water) modes in dyn3dmem:
- adapted calfis_loc.F and gcm.F
- removed unused routines divgrad_p.F and gradiv_p.F
- adapted iniacademic.F90 and sw_case_williamson.F to work in parallel and renamed them iniacademi_loc.F90 and sw_case_williamson_loc.F
- fixed bug in exner_milieu_loc.F (filtreg_p form mod_filtreg_p should be used)
EM
500 lines of code changed in 9 files:
OMP CRITICAL name must be between brackets (ifort specific?) or else compilation fails in OpenMP mode. Moreover it is recommended to use different names for different OMP CRITICAL sections.
EM
5 lines of code changed in 2 files:
When an MPI process is stopped by a Fortran stop, other MPI processes may not stop. Use mpi_abort instead.
15 lines of code changed in 2 files:
Correction de bug sur le calcul de l'eau condensee dans fisrtilp.F90
2 lines of code changed in 1 file:
Correction de bug (introduit depuis la version 1742) !
3 lines of code changed in 1 file:
Fix to avoid dividing by zero.
JYG
4 lines of code changed in 1 file:
1- Inclusion des developpements de la these de Romain Pilon sur le
lessivage des aerosols :
a/ par les pluies convectives (modifs cv30_routines et cv3_routines pour
sortir les champs nécessaires au calcul off-line ; modif cvltr)
b/ par les pluies stratiformes (modifs phytrac et introduction
lsc_scav).
2- Choix entre plusieurs schemas pour les pluies stratiformes, commande
par iflag_lscav.
3- Quelques corrections dans la convection "Nouvelle Physique" pour
assurer la conservation des traceurs (cv3p1_mixing et cva_driver) (travail
de Robin Locatelli).
2413 lines of code changed in 29 files:
Option mcmodel needs to be specified at compile-time, not link-time.
2 lines of code changed in 1 file:
Fixing bug from revision 1539 (see also revision 1604). Allocatable
arrays "tabijgcm", "longcm", "latgcm", "igcm", "jgcm" cannot be
arguments of "phys_output_open" if they have not been
allocated. Allocate them with zero size when iflag_con < 3.
Replaced non-ASCII characters in comments by ASCII sequences. There
are problems with non-ASCII characters, see for example line 2238:
some information is lost here.
23 lines of code changed in 1 file:
correction in the call of iniphysiq.F
1 lines of code changed in 1 file:
Modifications for numerical stability of the boundary layer parameterizations.
Concerning yamada4.F :
- option with asymptotic mixing length l0 imposed (iflag_pbl=8 and 9)
- option with a new temporal scheme (iflag_pbl=10 and 11)
- Correction for very stable PBLs (iflag_pbl=10 and 11)
iflag_pbl=8 converges numerically with NPv3.1
iflag_pbl=11 -> now starts with NP from start files created by ce0l
-> the model can run with longer time-steps.
Concerning thermals :
Introduction of an implicit computation of vertical advection in
the environment of thermal plumes in thermcell_dq
impl = 0 : explicit, 1 : implicit, -1 : old version
controled by iflag_thermals =
15, 16 run with impl=-1 : numerical convergence with NPv3
17, 18 run with impl=1 : more stable
15 and 17 correspond to the activation of the stratocumulus "bidouille"
Modified routines (phylmd/):
calltherm.F90 : for managing the various options of thermcell_dq
coef_diff_turb_mod.F90 : yamada4 called for iflag_pbl<= 18 instead of 11
physiq.F : desactivation of the vertical diffusion of TKE for iflag_pbl=10
thermcellV0_main.F90 : calling thermcell_dq with implicit scheme
thermcell_dq.F90 : thermcell_dq with optional explicit/implicit scheme
thermcell_main.F90 : various option for vertical transport by thermals
yamada4.F : Yamada scheme with new options
297 lines of code changed in 10 files: