# =================================================================== # A parameter file for defining a CNRM-CM experiment # =================================================================== # Once executed , this file actually "installs" (or setup) an experiment, # by sourcing the setup script INSTALLER (see file bottom ) with the parameters # Launching the experiment should thereafter be done by "relan EXPID", # Note : this file MUST be named according to the pattern : param_EXPID # where EXPID is the identifier for the experiment EXPID=$(basename $0| sed -r -e 's/.*param_//g') # Please refer to http://www.cnrm.meteo.fr/cm/spip.php?article1 # for further explanations on parameters below # Syntax is according to your the shell # Sections are : # DESCRIBING THE EXPERIMENT AND WHERE TO STORE IT # PERIOD TO PROCESS, AND JOBS GRANULARITY # JOBS CHARACTERISTICS # OUTPUT FILES : DIAGNOSTICS, RESTART, LISITINGS # NAMELISTS USED. # RESTART FILES. # FORCINGS AND THEIR EVOLUTION # FOR ADVANCED USE # The environment variable EXPID is a reserved word. # =================================================================== # DESCRIBING THE EXPERIMENT AND WHERE TO STORE IT # =================================================================== # If you want your experiment to be named EXPID, the current file # should be naled : "param_$EXPID" # The next string will be used to document diagnostic fields TITRE="Idem SPEK2 en 2 noeuds + correction runoff + correction eau/sel Glt + demarrage de Levitus pour oce, 1900 de SPEK2 pour le reste + ctte solaire 1361 + correction derive solaire + correction masque" # The experiment should belong to a group (corresponding parent directories # on archive machine and front-end will be automatically created; See section # "Advanced use" for archive location and relance location details). GROUP=tech # =================================================================== # PERIOD TO PROCESS, AND JOBS GRANULARITY # =================================================================== # The period will be processed in multiple jobs. Each job will process # a number of months. The first month of the first job may begin at any # date in the month, provided there is a valid ocean initial state for # that date (either climatology or a restart).Later months will be entire. # Initial and last date of the run INIDATE=18010130 ; ENDDATE=18010202 # Start time of the run INITIME=0 # How many months should we process in each job NMONTH=12 # =================================================================== # JOBS CHARACTERISTICS # =================================================================== # These characteristics apply to each of the re-launched jobs. They may # be changed during the experiment in file $RELDIR/$EXP_his # =================================================================== # Elapsed time and memory usage on the frontal computer ELAPSFRONT=00:30:00 ; MEMFRONT=8gb # Elapsed time and memory usage on the number crunching computer ELAPS=00:05:00 # Normally , 3h40 is enough for one year when using 6 procs for Nemo on tori # Memory usage on the number crunching computer # IDRIS : le para MEM vaut pour la zone "DATA", qui contient les tableaux # allocatable (heap), les DATA, .... Defaut=3.2Gb. MEM=60gb # AT IDRIS, you can specifiy a stack value, default = 0.5gb, data+stack <= 3.5Gb #STACK=0.5Gb # NFS queue on the number crunching computer (can be "test", or "vector"). # At IDRIS, queue is automatic QUEUE=vector # How many nodes and processors per node are used ? # At IDRIS, for now, you do not request nodes, but only processes NN=1 ; NNP=8 # Number of processors dedicated to Arpege, to the Ocean model NPROC_ARP=6 ; NPROC_OCE=2 ; # Number of processors dedicated to the IO server. Can be 0 at IDRIS, but not yet at MF NPROC_IOS=1 # E-mail to send error messages RUNMAIL=senesi@meteo.fr # =================================================================== # OUTPUT FILES: DIAGNOSTICS, RESTART, LISITINGS # =================================================================== # Listings are saved in the "relance" directory # Quoting the models which listing will be saved, by a key-letter or keyword: # all, A (Atm), O (ocean), R (River), G (Ice), C (Coupler) SAVE_LISTING=no # Diagnostics and restarts are saved on the archive machine according # (by default) to the standard CNRM-CM file organization . See section # "advanced use" and variables $(modele)[RES|OUT]ARCH for changing it. # To save restarts every 6h (for nudging or Aladin) put 'all' # Otherwise put the period of saving for restart: # if 1 ---> save restarts every month # if 12 ---> save restarts every year # Anyway, if the processed dates interval is less than this period, # you will get a restart corresponding to the day after the last # processed day. SAVE_RESTART_PER=12 # =================================================================== # NAMELISTS USED. # =================================================================== # Namelist files can be on any rcp-reachable remote machine or cougar. # If you want to share such a parameter file and the files it uses, # remember to use either absolute path or the form "~account/..." MACH_NAMELIST=tori ATMNAMREF=~mrgu548/public/namelists/namarp_ar5.2.gwdse4 # gwdse4 pour Ecume, a priori ... SFXNAMREF=~mrgu548/public/namelists/namsfx_ar5.1.ecume OCENAMREF=~cgie002/NEMO/Namelist/namelist_nemo3.3.gel_SPEK2 ICENAMREF=~mrgu548/public/namelists_ar5/namgel_ar5 #~mrgu548/public/namelists/namgel_ar5.2 CPLNAMREF=~mrgu548/public/namelists/namcpl_ar5.1 RIVNAMREF=~mrgu548/public/namelists/namtrp_ar5.1 IOXNAMREF=~mrgu548/public/namelists/namiox_ar5.1 IOSNAMREF=~mrgu548/public/namelists/namios_ar5.1 # =================================================================== # RESTART FILES. # =================================================================== # Input restart files can be on any rcp-reachable remote machine or cougar. # Default is cougar MACH_RESTART=cougar EXPREF=PICTL ; DATREF=18500101 hm=/home/m/mgie/mgie003/AR5/$EXPREF/ ATMREST=$hm/atm/restart/$EXPREF.rst.atm.P$DATREF SFXREST=$hm/sfx/restart/$EXPREF.rst.sfx.P$DATREF #If intended Ocean initial state is the climatology, LOCEREST should be set to 0 #LOCEREST=0 # Otherwise do not forget to quote an ocean restart file, e.g. OCEREST=$hm/oce/restart/$EXPREF.rst.oce.P$DATREF.nc ICEREST=$hm/oce/restart/$EXPREF.rst.ice.P$DATREF RIVREST=$hm/riv/restart/$EXPREF.rst.trp.P$DATREF CPLATMRES=$hm/cpl/restart/$EXPREF.cpl.atm.P$DATREF.nc CPLOCERES=$hm/cpl/restart/$EXPREF.cpl.oce.P$DATREF.nc CPLRIVRES=$hm/cpl/restart/$EXPREF.cpl.trp.P$DATREF.nc # =================================================================== # FORCINGS AND THEIR EVOLUTION # =================================================================== # # For each forcing 2 variables : # YEAR_XXX : defining how the forcing evolves depending on the year of the simulation # FORXXX : pattern for filename used for the forcing, which can includes keywords YYYY # and MM, where YYYY will be replaced by current year and MM will be replaced by current # month at execution time, using the scheme described below # # In case FORXXX does no contain the string YYYY then YEAR_XXX is not used! # In case FORXXX does no contain YYYY nor MM the forcing will be fixed # (except for GHG gases where the year is read in the file, not in the name of the file) # # In case FORXXX contains YYYY # if YEAR_XXX is less or equal to 4 characters # --> YYYY will be replaced by YEAR_XXX in FORXXX (always the same YYYY # during the simulation, but YYYY can be a string or a year number) # # if YEAR_XXX is more than 4 characters, it should be of the form : # SCEN_YYY1_YYY2 : in this case the forcing will evolve every year between # YYY1 and YYY2 and will be equal to YYY1 before year YYY1 # and will be equal to YYY2 afetr year YYY2 # SCED_YYY1_YYY2 : similar to previous case but the forcing evolves only # every decade (ex: from 1985 to 1994, YYYY=1990) # # Geometry for arpege - Used for computing forcing filenames GEOM=tl127l31r # Boundary conditions : orography, land use ... # By exception, for boundary conditions, XXX=BCD but pattern is named BCOND YEAR_BCD=no ; BCOND=~mrga561/bcond/nclim4_${GEOM}_mMM # GHG gases YEAR_GHG=1850 ; FORGHG=~mrgu548/forcages/ghg_et_solaire_cmip5/GHG_PICTL.dat # Sulfate aerosols YEAR_SUL=1855 ; FORSUL=~mrgu548/clim/aerosols_balkanski_$GEOM/aero_balkanski_SO4.YYYYMM.ieee # Volcanic aerosols YEAR_VOL=no ; FORVOL=~mrgu556/clim/aero_volcans_$GEOM/aero_Ammann_vo.globavg.ieee # Cas particulier ici YEAR_OZO=no ; FOROZO=~mrgs536/clim/ozone_sl_$GEOM/PROBnecCOEF_60_v3.2_coeff7.MM.ieee # FIXED AEROSOLS # Black Carbon YEAR_BCA=1855 ; FORBCA=~mrgu548/clim/aerosols_balkanski_$GEOM/aero_balkanski_BC.YYYYMM.ieee # Organic Carbon YEAR_ORA=1855 ; FORORA=~mrgu548/clim/aerosols_balkanski_$GEOM/aero_balkanski_POM.YYYYMM.ieee # Sand Dust YEAR_SDA=1855 ; FORSDA=~mrgu548/clim/aerosols_balkanski_$GEOM/aero_balkanski_CIDUST.YYYYMM.ieee # Sea Salt YEAR_SSA=1855 ; FORSSA=~mrgu548/clim/aerosols_SS_scales_tegen_$GEOM/aero_baltegen_SS.YYYYMM.ieee # =================================================================== # FOR ADVANCED USE # =================================================================== # Valid on tori as of feb 2012 : # Models binaries, their data and related tools ATMEXE=~mrga589/public/arp502/bin/ald502_export.01.exe ; ASCEXE=no # May indicate a binary complementing Arpege restarts with Surfex diag fields. e.g. ~mrga589/src/addts.sh UPDCLI=~mrgu556/arpege/library/updcli_gen_1 ; UPDOZO=~mrgs536/arpege/library/updozo5.1.v5 OCEEXE=~mrgu556/opa/opa9/NEMO3.3/bin/nemo3.3_o1.5.30.v10.6.2P.CPL ; DATA_OCE=~mrgu548/opa/opa9/NEMO3.3/data2 # An I/O server for Nemo should be quoted when using multi-proc Nemo, for assembling Nemo procs outputs # Otherwise, you'd better comment out next line to avoid an additionnal process IOSEXE=~mrgu548/opa/opa9/NEMO3.3/bin/ioserver RIVEXE=~mrgu556/trip/library/routeDBL.ipcc ; DATA_RIV=~mrgu556/trip/data CPLEXE=~mrgu556/oasis/v3.3.02/prism/SX8/bin/oasis3.MPI1.x ; DATA_CPL=~mrgu556/oasis/data/v3_1d_sfx_sophie.2m # Temporary location to store output data. STORE_LOC=$FTDIR # Archive directories on cougar used to save diagnostics and # restarts. Next lines follow the CNRM-CM5 standard file organization # for CMIP5 outputs # The experiment can belong to a group (corresponding parent # directories will be automatically created) EXPARCHDIR=${GROUP:-NOGROUP}/$EXPID # Will hold a copy of the current file ATMARCH=$EXPARCHDIR/atm; ATMRESARCH=$ATMARCH/restart ; ATMOUTARCH=$ATMARCH/output SFXARCH=$EXPARCHDIR/sfx; SFXRESARCH=$SFXARCH/restart ; SFXOUTARCH=$SFXARCH/output OCEARCH=$EXPARCHDIR/oce; OCERESARCH=$OCEARCH/restart ; OCEOUTARCH=$OCEARCH/output ICEARCH=$EXPARCHDIR/oce; ICERESARCH=$ICEARCH/restart ; ICEOUTARCH=$ICEARCH/output RIVARCH=$EXPARCHDIR/riv; RIVRESARCH=$RIVARCH/restart ; RIVOUTARCH=$RIVARCH/output CPLARCH=$EXPARCHDIR/cpl; CPLRESARCH=$CPLARCH/restart ; CPLOUTARCH=$CPLARCH/output # Directory for automatic relance (and hence job outputs). This default # value is fine for most uses RELDIR=$HOME/relances/$EXPARCHDIR # 'System' parameters ####################### # When should result files archiving be performed (for each set of NMONTH): # BEFORE processing next set, or DURING this processing (with a # control at its end) or in a totally UNCONTROLLED way. Last two # options only if you are in a hurry and ready to spoil some CPU for # re-running in case of archive machine down, and/or some time to # re-launch manually the archiving sub-jobs. ARCHIVING=DURING # BEFORE, DURING,or UNCONTROLLED # Which is the version of CNRM-CM experiment installer script you will use : # (by default, this will also drive the versions of mtool and relan) INSTALLER=~mrgu548/cnrm-cm/V5.4/scripts/instexp_couple # Where to find shell script tools. Default value is toolbox in the # parent directory of $INSTALLER, and is usually fine # TOOL=~mrgu548/cnrm-cm/V5.4/toolbox # Script implementing the coupled model. Default value is 'script_couple' # in the directory of $INSTALLER # script=~mrgu548/cnrm-cm/V5.4/script_couple.5.3 # Launching the experiment "installer" . $INSTALLER