| Version 37 (modified by dgoll, 8 years ago) (diff) |
|---|
SPINUP_ANALYTIC
1.1 create new experiment folder
cd ../../config/ORCHIDEE_OL/ cp -fr SPINUP_ANALYTIC/ MY_SPINUP_ANALYTIC/ cd MY_SPINUP_ANALYTIC/
1.2 tell libIGCM about the new input files (they are stored on the IMBALANCE-P disk on obelix)
modify the variable "ListNonDel?" in COMP/stomate.card:
ListNonDel= (${R_IN}/SRF/reftemp.nc, .),\
(/home/surface3/common/CNP_files/revXX/lithology.nc, .), \
(/home/users/dgoll/ORC_data/N_deposition/LMDzINCA_DEP_2000.nc, .), \
(/home/surface3/common/CNP_files/deposition/deposition_1850.nc, deposition.nc)
REMARK: make sure there are no blanks behind the "\"!
add to the list of "[InitialStateFiles?]" in COMP/sechiba.card:
(/home/surface3/common/CNP_files/revXX/soils_param.nc, .), \ (/home/surface3/common/CNP_files/revXX/soils_param_usda_new.nc.nc, .), \
You want account for inter-annual variability of deposition ?
You can loop over a couple for years of deposition to account for inter annual variability during a spinup using CyclicYear?. This done by adding deposition file to COMP/sechiba(!).card:
List= (${R_IN}/SRF/routing.nc, .), \
(/home/surface3/common/CNP_files/rev01/soils_param.nc, .), \
(/home/surface3/common/CNP_files/revXX/soils_param_usda_new.nc.nc, .), \
(${R_IN}/SRF/cartepente2d_15min.nc, .), \
(${R_IN}/SRF/floodplains.nc, .), \
(${R_IN}/SRF/PFTmap_1850to2005_AR5_LUHa.rc2/PFTmap_IPCC_${year}.nc, PFTmap.nc), \
(/ccc/work/cont003/dsm/p529vui/ORCHIDEE_trunk_CN/LMDzINCA_DEP_2000.nc, .),\
(/ccc/work/cont003/dsm/p529goll/NP_deposition/final/deposition_${CyclicYear}.nc, deposition.nc)
as well as(!) in COMP/stomate.card:
ListNonDel= (${R_IN}/SRF/reftemp.nc, .),\
(/home/surface3/common/CNP_files/rev00/lithology.nc, .), \
(/home/users/dgoll/ORC_data/N_deposition/LMDzINCA_DEP_2000.nc, .), \
(/home/surface3/common/CNP_files/rev00/deposition_${CyclicYear}.nc, .)
REMARK: Although the file with N deposition is used in sechiba we add the file to the stomate.card. This is needed to ensure it is copied when we restart from a simulation with impose_cn. In that case, the N deposition variable will not be present in the sechiba_restart_out.nc but the libIGCM assumes all sechiba variables are present.
WARNING: The double entry of deposition will cause a stop on obelix, as the "cp" option are set to not overwrite files (issue not solved yet; please ask Fabienne or Josefine how the cp option in libIGCM can be changed, so that you can overwrite files)
1.3 adjust the libIGCM
disable XIOS in COMP/orchidee_ol.card
# Use XIOS as output library instead of IOIPSL XIOS=n
make sure XIOS is set to "_AUTOBLOCKER_" (if you set it to "n", libIGCM will trigger a stop) in PARAM/run.def:
# Use XIOS for writing diagnostics file XIOS_ORCHIDEE_OK = _AUTOBLOCKER_
disable XIOS in config.card:
#IOS= (xios, XIOS)
[...]
#IOS= (xios_server.exe, xios.x, 1MPI)
set the IO frequency to avoid crashes in PARAM/run.def
# Writefrequency in seconds in sechiba_history.nc # default = 86400.0 WRITE_STEP = 1800.0 # Writefrequency in days in stomate_history.nc # default = 10. STOMATE_HIST_DT = 1
Set the 11-layer hydrology & scheme to explicitly to "y" ; this is the default hydrological scheme in ORCHIDEE-CN-P (r4166). If it is on "_AUTO_" it will cause a stop of the orchidee executable. For activating the 11-layer hydrology see here http://forge.ipsl.jussieu.fr/orchidee/wiki/DevelopmentActivities/ORCHIDEE-CNP/activate11HYDRO.
copy a run.def with parameter values for the ORCHIDEE-CAN components as well as for the nutrient components. Make sure you are using the newest version here:
cp /home/surface3/common/CNP_files/revXX/run.def PARAM/run.def
these flags are needed for the nutrient cycles in PARAM/run.def, so make sure they are in there and you set them according to your plans:
# to activate P cycle set =y STOMATE_OK_PCYCLE=y STOMATE_IMPOSE_NP=y # to activate N cycle set =y STOMATE_OK_NCYCLE=y STOMATE_IMPOSE_CN=y # make sure we dont fake LD_FAKE_HEIGHT=n # if you spinup the nitrogen and phosphorus; #the mineral N stocks can occasionally get over-depleted due to SOM accumulation; # to avoid that; set this true SPINUP_CNP=y # you can enable mass conservation and stoichiometry checks by =y STOMATE_MASS_CONSERVATION=y STOMATE_DSG_DEBUG=y Nammonium_FILE=LMDzINCA_DEP_2000.nc Nnitrate_FILE=LMDzINCA_DEP_2000.nc Nnitrate_VAR = NOY Nammonium_VAR = NHX Nfert_FILE=NONE Nbnf_FILE=NONE NINPUT_UPDATE=-1
disable river routing if you run on non-global scale in run.def
# Activate river routing RIVER_ROUTING = n
for more information on the manipulation of run.def: http://forge.ipsl.jussieu.fr/orchidee/wiki/Documentation/UserGuide/ManipulationDefFile
1.4a activate analytic solution
in COMP/stomate.card:
[UserChoices] # SPINUP_ANALYTIC=y/n : Activate the spinup analytic option to solve the carbon in soil balance SPINUP_ANALYTIC=y
1.4b adjust the job as usual
modify config.card as usual (see here for more information howto set these parameters in general: https://forge.ipsl.jussieu.fr/orchidee/wiki/Documentation/UserGuide/SpinupInJobBasic) I achieved an equilibrium for Europe using sets of 300-350 yr simulations (first step imposing NP concentration, second step prognostic NP concentrations restarting from step 1):
#-- leap(gregorian), noleap, 360d CalendarType=noleap #-- Start and End of Job #-- "YYYY-MM-DD" DateBegin=1960-01-01 DateEnd=2260-12-31 # Forcing data between 1901 and 1920 CyclicBegin=1901 CyclicEnd=1920
REMARK (curie only): If you want to run ORCHIDEE on spatial subdomain (here the example for a single grid box) you need to adjust the number before MPI line:
#OOL= (orchidee_ol, orchidee_ol, 32MPI) OOL= (orchidee_ol, orchidee_ol, 1MPI)
modify run.def as usual (see wiki) plus if you want to archive, for example in the IMBALANCE-P storage; add to config.card:
ARCHIVE=/home/surface3/dgoll
1.5 create job
../../../libIGCM/ins_job
1.6 adjust job_file (optional)
Header to run a site scale simulation on obelix
###################### ## OBELIX LSCE ## ###################### #PBS -N test #PBS -m a #PBS -j oe #PBS -q short #PBS -o Script_Output_test.000001 #PBS -S /bin/ksh #PBS -v BATCH_NUM_PROC_TOT=1 #PBS -l nodes=1:ppn=1
to automatically have existing files deleted, set JobType?=DEV
#D- Experience type : DEB(ug), DEV(elopment), RUN (default) JobType=DEV
to avoid queuing take full advantage of your job time; the value depends on #PBS -q short and computation time of orchidee
#D- Number of execution in one job PeriodNb=60
to have a temp directory you can access on obelix; exchange dgoll with your ID
#D- Define running directory
#D- Default=${TMPDIR} ie temporary batch directory
#D-
#RUN_DIR_PATH=/workdir/or/scratchdir/of/this/machine
RUN_DIR_PATH=/home/scratch01/dgoll/RUN_DIR
1.7 submit the job
on obelix:
qsub Job_ID
on curie
ccc_msub Job_ID
1.8 restart / rerun a job after crash
It is important that you stick to the libIGCM regulations. You /libIGCM/clean_Month to reset run.card or delete run.card if you want to start from scratch. To avoid headaches, it is recommended to always delete all file which you want to redo , for example in the experiment output in IGCM_OUT, etc
2.0 problems which might occur
2.1. executable causes a stop
In general, you should check the out_orchidee_ol and out_execution files for error messages.
known problems
if you can find in "out_orchidee_ol" a message that a LOGICAL has an inappropriate value, you must set in PARAM/run.def the LOGICAL to "y" or "n", for example:
# Activate the multi-layer diffusion scheme adapted from CWRR # by default the Choisnel hydrology is used. HYDROL_CWRR = n
