Changes between Version 4 and Version 5 of Documentation/UserGuide/FLUXNETValidation


Ignore:
Timestamp:
2018-11-08T08:53:56+01:00 (5 years ago)
Author:
mmcgrath
Comment:

--

Legend:

Unmodified
Added
Removed
Modified
  • Documentation/UserGuide/FLUXNETValidation

    v4 v5  
    33http://forge.ipsl.jussieu.fr/orchidee/wiki/Scripts/FluxnetValidation 
    44 
    5 The "HowTo" section is the first place people might look for this information, so this page just serves as a redirect to the real page. 
     5And then look at the README file in config/ORCHIDEE_OL/ENSEMBLE. 
    66 
    7 This is a location on asterix that he gave me.  When I remember what it's for, I'll update this page.  Possibly the location of data files? 
     7Now, if you are using ORCHIDEE-CN-CAN, you need to do some other things.  I am running on Obelix. 
    88 
    9 /home/orchidee01/mmancip/IGCM/BC/OOL/OL2/FLUXNET/  
     9Start from a clean SVN ENSEMBLE install.  Notice that Job\_ENSEMBLE is the main driver, and it should not be deleted!  It will create jobs based on ../SPINUP/SUBJOB/OOL\_SEC\_STO/. 
    1010 
    11 One change that I had to make is in libIGCM/libIGCM_sys/libIGCM_sys.ksh.  I had to add the following three lines 
     11First, a change in the IOIPSL source code needs to be made, with a recompile. 
     12In modeles/IOIPSL/src/histcom.f90, change nb\_zax\_max=20.  Recompile. 
    1213 
    1314{{{ 
    14    asterix*) 
    15         echo "sys source asterix/obelix lib." 
    16         . ${libIGCM}/libIGCM_sys/libIGCM_sys_obelix.ksh;; 
     15cd ~/ORCHIDEE-CN-CAN/config/ORCHIDEE_OL/ENSEMBLE 
     16vi config.card 
     17}}} 
     18Change the following lines 
     19{{{ 
     20JobName=FLUXNET 
     21  ARCHIVE=/home/scratch01/mmcgrath 
     22}}} 
     23then create the job scripts 
     24{{{ 
     25../../../libIGCM/ins_job 
     26}}} 
     27this creates Job\_FLUXNET.  Notice that this job will pull from the SPINUP directory as well.  ins\_job used to create Job files in every directory, but that functionally changed a while ago.  Therefore, the following is now necessary (OOL\_SEC\_STO because we will run a job with sechiba and stomate). 
     28{{{ 
     29cd ../SPINUP 
     30../../../libIGCM/ins_job 
     31cd SUBJOB/OOL_SEC_STO/ 
     32../../../../../libIGCM/ins_job 
     33cd ../../../ENSEMBLE 
     34}}} 
     35Now modify the run directory. 
     36{{{ 
     37vi Job_FLUXNET 
     38(change RUN_DIR_PATH=/home/scratch01/mmcgrath/RUN_DIR) 
     39mkdir /home/scratch01/mmcgrath/RUN_DIR 
     40}}} 
     41Now change the options for the sites to run against. 
     42{{{ 
     43vi fluxnet.card 
     44}}} 
     45Need to change the number of PFTs, which sites we are running against (do a small testcase with a single site first), and the PFT characteristics of that site. 
     46{{{ 
     47  NbPFTs= 28 
     48  Groups= ( TEST ) 
     49  TEST =           ( BR-Sa3 , BR-Sa3_2000-2003.nc , 2000 , 4 , 0,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0 )  
    1750}}} 
    1851 
    19 above the line "obelix*)".  This was because when the script was run, the hostname was coming up as asterix and there was no option for asterix so it went with the default options.  This caused some bad directory names to be used which caused the script the crash. 
     52For some reason, in SPINUP/SUBJOB/OOL\_SEC\_STO/ reftemp.nc is copied to the run directory in both sechiba.card and stomate.card; this results in a crash.  To fix: 
     53{{{ 
     54   vi ../SPINUP/SUBJOB/OOL_SEC_STO/COMP/sechiba.card 
     55}}} 
     56and remove the line 
     57{{{ 
     58            (${R_IN}/SRF/reftemp.nc, .), \ 
     59}}} 
    2060 
    21 In addition, make sure you run in the config/ORCHIDEE_OL/ directory.  The script looks for some files under the assumption that it is in that directory (for example, when it copies ../SPINUP). 
     61Now we need to copy the CAN-CN run.def that we want to use.  XXXXXX is the run.def file from whatever configuration directory you have that is currently working. 
     62{{{ 
     63   cp XXXXXX ../SPINUP/SUBJOB/OOL_SEC_STO/PARAM/ 
     64   cp XXXXXX ../ENSEMBLE/PARAM/ 
     65}}} 
     66It seems that AUTO values are not being removed in the run.def during processing.  I set XIOS\_ORCHIDEE\_OK=N, DRIVER\_reset\_time=y, taking values from a previous successful run I did (looking in the RUN\_DIR).  I only replaced a value when the code crashed with a complaint. 
     67 
     68{{{ 
     69RESTART_FILEIN= NONE 
     70SECHIBA_restart_in= NONE 
     71STOMATE_RESTART_FILEIN= NONE 
     72}}} 
     73 
     74Need to check the differences between the sechiba.card in our working configuration directory and that in SPINUP/SUBJOB/OOL\_SEC\_STO/COMP.  I had to change the following in SPINUP/SUBJOB/OOL\_SEC\_STO/COMP/sechiba.card 
     75{{{ 
     76[InitialStateFiles] 
     77List=       (${R_IN}/SRF/routing.nc, .), \ 
     78            (${R_IN}/SRF/cartepente2d_15min.nc, .), \ 
     79            (${R_IN}/SRF/floodplains.nc, .), \ 
     80            (${R_IN}/SRF/albedo/alb_bg_modisopt_2D.nc, alb_bg.nc), \ 
     81            (${R_IN}/SRF/PFTmap_1850to2005_AR5_LUHa.rc2/PFTmap_IPCC_2000.nc, PFTmap.nc) 
     82}}} 
     83 
     84in SPINUP/SUBJOB/OOL\_SEC\_STO/COMP/sechiba.card, I made the following change: 
     85{{{ 
     86[BoundaryFiles] 
     87List=   () 
     88ListNonDel= (${R_IN}/SRF/reftemp.nc, .), \ 
     89            (/home/orchidee04/aslanso/Files/soils_param.nc, soils_param.nc), \ 
     90            (${R_IN}/SRF/soils_param_usdatop.nc, soils_param_usda.nc) 
     91}}} 
     92 
     93I had to modify the run.def a bit due to crashes (AUTO values were not getting replaced). 
     94 
     95I removed the following lines from ENSEMBLE/PARAM/run.def, since those variables didn't appear anywhere in the working config directories for the TRUNK.   
     96 
     97{{{ 
     98TIME_LENGTH = _AUTO_ 
     99TIME_SKIP = _AUTO_ 
     100SECHIBA_HISTFILE2=_AUTO_ 
     101WRITE_STEP=_AUTO_ 
     102STOMATE_HIST_DT=_AUTO_ 
     103STOMATE_IPCC_HIST_DT=_AUTO_ 
     104ATM_CO2 =_AUTO_: DEFAULT = 350. 
     105STOMATE_CFORCING_NAME= _AUTO_ : DEFAULT = NONE 
     106}}} 
     107 
     108Even though the next ones weren't in the TRUNK run.def, I assigned a value since I don't like the default. 
     109{{{ 
     110STOMATE_OK_STOMATE=y 
     111}}} 
     112 
     113With my debug flags on, the code threw a runtime error on the following line for an unknown reason, so I just commented it out (it's a debugging write statement, and I had l\_dbg activated...no harm in removing for the moment). 
     114{{{ 
     115modeles/IOIPSL/src/calendar.f90 
     116 
     117    IF (l_dbg) THEN 
     118!      WRITE(ipslout,*) "isittime 2.0 ", & 
     119! &     date_next_check,date_next_act,ABS(dt_action-freq), & 
     120! &     ABS(dt_action+dt_check-freq),dt_action,dt_check, & 
     121! &     next_check_itau,do_action 
     122    ENDIF 
     123}}} 
     124 
     125out\_orchidee has become huge, seemingly because of histfile debugging output.  So I am turning off ioipsl\_debug in modeles/IOIPSL/src/errioipsl.f90. 
     126 
     127We need to make sure all of the following lines are commented out (or do not exist in the run.def, since the script will change the vegetation for each site by adding lines at the end of the run.def and if these lines are present they will override them: 
     128 
     129{{{ 
     130IMPOSE_VEG=y 
     131SECHIBA_VEG__01=0.0769230769231 
     132... 
     133SECHIBA_VEGMAX__01=0.0769230769231 
     134  ... 
     135}}} 
     136 
     137I often set the JobType=DEV in my Job file so that it will overwrite old output files in the archive.  Else, you have to delete the IGCM\_OUT/OL2/PROD/ensemble/ files for every site before every run.  Which is no problem if everything works, but it is annoying if you are not yet sure they will work. 
     138 
     139Launch the job (from the README file). 
     140 
     141{{{ 
     142   ./Job_ENSEMBLE fluxnet > out.Job_ENSEMBLE 
     143}}} 
     144 
     145This submits jobs to the queue which create new directories in the ENSEMBLE directory, each one having the name of the test site. 
     146 
     147First simulation finishes with no problems!  However, it complains of an error in copying the restart file.  I tried deleting the whole archive and rerunning. 
     148 
     149It takes 20 minutes or so to run the STOI (it runs over four years or so, the length of the forcing file for that site). 
     150 
     151 
     152