Changes between Version 6 and Version 7 of Documentation/UserGuide/FLUXNETValidation


Ignore:
Timestamp:
2018-12-06T15:14:14+01:00 (6 years ago)
Author:
mmcgrath
Comment:

Updating after the TRUNK/CN merge, since some things changed.

Legend:

Unmodified
Added
Removed
Modified

  • Documentation/UserGuide/FLUXNETValidation

    v6 v7  
    3939mkdir /home/scratch01/mmcgrath/RUN_DIR 
    4040}}} 
     41In addition, you may sometimes have to change the queue name in this file (I use medium on obelix, but sometimes long is required as the job won't finish). 
     42 
    4143Now change the options for the sites to run against. 
    4244{{{ 
    … …  
    4850  Groups= ( TEST ) 
    4951  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 )  
     52 
     53  n_iter=1 
     54  duree_sechiba=500 
     55  duree_stomate=0 
     56  duree_carbonsol=0 
     57  duree_final=1 
    5058}}} 
    5159 
    … …  
    5967}}} 
    6068 
    61 Now 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 }}} 
     69Some things to watch out for with ORCHIDEE-CN-CAN: best to use a run.def with no age classes (for easier comparison) and no forest management. 
     70 
    6671It 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. 
    6772 
    … …  
    7176STOMATE_RESTART_FILEIN= NONE 
    7277}}} 
    73  
    74 Need 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] 
    77 List=       (${R_IN}/SRF/routing.nc, .), \ 
    78             (${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             (${R_IN}/SRF/PFTmap_1850to2005_AR5_LUHa.rc2/PFTmap_IPCC_2000.nc, PFTmap.nc) 
    82 }}} 
    83  
    84 in SPINUP/SUBJOB/OOL\_SEC\_STO/COMP/sechiba.card, I made the following change: 
    85 {{{ 
    86 [BoundaryFiles] 
    87 List=   () 
    88 ListNonDel= (${R_IN}/SRF/reftemp.nc, .), \ 
    89             (/home/orchidee04/aslanso/Files/soils_param.nc, soils_param.nc), \ 
    90             (${R_IN}/SRF/soils_param_usdatop.nc, soils_param_usda.nc) 
    91 }}} 
    92  
    93 I had to modify the run.def a bit due to crashes (AUTO values were not getting replaced). 
    9478 
    9579I removed the following lines from ENSEMBLE/PARAM/run.def, since those variables didn't appear anywhere in the working config directories for the TRUNK.   
    … …  
    11195}}} 
    11296 
     97We need to make sure all of the following lines are commented out (or do not exist in the run.def, since the script will change the vegetation for each site by adding lines at the end of the run.def and if these lines are present they will override them: 
     98 
     99{{{ 
     100IMPOSE_VEG=y 
     101SECHIBA_VEG__01=0.0769230769231 
     102... 
     103SECHIBA_VEGMAX__01=0.0769230769231 
     104  ... 
     105}}} 
     106 
     107Make sure no limits are specified, either (i.e., remove lines like the following): 
     108{{{ 
     109LIMIT_WEST=4 
     110LIMIT_EAST=6 
     111LIMIT_NORTH=48 
     112LIMIT_SOUTH=46 
     113}}} 
     114 
     115Now we need to copy the CAN-CN run.def that we want to use.  XXXXXX is the run.def file from whatever configuration directory you have that is currently working. 
     116{{{ 
     117   cp XXXXXX ../SPINUP/SUBJOB/OOL_SEC_STO/PARAM/ 
     118   cp XXXXXX ../ENSEMBLE/PARAM/ 
     119}}} 
     120 
     121For ORCHIDEE-CN-CAN, I checked the differences between the sechiba.card in our working configuration directory and that in SPINUP/SUBJOB/OOL\_SEC\_STO/COMP.  I had to change the following in SPINUP/SUBJOB/OOL\_SEC\_STO/COMP/sechiba.card 
     122{{{ 
     123[InitialStateFiles] 
     124List=       (${R_IN}/SRF/routing.nc, .), \ 
     125            (${R_IN}/SRF/cartepente2d_15min.nc, .), \ 
     126            (${R_IN}/SRF/floodplains.nc, .), \ 
     127            (${R_IN}/SRF/albedo/alb_bg_modisopt_2D.nc, alb_bg.nc), \ 
     128            (${R_IN}/SRF/PFTmap_1850to2005_AR5_LUHa.rc2/PFTmap_IPCC_2000.nc, PFTmap.nc) 
     129}}} 
     130 
     131in SPINUP/SUBJOB/OOL\_SEC\_STO/COMP/stomate.card, I made the following change: 
     132{{{ 
     133[BoundaryFiles] 
     134List=   () 
     135ListNonDel= (${R_IN}/SRF/reftemp.nc, .), \ 
     136            (/home/orchidee04/aslanso/Files/soils_param.nc, soils_param.nc), \ 
     137            (${R_IN}/SRF/soils_param_usdatop.nc, soils_param_usda.nc) 
     138}}} 
     139 
    113140With 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). 
    114141{{{ 
    … …  
    125152out\_orchidee has become huge, seemingly because of histfile debugging output.  So I am turning off ioipsl\_debug in modeles/IOIPSL/src/errioipsl.f90. 
    126153 
    127 We 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 {{{ 
    130 IMPOSE_VEG=y 
    131 SECHIBA_VEG__01=0.0769230769231 
    132 ... 
    133 SECHIBA_VEGMAX__01=0.0769230769231 
    134   ... 
    135 }}} 
    136  
    137154I 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. 
    138155 
    … …  
    145162This submits jobs to the queue which create new directories in the ENSEMBLE directory, each one having the name of the test site. 
    146163 
    147 First simulation finishes with no problems!  However, it complains of an error in copying the restart file.  I tried deleting the whole archive and rerunning. 
     164After the trunk merged with CN, I got very quick crashes along the lines of: 
    148165 
    149 It 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). 
     166{{{ 
     167ERROR : parameter Nfert_cropland_FILE doesn't exist in run parameter file ! 
     168You must correct subjobparam option in your card file or add new parameter in run.def file 
     169We must STOP here. 
     170}}} 
     171For a variety of variables.  Every time this happened, I removed the variable from fluxnet.card and relaunched. 
    150172 
    151173 
    152174 
     175 
     176 
     177 
     178